{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.58468609665186,
  "eval_steps": 4000,
  "global_step": 36000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.8286386209540069,
      "epoch": 0.00016241280462551668,
      "grad_norm": 9.25,
      "learning_rate": 4.5e-07,
      "loss": 1.0799656867980958,
      "mean_token_accuracy": 0.7693958081305027,
      "num_tokens": 21078.0,
      "step": 10
    },
    {
      "entropy": 0.7109394183382391,
      "epoch": 0.00032482560925103335,
      "grad_norm": 12.0625,
      "learning_rate": 9.5e-07,
      "loss": 0.8955118179321289,
      "mean_token_accuracy": 0.8040434066206217,
      "num_tokens": 41157.0,
      "step": 20
    },
    {
      "entropy": 0.7656834285706282,
      "epoch": 0.00048723841387655003,
      "grad_norm": 9.875,
      "learning_rate": 1.45e-06,
      "loss": 1.0864809989929198,
      "mean_token_accuracy": 0.7730953447520733,
      "num_tokens": 61654.0,
      "step": 30
    },
    {
      "entropy": 0.72407943662256,
      "epoch": 0.0006496512185020667,
      "grad_norm": 11.0,
      "learning_rate": 1.95e-06,
      "loss": 0.9031203269958497,
      "mean_token_accuracy": 0.7893224690109492,
      "num_tokens": 82960.0,
      "step": 40
    },
    {
      "entropy": 0.8371759531088173,
      "epoch": 0.0008120640231275833,
      "grad_norm": 14.1875,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 1.013376522064209,
      "mean_token_accuracy": 0.7663579270243644,
      "num_tokens": 104980.0,
      "step": 50
    },
    {
      "entropy": 0.7614489119965583,
      "epoch": 0.0009744768277531001,
      "grad_norm": 11.75,
      "learning_rate": 2.95e-06,
      "loss": 0.9164009094238281,
      "mean_token_accuracy": 0.7984466932713985,
      "num_tokens": 125062.0,
      "step": 60
    },
    {
      "entropy": 0.6971589954569936,
      "epoch": 0.0011368896323786168,
      "grad_norm": 11.125,
      "learning_rate": 3.4500000000000004e-06,
      "loss": 0.9299216270446777,
      "mean_token_accuracy": 0.7987640958279372,
      "num_tokens": 144332.0,
      "step": 70
    },
    {
      "entropy": 0.7037347192876041,
      "epoch": 0.0012993024370041334,
      "grad_norm": 8.6875,
      "learning_rate": 3.95e-06,
      "loss": 0.8815269470214844,
      "mean_token_accuracy": 0.80237399302423,
      "num_tokens": 164793.0,
      "step": 80
    },
    {
      "entropy": 0.6618278024718165,
      "epoch": 0.00146171524162965,
      "grad_norm": 9.6875,
      "learning_rate": 4.45e-06,
      "loss": 0.7643122673034668,
      "mean_token_accuracy": 0.8236934401094913,
      "num_tokens": 185887.0,
      "step": 90
    },
    {
      "entropy": 0.7201404077932239,
      "epoch": 0.0016241280462551667,
      "grad_norm": 12.0,
      "learning_rate": 4.950000000000001e-06,
      "loss": 0.9489357948303223,
      "mean_token_accuracy": 0.7943450938910246,
      "num_tokens": 206736.0,
      "step": 100
    },
    {
      "entropy": 0.7261254206299782,
      "epoch": 0.0017865408508806835,
      "grad_norm": 15.0625,
      "learning_rate": 5.45e-06,
      "loss": 0.9126925468444824,
      "mean_token_accuracy": 0.7977610722184181,
      "num_tokens": 227243.0,
      "step": 110
    },
    {
      "entropy": 0.8010070217773319,
      "epoch": 0.0019489536555062001,
      "grad_norm": 13.875,
      "learning_rate": 5.95e-06,
      "loss": 1.0046897888183595,
      "mean_token_accuracy": 0.7712242294102907,
      "num_tokens": 247183.0,
      "step": 120
    },
    {
      "entropy": 0.7256525276228786,
      "epoch": 0.002111366460131717,
      "grad_norm": 8.6875,
      "learning_rate": 6.45e-06,
      "loss": 0.852382755279541,
      "mean_token_accuracy": 0.8060502171516418,
      "num_tokens": 267532.0,
      "step": 130
    },
    {
      "entropy": 0.7117265942506492,
      "epoch": 0.0022737792647572336,
      "grad_norm": 12.8125,
      "learning_rate": 6.950000000000001e-06,
      "loss": 0.8560203552246094,
      "mean_token_accuracy": 0.8086145672947168,
      "num_tokens": 289440.0,
      "step": 140
    },
    {
      "entropy": 0.7038921610452235,
      "epoch": 0.00243619206938275,
      "grad_norm": 9.375,
      "learning_rate": 7.45e-06,
      "loss": 0.8178808212280273,
      "mean_token_accuracy": 0.815917557477951,
      "num_tokens": 310714.0,
      "step": 150
    },
    {
      "entropy": 0.848792072199285,
      "epoch": 0.002598604874008267,
      "grad_norm": 13.0,
      "learning_rate": 7.95e-06,
      "loss": 0.9831900596618652,
      "mean_token_accuracy": 0.7895596921443939,
      "num_tokens": 329435.0,
      "step": 160
    },
    {
      "entropy": 0.6665874444413931,
      "epoch": 0.0027610176786337834,
      "grad_norm": 7.3125,
      "learning_rate": 8.45e-06,
      "loss": 0.7425652503967285,
      "mean_token_accuracy": 0.8268452137708664,
      "num_tokens": 351279.0,
      "step": 170
    },
    {
      "entropy": 0.7654673175886273,
      "epoch": 0.0029234304832593,
      "grad_norm": 10.625,
      "learning_rate": 8.95e-06,
      "loss": 0.859465217590332,
      "mean_token_accuracy": 0.7999506097286939,
      "num_tokens": 372012.0,
      "step": 180
    },
    {
      "entropy": 0.7425559777766466,
      "epoch": 0.0030858432878848167,
      "grad_norm": 9.5,
      "learning_rate": 9.450000000000001e-06,
      "loss": 0.8205994606018067,
      "mean_token_accuracy": 0.8093760065734387,
      "num_tokens": 392252.0,
      "step": 190
    },
    {
      "entropy": 0.6891162494663149,
      "epoch": 0.0032482560925103333,
      "grad_norm": 9.4375,
      "learning_rate": 9.950000000000001e-06,
      "loss": 0.6921127319335938,
      "mean_token_accuracy": 0.8340979300439357,
      "num_tokens": 412508.0,
      "step": 200
    },
    {
      "entropy": 0.7497544032521546,
      "epoch": 0.0034106688971358504,
      "grad_norm": 6.34375,
      "learning_rate": 1.045e-05,
      "loss": 0.7874168872833252,
      "mean_token_accuracy": 0.8073534950613975,
      "num_tokens": 432996.0,
      "step": 210
    },
    {
      "entropy": 0.8025054382160306,
      "epoch": 0.003573081701761367,
      "grad_norm": 10.4375,
      "learning_rate": 1.095e-05,
      "loss": 0.8000614166259765,
      "mean_token_accuracy": 0.8073957156389951,
      "num_tokens": 453540.0,
      "step": 220
    },
    {
      "entropy": 0.726252057030797,
      "epoch": 0.0037354945063868836,
      "grad_norm": 7.28125,
      "learning_rate": 1.145e-05,
      "loss": 0.699830961227417,
      "mean_token_accuracy": 0.8347548361867666,
      "num_tokens": 475088.0,
      "step": 230
    },
    {
      "entropy": 0.7629142621532082,
      "epoch": 0.0038979073110124002,
      "grad_norm": 10.375,
      "learning_rate": 1.195e-05,
      "loss": 0.7690689563751221,
      "mean_token_accuracy": 0.819105388969183,
      "num_tokens": 494855.0,
      "step": 240
    },
    {
      "entropy": 0.7240849315188825,
      "epoch": 0.004060320115637917,
      "grad_norm": 8.1875,
      "learning_rate": 1.2450000000000001e-05,
      "loss": 0.7371947765350342,
      "mean_token_accuracy": 0.8307590883225202,
      "num_tokens": 513907.0,
      "step": 250
    },
    {
      "entropy": 0.7596399324014783,
      "epoch": 0.004222732920263434,
      "grad_norm": 8.5,
      "learning_rate": 1.2950000000000001e-05,
      "loss": 0.7624147415161133,
      "mean_token_accuracy": 0.8173091482371092,
      "num_tokens": 533835.0,
      "step": 260
    },
    {
      "entropy": 0.7431545678526164,
      "epoch": 0.00438514572488895,
      "grad_norm": 6.53125,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 0.7522071361541748,
      "mean_token_accuracy": 0.8200604744255543,
      "num_tokens": 554020.0,
      "step": 270
    },
    {
      "entropy": 0.7573855429887771,
      "epoch": 0.004547558529514467,
      "grad_norm": 8.5,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 0.7579145431518555,
      "mean_token_accuracy": 0.8237630523741245,
      "num_tokens": 574522.0,
      "step": 280
    },
    {
      "entropy": 0.7901369592174887,
      "epoch": 0.004709971334139983,
      "grad_norm": 11.3125,
      "learning_rate": 1.4449999999999999e-05,
      "loss": 0.7745565891265869,
      "mean_token_accuracy": 0.8216199889779091,
      "num_tokens": 595133.0,
      "step": 290
    },
    {
      "entropy": 0.7269687445834279,
      "epoch": 0.0048723841387655,
      "grad_norm": 10.5625,
      "learning_rate": 1.4950000000000001e-05,
      "loss": 0.7418421745300293,
      "mean_token_accuracy": 0.828744861483574,
      "num_tokens": 614857.0,
      "step": 300
    },
    {
      "entropy": 0.7721429783850908,
      "epoch": 0.005034796943391017,
      "grad_norm": 8.125,
      "learning_rate": 1.545e-05,
      "loss": 0.8017391204833985,
      "mean_token_accuracy": 0.8127230260521173,
      "num_tokens": 635235.0,
      "step": 310
    },
    {
      "entropy": 0.7957702752202749,
      "epoch": 0.005197209748016534,
      "grad_norm": 8.4375,
      "learning_rate": 1.595e-05,
      "loss": 0.8230579376220704,
      "mean_token_accuracy": 0.7985282287001609,
      "num_tokens": 656197.0,
      "step": 320
    },
    {
      "entropy": 0.6822628723457456,
      "epoch": 0.005359622552642051,
      "grad_norm": 9.5625,
      "learning_rate": 1.645e-05,
      "loss": 0.6820508480072022,
      "mean_token_accuracy": 0.8312795639038086,
      "num_tokens": 676265.0,
      "step": 330
    },
    {
      "entropy": 0.6956167140975594,
      "epoch": 0.005522035357267567,
      "grad_norm": 5.46875,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 0.7457751750946044,
      "mean_token_accuracy": 0.8247509013861418,
      "num_tokens": 697557.0,
      "step": 340
    },
    {
      "entropy": 0.7414393374696374,
      "epoch": 0.005684448161893084,
      "grad_norm": 7.5625,
      "learning_rate": 1.745e-05,
      "loss": 0.7440137386322021,
      "mean_token_accuracy": 0.82403952665627,
      "num_tokens": 717764.0,
      "step": 350
    },
    {
      "entropy": 0.7625441277399659,
      "epoch": 0.0058468609665186,
      "grad_norm": 7.71875,
      "learning_rate": 1.795e-05,
      "loss": 0.7318333148956299,
      "mean_token_accuracy": 0.8224152456969023,
      "num_tokens": 737751.0,
      "step": 360
    },
    {
      "entropy": 0.734699510037899,
      "epoch": 0.006009273771144117,
      "grad_norm": 6.8125,
      "learning_rate": 1.845e-05,
      "loss": 0.7759143829345703,
      "mean_token_accuracy": 0.8186948299407959,
      "num_tokens": 758755.0,
      "step": 370
    },
    {
      "entropy": 0.8060480868443847,
      "epoch": 0.006171686575769633,
      "grad_norm": 7.46875,
      "learning_rate": 1.895e-05,
      "loss": 0.7542331218719482,
      "mean_token_accuracy": 0.8169570811092853,
      "num_tokens": 779110.0,
      "step": 380
    },
    {
      "entropy": 0.6588757059536874,
      "epoch": 0.00633409938039515,
      "grad_norm": 6.4375,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 0.5770083427429199,
      "mean_token_accuracy": 0.8503068417310715,
      "num_tokens": 799775.0,
      "step": 390
    },
    {
      "entropy": 0.6667641870677471,
      "epoch": 0.006496512185020667,
      "grad_norm": 11.8125,
      "learning_rate": 1.995e-05,
      "loss": 0.6269936084747314,
      "mean_token_accuracy": 0.8433299727737904,
      "num_tokens": 820320.0,
      "step": 400
    },
    {
      "entropy": 0.6915348025038839,
      "epoch": 0.006658924989646184,
      "grad_norm": 5.40625,
      "learning_rate": 2.045e-05,
      "loss": 0.6881805419921875,
      "mean_token_accuracy": 0.8291698649525643,
      "num_tokens": 840349.0,
      "step": 410
    },
    {
      "entropy": 0.740300617273897,
      "epoch": 0.006821337794271701,
      "grad_norm": 9.4375,
      "learning_rate": 2.095e-05,
      "loss": 0.7356808662414551,
      "mean_token_accuracy": 0.8218936193734407,
      "num_tokens": 861048.0,
      "step": 420
    },
    {
      "entropy": 0.6525337123312056,
      "epoch": 0.006983750598897217,
      "grad_norm": 8.4375,
      "learning_rate": 2.145e-05,
      "loss": 0.6163470268249511,
      "mean_token_accuracy": 0.8430293515324593,
      "num_tokens": 883800.0,
      "step": 430
    },
    {
      "entropy": 0.6596776175312697,
      "epoch": 0.007146163403522734,
      "grad_norm": 8.875,
      "learning_rate": 2.195e-05,
      "loss": 0.6624906063079834,
      "mean_token_accuracy": 0.8395840384066104,
      "num_tokens": 903944.0,
      "step": 440
    },
    {
      "entropy": 0.6519672375172376,
      "epoch": 0.00730857620814825,
      "grad_norm": 9.25,
      "learning_rate": 2.245e-05,
      "loss": 0.6771381855010986,
      "mean_token_accuracy": 0.8358573734760284,
      "num_tokens": 924769.0,
      "step": 450
    },
    {
      "entropy": 0.7214407650753856,
      "epoch": 0.007470989012773767,
      "grad_norm": 8.625,
      "learning_rate": 2.2950000000000002e-05,
      "loss": 0.7346301555633545,
      "mean_token_accuracy": 0.8196297101676464,
      "num_tokens": 945920.0,
      "step": 460
    },
    {
      "entropy": 0.7234843699261546,
      "epoch": 0.007633401817399283,
      "grad_norm": 12.375,
      "learning_rate": 2.345e-05,
      "loss": 0.7236921787261963,
      "mean_token_accuracy": 0.8303526669740677,
      "num_tokens": 967065.0,
      "step": 470
    },
    {
      "entropy": 0.70143166417256,
      "epoch": 0.0077958146220248005,
      "grad_norm": 6.625,
      "learning_rate": 2.395e-05,
      "loss": 0.6931093215942383,
      "mean_token_accuracy": 0.8317725829780102,
      "num_tokens": 986483.0,
      "step": 480
    },
    {
      "entropy": 0.6759544815868139,
      "epoch": 0.007958227426650318,
      "grad_norm": 7.40625,
      "learning_rate": 2.445e-05,
      "loss": 0.679323148727417,
      "mean_token_accuracy": 0.8361477315425873,
      "num_tokens": 1006208.0,
      "step": 490
    },
    {
      "entropy": 0.6707666153088212,
      "epoch": 0.008120640231275834,
      "grad_norm": 6.84375,
      "learning_rate": 2.495e-05,
      "loss": 0.6959585666656494,
      "mean_token_accuracy": 0.8390910148620605,
      "num_tokens": 1026627.0,
      "step": 500
    },
    {
      "entropy": 0.6844200716353953,
      "epoch": 0.00828305303590135,
      "grad_norm": 8.125,
      "learning_rate": 2.5450000000000002e-05,
      "loss": 0.6855648517608642,
      "mean_token_accuracy": 0.8376705981791019,
      "num_tokens": 1048389.0,
      "step": 510
    },
    {
      "entropy": 0.7058569299057126,
      "epoch": 0.008445465840526868,
      "grad_norm": 8.125,
      "learning_rate": 2.595e-05,
      "loss": 0.726499366760254,
      "mean_token_accuracy": 0.81929427459836,
      "num_tokens": 1070137.0,
      "step": 520
    },
    {
      "entropy": 0.6845117727294564,
      "epoch": 0.008607878645152384,
      "grad_norm": 9.0,
      "learning_rate": 2.6450000000000003e-05,
      "loss": 0.6557787418365478,
      "mean_token_accuracy": 0.8437227971851826,
      "num_tokens": 1090180.0,
      "step": 530
    },
    {
      "entropy": 0.7029878507833928,
      "epoch": 0.0087702914497779,
      "grad_norm": 8.6875,
      "learning_rate": 2.6950000000000005e-05,
      "loss": 0.6765800952911377,
      "mean_token_accuracy": 0.8378778167068959,
      "num_tokens": 1109304.0,
      "step": 540
    },
    {
      "entropy": 0.6664782756008207,
      "epoch": 0.008932704254403416,
      "grad_norm": 10.0625,
      "learning_rate": 2.7450000000000003e-05,
      "loss": 0.7219091892242432,
      "mean_token_accuracy": 0.8199756242334842,
      "num_tokens": 1129567.0,
      "step": 550
    },
    {
      "entropy": 0.6574279968626797,
      "epoch": 0.009095117059028934,
      "grad_norm": 7.34375,
      "learning_rate": 2.7950000000000005e-05,
      "loss": 0.6412994384765625,
      "mean_token_accuracy": 0.847743783891201,
      "num_tokens": 1149958.0,
      "step": 560
    },
    {
      "entropy": 0.6927648137323559,
      "epoch": 0.00925752986365445,
      "grad_norm": 7.5,
      "learning_rate": 2.845e-05,
      "loss": 0.682568359375,
      "mean_token_accuracy": 0.8263876050710678,
      "num_tokens": 1169775.0,
      "step": 570
    },
    {
      "entropy": 0.6439135386608541,
      "epoch": 0.009419942668279967,
      "grad_norm": 7.5625,
      "learning_rate": 2.895e-05,
      "loss": 0.5950417518615723,
      "mean_token_accuracy": 0.8508162133395671,
      "num_tokens": 1189853.0,
      "step": 580
    },
    {
      "entropy": 0.712180577032268,
      "epoch": 0.009582355472905485,
      "grad_norm": 6.90625,
      "learning_rate": 2.945e-05,
      "loss": 0.7307719707489013,
      "mean_token_accuracy": 0.8247457049787045,
      "num_tokens": 1210168.0,
      "step": 590
    },
    {
      "entropy": 0.7006784784607589,
      "epoch": 0.009744768277531,
      "grad_norm": 8.75,
      "learning_rate": 2.995e-05,
      "loss": 0.6609793663024902,
      "mean_token_accuracy": 0.8408290341496467,
      "num_tokens": 1230776.0,
      "step": 600
    },
    {
      "entropy": 0.7163872819393873,
      "epoch": 0.009907181082156517,
      "grad_norm": 8.5625,
      "learning_rate": 3.045e-05,
      "loss": 0.7742831230163574,
      "mean_token_accuracy": 0.822585117071867,
      "num_tokens": 1251914.0,
      "step": 610
    },
    {
      "entropy": 0.6246914233081042,
      "epoch": 0.010069593886782033,
      "grad_norm": 6.5,
      "learning_rate": 3.095e-05,
      "loss": 0.615490198135376,
      "mean_token_accuracy": 0.8435915000736713,
      "num_tokens": 1272275.0,
      "step": 620
    },
    {
      "entropy": 0.6043330681510269,
      "epoch": 0.010232006691407551,
      "grad_norm": 7.78125,
      "learning_rate": 3.145e-05,
      "loss": 0.5805882930755615,
      "mean_token_accuracy": 0.8595341272652149,
      "num_tokens": 1291733.0,
      "step": 630
    },
    {
      "entropy": 0.6278383932542055,
      "epoch": 0.010394419496033067,
      "grad_norm": 7.5,
      "learning_rate": 3.1950000000000004e-05,
      "loss": 0.6627034664154052,
      "mean_token_accuracy": 0.8372619360685348,
      "num_tokens": 1311579.0,
      "step": 640
    },
    {
      "entropy": 0.7058345880359411,
      "epoch": 0.010556832300658583,
      "grad_norm": 7.40625,
      "learning_rate": 3.245e-05,
      "loss": 0.6865080833435059,
      "mean_token_accuracy": 0.833170697838068,
      "num_tokens": 1331163.0,
      "step": 650
    },
    {
      "entropy": 0.6692308727651834,
      "epoch": 0.010719245105284101,
      "grad_norm": 6.84375,
      "learning_rate": 3.295e-05,
      "loss": 0.6427846431732178,
      "mean_token_accuracy": 0.848775465041399,
      "num_tokens": 1350229.0,
      "step": 660
    },
    {
      "entropy": 0.7206677223555744,
      "epoch": 0.010881657909909618,
      "grad_norm": 9.375,
      "learning_rate": 3.345000000000001e-05,
      "loss": 0.7576824188232422,
      "mean_token_accuracy": 0.8148382417857647,
      "num_tokens": 1370212.0,
      "step": 670
    },
    {
      "entropy": 0.6522296311333775,
      "epoch": 0.011044070714535134,
      "grad_norm": 8.5,
      "learning_rate": 3.3950000000000005e-05,
      "loss": 0.6881757736206054,
      "mean_token_accuracy": 0.8299212366342544,
      "num_tokens": 1390679.0,
      "step": 680
    },
    {
      "entropy": 0.6086408912204206,
      "epoch": 0.01120648351916065,
      "grad_norm": 8.4375,
      "learning_rate": 3.445e-05,
      "loss": 0.6275379180908203,
      "mean_token_accuracy": 0.8528399422764779,
      "num_tokens": 1410597.0,
      "step": 690
    },
    {
      "entropy": 0.7196457803249359,
      "epoch": 0.011368896323786168,
      "grad_norm": 6.65625,
      "learning_rate": 3.495e-05,
      "loss": 0.7207755088806153,
      "mean_token_accuracy": 0.8347064822912216,
      "num_tokens": 1431528.0,
      "step": 700
    },
    {
      "entropy": 0.6606773607432842,
      "epoch": 0.011531309128411684,
      "grad_norm": 8.875,
      "learning_rate": 3.545e-05,
      "loss": 0.6552667617797852,
      "mean_token_accuracy": 0.8414863862097264,
      "num_tokens": 1451109.0,
      "step": 710
    },
    {
      "entropy": 0.6706155283376575,
      "epoch": 0.0116937219330372,
      "grad_norm": 8.3125,
      "learning_rate": 3.595e-05,
      "loss": 0.6719372272491455,
      "mean_token_accuracy": 0.8321613404899836,
      "num_tokens": 1471205.0,
      "step": 720
    },
    {
      "entropy": 0.6477373008616268,
      "epoch": 0.011856134737662718,
      "grad_norm": 9.0625,
      "learning_rate": 3.645e-05,
      "loss": 0.6495148181915283,
      "mean_token_accuracy": 0.836396151036024,
      "num_tokens": 1491334.0,
      "step": 730
    },
    {
      "entropy": 0.6897125015035271,
      "epoch": 0.012018547542288234,
      "grad_norm": 7.71875,
      "learning_rate": 3.6950000000000004e-05,
      "loss": 0.6889693737030029,
      "mean_token_accuracy": 0.8305387049913406,
      "num_tokens": 1511805.0,
      "step": 740
    },
    {
      "entropy": 0.6438182471320033,
      "epoch": 0.01218096034691375,
      "grad_norm": 8.25,
      "learning_rate": 3.745e-05,
      "loss": 0.6293457508087158,
      "mean_token_accuracy": 0.8460224624723196,
      "num_tokens": 1531375.0,
      "step": 750
    },
    {
      "entropy": 0.6562992895022035,
      "epoch": 0.012343373151539267,
      "grad_norm": 6.8125,
      "learning_rate": 3.795e-05,
      "loss": 0.6884897708892822,
      "mean_token_accuracy": 0.8328377775847912,
      "num_tokens": 1551968.0,
      "step": 760
    },
    {
      "entropy": 0.6044487936422229,
      "epoch": 0.012505785956164785,
      "grad_norm": 7.0625,
      "learning_rate": 3.845e-05,
      "loss": 0.6261463165283203,
      "mean_token_accuracy": 0.8525740645825863,
      "num_tokens": 1571824.0,
      "step": 770
    },
    {
      "entropy": 0.6126528047956527,
      "epoch": 0.0126681987607903,
      "grad_norm": 7.625,
      "learning_rate": 3.8950000000000005e-05,
      "loss": 0.6340300083160401,
      "mean_token_accuracy": 0.8427666001021862,
      "num_tokens": 1592653.0,
      "step": 780
    },
    {
      "entropy": 0.5936457801610231,
      "epoch": 0.012830611565415817,
      "grad_norm": 10.5625,
      "learning_rate": 3.9450000000000003e-05,
      "loss": 0.60413498878479,
      "mean_token_accuracy": 0.8519487999379635,
      "num_tokens": 1613628.0,
      "step": 790
    },
    {
      "entropy": 0.6624419002793729,
      "epoch": 0.012993024370041333,
      "grad_norm": 7.5,
      "learning_rate": 3.995e-05,
      "loss": 0.66268310546875,
      "mean_token_accuracy": 0.8383010856807231,
      "num_tokens": 1634109.0,
      "step": 800
    },
    {
      "entropy": 0.7716493745334446,
      "epoch": 0.013155437174666851,
      "grad_norm": 11.625,
      "learning_rate": 4.045000000000001e-05,
      "loss": 0.7763179779052735,
      "mean_token_accuracy": 0.8299281183630228,
      "num_tokens": 1654487.0,
      "step": 810
    },
    {
      "entropy": 0.7061418435536325,
      "epoch": 0.013317849979292367,
      "grad_norm": 10.75,
      "learning_rate": 4.095e-05,
      "loss": 0.7221976280212402,
      "mean_token_accuracy": 0.8228506475687027,
      "num_tokens": 1674688.0,
      "step": 820
    },
    {
      "entropy": 0.6204464551992714,
      "epoch": 0.013480262783917884,
      "grad_norm": 12.4375,
      "learning_rate": 4.145e-05,
      "loss": 0.6605092525482178,
      "mean_token_accuracy": 0.8365028351545334,
      "num_tokens": 1695100.0,
      "step": 830
    },
    {
      "entropy": 0.6647799324244261,
      "epoch": 0.013642675588543401,
      "grad_norm": 7.1875,
      "learning_rate": 4.195e-05,
      "loss": 0.6104082107543946,
      "mean_token_accuracy": 0.8425073504447937,
      "num_tokens": 1715477.0,
      "step": 840
    },
    {
      "entropy": 0.6345407022163272,
      "epoch": 0.013805088393168918,
      "grad_norm": 10.375,
      "learning_rate": 4.245e-05,
      "loss": 0.6713770389556885,
      "mean_token_accuracy": 0.8404051989316941,
      "num_tokens": 1734868.0,
      "step": 850
    },
    {
      "entropy": 0.7121429244987667,
      "epoch": 0.013967501197794434,
      "grad_norm": 7.21875,
      "learning_rate": 4.295e-05,
      "loss": 0.7055813312530518,
      "mean_token_accuracy": 0.8333683721721172,
      "num_tokens": 1755207.0,
      "step": 860
    },
    {
      "entropy": 0.6758566972799599,
      "epoch": 0.01412991400241995,
      "grad_norm": 9.5,
      "learning_rate": 4.345e-05,
      "loss": 0.7095383167266845,
      "mean_token_accuracy": 0.833843195810914,
      "num_tokens": 1776283.0,
      "step": 870
    },
    {
      "entropy": 0.6648086906876415,
      "epoch": 0.014292326807045468,
      "grad_norm": 7.65625,
      "learning_rate": 4.3950000000000004e-05,
      "loss": 0.6750187397003173,
      "mean_token_accuracy": 0.836848222091794,
      "num_tokens": 1797574.0,
      "step": 880
    },
    {
      "entropy": 0.6431586857885122,
      "epoch": 0.014454739611670984,
      "grad_norm": 8.0625,
      "learning_rate": 4.445e-05,
      "loss": 0.6321547985076904,
      "mean_token_accuracy": 0.8366352871060372,
      "num_tokens": 1817879.0,
      "step": 890
    },
    {
      "entropy": 0.5857367699034512,
      "epoch": 0.0146171524162965,
      "grad_norm": 8.9375,
      "learning_rate": 4.495e-05,
      "loss": 0.605044174194336,
      "mean_token_accuracy": 0.8495098151266575,
      "num_tokens": 1837049.0,
      "step": 900
    },
    {
      "entropy": 0.5559913588687777,
      "epoch": 0.014779565220922018,
      "grad_norm": 6.5,
      "learning_rate": 4.545000000000001e-05,
      "loss": 0.5686515808105469,
      "mean_token_accuracy": 0.860679442435503,
      "num_tokens": 1856738.0,
      "step": 910
    },
    {
      "entropy": 0.66058643553406,
      "epoch": 0.014941978025547534,
      "grad_norm": 6.09375,
      "learning_rate": 4.5950000000000006e-05,
      "loss": 0.6703020572662354,
      "mean_token_accuracy": 0.8388163961470128,
      "num_tokens": 1877866.0,
      "step": 920
    },
    {
      "entropy": 0.6154505412094295,
      "epoch": 0.01510439083017305,
      "grad_norm": 6.90625,
      "learning_rate": 4.6450000000000004e-05,
      "loss": 0.614939832687378,
      "mean_token_accuracy": 0.8522789657115937,
      "num_tokens": 1897917.0,
      "step": 930
    },
    {
      "entropy": 0.5696395044215024,
      "epoch": 0.015266803634798567,
      "grad_norm": 5.9375,
      "learning_rate": 4.695e-05,
      "loss": 0.5447560787200928,
      "mean_token_accuracy": 0.8636209405958652,
      "num_tokens": 1916760.0,
      "step": 940
    },
    {
      "entropy": 0.6210475713945925,
      "epoch": 0.015429216439424085,
      "grad_norm": 8.125,
      "learning_rate": 4.745e-05,
      "loss": 0.6639426708221435,
      "mean_token_accuracy": 0.8451591573655606,
      "num_tokens": 1937032.0,
      "step": 950
    },
    {
      "entropy": 0.6005269814282655,
      "epoch": 0.015591629244049601,
      "grad_norm": 6.65625,
      "learning_rate": 4.795e-05,
      "loss": 0.6444539546966552,
      "mean_token_accuracy": 0.8426235385239125,
      "num_tokens": 1957253.0,
      "step": 960
    },
    {
      "entropy": 0.6851747852051631,
      "epoch": 0.01575404204867512,
      "grad_norm": 8.75,
      "learning_rate": 4.845e-05,
      "loss": 0.6680282115936279,
      "mean_token_accuracy": 0.8380367375910283,
      "num_tokens": 1976829.0,
      "step": 970
    },
    {
      "entropy": 0.7627738047856838,
      "epoch": 0.015916454853300635,
      "grad_norm": 8.875,
      "learning_rate": 4.8950000000000004e-05,
      "loss": 0.7941970825195312,
      "mean_token_accuracy": 0.8154043816030025,
      "num_tokens": 1997973.0,
      "step": 980
    },
    {
      "entropy": 0.6788709055166692,
      "epoch": 0.01607886765792615,
      "grad_norm": 5.90625,
      "learning_rate": 4.945e-05,
      "loss": 0.6722760200500488,
      "mean_token_accuracy": 0.838634318113327,
      "num_tokens": 2020914.0,
      "step": 990
    },
    {
      "entropy": 0.7065723296254873,
      "epoch": 0.016241280462551667,
      "grad_norm": 11.875,
      "learning_rate": 4.995e-05,
      "loss": 0.7217304706573486,
      "mean_token_accuracy": 0.8208229046314954,
      "num_tokens": 2041016.0,
      "step": 1000
    },
    {
      "entropy": 0.6567716733086855,
      "epoch": 0.016403693267177184,
      "grad_norm": 7.21875,
      "learning_rate": 4.999999727635224e-05,
      "loss": 0.6528256893157959,
      "mean_token_accuracy": 0.8417807046324015,
      "num_tokens": 2061456.0,
      "step": 1010
    },
    {
      "entropy": 0.67735909326002,
      "epoch": 0.0165661060718027,
      "grad_norm": 7.53125,
      "learning_rate": 4.99999878612743e-05,
      "loss": 0.6576723098754883,
      "mean_token_accuracy": 0.8372169874608517,
      "num_tokens": 2082536.0,
      "step": 1020
    },
    {
      "entropy": 0.6566119282040745,
      "epoch": 0.016728518876428216,
      "grad_norm": 7.53125,
      "learning_rate": 4.9999971721143444e-05,
      "loss": 0.7015716075897217,
      "mean_token_accuracy": 0.82717998996377,
      "num_tokens": 2103485.0,
      "step": 1030
    },
    {
      "entropy": 0.6743375883437693,
      "epoch": 0.016890931681053736,
      "grad_norm": 8.1875,
      "learning_rate": 4.9999948855964015e-05,
      "loss": 0.6762298107147217,
      "mean_token_accuracy": 0.8294478237628937,
      "num_tokens": 2124022.0,
      "step": 1040
    },
    {
      "entropy": 0.6313844756223261,
      "epoch": 0.017053344485679252,
      "grad_norm": 7.09375,
      "learning_rate": 4.9999919265742155e-05,
      "loss": 0.6340336799621582,
      "mean_token_accuracy": 0.8431316096335649,
      "num_tokens": 2144043.0,
      "step": 1050
    },
    {
      "entropy": 0.6607924028299749,
      "epoch": 0.017215757290304768,
      "grad_norm": 7.53125,
      "learning_rate": 4.999988295048583e-05,
      "loss": 0.6779970169067383,
      "mean_token_accuracy": 0.8276824347674847,
      "num_tokens": 2164537.0,
      "step": 1060
    },
    {
      "entropy": 0.6185945854289457,
      "epoch": 0.017378170094930284,
      "grad_norm": 8.1875,
      "learning_rate": 4.9999839910204816e-05,
      "loss": 0.6573071479797363,
      "mean_token_accuracy": 0.8407052546739578,
      "num_tokens": 2186031.0,
      "step": 1070
    },
    {
      "entropy": 0.674162417370826,
      "epoch": 0.0175405828995558,
      "grad_norm": 6.4375,
      "learning_rate": 4.9999790144910676e-05,
      "loss": 0.6554736614227294,
      "mean_token_accuracy": 0.8369926974177361,
      "num_tokens": 2207052.0,
      "step": 1080
    },
    {
      "entropy": 0.7422868219204247,
      "epoch": 0.017702995704181317,
      "grad_norm": 8.25,
      "learning_rate": 4.99997336546168e-05,
      "loss": 0.720973539352417,
      "mean_token_accuracy": 0.8286464180797338,
      "num_tokens": 2227432.0,
      "step": 1090
    },
    {
      "entropy": 0.6908392930403352,
      "epoch": 0.017865408508806833,
      "grad_norm": 7.34375,
      "learning_rate": 4.9999670439338385e-05,
      "loss": 0.6983513832092285,
      "mean_token_accuracy": 0.8333748318254948,
      "num_tokens": 2248598.0,
      "step": 1100
    },
    {
      "entropy": 0.603192143375054,
      "epoch": 0.018027821313432352,
      "grad_norm": 10.0625,
      "learning_rate": 4.999960049909245e-05,
      "loss": 0.5810305118560791,
      "mean_token_accuracy": 0.8501943588256836,
      "num_tokens": 2269864.0,
      "step": 1110
    },
    {
      "entropy": 0.6362513208761811,
      "epoch": 0.01819023411805787,
      "grad_norm": 5.625,
      "learning_rate": 4.999952383389779e-05,
      "loss": 0.6297592639923095,
      "mean_token_accuracy": 0.8458648063242435,
      "num_tokens": 2290890.0,
      "step": 1120
    },
    {
      "entropy": 0.7054472896736115,
      "epoch": 0.018352646922683385,
      "grad_norm": 7.40625,
      "learning_rate": 4.999944044377504e-05,
      "loss": 0.7286113739013672,
      "mean_token_accuracy": 0.8242336392402649,
      "num_tokens": 2311861.0,
      "step": 1130
    },
    {
      "entropy": 0.5913792917504906,
      "epoch": 0.0185150597273089,
      "grad_norm": 10.1875,
      "learning_rate": 4.999935032874662e-05,
      "loss": 0.5708553314208984,
      "mean_token_accuracy": 0.8517396159470081,
      "num_tokens": 2332486.0,
      "step": 1140
    },
    {
      "entropy": 0.6411094740964473,
      "epoch": 0.018677472531934417,
      "grad_norm": 8.4375,
      "learning_rate": 4.999925348883678e-05,
      "loss": 0.6693879127502441,
      "mean_token_accuracy": 0.8454736657440662,
      "num_tokens": 2353486.0,
      "step": 1150
    },
    {
      "entropy": 0.6356449664570392,
      "epoch": 0.018839885336559933,
      "grad_norm": 5.8125,
      "learning_rate": 4.9999149924071584e-05,
      "loss": 0.698000717163086,
      "mean_token_accuracy": 0.83321433365345,
      "num_tokens": 2374007.0,
      "step": 1160
    },
    {
      "entropy": 0.6261114864144475,
      "epoch": 0.01900229814118545,
      "grad_norm": 7.96875,
      "learning_rate": 4.999903963447886e-05,
      "loss": 0.6136640071868896,
      "mean_token_accuracy": 0.8442231640219688,
      "num_tokens": 2393752.0,
      "step": 1170
    },
    {
      "entropy": 0.6122575618326664,
      "epoch": 0.01916471094581097,
      "grad_norm": 6.0,
      "learning_rate": 4.9998922620088304e-05,
      "loss": 0.555076265335083,
      "mean_token_accuracy": 0.8535819552838803,
      "num_tokens": 2414294.0,
      "step": 1180
    },
    {
      "entropy": 0.628760457225144,
      "epoch": 0.019327123750436485,
      "grad_norm": 8.0,
      "learning_rate": 4.9998798880931376e-05,
      "loss": 0.6284883975982666,
      "mean_token_accuracy": 0.8457021541893482,
      "num_tokens": 2434025.0,
      "step": 1190
    },
    {
      "entropy": 0.5905484580434859,
      "epoch": 0.019489536555062,
      "grad_norm": 9.0,
      "learning_rate": 4.999866841704136e-05,
      "loss": 0.6752679347991943,
      "mean_token_accuracy": 0.8277235560119152,
      "num_tokens": 2454698.0,
      "step": 1200
    },
    {
      "entropy": 0.6085076556541026,
      "epoch": 0.019651949359687518,
      "grad_norm": 9.5625,
      "learning_rate": 4.999853122845337e-05,
      "loss": 0.5819586753845215,
      "mean_token_accuracy": 0.8564575485885143,
      "num_tokens": 2475017.0,
      "step": 1210
    },
    {
      "entropy": 0.7193897894583643,
      "epoch": 0.019814362164313034,
      "grad_norm": 7.28125,
      "learning_rate": 4.99983873152043e-05,
      "loss": 0.7372403144836426,
      "mean_token_accuracy": 0.8103011205792428,
      "num_tokens": 2495417.0,
      "step": 1220
    },
    {
      "entropy": 0.6662851992063225,
      "epoch": 0.01997677496893855,
      "grad_norm": 8.6875,
      "learning_rate": 4.9998236677332854e-05,
      "loss": 0.7223872661590576,
      "mean_token_accuracy": 0.8314795598387719,
      "num_tokens": 2515771.0,
      "step": 1230
    },
    {
      "entropy": 0.7223233181983233,
      "epoch": 0.020139187773564066,
      "grad_norm": 10.0,
      "learning_rate": 4.9998079314879566e-05,
      "loss": 0.655531358718872,
      "mean_token_accuracy": 0.8460821695625782,
      "num_tokens": 2536086.0,
      "step": 1240
    },
    {
      "entropy": 0.6147067807614803,
      "epoch": 0.020301600578189586,
      "grad_norm": 6.8125,
      "learning_rate": 4.999791522788676e-05,
      "loss": 0.6075562953948974,
      "mean_token_accuracy": 0.8397103898227215,
      "num_tokens": 2557128.0,
      "step": 1250
    },
    {
      "entropy": 0.6335235115140676,
      "epoch": 0.020464013382815102,
      "grad_norm": 9.8125,
      "learning_rate": 4.9997744416398576e-05,
      "loss": 0.6286555767059326,
      "mean_token_accuracy": 0.8338091671466827,
      "num_tokens": 2578115.0,
      "step": 1260
    },
    {
      "entropy": 0.5821547092869878,
      "epoch": 0.02062642618744062,
      "grad_norm": 9.125,
      "learning_rate": 4.999756688046097e-05,
      "loss": 0.6697710037231446,
      "mean_token_accuracy": 0.8428331598639488,
      "num_tokens": 2598353.0,
      "step": 1270
    },
    {
      "entropy": 0.6776509778574109,
      "epoch": 0.020788838992066135,
      "grad_norm": 13.5,
      "learning_rate": 4.99973826201217e-05,
      "loss": 0.674089002609253,
      "mean_token_accuracy": 0.8287919774651528,
      "num_tokens": 2618104.0,
      "step": 1280
    },
    {
      "entropy": 0.6259799963794649,
      "epoch": 0.02095125179669165,
      "grad_norm": 5.34375,
      "learning_rate": 4.999719163543032e-05,
      "loss": 0.6220345973968506,
      "mean_token_accuracy": 0.8407798357307911,
      "num_tokens": 2639320.0,
      "step": 1290
    },
    {
      "entropy": 0.6262943460606039,
      "epoch": 0.021113664601317167,
      "grad_norm": 5.96875,
      "learning_rate": 4.999699392643821e-05,
      "loss": 0.6192759513854981,
      "mean_token_accuracy": 0.8404114864766598,
      "num_tokens": 2659183.0,
      "step": 1300
    },
    {
      "entropy": 0.6027031498961151,
      "epoch": 0.021276077405942683,
      "grad_norm": 6.625,
      "learning_rate": 4.999678949319856e-05,
      "loss": 0.6081377983093261,
      "mean_token_accuracy": 0.850033626705408,
      "num_tokens": 2680174.0,
      "step": 1310
    },
    {
      "entropy": 0.6384805213660002,
      "epoch": 0.021438490210568203,
      "grad_norm": 7.15625,
      "learning_rate": 4.9996578335766355e-05,
      "loss": 0.6952379703521728,
      "mean_token_accuracy": 0.8259844411164522,
      "num_tokens": 2701385.0,
      "step": 1320
    },
    {
      "entropy": 0.6324350102804601,
      "epoch": 0.02160090301519372,
      "grad_norm": 11.8125,
      "learning_rate": 4.999636045419841e-05,
      "loss": 0.6400077819824219,
      "mean_token_accuracy": 0.845042721927166,
      "num_tokens": 2722502.0,
      "step": 1330
    },
    {
      "entropy": 0.6273757327347994,
      "epoch": 0.021763315819819235,
      "grad_norm": 7.125,
      "learning_rate": 4.999613584855333e-05,
      "loss": 0.6311779499053956,
      "mean_token_accuracy": 0.8395951971411705,
      "num_tokens": 2741959.0,
      "step": 1340
    },
    {
      "entropy": 0.6877656266093254,
      "epoch": 0.02192572862444475,
      "grad_norm": 6.65625,
      "learning_rate": 4.999590451889152e-05,
      "loss": 0.6830215930938721,
      "mean_token_accuracy": 0.8293817102909088,
      "num_tokens": 2763514.0,
      "step": 1350
    },
    {
      "entropy": 0.6326506265439094,
      "epoch": 0.022088141429070268,
      "grad_norm": 8.8125,
      "learning_rate": 4.999566646527522e-05,
      "loss": 0.6457833290100098,
      "mean_token_accuracy": 0.8413306251168251,
      "num_tokens": 2783106.0,
      "step": 1360
    },
    {
      "entropy": 0.7112111530266703,
      "epoch": 0.022250554233695784,
      "grad_norm": 11.25,
      "learning_rate": 4.9995421687768476e-05,
      "loss": 0.7356928825378418,
      "mean_token_accuracy": 0.8242002084851265,
      "num_tokens": 2805110.0,
      "step": 1370
    },
    {
      "entropy": 0.6938821950927376,
      "epoch": 0.0224129670383213,
      "grad_norm": 8.875,
      "learning_rate": 4.999517018643711e-05,
      "loss": 0.6780797958374023,
      "mean_token_accuracy": 0.8318148702383041,
      "num_tokens": 2825434.0,
      "step": 1380
    },
    {
      "entropy": 0.660740125272423,
      "epoch": 0.02257537984294682,
      "grad_norm": 7.59375,
      "learning_rate": 4.99949119613488e-05,
      "loss": 0.6329428195953369,
      "mean_token_accuracy": 0.8385577112436294,
      "num_tokens": 2845972.0,
      "step": 1390
    },
    {
      "entropy": 0.6460150822065771,
      "epoch": 0.022737792647572336,
      "grad_norm": 7.25,
      "learning_rate": 4.9994647012573004e-05,
      "loss": 0.6976537227630615,
      "mean_token_accuracy": 0.8317453965544701,
      "num_tokens": 2866846.0,
      "step": 1400
    },
    {
      "entropy": 0.6826882255263627,
      "epoch": 0.022900205452197852,
      "grad_norm": 9.25,
      "learning_rate": 4.999437534018099e-05,
      "loss": 0.7186339855194092,
      "mean_token_accuracy": 0.826085914671421,
      "num_tokens": 2886897.0,
      "step": 1410
    },
    {
      "entropy": 0.7203893756493926,
      "epoch": 0.023062618256823368,
      "grad_norm": 9.5625,
      "learning_rate": 4.9994096944245826e-05,
      "loss": 0.69368896484375,
      "mean_token_accuracy": 0.8317368142306805,
      "num_tokens": 2907810.0,
      "step": 1420
    },
    {
      "entropy": 0.6474122636020183,
      "epoch": 0.023225031061448884,
      "grad_norm": 6.5625,
      "learning_rate": 4.999381182484242e-05,
      "loss": 0.654993724822998,
      "mean_token_accuracy": 0.8460724711418152,
      "num_tokens": 2929364.0,
      "step": 1430
    },
    {
      "entropy": 0.6712787088006735,
      "epoch": 0.0233874438660744,
      "grad_norm": 7.5,
      "learning_rate": 4.9993519982047465e-05,
      "loss": 0.6528678417205811,
      "mean_token_accuracy": 0.8440250113606453,
      "num_tokens": 2949771.0,
      "step": 1440
    },
    {
      "entropy": 0.6068566062487661,
      "epoch": 0.023549856670699917,
      "grad_norm": 8.5,
      "learning_rate": 4.9993221415939464e-05,
      "loss": 0.6209955215454102,
      "mean_token_accuracy": 0.8436641719192266,
      "num_tokens": 2970708.0,
      "step": 1450
    },
    {
      "entropy": 0.6884843144565821,
      "epoch": 0.023712269475325436,
      "grad_norm": 7.375,
      "learning_rate": 4.9992916126598734e-05,
      "loss": 0.6602507591247558,
      "mean_token_accuracy": 0.8337929643690586,
      "num_tokens": 2991126.0,
      "step": 1460
    },
    {
      "entropy": 0.7101544714532793,
      "epoch": 0.023874682279950953,
      "grad_norm": 10.1875,
      "learning_rate": 4.99926041141074e-05,
      "loss": 0.7643000602722168,
      "mean_token_accuracy": 0.8093240343034267,
      "num_tokens": 3010976.0,
      "step": 1470
    },
    {
      "entropy": 0.6422369550913573,
      "epoch": 0.02403709508457647,
      "grad_norm": 9.5625,
      "learning_rate": 4.999228537854938e-05,
      "loss": 0.6396332263946534,
      "mean_token_accuracy": 0.840703946352005,
      "num_tokens": 3031020.0,
      "step": 1480
    },
    {
      "entropy": 0.6317154383286834,
      "epoch": 0.024199507889201985,
      "grad_norm": 5.875,
      "learning_rate": 4.999195992001044e-05,
      "loss": 0.6271434783935547,
      "mean_token_accuracy": 0.8460198536515235,
      "num_tokens": 3051194.0,
      "step": 1490
    },
    {
      "entropy": 0.6059641180559993,
      "epoch": 0.0243619206938275,
      "grad_norm": 8.6875,
      "learning_rate": 4.9991627738578115e-05,
      "loss": 0.5861680030822753,
      "mean_token_accuracy": 0.855878384411335,
      "num_tokens": 3073657.0,
      "step": 1500
    },
    {
      "entropy": 0.6576013986952602,
      "epoch": 0.024524333498453017,
      "grad_norm": 10.0625,
      "learning_rate": 4.9991288834341755e-05,
      "loss": 0.6853253364562988,
      "mean_token_accuracy": 0.8249302089214325,
      "num_tokens": 3093966.0,
      "step": 1510
    },
    {
      "entropy": 0.6546165156178176,
      "epoch": 0.024686746303078533,
      "grad_norm": 10.125,
      "learning_rate": 4.9990943207392534e-05,
      "loss": 0.6464301586151123,
      "mean_token_accuracy": 0.848953840136528,
      "num_tokens": 3114787.0,
      "step": 1520
    },
    {
      "entropy": 0.6702816617675126,
      "epoch": 0.02484915910770405,
      "grad_norm": 9.5625,
      "learning_rate": 4.999059085782344e-05,
      "loss": 0.7352957725524902,
      "mean_token_accuracy": 0.8313984517008066,
      "num_tokens": 3135298.0,
      "step": 1530
    },
    {
      "entropy": 0.6151116089895368,
      "epoch": 0.02501157191232957,
      "grad_norm": 7.90625,
      "learning_rate": 4.999023178572922e-05,
      "loss": 0.6389631271362305,
      "mean_token_accuracy": 0.8396187096834182,
      "num_tokens": 3154742.0,
      "step": 1540
    },
    {
      "entropy": 0.6610760375857353,
      "epoch": 0.025173984716955086,
      "grad_norm": 6.53125,
      "learning_rate": 4.998986599120651e-05,
      "loss": 0.6815928936004638,
      "mean_token_accuracy": 0.8293173111975193,
      "num_tokens": 3176582.0,
      "step": 1550
    },
    {
      "entropy": 0.7457611185498536,
      "epoch": 0.0253363975215806,
      "grad_norm": 9.8125,
      "learning_rate": 4.998949347435368e-05,
      "loss": 0.7027628421783447,
      "mean_token_accuracy": 0.8341170594096183,
      "num_tokens": 3197288.0,
      "step": 1560
    },
    {
      "entropy": 0.6536169507540762,
      "epoch": 0.025498810326206118,
      "grad_norm": 7.0,
      "learning_rate": 4.998911423527094e-05,
      "loss": 0.7027228832244873,
      "mean_token_accuracy": 0.837767268717289,
      "num_tokens": 3217391.0,
      "step": 1570
    },
    {
      "entropy": 0.6839637429453432,
      "epoch": 0.025661223130831634,
      "grad_norm": 8.0625,
      "learning_rate": 4.998872827406032e-05,
      "loss": 0.6732037544250489,
      "mean_token_accuracy": 0.8272624306380749,
      "num_tokens": 3237450.0,
      "step": 1580
    },
    {
      "entropy": 0.6068074489943683,
      "epoch": 0.02582363593545715,
      "grad_norm": 8.375,
      "learning_rate": 4.9988335590825626e-05,
      "loss": 0.5995161056518554,
      "mean_token_accuracy": 0.8518301270902157,
      "num_tokens": 3257823.0,
      "step": 1590
    },
    {
      "entropy": 0.7045330876484514,
      "epoch": 0.025986048740082666,
      "grad_norm": 6.5625,
      "learning_rate": 4.99879361856725e-05,
      "loss": 0.7300395965576172,
      "mean_token_accuracy": 0.8216715585440397,
      "num_tokens": 3278658.0,
      "step": 1600
    },
    {
      "entropy": 0.7003553207963705,
      "epoch": 0.026148461544708186,
      "grad_norm": 9.375,
      "learning_rate": 4.9987530058708386e-05,
      "loss": 0.7295712947845459,
      "mean_token_accuracy": 0.827605814114213,
      "num_tokens": 3300790.0,
      "step": 1610
    },
    {
      "entropy": 0.6379376510158181,
      "epoch": 0.026310874349333702,
      "grad_norm": 7.875,
      "learning_rate": 4.9987117210042535e-05,
      "loss": 0.6378554821014404,
      "mean_token_accuracy": 0.8439707070589065,
      "num_tokens": 3322718.0,
      "step": 1620
    },
    {
      "entropy": 0.6577769642695784,
      "epoch": 0.02647328715395922,
      "grad_norm": 14.125,
      "learning_rate": 4.998669763978599e-05,
      "loss": 0.6310774326324463,
      "mean_token_accuracy": 0.8462517261505127,
      "num_tokens": 3342580.0,
      "step": 1630
    },
    {
      "entropy": 0.686842864844948,
      "epoch": 0.026635699958584735,
      "grad_norm": 8.5,
      "learning_rate": 4.998627134805164e-05,
      "loss": 0.7196121215820312,
      "mean_token_accuracy": 0.8283118709921837,
      "num_tokens": 3362778.0,
      "step": 1640
    },
    {
      "entropy": 0.7469226297922432,
      "epoch": 0.02679811276321025,
      "grad_norm": 7.8125,
      "learning_rate": 4.998583833495414e-05,
      "loss": 0.717543077468872,
      "mean_token_accuracy": 0.8232828296720982,
      "num_tokens": 3382935.0,
      "step": 1650
    },
    {
      "entropy": 0.6493885385803878,
      "epoch": 0.026960525567835767,
      "grad_norm": 9.4375,
      "learning_rate": 4.998539860060997e-05,
      "loss": 0.642906665802002,
      "mean_token_accuracy": 0.8445032376796007,
      "num_tokens": 3402898.0,
      "step": 1660
    },
    {
      "entropy": 0.6137605054304004,
      "epoch": 0.027122938372461283,
      "grad_norm": 8.375,
      "learning_rate": 4.998495214513742e-05,
      "loss": 0.6090366840362549,
      "mean_token_accuracy": 0.8509395346045494,
      "num_tokens": 3423369.0,
      "step": 1670
    },
    {
      "entropy": 0.6584057150408625,
      "epoch": 0.027285351177086803,
      "grad_norm": 9.9375,
      "learning_rate": 4.99844989686566e-05,
      "loss": 0.684459114074707,
      "mean_token_accuracy": 0.8399760983884335,
      "num_tokens": 3443150.0,
      "step": 1680
    },
    {
      "entropy": 0.5983063196763396,
      "epoch": 0.02744776398171232,
      "grad_norm": 9.0625,
      "learning_rate": 4.998403907128941e-05,
      "loss": 0.617271614074707,
      "mean_token_accuracy": 0.8493560530245304,
      "num_tokens": 3463689.0,
      "step": 1690
    },
    {
      "entropy": 0.7433805920183658,
      "epoch": 0.027610176786337835,
      "grad_norm": 8.6875,
      "learning_rate": 4.998357245315956e-05,
      "loss": 0.7538199424743652,
      "mean_token_accuracy": 0.8160848811268806,
      "num_tokens": 3483373.0,
      "step": 1700
    },
    {
      "entropy": 0.6546157393604517,
      "epoch": 0.02777258959096335,
      "grad_norm": 7.46875,
      "learning_rate": 4.9983099114392575e-05,
      "loss": 0.652245569229126,
      "mean_token_accuracy": 0.8428901672363281,
      "num_tokens": 3503738.0,
      "step": 1710
    },
    {
      "entropy": 0.6344511146657169,
      "epoch": 0.027935002395588868,
      "grad_norm": 7.125,
      "learning_rate": 4.998261905511577e-05,
      "loss": 0.6381659984588623,
      "mean_token_accuracy": 0.839777535200119,
      "num_tokens": 3524489.0,
      "step": 1720
    },
    {
      "entropy": 0.7599488878622651,
      "epoch": 0.028097415200214384,
      "grad_norm": 9.1875,
      "learning_rate": 4.998213227545832e-05,
      "loss": 0.7533522605895996,
      "mean_token_accuracy": 0.8187672339379788,
      "num_tokens": 3545698.0,
      "step": 1730
    },
    {
      "entropy": 0.6055545805953443,
      "epoch": 0.0282598280048399,
      "grad_norm": 4.71875,
      "learning_rate": 4.998163877555112e-05,
      "loss": 0.5821003437042236,
      "mean_token_accuracy": 0.856587091833353,
      "num_tokens": 3565365.0,
      "step": 1740
    },
    {
      "entropy": 0.4880189746618271,
      "epoch": 0.02842224080946542,
      "grad_norm": 7.5,
      "learning_rate": 4.998113855552696e-05,
      "loss": 0.5201074600219726,
      "mean_token_accuracy": 0.8660154804587364,
      "num_tokens": 3584203.0,
      "step": 1750
    },
    {
      "entropy": 0.566974843107164,
      "epoch": 0.028584653614090936,
      "grad_norm": 8.25,
      "learning_rate": 4.9980631615520376e-05,
      "loss": 0.5837238788604736,
      "mean_token_accuracy": 0.8564683832228184,
      "num_tokens": 3604865.0,
      "step": 1760
    },
    {
      "entropy": 0.6121708530932665,
      "epoch": 0.028747066418716452,
      "grad_norm": 7.40625,
      "learning_rate": 4.998011795566776e-05,
      "loss": 0.6285966396331787,
      "mean_token_accuracy": 0.8450384791940451,
      "num_tokens": 3624628.0,
      "step": 1770
    },
    {
      "entropy": 0.6610128439962863,
      "epoch": 0.028909479223341968,
      "grad_norm": 7.3125,
      "learning_rate": 4.997959757610726e-05,
      "loss": 0.6536825656890869,
      "mean_token_accuracy": 0.8381315685808659,
      "num_tokens": 3645530.0,
      "step": 1780
    },
    {
      "entropy": 0.650514520611614,
      "epoch": 0.029071892027967484,
      "grad_norm": 7.125,
      "learning_rate": 4.9979070476978884e-05,
      "loss": 0.6949801445007324,
      "mean_token_accuracy": 0.8273599307984114,
      "num_tokens": 3666153.0,
      "step": 1790
    },
    {
      "entropy": 0.6403517935425043,
      "epoch": 0.029234304832593,
      "grad_norm": 6.125,
      "learning_rate": 4.9978536658424413e-05,
      "loss": 0.6089488506317139,
      "mean_token_accuracy": 0.845673780143261,
      "num_tokens": 3686500.0,
      "step": 1800
    },
    {
      "entropy": 0.7061127666383982,
      "epoch": 0.029396717637218517,
      "grad_norm": 9.3125,
      "learning_rate": 4.997799612058744e-05,
      "loss": 0.7406625270843505,
      "mean_token_accuracy": 0.8216935351490975,
      "num_tokens": 3706767.0,
      "step": 1810
    },
    {
      "entropy": 0.6842590109445155,
      "epoch": 0.029559130441844036,
      "grad_norm": 8.1875,
      "learning_rate": 4.997744886361338e-05,
      "loss": 0.6931375980377197,
      "mean_token_accuracy": 0.8345143616199493,
      "num_tokens": 3727303.0,
      "step": 1820
    },
    {
      "entropy": 0.6869341291487217,
      "epoch": 0.029721543246469553,
      "grad_norm": 8.8125,
      "learning_rate": 4.997689488764944e-05,
      "loss": 0.7084289073944092,
      "mean_token_accuracy": 0.8276141107082366,
      "num_tokens": 3748063.0,
      "step": 1830
    },
    {
      "entropy": 0.6600330295041203,
      "epoch": 0.02988395605109507,
      "grad_norm": 8.875,
      "learning_rate": 4.9976334192844646e-05,
      "loss": 0.633921480178833,
      "mean_token_accuracy": 0.8415381994098425,
      "num_tokens": 3769480.0,
      "step": 1840
    },
    {
      "entropy": 0.7443829724565149,
      "epoch": 0.030046368855720585,
      "grad_norm": 7.0625,
      "learning_rate": 4.997576677934982e-05,
      "loss": 0.7611069679260254,
      "mean_token_accuracy": 0.8150496128946543,
      "num_tokens": 3791311.0,
      "step": 1850
    },
    {
      "entropy": 0.7486577089177444,
      "epoch": 0.0302087816603461,
      "grad_norm": 12.5625,
      "learning_rate": 4.99751926473176e-05,
      "loss": 0.7113588333129883,
      "mean_token_accuracy": 0.8199596181511879,
      "num_tokens": 3813378.0,
      "step": 1860
    },
    {
      "entropy": 0.6778116539120674,
      "epoch": 0.030371194464971617,
      "grad_norm": 6.5,
      "learning_rate": 4.997461179690244e-05,
      "loss": 0.7105162620544434,
      "mean_token_accuracy": 0.8218120686709881,
      "num_tokens": 3833517.0,
      "step": 1870
    },
    {
      "entropy": 0.6684778113849461,
      "epoch": 0.030533607269597134,
      "grad_norm": 7.75,
      "learning_rate": 4.9974024228260565e-05,
      "loss": 0.6603516578674317,
      "mean_token_accuracy": 0.828897763043642,
      "num_tokens": 3853646.0,
      "step": 1880
    },
    {
      "entropy": 0.6688979282975197,
      "epoch": 0.030696020074222653,
      "grad_norm": 10.25,
      "learning_rate": 4.997342994155005e-05,
      "loss": 0.67752685546875,
      "mean_token_accuracy": 0.8379020355641842,
      "num_tokens": 3874335.0,
      "step": 1890
    },
    {
      "entropy": 0.6140990403946489,
      "epoch": 0.03085843287884817,
      "grad_norm": 9.625,
      "learning_rate": 4.997282893693076e-05,
      "loss": 0.6682180404663086,
      "mean_token_accuracy": 0.829295140132308,
      "num_tokens": 3893705.0,
      "step": 1900
    },
    {
      "entropy": 0.6512978106737137,
      "epoch": 0.031020845683473686,
      "grad_norm": 10.375,
      "learning_rate": 4.997222121456436e-05,
      "loss": 0.6616105556488037,
      "mean_token_accuracy": 0.8433177877217531,
      "num_tokens": 3913957.0,
      "step": 1910
    },
    {
      "entropy": 0.6472589701414109,
      "epoch": 0.031183258488099202,
      "grad_norm": 6.8125,
      "learning_rate": 4.9971606774614345e-05,
      "loss": 0.5852319240570069,
      "mean_token_accuracy": 0.8498035192489624,
      "num_tokens": 3934278.0,
      "step": 1920
    },
    {
      "entropy": 0.7154510764405131,
      "epoch": 0.03134567129272472,
      "grad_norm": 10.1875,
      "learning_rate": 4.997098561724597e-05,
      "loss": 0.7976554870605469,
      "mean_token_accuracy": 0.8133331589400769,
      "num_tokens": 3954782.0,
      "step": 1930
    },
    {
      "entropy": 0.6751969298813492,
      "epoch": 0.03150808409735024,
      "grad_norm": 14.0625,
      "learning_rate": 4.997035774262635e-05,
      "loss": 0.7179558753967286,
      "mean_token_accuracy": 0.8260372389107943,
      "num_tokens": 3975619.0,
      "step": 1940
    },
    {
      "entropy": 0.5933005978353322,
      "epoch": 0.03167049690197575,
      "grad_norm": 7.875,
      "learning_rate": 4.9969723150924385e-05,
      "loss": 0.5667738914489746,
      "mean_token_accuracy": 0.848986653983593,
      "num_tokens": 3996094.0,
      "step": 1950
    },
    {
      "entropy": 0.6796863836236298,
      "epoch": 0.03183290970660127,
      "grad_norm": 8.6875,
      "learning_rate": 4.9969081842310776e-05,
      "loss": 0.7167119979858398,
      "mean_token_accuracy": 0.8279650837182999,
      "num_tokens": 4016937.0,
      "step": 1960
    },
    {
      "entropy": 0.6594563874881715,
      "epoch": 0.03199532251122678,
      "grad_norm": 10.625,
      "learning_rate": 4.996843381695804e-05,
      "loss": 0.680524492263794,
      "mean_token_accuracy": 0.8339626871049404,
      "num_tokens": 4037390.0,
      "step": 1970
    },
    {
      "entropy": 0.7767967276275158,
      "epoch": 0.0321577353158523,
      "grad_norm": 7.28125,
      "learning_rate": 4.996777907504048e-05,
      "loss": 0.7757351398468018,
      "mean_token_accuracy": 0.8121193595230579,
      "num_tokens": 4058914.0,
      "step": 1980
    },
    {
      "entropy": 0.6791807080619037,
      "epoch": 0.032320148120477815,
      "grad_norm": 8.625,
      "learning_rate": 4.996711761673425e-05,
      "loss": 0.6949872493743896,
      "mean_token_accuracy": 0.8287593573331833,
      "num_tokens": 4080202.0,
      "step": 1990
    },
    {
      "entropy": 0.5941385162062943,
      "epoch": 0.032482560925103335,
      "grad_norm": 7.09375,
      "learning_rate": 4.9966449442217264e-05,
      "loss": 0.5493052959442138,
      "mean_token_accuracy": 0.8565529599785805,
      "num_tokens": 4100583.0,
      "step": 2000
    },
    {
      "entropy": 0.6329201754182577,
      "epoch": 0.032644973729728854,
      "grad_norm": 6.8125,
      "learning_rate": 4.9965774551669276e-05,
      "loss": 0.6539619922637939,
      "mean_token_accuracy": 0.8407168544828891,
      "num_tokens": 4121317.0,
      "step": 2010
    },
    {
      "entropy": 0.6138559967279434,
      "epoch": 0.03280738653435437,
      "grad_norm": 8.125,
      "learning_rate": 4.996509294527182e-05,
      "loss": 0.6240595817565918,
      "mean_token_accuracy": 0.8450943857431412,
      "num_tokens": 4140819.0,
      "step": 2020
    },
    {
      "entropy": 0.6662635295651853,
      "epoch": 0.03296979933897989,
      "grad_norm": 9.4375,
      "learning_rate": 4.996440462320826e-05,
      "loss": 0.6726973056793213,
      "mean_token_accuracy": 0.841603733971715,
      "num_tokens": 4162343.0,
      "step": 2030
    },
    {
      "entropy": 0.623607392469421,
      "epoch": 0.0331322121436054,
      "grad_norm": 6.6875,
      "learning_rate": 4.9963709585663756e-05,
      "loss": 0.6472007274627686,
      "mean_token_accuracy": 0.8434355542063713,
      "num_tokens": 4183537.0,
      "step": 2040
    },
    {
      "entropy": 0.6435566311702132,
      "epoch": 0.03329462494823092,
      "grad_norm": 9.0,
      "learning_rate": 4.9963007832825266e-05,
      "loss": 0.6425178050994873,
      "mean_token_accuracy": 0.8408272985368968,
      "num_tokens": 4203313.0,
      "step": 2050
    },
    {
      "entropy": 0.6491169307380915,
      "epoch": 0.03345703775285643,
      "grad_norm": 10.8125,
      "learning_rate": 4.9962299364881565e-05,
      "loss": 0.6958297729492188,
      "mean_token_accuracy": 0.8402622133493424,
      "num_tokens": 4222877.0,
      "step": 2060
    },
    {
      "entropy": 0.6941154435276985,
      "epoch": 0.03361945055748195,
      "grad_norm": 9.25,
      "learning_rate": 4.996158418202324e-05,
      "loss": 0.7042751789093018,
      "mean_token_accuracy": 0.839061988145113,
      "num_tokens": 4242751.0,
      "step": 2070
    },
    {
      "entropy": 0.5966043720021844,
      "epoch": 0.03378186336210747,
      "grad_norm": 7.59375,
      "learning_rate": 4.996086228444267e-05,
      "loss": 0.5783230304718018,
      "mean_token_accuracy": 0.8504262544214726,
      "num_tokens": 4264156.0,
      "step": 2080
    },
    {
      "entropy": 0.6556655721738934,
      "epoch": 0.033944276166732984,
      "grad_norm": 6.03125,
      "learning_rate": 4.9960133672334055e-05,
      "loss": 0.6465120792388916,
      "mean_token_accuracy": 0.8351392693817615,
      "num_tokens": 4285366.0,
      "step": 2090
    },
    {
      "entropy": 0.7022624588571489,
      "epoch": 0.034106688971358504,
      "grad_norm": 8.375,
      "learning_rate": 4.995939834589338e-05,
      "loss": 0.6974969387054444,
      "mean_token_accuracy": 0.8283555507659912,
      "num_tokens": 4306996.0,
      "step": 2100
    },
    {
      "entropy": 0.6127382433973253,
      "epoch": 0.034269101775984016,
      "grad_norm": 6.03125,
      "learning_rate": 4.995865630531846e-05,
      "loss": 0.6397301197052002,
      "mean_token_accuracy": 0.8459843885153532,
      "num_tokens": 4326392.0,
      "step": 2110
    },
    {
      "entropy": 0.6887777224183083,
      "epoch": 0.034431514580609536,
      "grad_norm": 9.6875,
      "learning_rate": 4.995790755080891e-05,
      "loss": 0.6962561130523681,
      "mean_token_accuracy": 0.8213602989912033,
      "num_tokens": 4346742.0,
      "step": 2120
    },
    {
      "entropy": 0.5878725726157427,
      "epoch": 0.03459392738523505,
      "grad_norm": 9.5625,
      "learning_rate": 4.995715208256613e-05,
      "loss": 0.5715918064117431,
      "mean_token_accuracy": 0.8581021271646023,
      "num_tokens": 4366715.0,
      "step": 2130
    },
    {
      "entropy": 0.620410893857479,
      "epoch": 0.03475634018986057,
      "grad_norm": 6.78125,
      "learning_rate": 4.9956389900793354e-05,
      "loss": 0.6760702133178711,
      "mean_token_accuracy": 0.835820921510458,
      "num_tokens": 4387373.0,
      "step": 2140
    },
    {
      "entropy": 0.5778453190810978,
      "epoch": 0.03491875299448609,
      "grad_norm": 8.5,
      "learning_rate": 4.995562100569561e-05,
      "loss": 0.6110926628112793,
      "mean_token_accuracy": 0.8419410035014152,
      "num_tokens": 4407378.0,
      "step": 2150
    },
    {
      "entropy": 0.6965666376054287,
      "epoch": 0.0350811657991116,
      "grad_norm": 6.625,
      "learning_rate": 4.995484539747973e-05,
      "loss": 0.7376019477844238,
      "mean_token_accuracy": 0.8128385357558727,
      "num_tokens": 4428563.0,
      "step": 2160
    },
    {
      "entropy": 0.6876415701583027,
      "epoch": 0.03524357860373712,
      "grad_norm": 6.5,
      "learning_rate": 4.9954063076354364e-05,
      "loss": 0.6575036525726319,
      "mean_token_accuracy": 0.828713221848011,
      "num_tokens": 4448744.0,
      "step": 2170
    },
    {
      "entropy": 0.6416564293205738,
      "epoch": 0.03540599140836263,
      "grad_norm": 6.34375,
      "learning_rate": 4.995327404252994e-05,
      "loss": 0.621129846572876,
      "mean_token_accuracy": 0.8493403427302837,
      "num_tokens": 4468329.0,
      "step": 2180
    },
    {
      "entropy": 0.6869419689755887,
      "epoch": 0.03556840421298815,
      "grad_norm": 8.0625,
      "learning_rate": 4.995247829621872e-05,
      "loss": 0.6869090080261231,
      "mean_token_accuracy": 0.8322326209396124,
      "num_tokens": 4489087.0,
      "step": 2190
    },
    {
      "entropy": 0.5583072192966938,
      "epoch": 0.035730817017613666,
      "grad_norm": 6.375,
      "learning_rate": 4.995167583763476e-05,
      "loss": 0.6015540599822998,
      "mean_token_accuracy": 0.8491132784634828,
      "num_tokens": 4509204.0,
      "step": 2200
    },
    {
      "entropy": 0.563787151966244,
      "epoch": 0.035893229822239185,
      "grad_norm": 9.4375,
      "learning_rate": 4.9950866666993936e-05,
      "loss": 0.5742937564849854,
      "mean_token_accuracy": 0.8562870733439922,
      "num_tokens": 4529540.0,
      "step": 2210
    },
    {
      "entropy": 0.6526942936703563,
      "epoch": 0.036055642626864705,
      "grad_norm": 6.75,
      "learning_rate": 4.995005078451389e-05,
      "loss": 0.642384147644043,
      "mean_token_accuracy": 0.8385043375194072,
      "num_tokens": 4550570.0,
      "step": 2220
    },
    {
      "entropy": 0.686827337089926,
      "epoch": 0.03621805543149022,
      "grad_norm": 8.0625,
      "learning_rate": 4.994922819041412e-05,
      "loss": 0.7027174472808838,
      "mean_token_accuracy": 0.8330831401050091,
      "num_tokens": 4570321.0,
      "step": 2230
    },
    {
      "entropy": 0.6269988880492747,
      "epoch": 0.03638046823611574,
      "grad_norm": 8.4375,
      "learning_rate": 4.994839888491589e-05,
      "loss": 0.6209956169128418,
      "mean_token_accuracy": 0.8434634730219841,
      "num_tokens": 4590661.0,
      "step": 2240
    },
    {
      "entropy": 0.6285771703347564,
      "epoch": 0.03654288104074125,
      "grad_norm": 8.375,
      "learning_rate": 4.9947562868242305e-05,
      "loss": 0.6264989852905274,
      "mean_token_accuracy": 0.8432045605033636,
      "num_tokens": 4611286.0,
      "step": 2250
    },
    {
      "entropy": 0.6674019176512956,
      "epoch": 0.03670529384536677,
      "grad_norm": 8.0625,
      "learning_rate": 4.994672014061823e-05,
      "loss": 0.6920874118804932,
      "mean_token_accuracy": 0.8319635316729546,
      "num_tokens": 4631076.0,
      "step": 2260
    },
    {
      "entropy": 0.6526093812193722,
      "epoch": 0.03686770664999228,
      "grad_norm": 7.28125,
      "learning_rate": 4.994587070227038e-05,
      "loss": 0.6433675289154053,
      "mean_token_accuracy": 0.836441696062684,
      "num_tokens": 4652108.0,
      "step": 2270
    },
    {
      "entropy": 0.6292307368479669,
      "epoch": 0.0370301194546178,
      "grad_norm": 7.53125,
      "learning_rate": 4.994501455342724e-05,
      "loss": 0.6888304233551026,
      "mean_token_accuracy": 0.8364014320075512,
      "num_tokens": 4671534.0,
      "step": 2280
    },
    {
      "entropy": 0.6040578121319413,
      "epoch": 0.03719253225924332,
      "grad_norm": 4.71875,
      "learning_rate": 4.994415169431913e-05,
      "loss": 0.6241706371307373,
      "mean_token_accuracy": 0.8489211052656174,
      "num_tokens": 4691460.0,
      "step": 2290
    },
    {
      "entropy": 0.6578195394948125,
      "epoch": 0.037354945063868834,
      "grad_norm": 7.375,
      "learning_rate": 4.994328212517816e-05,
      "loss": 0.6137680053710938,
      "mean_token_accuracy": 0.8434836030006408,
      "num_tokens": 4711957.0,
      "step": 2300
    },
    {
      "entropy": 0.6509777183644474,
      "epoch": 0.037517357868494354,
      "grad_norm": 8.75,
      "learning_rate": 4.994240584623824e-05,
      "loss": 0.6698960304260254,
      "mean_token_accuracy": 0.8323767930269241,
      "num_tokens": 4732336.0,
      "step": 2310
    },
    {
      "entropy": 0.6516154005192221,
      "epoch": 0.03767977067311987,
      "grad_norm": 7.46875,
      "learning_rate": 4.994152285773509e-05,
      "loss": 0.6755148410797119,
      "mean_token_accuracy": 0.8318980149924755,
      "num_tokens": 4751417.0,
      "step": 2320
    },
    {
      "entropy": 0.6206552578136325,
      "epoch": 0.037842183477745386,
      "grad_norm": 8.3125,
      "learning_rate": 4.994063315990624e-05,
      "loss": 0.6169948101043701,
      "mean_token_accuracy": 0.8539159312844277,
      "num_tokens": 4771601.0,
      "step": 2330
    },
    {
      "entropy": 0.6806374900974334,
      "epoch": 0.0380045962823709,
      "grad_norm": 7.625,
      "learning_rate": 4.993973675299102e-05,
      "loss": 0.7160804748535157,
      "mean_token_accuracy": 0.8221852160990238,
      "num_tokens": 4792077.0,
      "step": 2340
    },
    {
      "entropy": 0.6235373409464955,
      "epoch": 0.03816700908699642,
      "grad_norm": 6.65625,
      "learning_rate": 4.993883363723056e-05,
      "loss": 0.637391996383667,
      "mean_token_accuracy": 0.8426662363111973,
      "num_tokens": 4812432.0,
      "step": 2350
    },
    {
      "entropy": 0.6373634625226259,
      "epoch": 0.03832942189162194,
      "grad_norm": 14.0625,
      "learning_rate": 4.9937923812867804e-05,
      "loss": 0.623402452468872,
      "mean_token_accuracy": 0.8462376371026039,
      "num_tokens": 4832913.0,
      "step": 2360
    },
    {
      "entropy": 0.6548450456932187,
      "epoch": 0.03849183469624745,
      "grad_norm": 8.8125,
      "learning_rate": 4.9937007280147506e-05,
      "loss": 0.6566375255584717,
      "mean_token_accuracy": 0.8474233612418175,
      "num_tokens": 4852875.0,
      "step": 2370
    },
    {
      "entropy": 0.6077595518901944,
      "epoch": 0.03865424750087297,
      "grad_norm": 6.21875,
      "learning_rate": 4.99360840393162e-05,
      "loss": 0.5916748046875,
      "mean_token_accuracy": 0.8549061041325331,
      "num_tokens": 4872657.0,
      "step": 2380
    },
    {
      "entropy": 0.6320307263173163,
      "epoch": 0.038816660305498484,
      "grad_norm": 7.90625,
      "learning_rate": 4.9935154090622256e-05,
      "loss": 0.6422575950622559,
      "mean_token_accuracy": 0.8447648040950299,
      "num_tokens": 4893851.0,
      "step": 2390
    },
    {
      "entropy": 0.6536320655606687,
      "epoch": 0.038979073110124,
      "grad_norm": 7.15625,
      "learning_rate": 4.993421743431582e-05,
      "loss": 0.6682426929473877,
      "mean_token_accuracy": 0.8385644659399987,
      "num_tokens": 4913847.0,
      "step": 2400
    },
    {
      "entropy": 0.6405682291835546,
      "epoch": 0.039141485914749516,
      "grad_norm": 8.5,
      "learning_rate": 4.993327407064885e-05,
      "loss": 0.6972053050994873,
      "mean_token_accuracy": 0.832927867770195,
      "num_tokens": 4934444.0,
      "step": 2410
    },
    {
      "entropy": 0.689426377415657,
      "epoch": 0.039303898719375036,
      "grad_norm": 10.75,
      "learning_rate": 4.993232399987513e-05,
      "loss": 0.6514886856079102,
      "mean_token_accuracy": 0.836055675148964,
      "num_tokens": 4955378.0,
      "step": 2420
    },
    {
      "entropy": 0.6522393541410565,
      "epoch": 0.039466311524000555,
      "grad_norm": 6.4375,
      "learning_rate": 4.993136722225022e-05,
      "loss": 0.6825310230255127,
      "mean_token_accuracy": 0.8419976815581321,
      "num_tokens": 4974551.0,
      "step": 2430
    },
    {
      "entropy": 0.6360483765602112,
      "epoch": 0.03962872432862607,
      "grad_norm": 9.0,
      "learning_rate": 4.9930403738031496e-05,
      "loss": 0.6411428451538086,
      "mean_token_accuracy": 0.8455651920288801,
      "num_tokens": 4993679.0,
      "step": 2440
    },
    {
      "entropy": 0.6481083652935922,
      "epoch": 0.03979113713325159,
      "grad_norm": 8.4375,
      "learning_rate": 4.9929433547478147e-05,
      "loss": 0.6905539035797119,
      "mean_token_accuracy": 0.8374447442591191,
      "num_tokens": 5014197.0,
      "step": 2450
    },
    {
      "entropy": 0.5613165777176619,
      "epoch": 0.0399535499378771,
      "grad_norm": 6.15625,
      "learning_rate": 4.992845665085113e-05,
      "loss": 0.5980227947235107,
      "mean_token_accuracy": 0.8497913345694542,
      "num_tokens": 5033478.0,
      "step": 2460
    },
    {
      "entropy": 0.6386135086882859,
      "epoch": 0.04011596274250262,
      "grad_norm": 6.65625,
      "learning_rate": 4.9927473048413265e-05,
      "loss": 0.6227917671203613,
      "mean_token_accuracy": 0.8425655208528042,
      "num_tokens": 5054241.0,
      "step": 2470
    },
    {
      "entropy": 0.6724992427043617,
      "epoch": 0.04027837554712813,
      "grad_norm": 8.0625,
      "learning_rate": 4.992648274042913e-05,
      "loss": 0.665386152267456,
      "mean_token_accuracy": 0.8398520104587078,
      "num_tokens": 5074465.0,
      "step": 2480
    },
    {
      "entropy": 0.5983042795211077,
      "epoch": 0.04044078835175365,
      "grad_norm": 6.59375,
      "learning_rate": 4.9925485727165114e-05,
      "loss": 0.5984559059143066,
      "mean_token_accuracy": 0.8589447520673275,
      "num_tokens": 5094733.0,
      "step": 2490
    },
    {
      "entropy": 0.6814679374918342,
      "epoch": 0.04060320115637917,
      "grad_norm": 5.59375,
      "learning_rate": 4.992448200888942e-05,
      "loss": 0.6941049098968506,
      "mean_token_accuracy": 0.8306655757129192,
      "num_tokens": 5113676.0,
      "step": 2500
    },
    {
      "entropy": 0.6662566600367427,
      "epoch": 0.040765613961004685,
      "grad_norm": 10.75,
      "learning_rate": 4.992347158587205e-05,
      "loss": 0.677312707901001,
      "mean_token_accuracy": 0.8293142408132553,
      "num_tokens": 5133935.0,
      "step": 2510
    },
    {
      "entropy": 0.6669080972671508,
      "epoch": 0.040928026765630204,
      "grad_norm": 11.25,
      "learning_rate": 4.992245445838482e-05,
      "loss": 0.6824161529541015,
      "mean_token_accuracy": 0.8259681671857834,
      "num_tokens": 5153663.0,
      "step": 2520
    },
    {
      "entropy": 0.6314652482979,
      "epoch": 0.04109043957025572,
      "grad_norm": 7.9375,
      "learning_rate": 4.992143062670132e-05,
      "loss": 0.5916683197021484,
      "mean_token_accuracy": 0.850905179977417,
      "num_tokens": 5174098.0,
      "step": 2530
    },
    {
      "entropy": 0.7374189527705312,
      "epoch": 0.04125285237488124,
      "grad_norm": 9.6875,
      "learning_rate": 4.992040009109698e-05,
      "loss": 0.7898959159851074,
      "mean_token_accuracy": 0.8092671103775502,
      "num_tokens": 5194690.0,
      "step": 2540
    },
    {
      "entropy": 0.70632503782399,
      "epoch": 0.04141526517950675,
      "grad_norm": 7.90625,
      "learning_rate": 4.991936285184901e-05,
      "loss": 0.6878218650817871,
      "mean_token_accuracy": 0.8274204213172197,
      "num_tokens": 5215001.0,
      "step": 2550
    },
    {
      "entropy": 0.679837900120765,
      "epoch": 0.04157767798413227,
      "grad_norm": 8.0,
      "learning_rate": 4.9918318909236427e-05,
      "loss": 0.6833276271820068,
      "mean_token_accuracy": 0.8315151773393155,
      "num_tokens": 5235591.0,
      "step": 2560
    },
    {
      "entropy": 0.6829681197181344,
      "epoch": 0.04174009078875779,
      "grad_norm": 8.4375,
      "learning_rate": 4.991726826354006e-05,
      "loss": 0.7016257286071778,
      "mean_token_accuracy": 0.8277913201600313,
      "num_tokens": 5257601.0,
      "step": 2570
    },
    {
      "entropy": 0.595448627974838,
      "epoch": 0.0419025035933833,
      "grad_norm": 5.15625,
      "learning_rate": 4.991621091504254e-05,
      "loss": 0.5617449283599854,
      "mean_token_accuracy": 0.8588613517582416,
      "num_tokens": 5277839.0,
      "step": 2580
    },
    {
      "entropy": 0.5623145036399364,
      "epoch": 0.04206491639800882,
      "grad_norm": 10.5625,
      "learning_rate": 4.9915146864028275e-05,
      "loss": 0.5626225948333741,
      "mean_token_accuracy": 0.8517605714499951,
      "num_tokens": 5298223.0,
      "step": 2590
    },
    {
      "entropy": 0.5724875956773758,
      "epoch": 0.042227329202634334,
      "grad_norm": 6.78125,
      "learning_rate": 4.991407611078351e-05,
      "loss": 0.5911717891693116,
      "mean_token_accuracy": 0.8581851676106453,
      "num_tokens": 5318362.0,
      "step": 2600
    },
    {
      "entropy": 0.6032400705851615,
      "epoch": 0.042389742007259854,
      "grad_norm": 8.6875,
      "learning_rate": 4.991299865559629e-05,
      "loss": 0.5904241561889648,
      "mean_token_accuracy": 0.851003035530448,
      "num_tokens": 5339217.0,
      "step": 2610
    },
    {
      "entropy": 0.6398322516586632,
      "epoch": 0.042552154811885366,
      "grad_norm": 7.15625,
      "learning_rate": 4.991191449875644e-05,
      "loss": 0.6712616920471192,
      "mean_token_accuracy": 0.8360833860933781,
      "num_tokens": 5359712.0,
      "step": 2620
    },
    {
      "entropy": 0.7209280911833048,
      "epoch": 0.042714567616510886,
      "grad_norm": 9.4375,
      "learning_rate": 4.99108236405556e-05,
      "loss": 0.7008921146392822,
      "mean_token_accuracy": 0.8279690191149711,
      "num_tokens": 5381308.0,
      "step": 2630
    },
    {
      "entropy": 0.7144496374763548,
      "epoch": 0.042876980421136406,
      "grad_norm": 9.4375,
      "learning_rate": 4.990972608128722e-05,
      "loss": 0.7088531970977783,
      "mean_token_accuracy": 0.8318045917898417,
      "num_tokens": 5401685.0,
      "step": 2640
    },
    {
      "entropy": 0.6163334139157086,
      "epoch": 0.04303939322576192,
      "grad_norm": 6.96875,
      "learning_rate": 4.990862182124655e-05,
      "loss": 0.6124753952026367,
      "mean_token_accuracy": 0.8464390028268098,
      "num_tokens": 5421394.0,
      "step": 2650
    },
    {
      "entropy": 0.6740220096893609,
      "epoch": 0.04320180603038744,
      "grad_norm": 9.4375,
      "learning_rate": 4.990751086073062e-05,
      "loss": 0.7052641868591308,
      "mean_token_accuracy": 0.8279772713780403,
      "num_tokens": 5441651.0,
      "step": 2660
    },
    {
      "entropy": 0.6317045450210571,
      "epoch": 0.04336421883501295,
      "grad_norm": 5.1875,
      "learning_rate": 4.990639320003829e-05,
      "loss": 0.6352618217468262,
      "mean_token_accuracy": 0.8391068048775197,
      "num_tokens": 5462467.0,
      "step": 2670
    },
    {
      "entropy": 0.5632481563836336,
      "epoch": 0.04352663163963847,
      "grad_norm": 5.84375,
      "learning_rate": 4.9905268839470234e-05,
      "loss": 0.543444013595581,
      "mean_token_accuracy": 0.8620394960045814,
      "num_tokens": 5483588.0,
      "step": 2680
    },
    {
      "entropy": 0.6363900216296315,
      "epoch": 0.04368904444426398,
      "grad_norm": 6.25,
      "learning_rate": 4.9904137779328875e-05,
      "loss": 0.7050906181335449,
      "mean_token_accuracy": 0.8380511701107025,
      "num_tokens": 5503456.0,
      "step": 2690
    },
    {
      "entropy": 0.6352678050287068,
      "epoch": 0.0438514572488895,
      "grad_norm": 8.0625,
      "learning_rate": 4.990300001991848e-05,
      "loss": 0.6055617332458496,
      "mean_token_accuracy": 0.8458418905735016,
      "num_tokens": 5523863.0,
      "step": 2700
    },
    {
      "entropy": 0.6059881245717407,
      "epoch": 0.04401387005351502,
      "grad_norm": 9.0625,
      "learning_rate": 4.990185556154512e-05,
      "loss": 0.5839466094970703,
      "mean_token_accuracy": 0.8468217875808477,
      "num_tokens": 5543824.0,
      "step": 2710
    },
    {
      "entropy": 0.6353316034190357,
      "epoch": 0.044176282858140535,
      "grad_norm": 8.8125,
      "learning_rate": 4.9900704404516656e-05,
      "loss": 0.6560567378997803,
      "mean_token_accuracy": 0.8402208670973778,
      "num_tokens": 5563228.0,
      "step": 2720
    },
    {
      "entropy": 0.6160068994387984,
      "epoch": 0.044338695662766055,
      "grad_norm": 10.125,
      "learning_rate": 4.989954654914274e-05,
      "loss": 0.6391449451446534,
      "mean_token_accuracy": 0.846841161698103,
      "num_tokens": 5581925.0,
      "step": 2730
    },
    {
      "entropy": 0.5835158546920866,
      "epoch": 0.04450110846739157,
      "grad_norm": 5.75,
      "learning_rate": 4.9898381995734844e-05,
      "loss": 0.6077471256256104,
      "mean_token_accuracy": 0.8472371194511652,
      "num_tokens": 5602456.0,
      "step": 2740
    },
    {
      "entropy": 0.7187348742038011,
      "epoch": 0.04466352127201709,
      "grad_norm": 11.3125,
      "learning_rate": 4.9897210744606235e-05,
      "loss": 0.7577851772308349,
      "mean_token_accuracy": 0.821418435126543,
      "num_tokens": 5623190.0,
      "step": 2750
    },
    {
      "entropy": 0.7146253853105009,
      "epoch": 0.0448259340766426,
      "grad_norm": 5.59375,
      "learning_rate": 4.989603279607199e-05,
      "loss": 0.6649045944213867,
      "mean_token_accuracy": 0.8378861252218485,
      "num_tokens": 5643660.0,
      "step": 2760
    },
    {
      "entropy": 0.7270177016500383,
      "epoch": 0.04498834688126812,
      "grad_norm": 10.3125,
      "learning_rate": 4.989484815044897e-05,
      "loss": 0.7664875984191895,
      "mean_token_accuracy": 0.8206714265048504,
      "num_tokens": 5663671.0,
      "step": 2770
    },
    {
      "entropy": 0.6373738550581038,
      "epoch": 0.04515075968589364,
      "grad_norm": 11.5625,
      "learning_rate": 4.9893656808055845e-05,
      "loss": 0.6514851570129394,
      "mean_token_accuracy": 0.8311269264668226,
      "num_tokens": 5684474.0,
      "step": 2780
    },
    {
      "entropy": 0.6361436916980893,
      "epoch": 0.04531317249051915,
      "grad_norm": 7.40625,
      "learning_rate": 4.98924587692131e-05,
      "loss": 0.6313751220703125,
      "mean_token_accuracy": 0.8405101835727692,
      "num_tokens": 5703791.0,
      "step": 2790
    },
    {
      "entropy": 0.6087742246687412,
      "epoch": 0.04547558529514467,
      "grad_norm": 7.90625,
      "learning_rate": 4.9891254034242996e-05,
      "loss": 0.5916975498199463,
      "mean_token_accuracy": 0.8437209621071815,
      "num_tokens": 5723794.0,
      "step": 2800
    },
    {
      "entropy": 0.6337647124193608,
      "epoch": 0.045637998099770184,
      "grad_norm": 8.25,
      "learning_rate": 4.989004260346962e-05,
      "loss": 0.6329709053039551,
      "mean_token_accuracy": 0.8375658243894577,
      "num_tokens": 5744769.0,
      "step": 2810
    },
    {
      "entropy": 0.6656116365455091,
      "epoch": 0.045800410904395704,
      "grad_norm": 10.0,
      "learning_rate": 4.9888824477218845e-05,
      "loss": 0.6892581939697265,
      "mean_token_accuracy": 0.8408070728182793,
      "num_tokens": 5766321.0,
      "step": 2820
    },
    {
      "entropy": 0.5909729884937406,
      "epoch": 0.04596282370902122,
      "grad_norm": 7.59375,
      "learning_rate": 4.988759965581836e-05,
      "loss": 0.6100039482116699,
      "mean_token_accuracy": 0.8441951796412468,
      "num_tokens": 5786765.0,
      "step": 2830
    },
    {
      "entropy": 0.6329749719239771,
      "epoch": 0.046125236513646736,
      "grad_norm": 8.4375,
      "learning_rate": 4.9886368139597626e-05,
      "loss": 0.5884695053100586,
      "mean_token_accuracy": 0.8480496771633625,
      "num_tokens": 5808063.0,
      "step": 2840
    },
    {
      "entropy": 0.5918437451589853,
      "epoch": 0.046287649318272256,
      "grad_norm": 9.375,
      "learning_rate": 4.988512992888794e-05,
      "loss": 0.595492935180664,
      "mean_token_accuracy": 0.845385817438364,
      "num_tokens": 5828793.0,
      "step": 2850
    },
    {
      "entropy": 0.6279665031470358,
      "epoch": 0.04645006212289777,
      "grad_norm": 8.3125,
      "learning_rate": 4.9883885024022374e-05,
      "loss": 0.670698881149292,
      "mean_token_accuracy": 0.8347293566912413,
      "num_tokens": 5848752.0,
      "step": 2860
    },
    {
      "entropy": 0.6289356740191578,
      "epoch": 0.04661247492752329,
      "grad_norm": 8.9375,
      "learning_rate": 4.988263342533582e-05,
      "loss": 0.693397331237793,
      "mean_token_accuracy": 0.8231019467115402,
      "num_tokens": 5871221.0,
      "step": 2870
    },
    {
      "entropy": 0.587257741112262,
      "epoch": 0.0467748877321488,
      "grad_norm": 11.5,
      "learning_rate": 4.988137513316494e-05,
      "loss": 0.572995376586914,
      "mean_token_accuracy": 0.8613566741347313,
      "num_tokens": 5889998.0,
      "step": 2880
    },
    {
      "entropy": 0.7090319300070405,
      "epoch": 0.04693730053677432,
      "grad_norm": 5.6875,
      "learning_rate": 4.988011014784825e-05,
      "loss": 0.7409762382507324,
      "mean_token_accuracy": 0.8242812421172857,
      "num_tokens": 5910277.0,
      "step": 2890
    },
    {
      "entropy": 0.6868734234943986,
      "epoch": 0.04709971334139983,
      "grad_norm": 6.625,
      "learning_rate": 4.9878838469726e-05,
      "loss": 0.7320771217346191,
      "mean_token_accuracy": 0.8271150484681129,
      "num_tokens": 5930948.0,
      "step": 2900
    },
    {
      "entropy": 0.6431249202229082,
      "epoch": 0.04726212614602535,
      "grad_norm": 5.4375,
      "learning_rate": 4.9877560099140296e-05,
      "loss": 0.661271333694458,
      "mean_token_accuracy": 0.8406444422900676,
      "num_tokens": 5951340.0,
      "step": 2910
    },
    {
      "entropy": 0.7424022488296032,
      "epoch": 0.04742453895065087,
      "grad_norm": 5.5,
      "learning_rate": 4.987627503643502e-05,
      "loss": 0.7327793598175049,
      "mean_token_accuracy": 0.8177293986082077,
      "num_tokens": 5973098.0,
      "step": 2920
    },
    {
      "entropy": 0.6934476930648088,
      "epoch": 0.047586951755276385,
      "grad_norm": 9.25,
      "learning_rate": 4.987498328195585e-05,
      "loss": 0.6487076759338379,
      "mean_token_accuracy": 0.8421486996114254,
      "num_tokens": 5992652.0,
      "step": 2930
    },
    {
      "entropy": 0.661999425292015,
      "epoch": 0.047749364559901905,
      "grad_norm": 8.0625,
      "learning_rate": 4.987368483605028e-05,
      "loss": 0.6471284389495849,
      "mean_token_accuracy": 0.8400127746164798,
      "num_tokens": 6013071.0,
      "step": 2940
    },
    {
      "entropy": 0.6999843327328563,
      "epoch": 0.04791177736452742,
      "grad_norm": 5.65625,
      "learning_rate": 4.9872379699067584e-05,
      "loss": 0.7519407272338867,
      "mean_token_accuracy": 0.8198323249816895,
      "num_tokens": 6034594.0,
      "step": 2950
    },
    {
      "entropy": 0.7032069018110633,
      "epoch": 0.04807419016915294,
      "grad_norm": 8.8125,
      "learning_rate": 4.9871067871358847e-05,
      "loss": 0.645330810546875,
      "mean_token_accuracy": 0.8383244939148426,
      "num_tokens": 6055560.0,
      "step": 2960
    },
    {
      "entropy": 0.662582057993859,
      "epoch": 0.04823660297377845,
      "grad_norm": 7.21875,
      "learning_rate": 4.986974935327696e-05,
      "loss": 0.6841488838195801,
      "mean_token_accuracy": 0.8291450373828411,
      "num_tokens": 6076928.0,
      "step": 2970
    },
    {
      "entropy": 0.5949266533367336,
      "epoch": 0.04839901577840397,
      "grad_norm": 7.34375,
      "learning_rate": 4.986842414517661e-05,
      "loss": 0.6104370594024658,
      "mean_token_accuracy": 0.8429790183901786,
      "num_tokens": 6097585.0,
      "step": 2980
    },
    {
      "entropy": 0.6268146784044802,
      "epoch": 0.04856142858302949,
      "grad_norm": 7.625,
      "learning_rate": 4.986709224741428e-05,
      "loss": 0.6487352848052979,
      "mean_token_accuracy": 0.8405282892286777,
      "num_tokens": 6117450.0,
      "step": 2990
    },
    {
      "entropy": 0.6422375218011439,
      "epoch": 0.048723841387655,
      "grad_norm": 6.3125,
      "learning_rate": 4.986575366034824e-05,
      "loss": 0.6440940380096436,
      "mean_token_accuracy": 0.8349718920886516,
      "num_tokens": 6138334.0,
      "step": 3000
    },
    {
      "entropy": 0.6935611752793193,
      "epoch": 0.04888625419228052,
      "grad_norm": 8.6875,
      "learning_rate": 4.986440838433859e-05,
      "loss": 0.6782209396362304,
      "mean_token_accuracy": 0.8318407140672207,
      "num_tokens": 6158947.0,
      "step": 3010
    },
    {
      "entropy": 0.6235702277161181,
      "epoch": 0.049048666996906035,
      "grad_norm": 6.34375,
      "learning_rate": 4.98630564197472e-05,
      "loss": 0.5967243671417236,
      "mean_token_accuracy": 0.8466006837785244,
      "num_tokens": 6178603.0,
      "step": 3020
    },
    {
      "entropy": 0.7022992140147835,
      "epoch": 0.049211079801531554,
      "grad_norm": 6.59375,
      "learning_rate": 4.9861697766937765e-05,
      "loss": 0.7170434951782226,
      "mean_token_accuracy": 0.8257182449102402,
      "num_tokens": 6200715.0,
      "step": 3030
    },
    {
      "entropy": 0.7003879647701978,
      "epoch": 0.04937349260615707,
      "grad_norm": 6.125,
      "learning_rate": 4.986033242627575e-05,
      "loss": 0.7401596546173096,
      "mean_token_accuracy": 0.8131972953677178,
      "num_tokens": 6222628.0,
      "step": 3040
    },
    {
      "entropy": 0.6100624246522784,
      "epoch": 0.04953590541078259,
      "grad_norm": 10.3125,
      "learning_rate": 4.985896039812845e-05,
      "loss": 0.6422460079193115,
      "mean_token_accuracy": 0.8437771417200566,
      "num_tokens": 6242870.0,
      "step": 3050
    },
    {
      "entropy": 0.6766166457906365,
      "epoch": 0.0496983182154081,
      "grad_norm": 7.53125,
      "learning_rate": 4.985758168286493e-05,
      "loss": 0.6614872932434082,
      "mean_token_accuracy": 0.8315763741731643,
      "num_tokens": 6265029.0,
      "step": 3060
    },
    {
      "entropy": 0.5662491852068342,
      "epoch": 0.04986073102003362,
      "grad_norm": 8.8125,
      "learning_rate": 4.985619628085607e-05,
      "loss": 0.5579596042633057,
      "mean_token_accuracy": 0.8544669635593891,
      "num_tokens": 6285111.0,
      "step": 3070
    },
    {
      "entropy": 0.6171354637481272,
      "epoch": 0.05002314382465914,
      "grad_norm": 8.625,
      "learning_rate": 4.9854804192474556e-05,
      "loss": 0.6334100246429444,
      "mean_token_accuracy": 0.840851490944624,
      "num_tokens": 6305331.0,
      "step": 3080
    },
    {
      "entropy": 0.6492121773771942,
      "epoch": 0.05018555662928465,
      "grad_norm": 9.6875,
      "learning_rate": 4.985340541809486e-05,
      "loss": 0.6877458572387696,
      "mean_token_accuracy": 0.8281893685460091,
      "num_tokens": 6325641.0,
      "step": 3090
    },
    {
      "entropy": 0.6027677514590323,
      "epoch": 0.05034796943391017,
      "grad_norm": 7.4375,
      "learning_rate": 4.9851999958093245e-05,
      "loss": 0.5979518890380859,
      "mean_token_accuracy": 0.8514004915952682,
      "num_tokens": 6346572.0,
      "step": 3100
    },
    {
      "entropy": 0.6478552164975554,
      "epoch": 0.050510382238535684,
      "grad_norm": 7.03125,
      "learning_rate": 4.985058781284779e-05,
      "loss": 0.6190874099731445,
      "mean_token_accuracy": 0.8382560908794403,
      "num_tokens": 6366039.0,
      "step": 3110
    },
    {
      "entropy": 0.615083362814039,
      "epoch": 0.0506727950431612,
      "grad_norm": 10.1875,
      "learning_rate": 4.984916898273837e-05,
      "loss": 0.6345150470733643,
      "mean_token_accuracy": 0.8417654484510422,
      "num_tokens": 6385456.0,
      "step": 3120
    },
    {
      "entropy": 0.6141808737069369,
      "epoch": 0.050835207847786716,
      "grad_norm": 7.5625,
      "learning_rate": 4.9847743468146644e-05,
      "loss": 0.6316059112548829,
      "mean_token_accuracy": 0.8432671047747136,
      "num_tokens": 6406968.0,
      "step": 3130
    },
    {
      "entropy": 0.6828809939324856,
      "epoch": 0.050997620652412236,
      "grad_norm": 8.3125,
      "learning_rate": 4.984631126945608e-05,
      "loss": 0.6829803466796875,
      "mean_token_accuracy": 0.8207681879401207,
      "num_tokens": 6427228.0,
      "step": 3140
    },
    {
      "entropy": 0.6489829713478684,
      "epoch": 0.051160033457037755,
      "grad_norm": 7.5,
      "learning_rate": 4.9844872387051954e-05,
      "loss": 0.6350878238677978,
      "mean_token_accuracy": 0.8414121761918067,
      "num_tokens": 6446851.0,
      "step": 3150
    },
    {
      "entropy": 0.5352966155856848,
      "epoch": 0.05132244626166327,
      "grad_norm": 8.6875,
      "learning_rate": 4.984342682132131e-05,
      "loss": 0.5481082916259765,
      "mean_token_accuracy": 0.8659876964986324,
      "num_tokens": 6467984.0,
      "step": 3160
    },
    {
      "entropy": 0.6254248144570738,
      "epoch": 0.05148485906628879,
      "grad_norm": 11.8125,
      "learning_rate": 4.984197457265303e-05,
      "loss": 0.683281135559082,
      "mean_token_accuracy": 0.8344996437430382,
      "num_tokens": 6488577.0,
      "step": 3170
    },
    {
      "entropy": 0.6474242734722793,
      "epoch": 0.0516472718709143,
      "grad_norm": 8.375,
      "learning_rate": 4.9840515641437755e-05,
      "loss": 0.6612156867980957,
      "mean_token_accuracy": 0.8423461005091667,
      "num_tokens": 6509074.0,
      "step": 3180
    },
    {
      "entropy": 0.6262621279805899,
      "epoch": 0.05180968467553982,
      "grad_norm": 7.4375,
      "learning_rate": 4.983905002806796e-05,
      "loss": 0.6439118385314941,
      "mean_token_accuracy": 0.8551285330206155,
      "num_tokens": 6529124.0,
      "step": 3190
    },
    {
      "entropy": 0.6742719715461135,
      "epoch": 0.05197209748016533,
      "grad_norm": 6.65625,
      "learning_rate": 4.983757773293787e-05,
      "loss": 0.6999811172485352,
      "mean_token_accuracy": 0.8285171501338482,
      "num_tokens": 6549018.0,
      "step": 3200
    },
    {
      "entropy": 0.6714853652752936,
      "epoch": 0.05213451028479085,
      "grad_norm": 9.3125,
      "learning_rate": 4.983609875644355e-05,
      "loss": 0.6614511966705322,
      "mean_token_accuracy": 0.8380776554346084,
      "num_tokens": 6570060.0,
      "step": 3210
    },
    {
      "entropy": 0.6439214648678899,
      "epoch": 0.05229692308941637,
      "grad_norm": 10.5625,
      "learning_rate": 4.9834613098982855e-05,
      "loss": 0.6004345893859864,
      "mean_token_accuracy": 0.8492955140769481,
      "num_tokens": 6590034.0,
      "step": 3220
    },
    {
      "entropy": 0.6454493676312267,
      "epoch": 0.052459335894041885,
      "grad_norm": 8.9375,
      "learning_rate": 4.983312076095542e-05,
      "loss": 0.6671023368835449,
      "mean_token_accuracy": 0.8310371153056622,
      "num_tokens": 6610627.0,
      "step": 3230
    },
    {
      "entropy": 0.6315222263336182,
      "epoch": 0.052621748698667405,
      "grad_norm": 6.5,
      "learning_rate": 4.98316217427627e-05,
      "loss": 0.6295682430267334,
      "mean_token_accuracy": 0.8411233454942704,
      "num_tokens": 6631467.0,
      "step": 3240
    },
    {
      "entropy": 0.6201523557305336,
      "epoch": 0.05278416150329292,
      "grad_norm": 6.8125,
      "learning_rate": 4.983011604480792e-05,
      "loss": 0.6581500053405762,
      "mean_token_accuracy": 0.8345489218831063,
      "num_tokens": 6652253.0,
      "step": 3250
    },
    {
      "entropy": 0.6057658595498652,
      "epoch": 0.05294657430791844,
      "grad_norm": 7.8125,
      "learning_rate": 4.982860366749612e-05,
      "loss": 0.6194047451019287,
      "mean_token_accuracy": 0.8457935705780983,
      "num_tokens": 6674023.0,
      "step": 3260
    },
    {
      "entropy": 0.6321124321315438,
      "epoch": 0.05310898711254395,
      "grad_norm": 9.5625,
      "learning_rate": 4.982708461123414e-05,
      "loss": 0.6193779468536377,
      "mean_token_accuracy": 0.8444336891174317,
      "num_tokens": 6693482.0,
      "step": 3270
    },
    {
      "entropy": 0.6074628360569477,
      "epoch": 0.05327139991716947,
      "grad_norm": 12.25,
      "learning_rate": 4.98255588764306e-05,
      "loss": 0.6265405654907227,
      "mean_token_accuracy": 0.8478332698345185,
      "num_tokens": 6713510.0,
      "step": 3280
    },
    {
      "entropy": 0.6593122572638095,
      "epoch": 0.05343381272179499,
      "grad_norm": 6.6875,
      "learning_rate": 4.9824026463495935e-05,
      "loss": 0.6423667430877685,
      "mean_token_accuracy": 0.8253486692905426,
      "num_tokens": 6733819.0,
      "step": 3290
    },
    {
      "entropy": 0.6318762991577387,
      "epoch": 0.0535962255264205,
      "grad_norm": 7.40625,
      "learning_rate": 4.982248737284235e-05,
      "loss": 0.6779256343841553,
      "mean_token_accuracy": 0.8331849966198206,
      "num_tokens": 6755317.0,
      "step": 3300
    },
    {
      "entropy": 0.6566152520477772,
      "epoch": 0.05375863833104602,
      "grad_norm": 9.1875,
      "learning_rate": 4.982094160488389e-05,
      "loss": 0.6812151432037353,
      "mean_token_accuracy": 0.8303148686885834,
      "num_tokens": 6775917.0,
      "step": 3310
    },
    {
      "entropy": 0.6755422582384198,
      "epoch": 0.053921051135671534,
      "grad_norm": 6.15625,
      "learning_rate": 4.981938916003635e-05,
      "loss": 0.7052660942077636,
      "mean_token_accuracy": 0.8284700527787209,
      "num_tokens": 6796904.0,
      "step": 3320
    },
    {
      "entropy": 0.6428181540220976,
      "epoch": 0.054083463940297054,
      "grad_norm": 5.5,
      "learning_rate": 4.981783003871735e-05,
      "loss": 0.6236994743347168,
      "mean_token_accuracy": 0.842438717931509,
      "num_tokens": 6816656.0,
      "step": 3330
    },
    {
      "entropy": 0.6446730414405465,
      "epoch": 0.054245876744922567,
      "grad_norm": 7.65625,
      "learning_rate": 4.981626424134629e-05,
      "loss": 0.6383639812469483,
      "mean_token_accuracy": 0.839545851200819,
      "num_tokens": 6836337.0,
      "step": 3340
    },
    {
      "entropy": 0.673233937472105,
      "epoch": 0.054408289549548086,
      "grad_norm": 8.1875,
      "learning_rate": 4.981469176834439e-05,
      "loss": 0.7100165843963623,
      "mean_token_accuracy": 0.8334970086812973,
      "num_tokens": 6856421.0,
      "step": 3350
    },
    {
      "entropy": 0.6346360729075968,
      "epoch": 0.054570702354173606,
      "grad_norm": 10.125,
      "learning_rate": 4.9813112620134625e-05,
      "loss": 0.6407549858093262,
      "mean_token_accuracy": 0.8395348787307739,
      "num_tokens": 6876054.0,
      "step": 3360
    },
    {
      "entropy": 0.684779678657651,
      "epoch": 0.05473311515879912,
      "grad_norm": 7.25,
      "learning_rate": 4.9811526797141806e-05,
      "loss": 0.6471865177154541,
      "mean_token_accuracy": 0.8374948419630528,
      "num_tokens": 6897223.0,
      "step": 3370
    },
    {
      "entropy": 0.7285437794402242,
      "epoch": 0.05489552796342464,
      "grad_norm": 7.28125,
      "learning_rate": 4.9809934299792524e-05,
      "loss": 0.7308565139770508,
      "mean_token_accuracy": 0.8201438646763564,
      "num_tokens": 6917348.0,
      "step": 3380
    },
    {
      "entropy": 0.5953182641416788,
      "epoch": 0.05505794076805015,
      "grad_norm": 6.75,
      "learning_rate": 4.980833512851515e-05,
      "loss": 0.5936457633972168,
      "mean_token_accuracy": 0.85238401517272,
      "num_tokens": 6937214.0,
      "step": 3390
    },
    {
      "entropy": 0.6735324635636062,
      "epoch": 0.05522035357267567,
      "grad_norm": 8.875,
      "learning_rate": 4.9806729283739874e-05,
      "loss": 0.7508646488189697,
      "mean_token_accuracy": 0.8197361096739769,
      "num_tokens": 6957456.0,
      "step": 3400
    },
    {
      "entropy": 0.6712119284085929,
      "epoch": 0.05538276637730118,
      "grad_norm": 10.625,
      "learning_rate": 4.9805116765898676e-05,
      "loss": 0.6719058990478516,
      "mean_token_accuracy": 0.8326025024056435,
      "num_tokens": 6977485.0,
      "step": 3410
    },
    {
      "entropy": 0.7157453505322338,
      "epoch": 0.0555451791819267,
      "grad_norm": 7.28125,
      "learning_rate": 4.980349757542532e-05,
      "loss": 0.6884281158447265,
      "mean_token_accuracy": 0.8272968545556069,
      "num_tokens": 6998327.0,
      "step": 3420
    },
    {
      "entropy": 0.6955807664431631,
      "epoch": 0.05570759198655222,
      "grad_norm": 6.65625,
      "learning_rate": 4.9801871712755364e-05,
      "loss": 0.6830050468444824,
      "mean_token_accuracy": 0.8290074713528156,
      "num_tokens": 7018742.0,
      "step": 3430
    },
    {
      "entropy": 0.6512462690472602,
      "epoch": 0.055870004791177735,
      "grad_norm": 5.71875,
      "learning_rate": 4.980023917832619e-05,
      "loss": 0.639318323135376,
      "mean_token_accuracy": 0.8357956558465958,
      "num_tokens": 7038824.0,
      "step": 3440
    },
    {
      "entropy": 0.7076734066009521,
      "epoch": 0.056032417595803255,
      "grad_norm": 10.5,
      "learning_rate": 4.9798599972576944e-05,
      "loss": 0.6960817337036133,
      "mean_token_accuracy": 0.8302401900291443,
      "num_tokens": 7059257.0,
      "step": 3450
    },
    {
      "entropy": 0.7038292715325951,
      "epoch": 0.05619483040042877,
      "grad_norm": 8.0,
      "learning_rate": 4.9796954095948564e-05,
      "loss": 0.6924526691436768,
      "mean_token_accuracy": 0.8354487463831901,
      "num_tokens": 7081574.0,
      "step": 3460
    },
    {
      "entropy": 0.6379551439546048,
      "epoch": 0.05635724320505429,
      "grad_norm": 7.0,
      "learning_rate": 4.979530154888381e-05,
      "loss": 0.6768503189086914,
      "mean_token_accuracy": 0.8333509303629398,
      "num_tokens": 7101787.0,
      "step": 3470
    },
    {
      "entropy": 0.589049381390214,
      "epoch": 0.0565196560096798,
      "grad_norm": 6.46875,
      "learning_rate": 4.9793642331827216e-05,
      "loss": 0.6215909957885742,
      "mean_token_accuracy": 0.8486528433859348,
      "num_tokens": 7120899.0,
      "step": 3480
    },
    {
      "entropy": 0.6488440335728228,
      "epoch": 0.05668206881430532,
      "grad_norm": 10.5,
      "learning_rate": 4.9791976445225116e-05,
      "loss": 0.6715599536895752,
      "mean_token_accuracy": 0.8359510585665703,
      "num_tokens": 7140551.0,
      "step": 3490
    },
    {
      "entropy": 0.6717993238940835,
      "epoch": 0.05684448161893084,
      "grad_norm": 8.25,
      "learning_rate": 4.979030388952563e-05,
      "loss": 0.7009417533874511,
      "mean_token_accuracy": 0.831206327676773,
      "num_tokens": 7162405.0,
      "step": 3500
    },
    {
      "entropy": 0.6181788121350109,
      "epoch": 0.05700689442355635,
      "grad_norm": 8.75,
      "learning_rate": 4.978862466517869e-05,
      "loss": 0.6091383934020996,
      "mean_token_accuracy": 0.8513526760041714,
      "num_tokens": 7182595.0,
      "step": 3510
    },
    {
      "entropy": 0.6840516184456646,
      "epoch": 0.05716930722818187,
      "grad_norm": 8.3125,
      "learning_rate": 4.9786938772636004e-05,
      "loss": 0.6596336841583252,
      "mean_token_accuracy": 0.8308140747249126,
      "num_tokens": 7205469.0,
      "step": 3520
    },
    {
      "entropy": 0.5835915111005306,
      "epoch": 0.057331720032807384,
      "grad_norm": 6.5625,
      "learning_rate": 4.9785246212351083e-05,
      "loss": 0.6211875915527344,
      "mean_token_accuracy": 0.847928524017334,
      "num_tokens": 7226037.0,
      "step": 3530
    },
    {
      "entropy": 0.6607072552666068,
      "epoch": 0.057494132837432904,
      "grad_norm": 10.25,
      "learning_rate": 4.978354698477923e-05,
      "loss": 0.6849515438079834,
      "mean_token_accuracy": 0.8259208116680383,
      "num_tokens": 7246327.0,
      "step": 3540
    },
    {
      "entropy": 0.6076311991550029,
      "epoch": 0.05765654564205842,
      "grad_norm": 7.8125,
      "learning_rate": 4.978184109037754e-05,
      "loss": 0.5726810932159424,
      "mean_token_accuracy": 0.8491845276206732,
      "num_tokens": 7266171.0,
      "step": 3550
    },
    {
      "entropy": 0.7481310613453388,
      "epoch": 0.057818958446683937,
      "grad_norm": 7.125,
      "learning_rate": 4.978012852960491e-05,
      "loss": 0.7897719860076904,
      "mean_token_accuracy": 0.8150894939899445,
      "num_tokens": 7287541.0,
      "step": 3560
    },
    {
      "entropy": 0.7153817551210523,
      "epoch": 0.057981371251309456,
      "grad_norm": 8.5625,
      "learning_rate": 4.9778409302922e-05,
      "loss": 0.750181007385254,
      "mean_token_accuracy": 0.8134925406426191,
      "num_tokens": 7309214.0,
      "step": 3570
    },
    {
      "entropy": 0.65593605004251,
      "epoch": 0.05814378405593497,
      "grad_norm": 10.9375,
      "learning_rate": 4.9776683410791316e-05,
      "loss": 0.6135035037994385,
      "mean_token_accuracy": 0.8380598038434982,
      "num_tokens": 7330411.0,
      "step": 3580
    },
    {
      "entropy": 0.665345411747694,
      "epoch": 0.05830619686056049,
      "grad_norm": 8.3125,
      "learning_rate": 4.977495085367711e-05,
      "loss": 0.6529447078704834,
      "mean_token_accuracy": 0.8338006265461445,
      "num_tokens": 7350502.0,
      "step": 3590
    },
    {
      "entropy": 0.6232823869213462,
      "epoch": 0.058468609665186,
      "grad_norm": 8.125,
      "learning_rate": 4.977321163204544e-05,
      "loss": 0.6381690025329589,
      "mean_token_accuracy": 0.8272351827472448,
      "num_tokens": 7370261.0,
      "step": 3600
    },
    {
      "entropy": 0.6225099029019475,
      "epoch": 0.05863102246981152,
      "grad_norm": 6.125,
      "learning_rate": 4.9771465746364166e-05,
      "loss": 0.6379477977752686,
      "mean_token_accuracy": 0.8398970864713192,
      "num_tokens": 7390120.0,
      "step": 3610
    },
    {
      "entropy": 0.707944474555552,
      "epoch": 0.058793435274437034,
      "grad_norm": 6.1875,
      "learning_rate": 4.976971319710294e-05,
      "loss": 0.735454511642456,
      "mean_token_accuracy": 0.8203855894505978,
      "num_tokens": 7410105.0,
      "step": 3620
    },
    {
      "entropy": 0.6314360270276665,
      "epoch": 0.05895584807906255,
      "grad_norm": 6.65625,
      "learning_rate": 4.9767953984733196e-05,
      "loss": 0.6023908615112304,
      "mean_token_accuracy": 0.8484996825456619,
      "num_tokens": 7429911.0,
      "step": 3630
    },
    {
      "entropy": 0.6067619973793625,
      "epoch": 0.05911826088368807,
      "grad_norm": 8.0625,
      "learning_rate": 4.976618810972817e-05,
      "loss": 0.6397185325622559,
      "mean_token_accuracy": 0.8350897818803787,
      "num_tokens": 7449056.0,
      "step": 3640
    },
    {
      "entropy": 0.7075800098478794,
      "epoch": 0.059280673688313586,
      "grad_norm": 7.03125,
      "learning_rate": 4.9764415572562887e-05,
      "loss": 0.7227832317352295,
      "mean_token_accuracy": 0.8289556048810482,
      "num_tokens": 7469969.0,
      "step": 3650
    },
    {
      "entropy": 0.6366183378733694,
      "epoch": 0.059443086492939105,
      "grad_norm": 8.8125,
      "learning_rate": 4.976263637371416e-05,
      "loss": 0.673203992843628,
      "mean_token_accuracy": 0.8265930786728859,
      "num_tokens": 7490360.0,
      "step": 3660
    },
    {
      "entropy": 0.6385518738068641,
      "epoch": 0.05960549929756462,
      "grad_norm": 7.59375,
      "learning_rate": 4.976085051366058e-05,
      "loss": 0.6257910251617431,
      "mean_token_accuracy": 0.8463421709835529,
      "num_tokens": 7512677.0,
      "step": 3670
    },
    {
      "entropy": 0.6634161598049104,
      "epoch": 0.05976791210219014,
      "grad_norm": 10.0625,
      "learning_rate": 4.9759057992882587e-05,
      "loss": 0.6892859458923339,
      "mean_token_accuracy": 0.8272095255553722,
      "num_tokens": 7531436.0,
      "step": 3680
    },
    {
      "entropy": 0.6745593910571188,
      "epoch": 0.05993032490681565,
      "grad_norm": 7.0625,
      "learning_rate": 4.975725881186234e-05,
      "loss": 0.6488647937774659,
      "mean_token_accuracy": 0.8441109918057919,
      "num_tokens": 7551673.0,
      "step": 3690
    },
    {
      "entropy": 0.616761663928628,
      "epoch": 0.06009273771144117,
      "grad_norm": 9.1875,
      "learning_rate": 4.975545297108384e-05,
      "loss": 0.607551097869873,
      "mean_token_accuracy": 0.8412525683641434,
      "num_tokens": 7571978.0,
      "step": 3700
    },
    {
      "entropy": 0.6694912530481816,
      "epoch": 0.06025515051606669,
      "grad_norm": 8.125,
      "learning_rate": 4.975364047103286e-05,
      "loss": 0.7068482875823975,
      "mean_token_accuracy": 0.8315350703895092,
      "num_tokens": 7592049.0,
      "step": 3710
    },
    {
      "entropy": 0.6048460864461959,
      "epoch": 0.0604175633206922,
      "grad_norm": 8.625,
      "learning_rate": 4.9751821312196955e-05,
      "loss": 0.6369690418243408,
      "mean_token_accuracy": 0.8351354736834764,
      "num_tokens": 7612880.0,
      "step": 3720
    },
    {
      "entropy": 0.7358152823522687,
      "epoch": 0.06057997612531772,
      "grad_norm": 7.96875,
      "learning_rate": 4.974999549506549e-05,
      "loss": 0.7889390468597413,
      "mean_token_accuracy": 0.8173042796552181,
      "num_tokens": 7633734.0,
      "step": 3730
    },
    {
      "entropy": 0.6880794854834675,
      "epoch": 0.060742388929943235,
      "grad_norm": 7.4375,
      "learning_rate": 4.974816302012962e-05,
      "loss": 0.684275770187378,
      "mean_token_accuracy": 0.8307307705283165,
      "num_tokens": 7656054.0,
      "step": 3740
    },
    {
      "entropy": 0.6932039026170969,
      "epoch": 0.060904801734568755,
      "grad_norm": 10.875,
      "learning_rate": 4.9746323887882275e-05,
      "loss": 0.6498753070831299,
      "mean_token_accuracy": 0.8387250781059266,
      "num_tokens": 7677207.0,
      "step": 3750
    },
    {
      "entropy": 0.6073678065091371,
      "epoch": 0.06106721453919427,
      "grad_norm": 7.0625,
      "learning_rate": 4.9744478098818195e-05,
      "loss": 0.5961812973022461,
      "mean_token_accuracy": 0.8385385982692242,
      "num_tokens": 7697083.0,
      "step": 3760
    },
    {
      "entropy": 0.6104508604854345,
      "epoch": 0.06122962734381979,
      "grad_norm": 5.96875,
      "learning_rate": 4.974262565343389e-05,
      "loss": 0.5813981056213379,
      "mean_token_accuracy": 0.8522955946624279,
      "num_tokens": 7718347.0,
      "step": 3770
    },
    {
      "entropy": 0.6045610118657351,
      "epoch": 0.06139204014844531,
      "grad_norm": 7.09375,
      "learning_rate": 4.974076655222769e-05,
      "loss": 0.6310525417327881,
      "mean_token_accuracy": 0.8386100128293037,
      "num_tokens": 7738530.0,
      "step": 3780
    },
    {
      "entropy": 0.5910926104523242,
      "epoch": 0.06155445295307082,
      "grad_norm": 9.0625,
      "learning_rate": 4.973890079569967e-05,
      "loss": 0.6282540321350097,
      "mean_token_accuracy": 0.8489861045032739,
      "num_tokens": 7758434.0,
      "step": 3790
    },
    {
      "entropy": 0.5710391078144312,
      "epoch": 0.06171686575769634,
      "grad_norm": 13.0,
      "learning_rate": 4.9737028384351755e-05,
      "loss": 0.6210896968841553,
      "mean_token_accuracy": 0.8436822183430195,
      "num_tokens": 7779365.0,
      "step": 3800
    },
    {
      "entropy": 0.7116972275078297,
      "epoch": 0.06187927856232185,
      "grad_norm": 8.0,
      "learning_rate": 4.97351493186876e-05,
      "loss": 0.766285514831543,
      "mean_token_accuracy": 0.8200442895293236,
      "num_tokens": 7799334.0,
      "step": 3810
    },
    {
      "entropy": 0.67995271733962,
      "epoch": 0.06204169136694737,
      "grad_norm": 6.3125,
      "learning_rate": 4.9733263599212695e-05,
      "loss": 0.6566415786743164,
      "mean_token_accuracy": 0.8402710676193237,
      "num_tokens": 7819052.0,
      "step": 3820
    },
    {
      "entropy": 0.6811694384552538,
      "epoch": 0.062204104171572884,
      "grad_norm": 6.4375,
      "learning_rate": 4.97313712264343e-05,
      "loss": 0.6304548740386963,
      "mean_token_accuracy": 0.8392180155962705,
      "num_tokens": 7839838.0,
      "step": 3830
    },
    {
      "entropy": 0.6380248936824501,
      "epoch": 0.062366516976198404,
      "grad_norm": 11.6875,
      "learning_rate": 4.972947220086146e-05,
      "loss": 0.6095435142517089,
      "mean_token_accuracy": 0.846200242638588,
      "num_tokens": 7859579.0,
      "step": 3840
    },
    {
      "entropy": 0.5935035164933652,
      "epoch": 0.06252892978082392,
      "grad_norm": 4.9375,
      "learning_rate": 4.972756652300502e-05,
      "loss": 0.5667986392974853,
      "mean_token_accuracy": 0.8525295637547969,
      "num_tokens": 7879808.0,
      "step": 3850
    },
    {
      "entropy": 0.6477035215124488,
      "epoch": 0.06269134258544944,
      "grad_norm": 7.5625,
      "learning_rate": 4.972565419337761e-05,
      "loss": 0.7064467430114746,
      "mean_token_accuracy": 0.8296240210533142,
      "num_tokens": 7901563.0,
      "step": 3860
    },
    {
      "entropy": 0.6724066206254065,
      "epoch": 0.06285375539007496,
      "grad_norm": 6.9375,
      "learning_rate": 4.9723735212493674e-05,
      "loss": 0.6806609153747558,
      "mean_token_accuracy": 0.833264297246933,
      "num_tokens": 7922549.0,
      "step": 3870
    },
    {
      "entropy": 0.622235485073179,
      "epoch": 0.06301616819470048,
      "grad_norm": 7.25,
      "learning_rate": 4.9721809580869385e-05,
      "loss": 0.6572322368621826,
      "mean_token_accuracy": 0.8402150720357895,
      "num_tokens": 7942661.0,
      "step": 3880
    },
    {
      "entropy": 0.717705905623734,
      "epoch": 0.06317858099932598,
      "grad_norm": 7.34375,
      "learning_rate": 4.9719877299022765e-05,
      "loss": 0.7567597389221191,
      "mean_token_accuracy": 0.8185842230916023,
      "num_tokens": 7963771.0,
      "step": 3890
    },
    {
      "entropy": 0.5887428083457052,
      "epoch": 0.0633409938039515,
      "grad_norm": 6.75,
      "learning_rate": 4.9717938367473593e-05,
      "loss": 0.5570251941680908,
      "mean_token_accuracy": 0.8605486780405045,
      "num_tokens": 7983205.0,
      "step": 3900
    },
    {
      "entropy": 0.54711245726794,
      "epoch": 0.06350340660857702,
      "grad_norm": 7.28125,
      "learning_rate": 4.971599278674346e-05,
      "loss": 0.564736270904541,
      "mean_token_accuracy": 0.8536185957491398,
      "num_tokens": 8003663.0,
      "step": 3910
    },
    {
      "entropy": 0.6282231878489256,
      "epoch": 0.06366581941320254,
      "grad_norm": 7.0625,
      "learning_rate": 4.971404055735571e-05,
      "loss": 0.672048807144165,
      "mean_token_accuracy": 0.8340695530176163,
      "num_tokens": 8024170.0,
      "step": 3920
    },
    {
      "entropy": 0.5719426597468555,
      "epoch": 0.06382823221782806,
      "grad_norm": 8.5,
      "learning_rate": 4.971208167983552e-05,
      "loss": 0.5595303058624268,
      "mean_token_accuracy": 0.8493384972214699,
      "num_tokens": 8045765.0,
      "step": 3930
    },
    {
      "entropy": 0.6285661475732922,
      "epoch": 0.06399064502245357,
      "grad_norm": 6.6875,
      "learning_rate": 4.971011615470981e-05,
      "loss": 0.6237983703613281,
      "mean_token_accuracy": 0.8351557396352292,
      "num_tokens": 8065808.0,
      "step": 3940
    },
    {
      "entropy": 0.556355993822217,
      "epoch": 0.06415305782707909,
      "grad_norm": 7.6875,
      "learning_rate": 4.9708143982507326e-05,
      "loss": 0.5424407958984375,
      "mean_token_accuracy": 0.8569916449487209,
      "num_tokens": 8086332.0,
      "step": 3950
    },
    {
      "entropy": 0.6120845274999738,
      "epoch": 0.0643154706317046,
      "grad_norm": 6.84375,
      "learning_rate": 4.970616516375859e-05,
      "loss": 0.6195963859558106,
      "mean_token_accuracy": 0.8409288100898266,
      "num_tokens": 8107618.0,
      "step": 3960
    },
    {
      "entropy": 0.6709695937111974,
      "epoch": 0.06447788343633012,
      "grad_norm": 8.9375,
      "learning_rate": 4.970417969899589e-05,
      "loss": 0.6882740974426269,
      "mean_token_accuracy": 0.8349254507571459,
      "num_tokens": 8127660.0,
      "step": 3970
    },
    {
      "entropy": 0.6144146902486682,
      "epoch": 0.06464029624095563,
      "grad_norm": 7.90625,
      "learning_rate": 4.9702187588753334e-05,
      "loss": 0.6158870697021485,
      "mean_token_accuracy": 0.8434836991131306,
      "num_tokens": 8147886.0,
      "step": 3980
    },
    {
      "entropy": 0.5471114906016737,
      "epoch": 0.06480270904558115,
      "grad_norm": 5.46875,
      "learning_rate": 4.970018883356681e-05,
      "loss": 0.5272904396057129,
      "mean_token_accuracy": 0.8653822094202042,
      "num_tokens": 8167565.0,
      "step": 3990
    },
    {
      "epoch": 0.06496512185020667,
      "eval_entropy": 0.6584339497089386,
      "eval_loss": 0.687670111656189,
      "eval_mean_token_accuracy": 0.8280174841880799,
      "eval_num_tokens": 8188940.0,
      "eval_runtime": 40.4592,
      "eval_samples_per_second": 49.432,
      "eval_steps_per_second": 6.179,
      "step": 4000
    },
    {
      "entropy": 0.59657037188299,
      "epoch": 0.06512753465483219,
      "grad_norm": 18.375,
      "learning_rate": 4.969617139051429e-05,
      "loss": 0.6662103652954101,
      "mean_token_accuracy": 0.8357427053153514,
      "num_tokens": 8208086.0,
      "step": 4010
    },
    {
      "entropy": 0.6556566601619125,
      "epoch": 0.06528994745945771,
      "grad_norm": 8.6875,
      "learning_rate": 4.9694152703729004e-05,
      "loss": 0.6410644054412842,
      "mean_token_accuracy": 0.8400427147746086,
      "num_tokens": 8227787.0,
      "step": 4020
    },
    {
      "entropy": 0.6485343052074313,
      "epoch": 0.06545236026408321,
      "grad_norm": 7.6875,
      "learning_rate": 4.9692127374161136e-05,
      "loss": 0.6166937351226807,
      "mean_token_accuracy": 0.8426643446087837,
      "num_tokens": 8248157.0,
      "step": 4030
    },
    {
      "entropy": 0.6764893119223416,
      "epoch": 0.06561477306870873,
      "grad_norm": 7.09375,
      "learning_rate": 4.969009540235551e-05,
      "loss": 0.6841977119445801,
      "mean_token_accuracy": 0.8340508855879307,
      "num_tokens": 8268849.0,
      "step": 4040
    },
    {
      "entropy": 0.6103664681315422,
      "epoch": 0.06577718587333425,
      "grad_norm": 6.90625,
      "learning_rate": 4.968805678885875e-05,
      "loss": 0.6078847885131836,
      "mean_token_accuracy": 0.8430127948522568,
      "num_tokens": 8288273.0,
      "step": 4050
    },
    {
      "entropy": 0.6488130383193493,
      "epoch": 0.06593959867795977,
      "grad_norm": 6.375,
      "learning_rate": 4.968601153421921e-05,
      "loss": 0.6693777561187744,
      "mean_token_accuracy": 0.8321704961359501,
      "num_tokens": 8309094.0,
      "step": 4060
    },
    {
      "entropy": 0.6404385806061328,
      "epoch": 0.0661020114825853,
      "grad_norm": 8.1875,
      "learning_rate": 4.96839596389871e-05,
      "loss": 0.6492941856384278,
      "mean_token_accuracy": 0.8379133835434913,
      "num_tokens": 8328906.0,
      "step": 4070
    },
    {
      "entropy": 0.6383732973597944,
      "epoch": 0.0662644242872108,
      "grad_norm": 7.875,
      "learning_rate": 4.968190110371438e-05,
      "loss": 0.683575439453125,
      "mean_token_accuracy": 0.8337679870426655,
      "num_tokens": 8350405.0,
      "step": 4080
    },
    {
      "entropy": 0.7279273853637278,
      "epoch": 0.06642683709183632,
      "grad_norm": 6.1875,
      "learning_rate": 4.967983592895478e-05,
      "loss": 0.7294415473937989,
      "mean_token_accuracy": 0.8200351405888796,
      "num_tokens": 8371903.0,
      "step": 4090
    },
    {
      "entropy": 0.6089258077554405,
      "epoch": 0.06658924989646184,
      "grad_norm": 8.4375,
      "learning_rate": 4.967776411526386e-05,
      "loss": 0.6082272529602051,
      "mean_token_accuracy": 0.8504539206624031,
      "num_tokens": 8392929.0,
      "step": 4100
    },
    {
      "entropy": 0.6479860631749034,
      "epoch": 0.06675166270108736,
      "grad_norm": 7.53125,
      "learning_rate": 4.967568566319893e-05,
      "loss": 0.6464333534240723,
      "mean_token_accuracy": 0.8407898757606744,
      "num_tokens": 8413921.0,
      "step": 4110
    },
    {
      "entropy": 0.6782472599297762,
      "epoch": 0.06691407550571286,
      "grad_norm": 8.375,
      "learning_rate": 4.96736005733191e-05,
      "loss": 0.7056226253509521,
      "mean_token_accuracy": 0.8367070361971856,
      "num_tokens": 8433785.0,
      "step": 4120
    },
    {
      "entropy": 0.6130849519744516,
      "epoch": 0.06707648831033838,
      "grad_norm": 6.9375,
      "learning_rate": 4.9671508846185266e-05,
      "loss": 0.6041444778442383,
      "mean_token_accuracy": 0.8490801520645619,
      "num_tokens": 8453684.0,
      "step": 4130
    },
    {
      "entropy": 0.6525204100646078,
      "epoch": 0.0672389011149639,
      "grad_norm": 8.6875,
      "learning_rate": 4.966941048236011e-05,
      "loss": 0.6274653434753418,
      "mean_token_accuracy": 0.844075383245945,
      "num_tokens": 8473500.0,
      "step": 4140
    },
    {
      "entropy": 0.6187574036419392,
      "epoch": 0.06740131391958942,
      "grad_norm": 5.75,
      "learning_rate": 4.966730548240809e-05,
      "loss": 0.6418675422668457,
      "mean_token_accuracy": 0.8527890175580979,
      "num_tokens": 8492844.0,
      "step": 4150
    },
    {
      "entropy": 0.5923375618644059,
      "epoch": 0.06756372672421494,
      "grad_norm": 5.90625,
      "learning_rate": 4.9665193846895465e-05,
      "loss": 0.6104167938232422,
      "mean_token_accuracy": 0.845196595042944,
      "num_tokens": 8514070.0,
      "step": 4160
    },
    {
      "entropy": 0.5581407878547907,
      "epoch": 0.06772613952884045,
      "grad_norm": 7.8125,
      "learning_rate": 4.9663075576390254e-05,
      "loss": 0.5616621017456055,
      "mean_token_accuracy": 0.8598431661725044,
      "num_tokens": 8533457.0,
      "step": 4170
    },
    {
      "entropy": 0.5695889973081648,
      "epoch": 0.06788855233346597,
      "grad_norm": 8.75,
      "learning_rate": 4.966095067146229e-05,
      "loss": 0.588599157333374,
      "mean_token_accuracy": 0.846741109341383,
      "num_tokens": 8553780.0,
      "step": 4180
    },
    {
      "entropy": 0.7367532709613442,
      "epoch": 0.06805096513809149,
      "grad_norm": 7.40625,
      "learning_rate": 4.9658819132683176e-05,
      "loss": 0.8038776397705079,
      "mean_token_accuracy": 0.8104439463466406,
      "num_tokens": 8573445.0,
      "step": 4190
    },
    {
      "entropy": 0.6487067444249988,
      "epoch": 0.06821337794271701,
      "grad_norm": 10.375,
      "learning_rate": 4.965668096062629e-05,
      "loss": 0.6201260566711426,
      "mean_token_accuracy": 0.8406964227557182,
      "num_tokens": 8593615.0,
      "step": 4200
    },
    {
      "entropy": 0.7759757129475474,
      "epoch": 0.06837579074734253,
      "grad_norm": 6.71875,
      "learning_rate": 4.9654536155866824e-05,
      "loss": 0.749416732788086,
      "mean_token_accuracy": 0.8167964987456798,
      "num_tokens": 8617507.0,
      "step": 4210
    },
    {
      "entropy": 0.6116909880191088,
      "epoch": 0.06853820355196803,
      "grad_norm": 8.5625,
      "learning_rate": 4.965238471898172e-05,
      "loss": 0.6042019367218018,
      "mean_token_accuracy": 0.8502664744853974,
      "num_tokens": 8637588.0,
      "step": 4220
    },
    {
      "entropy": 0.6624170936644077,
      "epoch": 0.06870061635659355,
      "grad_norm": 10.3125,
      "learning_rate": 4.965022665054972e-05,
      "loss": 0.7312849521636963,
      "mean_token_accuracy": 0.8316965609788894,
      "num_tokens": 8657565.0,
      "step": 4230
    },
    {
      "entropy": 0.5678249921649694,
      "epoch": 0.06886302916121907,
      "grad_norm": 8.5,
      "learning_rate": 4.964806195115136e-05,
      "loss": 0.5588680744171143,
      "mean_token_accuracy": 0.8551277458667755,
      "num_tokens": 8677706.0,
      "step": 4240
    },
    {
      "entropy": 0.6550601082853973,
      "epoch": 0.06902544196584459,
      "grad_norm": 6.6875,
      "learning_rate": 4.964589062136894e-05,
      "loss": 0.6155716896057128,
      "mean_token_accuracy": 0.8475619602948428,
      "num_tokens": 8698804.0,
      "step": 4250
    },
    {
      "entropy": 0.6035366371273995,
      "epoch": 0.0691878547704701,
      "grad_norm": 9.6875,
      "learning_rate": 4.964371266178656e-05,
      "loss": 0.6050459384918213,
      "mean_token_accuracy": 0.8467766128480434,
      "num_tokens": 8719528.0,
      "step": 4260
    },
    {
      "entropy": 0.6342330690473318,
      "epoch": 0.06935026757509562,
      "grad_norm": 10.0625,
      "learning_rate": 4.964152807299008e-05,
      "loss": 0.6773311614990234,
      "mean_token_accuracy": 0.8325214240700006,
      "num_tokens": 8740859.0,
      "step": 4270
    },
    {
      "entropy": 0.6128977849148214,
      "epoch": 0.06951268037972114,
      "grad_norm": 7.71875,
      "learning_rate": 4.963933685556717e-05,
      "loss": 0.6456241607666016,
      "mean_token_accuracy": 0.8447628572583199,
      "num_tokens": 8761760.0,
      "step": 4280
    },
    {
      "entropy": 0.5949186106212437,
      "epoch": 0.06967509318434666,
      "grad_norm": 6.3125,
      "learning_rate": 4.963713901010728e-05,
      "loss": 0.5786555767059326,
      "mean_token_accuracy": 0.8502746544778347,
      "num_tokens": 8782382.0,
      "step": 4290
    },
    {
      "entropy": 0.7527461871504784,
      "epoch": 0.06983750598897218,
      "grad_norm": 6.28125,
      "learning_rate": 4.963493453720162e-05,
      "loss": 0.7537468910217285,
      "mean_token_accuracy": 0.8092567604035139,
      "num_tokens": 8802014.0,
      "step": 4300
    },
    {
      "entropy": 0.6926262227818369,
      "epoch": 0.06999991879359768,
      "grad_norm": 9.3125,
      "learning_rate": 4.963272343744321e-05,
      "loss": 0.7002501964569092,
      "mean_token_accuracy": 0.8282694987952709,
      "num_tokens": 8823176.0,
      "step": 4310
    },
    {
      "entropy": 0.6235510830767452,
      "epoch": 0.0701623315982232,
      "grad_norm": 7.46875,
      "learning_rate": 4.963050571142684e-05,
      "loss": 0.6615278720855713,
      "mean_token_accuracy": 0.8476155772805214,
      "num_tokens": 8843165.0,
      "step": 4320
    },
    {
      "entropy": 0.7028528685681522,
      "epoch": 0.07032474440284872,
      "grad_norm": 8.125,
      "learning_rate": 4.962828135974907e-05,
      "loss": 0.694635534286499,
      "mean_token_accuracy": 0.8282914761453867,
      "num_tokens": 8862884.0,
      "step": 4330
    },
    {
      "entropy": 0.6162197687663138,
      "epoch": 0.07048715720747424,
      "grad_norm": 6.8125,
      "learning_rate": 4.962605038300828e-05,
      "loss": 0.6006672382354736,
      "mean_token_accuracy": 0.8466765988618136,
      "num_tokens": 8883222.0,
      "step": 4340
    },
    {
      "entropy": 0.6671367372386158,
      "epoch": 0.07064957001209976,
      "grad_norm": 8.0625,
      "learning_rate": 4.9623812781804584e-05,
      "loss": 0.6649251461029053,
      "mean_token_accuracy": 0.831134419888258,
      "num_tokens": 8905739.0,
      "step": 4350
    },
    {
      "entropy": 0.6145354522392154,
      "epoch": 0.07081198281672527,
      "grad_norm": 7.1875,
      "learning_rate": 4.9621568556739915e-05,
      "loss": 0.6461136817932129,
      "mean_token_accuracy": 0.8431534878909588,
      "num_tokens": 8925618.0,
      "step": 4360
    },
    {
      "entropy": 0.6345865175127983,
      "epoch": 0.07097439562135079,
      "grad_norm": 10.625,
      "learning_rate": 4.9619317708417965e-05,
      "loss": 0.6354070663452148,
      "mean_token_accuracy": 0.8425127878785134,
      "num_tokens": 8946401.0,
      "step": 4370
    },
    {
      "entropy": 0.6136418614536524,
      "epoch": 0.0711368084259763,
      "grad_norm": 7.53125,
      "learning_rate": 4.9617060237444227e-05,
      "loss": 0.6376131534576416,
      "mean_token_accuracy": 0.8425911530852318,
      "num_tokens": 8966725.0,
      "step": 4380
    },
    {
      "entropy": 0.6740480018779635,
      "epoch": 0.07129922123060183,
      "grad_norm": 8.1875,
      "learning_rate": 4.961479614442597e-05,
      "loss": 0.675442123413086,
      "mean_token_accuracy": 0.8270412348210812,
      "num_tokens": 8988854.0,
      "step": 4390
    },
    {
      "entropy": 0.6035504080355167,
      "epoch": 0.07146163403522733,
      "grad_norm": 10.5625,
      "learning_rate": 4.9612525429972215e-05,
      "loss": 0.6455621242523193,
      "mean_token_accuracy": 0.8451152667403221,
      "num_tokens": 9009574.0,
      "step": 4400
    },
    {
      "entropy": 0.594661158695817,
      "epoch": 0.07162404683985285,
      "grad_norm": 8.625,
      "learning_rate": 4.961024809469382e-05,
      "loss": 0.6251354694366456,
      "mean_token_accuracy": 0.8522915355861187,
      "num_tokens": 9028996.0,
      "step": 4410
    },
    {
      "entropy": 0.6308258430100977,
      "epoch": 0.07178645964447837,
      "grad_norm": 8.3125,
      "learning_rate": 4.960796413920337e-05,
      "loss": 0.6239512920379638,
      "mean_token_accuracy": 0.8413037165999413,
      "num_tokens": 9048121.0,
      "step": 4420
    },
    {
      "entropy": 0.6547151442617178,
      "epoch": 0.07194887244910389,
      "grad_norm": 9.375,
      "learning_rate": 4.9605673564115274e-05,
      "loss": 0.6519896507263183,
      "mean_token_accuracy": 0.8345177296549082,
      "num_tokens": 9068131.0,
      "step": 4430
    },
    {
      "entropy": 0.6199771880172193,
      "epoch": 0.07211128525372941,
      "grad_norm": 5.3125,
      "learning_rate": 4.9603376370045684e-05,
      "loss": 0.6159055233001709,
      "mean_token_accuracy": 0.8393093034625053,
      "num_tokens": 9088272.0,
      "step": 4440
    },
    {
      "entropy": 0.6139411455951631,
      "epoch": 0.07227369805835492,
      "grad_norm": 7.875,
      "learning_rate": 4.960107255761256e-05,
      "loss": 0.602547025680542,
      "mean_token_accuracy": 0.8453815996646881,
      "num_tokens": 9108377.0,
      "step": 4450
    },
    {
      "entropy": 0.7398131346330047,
      "epoch": 0.07243611086298044,
      "grad_norm": 9.8125,
      "learning_rate": 4.9598762127435625e-05,
      "loss": 0.7744777202606201,
      "mean_token_accuracy": 0.8173283323645592,
      "num_tokens": 9128649.0,
      "step": 4460
    },
    {
      "entropy": 0.6618171680718661,
      "epoch": 0.07259852366760595,
      "grad_norm": 9.5,
      "learning_rate": 4.95964450801364e-05,
      "loss": 0.6518251419067382,
      "mean_token_accuracy": 0.8385445035994052,
      "num_tokens": 9150860.0,
      "step": 4470
    },
    {
      "entropy": 0.6526267222128809,
      "epoch": 0.07276093647223147,
      "grad_norm": 8.3125,
      "learning_rate": 4.959412141633817e-05,
      "loss": 0.6878862380981445,
      "mean_token_accuracy": 0.8334402985870838,
      "num_tokens": 9172263.0,
      "step": 4480
    },
    {
      "entropy": 0.6544343112036586,
      "epoch": 0.072923349276857,
      "grad_norm": 8.4375,
      "learning_rate": 4.959179113666601e-05,
      "loss": 0.6764401435852051,
      "mean_token_accuracy": 0.8297100096940995,
      "num_tokens": 9192713.0,
      "step": 4490
    },
    {
      "entropy": 0.6737438500858843,
      "epoch": 0.0730857620814825,
      "grad_norm": 6.09375,
      "learning_rate": 4.958945424174677e-05,
      "loss": 0.6143169403076172,
      "mean_token_accuracy": 0.8463603287935257,
      "num_tokens": 9212843.0,
      "step": 4500
    },
    {
      "entropy": 0.608481558971107,
      "epoch": 0.07324817488610802,
      "grad_norm": 7.59375,
      "learning_rate": 4.958711073220907e-05,
      "loss": 0.6361787796020508,
      "mean_token_accuracy": 0.8443557702004909,
      "num_tokens": 9233124.0,
      "step": 4510
    },
    {
      "entropy": 0.5904541192110628,
      "epoch": 0.07341058769073354,
      "grad_norm": 8.125,
      "learning_rate": 4.9584760608683335e-05,
      "loss": 0.6713534832000733,
      "mean_token_accuracy": 0.8394067980349064,
      "num_tokens": 9253776.0,
      "step": 4520
    },
    {
      "entropy": 0.7111636479385197,
      "epoch": 0.07357300049535906,
      "grad_norm": 7.625,
      "learning_rate": 4.958240387180174e-05,
      "loss": 0.7027684688568115,
      "mean_token_accuracy": 0.8293249811977148,
      "num_tokens": 9274460.0,
      "step": 4530
    },
    {
      "entropy": 0.653181302268058,
      "epoch": 0.07373541329998456,
      "grad_norm": 7.6875,
      "learning_rate": 4.958004052219826e-05,
      "loss": 0.6477846622467041,
      "mean_token_accuracy": 0.8470566391944885,
      "num_tokens": 9294755.0,
      "step": 4540
    },
    {
      "entropy": 0.6654880329966545,
      "epoch": 0.07389782610461008,
      "grad_norm": 5.21875,
      "learning_rate": 4.957767056050864e-05,
      "loss": 0.6708653450012207,
      "mean_token_accuracy": 0.8277440384030342,
      "num_tokens": 9315338.0,
      "step": 4550
    },
    {
      "entropy": 0.5965668763965368,
      "epoch": 0.0740602389092356,
      "grad_norm": 6.125,
      "learning_rate": 4.9575293987370396e-05,
      "loss": 0.5827322006225586,
      "mean_token_accuracy": 0.851749524474144,
      "num_tokens": 9335605.0,
      "step": 4560
    },
    {
      "entropy": 0.5778095096349716,
      "epoch": 0.07422265171386112,
      "grad_norm": 6.46875,
      "learning_rate": 4.957291080342285e-05,
      "loss": 0.598599910736084,
      "mean_token_accuracy": 0.8527276519685983,
      "num_tokens": 9356232.0,
      "step": 4570
    },
    {
      "entropy": 0.6098885674029588,
      "epoch": 0.07438506451848664,
      "grad_norm": 6.875,
      "learning_rate": 4.957052100930707e-05,
      "loss": 0.6338545799255371,
      "mean_token_accuracy": 0.8488633498549462,
      "num_tokens": 9375932.0,
      "step": 4580
    },
    {
      "entropy": 0.6734701137989759,
      "epoch": 0.07454747732311215,
      "grad_norm": 7.1875,
      "learning_rate": 4.9568124605665925e-05,
      "loss": 0.6525713920593261,
      "mean_token_accuracy": 0.8386058017611504,
      "num_tokens": 9396996.0,
      "step": 4590
    },
    {
      "entropy": 0.5879238824360072,
      "epoch": 0.07470989012773767,
      "grad_norm": 7.40625,
      "learning_rate": 4.956572159314404e-05,
      "loss": 0.6008066177368164,
      "mean_token_accuracy": 0.8475782603025437,
      "num_tokens": 9416377.0,
      "step": 4600
    },
    {
      "entropy": 0.608251191303134,
      "epoch": 0.07487230293236319,
      "grad_norm": 11.4375,
      "learning_rate": 4.956331197238784e-05,
      "loss": 0.661914587020874,
      "mean_token_accuracy": 0.8446639597415924,
      "num_tokens": 9435692.0,
      "step": 4610
    },
    {
      "entropy": 0.7311778465285897,
      "epoch": 0.07503471573698871,
      "grad_norm": 8.125,
      "learning_rate": 4.956089574404551e-05,
      "loss": 0.7442729473114014,
      "mean_token_accuracy": 0.8161278996616602,
      "num_tokens": 9457501.0,
      "step": 4620
    },
    {
      "entropy": 0.5746764669194817,
      "epoch": 0.07519712854161423,
      "grad_norm": 5.5625,
      "learning_rate": 4.955847290876704e-05,
      "loss": 0.5439295291900634,
      "mean_token_accuracy": 0.8540162913501262,
      "num_tokens": 9477210.0,
      "step": 4630
    },
    {
      "entropy": 0.6886419186368584,
      "epoch": 0.07535954134623973,
      "grad_norm": 6.84375,
      "learning_rate": 4.955604346720416e-05,
      "loss": 0.7328068256378174,
      "mean_token_accuracy": 0.8253999069333077,
      "num_tokens": 9497196.0,
      "step": 4640
    },
    {
      "entropy": 0.5974585706368089,
      "epoch": 0.07552195415086525,
      "grad_norm": 7.3125,
      "learning_rate": 4.9553607420010406e-05,
      "loss": 0.6019351482391357,
      "mean_token_accuracy": 0.8427147604525089,
      "num_tokens": 9518102.0,
      "step": 4650
    },
    {
      "entropy": 0.6890899701043963,
      "epoch": 0.07568436695549077,
      "grad_norm": 8.5625,
      "learning_rate": 4.9551164767841075e-05,
      "loss": 0.6990561485290527,
      "mean_token_accuracy": 0.8328306473791599,
      "num_tokens": 9537523.0,
      "step": 4660
    },
    {
      "entropy": 0.6631820605136454,
      "epoch": 0.07584677976011629,
      "grad_norm": 8.3125,
      "learning_rate": 4.954871551135324e-05,
      "loss": 0.625786542892456,
      "mean_token_accuracy": 0.841561759263277,
      "num_tokens": 9557823.0,
      "step": 4670
    },
    {
      "entropy": 0.6061957945115864,
      "epoch": 0.0760091925647418,
      "grad_norm": 7.5,
      "learning_rate": 4.9546259651205765e-05,
      "loss": 0.617872953414917,
      "mean_token_accuracy": 0.8374070636928082,
      "num_tokens": 9577709.0,
      "step": 4680
    },
    {
      "entropy": 0.6638525852933526,
      "epoch": 0.07617160536936732,
      "grad_norm": 7.5625,
      "learning_rate": 4.954379718805928e-05,
      "loss": 0.6399277210235595,
      "mean_token_accuracy": 0.8351655241101981,
      "num_tokens": 9598351.0,
      "step": 4690
    },
    {
      "entropy": 0.6586564770899713,
      "epoch": 0.07633401817399284,
      "grad_norm": 7.46875,
      "learning_rate": 4.9541328122576194e-05,
      "loss": 0.6815990924835205,
      "mean_token_accuracy": 0.8324289739131927,
      "num_tokens": 9619117.0,
      "step": 4700
    },
    {
      "entropy": 0.625049875676632,
      "epoch": 0.07649643097861836,
      "grad_norm": 6.75,
      "learning_rate": 4.953885245542069e-05,
      "loss": 0.6064480304718017,
      "mean_token_accuracy": 0.8460681267082691,
      "num_tokens": 9639292.0,
      "step": 4710
    },
    {
      "entropy": 0.6891991007141769,
      "epoch": 0.07665884378324388,
      "grad_norm": 7.96875,
      "learning_rate": 4.953637018725873e-05,
      "loss": 0.7469194889068603,
      "mean_token_accuracy": 0.8167282767593861,
      "num_tokens": 9659694.0,
      "step": 4720
    },
    {
      "entropy": 0.644246906042099,
      "epoch": 0.07682125658786938,
      "grad_norm": 7.15625,
      "learning_rate": 4.953388131875804e-05,
      "loss": 0.6113909244537353,
      "mean_token_accuracy": 0.8501339592039585,
      "num_tokens": 9680290.0,
      "step": 4730
    },
    {
      "entropy": 0.6095455882139504,
      "epoch": 0.0769836693924949,
      "grad_norm": 9.5625,
      "learning_rate": 4.953138585058814e-05,
      "loss": 0.6163639068603516,
      "mean_token_accuracy": 0.8464456975460053,
      "num_tokens": 9699993.0,
      "step": 4740
    },
    {
      "entropy": 0.5748913143761456,
      "epoch": 0.07714608219712042,
      "grad_norm": 8.5,
      "learning_rate": 4.952888378342032e-05,
      "loss": 0.5689167499542236,
      "mean_token_accuracy": 0.8654819943010807,
      "num_tokens": 9719782.0,
      "step": 4750
    },
    {
      "entropy": 0.6310321614146233,
      "epoch": 0.07730849500174594,
      "grad_norm": 12.5,
      "learning_rate": 4.952637511792764e-05,
      "loss": 0.6626490116119385,
      "mean_token_accuracy": 0.8381782792508602,
      "num_tokens": 9741155.0,
      "step": 4760
    },
    {
      "entropy": 0.6368576645851135,
      "epoch": 0.07747090780637146,
      "grad_norm": 9.8125,
      "learning_rate": 4.952385985478493e-05,
      "loss": 0.6029601097106934,
      "mean_token_accuracy": 0.8416288785636425,
      "num_tokens": 9761043.0,
      "step": 4770
    },
    {
      "entropy": 0.6221295884344726,
      "epoch": 0.07763332061099697,
      "grad_norm": 8.6875,
      "learning_rate": 4.95213379946688e-05,
      "loss": 0.6285975456237793,
      "mean_token_accuracy": 0.8402429152280092,
      "num_tokens": 9781303.0,
      "step": 4780
    },
    {
      "entropy": 0.5921527820639312,
      "epoch": 0.07779573341562249,
      "grad_norm": 10.5625,
      "learning_rate": 4.9518809538257646e-05,
      "loss": 0.659816837310791,
      "mean_token_accuracy": 0.8424258559942246,
      "num_tokens": 9801989.0,
      "step": 4790
    },
    {
      "entropy": 0.589257238060236,
      "epoch": 0.077958146220248,
      "grad_norm": 7.46875,
      "learning_rate": 4.951627448623162e-05,
      "loss": 0.6013991832733154,
      "mean_token_accuracy": 0.8496914029121398,
      "num_tokens": 9822856.0,
      "step": 4800
    },
    {
      "entropy": 0.6988964445888997,
      "epoch": 0.07812055902487353,
      "grad_norm": 6.96875,
      "learning_rate": 4.951373283927266e-05,
      "loss": 0.7083846092224121,
      "mean_token_accuracy": 0.8164467625319958,
      "num_tokens": 9843361.0,
      "step": 4810
    },
    {
      "entropy": 0.6911227965727449,
      "epoch": 0.07828297182949903,
      "grad_norm": 8.6875,
      "learning_rate": 4.951118459806449e-05,
      "loss": 0.7014126777648926,
      "mean_token_accuracy": 0.8366714313626289,
      "num_tokens": 9863157.0,
      "step": 4820
    },
    {
      "entropy": 0.6688828192185611,
      "epoch": 0.07844538463412455,
      "grad_norm": 6.03125,
      "learning_rate": 4.9508629763292554e-05,
      "loss": 0.6663030624389649,
      "mean_token_accuracy": 0.825884073972702,
      "num_tokens": 9883876.0,
      "step": 4830
    },
    {
      "entropy": 0.7494215987622738,
      "epoch": 0.07860779743875007,
      "grad_norm": 5.84375,
      "learning_rate": 4.9506068335644154e-05,
      "loss": 0.7448901653289794,
      "mean_token_accuracy": 0.8165868297219276,
      "num_tokens": 9903546.0,
      "step": 4840
    },
    {
      "entropy": 0.6206991045735777,
      "epoch": 0.07877021024337559,
      "grad_norm": 7.71875,
      "learning_rate": 4.9503500315808284e-05,
      "loss": 0.632866621017456,
      "mean_token_accuracy": 0.8432051688432693,
      "num_tokens": 9923466.0,
      "step": 4850
    },
    {
      "entropy": 0.596147621795535,
      "epoch": 0.07893262304800111,
      "grad_norm": 7.78125,
      "learning_rate": 4.9500925704475765e-05,
      "loss": 0.5856470108032227,
      "mean_token_accuracy": 0.848544605076313,
      "num_tokens": 9946533.0,
      "step": 4860
    },
    {
      "entropy": 0.6550322595983744,
      "epoch": 0.07909503585262662,
      "grad_norm": 5.65625,
      "learning_rate": 4.9498344502339165e-05,
      "loss": 0.6990450382232666,
      "mean_token_accuracy": 0.8321107421070337,
      "num_tokens": 9966961.0,
      "step": 4870
    },
    {
      "entropy": 0.6440950100310147,
      "epoch": 0.07925744865725214,
      "grad_norm": 7.53125,
      "learning_rate": 4.949575671009285e-05,
      "loss": 0.679059362411499,
      "mean_token_accuracy": 0.8301635667681694,
      "num_tokens": 9986555.0,
      "step": 4880
    },
    {
      "entropy": 0.6795095540583134,
      "epoch": 0.07941986146187766,
      "grad_norm": 7.125,
      "learning_rate": 4.949316232843292e-05,
      "loss": 0.6497490406036377,
      "mean_token_accuracy": 0.8340746454894543,
      "num_tokens": 10007179.0,
      "step": 4890
    },
    {
      "entropy": 0.614184751920402,
      "epoch": 0.07958227426650318,
      "grad_norm": 8.125,
      "learning_rate": 4.9490561358057276e-05,
      "loss": 0.6331201553344726,
      "mean_token_accuracy": 0.8490272544324398,
      "num_tokens": 10027530.0,
      "step": 4900
    },
    {
      "entropy": 0.6610461741220206,
      "epoch": 0.07974468707112868,
      "grad_norm": 9.9375,
      "learning_rate": 4.948795379966559e-05,
      "loss": 0.6990130424499512,
      "mean_token_accuracy": 0.8327658340334892,
      "num_tokens": 10046684.0,
      "step": 4910
    },
    {
      "entropy": 0.5885923689231276,
      "epoch": 0.0799070998757542,
      "grad_norm": 7.5625,
      "learning_rate": 4.9485339653959294e-05,
      "loss": 0.557637071609497,
      "mean_token_accuracy": 0.8561469249427318,
      "num_tokens": 10067042.0,
      "step": 4920
    },
    {
      "entropy": 0.6668364732526243,
      "epoch": 0.08006951268037972,
      "grad_norm": 8.8125,
      "learning_rate": 4.948271892164161e-05,
      "loss": 0.7047749519348144,
      "mean_token_accuracy": 0.8359684132039547,
      "num_tokens": 10088330.0,
      "step": 4930
    },
    {
      "entropy": 0.658598153758794,
      "epoch": 0.08023192548500524,
      "grad_norm": 7.34375,
      "learning_rate": 4.948009160341752e-05,
      "loss": 0.6554466724395752,
      "mean_token_accuracy": 0.8344687484204769,
      "num_tokens": 10107684.0,
      "step": 4940
    },
    {
      "entropy": 0.6474485736340284,
      "epoch": 0.08039433828963076,
      "grad_norm": 8.75,
      "learning_rate": 4.9477457699993765e-05,
      "loss": 0.6534360408782959,
      "mean_token_accuracy": 0.8392691500484943,
      "num_tokens": 10128773.0,
      "step": 4950
    },
    {
      "entropy": 0.6443928050808608,
      "epoch": 0.08055675109425627,
      "grad_norm": 7.34375,
      "learning_rate": 4.9474817212078885e-05,
      "loss": 0.6535509109497071,
      "mean_token_accuracy": 0.8415901988744736,
      "num_tokens": 10149077.0,
      "step": 4960
    },
    {
      "entropy": 0.6375644318759441,
      "epoch": 0.08071916389888179,
      "grad_norm": 9.375,
      "learning_rate": 4.947217014038316e-05,
      "loss": 0.6529776573181152,
      "mean_token_accuracy": 0.8371326476335526,
      "num_tokens": 10168255.0,
      "step": 4970
    },
    {
      "entropy": 0.5736218354664743,
      "epoch": 0.0808815767035073,
      "grad_norm": 6.1875,
      "learning_rate": 4.9469516485618675e-05,
      "loss": 0.5750738143920898,
      "mean_token_accuracy": 0.8604113213717938,
      "num_tokens": 10188814.0,
      "step": 4980
    },
    {
      "entropy": 0.6259093351662159,
      "epoch": 0.08104398950813282,
      "grad_norm": 8.875,
      "learning_rate": 4.946685624849927e-05,
      "loss": 0.6583810806274414,
      "mean_token_accuracy": 0.8400992691516876,
      "num_tokens": 10208281.0,
      "step": 4990
    },
    {
      "entropy": 0.5598170708399266,
      "epoch": 0.08120640231275834,
      "grad_norm": 5.84375,
      "learning_rate": 4.946418942974054e-05,
      "loss": 0.5375143527984619,
      "mean_token_accuracy": 0.8721061527729035,
      "num_tokens": 10227802.0,
      "step": 5000
    },
    {
      "entropy": 0.6057121331803501,
      "epoch": 0.08136881511738385,
      "grad_norm": 7.96875,
      "learning_rate": 4.9461516030059876e-05,
      "loss": 0.6155375480651856,
      "mean_token_accuracy": 0.8383884519338608,
      "num_tokens": 10249326.0,
      "step": 5010
    },
    {
      "entropy": 0.5951614056713879,
      "epoch": 0.08153122792200937,
      "grad_norm": 7.875,
      "learning_rate": 4.945883605017643e-05,
      "loss": 0.5660048961639405,
      "mean_token_accuracy": 0.8564432084560394,
      "num_tokens": 10269990.0,
      "step": 5020
    },
    {
      "entropy": 0.650169312208891,
      "epoch": 0.08169364072663489,
      "grad_norm": 5.0,
      "learning_rate": 4.945614949081112e-05,
      "loss": 0.6910473823547363,
      "mean_token_accuracy": 0.830925741046667,
      "num_tokens": 10291079.0,
      "step": 5030
    },
    {
      "entropy": 0.5980288774706424,
      "epoch": 0.08185605353126041,
      "grad_norm": 9.75,
      "learning_rate": 4.9453456352686626e-05,
      "loss": 0.6289026737213135,
      "mean_token_accuracy": 0.8433196566998958,
      "num_tokens": 10311387.0,
      "step": 5040
    },
    {
      "entropy": 0.6629435306414961,
      "epoch": 0.08201846633588591,
      "grad_norm": 5.625,
      "learning_rate": 4.945075663652742e-05,
      "loss": 0.6363174438476562,
      "mean_token_accuracy": 0.8429024174809456,
      "num_tokens": 10332262.0,
      "step": 5050
    },
    {
      "entropy": 0.6410766674671322,
      "epoch": 0.08218087914051143,
      "grad_norm": 11.0625,
      "learning_rate": 4.9448050343059736e-05,
      "loss": 0.6726670742034913,
      "mean_token_accuracy": 0.8403808742761611,
      "num_tokens": 10352804.0,
      "step": 5060
    },
    {
      "entropy": 0.6930725152604282,
      "epoch": 0.08234329194513695,
      "grad_norm": 8.875,
      "learning_rate": 4.9445337473011563e-05,
      "loss": 0.7510185718536377,
      "mean_token_accuracy": 0.8295759953558445,
      "num_tokens": 10373034.0,
      "step": 5070
    },
    {
      "entropy": 0.5894734486937523,
      "epoch": 0.08250570474976247,
      "grad_norm": 6.53125,
      "learning_rate": 4.9442618027112667e-05,
      "loss": 0.5793861865997314,
      "mean_token_accuracy": 0.8555416770279407,
      "num_tokens": 10394727.0,
      "step": 5080
    },
    {
      "entropy": 0.7120319345965982,
      "epoch": 0.082668117554388,
      "grad_norm": 7.03125,
      "learning_rate": 4.94398920060946e-05,
      "loss": 0.7151927471160888,
      "mean_token_accuracy": 0.8383033473044634,
      "num_tokens": 10415359.0,
      "step": 5090
    },
    {
      "entropy": 0.6719475731253624,
      "epoch": 0.0828305303590135,
      "grad_norm": 7.9375,
      "learning_rate": 4.943715941069065e-05,
      "loss": 0.6789883136749267,
      "mean_token_accuracy": 0.8341887325048447,
      "num_tokens": 10434865.0,
      "step": 5100
    },
    {
      "entropy": 0.5589958894997835,
      "epoch": 0.08299294316363902,
      "grad_norm": 8.75,
      "learning_rate": 4.9434420241635905e-05,
      "loss": 0.5733313083648681,
      "mean_token_accuracy": 0.8563144087791443,
      "num_tokens": 10455073.0,
      "step": 5110
    },
    {
      "entropy": 0.6855927634984255,
      "epoch": 0.08315535596826454,
      "grad_norm": 7.21875,
      "learning_rate": 4.9431674499667194e-05,
      "loss": 0.6664769172668457,
      "mean_token_accuracy": 0.8326185751706362,
      "num_tokens": 10477025.0,
      "step": 5120
    },
    {
      "entropy": 0.5816048513166606,
      "epoch": 0.08331776877289006,
      "grad_norm": 7.5,
      "learning_rate": 4.9428922185523143e-05,
      "loss": 0.5941498279571533,
      "mean_token_accuracy": 0.8555799119174481,
      "num_tokens": 10496949.0,
      "step": 5130
    },
    {
      "entropy": 0.6426875446923077,
      "epoch": 0.08348018157751558,
      "grad_norm": 6.5625,
      "learning_rate": 4.9426163299944114e-05,
      "loss": 0.6639914989471436,
      "mean_token_accuracy": 0.8337427839636803,
      "num_tokens": 10516984.0,
      "step": 5140
    },
    {
      "entropy": 0.5499494080431759,
      "epoch": 0.08364259438214108,
      "grad_norm": 7.1875,
      "learning_rate": 4.9423397843672277e-05,
      "loss": 0.553646469116211,
      "mean_token_accuracy": 0.8541104160249233,
      "num_tokens": 10536896.0,
      "step": 5150
    },
    {
      "entropy": 0.6421358630061149,
      "epoch": 0.0838050071867666,
      "grad_norm": 4.78125,
      "learning_rate": 4.942062581745153e-05,
      "loss": 0.6216344833374023,
      "mean_token_accuracy": 0.8433695696294308,
      "num_tokens": 10559031.0,
      "step": 5160
    },
    {
      "entropy": 0.6411819517612457,
      "epoch": 0.08396741999139212,
      "grad_norm": 11.1875,
      "learning_rate": 4.9417847222027544e-05,
      "loss": 0.6660083770751953,
      "mean_token_accuracy": 0.8391093079000711,
      "num_tokens": 10579138.0,
      "step": 5170
    },
    {
      "entropy": 0.5786192860454321,
      "epoch": 0.08412983279601764,
      "grad_norm": 8.0625,
      "learning_rate": 4.9415062058147785e-05,
      "loss": 0.5851787090301513,
      "mean_token_accuracy": 0.8527331590652466,
      "num_tokens": 10599403.0,
      "step": 5180
    },
    {
      "entropy": 0.5920738274231553,
      "epoch": 0.08429224560064315,
      "grad_norm": 6.625,
      "learning_rate": 4.941227032656147e-05,
      "loss": 0.6059837341308594,
      "mean_token_accuracy": 0.8455815583467483,
      "num_tokens": 10618721.0,
      "step": 5190
    },
    {
      "entropy": 0.6267990993335844,
      "epoch": 0.08445465840526867,
      "grad_norm": 6.46875,
      "learning_rate": 4.9409472028019564e-05,
      "loss": 0.6119988441467286,
      "mean_token_accuracy": 0.847959142178297,
      "num_tokens": 10639705.0,
      "step": 5200
    },
    {
      "entropy": 0.5783751524984837,
      "epoch": 0.08461707120989419,
      "grad_norm": 5.9375,
      "learning_rate": 4.9406667163274835e-05,
      "loss": 0.5898916244506835,
      "mean_token_accuracy": 0.8498507089912891,
      "num_tokens": 10660051.0,
      "step": 5210
    },
    {
      "entropy": 0.6499343666248023,
      "epoch": 0.08477948401451971,
      "grad_norm": 6.46875,
      "learning_rate": 4.940385573308178e-05,
      "loss": 0.6796340942382812,
      "mean_token_accuracy": 0.8294782746583224,
      "num_tokens": 10680245.0,
      "step": 5220
    },
    {
      "entropy": 0.6164681887254119,
      "epoch": 0.08494189681914523,
      "grad_norm": 4.875,
      "learning_rate": 4.94010377381967e-05,
      "loss": 0.6364531993865967,
      "mean_token_accuracy": 0.8416301932185888,
      "num_tokens": 10699589.0,
      "step": 5230
    },
    {
      "entropy": 0.5570803395938129,
      "epoch": 0.08510430962377073,
      "grad_norm": 7.875,
      "learning_rate": 4.939821317937762e-05,
      "loss": 0.5307340145111084,
      "mean_token_accuracy": 0.8631787493824958,
      "num_tokens": 10719661.0,
      "step": 5240
    },
    {
      "entropy": 0.6254891388118267,
      "epoch": 0.08526672242839625,
      "grad_norm": 8.8125,
      "learning_rate": 4.9395382057384363e-05,
      "loss": 0.659565544128418,
      "mean_token_accuracy": 0.8414617232978344,
      "num_tokens": 10739598.0,
      "step": 5250
    },
    {
      "entropy": 0.6107869914732873,
      "epoch": 0.08542913523302177,
      "grad_norm": 7.21875,
      "learning_rate": 4.939254437297851e-05,
      "loss": 0.6181818962097168,
      "mean_token_accuracy": 0.8477704729884863,
      "num_tokens": 10760220.0,
      "step": 5260
    },
    {
      "entropy": 0.6104566549183801,
      "epoch": 0.08559154803764729,
      "grad_norm": 9.6875,
      "learning_rate": 4.938970012692341e-05,
      "loss": 0.6153532981872558,
      "mean_token_accuracy": 0.8421719916164875,
      "num_tokens": 10780130.0,
      "step": 5270
    },
    {
      "entropy": 0.6349617819301784,
      "epoch": 0.08575396084227281,
      "grad_norm": 6.53125,
      "learning_rate": 4.938684931998415e-05,
      "loss": 0.6742961406707764,
      "mean_token_accuracy": 0.839853010326624,
      "num_tokens": 10800374.0,
      "step": 5280
    },
    {
      "entropy": 0.6702006112784147,
      "epoch": 0.08591637364689832,
      "grad_norm": 7.53125,
      "learning_rate": 4.938399195292762e-05,
      "loss": 0.6885315418243408,
      "mean_token_accuracy": 0.8360910423099994,
      "num_tokens": 10820651.0,
      "step": 5290
    },
    {
      "entropy": 0.6499825142323971,
      "epoch": 0.08607878645152384,
      "grad_norm": 7.6875,
      "learning_rate": 4.938112802652246e-05,
      "loss": 0.6884824752807617,
      "mean_token_accuracy": 0.833566851168871,
      "num_tokens": 10840416.0,
      "step": 5300
    },
    {
      "entropy": 0.7190689125098289,
      "epoch": 0.08624119925614936,
      "grad_norm": 9.3125,
      "learning_rate": 4.9378257541539067e-05,
      "loss": 0.6715456962585449,
      "mean_token_accuracy": 0.8415072001516819,
      "num_tokens": 10860558.0,
      "step": 5310
    },
    {
      "entropy": 0.6251872099936009,
      "epoch": 0.08640361206077488,
      "grad_norm": 7.40625,
      "learning_rate": 4.937538049874961e-05,
      "loss": 0.6179144859313965,
      "mean_token_accuracy": 0.8462849885225296,
      "num_tokens": 10880340.0,
      "step": 5320
    },
    {
      "entropy": 0.6427540705539286,
      "epoch": 0.08656602486540038,
      "grad_norm": 6.84375,
      "learning_rate": 4.937249689892801e-05,
      "loss": 0.6948603630065918,
      "mean_token_accuracy": 0.8310207940638066,
      "num_tokens": 10899272.0,
      "step": 5330
    },
    {
      "entropy": 0.6112544117495418,
      "epoch": 0.0867284376700259,
      "grad_norm": 7.4375,
      "learning_rate": 4.936960674284998e-05,
      "loss": 0.604194164276123,
      "mean_token_accuracy": 0.847838393971324,
      "num_tokens": 10919262.0,
      "step": 5340
    },
    {
      "entropy": 0.6243371238466352,
      "epoch": 0.08689085047465142,
      "grad_norm": 6.65625,
      "learning_rate": 4.9366710031292965e-05,
      "loss": 0.6522641181945801,
      "mean_token_accuracy": 0.8375919163227081,
      "num_tokens": 10939919.0,
      "step": 5350
    },
    {
      "entropy": 0.5814169902354479,
      "epoch": 0.08705326327927694,
      "grad_norm": 8.125,
      "learning_rate": 4.9363806765036194e-05,
      "loss": 0.5493687152862549,
      "mean_token_accuracy": 0.8554010517895222,
      "num_tokens": 10959080.0,
      "step": 5360
    },
    {
      "entropy": 0.642770760692656,
      "epoch": 0.08721567608390246,
      "grad_norm": 6.0625,
      "learning_rate": 4.9360896944860644e-05,
      "loss": 0.6677518844604492,
      "mean_token_accuracy": 0.8433307923376561,
      "num_tokens": 10978344.0,
      "step": 5370
    },
    {
      "entropy": 0.6084460916928947,
      "epoch": 0.08737808888852797,
      "grad_norm": 9.5,
      "learning_rate": 4.935798057154907e-05,
      "loss": 0.6008291721343995,
      "mean_token_accuracy": 0.8442369610071182,
      "num_tokens": 10998474.0,
      "step": 5380
    },
    {
      "entropy": 0.7281197951640934,
      "epoch": 0.08754050169315349,
      "grad_norm": 7.84375,
      "learning_rate": 4.935505764588598e-05,
      "loss": 0.7159610271453858,
      "mean_token_accuracy": 0.8228184826672077,
      "num_tokens": 11019736.0,
      "step": 5390
    },
    {
      "entropy": 0.6604039056226612,
      "epoch": 0.087702914497779,
      "grad_norm": 8.1875,
      "learning_rate": 4.935212816865765e-05,
      "loss": 0.6889187812805175,
      "mean_token_accuracy": 0.8336355306208134,
      "num_tokens": 11040722.0,
      "step": 5400
    },
    {
      "entropy": 0.5105343041010201,
      "epoch": 0.08786532730240453,
      "grad_norm": 7.78125,
      "learning_rate": 4.9349192140652115e-05,
      "loss": 0.48794264793395997,
      "mean_token_accuracy": 0.8761614501476288,
      "num_tokens": 11061966.0,
      "step": 5410
    },
    {
      "entropy": 0.6003414323553443,
      "epoch": 0.08802774010703004,
      "grad_norm": 7.78125,
      "learning_rate": 4.934624956265917e-05,
      "loss": 0.6512250423431396,
      "mean_token_accuracy": 0.8346175529062748,
      "num_tokens": 11082417.0,
      "step": 5420
    },
    {
      "entropy": 0.6423029700294137,
      "epoch": 0.08819015291165555,
      "grad_norm": 11.75,
      "learning_rate": 4.9343300435470384e-05,
      "loss": 0.7145370006561279,
      "mean_token_accuracy": 0.8275042608380317,
      "num_tokens": 11104089.0,
      "step": 5430
    },
    {
      "entropy": 0.6451686321292073,
      "epoch": 0.08835256571628107,
      "grad_norm": 6.3125,
      "learning_rate": 4.9340344759879064e-05,
      "loss": 0.6304905414581299,
      "mean_token_accuracy": 0.8430520944297314,
      "num_tokens": 11124864.0,
      "step": 5440
    },
    {
      "entropy": 0.6334490654990077,
      "epoch": 0.08851497852090659,
      "grad_norm": 6.46875,
      "learning_rate": 4.933738253668031e-05,
      "loss": 0.6170599937438965,
      "mean_token_accuracy": 0.837143412232399,
      "num_tokens": 11144364.0,
      "step": 5450
    },
    {
      "entropy": 0.6064558143727481,
      "epoch": 0.08867739132553211,
      "grad_norm": 7.46875,
      "learning_rate": 4.9334413766670956e-05,
      "loss": 0.6072060585021972,
      "mean_token_accuracy": 0.8475678980350494,
      "num_tokens": 11164317.0,
      "step": 5460
    },
    {
      "entropy": 0.6237148152664304,
      "epoch": 0.08883980413015762,
      "grad_norm": 9.125,
      "learning_rate": 4.9331438450649606e-05,
      "loss": 0.6093404769897461,
      "mean_token_accuracy": 0.8471099127084016,
      "num_tokens": 11184928.0,
      "step": 5470
    },
    {
      "entropy": 0.5757282917387784,
      "epoch": 0.08900221693478313,
      "grad_norm": 10.125,
      "learning_rate": 4.932845658941664e-05,
      "loss": 0.6404849529266358,
      "mean_token_accuracy": 0.8386787280440331,
      "num_tokens": 11206130.0,
      "step": 5480
    },
    {
      "entropy": 0.6905780477449298,
      "epoch": 0.08916462973940865,
      "grad_norm": 7.46875,
      "learning_rate": 4.932546818377417e-05,
      "loss": 0.702696418762207,
      "mean_token_accuracy": 0.8250477857887745,
      "num_tokens": 11228674.0,
      "step": 5490
    },
    {
      "entropy": 0.6459718506783247,
      "epoch": 0.08932704254403417,
      "grad_norm": 8.0625,
      "learning_rate": 4.9322473234526096e-05,
      "loss": 0.6440560340881347,
      "mean_token_accuracy": 0.8414949037134647,
      "num_tokens": 11248476.0,
      "step": 5500
    },
    {
      "entropy": 0.5802592306397856,
      "epoch": 0.0894894553486597,
      "grad_norm": 7.65625,
      "learning_rate": 4.9319471742478055e-05,
      "loss": 0.6213495254516601,
      "mean_token_accuracy": 0.848654656112194,
      "num_tokens": 11268080.0,
      "step": 5510
    },
    {
      "entropy": 0.6442588407546281,
      "epoch": 0.0896518681532852,
      "grad_norm": 7.09375,
      "learning_rate": 4.9316463708437466e-05,
      "loss": 0.6078289985656739,
      "mean_token_accuracy": 0.8529248848557472,
      "num_tokens": 11288752.0,
      "step": 5520
    },
    {
      "entropy": 0.5800073832273483,
      "epoch": 0.08981428095791072,
      "grad_norm": 6.09375,
      "learning_rate": 4.9313449133213483e-05,
      "loss": 0.5902406692504882,
      "mean_token_accuracy": 0.8432722076773643,
      "num_tokens": 11309193.0,
      "step": 5530
    },
    {
      "entropy": 0.6451068416237831,
      "epoch": 0.08997669376253624,
      "grad_norm": 9.9375,
      "learning_rate": 4.931042801761705e-05,
      "loss": 0.6964611530303955,
      "mean_token_accuracy": 0.8266104102134705,
      "num_tokens": 11328028.0,
      "step": 5540
    },
    {
      "entropy": 0.7333416105248034,
      "epoch": 0.09013910656716176,
      "grad_norm": 7.5625,
      "learning_rate": 4.930740036246084e-05,
      "loss": 0.75954270362854,
      "mean_token_accuracy": 0.8153920006006956,
      "num_tokens": 11348913.0,
      "step": 5550
    },
    {
      "entropy": 0.5992802513763309,
      "epoch": 0.09030151937178728,
      "grad_norm": 6.9375,
      "learning_rate": 4.9304366168559315e-05,
      "loss": 0.5746379852294922,
      "mean_token_accuracy": 0.8538239255547524,
      "num_tokens": 11368667.0,
      "step": 5560
    },
    {
      "entropy": 0.617322726920247,
      "epoch": 0.09046393217641278,
      "grad_norm": 5.78125,
      "learning_rate": 4.930132543672866e-05,
      "loss": 0.607831335067749,
      "mean_token_accuracy": 0.8455162569880486,
      "num_tokens": 11388742.0,
      "step": 5570
    },
    {
      "entropy": 0.6373940145596861,
      "epoch": 0.0906263449810383,
      "grad_norm": 6.1875,
      "learning_rate": 4.9298278167786854e-05,
      "loss": 0.6336771011352539,
      "mean_token_accuracy": 0.8423676803708077,
      "num_tokens": 11409260.0,
      "step": 5580
    },
    {
      "entropy": 0.6668388342484832,
      "epoch": 0.09078875778566382,
      "grad_norm": 7.53125,
      "learning_rate": 4.9295224362553616e-05,
      "loss": 0.7081441879272461,
      "mean_token_accuracy": 0.8305798426270485,
      "num_tokens": 11431389.0,
      "step": 5590
    },
    {
      "entropy": 0.6991683353669942,
      "epoch": 0.09095117059028934,
      "grad_norm": 7.90625,
      "learning_rate": 4.929216402185042e-05,
      "loss": 0.715422534942627,
      "mean_token_accuracy": 0.8236151970922947,
      "num_tokens": 11451943.0,
      "step": 5600
    },
    {
      "entropy": 0.6517792828381062,
      "epoch": 0.09111358339491485,
      "grad_norm": 6.96875,
      "learning_rate": 4.928909714650051e-05,
      "loss": 0.6366158962249756,
      "mean_token_accuracy": 0.8373116962611675,
      "num_tokens": 11473005.0,
      "step": 5610
    },
    {
      "entropy": 0.6652334915474057,
      "epoch": 0.09127599619954037,
      "grad_norm": 8.6875,
      "learning_rate": 4.928602373732888e-05,
      "loss": 0.6986649990081787,
      "mean_token_accuracy": 0.8397133007645607,
      "num_tokens": 11494121.0,
      "step": 5620
    },
    {
      "entropy": 0.5822399449534714,
      "epoch": 0.09143840900416589,
      "grad_norm": 8.25,
      "learning_rate": 4.928294379516228e-05,
      "loss": 0.5544229030609131,
      "mean_token_accuracy": 0.8519444763660431,
      "num_tokens": 11514745.0,
      "step": 5630
    },
    {
      "entropy": 0.5854098604992032,
      "epoch": 0.09160082180879141,
      "grad_norm": 7.1875,
      "learning_rate": 4.927985732082924e-05,
      "loss": 0.584111499786377,
      "mean_token_accuracy": 0.8497672028839588,
      "num_tokens": 11535310.0,
      "step": 5640
    },
    {
      "entropy": 0.5732845838181675,
      "epoch": 0.09176323461341693,
      "grad_norm": 6.46875,
      "learning_rate": 4.9276764315160005e-05,
      "loss": 0.5545302391052246,
      "mean_token_accuracy": 0.8615542069077492,
      "num_tokens": 11555854.0,
      "step": 5650
    },
    {
      "entropy": 0.6476027883589268,
      "epoch": 0.09192564741804243,
      "grad_norm": 4.96875,
      "learning_rate": 4.9273664778986613e-05,
      "loss": 0.7048705101013184,
      "mean_token_accuracy": 0.8281946685165167,
      "num_tokens": 11575133.0,
      "step": 5660
    },
    {
      "entropy": 0.7215040531009436,
      "epoch": 0.09208806022266795,
      "grad_norm": 7.96875,
      "learning_rate": 4.927055871314284e-05,
      "loss": 0.7234652996063232,
      "mean_token_accuracy": 0.8235920663923025,
      "num_tokens": 11596195.0,
      "step": 5670
    },
    {
      "entropy": 0.5627514095511288,
      "epoch": 0.09225047302729347,
      "grad_norm": 6.53125,
      "learning_rate": 4.9267446118464233e-05,
      "loss": 0.5749897003173828,
      "mean_token_accuracy": 0.857719873636961,
      "num_tokens": 11617287.0,
      "step": 5680
    },
    {
      "entropy": 0.6435061246156693,
      "epoch": 0.09241288583191899,
      "grad_norm": 9.625,
      "learning_rate": 4.926432699578809e-05,
      "loss": 0.6815619945526123,
      "mean_token_accuracy": 0.8332644391804933,
      "num_tokens": 11637975.0,
      "step": 5690
    },
    {
      "entropy": 0.6932725168764591,
      "epoch": 0.09257529863654451,
      "grad_norm": 8.0625,
      "learning_rate": 4.926120134595345e-05,
      "loss": 0.6874166488647461,
      "mean_token_accuracy": 0.8254999034106731,
      "num_tokens": 11658955.0,
      "step": 5700
    },
    {
      "entropy": 0.6644466552883387,
      "epoch": 0.09273771144117002,
      "grad_norm": 8.875,
      "learning_rate": 4.9258069169801114e-05,
      "loss": 0.6455849170684814,
      "mean_token_accuracy": 0.8269536457955837,
      "num_tokens": 11679577.0,
      "step": 5710
    },
    {
      "entropy": 0.6981634385883808,
      "epoch": 0.09290012424579554,
      "grad_norm": 7.90625,
      "learning_rate": 4.925493046817367e-05,
      "loss": 0.7206590652465821,
      "mean_token_accuracy": 0.8137824270874262,
      "num_tokens": 11701631.0,
      "step": 5720
    },
    {
      "entropy": 0.6097036791965366,
      "epoch": 0.09306253705042106,
      "grad_norm": 9.3125,
      "learning_rate": 4.92517852419154e-05,
      "loss": 0.6002647876739502,
      "mean_token_accuracy": 0.842541278898716,
      "num_tokens": 11722057.0,
      "step": 5730
    },
    {
      "entropy": 0.6781312499195338,
      "epoch": 0.09322494985504658,
      "grad_norm": 7.90625,
      "learning_rate": 4.924863349187242e-05,
      "loss": 0.6763605594635009,
      "mean_token_accuracy": 0.8276946157217026,
      "num_tokens": 11742861.0,
      "step": 5740
    },
    {
      "entropy": 0.6410096267703921,
      "epoch": 0.09338736265967208,
      "grad_norm": 9.875,
      "learning_rate": 4.9245475218892524e-05,
      "loss": 0.7083607673645019,
      "mean_token_accuracy": 0.8270403660833836,
      "num_tokens": 11763285.0,
      "step": 5750
    },
    {
      "entropy": 0.6320790896192193,
      "epoch": 0.0935497754642976,
      "grad_norm": 12.4375,
      "learning_rate": 4.92423104238253e-05,
      "loss": 0.5950321674346923,
      "mean_token_accuracy": 0.8523061156272889,
      "num_tokens": 11784748.0,
      "step": 5760
    },
    {
      "entropy": 0.6549301604740322,
      "epoch": 0.09371218826892312,
      "grad_norm": 8.75,
      "learning_rate": 4.9239139107522106e-05,
      "loss": 0.6309489250183106,
      "mean_token_accuracy": 0.8387120313942432,
      "num_tokens": 11805146.0,
      "step": 5770
    },
    {
      "entropy": 0.5753857920877635,
      "epoch": 0.09387460107354864,
      "grad_norm": 4.90625,
      "learning_rate": 4.923596127083601e-05,
      "loss": 0.5750814914703369,
      "mean_token_accuracy": 0.8526087913662195,
      "num_tokens": 11825359.0,
      "step": 5780
    },
    {
      "entropy": 0.5677464101463556,
      "epoch": 0.09403701387817416,
      "grad_norm": 7.8125,
      "learning_rate": 4.923277691462186e-05,
      "loss": 0.5646823406219482,
      "mean_token_accuracy": 0.8607584916055202,
      "num_tokens": 11845409.0,
      "step": 5790
    },
    {
      "entropy": 0.6200788105838001,
      "epoch": 0.09419942668279967,
      "grad_norm": 7.46875,
      "learning_rate": 4.922958603973626e-05,
      "loss": 0.6502737045288086,
      "mean_token_accuracy": 0.8411554090678692,
      "num_tokens": 11866489.0,
      "step": 5800
    },
    {
      "entropy": 0.6292120623402297,
      "epoch": 0.09436183948742519,
      "grad_norm": 6.9375,
      "learning_rate": 4.9226388647037566e-05,
      "loss": 0.6439945220947265,
      "mean_token_accuracy": 0.8383519954979419,
      "num_tokens": 11887006.0,
      "step": 5810
    },
    {
      "entropy": 0.677078590542078,
      "epoch": 0.0945242522920507,
      "grad_norm": 7.1875,
      "learning_rate": 4.9223184737385884e-05,
      "loss": 0.7318645477294922,
      "mean_token_accuracy": 0.823764257133007,
      "num_tokens": 11906871.0,
      "step": 5820
    },
    {
      "entropy": 0.6257118264213204,
      "epoch": 0.09468666509667623,
      "grad_norm": 5.5625,
      "learning_rate": 4.921997431164307e-05,
      "loss": 0.6435337066650391,
      "mean_token_accuracy": 0.8396163575351239,
      "num_tokens": 11926335.0,
      "step": 5830
    },
    {
      "entropy": 0.7018745961599052,
      "epoch": 0.09484907790130175,
      "grad_norm": 11.25,
      "learning_rate": 4.921675737067273e-05,
      "loss": 0.6689714908599853,
      "mean_token_accuracy": 0.8301255956292153,
      "num_tokens": 11947027.0,
      "step": 5840
    },
    {
      "entropy": 0.7136979087255895,
      "epoch": 0.09501149070592725,
      "grad_norm": 9.8125,
      "learning_rate": 4.9213533915340235e-05,
      "loss": 0.6811447143554688,
      "mean_token_accuracy": 0.8385517653077841,
      "num_tokens": 11966872.0,
      "step": 5850
    },
    {
      "entropy": 0.6285030204802752,
      "epoch": 0.09517390351055277,
      "grad_norm": 7.84375,
      "learning_rate": 4.9210303946512705e-05,
      "loss": 0.6226911067962646,
      "mean_token_accuracy": 0.8465060502290725,
      "num_tokens": 11987527.0,
      "step": 5860
    },
    {
      "entropy": 0.609409558121115,
      "epoch": 0.09533631631517829,
      "grad_norm": 6.5625,
      "learning_rate": 4.9207067465058996e-05,
      "loss": 0.6263372421264648,
      "mean_token_accuracy": 0.8349673487246037,
      "num_tokens": 12006871.0,
      "step": 5870
    },
    {
      "entropy": 0.6607967047020793,
      "epoch": 0.09549872911980381,
      "grad_norm": 12.1875,
      "learning_rate": 4.920382447184975e-05,
      "loss": 0.6952318668365478,
      "mean_token_accuracy": 0.8282220892608165,
      "num_tokens": 12027045.0,
      "step": 5880
    },
    {
      "entropy": 0.6388654561713338,
      "epoch": 0.09566114192442932,
      "grad_norm": 7.28125,
      "learning_rate": 4.920057496775732e-05,
      "loss": 0.6413753986358642,
      "mean_token_accuracy": 0.8395051188766957,
      "num_tokens": 12047663.0,
      "step": 5890
    },
    {
      "entropy": 0.6895679261535406,
      "epoch": 0.09582355472905484,
      "grad_norm": 11.125,
      "learning_rate": 4.919731895365583e-05,
      "loss": 0.6976691246032715,
      "mean_token_accuracy": 0.834809473529458,
      "num_tokens": 12067909.0,
      "step": 5900
    },
    {
      "entropy": 0.6663657059893012,
      "epoch": 0.09598596753368036,
      "grad_norm": 6.5625,
      "learning_rate": 4.9194056430421164e-05,
      "loss": 0.6548609733581543,
      "mean_token_accuracy": 0.8349889636039733,
      "num_tokens": 12087003.0,
      "step": 5910
    },
    {
      "entropy": 0.6139764716848731,
      "epoch": 0.09614838033830587,
      "grad_norm": 9.1875,
      "learning_rate": 4.919078739893095e-05,
      "loss": 0.6343098163604737,
      "mean_token_accuracy": 0.8475162290036679,
      "num_tokens": 12107473.0,
      "step": 5920
    },
    {
      "entropy": 0.5921707196161151,
      "epoch": 0.0963107931429314,
      "grad_norm": 10.0625,
      "learning_rate": 4.9187511860064556e-05,
      "loss": 0.5860202312469482,
      "mean_token_accuracy": 0.8487659100443125,
      "num_tokens": 12127591.0,
      "step": 5930
    },
    {
      "entropy": 0.6342114270664752,
      "epoch": 0.0964732059475569,
      "grad_norm": 7.125,
      "learning_rate": 4.918422981470312e-05,
      "loss": 0.6286733627319336,
      "mean_token_accuracy": 0.8476006545126438,
      "num_tokens": 12146983.0,
      "step": 5940
    },
    {
      "entropy": 0.7546248151455075,
      "epoch": 0.09663561875218242,
      "grad_norm": 9.9375,
      "learning_rate": 4.918094126372951e-05,
      "loss": 0.7934096813201904,
      "mean_token_accuracy": 0.8111888900399208,
      "num_tokens": 12167492.0,
      "step": 5950
    },
    {
      "entropy": 0.6676985073834658,
      "epoch": 0.09679803155680794,
      "grad_norm": 7.90625,
      "learning_rate": 4.917764620802835e-05,
      "loss": 0.6434507369995117,
      "mean_token_accuracy": 0.8340744532644748,
      "num_tokens": 12187634.0,
      "step": 5960
    },
    {
      "entropy": 0.6119903038721531,
      "epoch": 0.09696044436143346,
      "grad_norm": 7.0625,
      "learning_rate": 4.917434464848603e-05,
      "loss": 0.6177764892578125,
      "mean_token_accuracy": 0.8404625453054905,
      "num_tokens": 12208070.0,
      "step": 5970
    },
    {
      "entropy": 0.580264858994633,
      "epoch": 0.09712285716605898,
      "grad_norm": 8.75,
      "learning_rate": 4.9171036585990674e-05,
      "loss": 0.6094505786895752,
      "mean_token_accuracy": 0.8474940560758114,
      "num_tokens": 12228907.0,
      "step": 5980
    },
    {
      "entropy": 0.5836649311706423,
      "epoch": 0.09728526997068448,
      "grad_norm": 8.0,
      "learning_rate": 4.916772202143215e-05,
      "loss": 0.5997894287109375,
      "mean_token_accuracy": 0.8503757081925869,
      "num_tokens": 12249368.0,
      "step": 5990
    },
    {
      "entropy": 0.6027049476280808,
      "epoch": 0.09744768277531,
      "grad_norm": 9.5625,
      "learning_rate": 4.9164400955702084e-05,
      "loss": 0.5994089603424072,
      "mean_token_accuracy": 0.8458671323955059,
      "num_tokens": 12269719.0,
      "step": 6000
    },
    {
      "entropy": 0.6554845773614943,
      "epoch": 0.09761009557993552,
      "grad_norm": 10.5,
      "learning_rate": 4.9161073389693864e-05,
      "loss": 0.6659326076507568,
      "mean_token_accuracy": 0.8357393354177475,
      "num_tokens": 12289469.0,
      "step": 6010
    },
    {
      "entropy": 0.5960832364857197,
      "epoch": 0.09777250838456104,
      "grad_norm": 9.6875,
      "learning_rate": 4.91577393243026e-05,
      "loss": 0.5783562660217285,
      "mean_token_accuracy": 0.8506297983229161,
      "num_tokens": 12308690.0,
      "step": 6020
    },
    {
      "entropy": 0.5752148922532797,
      "epoch": 0.09793492118918655,
      "grad_norm": 9.5,
      "learning_rate": 4.915439876042517e-05,
      "loss": 0.6113204002380371,
      "mean_token_accuracy": 0.839603129029274,
      "num_tokens": 12329583.0,
      "step": 6030
    },
    {
      "entropy": 0.6065985410474241,
      "epoch": 0.09809733399381207,
      "grad_norm": 7.84375,
      "learning_rate": 4.915105169896018e-05,
      "loss": 0.593541955947876,
      "mean_token_accuracy": 0.8497001029551029,
      "num_tokens": 12350237.0,
      "step": 6040
    },
    {
      "entropy": 0.6282325727399438,
      "epoch": 0.09825974679843759,
      "grad_norm": 6.90625,
      "learning_rate": 4.914769814080802e-05,
      "loss": 0.632943058013916,
      "mean_token_accuracy": 0.8353254850953817,
      "num_tokens": 12370420.0,
      "step": 6050
    },
    {
      "entropy": 0.5607091772370041,
      "epoch": 0.09842215960306311,
      "grad_norm": 7.3125,
      "learning_rate": 4.914433808687079e-05,
      "loss": 0.5594011783599854,
      "mean_token_accuracy": 0.8519423916935921,
      "num_tokens": 12389920.0,
      "step": 6060
    },
    {
      "entropy": 0.6183028971310705,
      "epoch": 0.09858457240768863,
      "grad_norm": 8.8125,
      "learning_rate": 4.914097153805235e-05,
      "loss": 0.602258825302124,
      "mean_token_accuracy": 0.853596031665802,
      "num_tokens": 12410944.0,
      "step": 6070
    },
    {
      "entropy": 0.6815085278823971,
      "epoch": 0.09874698521231413,
      "grad_norm": 7.125,
      "learning_rate": 4.913759849525831e-05,
      "loss": 0.697646951675415,
      "mean_token_accuracy": 0.8308833621442318,
      "num_tokens": 12431751.0,
      "step": 6080
    },
    {
      "entropy": 0.6059432869311422,
      "epoch": 0.09890939801693965,
      "grad_norm": 10.8125,
      "learning_rate": 4.9134218959396025e-05,
      "loss": 0.6104983329772949,
      "mean_token_accuracy": 0.8482199154794217,
      "num_tokens": 12452106.0,
      "step": 6090
    },
    {
      "entropy": 0.6143333061598242,
      "epoch": 0.09907181082156517,
      "grad_norm": 6.96875,
      "learning_rate": 4.913083293137461e-05,
      "loss": 0.6225437641143798,
      "mean_token_accuracy": 0.8433362126350403,
      "num_tokens": 12474058.0,
      "step": 6100
    },
    {
      "entropy": 0.612652269564569,
      "epoch": 0.09923422362619069,
      "grad_norm": 9.5,
      "learning_rate": 4.912744041210491e-05,
      "loss": 0.6343034744262696,
      "mean_token_accuracy": 0.840189841017127,
      "num_tokens": 12493704.0,
      "step": 6110
    },
    {
      "entropy": 0.5554334382526577,
      "epoch": 0.0993966364308162,
      "grad_norm": 7.25,
      "learning_rate": 4.9124041402499504e-05,
      "loss": 0.6024789810180664,
      "mean_token_accuracy": 0.8435857474803925,
      "num_tokens": 12513136.0,
      "step": 6120
    },
    {
      "entropy": 0.6666329228319228,
      "epoch": 0.09955904923544172,
      "grad_norm": 6.4375,
      "learning_rate": 4.912063590347275e-05,
      "loss": 0.6778152465820313,
      "mean_token_accuracy": 0.827507147192955,
      "num_tokens": 12534285.0,
      "step": 6130
    },
    {
      "entropy": 0.5980260455049574,
      "epoch": 0.09972146204006724,
      "grad_norm": 8.8125,
      "learning_rate": 4.911722391594074e-05,
      "loss": 0.5889979362487793,
      "mean_token_accuracy": 0.8478504225611687,
      "num_tokens": 12555325.0,
      "step": 6140
    },
    {
      "entropy": 0.6120086058042944,
      "epoch": 0.09988387484469276,
      "grad_norm": 7.125,
      "learning_rate": 4.911380544082128e-05,
      "loss": 0.5923365592956543,
      "mean_token_accuracy": 0.8444542773067951,
      "num_tokens": 12575992.0,
      "step": 6150
    },
    {
      "entropy": 0.6001654099673033,
      "epoch": 0.10004628764931828,
      "grad_norm": 8.25,
      "learning_rate": 4.911038047903397e-05,
      "loss": 0.6027307033538818,
      "mean_token_accuracy": 0.8473360680043698,
      "num_tokens": 12597334.0,
      "step": 6160
    },
    {
      "entropy": 0.6207100682891905,
      "epoch": 0.10020870045394378,
      "grad_norm": 7.03125,
      "learning_rate": 4.910694903150013e-05,
      "loss": 0.6054044246673584,
      "mean_token_accuracy": 0.8388069026172161,
      "num_tokens": 12617418.0,
      "step": 6170
    },
    {
      "entropy": 0.6936485547572374,
      "epoch": 0.1003711132585693,
      "grad_norm": 6.03125,
      "learning_rate": 4.910351109914282e-05,
      "loss": 0.7291698932647706,
      "mean_token_accuracy": 0.8230216655880213,
      "num_tokens": 12636690.0,
      "step": 6180
    },
    {
      "entropy": 0.6112619993276894,
      "epoch": 0.10053352606319482,
      "grad_norm": 7.8125,
      "learning_rate": 4.910006668288686e-05,
      "loss": 0.590873384475708,
      "mean_token_accuracy": 0.8559373661875724,
      "num_tokens": 12656389.0,
      "step": 6190
    },
    {
      "entropy": 0.5967893553897738,
      "epoch": 0.10069593886782034,
      "grad_norm": 6.25,
      "learning_rate": 4.90966157836588e-05,
      "loss": 0.5933689594268798,
      "mean_token_accuracy": 0.8468652173876763,
      "num_tokens": 12677105.0,
      "step": 6200
    },
    {
      "entropy": 0.5689596822950989,
      "epoch": 0.10085835167244586,
      "grad_norm": 8.875,
      "learning_rate": 4.9093158402386937e-05,
      "loss": 0.5938102245330811,
      "mean_token_accuracy": 0.8549989901483059,
      "num_tokens": 12696883.0,
      "step": 6210
    },
    {
      "entropy": 0.6307418226730078,
      "epoch": 0.10102076447707137,
      "grad_norm": 6.71875,
      "learning_rate": 4.908969454000132e-05,
      "loss": 0.6243972778320312,
      "mean_token_accuracy": 0.8452994950115681,
      "num_tokens": 12716908.0,
      "step": 6220
    },
    {
      "entropy": 0.6778909185901284,
      "epoch": 0.10118317728169689,
      "grad_norm": 8.5625,
      "learning_rate": 4.908622419743373e-05,
      "loss": 0.7170105457305909,
      "mean_token_accuracy": 0.8166264861822128,
      "num_tokens": 12737328.0,
      "step": 6230
    },
    {
      "entropy": 0.621736378595233,
      "epoch": 0.1013455900863224,
      "grad_norm": 8.3125,
      "learning_rate": 4.90827473756177e-05,
      "loss": 0.6073185443878174,
      "mean_token_accuracy": 0.8497686441987753,
      "num_tokens": 12757787.0,
      "step": 6240
    },
    {
      "entropy": 0.6796503243967891,
      "epoch": 0.10150800289094793,
      "grad_norm": 8.0,
      "learning_rate": 4.907926407548851e-05,
      "loss": 0.7197381973266601,
      "mean_token_accuracy": 0.8208652697503567,
      "num_tokens": 12777607.0,
      "step": 6250
    },
    {
      "entropy": 0.6326322045177222,
      "epoch": 0.10167041569557343,
      "grad_norm": 8.375,
      "learning_rate": 4.907577429798316e-05,
      "loss": 0.6549229145050048,
      "mean_token_accuracy": 0.8428358383476734,
      "num_tokens": 12797653.0,
      "step": 6260
    },
    {
      "entropy": 0.6525430457666517,
      "epoch": 0.10183282850019895,
      "grad_norm": 5.46875,
      "learning_rate": 4.907227804404042e-05,
      "loss": 0.662497615814209,
      "mean_token_accuracy": 0.8433516226708889,
      "num_tokens": 12818394.0,
      "step": 6270
    },
    {
      "entropy": 0.6628718571737409,
      "epoch": 0.10199524130482447,
      "grad_norm": 6.4375,
      "learning_rate": 4.906877531460079e-05,
      "loss": 0.6757307052612305,
      "mean_token_accuracy": 0.8292798720300197,
      "num_tokens": 12839235.0,
      "step": 6280
    },
    {
      "entropy": 0.6136347286403179,
      "epoch": 0.10215765410944999,
      "grad_norm": 8.4375,
      "learning_rate": 4.906526611060651e-05,
      "loss": 0.596739387512207,
      "mean_token_accuracy": 0.8457308225333691,
      "num_tokens": 12859680.0,
      "step": 6290
    },
    {
      "entropy": 0.6096434370614588,
      "epoch": 0.10232006691407551,
      "grad_norm": 6.5625,
      "learning_rate": 4.9061750433001564e-05,
      "loss": 0.6274813652038574,
      "mean_token_accuracy": 0.8374573700129986,
      "num_tokens": 12881118.0,
      "step": 6300
    },
    {
      "entropy": 0.6869486537761986,
      "epoch": 0.10248247971870102,
      "grad_norm": 10.0,
      "learning_rate": 4.905822828273167e-05,
      "loss": 0.726727819442749,
      "mean_token_accuracy": 0.8138910997658968,
      "num_tokens": 12902968.0,
      "step": 6310
    },
    {
      "entropy": 0.6905327472835779,
      "epoch": 0.10264489252332654,
      "grad_norm": 8.375,
      "learning_rate": 4.905469966074431e-05,
      "loss": 0.6612315654754639,
      "mean_token_accuracy": 0.8297219827771187,
      "num_tokens": 12923854.0,
      "step": 6320
    },
    {
      "entropy": 0.6386252098716796,
      "epoch": 0.10280730532795206,
      "grad_norm": 8.5625,
      "learning_rate": 4.9051164567988673e-05,
      "loss": 0.6311389923095703,
      "mean_token_accuracy": 0.8365753784775734,
      "num_tokens": 12943621.0,
      "step": 6330
    },
    {
      "entropy": 0.5903109846636653,
      "epoch": 0.10296971813257758,
      "grad_norm": 7.5,
      "learning_rate": 4.9047623005415716e-05,
      "loss": 0.5930712699890137,
      "mean_token_accuracy": 0.8421047933399677,
      "num_tokens": 12963481.0,
      "step": 6340
    },
    {
      "entropy": 0.677538734395057,
      "epoch": 0.1031321309372031,
      "grad_norm": 6.53125,
      "learning_rate": 4.904407497397813e-05,
      "loss": 0.660320520401001,
      "mean_token_accuracy": 0.8387979425489902,
      "num_tokens": 12984158.0,
      "step": 6350
    },
    {
      "entropy": 0.6726973691955209,
      "epoch": 0.1032945437418286,
      "grad_norm": 7.875,
      "learning_rate": 4.904052047463034e-05,
      "loss": 0.7057188034057618,
      "mean_token_accuracy": 0.8199150513857603,
      "num_tokens": 13003340.0,
      "step": 6360
    },
    {
      "entropy": 0.5950033149216324,
      "epoch": 0.10345695654645412,
      "grad_norm": 5.71875,
      "learning_rate": 4.903695950832852e-05,
      "loss": 0.635451078414917,
      "mean_token_accuracy": 0.8366922959685326,
      "num_tokens": 13024184.0,
      "step": 6370
    },
    {
      "entropy": 0.6123795830644667,
      "epoch": 0.10361936935107964,
      "grad_norm": 9.6875,
      "learning_rate": 4.903339207603056e-05,
      "loss": 0.63008713722229,
      "mean_token_accuracy": 0.8346828334033489,
      "num_tokens": 13043670.0,
      "step": 6380
    },
    {
      "entropy": 0.6219707918353379,
      "epoch": 0.10378178215570516,
      "grad_norm": 7.84375,
      "learning_rate": 4.902981817869613e-05,
      "loss": 0.6600622653961181,
      "mean_token_accuracy": 0.8456276528537273,
      "num_tokens": 13063421.0,
      "step": 6390
    },
    {
      "entropy": 0.6061103956773877,
      "epoch": 0.10394419496033067,
      "grad_norm": 5.90625,
      "learning_rate": 4.90262378172866e-05,
      "loss": 0.5751877307891846,
      "mean_token_accuracy": 0.8529136382043362,
      "num_tokens": 13084355.0,
      "step": 6400
    },
    {
      "entropy": 0.634320900356397,
      "epoch": 0.10410660776495619,
      "grad_norm": 8.0625,
      "learning_rate": 4.9022650992765105e-05,
      "loss": 0.6343759059906006,
      "mean_token_accuracy": 0.8439286552369595,
      "num_tokens": 13104810.0,
      "step": 6410
    },
    {
      "entropy": 0.5623653528280557,
      "epoch": 0.1042690205695817,
      "grad_norm": 7.78125,
      "learning_rate": 4.9019057706096506e-05,
      "loss": 0.5704917907714844,
      "mean_token_accuracy": 0.8563205279409886,
      "num_tokens": 13123974.0,
      "step": 6420
    },
    {
      "entropy": 0.6860338868573308,
      "epoch": 0.10443143337420722,
      "grad_norm": 8.8125,
      "learning_rate": 4.90154579582474e-05,
      "loss": 0.6964627265930176,
      "mean_token_accuracy": 0.8305376373231411,
      "num_tokens": 13143856.0,
      "step": 6430
    },
    {
      "entropy": 0.5709740391001106,
      "epoch": 0.10459384617883274,
      "grad_norm": 7.34375,
      "learning_rate": 4.901185175018613e-05,
      "loss": 0.6113903999328614,
      "mean_token_accuracy": 0.8465806435793638,
      "num_tokens": 13163545.0,
      "step": 6440
    },
    {
      "entropy": 0.6497429087758064,
      "epoch": 0.10475625898345825,
      "grad_norm": 11.3125,
      "learning_rate": 4.900823908288279e-05,
      "loss": 0.6574363708496094,
      "mean_token_accuracy": 0.8371558859944344,
      "num_tokens": 13184840.0,
      "step": 6450
    },
    {
      "entropy": 0.6203077908605337,
      "epoch": 0.10491867178808377,
      "grad_norm": 9.3125,
      "learning_rate": 4.900461995730917e-05,
      "loss": 0.5649062156677246,
      "mean_token_accuracy": 0.8436340287327766,
      "num_tokens": 13204754.0,
      "step": 6460
    },
    {
      "entropy": 0.6261931812390685,
      "epoch": 0.10508108459270929,
      "grad_norm": 8.4375,
      "learning_rate": 4.900099437443883e-05,
      "loss": 0.6290972709655762,
      "mean_token_accuracy": 0.8385742295533418,
      "num_tokens": 13224445.0,
      "step": 6470
    },
    {
      "entropy": 0.6390118858776986,
      "epoch": 0.10524349739733481,
      "grad_norm": 6.9375,
      "learning_rate": 4.899736233524708e-05,
      "loss": 0.6686450958251953,
      "mean_token_accuracy": 0.8302474655210972,
      "num_tokens": 13243995.0,
      "step": 6480
    },
    {
      "entropy": 0.6340270144864917,
      "epoch": 0.10540591020196033,
      "grad_norm": 10.0,
      "learning_rate": 4.899372384071092e-05,
      "loss": 0.6745894908905029,
      "mean_token_accuracy": 0.841634188964963,
      "num_tokens": 13263468.0,
      "step": 6490
    },
    {
      "entropy": 0.6341075411066412,
      "epoch": 0.10556832300658583,
      "grad_norm": 6.3125,
      "learning_rate": 4.8990078891809125e-05,
      "loss": 0.6327983379364014,
      "mean_token_accuracy": 0.8408970452845097,
      "num_tokens": 13283929.0,
      "step": 6500
    },
    {
      "entropy": 0.6664155066013336,
      "epoch": 0.10573073581121135,
      "grad_norm": 7.625,
      "learning_rate": 4.89864274895222e-05,
      "loss": 0.6782182216644287,
      "mean_token_accuracy": 0.833850584179163,
      "num_tokens": 13303313.0,
      "step": 6510
    },
    {
      "entropy": 0.6132186763454228,
      "epoch": 0.10589314861583687,
      "grad_norm": 5.375,
      "learning_rate": 4.898276963483237e-05,
      "loss": 0.6186100482940674,
      "mean_token_accuracy": 0.8431302443146705,
      "num_tokens": 13324373.0,
      "step": 6520
    },
    {
      "entropy": 0.6714445546269416,
      "epoch": 0.1060555614204624,
      "grad_norm": 5.46875,
      "learning_rate": 4.897910532872361e-05,
      "loss": 0.6207530975341797,
      "mean_token_accuracy": 0.8447301346808672,
      "num_tokens": 13345239.0,
      "step": 6530
    },
    {
      "entropy": 0.6688847640994936,
      "epoch": 0.1062179742250879,
      "grad_norm": 7.09375,
      "learning_rate": 4.897543457218162e-05,
      "loss": 0.6689769268035889,
      "mean_token_accuracy": 0.8340058982372284,
      "num_tokens": 13365930.0,
      "step": 6540
    },
    {
      "entropy": 0.592306956090033,
      "epoch": 0.10638038702971342,
      "grad_norm": 9.1875,
      "learning_rate": 4.897175736619386e-05,
      "loss": 0.6057229518890381,
      "mean_token_accuracy": 0.8451463550329208,
      "num_tokens": 13386363.0,
      "step": 6550
    },
    {
      "entropy": 0.5931718320585787,
      "epoch": 0.10654279983433894,
      "grad_norm": 8.75,
      "learning_rate": 4.8968073711749496e-05,
      "loss": 0.6492452621459961,
      "mean_token_accuracy": 0.8454679928719997,
      "num_tokens": 13407512.0,
      "step": 6560
    },
    {
      "entropy": 0.659595332108438,
      "epoch": 0.10670521263896446,
      "grad_norm": 7.0625,
      "learning_rate": 4.8964383609839424e-05,
      "loss": 0.6515290260314941,
      "mean_token_accuracy": 0.8403771601617336,
      "num_tokens": 13428265.0,
      "step": 6570
    },
    {
      "entropy": 0.5930665609426796,
      "epoch": 0.10686762544358998,
      "grad_norm": 6.15625,
      "learning_rate": 4.8960687061456324e-05,
      "loss": 0.5729175090789795,
      "mean_token_accuracy": 0.8499673165380954,
      "num_tokens": 13447885.0,
      "step": 6580
    },
    {
      "entropy": 0.6070752565748989,
      "epoch": 0.10703003824821548,
      "grad_norm": 6.84375,
      "learning_rate": 4.895698406759454e-05,
      "loss": 0.6022463321685791,
      "mean_token_accuracy": 0.8481736466288566,
      "num_tokens": 13468551.0,
      "step": 6590
    },
    {
      "entropy": 0.5811776736751199,
      "epoch": 0.107192451052841,
      "grad_norm": 7.4375,
      "learning_rate": 4.895327462925021e-05,
      "loss": 0.5739788055419922,
      "mean_token_accuracy": 0.8505115218460559,
      "num_tokens": 13489971.0,
      "step": 6600
    },
    {
      "entropy": 0.6433970564976335,
      "epoch": 0.10735486385746652,
      "grad_norm": 6.90625,
      "learning_rate": 4.894955874742117e-05,
      "loss": 0.6606588363647461,
      "mean_token_accuracy": 0.8362593173980712,
      "num_tokens": 13510937.0,
      "step": 6610
    },
    {
      "entropy": 0.5758526429533959,
      "epoch": 0.10751727666209204,
      "grad_norm": 6.65625,
      "learning_rate": 4.8945836423107004e-05,
      "loss": 0.6286211967468261,
      "mean_token_accuracy": 0.8415801376104355,
      "num_tokens": 13531274.0,
      "step": 6620
    },
    {
      "entropy": 0.5736963151022791,
      "epoch": 0.10767968946671756,
      "grad_norm": 6.65625,
      "learning_rate": 4.894210765730903e-05,
      "loss": 0.6679504871368408,
      "mean_token_accuracy": 0.8332979448139668,
      "num_tokens": 13552115.0,
      "step": 6630
    },
    {
      "entropy": 0.6388195671141148,
      "epoch": 0.10784210227134307,
      "grad_norm": 6.96875,
      "learning_rate": 4.893837245103028e-05,
      "loss": 0.6431354999542236,
      "mean_token_accuracy": 0.8358840674161911,
      "num_tokens": 13572359.0,
      "step": 6640
    },
    {
      "entropy": 0.6485201071016491,
      "epoch": 0.10800451507596859,
      "grad_norm": 7.71875,
      "learning_rate": 4.893463080527555e-05,
      "loss": 0.6077786445617676,
      "mean_token_accuracy": 0.8512351542711258,
      "num_tokens": 13593733.0,
      "step": 6650
    },
    {
      "entropy": 0.6238375777378679,
      "epoch": 0.10816692788059411,
      "grad_norm": 10.5625,
      "learning_rate": 4.8930882721051345e-05,
      "loss": 0.5886755466461182,
      "mean_token_accuracy": 0.8462965518236161,
      "num_tokens": 13614714.0,
      "step": 6660
    },
    {
      "entropy": 0.669042232260108,
      "epoch": 0.10832934068521963,
      "grad_norm": 7.21875,
      "learning_rate": 4.89271281993659e-05,
      "loss": 0.6709877967834472,
      "mean_token_accuracy": 0.8353570513427258,
      "num_tokens": 13634407.0,
      "step": 6670
    },
    {
      "entropy": 0.5910451102070511,
      "epoch": 0.10849175348984513,
      "grad_norm": 8.4375,
      "learning_rate": 4.89233672412292e-05,
      "loss": 0.6187312126159668,
      "mean_token_accuracy": 0.842661264538765,
      "num_tokens": 13656375.0,
      "step": 6680
    },
    {
      "entropy": 0.606962144561112,
      "epoch": 0.10865416629447065,
      "grad_norm": 8.5,
      "learning_rate": 4.891959984765295e-05,
      "loss": 0.6096054553985596,
      "mean_token_accuracy": 0.8443181097507477,
      "num_tokens": 13678261.0,
      "step": 6690
    },
    {
      "entropy": 0.620755261555314,
      "epoch": 0.10881657909909617,
      "grad_norm": 7.375,
      "learning_rate": 4.891582601965059e-05,
      "loss": 0.6147064208984375,
      "mean_token_accuracy": 0.8440775915980339,
      "num_tokens": 13699279.0,
      "step": 6700
    },
    {
      "entropy": 0.56182921952568,
      "epoch": 0.10897899190372169,
      "grad_norm": 9.5,
      "learning_rate": 4.8912045758237276e-05,
      "loss": 0.6163622379302979,
      "mean_token_accuracy": 0.8419118158519268,
      "num_tokens": 13718771.0,
      "step": 6710
    },
    {
      "entropy": 0.614237979799509,
      "epoch": 0.10914140470834721,
      "grad_norm": 5.59375,
      "learning_rate": 4.8908259064429915e-05,
      "loss": 0.5982274055480957,
      "mean_token_accuracy": 0.8418500125408173,
      "num_tokens": 13738663.0,
      "step": 6720
    },
    {
      "entropy": 0.6365445976145565,
      "epoch": 0.10930381751297272,
      "grad_norm": 7.75,
      "learning_rate": 4.890446593924714e-05,
      "loss": 0.6637476444244385,
      "mean_token_accuracy": 0.8420601941645145,
      "num_tokens": 13758053.0,
      "step": 6730
    },
    {
      "entropy": 0.6402726686559618,
      "epoch": 0.10946623031759824,
      "grad_norm": 8.75,
      "learning_rate": 4.890066638370929e-05,
      "loss": 0.6488207817077637,
      "mean_token_accuracy": 0.8352530241012573,
      "num_tokens": 13777372.0,
      "step": 6740
    },
    {
      "entropy": 0.6088193757459521,
      "epoch": 0.10962864312222376,
      "grad_norm": 9.4375,
      "learning_rate": 4.889686039883849e-05,
      "loss": 0.5872569561004639,
      "mean_token_accuracy": 0.8487405680119992,
      "num_tokens": 13798195.0,
      "step": 6750
    },
    {
      "entropy": 0.5976423224434256,
      "epoch": 0.10979105592684928,
      "grad_norm": 5.34375,
      "learning_rate": 4.889304798565852e-05,
      "loss": 0.6092954158782959,
      "mean_token_accuracy": 0.8431301355361939,
      "num_tokens": 13819754.0,
      "step": 6760
    },
    {
      "entropy": 0.6678327482193709,
      "epoch": 0.1099534687314748,
      "grad_norm": 6.8125,
      "learning_rate": 4.888922914519495e-05,
      "loss": 0.6916511058807373,
      "mean_token_accuracy": 0.8308367826044559,
      "num_tokens": 13840842.0,
      "step": 6770
    },
    {
      "entropy": 0.6302380417473614,
      "epoch": 0.1101158815361003,
      "grad_norm": 10.125,
      "learning_rate": 4.8885403878475054e-05,
      "loss": 0.611336612701416,
      "mean_token_accuracy": 0.8436132676899433,
      "num_tokens": 13860966.0,
      "step": 6780
    },
    {
      "entropy": 0.6915411301422865,
      "epoch": 0.11027829434072582,
      "grad_norm": 10.4375,
      "learning_rate": 4.8881572186527836e-05,
      "loss": 0.7231813430786133,
      "mean_token_accuracy": 0.8216905999928713,
      "num_tokens": 13882550.0,
      "step": 6790
    },
    {
      "entropy": 0.632090959046036,
      "epoch": 0.11044070714535134,
      "grad_norm": 7.5,
      "learning_rate": 4.887773407038403e-05,
      "loss": 0.6470117092132568,
      "mean_token_accuracy": 0.8358137283474207,
      "num_tokens": 13903085.0,
      "step": 6800
    },
    {
      "entropy": 0.6490141788963228,
      "epoch": 0.11060311994997686,
      "grad_norm": 8.875,
      "learning_rate": 4.88738895310761e-05,
      "loss": 0.6354988574981689,
      "mean_token_accuracy": 0.850813964009285,
      "num_tokens": 13922823.0,
      "step": 6810
    },
    {
      "entropy": 0.669668807182461,
      "epoch": 0.11076553275460237,
      "grad_norm": 8.8125,
      "learning_rate": 4.887003856963823e-05,
      "loss": 0.6762049674987793,
      "mean_token_accuracy": 0.8345622800290584,
      "num_tokens": 13942950.0,
      "step": 6820
    },
    {
      "entropy": 0.5729212189093232,
      "epoch": 0.11092794555922789,
      "grad_norm": 8.3125,
      "learning_rate": 4.8866181187106344e-05,
      "loss": 0.5701552867889405,
      "mean_token_accuracy": 0.8569412715733051,
      "num_tokens": 13963871.0,
      "step": 6830
    },
    {
      "entropy": 0.6246520292013884,
      "epoch": 0.1110903583638534,
      "grad_norm": 6.9375,
      "learning_rate": 4.8862317384518075e-05,
      "loss": 0.6252270221710206,
      "mean_token_accuracy": 0.8386704131960869,
      "num_tokens": 13984529.0,
      "step": 6840
    },
    {
      "entropy": 0.56360171912238,
      "epoch": 0.11125277116847893,
      "grad_norm": 7.03125,
      "learning_rate": 4.885844716291281e-05,
      "loss": 0.5482209682464599,
      "mean_token_accuracy": 0.8529964804649353,
      "num_tokens": 14004812.0,
      "step": 6850
    },
    {
      "entropy": 0.6201188803184777,
      "epoch": 0.11141518397310445,
      "grad_norm": 8.5625,
      "learning_rate": 4.885457052333163e-05,
      "loss": 0.6806389808654785,
      "mean_token_accuracy": 0.8331883233040571,
      "num_tokens": 14025039.0,
      "step": 6860
    },
    {
      "entropy": 0.675892140995711,
      "epoch": 0.11157759677772995,
      "grad_norm": 7.09375,
      "learning_rate": 4.885068746681738e-05,
      "loss": 0.6782083511352539,
      "mean_token_accuracy": 0.836098014190793,
      "num_tokens": 14044706.0,
      "step": 6870
    },
    {
      "entropy": 0.6391027141362429,
      "epoch": 0.11174000958235547,
      "grad_norm": 9.125,
      "learning_rate": 4.884679799441459e-05,
      "loss": 0.663369607925415,
      "mean_token_accuracy": 0.8380234383046627,
      "num_tokens": 14065036.0,
      "step": 6880
    },
    {
      "entropy": 0.6439007128588855,
      "epoch": 0.11190242238698099,
      "grad_norm": 7.125,
      "learning_rate": 4.884290210716956e-05,
      "loss": 0.6571524620056153,
      "mean_token_accuracy": 0.8414820030331611,
      "num_tokens": 14085665.0,
      "step": 6890
    },
    {
      "entropy": 0.6176325906068086,
      "epoch": 0.11206483519160651,
      "grad_norm": 7.125,
      "learning_rate": 4.883899980613027e-05,
      "loss": 0.593321704864502,
      "mean_token_accuracy": 0.8552440814673901,
      "num_tokens": 14106260.0,
      "step": 6900
    },
    {
      "entropy": 0.6730373224243522,
      "epoch": 0.11222724799623203,
      "grad_norm": 6.21875,
      "learning_rate": 4.8835091092346464e-05,
      "loss": 0.6489469051361084,
      "mean_token_accuracy": 0.8356033369898797,
      "num_tokens": 14126429.0,
      "step": 6910
    },
    {
      "entropy": 0.6162665181793272,
      "epoch": 0.11238966080085754,
      "grad_norm": 7.90625,
      "learning_rate": 4.883117596686958e-05,
      "loss": 0.6322609424591065,
      "mean_token_accuracy": 0.8343748074024916,
      "num_tokens": 14145408.0,
      "step": 6920
    },
    {
      "entropy": 0.5378955965861678,
      "epoch": 0.11255207360548306,
      "grad_norm": 7.375,
      "learning_rate": 4.882725443075281e-05,
      "loss": 0.5281177043914795,
      "mean_token_accuracy": 0.8669402003288269,
      "num_tokens": 14164958.0,
      "step": 6930
    },
    {
      "entropy": 0.6279783669859171,
      "epoch": 0.11271448641010857,
      "grad_norm": 10.75,
      "learning_rate": 4.882332648505105e-05,
      "loss": 0.7311088562011718,
      "mean_token_accuracy": 0.8274219367653132,
      "num_tokens": 14185423.0,
      "step": 6940
    },
    {
      "entropy": 0.659845559950918,
      "epoch": 0.1128768992147341,
      "grad_norm": 8.1875,
      "learning_rate": 4.8819392130820925e-05,
      "loss": 0.6503970623016357,
      "mean_token_accuracy": 0.8358858600258827,
      "num_tokens": 14206347.0,
      "step": 6950
    },
    {
      "entropy": 0.6662270731292665,
      "epoch": 0.1130393120193596,
      "grad_norm": 9.6875,
      "learning_rate": 4.8815451369120776e-05,
      "loss": 0.6333790302276612,
      "mean_token_accuracy": 0.8423026278614998,
      "num_tokens": 14226947.0,
      "step": 6960
    },
    {
      "entropy": 0.6885839609894902,
      "epoch": 0.11320172482398512,
      "grad_norm": 8.5625,
      "learning_rate": 4.8811504201010706e-05,
      "loss": 0.7121707916259765,
      "mean_token_accuracy": 0.8277998588979244,
      "num_tokens": 14246980.0,
      "step": 6970
    },
    {
      "entropy": 0.7000529564917087,
      "epoch": 0.11336413762861064,
      "grad_norm": 6.875,
      "learning_rate": 4.880755062755248e-05,
      "loss": 0.6723496913909912,
      "mean_token_accuracy": 0.832783379778266,
      "num_tokens": 14268163.0,
      "step": 6980
    },
    {
      "entropy": 0.583333833143115,
      "epoch": 0.11352655043323616,
      "grad_norm": 6.90625,
      "learning_rate": 4.8803590649809634e-05,
      "loss": 0.5970757007598877,
      "mean_token_accuracy": 0.8468028523027897,
      "num_tokens": 14288440.0,
      "step": 6990
    },
    {
      "entropy": 0.6454388303682208,
      "epoch": 0.11368896323786168,
      "grad_norm": 6.84375,
      "learning_rate": 4.87996242688474e-05,
      "loss": 0.6254180431365967,
      "mean_token_accuracy": 0.8469494134187698,
      "num_tokens": 14309122.0,
      "step": 7000
    },
    {
      "entropy": 0.6867518450133503,
      "epoch": 0.11385137604248718,
      "grad_norm": 8.875,
      "learning_rate": 4.879565148573276e-05,
      "loss": 0.7179302215576172,
      "mean_token_accuracy": 0.8309253677725792,
      "num_tokens": 14329578.0,
      "step": 7010
    },
    {
      "entropy": 0.5841545347124338,
      "epoch": 0.1140137888471127,
      "grad_norm": 11.5,
      "learning_rate": 4.87916723015344e-05,
      "loss": 0.559435510635376,
      "mean_token_accuracy": 0.8564340442419052,
      "num_tokens": 14349107.0,
      "step": 7020
    },
    {
      "entropy": 0.6947998816147447,
      "epoch": 0.11417620165173822,
      "grad_norm": 9.4375,
      "learning_rate": 4.878768671732271e-05,
      "loss": 0.6578237056732178,
      "mean_token_accuracy": 0.8378112848848105,
      "num_tokens": 14370523.0,
      "step": 7030
    },
    {
      "entropy": 0.6449729646556079,
      "epoch": 0.11433861445636374,
      "grad_norm": 7.84375,
      "learning_rate": 4.878369473416984e-05,
      "loss": 0.7106992244720459,
      "mean_token_accuracy": 0.8284663043916225,
      "num_tokens": 14391019.0,
      "step": 7040
    },
    {
      "entropy": 0.5898728968575597,
      "epoch": 0.11450102726098926,
      "grad_norm": 9.1875,
      "learning_rate": 4.877969635314963e-05,
      "loss": 0.6204934597015381,
      "mean_token_accuracy": 0.8463316172361374,
      "num_tokens": 14410755.0,
      "step": 7050
    },
    {
      "entropy": 0.7000406663864851,
      "epoch": 0.11466344006561477,
      "grad_norm": 10.0,
      "learning_rate": 4.8775691575337666e-05,
      "loss": 0.6957732677459717,
      "mean_token_accuracy": 0.8246676102280617,
      "num_tokens": 14431723.0,
      "step": 7060
    },
    {
      "entropy": 0.6235147194005549,
      "epoch": 0.11482585287024029,
      "grad_norm": 9.25,
      "learning_rate": 4.877168040181124e-05,
      "loss": 0.6475462436676025,
      "mean_token_accuracy": 0.8401832170784473,
      "num_tokens": 14451231.0,
      "step": 7070
    },
    {
      "entropy": 0.6663553392514586,
      "epoch": 0.11498826567486581,
      "grad_norm": 5.40625,
      "learning_rate": 4.876766283364935e-05,
      "loss": 0.6326309204101562,
      "mean_token_accuracy": 0.8428232319653034,
      "num_tokens": 14473701.0,
      "step": 7080
    },
    {
      "entropy": 0.6804873918183147,
      "epoch": 0.11515067847949133,
      "grad_norm": 9.75,
      "learning_rate": 4.876363887193275e-05,
      "loss": 0.7120452880859375,
      "mean_token_accuracy": 0.8237445276230574,
      "num_tokens": 14493949.0,
      "step": 7090
    },
    {
      "entropy": 0.6549240319058299,
      "epoch": 0.11531309128411683,
      "grad_norm": 9.6875,
      "learning_rate": 4.875960851774388e-05,
      "loss": 0.6439688205718994,
      "mean_token_accuracy": 0.839663740992546,
      "num_tokens": 14514394.0,
      "step": 7100
    },
    {
      "entropy": 0.6700948871672153,
      "epoch": 0.11547550408874235,
      "grad_norm": 6.625,
      "learning_rate": 4.875557177216693e-05,
      "loss": 0.7211855411529541,
      "mean_token_accuracy": 0.8280825346708298,
      "num_tokens": 14535955.0,
      "step": 7110
    },
    {
      "entropy": 0.6330760207027197,
      "epoch": 0.11563791689336787,
      "grad_norm": 8.8125,
      "learning_rate": 4.875152863628779e-05,
      "loss": 0.6327279567718506,
      "mean_token_accuracy": 0.8427125945687294,
      "num_tokens": 14554819.0,
      "step": 7120
    },
    {
      "entropy": 0.6690751980990172,
      "epoch": 0.11580032969799339,
      "grad_norm": 9.5,
      "learning_rate": 4.8747479111194063e-05,
      "loss": 0.6943010330200196,
      "mean_token_accuracy": 0.8277784004807472,
      "num_tokens": 14574662.0,
      "step": 7130
    },
    {
      "entropy": 0.6235324409790337,
      "epoch": 0.11596274250261891,
      "grad_norm": 7.6875,
      "learning_rate": 4.874342319797509e-05,
      "loss": 0.6451843738555908,
      "mean_token_accuracy": 0.8397147111594677,
      "num_tokens": 14593859.0,
      "step": 7140
    },
    {
      "entropy": 0.6349905861541629,
      "epoch": 0.11612515530724442,
      "grad_norm": 6.84375,
      "learning_rate": 4.8739360897721906e-05,
      "loss": 0.6042563438415527,
      "mean_token_accuracy": 0.8480540368705988,
      "num_tokens": 14615210.0,
      "step": 7150
    },
    {
      "entropy": 0.6056862762197852,
      "epoch": 0.11628756811186994,
      "grad_norm": 6.875,
      "learning_rate": 4.87352922115273e-05,
      "loss": 0.5884300231933594,
      "mean_token_accuracy": 0.8559227049350738,
      "num_tokens": 14635612.0,
      "step": 7160
    },
    {
      "entropy": 0.6033580385148525,
      "epoch": 0.11644998091649546,
      "grad_norm": 7.125,
      "learning_rate": 4.873121714048574e-05,
      "loss": 0.6091251850128174,
      "mean_token_accuracy": 0.8353392295539379,
      "num_tokens": 14655003.0,
      "step": 7170
    },
    {
      "entropy": 0.6596353005617857,
      "epoch": 0.11661239372112098,
      "grad_norm": 6.25,
      "learning_rate": 4.872713568569345e-05,
      "loss": 0.6886902332305909,
      "mean_token_accuracy": 0.8290299266576767,
      "num_tokens": 14675284.0,
      "step": 7180
    },
    {
      "entropy": 0.5943818567320704,
      "epoch": 0.1167748065257465,
      "grad_norm": 11.4375,
      "learning_rate": 4.872304784824833e-05,
      "loss": 0.6276145458221436,
      "mean_token_accuracy": 0.8399595737457275,
      "num_tokens": 14695261.0,
      "step": 7190
    },
    {
      "entropy": 0.5994626818224787,
      "epoch": 0.116937219330372,
      "grad_norm": 12.375,
      "learning_rate": 4.871895362925002e-05,
      "loss": 0.6485479354858399,
      "mean_token_accuracy": 0.8366317495703697,
      "num_tokens": 14716642.0,
      "step": 7200
    },
    {
      "entropy": 0.6693848886527121,
      "epoch": 0.11709963213499752,
      "grad_norm": 7.09375,
      "learning_rate": 4.871485302979988e-05,
      "loss": 0.7286888599395752,
      "mean_token_accuracy": 0.8176114059984684,
      "num_tokens": 14736843.0,
      "step": 7210
    },
    {
      "entropy": 0.6585784211754799,
      "epoch": 0.11726204493962304,
      "grad_norm": 5.9375,
      "learning_rate": 4.871074605100098e-05,
      "loss": 0.6013680458068847,
      "mean_token_accuracy": 0.8484255261719227,
      "num_tokens": 14758963.0,
      "step": 7220
    },
    {
      "entropy": 0.5930521002039313,
      "epoch": 0.11742445774424856,
      "grad_norm": 7.8125,
      "learning_rate": 4.87066326939581e-05,
      "loss": 0.5705056190490723,
      "mean_token_accuracy": 0.8507889948785305,
      "num_tokens": 14779682.0,
      "step": 7230
    },
    {
      "entropy": 0.585646181087941,
      "epoch": 0.11758687054887407,
      "grad_norm": 8.3125,
      "learning_rate": 4.870251295977776e-05,
      "loss": 0.5886154651641846,
      "mean_token_accuracy": 0.8511557713150978,
      "num_tokens": 14799732.0,
      "step": 7240
    },
    {
      "entropy": 0.556627737544477,
      "epoch": 0.11774928335349959,
      "grad_norm": 9.125,
      "learning_rate": 4.8698386849568154e-05,
      "loss": 0.5919708251953125,
      "mean_token_accuracy": 0.8468708410859108,
      "num_tokens": 14818347.0,
      "step": 7250
    },
    {
      "entropy": 0.5633593716658651,
      "epoch": 0.1179116961581251,
      "grad_norm": 8.375,
      "learning_rate": 4.869425436443924e-05,
      "loss": 0.6168779850006103,
      "mean_token_accuracy": 0.8454789571464062,
      "num_tokens": 14838085.0,
      "step": 7260
    },
    {
      "entropy": 0.6732808827189729,
      "epoch": 0.11807410896275063,
      "grad_norm": 9.4375,
      "learning_rate": 4.869011550550264e-05,
      "loss": 0.7014983654022217,
      "mean_token_accuracy": 0.830406354367733,
      "num_tokens": 14859088.0,
      "step": 7270
    },
    {
      "entropy": 0.5778659253381193,
      "epoch": 0.11823652176737615,
      "grad_norm": 6.625,
      "learning_rate": 4.868597027387174e-05,
      "loss": 0.5551287651062011,
      "mean_token_accuracy": 0.8569247588515282,
      "num_tokens": 14879448.0,
      "step": 7280
    },
    {
      "entropy": 0.6157184482552112,
      "epoch": 0.11839893457200165,
      "grad_norm": 12.25,
      "learning_rate": 4.8681818670661593e-05,
      "loss": 0.6244954586029052,
      "mean_token_accuracy": 0.8459557637572288,
      "num_tokens": 14900882.0,
      "step": 7290
    },
    {
      "entropy": 0.6667919580824673,
      "epoch": 0.11856134737662717,
      "grad_norm": 7.0625,
      "learning_rate": 4.8677660696989016e-05,
      "loss": 0.6828048706054688,
      "mean_token_accuracy": 0.8306096665561199,
      "num_tokens": 14921907.0,
      "step": 7300
    },
    {
      "entropy": 0.7016960868611932,
      "epoch": 0.11872376018125269,
      "grad_norm": 7.34375,
      "learning_rate": 4.867349635397248e-05,
      "loss": 0.716359806060791,
      "mean_token_accuracy": 0.8232195496559143,
      "num_tokens": 14942105.0,
      "step": 7310
    },
    {
      "entropy": 0.6506758023053407,
      "epoch": 0.11888617298587821,
      "grad_norm": 6.5625,
      "learning_rate": 4.8669325642732246e-05,
      "loss": 0.6265828132629394,
      "mean_token_accuracy": 0.8364840589463711,
      "num_tokens": 14962347.0,
      "step": 7320
    },
    {
      "entropy": 0.6793728331103921,
      "epoch": 0.11904858579050372,
      "grad_norm": 10.8125,
      "learning_rate": 4.86651485643902e-05,
      "loss": 0.6768495082855225,
      "mean_token_accuracy": 0.8394718952476978,
      "num_tokens": 14982923.0,
      "step": 7330
    },
    {
      "entropy": 0.6176534604281188,
      "epoch": 0.11921099859512924,
      "grad_norm": 8.875,
      "learning_rate": 4.8660965120070015e-05,
      "loss": 0.6288275241851806,
      "mean_token_accuracy": 0.8474026873707772,
      "num_tokens": 15003266.0,
      "step": 7340
    },
    {
      "entropy": 0.5955722751095891,
      "epoch": 0.11937341139975476,
      "grad_norm": 8.1875,
      "learning_rate": 4.865677531089704e-05,
      "loss": 0.6159415721893311,
      "mean_token_accuracy": 0.856735248491168,
      "num_tokens": 15023046.0,
      "step": 7350
    },
    {
      "entropy": 0.5504735384136439,
      "epoch": 0.11953582420438028,
      "grad_norm": 10.8125,
      "learning_rate": 4.865257913799834e-05,
      "loss": 0.5912277698516846,
      "mean_token_accuracy": 0.8499266438186168,
      "num_tokens": 15043325.0,
      "step": 7360
    },
    {
      "entropy": 0.6680921972263605,
      "epoch": 0.1196982370090058,
      "grad_norm": 7.71875,
      "learning_rate": 4.8648376602502695e-05,
      "loss": 0.6677202701568603,
      "mean_token_accuracy": 0.8364586301147938,
      "num_tokens": 15063912.0,
      "step": 7370
    },
    {
      "entropy": 0.6183457057923079,
      "epoch": 0.1198606498136313,
      "grad_norm": 9.5625,
      "learning_rate": 4.8644167705540593e-05,
      "loss": 0.6339645385742188,
      "mean_token_accuracy": 0.8400274563580752,
      "num_tokens": 15083557.0,
      "step": 7380
    },
    {
      "entropy": 0.6802886843681335,
      "epoch": 0.12002306261825682,
      "grad_norm": 8.5,
      "learning_rate": 4.863995244824425e-05,
      "loss": 0.6425731182098389,
      "mean_token_accuracy": 0.8400769006460905,
      "num_tokens": 15103372.0,
      "step": 7390
    },
    {
      "entropy": 0.610713183064945,
      "epoch": 0.12018547542288234,
      "grad_norm": 6.0625,
      "learning_rate": 4.863573083174756e-05,
      "loss": 0.604921531677246,
      "mean_token_accuracy": 0.8408932879567146,
      "num_tokens": 15124074.0,
      "step": 7400
    },
    {
      "entropy": 0.6164196205325425,
      "epoch": 0.12034788822750786,
      "grad_norm": 6.375,
      "learning_rate": 4.8631502857186164e-05,
      "loss": 0.583981704711914,
      "mean_token_accuracy": 0.8540714737027884,
      "num_tokens": 15144951.0,
      "step": 7410
    },
    {
      "entropy": 0.6007607446517795,
      "epoch": 0.12051030103213338,
      "grad_norm": 8.75,
      "learning_rate": 4.86272685256974e-05,
      "loss": 0.6579561710357666,
      "mean_token_accuracy": 0.8367973990738392,
      "num_tokens": 15166024.0,
      "step": 7420
    },
    {
      "entropy": 0.640202040784061,
      "epoch": 0.12067271383675889,
      "grad_norm": 5.875,
      "learning_rate": 4.862302783842029e-05,
      "loss": 0.6547267436981201,
      "mean_token_accuracy": 0.843694356828928,
      "num_tokens": 15186140.0,
      "step": 7430
    },
    {
      "entropy": 0.6304908854886889,
      "epoch": 0.1208351266413844,
      "grad_norm": 10.0,
      "learning_rate": 4.861878079649561e-05,
      "loss": 0.6184782981872559,
      "mean_token_accuracy": 0.8437280673533678,
      "num_tokens": 15206838.0,
      "step": 7440
    },
    {
      "entropy": 0.567953909188509,
      "epoch": 0.12099753944600992,
      "grad_norm": 10.5625,
      "learning_rate": 4.861452740106581e-05,
      "loss": 0.581709623336792,
      "mean_token_accuracy": 0.8579314887523651,
      "num_tokens": 15226190.0,
      "step": 7450
    },
    {
      "entropy": 0.560838807746768,
      "epoch": 0.12115995225063544,
      "grad_norm": 6.28125,
      "learning_rate": 4.8610267653275076e-05,
      "loss": 0.5765435695648193,
      "mean_token_accuracy": 0.8543419755995274,
      "num_tokens": 15246656.0,
      "step": 7460
    },
    {
      "entropy": 0.6734841542318464,
      "epoch": 0.12132236505526095,
      "grad_norm": 10.0,
      "learning_rate": 4.8606001554269275e-05,
      "loss": 0.6710083961486817,
      "mean_token_accuracy": 0.841570220515132,
      "num_tokens": 15268922.0,
      "step": 7470
    },
    {
      "entropy": 0.6127755543217063,
      "epoch": 0.12148477785988647,
      "grad_norm": 9.0625,
      "learning_rate": 4.860172910519601e-05,
      "loss": 0.6012803077697754,
      "mean_token_accuracy": 0.8373278252780437,
      "num_tokens": 15289142.0,
      "step": 7480
    },
    {
      "entropy": 0.5476171324029565,
      "epoch": 0.12164719066451199,
      "grad_norm": 6.75,
      "learning_rate": 4.859745030720457e-05,
      "loss": 0.5680364608764649,
      "mean_token_accuracy": 0.8531597897410392,
      "num_tokens": 15309367.0,
      "step": 7490
    },
    {
      "entropy": 0.6312498510349542,
      "epoch": 0.12180960346913751,
      "grad_norm": 11.5,
      "learning_rate": 4.859316516144597e-05,
      "loss": 0.6221071720123291,
      "mean_token_accuracy": 0.8452431313693524,
      "num_tokens": 15329615.0,
      "step": 7500
    },
    {
      "entropy": 0.7319263068959116,
      "epoch": 0.12197201627376303,
      "grad_norm": 8.25,
      "learning_rate": 4.8588873669072915e-05,
      "loss": 0.7496577262878418,
      "mean_token_accuracy": 0.8232697162777185,
      "num_tokens": 15350838.0,
      "step": 7510
    },
    {
      "entropy": 0.5956805900670588,
      "epoch": 0.12213442907838853,
      "grad_norm": 9.125,
      "learning_rate": 4.858457583123983e-05,
      "loss": 0.6017590045928956,
      "mean_token_accuracy": 0.8455284662544728,
      "num_tokens": 15369977.0,
      "step": 7520
    },
    {
      "entropy": 0.6103997309692204,
      "epoch": 0.12229684188301405,
      "grad_norm": 6.5625,
      "learning_rate": 4.858027164910285e-05,
      "loss": 0.5640773773193359,
      "mean_token_accuracy": 0.8637897662818432,
      "num_tokens": 15389758.0,
      "step": 7530
    },
    {
      "entropy": 0.5763955962378532,
      "epoch": 0.12245925468763957,
      "grad_norm": 8.0625,
      "learning_rate": 4.8575961123819795e-05,
      "loss": 0.5918526172637939,
      "mean_token_accuracy": 0.8524938590824604,
      "num_tokens": 15411488.0,
      "step": 7540
    },
    {
      "entropy": 0.5492710906080902,
      "epoch": 0.1226216674922651,
      "grad_norm": 7.5,
      "learning_rate": 4.857164425655022e-05,
      "loss": 0.6020978450775146,
      "mean_token_accuracy": 0.8505723290145397,
      "num_tokens": 15431645.0,
      "step": 7550
    },
    {
      "entropy": 0.6210299440659582,
      "epoch": 0.12278408029689061,
      "grad_norm": 7.34375,
      "learning_rate": 4.856732104845536e-05,
      "loss": 0.6359973907470703,
      "mean_token_accuracy": 0.8416826289147139,
      "num_tokens": 15451539.0,
      "step": 7560
    },
    {
      "entropy": 0.6309370721690357,
      "epoch": 0.12294649310151612,
      "grad_norm": 7.84375,
      "learning_rate": 4.856299150069818e-05,
      "loss": 0.6386836051940918,
      "mean_token_accuracy": 0.8368857085704804,
      "num_tokens": 15472379.0,
      "step": 7570
    },
    {
      "entropy": 0.6273793512955308,
      "epoch": 0.12310890590614164,
      "grad_norm": 7.40625,
      "learning_rate": 4.855865561444333e-05,
      "loss": 0.6163985252380371,
      "mean_token_accuracy": 0.847382290661335,
      "num_tokens": 15493319.0,
      "step": 7580
    },
    {
      "entropy": 0.6242604058235883,
      "epoch": 0.12327131871076716,
      "grad_norm": 7.625,
      "learning_rate": 4.855431339085717e-05,
      "loss": 0.6415502548217773,
      "mean_token_accuracy": 0.8408247478306293,
      "num_tokens": 15515465.0,
      "step": 7590
    },
    {
      "entropy": 0.5763382008764892,
      "epoch": 0.12343373151539268,
      "grad_norm": 8.1875,
      "learning_rate": 4.854996483110778e-05,
      "loss": 0.5750430583953857,
      "mean_token_accuracy": 0.8425360545516014,
      "num_tokens": 15535541.0,
      "step": 7600
    },
    {
      "entropy": 0.6822948658838868,
      "epoch": 0.12359614432001818,
      "grad_norm": 8.75,
      "learning_rate": 4.854560993636492e-05,
      "loss": 0.6955759048461914,
      "mean_token_accuracy": 0.8186479087918996,
      "num_tokens": 15556040.0,
      "step": 7610
    },
    {
      "entropy": 0.609074542298913,
      "epoch": 0.1237585571246437,
      "grad_norm": 10.625,
      "learning_rate": 4.854124870780008e-05,
      "loss": 0.6055562019348144,
      "mean_token_accuracy": 0.8485120698809624,
      "num_tokens": 15576238.0,
      "step": 7620
    },
    {
      "entropy": 0.6269749568775296,
      "epoch": 0.12392096992926922,
      "grad_norm": 7.4375,
      "learning_rate": 4.853688114658642e-05,
      "loss": 0.6646974086761475,
      "mean_token_accuracy": 0.8391705103218555,
      "num_tokens": 15596646.0,
      "step": 7630
    },
    {
      "entropy": 0.6298165768384933,
      "epoch": 0.12408338273389474,
      "grad_norm": 5.875,
      "learning_rate": 4.853250725389884e-05,
      "loss": 0.6116808891296387,
      "mean_token_accuracy": 0.8425516501069069,
      "num_tokens": 15617354.0,
      "step": 7640
    },
    {
      "entropy": 0.5775593463331461,
      "epoch": 0.12424579553852026,
      "grad_norm": 9.875,
      "learning_rate": 4.8528127030913925e-05,
      "loss": 0.6081920146942139,
      "mean_token_accuracy": 0.8469698771834373,
      "num_tokens": 15637242.0,
      "step": 7650
    },
    {
      "entropy": 0.6062617841176688,
      "epoch": 0.12440820834314577,
      "grad_norm": 6.5,
      "learning_rate": 4.8523740478809965e-05,
      "loss": 0.6285706520080566,
      "mean_token_accuracy": 0.8504937168210744,
      "num_tokens": 15657128.0,
      "step": 7660
    },
    {
      "entropy": 0.5479859788902104,
      "epoch": 0.12457062114777129,
      "grad_norm": 5.96875,
      "learning_rate": 4.851934759876694e-05,
      "loss": 0.5487621784210205,
      "mean_token_accuracy": 0.8601790960878134,
      "num_tokens": 15677076.0,
      "step": 7670
    },
    {
      "entropy": 0.6316460096044466,
      "epoch": 0.12473303395239681,
      "grad_norm": 11.25,
      "learning_rate": 4.8514948391966566e-05,
      "loss": 0.6669494152069092,
      "mean_token_accuracy": 0.8348216570913791,
      "num_tokens": 15698066.0,
      "step": 7680
    },
    {
      "entropy": 0.6628032744396478,
      "epoch": 0.12489544675702233,
      "grad_norm": 8.625,
      "learning_rate": 4.851054285959222e-05,
      "loss": 0.6674529552459717,
      "mean_token_accuracy": 0.8374891072511673,
      "num_tokens": 15718577.0,
      "step": 7690
    },
    {
      "entropy": 0.6248521354049444,
      "epoch": 0.12505785956164783,
      "grad_norm": 5.46875,
      "learning_rate": 4.850613100282901e-05,
      "loss": 0.6675062656402588,
      "mean_token_accuracy": 0.8277590230107308,
      "num_tokens": 15738470.0,
      "step": 7700
    },
    {
      "entropy": 0.6901901179924608,
      "epoch": 0.12522027236627337,
      "grad_norm": 6.6875,
      "learning_rate": 4.8501712822863746e-05,
      "loss": 0.6707696914672852,
      "mean_token_accuracy": 0.8383807621896266,
      "num_tokens": 15759479.0,
      "step": 7710
    },
    {
      "entropy": 0.5925189642235636,
      "epoch": 0.12538268517089887,
      "grad_norm": 9.0,
      "learning_rate": 4.84972883208849e-05,
      "loss": 0.5746827602386475,
      "mean_token_accuracy": 0.8556260541081429,
      "num_tokens": 15779705.0,
      "step": 7720
    },
    {
      "entropy": 0.6398994821123779,
      "epoch": 0.12554509797552438,
      "grad_norm": 6.5625,
      "learning_rate": 4.849285749808269e-05,
      "loss": 0.615979528427124,
      "mean_token_accuracy": 0.840717040002346,
      "num_tokens": 15800617.0,
      "step": 7730
    },
    {
      "entropy": 0.5638362001627684,
      "epoch": 0.1257075107801499,
      "grad_norm": 7.53125,
      "learning_rate": 4.8488420355649025e-05,
      "loss": 0.5557899475097656,
      "mean_token_accuracy": 0.863019373267889,
      "num_tokens": 15819646.0,
      "step": 7740
    },
    {
      "entropy": 0.6186658089514822,
      "epoch": 0.12586992358477542,
      "grad_norm": 9.6875,
      "learning_rate": 4.848397689477749e-05,
      "loss": 0.6154867649078369,
      "mean_token_accuracy": 0.8464500948786735,
      "num_tokens": 15840857.0,
      "step": 7750
    },
    {
      "entropy": 0.5703591752797366,
      "epoch": 0.12603233638940095,
      "grad_norm": 9.25,
      "learning_rate": 4.8479527116663396e-05,
      "loss": 0.5840265274047851,
      "mean_token_accuracy": 0.8546222925186158,
      "num_tokens": 15860373.0,
      "step": 7760
    },
    {
      "entropy": 0.6932855471968651,
      "epoch": 0.12619474919402646,
      "grad_norm": 7.4375,
      "learning_rate": 4.8475071022503734e-05,
      "loss": 0.6869179725646972,
      "mean_token_accuracy": 0.8237015314400196,
      "num_tokens": 15880774.0,
      "step": 7770
    },
    {
      "entropy": 0.6024833154398948,
      "epoch": 0.12635716199865196,
      "grad_norm": 7.28125,
      "learning_rate": 4.847060861349722e-05,
      "loss": 0.6251701831817627,
      "mean_token_accuracy": 0.8417396977543831,
      "num_tokens": 15901455.0,
      "step": 7780
    },
    {
      "entropy": 0.6365250603761524,
      "epoch": 0.1265195748032775,
      "grad_norm": 7.4375,
      "learning_rate": 4.8466139890844236e-05,
      "loss": 0.6903351306915283,
      "mean_token_accuracy": 0.8256426963955164,
      "num_tokens": 15923308.0,
      "step": 7790
    },
    {
      "entropy": 0.5827832067385316,
      "epoch": 0.126681987607903,
      "grad_norm": 7.71875,
      "learning_rate": 4.846166485574689e-05,
      "loss": 0.5857896327972412,
      "mean_token_accuracy": 0.8434395104646683,
      "num_tokens": 15943225.0,
      "step": 7800
    },
    {
      "entropy": 0.6343667877838015,
      "epoch": 0.12684440041252854,
      "grad_norm": 8.0625,
      "learning_rate": 4.845718350940896e-05,
      "loss": 0.580354642868042,
      "mean_token_accuracy": 0.8512080490589142,
      "num_tokens": 15963343.0,
      "step": 7810
    },
    {
      "entropy": 0.6492784551344812,
      "epoch": 0.12700681321715404,
      "grad_norm": 7.4375,
      "learning_rate": 4.845269585303596e-05,
      "loss": 0.630829906463623,
      "mean_token_accuracy": 0.8393112353980541,
      "num_tokens": 15984076.0,
      "step": 7820
    },
    {
      "entropy": 0.6305854475125671,
      "epoch": 0.12716922602177955,
      "grad_norm": 7.5,
      "learning_rate": 4.8448201887835064e-05,
      "loss": 0.6465327739715576,
      "mean_token_accuracy": 0.835585230588913,
      "num_tokens": 16004111.0,
      "step": 7830
    },
    {
      "entropy": 0.5650284240022302,
      "epoch": 0.12733163882640508,
      "grad_norm": 7.6875,
      "learning_rate": 4.844370161501517e-05,
      "loss": 0.5929896831512451,
      "mean_token_accuracy": 0.8557569023221732,
      "num_tokens": 16023767.0,
      "step": 7840
    },
    {
      "entropy": 0.6274729921482504,
      "epoch": 0.12749405163103059,
      "grad_norm": 7.59375,
      "learning_rate": 4.843919503578685e-05,
      "loss": 0.6637857913970947,
      "mean_token_accuracy": 0.834445309638977,
      "num_tokens": 16044399.0,
      "step": 7850
    },
    {
      "entropy": 0.6519081751815975,
      "epoch": 0.12765646443565612,
      "grad_norm": 8.375,
      "learning_rate": 4.8434682151362385e-05,
      "loss": 0.6519596099853515,
      "mean_token_accuracy": 0.8383948385715485,
      "num_tokens": 16064677.0,
      "step": 7860
    },
    {
      "entropy": 0.6375115774571896,
      "epoch": 0.12781887724028163,
      "grad_norm": 5.71875,
      "learning_rate": 4.843016296295576e-05,
      "loss": 0.6133630275726318,
      "mean_token_accuracy": 0.8395615514367819,
      "num_tokens": 16085253.0,
      "step": 7870
    },
    {
      "entropy": 0.5950665302574635,
      "epoch": 0.12798129004490713,
      "grad_norm": 12.625,
      "learning_rate": 4.842563747178265e-05,
      "loss": 0.600170373916626,
      "mean_token_accuracy": 0.8479973211884498,
      "num_tokens": 16104421.0,
      "step": 7880
    },
    {
      "entropy": 0.6079981143586337,
      "epoch": 0.12814370284953266,
      "grad_norm": 8.0,
      "learning_rate": 4.84211056790604e-05,
      "loss": 0.6188004970550537,
      "mean_token_accuracy": 0.844696868211031,
      "num_tokens": 16125818.0,
      "step": 7890
    },
    {
      "entropy": 0.6150463140569628,
      "epoch": 0.12830611565415817,
      "grad_norm": 6.09375,
      "learning_rate": 4.841656758600809e-05,
      "loss": 0.6155622482299805,
      "mean_token_accuracy": 0.848316814750433,
      "num_tokens": 16146289.0,
      "step": 7900
    },
    {
      "entropy": 0.6402594113722444,
      "epoch": 0.1284685284587837,
      "grad_norm": 7.4375,
      "learning_rate": 4.841202319384648e-05,
      "loss": 0.6352404117584228,
      "mean_token_accuracy": 0.8428412113338709,
      "num_tokens": 16168690.0,
      "step": 7910
    },
    {
      "entropy": 0.6167690481990575,
      "epoch": 0.1286309412634092,
      "grad_norm": 17.0,
      "learning_rate": 4.840747250379801e-05,
      "loss": 0.6364299297332764,
      "mean_token_accuracy": 0.8411181882023812,
      "num_tokens": 16189736.0,
      "step": 7920
    },
    {
      "entropy": 0.6205445756204426,
      "epoch": 0.12879335406803472,
      "grad_norm": 6.875,
      "learning_rate": 4.840291551708684e-05,
      "loss": 0.6238097667694091,
      "mean_token_accuracy": 0.8449044957756996,
      "num_tokens": 16210965.0,
      "step": 7930
    },
    {
      "entropy": 0.5817910208366811,
      "epoch": 0.12895576687266025,
      "grad_norm": 8.875,
      "learning_rate": 4.8398352234938794e-05,
      "loss": 0.6258269309997558,
      "mean_token_accuracy": 0.8443170059472322,
      "num_tokens": 16230942.0,
      "step": 7940
    },
    {
      "entropy": 0.578250830154866,
      "epoch": 0.12911817967728575,
      "grad_norm": 7.9375,
      "learning_rate": 4.839378265858141e-05,
      "loss": 0.5993615627288819,
      "mean_token_accuracy": 0.8485110260546207,
      "num_tokens": 16251667.0,
      "step": 7950
    },
    {
      "entropy": 0.6684277079999447,
      "epoch": 0.12928059248191126,
      "grad_norm": 8.375,
      "learning_rate": 4.8389206789243925e-05,
      "loss": 0.6276770114898682,
      "mean_token_accuracy": 0.8393541127443314,
      "num_tokens": 16272430.0,
      "step": 7960
    },
    {
      "entropy": 0.5910617119632662,
      "epoch": 0.1294430052865368,
      "grad_norm": 7.5625,
      "learning_rate": 4.838462462815724e-05,
      "loss": 0.5928113460540771,
      "mean_token_accuracy": 0.8496321134269238,
      "num_tokens": 16292443.0,
      "step": 7970
    },
    {
      "entropy": 0.6718311110511422,
      "epoch": 0.1296054180911623,
      "grad_norm": 8.9375,
      "learning_rate": 4.8380036176553975e-05,
      "loss": 0.6581284523010253,
      "mean_token_accuracy": 0.8403376929461956,
      "num_tokens": 16313001.0,
      "step": 7980
    },
    {
      "entropy": 0.6199390255846083,
      "epoch": 0.12976783089578783,
      "grad_norm": 10.375,
      "learning_rate": 4.8375441435668445e-05,
      "loss": 0.6699905395507812,
      "mean_token_accuracy": 0.8359346866607666,
      "num_tokens": 16333028.0,
      "step": 7990
    },
    {
      "epoch": 0.12993024370041334,
      "eval_entropy": 0.6512044446468354,
      "eval_loss": 0.6597878336906433,
      "eval_mean_token_accuracy": 0.8333811078071595,
      "eval_num_tokens": 16353816.0,
      "eval_runtime": 40.1175,
      "eval_samples_per_second": 49.854,
      "eval_steps_per_second": 6.232,
      "step": 8000
    },
    {
      "entropy": 0.6002837738138623,
      "epoch": 0.13009265650503885,
      "grad_norm": 7.65625,
      "learning_rate": 4.836623309099623e-05,
      "loss": 0.6236101150512695,
      "mean_token_accuracy": 0.8401299823075533,
      "num_tokens": 16374463.0,
      "step": 8010
    },
    {
      "entropy": 0.6160245701670647,
      "epoch": 0.13025506930966438,
      "grad_norm": 7.6875,
      "learning_rate": 4.8361619489686606e-05,
      "loss": 0.6124776363372803,
      "mean_token_accuracy": 0.841727401316166,
      "num_tokens": 16395242.0,
      "step": 8020
    },
    {
      "entropy": 0.6377709816209972,
      "epoch": 0.13041748211428988,
      "grad_norm": 6.53125,
      "learning_rate": 4.835699960404885e-05,
      "loss": 0.6485124588012695,
      "mean_token_accuracy": 0.8398052036762238,
      "num_tokens": 16416702.0,
      "step": 8030
    },
    {
      "entropy": 0.6418528018519283,
      "epoch": 0.13057989491891542,
      "grad_norm": 6.65625,
      "learning_rate": 4.835237343532569e-05,
      "loss": 0.665829086303711,
      "mean_token_accuracy": 0.8453404977917671,
      "num_tokens": 16438032.0,
      "step": 8040
    },
    {
      "entropy": 0.586878748331219,
      "epoch": 0.13074230772354092,
      "grad_norm": 6.40625,
      "learning_rate": 4.8347740984761616e-05,
      "loss": 0.5672257423400879,
      "mean_token_accuracy": 0.8515409275889396,
      "num_tokens": 16457970.0,
      "step": 8050
    },
    {
      "entropy": 0.6691286198794841,
      "epoch": 0.13090472052816643,
      "grad_norm": 7.90625,
      "learning_rate": 4.8343102253602735e-05,
      "loss": 0.7103353977203369,
      "mean_token_accuracy": 0.8292105980217457,
      "num_tokens": 16479122.0,
      "step": 8060
    },
    {
      "entropy": 0.6820785360410809,
      "epoch": 0.13106713333279196,
      "grad_norm": 10.6875,
      "learning_rate": 4.8338457243096884e-05,
      "loss": 0.6687970161437988,
      "mean_token_accuracy": 0.8260152503848076,
      "num_tokens": 16498423.0,
      "step": 8070
    },
    {
      "entropy": 0.6616665316745639,
      "epoch": 0.13122954613741747,
      "grad_norm": 11.0625,
      "learning_rate": 4.833380595449359e-05,
      "loss": 0.6667894840240478,
      "mean_token_accuracy": 0.8441867493093014,
      "num_tokens": 16519794.0,
      "step": 8080
    },
    {
      "entropy": 0.6193872195668518,
      "epoch": 0.131391958942043,
      "grad_norm": 8.5,
      "learning_rate": 4.832914838904405e-05,
      "loss": 0.5703090190887451,
      "mean_token_accuracy": 0.8525636248290539,
      "num_tokens": 16538888.0,
      "step": 8090
    },
    {
      "entropy": 0.6712438145652413,
      "epoch": 0.1315543717466685,
      "grad_norm": 12.375,
      "learning_rate": 4.832448454800117e-05,
      "loss": 0.7059022426605225,
      "mean_token_accuracy": 0.8354361575096846,
      "num_tokens": 16558517.0,
      "step": 8100
    },
    {
      "entropy": 0.6247339708730578,
      "epoch": 0.131716784551294,
      "grad_norm": 8.3125,
      "learning_rate": 4.8319814432619526e-05,
      "loss": 0.6660276412963867,
      "mean_token_accuracy": 0.8404201753437519,
      "num_tokens": 16577964.0,
      "step": 8110
    },
    {
      "entropy": 0.6482763753272593,
      "epoch": 0.13187919735591955,
      "grad_norm": 6.75,
      "learning_rate": 4.831513804415539e-05,
      "loss": 0.6290378570556641,
      "mean_token_accuracy": 0.8450009893625975,
      "num_tokens": 16598015.0,
      "step": 8120
    },
    {
      "entropy": 0.5352983921766281,
      "epoch": 0.13204161016054505,
      "grad_norm": 7.5625,
      "learning_rate": 4.8310455383866734e-05,
      "loss": 0.5190680503845215,
      "mean_token_accuracy": 0.8622370228171349,
      "num_tokens": 16618662.0,
      "step": 8130
    },
    {
      "entropy": 0.5972808548249304,
      "epoch": 0.1322040229651706,
      "grad_norm": 10.4375,
      "learning_rate": 4.830576645301318e-05,
      "loss": 0.6521633625030517,
      "mean_token_accuracy": 0.83751914575696,
      "num_tokens": 16639387.0,
      "step": 8140
    },
    {
      "entropy": 0.5835219390690327,
      "epoch": 0.1323664357697961,
      "grad_norm": 4.96875,
      "learning_rate": 4.830107125285608e-05,
      "loss": 0.6134262084960938,
      "mean_token_accuracy": 0.8468405909836292,
      "num_tokens": 16660264.0,
      "step": 8150
    },
    {
      "entropy": 0.6249120375141501,
      "epoch": 0.1325288485744216,
      "grad_norm": 9.125,
      "learning_rate": 4.829636978465844e-05,
      "loss": 0.6265413761138916,
      "mean_token_accuracy": 0.8365109220147133,
      "num_tokens": 16680610.0,
      "step": 8160
    },
    {
      "entropy": 0.6322290762327611,
      "epoch": 0.13269126137904713,
      "grad_norm": 6.75,
      "learning_rate": 4.829166204968498e-05,
      "loss": 0.596568489074707,
      "mean_token_accuracy": 0.8476288914680481,
      "num_tokens": 16701930.0,
      "step": 8170
    },
    {
      "entropy": 0.5928081708028913,
      "epoch": 0.13285367418367264,
      "grad_norm": 7.3125,
      "learning_rate": 4.828694804920208e-05,
      "loss": 0.646027660369873,
      "mean_token_accuracy": 0.8451593883335591,
      "num_tokens": 16722190.0,
      "step": 8180
    },
    {
      "entropy": 0.6093266423791647,
      "epoch": 0.13301608698829817,
      "grad_norm": 6.15625,
      "learning_rate": 4.828222778447782e-05,
      "loss": 0.6057221412658691,
      "mean_token_accuracy": 0.8497763186693191,
      "num_tokens": 16742276.0,
      "step": 8190
    },
    {
      "entropy": 0.6136107521131635,
      "epoch": 0.13317849979292368,
      "grad_norm": 7.625,
      "learning_rate": 4.8277501256781964e-05,
      "loss": 0.5997648715972901,
      "mean_token_accuracy": 0.8429534651339055,
      "num_tokens": 16763073.0,
      "step": 8200
    },
    {
      "entropy": 0.608992308191955,
      "epoch": 0.13334091259754918,
      "grad_norm": 8.5625,
      "learning_rate": 4.827276846738596e-05,
      "loss": 0.6207221984863281,
      "mean_token_accuracy": 0.8499386459589005,
      "num_tokens": 16782752.0,
      "step": 8210
    },
    {
      "entropy": 0.6417168821208179,
      "epoch": 0.13350332540217472,
      "grad_norm": 7.75,
      "learning_rate": 4.8268029417562935e-05,
      "loss": 0.6070938587188721,
      "mean_token_accuracy": 0.8402719169855117,
      "num_tokens": 16801792.0,
      "step": 8220
    },
    {
      "entropy": 0.6015859568491578,
      "epoch": 0.13366573820680022,
      "grad_norm": 7.5,
      "learning_rate": 4.82632841085877e-05,
      "loss": 0.6316658020019531,
      "mean_token_accuracy": 0.8397339679300785,
      "num_tokens": 16821149.0,
      "step": 8230
    },
    {
      "entropy": 0.632970084855333,
      "epoch": 0.13382815101142573,
      "grad_norm": 4.875,
      "learning_rate": 4.825853254173675e-05,
      "loss": 0.6383554935455322,
      "mean_token_accuracy": 0.8498728990554809,
      "num_tokens": 16840958.0,
      "step": 8240
    },
    {
      "entropy": 0.6239515744149685,
      "epoch": 0.13399056381605126,
      "grad_norm": 8.6875,
      "learning_rate": 4.8253774718288294e-05,
      "loss": 0.6052073955535888,
      "mean_token_accuracy": 0.8569794990122318,
      "num_tokens": 16861430.0,
      "step": 8250
    },
    {
      "entropy": 0.6152473001740872,
      "epoch": 0.13415297662067677,
      "grad_norm": 9.75,
      "learning_rate": 4.824901063952217e-05,
      "loss": 0.6460752964019776,
      "mean_token_accuracy": 0.8433881383389235,
      "num_tokens": 16881447.0,
      "step": 8260
    },
    {
      "entropy": 0.647532626800239,
      "epoch": 0.1343153894253023,
      "grad_norm": 8.6875,
      "learning_rate": 4.8244240306719945e-05,
      "loss": 0.707727575302124,
      "mean_token_accuracy": 0.827545964717865,
      "num_tokens": 16901273.0,
      "step": 8270
    },
    {
      "entropy": 0.6197872122749686,
      "epoch": 0.1344778022299278,
      "grad_norm": 8.5625,
      "learning_rate": 4.823946372116484e-05,
      "loss": 0.5668740749359131,
      "mean_token_accuracy": 0.849470553547144,
      "num_tokens": 16922359.0,
      "step": 8280
    },
    {
      "entropy": 0.5717398472130298,
      "epoch": 0.1346402150345533,
      "grad_norm": 11.6875,
      "learning_rate": 4.8234680884141745e-05,
      "loss": 0.5939699172973633,
      "mean_token_accuracy": 0.8414655283093453,
      "num_tokens": 16942434.0,
      "step": 8290
    },
    {
      "entropy": 0.6043474021367728,
      "epoch": 0.13480262783917885,
      "grad_norm": 8.9375,
      "learning_rate": 4.82298917969373e-05,
      "loss": 0.5975978374481201,
      "mean_token_accuracy": 0.8487195003777742,
      "num_tokens": 16964170.0,
      "step": 8300
    },
    {
      "entropy": 0.58698799489066,
      "epoch": 0.13496504064380435,
      "grad_norm": 11.125,
      "learning_rate": 4.822509646083974e-05,
      "loss": 0.6196430683135986,
      "mean_token_accuracy": 0.850944322347641,
      "num_tokens": 16985103.0,
      "step": 8310
    },
    {
      "entropy": 0.6047907034866512,
      "epoch": 0.13512745344842989,
      "grad_norm": 8.25,
      "learning_rate": 4.822029487713904e-05,
      "loss": 0.6213749408721924,
      "mean_token_accuracy": 0.8484115842729807,
      "num_tokens": 17004827.0,
      "step": 8320
    },
    {
      "entropy": 0.6345514814369381,
      "epoch": 0.1352898662530554,
      "grad_norm": 9.5,
      "learning_rate": 4.821548704712684e-05,
      "loss": 0.6249322891235352,
      "mean_token_accuracy": 0.8428008608520031,
      "num_tokens": 17025021.0,
      "step": 8330
    },
    {
      "entropy": 0.5765104973688722,
      "epoch": 0.1354522790576809,
      "grad_norm": 7.59375,
      "learning_rate": 4.8210672972096446e-05,
      "loss": 0.5767984390258789,
      "mean_token_accuracy": 0.8518035106360913,
      "num_tokens": 17044124.0,
      "step": 8340
    },
    {
      "entropy": 0.5617858508601785,
      "epoch": 0.13561469186230643,
      "grad_norm": 7.71875,
      "learning_rate": 4.820585265334286e-05,
      "loss": 0.5512345790863037,
      "mean_token_accuracy": 0.8665278743952513,
      "num_tokens": 17065164.0,
      "step": 8350
    },
    {
      "entropy": 0.7074656657874584,
      "epoch": 0.13577710466693194,
      "grad_norm": 6.15625,
      "learning_rate": 4.820102609216276e-05,
      "loss": 0.7114555358886718,
      "mean_token_accuracy": 0.8246616054326296,
      "num_tokens": 17085388.0,
      "step": 8360
    },
    {
      "entropy": 0.5556575623340905,
      "epoch": 0.13593951747155747,
      "grad_norm": 9.8125,
      "learning_rate": 4.81961932898545e-05,
      "loss": 0.5528528213500976,
      "mean_token_accuracy": 0.8575867399573326,
      "num_tokens": 17105465.0,
      "step": 8370
    },
    {
      "entropy": 0.5605721755884587,
      "epoch": 0.13610193027618298,
      "grad_norm": 10.6875,
      "learning_rate": 4.8191354247718114e-05,
      "loss": 0.596155834197998,
      "mean_token_accuracy": 0.8503992758691311,
      "num_tokens": 17126407.0,
      "step": 8380
    },
    {
      "entropy": 0.5901248413138092,
      "epoch": 0.13626434308080848,
      "grad_norm": 9.4375,
      "learning_rate": 4.818650896705531e-05,
      "loss": 0.6547510623931885,
      "mean_token_accuracy": 0.8477352254092694,
      "num_tokens": 17147632.0,
      "step": 8390
    },
    {
      "entropy": 0.6591494657099247,
      "epoch": 0.13642675588543401,
      "grad_norm": 5.4375,
      "learning_rate": 4.818165744916949e-05,
      "loss": 0.6659523963928222,
      "mean_token_accuracy": 0.8445221774280072,
      "num_tokens": 17168117.0,
      "step": 8400
    },
    {
      "entropy": 0.672949877474457,
      "epoch": 0.13658916869005952,
      "grad_norm": 7.90625,
      "learning_rate": 4.817679969536572e-05,
      "loss": 0.6292330741882324,
      "mean_token_accuracy": 0.8421241238713264,
      "num_tokens": 17189327.0,
      "step": 8410
    },
    {
      "entropy": 0.6465920623391866,
      "epoch": 0.13675158149468505,
      "grad_norm": 5.34375,
      "learning_rate": 4.817193570695074e-05,
      "loss": 0.6280487060546875,
      "mean_token_accuracy": 0.8433338016271591,
      "num_tokens": 17211229.0,
      "step": 8420
    },
    {
      "entropy": 0.647976403683424,
      "epoch": 0.13691399429931056,
      "grad_norm": 6.3125,
      "learning_rate": 4.8167065485232994e-05,
      "loss": 0.6501492500305176,
      "mean_token_accuracy": 0.8366633869707585,
      "num_tokens": 17232095.0,
      "step": 8430
    },
    {
      "entropy": 0.6474811005406081,
      "epoch": 0.13707640710393607,
      "grad_norm": 8.75,
      "learning_rate": 4.8162189031522556e-05,
      "loss": 0.662212610244751,
      "mean_token_accuracy": 0.8346234235912562,
      "num_tokens": 17252545.0,
      "step": 8440
    },
    {
      "entropy": 0.6744946873746812,
      "epoch": 0.1372388199085616,
      "grad_norm": 12.0625,
      "learning_rate": 4.8157306347131216e-05,
      "loss": 0.7001089572906494,
      "mean_token_accuracy": 0.8311279579997063,
      "num_tokens": 17272984.0,
      "step": 8450
    },
    {
      "entropy": 0.557199427112937,
      "epoch": 0.1374012327131871,
      "grad_norm": 8.375,
      "learning_rate": 4.8152417433372423e-05,
      "loss": 0.5723966121673584,
      "mean_token_accuracy": 0.8493536658585071,
      "num_tokens": 17292973.0,
      "step": 8460
    },
    {
      "entropy": 0.7297378906048835,
      "epoch": 0.1375636455178126,
      "grad_norm": 7.625,
      "learning_rate": 4.814752229156132e-05,
      "loss": 0.7361185073852539,
      "mean_token_accuracy": 0.822200732678175,
      "num_tokens": 17313628.0,
      "step": 8470
    },
    {
      "entropy": 0.6316851863637567,
      "epoch": 0.13772605832243814,
      "grad_norm": 5.5,
      "learning_rate": 4.814262092301469e-05,
      "loss": 0.6168622493743896,
      "mean_token_accuracy": 0.8409944657236338,
      "num_tokens": 17333527.0,
      "step": 8480
    },
    {
      "entropy": 0.677087564393878,
      "epoch": 0.13788847112706365,
      "grad_norm": 7.90625,
      "learning_rate": 4.813771332905102e-05,
      "loss": 0.674922513961792,
      "mean_token_accuracy": 0.8243818216025829,
      "num_tokens": 17353329.0,
      "step": 8490
    },
    {
      "entropy": 0.6535451786592603,
      "epoch": 0.13805088393168918,
      "grad_norm": 7.15625,
      "learning_rate": 4.813279951099047e-05,
      "loss": 0.6694985389709472,
      "mean_token_accuracy": 0.8277546979486943,
      "num_tokens": 17372939.0,
      "step": 8500
    },
    {
      "entropy": 0.6181755037046969,
      "epoch": 0.1382132967363147,
      "grad_norm": 14.3125,
      "learning_rate": 4.8127879470154866e-05,
      "loss": 0.6153836727142334,
      "mean_token_accuracy": 0.8436597272753715,
      "num_tokens": 17394086.0,
      "step": 8510
    },
    {
      "entropy": 0.5728129694238305,
      "epoch": 0.1383757095409402,
      "grad_norm": 6.84375,
      "learning_rate": 4.8122953207867704e-05,
      "loss": 0.6187980651855469,
      "mean_token_accuracy": 0.8469486124813557,
      "num_tokens": 17415331.0,
      "step": 8520
    },
    {
      "entropy": 0.6382316127419472,
      "epoch": 0.13853812234556573,
      "grad_norm": 7.375,
      "learning_rate": 4.8118020725454165e-05,
      "loss": 0.6913145542144775,
      "mean_token_accuracy": 0.833270663022995,
      "num_tokens": 17435921.0,
      "step": 8530
    },
    {
      "entropy": 0.6189044157974422,
      "epoch": 0.13870053515019123,
      "grad_norm": 7.5,
      "learning_rate": 4.8113082024241096e-05,
      "loss": 0.6309364795684814,
      "mean_token_accuracy": 0.8435577481985093,
      "num_tokens": 17456980.0,
      "step": 8540
    },
    {
      "entropy": 0.6040691871196031,
      "epoch": 0.13886294795481677,
      "grad_norm": 5.25,
      "learning_rate": 4.810813710555702e-05,
      "loss": 0.5949012756347656,
      "mean_token_accuracy": 0.8508635856211185,
      "num_tokens": 17477910.0,
      "step": 8550
    },
    {
      "entropy": 0.6506683127954602,
      "epoch": 0.13902536075944227,
      "grad_norm": 8.625,
      "learning_rate": 4.810318597073213e-05,
      "loss": 0.6433038711547852,
      "mean_token_accuracy": 0.843486038595438,
      "num_tokens": 17498053.0,
      "step": 8560
    },
    {
      "entropy": 0.6198129962198436,
      "epoch": 0.13918777356406778,
      "grad_norm": 8.8125,
      "learning_rate": 4.8098228621098296e-05,
      "loss": 0.6160755157470703,
      "mean_token_accuracy": 0.8377025984227657,
      "num_tokens": 17519206.0,
      "step": 8570
    },
    {
      "entropy": 0.7188583522103726,
      "epoch": 0.1393501863686933,
      "grad_norm": 5.8125,
      "learning_rate": 4.809326505798905e-05,
      "loss": 0.7471960544586181,
      "mean_token_accuracy": 0.818550718203187,
      "num_tokens": 17539265.0,
      "step": 8580
    },
    {
      "entropy": 0.6336034391541034,
      "epoch": 0.13951259917331882,
      "grad_norm": 7.0,
      "learning_rate": 4.8088295282739596e-05,
      "loss": 0.6573149681091308,
      "mean_token_accuracy": 0.8382480029016733,
      "num_tokens": 17559723.0,
      "step": 8590
    },
    {
      "entropy": 0.6569077915512025,
      "epoch": 0.13967501197794435,
      "grad_norm": 7.1875,
      "learning_rate": 4.8083319296686835e-05,
      "loss": 0.668522310256958,
      "mean_token_accuracy": 0.8360339254140854,
      "num_tokens": 17579817.0,
      "step": 8600
    },
    {
      "entropy": 0.5352107850834728,
      "epoch": 0.13983742478256986,
      "grad_norm": 8.125,
      "learning_rate": 4.80783371011693e-05,
      "loss": 0.5300621509552002,
      "mean_token_accuracy": 0.8656205184757709,
      "num_tokens": 17600407.0,
      "step": 8610
    },
    {
      "entropy": 0.6283737347461283,
      "epoch": 0.13999983758719536,
      "grad_norm": 9.8125,
      "learning_rate": 4.807334869752722e-05,
      "loss": 0.6202961444854737,
      "mean_token_accuracy": 0.8459905974566937,
      "num_tokens": 17620507.0,
      "step": 8620
    },
    {
      "entropy": 0.5526928183622658,
      "epoch": 0.1401622503918209,
      "grad_norm": 6.78125,
      "learning_rate": 4.806835408710249e-05,
      "loss": 0.5597908020019531,
      "mean_token_accuracy": 0.8588463172316552,
      "num_tokens": 17640900.0,
      "step": 8630
    },
    {
      "entropy": 0.6468131605535745,
      "epoch": 0.1403246631964464,
      "grad_norm": 7.90625,
      "learning_rate": 4.806335327123866e-05,
      "loss": 0.6744547843933105,
      "mean_token_accuracy": 0.8320296593010426,
      "num_tokens": 17660891.0,
      "step": 8640
    },
    {
      "entropy": 0.6314097859431058,
      "epoch": 0.14048707600107194,
      "grad_norm": 7.65625,
      "learning_rate": 4.805834625128098e-05,
      "loss": 0.6375339984893799,
      "mean_token_accuracy": 0.8433954488486052,
      "num_tokens": 17681418.0,
      "step": 8650
    },
    {
      "entropy": 0.5357688201125711,
      "epoch": 0.14064948880569744,
      "grad_norm": 8.5,
      "learning_rate": 4.805333302857633e-05,
      "loss": 0.5254237651824951,
      "mean_token_accuracy": 0.8610177971422672,
      "num_tokens": 17702678.0,
      "step": 8660
    },
    {
      "entropy": 0.5494879598729312,
      "epoch": 0.14081190161032295,
      "grad_norm": 8.6875,
      "learning_rate": 4.8048313604473293e-05,
      "loss": 0.5278853416442871,
      "mean_token_accuracy": 0.8578646138310433,
      "num_tokens": 17722471.0,
      "step": 8670
    },
    {
      "entropy": 0.6348255965858698,
      "epoch": 0.14097431441494848,
      "grad_norm": 7.03125,
      "learning_rate": 4.804328798032209e-05,
      "loss": 0.6691157817840576,
      "mean_token_accuracy": 0.836989164352417,
      "num_tokens": 17742749.0,
      "step": 8680
    },
    {
      "entropy": 0.6216359496116638,
      "epoch": 0.141136727219574,
      "grad_norm": 11.6875,
      "learning_rate": 4.803825615747464e-05,
      "loss": 0.6069502353668212,
      "mean_token_accuracy": 0.8472615905106068,
      "num_tokens": 17762555.0,
      "step": 8690
    },
    {
      "entropy": 0.6245265663601458,
      "epoch": 0.14129914002419952,
      "grad_norm": 6.65625,
      "learning_rate": 4.803321813728451e-05,
      "loss": 0.6186189174652099,
      "mean_token_accuracy": 0.8466506563127041,
      "num_tokens": 17781825.0,
      "step": 8700
    },
    {
      "entropy": 0.603897444717586,
      "epoch": 0.14146155282882503,
      "grad_norm": 8.125,
      "learning_rate": 4.8028173921106935e-05,
      "loss": 0.5964725017547607,
      "mean_token_accuracy": 0.8599480934441089,
      "num_tokens": 17801791.0,
      "step": 8710
    },
    {
      "entropy": 0.6199798761401325,
      "epoch": 0.14162396563345053,
      "grad_norm": 11.0,
      "learning_rate": 4.8023123510298815e-05,
      "loss": 0.6387128353118896,
      "mean_token_accuracy": 0.844667038321495,
      "num_tokens": 17822323.0,
      "step": 8720
    },
    {
      "entropy": 0.5863169393502176,
      "epoch": 0.14178637843807607,
      "grad_norm": 7.65625,
      "learning_rate": 4.801806690621874e-05,
      "loss": 0.6053766727447509,
      "mean_token_accuracy": 0.8421031963080168,
      "num_tokens": 17842553.0,
      "step": 8730
    },
    {
      "entropy": 0.6366223342716694,
      "epoch": 0.14194879124270157,
      "grad_norm": 6.0625,
      "learning_rate": 4.801300411022693e-05,
      "loss": 0.5729235649108887,
      "mean_token_accuracy": 0.8551762863993645,
      "num_tokens": 17862821.0,
      "step": 8740
    },
    {
      "entropy": 0.5494275813922286,
      "epoch": 0.14211120404732708,
      "grad_norm": 9.625,
      "learning_rate": 4.8007935123685297e-05,
      "loss": 0.561906099319458,
      "mean_token_accuracy": 0.8588738687336445,
      "num_tokens": 17883838.0,
      "step": 8750
    },
    {
      "entropy": 0.6048988094553351,
      "epoch": 0.1422736168519526,
      "grad_norm": 8.25,
      "learning_rate": 4.800285994795741e-05,
      "loss": 0.6365787029266358,
      "mean_token_accuracy": 0.8365307912230492,
      "num_tokens": 17904520.0,
      "step": 8760
    },
    {
      "entropy": 0.5896524689160287,
      "epoch": 0.14243602965657812,
      "grad_norm": 8.25,
      "learning_rate": 4.79977785844085e-05,
      "loss": 0.6091255187988281,
      "mean_token_accuracy": 0.8466038279235363,
      "num_tokens": 17924765.0,
      "step": 8770
    },
    {
      "entropy": 0.6075211567804217,
      "epoch": 0.14259844246120365,
      "grad_norm": 10.8125,
      "learning_rate": 4.7992691034405476e-05,
      "loss": 0.6383663177490234,
      "mean_token_accuracy": 0.8371895730495453,
      "num_tokens": 17946877.0,
      "step": 8780
    },
    {
      "entropy": 0.535763907013461,
      "epoch": 0.14276085526582916,
      "grad_norm": 8.625,
      "learning_rate": 4.798759729931688e-05,
      "loss": 0.493394660949707,
      "mean_token_accuracy": 0.8665234051644802,
      "num_tokens": 17966619.0,
      "step": 8790
    },
    {
      "entropy": 0.590048149973154,
      "epoch": 0.14292326807045466,
      "grad_norm": 10.3125,
      "learning_rate": 4.7982497380512945e-05,
      "loss": 0.5950063228607178,
      "mean_token_accuracy": 0.8475125938653946,
      "num_tokens": 17987381.0,
      "step": 8800
    },
    {
      "entropy": 0.6200922342017293,
      "epoch": 0.1430856808750802,
      "grad_norm": 8.5,
      "learning_rate": 4.797739127936558e-05,
      "loss": 0.6113042831420898,
      "mean_token_accuracy": 0.845060658082366,
      "num_tokens": 18007954.0,
      "step": 8810
    },
    {
      "entropy": 0.6321099965833128,
      "epoch": 0.1432480936797057,
      "grad_norm": 8.25,
      "learning_rate": 4.797227899724831e-05,
      "loss": 0.6427988052368164,
      "mean_token_accuracy": 0.8395305521786213,
      "num_tokens": 18027573.0,
      "step": 8820
    },
    {
      "entropy": 0.5789375138469041,
      "epoch": 0.14341050648433123,
      "grad_norm": 5.65625,
      "learning_rate": 4.796716053553637e-05,
      "loss": 0.5639959812164307,
      "mean_token_accuracy": 0.8546579889953136,
      "num_tokens": 18047771.0,
      "step": 8830
    },
    {
      "entropy": 0.5127344804815948,
      "epoch": 0.14357291928895674,
      "grad_norm": 8.125,
      "learning_rate": 4.796203589560663e-05,
      "loss": 0.5762393474578857,
      "mean_token_accuracy": 0.8536888927221298,
      "num_tokens": 18067502.0,
      "step": 8840
    },
    {
      "entropy": 0.638441712781787,
      "epoch": 0.14373533209358225,
      "grad_norm": 7.6875,
      "learning_rate": 4.795690507883762e-05,
      "loss": 0.6196602821350098,
      "mean_token_accuracy": 0.8432079963386059,
      "num_tokens": 18086834.0,
      "step": 8850
    },
    {
      "entropy": 0.6633912310004234,
      "epoch": 0.14389774489820778,
      "grad_norm": 9.75,
      "learning_rate": 4.795176808660956e-05,
      "loss": 0.6984458446502686,
      "mean_token_accuracy": 0.8249208971858024,
      "num_tokens": 18105943.0,
      "step": 8860
    },
    {
      "entropy": 0.7260170597583055,
      "epoch": 0.14406015770283329,
      "grad_norm": 8.0625,
      "learning_rate": 4.7946624920304304e-05,
      "loss": 0.739480447769165,
      "mean_token_accuracy": 0.823813508450985,
      "num_tokens": 18126554.0,
      "step": 8870
    },
    {
      "entropy": 0.5872039915062487,
      "epoch": 0.14422257050745882,
      "grad_norm": 9.125,
      "learning_rate": 4.7941475581305376e-05,
      "loss": 0.6195332527160644,
      "mean_token_accuracy": 0.8462771445512771,
      "num_tokens": 18145758.0,
      "step": 8880
    },
    {
      "entropy": 0.6454668801277876,
      "epoch": 0.14438498331208433,
      "grad_norm": 7.9375,
      "learning_rate": 4.793632007099796e-05,
      "loss": 0.6371306419372559,
      "mean_token_accuracy": 0.8479496099054813,
      "num_tokens": 18166369.0,
      "step": 8890
    },
    {
      "entropy": 0.6207818016409874,
      "epoch": 0.14454739611670983,
      "grad_norm": 7.71875,
      "learning_rate": 4.793115839076889e-05,
      "loss": 0.6086225032806396,
      "mean_token_accuracy": 0.8488855518400669,
      "num_tokens": 18187631.0,
      "step": 8900
    },
    {
      "entropy": 0.647044169716537,
      "epoch": 0.14470980892133536,
      "grad_norm": 7.65625,
      "learning_rate": 4.792599054200669e-05,
      "loss": 0.630368995666504,
      "mean_token_accuracy": 0.8473310865461826,
      "num_tokens": 18206494.0,
      "step": 8910
    },
    {
      "entropy": 0.6967596558853983,
      "epoch": 0.14487222172596087,
      "grad_norm": 11.0,
      "learning_rate": 4.79208165261015e-05,
      "loss": 0.7384674072265625,
      "mean_token_accuracy": 0.8266443505883216,
      "num_tokens": 18227930.0,
      "step": 8920
    },
    {
      "entropy": 0.5934607520699501,
      "epoch": 0.1450346345305864,
      "grad_norm": 5.90625,
      "learning_rate": 4.791563634444516e-05,
      "loss": 0.5694482803344727,
      "mean_token_accuracy": 0.8540064878761768,
      "num_tokens": 18249285.0,
      "step": 8930
    },
    {
      "entropy": 0.6964512680657208,
      "epoch": 0.1451970473352119,
      "grad_norm": 6.65625,
      "learning_rate": 4.791044999843114e-05,
      "loss": 0.7183862209320069,
      "mean_token_accuracy": 0.8237753346562385,
      "num_tokens": 18270855.0,
      "step": 8940
    },
    {
      "entropy": 0.7606234382838011,
      "epoch": 0.14535946013983742,
      "grad_norm": 7.75,
      "learning_rate": 4.790525748945459e-05,
      "loss": 0.7820095539093017,
      "mean_token_accuracy": 0.8059801030904055,
      "num_tokens": 18292495.0,
      "step": 8950
    },
    {
      "entropy": 0.642435679025948,
      "epoch": 0.14552187294446295,
      "grad_norm": 7.375,
      "learning_rate": 4.790005881891229e-05,
      "loss": 0.6539732933044433,
      "mean_token_accuracy": 0.8353834673762321,
      "num_tokens": 18313284.0,
      "step": 8960
    },
    {
      "entropy": 0.5751558568328619,
      "epoch": 0.14568428574908845,
      "grad_norm": 7.09375,
      "learning_rate": 4.78948539882027e-05,
      "loss": 0.550535774230957,
      "mean_token_accuracy": 0.8527534812688827,
      "num_tokens": 18333872.0,
      "step": 8970
    },
    {
      "entropy": 0.6356768487021327,
      "epoch": 0.145846698553714,
      "grad_norm": 10.5,
      "learning_rate": 4.788964299872595e-05,
      "loss": 0.6564856052398682,
      "mean_token_accuracy": 0.8408609978854656,
      "num_tokens": 18355042.0,
      "step": 8980
    },
    {
      "entropy": 0.5898800482973456,
      "epoch": 0.1460091113583395,
      "grad_norm": 7.53125,
      "learning_rate": 4.7884425851883776e-05,
      "loss": 0.5684184074401856,
      "mean_token_accuracy": 0.8519672967493535,
      "num_tokens": 18375349.0,
      "step": 8990
    },
    {
      "entropy": 0.5975275709759444,
      "epoch": 0.146171524162965,
      "grad_norm": 8.0,
      "learning_rate": 4.787920254907963e-05,
      "loss": 0.6020020008087158,
      "mean_token_accuracy": 0.8505015857517719,
      "num_tokens": 18395664.0,
      "step": 9000
    },
    {
      "entropy": 0.6543949369341135,
      "epoch": 0.14633393696759053,
      "grad_norm": 8.1875,
      "learning_rate": 4.787397309171857e-05,
      "loss": 0.6423961639404296,
      "mean_token_accuracy": 0.8405837953090668,
      "num_tokens": 18414997.0,
      "step": 9010
    },
    {
      "entropy": 0.6199060659855604,
      "epoch": 0.14649634977221604,
      "grad_norm": 10.125,
      "learning_rate": 4.786873748120735e-05,
      "loss": 0.6181083679199219,
      "mean_token_accuracy": 0.8354721985757351,
      "num_tokens": 18435259.0,
      "step": 9020
    },
    {
      "entropy": 0.5546169493347406,
      "epoch": 0.14665876257684154,
      "grad_norm": 8.5,
      "learning_rate": 4.7863495718954355e-05,
      "loss": 0.575148344039917,
      "mean_token_accuracy": 0.8507708147168159,
      "num_tokens": 18455992.0,
      "step": 9030
    },
    {
      "entropy": 0.5520155111327767,
      "epoch": 0.14682117538146708,
      "grad_norm": 7.9375,
      "learning_rate": 4.785824780636963e-05,
      "loss": 0.561829662322998,
      "mean_token_accuracy": 0.8581220734864473,
      "num_tokens": 18476478.0,
      "step": 9040
    },
    {
      "entropy": 0.6744874310679734,
      "epoch": 0.14698358818609258,
      "grad_norm": 7.03125,
      "learning_rate": 4.785299374486486e-05,
      "loss": 0.6754958629608154,
      "mean_token_accuracy": 0.8356623984873295,
      "num_tokens": 18497302.0,
      "step": 9050
    },
    {
      "entropy": 0.6525230835191905,
      "epoch": 0.14714600099071812,
      "grad_norm": 6.03125,
      "learning_rate": 4.784773353585343e-05,
      "loss": 0.6599681854248047,
      "mean_token_accuracy": 0.8379609428346158,
      "num_tokens": 18519168.0,
      "step": 9060
    },
    {
      "entropy": 0.6066897181794048,
      "epoch": 0.14730841379534362,
      "grad_norm": 8.1875,
      "learning_rate": 4.7842467180750324e-05,
      "loss": 0.573602819442749,
      "mean_token_accuracy": 0.8542135834693909,
      "num_tokens": 18538875.0,
      "step": 9070
    },
    {
      "entropy": 0.585414238832891,
      "epoch": 0.14747082659996913,
      "grad_norm": 8.5625,
      "learning_rate": 4.7837194680972206e-05,
      "loss": 0.5836697578430176,
      "mean_token_accuracy": 0.850047605484724,
      "num_tokens": 18559647.0,
      "step": 9080
    },
    {
      "entropy": 0.5991341841407121,
      "epoch": 0.14763323940459466,
      "grad_norm": 5.625,
      "learning_rate": 4.783191603793741e-05,
      "loss": 0.6530519485473633,
      "mean_token_accuracy": 0.8373940519988536,
      "num_tokens": 18581607.0,
      "step": 9090
    },
    {
      "entropy": 0.5853006782010197,
      "epoch": 0.14779565220922017,
      "grad_norm": 7.90625,
      "learning_rate": 4.782663125306588e-05,
      "loss": 0.6276863574981689,
      "mean_token_accuracy": 0.8422775700688362,
      "num_tokens": 18600924.0,
      "step": 9100
    },
    {
      "entropy": 0.6123523896560072,
      "epoch": 0.1479580650138457,
      "grad_norm": 6.21875,
      "learning_rate": 4.782134032777924e-05,
      "loss": 0.6104051113128662,
      "mean_token_accuracy": 0.8474269192665815,
      "num_tokens": 18622314.0,
      "step": 9110
    },
    {
      "entropy": 0.5908287854865193,
      "epoch": 0.1481204778184712,
      "grad_norm": 9.25,
      "learning_rate": 4.7816043263500764e-05,
      "loss": 0.5948742389678955,
      "mean_token_accuracy": 0.8567447409033775,
      "num_tokens": 18642725.0,
      "step": 9120
    },
    {
      "entropy": 0.6689820470288396,
      "epoch": 0.1482828906230967,
      "grad_norm": 10.125,
      "learning_rate": 4.781074006165538e-05,
      "loss": 0.6869154930114746,
      "mean_token_accuracy": 0.8300175409764051,
      "num_tokens": 18663687.0,
      "step": 9130
    },
    {
      "entropy": 0.624483793694526,
      "epoch": 0.14844530342772225,
      "grad_norm": 7.125,
      "learning_rate": 4.7805430723669656e-05,
      "loss": 0.6391027927398681,
      "mean_token_accuracy": 0.8439335450530052,
      "num_tokens": 18683561.0,
      "step": 9140
    },
    {
      "entropy": 0.606061231624335,
      "epoch": 0.14860771623234775,
      "grad_norm": 10.5625,
      "learning_rate": 4.780011525097181e-05,
      "loss": 0.5734774112701416,
      "mean_token_accuracy": 0.8566057853400707,
      "num_tokens": 18703962.0,
      "step": 9150
    },
    {
      "entropy": 0.6509995698928833,
      "epoch": 0.1487701290369733,
      "grad_norm": 8.75,
      "learning_rate": 4.779479364499172e-05,
      "loss": 0.6732940196990966,
      "mean_token_accuracy": 0.8378686271607876,
      "num_tokens": 18725401.0,
      "step": 9160
    },
    {
      "entropy": 0.6176853084936738,
      "epoch": 0.1489325418415988,
      "grad_norm": 12.0,
      "learning_rate": 4.778946590716092e-05,
      "loss": 0.6686935424804688,
      "mean_token_accuracy": 0.8384677097201347,
      "num_tokens": 18745900.0,
      "step": 9170
    },
    {
      "entropy": 0.6305993126705289,
      "epoch": 0.1490949546462243,
      "grad_norm": 8.4375,
      "learning_rate": 4.778413203891257e-05,
      "loss": 0.669380521774292,
      "mean_token_accuracy": 0.8343824103474617,
      "num_tokens": 18766879.0,
      "step": 9180
    },
    {
      "entropy": 0.6602992217056454,
      "epoch": 0.14925736745084983,
      "grad_norm": 7.875,
      "learning_rate": 4.77787920416815e-05,
      "loss": 0.6773960590362549,
      "mean_token_accuracy": 0.836940386146307,
      "num_tokens": 18786710.0,
      "step": 9190
    },
    {
      "entropy": 0.6405679081566632,
      "epoch": 0.14941978025547534,
      "grad_norm": 5.46875,
      "learning_rate": 4.7773445916904175e-05,
      "loss": 0.6209536552429199,
      "mean_token_accuracy": 0.8485879246145487,
      "num_tokens": 18807377.0,
      "step": 9200
    },
    {
      "entropy": 0.6160450478084385,
      "epoch": 0.14958219306010087,
      "grad_norm": 7.375,
      "learning_rate": 4.776809366601872e-05,
      "loss": 0.6320612907409668,
      "mean_token_accuracy": 0.843862647563219,
      "num_tokens": 18828217.0,
      "step": 9210
    },
    {
      "entropy": 0.6477995933033526,
      "epoch": 0.14974460586472638,
      "grad_norm": 7.09375,
      "learning_rate": 4.77627352904649e-05,
      "loss": 0.638882064819336,
      "mean_token_accuracy": 0.8336539629846811,
      "num_tokens": 18848628.0,
      "step": 9220
    },
    {
      "entropy": 0.5719367106445133,
      "epoch": 0.14990701866935188,
      "grad_norm": 7.1875,
      "learning_rate": 4.775737079168413e-05,
      "loss": 0.5550568580627442,
      "mean_token_accuracy": 0.8616848662495613,
      "num_tokens": 18869336.0,
      "step": 9230
    },
    {
      "entropy": 0.6118508036248386,
      "epoch": 0.15006943147397742,
      "grad_norm": 10.375,
      "learning_rate": 4.775200017111947e-05,
      "loss": 0.6317952632904053,
      "mean_token_accuracy": 0.8357299461960792,
      "num_tokens": 18890199.0,
      "step": 9240
    },
    {
      "entropy": 0.6451764466241002,
      "epoch": 0.15023184427860292,
      "grad_norm": 7.0625,
      "learning_rate": 4.774662343021564e-05,
      "loss": 0.6544273376464844,
      "mean_token_accuracy": 0.8358705043792725,
      "num_tokens": 18911713.0,
      "step": 9250
    },
    {
      "entropy": 0.5338861975818873,
      "epoch": 0.15039425708322846,
      "grad_norm": 8.5,
      "learning_rate": 4.774124057041898e-05,
      "loss": 0.5663548469543457,
      "mean_token_accuracy": 0.8531091459095478,
      "num_tokens": 18931721.0,
      "step": 9260
    },
    {
      "entropy": 0.6365612098947168,
      "epoch": 0.15055666988785396,
      "grad_norm": 8.6875,
      "learning_rate": 4.77358515931775e-05,
      "loss": 0.626079797744751,
      "mean_token_accuracy": 0.8469805810600519,
      "num_tokens": 18951837.0,
      "step": 9270
    },
    {
      "entropy": 0.6128804503940046,
      "epoch": 0.15071908269247947,
      "grad_norm": 7.96875,
      "learning_rate": 4.7730456499940854e-05,
      "loss": 0.6717112541198731,
      "mean_token_accuracy": 0.8364338502287865,
      "num_tokens": 18972136.0,
      "step": 9280
    },
    {
      "entropy": 0.671223658695817,
      "epoch": 0.150881495497105,
      "grad_norm": 8.125,
      "learning_rate": 4.772505529216032e-05,
      "loss": 0.6622696876525879,
      "mean_token_accuracy": 0.8370720207691192,
      "num_tokens": 18992215.0,
      "step": 9290
    },
    {
      "entropy": 0.6786268007010221,
      "epoch": 0.1510439083017305,
      "grad_norm": 9.5625,
      "learning_rate": 4.771964797128884e-05,
      "loss": 0.6802579402923584,
      "mean_token_accuracy": 0.8324664890766144,
      "num_tokens": 19013386.0,
      "step": 9300
    },
    {
      "entropy": 0.6320101212710142,
      "epoch": 0.151206321106356,
      "grad_norm": 8.625,
      "learning_rate": 4.7714234538781e-05,
      "loss": 0.641576623916626,
      "mean_token_accuracy": 0.841977807879448,
      "num_tokens": 19033413.0,
      "step": 9310
    },
    {
      "entropy": 0.6837564260698855,
      "epoch": 0.15136873391098155,
      "grad_norm": 7.8125,
      "learning_rate": 4.7708814996093024e-05,
      "loss": 0.6872798919677734,
      "mean_token_accuracy": 0.8306986182928086,
      "num_tokens": 19053777.0,
      "step": 9320
    },
    {
      "entropy": 0.6602279188111424,
      "epoch": 0.15153114671560705,
      "grad_norm": 8.75,
      "learning_rate": 4.770338934468277e-05,
      "loss": 0.6122377872467041,
      "mean_token_accuracy": 0.8492344062775373,
      "num_tokens": 19074150.0,
      "step": 9330
    },
    {
      "entropy": 0.5828174193389714,
      "epoch": 0.15169355952023258,
      "grad_norm": 7.4375,
      "learning_rate": 4.769795758600976e-05,
      "loss": 0.5968486309051514,
      "mean_token_accuracy": 0.8481022179126739,
      "num_tokens": 19095472.0,
      "step": 9340
    },
    {
      "entropy": 0.586430444708094,
      "epoch": 0.1518559723248581,
      "grad_norm": 7.5625,
      "learning_rate": 4.769251972153515e-05,
      "loss": 0.6035156726837159,
      "mean_token_accuracy": 0.8519285053014756,
      "num_tokens": 19115103.0,
      "step": 9350
    },
    {
      "entropy": 0.6085972970351576,
      "epoch": 0.1520183851294836,
      "grad_norm": 7.9375,
      "learning_rate": 4.768707575272174e-05,
      "loss": 0.6127945899963378,
      "mean_token_accuracy": 0.8465140715241433,
      "num_tokens": 19135329.0,
      "step": 9360
    },
    {
      "entropy": 0.6286271648481488,
      "epoch": 0.15218079793410913,
      "grad_norm": 8.625,
      "learning_rate": 4.768162568103397e-05,
      "loss": 0.5919161319732666,
      "mean_token_accuracy": 0.8485459983348846,
      "num_tokens": 19155481.0,
      "step": 9370
    },
    {
      "entropy": 0.5644595285877585,
      "epoch": 0.15234321073873464,
      "grad_norm": 6.4375,
      "learning_rate": 4.767616950793792e-05,
      "loss": 0.5931994915008545,
      "mean_token_accuracy": 0.8527758210897446,
      "num_tokens": 19176477.0,
      "step": 9380
    },
    {
      "entropy": 0.595355412364006,
      "epoch": 0.15250562354336017,
      "grad_norm": 7.53125,
      "learning_rate": 4.76707072349013e-05,
      "loss": 0.6290109634399415,
      "mean_token_accuracy": 0.849372398853302,
      "num_tokens": 19198585.0,
      "step": 9390
    },
    {
      "entropy": 0.6382844749838114,
      "epoch": 0.15266803634798568,
      "grad_norm": 7.0,
      "learning_rate": 4.766523886339349e-05,
      "loss": 0.6386823177337646,
      "mean_token_accuracy": 0.845276989787817,
      "num_tokens": 19218942.0,
      "step": 9400
    },
    {
      "entropy": 0.5796367482282221,
      "epoch": 0.15283044915261118,
      "grad_norm": 8.0,
      "learning_rate": 4.7659764394885495e-05,
      "loss": 0.5816111087799072,
      "mean_token_accuracy": 0.8576555110514164,
      "num_tokens": 19239570.0,
      "step": 9410
    },
    {
      "entropy": 0.614752927236259,
      "epoch": 0.15299286195723671,
      "grad_norm": 13.1875,
      "learning_rate": 4.7654283830849945e-05,
      "loss": 0.6411834716796875,
      "mean_token_accuracy": 0.8368097584694624,
      "num_tokens": 19259358.0,
      "step": 9420
    },
    {
      "entropy": 0.5999226151034236,
      "epoch": 0.15315527476186222,
      "grad_norm": 6.59375,
      "learning_rate": 4.7648797172761127e-05,
      "loss": 0.6217093467712402,
      "mean_token_accuracy": 0.8428853027522564,
      "num_tokens": 19279803.0,
      "step": 9430
    },
    {
      "entropy": 0.6397399811074138,
      "epoch": 0.15331768756648775,
      "grad_norm": 7.5625,
      "learning_rate": 4.764330442209498e-05,
      "loss": 0.6770953178405762,
      "mean_token_accuracy": 0.8353389024734497,
      "num_tokens": 19300548.0,
      "step": 9440
    },
    {
      "entropy": 0.6582180628553033,
      "epoch": 0.15348010037111326,
      "grad_norm": 8.5,
      "learning_rate": 4.763780558032905e-05,
      "loss": 0.6480880737304687,
      "mean_token_accuracy": 0.8370811097323895,
      "num_tokens": 19320983.0,
      "step": 9450
    },
    {
      "entropy": 0.6426882747560739,
      "epoch": 0.15364251317573877,
      "grad_norm": 7.46875,
      "learning_rate": 4.763230064894255e-05,
      "loss": 0.6488073348999024,
      "mean_token_accuracy": 0.8393580108880997,
      "num_tokens": 19340646.0,
      "step": 9460
    },
    {
      "entropy": 0.568936332874,
      "epoch": 0.1538049259803643,
      "grad_norm": 9.5,
      "learning_rate": 4.7626789629416305e-05,
      "loss": 0.5239686489105224,
      "mean_token_accuracy": 0.8632427386939525,
      "num_tokens": 19361509.0,
      "step": 9470
    },
    {
      "entropy": 0.5813541390933097,
      "epoch": 0.1539673387849898,
      "grad_norm": 8.75,
      "learning_rate": 4.7621272523232805e-05,
      "loss": 0.5929550647735595,
      "mean_token_accuracy": 0.8509740315377712,
      "num_tokens": 19381922.0,
      "step": 9480
    },
    {
      "entropy": 0.6374801522120833,
      "epoch": 0.15412975158961534,
      "grad_norm": 9.8125,
      "learning_rate": 4.761574933187616e-05,
      "loss": 0.6530354499816895,
      "mean_token_accuracy": 0.842949989438057,
      "num_tokens": 19402294.0,
      "step": 9490
    },
    {
      "entropy": 0.5915313007310032,
      "epoch": 0.15429216439424084,
      "grad_norm": 8.875,
      "learning_rate": 4.761022005683211e-05,
      "loss": 0.623656940460205,
      "mean_token_accuracy": 0.8449139714241027,
      "num_tokens": 19423228.0,
      "step": 9500
    },
    {
      "entropy": 0.6307830279693007,
      "epoch": 0.15445457719886635,
      "grad_norm": 9.8125,
      "learning_rate": 4.760468469958806e-05,
      "loss": 0.6370206832885742,
      "mean_token_accuracy": 0.8410321936011315,
      "num_tokens": 19443385.0,
      "step": 9510
    },
    {
      "entropy": 0.5717444282956421,
      "epoch": 0.15461699000349188,
      "grad_norm": 7.09375,
      "learning_rate": 4.7599143261633015e-05,
      "loss": 0.567040205001831,
      "mean_token_accuracy": 0.8576523587107658,
      "num_tokens": 19465376.0,
      "step": 9520
    },
    {
      "entropy": 0.5838196300901473,
      "epoch": 0.1547794028081174,
      "grad_norm": 9.75,
      "learning_rate": 4.7593595744457655e-05,
      "loss": 0.6029806137084961,
      "mean_token_accuracy": 0.8475877128541469,
      "num_tokens": 19486617.0,
      "step": 9530
    },
    {
      "entropy": 0.6519259840250016,
      "epoch": 0.15494181561274292,
      "grad_norm": 8.375,
      "learning_rate": 4.758804214955426e-05,
      "loss": 0.6817457199096679,
      "mean_token_accuracy": 0.8346240479499102,
      "num_tokens": 19508004.0,
      "step": 9540
    },
    {
      "entropy": 0.6889466105028987,
      "epoch": 0.15510422841736843,
      "grad_norm": 8.375,
      "learning_rate": 4.758248247841676e-05,
      "loss": 0.7079968452453613,
      "mean_token_accuracy": 0.826197998598218,
      "num_tokens": 19529802.0,
      "step": 9550
    },
    {
      "entropy": 0.6728490669745952,
      "epoch": 0.15526664122199393,
      "grad_norm": 8.6875,
      "learning_rate": 4.757691673254073e-05,
      "loss": 0.7061266899108887,
      "mean_token_accuracy": 0.8264325864613056,
      "num_tokens": 19550230.0,
      "step": 9560
    },
    {
      "entropy": 0.6599953419528901,
      "epoch": 0.15542905402661947,
      "grad_norm": 5.34375,
      "learning_rate": 4.757134491342335e-05,
      "loss": 0.6428809642791748,
      "mean_token_accuracy": 0.8317606687545777,
      "num_tokens": 19571868.0,
      "step": 9570
    },
    {
      "entropy": 0.6808361680712551,
      "epoch": 0.15559146683124497,
      "grad_norm": 10.25,
      "learning_rate": 4.756576702256347e-05,
      "loss": 0.6610795021057129,
      "mean_token_accuracy": 0.8325070716440678,
      "num_tokens": 19591765.0,
      "step": 9580
    },
    {
      "entropy": 0.6286493331193924,
      "epoch": 0.15575387963587048,
      "grad_norm": 7.28125,
      "learning_rate": 4.7560183061461543e-05,
      "loss": 0.5736021518707275,
      "mean_token_accuracy": 0.8557942725718022,
      "num_tokens": 19612175.0,
      "step": 9590
    },
    {
      "entropy": 0.5353074976243078,
      "epoch": 0.155916292440496,
      "grad_norm": 7.71875,
      "learning_rate": 4.7554593031619673e-05,
      "loss": 0.6169938087463379,
      "mean_token_accuracy": 0.8495191283524036,
      "num_tokens": 19632064.0,
      "step": 9600
    },
    {
      "entropy": 0.6597498061135412,
      "epoch": 0.15607870524512152,
      "grad_norm": 11.6875,
      "learning_rate": 4.754899693454159e-05,
      "loss": 0.6788104057312012,
      "mean_token_accuracy": 0.8402855955064297,
      "num_tokens": 19653128.0,
      "step": 9610
    },
    {
      "entropy": 0.6469740585424006,
      "epoch": 0.15624111804974705,
      "grad_norm": 7.65625,
      "learning_rate": 4.754339477173267e-05,
      "loss": 0.6675477504730225,
      "mean_token_accuracy": 0.8299224019050598,
      "num_tokens": 19673326.0,
      "step": 9620
    },
    {
      "entropy": 0.5688302397727967,
      "epoch": 0.15640353085437256,
      "grad_norm": 6.21875,
      "learning_rate": 4.753778654469987e-05,
      "loss": 0.5748984813690186,
      "mean_token_accuracy": 0.8528425954282284,
      "num_tokens": 19693282.0,
      "step": 9630
    },
    {
      "entropy": 0.6028886201791466,
      "epoch": 0.15656594365899806,
      "grad_norm": 7.90625,
      "learning_rate": 4.7532172254951844e-05,
      "loss": 0.6257241249084473,
      "mean_token_accuracy": 0.8434146121144295,
      "num_tokens": 19713749.0,
      "step": 9640
    },
    {
      "entropy": 0.6426247789524495,
      "epoch": 0.1567283564636236,
      "grad_norm": 7.03125,
      "learning_rate": 4.752655190399885e-05,
      "loss": 0.6677342414855957,
      "mean_token_accuracy": 0.832595519721508,
      "num_tokens": 19734272.0,
      "step": 9650
    },
    {
      "entropy": 0.6408684544265271,
      "epoch": 0.1568907692682491,
      "grad_norm": 6.4375,
      "learning_rate": 4.7520925493352764e-05,
      "loss": 0.6047965526580811,
      "mean_token_accuracy": 0.8512013837695122,
      "num_tokens": 19754999.0,
      "step": 9660
    },
    {
      "entropy": 0.6145822267513722,
      "epoch": 0.15705318207287464,
      "grad_norm": 10.0625,
      "learning_rate": 4.7515293024527116e-05,
      "loss": 0.6732779026031495,
      "mean_token_accuracy": 0.8434772111475468,
      "num_tokens": 19775900.0,
      "step": 9670
    },
    {
      "entropy": 0.6563623644411564,
      "epoch": 0.15721559487750014,
      "grad_norm": 8.3125,
      "learning_rate": 4.7509654499037034e-05,
      "loss": 0.6312705039978027,
      "mean_token_accuracy": 0.8425785824656487,
      "num_tokens": 19795762.0,
      "step": 9680
    },
    {
      "entropy": 0.5867878974415361,
      "epoch": 0.15737800768212565,
      "grad_norm": 6.78125,
      "learning_rate": 4.750400991839931e-05,
      "loss": 0.5712440013885498,
      "mean_token_accuracy": 0.8634155042469501,
      "num_tokens": 19815569.0,
      "step": 9690
    },
    {
      "entropy": 0.603073606826365,
      "epoch": 0.15754042048675118,
      "grad_norm": 4.9375,
      "learning_rate": 4.7498359284132344e-05,
      "loss": 0.5984757900238037,
      "mean_token_accuracy": 0.8530003942549229,
      "num_tokens": 19835166.0,
      "step": 9700
    },
    {
      "entropy": 0.6206632005982101,
      "epoch": 0.1577028332913767,
      "grad_norm": 7.96875,
      "learning_rate": 4.7492702597756164e-05,
      "loss": 0.6072278499603272,
      "mean_token_accuracy": 0.8495746836066246,
      "num_tokens": 19855187.0,
      "step": 9710
    },
    {
      "entropy": 0.6478312129154802,
      "epoch": 0.15786524609600222,
      "grad_norm": 6.4375,
      "learning_rate": 4.7487039860792436e-05,
      "loss": 0.6670050621032715,
      "mean_token_accuracy": 0.8348179310560226,
      "num_tokens": 19876271.0,
      "step": 9720
    },
    {
      "entropy": 0.5571336104068905,
      "epoch": 0.15802765890062773,
      "grad_norm": 7.0,
      "learning_rate": 4.748137107476446e-05,
      "loss": 0.5447845935821534,
      "mean_token_accuracy": 0.8595463387668133,
      "num_tokens": 19896238.0,
      "step": 9730
    },
    {
      "entropy": 0.6005212897434831,
      "epoch": 0.15819007170525323,
      "grad_norm": 8.5,
      "learning_rate": 4.747569624119713e-05,
      "loss": 0.6174000263214111,
      "mean_token_accuracy": 0.8469960875809193,
      "num_tokens": 19916850.0,
      "step": 9740
    },
    {
      "entropy": 0.6102871877141297,
      "epoch": 0.15835248450987877,
      "grad_norm": 7.8125,
      "learning_rate": 4.7470015361617006e-05,
      "loss": 0.6289696216583252,
      "mean_token_accuracy": 0.8428732868283987,
      "num_tokens": 19937466.0,
      "step": 9750
    },
    {
      "entropy": 0.5997185956686735,
      "epoch": 0.15851489731450427,
      "grad_norm": 7.90625,
      "learning_rate": 4.7464328437552245e-05,
      "loss": 0.6241291046142579,
      "mean_token_accuracy": 0.8472841091454029,
      "num_tokens": 19958155.0,
      "step": 9760
    },
    {
      "entropy": 0.6648338919505477,
      "epoch": 0.1586773101191298,
      "grad_norm": 11.125,
      "learning_rate": 4.745863547053265e-05,
      "loss": 0.6752971649169922,
      "mean_token_accuracy": 0.8366879358887672,
      "num_tokens": 19977324.0,
      "step": 9770
    },
    {
      "entropy": 0.5789738723542541,
      "epoch": 0.1588397229237553,
      "grad_norm": 9.1875,
      "learning_rate": 4.745293646208965e-05,
      "loss": 0.5664704322814942,
      "mean_token_accuracy": 0.8487032026052475,
      "num_tokens": 19997076.0,
      "step": 9780
    },
    {
      "entropy": 0.6300338912289589,
      "epoch": 0.15900213572838082,
      "grad_norm": 7.3125,
      "learning_rate": 4.7447231413756276e-05,
      "loss": 0.6146106719970703,
      "mean_token_accuracy": 0.8390329584479332,
      "num_tokens": 20017729.0,
      "step": 9790
    },
    {
      "entropy": 0.6204449416138231,
      "epoch": 0.15916454853300635,
      "grad_norm": 8.3125,
      "learning_rate": 4.74415203270672e-05,
      "loss": 0.6354228496551514,
      "mean_token_accuracy": 0.8435870558023453,
      "num_tokens": 20037343.0,
      "step": 9800
    },
    {
      "entropy": 0.5659003305714577,
      "epoch": 0.15932696133763186,
      "grad_norm": 6.75,
      "learning_rate": 4.7435803203558726e-05,
      "loss": 0.5645187377929688,
      "mean_token_accuracy": 0.8562339976429939,
      "num_tokens": 20059193.0,
      "step": 9810
    },
    {
      "entropy": 0.5824140844400972,
      "epoch": 0.15948937414225736,
      "grad_norm": 8.6875,
      "learning_rate": 4.7430080044768765e-05,
      "loss": 0.6013901710510254,
      "mean_token_accuracy": 0.8432653449475765,
      "num_tokens": 20080738.0,
      "step": 9820
    },
    {
      "entropy": 0.6200696306303144,
      "epoch": 0.1596517869468829,
      "grad_norm": 6.75,
      "learning_rate": 4.7424350852236876e-05,
      "loss": 0.6118596076965332,
      "mean_token_accuracy": 0.8418068371713161,
      "num_tokens": 20101094.0,
      "step": 9830
    },
    {
      "entropy": 0.5530235239304602,
      "epoch": 0.1598141997515084,
      "grad_norm": 8.375,
      "learning_rate": 4.74186156275042e-05,
      "loss": 0.5241657733917237,
      "mean_token_accuracy": 0.8719382792711258,
      "num_tokens": 20120581.0,
      "step": 9840
    },
    {
      "entropy": 0.5808429464697837,
      "epoch": 0.15997661255613393,
      "grad_norm": 7.28125,
      "learning_rate": 4.7412874372113544e-05,
      "loss": 0.5879287242889404,
      "mean_token_accuracy": 0.847440917044878,
      "num_tokens": 20142219.0,
      "step": 9850
    },
    {
      "entropy": 0.5360729384236038,
      "epoch": 0.16013902536075944,
      "grad_norm": 7.1875,
      "learning_rate": 4.7407127087609305e-05,
      "loss": 0.5218863487243652,
      "mean_token_accuracy": 0.8638460032641888,
      "num_tokens": 20163259.0,
      "step": 9860
    },
    {
      "entropy": 0.6425107331946492,
      "epoch": 0.16030143816538495,
      "grad_norm": 6.53125,
      "learning_rate": 4.740137377553753e-05,
      "loss": 0.6645955562591552,
      "mean_token_accuracy": 0.8458774819970131,
      "num_tokens": 20182185.0,
      "step": 9870
    },
    {
      "entropy": 0.5561765970662236,
      "epoch": 0.16046385097001048,
      "grad_norm": 5.96875,
      "learning_rate": 4.739561443744586e-05,
      "loss": 0.5652360439300537,
      "mean_token_accuracy": 0.8513098962604999,
      "num_tokens": 20202062.0,
      "step": 9880
    },
    {
      "entropy": 0.6039927275385708,
      "epoch": 0.16062626377463599,
      "grad_norm": 6.25,
      "learning_rate": 4.738984907488357e-05,
      "loss": 0.6235170841217041,
      "mean_token_accuracy": 0.8409020751714706,
      "num_tokens": 20222496.0,
      "step": 9890
    },
    {
      "entropy": 0.5775435818359256,
      "epoch": 0.16078867657926152,
      "grad_norm": 5.90625,
      "learning_rate": 4.7384077689401576e-05,
      "loss": 0.5950204849243164,
      "mean_token_accuracy": 0.8477249838411808,
      "num_tokens": 20242251.0,
      "step": 9900
    },
    {
      "entropy": 0.6783268143422901,
      "epoch": 0.16095108938388702,
      "grad_norm": 4.90625,
      "learning_rate": 4.737830028255237e-05,
      "loss": 0.6445770263671875,
      "mean_token_accuracy": 0.846524391695857,
      "num_tokens": 20262155.0,
      "step": 9910
    },
    {
      "entropy": 0.6511921867728233,
      "epoch": 0.16111350218851253,
      "grad_norm": 6.1875,
      "learning_rate": 4.73725168558901e-05,
      "loss": 0.6584742069244385,
      "mean_token_accuracy": 0.8270201563835144,
      "num_tokens": 20282916.0,
      "step": 9920
    },
    {
      "entropy": 0.5804347527679056,
      "epoch": 0.16127591499313806,
      "grad_norm": 7.1875,
      "learning_rate": 4.7366727410970515e-05,
      "loss": 0.5705376625061035,
      "mean_token_accuracy": 0.8526408389210701,
      "num_tokens": 20303044.0,
      "step": 9930
    },
    {
      "entropy": 0.5436915442347526,
      "epoch": 0.16143832779776357,
      "grad_norm": 8.0625,
      "learning_rate": 4.7360931949350984e-05,
      "loss": 0.5509803771972657,
      "mean_token_accuracy": 0.8655401818454266,
      "num_tokens": 20322781.0,
      "step": 9940
    },
    {
      "entropy": 0.5842553658410907,
      "epoch": 0.1616007406023891,
      "grad_norm": 6.46875,
      "learning_rate": 4.735513047259051e-05,
      "loss": 0.6214926719665528,
      "mean_token_accuracy": 0.8449900954961777,
      "num_tokens": 20343783.0,
      "step": 9950
    },
    {
      "entropy": 0.5747113563120365,
      "epoch": 0.1617631534070146,
      "grad_norm": 10.0,
      "learning_rate": 4.73493229822497e-05,
      "loss": 0.5740239143371582,
      "mean_token_accuracy": 0.8648852463811636,
      "num_tokens": 20364671.0,
      "step": 9960
    },
    {
      "entropy": 0.5857602461241186,
      "epoch": 0.16192556621164012,
      "grad_norm": 10.4375,
      "learning_rate": 4.734350947989078e-05,
      "loss": 0.6017921924591064,
      "mean_token_accuracy": 0.849933621659875,
      "num_tokens": 20385364.0,
      "step": 9970
    },
    {
      "entropy": 0.5855033164843917,
      "epoch": 0.16208797901626565,
      "grad_norm": 6.90625,
      "learning_rate": 4.733768996707759e-05,
      "loss": 0.6099676609039306,
      "mean_token_accuracy": 0.84154096506536,
      "num_tokens": 20404966.0,
      "step": 9980
    },
    {
      "entropy": 0.613588894950226,
      "epoch": 0.16225039182089115,
      "grad_norm": 11.0625,
      "learning_rate": 4.7331864445375604e-05,
      "loss": 0.673856782913208,
      "mean_token_accuracy": 0.836117721349001,
      "num_tokens": 20424758.0,
      "step": 9990
    },
    {
      "entropy": 0.6387387309223413,
      "epoch": 0.1624128046255167,
      "grad_norm": 8.8125,
      "learning_rate": 4.73260329163519e-05,
      "loss": 0.6230509757995606,
      "mean_token_accuracy": 0.8452924638986588,
      "num_tokens": 20444932.0,
      "step": 10000
    },
    {
      "entropy": 0.6292905842885375,
      "epoch": 0.1625752174301422,
      "grad_norm": 10.375,
      "learning_rate": 4.7320195381575164e-05,
      "loss": 0.6288307189941407,
      "mean_token_accuracy": 0.8423650063574314,
      "num_tokens": 20465552.0,
      "step": 10010
    },
    {
      "entropy": 0.5735433204099536,
      "epoch": 0.1627376302347677,
      "grad_norm": 8.875,
      "learning_rate": 4.73143518426157e-05,
      "loss": 0.564060640335083,
      "mean_token_accuracy": 0.8548333950340747,
      "num_tokens": 20484946.0,
      "step": 10020
    },
    {
      "entropy": 0.640919862035662,
      "epoch": 0.16290004303939323,
      "grad_norm": 6.75,
      "learning_rate": 4.730850230104545e-05,
      "loss": 0.6143383026123047,
      "mean_token_accuracy": 0.8440393216907978,
      "num_tokens": 20507176.0,
      "step": 10030
    },
    {
      "entropy": 0.6496960234828293,
      "epoch": 0.16306245584401874,
      "grad_norm": 5.90625,
      "learning_rate": 4.7302646758437945e-05,
      "loss": 0.6794010639190674,
      "mean_token_accuracy": 0.8283379197120666,
      "num_tokens": 20527671.0,
      "step": 10040
    },
    {
      "entropy": 0.6454533107578755,
      "epoch": 0.16322486864864427,
      "grad_norm": 9.0,
      "learning_rate": 4.729678521636834e-05,
      "loss": 0.6555508136749267,
      "mean_token_accuracy": 0.8326481461524964,
      "num_tokens": 20548818.0,
      "step": 10050
    },
    {
      "entropy": 0.5529184692539275,
      "epoch": 0.16338728145326978,
      "grad_norm": 8.875,
      "learning_rate": 4.7290917676413405e-05,
      "loss": 0.5706881999969482,
      "mean_token_accuracy": 0.8554768905043602,
      "num_tokens": 20568545.0,
      "step": 10060
    },
    {
      "entropy": 0.6126372271217406,
      "epoch": 0.16354969425789528,
      "grad_norm": 8.125,
      "learning_rate": 4.7285044140151514e-05,
      "loss": 0.5821037769317627,
      "mean_token_accuracy": 0.8512091115117073,
      "num_tokens": 20589653.0,
      "step": 10070
    },
    {
      "entropy": 0.5499625711236149,
      "epoch": 0.16371210706252082,
      "grad_norm": 6.21875,
      "learning_rate": 4.7279164609162664e-05,
      "loss": 0.5796096801757813,
      "mean_token_accuracy": 0.8545595355331898,
      "num_tokens": 20609503.0,
      "step": 10080
    },
    {
      "entropy": 0.5897466137073935,
      "epoch": 0.16387451986714632,
      "grad_norm": 10.125,
      "learning_rate": 4.727327908502848e-05,
      "loss": 0.5967008590698242,
      "mean_token_accuracy": 0.8452329732477665,
      "num_tokens": 20629082.0,
      "step": 10090
    },
    {
      "entropy": 0.6203928643837571,
      "epoch": 0.16403693267177183,
      "grad_norm": 9.1875,
      "learning_rate": 4.7267387569332144e-05,
      "loss": 0.6630392074584961,
      "mean_token_accuracy": 0.8498360618948937,
      "num_tokens": 20650120.0,
      "step": 10100
    },
    {
      "entropy": 0.6710710220038891,
      "epoch": 0.16419934547639736,
      "grad_norm": 10.375,
      "learning_rate": 4.7261490063658525e-05,
      "loss": 0.6594163417816162,
      "mean_token_accuracy": 0.8320409130305052,
      "num_tokens": 20670294.0,
      "step": 10110
    },
    {
      "entropy": 0.7260538349393755,
      "epoch": 0.16436175828102287,
      "grad_norm": 8.1875,
      "learning_rate": 4.7255586569594045e-05,
      "loss": 0.7361472129821778,
      "mean_token_accuracy": 0.8247964542359114,
      "num_tokens": 20690975.0,
      "step": 10120
    },
    {
      "entropy": 0.5999591907486319,
      "epoch": 0.1645241710856484,
      "grad_norm": 6.40625,
      "learning_rate": 4.724967708872676e-05,
      "loss": 0.5981400966644287,
      "mean_token_accuracy": 0.8527772448956966,
      "num_tokens": 20711690.0,
      "step": 10130
    },
    {
      "entropy": 0.5683107278309762,
      "epoch": 0.1646865838902739,
      "grad_norm": 8.625,
      "learning_rate": 4.724376162264634e-05,
      "loss": 0.5460116863250732,
      "mean_token_accuracy": 0.8584129229187966,
      "num_tokens": 20731670.0,
      "step": 10140
    },
    {
      "entropy": 0.6354150795843452,
      "epoch": 0.1648489966948994,
      "grad_norm": 7.75,
      "learning_rate": 4.723784017294405e-05,
      "loss": 0.6727088451385498,
      "mean_token_accuracy": 0.8353624798357486,
      "num_tokens": 20752352.0,
      "step": 10150
    },
    {
      "entropy": 0.6236502299085259,
      "epoch": 0.16501140949952495,
      "grad_norm": 9.3125,
      "learning_rate": 4.723191274121278e-05,
      "loss": 0.5961717128753662,
      "mean_token_accuracy": 0.8429586805403233,
      "num_tokens": 20771883.0,
      "step": 10160
    },
    {
      "entropy": 0.7304754287004471,
      "epoch": 0.16517382230415045,
      "grad_norm": 8.5,
      "learning_rate": 4.722597932904702e-05,
      "loss": 0.7665973663330078,
      "mean_token_accuracy": 0.8197622079402208,
      "num_tokens": 20792730.0,
      "step": 10170
    },
    {
      "entropy": 0.6256721806246788,
      "epoch": 0.165336235108776,
      "grad_norm": 9.125,
      "learning_rate": 4.722003993804286e-05,
      "loss": 0.6191418647766114,
      "mean_token_accuracy": 0.8493439488112926,
      "num_tokens": 20813482.0,
      "step": 10180
    },
    {
      "entropy": 0.6137599318288267,
      "epoch": 0.1654986479134015,
      "grad_norm": 8.25,
      "learning_rate": 4.721409456979803e-05,
      "loss": 0.6372104167938233,
      "mean_token_accuracy": 0.843912660330534,
      "num_tokens": 20834396.0,
      "step": 10190
    },
    {
      "entropy": 0.5676281632855534,
      "epoch": 0.165661060718027,
      "grad_norm": 5.46875,
      "learning_rate": 4.7208143225911836e-05,
      "loss": 0.5307728767395019,
      "mean_token_accuracy": 0.8583210855722427,
      "num_tokens": 20856038.0,
      "step": 10200
    },
    {
      "entropy": 0.5808394033461809,
      "epoch": 0.16582347352265253,
      "grad_norm": 7.0,
      "learning_rate": 4.720218590798521e-05,
      "loss": 0.5605245590209961,
      "mean_token_accuracy": 0.8576077461242676,
      "num_tokens": 20876577.0,
      "step": 10210
    },
    {
      "entropy": 0.5972050973214209,
      "epoch": 0.16598588632727804,
      "grad_norm": 8.625,
      "learning_rate": 4.719622261762067e-05,
      "loss": 0.592203950881958,
      "mean_token_accuracy": 0.8461033940315247,
      "num_tokens": 20896662.0,
      "step": 10220
    },
    {
      "entropy": 0.6008110371418297,
      "epoch": 0.16614829913190357,
      "grad_norm": 9.0625,
      "learning_rate": 4.719025335642236e-05,
      "loss": 0.6636116981506348,
      "mean_token_accuracy": 0.8344780251383781,
      "num_tokens": 20916892.0,
      "step": 10230
    },
    {
      "entropy": 0.6933499976992608,
      "epoch": 0.16631071193652908,
      "grad_norm": 7.5,
      "learning_rate": 4.718427812599604e-05,
      "loss": 0.67960205078125,
      "mean_token_accuracy": 0.8183411508798599,
      "num_tokens": 20936683.0,
      "step": 10240
    },
    {
      "entropy": 0.6092169320210814,
      "epoch": 0.16647312474115458,
      "grad_norm": 7.78125,
      "learning_rate": 4.717829692794904e-05,
      "loss": 0.6351668357849121,
      "mean_token_accuracy": 0.8445023156702518,
      "num_tokens": 20957178.0,
      "step": 10250
    },
    {
      "entropy": 0.6240492655895651,
      "epoch": 0.16663553754578012,
      "grad_norm": 8.1875,
      "learning_rate": 4.717230976389032e-05,
      "loss": 0.5972093105316162,
      "mean_token_accuracy": 0.8557727560400963,
      "num_tokens": 20976808.0,
      "step": 10260
    },
    {
      "entropy": 0.5924633769318461,
      "epoch": 0.16679795035040562,
      "grad_norm": 4.8125,
      "learning_rate": 4.716631663543045e-05,
      "loss": 0.575700330734253,
      "mean_token_accuracy": 0.8456524550914765,
      "num_tokens": 20998818.0,
      "step": 10270
    },
    {
      "entropy": 0.5724406103603542,
      "epoch": 0.16696036315503116,
      "grad_norm": 8.6875,
      "learning_rate": 4.7160317544181585e-05,
      "loss": 0.5980769634246826,
      "mean_token_accuracy": 0.8465342834591866,
      "num_tokens": 21019352.0,
      "step": 10280
    },
    {
      "entropy": 0.49440871328115465,
      "epoch": 0.16712277595965666,
      "grad_norm": 5.46875,
      "learning_rate": 4.715431249175751e-05,
      "loss": 0.5427539348602295,
      "mean_token_accuracy": 0.858744015544653,
      "num_tokens": 21039076.0,
      "step": 10290
    },
    {
      "entropy": 0.6056377184577286,
      "epoch": 0.16728518876428217,
      "grad_norm": 8.5625,
      "learning_rate": 4.7148301479773576e-05,
      "loss": 0.6022718906402588,
      "mean_token_accuracy": 0.8471645466983319,
      "num_tokens": 21058334.0,
      "step": 10300
    },
    {
      "entropy": 0.5987233026884496,
      "epoch": 0.1674476015689077,
      "grad_norm": 9.1875,
      "learning_rate": 4.714228450984677e-05,
      "loss": 0.6375426769256591,
      "mean_token_accuracy": 0.8418488562107086,
      "num_tokens": 21079128.0,
      "step": 10310
    },
    {
      "entropy": 0.610554994083941,
      "epoch": 0.1676100143735332,
      "grad_norm": 6.65625,
      "learning_rate": 4.7136261583595676e-05,
      "loss": 0.5745908737182617,
      "mean_token_accuracy": 0.8553778603672981,
      "num_tokens": 21098678.0,
      "step": 10320
    },
    {
      "entropy": 0.6034529755823315,
      "epoch": 0.16777242717815874,
      "grad_norm": 7.9375,
      "learning_rate": 4.713023270264045e-05,
      "loss": 0.6157416343688965,
      "mean_token_accuracy": 0.8498106732964515,
      "num_tokens": 21119826.0,
      "step": 10330
    },
    {
      "entropy": 0.5748192287981511,
      "epoch": 0.16793483998278425,
      "grad_norm": 8.3125,
      "learning_rate": 4.7124197868602916e-05,
      "loss": 0.5943433761596679,
      "mean_token_accuracy": 0.856507969275117,
      "num_tokens": 21140555.0,
      "step": 10340
    },
    {
      "entropy": 0.5567594275809824,
      "epoch": 0.16809725278740975,
      "grad_norm": 8.125,
      "learning_rate": 4.7118157083106415e-05,
      "loss": 0.5686344623565673,
      "mean_token_accuracy": 0.8534098178148269,
      "num_tokens": 21162129.0,
      "step": 10350
    },
    {
      "entropy": 0.585849080979824,
      "epoch": 0.16825966559203528,
      "grad_norm": 6.03125,
      "learning_rate": 4.711211034777596e-05,
      "loss": 0.585263442993164,
      "mean_token_accuracy": 0.8598802357912063,
      "num_tokens": 21183409.0,
      "step": 10360
    },
    {
      "entropy": 0.5705983297433704,
      "epoch": 0.1684220783966608,
      "grad_norm": 8.1875,
      "learning_rate": 4.710605766423813e-05,
      "loss": 0.5681751251220704,
      "mean_token_accuracy": 0.8627011843025685,
      "num_tokens": 21204471.0,
      "step": 10370
    },
    {
      "entropy": 0.6004733739420771,
      "epoch": 0.1685844912012863,
      "grad_norm": 8.8125,
      "learning_rate": 4.7099999034121106e-05,
      "loss": 0.5981147766113282,
      "mean_token_accuracy": 0.8471820540726185,
      "num_tokens": 21225477.0,
      "step": 10380
    },
    {
      "entropy": 0.5937882238999009,
      "epoch": 0.16874690400591183,
      "grad_norm": 9.6875,
      "learning_rate": 4.7093934459054676e-05,
      "loss": 0.6099638938903809,
      "mean_token_accuracy": 0.8439375191926957,
      "num_tokens": 21244982.0,
      "step": 10390
    },
    {
      "entropy": 0.6564030193723738,
      "epoch": 0.16890931681053734,
      "grad_norm": 8.8125,
      "learning_rate": 4.7087863940670216e-05,
      "loss": 0.7038453578948974,
      "mean_token_accuracy": 0.8276472575962543,
      "num_tokens": 21265465.0,
      "step": 10400
    },
    {
      "entropy": 0.5750477065565065,
      "epoch": 0.16907172961516287,
      "grad_norm": 7.21875,
      "learning_rate": 4.7081787480600726e-05,
      "loss": 0.5419651985168457,
      "mean_token_accuracy": 0.8617801934480667,
      "num_tokens": 21286204.0,
      "step": 10410
    },
    {
      "entropy": 0.5551675577647984,
      "epoch": 0.16923414241978837,
      "grad_norm": 7.5625,
      "learning_rate": 4.707570508048077e-05,
      "loss": 0.5645916938781739,
      "mean_token_accuracy": 0.8483758747577668,
      "num_tokens": 21306521.0,
      "step": 10420
    },
    {
      "entropy": 0.5960325178690254,
      "epoch": 0.16939655522441388,
      "grad_norm": 7.15625,
      "learning_rate": 4.706961674194654e-05,
      "loss": 0.5771548748016357,
      "mean_token_accuracy": 0.8501707784831524,
      "num_tokens": 21326752.0,
      "step": 10430
    },
    {
      "entropy": 0.6105166886933148,
      "epoch": 0.16955896802903941,
      "grad_norm": 9.5,
      "learning_rate": 4.706352246663581e-05,
      "loss": 0.6537672996520996,
      "mean_token_accuracy": 0.8330821670591831,
      "num_tokens": 21347222.0,
      "step": 10440
    },
    {
      "entropy": 0.6188543607946485,
      "epoch": 0.16972138083366492,
      "grad_norm": 7.53125,
      "learning_rate": 4.705742225618794e-05,
      "loss": 0.6583258152008057,
      "mean_token_accuracy": 0.8370934624224902,
      "num_tokens": 21366553.0,
      "step": 10450
    },
    {
      "entropy": 0.6345521989744156,
      "epoch": 0.16988379363829045,
      "grad_norm": 8.1875,
      "learning_rate": 4.7051316112243926e-05,
      "loss": 0.6178805351257324,
      "mean_token_accuracy": 0.8467279814183712,
      "num_tokens": 21387074.0,
      "step": 10460
    },
    {
      "entropy": 0.668431849218905,
      "epoch": 0.17004620644291596,
      "grad_norm": 7.25,
      "learning_rate": 4.704520403644632e-05,
      "loss": 0.6284631729125977,
      "mean_token_accuracy": 0.8482674896717072,
      "num_tokens": 21407276.0,
      "step": 10470
    },
    {
      "entropy": 0.5915845348499715,
      "epoch": 0.17020861924754146,
      "grad_norm": 8.75,
      "learning_rate": 4.703908603043928e-05,
      "loss": 0.64119873046875,
      "mean_token_accuracy": 0.8467101413756609,
      "num_tokens": 21428457.0,
      "step": 10480
    },
    {
      "entropy": 0.5231925231404603,
      "epoch": 0.170371032052167,
      "grad_norm": 8.25,
      "learning_rate": 4.703296209586857e-05,
      "loss": 0.5436205863952637,
      "mean_token_accuracy": 0.858207831531763,
      "num_tokens": 21448090.0,
      "step": 10490
    },
    {
      "entropy": 0.5917032637633384,
      "epoch": 0.1705334448567925,
      "grad_norm": 8.3125,
      "learning_rate": 4.7026832234381534e-05,
      "loss": 0.5627042770385742,
      "mean_token_accuracy": 0.8518933445215225,
      "num_tokens": 21469051.0,
      "step": 10500
    },
    {
      "entropy": 0.5583326623775065,
      "epoch": 0.17069585766141804,
      "grad_norm": 7.0625,
      "learning_rate": 4.7020696447627135e-05,
      "loss": 0.5373233795166016,
      "mean_token_accuracy": 0.8591613344848156,
      "num_tokens": 21489212.0,
      "step": 10510
    },
    {
      "entropy": 0.6639450607821346,
      "epoch": 0.17085827046604354,
      "grad_norm": 8.5,
      "learning_rate": 4.7014554737255895e-05,
      "loss": 0.7171398162841797,
      "mean_token_accuracy": 0.8222590766847133,
      "num_tokens": 21509362.0,
      "step": 10520
    },
    {
      "entropy": 0.6761670219711959,
      "epoch": 0.17102068327066905,
      "grad_norm": 9.125,
      "learning_rate": 4.700840710491996e-05,
      "loss": 0.6910423278808594,
      "mean_token_accuracy": 0.8262012612074614,
      "num_tokens": 21529451.0,
      "step": 10530
    },
    {
      "entropy": 0.6781721117906272,
      "epoch": 0.17118309607529458,
      "grad_norm": 6.46875,
      "learning_rate": 4.700225355227305e-05,
      "loss": 0.6644456863403321,
      "mean_token_accuracy": 0.8337820455431938,
      "num_tokens": 21550961.0,
      "step": 10540
    },
    {
      "entropy": 0.6340088805183768,
      "epoch": 0.1713455088799201,
      "grad_norm": 8.3125,
      "learning_rate": 4.699609408097049e-05,
      "loss": 0.6273257255554199,
      "mean_token_accuracy": 0.8420660763978958,
      "num_tokens": 21572266.0,
      "step": 10550
    },
    {
      "entropy": 0.6510842032730579,
      "epoch": 0.17150792168454562,
      "grad_norm": 9.0625,
      "learning_rate": 4.698992869266919e-05,
      "loss": 0.6351901054382324,
      "mean_token_accuracy": 0.8334381133317947,
      "num_tokens": 21593594.0,
      "step": 10560
    },
    {
      "entropy": 0.6943551875650883,
      "epoch": 0.17167033448917113,
      "grad_norm": 11.0625,
      "learning_rate": 4.698375738902765e-05,
      "loss": 0.7374504566192627,
      "mean_token_accuracy": 0.821067775785923,
      "num_tokens": 21614266.0,
      "step": 10570
    },
    {
      "entropy": 0.6836581083014608,
      "epoch": 0.17183274729379663,
      "grad_norm": 7.625,
      "learning_rate": 4.6977580171705966e-05,
      "loss": 0.6874443054199219,
      "mean_token_accuracy": 0.8279351517558098,
      "num_tokens": 21634106.0,
      "step": 10580
    },
    {
      "entropy": 0.6171251626685261,
      "epoch": 0.17199516009842217,
      "grad_norm": 7.78125,
      "learning_rate": 4.697139704236583e-05,
      "loss": 0.6435835838317872,
      "mean_token_accuracy": 0.8442153163254261,
      "num_tokens": 21654421.0,
      "step": 10590
    },
    {
      "entropy": 0.6085685432888568,
      "epoch": 0.17215757290304767,
      "grad_norm": 6.84375,
      "learning_rate": 4.69652080026705e-05,
      "loss": 0.6002433776855469,
      "mean_token_accuracy": 0.8494298763573169,
      "num_tokens": 21675967.0,
      "step": 10600
    },
    {
      "entropy": 0.5912724924273789,
      "epoch": 0.1723199857076732,
      "grad_norm": 8.8125,
      "learning_rate": 4.695901305428486e-05,
      "loss": 0.5677783966064454,
      "mean_token_accuracy": 0.8532762497663497,
      "num_tokens": 21695098.0,
      "step": 10610
    },
    {
      "entropy": 0.6128319635055959,
      "epoch": 0.1724823985122987,
      "grad_norm": 7.46875,
      "learning_rate": 4.695281219887536e-05,
      "loss": 0.6275811672210694,
      "mean_token_accuracy": 0.8459358856081962,
      "num_tokens": 21716548.0,
      "step": 10620
    },
    {
      "entropy": 0.572039258480072,
      "epoch": 0.17264481131692422,
      "grad_norm": 8.6875,
      "learning_rate": 4.694660543811005e-05,
      "loss": 0.6030188083648682,
      "mean_token_accuracy": 0.8532883554697037,
      "num_tokens": 21736454.0,
      "step": 10630
    },
    {
      "entropy": 0.6417946837842464,
      "epoch": 0.17280722412154975,
      "grad_norm": 8.8125,
      "learning_rate": 4.694039277365855e-05,
      "loss": 0.6691797256469727,
      "mean_token_accuracy": 0.8340727329254151,
      "num_tokens": 21756074.0,
      "step": 10640
    },
    {
      "entropy": 0.6199332669377327,
      "epoch": 0.17296963692617526,
      "grad_norm": 5.34375,
      "learning_rate": 4.693417420719208e-05,
      "loss": 0.5659583568572998,
      "mean_token_accuracy": 0.8639690816402436,
      "num_tokens": 21775830.0,
      "step": 10650
    },
    {
      "entropy": 0.6786440270021558,
      "epoch": 0.17313204973080076,
      "grad_norm": 8.3125,
      "learning_rate": 4.692794974038346e-05,
      "loss": 0.7215734958648682,
      "mean_token_accuracy": 0.8249361164867878,
      "num_tokens": 21796753.0,
      "step": 10660
    },
    {
      "entropy": 0.6814297997392714,
      "epoch": 0.1732944625354263,
      "grad_norm": 8.125,
      "learning_rate": 4.6921719374907076e-05,
      "loss": 0.6739514350891114,
      "mean_token_accuracy": 0.8306961540132761,
      "num_tokens": 21817382.0,
      "step": 10670
    },
    {
      "entropy": 0.6765802304260433,
      "epoch": 0.1734568753400518,
      "grad_norm": 7.59375,
      "learning_rate": 4.6915483112438916e-05,
      "loss": 0.7113444805145264,
      "mean_token_accuracy": 0.8271530359983444,
      "num_tokens": 21837949.0,
      "step": 10680
    },
    {
      "entropy": 0.6123078907839954,
      "epoch": 0.17361928814467734,
      "grad_norm": 8.8125,
      "learning_rate": 4.6909240954656544e-05,
      "loss": 0.596946907043457,
      "mean_token_accuracy": 0.8530781116336584,
      "num_tokens": 21859921.0,
      "step": 10690
    },
    {
      "entropy": 0.6887896069325506,
      "epoch": 0.17378170094930284,
      "grad_norm": 12.25,
      "learning_rate": 4.6902992903239126e-05,
      "loss": 0.6963600158691406,
      "mean_token_accuracy": 0.8260877415537834,
      "num_tokens": 21880022.0,
      "step": 10700
    },
    {
      "entropy": 0.5960656862240284,
      "epoch": 0.17394411375392835,
      "grad_norm": 8.375,
      "learning_rate": 4.6896738959867384e-05,
      "loss": 0.6011978149414062,
      "mean_token_accuracy": 0.8459673263132572,
      "num_tokens": 21901668.0,
      "step": 10710
    },
    {
      "entropy": 0.5453526728088036,
      "epoch": 0.17410652655855388,
      "grad_norm": 5.75,
      "learning_rate": 4.6890479126223654e-05,
      "loss": 0.5343749523162842,
      "mean_token_accuracy": 0.8527902200818062,
      "num_tokens": 21922822.0,
      "step": 10720
    },
    {
      "entropy": 0.6030847841873765,
      "epoch": 0.1742689393631794,
      "grad_norm": 11.0,
      "learning_rate": 4.688421340399185e-05,
      "loss": 0.6199338436126709,
      "mean_token_accuracy": 0.8487144328653813,
      "num_tokens": 21943078.0,
      "step": 10730
    },
    {
      "entropy": 0.5403317713178695,
      "epoch": 0.17443135216780492,
      "grad_norm": 10.875,
      "learning_rate": 4.687794179485745e-05,
      "loss": 0.6082958698272705,
      "mean_token_accuracy": 0.8525116287171841,
      "num_tokens": 21963036.0,
      "step": 10740
    },
    {
      "entropy": 0.6356238946318626,
      "epoch": 0.17459376497243043,
      "grad_norm": 5.75,
      "learning_rate": 4.6871664300507546e-05,
      "loss": 0.6417235851287841,
      "mean_token_accuracy": 0.847812169417739,
      "num_tokens": 21984065.0,
      "step": 10750
    },
    {
      "entropy": 0.635120618995279,
      "epoch": 0.17475617777705593,
      "grad_norm": 7.75,
      "learning_rate": 4.6865380922630786e-05,
      "loss": 0.6134832382202149,
      "mean_token_accuracy": 0.8507581070065499,
      "num_tokens": 22005273.0,
      "step": 10760
    },
    {
      "entropy": 0.6795597025193274,
      "epoch": 0.17491859058168147,
      "grad_norm": 7.09375,
      "learning_rate": 4.685909166291742e-05,
      "loss": 0.6947340965270996,
      "mean_token_accuracy": 0.8323920376598835,
      "num_tokens": 22025960.0,
      "step": 10770
    },
    {
      "entropy": 0.6110397621057928,
      "epoch": 0.17508100338630697,
      "grad_norm": 12.3125,
      "learning_rate": 4.685279652305927e-05,
      "loss": 0.638221263885498,
      "mean_token_accuracy": 0.8368886917829513,
      "num_tokens": 22049193.0,
      "step": 10780
    },
    {
      "entropy": 0.6113326961174608,
      "epoch": 0.1752434161909325,
      "grad_norm": 5.3125,
      "learning_rate": 4.684649550474975e-05,
      "loss": 0.614938497543335,
      "mean_token_accuracy": 0.8524614930152893,
      "num_tokens": 22069948.0,
      "step": 10790
    },
    {
      "entropy": 0.6313136884942651,
      "epoch": 0.175405828995558,
      "grad_norm": 6.8125,
      "learning_rate": 4.684018860968384e-05,
      "loss": 0.6065897941589355,
      "mean_token_accuracy": 0.8557863235473633,
      "num_tokens": 22090506.0,
      "step": 10800
    },
    {
      "entropy": 0.5972631795331835,
      "epoch": 0.17556824180018352,
      "grad_norm": 7.65625,
      "learning_rate": 4.68338758395581e-05,
      "loss": 0.6023406982421875,
      "mean_token_accuracy": 0.8473982945084572,
      "num_tokens": 22110431.0,
      "step": 10810
    },
    {
      "entropy": 0.5519067537970841,
      "epoch": 0.17573065460480905,
      "grad_norm": 7.6875,
      "learning_rate": 4.682755719607071e-05,
      "loss": 0.5770535945892334,
      "mean_token_accuracy": 0.8571524303406477,
      "num_tokens": 22130202.0,
      "step": 10820
    },
    {
      "entropy": 0.5809687945991755,
      "epoch": 0.17589306740943456,
      "grad_norm": 8.3125,
      "learning_rate": 4.682123268092137e-05,
      "loss": 0.6151505470275879,
      "mean_token_accuracy": 0.8523137882351876,
      "num_tokens": 22150108.0,
      "step": 10830
    },
    {
      "entropy": 0.5806112525984645,
      "epoch": 0.1760554802140601,
      "grad_norm": 7.9375,
      "learning_rate": 4.68149022958114e-05,
      "loss": 0.5730878353118897,
      "mean_token_accuracy": 0.8575144052505493,
      "num_tokens": 22169379.0,
      "step": 10840
    },
    {
      "entropy": 0.6602497113868594,
      "epoch": 0.1762178930186856,
      "grad_norm": 8.25,
      "learning_rate": 4.680856604244369e-05,
      "loss": 0.6340813159942627,
      "mean_token_accuracy": 0.8443503774702549,
      "num_tokens": 22189675.0,
      "step": 10850
    },
    {
      "entropy": 0.6732729093171657,
      "epoch": 0.1763803058233111,
      "grad_norm": 9.0,
      "learning_rate": 4.680222392252269e-05,
      "loss": 0.6872151851654053,
      "mean_token_accuracy": 0.835339380055666,
      "num_tokens": 22210795.0,
      "step": 10860
    },
    {
      "entropy": 0.5664860108867288,
      "epoch": 0.17654271862793663,
      "grad_norm": 10.1875,
      "learning_rate": 4.679587593775447e-05,
      "loss": 0.5525238990783692,
      "mean_token_accuracy": 0.8540775731205941,
      "num_tokens": 22231794.0,
      "step": 10870
    },
    {
      "entropy": 0.5827737994492054,
      "epoch": 0.17670513143256214,
      "grad_norm": 7.21875,
      "learning_rate": 4.678952208984664e-05,
      "loss": 0.5807932376861572,
      "mean_token_accuracy": 0.8526519849896431,
      "num_tokens": 22252305.0,
      "step": 10880
    },
    {
      "entropy": 0.6414047250524163,
      "epoch": 0.17686754423718765,
      "grad_norm": 8.4375,
      "learning_rate": 4.678316238050839e-05,
      "loss": 0.6418270587921142,
      "mean_token_accuracy": 0.845734903216362,
      "num_tokens": 22272523.0,
      "step": 10890
    },
    {
      "entropy": 0.6352237395010889,
      "epoch": 0.17702995704181318,
      "grad_norm": 7.84375,
      "learning_rate": 4.677679681145051e-05,
      "loss": 0.6773062705993652,
      "mean_token_accuracy": 0.8409270659089089,
      "num_tokens": 22294076.0,
      "step": 10900
    },
    {
      "entropy": 0.5979948588646948,
      "epoch": 0.17719236984643869,
      "grad_norm": 6.71875,
      "learning_rate": 4.6770425384385344e-05,
      "loss": 0.5737119197845459,
      "mean_token_accuracy": 0.8464342504739761,
      "num_tokens": 22314009.0,
      "step": 10910
    },
    {
      "entropy": 0.6170298194512724,
      "epoch": 0.17735478265106422,
      "grad_norm": 7.75,
      "learning_rate": 4.6764048101026825e-05,
      "loss": 0.6047065734863282,
      "mean_token_accuracy": 0.8543164685368538,
      "num_tokens": 22336064.0,
      "step": 10920
    },
    {
      "entropy": 0.6079199057072401,
      "epoch": 0.17751719545568972,
      "grad_norm": 7.96875,
      "learning_rate": 4.675766496309046e-05,
      "loss": 0.6036756038665771,
      "mean_token_accuracy": 0.8476140175014735,
      "num_tokens": 22355096.0,
      "step": 10930
    },
    {
      "entropy": 0.6273944883607327,
      "epoch": 0.17767960826031523,
      "grad_norm": 9.3125,
      "learning_rate": 4.675127597229332e-05,
      "loss": 0.6007763385772705,
      "mean_token_accuracy": 0.8472405135631561,
      "num_tokens": 22376721.0,
      "step": 10940
    },
    {
      "entropy": 0.6312616460025311,
      "epoch": 0.17784202106494076,
      "grad_norm": 11.875,
      "learning_rate": 4.6744881130354056e-05,
      "loss": 0.6448246955871582,
      "mean_token_accuracy": 0.8414410702884197,
      "num_tokens": 22397526.0,
      "step": 10950
    },
    {
      "entropy": 0.6069944577291608,
      "epoch": 0.17800443386956627,
      "grad_norm": 9.125,
      "learning_rate": 4.67384804389929e-05,
      "loss": 0.6356436729431152,
      "mean_token_accuracy": 0.8389561884105206,
      "num_tokens": 22418091.0,
      "step": 10960
    },
    {
      "entropy": 0.5530365534126759,
      "epoch": 0.1781668466741918,
      "grad_norm": 7.0625,
      "learning_rate": 4.673207389993166e-05,
      "loss": 0.5295453071594238,
      "mean_token_accuracy": 0.8684907328337431,
      "num_tokens": 22438771.0,
      "step": 10970
    },
    {
      "entropy": 0.5843997832387686,
      "epoch": 0.1783292594788173,
      "grad_norm": 7.71875,
      "learning_rate": 4.6725661514893696e-05,
      "loss": 0.5798361301422119,
      "mean_token_accuracy": 0.8562892414629459,
      "num_tokens": 22459254.0,
      "step": 10980
    },
    {
      "entropy": 0.5853789418935775,
      "epoch": 0.17849167228344281,
      "grad_norm": 5.5,
      "learning_rate": 4.671924328560396e-05,
      "loss": 0.6005899429321289,
      "mean_token_accuracy": 0.8469374477863312,
      "num_tokens": 22480369.0,
      "step": 10990
    },
    {
      "entropy": 0.589882192388177,
      "epoch": 0.17865408508806835,
      "grad_norm": 5.875,
      "learning_rate": 4.671281921378897e-05,
      "loss": 0.5855106353759766,
      "mean_token_accuracy": 0.8496033824980259,
      "num_tokens": 22501009.0,
      "step": 11000
    },
    {
      "entropy": 0.5323560175485909,
      "epoch": 0.17881649789269385,
      "grad_norm": 7.96875,
      "learning_rate": 4.670638930117681e-05,
      "loss": 0.5418843746185302,
      "mean_token_accuracy": 0.8670040659606457,
      "num_tokens": 22520545.0,
      "step": 11010
    },
    {
      "entropy": 0.608418776281178,
      "epoch": 0.1789789106973194,
      "grad_norm": 9.0,
      "learning_rate": 4.6699953549497156e-05,
      "loss": 0.6710271835327148,
      "mean_token_accuracy": 0.8405123308300972,
      "num_tokens": 22542378.0,
      "step": 11020
    },
    {
      "entropy": 0.5733260607346893,
      "epoch": 0.1791413235019449,
      "grad_norm": 8.4375,
      "learning_rate": 4.669351196048122e-05,
      "loss": 0.5523947715759278,
      "mean_token_accuracy": 0.8474179647862912,
      "num_tokens": 22562329.0,
      "step": 11030
    },
    {
      "entropy": 0.6337099796161055,
      "epoch": 0.1793037363065704,
      "grad_norm": 8.125,
      "learning_rate": 4.668706453586181e-05,
      "loss": 0.6196693420410156,
      "mean_token_accuracy": 0.8471325434744358,
      "num_tokens": 22584193.0,
      "step": 11040
    },
    {
      "entropy": 0.6362235917709768,
      "epoch": 0.17946614911119593,
      "grad_norm": 6.21875,
      "learning_rate": 4.668061127737332e-05,
      "loss": 0.6721042633056641,
      "mean_token_accuracy": 0.8359228141605854,
      "num_tokens": 22604041.0,
      "step": 11050
    },
    {
      "entropy": 0.6092807310633361,
      "epoch": 0.17962856191582144,
      "grad_norm": 9.125,
      "learning_rate": 4.667415218675165e-05,
      "loss": 0.5539704322814941,
      "mean_token_accuracy": 0.8553075969219208,
      "num_tokens": 22624723.0,
      "step": 11060
    },
    {
      "entropy": 0.5759239071980119,
      "epoch": 0.17979097472044697,
      "grad_norm": 12.75,
      "learning_rate": 4.666768726573435e-05,
      "loss": 0.5975768089294433,
      "mean_token_accuracy": 0.8472062081098557,
      "num_tokens": 22645678.0,
      "step": 11070
    },
    {
      "entropy": 0.6080147368367761,
      "epoch": 0.17995338752507248,
      "grad_norm": 10.25,
      "learning_rate": 4.666121651606047e-05,
      "loss": 0.6572309970855713,
      "mean_token_accuracy": 0.8465596012771129,
      "num_tokens": 22667241.0,
      "step": 11080
    },
    {
      "entropy": 0.6499949716031551,
      "epoch": 0.18011580032969798,
      "grad_norm": 11.125,
      "learning_rate": 4.665473993947067e-05,
      "loss": 0.6843333721160889,
      "mean_token_accuracy": 0.8348166979849339,
      "num_tokens": 22686897.0,
      "step": 11090
    },
    {
      "entropy": 0.6501326019875705,
      "epoch": 0.18027821313432352,
      "grad_norm": 6.84375,
      "learning_rate": 4.6648257537707155e-05,
      "loss": 0.6440104007720947,
      "mean_token_accuracy": 0.849724854528904,
      "num_tokens": 22707449.0,
      "step": 11100
    },
    {
      "entropy": 0.6061643383465707,
      "epoch": 0.18044062593894902,
      "grad_norm": 6.75,
      "learning_rate": 4.6641769312513714e-05,
      "loss": 0.5639350414276123,
      "mean_token_accuracy": 0.857370761781931,
      "num_tokens": 22727456.0,
      "step": 11110
    },
    {
      "entropy": 0.5360599214211106,
      "epoch": 0.18060303874357456,
      "grad_norm": 9.5,
      "learning_rate": 4.6635275265635694e-05,
      "loss": 0.5453049659729003,
      "mean_token_accuracy": 0.8583155244588851,
      "num_tokens": 22747733.0,
      "step": 11120
    },
    {
      "entropy": 0.622758490499109,
      "epoch": 0.18076545154820006,
      "grad_norm": 10.875,
      "learning_rate": 4.662877539882001e-05,
      "loss": 0.6537322521209716,
      "mean_token_accuracy": 0.8441423065960407,
      "num_tokens": 22767066.0,
      "step": 11130
    },
    {
      "entropy": 0.5657371178735048,
      "epoch": 0.18092786435282557,
      "grad_norm": 9.6875,
      "learning_rate": 4.662226971381513e-05,
      "loss": 0.5762892246246338,
      "mean_token_accuracy": 0.8549931004643441,
      "num_tokens": 22786607.0,
      "step": 11140
    },
    {
      "entropy": 0.5394769676961004,
      "epoch": 0.1810902771574511,
      "grad_norm": 8.125,
      "learning_rate": 4.66157582123711e-05,
      "loss": 0.5325759410858154,
      "mean_token_accuracy": 0.8651075452566147,
      "num_tokens": 22806453.0,
      "step": 11150
    },
    {
      "entropy": 0.6356092429719865,
      "epoch": 0.1812526899620766,
      "grad_norm": 6.78125,
      "learning_rate": 4.6609240896239534e-05,
      "loss": 0.637923002243042,
      "mean_token_accuracy": 0.8396657794713974,
      "num_tokens": 22826619.0,
      "step": 11160
    },
    {
      "entropy": 0.6450096301734447,
      "epoch": 0.1814151027667021,
      "grad_norm": 6.78125,
      "learning_rate": 4.660271776717361e-05,
      "loss": 0.6121939182281494,
      "mean_token_accuracy": 0.8498387351632118,
      "num_tokens": 22846502.0,
      "step": 11170
    },
    {
      "entropy": 0.6279677834361792,
      "epoch": 0.18157751557132765,
      "grad_norm": 7.875,
      "learning_rate": 4.659618882692805e-05,
      "loss": 0.6552072525024414,
      "mean_token_accuracy": 0.8420822951942682,
      "num_tokens": 22866502.0,
      "step": 11180
    },
    {
      "entropy": 0.5738108268007636,
      "epoch": 0.18173992837595315,
      "grad_norm": 8.4375,
      "learning_rate": 4.658965407725916e-05,
      "loss": 0.5637845993041992,
      "mean_token_accuracy": 0.845904940366745,
      "num_tokens": 22887661.0,
      "step": 11190
    },
    {
      "entropy": 0.518202861584723,
      "epoch": 0.18190234118057869,
      "grad_norm": 9.25,
      "learning_rate": 4.6583113519924815e-05,
      "loss": 0.5322105884552002,
      "mean_token_accuracy": 0.8660988800227643,
      "num_tokens": 22907126.0,
      "step": 11200
    },
    {
      "entropy": 0.6461315168417059,
      "epoch": 0.1820647539852042,
      "grad_norm": 5.53125,
      "learning_rate": 4.657656715668442e-05,
      "loss": 0.676042127609253,
      "mean_token_accuracy": 0.8293530806899071,
      "num_tokens": 22929119.0,
      "step": 11210
    },
    {
      "entropy": 0.6140487620607018,
      "epoch": 0.1822271667898297,
      "grad_norm": 7.21875,
      "learning_rate": 4.657001498929897e-05,
      "loss": 0.6162654399871826,
      "mean_token_accuracy": 0.8537043545395135,
      "num_tokens": 22950175.0,
      "step": 11220
    },
    {
      "entropy": 0.5411909286864102,
      "epoch": 0.18238957959445523,
      "grad_norm": 7.1875,
      "learning_rate": 4.656345701953101e-05,
      "loss": 0.5150916099548339,
      "mean_token_accuracy": 0.8698645062744618,
      "num_tokens": 22971493.0,
      "step": 11230
    },
    {
      "entropy": 0.6187210272997617,
      "epoch": 0.18255199239908074,
      "grad_norm": 9.0625,
      "learning_rate": 4.6556893249144654e-05,
      "loss": 0.6413628101348877,
      "mean_token_accuracy": 0.8356572844088077,
      "num_tokens": 22991445.0,
      "step": 11240
    },
    {
      "entropy": 0.5899793728254735,
      "epoch": 0.18271440520370627,
      "grad_norm": 10.4375,
      "learning_rate": 4.655032367990557e-05,
      "loss": 0.6310328960418701,
      "mean_token_accuracy": 0.840613280236721,
      "num_tokens": 23010450.0,
      "step": 11250
    },
    {
      "entropy": 0.6489702215418219,
      "epoch": 0.18287681800833178,
      "grad_norm": 9.9375,
      "learning_rate": 4.654374831358098e-05,
      "loss": 0.6480605602264404,
      "mean_token_accuracy": 0.8406102821230889,
      "num_tokens": 23032489.0,
      "step": 11260
    },
    {
      "entropy": 0.5562030443921685,
      "epoch": 0.18303923081295728,
      "grad_norm": 7.1875,
      "learning_rate": 4.653716715193968e-05,
      "loss": 0.6100626945495605,
      "mean_token_accuracy": 0.8498644292354584,
      "num_tokens": 23051978.0,
      "step": 11270
    },
    {
      "entropy": 0.5883121547289193,
      "epoch": 0.18320164361758282,
      "grad_norm": 8.0,
      "learning_rate": 4.6530580196752015e-05,
      "loss": 0.5543926239013672,
      "mean_token_accuracy": 0.8623779296875,
      "num_tokens": 23071896.0,
      "step": 11280
    },
    {
      "entropy": 0.5047340583987534,
      "epoch": 0.18336405642220832,
      "grad_norm": 4.96875,
      "learning_rate": 4.6523987449789885e-05,
      "loss": 0.5033855438232422,
      "mean_token_accuracy": 0.8689239278435708,
      "num_tokens": 23093232.0,
      "step": 11290
    },
    {
      "entropy": 0.5638967530801893,
      "epoch": 0.18352646922683385,
      "grad_norm": 9.375,
      "learning_rate": 4.6517388912826763e-05,
      "loss": 0.5758753776550293,
      "mean_token_accuracy": 0.8549058258533477,
      "num_tokens": 23113377.0,
      "step": 11300
    },
    {
      "entropy": 0.5715846370905637,
      "epoch": 0.18368888203145936,
      "grad_norm": 8.625,
      "learning_rate": 4.651078458763767e-05,
      "loss": 0.6035704135894775,
      "mean_token_accuracy": 0.8475138381123543,
      "num_tokens": 23135211.0,
      "step": 11310
    },
    {
      "entropy": 0.5947141963988543,
      "epoch": 0.18385129483608487,
      "grad_norm": 8.0625,
      "learning_rate": 4.6504174475999174e-05,
      "loss": 0.6079383850097656,
      "mean_token_accuracy": 0.8471342951059342,
      "num_tokens": 23156141.0,
      "step": 11320
    },
    {
      "entropy": 0.6095105501823127,
      "epoch": 0.1840137076407104,
      "grad_norm": 6.6875,
      "learning_rate": 4.649755857968942e-05,
      "loss": 0.608941650390625,
      "mean_token_accuracy": 0.8421090051531792,
      "num_tokens": 23176547.0,
      "step": 11330
    },
    {
      "entropy": 0.5989863969385624,
      "epoch": 0.1841761204453359,
      "grad_norm": 6.625,
      "learning_rate": 4.6490936900488096e-05,
      "loss": 0.6429137706756591,
      "mean_token_accuracy": 0.8409624509513378,
      "num_tokens": 23196869.0,
      "step": 11340
    },
    {
      "entropy": 0.5538722459226847,
      "epoch": 0.18433853324996144,
      "grad_norm": 5.59375,
      "learning_rate": 4.648430944017644e-05,
      "loss": 0.5704489707946777,
      "mean_token_accuracy": 0.8609937012195588,
      "num_tokens": 23217310.0,
      "step": 11350
    },
    {
      "entropy": 0.6208458315581084,
      "epoch": 0.18450094605458695,
      "grad_norm": 8.6875,
      "learning_rate": 4.647767620053727e-05,
      "loss": 0.6122315406799317,
      "mean_token_accuracy": 0.8450802378356457,
      "num_tokens": 23237880.0,
      "step": 11360
    },
    {
      "entropy": 0.5913416612893343,
      "epoch": 0.18466335885921245,
      "grad_norm": 5.15625,
      "learning_rate": 4.6471037183354936e-05,
      "loss": 0.5684560775756836,
      "mean_token_accuracy": 0.8536857448518276,
      "num_tokens": 23258935.0,
      "step": 11370
    },
    {
      "entropy": 0.6448482762090861,
      "epoch": 0.18482577166383798,
      "grad_norm": 6.53125,
      "learning_rate": 4.646439239041534e-05,
      "loss": 0.6638885974884033,
      "mean_token_accuracy": 0.8385859522968531,
      "num_tokens": 23280937.0,
      "step": 11380
    },
    {
      "entropy": 0.6316696994006634,
      "epoch": 0.1849881844684635,
      "grad_norm": 8.125,
      "learning_rate": 4.6457741823505956e-05,
      "loss": 0.6612439155578613,
      "mean_token_accuracy": 0.8354389876127243,
      "num_tokens": 23302581.0,
      "step": 11390
    },
    {
      "entropy": 0.6395163542591036,
      "epoch": 0.18515059727308902,
      "grad_norm": 7.25,
      "learning_rate": 4.64510854844158e-05,
      "loss": 0.6510869026184082,
      "mean_token_accuracy": 0.8346766863018275,
      "num_tokens": 23323339.0,
      "step": 11400
    },
    {
      "entropy": 0.5586938069202005,
      "epoch": 0.18531301007771453,
      "grad_norm": 7.46875,
      "learning_rate": 4.6444423374935436e-05,
      "loss": 0.507680606842041,
      "mean_token_accuracy": 0.8696677081286908,
      "num_tokens": 23343155.0,
      "step": 11410
    },
    {
      "entropy": 0.5520674834027887,
      "epoch": 0.18547542288234004,
      "grad_norm": 8.8125,
      "learning_rate": 4.643775549685699e-05,
      "loss": 0.5349523067474365,
      "mean_token_accuracy": 0.8678378507494926,
      "num_tokens": 23363569.0,
      "step": 11420
    },
    {
      "entropy": 0.6324312012642622,
      "epoch": 0.18563783568696557,
      "grad_norm": 8.5,
      "learning_rate": 4.643108185197414e-05,
      "loss": 0.6364957332611084,
      "mean_token_accuracy": 0.8420477129518986,
      "num_tokens": 23383718.0,
      "step": 11430
    },
    {
      "entropy": 0.5789354477077723,
      "epoch": 0.18580024849159107,
      "grad_norm": 8.875,
      "learning_rate": 4.64244024420821e-05,
      "loss": 0.6040971279144287,
      "mean_token_accuracy": 0.8514778058975935,
      "num_tokens": 23403636.0,
      "step": 11440
    },
    {
      "entropy": 0.5714084260165692,
      "epoch": 0.18596266129621658,
      "grad_norm": 6.625,
      "learning_rate": 4.6417717268977665e-05,
      "loss": 0.5539509296417237,
      "mean_token_accuracy": 0.8528409779071808,
      "num_tokens": 23425599.0,
      "step": 11450
    },
    {
      "entropy": 0.6188283680006862,
      "epoch": 0.18612507410084211,
      "grad_norm": 10.625,
      "learning_rate": 4.641102633445915e-05,
      "loss": 0.6770379066467285,
      "mean_token_accuracy": 0.8255448900163174,
      "num_tokens": 23446559.0,
      "step": 11460
    },
    {
      "entropy": 0.5487773073837161,
      "epoch": 0.18628748690546762,
      "grad_norm": 4.8125,
      "learning_rate": 4.640432964032644e-05,
      "loss": 0.5427996158599854,
      "mean_token_accuracy": 0.8605930663645267,
      "num_tokens": 23466608.0,
      "step": 11470
    },
    {
      "entropy": 0.5994386035948992,
      "epoch": 0.18644989971009315,
      "grad_norm": 9.3125,
      "learning_rate": 4.6397627188380946e-05,
      "loss": 0.6010112762451172,
      "mean_token_accuracy": 0.8474491968750953,
      "num_tokens": 23486326.0,
      "step": 11480
    },
    {
      "entropy": 0.5763688830658793,
      "epoch": 0.18661231251471866,
      "grad_norm": 8.5625,
      "learning_rate": 4.6390918980425647e-05,
      "loss": 0.5786778926849365,
      "mean_token_accuracy": 0.8502227276563644,
      "num_tokens": 23505775.0,
      "step": 11490
    },
    {
      "entropy": 0.6072590574622154,
      "epoch": 0.18677472531934416,
      "grad_norm": 7.65625,
      "learning_rate": 4.638420501826507e-05,
      "loss": 0.5698222160339356,
      "mean_token_accuracy": 0.8524160712957383,
      "num_tokens": 23525161.0,
      "step": 11500
    },
    {
      "entropy": 0.5710085694678128,
      "epoch": 0.1869371381239697,
      "grad_norm": 6.6875,
      "learning_rate": 4.637748530370529e-05,
      "loss": 0.6042403697967529,
      "mean_token_accuracy": 0.8427183195948601,
      "num_tokens": 23544901.0,
      "step": 11510
    },
    {
      "entropy": 0.6102345215156675,
      "epoch": 0.1870995509285952,
      "grad_norm": 6.59375,
      "learning_rate": 4.637075983855391e-05,
      "loss": 0.6511147499084473,
      "mean_token_accuracy": 0.8361846413463354,
      "num_tokens": 23564850.0,
      "step": 11520
    },
    {
      "entropy": 0.6077489078976214,
      "epoch": 0.18726196373322074,
      "grad_norm": 6.84375,
      "learning_rate": 4.6364028624620115e-05,
      "loss": 0.6187078952789307,
      "mean_token_accuracy": 0.845134174823761,
      "num_tokens": 23585786.0,
      "step": 11530
    },
    {
      "entropy": 0.5890426313504576,
      "epoch": 0.18742437653784624,
      "grad_norm": 8.0625,
      "learning_rate": 4.635729166371461e-05,
      "loss": 0.5800262451171875,
      "mean_token_accuracy": 0.8502888932824135,
      "num_tokens": 23605510.0,
      "step": 11540
    },
    {
      "entropy": 0.572234072163701,
      "epoch": 0.18758678934247175,
      "grad_norm": 5.8125,
      "learning_rate": 4.6350548957649644e-05,
      "loss": 0.5232526779174804,
      "mean_token_accuracy": 0.861703735589981,
      "num_tokens": 23624877.0,
      "step": 11550
    },
    {
      "entropy": 0.6060842154547572,
      "epoch": 0.18774920214709728,
      "grad_norm": 7.1875,
      "learning_rate": 4.634380050823902e-05,
      "loss": 0.6746888160705566,
      "mean_token_accuracy": 0.8327226176857948,
      "num_tokens": 23645022.0,
      "step": 11560
    },
    {
      "entropy": 0.5589473262429238,
      "epoch": 0.1879116149517228,
      "grad_norm": 5.78125,
      "learning_rate": 4.633704631729811e-05,
      "loss": 0.5405016422271729,
      "mean_token_accuracy": 0.8564403526484966,
      "num_tokens": 23664351.0,
      "step": 11570
    },
    {
      "entropy": 0.5991588875651359,
      "epoch": 0.18807402775634832,
      "grad_norm": 7.4375,
      "learning_rate": 4.6330286386643775e-05,
      "loss": 0.6078540325164795,
      "mean_token_accuracy": 0.8473441913723946,
      "num_tokens": 23685274.0,
      "step": 11580
    },
    {
      "entropy": 0.5984087789431214,
      "epoch": 0.18823644056097383,
      "grad_norm": 8.25,
      "learning_rate": 4.632352071809446e-05,
      "loss": 0.6191217422485351,
      "mean_token_accuracy": 0.838365388661623,
      "num_tokens": 23705371.0,
      "step": 11590
    },
    {
      "entropy": 0.6619485140778124,
      "epoch": 0.18839885336559933,
      "grad_norm": 5.71875,
      "learning_rate": 4.631674931347017e-05,
      "loss": 0.6891249179840088,
      "mean_token_accuracy": 0.8329481985419989,
      "num_tokens": 23726303.0,
      "step": 11600
    },
    {
      "entropy": 0.5806306603364646,
      "epoch": 0.18856126617022487,
      "grad_norm": 7.15625,
      "learning_rate": 4.630997217459239e-05,
      "loss": 0.5531257152557373,
      "mean_token_accuracy": 0.8558954007923603,
      "num_tokens": 23745604.0,
      "step": 11610
    },
    {
      "entropy": 0.5989479161798954,
      "epoch": 0.18872367897485037,
      "grad_norm": 8.5625,
      "learning_rate": 4.630318930328421e-05,
      "loss": 0.561142873764038,
      "mean_token_accuracy": 0.8608078248798847,
      "num_tokens": 23765378.0,
      "step": 11620
    },
    {
      "entropy": 0.5502746394835413,
      "epoch": 0.1888860917794759,
      "grad_norm": 9.3125,
      "learning_rate": 4.629640070137023e-05,
      "loss": 0.6189996719360351,
      "mean_token_accuracy": 0.8484484314918518,
      "num_tokens": 23786104.0,
      "step": 11630
    },
    {
      "entropy": 0.6045003263279796,
      "epoch": 0.1890485045841014,
      "grad_norm": 8.375,
      "learning_rate": 4.62896063706766e-05,
      "loss": 0.639933967590332,
      "mean_token_accuracy": 0.8455297954380512,
      "num_tokens": 23807812.0,
      "step": 11640
    },
    {
      "entropy": 0.5947369188070297,
      "epoch": 0.18921091738872692,
      "grad_norm": 7.3125,
      "learning_rate": 4.6282806313031014e-05,
      "loss": 0.6229263782501221,
      "mean_token_accuracy": 0.84931422136724,
      "num_tokens": 23828831.0,
      "step": 11650
    },
    {
      "entropy": 0.5550688760355115,
      "epoch": 0.18937333019335245,
      "grad_norm": 7.34375,
      "learning_rate": 4.62760005302627e-05,
      "loss": 0.5538161277770997,
      "mean_token_accuracy": 0.8545195765793323,
      "num_tokens": 23849000.0,
      "step": 11660
    },
    {
      "entropy": 0.5729357229545713,
      "epoch": 0.18953574299797796,
      "grad_norm": 10.375,
      "learning_rate": 4.626918902420243e-05,
      "loss": 0.5721407890319824,
      "mean_token_accuracy": 0.8482706271111965,
      "num_tokens": 23871039.0,
      "step": 11670
    },
    {
      "entropy": 0.5566247878596187,
      "epoch": 0.1896981558026035,
      "grad_norm": 10.25,
      "learning_rate": 4.626237179668251e-05,
      "loss": 0.5426021099090577,
      "mean_token_accuracy": 0.855032379925251,
      "num_tokens": 23890811.0,
      "step": 11680
    },
    {
      "entropy": 0.5885805001482367,
      "epoch": 0.189860568607229,
      "grad_norm": 7.25,
      "learning_rate": 4.625554884953679e-05,
      "loss": 0.5761745929718017,
      "mean_token_accuracy": 0.8414988122880459,
      "num_tokens": 23911690.0,
      "step": 11690
    },
    {
      "entropy": 0.5638963340781629,
      "epoch": 0.1900229814118545,
      "grad_norm": 7.71875,
      "learning_rate": 4.6248720184600664e-05,
      "loss": 0.5529824256896972,
      "mean_token_accuracy": 0.8537318170070648,
      "num_tokens": 23932474.0,
      "step": 11700
    },
    {
      "entropy": 0.7039069826714694,
      "epoch": 0.19018539421648004,
      "grad_norm": 7.84375,
      "learning_rate": 4.6241885803711055e-05,
      "loss": 0.7582479000091553,
      "mean_token_accuracy": 0.8250743001699448,
      "num_tokens": 23953149.0,
      "step": 11710
    },
    {
      "entropy": 0.6131513524800539,
      "epoch": 0.19034780702110554,
      "grad_norm": 8.8125,
      "learning_rate": 4.6235045708706426e-05,
      "loss": 0.6371394157409668,
      "mean_token_accuracy": 0.841522041708231,
      "num_tokens": 23973637.0,
      "step": 11720
    },
    {
      "entropy": 0.7674106302438304,
      "epoch": 0.19051021982573105,
      "grad_norm": 10.9375,
      "learning_rate": 4.622819990142678e-05,
      "loss": 0.7631343841552735,
      "mean_token_accuracy": 0.8224827516824007,
      "num_tokens": 23993107.0,
      "step": 11730
    },
    {
      "entropy": 0.6727558286860585,
      "epoch": 0.19067263263035658,
      "grad_norm": 6.875,
      "learning_rate": 4.6221348383713645e-05,
      "loss": 0.6429602146148682,
      "mean_token_accuracy": 0.8432405613362789,
      "num_tokens": 24013651.0,
      "step": 11740
    },
    {
      "entropy": 0.6304172269999981,
      "epoch": 0.1908350454349821,
      "grad_norm": 8.25,
      "learning_rate": 4.6214491157410114e-05,
      "loss": 0.6969801902770996,
      "mean_token_accuracy": 0.8280431032180786,
      "num_tokens": 24033391.0,
      "step": 11750
    },
    {
      "entropy": 0.5995217938907444,
      "epoch": 0.19099745823960762,
      "grad_norm": 10.0625,
      "learning_rate": 4.6207628224360784e-05,
      "loss": 0.6132400512695313,
      "mean_token_accuracy": 0.8412525109946728,
      "num_tokens": 24053390.0,
      "step": 11760
    },
    {
      "entropy": 0.6190703211352229,
      "epoch": 0.19115987104423313,
      "grad_norm": 7.9375,
      "learning_rate": 4.620075958641179e-05,
      "loss": 0.6451432704925537,
      "mean_token_accuracy": 0.8455492798238993,
      "num_tokens": 24073824.0,
      "step": 11770
    },
    {
      "entropy": 0.6562349705956876,
      "epoch": 0.19132228384885863,
      "grad_norm": 7.59375,
      "learning_rate": 4.6193885245410826e-05,
      "loss": 0.6470342636108398,
      "mean_token_accuracy": 0.8407332591712475,
      "num_tokens": 24094013.0,
      "step": 11780
    },
    {
      "entropy": 0.5651896812021733,
      "epoch": 0.19148469665348417,
      "grad_norm": 8.0,
      "learning_rate": 4.618700520320711e-05,
      "loss": 0.5558263301849365,
      "mean_token_accuracy": 0.8527702510356903,
      "num_tokens": 24113113.0,
      "step": 11790
    },
    {
      "entropy": 0.7047504777088761,
      "epoch": 0.19164710945810967,
      "grad_norm": 5.75,
      "learning_rate": 4.6180119461651375e-05,
      "loss": 0.6877196788787842,
      "mean_token_accuracy": 0.8346897013485431,
      "num_tokens": 24134457.0,
      "step": 11800
    },
    {
      "entropy": 0.4878023267257959,
      "epoch": 0.1918095222627352,
      "grad_norm": 8.5,
      "learning_rate": 4.61732280225959e-05,
      "loss": 0.4870959758758545,
      "mean_token_accuracy": 0.8794792555272579,
      "num_tokens": 24154058.0,
      "step": 11810
    },
    {
      "entropy": 0.5612718896009028,
      "epoch": 0.1919719350673607,
      "grad_norm": 6.71875,
      "learning_rate": 4.61663308878945e-05,
      "loss": 0.5664602279663086,
      "mean_token_accuracy": 0.8527847126126289,
      "num_tokens": 24175371.0,
      "step": 11820
    },
    {
      "entropy": 0.5913427806459367,
      "epoch": 0.19213434787198622,
      "grad_norm": 9.8125,
      "learning_rate": 4.6159428059402536e-05,
      "loss": 0.612132215499878,
      "mean_token_accuracy": 0.8464497491717339,
      "num_tokens": 24195440.0,
      "step": 11830
    },
    {
      "entropy": 0.622787051834166,
      "epoch": 0.19229676067661175,
      "grad_norm": 6.03125,
      "learning_rate": 4.615251953897687e-05,
      "loss": 0.5943440437316895,
      "mean_token_accuracy": 0.8490841206163168,
      "num_tokens": 24216256.0,
      "step": 11840
    },
    {
      "entropy": 0.6061259699054062,
      "epoch": 0.19245917348123726,
      "grad_norm": 12.75,
      "learning_rate": 4.61456053284759e-05,
      "loss": 0.62724928855896,
      "mean_token_accuracy": 0.8399424858391284,
      "num_tokens": 24237488.0,
      "step": 11850
    },
    {
      "entropy": 0.5865895381197334,
      "epoch": 0.1926215862858628,
      "grad_norm": 9.4375,
      "learning_rate": 4.613868542975958e-05,
      "loss": 0.6027060508728027,
      "mean_token_accuracy": 0.8473015904426575,
      "num_tokens": 24259264.0,
      "step": 11860
    },
    {
      "entropy": 0.621085929311812,
      "epoch": 0.1927839990904883,
      "grad_norm": 6.75,
      "learning_rate": 4.613175984468937e-05,
      "loss": 0.6310685157775879,
      "mean_token_accuracy": 0.832755483686924,
      "num_tokens": 24281295.0,
      "step": 11870
    },
    {
      "entropy": 0.6070008102804423,
      "epoch": 0.1929464118951138,
      "grad_norm": 9.0,
      "learning_rate": 4.6124828575128276e-05,
      "loss": 0.5890292644500732,
      "mean_token_accuracy": 0.8455547466874123,
      "num_tokens": 24300892.0,
      "step": 11880
    },
    {
      "entropy": 0.5771111030131578,
      "epoch": 0.19310882469973933,
      "grad_norm": 8.1875,
      "learning_rate": 4.611789162294082e-05,
      "loss": 0.6208341121673584,
      "mean_token_accuracy": 0.8475520819425583,
      "num_tokens": 24320874.0,
      "step": 11890
    },
    {
      "entropy": 0.6597407171502709,
      "epoch": 0.19327123750436484,
      "grad_norm": 8.25,
      "learning_rate": 4.611094898999305e-05,
      "loss": 0.6532712936401367,
      "mean_token_accuracy": 0.8465806555002928,
      "num_tokens": 24341321.0,
      "step": 11900
    },
    {
      "entropy": 0.5456974997185171,
      "epoch": 0.19343365030899037,
      "grad_norm": 8.5625,
      "learning_rate": 4.610400067815257e-05,
      "loss": 0.5554890155792236,
      "mean_token_accuracy": 0.8589973010122776,
      "num_tokens": 24360405.0,
      "step": 11910
    },
    {
      "entropy": 0.6638603099156171,
      "epoch": 0.19359606311361588,
      "grad_norm": 6.3125,
      "learning_rate": 4.609704668928848e-05,
      "loss": 0.661817979812622,
      "mean_token_accuracy": 0.8369853772222996,
      "num_tokens": 24380332.0,
      "step": 11920
    },
    {
      "entropy": 0.614758375287056,
      "epoch": 0.19375847591824139,
      "grad_norm": 9.875,
      "learning_rate": 4.60900870252714e-05,
      "loss": 0.6699307918548584,
      "mean_token_accuracy": 0.8342101223766804,
      "num_tokens": 24402414.0,
      "step": 11930
    },
    {
      "entropy": 0.7106262544635683,
      "epoch": 0.19392088872286692,
      "grad_norm": 10.5625,
      "learning_rate": 4.608312168797353e-05,
      "loss": 0.7082159519195557,
      "mean_token_accuracy": 0.832688856124878,
      "num_tokens": 24423182.0,
      "step": 11940
    },
    {
      "entropy": 0.650323319947347,
      "epoch": 0.19408330152749242,
      "grad_norm": 7.09375,
      "learning_rate": 4.607615067926854e-05,
      "loss": 0.5938681125640869,
      "mean_token_accuracy": 0.8429953418672085,
      "num_tokens": 24443395.0,
      "step": 11950
    },
    {
      "entropy": 0.6124298573471606,
      "epoch": 0.19424571433211796,
      "grad_norm": 7.875,
      "learning_rate": 4.6069174001031644e-05,
      "loss": 0.6113751888275146,
      "mean_token_accuracy": 0.8444213792681694,
      "num_tokens": 24464007.0,
      "step": 11960
    },
    {
      "entropy": 0.6030102519318461,
      "epoch": 0.19440812713674346,
      "grad_norm": 5.96875,
      "learning_rate": 4.60621916551396e-05,
      "loss": 0.601668119430542,
      "mean_token_accuracy": 0.8506322573870421,
      "num_tokens": 24484337.0,
      "step": 11970
    },
    {
      "entropy": 0.6361741611734033,
      "epoch": 0.19457053994136897,
      "grad_norm": 6.625,
      "learning_rate": 4.605520364347067e-05,
      "loss": 0.6812771797180176,
      "mean_token_accuracy": 0.8322399992495775,
      "num_tokens": 24505336.0,
      "step": 11980
    },
    {
      "entropy": 0.5935441061388701,
      "epoch": 0.1947329527459945,
      "grad_norm": 8.75,
      "learning_rate": 4.604820996790463e-05,
      "loss": 0.6081809997558594,
      "mean_token_accuracy": 0.8515283331274986,
      "num_tokens": 24525417.0,
      "step": 11990
    },
    {
      "epoch": 0.19489536555062,
      "eval_entropy": 0.6352745468616485,
      "eval_loss": 0.6333270072937012,
      "eval_mean_token_accuracy": 0.8394083044528962,
      "eval_num_tokens": 24545303.0,
      "eval_runtime": 40.3085,
      "eval_samples_per_second": 49.617,
      "eval_steps_per_second": 6.202,
      "step": 12000
    },
    {
      "entropy": 0.5842212397838011,
      "epoch": 0.19505777835524551,
      "grad_norm": 6.84375,
      "learning_rate": 4.6034205632608055e-05,
      "loss": 0.582798957824707,
      "mean_token_accuracy": 0.8542703567072749,
      "num_tokens": 24566632.0,
      "step": 12010
    },
    {
      "entropy": 0.5612338077276945,
      "epoch": 0.19522019115987105,
      "grad_norm": 7.3125,
      "learning_rate": 4.60271949766447e-05,
      "loss": 0.5664519309997559,
      "mean_token_accuracy": 0.850143076479435,
      "num_tokens": 24586892.0,
      "step": 12020
    },
    {
      "entropy": 0.6560438976623117,
      "epoch": 0.19538260396449655,
      "grad_norm": 8.8125,
      "learning_rate": 4.602017866431865e-05,
      "loss": 0.6273918151855469,
      "mean_token_accuracy": 0.8384039390832185,
      "num_tokens": 24607205.0,
      "step": 12030
    },
    {
      "entropy": 0.5763902154751122,
      "epoch": 0.1955450167691221,
      "grad_norm": 7.1875,
      "learning_rate": 4.601315669751729e-05,
      "loss": 0.5894418239593506,
      "mean_token_accuracy": 0.8517339989542961,
      "num_tokens": 24626219.0,
      "step": 12040
    },
    {
      "entropy": 0.6443814169615507,
      "epoch": 0.1957074295737476,
      "grad_norm": 8.375,
      "learning_rate": 4.600612907812956e-05,
      "loss": 0.6380340099334717,
      "mean_token_accuracy": 0.8465047001838684,
      "num_tokens": 24645825.0,
      "step": 12050
    },
    {
      "entropy": 0.6643472737632692,
      "epoch": 0.1958698423783731,
      "grad_norm": 7.3125,
      "learning_rate": 4.599909580804589e-05,
      "loss": 0.7097051620483399,
      "mean_token_accuracy": 0.8234158299863339,
      "num_tokens": 24665422.0,
      "step": 12060
    },
    {
      "entropy": 0.5927482888102531,
      "epoch": 0.19603225518299863,
      "grad_norm": 7.65625,
      "learning_rate": 4.599205688915826e-05,
      "loss": 0.5924841880798339,
      "mean_token_accuracy": 0.8567998267710208,
      "num_tokens": 24687497.0,
      "step": 12070
    },
    {
      "entropy": 0.6623673619702458,
      "epoch": 0.19619466798762414,
      "grad_norm": 6.5,
      "learning_rate": 4.5985012323360154e-05,
      "loss": 0.6647514343261719,
      "mean_token_accuracy": 0.8347026467323303,
      "num_tokens": 24709193.0,
      "step": 12080
    },
    {
      "entropy": 0.5642630288377404,
      "epoch": 0.19635708079224967,
      "grad_norm": 6.90625,
      "learning_rate": 4.597796211254658e-05,
      "loss": 0.5989195823669433,
      "mean_token_accuracy": 0.8474084064364433,
      "num_tokens": 24730282.0,
      "step": 12090
    },
    {
      "entropy": 0.5500552109675482,
      "epoch": 0.19651949359687518,
      "grad_norm": 6.1875,
      "learning_rate": 4.597090625861404e-05,
      "loss": 0.6277587890625,
      "mean_token_accuracy": 0.8525433756411076,
      "num_tokens": 24751481.0,
      "step": 12100
    },
    {
      "entropy": 0.5770912560634315,
      "epoch": 0.19668190640150068,
      "grad_norm": 9.4375,
      "learning_rate": 4.5963844763460586e-05,
      "loss": 0.5572945594787597,
      "mean_token_accuracy": 0.8621942054480314,
      "num_tokens": 24772257.0,
      "step": 12110
    },
    {
      "entropy": 0.6638257725164294,
      "epoch": 0.19684431920612622,
      "grad_norm": 9.9375,
      "learning_rate": 4.5956777628985786e-05,
      "loss": 0.6389250755310059,
      "mean_token_accuracy": 0.8278104931116104,
      "num_tokens": 24792250.0,
      "step": 12120
    },
    {
      "entropy": 0.6357780814170837,
      "epoch": 0.19700673201075172,
      "grad_norm": 8.25,
      "learning_rate": 4.59497048570907e-05,
      "loss": 0.6687206268310547,
      "mean_token_accuracy": 0.8391313157975674,
      "num_tokens": 24812178.0,
      "step": 12130
    },
    {
      "entropy": 0.5939867103472352,
      "epoch": 0.19716914481537726,
      "grad_norm": 14.625,
      "learning_rate": 4.594262644967793e-05,
      "loss": 0.5968604564666748,
      "mean_token_accuracy": 0.8549564704298973,
      "num_tokens": 24832205.0,
      "step": 12140
    },
    {
      "entropy": 0.563919450622052,
      "epoch": 0.19733155762000276,
      "grad_norm": 8.5625,
      "learning_rate": 4.5935542408651576e-05,
      "loss": 0.5433680057525635,
      "mean_token_accuracy": 0.8539632156491279,
      "num_tokens": 24852209.0,
      "step": 12150
    },
    {
      "entropy": 0.677375310845673,
      "epoch": 0.19749397042462827,
      "grad_norm": 7.3125,
      "learning_rate": 4.592845273591727e-05,
      "loss": 0.6873537540435791,
      "mean_token_accuracy": 0.8329411759972573,
      "num_tokens": 24874039.0,
      "step": 12160
    },
    {
      "entropy": 0.5796006670221686,
      "epoch": 0.1976563832292538,
      "grad_norm": 9.625,
      "learning_rate": 4.592135743338214e-05,
      "loss": 0.5948408126831055,
      "mean_token_accuracy": 0.8498248875141143,
      "num_tokens": 24894232.0,
      "step": 12170
    },
    {
      "entropy": 0.5285936133936048,
      "epoch": 0.1978187960338793,
      "grad_norm": 7.4375,
      "learning_rate": 4.5914256502954843e-05,
      "loss": 0.5077470779418946,
      "mean_token_accuracy": 0.8718863628804684,
      "num_tokens": 24914862.0,
      "step": 12180
    },
    {
      "entropy": 0.5790426744148135,
      "epoch": 0.19798120883850484,
      "grad_norm": 5.6875,
      "learning_rate": 4.590714994654555e-05,
      "loss": 0.6062549114227295,
      "mean_token_accuracy": 0.8500985227525234,
      "num_tokens": 24935027.0,
      "step": 12190
    },
    {
      "entropy": 0.6258042281493544,
      "epoch": 0.19814362164313035,
      "grad_norm": 11.375,
      "learning_rate": 4.590003776606593e-05,
      "loss": 0.5895543098449707,
      "mean_token_accuracy": 0.8503380060195923,
      "num_tokens": 24954785.0,
      "step": 12200
    },
    {
      "entropy": 0.6277354843914509,
      "epoch": 0.19830603444775585,
      "grad_norm": 9.4375,
      "learning_rate": 4.589291996342917e-05,
      "loss": 0.6376527786254883,
      "mean_token_accuracy": 0.839674511179328,
      "num_tokens": 24974983.0,
      "step": 12210
    },
    {
      "entropy": 0.5727580291219055,
      "epoch": 0.19846844725238139,
      "grad_norm": 8.6875,
      "learning_rate": 4.588579654055001e-05,
      "loss": 0.5452008724212647,
      "mean_token_accuracy": 0.8614663556218147,
      "num_tokens": 24995188.0,
      "step": 12220
    },
    {
      "entropy": 0.5996033737435937,
      "epoch": 0.1986308600570069,
      "grad_norm": 5.90625,
      "learning_rate": 4.5878667499344626e-05,
      "loss": 0.6016168117523193,
      "mean_token_accuracy": 0.8393660090863705,
      "num_tokens": 25016571.0,
      "step": 12230
    },
    {
      "entropy": 0.5416421694681048,
      "epoch": 0.1987932728616324,
      "grad_norm": 8.875,
      "learning_rate": 4.587153284173077e-05,
      "loss": 0.5561915874481201,
      "mean_token_accuracy": 0.8631891179829836,
      "num_tokens": 25037971.0,
      "step": 12240
    },
    {
      "entropy": 0.6198043281212449,
      "epoch": 0.19895568566625793,
      "grad_norm": 6.03125,
      "learning_rate": 4.586439256962767e-05,
      "loss": 0.6787820816040039,
      "mean_token_accuracy": 0.8327558744698763,
      "num_tokens": 25058284.0,
      "step": 12250
    },
    {
      "entropy": 0.5668345962651073,
      "epoch": 0.19911809847088344,
      "grad_norm": 6.8125,
      "learning_rate": 4.585724668495608e-05,
      "loss": 0.5951244831085205,
      "mean_token_accuracy": 0.8487914100289344,
      "num_tokens": 25077610.0,
      "step": 12260
    },
    {
      "entropy": 0.5841509610414505,
      "epoch": 0.19928051127550897,
      "grad_norm": 7.46875,
      "learning_rate": 4.585009518963825e-05,
      "loss": 0.6184364318847656,
      "mean_token_accuracy": 0.8462065104395151,
      "num_tokens": 25098169.0,
      "step": 12270
    },
    {
      "entropy": 0.5929702019318939,
      "epoch": 0.19944292408013448,
      "grad_norm": 6.71875,
      "learning_rate": 4.584293808559797e-05,
      "loss": 0.5947806358337402,
      "mean_token_accuracy": 0.8467857114970684,
      "num_tokens": 25118267.0,
      "step": 12280
    },
    {
      "entropy": 0.5853442711755633,
      "epoch": 0.19960533688475998,
      "grad_norm": 8.5,
      "learning_rate": 4.5835775374760496e-05,
      "loss": 0.5717984199523926,
      "mean_token_accuracy": 0.8507145315408706,
      "num_tokens": 25138674.0,
      "step": 12290
    },
    {
      "entropy": 0.6651414018124342,
      "epoch": 0.19976774968938552,
      "grad_norm": 8.125,
      "learning_rate": 4.582860705905262e-05,
      "loss": 0.6717865943908692,
      "mean_token_accuracy": 0.8396883226931096,
      "num_tokens": 25159648.0,
      "step": 12300
    },
    {
      "entropy": 0.5973228931427002,
      "epoch": 0.19993016249401102,
      "grad_norm": 5.75,
      "learning_rate": 4.5821433140402635e-05,
      "loss": 0.5680655479431153,
      "mean_token_accuracy": 0.8569359742105007,
      "num_tokens": 25180509.0,
      "step": 12310
    },
    {
      "entropy": 0.5540860029170289,
      "epoch": 0.20009257529863655,
      "grad_norm": 8.0625,
      "learning_rate": 4.581425362074035e-05,
      "loss": 0.5690892696380615,
      "mean_token_accuracy": 0.859893174469471,
      "num_tokens": 25200902.0,
      "step": 12320
    },
    {
      "entropy": 0.6248894959688187,
      "epoch": 0.20025498810326206,
      "grad_norm": 8.375,
      "learning_rate": 4.580706850199706e-05,
      "loss": 0.6312798500061035,
      "mean_token_accuracy": 0.8403501480817794,
      "num_tokens": 25220635.0,
      "step": 12330
    },
    {
      "entropy": 0.5906405672896653,
      "epoch": 0.20041740090788757,
      "grad_norm": 8.4375,
      "learning_rate": 4.579987778610558e-05,
      "loss": 0.6269954681396485,
      "mean_token_accuracy": 0.8447614349424839,
      "num_tokens": 25241666.0,
      "step": 12340
    },
    {
      "entropy": 0.6615077394992113,
      "epoch": 0.2005798137125131,
      "grad_norm": 6.84375,
      "learning_rate": 4.5792681475000235e-05,
      "loss": 0.6507661819458008,
      "mean_token_accuracy": 0.8414684787392617,
      "num_tokens": 25262934.0,
      "step": 12350
    },
    {
      "entropy": 0.5459652805700899,
      "epoch": 0.2007422265171386,
      "grad_norm": 7.90625,
      "learning_rate": 4.578547957061684e-05,
      "loss": 0.5598945617675781,
      "mean_token_accuracy": 0.8539892956614494,
      "num_tokens": 25282350.0,
      "step": 12360
    },
    {
      "entropy": 0.6762089921161533,
      "epoch": 0.20090463932176414,
      "grad_norm": 7.1875,
      "learning_rate": 4.577827207489273e-05,
      "loss": 0.692571496963501,
      "mean_token_accuracy": 0.8281513914465904,
      "num_tokens": 25303253.0,
      "step": 12370
    },
    {
      "entropy": 0.544951643422246,
      "epoch": 0.20106705212638964,
      "grad_norm": 5.65625,
      "learning_rate": 4.577105898976673e-05,
      "loss": 0.5579776287078857,
      "mean_token_accuracy": 0.8642911426723003,
      "num_tokens": 25323221.0,
      "step": 12380
    },
    {
      "entropy": 0.6526072254404426,
      "epoch": 0.20122946493101515,
      "grad_norm": 9.6875,
      "learning_rate": 4.5763840317179185e-05,
      "loss": 0.6545165061950684,
      "mean_token_accuracy": 0.8361254878342151,
      "num_tokens": 25343338.0,
      "step": 12390
    },
    {
      "entropy": 0.6418019724078476,
      "epoch": 0.20139187773564068,
      "grad_norm": 9.0625,
      "learning_rate": 4.5756616059071934e-05,
      "loss": 0.6175976276397706,
      "mean_token_accuracy": 0.8409507911652326,
      "num_tokens": 25363911.0,
      "step": 12400
    },
    {
      "entropy": 0.6676965657621622,
      "epoch": 0.2015542905402662,
      "grad_norm": 5.625,
      "learning_rate": 4.574938621738831e-05,
      "loss": 0.6540782451629639,
      "mean_token_accuracy": 0.8318482153117657,
      "num_tokens": 25384986.0,
      "step": 12410
    },
    {
      "entropy": 0.6576505059376359,
      "epoch": 0.20171670334489172,
      "grad_norm": 7.90625,
      "learning_rate": 4.574215079407317e-05,
      "loss": 0.6810944080352783,
      "mean_token_accuracy": 0.8311419367790223,
      "num_tokens": 25405111.0,
      "step": 12420
    },
    {
      "entropy": 0.6235237221233547,
      "epoch": 0.20187911614951723,
      "grad_norm": 9.375,
      "learning_rate": 4.573490979107285e-05,
      "loss": 0.6142261505126954,
      "mean_token_accuracy": 0.8412287473678589,
      "num_tokens": 25424890.0,
      "step": 12430
    },
    {
      "entropy": 0.6629935150966049,
      "epoch": 0.20204152895414274,
      "grad_norm": 8.1875,
      "learning_rate": 4.572766321033519e-05,
      "loss": 0.6362610816955566,
      "mean_token_accuracy": 0.839865130931139,
      "num_tokens": 25445756.0,
      "step": 12440
    },
    {
      "entropy": 0.6212058047764003,
      "epoch": 0.20220394175876827,
      "grad_norm": 8.8125,
      "learning_rate": 4.572041105380955e-05,
      "loss": 0.627192211151123,
      "mean_token_accuracy": 0.8364308565855026,
      "num_tokens": 25466700.0,
      "step": 12450
    },
    {
      "entropy": 0.6234149986878037,
      "epoch": 0.20236635456339377,
      "grad_norm": 5.59375,
      "learning_rate": 4.571315332344677e-05,
      "loss": 0.6454553604125977,
      "mean_token_accuracy": 0.84310552328825,
      "num_tokens": 25487474.0,
      "step": 12460
    },
    {
      "entropy": 0.5880932228639721,
      "epoch": 0.2025287673680193,
      "grad_norm": 6.75,
      "learning_rate": 4.570589002119919e-05,
      "loss": 0.5909905433654785,
      "mean_token_accuracy": 0.8543611988425255,
      "num_tokens": 25507920.0,
      "step": 12470
    },
    {
      "entropy": 0.5964352490380407,
      "epoch": 0.2026911801726448,
      "grad_norm": 7.4375,
      "learning_rate": 4.569862114902067e-05,
      "loss": 0.6112945079803467,
      "mean_token_accuracy": 0.8511624678969383,
      "num_tokens": 25528521.0,
      "step": 12480
    },
    {
      "entropy": 0.5483819904737175,
      "epoch": 0.20285359297727032,
      "grad_norm": 5.625,
      "learning_rate": 4.5691346708866546e-05,
      "loss": 0.5642324924468994,
      "mean_token_accuracy": 0.8618316642940045,
      "num_tokens": 25549204.0,
      "step": 12490
    },
    {
      "entropy": 0.5949614201206714,
      "epoch": 0.20301600578189585,
      "grad_norm": 8.125,
      "learning_rate": 4.5684066702693646e-05,
      "loss": 0.6095510005950928,
      "mean_token_accuracy": 0.845836928486824,
      "num_tokens": 25570451.0,
      "step": 12500
    },
    {
      "entropy": 0.5709110877476633,
      "epoch": 0.20317841858652136,
      "grad_norm": 6.65625,
      "learning_rate": 4.5676781132460324e-05,
      "loss": 0.598017692565918,
      "mean_token_accuracy": 0.8556294903159142,
      "num_tokens": 25591357.0,
      "step": 12510
    },
    {
      "entropy": 0.5978128618560732,
      "epoch": 0.20334083139114686,
      "grad_norm": 7.8125,
      "learning_rate": 4.5669490000126404e-05,
      "loss": 0.5989922523498535,
      "mean_token_accuracy": 0.8496179550886154,
      "num_tokens": 25611389.0,
      "step": 12520
    },
    {
      "entropy": 0.5824659419246018,
      "epoch": 0.2035032441957724,
      "grad_norm": 8.9375,
      "learning_rate": 4.566219330765323e-05,
      "loss": 0.6109497547149658,
      "mean_token_accuracy": 0.8479265555739403,
      "num_tokens": 25632320.0,
      "step": 12530
    },
    {
      "entropy": 0.6336340075358748,
      "epoch": 0.2036656570003979,
      "grad_norm": 9.5625,
      "learning_rate": 4.5654891057003614e-05,
      "loss": 0.6162057399749756,
      "mean_token_accuracy": 0.8453248865902424,
      "num_tokens": 25652972.0,
      "step": 12540
    },
    {
      "entropy": 0.6921040973626077,
      "epoch": 0.20382806980502344,
      "grad_norm": 9.25,
      "learning_rate": 4.564758325014189e-05,
      "loss": 0.6840372085571289,
      "mean_token_accuracy": 0.8254360154271125,
      "num_tokens": 25674637.0,
      "step": 12550
    },
    {
      "entropy": 0.5493938840925694,
      "epoch": 0.20399048260964894,
      "grad_norm": 7.84375,
      "learning_rate": 4.564026988903387e-05,
      "loss": 0.5005664348602294,
      "mean_token_accuracy": 0.8690766595304013,
      "num_tokens": 25693758.0,
      "step": 12560
    },
    {
      "entropy": 0.5918134281411767,
      "epoch": 0.20415289541427445,
      "grad_norm": 8.0625,
      "learning_rate": 4.5632950975646855e-05,
      "loss": 0.6353377342224121,
      "mean_token_accuracy": 0.8374360002577305,
      "num_tokens": 25713942.0,
      "step": 12570
    },
    {
      "entropy": 0.627518549002707,
      "epoch": 0.20431530821889998,
      "grad_norm": 11.4375,
      "learning_rate": 4.562562651194966e-05,
      "loss": 0.6557719230651855,
      "mean_token_accuracy": 0.8441725812852383,
      "num_tokens": 25733785.0,
      "step": 12580
    },
    {
      "entropy": 0.6129566465504468,
      "epoch": 0.2044777210235255,
      "grad_norm": 12.0,
      "learning_rate": 4.561829649991258e-05,
      "loss": 0.6014132499694824,
      "mean_token_accuracy": 0.8462629064917564,
      "num_tokens": 25753691.0,
      "step": 12590
    },
    {
      "entropy": 0.6471533371135593,
      "epoch": 0.20464013382815102,
      "grad_norm": 8.8125,
      "learning_rate": 4.561096094150742e-05,
      "loss": 0.6440921783447265,
      "mean_token_accuracy": 0.8424212314188481,
      "num_tokens": 25773465.0,
      "step": 12600
    },
    {
      "entropy": 0.5994148073717952,
      "epoch": 0.20480254663277653,
      "grad_norm": 8.75,
      "learning_rate": 4.560361983870743e-05,
      "loss": 0.5762141704559326,
      "mean_token_accuracy": 0.8506194092333317,
      "num_tokens": 25793534.0,
      "step": 12610
    },
    {
      "entropy": 0.5853875853121281,
      "epoch": 0.20496495943740203,
      "grad_norm": 10.0625,
      "learning_rate": 4.5596273193487406e-05,
      "loss": 0.5767448425292969,
      "mean_token_accuracy": 0.8518157228827477,
      "num_tokens": 25813729.0,
      "step": 12620
    },
    {
      "entropy": 0.574063543882221,
      "epoch": 0.20512737224202757,
      "grad_norm": 6.28125,
      "learning_rate": 4.558892100782361e-05,
      "loss": 0.5954799652099609,
      "mean_token_accuracy": 0.8461894355714321,
      "num_tokens": 25834794.0,
      "step": 12630
    },
    {
      "entropy": 0.573026316286996,
      "epoch": 0.20528978504665307,
      "grad_norm": 5.53125,
      "learning_rate": 4.558156328369377e-05,
      "loss": 0.581175708770752,
      "mean_token_accuracy": 0.852231876552105,
      "num_tokens": 25854624.0,
      "step": 12640
    },
    {
      "entropy": 0.6256012193858623,
      "epoch": 0.2054521978512786,
      "grad_norm": 11.375,
      "learning_rate": 4.557420002307718e-05,
      "loss": 0.7018474578857422,
      "mean_token_accuracy": 0.8344818133860826,
      "num_tokens": 25874297.0,
      "step": 12650
    },
    {
      "entropy": 0.6324831193313003,
      "epoch": 0.2056146106559041,
      "grad_norm": 8.0625,
      "learning_rate": 4.5566831227954534e-05,
      "loss": 0.6324845790863037,
      "mean_token_accuracy": 0.8384547051042318,
      "num_tokens": 25895409.0,
      "step": 12660
    },
    {
      "entropy": 0.565849128831178,
      "epoch": 0.20577702346052962,
      "grad_norm": 6.3125,
      "learning_rate": 4.555945690030806e-05,
      "loss": 0.5494294166564941,
      "mean_token_accuracy": 0.8569397337734699,
      "num_tokens": 25915064.0,
      "step": 12670
    },
    {
      "entropy": 0.6091129289939999,
      "epoch": 0.20593943626515515,
      "grad_norm": 7.40625,
      "learning_rate": 4.555207704212149e-05,
      "loss": 0.5675833702087403,
      "mean_token_accuracy": 0.8546780467033386,
      "num_tokens": 25935944.0,
      "step": 12680
    },
    {
      "entropy": 0.6037541112396866,
      "epoch": 0.20610184906978066,
      "grad_norm": 7.875,
      "learning_rate": 4.554469165538e-05,
      "loss": 0.6208678245544433,
      "mean_token_accuracy": 0.8511412315070629,
      "num_tokens": 25956399.0,
      "step": 12690
    },
    {
      "entropy": 0.5678708024322987,
      "epoch": 0.2062642618744062,
      "grad_norm": 8.8125,
      "learning_rate": 4.553730074207028e-05,
      "loss": 0.5802405834197998,
      "mean_token_accuracy": 0.8421403937041759,
      "num_tokens": 25977047.0,
      "step": 12700
    },
    {
      "entropy": 0.5321389018557966,
      "epoch": 0.2064266746790317,
      "grad_norm": 8.625,
      "learning_rate": 4.552990430418051e-05,
      "loss": 0.5730510711669922,
      "mean_token_accuracy": 0.8496560178697109,
      "num_tokens": 25997649.0,
      "step": 12710
    },
    {
      "entropy": 0.5172807401046157,
      "epoch": 0.2065890874836572,
      "grad_norm": 6.4375,
      "learning_rate": 4.5522502343700344e-05,
      "loss": 0.5073169231414795,
      "mean_token_accuracy": 0.8676378220319748,
      "num_tokens": 26017249.0,
      "step": 12720
    },
    {
      "entropy": 0.5905818049795926,
      "epoch": 0.20675150028828274,
      "grad_norm": 6.84375,
      "learning_rate": 4.551509486262092e-05,
      "loss": 0.6217871189117432,
      "mean_token_accuracy": 0.8470325686037541,
      "num_tokens": 26037807.0,
      "step": 12730
    },
    {
      "entropy": 0.5386319880373776,
      "epoch": 0.20691391309290824,
      "grad_norm": 7.15625,
      "learning_rate": 4.550768186293488e-05,
      "loss": 0.5165777683258057,
      "mean_token_accuracy": 0.8650362238287925,
      "num_tokens": 26058342.0,
      "step": 12740
    },
    {
      "entropy": 0.6223049905151129,
      "epoch": 0.20707632589753378,
      "grad_norm": 9.4375,
      "learning_rate": 4.5500263346636335e-05,
      "loss": 0.6315126895904541,
      "mean_token_accuracy": 0.8455738827586174,
      "num_tokens": 26079612.0,
      "step": 12750
    },
    {
      "entropy": 0.5634799910709262,
      "epoch": 0.20723873870215928,
      "grad_norm": 7.0,
      "learning_rate": 4.549283931572087e-05,
      "loss": 0.5063044548034668,
      "mean_token_accuracy": 0.859919936209917,
      "num_tokens": 26099959.0,
      "step": 12760
    },
    {
      "entropy": 0.634121502796188,
      "epoch": 0.2074011515067848,
      "grad_norm": 10.25,
      "learning_rate": 4.548540977218558e-05,
      "loss": 0.6917435646057128,
      "mean_token_accuracy": 0.8359726808965207,
      "num_tokens": 26120413.0,
      "step": 12770
    },
    {
      "entropy": 0.5698966154828667,
      "epoch": 0.20756356431141032,
      "grad_norm": 8.25,
      "learning_rate": 4.547797471802902e-05,
      "loss": 0.5422182083129883,
      "mean_token_accuracy": 0.8577943921089173,
      "num_tokens": 26140404.0,
      "step": 12780
    },
    {
      "entropy": 0.5425614455714822,
      "epoch": 0.20772597711603583,
      "grad_norm": 6.25,
      "learning_rate": 4.547053415525124e-05,
      "loss": 0.5776753902435303,
      "mean_token_accuracy": 0.8568359181284905,
      "num_tokens": 26159751.0,
      "step": 12790
    },
    {
      "entropy": 0.5403234341181815,
      "epoch": 0.20788838992066133,
      "grad_norm": 9.125,
      "learning_rate": 4.5463088085853764e-05,
      "loss": 0.5276986598968506,
      "mean_token_accuracy": 0.8671185091137886,
      "num_tokens": 26181101.0,
      "step": 12800
    },
    {
      "entropy": 0.555131318513304,
      "epoch": 0.20805080272528687,
      "grad_norm": 7.34375,
      "learning_rate": 4.545563651183961e-05,
      "loss": 0.5465211391448974,
      "mean_token_accuracy": 0.8508646734058857,
      "num_tokens": 26201244.0,
      "step": 12810
    },
    {
      "entropy": 0.5831199629232288,
      "epoch": 0.20821321552991237,
      "grad_norm": 11.125,
      "learning_rate": 4.5448179435213266e-05,
      "loss": 0.6354680538177491,
      "mean_token_accuracy": 0.8401526510715485,
      "num_tokens": 26221019.0,
      "step": 12820
    },
    {
      "entropy": 0.6137001304887235,
      "epoch": 0.2083756283345379,
      "grad_norm": 8.25,
      "learning_rate": 4.544071685798069e-05,
      "loss": 0.6217742919921875,
      "mean_token_accuracy": 0.845080553740263,
      "num_tokens": 26242519.0,
      "step": 12830
    },
    {
      "entropy": 0.6177702963352203,
      "epoch": 0.2085380411391634,
      "grad_norm": 6.6875,
      "learning_rate": 4.543324878214935e-05,
      "loss": 0.6535583972930908,
      "mean_token_accuracy": 0.8384648930281401,
      "num_tokens": 26262102.0,
      "step": 12840
    },
    {
      "entropy": 0.5606361345387996,
      "epoch": 0.20870045394378892,
      "grad_norm": 5.75,
      "learning_rate": 4.542577520972816e-05,
      "loss": 0.5556678295135498,
      "mean_token_accuracy": 0.8576150618493557,
      "num_tokens": 26282728.0,
      "step": 12850
    },
    {
      "entropy": 0.5913861093111337,
      "epoch": 0.20886286674841445,
      "grad_norm": 8.375,
      "learning_rate": 4.5418296142727546e-05,
      "loss": 0.587209415435791,
      "mean_token_accuracy": 0.8512399435043335,
      "num_tokens": 26303066.0,
      "step": 12860
    },
    {
      "entropy": 0.5728941245004535,
      "epoch": 0.20902527955303996,
      "grad_norm": 8.625,
      "learning_rate": 4.541081158315937e-05,
      "loss": 0.5922245979309082,
      "mean_token_accuracy": 0.8495579227805138,
      "num_tokens": 26324145.0,
      "step": 12870
    },
    {
      "entropy": 0.6242799975909292,
      "epoch": 0.2091876923576655,
      "grad_norm": 10.1875,
      "learning_rate": 4.540332153303701e-05,
      "loss": 0.6024978160858154,
      "mean_token_accuracy": 0.8456556364893913,
      "num_tokens": 26345287.0,
      "step": 12880
    },
    {
      "entropy": 0.5608430268242955,
      "epoch": 0.209350105162291,
      "grad_norm": 7.96875,
      "learning_rate": 4.539582599437531e-05,
      "loss": 0.5367940425872803,
      "mean_token_accuracy": 0.8634302474558353,
      "num_tokens": 26366246.0,
      "step": 12890
    },
    {
      "entropy": 0.6002460293471813,
      "epoch": 0.2095125179669165,
      "grad_norm": 5.75,
      "learning_rate": 4.5388324969190575e-05,
      "loss": 0.595156478881836,
      "mean_token_accuracy": 0.856720020622015,
      "num_tokens": 26386159.0,
      "step": 12900
    },
    {
      "entropy": 0.5566518366336822,
      "epoch": 0.20967493077154203,
      "grad_norm": 6.34375,
      "learning_rate": 4.538081845950061e-05,
      "loss": 0.5688837051391602,
      "mean_token_accuracy": 0.8517802715301513,
      "num_tokens": 26407716.0,
      "step": 12910
    },
    {
      "entropy": 0.6460247313603759,
      "epoch": 0.20983734357616754,
      "grad_norm": 10.75,
      "learning_rate": 4.537330646732467e-05,
      "loss": 0.6878816604614257,
      "mean_token_accuracy": 0.8275242324918508,
      "num_tokens": 26428923.0,
      "step": 12920
    },
    {
      "entropy": 0.5795514481142163,
      "epoch": 0.20999975638079307,
      "grad_norm": 6.09375,
      "learning_rate": 4.53657889946835e-05,
      "loss": 0.5784704208374023,
      "mean_token_accuracy": 0.854938293620944,
      "num_tokens": 26450044.0,
      "step": 12930
    },
    {
      "entropy": 0.564535618852824,
      "epoch": 0.21016216918541858,
      "grad_norm": 8.125,
      "learning_rate": 4.535826604359933e-05,
      "loss": 0.6065794467926026,
      "mean_token_accuracy": 0.8490377325564623,
      "num_tokens": 26469650.0,
      "step": 12940
    },
    {
      "entropy": 0.6050719400402158,
      "epoch": 0.21032458199004408,
      "grad_norm": 8.625,
      "learning_rate": 4.535073761609584e-05,
      "loss": 0.6083506107330322,
      "mean_token_accuracy": 0.8514999642968177,
      "num_tokens": 26491114.0,
      "step": 12950
    },
    {
      "entropy": 0.6171657664701342,
      "epoch": 0.21048699479466962,
      "grad_norm": 13.125,
      "learning_rate": 4.534320371419819e-05,
      "loss": 0.5959738731384278,
      "mean_token_accuracy": 0.8469117738306522,
      "num_tokens": 26509790.0,
      "step": 12960
    },
    {
      "entropy": 0.5391187734901906,
      "epoch": 0.21064940759929512,
      "grad_norm": 8.0,
      "learning_rate": 4.5335664339933026e-05,
      "loss": 0.5375486373901367,
      "mean_token_accuracy": 0.8594507731497287,
      "num_tokens": 26529014.0,
      "step": 12970
    },
    {
      "entropy": 0.5697816568426788,
      "epoch": 0.21081182040392066,
      "grad_norm": 7.375,
      "learning_rate": 4.532811949532846e-05,
      "loss": 0.6093725204467774,
      "mean_token_accuracy": 0.8468419961631298,
      "num_tokens": 26548741.0,
      "step": 12980
    },
    {
      "entropy": 0.6318000695668161,
      "epoch": 0.21097423320854616,
      "grad_norm": 12.875,
      "learning_rate": 4.532056918241405e-05,
      "loss": 0.6285727024078369,
      "mean_token_accuracy": 0.8420143134891986,
      "num_tokens": 26569622.0,
      "step": 12990
    },
    {
      "entropy": 0.5309915577992796,
      "epoch": 0.21113664601317167,
      "grad_norm": 8.1875,
      "learning_rate": 4.531301340322088e-05,
      "loss": 0.5215253829956055,
      "mean_token_accuracy": 0.8714127957820892,
      "num_tokens": 26588775.0,
      "step": 13000
    },
    {
      "entropy": 0.5509628580883146,
      "epoch": 0.2112990588177972,
      "grad_norm": 6.15625,
      "learning_rate": 4.5305452159781446e-05,
      "loss": 0.557252311706543,
      "mean_token_accuracy": 0.8641130894422531,
      "num_tokens": 26607611.0,
      "step": 13010
    },
    {
      "entropy": 0.559067863272503,
      "epoch": 0.2114614716224227,
      "grad_norm": 8.1875,
      "learning_rate": 4.529788545412974e-05,
      "loss": 0.5459411144256592,
      "mean_token_accuracy": 0.8546712540090085,
      "num_tokens": 26626936.0,
      "step": 13020
    },
    {
      "entropy": 0.47954417162109164,
      "epoch": 0.21162388442704824,
      "grad_norm": 8.75,
      "learning_rate": 4.529031328830125e-05,
      "loss": 0.4769723415374756,
      "mean_token_accuracy": 0.8765573270618916,
      "num_tokens": 26645804.0,
      "step": 13030
    },
    {
      "entropy": 0.6317354856058955,
      "epoch": 0.21178629723167375,
      "grad_norm": 8.9375,
      "learning_rate": 4.528273566433286e-05,
      "loss": 0.6989704608917237,
      "mean_token_accuracy": 0.8353893034160137,
      "num_tokens": 26665657.0,
      "step": 13040
    },
    {
      "entropy": 0.5908635720610619,
      "epoch": 0.21194871003629925,
      "grad_norm": 8.375,
      "learning_rate": 4.527515258426302e-05,
      "loss": 0.6051485538482666,
      "mean_token_accuracy": 0.842450650781393,
      "num_tokens": 26684975.0,
      "step": 13050
    },
    {
      "entropy": 0.5843564012087882,
      "epoch": 0.2121111228409248,
      "grad_norm": 10.375,
      "learning_rate": 4.5267564050131545e-05,
      "loss": 0.5552528381347657,
      "mean_token_accuracy": 0.8510452441871166,
      "num_tokens": 26705812.0,
      "step": 13060
    },
    {
      "entropy": 0.5904080636799336,
      "epoch": 0.2122735356455503,
      "grad_norm": 6.65625,
      "learning_rate": 4.525997006397981e-05,
      "loss": 0.5950635433197021,
      "mean_token_accuracy": 0.8481556177139282,
      "num_tokens": 26727312.0,
      "step": 13070
    },
    {
      "entropy": 0.5930886968970299,
      "epoch": 0.2124359484501758,
      "grad_norm": 5.84375,
      "learning_rate": 4.525237062785058e-05,
      "loss": 0.5662363529205322,
      "mean_token_accuracy": 0.8589229010045528,
      "num_tokens": 26747351.0,
      "step": 13080
    },
    {
      "entropy": 0.6134946128353477,
      "epoch": 0.21259836125480133,
      "grad_norm": 10.8125,
      "learning_rate": 4.524476574378815e-05,
      "loss": 0.6202672481536865,
      "mean_token_accuracy": 0.8436128254979849,
      "num_tokens": 26767330.0,
      "step": 13090
    },
    {
      "entropy": 0.5631523092277348,
      "epoch": 0.21276077405942684,
      "grad_norm": 5.5625,
      "learning_rate": 4.523715541383823e-05,
      "loss": 0.6167941093444824,
      "mean_token_accuracy": 0.8495938830077648,
      "num_tokens": 26787762.0,
      "step": 13100
    },
    {
      "entropy": 0.5860170730389654,
      "epoch": 0.21292318686405237,
      "grad_norm": 5.53125,
      "learning_rate": 4.522953964004803e-05,
      "loss": 0.5743618011474609,
      "mean_token_accuracy": 0.8501062326133251,
      "num_tokens": 26806907.0,
      "step": 13110
    },
    {
      "entropy": 0.6544931184500455,
      "epoch": 0.21308559966867788,
      "grad_norm": 8.125,
      "learning_rate": 4.5221918424466205e-05,
      "loss": 0.7203569889068604,
      "mean_token_accuracy": 0.8309626929461956,
      "num_tokens": 26826584.0,
      "step": 13120
    },
    {
      "entropy": 0.5908799624070525,
      "epoch": 0.21324801247330338,
      "grad_norm": 6.96875,
      "learning_rate": 4.521429176914287e-05,
      "loss": 0.598411750793457,
      "mean_token_accuracy": 0.8551476657390594,
      "num_tokens": 26845347.0,
      "step": 13130
    },
    {
      "entropy": 0.5607292917557061,
      "epoch": 0.21341042527792892,
      "grad_norm": 16.875,
      "learning_rate": 4.520665967612963e-05,
      "loss": 0.5665550708770752,
      "mean_token_accuracy": 0.8568595871329308,
      "num_tokens": 26865107.0,
      "step": 13140
    },
    {
      "entropy": 0.5846299070864915,
      "epoch": 0.21357283808255442,
      "grad_norm": 8.9375,
      "learning_rate": 4.519902214747951e-05,
      "loss": 0.5735989570617676,
      "mean_token_accuracy": 0.8532745398581028,
      "num_tokens": 26885825.0,
      "step": 13150
    },
    {
      "entropy": 0.5879601422231644,
      "epoch": 0.21373525088717996,
      "grad_norm": 7.21875,
      "learning_rate": 4.519137918524705e-05,
      "loss": 0.6088402748107911,
      "mean_token_accuracy": 0.8364171959459782,
      "num_tokens": 26906327.0,
      "step": 13160
    },
    {
      "entropy": 0.6215938205365091,
      "epoch": 0.21389766369180546,
      "grad_norm": 6.15625,
      "learning_rate": 4.518373079148821e-05,
      "loss": 0.6142525196075439,
      "mean_token_accuracy": 0.847584730386734,
      "num_tokens": 26926813.0,
      "step": 13170
    },
    {
      "entropy": 0.6777175518684089,
      "epoch": 0.21406007649643097,
      "grad_norm": 5.6875,
      "learning_rate": 4.517607696826043e-05,
      "loss": 0.6888678550720215,
      "mean_token_accuracy": 0.8328000564128161,
      "num_tokens": 26947930.0,
      "step": 13180
    },
    {
      "entropy": 0.5537998434156179,
      "epoch": 0.2142224893010565,
      "grad_norm": 5.90625,
      "learning_rate": 4.516841771762259e-05,
      "loss": 0.554570484161377,
      "mean_token_accuracy": 0.855620163679123,
      "num_tokens": 26969037.0,
      "step": 13190
    },
    {
      "entropy": 0.6061325184069574,
      "epoch": 0.214384902105682,
      "grad_norm": 8.0625,
      "learning_rate": 4.516075304163507e-05,
      "loss": 0.5850739479064941,
      "mean_token_accuracy": 0.8552710197865963,
      "num_tokens": 26988595.0,
      "step": 13200
    },
    {
      "entropy": 0.5684152217581868,
      "epoch": 0.21454731491030754,
      "grad_norm": 7.21875,
      "learning_rate": 4.515308294235967e-05,
      "loss": 0.5409348011016846,
      "mean_token_accuracy": 0.8607885114848614,
      "num_tokens": 27008679.0,
      "step": 13210
    },
    {
      "entropy": 0.5725357919931412,
      "epoch": 0.21470972771493305,
      "grad_norm": 13.375,
      "learning_rate": 4.514540742185967e-05,
      "loss": 0.6376799583435059,
      "mean_token_accuracy": 0.8449529021978378,
      "num_tokens": 27028110.0,
      "step": 13220
    },
    {
      "entropy": 0.5455437148921192,
      "epoch": 0.21487214051955855,
      "grad_norm": 6.53125,
      "learning_rate": 4.5137726482199795e-05,
      "loss": 0.5551892757415772,
      "mean_token_accuracy": 0.8614797934889793,
      "num_tokens": 27048075.0,
      "step": 13230
    },
    {
      "entropy": 0.6112853334285319,
      "epoch": 0.21503455332418409,
      "grad_norm": 10.0,
      "learning_rate": 4.513004012544624e-05,
      "loss": 0.6099318027496338,
      "mean_token_accuracy": 0.8447923831641674,
      "num_tokens": 27068804.0,
      "step": 13240
    },
    {
      "entropy": 0.561993311997503,
      "epoch": 0.2151969661288096,
      "grad_norm": 6.96875,
      "learning_rate": 4.512234835366665e-05,
      "loss": 0.5408859729766846,
      "mean_token_accuracy": 0.8612990379333496,
      "num_tokens": 27088007.0,
      "step": 13250
    },
    {
      "entropy": 0.5996526384260505,
      "epoch": 0.21535937893343512,
      "grad_norm": 11.75,
      "learning_rate": 4.5114651168930146e-05,
      "loss": 0.6766511917114257,
      "mean_token_accuracy": 0.8420096188783646,
      "num_tokens": 27108241.0,
      "step": 13260
    },
    {
      "entropy": 0.6187214515171945,
      "epoch": 0.21552179173806063,
      "grad_norm": 8.8125,
      "learning_rate": 4.510694857330726e-05,
      "loss": 0.6370086669921875,
      "mean_token_accuracy": 0.849754985421896,
      "num_tokens": 27128670.0,
      "step": 13270
    },
    {
      "entropy": 0.5616675010882318,
      "epoch": 0.21568420454268614,
      "grad_norm": 6.75,
      "learning_rate": 4.509924056887002e-05,
      "loss": 0.5633458614349365,
      "mean_token_accuracy": 0.857379600405693,
      "num_tokens": 27149453.0,
      "step": 13280
    },
    {
      "entropy": 0.5845215354114771,
      "epoch": 0.21584661734731167,
      "grad_norm": 8.1875,
      "learning_rate": 4.50915271576919e-05,
      "loss": 0.5596214771270752,
      "mean_token_accuracy": 0.8613359294831753,
      "num_tokens": 27169482.0,
      "step": 13290
    },
    {
      "entropy": 0.6241718929260969,
      "epoch": 0.21600903015193718,
      "grad_norm": 8.375,
      "learning_rate": 4.508380834184782e-05,
      "loss": 0.6312120437622071,
      "mean_token_accuracy": 0.849313897639513,
      "num_tokens": 27190053.0,
      "step": 13300
    },
    {
      "entropy": 0.5954360147938133,
      "epoch": 0.21617144295656268,
      "grad_norm": 8.1875,
      "learning_rate": 4.507608412341416e-05,
      "loss": 0.5822605133056641,
      "mean_token_accuracy": 0.8527102522552014,
      "num_tokens": 27209705.0,
      "step": 13310
    },
    {
      "entropy": 0.6242711815051735,
      "epoch": 0.21633385576118822,
      "grad_norm": 8.625,
      "learning_rate": 4.5068354504468745e-05,
      "loss": 0.5923354625701904,
      "mean_token_accuracy": 0.8470178112387657,
      "num_tokens": 27229733.0,
      "step": 13320
    },
    {
      "entropy": 0.7092706411611289,
      "epoch": 0.21649626856581372,
      "grad_norm": 7.09375,
      "learning_rate": 4.506061948709088e-05,
      "loss": 0.7210259914398194,
      "mean_token_accuracy": 0.8336750157177448,
      "num_tokens": 27251310.0,
      "step": 13330
    },
    {
      "entropy": 0.5234185881912709,
      "epoch": 0.21665868137043925,
      "grad_norm": 6.96875,
      "learning_rate": 4.5052879073361286e-05,
      "loss": 0.5401426315307617,
      "mean_token_accuracy": 0.8577727094292641,
      "num_tokens": 27271976.0,
      "step": 13340
    },
    {
      "entropy": 0.5632377143017948,
      "epoch": 0.21682109417506476,
      "grad_norm": 7.15625,
      "learning_rate": 4.504513326536216e-05,
      "loss": 0.555366039276123,
      "mean_token_accuracy": 0.848810525238514,
      "num_tokens": 27291720.0,
      "step": 13350
    },
    {
      "entropy": 0.6184621044434607,
      "epoch": 0.21698350697969027,
      "grad_norm": 7.8125,
      "learning_rate": 4.5037382065177126e-05,
      "loss": 0.6270238876342773,
      "mean_token_accuracy": 0.8402108550071716,
      "num_tokens": 27313075.0,
      "step": 13360
    },
    {
      "entropy": 0.6369271812960505,
      "epoch": 0.2171459197843158,
      "grad_norm": 9.8125,
      "learning_rate": 4.5029625474891294e-05,
      "loss": 0.5924434185028076,
      "mean_token_accuracy": 0.8406204037368298,
      "num_tokens": 27333762.0,
      "step": 13370
    },
    {
      "entropy": 0.5673545027151704,
      "epoch": 0.2173083325889413,
      "grad_norm": 8.1875,
      "learning_rate": 4.502186349659119e-05,
      "loss": 0.6143340587615966,
      "mean_token_accuracy": 0.8496801681816578,
      "num_tokens": 27355200.0,
      "step": 13380
    },
    {
      "entropy": 0.5819858834147453,
      "epoch": 0.21747074539356684,
      "grad_norm": 7.5,
      "learning_rate": 4.501409613236481e-05,
      "loss": 0.5708133697509765,
      "mean_token_accuracy": 0.8557849057018757,
      "num_tokens": 27374582.0,
      "step": 13390
    },
    {
      "entropy": 0.5533765202388168,
      "epoch": 0.21763315819819234,
      "grad_norm": 6.90625,
      "learning_rate": 4.500632338430159e-05,
      "loss": 0.6125475883483886,
      "mean_token_accuracy": 0.8456336699426175,
      "num_tokens": 27395773.0,
      "step": 13400
    },
    {
      "entropy": 0.6313641514629126,
      "epoch": 0.21779557100281785,
      "grad_norm": 8.75,
      "learning_rate": 4.499854525449242e-05,
      "loss": 0.6225035190582275,
      "mean_token_accuracy": 0.8393168538808823,
      "num_tokens": 27416111.0,
      "step": 13410
    },
    {
      "entropy": 0.6179720372892916,
      "epoch": 0.21795798380744338,
      "grad_norm": 6.8125,
      "learning_rate": 4.499076174502964e-05,
      "loss": 0.6161185264587402,
      "mean_token_accuracy": 0.8513596564531326,
      "num_tokens": 27436254.0,
      "step": 13420
    },
    {
      "entropy": 0.590073402132839,
      "epoch": 0.2181203966120689,
      "grad_norm": 10.4375,
      "learning_rate": 4.4982972858007014e-05,
      "loss": 0.6111728191375733,
      "mean_token_accuracy": 0.8455126576125622,
      "num_tokens": 27457682.0,
      "step": 13430
    },
    {
      "entropy": 0.6442305687814951,
      "epoch": 0.21828280941669442,
      "grad_norm": 8.1875,
      "learning_rate": 4.4975178595519784e-05,
      "loss": 0.6590892791748046,
      "mean_token_accuracy": 0.8325745470821857,
      "num_tokens": 27479258.0,
      "step": 13440
    },
    {
      "entropy": 0.604471187852323,
      "epoch": 0.21844522222131993,
      "grad_norm": 9.3125,
      "learning_rate": 4.496737895966462e-05,
      "loss": 0.6176891803741456,
      "mean_token_accuracy": 0.8488929964601993,
      "num_tokens": 27499221.0,
      "step": 13450
    },
    {
      "entropy": 0.6158365201205015,
      "epoch": 0.21860763502594543,
      "grad_norm": 7.625,
      "learning_rate": 4.4959573952539644e-05,
      "loss": 0.5916090965270996,
      "mean_token_accuracy": 0.8533323884010315,
      "num_tokens": 27519600.0,
      "step": 13460
    },
    {
      "entropy": 0.6057021263055503,
      "epoch": 0.21877004783057097,
      "grad_norm": 11.5,
      "learning_rate": 4.495176357624441e-05,
      "loss": 0.6100963592529297,
      "mean_token_accuracy": 0.8426780834794044,
      "num_tokens": 27538892.0,
      "step": 13470
    },
    {
      "entropy": 0.6184348238864914,
      "epoch": 0.21893246063519647,
      "grad_norm": 5.625,
      "learning_rate": 4.4943947832879945e-05,
      "loss": 0.6120424747467041,
      "mean_token_accuracy": 0.8492647856473923,
      "num_tokens": 27558588.0,
      "step": 13480
    },
    {
      "entropy": 0.5870922184083611,
      "epoch": 0.219094873439822,
      "grad_norm": 7.4375,
      "learning_rate": 4.4936126724548686e-05,
      "loss": 0.5942336082458496,
      "mean_token_accuracy": 0.85293108522892,
      "num_tokens": 27579238.0,
      "step": 13490
    },
    {
      "entropy": 0.6466701513156294,
      "epoch": 0.2192572862444475,
      "grad_norm": 6.75,
      "learning_rate": 4.4928300253354536e-05,
      "loss": 0.6805175304412842,
      "mean_token_accuracy": 0.8288794130086898,
      "num_tokens": 27598014.0,
      "step": 13500
    },
    {
      "entropy": 0.5830862953327596,
      "epoch": 0.21941969904907302,
      "grad_norm": 6.96875,
      "learning_rate": 4.492046842140284e-05,
      "loss": 0.6143934726715088,
      "mean_token_accuracy": 0.8496368303894997,
      "num_tokens": 27617769.0,
      "step": 13510
    },
    {
      "entropy": 0.7295439038425684,
      "epoch": 0.21958211185369855,
      "grad_norm": 5.53125,
      "learning_rate": 4.491263123080036e-05,
      "loss": 0.7626360416412353,
      "mean_token_accuracy": 0.8090497124940157,
      "num_tokens": 27638735.0,
      "step": 13520
    },
    {
      "entropy": 0.6053784744814038,
      "epoch": 0.21974452465832406,
      "grad_norm": 7.09375,
      "learning_rate": 4.490478868365533e-05,
      "loss": 0.5695422649383545,
      "mean_token_accuracy": 0.8540315806865693,
      "num_tokens": 27658683.0,
      "step": 13530
    },
    {
      "entropy": 0.6701673232018948,
      "epoch": 0.2199069374629496,
      "grad_norm": 11.1875,
      "learning_rate": 4.4896940782077416e-05,
      "loss": 0.65236496925354,
      "mean_token_accuracy": 0.8288718581199646,
      "num_tokens": 27679782.0,
      "step": 13540
    },
    {
      "entropy": 0.5447944714687765,
      "epoch": 0.2200693502675751,
      "grad_norm": 5.375,
      "learning_rate": 4.488908752817771e-05,
      "loss": 0.503563928604126,
      "mean_token_accuracy": 0.8681561291217804,
      "num_tokens": 27699878.0,
      "step": 13550
    },
    {
      "entropy": 0.5631998360157013,
      "epoch": 0.2202317630722006,
      "grad_norm": 9.0,
      "learning_rate": 4.488122892406876e-05,
      "loss": 0.5946437835693359,
      "mean_token_accuracy": 0.8484135314822197,
      "num_tokens": 27719670.0,
      "step": 13560
    },
    {
      "entropy": 0.5773196968249976,
      "epoch": 0.22039417587682614,
      "grad_norm": 6.53125,
      "learning_rate": 4.4873364971864554e-05,
      "loss": 0.604633378982544,
      "mean_token_accuracy": 0.8464183278381825,
      "num_tokens": 27740124.0,
      "step": 13570
    },
    {
      "entropy": 0.550227849278599,
      "epoch": 0.22055658868145164,
      "grad_norm": 7.125,
      "learning_rate": 4.4865495673680505e-05,
      "loss": 0.5245660305023193,
      "mean_token_accuracy": 0.8624074079096318,
      "num_tokens": 27759341.0,
      "step": 13580
    },
    {
      "entropy": 0.5853316965512931,
      "epoch": 0.22071900148607715,
      "grad_norm": 6.9375,
      "learning_rate": 4.485762103163348e-05,
      "loss": 0.6123684883117676,
      "mean_token_accuracy": 0.857279871404171,
      "num_tokens": 27780094.0,
      "step": 13590
    },
    {
      "entropy": 0.6120886289514601,
      "epoch": 0.22088141429070268,
      "grad_norm": 7.46875,
      "learning_rate": 4.484974104784177e-05,
      "loss": 0.6310450553894043,
      "mean_token_accuracy": 0.8393819686025381,
      "num_tokens": 27800206.0,
      "step": 13600
    },
    {
      "entropy": 0.6180294941645116,
      "epoch": 0.2210438270953282,
      "grad_norm": 6.75,
      "learning_rate": 4.4841855724425096e-05,
      "loss": 0.6286464691162109,
      "mean_token_accuracy": 0.8329641379415988,
      "num_tokens": 27821152.0,
      "step": 13610
    },
    {
      "entropy": 0.5798514286056161,
      "epoch": 0.22120623989995372,
      "grad_norm": 7.34375,
      "learning_rate": 4.483396506350466e-05,
      "loss": 0.6167507171630859,
      "mean_token_accuracy": 0.8508647918701172,
      "num_tokens": 27842506.0,
      "step": 13620
    },
    {
      "entropy": 0.5403699401766062,
      "epoch": 0.22136865270457923,
      "grad_norm": 7.3125,
      "learning_rate": 4.4826069067203034e-05,
      "loss": 0.5899118423461914,
      "mean_token_accuracy": 0.8516736440360546,
      "num_tokens": 27863904.0,
      "step": 13630
    },
    {
      "entropy": 0.5831394665874541,
      "epoch": 0.22153106550920473,
      "grad_norm": 10.8125,
      "learning_rate": 4.4818167737644274e-05,
      "loss": 0.5981091976165771,
      "mean_token_accuracy": 0.8512608230113983,
      "num_tokens": 27884047.0,
      "step": 13640
    },
    {
      "entropy": 0.6130176941864193,
      "epoch": 0.22169347831383027,
      "grad_norm": 8.5,
      "learning_rate": 4.481026107695385e-05,
      "loss": 0.5933070182800293,
      "mean_token_accuracy": 0.8522829137742519,
      "num_tokens": 27903916.0,
      "step": 13650
    },
    {
      "entropy": 0.6087853613309562,
      "epoch": 0.22185589111845577,
      "grad_norm": 8.375,
      "learning_rate": 4.480234908725869e-05,
      "loss": 0.5607362270355225,
      "mean_token_accuracy": 0.8576692640781403,
      "num_tokens": 27923521.0,
      "step": 13660
    },
    {
      "entropy": 0.5609751909971237,
      "epoch": 0.2220183039230813,
      "grad_norm": 7.90625,
      "learning_rate": 4.479443177068711e-05,
      "loss": 0.5735602378845215,
      "mean_token_accuracy": 0.8542346373200417,
      "num_tokens": 27943921.0,
      "step": 13670
    },
    {
      "entropy": 0.529983954783529,
      "epoch": 0.2221807167277068,
      "grad_norm": 8.25,
      "learning_rate": 4.47865091293689e-05,
      "loss": 0.5503905296325684,
      "mean_token_accuracy": 0.8610822096467018,
      "num_tokens": 27962733.0,
      "step": 13680
    },
    {
      "entropy": 0.5485262484289706,
      "epoch": 0.22234312953233232,
      "grad_norm": 6.46875,
      "learning_rate": 4.477858116543527e-05,
      "loss": 0.5694108009338379,
      "mean_token_accuracy": 0.8507606893777847,
      "num_tokens": 27983373.0,
      "step": 13690
    },
    {
      "entropy": 0.5920074631460011,
      "epoch": 0.22250554233695785,
      "grad_norm": 8.5,
      "learning_rate": 4.477064788101886e-05,
      "loss": 0.6683462619781494,
      "mean_token_accuracy": 0.8407866597175598,
      "num_tokens": 28003489.0,
      "step": 13700
    },
    {
      "entropy": 0.6349648659117519,
      "epoch": 0.22266795514158336,
      "grad_norm": 7.8125,
      "learning_rate": 4.4762709278253734e-05,
      "loss": 0.6356976509094239,
      "mean_token_accuracy": 0.8409346073865891,
      "num_tokens": 28023105.0,
      "step": 13710
    },
    {
      "entropy": 0.5340277361683547,
      "epoch": 0.2228303679462089,
      "grad_norm": 7.125,
      "learning_rate": 4.47547653592754e-05,
      "loss": 0.5230738639831543,
      "mean_token_accuracy": 0.8633794024586677,
      "num_tokens": 28044186.0,
      "step": 13720
    },
    {
      "entropy": 0.6415080146864056,
      "epoch": 0.2229927807508344,
      "grad_norm": 9.0625,
      "learning_rate": 4.474681612622078e-05,
      "loss": 0.6102344036102295,
      "mean_token_accuracy": 0.8445476673543453,
      "num_tokens": 28065539.0,
      "step": 13730
    },
    {
      "entropy": 0.616783452546224,
      "epoch": 0.2231551935554599,
      "grad_norm": 8.375,
      "learning_rate": 4.4738861581228255e-05,
      "loss": 0.6026972770690918,
      "mean_token_accuracy": 0.8515866793692112,
      "num_tokens": 28086505.0,
      "step": 13740
    },
    {
      "entropy": 0.5939816577825695,
      "epoch": 0.22331760636008544,
      "grad_norm": 8.9375,
      "learning_rate": 4.47309017264376e-05,
      "loss": 0.5734948635101318,
      "mean_token_accuracy": 0.8500930014997721,
      "num_tokens": 28106588.0,
      "step": 13750
    },
    {
      "entropy": 0.5319269943749532,
      "epoch": 0.22348001916471094,
      "grad_norm": 12.0,
      "learning_rate": 4.4722936563990034e-05,
      "loss": 0.5489365100860596,
      "mean_token_accuracy": 0.8652431167662143,
      "num_tokens": 28126403.0,
      "step": 13760
    },
    {
      "entropy": 0.61227259747684,
      "epoch": 0.22364243196933647,
      "grad_norm": 6.5625,
      "learning_rate": 4.471496609602821e-05,
      "loss": 0.6666303157806397,
      "mean_token_accuracy": 0.834812230616808,
      "num_tokens": 28146706.0,
      "step": 13770
    },
    {
      "entropy": 0.5662860034964978,
      "epoch": 0.22380484477396198,
      "grad_norm": 8.1875,
      "learning_rate": 4.4706990324696206e-05,
      "loss": 0.5735319137573243,
      "mean_token_accuracy": 0.849877105653286,
      "num_tokens": 28168122.0,
      "step": 13780
    },
    {
      "entropy": 0.5746642330661416,
      "epoch": 0.2239672575785875,
      "grad_norm": 6.84375,
      "learning_rate": 4.4699009252139514e-05,
      "loss": 0.531390380859375,
      "mean_token_accuracy": 0.864101642370224,
      "num_tokens": 28189353.0,
      "step": 13790
    },
    {
      "entropy": 0.5412967998534441,
      "epoch": 0.22412967038321302,
      "grad_norm": 9.125,
      "learning_rate": 4.469102288050506e-05,
      "loss": 0.6026793956756592,
      "mean_token_accuracy": 0.8498197734355927,
      "num_tokens": 28210693.0,
      "step": 13800
    },
    {
      "entropy": 0.5952402434311808,
      "epoch": 0.22429208318783853,
      "grad_norm": 7.5625,
      "learning_rate": 4.468303121194119e-05,
      "loss": 0.644437599182129,
      "mean_token_accuracy": 0.8459443241357804,
      "num_tokens": 28230354.0,
      "step": 13810
    },
    {
      "entropy": 0.618439129460603,
      "epoch": 0.22445449599246406,
      "grad_norm": 6.40625,
      "learning_rate": 4.46750342485977e-05,
      "loss": 0.5595462322235107,
      "mean_token_accuracy": 0.8510464958846569,
      "num_tokens": 28250690.0,
      "step": 13820
    },
    {
      "entropy": 0.5725077806739136,
      "epoch": 0.22461690879708957,
      "grad_norm": 7.28125,
      "learning_rate": 4.466703199262578e-05,
      "loss": 0.5617525100708007,
      "mean_token_accuracy": 0.8542746573686599,
      "num_tokens": 28270725.0,
      "step": 13830
    },
    {
      "entropy": 0.6076036158949136,
      "epoch": 0.22477932160171507,
      "grad_norm": 6.53125,
      "learning_rate": 4.4659024446178054e-05,
      "loss": 0.6088498592376709,
      "mean_token_accuracy": 0.8498963922262192,
      "num_tokens": 28290057.0,
      "step": 13840
    },
    {
      "entropy": 0.6049760754220188,
      "epoch": 0.2249417344063406,
      "grad_norm": 7.03125,
      "learning_rate": 4.465101161140857e-05,
      "loss": 0.6366272449493409,
      "mean_token_accuracy": 0.8464096069335938,
      "num_tokens": 28311388.0,
      "step": 13850
    },
    {
      "entropy": 0.5702020277967677,
      "epoch": 0.2251041472109661,
      "grad_norm": 6.5625,
      "learning_rate": 4.4642993490472795e-05,
      "loss": 0.5936555385589599,
      "mean_token_accuracy": 0.8497156903147698,
      "num_tokens": 28332070.0,
      "step": 13860
    },
    {
      "entropy": 0.5557423291727901,
      "epoch": 0.22526656001559162,
      "grad_norm": 8.6875,
      "learning_rate": 4.4634970085527634e-05,
      "loss": 0.5669330596923828,
      "mean_token_accuracy": 0.861674465239048,
      "num_tokens": 28351696.0,
      "step": 13870
    },
    {
      "entropy": 0.6310632089152932,
      "epoch": 0.22542897282021715,
      "grad_norm": 8.0,
      "learning_rate": 4.462694139873139e-05,
      "loss": 0.6187809944152832,
      "mean_token_accuracy": 0.8501496009528637,
      "num_tokens": 28371205.0,
      "step": 13880
    },
    {
      "entropy": 0.5907484350726009,
      "epoch": 0.22559138562484266,
      "grad_norm": 9.375,
      "learning_rate": 4.46189074322438e-05,
      "loss": 0.5945881366729736,
      "mean_token_accuracy": 0.8455508224666118,
      "num_tokens": 28391529.0,
      "step": 13890
    },
    {
      "entropy": 0.5764099372550845,
      "epoch": 0.2257537984294682,
      "grad_norm": 7.625,
      "learning_rate": 4.461086818822602e-05,
      "loss": 0.565182113647461,
      "mean_token_accuracy": 0.8497063960880041,
      "num_tokens": 28411163.0,
      "step": 13900
    },
    {
      "entropy": 0.5862484039738775,
      "epoch": 0.2259162112340937,
      "grad_norm": 10.6875,
      "learning_rate": 4.460282366884063e-05,
      "loss": 0.5830499649047851,
      "mean_token_accuracy": 0.8527098469436168,
      "num_tokens": 28432416.0,
      "step": 13910
    },
    {
      "entropy": 0.5945220396388322,
      "epoch": 0.2260786240387192,
      "grad_norm": 7.3125,
      "learning_rate": 4.459477387625161e-05,
      "loss": 0.6214711189270019,
      "mean_token_accuracy": 0.8454022675752639,
      "num_tokens": 28451930.0,
      "step": 13920
    },
    {
      "entropy": 0.6161815460771323,
      "epoch": 0.22624103684334473,
      "grad_norm": 8.9375,
      "learning_rate": 4.458671881262438e-05,
      "loss": 0.6246932983398438,
      "mean_token_accuracy": 0.8488273587077856,
      "num_tokens": 28472505.0,
      "step": 13930
    },
    {
      "entropy": 0.5473623688332736,
      "epoch": 0.22640344964797024,
      "grad_norm": 11.75,
      "learning_rate": 4.457865848012578e-05,
      "loss": 0.5360188484191895,
      "mean_token_accuracy": 0.8589949645102024,
      "num_tokens": 28491661.0,
      "step": 13940
    },
    {
      "entropy": 0.5951306293718517,
      "epoch": 0.22656586245259577,
      "grad_norm": 6.53125,
      "learning_rate": 4.4570592880924045e-05,
      "loss": 0.6291618347167969,
      "mean_token_accuracy": 0.8427003055810929,
      "num_tokens": 28512636.0,
      "step": 13950
    },
    {
      "entropy": 0.5335546204354614,
      "epoch": 0.22672827525722128,
      "grad_norm": 7.40625,
      "learning_rate": 4.4562522017188846e-05,
      "loss": 0.5278321743011475,
      "mean_token_accuracy": 0.8717438161373139,
      "num_tokens": 28533536.0,
      "step": 13960
    },
    {
      "entropy": 0.563526123133488,
      "epoch": 0.22689068806184678,
      "grad_norm": 8.25,
      "learning_rate": 4.4554445891091256e-05,
      "loss": 0.5699947834014892,
      "mean_token_accuracy": 0.858772262185812,
      "num_tokens": 28554248.0,
      "step": 13970
    },
    {
      "entropy": 0.5539604137651623,
      "epoch": 0.22705310086647232,
      "grad_norm": 10.125,
      "learning_rate": 4.454636450480377e-05,
      "loss": 0.536622142791748,
      "mean_token_accuracy": 0.8501306846737862,
      "num_tokens": 28573328.0,
      "step": 13980
    },
    {
      "entropy": 0.5269551783800125,
      "epoch": 0.22721551367109782,
      "grad_norm": 9.125,
      "learning_rate": 4.453827786050031e-05,
      "loss": 0.5000077724456787,
      "mean_token_accuracy": 0.8657857678830624,
      "num_tokens": 28594484.0,
      "step": 13990
    },
    {
      "entropy": 0.6549689246341586,
      "epoch": 0.22737792647572336,
      "grad_norm": 6.25,
      "learning_rate": 4.45301859603562e-05,
      "loss": 0.7027140140533448,
      "mean_token_accuracy": 0.8267220567911864,
      "num_tokens": 28615269.0,
      "step": 14000
    },
    {
      "entropy": 0.6449488624930382,
      "epoch": 0.22754033928034886,
      "grad_norm": 7.75,
      "learning_rate": 4.4522088806548166e-05,
      "loss": 0.6412299633026123,
      "mean_token_accuracy": 0.840188154950738,
      "num_tokens": 28634914.0,
      "step": 14010
    },
    {
      "entropy": 0.6314699687995017,
      "epoch": 0.22770275208497437,
      "grad_norm": 7.84375,
      "learning_rate": 4.4513986401254384e-05,
      "loss": 0.6465074062347412,
      "mean_token_accuracy": 0.8414154484868049,
      "num_tokens": 28656289.0,
      "step": 14020
    },
    {
      "entropy": 0.5783427474088967,
      "epoch": 0.2278651648895999,
      "grad_norm": 7.78125,
      "learning_rate": 4.4505878746654396e-05,
      "loss": 0.5527902126312256,
      "mean_token_accuracy": 0.8667437300086022,
      "num_tokens": 28676809.0,
      "step": 14030
    },
    {
      "entropy": 0.5383872201666235,
      "epoch": 0.2280275776942254,
      "grad_norm": 7.40625,
      "learning_rate": 4.4497765844929185e-05,
      "loss": 0.5562599658966064,
      "mean_token_accuracy": 0.8558428712189198,
      "num_tokens": 28697767.0,
      "step": 14040
    },
    {
      "entropy": 0.5104854839853943,
      "epoch": 0.22818999049885094,
      "grad_norm": 6.625,
      "learning_rate": 4.448964769826115e-05,
      "loss": 0.5210018634796143,
      "mean_token_accuracy": 0.8615947142243385,
      "num_tokens": 28717567.0,
      "step": 14050
    },
    {
      "entropy": 0.5568932576104999,
      "epoch": 0.22835240330347645,
      "grad_norm": 11.25,
      "learning_rate": 4.448152430883408e-05,
      "loss": 0.5360284805297851,
      "mean_token_accuracy": 0.8587586566805839,
      "num_tokens": 28736710.0,
      "step": 14060
    },
    {
      "entropy": 0.537816553749144,
      "epoch": 0.22851481610810195,
      "grad_norm": 7.53125,
      "learning_rate": 4.447339567883319e-05,
      "loss": 0.5984487533569336,
      "mean_token_accuracy": 0.8550431087613106,
      "num_tokens": 28756469.0,
      "step": 14070
    },
    {
      "entropy": 0.5987397157587111,
      "epoch": 0.2286772289127275,
      "grad_norm": 5.625,
      "learning_rate": 4.4465261810445104e-05,
      "loss": 0.6542885780334473,
      "mean_token_accuracy": 0.8517462208867073,
      "num_tokens": 28776940.0,
      "step": 14080
    },
    {
      "entropy": 0.5673832355998456,
      "epoch": 0.228839641717353,
      "grad_norm": 9.375,
      "learning_rate": 4.4457122705857836e-05,
      "loss": 0.5409218311309815,
      "mean_token_accuracy": 0.8584367960691452,
      "num_tokens": 28796051.0,
      "step": 14090
    },
    {
      "entropy": 0.6524776551872492,
      "epoch": 0.22900205452197853,
      "grad_norm": 8.5,
      "learning_rate": 4.444897836726084e-05,
      "loss": 0.7008672714233398,
      "mean_token_accuracy": 0.837493521720171,
      "num_tokens": 28816015.0,
      "step": 14100
    },
    {
      "entropy": 0.6407729657832533,
      "epoch": 0.22916446732660403,
      "grad_norm": 7.21875,
      "learning_rate": 4.4440828796844944e-05,
      "loss": 0.6540714263916015,
      "mean_token_accuracy": 0.8395816773176193,
      "num_tokens": 28836752.0,
      "step": 14110
    },
    {
      "entropy": 0.6154855888802558,
      "epoch": 0.22932688013122954,
      "grad_norm": 7.03125,
      "learning_rate": 4.443267399680242e-05,
      "loss": 0.6000308990478516,
      "mean_token_accuracy": 0.852210120856762,
      "num_tokens": 28858320.0,
      "step": 14120
    },
    {
      "entropy": 0.6484136031009257,
      "epoch": 0.22948929293585507,
      "grad_norm": 8.5625,
      "learning_rate": 4.442451396932692e-05,
      "loss": 0.637156629562378,
      "mean_token_accuracy": 0.843451027572155,
      "num_tokens": 28877612.0,
      "step": 14130
    },
    {
      "entropy": 0.5736938045360148,
      "epoch": 0.22965170574048058,
      "grad_norm": 9.4375,
      "learning_rate": 4.441634871661351e-05,
      "loss": 0.5833540916442871,
      "mean_token_accuracy": 0.8545294493436814,
      "num_tokens": 28897519.0,
      "step": 14140
    },
    {
      "entropy": 0.6414812512695789,
      "epoch": 0.22981411854510608,
      "grad_norm": 10.0,
      "learning_rate": 4.4408178240858656e-05,
      "loss": 0.6741921424865722,
      "mean_token_accuracy": 0.837662261724472,
      "num_tokens": 28917929.0,
      "step": 14150
    },
    {
      "entropy": 0.5852225442416966,
      "epoch": 0.22997653134973162,
      "grad_norm": 9.25,
      "learning_rate": 4.4400002544260235e-05,
      "loss": 0.582069206237793,
      "mean_token_accuracy": 0.8444306261837482,
      "num_tokens": 28938370.0,
      "step": 14160
    },
    {
      "entropy": 0.5976995093282312,
      "epoch": 0.23013894415435712,
      "grad_norm": 10.5,
      "learning_rate": 4.439182162901753e-05,
      "loss": 0.6008153915405273,
      "mean_token_accuracy": 0.8453587353229522,
      "num_tokens": 28958406.0,
      "step": 14170
    },
    {
      "entropy": 0.6024293678812682,
      "epoch": 0.23030135695898266,
      "grad_norm": 5.84375,
      "learning_rate": 4.438363549733123e-05,
      "loss": 0.6366767406463623,
      "mean_token_accuracy": 0.8400787223130465,
      "num_tokens": 28980534.0,
      "step": 14180
    },
    {
      "entropy": 0.5903919907286763,
      "epoch": 0.23046376976360816,
      "grad_norm": 6.53125,
      "learning_rate": 4.437544415140342e-05,
      "loss": 0.5683595657348632,
      "mean_token_accuracy": 0.8524060562252999,
      "num_tokens": 29001215.0,
      "step": 14190
    },
    {
      "entropy": 0.6237150091677904,
      "epoch": 0.23062618256823367,
      "grad_norm": 7.125,
      "learning_rate": 4.436724759343759e-05,
      "loss": 0.6340179920196534,
      "mean_token_accuracy": 0.8495498407632113,
      "num_tokens": 29023828.0,
      "step": 14200
    },
    {
      "entropy": 0.6070982997305692,
      "epoch": 0.2307885953728592,
      "grad_norm": 5.6875,
      "learning_rate": 4.435904582563863e-05,
      "loss": 0.5932548522949219,
      "mean_token_accuracy": 0.8436196200549603,
      "num_tokens": 29043421.0,
      "step": 14210
    },
    {
      "entropy": 0.5533425382804126,
      "epoch": 0.2309510081774847,
      "grad_norm": 7.3125,
      "learning_rate": 4.435083885021284e-05,
      "loss": 0.5561661720275879,
      "mean_token_accuracy": 0.8636214017868042,
      "num_tokens": 29063319.0,
      "step": 14220
    },
    {
      "entropy": 0.5582830990664661,
      "epoch": 0.23111342098211024,
      "grad_norm": 10.5625,
      "learning_rate": 4.4342626669367904e-05,
      "loss": 0.5643480777740478,
      "mean_token_accuracy": 0.8621426865458488,
      "num_tokens": 29083152.0,
      "step": 14230
    },
    {
      "entropy": 0.5823808816727251,
      "epoch": 0.23127583378673575,
      "grad_norm": 7.25,
      "learning_rate": 4.433440928531293e-05,
      "loss": 0.6051177501678466,
      "mean_token_accuracy": 0.8514732748270035,
      "num_tokens": 29103696.0,
      "step": 14240
    },
    {
      "entropy": 0.5406205859966576,
      "epoch": 0.23143824659136125,
      "grad_norm": 6.9375,
      "learning_rate": 4.432618670025841e-05,
      "loss": 0.5925899982452393,
      "mean_token_accuracy": 0.8534427121281624,
      "num_tokens": 29124042.0,
      "step": 14250
    },
    {
      "entropy": 0.6258663108572364,
      "epoch": 0.23160065939598679,
      "grad_norm": 9.1875,
      "learning_rate": 4.4317958916416225e-05,
      "loss": 0.6602246284484863,
      "mean_token_accuracy": 0.8395012296736241,
      "num_tokens": 29144428.0,
      "step": 14260
    },
    {
      "entropy": 0.5573412707075477,
      "epoch": 0.2317630722006123,
      "grad_norm": 6.5625,
      "learning_rate": 4.430972593599968e-05,
      "loss": 0.5324358940124512,
      "mean_token_accuracy": 0.8588787525892257,
      "num_tokens": 29166001.0,
      "step": 14270
    },
    {
      "entropy": 0.6020629708655179,
      "epoch": 0.23192548500523782,
      "grad_norm": 7.28125,
      "learning_rate": 4.430148776122347e-05,
      "loss": 0.5773401260375977,
      "mean_token_accuracy": 0.8549735575914383,
      "num_tokens": 29185809.0,
      "step": 14280
    },
    {
      "entropy": 0.5154556749388576,
      "epoch": 0.23208789780986333,
      "grad_norm": 11.8125,
      "learning_rate": 4.429324439430367e-05,
      "loss": 0.5672462463378907,
      "mean_token_accuracy": 0.8564390800893307,
      "num_tokens": 29206174.0,
      "step": 14290
    },
    {
      "entropy": 0.6067350013181567,
      "epoch": 0.23225031061448884,
      "grad_norm": 8.625,
      "learning_rate": 4.428499583745777e-05,
      "loss": 0.6073486328125,
      "mean_token_accuracy": 0.8483822628855705,
      "num_tokens": 29226890.0,
      "step": 14300
    },
    {
      "entropy": 0.621578310802579,
      "epoch": 0.23241272341911437,
      "grad_norm": 9.5,
      "learning_rate": 4.4276742092904644e-05,
      "loss": 0.5912760257720947,
      "mean_token_accuracy": 0.8459760159254074,
      "num_tokens": 29245624.0,
      "step": 14310
    },
    {
      "entropy": 0.5951930285431445,
      "epoch": 0.23257513622373988,
      "grad_norm": 8.0625,
      "learning_rate": 4.426848316286457e-05,
      "loss": 0.632740592956543,
      "mean_token_accuracy": 0.8413855493068695,
      "num_tokens": 29265252.0,
      "step": 14320
    },
    {
      "entropy": 0.6244874449912459,
      "epoch": 0.2327375490283654,
      "grad_norm": 7.75,
      "learning_rate": 4.426021904955924e-05,
      "loss": 0.6067125797271729,
      "mean_token_accuracy": 0.8513967491686344,
      "num_tokens": 29285241.0,
      "step": 14330
    },
    {
      "entropy": 0.6195795141160488,
      "epoch": 0.23289996183299091,
      "grad_norm": 7.0,
      "learning_rate": 4.4251949755211685e-05,
      "loss": 0.6409644603729248,
      "mean_token_accuracy": 0.8410849574953317,
      "num_tokens": 29305003.0,
      "step": 14340
    },
    {
      "entropy": 0.5860907919704914,
      "epoch": 0.23306237463761642,
      "grad_norm": 7.75,
      "learning_rate": 4.424367528204638e-05,
      "loss": 0.5434869766235352,
      "mean_token_accuracy": 0.8593247652053833,
      "num_tokens": 29326024.0,
      "step": 14350
    },
    {
      "entropy": 0.6002020202111453,
      "epoch": 0.23322478744224195,
      "grad_norm": 10.6875,
      "learning_rate": 4.423539563228918e-05,
      "loss": 0.6290452957153321,
      "mean_token_accuracy": 0.8493446066975594,
      "num_tokens": 29347159.0,
      "step": 14360
    },
    {
      "entropy": 0.5403524368070066,
      "epoch": 0.23338720024686746,
      "grad_norm": 10.5625,
      "learning_rate": 4.422711080816733e-05,
      "loss": 0.567730188369751,
      "mean_token_accuracy": 0.8518587246537208,
      "num_tokens": 29367947.0,
      "step": 14370
    },
    {
      "entropy": 0.5872438973747194,
      "epoch": 0.233549613051493,
      "grad_norm": 7.09375,
      "learning_rate": 4.4218820811909465e-05,
      "loss": 0.6316866874694824,
      "mean_token_accuracy": 0.8442548219114542,
      "num_tokens": 29386509.0,
      "step": 14380
    },
    {
      "entropy": 0.5720058725215494,
      "epoch": 0.2337120258561185,
      "grad_norm": 10.0625,
      "learning_rate": 4.4210525645745604e-05,
      "loss": 0.5692018032073974,
      "mean_token_accuracy": 0.8552003227174282,
      "num_tokens": 29407702.0,
      "step": 14390
    },
    {
      "entropy": 0.5787298213690519,
      "epoch": 0.233874438660744,
      "grad_norm": 8.0,
      "learning_rate": 4.420222531190717e-05,
      "loss": 0.6025168418884277,
      "mean_token_accuracy": 0.8427248656749725,
      "num_tokens": 29428029.0,
      "step": 14400
    },
    {
      "entropy": 0.5485108428634703,
      "epoch": 0.23403685146536954,
      "grad_norm": 8.1875,
      "learning_rate": 4.419391981262697e-05,
      "loss": 0.5677234172821045,
      "mean_token_accuracy": 0.8596973657608032,
      "num_tokens": 29448232.0,
      "step": 14410
    },
    {
      "entropy": 0.7251012403517961,
      "epoch": 0.23419926426999504,
      "grad_norm": 9.375,
      "learning_rate": 4.4185609150139216e-05,
      "loss": 0.7318049430847168,
      "mean_token_accuracy": 0.8250875841826201,
      "num_tokens": 29470253.0,
      "step": 14420
    },
    {
      "entropy": 0.5567233490757644,
      "epoch": 0.23436167707462055,
      "grad_norm": 8.25,
      "learning_rate": 4.4177293326679475e-05,
      "loss": 0.5465403079986573,
      "mean_token_accuracy": 0.8551524549722671,
      "num_tokens": 29490684.0,
      "step": 14430
    },
    {
      "entropy": 0.5763582966290415,
      "epoch": 0.23452408987924608,
      "grad_norm": 5.6875,
      "learning_rate": 4.4168972344484736e-05,
      "loss": 0.6043893337249756,
      "mean_token_accuracy": 0.8444477461278439,
      "num_tokens": 29512303.0,
      "step": 14440
    },
    {
      "entropy": 0.5581884070299565,
      "epoch": 0.2346865026838716,
      "grad_norm": 10.5625,
      "learning_rate": 4.4160646205793354e-05,
      "loss": 0.5740495204925538,
      "mean_token_accuracy": 0.8584456026554108,
      "num_tokens": 29532583.0,
      "step": 14450
    },
    {
      "entropy": 0.5767486054450274,
      "epoch": 0.23484891548849712,
      "grad_norm": 6.96875,
      "learning_rate": 4.415231491284508e-05,
      "loss": 0.5999647617340088,
      "mean_token_accuracy": 0.8478722974658013,
      "num_tokens": 29552523.0,
      "step": 14460
    },
    {
      "entropy": 0.5878359829541295,
      "epoch": 0.23501132829312263,
      "grad_norm": 9.6875,
      "learning_rate": 4.4143978467881053e-05,
      "loss": 0.5972198009490967,
      "mean_token_accuracy": 0.840363047271967,
      "num_tokens": 29573147.0,
      "step": 14470
    },
    {
      "entropy": 0.6153153183870017,
      "epoch": 0.23517374109774813,
      "grad_norm": 6.71875,
      "learning_rate": 4.41356368731438e-05,
      "loss": 0.6012801170349121,
      "mean_token_accuracy": 0.8452833637595176,
      "num_tokens": 29593853.0,
      "step": 14480
    },
    {
      "entropy": 0.5837062304839492,
      "epoch": 0.23533615390237367,
      "grad_norm": 8.75,
      "learning_rate": 4.412729013087722e-05,
      "loss": 0.5403291702270507,
      "mean_token_accuracy": 0.8603835113346576,
      "num_tokens": 29613245.0,
      "step": 14490
    },
    {
      "entropy": 0.5337497735396027,
      "epoch": 0.23549856670699917,
      "grad_norm": 9.9375,
      "learning_rate": 4.4118938243326605e-05,
      "loss": 0.5492832660675049,
      "mean_token_accuracy": 0.8583649463951588,
      "num_tokens": 29633775.0,
      "step": 14500
    },
    {
      "entropy": 0.597385904379189,
      "epoch": 0.2356609795116247,
      "grad_norm": 10.5625,
      "learning_rate": 4.411058121273863e-05,
      "loss": 0.5671964645385742,
      "mean_token_accuracy": 0.8593403846025467,
      "num_tokens": 29655573.0,
      "step": 14510
    },
    {
      "entropy": 0.5606032523326576,
      "epoch": 0.2358233923162502,
      "grad_norm": 6.21875,
      "learning_rate": 4.4102219041361376e-05,
      "loss": 0.5972947120666504,
      "mean_token_accuracy": 0.843195577710867,
      "num_tokens": 29675104.0,
      "step": 14520
    },
    {
      "entropy": 0.5573966089636088,
      "epoch": 0.23598580512087572,
      "grad_norm": 7.1875,
      "learning_rate": 4.4093851731444264e-05,
      "loss": 0.5859920978546143,
      "mean_token_accuracy": 0.8542788952589035,
      "num_tokens": 29694977.0,
      "step": 14530
    },
    {
      "entropy": 0.5382746668532491,
      "epoch": 0.23614821792550125,
      "grad_norm": 5.46875,
      "learning_rate": 4.408547928523812e-05,
      "loss": 0.5236104488372803,
      "mean_token_accuracy": 0.8660301841795445,
      "num_tokens": 29714440.0,
      "step": 14540
    },
    {
      "entropy": 0.5802041619084776,
      "epoch": 0.23631063073012676,
      "grad_norm": 5.5,
      "learning_rate": 4.4077101704995166e-05,
      "loss": 0.5348324298858642,
      "mean_token_accuracy": 0.862752340734005,
      "num_tokens": 29734946.0,
      "step": 14550
    },
    {
      "entropy": 0.5489219477400183,
      "epoch": 0.2364730435347523,
      "grad_norm": 9.4375,
      "learning_rate": 4.4068718992968975e-05,
      "loss": 0.5289930820465087,
      "mean_token_accuracy": 0.8595574207603931,
      "num_tokens": 29754165.0,
      "step": 14560
    },
    {
      "entropy": 0.5149812425952405,
      "epoch": 0.2366354563393778,
      "grad_norm": 6.625,
      "learning_rate": 4.4060331151414526e-05,
      "loss": 0.5386037826538086,
      "mean_token_accuracy": 0.8643578045070172,
      "num_tokens": 29775100.0,
      "step": 14570
    },
    {
      "entropy": 0.5464812146499753,
      "epoch": 0.2367978691440033,
      "grad_norm": 6.3125,
      "learning_rate": 4.405193818258815e-05,
      "loss": 0.5282162666320801,
      "mean_token_accuracy": 0.8610968567430973,
      "num_tokens": 29795599.0,
      "step": 14580
    },
    {
      "entropy": 0.5647190879099071,
      "epoch": 0.23696028194862884,
      "grad_norm": 10.0,
      "learning_rate": 4.4043540088747595e-05,
      "loss": 0.6285149574279785,
      "mean_token_accuracy": 0.8476939555257559,
      "num_tokens": 29816060.0,
      "step": 14590
    },
    {
      "entropy": 0.5409979239106179,
      "epoch": 0.23712269475325434,
      "grad_norm": 7.75,
      "learning_rate": 4.403513687215195e-05,
      "loss": 0.5370771884918213,
      "mean_token_accuracy": 0.8598192229866981,
      "num_tokens": 29837333.0,
      "step": 14600
    },
    {
      "entropy": 0.5457262203097344,
      "epoch": 0.23728510755787988,
      "grad_norm": 6.4375,
      "learning_rate": 4.402672853506172e-05,
      "loss": 0.5339758396148682,
      "mean_token_accuracy": 0.8637374684214592,
      "num_tokens": 29858998.0,
      "step": 14610
    },
    {
      "entropy": 0.619594536907971,
      "epoch": 0.23744752036250538,
      "grad_norm": 8.625,
      "learning_rate": 4.4018315079738746e-05,
      "loss": 0.6460004329681397,
      "mean_token_accuracy": 0.8417407542467117,
      "num_tokens": 29878870.0,
      "step": 14620
    },
    {
      "entropy": 0.6194007402285934,
      "epoch": 0.2376099331671309,
      "grad_norm": 10.75,
      "learning_rate": 4.400989650844628e-05,
      "loss": 0.6005556583404541,
      "mean_token_accuracy": 0.8471309795975686,
      "num_tokens": 29900033.0,
      "step": 14630
    },
    {
      "entropy": 0.6226085294503718,
      "epoch": 0.23777234597175642,
      "grad_norm": 8.5,
      "learning_rate": 4.400147282344893e-05,
      "loss": 0.6639465808868408,
      "mean_token_accuracy": 0.8335112281143665,
      "num_tokens": 29921148.0,
      "step": 14640
    },
    {
      "entropy": 0.5869026051834225,
      "epoch": 0.23793475877638193,
      "grad_norm": 7.0625,
      "learning_rate": 4.3993044027012684e-05,
      "loss": 0.5739975452423096,
      "mean_token_accuracy": 0.8650178827345372,
      "num_tokens": 29940508.0,
      "step": 14650
    },
    {
      "entropy": 0.677211913280189,
      "epoch": 0.23809717158100743,
      "grad_norm": 8.6875,
      "learning_rate": 4.3984610121404914e-05,
      "loss": 0.6504720211029053,
      "mean_token_accuracy": 0.8356803823262453,
      "num_tokens": 29963895.0,
      "step": 14660
    },
    {
      "entropy": 0.5961818290874362,
      "epoch": 0.23825958438563297,
      "grad_norm": 9.5625,
      "learning_rate": 4.397617110889435e-05,
      "loss": 0.6467212200164795,
      "mean_token_accuracy": 0.8445644795894622,
      "num_tokens": 29984981.0,
      "step": 14670
    },
    {
      "entropy": 0.5332651942968368,
      "epoch": 0.23842199719025847,
      "grad_norm": 6.90625,
      "learning_rate": 4.3967726991751115e-05,
      "loss": 0.5183073520660401,
      "mean_token_accuracy": 0.8592260695993901,
      "num_tokens": 30005065.0,
      "step": 14680
    },
    {
      "entropy": 0.6033140605315566,
      "epoch": 0.238584409994884,
      "grad_norm": 8.9375,
      "learning_rate": 4.3959277772246695e-05,
      "loss": 0.5964116096496582,
      "mean_token_accuracy": 0.8468216463923455,
      "num_tokens": 30025215.0,
      "step": 14690
    },
    {
      "entropy": 0.542246154230088,
      "epoch": 0.2387468227995095,
      "grad_norm": 7.40625,
      "learning_rate": 4.395082345265394e-05,
      "loss": 0.539665699005127,
      "mean_token_accuracy": 0.8643945179879665,
      "num_tokens": 30043474.0,
      "step": 14700
    },
    {
      "entropy": 0.5461127854883671,
      "epoch": 0.23890923560413502,
      "grad_norm": 7.34375,
      "learning_rate": 4.39423640352471e-05,
      "loss": 0.5873907089233399,
      "mean_token_accuracy": 0.8533617570996285,
      "num_tokens": 30063958.0,
      "step": 14710
    },
    {
      "entropy": 0.6103121068328619,
      "epoch": 0.23907164840876055,
      "grad_norm": 9.375,
      "learning_rate": 4.3933899522301746e-05,
      "loss": 0.6296722888946533,
      "mean_token_accuracy": 0.8477992609143257,
      "num_tokens": 30084235.0,
      "step": 14720
    },
    {
      "entropy": 0.569952345918864,
      "epoch": 0.23923406121338606,
      "grad_norm": 11.0,
      "learning_rate": 4.392542991609488e-05,
      "loss": 0.5820718288421631,
      "mean_token_accuracy": 0.854785966873169,
      "num_tokens": 30104669.0,
      "step": 14730
    },
    {
      "entropy": 0.6363669512793422,
      "epoch": 0.2393964740180116,
      "grad_norm": 9.25,
      "learning_rate": 4.3916955218904833e-05,
      "loss": 0.6531990528106689,
      "mean_token_accuracy": 0.8345799945294857,
      "num_tokens": 30125093.0,
      "step": 14740
    },
    {
      "entropy": 0.609832201898098,
      "epoch": 0.2395588868226371,
      "grad_norm": 10.375,
      "learning_rate": 4.390847543301132e-05,
      "loss": 0.5841368675231934,
      "mean_token_accuracy": 0.8438720520585775,
      "num_tokens": 30145441.0,
      "step": 14750
    },
    {
      "entropy": 0.5882110844366253,
      "epoch": 0.2397212996272626,
      "grad_norm": 9.3125,
      "learning_rate": 4.3899990560695414e-05,
      "loss": 0.5855683326721192,
      "mean_token_accuracy": 0.8536567606031895,
      "num_tokens": 30165982.0,
      "step": 14760
    },
    {
      "entropy": 0.5687101386953145,
      "epoch": 0.23988371243188814,
      "grad_norm": 8.625,
      "learning_rate": 4.389150060423958e-05,
      "loss": 0.5624008178710938,
      "mean_token_accuracy": 0.8569696247577667,
      "num_tokens": 30186338.0,
      "step": 14770
    },
    {
      "entropy": 0.5845256824977696,
      "epoch": 0.24004612523651364,
      "grad_norm": 8.9375,
      "learning_rate": 4.388300556592763e-05,
      "loss": 0.6297461032867432,
      "mean_token_accuracy": 0.8434038307517767,
      "num_tokens": 30206993.0,
      "step": 14780
    },
    {
      "entropy": 0.5873375245369971,
      "epoch": 0.24020853804113917,
      "grad_norm": 6.78125,
      "learning_rate": 4.3874505448044734e-05,
      "loss": 0.5527103424072266,
      "mean_token_accuracy": 0.8617402002215385,
      "num_tokens": 30228128.0,
      "step": 14790
    },
    {
      "entropy": 0.687246706802398,
      "epoch": 0.24037095084576468,
      "grad_norm": 10.1875,
      "learning_rate": 4.386600025287746e-05,
      "loss": 0.6868251323699951,
      "mean_token_accuracy": 0.8385193429887294,
      "num_tokens": 30248950.0,
      "step": 14800
    },
    {
      "entropy": 0.6285386463627219,
      "epoch": 0.2405333636503902,
      "grad_norm": 7.78125,
      "learning_rate": 4.385748998271372e-05,
      "loss": 0.6195930004119873,
      "mean_token_accuracy": 0.8385484501719475,
      "num_tokens": 30268246.0,
      "step": 14810
    },
    {
      "entropy": 0.5239374488592148,
      "epoch": 0.24069577645501572,
      "grad_norm": 6.75,
      "learning_rate": 4.38489746398428e-05,
      "loss": 0.5067069053649902,
      "mean_token_accuracy": 0.8702627994120121,
      "num_tokens": 30287504.0,
      "step": 14820
    },
    {
      "entropy": 0.5879964422434568,
      "epoch": 0.24085818925964123,
      "grad_norm": 6.625,
      "learning_rate": 4.3840454226555336e-05,
      "loss": 0.595723295211792,
      "mean_token_accuracy": 0.8532620728015899,
      "num_tokens": 30308618.0,
      "step": 14830
    },
    {
      "entropy": 0.5661723539233208,
      "epoch": 0.24102060206426676,
      "grad_norm": 8.375,
      "learning_rate": 4.383192874514335e-05,
      "loss": 0.5768740177154541,
      "mean_token_accuracy": 0.8580047070980072,
      "num_tokens": 30329468.0,
      "step": 14840
    },
    {
      "entropy": 0.560481853177771,
      "epoch": 0.24118301486889226,
      "grad_norm": 9.625,
      "learning_rate": 4.38233981979002e-05,
      "loss": 0.5720639705657959,
      "mean_token_accuracy": 0.8582237392663956,
      "num_tokens": 30348891.0,
      "step": 14850
    },
    {
      "entropy": 0.5252339316997677,
      "epoch": 0.24134542767351777,
      "grad_norm": 6.125,
      "learning_rate": 4.381486258712064e-05,
      "loss": 0.5085159301757812,
      "mean_token_accuracy": 0.8672223754227162,
      "num_tokens": 30367424.0,
      "step": 14860
    },
    {
      "entropy": 0.6448436572682112,
      "epoch": 0.2415078404781433,
      "grad_norm": 8.25,
      "learning_rate": 4.380632191510075e-05,
      "loss": 0.6645538806915283,
      "mean_token_accuracy": 0.8306098140776157,
      "num_tokens": 30387894.0,
      "step": 14870
    },
    {
      "entropy": 0.5036853117868304,
      "epoch": 0.2416702532827688,
      "grad_norm": 9.25,
      "learning_rate": 4.379777618413801e-05,
      "loss": 0.5116088390350342,
      "mean_token_accuracy": 0.8658053562045097,
      "num_tokens": 30407310.0,
      "step": 14880
    },
    {
      "entropy": 0.5769014475867152,
      "epoch": 0.24183266608739434,
      "grad_norm": 6.8125,
      "learning_rate": 4.378922539653123e-05,
      "loss": 0.5549660682678222,
      "mean_token_accuracy": 0.856963224709034,
      "num_tokens": 30428070.0,
      "step": 14890
    },
    {
      "entropy": 0.5755767674185336,
      "epoch": 0.24199507889201985,
      "grad_norm": 11.3125,
      "learning_rate": 4.3780669554580586e-05,
      "loss": 0.6151656150817871,
      "mean_token_accuracy": 0.8406831540167332,
      "num_tokens": 30448027.0,
      "step": 14900
    },
    {
      "entropy": 0.5615435445681214,
      "epoch": 0.24215749169664536,
      "grad_norm": 9.1875,
      "learning_rate": 4.377210866058763e-05,
      "loss": 0.5779373645782471,
      "mean_token_accuracy": 0.8570484668016434,
      "num_tokens": 30467607.0,
      "step": 14910
    },
    {
      "entropy": 0.5238499933853745,
      "epoch": 0.2423199045012709,
      "grad_norm": 9.0,
      "learning_rate": 4.3763542716855256e-05,
      "loss": 0.5502255916595459,
      "mean_token_accuracy": 0.8568500109016896,
      "num_tokens": 30487259.0,
      "step": 14920
    },
    {
      "entropy": 0.6081096332520246,
      "epoch": 0.2424823173058964,
      "grad_norm": 9.5,
      "learning_rate": 4.375497172568772e-05,
      "loss": 0.6475711345672608,
      "mean_token_accuracy": 0.8418792761862278,
      "num_tokens": 30508147.0,
      "step": 14930
    },
    {
      "entropy": 0.5225677321199328,
      "epoch": 0.2426447301105219,
      "grad_norm": 6.1875,
      "learning_rate": 4.374639568939065e-05,
      "loss": 0.4899734020233154,
      "mean_token_accuracy": 0.8770474031567573,
      "num_tokens": 30528876.0,
      "step": 14940
    },
    {
      "entropy": 0.6693194425664842,
      "epoch": 0.24280714291514743,
      "grad_norm": 7.40625,
      "learning_rate": 4.3737814610271e-05,
      "loss": 0.6949044227600097,
      "mean_token_accuracy": 0.836475407704711,
      "num_tokens": 30550620.0,
      "step": 14950
    },
    {
      "entropy": 0.6108877833932638,
      "epoch": 0.24296955571977294,
      "grad_norm": 6.6875,
      "learning_rate": 4.372922849063711e-05,
      "loss": 0.589052677154541,
      "mean_token_accuracy": 0.8511971358209849,
      "num_tokens": 30571127.0,
      "step": 14960
    },
    {
      "entropy": 0.585554796922952,
      "epoch": 0.24313196852439847,
      "grad_norm": 11.6875,
      "learning_rate": 4.3720637332798666e-05,
      "loss": 0.590482521057129,
      "mean_token_accuracy": 0.8527560845017433,
      "num_tokens": 30591014.0,
      "step": 14970
    },
    {
      "entropy": 0.6299329937435687,
      "epoch": 0.24329438132902398,
      "grad_norm": 8.625,
      "learning_rate": 4.3712041139066706e-05,
      "loss": 0.6605061054229736,
      "mean_token_accuracy": 0.8374121967703104,
      "num_tokens": 30611768.0,
      "step": 14980
    },
    {
      "entropy": 0.5743547551333904,
      "epoch": 0.24345679413364948,
      "grad_norm": 8.125,
      "learning_rate": 4.370343991175363e-05,
      "loss": 0.5842247486114502,
      "mean_token_accuracy": 0.8541637055575848,
      "num_tokens": 30632433.0,
      "step": 14990
    },
    {
      "entropy": 0.5950376314111054,
      "epoch": 0.24361920693827502,
      "grad_norm": 7.96875,
      "learning_rate": 4.3694833653173185e-05,
      "loss": 0.5478064060211182,
      "mean_token_accuracy": 0.8641412816941738,
      "num_tokens": 30654015.0,
      "step": 15000
    },
    {
      "entropy": 0.6485071200877428,
      "epoch": 0.24378161974290052,
      "grad_norm": 11.3125,
      "learning_rate": 4.368622236564047e-05,
      "loss": 0.606068468093872,
      "mean_token_accuracy": 0.8489415943622589,
      "num_tokens": 30674123.0,
      "step": 15010
    },
    {
      "entropy": 0.5891464309301228,
      "epoch": 0.24394403254752606,
      "grad_norm": 20.875,
      "learning_rate": 4.367760605147195e-05,
      "loss": 0.653269910812378,
      "mean_token_accuracy": 0.839951304346323,
      "num_tokens": 30694037.0,
      "step": 15020
    },
    {
      "entropy": 0.6294226916506886,
      "epoch": 0.24410644535215156,
      "grad_norm": 7.96875,
      "learning_rate": 4.3668984712985424e-05,
      "loss": 0.6595374584197998,
      "mean_token_accuracy": 0.8356666781008244,
      "num_tokens": 30714350.0,
      "step": 15030
    },
    {
      "entropy": 0.538958351034671,
      "epoch": 0.24426885815677707,
      "grad_norm": 6.71875,
      "learning_rate": 4.366035835250005e-05,
      "loss": 0.5581567764282227,
      "mean_token_accuracy": 0.8548636771738529,
      "num_tokens": 30735217.0,
      "step": 15040
    },
    {
      "entropy": 0.6154553745873272,
      "epoch": 0.2444312709614026,
      "grad_norm": 5.40625,
      "learning_rate": 4.3651726972336346e-05,
      "loss": 0.572450828552246,
      "mean_token_accuracy": 0.8533853113651275,
      "num_tokens": 30755471.0,
      "step": 15050
    },
    {
      "entropy": 0.5973708195611834,
      "epoch": 0.2445936837660281,
      "grad_norm": 5.15625,
      "learning_rate": 4.364309057481617e-05,
      "loss": 0.6325802803039551,
      "mean_token_accuracy": 0.8517468959093094,
      "num_tokens": 30776688.0,
      "step": 15060
    },
    {
      "entropy": 0.5826473977416754,
      "epoch": 0.24475609657065364,
      "grad_norm": 7.375,
      "learning_rate": 4.3634449162262725e-05,
      "loss": 0.5725443363189697,
      "mean_token_accuracy": 0.856584595143795,
      "num_tokens": 30795605.0,
      "step": 15070
    },
    {
      "entropy": 0.5394411638379097,
      "epoch": 0.24491850937527915,
      "grad_norm": 7.0,
      "learning_rate": 4.3625802737000586e-05,
      "loss": 0.5030874729156494,
      "mean_token_accuracy": 0.8665475234389305,
      "num_tokens": 30815429.0,
      "step": 15080
    },
    {
      "entropy": 0.5412048373371363,
      "epoch": 0.24508092217990465,
      "grad_norm": 7.34375,
      "learning_rate": 4.361715130135565e-05,
      "loss": 0.5371362686157226,
      "mean_token_accuracy": 0.8615266591310501,
      "num_tokens": 30835451.0,
      "step": 15090
    },
    {
      "entropy": 0.540456778742373,
      "epoch": 0.2452433349845302,
      "grad_norm": 6.8125,
      "learning_rate": 4.360849485765517e-05,
      "loss": 0.5781017303466797,
      "mean_token_accuracy": 0.8527266636490822,
      "num_tokens": 30855324.0,
      "step": 15100
    },
    {
      "entropy": 0.6495681690052152,
      "epoch": 0.2454057477891557,
      "grad_norm": 5.21875,
      "learning_rate": 4.359983340822775e-05,
      "loss": 0.6465807914733886,
      "mean_token_accuracy": 0.8363168716430665,
      "num_tokens": 30876397.0,
      "step": 15110
    },
    {
      "entropy": 0.5944933845661581,
      "epoch": 0.24556816059378123,
      "grad_norm": 8.0,
      "learning_rate": 4.3591166955403356e-05,
      "loss": 0.5981156349182128,
      "mean_token_accuracy": 0.8489713847637177,
      "num_tokens": 30897124.0,
      "step": 15120
    },
    {
      "entropy": 0.6534764861688018,
      "epoch": 0.24573057339840673,
      "grad_norm": 10.3125,
      "learning_rate": 4.3582495501513264e-05,
      "loss": 0.6767714500427247,
      "mean_token_accuracy": 0.8347577914595604,
      "num_tokens": 30918376.0,
      "step": 15130
    },
    {
      "entropy": 0.6362490458413959,
      "epoch": 0.24589298620303224,
      "grad_norm": 8.375,
      "learning_rate": 4.3573819048890116e-05,
      "loss": 0.6539965152740479,
      "mean_token_accuracy": 0.8312241673469544,
      "num_tokens": 30939832.0,
      "step": 15140
    },
    {
      "entropy": 0.5229353215312585,
      "epoch": 0.24605539900765777,
      "grad_norm": 5.28125,
      "learning_rate": 4.356513759986791e-05,
      "loss": 0.5273227691650391,
      "mean_token_accuracy": 0.8652983829379082,
      "num_tokens": 30959956.0,
      "step": 15150
    },
    {
      "entropy": 0.6362977486103774,
      "epoch": 0.24621781181228328,
      "grad_norm": 5.625,
      "learning_rate": 4.3556451156781965e-05,
      "loss": 0.6242793560028076,
      "mean_token_accuracy": 0.8456807862967253,
      "num_tokens": 30980893.0,
      "step": 15160
    },
    {
      "entropy": 0.5384583485312759,
      "epoch": 0.2463802246169088,
      "grad_norm": 8.8125,
      "learning_rate": 4.354775972196895e-05,
      "loss": 0.5819664001464844,
      "mean_token_accuracy": 0.8562564298510551,
      "num_tokens": 31000691.0,
      "step": 15170
    },
    {
      "entropy": 0.6258472930639982,
      "epoch": 0.24654263742153432,
      "grad_norm": 8.6875,
      "learning_rate": 4.3539063297766894e-05,
      "loss": 0.6226630210876465,
      "mean_token_accuracy": 0.8435637213289737,
      "num_tokens": 31021095.0,
      "step": 15180
    },
    {
      "entropy": 0.5226189535111189,
      "epoch": 0.24670505022615982,
      "grad_norm": 7.625,
      "learning_rate": 4.353036188651514e-05,
      "loss": 0.4939117908477783,
      "mean_token_accuracy": 0.8774497926235199,
      "num_tokens": 31040908.0,
      "step": 15190
    },
    {
      "entropy": 0.5515336023643613,
      "epoch": 0.24686746303078536,
      "grad_norm": 8.1875,
      "learning_rate": 4.352165549055439e-05,
      "loss": 0.560255765914917,
      "mean_token_accuracy": 0.8591833896934986,
      "num_tokens": 31061562.0,
      "step": 15200
    },
    {
      "entropy": 0.46942175440490247,
      "epoch": 0.24702987583541086,
      "grad_norm": 10.375,
      "learning_rate": 4.35129441122267e-05,
      "loss": 0.4594561100006104,
      "mean_token_accuracy": 0.8757174961268902,
      "num_tokens": 31080682.0,
      "step": 15210
    },
    {
      "entropy": 0.5413955338299274,
      "epoch": 0.24719228864003637,
      "grad_norm": 7.34375,
      "learning_rate": 4.350422775387543e-05,
      "loss": 0.5352540493011475,
      "mean_token_accuracy": 0.8671479307115078,
      "num_tokens": 31102879.0,
      "step": 15220
    },
    {
      "entropy": 0.6085431596264244,
      "epoch": 0.2473547014446619,
      "grad_norm": 5.40625,
      "learning_rate": 4.349550641784531e-05,
      "loss": 0.6637259960174561,
      "mean_token_accuracy": 0.8344017527997494,
      "num_tokens": 31122888.0,
      "step": 15230
    },
    {
      "entropy": 0.5345024525653571,
      "epoch": 0.2475171142492874,
      "grad_norm": 9.8125,
      "learning_rate": 4.34867801064824e-05,
      "loss": 0.5595914840698242,
      "mean_token_accuracy": 0.8488235265016556,
      "num_tokens": 31142442.0,
      "step": 15240
    },
    {
      "entropy": 0.5410179506987334,
      "epoch": 0.24767952705391294,
      "grad_norm": 5.25,
      "learning_rate": 4.34780488221341e-05,
      "loss": 0.5111274242401123,
      "mean_token_accuracy": 0.8741055347025395,
      "num_tokens": 31163165.0,
      "step": 15250
    },
    {
      "entropy": 0.5568346183747053,
      "epoch": 0.24784193985853845,
      "grad_norm": 8.6875,
      "learning_rate": 4.346931256714913e-05,
      "loss": 0.5668274879455566,
      "mean_token_accuracy": 0.8567698888480664,
      "num_tokens": 31182327.0,
      "step": 15260
    },
    {
      "entropy": 0.5658491785638035,
      "epoch": 0.24800435266316395,
      "grad_norm": 9.25,
      "learning_rate": 4.346057134387758e-05,
      "loss": 0.5809285163879394,
      "mean_token_accuracy": 0.8494610711932182,
      "num_tokens": 31201584.0,
      "step": 15270
    },
    {
      "entropy": 0.5550103968475014,
      "epoch": 0.24816676546778949,
      "grad_norm": 8.875,
      "learning_rate": 4.345182515467084e-05,
      "loss": 0.5920053005218506,
      "mean_token_accuracy": 0.8491007022559642,
      "num_tokens": 31221505.0,
      "step": 15280
    },
    {
      "entropy": 0.5834278243593871,
      "epoch": 0.248329178272415,
      "grad_norm": 6.53125,
      "learning_rate": 4.344307400188168e-05,
      "loss": 0.6301911354064942,
      "mean_token_accuracy": 0.8393901444971561,
      "num_tokens": 31241491.0,
      "step": 15290
    },
    {
      "entropy": 0.5348898220341652,
      "epoch": 0.24849159107704052,
      "grad_norm": 9.9375,
      "learning_rate": 4.343431788786415e-05,
      "loss": 0.5553106307983399,
      "mean_token_accuracy": 0.8565469086170197,
      "num_tokens": 31262902.0,
      "step": 15300
    },
    {
      "entropy": 0.5247830992564559,
      "epoch": 0.24865400388166603,
      "grad_norm": 7.65625,
      "learning_rate": 4.342555681497369e-05,
      "loss": 0.5062832355499267,
      "mean_token_accuracy": 0.8671508803963661,
      "num_tokens": 31283911.0,
      "step": 15310
    },
    {
      "entropy": 0.5091532193124294,
      "epoch": 0.24881641668629154,
      "grad_norm": 5.28125,
      "learning_rate": 4.341679078556704e-05,
      "loss": 0.5440468311309814,
      "mean_token_accuracy": 0.861503129452467,
      "num_tokens": 31303941.0,
      "step": 15320
    },
    {
      "entropy": 0.6267969204112888,
      "epoch": 0.24897882949091707,
      "grad_norm": 11.0,
      "learning_rate": 4.340801980200228e-05,
      "loss": 0.6562600135803223,
      "mean_token_accuracy": 0.8411119185388088,
      "num_tokens": 31323761.0,
      "step": 15330
    },
    {
      "entropy": 0.6203807605430484,
      "epoch": 0.24914124229554258,
      "grad_norm": 9.25,
      "learning_rate": 4.339924386663882e-05,
      "loss": 0.5832053184509277,
      "mean_token_accuracy": 0.8573539555072784,
      "num_tokens": 31344120.0,
      "step": 15340
    },
    {
      "entropy": 0.5713986282236874,
      "epoch": 0.2493036551001681,
      "grad_norm": 10.4375,
      "learning_rate": 4.339046298183741e-05,
      "loss": 0.5910331249237061,
      "mean_token_accuracy": 0.8455503262579441,
      "num_tokens": 31364297.0,
      "step": 15350
    },
    {
      "entropy": 0.5738907265011222,
      "epoch": 0.24946606790479361,
      "grad_norm": 7.21875,
      "learning_rate": 4.338167714996012e-05,
      "loss": 0.5743090152740479,
      "mean_token_accuracy": 0.8490411575883627,
      "num_tokens": 31385435.0,
      "step": 15360
    },
    {
      "entropy": 0.6226803259924054,
      "epoch": 0.24962848070941912,
      "grad_norm": 6.96875,
      "learning_rate": 4.3372886373370376e-05,
      "loss": 0.662813663482666,
      "mean_token_accuracy": 0.8439899906516075,
      "num_tokens": 31407684.0,
      "step": 15370
    },
    {
      "entropy": 0.5530667746439576,
      "epoch": 0.24979089351404465,
      "grad_norm": 9.9375,
      "learning_rate": 4.33640906544329e-05,
      "loss": 0.5643144607543945,
      "mean_token_accuracy": 0.8632739603519439,
      "num_tokens": 31427250.0,
      "step": 15380
    },
    {
      "entropy": 0.6290823282208293,
      "epoch": 0.24995330631867016,
      "grad_norm": 9.5625,
      "learning_rate": 4.335528999551377e-05,
      "loss": 0.6765377998352051,
      "mean_token_accuracy": 0.8345309391617775,
      "num_tokens": 31448177.0,
      "step": 15390
    },
    {
      "entropy": 0.6678023953456431,
      "epoch": 0.25011571912329567,
      "grad_norm": 7.875,
      "learning_rate": 4.3346484398980394e-05,
      "loss": 0.679468297958374,
      "mean_token_accuracy": 0.833036357909441,
      "num_tokens": 31469428.0,
      "step": 15400
    },
    {
      "entropy": 0.4688108302652836,
      "epoch": 0.2502781319279212,
      "grad_norm": 8.6875,
      "learning_rate": 4.333767386720147e-05,
      "loss": 0.4608167171478271,
      "mean_token_accuracy": 0.8769662849605083,
      "num_tokens": 31489377.0,
      "step": 15410
    },
    {
      "entropy": 0.6404851194471121,
      "epoch": 0.25044054473254673,
      "grad_norm": 10.4375,
      "learning_rate": 4.3328858402547065e-05,
      "loss": 0.6063575267791748,
      "mean_token_accuracy": 0.8455529227852822,
      "num_tokens": 31509416.0,
      "step": 15420
    },
    {
      "entropy": 0.503203792963177,
      "epoch": 0.2506029575371722,
      "grad_norm": 6.8125,
      "learning_rate": 4.3320038007388556e-05,
      "loss": 0.5147240161895752,
      "mean_token_accuracy": 0.8650565199553967,
      "num_tokens": 31529232.0,
      "step": 15430
    },
    {
      "entropy": 0.5574098563753068,
      "epoch": 0.25076537034179774,
      "grad_norm": 9.375,
      "learning_rate": 4.331121268409866e-05,
      "loss": 0.5542675971984863,
      "mean_token_accuracy": 0.8587027601897717,
      "num_tokens": 31551073.0,
      "step": 15440
    },
    {
      "entropy": 0.5560150746256113,
      "epoch": 0.2509277831464233,
      "grad_norm": 9.9375,
      "learning_rate": 4.3302382435051393e-05,
      "loss": 0.5500722885131836,
      "mean_token_accuracy": 0.8638024650514126,
      "num_tokens": 31572091.0,
      "step": 15450
    },
    {
      "entropy": 0.6067171305418014,
      "epoch": 0.25109019595104876,
      "grad_norm": 13.3125,
      "learning_rate": 4.3293547262622115e-05,
      "loss": 0.6475485324859619,
      "mean_token_accuracy": 0.8446847170591354,
      "num_tokens": 31592189.0,
      "step": 15460
    },
    {
      "entropy": 0.5441944845020771,
      "epoch": 0.2512526087556743,
      "grad_norm": 8.0,
      "learning_rate": 4.328470716918752e-05,
      "loss": 0.5204062461853027,
      "mean_token_accuracy": 0.8664009988307952,
      "num_tokens": 31612449.0,
      "step": 15470
    },
    {
      "entropy": 0.5215622836723923,
      "epoch": 0.2514150215602998,
      "grad_norm": 8.625,
      "learning_rate": 4.32758621571256e-05,
      "loss": 0.5645112037658692,
      "mean_token_accuracy": 0.8565972276031971,
      "num_tokens": 31633348.0,
      "step": 15480
    },
    {
      "entropy": 0.560841653496027,
      "epoch": 0.25157743436492536,
      "grad_norm": 7.5,
      "learning_rate": 4.326701222881568e-05,
      "loss": 0.5883710861206055,
      "mean_token_accuracy": 0.8478001751005649,
      "num_tokens": 31653528.0,
      "step": 15490
    },
    {
      "entropy": 0.6301664214581251,
      "epoch": 0.25173984716955083,
      "grad_norm": 6.40625,
      "learning_rate": 4.325815738663843e-05,
      "loss": 0.6019854545593262,
      "mean_token_accuracy": 0.8453720815479755,
      "num_tokens": 31674539.0,
      "step": 15500
    },
    {
      "entropy": 0.5082409573718906,
      "epoch": 0.25190225997417637,
      "grad_norm": 9.8125,
      "learning_rate": 4.3249297632975813e-05,
      "loss": 0.49705934524536133,
      "mean_token_accuracy": 0.8691798731684685,
      "num_tokens": 31694592.0,
      "step": 15510
    },
    {
      "entropy": 0.5715178216807544,
      "epoch": 0.2520646727788019,
      "grad_norm": 7.1875,
      "learning_rate": 4.324043297021111e-05,
      "loss": 0.5892731666564941,
      "mean_token_accuracy": 0.8504037477076054,
      "num_tokens": 31714741.0,
      "step": 15520
    },
    {
      "entropy": 0.575499612186104,
      "epoch": 0.2522270855834274,
      "grad_norm": 7.9375,
      "learning_rate": 4.323156340072895e-05,
      "loss": 0.5749993324279785,
      "mean_token_accuracy": 0.8490799978375435,
      "num_tokens": 31734416.0,
      "step": 15530
    },
    {
      "entropy": 0.6137212052941322,
      "epoch": 0.2523894983880529,
      "grad_norm": 6.0,
      "learning_rate": 4.322268892691526e-05,
      "loss": 0.602094316482544,
      "mean_token_accuracy": 0.8485502868890762,
      "num_tokens": 31755722.0,
      "step": 15540
    },
    {
      "entropy": 0.6559651636518538,
      "epoch": 0.25255191119267845,
      "grad_norm": 7.09375,
      "learning_rate": 4.32138095511573e-05,
      "loss": 0.6765952110290527,
      "mean_token_accuracy": 0.8346683736890554,
      "num_tokens": 31775135.0,
      "step": 15550
    },
    {
      "entropy": 0.5045758728403598,
      "epoch": 0.2527143239973039,
      "grad_norm": 9.25,
      "learning_rate": 4.320492527584363e-05,
      "loss": 0.46181230545043944,
      "mean_token_accuracy": 0.8868521563708782,
      "num_tokens": 31795722.0,
      "step": 15560
    },
    {
      "entropy": 0.5339118548668921,
      "epoch": 0.25287673680192946,
      "grad_norm": 6.53125,
      "learning_rate": 4.319603610336416e-05,
      "loss": 0.5583122730255127,
      "mean_token_accuracy": 0.8545616641640663,
      "num_tokens": 31814996.0,
      "step": 15570
    },
    {
      "entropy": 0.4760604371316731,
      "epoch": 0.253039149606555,
      "grad_norm": 6.5,
      "learning_rate": 4.318714203611007e-05,
      "loss": 0.49817795753479005,
      "mean_token_accuracy": 0.8684930555522442,
      "num_tokens": 31834454.0,
      "step": 15580
    },
    {
      "entropy": 0.5157092433888465,
      "epoch": 0.2532015624111805,
      "grad_norm": 8.6875,
      "learning_rate": 4.317824307647391e-05,
      "loss": 0.549840259552002,
      "mean_token_accuracy": 0.8601284757256508,
      "num_tokens": 31854369.0,
      "step": 15590
    },
    {
      "entropy": 0.5532092477194965,
      "epoch": 0.253363975215806,
      "grad_norm": 8.1875,
      "learning_rate": 4.3169339226849505e-05,
      "loss": 0.5727435111999511,
      "mean_token_accuracy": 0.8526933941990137,
      "num_tokens": 31875676.0,
      "step": 15600
    },
    {
      "entropy": 0.5453777751885355,
      "epoch": 0.25352638802043154,
      "grad_norm": 7.625,
      "learning_rate": 4.316043048963202e-05,
      "loss": 0.5265498638153077,
      "mean_token_accuracy": 0.8595675863325596,
      "num_tokens": 31895612.0,
      "step": 15610
    },
    {
      "entropy": 0.5747438267804682,
      "epoch": 0.25368880082505707,
      "grad_norm": 9.0625,
      "learning_rate": 4.315151686721791e-05,
      "loss": 0.5577717781066894,
      "mean_token_accuracy": 0.8612784147262573,
      "num_tokens": 31915229.0,
      "step": 15620
    },
    {
      "entropy": 0.5240500099025667,
      "epoch": 0.25385121362968255,
      "grad_norm": 6.65625,
      "learning_rate": 4.3142598362004985e-05,
      "loss": 0.5302580833435059,
      "mean_token_accuracy": 0.86611467897892,
      "num_tokens": 31934941.0,
      "step": 15630
    },
    {
      "entropy": 0.609642923809588,
      "epoch": 0.2540136264343081,
      "grad_norm": 6.15625,
      "learning_rate": 4.313367497639231e-05,
      "loss": 0.6148351192474365,
      "mean_token_accuracy": 0.8453241415321827,
      "num_tokens": 31955323.0,
      "step": 15640
    },
    {
      "entropy": 0.568718890612945,
      "epoch": 0.2541760392389336,
      "grad_norm": 8.125,
      "learning_rate": 4.312474671278033e-05,
      "loss": 0.5747434616088867,
      "mean_token_accuracy": 0.8550405524671078,
      "num_tokens": 31975137.0,
      "step": 15650
    },
    {
      "entropy": 0.5473157517611981,
      "epoch": 0.2543384520435591,
      "grad_norm": 7.1875,
      "learning_rate": 4.311581357357074e-05,
      "loss": 0.5780817985534668,
      "mean_token_accuracy": 0.8631300315260887,
      "num_tokens": 31994772.0,
      "step": 15660
    },
    {
      "entropy": 0.547796099120751,
      "epoch": 0.2545008648481846,
      "grad_norm": 7.21875,
      "learning_rate": 4.3106875561166596e-05,
      "loss": 0.5117413997650146,
      "mean_token_accuracy": 0.8690782651305199,
      "num_tokens": 32014173.0,
      "step": 15670
    },
    {
      "entropy": 0.5009539962280541,
      "epoch": 0.25466327765281016,
      "grad_norm": 5.6875,
      "learning_rate": 4.309793267797222e-05,
      "loss": 0.48561906814575195,
      "mean_token_accuracy": 0.876363368332386,
      "num_tokens": 32034593.0,
      "step": 15680
    },
    {
      "entropy": 0.6489931442774832,
      "epoch": 0.25482569045743564,
      "grad_norm": 9.1875,
      "learning_rate": 4.308898492639329e-05,
      "loss": 0.6516633987426758,
      "mean_token_accuracy": 0.8318918745964765,
      "num_tokens": 32054941.0,
      "step": 15690
    },
    {
      "entropy": 0.5608673102688044,
      "epoch": 0.25498810326206117,
      "grad_norm": 7.75,
      "learning_rate": 4.308003230883676e-05,
      "loss": 0.5833994388580322,
      "mean_token_accuracy": 0.8530357964336872,
      "num_tokens": 32074973.0,
      "step": 15700
    },
    {
      "entropy": 0.5728492319583893,
      "epoch": 0.2551505160666867,
      "grad_norm": 7.25,
      "learning_rate": 4.30710748277109e-05,
      "loss": 0.5899284362792969,
      "mean_token_accuracy": 0.8559984862804413,
      "num_tokens": 32096871.0,
      "step": 15710
    },
    {
      "entropy": 0.5860084610059857,
      "epoch": 0.25531292887131224,
      "grad_norm": 7.8125,
      "learning_rate": 4.306211248542531e-05,
      "loss": 0.5921046257019043,
      "mean_token_accuracy": 0.856103340536356,
      "num_tokens": 32115329.0,
      "step": 15720
    },
    {
      "entropy": 0.6470977185294032,
      "epoch": 0.2554753416759377,
      "grad_norm": 10.0,
      "learning_rate": 4.3053145284390864e-05,
      "loss": 0.6689095020294189,
      "mean_token_accuracy": 0.825097155943513,
      "num_tokens": 32136804.0,
      "step": 15730
    },
    {
      "entropy": 0.5800309094134718,
      "epoch": 0.25563775448056325,
      "grad_norm": 7.0625,
      "learning_rate": 4.304417322701977e-05,
      "loss": 0.5901936531066895,
      "mean_token_accuracy": 0.8507696568965912,
      "num_tokens": 32157432.0,
      "step": 15740
    },
    {
      "entropy": 0.5346530498936772,
      "epoch": 0.2558001672851888,
      "grad_norm": 11.75,
      "learning_rate": 4.303519631572552e-05,
      "loss": 0.5190016746520996,
      "mean_token_accuracy": 0.8742322601377964,
      "num_tokens": 32177944.0,
      "step": 15750
    },
    {
      "entropy": 0.61239367374219,
      "epoch": 0.25596258008981426,
      "grad_norm": 7.78125,
      "learning_rate": 4.302621455292294e-05,
      "loss": 0.6012380599975586,
      "mean_token_accuracy": 0.8453654490411282,
      "num_tokens": 32199357.0,
      "step": 15760
    },
    {
      "entropy": 0.5522593474946916,
      "epoch": 0.2561249928944398,
      "grad_norm": 6.96875,
      "learning_rate": 4.301722794102812e-05,
      "loss": 0.5532767295837402,
      "mean_token_accuracy": 0.8587029751390218,
      "num_tokens": 32218829.0,
      "step": 15770
    },
    {
      "entropy": 0.5581951828673481,
      "epoch": 0.25628740569906533,
      "grad_norm": 7.03125,
      "learning_rate": 4.30082364824585e-05,
      "loss": 0.5771262168884277,
      "mean_token_accuracy": 0.8547569043934345,
      "num_tokens": 32239384.0,
      "step": 15780
    },
    {
      "entropy": 0.6390960846096277,
      "epoch": 0.2564498185036908,
      "grad_norm": 7.09375,
      "learning_rate": 4.2999240179632794e-05,
      "loss": 0.6417674064636231,
      "mean_token_accuracy": 0.844633562117815,
      "num_tokens": 32259644.0,
      "step": 15790
    },
    {
      "entropy": 0.5922097467817367,
      "epoch": 0.25661223130831634,
      "grad_norm": 8.9375,
      "learning_rate": 4.299023903497102e-05,
      "loss": 0.5955427169799805,
      "mean_token_accuracy": 0.847586727142334,
      "num_tokens": 32280235.0,
      "step": 15800
    },
    {
      "entropy": 0.594960379647091,
      "epoch": 0.2567746441129419,
      "grad_norm": 7.375,
      "learning_rate": 4.298123305089453e-05,
      "loss": 0.5736542701721191,
      "mean_token_accuracy": 0.8510989397764206,
      "num_tokens": 32299739.0,
      "step": 15810
    },
    {
      "entropy": 0.5179244023747742,
      "epoch": 0.2569370569175674,
      "grad_norm": 5.9375,
      "learning_rate": 4.297222222982593e-05,
      "loss": 0.4979219436645508,
      "mean_token_accuracy": 0.8639967389404773,
      "num_tokens": 32322457.0,
      "step": 15820
    },
    {
      "entropy": 0.5297730300575495,
      "epoch": 0.2570994697221929,
      "grad_norm": 6.65625,
      "learning_rate": 4.2963206574189166e-05,
      "loss": 0.5702173233032226,
      "mean_token_accuracy": 0.856246829777956,
      "num_tokens": 32342122.0,
      "step": 15830
    },
    {
      "entropy": 0.6159854920580983,
      "epoch": 0.2572618825268184,
      "grad_norm": 7.71875,
      "learning_rate": 4.295418608640947e-05,
      "loss": 0.5980271339416504,
      "mean_token_accuracy": 0.8491918802261352,
      "num_tokens": 32363420.0,
      "step": 15840
    },
    {
      "entropy": 0.52522870588582,
      "epoch": 0.25742429533144395,
      "grad_norm": 6.65625,
      "learning_rate": 4.294516076891335e-05,
      "loss": 0.5282418251037597,
      "mean_token_accuracy": 0.8686798684298992,
      "num_tokens": 32383569.0,
      "step": 15850
    },
    {
      "entropy": 0.5813529073260725,
      "epoch": 0.25758670813606943,
      "grad_norm": 8.6875,
      "learning_rate": 4.293613062412866e-05,
      "loss": 0.6494122505187988,
      "mean_token_accuracy": 0.8402756504714489,
      "num_tokens": 32403662.0,
      "step": 15860
    },
    {
      "entropy": 0.6148645138368011,
      "epoch": 0.25774912094069496,
      "grad_norm": 8.5,
      "learning_rate": 4.292709565448452e-05,
      "loss": 0.6143770217895508,
      "mean_token_accuracy": 0.8451914139091968,
      "num_tokens": 32425203.0,
      "step": 15870
    },
    {
      "entropy": 0.6489903093315661,
      "epoch": 0.2579115337453205,
      "grad_norm": 10.75,
      "learning_rate": 4.291805586241137e-05,
      "loss": 0.729645824432373,
      "mean_token_accuracy": 0.8307115159928798,
      "num_tokens": 32445964.0,
      "step": 15880
    },
    {
      "entropy": 0.5806382352486252,
      "epoch": 0.258073946549946,
      "grad_norm": 6.0625,
      "learning_rate": 4.290901125034092e-05,
      "loss": 0.5491117477416992,
      "mean_token_accuracy": 0.8596501752734185,
      "num_tokens": 32467185.0,
      "step": 15890
    },
    {
      "entropy": 0.6109069483354688,
      "epoch": 0.2582363593545715,
      "grad_norm": 7.375,
      "learning_rate": 4.2899961820706193e-05,
      "loss": 0.5586971282958985,
      "mean_token_accuracy": 0.8523763608187437,
      "num_tokens": 32487569.0,
      "step": 15900
    },
    {
      "entropy": 0.5413804239593446,
      "epoch": 0.25839877215919704,
      "grad_norm": 6.15625,
      "learning_rate": 4.2890907575941515e-05,
      "loss": 0.5060252666473388,
      "mean_token_accuracy": 0.8706704676151276,
      "num_tokens": 32508128.0,
      "step": 15910
    },
    {
      "entropy": 0.5825108227320015,
      "epoch": 0.2585611849638225,
      "grad_norm": 9.4375,
      "learning_rate": 4.2881848518482485e-05,
      "loss": 0.5634130001068115,
      "mean_token_accuracy": 0.8502587854862214,
      "num_tokens": 32528795.0,
      "step": 15920
    },
    {
      "entropy": 0.48144742781296374,
      "epoch": 0.25872359776844805,
      "grad_norm": 10.1875,
      "learning_rate": 4.287278465076602e-05,
      "loss": 0.49934802055358884,
      "mean_token_accuracy": 0.8735240899026394,
      "num_tokens": 32549627.0,
      "step": 15930
    },
    {
      "entropy": 0.5593880249187351,
      "epoch": 0.2588860105730736,
      "grad_norm": 6.6875,
      "learning_rate": 4.2863715975230315e-05,
      "loss": 0.5780657291412353,
      "mean_token_accuracy": 0.8456660892814398,
      "num_tokens": 32570728.0,
      "step": 15940
    },
    {
      "entropy": 0.5743173694703728,
      "epoch": 0.2590484233776991,
      "grad_norm": 8.0,
      "learning_rate": 4.2854642494314865e-05,
      "loss": 0.5898411273956299,
      "mean_token_accuracy": 0.8542416028678417,
      "num_tokens": 32591104.0,
      "step": 15950
    },
    {
      "entropy": 0.601020597666502,
      "epoch": 0.2592108361823246,
      "grad_norm": 12.5,
      "learning_rate": 4.2845564210460464e-05,
      "loss": 0.6554780960083008,
      "mean_token_accuracy": 0.8408015370368958,
      "num_tokens": 32611629.0,
      "step": 15960
    },
    {
      "entropy": 0.5835861927829683,
      "epoch": 0.25937324898695013,
      "grad_norm": 8.5,
      "learning_rate": 4.283648112610918e-05,
      "loss": 0.5824780464172363,
      "mean_token_accuracy": 0.8554405227303505,
      "num_tokens": 32631774.0,
      "step": 15970
    },
    {
      "entropy": 0.5203159669414162,
      "epoch": 0.25953566179157567,
      "grad_norm": 7.25,
      "learning_rate": 4.282739324370439e-05,
      "loss": 0.5492111682891846,
      "mean_token_accuracy": 0.854703713953495,
      "num_tokens": 32652251.0,
      "step": 15980
    },
    {
      "entropy": 0.6131322188302875,
      "epoch": 0.25969807459620114,
      "grad_norm": 5.59375,
      "learning_rate": 4.281830056569076e-05,
      "loss": 0.6020207405090332,
      "mean_token_accuracy": 0.8518664203584194,
      "num_tokens": 32674966.0,
      "step": 15990
    },
    {
      "epoch": 0.2598604874008267,
      "eval_entropy": 0.6120449800491333,
      "eval_loss": 0.613021194934845,
      "eval_mean_token_accuracy": 0.843826699256897,
      "eval_num_tokens": 32694423.0,
      "eval_runtime": 40.1626,
      "eval_samples_per_second": 49.798,
      "eval_steps_per_second": 6.225,
      "step": 16000
    },
    {
      "entropy": 0.5872263912344351,
      "epoch": 0.2600229002054522,
      "grad_norm": 7.34375,
      "learning_rate": 4.2800100832622034e-05,
      "loss": 0.5916765689849853,
      "mean_token_accuracy": 0.8476166643202305,
      "num_tokens": 32716875.0,
      "step": 16010
    },
    {
      "entropy": 0.5311224710196256,
      "epoch": 0.2601853130100777,
      "grad_norm": 9.9375,
      "learning_rate": 4.279099378246272e-05,
      "loss": 0.5223794460296631,
      "mean_token_accuracy": 0.861445589736104,
      "num_tokens": 32737111.0,
      "step": 16020
    },
    {
      "entropy": 0.5683638142421842,
      "epoch": 0.2603477258147032,
      "grad_norm": 6.5,
      "learning_rate": 4.2781881946486093e-05,
      "loss": 0.5886410236358642,
      "mean_token_accuracy": 0.8489915072917938,
      "num_tokens": 32756461.0,
      "step": 16030
    },
    {
      "entropy": 0.5772123669274152,
      "epoch": 0.26051013861932876,
      "grad_norm": 10.0625,
      "learning_rate": 4.277276532714326e-05,
      "loss": 0.5594343185424805,
      "mean_token_accuracy": 0.8571921654045582,
      "num_tokens": 32776649.0,
      "step": 16040
    },
    {
      "entropy": 0.5909796336665749,
      "epoch": 0.2606725514239543,
      "grad_norm": 7.25,
      "learning_rate": 4.2763643926886615e-05,
      "loss": 0.6415755271911621,
      "mean_token_accuracy": 0.8333929240703583,
      "num_tokens": 32796688.0,
      "step": 16050
    },
    {
      "entropy": 0.5632294069975614,
      "epoch": 0.26083496422857977,
      "grad_norm": 9.3125,
      "learning_rate": 4.275451774816983e-05,
      "loss": 0.5782293796539306,
      "mean_token_accuracy": 0.8559122875332832,
      "num_tokens": 32817121.0,
      "step": 16060
    },
    {
      "entropy": 0.5966997879557312,
      "epoch": 0.2609973770332053,
      "grad_norm": 7.75,
      "learning_rate": 4.2745386793447875e-05,
      "loss": 0.5659688949584961,
      "mean_token_accuracy": 0.8616336017847062,
      "num_tokens": 32836848.0,
      "step": 16070
    },
    {
      "entropy": 0.5465869106352329,
      "epoch": 0.26115978983783084,
      "grad_norm": 5.09375,
      "learning_rate": 4.273625106517699e-05,
      "loss": 0.5423137664794921,
      "mean_token_accuracy": 0.8634883366525173,
      "num_tokens": 32857559.0,
      "step": 16080
    },
    {
      "entropy": 0.5609927339479327,
      "epoch": 0.2613222026424563,
      "grad_norm": 7.125,
      "learning_rate": 4.2727110565814695e-05,
      "loss": 0.5616235256195068,
      "mean_token_accuracy": 0.8476842626929283,
      "num_tokens": 32878441.0,
      "step": 16090
    },
    {
      "entropy": 0.6178541194181889,
      "epoch": 0.26148461544708185,
      "grad_norm": 10.4375,
      "learning_rate": 4.271796529781984e-05,
      "loss": 0.6368340015411377,
      "mean_token_accuracy": 0.842853956669569,
      "num_tokens": 32899307.0,
      "step": 16100
    },
    {
      "entropy": 0.6189182843081653,
      "epoch": 0.2616470282517074,
      "grad_norm": 7.84375,
      "learning_rate": 4.270881526365249e-05,
      "loss": 0.607978105545044,
      "mean_token_accuracy": 0.8537335358560085,
      "num_tokens": 32920933.0,
      "step": 16110
    },
    {
      "entropy": 0.6145026623271406,
      "epoch": 0.26180944105633286,
      "grad_norm": 9.75,
      "learning_rate": 4.269966046577404e-05,
      "loss": 0.625002384185791,
      "mean_token_accuracy": 0.8482311524450779,
      "num_tokens": 32940610.0,
      "step": 16120
    },
    {
      "entropy": 0.5594085396733135,
      "epoch": 0.2619718538609584,
      "grad_norm": 7.125,
      "learning_rate": 4.2690500906647133e-05,
      "loss": 0.556900930404663,
      "mean_token_accuracy": 0.8569698378443718,
      "num_tokens": 32960174.0,
      "step": 16130
    },
    {
      "entropy": 0.5490311951376498,
      "epoch": 0.2621342666655839,
      "grad_norm": 8.875,
      "learning_rate": 4.2681336588735745e-05,
      "loss": 0.5387625694274902,
      "mean_token_accuracy": 0.8680584646761418,
      "num_tokens": 32979413.0,
      "step": 16140
    },
    {
      "entropy": 0.5449624948669225,
      "epoch": 0.2622966794702094,
      "grad_norm": 7.625,
      "learning_rate": 4.267216751450507e-05,
      "loss": 0.5729319095611572,
      "mean_token_accuracy": 0.8505830615758896,
      "num_tokens": 33000883.0,
      "step": 16150
    },
    {
      "entropy": 0.5244334993883968,
      "epoch": 0.26245909227483494,
      "grad_norm": 7.59375,
      "learning_rate": 4.266299368642162e-05,
      "loss": 0.5112378120422363,
      "mean_token_accuracy": 0.8633287761360406,
      "num_tokens": 33022034.0,
      "step": 16160
    },
    {
      "entropy": 0.5062373321503401,
      "epoch": 0.26262150507946047,
      "grad_norm": 7.625,
      "learning_rate": 4.2653815106953165e-05,
      "loss": 0.46835908889770506,
      "mean_token_accuracy": 0.8749705143272877,
      "num_tokens": 33042733.0,
      "step": 16170
    },
    {
      "entropy": 0.571616944577545,
      "epoch": 0.262783917884086,
      "grad_norm": 5.78125,
      "learning_rate": 4.264463177856878e-05,
      "loss": 0.5439795017242431,
      "mean_token_accuracy": 0.8584276922047138,
      "num_tokens": 33062510.0,
      "step": 16180
    },
    {
      "entropy": 0.5502996761351824,
      "epoch": 0.2629463306887115,
      "grad_norm": 8.125,
      "learning_rate": 4.2635443703738784e-05,
      "loss": 0.5287948608398437,
      "mean_token_accuracy": 0.8635528326034546,
      "num_tokens": 33082386.0,
      "step": 16190
    },
    {
      "entropy": 0.5331251315306872,
      "epoch": 0.263108743493337,
      "grad_norm": 10.0625,
      "learning_rate": 4.262625088493481e-05,
      "loss": 0.5483633518218994,
      "mean_token_accuracy": 0.8582705166190863,
      "num_tokens": 33102055.0,
      "step": 16200
    },
    {
      "entropy": 0.5693039912730455,
      "epoch": 0.26327115629796255,
      "grad_norm": 7.25,
      "learning_rate": 4.261705332462973e-05,
      "loss": 0.5953843593597412,
      "mean_token_accuracy": 0.8560689210891723,
      "num_tokens": 33122732.0,
      "step": 16210
    },
    {
      "entropy": 0.5127101918216794,
      "epoch": 0.263433569102588,
      "grad_norm": 10.0625,
      "learning_rate": 4.260785102529772e-05,
      "loss": 0.5462515830993653,
      "mean_token_accuracy": 0.8586858376860619,
      "num_tokens": 33142369.0,
      "step": 16220
    },
    {
      "entropy": 0.5132590401917696,
      "epoch": 0.26359598190721356,
      "grad_norm": 9.5625,
      "learning_rate": 4.25986439894142e-05,
      "loss": 0.5183812141418457,
      "mean_token_accuracy": 0.857141500711441,
      "num_tokens": 33162384.0,
      "step": 16230
    },
    {
      "entropy": 0.6678705089259893,
      "epoch": 0.2637583947118391,
      "grad_norm": 13.0,
      "learning_rate": 4.2589432219455906e-05,
      "loss": 0.7182539939880371,
      "mean_token_accuracy": 0.826000128313899,
      "num_tokens": 33185201.0,
      "step": 16240
    },
    {
      "entropy": 0.6287650496698916,
      "epoch": 0.2639208075164646,
      "grad_norm": 13.125,
      "learning_rate": 4.258021571790081e-05,
      "loss": 0.6098937511444091,
      "mean_token_accuracy": 0.8485548622906208,
      "num_tokens": 33204868.0,
      "step": 16250
    },
    {
      "entropy": 0.5461895851418376,
      "epoch": 0.2640832203210901,
      "grad_norm": 8.375,
      "learning_rate": 4.2570994487228184e-05,
      "loss": 0.576033639907837,
      "mean_token_accuracy": 0.8543567016720772,
      "num_tokens": 33225037.0,
      "step": 16260
    },
    {
      "entropy": 0.5336149439914152,
      "epoch": 0.26424563312571564,
      "grad_norm": 8.6875,
      "learning_rate": 4.256176852991855e-05,
      "loss": 0.523495101928711,
      "mean_token_accuracy": 0.8604555539786816,
      "num_tokens": 33246194.0,
      "step": 16270
    },
    {
      "entropy": 0.5963545046746731,
      "epoch": 0.2644080459303412,
      "grad_norm": 7.53125,
      "learning_rate": 4.255253784845371e-05,
      "loss": 0.6163718223571777,
      "mean_token_accuracy": 0.8453723147511483,
      "num_tokens": 33266049.0,
      "step": 16280
    },
    {
      "entropy": 0.5250309262424707,
      "epoch": 0.26457045873496665,
      "grad_norm": 7.6875,
      "learning_rate": 4.2543302445316745e-05,
      "loss": 0.5137829303741455,
      "mean_token_accuracy": 0.8653454124927521,
      "num_tokens": 33285186.0,
      "step": 16290
    },
    {
      "entropy": 0.5527250727638602,
      "epoch": 0.2647328715395922,
      "grad_norm": 8.0625,
      "learning_rate": 4.2534062322992e-05,
      "loss": 0.5296201229095459,
      "mean_token_accuracy": 0.8618603713810444,
      "num_tokens": 33305524.0,
      "step": 16300
    },
    {
      "entropy": 0.5441329414024949,
      "epoch": 0.2648952843442177,
      "grad_norm": 11.3125,
      "learning_rate": 4.252481748396507e-05,
      "loss": 0.5337966918945313,
      "mean_token_accuracy": 0.8627076774835587,
      "num_tokens": 33326671.0,
      "step": 16310
    },
    {
      "entropy": 0.6234054951928556,
      "epoch": 0.2650576971488432,
      "grad_norm": 6.8125,
      "learning_rate": 4.2515567930722865e-05,
      "loss": 0.6547018051147461,
      "mean_token_accuracy": 0.8392869472503662,
      "num_tokens": 33347013.0,
      "step": 16320
    },
    {
      "entropy": 0.582083504460752,
      "epoch": 0.26522010995346873,
      "grad_norm": 8.3125,
      "learning_rate": 4.2506313665753516e-05,
      "loss": 0.5876821517944336,
      "mean_token_accuracy": 0.85640790425241,
      "num_tokens": 33368195.0,
      "step": 16330
    },
    {
      "entropy": 0.5619365168735385,
      "epoch": 0.26538252275809426,
      "grad_norm": 8.5,
      "learning_rate": 4.2497054691546454e-05,
      "loss": 0.5679545402526855,
      "mean_token_accuracy": 0.8563005596399307,
      "num_tokens": 33388153.0,
      "step": 16340
    },
    {
      "entropy": 0.6284754982218146,
      "epoch": 0.26554493556271974,
      "grad_norm": 6.40625,
      "learning_rate": 4.248779101059236e-05,
      "loss": 0.6104865074157715,
      "mean_token_accuracy": 0.8421074993908405,
      "num_tokens": 33408692.0,
      "step": 16350
    },
    {
      "entropy": 0.5835771359503269,
      "epoch": 0.2657073483673453,
      "grad_norm": 10.6875,
      "learning_rate": 4.247852262538318e-05,
      "loss": 0.604541540145874,
      "mean_token_accuracy": 0.8534373760223388,
      "num_tokens": 33429638.0,
      "step": 16360
    },
    {
      "entropy": 0.5364490568637847,
      "epoch": 0.2658697611719708,
      "grad_norm": 7.65625,
      "learning_rate": 4.2469249538412126e-05,
      "loss": 0.5276559352874756,
      "mean_token_accuracy": 0.8644181504845619,
      "num_tokens": 33450595.0,
      "step": 16370
    },
    {
      "entropy": 0.5514813918620348,
      "epoch": 0.26603217397659634,
      "grad_norm": 10.875,
      "learning_rate": 4.245997175217369e-05,
      "loss": 0.5529515743255615,
      "mean_token_accuracy": 0.8537359863519669,
      "num_tokens": 33470948.0,
      "step": 16380
    },
    {
      "entropy": 0.5129128593951464,
      "epoch": 0.2661945867812218,
      "grad_norm": 6.4375,
      "learning_rate": 4.2450689269163625e-05,
      "loss": 0.5420063972473145,
      "mean_token_accuracy": 0.8648550517857074,
      "num_tokens": 33490824.0,
      "step": 16390
    },
    {
      "entropy": 0.5412522336468101,
      "epoch": 0.26635699958584735,
      "grad_norm": 7.53125,
      "learning_rate": 4.244140209187891e-05,
      "loss": 0.5616102695465088,
      "mean_token_accuracy": 0.8605915065854788,
      "num_tokens": 33510489.0,
      "step": 16400
    },
    {
      "entropy": 0.5327125217765569,
      "epoch": 0.2665194123904729,
      "grad_norm": 10.9375,
      "learning_rate": 4.243211022281785e-05,
      "loss": 0.5690330982208252,
      "mean_token_accuracy": 0.8564606696367264,
      "num_tokens": 33530612.0,
      "step": 16410
    },
    {
      "entropy": 0.5826835228130222,
      "epoch": 0.26668182519509837,
      "grad_norm": 9.5625,
      "learning_rate": 4.242281366447996e-05,
      "loss": 0.6051537990570068,
      "mean_token_accuracy": 0.8548326402902603,
      "num_tokens": 33550459.0,
      "step": 16420
    },
    {
      "entropy": 0.604815945122391,
      "epoch": 0.2668442379997239,
      "grad_norm": 5.0625,
      "learning_rate": 4.241351241936604e-05,
      "loss": 0.590269136428833,
      "mean_token_accuracy": 0.8574791558086872,
      "num_tokens": 33570656.0,
      "step": 16430
    },
    {
      "entropy": 0.5160299823153764,
      "epoch": 0.26700665080434943,
      "grad_norm": 4.875,
      "learning_rate": 4.240420648997814e-05,
      "loss": 0.48104586601257326,
      "mean_token_accuracy": 0.868197463452816,
      "num_tokens": 33590302.0,
      "step": 16440
    },
    {
      "entropy": 0.5970751129090786,
      "epoch": 0.2671690636089749,
      "grad_norm": 10.25,
      "learning_rate": 4.239489587881958e-05,
      "loss": 0.6419044971466065,
      "mean_token_accuracy": 0.8423970334231854,
      "num_tokens": 33611273.0,
      "step": 16450
    },
    {
      "entropy": 0.5298323538154364,
      "epoch": 0.26733147641360044,
      "grad_norm": 6.0625,
      "learning_rate": 4.238558058839495e-05,
      "loss": 0.5215121269226074,
      "mean_token_accuracy": 0.863980320841074,
      "num_tokens": 33631449.0,
      "step": 16460
    },
    {
      "entropy": 0.5542154439724982,
      "epoch": 0.267493889218226,
      "grad_norm": 11.0,
      "learning_rate": 4.2376260621210054e-05,
      "loss": 0.6321946620941162,
      "mean_token_accuracy": 0.84544732645154,
      "num_tokens": 33651168.0,
      "step": 16470
    },
    {
      "entropy": 0.5517195663414896,
      "epoch": 0.26765630202285146,
      "grad_norm": 7.90625,
      "learning_rate": 4.236693597977201e-05,
      "loss": 0.5449804306030274,
      "mean_token_accuracy": 0.8584469564259052,
      "num_tokens": 33671619.0,
      "step": 16480
    },
    {
      "entropy": 0.6310051124542951,
      "epoch": 0.267818714827477,
      "grad_norm": 9.75,
      "learning_rate": 4.2357606666589145e-05,
      "loss": 0.6404865741729736,
      "mean_token_accuracy": 0.8426645874977112,
      "num_tokens": 33692472.0,
      "step": 16490
    },
    {
      "entropy": 0.5884273577481508,
      "epoch": 0.2679811276321025,
      "grad_norm": 11.9375,
      "learning_rate": 4.234827268417109e-05,
      "loss": 0.5502813339233399,
      "mean_token_accuracy": 0.8546661324799061,
      "num_tokens": 33715309.0,
      "step": 16500
    },
    {
      "entropy": 0.5015754324384034,
      "epoch": 0.26814354043672806,
      "grad_norm": 6.71875,
      "learning_rate": 4.233893403502869e-05,
      "loss": 0.5132555961608887,
      "mean_token_accuracy": 0.8690888926386833,
      "num_tokens": 33736855.0,
      "step": 16510
    },
    {
      "entropy": 0.6149736329913139,
      "epoch": 0.26830595324135353,
      "grad_norm": 9.8125,
      "learning_rate": 4.2329590721674065e-05,
      "loss": 0.6158228874206543,
      "mean_token_accuracy": 0.8513496980071068,
      "num_tokens": 33758831.0,
      "step": 16520
    },
    {
      "entropy": 0.5170143836177885,
      "epoch": 0.26846836604597907,
      "grad_norm": 11.125,
      "learning_rate": 4.2320242746620596e-05,
      "loss": 0.5161044597625732,
      "mean_token_accuracy": 0.8749561227858067,
      "num_tokens": 33778378.0,
      "step": 16530
    },
    {
      "entropy": 0.5795797571539879,
      "epoch": 0.2686307788506046,
      "grad_norm": 5.5625,
      "learning_rate": 4.23108901123829e-05,
      "loss": 0.6109806537628174,
      "mean_token_accuracy": 0.8498612947762012,
      "num_tokens": 33799531.0,
      "step": 16540
    },
    {
      "entropy": 0.600351981446147,
      "epoch": 0.2687931916552301,
      "grad_norm": 7.90625,
      "learning_rate": 4.230153282147686e-05,
      "loss": 0.6233660697937011,
      "mean_token_accuracy": 0.8492235220968724,
      "num_tokens": 33819666.0,
      "step": 16550
    },
    {
      "entropy": 0.5845697808545083,
      "epoch": 0.2689556044598556,
      "grad_norm": 11.4375,
      "learning_rate": 4.229217087641961e-05,
      "loss": 0.5883178234100341,
      "mean_token_accuracy": 0.8510662697255611,
      "num_tokens": 33839301.0,
      "step": 16560
    },
    {
      "entropy": 0.510821298789233,
      "epoch": 0.26911801726448115,
      "grad_norm": 6.6875,
      "learning_rate": 4.228280427972954e-05,
      "loss": 0.5181769371032715,
      "mean_token_accuracy": 0.8714165113866329,
      "num_tokens": 33858289.0,
      "step": 16570
    },
    {
      "entropy": 0.6021355765871703,
      "epoch": 0.2692804300691066,
      "grad_norm": 6.28125,
      "learning_rate": 4.227343303392627e-05,
      "loss": 0.6193072319030761,
      "mean_token_accuracy": 0.8441445380449295,
      "num_tokens": 33880149.0,
      "step": 16580
    },
    {
      "entropy": 0.6207169647328555,
      "epoch": 0.26944284287373216,
      "grad_norm": 6.28125,
      "learning_rate": 4.226405714153069e-05,
      "loss": 0.6047979354858398,
      "mean_token_accuracy": 0.8490476347506046,
      "num_tokens": 33900625.0,
      "step": 16590
    },
    {
      "entropy": 0.5802690184675157,
      "epoch": 0.2696052556783577,
      "grad_norm": 10.0625,
      "learning_rate": 4.225467660506495e-05,
      "loss": 0.57479829788208,
      "mean_token_accuracy": 0.8546000212430954,
      "num_tokens": 33922498.0,
      "step": 16600
    },
    {
      "entropy": 0.5623771105427295,
      "epoch": 0.2697676684829832,
      "grad_norm": 5.28125,
      "learning_rate": 4.224529142705241e-05,
      "loss": 0.5789182662963868,
      "mean_token_accuracy": 0.8512364804744721,
      "num_tokens": 33943684.0,
      "step": 16610
    },
    {
      "entropy": 0.564122330583632,
      "epoch": 0.2699300812876087,
      "grad_norm": 9.1875,
      "learning_rate": 4.2235901610017736e-05,
      "loss": 0.5522418022155762,
      "mean_token_accuracy": 0.8605771139264107,
      "num_tokens": 33964202.0,
      "step": 16620
    },
    {
      "entropy": 0.5890005382709205,
      "epoch": 0.27009249409223424,
      "grad_norm": 5.65625,
      "learning_rate": 4.222650715648678e-05,
      "loss": 0.5731732368469238,
      "mean_token_accuracy": 0.8579270958900451,
      "num_tokens": 33985023.0,
      "step": 16630
    },
    {
      "entropy": 0.5548042587935924,
      "epoch": 0.27025490689685977,
      "grad_norm": 5.4375,
      "learning_rate": 4.22171080689867e-05,
      "loss": 0.5435409069061279,
      "mean_token_accuracy": 0.8553012430667877,
      "num_tokens": 34005510.0,
      "step": 16640
    },
    {
      "entropy": 0.5324314016848802,
      "epoch": 0.27041731970148525,
      "grad_norm": 8.8125,
      "learning_rate": 4.220770435004585e-05,
      "loss": 0.5271359443664551,
      "mean_token_accuracy": 0.864433404058218,
      "num_tokens": 34025083.0,
      "step": 16650
    },
    {
      "entropy": 0.5429102313704789,
      "epoch": 0.2705797325061108,
      "grad_norm": 8.25,
      "learning_rate": 4.219829600219386e-05,
      "loss": 0.5883169174194336,
      "mean_token_accuracy": 0.8542148374021054,
      "num_tokens": 34045486.0,
      "step": 16660
    },
    {
      "entropy": 0.615102867782116,
      "epoch": 0.2707421453107363,
      "grad_norm": 7.375,
      "learning_rate": 4.218888302796159e-05,
      "loss": 0.6614723205566406,
      "mean_token_accuracy": 0.8445832528173923,
      "num_tokens": 34064920.0,
      "step": 16670
    },
    {
      "entropy": 0.5798448003828526,
      "epoch": 0.2709045581153618,
      "grad_norm": 6.1875,
      "learning_rate": 4.217946542988116e-05,
      "loss": 0.5739744663238525,
      "mean_token_accuracy": 0.852310837060213,
      "num_tokens": 34087385.0,
      "step": 16680
    },
    {
      "entropy": 0.6251213777810335,
      "epoch": 0.2710669709199873,
      "grad_norm": 9.9375,
      "learning_rate": 4.2170043210485914e-05,
      "loss": 0.6683781147003174,
      "mean_token_accuracy": 0.8383262310177088,
      "num_tokens": 34108974.0,
      "step": 16690
    },
    {
      "entropy": 0.5418097000569105,
      "epoch": 0.27122938372461286,
      "grad_norm": 10.625,
      "learning_rate": 4.216061637231046e-05,
      "loss": 0.49447288513183596,
      "mean_token_accuracy": 0.8673963028937578,
      "num_tokens": 34130178.0,
      "step": 16700
    },
    {
      "entropy": 0.5465645428746939,
      "epoch": 0.27139179652923834,
      "grad_norm": 9.3125,
      "learning_rate": 4.215118491789064e-05,
      "loss": 0.5338427543640136,
      "mean_token_accuracy": 0.8583656795322895,
      "num_tokens": 34150770.0,
      "step": 16710
    },
    {
      "entropy": 0.5824407969601453,
      "epoch": 0.27155420933386387,
      "grad_norm": 12.125,
      "learning_rate": 4.214174884976353e-05,
      "loss": 0.5812499523162842,
      "mean_token_accuracy": 0.860411973297596,
      "num_tokens": 34170714.0,
      "step": 16720
    },
    {
      "entropy": 0.49869737653061746,
      "epoch": 0.2717166221384894,
      "grad_norm": 6.78125,
      "learning_rate": 4.213230817046745e-05,
      "loss": 0.48421783447265626,
      "mean_token_accuracy": 0.8789186835289001,
      "num_tokens": 34191449.0,
      "step": 16730
    },
    {
      "entropy": 0.5124646240845323,
      "epoch": 0.27187903494311494,
      "grad_norm": 8.4375,
      "learning_rate": 4.2122862882541984e-05,
      "loss": 0.5075724601745606,
      "mean_token_accuracy": 0.8635381847620011,
      "num_tokens": 34210849.0,
      "step": 16740
    },
    {
      "entropy": 0.5456450674682856,
      "epoch": 0.2720414477477404,
      "grad_norm": 9.125,
      "learning_rate": 4.2113412988527906e-05,
      "loss": 0.5643913745880127,
      "mean_token_accuracy": 0.8609402634203434,
      "num_tokens": 34231095.0,
      "step": 16750
    },
    {
      "entropy": 0.5326840508729219,
      "epoch": 0.27220386055236595,
      "grad_norm": 7.0,
      "learning_rate": 4.210395849096728e-05,
      "loss": 0.5703724384307861,
      "mean_token_accuracy": 0.8605490572750568,
      "num_tokens": 34251363.0,
      "step": 16760
    },
    {
      "entropy": 0.6256217143498362,
      "epoch": 0.2723662733569915,
      "grad_norm": 8.375,
      "learning_rate": 4.209449939240339e-05,
      "loss": 0.6251783847808838,
      "mean_token_accuracy": 0.8438511416316032,
      "num_tokens": 34272421.0,
      "step": 16770
    },
    {
      "entropy": 0.5645361507311464,
      "epoch": 0.27252868616161696,
      "grad_norm": 7.84375,
      "learning_rate": 4.2085035695380734e-05,
      "loss": 0.5741783618927002,
      "mean_token_accuracy": 0.8586551181972026,
      "num_tokens": 34293010.0,
      "step": 16780
    },
    {
      "entropy": 0.49915633974596857,
      "epoch": 0.2726910989662425,
      "grad_norm": 8.25,
      "learning_rate": 4.207556740244508e-05,
      "loss": 0.510115098953247,
      "mean_token_accuracy": 0.8660508673638105,
      "num_tokens": 34313705.0,
      "step": 16790
    },
    {
      "entropy": 0.5490601538680494,
      "epoch": 0.27285351177086803,
      "grad_norm": 8.0,
      "learning_rate": 4.206609451614342e-05,
      "loss": 0.5408076763153076,
      "mean_token_accuracy": 0.8670002140104771,
      "num_tokens": 34333963.0,
      "step": 16800
    },
    {
      "entropy": 0.528739502467215,
      "epoch": 0.2730159245754935,
      "grad_norm": 6.65625,
      "learning_rate": 4.205661703902399e-05,
      "loss": 0.5205760955810547,
      "mean_token_accuracy": 0.8645144209265709,
      "num_tokens": 34353673.0,
      "step": 16810
    },
    {
      "entropy": 0.5508116657845676,
      "epoch": 0.27317833738011904,
      "grad_norm": 8.625,
      "learning_rate": 4.204713497363624e-05,
      "loss": 0.5400552749633789,
      "mean_token_accuracy": 0.8547159381210804,
      "num_tokens": 34375454.0,
      "step": 16820
    },
    {
      "entropy": 0.5627385311760008,
      "epoch": 0.2733407501847446,
      "grad_norm": 6.75,
      "learning_rate": 4.2037648322530865e-05,
      "loss": 0.6165915489196777,
      "mean_token_accuracy": 0.8455269411206245,
      "num_tokens": 34396146.0,
      "step": 16830
    },
    {
      "entropy": 0.604011891130358,
      "epoch": 0.2735031629893701,
      "grad_norm": 7.0625,
      "learning_rate": 4.202815708825981e-05,
      "loss": 0.6942059993743896,
      "mean_token_accuracy": 0.8383936524391175,
      "num_tokens": 34417549.0,
      "step": 16840
    },
    {
      "entropy": 0.5634883895516396,
      "epoch": 0.2736655757939956,
      "grad_norm": 8.25,
      "learning_rate": 4.201866127337622e-05,
      "loss": 0.5792362689971924,
      "mean_token_accuracy": 0.857165378332138,
      "num_tokens": 34437902.0,
      "step": 16850
    },
    {
      "entropy": 0.6638612424023449,
      "epoch": 0.2738279885986211,
      "grad_norm": 7.6875,
      "learning_rate": 4.2009160880434514e-05,
      "loss": 0.603727912902832,
      "mean_token_accuracy": 0.8444490224123001,
      "num_tokens": 34458536.0,
      "step": 16860
    },
    {
      "entropy": 0.6491080243140459,
      "epoch": 0.27399040140324665,
      "grad_norm": 7.5,
      "learning_rate": 4.19996559119903e-05,
      "loss": 0.6011486530303956,
      "mean_token_accuracy": 0.8444184988737107,
      "num_tokens": 34478671.0,
      "step": 16870
    },
    {
      "entropy": 0.5894396848045289,
      "epoch": 0.27415281420787213,
      "grad_norm": 7.65625,
      "learning_rate": 4.199014637060044e-05,
      "loss": 0.5975615024566651,
      "mean_token_accuracy": 0.8498446650803089,
      "num_tokens": 34499308.0,
      "step": 16880
    },
    {
      "entropy": 0.5785379224922508,
      "epoch": 0.27431522701249766,
      "grad_norm": 6.28125,
      "learning_rate": 4.1980632258823016e-05,
      "loss": 0.5998688220977784,
      "mean_token_accuracy": 0.8473719231784343,
      "num_tokens": 34521074.0,
      "step": 16890
    },
    {
      "entropy": 0.5925297302193939,
      "epoch": 0.2744776398171232,
      "grad_norm": 12.875,
      "learning_rate": 4.1971113579217356e-05,
      "loss": 0.6416069984436035,
      "mean_token_accuracy": 0.8358314383774996,
      "num_tokens": 34541899.0,
      "step": 16900
    },
    {
      "entropy": 0.5447445314377546,
      "epoch": 0.2746400526217487,
      "grad_norm": 6.1875,
      "learning_rate": 4.196159033434402e-05,
      "loss": 0.5301802158355713,
      "mean_token_accuracy": 0.8715900272130966,
      "num_tokens": 34564169.0,
      "step": 16910
    },
    {
      "entropy": 0.5589986986480653,
      "epoch": 0.2748024654263742,
      "grad_norm": 9.375,
      "learning_rate": 4.1952062526764745e-05,
      "loss": 0.5843135356903076,
      "mean_token_accuracy": 0.8539220787584781,
      "num_tokens": 34585052.0,
      "step": 16920
    },
    {
      "entropy": 0.650121925584972,
      "epoch": 0.27496487823099974,
      "grad_norm": 7.9375,
      "learning_rate": 4.194253015904256e-05,
      "loss": 0.6756893634796143,
      "mean_token_accuracy": 0.8363918952643872,
      "num_tokens": 34605493.0,
      "step": 16930
    },
    {
      "entropy": 0.5356220480054616,
      "epoch": 0.2751272910356252,
      "grad_norm": 7.125,
      "learning_rate": 4.1932993233741683e-05,
      "loss": 0.5589962482452393,
      "mean_token_accuracy": 0.8615595169365406,
      "num_tokens": 34625675.0,
      "step": 16940
    },
    {
      "entropy": 0.6288696965202689,
      "epoch": 0.27528970384025075,
      "grad_norm": 7.84375,
      "learning_rate": 4.192345175342758e-05,
      "loss": 0.5986683368682861,
      "mean_token_accuracy": 0.8498246043920517,
      "num_tokens": 34647371.0,
      "step": 16950
    },
    {
      "entropy": 0.5831231956370175,
      "epoch": 0.2754521166448763,
      "grad_norm": 5.84375,
      "learning_rate": 4.191390572066691e-05,
      "loss": 0.6215912342071533,
      "mean_token_accuracy": 0.8437368616461753,
      "num_tokens": 34668184.0,
      "step": 16960
    },
    {
      "entropy": 0.6514689784497023,
      "epoch": 0.2756145294495018,
      "grad_norm": 7.0,
      "learning_rate": 4.19043551380276e-05,
      "loss": 0.6273575305938721,
      "mean_token_accuracy": 0.84620710760355,
      "num_tokens": 34690032.0,
      "step": 16970
    },
    {
      "entropy": 0.6574565122369677,
      "epoch": 0.2757769422541273,
      "grad_norm": 8.125,
      "learning_rate": 4.189480000807876e-05,
      "loss": 0.6485837936401367,
      "mean_token_accuracy": 0.8359887711703777,
      "num_tokens": 34711887.0,
      "step": 16980
    },
    {
      "entropy": 0.5792182378470898,
      "epoch": 0.27593935505875283,
      "grad_norm": 6.53125,
      "learning_rate": 4.188524033339076e-05,
      "loss": 0.6029642581939697,
      "mean_token_accuracy": 0.8530496574938298,
      "num_tokens": 34732237.0,
      "step": 16990
    },
    {
      "entropy": 0.5870880537666381,
      "epoch": 0.27610176786337837,
      "grad_norm": 10.25,
      "learning_rate": 4.187567611653515e-05,
      "loss": 0.6324345111846924,
      "mean_token_accuracy": 0.8510872669517994,
      "num_tokens": 34751892.0,
      "step": 17000
    },
    {
      "entropy": 0.5934215808287263,
      "epoch": 0.27626418066800384,
      "grad_norm": 7.8125,
      "learning_rate": 4.186610736008475e-05,
      "loss": 0.6163208961486817,
      "mean_token_accuracy": 0.8436141848564148,
      "num_tokens": 34773308.0,
      "step": 17010
    },
    {
      "entropy": 0.6300136646255851,
      "epoch": 0.2764265934726294,
      "grad_norm": 8.25,
      "learning_rate": 4.185653406661356e-05,
      "loss": 0.6235562324523926,
      "mean_token_accuracy": 0.8438818752765656,
      "num_tokens": 34793468.0,
      "step": 17020
    },
    {
      "entropy": 0.5535814927890896,
      "epoch": 0.2765890062772549,
      "grad_norm": 6.59375,
      "learning_rate": 4.1846956238696826e-05,
      "loss": 0.503101110458374,
      "mean_token_accuracy": 0.8690377783030272,
      "num_tokens": 34813889.0,
      "step": 17030
    },
    {
      "entropy": 0.5688198977615684,
      "epoch": 0.2767514190818804,
      "grad_norm": 8.375,
      "learning_rate": 4.1837373878911e-05,
      "loss": 0.5583671569824219,
      "mean_token_accuracy": 0.8567825440317393,
      "num_tokens": 34833380.0,
      "step": 17040
    },
    {
      "entropy": 0.5855814252980054,
      "epoch": 0.2769138318865059,
      "grad_norm": 6.0625,
      "learning_rate": 4.182778698983376e-05,
      "loss": 0.5642076015472413,
      "mean_token_accuracy": 0.8621127642691135,
      "num_tokens": 34854976.0,
      "step": 17050
    },
    {
      "entropy": 0.5549559784121811,
      "epoch": 0.27707624469113146,
      "grad_norm": 6.0,
      "learning_rate": 4.1818195574044005e-05,
      "loss": 0.5466836452484131,
      "mean_token_accuracy": 0.8516192182898521,
      "num_tokens": 34875440.0,
      "step": 17060
    },
    {
      "entropy": 0.5718593508936465,
      "epoch": 0.277238657495757,
      "grad_norm": 8.5625,
      "learning_rate": 4.1808599634121847e-05,
      "loss": 0.5986371517181397,
      "mean_token_accuracy": 0.8434766665101051,
      "num_tokens": 34895789.0,
      "step": 17070
    },
    {
      "entropy": 0.5525220753159374,
      "epoch": 0.27740107030038247,
      "grad_norm": 7.40625,
      "learning_rate": 4.17989991726486e-05,
      "loss": 0.5577431678771972,
      "mean_token_accuracy": 0.8547557480633259,
      "num_tokens": 34915986.0,
      "step": 17080
    },
    {
      "entropy": 0.5904794835485518,
      "epoch": 0.277563483105008,
      "grad_norm": 7.875,
      "learning_rate": 4.178939419220683e-05,
      "loss": 0.5893994808197022,
      "mean_token_accuracy": 0.8502130255103111,
      "num_tokens": 34936538.0,
      "step": 17090
    },
    {
      "entropy": 0.5207047936972231,
      "epoch": 0.27772589590963354,
      "grad_norm": 7.25,
      "learning_rate": 4.1779784695380296e-05,
      "loss": 0.5168735980987549,
      "mean_token_accuracy": 0.8733654268085956,
      "num_tokens": 34957085.0,
      "step": 17100
    },
    {
      "entropy": 0.519741227477789,
      "epoch": 0.277888308714259,
      "grad_norm": 7.09375,
      "learning_rate": 4.177017068475397e-05,
      "loss": 0.5618374824523926,
      "mean_token_accuracy": 0.8610585384070873,
      "num_tokens": 34977462.0,
      "step": 17110
    },
    {
      "entropy": 0.5616461067460478,
      "epoch": 0.27805072151888455,
      "grad_norm": 6.375,
      "learning_rate": 4.176055216291403e-05,
      "loss": 0.5823271751403809,
      "mean_token_accuracy": 0.856568917632103,
      "num_tokens": 34998017.0,
      "step": 17120
    },
    {
      "entropy": 0.624355851393193,
      "epoch": 0.2782131343235101,
      "grad_norm": 8.0625,
      "learning_rate": 4.17509291324479e-05,
      "loss": 0.619963550567627,
      "mean_token_accuracy": 0.8428517036139965,
      "num_tokens": 35018636.0,
      "step": 17130
    },
    {
      "entropy": 0.6934627486392856,
      "epoch": 0.27837554712813556,
      "grad_norm": 7.5625,
      "learning_rate": 4.174130159594418e-05,
      "loss": 0.6825096130371093,
      "mean_token_accuracy": 0.8393096674233675,
      "num_tokens": 35040798.0,
      "step": 17140
    },
    {
      "entropy": 0.6589559866115451,
      "epoch": 0.2785379599327611,
      "grad_norm": 7.5,
      "learning_rate": 4.173166955599271e-05,
      "loss": 0.6742685794830322,
      "mean_token_accuracy": 0.8311224080622196,
      "num_tokens": 35062688.0,
      "step": 17150
    },
    {
      "entropy": 0.5310449233278632,
      "epoch": 0.2787003727373866,
      "grad_norm": 5.90625,
      "learning_rate": 4.1722033015184524e-05,
      "loss": 0.5447644233703614,
      "mean_token_accuracy": 0.8583401057869195,
      "num_tokens": 35082848.0,
      "step": 17160
    },
    {
      "entropy": 0.5440952454693615,
      "epoch": 0.27886278554201216,
      "grad_norm": 8.5625,
      "learning_rate": 4.1712391976111884e-05,
      "loss": 0.532557487487793,
      "mean_token_accuracy": 0.8669322341680527,
      "num_tokens": 35103224.0,
      "step": 17170
    },
    {
      "entropy": 0.5400978217832744,
      "epoch": 0.27902519834663764,
      "grad_norm": 8.6875,
      "learning_rate": 4.1702746441368244e-05,
      "loss": 0.5274088382720947,
      "mean_token_accuracy": 0.8641109816730023,
      "num_tokens": 35123850.0,
      "step": 17180
    },
    {
      "entropy": 0.5854986840859056,
      "epoch": 0.27918761115126317,
      "grad_norm": 8.8125,
      "learning_rate": 4.169309641354827e-05,
      "loss": 0.628855562210083,
      "mean_token_accuracy": 0.8477097854018212,
      "num_tokens": 35144296.0,
      "step": 17190
    },
    {
      "entropy": 0.5581474377773702,
      "epoch": 0.2793500239558887,
      "grad_norm": 5.71875,
      "learning_rate": 4.1683441895247845e-05,
      "loss": 0.5336379528045654,
      "mean_token_accuracy": 0.8676164999604226,
      "num_tokens": 35165064.0,
      "step": 17200
    },
    {
      "entropy": 0.6300271459389478,
      "epoch": 0.2795124367605142,
      "grad_norm": 9.0,
      "learning_rate": 4.167378288906406e-05,
      "loss": 0.6607104778289795,
      "mean_token_accuracy": 0.8378219157457352,
      "num_tokens": 35185703.0,
      "step": 17210
    },
    {
      "entropy": 0.578103550337255,
      "epoch": 0.2796748495651397,
      "grad_norm": 6.78125,
      "learning_rate": 4.1664119397595204e-05,
      "loss": 0.5638640880584717,
      "mean_token_accuracy": 0.8464062280952931,
      "num_tokens": 35205799.0,
      "step": 17220
    },
    {
      "entropy": 0.5622347308322787,
      "epoch": 0.27983726236976525,
      "grad_norm": 7.46875,
      "learning_rate": 4.1654451423440774e-05,
      "loss": 0.5247482299804688,
      "mean_token_accuracy": 0.8586546801030636,
      "num_tokens": 35226867.0,
      "step": 17230
    },
    {
      "entropy": 0.5682780356612056,
      "epoch": 0.2799996751743907,
      "grad_norm": 9.375,
      "learning_rate": 4.164477896920149e-05,
      "loss": 0.6097508430480957,
      "mean_token_accuracy": 0.8518507562577724,
      "num_tokens": 35246731.0,
      "step": 17240
    },
    {
      "entropy": 0.6013665796723217,
      "epoch": 0.28016208797901626,
      "grad_norm": 8.3125,
      "learning_rate": 4.163510203747926e-05,
      "loss": 0.5910921096801758,
      "mean_token_accuracy": 0.8462337102741003,
      "num_tokens": 35266318.0,
      "step": 17250
    },
    {
      "entropy": 0.5779182584956288,
      "epoch": 0.2803245007836418,
      "grad_norm": 7.78125,
      "learning_rate": 4.162542063087719e-05,
      "loss": 0.5700791358947754,
      "mean_token_accuracy": 0.848541832715273,
      "num_tokens": 35287564.0,
      "step": 17260
    },
    {
      "entropy": 0.5899602007120848,
      "epoch": 0.2804869135882673,
      "grad_norm": 7.03125,
      "learning_rate": 4.16157347519996e-05,
      "loss": 0.6195521354675293,
      "mean_token_accuracy": 0.8446982979774476,
      "num_tokens": 35308302.0,
      "step": 17270
    },
    {
      "entropy": 0.5693035269156098,
      "epoch": 0.2806493263928928,
      "grad_norm": 9.875,
      "learning_rate": 4.160604440345204e-05,
      "loss": 0.5649159908294678,
      "mean_token_accuracy": 0.8521599486470223,
      "num_tokens": 35327306.0,
      "step": 17280
    },
    {
      "entropy": 0.6129769959487021,
      "epoch": 0.28081173919751834,
      "grad_norm": 11.875,
      "learning_rate": 4.15963495878412e-05,
      "loss": 0.6063143730163574,
      "mean_token_accuracy": 0.8516183126717806,
      "num_tokens": 35347327.0,
      "step": 17290
    },
    {
      "entropy": 0.5757642877288163,
      "epoch": 0.2809741520021439,
      "grad_norm": 10.25,
      "learning_rate": 4.158665030777503e-05,
      "loss": 0.5601007461547851,
      "mean_token_accuracy": 0.8558866009116173,
      "num_tokens": 35367302.0,
      "step": 17300
    },
    {
      "entropy": 0.5465124701149762,
      "epoch": 0.28113656480676935,
      "grad_norm": 7.125,
      "learning_rate": 4.157694656586264e-05,
      "loss": 0.5488223552703857,
      "mean_token_accuracy": 0.8587705932557583,
      "num_tokens": 35388511.0,
      "step": 17310
    },
    {
      "entropy": 0.5497525088489056,
      "epoch": 0.2812989776113949,
      "grad_norm": 9.625,
      "learning_rate": 4.156723836471438e-05,
      "loss": 0.5553276538848877,
      "mean_token_accuracy": 0.8568070910871028,
      "num_tokens": 35409351.0,
      "step": 17320
    },
    {
      "entropy": 0.4744534857571125,
      "epoch": 0.2814613904160204,
      "grad_norm": 6.34375,
      "learning_rate": 4.1557525706941766e-05,
      "loss": 0.5034657001495362,
      "mean_token_accuracy": 0.8700916402041912,
      "num_tokens": 35428953.0,
      "step": 17330
    },
    {
      "entropy": 0.5527645528316498,
      "epoch": 0.2816238032206459,
      "grad_norm": 8.875,
      "learning_rate": 4.154780859515752e-05,
      "loss": 0.5486737728118897,
      "mean_token_accuracy": 0.8577717334032059,
      "num_tokens": 35447996.0,
      "step": 17340
    },
    {
      "entropy": 0.5527544550132006,
      "epoch": 0.28178621602527143,
      "grad_norm": 8.75,
      "learning_rate": 4.153808703197558e-05,
      "loss": 0.5449261665344238,
      "mean_token_accuracy": 0.8609203234314918,
      "num_tokens": 35468642.0,
      "step": 17350
    },
    {
      "entropy": 0.5800910275429487,
      "epoch": 0.28194862882989696,
      "grad_norm": 8.25,
      "learning_rate": 4.1528361020011066e-05,
      "loss": 0.5869035720825195,
      "mean_token_accuracy": 0.8520423315465451,
      "num_tokens": 35489466.0,
      "step": 17360
    },
    {
      "entropy": 0.5823664361611008,
      "epoch": 0.28211104163452244,
      "grad_norm": 7.59375,
      "learning_rate": 4.151863056188028e-05,
      "loss": 0.5862632751464844,
      "mean_token_accuracy": 0.8559168040752411,
      "num_tokens": 35510451.0,
      "step": 17370
    },
    {
      "entropy": 0.5499073902145029,
      "epoch": 0.282273454439148,
      "grad_norm": 6.1875,
      "learning_rate": 4.1508895660200753e-05,
      "loss": 0.5790492057800293,
      "mean_token_accuracy": 0.8581585083156824,
      "num_tokens": 35531384.0,
      "step": 17380
    },
    {
      "entropy": 0.49489111742004754,
      "epoch": 0.2824358672437735,
      "grad_norm": 8.0,
      "learning_rate": 4.1499156317591206e-05,
      "loss": 0.47258906364440917,
      "mean_token_accuracy": 0.8745524615049363,
      "num_tokens": 35551730.0,
      "step": 17390
    },
    {
      "entropy": 0.598347395658493,
      "epoch": 0.28259828004839904,
      "grad_norm": 7.25,
      "learning_rate": 4.148941253667151e-05,
      "loss": 0.5980254650115967,
      "mean_token_accuracy": 0.8490207344293594,
      "num_tokens": 35571934.0,
      "step": 17400
    },
    {
      "entropy": 0.5667345805559307,
      "epoch": 0.2827606928530245,
      "grad_norm": 7.3125,
      "learning_rate": 4.1479664320062794e-05,
      "loss": 0.592627477645874,
      "mean_token_accuracy": 0.8590881638228893,
      "num_tokens": 35592705.0,
      "step": 17410
    },
    {
      "entropy": 0.6416580906137824,
      "epoch": 0.28292310565765005,
      "grad_norm": 12.3125,
      "learning_rate": 4.146991167038734e-05,
      "loss": 0.719015645980835,
      "mean_token_accuracy": 0.8285266526043416,
      "num_tokens": 35612561.0,
      "step": 17420
    },
    {
      "entropy": 0.6247211523354054,
      "epoch": 0.2830855184622756,
      "grad_norm": 7.5625,
      "learning_rate": 4.146015459026863e-05,
      "loss": 0.5971425533294678,
      "mean_token_accuracy": 0.8441060222685337,
      "num_tokens": 35632509.0,
      "step": 17430
    },
    {
      "entropy": 0.6453419416211545,
      "epoch": 0.28324793126690107,
      "grad_norm": 9.0625,
      "learning_rate": 4.1450393082331345e-05,
      "loss": 0.6612844467163086,
      "mean_token_accuracy": 0.8376931887120008,
      "num_tokens": 35651998.0,
      "step": 17440
    },
    {
      "entropy": 0.5539188359864056,
      "epoch": 0.2834103440715266,
      "grad_norm": 6.53125,
      "learning_rate": 4.144062714920135e-05,
      "loss": 0.5210306167602539,
      "mean_token_accuracy": 0.8713847391307354,
      "num_tokens": 35673054.0,
      "step": 17450
    },
    {
      "entropy": 0.5958240246400237,
      "epoch": 0.28357275687615213,
      "grad_norm": 7.28125,
      "learning_rate": 4.14308567935057e-05,
      "loss": 0.5703432559967041,
      "mean_token_accuracy": 0.853314845263958,
      "num_tokens": 35694354.0,
      "step": 17460
    },
    {
      "entropy": 0.6154017019551248,
      "epoch": 0.2837351696807776,
      "grad_norm": 9.4375,
      "learning_rate": 4.142108201787266e-05,
      "loss": 0.6284646034240723,
      "mean_token_accuracy": 0.8475752905011177,
      "num_tokens": 35714786.0,
      "step": 17470
    },
    {
      "entropy": 0.5869739778339863,
      "epoch": 0.28389758248540314,
      "grad_norm": 7.40625,
      "learning_rate": 4.141130282493164e-05,
      "loss": 0.6405982494354248,
      "mean_token_accuracy": 0.8430444978177547,
      "num_tokens": 35736322.0,
      "step": 17480
    },
    {
      "entropy": 0.5373887211317196,
      "epoch": 0.2840599952900287,
      "grad_norm": 7.4375,
      "learning_rate": 4.140151921731328e-05,
      "loss": 0.5245596408843994,
      "mean_token_accuracy": 0.8653483852744103,
      "num_tokens": 35756708.0,
      "step": 17490
    },
    {
      "entropy": 0.5025547658558935,
      "epoch": 0.28422240809465416,
      "grad_norm": 7.09375,
      "learning_rate": 4.1391731197649396e-05,
      "loss": 0.5525405406951904,
      "mean_token_accuracy": 0.8613792434334755,
      "num_tokens": 35777159.0,
      "step": 17500
    },
    {
      "entropy": 0.5854659236036241,
      "epoch": 0.2843848208992797,
      "grad_norm": 8.875,
      "learning_rate": 4.138193876857298e-05,
      "loss": 0.5806197643280029,
      "mean_token_accuracy": 0.8589339584112168,
      "num_tokens": 35796565.0,
      "step": 17510
    },
    {
      "entropy": 0.5519607802852988,
      "epoch": 0.2845472337039052,
      "grad_norm": 7.46875,
      "learning_rate": 4.137214193271821e-05,
      "loss": 0.545901870727539,
      "mean_token_accuracy": 0.8574182726442814,
      "num_tokens": 35818235.0,
      "step": 17520
    },
    {
      "entropy": 0.6152186324819923,
      "epoch": 0.28470964650853076,
      "grad_norm": 8.3125,
      "learning_rate": 4.1362340692720476e-05,
      "loss": 0.6106796741485596,
      "mean_token_accuracy": 0.8448817804455757,
      "num_tokens": 35840010.0,
      "step": 17530
    },
    {
      "entropy": 0.5822362299077213,
      "epoch": 0.28487205931315623,
      "grad_norm": 7.0,
      "learning_rate": 4.1352535051216325e-05,
      "loss": 0.588127851486206,
      "mean_token_accuracy": 0.8579162038862705,
      "num_tokens": 35859662.0,
      "step": 17540
    },
    {
      "entropy": 0.5471317727118731,
      "epoch": 0.28503447211778177,
      "grad_norm": 9.125,
      "learning_rate": 4.134272501084349e-05,
      "loss": 0.5259416103363037,
      "mean_token_accuracy": 0.8706240721046925,
      "num_tokens": 35880234.0,
      "step": 17550
    },
    {
      "entropy": 0.5969628514256329,
      "epoch": 0.2851968849224073,
      "grad_norm": 6.59375,
      "learning_rate": 4.1332910574240904e-05,
      "loss": 0.56806960105896,
      "mean_token_accuracy": 0.8556464098393917,
      "num_tokens": 35901623.0,
      "step": 17560
    },
    {
      "entropy": 0.574421445466578,
      "epoch": 0.2853592977270328,
      "grad_norm": 9.9375,
      "learning_rate": 4.132309174404866e-05,
      "loss": 0.600691556930542,
      "mean_token_accuracy": 0.8485545732080937,
      "num_tokens": 35923096.0,
      "step": 17570
    },
    {
      "entropy": 0.5160778786055744,
      "epoch": 0.2855217105316583,
      "grad_norm": 6.6875,
      "learning_rate": 4.1313268522908054e-05,
      "loss": 0.5769938468933106,
      "mean_token_accuracy": 0.8572316449135542,
      "num_tokens": 35942435.0,
      "step": 17580
    },
    {
      "entropy": 0.5498999962583184,
      "epoch": 0.28568412333628385,
      "grad_norm": 7.15625,
      "learning_rate": 4.130344091346156e-05,
      "loss": 0.5378740787506103,
      "mean_token_accuracy": 0.8644469544291496,
      "num_tokens": 35961577.0,
      "step": 17590
    },
    {
      "entropy": 0.5533097469247877,
      "epoch": 0.2858465361409093,
      "grad_norm": 8.125,
      "learning_rate": 4.129360891835282e-05,
      "loss": 0.549836254119873,
      "mean_token_accuracy": 0.8656658448278904,
      "num_tokens": 35981559.0,
      "step": 17600
    },
    {
      "entropy": 0.5843121553771198,
      "epoch": 0.28600894894553486,
      "grad_norm": 8.875,
      "learning_rate": 4.128377254022666e-05,
      "loss": 0.5941105842590332,
      "mean_token_accuracy": 0.8556938670575619,
      "num_tokens": 36001462.0,
      "step": 17610
    },
    {
      "entropy": 0.5473244554363191,
      "epoch": 0.2861713617501604,
      "grad_norm": 7.5,
      "learning_rate": 4.127393178172909e-05,
      "loss": 0.5528611183166504,
      "mean_token_accuracy": 0.8558194275945425,
      "num_tokens": 36021045.0,
      "step": 17620
    },
    {
      "entropy": 0.553973457403481,
      "epoch": 0.2863337745547859,
      "grad_norm": 10.8125,
      "learning_rate": 4.12640866455073e-05,
      "loss": 0.5782723426818848,
      "mean_token_accuracy": 0.85633700825274,
      "num_tokens": 36040359.0,
      "step": 17630
    },
    {
      "entropy": 0.5864728387678042,
      "epoch": 0.2864961873594114,
      "grad_norm": 10.5,
      "learning_rate": 4.125423713420964e-05,
      "loss": 0.5553581714630127,
      "mean_token_accuracy": 0.8580593869090081,
      "num_tokens": 36060630.0,
      "step": 17640
    },
    {
      "entropy": 0.5424744283314794,
      "epoch": 0.28665860016403694,
      "grad_norm": 8.0,
      "learning_rate": 4.124438325048568e-05,
      "loss": 0.5376532077789307,
      "mean_token_accuracy": 0.857544032484293,
      "num_tokens": 36081055.0,
      "step": 17650
    },
    {
      "entropy": 0.6458050215616822,
      "epoch": 0.28682101296866247,
      "grad_norm": 5.9375,
      "learning_rate": 4.123452499698611e-05,
      "loss": 0.6484622001647949,
      "mean_token_accuracy": 0.8370225392282009,
      "num_tokens": 36103235.0,
      "step": 17660
    },
    {
      "entropy": 0.534138029627502,
      "epoch": 0.28698342577328795,
      "grad_norm": 6.71875,
      "learning_rate": 4.1224662376362835e-05,
      "loss": 0.525483512878418,
      "mean_token_accuracy": 0.8630974285304547,
      "num_tokens": 36123020.0,
      "step": 17670
    },
    {
      "entropy": 0.5162336615845561,
      "epoch": 0.2871458385779135,
      "grad_norm": 6.59375,
      "learning_rate": 4.121479539126891e-05,
      "loss": 0.541999626159668,
      "mean_token_accuracy": 0.8643454276025295,
      "num_tokens": 36143251.0,
      "step": 17680
    },
    {
      "entropy": 0.6323876513168216,
      "epoch": 0.287308251382539,
      "grad_norm": 6.75,
      "learning_rate": 4.120492404435859e-05,
      "loss": 0.6170620918273926,
      "mean_token_accuracy": 0.8446521505713462,
      "num_tokens": 36165943.0,
      "step": 17690
    },
    {
      "entropy": 0.5623032076284289,
      "epoch": 0.2874706641871645,
      "grad_norm": 9.0625,
      "learning_rate": 4.119504833828728e-05,
      "loss": 0.5820642948150635,
      "mean_token_accuracy": 0.8524319879710675,
      "num_tokens": 36186983.0,
      "step": 17700
    },
    {
      "entropy": 0.6092072970233857,
      "epoch": 0.28763307699179,
      "grad_norm": 7.96875,
      "learning_rate": 4.1185168275711575e-05,
      "loss": 0.6203063488006592,
      "mean_token_accuracy": 0.8448413096368312,
      "num_tokens": 36207530.0,
      "step": 17710
    },
    {
      "entropy": 0.5783728525508195,
      "epoch": 0.28779548979641556,
      "grad_norm": 7.84375,
      "learning_rate": 4.117528385928923e-05,
      "loss": 0.5510667324066162,
      "mean_token_accuracy": 0.8617670249193907,
      "num_tokens": 36226903.0,
      "step": 17720
    },
    {
      "entropy": 0.6476022410672158,
      "epoch": 0.2879579026010411,
      "grad_norm": 6.75,
      "learning_rate": 4.116539509167917e-05,
      "loss": 0.6388350486755371,
      "mean_token_accuracy": 0.8455386407673359,
      "num_tokens": 36247134.0,
      "step": 17730
    },
    {
      "entropy": 0.5517082770355046,
      "epoch": 0.28812031540566657,
      "grad_norm": 8.1875,
      "learning_rate": 4.115550197554151e-05,
      "loss": 0.5484023094177246,
      "mean_token_accuracy": 0.8587029784917831,
      "num_tokens": 36268517.0,
      "step": 17740
    },
    {
      "entropy": 0.562576531432569,
      "epoch": 0.2882827282102921,
      "grad_norm": 6.90625,
      "learning_rate": 4.114560451353751e-05,
      "loss": 0.602958869934082,
      "mean_token_accuracy": 0.8474968262016773,
      "num_tokens": 36289504.0,
      "step": 17750
    },
    {
      "entropy": 0.5603354729246348,
      "epoch": 0.28844514101491764,
      "grad_norm": 7.5,
      "learning_rate": 4.11357027083296e-05,
      "loss": 0.5313716888427734,
      "mean_token_accuracy": 0.8664328280836344,
      "num_tokens": 36311290.0,
      "step": 17760
    },
    {
      "entropy": 0.5495112643111497,
      "epoch": 0.2886075538195431,
      "grad_norm": 6.90625,
      "learning_rate": 4.112579656258141e-05,
      "loss": 0.5571826457977295,
      "mean_token_accuracy": 0.8597597695887089,
      "num_tokens": 36333031.0,
      "step": 17770
    },
    {
      "entropy": 0.6527315212413669,
      "epoch": 0.28876996662416865,
      "grad_norm": 9.9375,
      "learning_rate": 4.11158860789577e-05,
      "loss": 0.6422440052032471,
      "mean_token_accuracy": 0.8332031443715096,
      "num_tokens": 36352934.0,
      "step": 17780
    },
    {
      "entropy": 0.577813727222383,
      "epoch": 0.2889323794287942,
      "grad_norm": 9.4375,
      "learning_rate": 4.110597126012441e-05,
      "loss": 0.5936680793762207,
      "mean_token_accuracy": 0.8521417856216431,
      "num_tokens": 36373320.0,
      "step": 17790
    },
    {
      "entropy": 0.5374428056180477,
      "epoch": 0.28909479223341966,
      "grad_norm": 7.625,
      "learning_rate": 4.109605210874866e-05,
      "loss": 0.5424469947814942,
      "mean_token_accuracy": 0.8592448588460684,
      "num_tokens": 36393144.0,
      "step": 17800
    },
    {
      "entropy": 0.5207731116563081,
      "epoch": 0.2892572050380452,
      "grad_norm": 8.875,
      "learning_rate": 4.108612862749872e-05,
      "loss": 0.5289613723754882,
      "mean_token_accuracy": 0.8686387054622173,
      "num_tokens": 36413043.0,
      "step": 17810
    },
    {
      "entropy": 0.5566857509315014,
      "epoch": 0.28941961784267073,
      "grad_norm": 9.8125,
      "learning_rate": 4.107620081904403e-05,
      "loss": 0.5656529426574707,
      "mean_token_accuracy": 0.8588267520070076,
      "num_tokens": 36432857.0,
      "step": 17820
    },
    {
      "entropy": 0.5664749258663505,
      "epoch": 0.2895820306472962,
      "grad_norm": 8.125,
      "learning_rate": 4.1066268686055185e-05,
      "loss": 0.5908914566040039,
      "mean_token_accuracy": 0.843728294223547,
      "num_tokens": 36454438.0,
      "step": 17830
    },
    {
      "entropy": 0.537671027542092,
      "epoch": 0.28974444345192174,
      "grad_norm": 6.8125,
      "learning_rate": 4.105633223120396e-05,
      "loss": 0.5199490547180176,
      "mean_token_accuracy": 0.8649649038910866,
      "num_tokens": 36475197.0,
      "step": 17840
    },
    {
      "entropy": 0.568777548475191,
      "epoch": 0.2899068562565473,
      "grad_norm": 7.65625,
      "learning_rate": 4.104639145716329e-05,
      "loss": 0.590980339050293,
      "mean_token_accuracy": 0.8534414507448673,
      "num_tokens": 36497981.0,
      "step": 17850
    },
    {
      "entropy": 0.5720078336074949,
      "epoch": 0.2900692690611728,
      "grad_norm": 6.75,
      "learning_rate": 4.103644636660725e-05,
      "loss": 0.5618825912475586,
      "mean_token_accuracy": 0.8561365321278572,
      "num_tokens": 36518080.0,
      "step": 17860
    },
    {
      "entropy": 0.5823044904274866,
      "epoch": 0.2902316818657983,
      "grad_norm": 6.625,
      "learning_rate": 4.102649696221109e-05,
      "loss": 0.5696081638336181,
      "mean_token_accuracy": 0.8533410087227822,
      "num_tokens": 36537423.0,
      "step": 17870
    },
    {
      "entropy": 0.5678859127685427,
      "epoch": 0.2903940946704238,
      "grad_norm": 9.75,
      "learning_rate": 4.101654324665124e-05,
      "loss": 0.5897946357727051,
      "mean_token_accuracy": 0.8512251764535904,
      "num_tokens": 36557439.0,
      "step": 17880
    },
    {
      "entropy": 0.6353920376393944,
      "epoch": 0.29055650747504935,
      "grad_norm": 8.3125,
      "learning_rate": 4.100658522260525e-05,
      "loss": 0.6696792602539062,
      "mean_token_accuracy": 0.8361506476998329,
      "num_tokens": 36577983.0,
      "step": 17890
    },
    {
      "entropy": 0.5568199940957129,
      "epoch": 0.29071892027967483,
      "grad_norm": 6.78125,
      "learning_rate": 4.099662289275187e-05,
      "loss": 0.5541680335998536,
      "mean_token_accuracy": 0.8598075456917286,
      "num_tokens": 36598132.0,
      "step": 17900
    },
    {
      "entropy": 0.5291572362184525,
      "epoch": 0.29088133308430036,
      "grad_norm": 5.34375,
      "learning_rate": 4.098665625977098e-05,
      "loss": 0.5445431232452392,
      "mean_token_accuracy": 0.8656539022922516,
      "num_tokens": 36618896.0,
      "step": 17910
    },
    {
      "entropy": 0.6025902549037709,
      "epoch": 0.2910437458889259,
      "grad_norm": 7.59375,
      "learning_rate": 4.097668532634364e-05,
      "loss": 0.6150074958801269,
      "mean_token_accuracy": 0.8478502459824085,
      "num_tokens": 36639908.0,
      "step": 17920
    },
    {
      "entropy": 0.6107180405408144,
      "epoch": 0.2912061586935514,
      "grad_norm": 9.4375,
      "learning_rate": 4.0966710095152025e-05,
      "loss": 0.652470064163208,
      "mean_token_accuracy": 0.8425132732838392,
      "num_tokens": 36660160.0,
      "step": 17930
    },
    {
      "entropy": 0.5969099901616574,
      "epoch": 0.2913685714981769,
      "grad_norm": 7.65625,
      "learning_rate": 4.095673056887952e-05,
      "loss": 0.5847942352294921,
      "mean_token_accuracy": 0.8524424131959677,
      "num_tokens": 36679780.0,
      "step": 17940
    },
    {
      "entropy": 0.6449717089533806,
      "epoch": 0.29153098430280244,
      "grad_norm": 9.5625,
      "learning_rate": 4.094674675021062e-05,
      "loss": 0.6113699436187744,
      "mean_token_accuracy": 0.8545657977461815,
      "num_tokens": 36700959.0,
      "step": 17950
    },
    {
      "entropy": 0.5516521789133548,
      "epoch": 0.291693397107428,
      "grad_norm": 9.0,
      "learning_rate": 4.093675864183102e-05,
      "loss": 0.5698765277862549,
      "mean_token_accuracy": 0.8648547828197479,
      "num_tokens": 36721201.0,
      "step": 17960
    },
    {
      "entropy": 0.555703211389482,
      "epoch": 0.29185580991205345,
      "grad_norm": 8.9375,
      "learning_rate": 4.092676624642753e-05,
      "loss": 0.6030320644378662,
      "mean_token_accuracy": 0.8496562197804451,
      "num_tokens": 36740652.0,
      "step": 17970
    },
    {
      "entropy": 0.5522500654682517,
      "epoch": 0.292018222716679,
      "grad_norm": 9.3125,
      "learning_rate": 4.091676956668812e-05,
      "loss": 0.5329155921936035,
      "mean_token_accuracy": 0.8671309120953083,
      "num_tokens": 36761112.0,
      "step": 17980
    },
    {
      "entropy": 0.5402680757455528,
      "epoch": 0.2921806355213045,
      "grad_norm": 6.65625,
      "learning_rate": 4.0906768605301926e-05,
      "loss": 0.5264708995819092,
      "mean_token_accuracy": 0.8663390628993511,
      "num_tokens": 36781696.0,
      "step": 17990
    },
    {
      "entropy": 0.6154843703843653,
      "epoch": 0.29234304832593,
      "grad_norm": 9.625,
      "learning_rate": 4.0896763364959226e-05,
      "loss": 0.5661901950836181,
      "mean_token_accuracy": 0.8550937570631504,
      "num_tokens": 36801898.0,
      "step": 18000
    },
    {
      "entropy": 0.5111170560587197,
      "epoch": 0.29250546113055553,
      "grad_norm": 7.3125,
      "learning_rate": 4.088675384835146e-05,
      "loss": 0.5245023727416992,
      "mean_token_accuracy": 0.8623235650360584,
      "num_tokens": 36821759.0,
      "step": 18010
    },
    {
      "entropy": 0.6104663755279034,
      "epoch": 0.29266787393518107,
      "grad_norm": 8.0625,
      "learning_rate": 4.0876740058171205e-05,
      "loss": 0.6635880947113038,
      "mean_token_accuracy": 0.8331946823745966,
      "num_tokens": 36841356.0,
      "step": 18020
    },
    {
      "entropy": 0.5692138301208616,
      "epoch": 0.29283028673980654,
      "grad_norm": 6.34375,
      "learning_rate": 4.0866721997112187e-05,
      "loss": 0.6090588092803955,
      "mean_token_accuracy": 0.8457665733993054,
      "num_tokens": 36862269.0,
      "step": 18030
    },
    {
      "entropy": 0.5730566069483757,
      "epoch": 0.2929926995444321,
      "grad_norm": 9.125,
      "learning_rate": 4.0856699667869306e-05,
      "loss": 0.5650660514831543,
      "mean_token_accuracy": 0.8469328068196773,
      "num_tokens": 36883190.0,
      "step": 18040
    },
    {
      "entropy": 0.5944055709987879,
      "epoch": 0.2931551123490576,
      "grad_norm": 8.625,
      "learning_rate": 4.0846673073138574e-05,
      "loss": 0.5808224201202392,
      "mean_token_accuracy": 0.8544750109314918,
      "num_tokens": 36904410.0,
      "step": 18050
    },
    {
      "entropy": 0.5761209764517844,
      "epoch": 0.2933175251536831,
      "grad_norm": 7.28125,
      "learning_rate": 4.083664221561717e-05,
      "loss": 0.518587875366211,
      "mean_token_accuracy": 0.8651981003582477,
      "num_tokens": 36923996.0,
      "step": 18060
    },
    {
      "entropy": 0.5270165452267974,
      "epoch": 0.2934799379583086,
      "grad_norm": 9.625,
      "learning_rate": 4.0826607098003425e-05,
      "loss": 0.5439772605895996,
      "mean_token_accuracy": 0.8673811666667461,
      "num_tokens": 36944136.0,
      "step": 18070
    },
    {
      "entropy": 0.5486955150030554,
      "epoch": 0.29364235076293416,
      "grad_norm": 7.5,
      "learning_rate": 4.081656772299679e-05,
      "loss": 0.5619600772857666,
      "mean_token_accuracy": 0.8522886082530021,
      "num_tokens": 36964878.0,
      "step": 18080
    },
    {
      "entropy": 0.4929536744020879,
      "epoch": 0.2938047635675597,
      "grad_norm": 6.34375,
      "learning_rate": 4.08065240932979e-05,
      "loss": 0.5016175746917725,
      "mean_token_accuracy": 0.8689725793898105,
      "num_tokens": 36984441.0,
      "step": 18090
    },
    {
      "entropy": 0.5907150311395526,
      "epoch": 0.29396717637218517,
      "grad_norm": 8.75,
      "learning_rate": 4.079647621160851e-05,
      "loss": 0.6377906322479248,
      "mean_token_accuracy": 0.8404063172638416,
      "num_tokens": 37006031.0,
      "step": 18100
    },
    {
      "entropy": 0.5768464021384716,
      "epoch": 0.2941295891768107,
      "grad_norm": 8.5625,
      "learning_rate": 4.078642408063151e-05,
      "loss": 0.5937373638153076,
      "mean_token_accuracy": 0.8580769155174494,
      "num_tokens": 37027121.0,
      "step": 18110
    },
    {
      "entropy": 0.5411137895891442,
      "epoch": 0.29429200198143624,
      "grad_norm": 7.5625,
      "learning_rate": 4.0776367703070964e-05,
      "loss": 0.5420875549316406,
      "mean_token_accuracy": 0.860224773734808,
      "num_tokens": 37046699.0,
      "step": 18120
    },
    {
      "entropy": 0.5106237702071666,
      "epoch": 0.2944544147860617,
      "grad_norm": 6.125,
      "learning_rate": 4.0766307081632044e-05,
      "loss": 0.5123288631439209,
      "mean_token_accuracy": 0.8725985728204251,
      "num_tokens": 37067915.0,
      "step": 18130
    },
    {
      "entropy": 0.5638970719650388,
      "epoch": 0.29461682759068725,
      "grad_norm": 7.03125,
      "learning_rate": 4.075624221902109e-05,
      "loss": 0.590222692489624,
      "mean_token_accuracy": 0.8569910779595376,
      "num_tokens": 37087216.0,
      "step": 18140
    },
    {
      "entropy": 0.51576405800879,
      "epoch": 0.2947792403953128,
      "grad_norm": 5.4375,
      "learning_rate": 4.074617311794556e-05,
      "loss": 0.5310060024261475,
      "mean_token_accuracy": 0.8604837134480476,
      "num_tokens": 37107950.0,
      "step": 18150
    },
    {
      "entropy": 0.5134930270723999,
      "epoch": 0.29494165319993826,
      "grad_norm": 7.0625,
      "learning_rate": 4.073609978111409e-05,
      "loss": 0.5159493923187256,
      "mean_token_accuracy": 0.8667227417230606,
      "num_tokens": 37127416.0,
      "step": 18160
    },
    {
      "entropy": 0.522550730407238,
      "epoch": 0.2951040660045638,
      "grad_norm": 6.5,
      "learning_rate": 4.0726022211236406e-05,
      "loss": 0.5133474349975586,
      "mean_token_accuracy": 0.8685505025088787,
      "num_tokens": 37148291.0,
      "step": 18170
    },
    {
      "entropy": 0.5872115667909383,
      "epoch": 0.2952664788091893,
      "grad_norm": 6.21875,
      "learning_rate": 4.07159404110234e-05,
      "loss": 0.5909671306610107,
      "mean_token_accuracy": 0.8518031291663647,
      "num_tokens": 37169408.0,
      "step": 18180
    },
    {
      "entropy": 0.6117963010445238,
      "epoch": 0.29542889161381486,
      "grad_norm": 10.0,
      "learning_rate": 4.070585438318711e-05,
      "loss": 0.635195541381836,
      "mean_token_accuracy": 0.8412977695465088,
      "num_tokens": 37189834.0,
      "step": 18190
    },
    {
      "entropy": 0.5234861487522722,
      "epoch": 0.29559130441844034,
      "grad_norm": 10.4375,
      "learning_rate": 4.0695764130440694e-05,
      "loss": 0.559336519241333,
      "mean_token_accuracy": 0.855384674668312,
      "num_tokens": 37210210.0,
      "step": 18200
    },
    {
      "entropy": 0.5356620155274868,
      "epoch": 0.29575371722306587,
      "grad_norm": 7.15625,
      "learning_rate": 4.0685669655498446e-05,
      "loss": 0.5773856163024902,
      "mean_token_accuracy": 0.8580003976821899,
      "num_tokens": 37229609.0,
      "step": 18210
    },
    {
      "entropy": 0.5619220338761807,
      "epoch": 0.2959161300276914,
      "grad_norm": 9.0,
      "learning_rate": 4.067557096107582e-05,
      "loss": 0.5762789726257325,
      "mean_token_accuracy": 0.8572461754083633,
      "num_tokens": 37250673.0,
      "step": 18220
    },
    {
      "entropy": 0.5985932556912303,
      "epoch": 0.2960785428323169,
      "grad_norm": 9.75,
      "learning_rate": 4.066546804988938e-05,
      "loss": 0.5882694244384765,
      "mean_token_accuracy": 0.8536705445498228,
      "num_tokens": 37271854.0,
      "step": 18230
    },
    {
      "entropy": 0.6399203925393522,
      "epoch": 0.2962409556369424,
      "grad_norm": 10.4375,
      "learning_rate": 4.0655360924656816e-05,
      "loss": 0.6362391471862793,
      "mean_token_accuracy": 0.8352310247719288,
      "num_tokens": 37293019.0,
      "step": 18240
    },
    {
      "entropy": 0.5028411030769349,
      "epoch": 0.29640336844156795,
      "grad_norm": 16.25,
      "learning_rate": 4.064524958809698e-05,
      "loss": 0.4999791145324707,
      "mean_token_accuracy": 0.8723169803619385,
      "num_tokens": 37312412.0,
      "step": 18250
    },
    {
      "entropy": 0.5751097188331187,
      "epoch": 0.2965657812461934,
      "grad_norm": 9.875,
      "learning_rate": 4.0635134042929845e-05,
      "loss": 0.5989929676055908,
      "mean_token_accuracy": 0.851113050058484,
      "num_tokens": 37332487.0,
      "step": 18260
    },
    {
      "entropy": 0.5347581686452031,
      "epoch": 0.29672819405081896,
      "grad_norm": 8.0,
      "learning_rate": 4.0625014291876525e-05,
      "loss": 0.5251726150512696,
      "mean_token_accuracy": 0.8616917006671428,
      "num_tokens": 37352682.0,
      "step": 18270
    },
    {
      "entropy": 0.575373284239322,
      "epoch": 0.2968906068554445,
      "grad_norm": 9.1875,
      "learning_rate": 4.061489033765923e-05,
      "loss": 0.5845426559448242,
      "mean_token_accuracy": 0.8572585381567478,
      "num_tokens": 37373882.0,
      "step": 18280
    },
    {
      "entropy": 0.555945485830307,
      "epoch": 0.29705301966007,
      "grad_norm": 6.96875,
      "learning_rate": 4.060476218300136e-05,
      "loss": 0.5536273002624512,
      "mean_token_accuracy": 0.8602010656148196,
      "num_tokens": 37395146.0,
      "step": 18290
    },
    {
      "entropy": 0.5867072810418904,
      "epoch": 0.2972154324646955,
      "grad_norm": 6.0,
      "learning_rate": 4.059462983062738e-05,
      "loss": 0.592656421661377,
      "mean_token_accuracy": 0.8501064013689756,
      "num_tokens": 37415945.0,
      "step": 18300
    },
    {
      "entropy": 0.49287839424796404,
      "epoch": 0.29737784526932104,
      "grad_norm": 6.8125,
      "learning_rate": 4.0584493283262934e-05,
      "loss": 0.5179312705993653,
      "mean_token_accuracy": 0.8665899276733399,
      "num_tokens": 37435325.0,
      "step": 18310
    },
    {
      "entropy": 0.5457955406047403,
      "epoch": 0.2975402580739466,
      "grad_norm": 6.75,
      "learning_rate": 4.057435254363477e-05,
      "loss": 0.5549544334411621,
      "mean_token_accuracy": 0.8597264714539051,
      "num_tokens": 37456261.0,
      "step": 18320
    },
    {
      "entropy": 0.5304658240173012,
      "epoch": 0.29770267087857205,
      "grad_norm": 6.25,
      "learning_rate": 4.0564207614470774e-05,
      "loss": 0.5480710029602051,
      "mean_token_accuracy": 0.8660931870341301,
      "num_tokens": 37476561.0,
      "step": 18330
    },
    {
      "entropy": 0.5616916247876361,
      "epoch": 0.2978650836831976,
      "grad_norm": 6.34375,
      "learning_rate": 4.055405849849995e-05,
      "loss": 0.5445189952850342,
      "mean_token_accuracy": 0.8548055727034807,
      "num_tokens": 37497620.0,
      "step": 18340
    },
    {
      "entropy": 0.5865047479979694,
      "epoch": 0.2980274964878231,
      "grad_norm": 5.1875,
      "learning_rate": 4.0543905198452423e-05,
      "loss": 0.6317931652069092,
      "mean_token_accuracy": 0.847425140067935,
      "num_tokens": 37519010.0,
      "step": 18350
    },
    {
      "entropy": 0.5604267084971071,
      "epoch": 0.2981899092924486,
      "grad_norm": 7.65625,
      "learning_rate": 4.053374771705948e-05,
      "loss": 0.5366101741790772,
      "mean_token_accuracy": 0.8648546151816845,
      "num_tokens": 37540024.0,
      "step": 18360
    },
    {
      "entropy": 0.6291342639364302,
      "epoch": 0.29835232209707413,
      "grad_norm": 8.625,
      "learning_rate": 4.0523586057053484e-05,
      "loss": 0.6290030479431152,
      "mean_token_accuracy": 0.8334765449166298,
      "num_tokens": 37559561.0,
      "step": 18370
    },
    {
      "entropy": 0.4761138725094497,
      "epoch": 0.29851473490169966,
      "grad_norm": 6.53125,
      "learning_rate": 4.051342022116795e-05,
      "loss": 0.471433162689209,
      "mean_token_accuracy": 0.8751187555491924,
      "num_tokens": 37580476.0,
      "step": 18380
    },
    {
      "entropy": 0.591613108292222,
      "epoch": 0.29867714770632514,
      "grad_norm": 8.125,
      "learning_rate": 4.0503250212137523e-05,
      "loss": 0.6463563919067383,
      "mean_token_accuracy": 0.83936932310462,
      "num_tokens": 37601462.0,
      "step": 18390
    },
    {
      "entropy": 0.5198147079441696,
      "epoch": 0.2988395605109507,
      "grad_norm": 6.34375,
      "learning_rate": 4.049307603269793e-05,
      "loss": 0.5053734302520752,
      "mean_token_accuracy": 0.8686460986733436,
      "num_tokens": 37621076.0,
      "step": 18400
    },
    {
      "entropy": 0.5581785670481623,
      "epoch": 0.2990019733155762,
      "grad_norm": 7.6875,
      "learning_rate": 4.048289768558608e-05,
      "loss": 0.5392992973327637,
      "mean_token_accuracy": 0.8625946007668972,
      "num_tokens": 37641908.0,
      "step": 18410
    },
    {
      "entropy": 0.5275594906881451,
      "epoch": 0.29916438612020174,
      "grad_norm": 8.6875,
      "learning_rate": 4.047271517353995e-05,
      "loss": 0.49792022705078126,
      "mean_token_accuracy": 0.8674438059329986,
      "num_tokens": 37662683.0,
      "step": 18420
    },
    {
      "entropy": 0.5245737377554178,
      "epoch": 0.2993267989248272,
      "grad_norm": 8.3125,
      "learning_rate": 4.0462528499298666e-05,
      "loss": 0.5501349925994873,
      "mean_token_accuracy": 0.8588576681911946,
      "num_tokens": 37682302.0,
      "step": 18430
    },
    {
      "entropy": 0.480031169205904,
      "epoch": 0.29948921172945275,
      "grad_norm": 7.25,
      "learning_rate": 4.045233766560247e-05,
      "loss": 0.47205214500427245,
      "mean_token_accuracy": 0.8748327143490314,
      "num_tokens": 37704008.0,
      "step": 18440
    },
    {
      "entropy": 0.44854514035396276,
      "epoch": 0.2996516245340783,
      "grad_norm": 5.5,
      "learning_rate": 4.0442142675192715e-05,
      "loss": 0.4672853469848633,
      "mean_token_accuracy": 0.875875324010849,
      "num_tokens": 37724921.0,
      "step": 18450
    },
    {
      "entropy": 0.4931590172462165,
      "epoch": 0.29981403733870376,
      "grad_norm": 14.0625,
      "learning_rate": 4.043194353081188e-05,
      "loss": 0.5496779918670655,
      "mean_token_accuracy": 0.8608825527131557,
      "num_tokens": 37745873.0,
      "step": 18460
    },
    {
      "entropy": 0.5909688537940383,
      "epoch": 0.2999764501433293,
      "grad_norm": 7.90625,
      "learning_rate": 4.042174023520354e-05,
      "loss": 0.6300872802734375,
      "mean_token_accuracy": 0.8459713123738766,
      "num_tokens": 37766060.0,
      "step": 18470
    },
    {
      "entropy": 0.4987281003035605,
      "epoch": 0.30013886294795483,
      "grad_norm": 5.71875,
      "learning_rate": 4.041153279111243e-05,
      "loss": 0.4945223808288574,
      "mean_token_accuracy": 0.8740331277251243,
      "num_tokens": 37786082.0,
      "step": 18480
    },
    {
      "entropy": 0.6078218809794634,
      "epoch": 0.3003012757525803,
      "grad_norm": 6.3125,
      "learning_rate": 4.0401321201284356e-05,
      "loss": 0.5837239742279052,
      "mean_token_accuracy": 0.8519362024962902,
      "num_tokens": 37806640.0,
      "step": 18490
    },
    {
      "entropy": 0.6298030248843134,
      "epoch": 0.30046368855720584,
      "grad_norm": 7.46875,
      "learning_rate": 4.039110546846627e-05,
      "loss": 0.622196102142334,
      "mean_token_accuracy": 0.8506173595786095,
      "num_tokens": 37828216.0,
      "step": 18500
    },
    {
      "entropy": 0.526211079210043,
      "epoch": 0.3006261013618314,
      "grad_norm": 5.9375,
      "learning_rate": 4.038088559540623e-05,
      "loss": 0.5140679359436036,
      "mean_token_accuracy": 0.8674201183021069,
      "num_tokens": 37848437.0,
      "step": 18510
    },
    {
      "entropy": 0.6190586889162659,
      "epoch": 0.3007885141664569,
      "grad_norm": 8.4375,
      "learning_rate": 4.037066158485338e-05,
      "loss": 0.6683550357818604,
      "mean_token_accuracy": 0.8353213183581829,
      "num_tokens": 37869560.0,
      "step": 18520
    },
    {
      "entropy": 0.5134971441468223,
      "epoch": 0.3009509269710824,
      "grad_norm": 8.0625,
      "learning_rate": 4.036043343955804e-05,
      "loss": 0.5547324657440186,
      "mean_token_accuracy": 0.8696597218513489,
      "num_tokens": 37889486.0,
      "step": 18530
    },
    {
      "entropy": 0.5937203916721046,
      "epoch": 0.3011133397757079,
      "grad_norm": 7.84375,
      "learning_rate": 4.0350201162271585e-05,
      "loss": 0.5922856807708741,
      "mean_token_accuracy": 0.8509636022150516,
      "num_tokens": 37909604.0,
      "step": 18540
    },
    {
      "entropy": 0.6088982063345612,
      "epoch": 0.30127575258033346,
      "grad_norm": 8.1875,
      "learning_rate": 4.0339964755746504e-05,
      "loss": 0.6338249683380127,
      "mean_token_accuracy": 0.8501441366970539,
      "num_tokens": 37929060.0,
      "step": 18550
    },
    {
      "entropy": 0.6426794612780213,
      "epoch": 0.30143816538495893,
      "grad_norm": 7.78125,
      "learning_rate": 4.032972422273644e-05,
      "loss": 0.6053624153137207,
      "mean_token_accuracy": 0.8474758937954903,
      "num_tokens": 37949396.0,
      "step": 18560
    },
    {
      "entropy": 0.46357202297076583,
      "epoch": 0.30160057818958447,
      "grad_norm": 6.78125,
      "learning_rate": 4.031947956599611e-05,
      "loss": 0.40838823318481443,
      "mean_token_accuracy": 0.8943016901612282,
      "num_tokens": 37969557.0,
      "step": 18570
    },
    {
      "entropy": 0.514061052724719,
      "epoch": 0.30176299099421,
      "grad_norm": 9.0625,
      "learning_rate": 4.030923078828135e-05,
      "loss": 0.5570794582366944,
      "mean_token_accuracy": 0.8568534646183252,
      "num_tokens": 37989542.0,
      "step": 18580
    },
    {
      "entropy": 0.5405584103893488,
      "epoch": 0.3019254037988355,
      "grad_norm": 8.8125,
      "learning_rate": 4.029897789234909e-05,
      "loss": 0.5596442699432373,
      "mean_token_accuracy": 0.8651847392320633,
      "num_tokens": 38008726.0,
      "step": 18590
    },
    {
      "entropy": 0.4916353357024491,
      "epoch": 0.302087816603461,
      "grad_norm": 8.625,
      "learning_rate": 4.028872088095741e-05,
      "loss": 0.4604952812194824,
      "mean_token_accuracy": 0.8768504455685615,
      "num_tokens": 38029129.0,
      "step": 18600
    },
    {
      "entropy": 0.5155279682483525,
      "epoch": 0.30225022940808655,
      "grad_norm": 7.46875,
      "learning_rate": 4.027845975686544e-05,
      "loss": 0.523536205291748,
      "mean_token_accuracy": 0.8625660508871078,
      "num_tokens": 38048987.0,
      "step": 18610
    },
    {
      "entropy": 0.4857929261401296,
      "epoch": 0.302412642212712,
      "grad_norm": 5.65625,
      "learning_rate": 4.026819452283347e-05,
      "loss": 0.4623286247253418,
      "mean_token_accuracy": 0.8777495145797729,
      "num_tokens": 38069112.0,
      "step": 18620
    },
    {
      "entropy": 0.5810016373172402,
      "epoch": 0.30257505501733756,
      "grad_norm": 6.34375,
      "learning_rate": 4.025792518162285e-05,
      "loss": 0.6277034759521485,
      "mean_token_accuracy": 0.8362510047852993,
      "num_tokens": 38090019.0,
      "step": 18630
    },
    {
      "entropy": 0.5497262555174529,
      "epoch": 0.3027374678219631,
      "grad_norm": 6.125,
      "learning_rate": 4.024765173599607e-05,
      "loss": 0.5717017650604248,
      "mean_token_accuracy": 0.8614972122013569,
      "num_tokens": 38111190.0,
      "step": 18640
    },
    {
      "entropy": 0.5776823358610272,
      "epoch": 0.3028998806265886,
      "grad_norm": 8.3125,
      "learning_rate": 4.0237374188716716e-05,
      "loss": 0.5677416801452637,
      "mean_token_accuracy": 0.8574865818023681,
      "num_tokens": 38131531.0,
      "step": 18650
    },
    {
      "entropy": 0.5262401051819324,
      "epoch": 0.3030622934312141,
      "grad_norm": 6.875,
      "learning_rate": 4.022709254254945e-05,
      "loss": 0.49150938987731935,
      "mean_token_accuracy": 0.8737956151366234,
      "num_tokens": 38151996.0,
      "step": 18660
    },
    {
      "entropy": 0.5739308902528137,
      "epoch": 0.30322470623583964,
      "grad_norm": 7.4375,
      "learning_rate": 4.021680680026007e-05,
      "loss": 0.569417667388916,
      "mean_token_accuracy": 0.8551946900784969,
      "num_tokens": 38172399.0,
      "step": 18670
    },
    {
      "entropy": 0.5793012114241719,
      "epoch": 0.30338711904046517,
      "grad_norm": 6.5625,
      "learning_rate": 4.020651696461547e-05,
      "loss": 0.6440511226654053,
      "mean_token_accuracy": 0.8509455643594265,
      "num_tokens": 38192936.0,
      "step": 18680
    },
    {
      "entropy": 0.560908401152119,
      "epoch": 0.30354953184509065,
      "grad_norm": 6.28125,
      "learning_rate": 4.0196223038383626e-05,
      "loss": 0.5597870826721192,
      "mean_token_accuracy": 0.8571750685572624,
      "num_tokens": 38214802.0,
      "step": 18690
    },
    {
      "entropy": 0.5578504668548703,
      "epoch": 0.3037119446497162,
      "grad_norm": 6.53125,
      "learning_rate": 4.018592502433363e-05,
      "loss": 0.5760014057159424,
      "mean_token_accuracy": 0.8574293315410614,
      "num_tokens": 38236120.0,
      "step": 18700
    },
    {
      "entropy": 0.5486619657371193,
      "epoch": 0.3038743574543417,
      "grad_norm": 5.96875,
      "learning_rate": 4.017562292523568e-05,
      "loss": 0.5545848846435547,
      "mean_token_accuracy": 0.8563994932919741,
      "num_tokens": 38256453.0,
      "step": 18710
    },
    {
      "entropy": 0.5290350871160626,
      "epoch": 0.3040367702589672,
      "grad_norm": 6.0,
      "learning_rate": 4.016531674386106e-05,
      "loss": 0.5835164546966553,
      "mean_token_accuracy": 0.8578483976423741,
      "num_tokens": 38277192.0,
      "step": 18720
    },
    {
      "entropy": 0.6084001552313566,
      "epoch": 0.3041991830635927,
      "grad_norm": 7.125,
      "learning_rate": 4.015500648298215e-05,
      "loss": 0.5854540348052979,
      "mean_token_accuracy": 0.8465478666126728,
      "num_tokens": 38297855.0,
      "step": 18730
    },
    {
      "entropy": 0.565308470511809,
      "epoch": 0.30436159586821826,
      "grad_norm": 10.5625,
      "learning_rate": 4.014469214537244e-05,
      "loss": 0.607196855545044,
      "mean_token_accuracy": 0.8484242931008339,
      "num_tokens": 38318252.0,
      "step": 18740
    },
    {
      "entropy": 0.574339872598648,
      "epoch": 0.3045240086728438,
      "grad_norm": 7.75,
      "learning_rate": 4.013437373380651e-05,
      "loss": 0.5973798274993897,
      "mean_token_accuracy": 0.8446942731738091,
      "num_tokens": 38337633.0,
      "step": 18750
    },
    {
      "entropy": 0.5720671666786075,
      "epoch": 0.30468642147746927,
      "grad_norm": 8.625,
      "learning_rate": 4.012405125106003e-05,
      "loss": 0.5775179862976074,
      "mean_token_accuracy": 0.8548632331192494,
      "num_tokens": 38358263.0,
      "step": 18760
    },
    {
      "entropy": 0.5116192387416959,
      "epoch": 0.3048488342820948,
      "grad_norm": 6.5,
      "learning_rate": 4.011372469990977e-05,
      "loss": 0.5235109329223633,
      "mean_token_accuracy": 0.8610612653195858,
      "num_tokens": 38377842.0,
      "step": 18770
    },
    {
      "entropy": 0.5275371117517352,
      "epoch": 0.30501124708672034,
      "grad_norm": 8.375,
      "learning_rate": 4.01033940831336e-05,
      "loss": 0.5244380474090576,
      "mean_token_accuracy": 0.8636931829154492,
      "num_tokens": 38397175.0,
      "step": 18780
    },
    {
      "entropy": 0.5273407051339746,
      "epoch": 0.3051736598913458,
      "grad_norm": 9.3125,
      "learning_rate": 4.009305940351048e-05,
      "loss": 0.5104909896850586,
      "mean_token_accuracy": 0.8686032436788083,
      "num_tokens": 38417280.0,
      "step": 18790
    },
    {
      "entropy": 0.514790934510529,
      "epoch": 0.30533607269597135,
      "grad_norm": 11.3125,
      "learning_rate": 4.008272066382047e-05,
      "loss": 0.5102662563323974,
      "mean_token_accuracy": 0.8684730850160122,
      "num_tokens": 38438506.0,
      "step": 18800
    },
    {
      "entropy": 0.5778977618552744,
      "epoch": 0.3054984855005969,
      "grad_norm": 10.75,
      "learning_rate": 4.007237786684469e-05,
      "loss": 0.5812434196472168,
      "mean_token_accuracy": 0.8492145761847496,
      "num_tokens": 38460147.0,
      "step": 18810
    },
    {
      "entropy": 0.5289366538636386,
      "epoch": 0.30566089830522236,
      "grad_norm": 9.0,
      "learning_rate": 4.0062031015365396e-05,
      "loss": 0.5674821376800537,
      "mean_token_accuracy": 0.8563410319387913,
      "num_tokens": 38480880.0,
      "step": 18820
    },
    {
      "entropy": 0.6358812071383,
      "epoch": 0.3058233111098479,
      "grad_norm": 7.46875,
      "learning_rate": 4.005168011216592e-05,
      "loss": 0.6255525588989258,
      "mean_token_accuracy": 0.8439608298242092,
      "num_tokens": 38501894.0,
      "step": 18830
    },
    {
      "entropy": 0.5146015097387135,
      "epoch": 0.30598572391447343,
      "grad_norm": 5.75,
      "learning_rate": 4.004132516003066e-05,
      "loss": 0.4996341705322266,
      "mean_token_accuracy": 0.8677243966609239,
      "num_tokens": 38522925.0,
      "step": 18840
    },
    {
      "entropy": 0.5024016416631639,
      "epoch": 0.3061481367190989,
      "grad_norm": 8.375,
      "learning_rate": 4.003096616174512e-05,
      "loss": 0.4870443820953369,
      "mean_token_accuracy": 0.8678838357329368,
      "num_tokens": 38543956.0,
      "step": 18850
    },
    {
      "entropy": 0.5860246462281793,
      "epoch": 0.30631054952372444,
      "grad_norm": 7.1875,
      "learning_rate": 4.002060312009591e-05,
      "loss": 0.6507218360900879,
      "mean_token_accuracy": 0.8421036954969168,
      "num_tokens": 38565183.0,
      "step": 18860
    },
    {
      "entropy": 0.5225791018456221,
      "epoch": 0.30647296232835,
      "grad_norm": 7.5,
      "learning_rate": 4.0010236037870705e-05,
      "loss": 0.4806240081787109,
      "mean_token_accuracy": 0.872143467515707,
      "num_tokens": 38586255.0,
      "step": 18870
    },
    {
      "entropy": 0.5179202020633966,
      "epoch": 0.3066353751329755,
      "grad_norm": 7.71875,
      "learning_rate": 3.9999864917858276e-05,
      "loss": 0.5184689044952393,
      "mean_token_accuracy": 0.8641481995582581,
      "num_tokens": 38606584.0,
      "step": 18880
    },
    {
      "entropy": 0.5677586127072572,
      "epoch": 0.306797787937601,
      "grad_norm": 6.96875,
      "learning_rate": 3.998948976284847e-05,
      "loss": 0.5744168281555175,
      "mean_token_accuracy": 0.8513001285493373,
      "num_tokens": 38626528.0,
      "step": 18890
    },
    {
      "entropy": 0.515854315739125,
      "epoch": 0.3069602007422265,
      "grad_norm": 7.0,
      "learning_rate": 3.9979110575632226e-05,
      "loss": 0.5207403182983399,
      "mean_token_accuracy": 0.8589541107416153,
      "num_tokens": 38647371.0,
      "step": 18900
    },
    {
      "entropy": 0.6085777279455215,
      "epoch": 0.30712261354685205,
      "grad_norm": 7.125,
      "learning_rate": 3.996872735900158e-05,
      "loss": 0.6056053638458252,
      "mean_token_accuracy": 0.8429012976586818,
      "num_tokens": 38668004.0,
      "step": 18910
    },
    {
      "entropy": 0.5638324806466699,
      "epoch": 0.30728502635147753,
      "grad_norm": 9.125,
      "learning_rate": 3.995834011574963e-05,
      "loss": 0.5830959796905517,
      "mean_token_accuracy": 0.8498951315879821,
      "num_tokens": 38688512.0,
      "step": 18920
    },
    {
      "entropy": 0.5790169212035835,
      "epoch": 0.30744743915610306,
      "grad_norm": 6.71875,
      "learning_rate": 3.994794884867058e-05,
      "loss": 0.603140926361084,
      "mean_token_accuracy": 0.8551956124603748,
      "num_tokens": 38709657.0,
      "step": 18930
    },
    {
      "entropy": 0.5848646854050458,
      "epoch": 0.3076098519607286,
      "grad_norm": 7.25,
      "learning_rate": 3.993755356055969e-05,
      "loss": 0.5836717128753662,
      "mean_token_accuracy": 0.8537973761558533,
      "num_tokens": 38730011.0,
      "step": 18940
    },
    {
      "entropy": 0.5304347008001059,
      "epoch": 0.3077722647653541,
      "grad_norm": 10.3125,
      "learning_rate": 3.9927154254213324e-05,
      "loss": 0.5124943733215332,
      "mean_token_accuracy": 0.8593435332179069,
      "num_tokens": 38750360.0,
      "step": 18950
    },
    {
      "entropy": 0.5537048161029816,
      "epoch": 0.3079346775699796,
      "grad_norm": 7.3125,
      "learning_rate": 3.991675093242891e-05,
      "loss": 0.5733994960784912,
      "mean_token_accuracy": 0.853875944763422,
      "num_tokens": 38769989.0,
      "step": 18960
    },
    {
      "entropy": 0.646009061485529,
      "epoch": 0.30809709037460514,
      "grad_norm": 7.21875,
      "learning_rate": 3.990634359800498e-05,
      "loss": 0.6725071430206299,
      "mean_token_accuracy": 0.841486033052206,
      "num_tokens": 38791407.0,
      "step": 18970
    },
    {
      "entropy": 0.5538863874971867,
      "epoch": 0.3082595031792307,
      "grad_norm": 5.6875,
      "learning_rate": 3.989593225374112e-05,
      "loss": 0.5354800224304199,
      "mean_token_accuracy": 0.868000753968954,
      "num_tokens": 38812137.0,
      "step": 18980
    },
    {
      "entropy": 0.6242260855156928,
      "epoch": 0.30842191598385615,
      "grad_norm": 8.9375,
      "learning_rate": 3.9885516902438016e-05,
      "loss": 0.5793234825134277,
      "mean_token_accuracy": 0.8533586151897907,
      "num_tokens": 38832093.0,
      "step": 18990
    },
    {
      "entropy": 0.5957398721016943,
      "epoch": 0.3085843287884817,
      "grad_norm": 9.5625,
      "learning_rate": 3.98750975468974e-05,
      "loss": 0.6055310726165771,
      "mean_token_accuracy": 0.8475604251027107,
      "num_tokens": 38853016.0,
      "step": 19000
    },
    {
      "entropy": 0.5511787381023169,
      "epoch": 0.3087467415931072,
      "grad_norm": 7.21875,
      "learning_rate": 3.9864674189922114e-05,
      "loss": 0.5812894344329834,
      "mean_token_accuracy": 0.8532300375401973,
      "num_tokens": 38874452.0,
      "step": 19010
    },
    {
      "entropy": 0.5321362631861121,
      "epoch": 0.3089091543977327,
      "grad_norm": 7.75,
      "learning_rate": 3.985424683431607e-05,
      "loss": 0.4938765048980713,
      "mean_token_accuracy": 0.8735089272260665,
      "num_tokens": 38893395.0,
      "step": 19020
    },
    {
      "entropy": 0.5352374540176242,
      "epoch": 0.30907156720235823,
      "grad_norm": 8.4375,
      "learning_rate": 3.984381548288425e-05,
      "loss": 0.5238100051879883,
      "mean_token_accuracy": 0.863590133190155,
      "num_tokens": 38912921.0,
      "step": 19030
    },
    {
      "entropy": 0.6083398039452732,
      "epoch": 0.30923398000698377,
      "grad_norm": 8.375,
      "learning_rate": 3.983338013843269e-05,
      "loss": 0.6686630249023438,
      "mean_token_accuracy": 0.8335504595190286,
      "num_tokens": 38933240.0,
      "step": 19040
    },
    {
      "entropy": 0.4984588557155803,
      "epoch": 0.30939639281160924,
      "grad_norm": 5.4375,
      "learning_rate": 3.982294080376855e-05,
      "loss": 0.4973402500152588,
      "mean_token_accuracy": 0.8700132440775633,
      "num_tokens": 38954709.0,
      "step": 19050
    },
    {
      "entropy": 0.5311712921597064,
      "epoch": 0.3095588056162348,
      "grad_norm": 8.8125,
      "learning_rate": 3.981249748170002e-05,
      "loss": 0.535819673538208,
      "mean_token_accuracy": 0.8602229751646518,
      "num_tokens": 38974865.0,
      "step": 19060
    },
    {
      "entropy": 0.5581549669615924,
      "epoch": 0.3097212184208603,
      "grad_norm": 7.375,
      "learning_rate": 3.9802050175036377e-05,
      "loss": 0.5196681022644043,
      "mean_token_accuracy": 0.8629326976835727,
      "num_tokens": 38996017.0,
      "step": 19070
    },
    {
      "entropy": 0.6870385620277375,
      "epoch": 0.30988363122548584,
      "grad_norm": 8.875,
      "learning_rate": 3.9791598886587964e-05,
      "loss": 0.705416202545166,
      "mean_token_accuracy": 0.8247367113828659,
      "num_tokens": 39016907.0,
      "step": 19080
    },
    {
      "entropy": 0.5622619421221315,
      "epoch": 0.3100460440301113,
      "grad_norm": 9.1875,
      "learning_rate": 3.978114361916622e-05,
      "loss": 0.5847455501556397,
      "mean_token_accuracy": 0.8482499629259109,
      "num_tokens": 39037388.0,
      "step": 19090
    },
    {
      "entropy": 0.5848425036529079,
      "epoch": 0.31020845683473686,
      "grad_norm": 8.5625,
      "learning_rate": 3.977068437558361e-05,
      "loss": 0.6195621013641357,
      "mean_token_accuracy": 0.8549220554530621,
      "num_tokens": 39058198.0,
      "step": 19100
    },
    {
      "entropy": 0.5159534751437604,
      "epoch": 0.3103708696393624,
      "grad_norm": 11.0,
      "learning_rate": 3.976022115865372e-05,
      "loss": 0.5237014293670654,
      "mean_token_accuracy": 0.8680881977081298,
      "num_tokens": 39077955.0,
      "step": 19110
    },
    {
      "entropy": 0.6860589657910168,
      "epoch": 0.31053328244398787,
      "grad_norm": 7.1875,
      "learning_rate": 3.974975397119116e-05,
      "loss": 0.6303433895111084,
      "mean_token_accuracy": 0.84304124340415,
      "num_tokens": 39098155.0,
      "step": 19120
    },
    {
      "entropy": 0.5860568817937747,
      "epoch": 0.3106956952486134,
      "grad_norm": 7.96875,
      "learning_rate": 3.973928281601164e-05,
      "loss": 0.5881663799285889,
      "mean_token_accuracy": 0.8512168727815151,
      "num_tokens": 39118038.0,
      "step": 19130
    },
    {
      "entropy": 0.5705150796566159,
      "epoch": 0.31085810805323894,
      "grad_norm": 11.5625,
      "learning_rate": 3.9728807695931916e-05,
      "loss": 0.6196613788604737,
      "mean_token_accuracy": 0.8537365816533565,
      "num_tokens": 39138809.0,
      "step": 19140
    },
    {
      "entropy": 0.47885154485702514,
      "epoch": 0.3110205208578644,
      "grad_norm": 6.96875,
      "learning_rate": 3.9718328613769817e-05,
      "loss": 0.45858087539672854,
      "mean_token_accuracy": 0.8804532706737518,
      "num_tokens": 39157960.0,
      "step": 19150
    },
    {
      "entropy": 0.5192302588373423,
      "epoch": 0.31118293366248995,
      "grad_norm": 8.9375,
      "learning_rate": 3.9707845572344244e-05,
      "loss": 0.5124351978302002,
      "mean_token_accuracy": 0.8645814381539821,
      "num_tokens": 39178443.0,
      "step": 19160
    },
    {
      "entropy": 0.5327319487929344,
      "epoch": 0.3113453464671155,
      "grad_norm": 6.34375,
      "learning_rate": 3.969735857447516e-05,
      "loss": 0.5344436168670654,
      "mean_token_accuracy": 0.8677810408174992,
      "num_tokens": 39198807.0,
      "step": 19170
    },
    {
      "entropy": 0.5710339611396194,
      "epoch": 0.31150775927174096,
      "grad_norm": 7.15625,
      "learning_rate": 3.968686762298359e-05,
      "loss": 0.5479374408721924,
      "mean_token_accuracy": 0.855444285273552,
      "num_tokens": 39219293.0,
      "step": 19180
    },
    {
      "entropy": 0.43814735640771685,
      "epoch": 0.3116701720763665,
      "grad_norm": 6.96875,
      "learning_rate": 3.967637272069162e-05,
      "loss": 0.38538327217102053,
      "mean_token_accuracy": 0.8971199378371238,
      "num_tokens": 39238312.0,
      "step": 19190
    },
    {
      "entropy": 0.544992359355092,
      "epoch": 0.311832584880992,
      "grad_norm": 7.625,
      "learning_rate": 3.966587387042241e-05,
      "loss": 0.5848678588867188,
      "mean_token_accuracy": 0.8497995935380459,
      "num_tokens": 39258617.0,
      "step": 19200
    },
    {
      "entropy": 0.49279797710478307,
      "epoch": 0.31199499768561756,
      "grad_norm": 15.3125,
      "learning_rate": 3.965537107500017e-05,
      "loss": 0.5475634098052978,
      "mean_token_accuracy": 0.8686653397977352,
      "num_tokens": 39278534.0,
      "step": 19210
    },
    {
      "entropy": 0.5287778607103973,
      "epoch": 0.31215741049024304,
      "grad_norm": 9.0625,
      "learning_rate": 3.964486433725017e-05,
      "loss": 0.5726676464080811,
      "mean_token_accuracy": 0.8591269560158252,
      "num_tokens": 39298880.0,
      "step": 19220
    },
    {
      "entropy": 0.5519081845879554,
      "epoch": 0.31231982329486857,
      "grad_norm": 7.9375,
      "learning_rate": 3.963435365999875e-05,
      "loss": 0.5998354434967041,
      "mean_token_accuracy": 0.8481284163892269,
      "num_tokens": 39319667.0,
      "step": 19230
    },
    {
      "entropy": 0.6301240844652056,
      "epoch": 0.3124822360994941,
      "grad_norm": 8.8125,
      "learning_rate": 3.9623839046073305e-05,
      "loss": 0.642618465423584,
      "mean_token_accuracy": 0.837563606724143,
      "num_tokens": 39340429.0,
      "step": 19240
    },
    {
      "entropy": 0.4826443509198725,
      "epoch": 0.3126446489041196,
      "grad_norm": 9.75,
      "learning_rate": 3.9613320498302304e-05,
      "loss": 0.47468152046203616,
      "mean_token_accuracy": 0.876922146230936,
      "num_tokens": 39360430.0,
      "step": 19250
    },
    {
      "entropy": 0.5668265367858112,
      "epoch": 0.3128070617087451,
      "grad_norm": 9.4375,
      "learning_rate": 3.9602798019515233e-05,
      "loss": 0.5279725074768067,
      "mean_token_accuracy": 0.8610369861125946,
      "num_tokens": 39380891.0,
      "step": 19260
    },
    {
      "entropy": 0.5270667726173996,
      "epoch": 0.31296947451337065,
      "grad_norm": 9.375,
      "learning_rate": 3.959227161254269e-05,
      "loss": 0.5035872459411621,
      "mean_token_accuracy": 0.867455618083477,
      "num_tokens": 39401129.0,
      "step": 19270
    },
    {
      "entropy": 0.5017395663075149,
      "epoch": 0.3131318873179961,
      "grad_norm": 5.90625,
      "learning_rate": 3.958174128021628e-05,
      "loss": 0.511758279800415,
      "mean_token_accuracy": 0.8668456487357616,
      "num_tokens": 39421192.0,
      "step": 19280
    },
    {
      "entropy": 0.5174957217648626,
      "epoch": 0.31329430012262166,
      "grad_norm": 8.0625,
      "learning_rate": 3.9571207025368695e-05,
      "loss": 0.5325199604034424,
      "mean_token_accuracy": 0.8600674942135811,
      "num_tokens": 39441591.0,
      "step": 19290
    },
    {
      "entropy": 0.5682986180298031,
      "epoch": 0.3134567129272472,
      "grad_norm": 7.6875,
      "learning_rate": 3.9560668850833674e-05,
      "loss": 0.5601281642913818,
      "mean_token_accuracy": 0.8608564235270023,
      "num_tokens": 39460923.0,
      "step": 19300
    },
    {
      "entropy": 0.533686973946169,
      "epoch": 0.3136191257318727,
      "grad_norm": 6.59375,
      "learning_rate": 3.955012675944602e-05,
      "loss": 0.544423770904541,
      "mean_token_accuracy": 0.8653843514621258,
      "num_tokens": 39481678.0,
      "step": 19310
    },
    {
      "entropy": 0.5507425420219079,
      "epoch": 0.3137815385364982,
      "grad_norm": 9.6875,
      "learning_rate": 3.953958075404156e-05,
      "loss": 0.5740674495697021,
      "mean_token_accuracy": 0.8559039458632469,
      "num_tokens": 39502316.0,
      "step": 19320
    },
    {
      "entropy": 0.5801454225555063,
      "epoch": 0.31394395134112374,
      "grad_norm": 6.34375,
      "learning_rate": 3.95290308374572e-05,
      "loss": 0.5982348918914795,
      "mean_token_accuracy": 0.8555800009518861,
      "num_tokens": 39522596.0,
      "step": 19330
    },
    {
      "entropy": 0.6091916370671242,
      "epoch": 0.3141063641457493,
      "grad_norm": 6.625,
      "learning_rate": 3.951847701253089e-05,
      "loss": 0.6586312770843505,
      "mean_token_accuracy": 0.8406965442001819,
      "num_tokens": 39542470.0,
      "step": 19340
    },
    {
      "entropy": 0.5263452836312353,
      "epoch": 0.31426877695037475,
      "grad_norm": 7.0,
      "learning_rate": 3.950791928210164e-05,
      "loss": 0.5331189632415771,
      "mean_token_accuracy": 0.8639646083116531,
      "num_tokens": 39562192.0,
      "step": 19350
    },
    {
      "entropy": 0.5400573233608157,
      "epoch": 0.3144311897550003,
      "grad_norm": 6.125,
      "learning_rate": 3.94973576490095e-05,
      "loss": 0.5522817611694336,
      "mean_token_accuracy": 0.8622964017093182,
      "num_tokens": 39582132.0,
      "step": 19360
    },
    {
      "entropy": 0.6017205367796123,
      "epoch": 0.3145936025596258,
      "grad_norm": 10.0625,
      "learning_rate": 3.948679211609557e-05,
      "loss": 0.6157132148742676,
      "mean_token_accuracy": 0.8447596989572048,
      "num_tokens": 39602548.0,
      "step": 19370
    },
    {
      "entropy": 0.5592338665854186,
      "epoch": 0.3147560153642513,
      "grad_norm": 8.0,
      "learning_rate": 3.9476222686202004e-05,
      "loss": 0.6307149887084961,
      "mean_token_accuracy": 0.8462336950004101,
      "num_tokens": 39623627.0,
      "step": 19380
    },
    {
      "entropy": 0.5967822981532663,
      "epoch": 0.31491842816887683,
      "grad_norm": 4.9375,
      "learning_rate": 3.9465649362171994e-05,
      "loss": 0.5895767211914062,
      "mean_token_accuracy": 0.8498439226299525,
      "num_tokens": 39644228.0,
      "step": 19390
    },
    {
      "entropy": 0.5383216224145144,
      "epoch": 0.31508084097350236,
      "grad_norm": 7.90625,
      "learning_rate": 3.9455072146849815e-05,
      "loss": 0.5457468032836914,
      "mean_token_accuracy": 0.8613985605537892,
      "num_tokens": 39664139.0,
      "step": 19400
    },
    {
      "entropy": 0.6676898158621043,
      "epoch": 0.31524325377812784,
      "grad_norm": 6.625,
      "learning_rate": 3.944449104308072e-05,
      "loss": 0.6607512950897216,
      "mean_token_accuracy": 0.8377457454800605,
      "num_tokens": 39685399.0,
      "step": 19410
    },
    {
      "entropy": 0.6280208151321858,
      "epoch": 0.3154056665827534,
      "grad_norm": 9.3125,
      "learning_rate": 3.943390605371109e-05,
      "loss": 0.6138193130493164,
      "mean_token_accuracy": 0.847489869222045,
      "num_tokens": 39707706.0,
      "step": 19420
    },
    {
      "entropy": 0.628469105064869,
      "epoch": 0.3155680793873789,
      "grad_norm": 7.78125,
      "learning_rate": 3.9423317181588284e-05,
      "loss": 0.6331322193145752,
      "mean_token_accuracy": 0.8453275434672832,
      "num_tokens": 39728038.0,
      "step": 19430
    },
    {
      "entropy": 0.5060825611464679,
      "epoch": 0.31573049219200444,
      "grad_norm": 10.0,
      "learning_rate": 3.9412724429560754e-05,
      "loss": 0.5083190441131592,
      "mean_token_accuracy": 0.8726790450513363,
      "num_tokens": 39748514.0,
      "step": 19440
    },
    {
      "entropy": 0.4799587679095566,
      "epoch": 0.3158929049966299,
      "grad_norm": 6.5625,
      "learning_rate": 3.9402127800477944e-05,
      "loss": 0.47894110679626467,
      "mean_token_accuracy": 0.8799192920327187,
      "num_tokens": 39768195.0,
      "step": 19450
    },
    {
      "entropy": 0.5562040118500591,
      "epoch": 0.31605531780125545,
      "grad_norm": 11.125,
      "learning_rate": 3.93915272971904e-05,
      "loss": 0.5658151149749756,
      "mean_token_accuracy": 0.8553099315613508,
      "num_tokens": 39787899.0,
      "step": 19460
    },
    {
      "entropy": 0.5441144986078144,
      "epoch": 0.316217730605881,
      "grad_norm": 7.25,
      "learning_rate": 3.938092292254967e-05,
      "loss": 0.5577087879180909,
      "mean_token_accuracy": 0.853182328492403,
      "num_tokens": 39808082.0,
      "step": 19470
    },
    {
      "entropy": 0.48495374359190463,
      "epoch": 0.31638014341050646,
      "grad_norm": 7.9375,
      "learning_rate": 3.937031467940835e-05,
      "loss": 0.4723358154296875,
      "mean_token_accuracy": 0.8687118560075759,
      "num_tokens": 39828651.0,
      "step": 19480
    },
    {
      "entropy": 0.5390023849904537,
      "epoch": 0.316542556215132,
      "grad_norm": 6.6875,
      "learning_rate": 3.935970257062009e-05,
      "loss": 0.515623664855957,
      "mean_token_accuracy": 0.8600548416376114,
      "num_tokens": 39849089.0,
      "step": 19490
    },
    {
      "entropy": 0.534993429388851,
      "epoch": 0.31670496901975753,
      "grad_norm": 6.28125,
      "learning_rate": 3.934908659903956e-05,
      "loss": 0.5749815940856934,
      "mean_token_accuracy": 0.8567329987883567,
      "num_tokens": 39869834.0,
      "step": 19500
    },
    {
      "entropy": 0.5613266935572028,
      "epoch": 0.316867381824383,
      "grad_norm": 7.0625,
      "learning_rate": 3.9338466767522496e-05,
      "loss": 0.5985610961914063,
      "mean_token_accuracy": 0.8471956968307495,
      "num_tokens": 39890325.0,
      "step": 19510
    },
    {
      "entropy": 0.5950109224766493,
      "epoch": 0.31702979462900854,
      "grad_norm": 9.5625,
      "learning_rate": 3.932784307892565e-05,
      "loss": 0.6104081630706787,
      "mean_token_accuracy": 0.8508874930441379,
      "num_tokens": 39910724.0,
      "step": 19520
    },
    {
      "entropy": 0.5881017514504492,
      "epoch": 0.3171922074336341,
      "grad_norm": 8.25,
      "learning_rate": 3.931721553610681e-05,
      "loss": 0.5622600555419922,
      "mean_token_accuracy": 0.860868314653635,
      "num_tokens": 39931531.0,
      "step": 19530
    },
    {
      "entropy": 0.5962625299114734,
      "epoch": 0.3173546202382596,
      "grad_norm": 7.5625,
      "learning_rate": 3.930658414192481e-05,
      "loss": 0.5906525135040284,
      "mean_token_accuracy": 0.8515744864940643,
      "num_tokens": 39951722.0,
      "step": 19540
    },
    {
      "entropy": 0.5768448542803526,
      "epoch": 0.3175170330428851,
      "grad_norm": 7.375,
      "learning_rate": 3.9295948899239534e-05,
      "loss": 0.5756896495819092,
      "mean_token_accuracy": 0.8503908183425665,
      "num_tokens": 39972275.0,
      "step": 19550
    },
    {
      "entropy": 0.5503929348662495,
      "epoch": 0.3176794458475106,
      "grad_norm": 7.03125,
      "learning_rate": 3.928530981091187e-05,
      "loss": 0.5346224784851075,
      "mean_token_accuracy": 0.8624202698469162,
      "num_tokens": 39993761.0,
      "step": 19560
    },
    {
      "entropy": 0.6092671368271112,
      "epoch": 0.31784185865213616,
      "grad_norm": 8.375,
      "learning_rate": 3.927466687980376e-05,
      "loss": 0.6240594387054443,
      "mean_token_accuracy": 0.8406522646546364,
      "num_tokens": 40014892.0,
      "step": 19570
    },
    {
      "entropy": 0.5335901903919875,
      "epoch": 0.31800427145676163,
      "grad_norm": 5.125,
      "learning_rate": 3.926402010877819e-05,
      "loss": 0.5396511077880859,
      "mean_token_accuracy": 0.8637661635875702,
      "num_tokens": 40035232.0,
      "step": 19580
    },
    {
      "entropy": 0.5771766271442175,
      "epoch": 0.31816668426138717,
      "grad_norm": 5.90625,
      "learning_rate": 3.925336950069915e-05,
      "loss": 0.5831156730651855,
      "mean_token_accuracy": 0.8487849824130536,
      "num_tokens": 40056401.0,
      "step": 19590
    },
    {
      "entropy": 0.5180296842008829,
      "epoch": 0.3183290970660127,
      "grad_norm": 9.4375,
      "learning_rate": 3.924271505843167e-05,
      "loss": 0.5219240665435791,
      "mean_token_accuracy": 0.8592601247131825,
      "num_tokens": 40075233.0,
      "step": 19600
    },
    {
      "entropy": 0.56553725739941,
      "epoch": 0.3184915098706382,
      "grad_norm": 5.34375,
      "learning_rate": 3.923205678484184e-05,
      "loss": 0.5500072479248047,
      "mean_token_accuracy": 0.8595289554446935,
      "num_tokens": 40095390.0,
      "step": 19610
    },
    {
      "entropy": 0.5385913213714957,
      "epoch": 0.3186539226752637,
      "grad_norm": 6.96875,
      "learning_rate": 3.9221394682796756e-05,
      "loss": 0.5421299457550048,
      "mean_token_accuracy": 0.8707927107810974,
      "num_tokens": 40116661.0,
      "step": 19620
    },
    {
      "entropy": 0.5736958432011307,
      "epoch": 0.31881633547988925,
      "grad_norm": 6.78125,
      "learning_rate": 3.921072875516454e-05,
      "loss": 0.598740816116333,
      "mean_token_accuracy": 0.8546320587396622,
      "num_tokens": 40136935.0,
      "step": 19630
    },
    {
      "entropy": 0.6134945794008673,
      "epoch": 0.3189787482845147,
      "grad_norm": 10.25,
      "learning_rate": 3.920005900481434e-05,
      "loss": 0.6222496509552002,
      "mean_token_accuracy": 0.8509710840880871,
      "num_tokens": 40157665.0,
      "step": 19640
    },
    {
      "entropy": 0.6565963018685579,
      "epoch": 0.31914116108914026,
      "grad_norm": 7.5,
      "learning_rate": 3.9189385434616366e-05,
      "loss": 0.6266607284545899,
      "mean_token_accuracy": 0.8395815525203943,
      "num_tokens": 40179034.0,
      "step": 19650
    },
    {
      "entropy": 0.5256908962503075,
      "epoch": 0.3193035738937658,
      "grad_norm": 8.5625,
      "learning_rate": 3.917870804744182e-05,
      "loss": 0.5384593963623047,
      "mean_token_accuracy": 0.8644326299428939,
      "num_tokens": 40199909.0,
      "step": 19660
    },
    {
      "entropy": 0.6060989880934358,
      "epoch": 0.3194659866983913,
      "grad_norm": 9.75,
      "learning_rate": 3.916802684616294e-05,
      "loss": 0.6036511421203613,
      "mean_token_accuracy": 0.8500706322491169,
      "num_tokens": 40220992.0,
      "step": 19670
    },
    {
      "entropy": 0.575438390346244,
      "epoch": 0.3196283995030168,
      "grad_norm": 7.78125,
      "learning_rate": 3.9157341833653e-05,
      "loss": 0.606380033493042,
      "mean_token_accuracy": 0.8532156061381102,
      "num_tokens": 40242990.0,
      "step": 19680
    },
    {
      "entropy": 0.540601222589612,
      "epoch": 0.31979081230764234,
      "grad_norm": 5.96875,
      "learning_rate": 3.9146653012786295e-05,
      "loss": 0.5079649925231934,
      "mean_token_accuracy": 0.8659315899014473,
      "num_tokens": 40263202.0,
      "step": 19690
    },
    {
      "entropy": 0.5532876365818084,
      "epoch": 0.31995322511226787,
      "grad_norm": 6.84375,
      "learning_rate": 3.913596038643814e-05,
      "loss": 0.5773871898651123,
      "mean_token_accuracy": 0.8529730640351773,
      "num_tokens": 40284519.0,
      "step": 19700
    },
    {
      "entropy": 0.5379133305512369,
      "epoch": 0.32011563791689335,
      "grad_norm": 7.46875,
      "learning_rate": 3.912526395748486e-05,
      "loss": 0.5361634731292725,
      "mean_token_accuracy": 0.8628397792577743,
      "num_tokens": 40305255.0,
      "step": 19710
    },
    {
      "entropy": 0.5893181855790317,
      "epoch": 0.3202780507215189,
      "grad_norm": 10.6875,
      "learning_rate": 3.911456372880383e-05,
      "loss": 0.6547330856323242,
      "mean_token_accuracy": 0.8449287779629231,
      "num_tokens": 40326213.0,
      "step": 19720
    },
    {
      "entropy": 0.5179369143210352,
      "epoch": 0.3204404635261444,
      "grad_norm": 7.6875,
      "learning_rate": 3.910385970327345e-05,
      "loss": 0.537169075012207,
      "mean_token_accuracy": 0.860140572488308,
      "num_tokens": 40347883.0,
      "step": 19730
    },
    {
      "entropy": 0.5375458169262857,
      "epoch": 0.3206028763307699,
      "grad_norm": 7.375,
      "learning_rate": 3.90931518837731e-05,
      "loss": 0.5277934551239014,
      "mean_token_accuracy": 0.871111424267292,
      "num_tokens": 40367913.0,
      "step": 19740
    },
    {
      "entropy": 0.5792287923395634,
      "epoch": 0.3207652891353954,
      "grad_norm": 5.5,
      "learning_rate": 3.908244027318323e-05,
      "loss": 0.5967052936553955,
      "mean_token_accuracy": 0.8489834614098072,
      "num_tokens": 40387807.0,
      "step": 19750
    },
    {
      "entropy": 0.5935343234799803,
      "epoch": 0.32092770194002096,
      "grad_norm": 6.53125,
      "learning_rate": 3.9071724874385274e-05,
      "loss": 0.6210056781768799,
      "mean_token_accuracy": 0.8529943019151688,
      "num_tokens": 40407614.0,
      "step": 19760
    },
    {
      "entropy": 0.5722434962168336,
      "epoch": 0.3210901147446465,
      "grad_norm": 7.5625,
      "learning_rate": 3.906100569026171e-05,
      "loss": 0.5679896354675293,
      "mean_token_accuracy": 0.8505684830248356,
      "num_tokens": 40429289.0,
      "step": 19770
    },
    {
      "entropy": 0.545999848563224,
      "epoch": 0.32125252754927197,
      "grad_norm": 12.0625,
      "learning_rate": 3.905028272369601e-05,
      "loss": 0.5709371089935302,
      "mean_token_accuracy": 0.8496932595968246,
      "num_tokens": 40449467.0,
      "step": 19780
    },
    {
      "entropy": 0.5450378015637398,
      "epoch": 0.3214149403538975,
      "grad_norm": 9.125,
      "learning_rate": 3.903955597757268e-05,
      "loss": 0.5618586540222168,
      "mean_token_accuracy": 0.8544690035283565,
      "num_tokens": 40470102.0,
      "step": 19790
    },
    {
      "entropy": 0.580790686327964,
      "epoch": 0.32157735315852304,
      "grad_norm": 6.53125,
      "learning_rate": 3.902882545477724e-05,
      "loss": 0.5696881771087646,
      "mean_token_accuracy": 0.8549814328551293,
      "num_tokens": 40490578.0,
      "step": 19800
    },
    {
      "entropy": 0.6026866814587265,
      "epoch": 0.3217397659631485,
      "grad_norm": 5.5625,
      "learning_rate": 3.9018091158196224e-05,
      "loss": 0.5902295589447022,
      "mean_token_accuracy": 0.8537688009440899,
      "num_tokens": 40511071.0,
      "step": 19810
    },
    {
      "entropy": 0.5318947600200772,
      "epoch": 0.32190217876777405,
      "grad_norm": 9.375,
      "learning_rate": 3.900735309071719e-05,
      "loss": 0.5128041267395019,
      "mean_token_accuracy": 0.8675425566732884,
      "num_tokens": 40530925.0,
      "step": 19820
    },
    {
      "entropy": 0.5283474426716566,
      "epoch": 0.3220645915723996,
      "grad_norm": 8.625,
      "learning_rate": 3.899661125522868e-05,
      "loss": 0.5639280319213867,
      "mean_token_accuracy": 0.8646511472761631,
      "num_tokens": 40551964.0,
      "step": 19830
    },
    {
      "entropy": 0.6227468535304069,
      "epoch": 0.32222700437702506,
      "grad_norm": 10.3125,
      "learning_rate": 3.89858656546203e-05,
      "loss": 0.6629209518432617,
      "mean_token_accuracy": 0.8413042537868023,
      "num_tokens": 40573791.0,
      "step": 19840
    },
    {
      "entropy": 0.5881687111221254,
      "epoch": 0.3223894171816506,
      "grad_norm": 7.9375,
      "learning_rate": 3.897511629178262e-05,
      "loss": 0.5973340988159179,
      "mean_token_accuracy": 0.8471304021775723,
      "num_tokens": 40594664.0,
      "step": 19850
    },
    {
      "entropy": 0.5967919916380197,
      "epoch": 0.32255182998627613,
      "grad_norm": 8.5625,
      "learning_rate": 3.896436316960726e-05,
      "loss": 0.5865896224975586,
      "mean_token_accuracy": 0.8562688034027814,
      "num_tokens": 40614112.0,
      "step": 19860
    },
    {
      "entropy": 0.6108686328399926,
      "epoch": 0.32271424279090166,
      "grad_norm": 7.0625,
      "learning_rate": 3.8953606290986814e-05,
      "loss": 0.6083669185638427,
      "mean_token_accuracy": 0.8412919618189335,
      "num_tokens": 40635397.0,
      "step": 19870
    },
    {
      "entropy": 0.5203675724565983,
      "epoch": 0.32287665559552714,
      "grad_norm": 8.8125,
      "learning_rate": 3.894284565881493e-05,
      "loss": 0.49150424003601073,
      "mean_token_accuracy": 0.8730705507099629,
      "num_tokens": 40655704.0,
      "step": 19880
    },
    {
      "entropy": 0.5716422371566295,
      "epoch": 0.3230390684001527,
      "grad_norm": 7.875,
      "learning_rate": 3.893208127598622e-05,
      "loss": 0.5687380313873291,
      "mean_token_accuracy": 0.8547327250242234,
      "num_tokens": 40676354.0,
      "step": 19890
    },
    {
      "entropy": 0.5656626813579351,
      "epoch": 0.3232014812047782,
      "grad_norm": 10.5,
      "learning_rate": 3.8921313145396334e-05,
      "loss": 0.5647895336151123,
      "mean_token_accuracy": 0.8536812320351601,
      "num_tokens": 40696530.0,
      "step": 19900
    },
    {
      "entropy": 0.5409168111160397,
      "epoch": 0.3233638940094037,
      "grad_norm": 7.25,
      "learning_rate": 3.8910541269941924e-05,
      "loss": 0.5285530090332031,
      "mean_token_accuracy": 0.8732989732176065,
      "num_tokens": 40717601.0,
      "step": 19910
    },
    {
      "entropy": 0.6130751780699939,
      "epoch": 0.3235263068140292,
      "grad_norm": 8.25,
      "learning_rate": 3.889976565252066e-05,
      "loss": 0.6324321746826171,
      "mean_token_accuracy": 0.8532677531242371,
      "num_tokens": 40737814.0,
      "step": 19920
    },
    {
      "entropy": 0.5145825708284975,
      "epoch": 0.32368871961865475,
      "grad_norm": 10.125,
      "learning_rate": 3.888898629603119e-05,
      "loss": 0.5276144027709961,
      "mean_token_accuracy": 0.8675759010016918,
      "num_tokens": 40758135.0,
      "step": 19930
    },
    {
      "entropy": 0.5005684619769454,
      "epoch": 0.32385113242328023,
      "grad_norm": 6.84375,
      "learning_rate": 3.88782032033732e-05,
      "loss": 0.5135180473327636,
      "mean_token_accuracy": 0.8687723644077778,
      "num_tokens": 40778330.0,
      "step": 19940
    },
    {
      "entropy": 0.5062372504500672,
      "epoch": 0.32401354522790576,
      "grad_norm": 7.5625,
      "learning_rate": 3.886741637744735e-05,
      "loss": 0.5013963222503662,
      "mean_token_accuracy": 0.8653702154755593,
      "num_tokens": 40798835.0,
      "step": 19950
    },
    {
      "entropy": 0.5539556500501931,
      "epoch": 0.3241759580325313,
      "grad_norm": 6.5,
      "learning_rate": 3.885662582115534e-05,
      "loss": 0.5449463367462158,
      "mean_token_accuracy": 0.8528797946870327,
      "num_tokens": 40817518.0,
      "step": 19960
    },
    {
      "entropy": 0.5752821778412909,
      "epoch": 0.3243383708371568,
      "grad_norm": 8.5,
      "learning_rate": 3.8845831537399835e-05,
      "loss": 0.5447644710540771,
      "mean_token_accuracy": 0.8640910472720862,
      "num_tokens": 40838538.0,
      "step": 19970
    },
    {
      "entropy": 0.5251571651548147,
      "epoch": 0.3245007836417823,
      "grad_norm": 7.15625,
      "learning_rate": 3.883503352908453e-05,
      "loss": 0.5643019676208496,
      "mean_token_accuracy": 0.8531571350991726,
      "num_tokens": 40858324.0,
      "step": 19980
    },
    {
      "entropy": 0.528089985344559,
      "epoch": 0.32466319644640784,
      "grad_norm": 6.8125,
      "learning_rate": 3.8824231799114116e-05,
      "loss": 0.5477115154266358,
      "mean_token_accuracy": 0.866584275662899,
      "num_tokens": 40877730.0,
      "step": 19990
    },
    {
      "epoch": 0.3248256092510334,
      "eval_entropy": 0.5932083523869515,
      "eval_loss": 0.5908288955688477,
      "eval_mean_token_accuracy": 0.8494270906448365,
      "eval_num_tokens": 40897311.0,
      "eval_runtime": 40.1073,
      "eval_samples_per_second": 49.866,
      "eval_steps_per_second": 6.233,
      "step": 20000
    },
    {
      "entropy": 0.5258246382931248,
      "epoch": 0.32498802205565885,
      "grad_norm": 9.375,
      "learning_rate": 3.880261718583171e-05,
      "loss": 0.5176889896392822,
      "mean_token_accuracy": 0.8661935655400157,
      "num_tokens": 40916775.0,
      "step": 20010
    },
    {
      "entropy": 0.5707350197248161,
      "epoch": 0.3251504348602844,
      "grad_norm": 8.6875,
      "learning_rate": 3.87918043083341e-05,
      "loss": 0.5853331089019775,
      "mean_token_accuracy": 0.8564554944634437,
      "num_tokens": 40938589.0,
      "step": 20020
    },
    {
      "entropy": 0.5551344627514482,
      "epoch": 0.3253128476649099,
      "grad_norm": 12.3125,
      "learning_rate": 3.8780987720810135e-05,
      "loss": 0.5574994087219238,
      "mean_token_accuracy": 0.8442889615893364,
      "num_tokens": 40959538.0,
      "step": 20030
    },
    {
      "entropy": 0.539919011760503,
      "epoch": 0.3254752604695354,
      "grad_norm": 8.4375,
      "learning_rate": 3.877016742616951e-05,
      "loss": 0.5372710704803467,
      "mean_token_accuracy": 0.8631287947297096,
      "num_tokens": 40980152.0,
      "step": 20040
    },
    {
      "entropy": 0.5584120092913508,
      "epoch": 0.32563767327416093,
      "grad_norm": 7.6875,
      "learning_rate": 3.875934342732289e-05,
      "loss": 0.5717068195343018,
      "mean_token_accuracy": 0.8580597907304763,
      "num_tokens": 40999864.0,
      "step": 20050
    },
    {
      "entropy": 0.616081522218883,
      "epoch": 0.32580008607878647,
      "grad_norm": 6.6875,
      "learning_rate": 3.8748515727181964e-05,
      "loss": 0.6341357707977295,
      "mean_token_accuracy": 0.8412978053092957,
      "num_tokens": 41019418.0,
      "step": 20060
    },
    {
      "entropy": 0.5850926938466727,
      "epoch": 0.32596249888341194,
      "grad_norm": 9.4375,
      "learning_rate": 3.873768432865942e-05,
      "loss": 0.6184796333312989,
      "mean_token_accuracy": 0.8462235383689404,
      "num_tokens": 41039609.0,
      "step": 20070
    },
    {
      "entropy": 0.533634114637971,
      "epoch": 0.3261249116880375,
      "grad_norm": 7.03125,
      "learning_rate": 3.872684923466891e-05,
      "loss": 0.5075406074523926,
      "mean_token_accuracy": 0.8738115198910237,
      "num_tokens": 41059630.0,
      "step": 20080
    },
    {
      "entropy": 0.5464841176290065,
      "epoch": 0.326287324492663,
      "grad_norm": 8.5,
      "learning_rate": 3.8716010448125106e-05,
      "loss": 0.5541236400604248,
      "mean_token_accuracy": 0.862818793207407,
      "num_tokens": 41080850.0,
      "step": 20090
    },
    {
      "entropy": 0.569511732365936,
      "epoch": 0.32644973729728854,
      "grad_norm": 6.21875,
      "learning_rate": 3.8705167971943676e-05,
      "loss": 0.5877649307250976,
      "mean_token_accuracy": 0.857048773765564,
      "num_tokens": 41103159.0,
      "step": 20100
    },
    {
      "entropy": 0.6059314062818885,
      "epoch": 0.326612150101914,
      "grad_norm": 8.75,
      "learning_rate": 3.8694321809041254e-05,
      "loss": 0.5983298778533935,
      "mean_token_accuracy": 0.8511995449662209,
      "num_tokens": 41124357.0,
      "step": 20110
    },
    {
      "entropy": 0.5439999983180315,
      "epoch": 0.32677456290653956,
      "grad_norm": 7.71875,
      "learning_rate": 3.8683471962335494e-05,
      "loss": 0.5570600509643555,
      "mean_token_accuracy": 0.8575830176472664,
      "num_tokens": 41144581.0,
      "step": 20120
    },
    {
      "entropy": 0.5316380645148456,
      "epoch": 0.3269369757111651,
      "grad_norm": 8.125,
      "learning_rate": 3.8672618434745016e-05,
      "loss": 0.5397607326507569,
      "mean_token_accuracy": 0.8645428784191609,
      "num_tokens": 41164880.0,
      "step": 20130
    },
    {
      "entropy": 0.5089918345678598,
      "epoch": 0.32709938851579057,
      "grad_norm": 5.8125,
      "learning_rate": 3.866176122918946e-05,
      "loss": 0.5502643585205078,
      "mean_token_accuracy": 0.8622263945639134,
      "num_tokens": 41185854.0,
      "step": 20140
    },
    {
      "entropy": 0.5675724079832435,
      "epoch": 0.3272618013204161,
      "grad_norm": 9.1875,
      "learning_rate": 3.865090034858942e-05,
      "loss": 0.5718104839324951,
      "mean_token_accuracy": 0.8518867403268814,
      "num_tokens": 41206482.0,
      "step": 20150
    },
    {
      "entropy": 0.5881489543244243,
      "epoch": 0.32742421412504163,
      "grad_norm": 6.375,
      "learning_rate": 3.8640035795866525e-05,
      "loss": 0.5723332405090332,
      "mean_token_accuracy": 0.8669893182814121,
      "num_tokens": 41226597.0,
      "step": 20160
    },
    {
      "entropy": 0.5522896509617568,
      "epoch": 0.3275866269296671,
      "grad_norm": 7.40625,
      "learning_rate": 3.8629167573943334e-05,
      "loss": 0.5575553417205811,
      "mean_token_accuracy": 0.8607357569038868,
      "num_tokens": 41246492.0,
      "step": 20170
    },
    {
      "entropy": 0.5054645639844239,
      "epoch": 0.32774903973429265,
      "grad_norm": 8.5625,
      "learning_rate": 3.8618295685743436e-05,
      "loss": 0.5064489364624023,
      "mean_token_accuracy": 0.8735892698168755,
      "num_tokens": 41265524.0,
      "step": 20180
    },
    {
      "entropy": 0.5575888625346124,
      "epoch": 0.3279114525389182,
      "grad_norm": 9.1875,
      "learning_rate": 3.86074201341914e-05,
      "loss": 0.5426346778869628,
      "mean_token_accuracy": 0.8707714308053255,
      "num_tokens": 41287850.0,
      "step": 20190
    },
    {
      "entropy": 0.5646102683618665,
      "epoch": 0.32807386534354366,
      "grad_norm": 8.9375,
      "learning_rate": 3.859654092221276e-05,
      "loss": 0.5689468383789062,
      "mean_token_accuracy": 0.848051880300045,
      "num_tokens": 41307815.0,
      "step": 20200
    },
    {
      "entropy": 0.5392456836998463,
      "epoch": 0.3282362781481692,
      "grad_norm": 10.0625,
      "learning_rate": 3.858565805273406e-05,
      "loss": 0.5314042568206787,
      "mean_token_accuracy": 0.8643523506820202,
      "num_tokens": 41328445.0,
      "step": 20210
    },
    {
      "entropy": 0.6064204074442386,
      "epoch": 0.3283986909527947,
      "grad_norm": 8.1875,
      "learning_rate": 3.8574771528682806e-05,
      "loss": 0.6308775424957276,
      "mean_token_accuracy": 0.8465713232755661,
      "num_tokens": 41349405.0,
      "step": 20220
    },
    {
      "entropy": 0.502283345349133,
      "epoch": 0.32856110375742026,
      "grad_norm": 8.25,
      "learning_rate": 3.8563881352987516e-05,
      "loss": 0.5050107955932617,
      "mean_token_accuracy": 0.8683884404599667,
      "num_tokens": 41369445.0,
      "step": 20230
    },
    {
      "entropy": 0.5671620949171483,
      "epoch": 0.32872351656204574,
      "grad_norm": 4.90625,
      "learning_rate": 3.855298752857764e-05,
      "loss": 0.5647209167480469,
      "mean_token_accuracy": 0.8586450040340423,
      "num_tokens": 41391232.0,
      "step": 20240
    },
    {
      "entropy": 0.6342523289844394,
      "epoch": 0.32888592936667127,
      "grad_norm": 6.53125,
      "learning_rate": 3.854209005838368e-05,
      "loss": 0.6508600234985351,
      "mean_token_accuracy": 0.8338122446089983,
      "num_tokens": 41412527.0,
      "step": 20250
    },
    {
      "entropy": 0.6246680040843785,
      "epoch": 0.3290483421712968,
      "grad_norm": 11.8125,
      "learning_rate": 3.853118894533705e-05,
      "loss": 0.6561876773834229,
      "mean_token_accuracy": 0.8361747726798058,
      "num_tokens": 41433210.0,
      "step": 20260
    },
    {
      "entropy": 0.6017771938815712,
      "epoch": 0.3292107549759223,
      "grad_norm": 13.5625,
      "learning_rate": 3.8520284192370195e-05,
      "loss": 0.6240430355072022,
      "mean_token_accuracy": 0.8496897026896477,
      "num_tokens": 41453950.0,
      "step": 20270
    },
    {
      "entropy": 0.5229405462741852,
      "epoch": 0.3293731677805478,
      "grad_norm": 6.84375,
      "learning_rate": 3.850937580241649e-05,
      "loss": 0.5229180812835693,
      "mean_token_accuracy": 0.8722349815070629,
      "num_tokens": 41473968.0,
      "step": 20280
    },
    {
      "entropy": 0.5189528007991612,
      "epoch": 0.32953558058517335,
      "grad_norm": 6.25,
      "learning_rate": 3.8498463778410346e-05,
      "loss": 0.5390919208526611,
      "mean_token_accuracy": 0.8588165313005447,
      "num_tokens": 41495041.0,
      "step": 20290
    },
    {
      "entropy": 0.4963974079117179,
      "epoch": 0.3296979933897988,
      "grad_norm": 10.0,
      "learning_rate": 3.84875481232871e-05,
      "loss": 0.49689221382141113,
      "mean_token_accuracy": 0.8699768863618373,
      "num_tokens": 41514329.0,
      "step": 20300
    },
    {
      "entropy": 0.5069220548961312,
      "epoch": 0.32986040619442436,
      "grad_norm": 5.59375,
      "learning_rate": 3.8476628839983106e-05,
      "loss": 0.5296567440032959,
      "mean_token_accuracy": 0.8658665500581264,
      "num_tokens": 41534674.0,
      "step": 20310
    },
    {
      "entropy": 0.5749042716808617,
      "epoch": 0.3300228189990499,
      "grad_norm": 6.59375,
      "learning_rate": 3.846570593143566e-05,
      "loss": 0.5957942962646484,
      "mean_token_accuracy": 0.8556736081838607,
      "num_tokens": 41555557.0,
      "step": 20320
    },
    {
      "entropy": 0.562180996593088,
      "epoch": 0.3301852318036754,
      "grad_norm": 7.6875,
      "learning_rate": 3.845477940058307e-05,
      "loss": 0.587843656539917,
      "mean_token_accuracy": 0.8556431740522384,
      "num_tokens": 41576400.0,
      "step": 20330
    },
    {
      "entropy": 0.5460783205926418,
      "epoch": 0.3303476446083009,
      "grad_norm": 10.3125,
      "learning_rate": 3.844384925036457e-05,
      "loss": 0.529017162322998,
      "mean_token_accuracy": 0.8624743930995464,
      "num_tokens": 41597303.0,
      "step": 20340
    },
    {
      "entropy": 0.6169278368353843,
      "epoch": 0.33051005741292644,
      "grad_norm": 8.1875,
      "learning_rate": 3.843291548372041e-05,
      "loss": 0.614155387878418,
      "mean_token_accuracy": 0.8594807602465153,
      "num_tokens": 41616974.0,
      "step": 20350
    },
    {
      "entropy": 0.5893424484878779,
      "epoch": 0.330672470217552,
      "grad_norm": 11.0,
      "learning_rate": 3.84219781035918e-05,
      "loss": 0.5875109195709228,
      "mean_token_accuracy": 0.8550508454442024,
      "num_tokens": 41637824.0,
      "step": 20360
    },
    {
      "entropy": 0.6058869617059827,
      "epoch": 0.33083488302217745,
      "grad_norm": 5.71875,
      "learning_rate": 3.841103711292092e-05,
      "loss": 0.5774574279785156,
      "mean_token_accuracy": 0.8547391809523106,
      "num_tokens": 41659031.0,
      "step": 20370
    },
    {
      "entropy": 0.5102329127956182,
      "epoch": 0.330997295826803,
      "grad_norm": 7.0,
      "learning_rate": 3.8400092514650913e-05,
      "loss": 0.5139196872711181,
      "mean_token_accuracy": 0.8710854962468147,
      "num_tokens": 41679457.0,
      "step": 20380
    },
    {
      "entropy": 0.5354670776519924,
      "epoch": 0.3311597086314285,
      "grad_norm": 8.3125,
      "learning_rate": 3.83891443117259e-05,
      "loss": 0.5355747699737549,
      "mean_token_accuracy": 0.8689523614943028,
      "num_tokens": 41699438.0,
      "step": 20390
    },
    {
      "entropy": 0.5157699976116419,
      "epoch": 0.331322121436054,
      "grad_norm": 9.125,
      "learning_rate": 3.837819250709098e-05,
      "loss": 0.5157564640045166,
      "mean_token_accuracy": 0.8631924841552973,
      "num_tokens": 41719843.0,
      "step": 20400
    },
    {
      "entropy": 0.52330502262339,
      "epoch": 0.33148453424067953,
      "grad_norm": 8.0,
      "learning_rate": 3.836723710369222e-05,
      "loss": 0.5508649349212646,
      "mean_token_accuracy": 0.857424645870924,
      "num_tokens": 41739707.0,
      "step": 20410
    },
    {
      "entropy": 0.5078095159493387,
      "epoch": 0.33164694704530506,
      "grad_norm": 6.125,
      "learning_rate": 3.8356278104476625e-05,
      "loss": 0.48797106742858887,
      "mean_token_accuracy": 0.8688635788857937,
      "num_tokens": 41761207.0,
      "step": 20420
    },
    {
      "entropy": 0.5987143235281109,
      "epoch": 0.33180935984993054,
      "grad_norm": 7.71875,
      "learning_rate": 3.8345315512392207e-05,
      "loss": 0.614687728881836,
      "mean_token_accuracy": 0.8451770409941674,
      "num_tokens": 41782234.0,
      "step": 20430
    },
    {
      "entropy": 0.4992811969481409,
      "epoch": 0.3319717726545561,
      "grad_norm": 8.0,
      "learning_rate": 3.833434933038793e-05,
      "loss": 0.49220967292785645,
      "mean_token_accuracy": 0.8720801621675491,
      "num_tokens": 41801808.0,
      "step": 20440
    },
    {
      "entropy": 0.5571612041443587,
      "epoch": 0.3321341854591816,
      "grad_norm": 7.1875,
      "learning_rate": 3.832337956141371e-05,
      "loss": 0.5701161861419678,
      "mean_token_accuracy": 0.8564592488110065,
      "num_tokens": 41822603.0,
      "step": 20450
    },
    {
      "entropy": 0.5115193566307425,
      "epoch": 0.33229659826380714,
      "grad_norm": 9.0,
      "learning_rate": 3.8312406208420445e-05,
      "loss": 0.5746944427490235,
      "mean_token_accuracy": 0.862982577085495,
      "num_tokens": 41842810.0,
      "step": 20460
    },
    {
      "entropy": 0.5494257669895888,
      "epoch": 0.3324590110684326,
      "grad_norm": 6.15625,
      "learning_rate": 3.830142927435999e-05,
      "loss": 0.5464491367340087,
      "mean_token_accuracy": 0.8553936511278153,
      "num_tokens": 41862867.0,
      "step": 20470
    },
    {
      "entropy": 0.5794742895290256,
      "epoch": 0.33262142387305815,
      "grad_norm": 8.0625,
      "learning_rate": 3.829044876218517e-05,
      "loss": 0.5548861503601075,
      "mean_token_accuracy": 0.8602238148450851,
      "num_tokens": 41883249.0,
      "step": 20480
    },
    {
      "entropy": 0.5690852659754455,
      "epoch": 0.3327838366776837,
      "grad_norm": 6.5,
      "learning_rate": 3.827946467484978e-05,
      "loss": 0.5930018901824952,
      "mean_token_accuracy": 0.8515468046069146,
      "num_tokens": 41904232.0,
      "step": 20490
    },
    {
      "entropy": 0.5621062159538269,
      "epoch": 0.33294624948230916,
      "grad_norm": 9.5625,
      "learning_rate": 3.8268477015308526e-05,
      "loss": 0.5763182640075684,
      "mean_token_accuracy": 0.8545185789465904,
      "num_tokens": 41925240.0,
      "step": 20500
    },
    {
      "entropy": 0.5822457010392099,
      "epoch": 0.3331086622869347,
      "grad_norm": 8.0625,
      "learning_rate": 3.825748578651714e-05,
      "loss": 0.5645479679107666,
      "mean_token_accuracy": 0.8548313274979591,
      "num_tokens": 41944654.0,
      "step": 20510
    },
    {
      "entropy": 0.580738668749109,
      "epoch": 0.33327107509156023,
      "grad_norm": 8.0,
      "learning_rate": 3.824649099143229e-05,
      "loss": 0.594007921218872,
      "mean_token_accuracy": 0.85033935084939,
      "num_tokens": 41966380.0,
      "step": 20520
    },
    {
      "entropy": 0.5456235494464636,
      "epoch": 0.3334334878961857,
      "grad_norm": 6.1875,
      "learning_rate": 3.8235492633011596e-05,
      "loss": 0.517634105682373,
      "mean_token_accuracy": 0.869189316779375,
      "num_tokens": 41987528.0,
      "step": 20530
    },
    {
      "entropy": 0.546254996350035,
      "epoch": 0.33359590070081124,
      "grad_norm": 7.375,
      "learning_rate": 3.8224490714213626e-05,
      "loss": 0.5575330257415771,
      "mean_token_accuracy": 0.8636027194559575,
      "num_tokens": 42007339.0,
      "step": 20540
    },
    {
      "entropy": 0.47852977802976965,
      "epoch": 0.3337583135054368,
      "grad_norm": 10.0625,
      "learning_rate": 3.8213485237997944e-05,
      "loss": 0.4823010444641113,
      "mean_token_accuracy": 0.8742073811590672,
      "num_tokens": 42028145.0,
      "step": 20550
    },
    {
      "entropy": 0.581709171179682,
      "epoch": 0.3339207263100623,
      "grad_norm": 8.625,
      "learning_rate": 3.8202476207325036e-05,
      "loss": 0.6190417766571045,
      "mean_token_accuracy": 0.8381850473582745,
      "num_tokens": 42047891.0,
      "step": 20560
    },
    {
      "entropy": 0.5712052478455008,
      "epoch": 0.3340831391146878,
      "grad_norm": 7.625,
      "learning_rate": 3.8191463625156354e-05,
      "loss": 0.580459451675415,
      "mean_token_accuracy": 0.85219143666327,
      "num_tokens": 42067879.0,
      "step": 20570
    },
    {
      "entropy": 0.5726041273213923,
      "epoch": 0.3342455519193133,
      "grad_norm": 6.25,
      "learning_rate": 3.818044749445432e-05,
      "loss": 0.5721821308135986,
      "mean_token_accuracy": 0.8524363044649362,
      "num_tokens": 42088497.0,
      "step": 20580
    },
    {
      "entropy": 0.5534457415342331,
      "epoch": 0.33440796472393886,
      "grad_norm": 8.25,
      "learning_rate": 3.816942781818228e-05,
      "loss": 0.5509944438934327,
      "mean_token_accuracy": 0.8607300028204918,
      "num_tokens": 42109350.0,
      "step": 20590
    },
    {
      "entropy": 0.5477527287788689,
      "epoch": 0.33457037752856433,
      "grad_norm": 6.4375,
      "learning_rate": 3.815840459930456e-05,
      "loss": 0.5754600048065186,
      "mean_token_accuracy": 0.8637143053114414,
      "num_tokens": 42130375.0,
      "step": 20600
    },
    {
      "entropy": 0.5319512112531811,
      "epoch": 0.33473279033318987,
      "grad_norm": 6.15625,
      "learning_rate": 3.8147377840786435e-05,
      "loss": 0.49176802635192873,
      "mean_token_accuracy": 0.8773545492440462,
      "num_tokens": 42151380.0,
      "step": 20610
    },
    {
      "entropy": 0.5461848322302103,
      "epoch": 0.3348952031378154,
      "grad_norm": 4.6875,
      "learning_rate": 3.813634754559412e-05,
      "loss": 0.5768444061279296,
      "mean_token_accuracy": 0.8608327001333237,
      "num_tokens": 42172391.0,
      "step": 20620
    },
    {
      "entropy": 0.5714185600169003,
      "epoch": 0.3350576159424409,
      "grad_norm": 9.125,
      "learning_rate": 3.81253137166948e-05,
      "loss": 0.5906174182891846,
      "mean_token_accuracy": 0.8592486258596181,
      "num_tokens": 42192915.0,
      "step": 20630
    },
    {
      "entropy": 0.5490285479463637,
      "epoch": 0.3352200287470664,
      "grad_norm": 10.375,
      "learning_rate": 3.811427635705659e-05,
      "loss": 0.5687652587890625,
      "mean_token_accuracy": 0.8604401141405106,
      "num_tokens": 42213135.0,
      "step": 20640
    },
    {
      "entropy": 0.5097557628527284,
      "epoch": 0.33538244155169195,
      "grad_norm": 14.25,
      "learning_rate": 3.8103235469648565e-05,
      "loss": 0.5130792140960694,
      "mean_token_accuracy": 0.8737553738057613,
      "num_tokens": 42232997.0,
      "step": 20650
    },
    {
      "entropy": 0.5026414898224175,
      "epoch": 0.3355448543563175,
      "grad_norm": 8.5,
      "learning_rate": 3.809219105744075e-05,
      "loss": 0.5358746528625489,
      "mean_token_accuracy": 0.8633633837103843,
      "num_tokens": 42253622.0,
      "step": 20660
    },
    {
      "entropy": 0.5592889266088605,
      "epoch": 0.33570726716094296,
      "grad_norm": 8.0625,
      "learning_rate": 3.808114312340412e-05,
      "loss": 0.5412032127380371,
      "mean_token_accuracy": 0.8701852299273014,
      "num_tokens": 42273935.0,
      "step": 20670
    },
    {
      "entropy": 0.5344620172400028,
      "epoch": 0.3358696799655685,
      "grad_norm": 7.53125,
      "learning_rate": 3.807009167051059e-05,
      "loss": 0.5496950149536133,
      "mean_token_accuracy": 0.8611288610845804,
      "num_tokens": 42293479.0,
      "step": 20680
    },
    {
      "entropy": 0.5664816641714424,
      "epoch": 0.336032092770194,
      "grad_norm": 6.65625,
      "learning_rate": 3.805903670173303e-05,
      "loss": 0.5592621803283692,
      "mean_token_accuracy": 0.8634056583046913,
      "num_tokens": 42313415.0,
      "step": 20690
    },
    {
      "entropy": 0.5659895903430879,
      "epoch": 0.3361945055748195,
      "grad_norm": 7.21875,
      "learning_rate": 3.804797822004525e-05,
      "loss": 0.5717123508453369,
      "mean_token_accuracy": 0.8568331733345985,
      "num_tokens": 42335556.0,
      "step": 20700
    },
    {
      "entropy": 0.5611231119371951,
      "epoch": 0.33635691837944504,
      "grad_norm": 8.375,
      "learning_rate": 3.8036916228422e-05,
      "loss": 0.5660227775573731,
      "mean_token_accuracy": 0.8578096784651279,
      "num_tokens": 42355716.0,
      "step": 20710
    },
    {
      "entropy": 0.5736502340063453,
      "epoch": 0.33651933118407057,
      "grad_norm": 9.1875,
      "learning_rate": 3.8025850729838995e-05,
      "loss": 0.5795518398284912,
      "mean_token_accuracy": 0.8586622409522533,
      "num_tokens": 42376357.0,
      "step": 20720
    },
    {
      "entropy": 0.5585266957990825,
      "epoch": 0.33668174398869605,
      "grad_norm": 6.90625,
      "learning_rate": 3.801478172727286e-05,
      "loss": 0.5670055389404297,
      "mean_token_accuracy": 0.854182193800807,
      "num_tokens": 42396614.0,
      "step": 20730
    },
    {
      "entropy": 0.5456746683456004,
      "epoch": 0.3368441567933216,
      "grad_norm": 7.75,
      "learning_rate": 3.80037092237012e-05,
      "loss": 0.5159268856048584,
      "mean_token_accuracy": 0.8610098365694284,
      "num_tokens": 42417146.0,
      "step": 20740
    },
    {
      "entropy": 0.5509303160011768,
      "epoch": 0.3370065695979471,
      "grad_norm": 8.0,
      "learning_rate": 3.7992633222102525e-05,
      "loss": 0.5614491939544678,
      "mean_token_accuracy": 0.8674316138029099,
      "num_tokens": 42437178.0,
      "step": 20750
    },
    {
      "entropy": 0.5054333887062967,
      "epoch": 0.3371689824025726,
      "grad_norm": 8.5,
      "learning_rate": 3.798155372545632e-05,
      "loss": 0.5196519374847413,
      "mean_token_accuracy": 0.8699027933180332,
      "num_tokens": 42456628.0,
      "step": 20760
    },
    {
      "entropy": 0.5985231728293001,
      "epoch": 0.3373313952071981,
      "grad_norm": 10.5,
      "learning_rate": 3.797047073674299e-05,
      "loss": 0.5708291530609131,
      "mean_token_accuracy": 0.8559908423572778,
      "num_tokens": 42476580.0,
      "step": 20770
    },
    {
      "entropy": 0.5735493028070777,
      "epoch": 0.33749380801182366,
      "grad_norm": 11.8125,
      "learning_rate": 3.7959384258943885e-05,
      "loss": 0.6453959465026855,
      "mean_token_accuracy": 0.8479140192270279,
      "num_tokens": 42495898.0,
      "step": 20780
    },
    {
      "entropy": 0.5782978703733533,
      "epoch": 0.3376562208164492,
      "grad_norm": 6.4375,
      "learning_rate": 3.794829429504129e-05,
      "loss": 0.6096779346466065,
      "mean_token_accuracy": 0.8540595714002848,
      "num_tokens": 42516222.0,
      "step": 20790
    },
    {
      "entropy": 0.5651690520811826,
      "epoch": 0.33781863362107467,
      "grad_norm": 7.125,
      "learning_rate": 3.793720084801842e-05,
      "loss": 0.5304489612579346,
      "mean_token_accuracy": 0.86651796028018,
      "num_tokens": 42536450.0,
      "step": 20800
    },
    {
      "entropy": 0.6439645347185433,
      "epoch": 0.3379810464257002,
      "grad_norm": 7.25,
      "learning_rate": 3.792610392085945e-05,
      "loss": 0.6724597930908203,
      "mean_token_accuracy": 0.8360070049762726,
      "num_tokens": 42557052.0,
      "step": 20810
    },
    {
      "entropy": 0.5568502291571349,
      "epoch": 0.33814345923032574,
      "grad_norm": 6.71875,
      "learning_rate": 3.791500351654948e-05,
      "loss": 0.5513248443603516,
      "mean_token_accuracy": 0.8574566349387169,
      "num_tokens": 42579272.0,
      "step": 20820
    },
    {
      "entropy": 0.5631507073529065,
      "epoch": 0.3383058720349512,
      "grad_norm": 10.0625,
      "learning_rate": 3.790389963807454e-05,
      "loss": 0.5373784065246582,
      "mean_token_accuracy": 0.86322673112154,
      "num_tokens": 42599724.0,
      "step": 20830
    },
    {
      "entropy": 0.5501909313956276,
      "epoch": 0.33846828483957675,
      "grad_norm": 7.9375,
      "learning_rate": 3.78927922884216e-05,
      "loss": 0.5373705387115478,
      "mean_token_accuracy": 0.8663251385092735,
      "num_tokens": 42620249.0,
      "step": 20840
    },
    {
      "entropy": 0.5444812235422433,
      "epoch": 0.3386306976442023,
      "grad_norm": 8.1875,
      "learning_rate": 3.788168147057855e-05,
      "loss": 0.5472448348999024,
      "mean_token_accuracy": 0.8618907168507576,
      "num_tokens": 42642462.0,
      "step": 20850
    },
    {
      "entropy": 0.6104479443281889,
      "epoch": 0.33879311044882776,
      "grad_norm": 8.25,
      "learning_rate": 3.7870567187534246e-05,
      "loss": 0.6048952579498291,
      "mean_token_accuracy": 0.8479999653995037,
      "num_tokens": 42663095.0,
      "step": 20860
    },
    {
      "entropy": 0.5369756100699306,
      "epoch": 0.3389555232534533,
      "grad_norm": 8.5,
      "learning_rate": 3.7859449442278436e-05,
      "loss": 0.5424748420715332,
      "mean_token_accuracy": 0.8623659722507,
      "num_tokens": 42682475.0,
      "step": 20870
    },
    {
      "entropy": 0.5654821131378412,
      "epoch": 0.33911793605807883,
      "grad_norm": 7.8125,
      "learning_rate": 3.784832823780182e-05,
      "loss": 0.5445939540863037,
      "mean_token_accuracy": 0.8635365940630436,
      "num_tokens": 42703284.0,
      "step": 20880
    },
    {
      "entropy": 0.5717876829206944,
      "epoch": 0.33928034886270436,
      "grad_norm": 7.125,
      "learning_rate": 3.783720357709605e-05,
      "loss": 0.6143857955932617,
      "mean_token_accuracy": 0.8466533958911896,
      "num_tokens": 42724096.0,
      "step": 20890
    },
    {
      "entropy": 0.5461292976047843,
      "epoch": 0.33944276166732984,
      "grad_norm": 6.34375,
      "learning_rate": 3.782607546315366e-05,
      "loss": 0.5192225933074951,
      "mean_token_accuracy": 0.8664758950471878,
      "num_tokens": 42744813.0,
      "step": 20900
    },
    {
      "entropy": 0.5342408258002251,
      "epoch": 0.3396051744719554,
      "grad_norm": 7.125,
      "learning_rate": 3.781494389896814e-05,
      "loss": 0.5185364723205567,
      "mean_token_accuracy": 0.8655607536435127,
      "num_tokens": 42766427.0,
      "step": 20910
    },
    {
      "entropy": 0.4996810111217201,
      "epoch": 0.3397675872765809,
      "grad_norm": 7.3125,
      "learning_rate": 3.780380888753392e-05,
      "loss": 0.5234630584716797,
      "mean_token_accuracy": 0.8690463855862618,
      "num_tokens": 42787462.0,
      "step": 20920
    },
    {
      "entropy": 0.5432744802441448,
      "epoch": 0.3399300000812064,
      "grad_norm": 6.34375,
      "learning_rate": 3.779267043184633e-05,
      "loss": 0.5768684387207031,
      "mean_token_accuracy": 0.8520784880965948,
      "num_tokens": 42808282.0,
      "step": 20930
    },
    {
      "entropy": 0.5169015765190125,
      "epoch": 0.3400924128858319,
      "grad_norm": 6.40625,
      "learning_rate": 3.7781528534901646e-05,
      "loss": 0.5006703853607177,
      "mean_token_accuracy": 0.8721001140773297,
      "num_tokens": 42828619.0,
      "step": 20940
    },
    {
      "entropy": 0.5533454094082118,
      "epoch": 0.34025482569045745,
      "grad_norm": 11.75,
      "learning_rate": 3.777038319969706e-05,
      "loss": 0.5154829502105713,
      "mean_token_accuracy": 0.8614557649940252,
      "num_tokens": 42849112.0,
      "step": 20950
    },
    {
      "entropy": 0.4907979262061417,
      "epoch": 0.34041723849508293,
      "grad_norm": 7.53125,
      "learning_rate": 3.7759234429230696e-05,
      "loss": 0.5211287021636963,
      "mean_token_accuracy": 0.8582017160952091,
      "num_tokens": 42868762.0,
      "step": 20960
    },
    {
      "entropy": 0.4907255725003779,
      "epoch": 0.34057965129970846,
      "grad_norm": 7.53125,
      "learning_rate": 3.7748082226501593e-05,
      "loss": 0.49332122802734374,
      "mean_token_accuracy": 0.8721557080745697,
      "num_tokens": 42888764.0,
      "step": 20970
    },
    {
      "entropy": 0.5641098922118545,
      "epoch": 0.340742064104334,
      "grad_norm": 8.4375,
      "learning_rate": 3.773692659450972e-05,
      "loss": 0.5899402141571045,
      "mean_token_accuracy": 0.8526029475033283,
      "num_tokens": 42909490.0,
      "step": 20980
    },
    {
      "entropy": 0.5195209678262472,
      "epoch": 0.3409044769089595,
      "grad_norm": 7.875,
      "learning_rate": 3.7725767536255976e-05,
      "loss": 0.5608676910400391,
      "mean_token_accuracy": 0.8601963382214308,
      "num_tokens": 42930394.0,
      "step": 20990
    },
    {
      "entropy": 0.4971170451492071,
      "epoch": 0.341066889713585,
      "grad_norm": 8.4375,
      "learning_rate": 3.7714605054742154e-05,
      "loss": 0.5011800765991211,
      "mean_token_accuracy": 0.8665483504533767,
      "num_tokens": 42950870.0,
      "step": 21000
    },
    {
      "entropy": 0.6138926239684224,
      "epoch": 0.34122930251821054,
      "grad_norm": 9.0,
      "learning_rate": 3.770343915297101e-05,
      "loss": 0.6509011268615723,
      "mean_token_accuracy": 0.8378890212625265,
      "num_tokens": 42971682.0,
      "step": 21010
    },
    {
      "entropy": 0.4860070194117725,
      "epoch": 0.3413917153228361,
      "grad_norm": 6.03125,
      "learning_rate": 3.769226983394617e-05,
      "loss": 0.4814160346984863,
      "mean_token_accuracy": 0.8700296960771083,
      "num_tokens": 42993433.0,
      "step": 21020
    },
    {
      "entropy": 0.5010112621821463,
      "epoch": 0.34155412812746155,
      "grad_norm": 9.75,
      "learning_rate": 3.768109710067224e-05,
      "loss": 0.47664618492126465,
      "mean_token_accuracy": 0.8822697527706623,
      "num_tokens": 43013261.0,
      "step": 21030
    },
    {
      "entropy": 0.5282897782512009,
      "epoch": 0.3417165409320871,
      "grad_norm": 7.09375,
      "learning_rate": 3.7669920956154675e-05,
      "loss": 0.5546607971191406,
      "mean_token_accuracy": 0.8562342844903469,
      "num_tokens": 43032826.0,
      "step": 21040
    },
    {
      "entropy": 0.5296033099293709,
      "epoch": 0.3418789537367126,
      "grad_norm": 14.3125,
      "learning_rate": 3.7658741403399907e-05,
      "loss": 0.5320944786071777,
      "mean_token_accuracy": 0.861217838525772,
      "num_tokens": 43052675.0,
      "step": 21050
    },
    {
      "entropy": 0.48334149033762513,
      "epoch": 0.3420413665413381,
      "grad_norm": 8.1875,
      "learning_rate": 3.764755844541526e-05,
      "loss": 0.4809300422668457,
      "mean_token_accuracy": 0.8720012940466404,
      "num_tokens": 43072423.0,
      "step": 21060
    },
    {
      "entropy": 0.5509046551305801,
      "epoch": 0.34220377934596363,
      "grad_norm": 8.8125,
      "learning_rate": 3.763637208520894e-05,
      "loss": 0.5834864139556885,
      "mean_token_accuracy": 0.8508333615958691,
      "num_tokens": 43094223.0,
      "step": 21070
    },
    {
      "entropy": 0.5801546820439398,
      "epoch": 0.34236619215058917,
      "grad_norm": 10.4375,
      "learning_rate": 3.762518232579015e-05,
      "loss": 0.5965095043182373,
      "mean_token_accuracy": 0.8504482470452785,
      "num_tokens": 43114577.0,
      "step": 21080
    },
    {
      "entropy": 0.500376574555412,
      "epoch": 0.34252860495521464,
      "grad_norm": 8.4375,
      "learning_rate": 3.7613989170168945e-05,
      "loss": 0.47359442710876465,
      "mean_token_accuracy": 0.8767856165766716,
      "num_tokens": 43135200.0,
      "step": 21090
    },
    {
      "entropy": 0.5051122381351888,
      "epoch": 0.3426910177598402,
      "grad_norm": 8.25,
      "learning_rate": 3.76027926213563e-05,
      "loss": 0.4842665672302246,
      "mean_token_accuracy": 0.8753570772707462,
      "num_tokens": 43155354.0,
      "step": 21100
    },
    {
      "entropy": 0.5315281203947961,
      "epoch": 0.3428534305644657,
      "grad_norm": 6.875,
      "learning_rate": 3.759159268236412e-05,
      "loss": 0.5438448905944824,
      "mean_token_accuracy": 0.8609080240130424,
      "num_tokens": 43174756.0,
      "step": 21110
    },
    {
      "entropy": 0.5070532309822738,
      "epoch": 0.34301584336909124,
      "grad_norm": 6.5,
      "learning_rate": 3.7580389356205205e-05,
      "loss": 0.5295492649078369,
      "mean_token_accuracy": 0.8644339330494404,
      "num_tokens": 43195146.0,
      "step": 21120
    },
    {
      "entropy": 0.5052597049623728,
      "epoch": 0.3431782561737167,
      "grad_norm": 7.09375,
      "learning_rate": 3.756918264589329e-05,
      "loss": 0.5092920303344727,
      "mean_token_accuracy": 0.8674265153706073,
      "num_tokens": 43216168.0,
      "step": 21130
    },
    {
      "entropy": 0.5324864949099719,
      "epoch": 0.34334066897834226,
      "grad_norm": 9.4375,
      "learning_rate": 3.755797255444298e-05,
      "loss": 0.5399191379547119,
      "mean_token_accuracy": 0.8678036272525788,
      "num_tokens": 43235988.0,
      "step": 21140
    },
    {
      "entropy": 0.5012240679003298,
      "epoch": 0.3435030817829678,
      "grad_norm": 7.34375,
      "learning_rate": 3.7546759084869834e-05,
      "loss": 0.5233304500579834,
      "mean_token_accuracy": 0.859471607208252,
      "num_tokens": 43256656.0,
      "step": 21150
    },
    {
      "entropy": 0.574106779973954,
      "epoch": 0.34366549458759327,
      "grad_norm": 6.1875,
      "learning_rate": 3.75355422401903e-05,
      "loss": 0.5541740894317627,
      "mean_token_accuracy": 0.8537972375750542,
      "num_tokens": 43277141.0,
      "step": 21160
    },
    {
      "entropy": 0.5159168996848166,
      "epoch": 0.3438279073922188,
      "grad_norm": 7.46875,
      "learning_rate": 3.7524322023421744e-05,
      "loss": 0.5284635066986084,
      "mean_token_accuracy": 0.8674596942961216,
      "num_tokens": 43298387.0,
      "step": 21170
    },
    {
      "entropy": 0.5503235080279409,
      "epoch": 0.34399032019684433,
      "grad_norm": 12.875,
      "learning_rate": 3.75130984375824e-05,
      "loss": 0.5595190525054932,
      "mean_token_accuracy": 0.8490270659327507,
      "num_tokens": 43319510.0,
      "step": 21180
    },
    {
      "entropy": 0.570077950321138,
      "epoch": 0.3441527330014698,
      "grad_norm": 9.125,
      "learning_rate": 3.7501871485691465e-05,
      "loss": 0.5637058258056641,
      "mean_token_accuracy": 0.8543650805950165,
      "num_tokens": 43339645.0,
      "step": 21190
    },
    {
      "entropy": 0.5571671977639199,
      "epoch": 0.34431514580609535,
      "grad_norm": 9.4375,
      "learning_rate": 3.7490641170769e-05,
      "loss": 0.552903127670288,
      "mean_token_accuracy": 0.8631694979965687,
      "num_tokens": 43360985.0,
      "step": 21200
    },
    {
      "entropy": 0.5670518563594669,
      "epoch": 0.3444775586107209,
      "grad_norm": 7.8125,
      "learning_rate": 3.747940749583599e-05,
      "loss": 0.5864743709564209,
      "mean_token_accuracy": 0.8530047971755266,
      "num_tokens": 43382664.0,
      "step": 21210
    },
    {
      "entropy": 0.5779545023571699,
      "epoch": 0.3446399714153464,
      "grad_norm": 11.0,
      "learning_rate": 3.7468170463914314e-05,
      "loss": 0.6344746589660645,
      "mean_token_accuracy": 0.8477489799261093,
      "num_tokens": 43404657.0,
      "step": 21220
    },
    {
      "entropy": 0.5779235538095235,
      "epoch": 0.3448023842199719,
      "grad_norm": 5.59375,
      "learning_rate": 3.745693007802677e-05,
      "loss": 0.6117434024810791,
      "mean_token_accuracy": 0.8496381796896457,
      "num_tokens": 43425324.0,
      "step": 21230
    },
    {
      "entropy": 0.56191716035828,
      "epoch": 0.3449647970245974,
      "grad_norm": 9.8125,
      "learning_rate": 3.744568634119704e-05,
      "loss": 0.558140230178833,
      "mean_token_accuracy": 0.8561855968087911,
      "num_tokens": 43445987.0,
      "step": 21240
    },
    {
      "entropy": 0.5502116495743394,
      "epoch": 0.34512720982922296,
      "grad_norm": 5.9375,
      "learning_rate": 3.74344392564497e-05,
      "loss": 0.5209548473358154,
      "mean_token_accuracy": 0.8719334699213505,
      "num_tokens": 43465646.0,
      "step": 21250
    },
    {
      "entropy": 0.5691877705976367,
      "epoch": 0.34528962263384844,
      "grad_norm": 6.71875,
      "learning_rate": 3.742318882681027e-05,
      "loss": 0.5437555313110352,
      "mean_token_accuracy": 0.8581645138561725,
      "num_tokens": 43487006.0,
      "step": 21260
    },
    {
      "entropy": 0.5894635946024209,
      "epoch": 0.34545203543847397,
      "grad_norm": 9.125,
      "learning_rate": 3.741193505530513e-05,
      "loss": 0.5693212032318116,
      "mean_token_accuracy": 0.8525977946817875,
      "num_tokens": 43506851.0,
      "step": 21270
    },
    {
      "entropy": 0.4519010942429304,
      "epoch": 0.3456144482430995,
      "grad_norm": 6.78125,
      "learning_rate": 3.7400677944961557e-05,
      "loss": 0.4853959083557129,
      "mean_token_accuracy": 0.8798061981797218,
      "num_tokens": 43527005.0,
      "step": 21280
    },
    {
      "entropy": 0.5040416173171252,
      "epoch": 0.345776861047725,
      "grad_norm": 10.5,
      "learning_rate": 3.7389417498807746e-05,
      "loss": 0.5110989570617676,
      "mean_token_accuracy": 0.8601395882666111,
      "num_tokens": 43547377.0,
      "step": 21290
    },
    {
      "entropy": 0.47961400989443065,
      "epoch": 0.3459392738523505,
      "grad_norm": 9.6875,
      "learning_rate": 3.737815371987279e-05,
      "loss": 0.5103609085083007,
      "mean_token_accuracy": 0.871604447811842,
      "num_tokens": 43568198.0,
      "step": 21300
    },
    {
      "entropy": 0.529077723575756,
      "epoch": 0.34610168665697605,
      "grad_norm": 9.9375,
      "learning_rate": 3.736688661118666e-05,
      "loss": 0.538639497756958,
      "mean_token_accuracy": 0.8583280898630619,
      "num_tokens": 43588617.0,
      "step": 21310
    },
    {
      "entropy": 0.5430535471066833,
      "epoch": 0.3462640994616015,
      "grad_norm": 7.71875,
      "learning_rate": 3.735561617578024e-05,
      "loss": 0.5266341686248779,
      "mean_token_accuracy": 0.8575321979820728,
      "num_tokens": 43609496.0,
      "step": 21320
    },
    {
      "entropy": 0.5389159864746034,
      "epoch": 0.34642651226622706,
      "grad_norm": 7.0,
      "learning_rate": 3.7344342416685304e-05,
      "loss": 0.519633150100708,
      "mean_token_accuracy": 0.8588675782084465,
      "num_tokens": 43630704.0,
      "step": 21330
    },
    {
      "entropy": 0.5120542561169714,
      "epoch": 0.3465889250708526,
      "grad_norm": 6.40625,
      "learning_rate": 3.7333065336934515e-05,
      "loss": 0.5246307373046875,
      "mean_token_accuracy": 0.8666411295533181,
      "num_tokens": 43649862.0,
      "step": 21340
    },
    {
      "entropy": 0.6195092142559588,
      "epoch": 0.3467513378754781,
      "grad_norm": 7.34375,
      "learning_rate": 3.732178493956143e-05,
      "loss": 0.6885006427764893,
      "mean_token_accuracy": 0.8250632867217064,
      "num_tokens": 43669687.0,
      "step": 21350
    },
    {
      "entropy": 0.6005741259083152,
      "epoch": 0.3469137506801036,
      "grad_norm": 8.875,
      "learning_rate": 3.7310501227600515e-05,
      "loss": 0.5819031238555908,
      "mean_token_accuracy": 0.8553003333508968,
      "num_tokens": 43690942.0,
      "step": 21360
    },
    {
      "entropy": 0.5737603027839213,
      "epoch": 0.34707616348472914,
      "grad_norm": 8.3125,
      "learning_rate": 3.729921420408709e-05,
      "loss": 0.5572847366333008,
      "mean_token_accuracy": 0.8666542716324329,
      "num_tokens": 43712678.0,
      "step": 21370
    },
    {
      "entropy": 0.5865101933944971,
      "epoch": 0.3472385762893547,
      "grad_norm": 9.1875,
      "learning_rate": 3.728792387205741e-05,
      "loss": 0.6006796836853028,
      "mean_token_accuracy": 0.851875352114439,
      "num_tokens": 43732947.0,
      "step": 21380
    },
    {
      "entropy": 0.48164226887747646,
      "epoch": 0.34740098909398015,
      "grad_norm": 10.125,
      "learning_rate": 3.727663023454859e-05,
      "loss": 0.48984537124633787,
      "mean_token_accuracy": 0.868570027500391,
      "num_tokens": 43752766.0,
      "step": 21390
    },
    {
      "entropy": 0.48823147965595126,
      "epoch": 0.3475634018986057,
      "grad_norm": 7.78125,
      "learning_rate": 3.7265333294598656e-05,
      "loss": 0.4858410358428955,
      "mean_token_accuracy": 0.8801802046597004,
      "num_tokens": 43771542.0,
      "step": 21400
    },
    {
      "entropy": 0.5525647732429206,
      "epoch": 0.3477258147032312,
      "grad_norm": 8.8125,
      "learning_rate": 3.725403305524649e-05,
      "loss": 0.5387996196746826,
      "mean_token_accuracy": 0.8682037949562073,
      "num_tokens": 43790525.0,
      "step": 21410
    },
    {
      "entropy": 0.47912707990035414,
      "epoch": 0.3478882275078567,
      "grad_norm": 7.84375,
      "learning_rate": 3.72427295195319e-05,
      "loss": 0.4974337100982666,
      "mean_token_accuracy": 0.8760239653289318,
      "num_tokens": 43810732.0,
      "step": 21420
    },
    {
      "entropy": 0.4995779213262722,
      "epoch": 0.34805064031248223,
      "grad_norm": 5.84375,
      "learning_rate": 3.723142269049556e-05,
      "loss": 0.4735243797302246,
      "mean_token_accuracy": 0.8807542406022548,
      "num_tokens": 43831517.0,
      "step": 21430
    },
    {
      "entropy": 0.5473405677825213,
      "epoch": 0.34821305311710776,
      "grad_norm": 7.40625,
      "learning_rate": 3.7220112571179015e-05,
      "loss": 0.5550507545471192,
      "mean_token_accuracy": 0.8586175017058849,
      "num_tokens": 43851490.0,
      "step": 21440
    },
    {
      "entropy": 0.5400142607744783,
      "epoch": 0.3483754659217333,
      "grad_norm": 9.25,
      "learning_rate": 3.720879916462473e-05,
      "loss": 0.578197431564331,
      "mean_token_accuracy": 0.8527533613145352,
      "num_tokens": 43871548.0,
      "step": 21450
    },
    {
      "entropy": 0.5078492473810912,
      "epoch": 0.3485378787263588,
      "grad_norm": 6.21875,
      "learning_rate": 3.719748247387603e-05,
      "loss": 0.5176805973052978,
      "mean_token_accuracy": 0.8761091165244579,
      "num_tokens": 43891660.0,
      "step": 21460
    },
    {
      "entropy": 0.48810845194384456,
      "epoch": 0.3487002915309843,
      "grad_norm": 6.6875,
      "learning_rate": 3.718616250197714e-05,
      "loss": 0.5015100479125977,
      "mean_token_accuracy": 0.8648010976612568,
      "num_tokens": 43912005.0,
      "step": 21470
    },
    {
      "entropy": 0.47190925776958464,
      "epoch": 0.34886270433560984,
      "grad_norm": 7.78125,
      "learning_rate": 3.717483925197315e-05,
      "loss": 0.4817328453063965,
      "mean_token_accuracy": 0.876376535743475,
      "num_tokens": 43931396.0,
      "step": 21480
    },
    {
      "entropy": 0.529636730439961,
      "epoch": 0.3490251171402353,
      "grad_norm": 7.03125,
      "learning_rate": 3.716351272691003e-05,
      "loss": 0.5204188823699951,
      "mean_token_accuracy": 0.8676709145307541,
      "num_tokens": 43950996.0,
      "step": 21490
    },
    {
      "entropy": 0.5262589151039719,
      "epoch": 0.34918752994486085,
      "grad_norm": 7.25,
      "learning_rate": 3.715218292983466e-05,
      "loss": 0.543371057510376,
      "mean_token_accuracy": 0.8545373544096947,
      "num_tokens": 43972637.0,
      "step": 21500
    },
    {
      "entropy": 0.5064210678450763,
      "epoch": 0.3493499427494864,
      "grad_norm": 6.40625,
      "learning_rate": 3.714084986379477e-05,
      "loss": 0.4744619369506836,
      "mean_token_accuracy": 0.8743989668786526,
      "num_tokens": 43992250.0,
      "step": 21510
    },
    {
      "entropy": 0.5680611081887037,
      "epoch": 0.34951235555411186,
      "grad_norm": 9.875,
      "learning_rate": 3.712951353183898e-05,
      "loss": 0.5830663681030274,
      "mean_token_accuracy": 0.863859997689724,
      "num_tokens": 44013218.0,
      "step": 21520
    },
    {
      "entropy": 0.4360484509263188,
      "epoch": 0.3496747683587374,
      "grad_norm": 4.90625,
      "learning_rate": 3.711817393701679e-05,
      "loss": 0.4132223606109619,
      "mean_token_accuracy": 0.885017243027687,
      "num_tokens": 44032874.0,
      "step": 21530
    },
    {
      "entropy": 0.5024894583038986,
      "epoch": 0.34983718116336293,
      "grad_norm": 9.6875,
      "learning_rate": 3.710683108237858e-05,
      "loss": 0.4771139144897461,
      "mean_token_accuracy": 0.8795409981161356,
      "num_tokens": 44052713.0,
      "step": 21540
    },
    {
      "entropy": 0.5225233669392765,
      "epoch": 0.3499995939679884,
      "grad_norm": 7.28125,
      "learning_rate": 3.709548497097559e-05,
      "loss": 0.5286216735839844,
      "mean_token_accuracy": 0.8644353672862053,
      "num_tokens": 44073299.0,
      "step": 21550
    },
    {
      "entropy": 0.55095727564767,
      "epoch": 0.35016200677261394,
      "grad_norm": 7.0,
      "learning_rate": 3.708413560585997e-05,
      "loss": 0.5868868350982666,
      "mean_token_accuracy": 0.851996437460184,
      "num_tokens": 44094358.0,
      "step": 21560
    },
    {
      "entropy": 0.5585955049376935,
      "epoch": 0.3503244195772395,
      "grad_norm": 7.65625,
      "learning_rate": 3.707278299008471e-05,
      "loss": 0.5599225997924805,
      "mean_token_accuracy": 0.8577305674552917,
      "num_tokens": 44115192.0,
      "step": 21570
    },
    {
      "entropy": 0.5963627215474844,
      "epoch": 0.350486832381865,
      "grad_norm": 16.125,
      "learning_rate": 3.70614271267037e-05,
      "loss": 0.6331636428833007,
      "mean_token_accuracy": 0.8444083340466022,
      "num_tokens": 44135218.0,
      "step": 21580
    },
    {
      "entropy": 0.5134365015197545,
      "epoch": 0.3506492451864905,
      "grad_norm": 9.0625,
      "learning_rate": 3.7050068018771685e-05,
      "loss": 0.5529001235961915,
      "mean_token_accuracy": 0.8635319948196412,
      "num_tokens": 44154868.0,
      "step": 21590
    },
    {
      "entropy": 0.5415897570550442,
      "epoch": 0.350811657991116,
      "grad_norm": 7.6875,
      "learning_rate": 3.703870566934429e-05,
      "loss": 0.563785457611084,
      "mean_token_accuracy": 0.855938021093607,
      "num_tokens": 44174958.0,
      "step": 21600
    },
    {
      "entropy": 0.5919224286451936,
      "epoch": 0.35097407079574156,
      "grad_norm": 7.9375,
      "learning_rate": 3.702734008147801e-05,
      "loss": 0.5804090023040771,
      "mean_token_accuracy": 0.8549750946462155,
      "num_tokens": 44197653.0,
      "step": 21610
    },
    {
      "entropy": 0.498003351315856,
      "epoch": 0.35113648360036703,
      "grad_norm": 8.375,
      "learning_rate": 3.701597125823023e-05,
      "loss": 0.4994787693023682,
      "mean_token_accuracy": 0.8688813537359238,
      "num_tokens": 44218706.0,
      "step": 21620
    },
    {
      "entropy": 0.5473702850751578,
      "epoch": 0.35129889640499257,
      "grad_norm": 6.25,
      "learning_rate": 3.700459920265917e-05,
      "loss": 0.5011055946350098,
      "mean_token_accuracy": 0.8620822094380856,
      "num_tokens": 44239599.0,
      "step": 21630
    },
    {
      "entropy": 0.5481806642375886,
      "epoch": 0.3514613092096181,
      "grad_norm": 9.0,
      "learning_rate": 3.699322391782394e-05,
      "loss": 0.606500244140625,
      "mean_token_accuracy": 0.8552890524268151,
      "num_tokens": 44260030.0,
      "step": 21640
    },
    {
      "entropy": 0.5533314242959022,
      "epoch": 0.3516237220142436,
      "grad_norm": 6.25,
      "learning_rate": 3.698184540678453e-05,
      "loss": 0.5632325172424316,
      "mean_token_accuracy": 0.8551447935402393,
      "num_tokens": 44279360.0,
      "step": 21650
    },
    {
      "entropy": 0.5809630160685628,
      "epoch": 0.3517861348188691,
      "grad_norm": 9.375,
      "learning_rate": 3.697046367260178e-05,
      "loss": 0.6267180442810059,
      "mean_token_accuracy": 0.8508105248212814,
      "num_tokens": 44298809.0,
      "step": 21660
    },
    {
      "entropy": 0.606494342116639,
      "epoch": 0.35194854762349465,
      "grad_norm": 8.0625,
      "learning_rate": 3.69590787183374e-05,
      "loss": 0.5740711688995361,
      "mean_token_accuracy": 0.8548745211213827,
      "num_tokens": 44318321.0,
      "step": 21670
    },
    {
      "entropy": 0.48012208715081217,
      "epoch": 0.3521109604281202,
      "grad_norm": 11.0625,
      "learning_rate": 3.694769054705396e-05,
      "loss": 0.46072964668273925,
      "mean_token_accuracy": 0.8849034443497658,
      "num_tokens": 44338352.0,
      "step": 21680
    },
    {
      "entropy": 0.4617147091077641,
      "epoch": 0.35227337323274566,
      "grad_norm": 7.21875,
      "learning_rate": 3.693629916181492e-05,
      "loss": 0.47025418281555176,
      "mean_token_accuracy": 0.8807723551988602,
      "num_tokens": 44357534.0,
      "step": 21690
    },
    {
      "entropy": 0.5836778533644974,
      "epoch": 0.3524357860373712,
      "grad_norm": 7.03125,
      "learning_rate": 3.6924904565684585e-05,
      "loss": 0.6206705570220947,
      "mean_token_accuracy": 0.8555289160460233,
      "num_tokens": 44377290.0,
      "step": 21700
    },
    {
      "entropy": 0.4984380847774446,
      "epoch": 0.3525981988419967,
      "grad_norm": 5.34375,
      "learning_rate": 3.691350676172811e-05,
      "loss": 0.462026309967041,
      "mean_token_accuracy": 0.8787111185491085,
      "num_tokens": 44398962.0,
      "step": 21710
    },
    {
      "entropy": 0.5435928407125175,
      "epoch": 0.3527606116466222,
      "grad_norm": 6.28125,
      "learning_rate": 3.690210575301155e-05,
      "loss": 0.5441035747528076,
      "mean_token_accuracy": 0.8599406030029059,
      "num_tokens": 44419541.0,
      "step": 21720
    },
    {
      "entropy": 0.5423305165953934,
      "epoch": 0.35292302445124774,
      "grad_norm": 4.46875,
      "learning_rate": 3.689070154260179e-05,
      "loss": 0.5704238891601563,
      "mean_token_accuracy": 0.8553355723619461,
      "num_tokens": 44439035.0,
      "step": 21730
    },
    {
      "entropy": 0.552898257644847,
      "epoch": 0.35308543725587327,
      "grad_norm": 6.53125,
      "learning_rate": 3.68792941335666e-05,
      "loss": 0.5355385303497314,
      "mean_token_accuracy": 0.8645597096532583,
      "num_tokens": 44460569.0,
      "step": 21740
    },
    {
      "entropy": 0.5165185953956097,
      "epoch": 0.35324785006049875,
      "grad_norm": 8.0,
      "learning_rate": 3.6867883528974576e-05,
      "loss": 0.5000123500823974,
      "mean_token_accuracy": 0.87314523011446,
      "num_tokens": 44480905.0,
      "step": 21750
    },
    {
      "entropy": 0.49139151107519863,
      "epoch": 0.3534102628651243,
      "grad_norm": 8.0,
      "learning_rate": 3.685646973189523e-05,
      "loss": 0.5180537223815918,
      "mean_token_accuracy": 0.8720071502029896,
      "num_tokens": 44500235.0,
      "step": 21760
    },
    {
      "entropy": 0.55072255525738,
      "epoch": 0.3535726756697498,
      "grad_norm": 7.28125,
      "learning_rate": 3.6845052745398866e-05,
      "loss": 0.5549768924713134,
      "mean_token_accuracy": 0.8527305822819471,
      "num_tokens": 44521332.0,
      "step": 21770
    },
    {
      "entropy": 0.5192476293072105,
      "epoch": 0.3537350884743753,
      "grad_norm": 8.6875,
      "learning_rate": 3.6833632572556705e-05,
      "loss": 0.5043487071990966,
      "mean_token_accuracy": 0.8677250273525715,
      "num_tokens": 44540746.0,
      "step": 21780
    },
    {
      "entropy": 0.47438510404899714,
      "epoch": 0.3538975012790008,
      "grad_norm": 6.8125,
      "learning_rate": 3.682220921644077e-05,
      "loss": 0.4808980941772461,
      "mean_token_accuracy": 0.8842921815812588,
      "num_tokens": 44560244.0,
      "step": 21790
    },
    {
      "entropy": 0.5818553438410163,
      "epoch": 0.35405991408362636,
      "grad_norm": 7.90625,
      "learning_rate": 3.681078268012401e-05,
      "loss": 0.5823946952819824,
      "mean_token_accuracy": 0.8535633683204651,
      "num_tokens": 44580065.0,
      "step": 21800
    },
    {
      "entropy": 0.515645598853007,
      "epoch": 0.3542223268882519,
      "grad_norm": 9.5625,
      "learning_rate": 3.6799352966680147e-05,
      "loss": 0.5042791366577148,
      "mean_token_accuracy": 0.8658063814043999,
      "num_tokens": 44602214.0,
      "step": 21810
    },
    {
      "entropy": 0.5196126556023956,
      "epoch": 0.35438473969287737,
      "grad_norm": 5.90625,
      "learning_rate": 3.678792007918383e-05,
      "loss": 0.4982717990875244,
      "mean_token_accuracy": 0.8693961277604103,
      "num_tokens": 44622227.0,
      "step": 21820
    },
    {
      "entropy": 0.4864508622325957,
      "epoch": 0.3545471524975029,
      "grad_norm": 9.8125,
      "learning_rate": 3.677648402071051e-05,
      "loss": 0.5489206790924073,
      "mean_token_accuracy": 0.8616645157337188,
      "num_tokens": 44641713.0,
      "step": 21830
    },
    {
      "entropy": 0.5858467519748956,
      "epoch": 0.35470956530212844,
      "grad_norm": 8.125,
      "learning_rate": 3.6765044794336525e-05,
      "loss": 0.6453559398651123,
      "mean_token_accuracy": 0.8480042748153209,
      "num_tokens": 44662167.0,
      "step": 21840
    },
    {
      "entropy": 0.5292134121526033,
      "epoch": 0.3548719781067539,
      "grad_norm": 6.875,
      "learning_rate": 3.675360240313906e-05,
      "loss": 0.5062375068664551,
      "mean_token_accuracy": 0.8673783734440803,
      "num_tokens": 44681723.0,
      "step": 21850
    },
    {
      "entropy": 0.6032746971584857,
      "epoch": 0.35503439091137945,
      "grad_norm": 9.25,
      "learning_rate": 3.6742156850196116e-05,
      "loss": 0.5393080711364746,
      "mean_token_accuracy": 0.8594005361199379,
      "num_tokens": 44702561.0,
      "step": 21860
    },
    {
      "entropy": 0.49641904830932615,
      "epoch": 0.355196803716005,
      "grad_norm": 8.3125,
      "learning_rate": 3.673070813858661e-05,
      "loss": 0.4966739654541016,
      "mean_token_accuracy": 0.8760651707649231,
      "num_tokens": 44721952.0,
      "step": 21870
    },
    {
      "entropy": 0.5487524488009512,
      "epoch": 0.35535921652063046,
      "grad_norm": 9.1875,
      "learning_rate": 3.671925627139024e-05,
      "loss": 0.554964542388916,
      "mean_token_accuracy": 0.8561396688222885,
      "num_tokens": 44741788.0,
      "step": 21880
    },
    {
      "entropy": 0.48239508452825247,
      "epoch": 0.355521629325256,
      "grad_norm": 7.90625,
      "learning_rate": 3.6707801251687604e-05,
      "loss": 0.5059203147888184,
      "mean_token_accuracy": 0.8724339105188846,
      "num_tokens": 44761786.0,
      "step": 21890
    },
    {
      "entropy": 0.5832954145036637,
      "epoch": 0.35568404212988153,
      "grad_norm": 9.8125,
      "learning_rate": 3.6696343082560105e-05,
      "loss": 0.6032882690429687,
      "mean_token_accuracy": 0.8437954049557448,
      "num_tokens": 44783681.0,
      "step": 21900
    },
    {
      "entropy": 0.5574038908816874,
      "epoch": 0.35584645493450706,
      "grad_norm": 5.5625,
      "learning_rate": 3.6684881767090045e-05,
      "loss": 0.5643949508666992,
      "mean_token_accuracy": 0.857009730860591,
      "num_tokens": 44805127.0,
      "step": 21910
    },
    {
      "entropy": 0.4810190365416929,
      "epoch": 0.35600886773913254,
      "grad_norm": 10.375,
      "learning_rate": 3.667341730836052e-05,
      "loss": 0.5055337429046631,
      "mean_token_accuracy": 0.8722274832427501,
      "num_tokens": 44824876.0,
      "step": 21920
    },
    {
      "entropy": 0.5386395157314837,
      "epoch": 0.3561712805437581,
      "grad_norm": 10.1875,
      "learning_rate": 3.6661949709455514e-05,
      "loss": 0.5412801265716553,
      "mean_token_accuracy": 0.8553182676434516,
      "num_tokens": 44845063.0,
      "step": 21930
    },
    {
      "entropy": 0.587773175444454,
      "epoch": 0.3563336933483836,
      "grad_norm": 8.125,
      "learning_rate": 3.665047897345982e-05,
      "loss": 0.6137939453125,
      "mean_token_accuracy": 0.8439148873090744,
      "num_tokens": 44865656.0,
      "step": 21940
    },
    {
      "entropy": 0.5175199605058879,
      "epoch": 0.3564961061530091,
      "grad_norm": 5.625,
      "learning_rate": 3.6639005103459104e-05,
      "loss": 0.5095441341400146,
      "mean_token_accuracy": 0.8687964402139187,
      "num_tokens": 44885868.0,
      "step": 21950
    },
    {
      "entropy": 0.5835752455983311,
      "epoch": 0.3566585189576346,
      "grad_norm": 16.875,
      "learning_rate": 3.662752810253987e-05,
      "loss": 0.6248465538024902,
      "mean_token_accuracy": 0.8542846202850342,
      "num_tokens": 44906704.0,
      "step": 21960
    },
    {
      "entropy": 0.5835324994288384,
      "epoch": 0.35682093176226015,
      "grad_norm": 8.4375,
      "learning_rate": 3.661604797378942e-05,
      "loss": 0.5633364200592041,
      "mean_token_accuracy": 0.8538009770214557,
      "num_tokens": 44926676.0,
      "step": 21970
    },
    {
      "entropy": 0.53116120621562,
      "epoch": 0.35698334456688563,
      "grad_norm": 6.25,
      "learning_rate": 3.6604564720295975e-05,
      "loss": 0.5457303047180175,
      "mean_token_accuracy": 0.8618809126317502,
      "num_tokens": 44947249.0,
      "step": 21980
    },
    {
      "entropy": 0.5035348412580788,
      "epoch": 0.35714575737151116,
      "grad_norm": 7.375,
      "learning_rate": 3.659307834514853e-05,
      "loss": 0.5207488536834717,
      "mean_token_accuracy": 0.8654859304428101,
      "num_tokens": 44968072.0,
      "step": 21990
    },
    {
      "entropy": 0.5952181398868561,
      "epoch": 0.3573081701761367,
      "grad_norm": 9.75,
      "learning_rate": 3.6581588851436966e-05,
      "loss": 0.5836207866668701,
      "mean_token_accuracy": 0.8528221480548381,
      "num_tokens": 44988990.0,
      "step": 22000
    },
    {
      "entropy": 0.5420656684786082,
      "epoch": 0.35747058298076223,
      "grad_norm": 6.875,
      "learning_rate": 3.657009624225197e-05,
      "loss": 0.5425662040710449,
      "mean_token_accuracy": 0.8631796941161156,
      "num_tokens": 45009112.0,
      "step": 22010
    },
    {
      "entropy": 0.5318556288722902,
      "epoch": 0.3576329957853877,
      "grad_norm": 7.625,
      "learning_rate": 3.6558600520685084e-05,
      "loss": 0.5316891670227051,
      "mean_token_accuracy": 0.8674138702452183,
      "num_tokens": 45029108.0,
      "step": 22020
    },
    {
      "entropy": 0.4942887583747506,
      "epoch": 0.35779540859001324,
      "grad_norm": 9.75,
      "learning_rate": 3.6547101689828684e-05,
      "loss": 0.48763370513916016,
      "mean_token_accuracy": 0.8772669836878777,
      "num_tokens": 45050137.0,
      "step": 22030
    },
    {
      "entropy": 0.5007436298765242,
      "epoch": 0.3579578213946388,
      "grad_norm": 5.96875,
      "learning_rate": 3.6535599752775975e-05,
      "loss": 0.46999049186706543,
      "mean_token_accuracy": 0.8719119802117348,
      "num_tokens": 45070562.0,
      "step": 22040
    },
    {
      "entropy": 0.5411676256917417,
      "epoch": 0.35812023419926425,
      "grad_norm": 9.0625,
      "learning_rate": 3.6524094712621004e-05,
      "loss": 0.5984457492828369,
      "mean_token_accuracy": 0.8583813183009624,
      "num_tokens": 45090188.0,
      "step": 22050
    },
    {
      "entropy": 0.44545589475892483,
      "epoch": 0.3582826470038898,
      "grad_norm": 7.15625,
      "learning_rate": 3.651258657245866e-05,
      "loss": 0.44718003273010254,
      "mean_token_accuracy": 0.8888007670640945,
      "num_tokens": 45110803.0,
      "step": 22060
    },
    {
      "entropy": 0.6053970804437995,
      "epoch": 0.3584450598085153,
      "grad_norm": 6.625,
      "learning_rate": 3.6501075335384654e-05,
      "loss": 0.6317929744720459,
      "mean_token_accuracy": 0.8477165762335062,
      "num_tokens": 45130906.0,
      "step": 22070
    },
    {
      "entropy": 0.5168034635018557,
      "epoch": 0.3586074726131408,
      "grad_norm": 8.875,
      "learning_rate": 3.648956100449554e-05,
      "loss": 0.560523509979248,
      "mean_token_accuracy": 0.862342418730259,
      "num_tokens": 45151430.0,
      "step": 22080
    },
    {
      "entropy": 0.5511314554139972,
      "epoch": 0.35876988541776633,
      "grad_norm": 9.5,
      "learning_rate": 3.647804358288869e-05,
      "loss": 0.5162510871887207,
      "mean_token_accuracy": 0.8694643691182137,
      "num_tokens": 45172098.0,
      "step": 22090
    },
    {
      "entropy": 0.5397120059467853,
      "epoch": 0.35893229822239187,
      "grad_norm": 14.5,
      "learning_rate": 3.6466523073662325e-05,
      "loss": 0.5300876140594483,
      "mean_token_accuracy": 0.864759573340416,
      "num_tokens": 45191466.0,
      "step": 22100
    },
    {
      "entropy": 0.5278235613368452,
      "epoch": 0.35909471102701734,
      "grad_norm": 9.9375,
      "learning_rate": 3.6454999479915484e-05,
      "loss": 0.553561019897461,
      "mean_token_accuracy": 0.8631356805562973,
      "num_tokens": 45211959.0,
      "step": 22110
    },
    {
      "entropy": 0.5071602772455662,
      "epoch": 0.3592571238316429,
      "grad_norm": 7.65625,
      "learning_rate": 3.644347280474804e-05,
      "loss": 0.47554659843444824,
      "mean_token_accuracy": 0.8854575321078301,
      "num_tokens": 45232866.0,
      "step": 22120
    },
    {
      "entropy": 0.4931931376457214,
      "epoch": 0.3594195366362684,
      "grad_norm": 8.625,
      "learning_rate": 3.643194305126069e-05,
      "loss": 0.49465503692626955,
      "mean_token_accuracy": 0.8659982047975063,
      "num_tokens": 45253874.0,
      "step": 22130
    },
    {
      "entropy": 0.5244854310993106,
      "epoch": 0.35958194944089394,
      "grad_norm": 8.75,
      "learning_rate": 3.642041022255497e-05,
      "loss": 0.5578582286834717,
      "mean_token_accuracy": 0.8598717704415322,
      "num_tokens": 45275143.0,
      "step": 22140
    },
    {
      "entropy": 0.47107594572007655,
      "epoch": 0.3597443622455194,
      "grad_norm": 5.96875,
      "learning_rate": 3.640887432173323e-05,
      "loss": 0.4795197010040283,
      "mean_token_accuracy": 0.8728474907577037,
      "num_tokens": 45295254.0,
      "step": 22150
    },
    {
      "entropy": 0.5526194002944976,
      "epoch": 0.35990677505014496,
      "grad_norm": 8.0625,
      "learning_rate": 3.639733535189866e-05,
      "loss": 0.5327893257141113,
      "mean_token_accuracy": 0.8618099734187126,
      "num_tokens": 45315815.0,
      "step": 22160
    },
    {
      "entropy": 0.5084180941572413,
      "epoch": 0.3600691878547705,
      "grad_norm": 5.6875,
      "learning_rate": 3.638579331615527e-05,
      "loss": 0.5276590347290039,
      "mean_token_accuracy": 0.8709589824080467,
      "num_tokens": 45335335.0,
      "step": 22170
    },
    {
      "entropy": 0.5346804557368159,
      "epoch": 0.36023160065939597,
      "grad_norm": 9.8125,
      "learning_rate": 3.6374248217607886e-05,
      "loss": 0.5672584533691406,
      "mean_token_accuracy": 0.8569342024624348,
      "num_tokens": 45356385.0,
      "step": 22180
    },
    {
      "entropy": 0.5921129929833114,
      "epoch": 0.3603940134640215,
      "grad_norm": 9.0,
      "learning_rate": 3.636270005936217e-05,
      "loss": 0.5767473697662353,
      "mean_token_accuracy": 0.8461877532303334,
      "num_tokens": 45375413.0,
      "step": 22190
    },
    {
      "entropy": 0.6093559931963682,
      "epoch": 0.36055642626864703,
      "grad_norm": 8.1875,
      "learning_rate": 3.6351148844524594e-05,
      "loss": 0.6237957000732421,
      "mean_token_accuracy": 0.8486976057291031,
      "num_tokens": 45396329.0,
      "step": 22200
    },
    {
      "entropy": 0.6094029310159386,
      "epoch": 0.3607188390732725,
      "grad_norm": 6.53125,
      "learning_rate": 3.633959457620249e-05,
      "loss": 0.5926123142242432,
      "mean_token_accuracy": 0.8491690862923861,
      "num_tokens": 45416194.0,
      "step": 22210
    },
    {
      "entropy": 0.5944394648075104,
      "epoch": 0.36088125187789805,
      "grad_norm": 9.4375,
      "learning_rate": 3.632803725750394e-05,
      "loss": 0.6248032569885253,
      "mean_token_accuracy": 0.845386903360486,
      "num_tokens": 45437660.0,
      "step": 22220
    },
    {
      "entropy": 0.5811026097740978,
      "epoch": 0.3610436646825236,
      "grad_norm": 6.65625,
      "learning_rate": 3.631647689153792e-05,
      "loss": 0.6323313236236572,
      "mean_token_accuracy": 0.8482848394662141,
      "num_tokens": 45457885.0,
      "step": 22230
    },
    {
      "entropy": 0.5439787949435413,
      "epoch": 0.3612060774871491,
      "grad_norm": 9.3125,
      "learning_rate": 3.6304913481414185e-05,
      "loss": 0.5603982925415039,
      "mean_token_accuracy": 0.8516043558716774,
      "num_tokens": 45477832.0,
      "step": 22240
    },
    {
      "entropy": 0.5780838655307889,
      "epoch": 0.3613684902917746,
      "grad_norm": 11.125,
      "learning_rate": 3.629334703024331e-05,
      "loss": 0.5688809394836426,
      "mean_token_accuracy": 0.8537171684205532,
      "num_tokens": 45498859.0,
      "step": 22250
    },
    {
      "entropy": 0.5445112710818648,
      "epoch": 0.3615309030964001,
      "grad_norm": 8.5,
      "learning_rate": 3.628177754113671e-05,
      "loss": 0.5171935081481933,
      "mean_token_accuracy": 0.8676237478852272,
      "num_tokens": 45519765.0,
      "step": 22260
    },
    {
      "entropy": 0.5242340377066285,
      "epoch": 0.36169331590102566,
      "grad_norm": 7.0,
      "learning_rate": 3.627020501720659e-05,
      "loss": 0.546094560623169,
      "mean_token_accuracy": 0.8627755973488093,
      "num_tokens": 45540133.0,
      "step": 22270
    },
    {
      "entropy": 0.591062839794904,
      "epoch": 0.36185572870565114,
      "grad_norm": 9.8125,
      "learning_rate": 3.6258629461566005e-05,
      "loss": 0.6062967300415039,
      "mean_token_accuracy": 0.8515837341547012,
      "num_tokens": 45561583.0,
      "step": 22280
    },
    {
      "entropy": 0.5886002252344043,
      "epoch": 0.36201814151027667,
      "grad_norm": 6.3125,
      "learning_rate": 3.624705087732878e-05,
      "loss": 0.6218472480773926,
      "mean_token_accuracy": 0.8407300688326359,
      "num_tokens": 45581659.0,
      "step": 22290
    },
    {
      "entropy": 0.5773891171440482,
      "epoch": 0.3621805543149022,
      "grad_norm": 7.96875,
      "learning_rate": 3.62354692676096e-05,
      "loss": 0.543226671218872,
      "mean_token_accuracy": 0.856671379506588,
      "num_tokens": 45601467.0,
      "step": 22300
    },
    {
      "entropy": 0.5339827533811331,
      "epoch": 0.3623429671195277,
      "grad_norm": 8.0625,
      "learning_rate": 3.622388463552393e-05,
      "loss": 0.5364508152008056,
      "mean_token_accuracy": 0.8621521323919297,
      "num_tokens": 45621953.0,
      "step": 22310
    },
    {
      "entropy": 0.4878361980430782,
      "epoch": 0.3625053799241532,
      "grad_norm": 7.375,
      "learning_rate": 3.6212296984188074e-05,
      "loss": 0.5297654151916504,
      "mean_token_accuracy": 0.8705601416528225,
      "num_tokens": 45641677.0,
      "step": 22320
    },
    {
      "entropy": 0.5138225459493697,
      "epoch": 0.36266779272877875,
      "grad_norm": 6.03125,
      "learning_rate": 3.6200706316719135e-05,
      "loss": 0.5136440753936767,
      "mean_token_accuracy": 0.8723449550569058,
      "num_tokens": 45661073.0,
      "step": 22330
    },
    {
      "entropy": 0.5173530599102378,
      "epoch": 0.3628302055334042,
      "grad_norm": 11.1875,
      "learning_rate": 3.618911263623502e-05,
      "loss": 0.5069920539855957,
      "mean_token_accuracy": 0.8697887666523456,
      "num_tokens": 45682083.0,
      "step": 22340
    },
    {
      "entropy": 0.5886632575653493,
      "epoch": 0.36299261833802976,
      "grad_norm": 12.625,
      "learning_rate": 3.6177515945854457e-05,
      "loss": 0.6017391681671143,
      "mean_token_accuracy": 0.8439410604536534,
      "num_tokens": 45702286.0,
      "step": 22350
    },
    {
      "entropy": 0.543036094494164,
      "epoch": 0.3631550311426553,
      "grad_norm": 8.6875,
      "learning_rate": 3.616591624869699e-05,
      "loss": 0.5345306873321534,
      "mean_token_accuracy": 0.862287325412035,
      "num_tokens": 45721492.0,
      "step": 22360
    },
    {
      "entropy": 0.5217805022373796,
      "epoch": 0.3633174439472808,
      "grad_norm": 7.96875,
      "learning_rate": 3.6154313547882956e-05,
      "loss": 0.4941995143890381,
      "mean_token_accuracy": 0.8761439457535743,
      "num_tokens": 45740241.0,
      "step": 22370
    },
    {
      "entropy": 0.5209983649663628,
      "epoch": 0.3634798567519063,
      "grad_norm": 7.15625,
      "learning_rate": 3.614270784653351e-05,
      "loss": 0.5240054130554199,
      "mean_token_accuracy": 0.8622512258589268,
      "num_tokens": 45760393.0,
      "step": 22380
    },
    {
      "entropy": 0.5026755829341709,
      "epoch": 0.36364226955653184,
      "grad_norm": 10.0625,
      "learning_rate": 3.613109914777061e-05,
      "loss": 0.49358415603637695,
      "mean_token_accuracy": 0.8731410123407841,
      "num_tokens": 45780942.0,
      "step": 22390
    },
    {
      "entropy": 0.5014746790751815,
      "epoch": 0.36380468236115737,
      "grad_norm": 6.375,
      "learning_rate": 3.611948745471703e-05,
      "loss": 0.4699852466583252,
      "mean_token_accuracy": 0.8689093895256519,
      "num_tokens": 45801880.0,
      "step": 22400
    },
    {
      "entropy": 0.5131936501711607,
      "epoch": 0.36396709516578285,
      "grad_norm": 9.5,
      "learning_rate": 3.610787277049633e-05,
      "loss": 0.5164310932159424,
      "mean_token_accuracy": 0.8663072057068348,
      "num_tokens": 45821267.0,
      "step": 22410
    },
    {
      "entropy": 0.4994734394364059,
      "epoch": 0.3641295079704084,
      "grad_norm": 4.5625,
      "learning_rate": 3.609625509823289e-05,
      "loss": 0.5087766170501709,
      "mean_token_accuracy": 0.8696040697395802,
      "num_tokens": 45842073.0,
      "step": 22420
    },
    {
      "entropy": 0.5538337216712534,
      "epoch": 0.3642919207750339,
      "grad_norm": 8.8125,
      "learning_rate": 3.608463444105189e-05,
      "loss": 0.5491084098815918,
      "mean_token_accuracy": 0.8498035863041877,
      "num_tokens": 45862449.0,
      "step": 22430
    },
    {
      "entropy": 0.5388235297985375,
      "epoch": 0.3644543335796594,
      "grad_norm": 10.9375,
      "learning_rate": 3.607301080207932e-05,
      "loss": 0.6304543018341064,
      "mean_token_accuracy": 0.857686672359705,
      "num_tokens": 45883404.0,
      "step": 22440
    },
    {
      "entropy": 0.5203216661699116,
      "epoch": 0.36461674638428493,
      "grad_norm": 5.46875,
      "learning_rate": 3.6061384184441956e-05,
      "loss": 0.5590576648712158,
      "mean_token_accuracy": 0.8596357282251119,
      "num_tokens": 45903966.0,
      "step": 22450
    },
    {
      "entropy": 0.5473650400526822,
      "epoch": 0.36477915918891046,
      "grad_norm": 6.09375,
      "learning_rate": 3.604975459126739e-05,
      "loss": 0.55399169921875,
      "mean_token_accuracy": 0.8622149959206581,
      "num_tokens": 45924088.0,
      "step": 22460
    },
    {
      "entropy": 0.527728881035,
      "epoch": 0.364941571993536,
      "grad_norm": 8.125,
      "learning_rate": 3.6038122025683994e-05,
      "loss": 0.49657497406005857,
      "mean_token_accuracy": 0.8673058494925499,
      "num_tokens": 45944498.0,
      "step": 22470
    },
    {
      "entropy": 0.5301307275891304,
      "epoch": 0.3651039847981615,
      "grad_norm": 8.3125,
      "learning_rate": 3.6026486490820986e-05,
      "loss": 0.5738278388977051,
      "mean_token_accuracy": 0.8554430663585663,
      "num_tokens": 45963570.0,
      "step": 22480
    },
    {
      "entropy": 0.5484862302895636,
      "epoch": 0.365266397602787,
      "grad_norm": 6.875,
      "learning_rate": 3.601484798980832e-05,
      "loss": 0.5608875751495361,
      "mean_token_accuracy": 0.8544336929917336,
      "num_tokens": 45985113.0,
      "step": 22490
    },
    {
      "entropy": 0.549569896387402,
      "epoch": 0.36542881040741254,
      "grad_norm": 6.65625,
      "learning_rate": 3.600320652577679e-05,
      "loss": 0.5329383373260498,
      "mean_token_accuracy": 0.8592706963419914,
      "num_tokens": 46004958.0,
      "step": 22500
    },
    {
      "entropy": 0.5578990668058396,
      "epoch": 0.365591223212038,
      "grad_norm": 5.625,
      "learning_rate": 3.599156210185798e-05,
      "loss": 0.5637103080749511,
      "mean_token_accuracy": 0.8621811509132385,
      "num_tokens": 46024298.0,
      "step": 22510
    },
    {
      "entropy": 0.526423163432628,
      "epoch": 0.36575363601666355,
      "grad_norm": 8.375,
      "learning_rate": 3.597991472118426e-05,
      "loss": 0.5573925971984863,
      "mean_token_accuracy": 0.862083799391985,
      "num_tokens": 46044222.0,
      "step": 22520
    },
    {
      "entropy": 0.5596354202367365,
      "epoch": 0.3659160488212891,
      "grad_norm": 9.9375,
      "learning_rate": 3.596826438688881e-05,
      "loss": 0.5738032817840576,
      "mean_token_accuracy": 0.8538367539644242,
      "num_tokens": 46064472.0,
      "step": 22530
    },
    {
      "entropy": 0.5945897872559727,
      "epoch": 0.36607846162591456,
      "grad_norm": 8.3125,
      "learning_rate": 3.5956611102105584e-05,
      "loss": 0.6222953796386719,
      "mean_token_accuracy": 0.8442585721611977,
      "num_tokens": 46084666.0,
      "step": 22540
    },
    {
      "entropy": 0.6085175042040646,
      "epoch": 0.3662408744305401,
      "grad_norm": 8.1875,
      "learning_rate": 3.594495486996935e-05,
      "loss": 0.5987044334411621,
      "mean_token_accuracy": 0.8501650646328927,
      "num_tokens": 46106845.0,
      "step": 22550
    },
    {
      "entropy": 0.5357793137431145,
      "epoch": 0.36640328723516563,
      "grad_norm": 10.0,
      "learning_rate": 3.593329569361566e-05,
      "loss": 0.5084915637969971,
      "mean_token_accuracy": 0.8592867992818356,
      "num_tokens": 46127098.0,
      "step": 22560
    },
    {
      "entropy": 0.57223765111994,
      "epoch": 0.36656570003979116,
      "grad_norm": 9.1875,
      "learning_rate": 3.5921633576180856e-05,
      "loss": 0.5900027275085449,
      "mean_token_accuracy": 0.8568408265709877,
      "num_tokens": 46147138.0,
      "step": 22570
    },
    {
      "entropy": 0.498483234224841,
      "epoch": 0.36672811284441664,
      "grad_norm": 7.5,
      "learning_rate": 3.5909968520802076e-05,
      "loss": 0.504655408859253,
      "mean_token_accuracy": 0.8765921898186206,
      "num_tokens": 46167731.0,
      "step": 22580
    },
    {
      "entropy": 0.4863421001471579,
      "epoch": 0.3668905256490422,
      "grad_norm": 6.875,
      "learning_rate": 3.589830053061724e-05,
      "loss": 0.48332624435424804,
      "mean_token_accuracy": 0.8732934996485711,
      "num_tokens": 46188521.0,
      "step": 22590
    },
    {
      "entropy": 0.5965361991431564,
      "epoch": 0.3670529384536677,
      "grad_norm": 7.8125,
      "learning_rate": 3.5886629608765075e-05,
      "loss": 0.628386402130127,
      "mean_token_accuracy": 0.839428273960948,
      "num_tokens": 46208227.0,
      "step": 22600
    },
    {
      "entropy": 0.516195021988824,
      "epoch": 0.3672153512582932,
      "grad_norm": 10.25,
      "learning_rate": 3.587495575838508e-05,
      "loss": 0.4876420497894287,
      "mean_token_accuracy": 0.874594870954752,
      "num_tokens": 46228065.0,
      "step": 22610
    },
    {
      "entropy": 0.5001425674185157,
      "epoch": 0.3673777640629187,
      "grad_norm": 7.65625,
      "learning_rate": 3.5863278982617544e-05,
      "loss": 0.4988094806671143,
      "mean_token_accuracy": 0.8746064335107804,
      "num_tokens": 46247436.0,
      "step": 22620
    },
    {
      "entropy": 0.4953230457380414,
      "epoch": 0.36754017686754425,
      "grad_norm": 9.9375,
      "learning_rate": 3.585159928460354e-05,
      "loss": 0.47537813186645506,
      "mean_token_accuracy": 0.8794655568897725,
      "num_tokens": 46267756.0,
      "step": 22630
    },
    {
      "entropy": 0.5102482470683753,
      "epoch": 0.36770258967216973,
      "grad_norm": 6.0625,
      "learning_rate": 3.583991666748496e-05,
      "loss": 0.5246462345123291,
      "mean_token_accuracy": 0.8625853963196277,
      "num_tokens": 46288319.0,
      "step": 22640
    },
    {
      "entropy": 0.5063205063343048,
      "epoch": 0.36786500247679527,
      "grad_norm": 11.75,
      "learning_rate": 3.582823113440442e-05,
      "loss": 0.5079082012176513,
      "mean_token_accuracy": 0.868780754506588,
      "num_tokens": 46307549.0,
      "step": 22650
    },
    {
      "entropy": 0.5103387008653953,
      "epoch": 0.3680274152814208,
      "grad_norm": 7.125,
      "learning_rate": 3.581654268850537e-05,
      "loss": 0.5136359214782715,
      "mean_token_accuracy": 0.8635205499827862,
      "num_tokens": 46327819.0,
      "step": 22660
    },
    {
      "entropy": 0.5618640937842428,
      "epoch": 0.3681898280860463,
      "grad_norm": 12.5,
      "learning_rate": 3.580485133293203e-05,
      "loss": 0.576469898223877,
      "mean_token_accuracy": 0.8440437003970146,
      "num_tokens": 46349805.0,
      "step": 22670
    },
    {
      "entropy": 0.5614996634423732,
      "epoch": 0.3683522408906718,
      "grad_norm": 5.46875,
      "learning_rate": 3.579315707082941e-05,
      "loss": 0.5707251071929932,
      "mean_token_accuracy": 0.8505207046866416,
      "num_tokens": 46372002.0,
      "step": 22680
    },
    {
      "entropy": 0.5534514199011028,
      "epoch": 0.36851465369529735,
      "grad_norm": 7.09375,
      "learning_rate": 3.578145990534327e-05,
      "loss": 0.5328909873962402,
      "mean_token_accuracy": 0.8627839908003807,
      "num_tokens": 46391672.0,
      "step": 22690
    },
    {
      "entropy": 0.5272256378084421,
      "epoch": 0.3686770664999229,
      "grad_norm": 6.71875,
      "learning_rate": 3.576975983962019e-05,
      "loss": 0.5815550327301026,
      "mean_token_accuracy": 0.8539358582347631,
      "num_tokens": 46412585.0,
      "step": 22700
    },
    {
      "entropy": 0.5788405261933803,
      "epoch": 0.36883947930454836,
      "grad_norm": 7.65625,
      "learning_rate": 3.57580568768075e-05,
      "loss": 0.580315637588501,
      "mean_token_accuracy": 0.8456543140113354,
      "num_tokens": 46433911.0,
      "step": 22710
    },
    {
      "entropy": 0.5318998185917735,
      "epoch": 0.3690018921091739,
      "grad_norm": 6.03125,
      "learning_rate": 3.574635102005335e-05,
      "loss": 0.49951949119567873,
      "mean_token_accuracy": 0.8704735808074474,
      "num_tokens": 46453154.0,
      "step": 22720
    },
    {
      "entropy": 0.5594008721411228,
      "epoch": 0.3691643049137994,
      "grad_norm": 6.1875,
      "learning_rate": 3.5734642272506616e-05,
      "loss": 0.5615673542022706,
      "mean_token_accuracy": 0.8592425882816315,
      "num_tokens": 46474077.0,
      "step": 22730
    },
    {
      "entropy": 0.5833348056301475,
      "epoch": 0.3693267177184249,
      "grad_norm": 10.0,
      "learning_rate": 3.572293063731698e-05,
      "loss": 0.627398681640625,
      "mean_token_accuracy": 0.8437128119170666,
      "num_tokens": 46494452.0,
      "step": 22740
    },
    {
      "entropy": 0.5044870178215206,
      "epoch": 0.36948913052305044,
      "grad_norm": 6.4375,
      "learning_rate": 3.571121611763491e-05,
      "loss": 0.49809775352478025,
      "mean_token_accuracy": 0.8761745683848858,
      "num_tokens": 46515451.0,
      "step": 22750
    },
    {
      "entropy": 0.5683391018770635,
      "epoch": 0.36965154332767597,
      "grad_norm": 12.1875,
      "learning_rate": 3.569949871661163e-05,
      "loss": 0.5646000862121582,
      "mean_token_accuracy": 0.8612325817346573,
      "num_tokens": 46535715.0,
      "step": 22760
    },
    {
      "entropy": 0.5393053836422042,
      "epoch": 0.36981395613230145,
      "grad_norm": 8.625,
      "learning_rate": 3.5687778437399154e-05,
      "loss": 0.5669569969177246,
      "mean_token_accuracy": 0.8690717004239559,
      "num_tokens": 46556940.0,
      "step": 22770
    },
    {
      "entropy": 0.4696126381866634,
      "epoch": 0.369976368936927,
      "grad_norm": 8.125,
      "learning_rate": 3.567605528315026e-05,
      "loss": 0.506529426574707,
      "mean_token_accuracy": 0.8779649198055267,
      "num_tokens": 46577909.0,
      "step": 22780
    },
    {
      "entropy": 0.5180344953201711,
      "epoch": 0.3701387817415525,
      "grad_norm": 5.09375,
      "learning_rate": 3.5664329257018496e-05,
      "loss": 0.5046396255493164,
      "mean_token_accuracy": 0.8725584954023361,
      "num_tokens": 46598156.0,
      "step": 22790
    },
    {
      "entropy": 0.5245853875298053,
      "epoch": 0.37030119454617805,
      "grad_norm": 7.09375,
      "learning_rate": 3.56526003621582e-05,
      "loss": 0.5262390613555908,
      "mean_token_accuracy": 0.8684110846370459,
      "num_tokens": 46619036.0,
      "step": 22800
    },
    {
      "entropy": 0.5361057993024587,
      "epoch": 0.3704636073508035,
      "grad_norm": 6.59375,
      "learning_rate": 3.5640868601724466e-05,
      "loss": 0.49814157485961913,
      "mean_token_accuracy": 0.871850086003542,
      "num_tokens": 46639079.0,
      "step": 22810
    },
    {
      "entropy": 0.5254886455833911,
      "epoch": 0.37062602015542906,
      "grad_norm": 10.125,
      "learning_rate": 3.562913397887316e-05,
      "loss": 0.5756486892700196,
      "mean_token_accuracy": 0.8552132278680802,
      "num_tokens": 46659528.0,
      "step": 22820
    },
    {
      "entropy": 0.5132388005033135,
      "epoch": 0.3707884329600546,
      "grad_norm": 7.875,
      "learning_rate": 3.5617396496760935e-05,
      "loss": 0.5069278240203857,
      "mean_token_accuracy": 0.8765708640217781,
      "num_tokens": 46679889.0,
      "step": 22830
    },
    {
      "entropy": 0.5437707738019526,
      "epoch": 0.37095084576468007,
      "grad_norm": 8.75,
      "learning_rate": 3.5605656158545184e-05,
      "loss": 0.5513464927673339,
      "mean_token_accuracy": 0.8712418720126152,
      "num_tokens": 46702352.0,
      "step": 22840
    },
    {
      "entropy": 0.5607295256108046,
      "epoch": 0.3711132585693056,
      "grad_norm": 5.78125,
      "learning_rate": 3.559391296738409e-05,
      "loss": 0.5468845844268799,
      "mean_token_accuracy": 0.8700308963656426,
      "num_tokens": 46722400.0,
      "step": 22850
    },
    {
      "entropy": 0.5075481545645744,
      "epoch": 0.37127567137393114,
      "grad_norm": 12.875,
      "learning_rate": 3.55821669264366e-05,
      "loss": 0.51412353515625,
      "mean_token_accuracy": 0.8679320923984051,
      "num_tokens": 46742501.0,
      "step": 22860
    },
    {
      "entropy": 0.5388759735506028,
      "epoch": 0.3714380841785566,
      "grad_norm": 8.3125,
      "learning_rate": 3.557041803886244e-05,
      "loss": 0.5300612449645996,
      "mean_token_accuracy": 0.8614227347075939,
      "num_tokens": 46762113.0,
      "step": 22870
    },
    {
      "entropy": 0.5458118900191039,
      "epoch": 0.37160049698318215,
      "grad_norm": 8.3125,
      "learning_rate": 3.555866630782205e-05,
      "loss": 0.5642400741577148,
      "mean_token_accuracy": 0.8576182417571545,
      "num_tokens": 46782277.0,
      "step": 22880
    },
    {
      "entropy": 0.526531751290895,
      "epoch": 0.3717629097878077,
      "grad_norm": 6.53125,
      "learning_rate": 3.554691173647671e-05,
      "loss": 0.5234723567962647,
      "mean_token_accuracy": 0.863782000541687,
      "num_tokens": 46802951.0,
      "step": 22890
    },
    {
      "entropy": 0.5256036151666195,
      "epoch": 0.37192532259243316,
      "grad_norm": 8.25,
      "learning_rate": 3.5535154327988395e-05,
      "loss": 0.5279816150665283,
      "mean_token_accuracy": 0.8615976721048355,
      "num_tokens": 46824235.0,
      "step": 22900
    },
    {
      "entropy": 0.5613566390704363,
      "epoch": 0.3720877353970587,
      "grad_norm": 9.75,
      "learning_rate": 3.55233940855199e-05,
      "loss": 0.5567831993103027,
      "mean_token_accuracy": 0.8626518622040749,
      "num_tokens": 46844664.0,
      "step": 22910
    },
    {
      "entropy": 0.5466673552058637,
      "epoch": 0.37225014820168423,
      "grad_norm": 7.03125,
      "learning_rate": 3.551163101223474e-05,
      "loss": 0.5369855403900147,
      "mean_token_accuracy": 0.8599102776497602,
      "num_tokens": 46865476.0,
      "step": 22920
    },
    {
      "entropy": 0.5471957039088011,
      "epoch": 0.37241256100630976,
      "grad_norm": 6.90625,
      "learning_rate": 3.549986511129722e-05,
      "loss": 0.5098142623901367,
      "mean_token_accuracy": 0.8685966536402703,
      "num_tokens": 46886198.0,
      "step": 22930
    },
    {
      "entropy": 0.4645593389868736,
      "epoch": 0.37257497381093524,
      "grad_norm": 9.6875,
      "learning_rate": 3.548809638587239e-05,
      "loss": 0.4508031368255615,
      "mean_token_accuracy": 0.8803025424480438,
      "num_tokens": 46906434.0,
      "step": 22940
    },
    {
      "entropy": 0.5266782679595053,
      "epoch": 0.3727373866155608,
      "grad_norm": 8.25,
      "learning_rate": 3.5476324839126054e-05,
      "loss": 0.5376164436340332,
      "mean_token_accuracy": 0.8612623754888773,
      "num_tokens": 46926922.0,
      "step": 22950
    },
    {
      "entropy": 0.4882442941889167,
      "epoch": 0.3728997994201863,
      "grad_norm": 7.875,
      "learning_rate": 3.54645504742248e-05,
      "loss": 0.5712001323699951,
      "mean_token_accuracy": 0.865396112576127,
      "num_tokens": 46946238.0,
      "step": 22960
    },
    {
      "entropy": 0.496988816652447,
      "epoch": 0.3730622122248118,
      "grad_norm": 7.90625,
      "learning_rate": 3.5452773294335946e-05,
      "loss": 0.5029908657073975,
      "mean_token_accuracy": 0.8775486685335636,
      "num_tokens": 46966658.0,
      "step": 22970
    },
    {
      "entropy": 0.5280406025238336,
      "epoch": 0.3732246250294373,
      "grad_norm": 10.875,
      "learning_rate": 3.544099330262759e-05,
      "loss": 0.49698305130004883,
      "mean_token_accuracy": 0.8702978737652302,
      "num_tokens": 46987267.0,
      "step": 22980
    },
    {
      "entropy": 0.4936371401883662,
      "epoch": 0.37338703783406285,
      "grad_norm": 7.875,
      "learning_rate": 3.542921050226856e-05,
      "loss": 0.5295087814331054,
      "mean_token_accuracy": 0.861780758947134,
      "num_tokens": 47007524.0,
      "step": 22990
    },
    {
      "entropy": 0.47892022065352646,
      "epoch": 0.37354945063868833,
      "grad_norm": 9.5625,
      "learning_rate": 3.541742489642848e-05,
      "loss": 0.5456148624420166,
      "mean_token_accuracy": 0.8599766887724399,
      "num_tokens": 47027393.0,
      "step": 23000
    },
    {
      "entropy": 0.571288030827418,
      "epoch": 0.37371186344331386,
      "grad_norm": 7.96875,
      "learning_rate": 3.540563648827768e-05,
      "loss": 0.5712827205657959,
      "mean_token_accuracy": 0.8631726332008839,
      "num_tokens": 47049011.0,
      "step": 23010
    },
    {
      "entropy": 0.560052125249058,
      "epoch": 0.3738742762479394,
      "grad_norm": 7.96875,
      "learning_rate": 3.5393845280987286e-05,
      "loss": 0.5331133842468262,
      "mean_token_accuracy": 0.8646037772297859,
      "num_tokens": 47070858.0,
      "step": 23020
    },
    {
      "entropy": 0.6352939387783408,
      "epoch": 0.37403668905256493,
      "grad_norm": 7.96875,
      "learning_rate": 3.538205127772915e-05,
      "loss": 0.6375955104827881,
      "mean_token_accuracy": 0.8402499165385962,
      "num_tokens": 47090720.0,
      "step": 23030
    },
    {
      "entropy": 0.4966933771967888,
      "epoch": 0.3741991018571904,
      "grad_norm": 7.0,
      "learning_rate": 3.5370254481675894e-05,
      "loss": 0.4814156532287598,
      "mean_token_accuracy": 0.8715220600366592,
      "num_tokens": 47110745.0,
      "step": 23040
    },
    {
      "entropy": 0.6033111325930804,
      "epoch": 0.37436151466181594,
      "grad_norm": 7.71875,
      "learning_rate": 3.535845489600087e-05,
      "loss": 0.6000465869903564,
      "mean_token_accuracy": 0.851013221591711,
      "num_tokens": 47131882.0,
      "step": 23050
    },
    {
      "entropy": 0.5365425140596927,
      "epoch": 0.3745239274664415,
      "grad_norm": 9.5625,
      "learning_rate": 3.5346652523878196e-05,
      "loss": 0.5196335315704346,
      "mean_token_accuracy": 0.8638619355857372,
      "num_tokens": 47152617.0,
      "step": 23060
    },
    {
      "entropy": 0.6058875522576272,
      "epoch": 0.37468634027106695,
      "grad_norm": 5.21875,
      "learning_rate": 3.5334847368482745e-05,
      "loss": 0.6404666423797607,
      "mean_token_accuracy": 0.8379699695855379,
      "num_tokens": 47173228.0,
      "step": 23070
    },
    {
      "entropy": 0.4880610327934846,
      "epoch": 0.3748487530756925,
      "grad_norm": 8.625,
      "learning_rate": 3.532303943299014e-05,
      "loss": 0.49109439849853515,
      "mean_token_accuracy": 0.8789512477815151,
      "num_tokens": 47193733.0,
      "step": 23080
    },
    {
      "entropy": 0.5497254593297839,
      "epoch": 0.375011165880318,
      "grad_norm": 9.625,
      "learning_rate": 3.5311228720576707e-05,
      "loss": 0.551785659790039,
      "mean_token_accuracy": 0.8620145283639431,
      "num_tokens": 47214564.0,
      "step": 23090
    },
    {
      "entropy": 0.4589710982516408,
      "epoch": 0.3751735786849435,
      "grad_norm": 6.25,
      "learning_rate": 3.529941523441958e-05,
      "loss": 0.45055465698242186,
      "mean_token_accuracy": 0.8832620993256569,
      "num_tokens": 47234326.0,
      "step": 23100
    },
    {
      "entropy": 0.49950263244099913,
      "epoch": 0.37533599148956903,
      "grad_norm": 11.5,
      "learning_rate": 3.5287598977696614e-05,
      "loss": 0.5233136653900147,
      "mean_token_accuracy": 0.863804891705513,
      "num_tokens": 47253919.0,
      "step": 23110
    },
    {
      "entropy": 0.5818585760425776,
      "epoch": 0.37549840429419457,
      "grad_norm": 9.4375,
      "learning_rate": 3.52757799535864e-05,
      "loss": 0.5802332401275635,
      "mean_token_accuracy": 0.8521456062793732,
      "num_tokens": 47275355.0,
      "step": 23120
    },
    {
      "entropy": 0.48734450014308095,
      "epoch": 0.37566081709882004,
      "grad_norm": 8.4375,
      "learning_rate": 3.526395816526828e-05,
      "loss": 0.5121031761169433,
      "mean_token_accuracy": 0.8631136104464531,
      "num_tokens": 47295598.0,
      "step": 23130
    },
    {
      "entropy": 0.5018413669196888,
      "epoch": 0.3758232299034456,
      "grad_norm": 8.3125,
      "learning_rate": 3.525213361592236e-05,
      "loss": 0.5102596282958984,
      "mean_token_accuracy": 0.8629988119006157,
      "num_tokens": 47316093.0,
      "step": 23140
    },
    {
      "entropy": 0.578115433268249,
      "epoch": 0.3759856427080711,
      "grad_norm": 7.78125,
      "learning_rate": 3.524030630872944e-05,
      "loss": 0.6214663028717041,
      "mean_token_accuracy": 0.8461781218647957,
      "num_tokens": 47337298.0,
      "step": 23150
    },
    {
      "entropy": 0.5498658765107394,
      "epoch": 0.37614805551269664,
      "grad_norm": 6.125,
      "learning_rate": 3.522847624687113e-05,
      "loss": 0.5372975826263428,
      "mean_token_accuracy": 0.8585484951734543,
      "num_tokens": 47358181.0,
      "step": 23160
    },
    {
      "entropy": 0.5674035249277949,
      "epoch": 0.3763104683173221,
      "grad_norm": 9.3125,
      "learning_rate": 3.5216643433529706e-05,
      "loss": 0.5584349155426025,
      "mean_token_accuracy": 0.8618085101246834,
      "num_tokens": 47378835.0,
      "step": 23170
    },
    {
      "entropy": 0.5508882936555892,
      "epoch": 0.37647288112194766,
      "grad_norm": 6.9375,
      "learning_rate": 3.5204807871888246e-05,
      "loss": 0.558116340637207,
      "mean_token_accuracy": 0.8662411734461785,
      "num_tokens": 47398174.0,
      "step": 23180
    },
    {
      "entropy": 0.5321155557408929,
      "epoch": 0.3766352939265732,
      "grad_norm": 7.1875,
      "learning_rate": 3.519296956513053e-05,
      "loss": 0.5405116558074952,
      "mean_token_accuracy": 0.863373076915741,
      "num_tokens": 47418930.0,
      "step": 23190
    },
    {
      "entropy": 0.5763557997532189,
      "epoch": 0.37679770673119867,
      "grad_norm": 5.8125,
      "learning_rate": 3.5181128516441095e-05,
      "loss": 0.5841362953186036,
      "mean_token_accuracy": 0.8575020115822554,
      "num_tokens": 47439195.0,
      "step": 23200
    },
    {
      "entropy": 0.5498569095507264,
      "epoch": 0.3769601195358242,
      "grad_norm": 5.71875,
      "learning_rate": 3.516928472900521e-05,
      "loss": 0.5873496532440186,
      "mean_token_accuracy": 0.859566243737936,
      "num_tokens": 47460804.0,
      "step": 23210
    },
    {
      "entropy": 0.5625232204794883,
      "epoch": 0.37712253234044973,
      "grad_norm": 7.25,
      "learning_rate": 3.5157438206008876e-05,
      "loss": 0.55870680809021,
      "mean_token_accuracy": 0.8648183718323708,
      "num_tokens": 47481481.0,
      "step": 23220
    },
    {
      "entropy": 0.5258202247787267,
      "epoch": 0.3772849451450752,
      "grad_norm": 7.28125,
      "learning_rate": 3.5145588950638844e-05,
      "loss": 0.4839372158050537,
      "mean_token_accuracy": 0.8768865942955018,
      "num_tokens": 47501516.0,
      "step": 23230
    },
    {
      "entropy": 0.628549318946898,
      "epoch": 0.37744735794970075,
      "grad_norm": 7.125,
      "learning_rate": 3.513373696608258e-05,
      "loss": 0.644710111618042,
      "mean_token_accuracy": 0.842871118709445,
      "num_tokens": 47522128.0,
      "step": 23240
    },
    {
      "entropy": 0.5085605416446924,
      "epoch": 0.3776097707543263,
      "grad_norm": 9.8125,
      "learning_rate": 3.5121882255528296e-05,
      "loss": 0.5093700408935546,
      "mean_token_accuracy": 0.8702546983957291,
      "num_tokens": 47543002.0,
      "step": 23250
    },
    {
      "entropy": 0.5005460197338835,
      "epoch": 0.3777721835589518,
      "grad_norm": 6.28125,
      "learning_rate": 3.5110024822164934e-05,
      "loss": 0.48158607482910154,
      "mean_token_accuracy": 0.8776849590241909,
      "num_tokens": 47563548.0,
      "step": 23260
    },
    {
      "entropy": 0.5071444848086685,
      "epoch": 0.3779345963635773,
      "grad_norm": 10.875,
      "learning_rate": 3.5098164669182175e-05,
      "loss": 0.5122921466827393,
      "mean_token_accuracy": 0.8713502727448941,
      "num_tokens": 47583230.0,
      "step": 23270
    },
    {
      "entropy": 0.4995664938353002,
      "epoch": 0.3780970091682028,
      "grad_norm": 8.3125,
      "learning_rate": 3.508630179977043e-05,
      "loss": 0.5138347625732422,
      "mean_token_accuracy": 0.8662989065051079,
      "num_tokens": 47604014.0,
      "step": 23280
    },
    {
      "entropy": 0.5845841980539263,
      "epoch": 0.37825942197282836,
      "grad_norm": 5.90625,
      "learning_rate": 3.507443621712083e-05,
      "loss": 0.5698809146881103,
      "mean_token_accuracy": 0.8617636803537607,
      "num_tokens": 47624674.0,
      "step": 23290
    },
    {
      "entropy": 0.5725499807856977,
      "epoch": 0.37842183477745384,
      "grad_norm": 12.8125,
      "learning_rate": 3.506256792442525e-05,
      "loss": 0.5955534458160401,
      "mean_token_accuracy": 0.8502981245517731,
      "num_tokens": 47645742.0,
      "step": 23300
    },
    {
      "entropy": 0.5264549440238625,
      "epoch": 0.37858424758207937,
      "grad_norm": 8.625,
      "learning_rate": 3.505069692487628e-05,
      "loss": 0.5521557807922364,
      "mean_token_accuracy": 0.8687032990157604,
      "num_tokens": 47665677.0,
      "step": 23310
    },
    {
      "entropy": 0.54162556855008,
      "epoch": 0.3787466603867049,
      "grad_norm": 6.0625,
      "learning_rate": 3.503882322166726e-05,
      "loss": 0.5396908283233642,
      "mean_token_accuracy": 0.8575963176786899,
      "num_tokens": 47686692.0,
      "step": 23320
    },
    {
      "entropy": 0.49104399350471795,
      "epoch": 0.3789090731913304,
      "grad_norm": 6.3125,
      "learning_rate": 3.502694681799222e-05,
      "loss": 0.491621732711792,
      "mean_token_accuracy": 0.8782141469419003,
      "num_tokens": 47707130.0,
      "step": 23330
    },
    {
      "entropy": 0.49153111353516576,
      "epoch": 0.3790714859959559,
      "grad_norm": 10.25,
      "learning_rate": 3.5015067717045956e-05,
      "loss": 0.5104156494140625,
      "mean_token_accuracy": 0.8701667435467243,
      "num_tokens": 47726964.0,
      "step": 23340
    },
    {
      "entropy": 0.5356636739801616,
      "epoch": 0.37923389880058145,
      "grad_norm": 6.59375,
      "learning_rate": 3.5003185922023975e-05,
      "loss": 0.5055852890014648,
      "mean_token_accuracy": 0.8678868159651756,
      "num_tokens": 47747247.0,
      "step": 23350
    },
    {
      "entropy": 0.5401990255806595,
      "epoch": 0.379396311605207,
      "grad_norm": 5.53125,
      "learning_rate": 3.499130143612249e-05,
      "loss": 0.5481836318969726,
      "mean_token_accuracy": 0.8628500305116177,
      "num_tokens": 47767945.0,
      "step": 23360
    },
    {
      "entropy": 0.5417534617241472,
      "epoch": 0.37955872440983246,
      "grad_norm": 11.875,
      "learning_rate": 3.497941426253847e-05,
      "loss": 0.5466581821441651,
      "mean_token_accuracy": 0.8594752680510283,
      "num_tokens": 47789218.0,
      "step": 23370
    },
    {
      "entropy": 0.5258222323842346,
      "epoch": 0.379721137214458,
      "grad_norm": 7.5625,
      "learning_rate": 3.496752440446959e-05,
      "loss": 0.5128636360168457,
      "mean_token_accuracy": 0.8603849198669196,
      "num_tokens": 47808817.0,
      "step": 23380
    },
    {
      "entropy": 0.4762317748274654,
      "epoch": 0.3798835500190835,
      "grad_norm": 9.3125,
      "learning_rate": 3.495563186511424e-05,
      "loss": 0.5172730922698975,
      "mean_token_accuracy": 0.8748526282608509,
      "num_tokens": 47829391.0,
      "step": 23390
    },
    {
      "entropy": 0.45121222916059195,
      "epoch": 0.380045962823709,
      "grad_norm": 8.5625,
      "learning_rate": 3.4943736647671547e-05,
      "loss": 0.4369924068450928,
      "mean_token_accuracy": 0.885579739511013,
      "num_tokens": 47849416.0,
      "step": 23400
    },
    {
      "entropy": 0.5380913590546698,
      "epoch": 0.38020837562833454,
      "grad_norm": 5.3125,
      "learning_rate": 3.493183875534134e-05,
      "loss": 0.5485765933990479,
      "mean_token_accuracy": 0.8605882182717324,
      "num_tokens": 47870056.0,
      "step": 23410
    },
    {
      "entropy": 0.5652221563737839,
      "epoch": 0.38037078843296007,
      "grad_norm": 10.75,
      "learning_rate": 3.49199381913242e-05,
      "loss": 0.5633525848388672,
      "mean_token_accuracy": 0.8557614356279373,
      "num_tokens": 47890163.0,
      "step": 23420
    },
    {
      "entropy": 0.5265408675186336,
      "epoch": 0.38053320123758555,
      "grad_norm": 6.34375,
      "learning_rate": 3.490803495882138e-05,
      "loss": 0.5179901123046875,
      "mean_token_accuracy": 0.8720933947712183,
      "num_tokens": 47912021.0,
      "step": 23430
    },
    {
      "entropy": 0.5559795063454658,
      "epoch": 0.3806956140422111,
      "grad_norm": 11.0,
      "learning_rate": 3.4896129061034896e-05,
      "loss": 0.5901319503784179,
      "mean_token_accuracy": 0.8546281069517135,
      "num_tokens": 47932769.0,
      "step": 23440
    },
    {
      "entropy": 0.5480166509747505,
      "epoch": 0.3808580268468366,
      "grad_norm": 6.28125,
      "learning_rate": 3.4884220501167455e-05,
      "loss": 0.5394044399261475,
      "mean_token_accuracy": 0.8672267783433198,
      "num_tokens": 47953128.0,
      "step": 23450
    },
    {
      "entropy": 0.4782569863367826,
      "epoch": 0.3810204396514621,
      "grad_norm": 4.78125,
      "learning_rate": 3.487230928242249e-05,
      "loss": 0.4927670955657959,
      "mean_token_accuracy": 0.8648092590272427,
      "num_tokens": 47974467.0,
      "step": 23460
    },
    {
      "entropy": 0.501753849547822,
      "epoch": 0.38118285245608763,
      "grad_norm": 7.0,
      "learning_rate": 3.486039540800413e-05,
      "loss": 0.5144572734832764,
      "mean_token_accuracy": 0.8649049296975135,
      "num_tokens": 47995616.0,
      "step": 23470
    },
    {
      "entropy": 0.538450820185244,
      "epoch": 0.38134526526071316,
      "grad_norm": 7.65625,
      "learning_rate": 3.484847888111725e-05,
      "loss": 0.5712307453155517,
      "mean_token_accuracy": 0.8561130963265896,
      "num_tokens": 48016981.0,
      "step": 23480
    },
    {
      "entropy": 0.5631506662350148,
      "epoch": 0.3815076780653387,
      "grad_norm": 7.125,
      "learning_rate": 3.483655970496742e-05,
      "loss": 0.56732177734375,
      "mean_token_accuracy": 0.8528319109231234,
      "num_tokens": 48037335.0,
      "step": 23490
    },
    {
      "entropy": 0.5370853966567666,
      "epoch": 0.3816700908699642,
      "grad_norm": 8.5625,
      "learning_rate": 3.4824637882760926e-05,
      "loss": 0.5631498336791992,
      "mean_token_accuracy": 0.8590424001216889,
      "num_tokens": 48058034.0,
      "step": 23500
    },
    {
      "entropy": 0.5949568750336767,
      "epoch": 0.3818325036745897,
      "grad_norm": 6.75,
      "learning_rate": 3.481271341770476e-05,
      "loss": 0.6184279441833496,
      "mean_token_accuracy": 0.847149258852005,
      "num_tokens": 48078783.0,
      "step": 23510
    },
    {
      "entropy": 0.5055835673119873,
      "epoch": 0.38199491647921524,
      "grad_norm": 9.625,
      "learning_rate": 3.480078631300663e-05,
      "loss": 0.5114854335784912,
      "mean_token_accuracy": 0.8693468421697617,
      "num_tokens": 48100358.0,
      "step": 23520
    },
    {
      "entropy": 0.5151665612123907,
      "epoch": 0.3821573292838407,
      "grad_norm": 6.0,
      "learning_rate": 3.478885657187495e-05,
      "loss": 0.5543181896209717,
      "mean_token_accuracy": 0.8669786743819714,
      "num_tokens": 48120677.0,
      "step": 23530
    },
    {
      "entropy": 0.5973755480721593,
      "epoch": 0.38231974208846625,
      "grad_norm": 11.0,
      "learning_rate": 3.477692419751886e-05,
      "loss": 0.6188558578491211,
      "mean_token_accuracy": 0.8456725552678108,
      "num_tokens": 48141091.0,
      "step": 23540
    },
    {
      "entropy": 0.5443610266316682,
      "epoch": 0.3824821548930918,
      "grad_norm": 8.875,
      "learning_rate": 3.47649891931482e-05,
      "loss": 0.5470822334289551,
      "mean_token_accuracy": 0.8671425685286522,
      "num_tokens": 48161245.0,
      "step": 23550
    },
    {
      "entropy": 0.5164826216641814,
      "epoch": 0.38264456769771726,
      "grad_norm": 6.4375,
      "learning_rate": 3.475305156197348e-05,
      "loss": 0.47877206802368166,
      "mean_token_accuracy": 0.8691567055881023,
      "num_tokens": 48180732.0,
      "step": 23560
    },
    {
      "entropy": 0.5802039559464902,
      "epoch": 0.3828069805023428,
      "grad_norm": 8.0,
      "learning_rate": 3.4741111307205984e-05,
      "loss": 0.5927969455718994,
      "mean_token_accuracy": 0.8535598516464233,
      "num_tokens": 48201450.0,
      "step": 23570
    },
    {
      "entropy": 0.5903122657909989,
      "epoch": 0.38296939330696833,
      "grad_norm": 7.3125,
      "learning_rate": 3.4729168432057654e-05,
      "loss": 0.592861270904541,
      "mean_token_accuracy": 0.8502638541162014,
      "num_tokens": 48223020.0,
      "step": 23580
    },
    {
      "entropy": 0.5342993377707899,
      "epoch": 0.38313180611159386,
      "grad_norm": 7.8125,
      "learning_rate": 3.4717222939741144e-05,
      "loss": 0.5393466472625732,
      "mean_token_accuracy": 0.8666709829121828,
      "num_tokens": 48244420.0,
      "step": 23590
    },
    {
      "entropy": 0.5356148012448102,
      "epoch": 0.38329421891621934,
      "grad_norm": 10.0625,
      "learning_rate": 3.4705274833469826e-05,
      "loss": 0.518127155303955,
      "mean_token_accuracy": 0.8713029198348522,
      "num_tokens": 48264562.0,
      "step": 23600
    },
    {
      "entropy": 0.4990497679915279,
      "epoch": 0.3834566317208449,
      "grad_norm": 10.875,
      "learning_rate": 3.469332411645777e-05,
      "loss": 0.5320947647094727,
      "mean_token_accuracy": 0.8701103553175926,
      "num_tokens": 48283628.0,
      "step": 23610
    },
    {
      "entropy": 0.5612592508085072,
      "epoch": 0.3836190445254704,
      "grad_norm": 6.3125,
      "learning_rate": 3.468137079191975e-05,
      "loss": 0.567941427230835,
      "mean_token_accuracy": 0.856011351197958,
      "num_tokens": 48304254.0,
      "step": 23620
    },
    {
      "entropy": 0.5070313761942089,
      "epoch": 0.3837814573300959,
      "grad_norm": 8.6875,
      "learning_rate": 3.466941486307121e-05,
      "loss": 0.49141554832458495,
      "mean_token_accuracy": 0.8748818807303905,
      "num_tokens": 48324460.0,
      "step": 23630
    },
    {
      "entropy": 0.5520128135569393,
      "epoch": 0.3839438701347214,
      "grad_norm": 6.8125,
      "learning_rate": 3.465745633312836e-05,
      "loss": 0.5662581920623779,
      "mean_token_accuracy": 0.848382630944252,
      "num_tokens": 48345877.0,
      "step": 23640
    },
    {
      "entropy": 0.5857702340465039,
      "epoch": 0.38410628293934695,
      "grad_norm": 7.46875,
      "learning_rate": 3.4645495205308046e-05,
      "loss": 0.5871270179748536,
      "mean_token_accuracy": 0.8553826045244932,
      "num_tokens": 48365556.0,
      "step": 23650
    },
    {
      "entropy": 0.5530993167776614,
      "epoch": 0.38426869574397243,
      "grad_norm": 7.125,
      "learning_rate": 3.4633531482827846e-05,
      "loss": 0.5805275917053223,
      "mean_token_accuracy": 0.8547425717115402,
      "num_tokens": 48387098.0,
      "step": 23660
    },
    {
      "entropy": 0.5627901503816247,
      "epoch": 0.38443110854859797,
      "grad_norm": 11.875,
      "learning_rate": 3.4621565168906025e-05,
      "loss": 0.571068525314331,
      "mean_token_accuracy": 0.8591080777347088,
      "num_tokens": 48407647.0,
      "step": 23670
    },
    {
      "entropy": 0.5305544070899486,
      "epoch": 0.3845935213532235,
      "grad_norm": 8.5,
      "learning_rate": 3.460959626676155e-05,
      "loss": 0.47148799896240234,
      "mean_token_accuracy": 0.8746775396168232,
      "num_tokens": 48428915.0,
      "step": 23680
    },
    {
      "entropy": 0.5250191957689821,
      "epoch": 0.384755934157849,
      "grad_norm": 13.25,
      "learning_rate": 3.4597624779614086e-05,
      "loss": 0.5076956748962402,
      "mean_token_accuracy": 0.8702695377171039,
      "num_tokens": 48449356.0,
      "step": 23690
    },
    {
      "entropy": 0.42083700983785094,
      "epoch": 0.3849183469624745,
      "grad_norm": 10.25,
      "learning_rate": 3.458565071068398e-05,
      "loss": 0.45928044319152833,
      "mean_token_accuracy": 0.8880272328853607,
      "num_tokens": 48469611.0,
      "step": 23700
    },
    {
      "entropy": 0.5583666137652472,
      "epoch": 0.38508075976710004,
      "grad_norm": 8.6875,
      "learning_rate": 3.4573674063192296e-05,
      "loss": 0.6178882122039795,
      "mean_token_accuracy": 0.8414328459650278,
      "num_tokens": 48489744.0,
      "step": 23710
    },
    {
      "entropy": 0.5521074629854411,
      "epoch": 0.3852431725717256,
      "grad_norm": 10.0,
      "learning_rate": 3.4561694840360776e-05,
      "loss": 0.5744510650634765,
      "mean_token_accuracy": 0.8531186789274215,
      "num_tokens": 48509799.0,
      "step": 23720
    },
    {
      "entropy": 0.5598664331249893,
      "epoch": 0.38540558537635106,
      "grad_norm": 6.71875,
      "learning_rate": 3.454971304541185e-05,
      "loss": 0.5660371780395508,
      "mean_token_accuracy": 0.8550675347447395,
      "num_tokens": 48530507.0,
      "step": 23730
    },
    {
      "entropy": 0.5949550493620336,
      "epoch": 0.3855679981809766,
      "grad_norm": 7.84375,
      "learning_rate": 3.4537728681568655e-05,
      "loss": 0.5599926948547364,
      "mean_token_accuracy": 0.8594248220324516,
      "num_tokens": 48551102.0,
      "step": 23740
    },
    {
      "entropy": 0.4969536532647908,
      "epoch": 0.3857304109856021,
      "grad_norm": 9.5625,
      "learning_rate": 3.4525741752055013e-05,
      "loss": 0.4961082458496094,
      "mean_token_accuracy": 0.8666239030659199,
      "num_tokens": 48570677.0,
      "step": 23750
    },
    {
      "entropy": 0.5967108998447657,
      "epoch": 0.3858928237902276,
      "grad_norm": 14.0625,
      "learning_rate": 3.451375226009543e-05,
      "loss": 0.6018310070037842,
      "mean_token_accuracy": 0.8561515100300312,
      "num_tokens": 48590468.0,
      "step": 23760
    },
    {
      "entropy": 0.47868533805012703,
      "epoch": 0.38605523659485314,
      "grad_norm": 5.625,
      "learning_rate": 3.4501760208915106e-05,
      "loss": 0.47731552124023435,
      "mean_token_accuracy": 0.871814139187336,
      "num_tokens": 48611502.0,
      "step": 23770
    },
    {
      "entropy": 0.563042457215488,
      "epoch": 0.38621764939947867,
      "grad_norm": 8.6875,
      "learning_rate": 3.448976560173993e-05,
      "loss": 0.5868594169616699,
      "mean_token_accuracy": 0.8435235068202018,
      "num_tokens": 48631909.0,
      "step": 23780
    },
    {
      "entropy": 0.5629034579265862,
      "epoch": 0.38638006220410415,
      "grad_norm": 7.34375,
      "learning_rate": 3.4477768441796484e-05,
      "loss": 0.5925994396209717,
      "mean_token_accuracy": 0.8596901133656502,
      "num_tokens": 48651263.0,
      "step": 23790
    },
    {
      "entropy": 0.5777868398465216,
      "epoch": 0.3865424750087297,
      "grad_norm": 8.6875,
      "learning_rate": 3.446576873231203e-05,
      "loss": 0.569429349899292,
      "mean_token_accuracy": 0.8620200663805008,
      "num_tokens": 48671950.0,
      "step": 23800
    },
    {
      "entropy": 0.4785087413620204,
      "epoch": 0.3867048878133552,
      "grad_norm": 9.125,
      "learning_rate": 3.445376647651452e-05,
      "loss": 0.476682710647583,
      "mean_token_accuracy": 0.880969313532114,
      "num_tokens": 48691954.0,
      "step": 23810
    },
    {
      "entropy": 0.5805737585760653,
      "epoch": 0.38686730061798075,
      "grad_norm": 9.3125,
      "learning_rate": 3.4441761677632564e-05,
      "loss": 0.5855605125427246,
      "mean_token_accuracy": 0.8499524317681789,
      "num_tokens": 48712790.0,
      "step": 23820
    },
    {
      "entropy": 0.5424793984740972,
      "epoch": 0.3870297134226062,
      "grad_norm": 8.875,
      "learning_rate": 3.4429754338895505e-05,
      "loss": 0.5154288291931153,
      "mean_token_accuracy": 0.8689431592822074,
      "num_tokens": 48732732.0,
      "step": 23830
    },
    {
      "entropy": 0.5817006151657551,
      "epoch": 0.38719212622723176,
      "grad_norm": 7.34375,
      "learning_rate": 3.441774446353334e-05,
      "loss": 0.593597412109375,
      "mean_token_accuracy": 0.8492477923631668,
      "num_tokens": 48754507.0,
      "step": 23840
    },
    {
      "entropy": 0.5576871499419213,
      "epoch": 0.3873545390318573,
      "grad_norm": 12.0,
      "learning_rate": 3.440573205477675e-05,
      "loss": 0.5655125617980957,
      "mean_token_accuracy": 0.8645967841148376,
      "num_tokens": 48774367.0,
      "step": 23850
    },
    {
      "entropy": 0.4901885537430644,
      "epoch": 0.38751695183648277,
      "grad_norm": 6.625,
      "learning_rate": 3.4393717115857094e-05,
      "loss": 0.4879577159881592,
      "mean_token_accuracy": 0.8758647382259369,
      "num_tokens": 48794526.0,
      "step": 23860
    },
    {
      "entropy": 0.5840347717981785,
      "epoch": 0.3876793646411083,
      "grad_norm": 10.75,
      "learning_rate": 3.438169965000643e-05,
      "loss": 0.5695525169372558,
      "mean_token_accuracy": 0.8586090043187141,
      "num_tokens": 48815849.0,
      "step": 23870
    },
    {
      "entropy": 0.49606708558276297,
      "epoch": 0.38784177744573384,
      "grad_norm": 7.28125,
      "learning_rate": 3.436967966045747e-05,
      "loss": 0.49727258682250974,
      "mean_token_accuracy": 0.8652135841548443,
      "num_tokens": 48834996.0,
      "step": 23880
    },
    {
      "entropy": 0.5342682395130396,
      "epoch": 0.3880041902503593,
      "grad_norm": 7.0625,
      "learning_rate": 3.435765715044362e-05,
      "loss": 0.5263957500457763,
      "mean_token_accuracy": 0.8688271194696426,
      "num_tokens": 48855301.0,
      "step": 23890
    },
    {
      "entropy": 0.5678919623605907,
      "epoch": 0.38816660305498485,
      "grad_norm": 8.5,
      "learning_rate": 3.434563212319897e-05,
      "loss": 0.5671892166137695,
      "mean_token_accuracy": 0.8622882030904293,
      "num_tokens": 48875417.0,
      "step": 23900
    },
    {
      "entropy": 0.5210700994823128,
      "epoch": 0.3883290158596104,
      "grad_norm": 7.84375,
      "learning_rate": 3.4333604581958286e-05,
      "loss": 0.5360320091247559,
      "mean_token_accuracy": 0.8665295284241438,
      "num_tokens": 48895209.0,
      "step": 23910
    },
    {
      "entropy": 0.5094985549338162,
      "epoch": 0.3884914286642359,
      "grad_norm": 5.40625,
      "learning_rate": 3.432157452995699e-05,
      "loss": 0.48364906311035155,
      "mean_token_accuracy": 0.8690386734902859,
      "num_tokens": 48915045.0,
      "step": 23920
    },
    {
      "entropy": 0.48963782461360095,
      "epoch": 0.3886538414688614,
      "grad_norm": 7.875,
      "learning_rate": 3.430954197043119e-05,
      "loss": 0.5120352745056153,
      "mean_token_accuracy": 0.870234988629818,
      "num_tokens": 48935903.0,
      "step": 23930
    },
    {
      "entropy": 0.562684582080692,
      "epoch": 0.3888162542734869,
      "grad_norm": 8.6875,
      "learning_rate": 3.4297506906617674e-05,
      "loss": 0.5709455013275146,
      "mean_token_accuracy": 0.857391194626689,
      "num_tokens": 48955559.0,
      "step": 23940
    },
    {
      "entropy": 0.47424800116568805,
      "epoch": 0.38897866707811246,
      "grad_norm": 7.5625,
      "learning_rate": 3.428546934175391e-05,
      "loss": 0.4597425937652588,
      "mean_token_accuracy": 0.8867179691791535,
      "num_tokens": 48977027.0,
      "step": 23950
    },
    {
      "entropy": 0.5251229098532348,
      "epoch": 0.38914107988273794,
      "grad_norm": 7.8125,
      "learning_rate": 3.4273429279078026e-05,
      "loss": 0.5439733505249024,
      "mean_token_accuracy": 0.8640069529414177,
      "num_tokens": 48997590.0,
      "step": 23960
    },
    {
      "entropy": 0.5697257700841873,
      "epoch": 0.3893034926873635,
      "grad_norm": 10.0,
      "learning_rate": 3.4261386721828825e-05,
      "loss": 0.5587650299072265,
      "mean_token_accuracy": 0.8628555290400982,
      "num_tokens": 49016902.0,
      "step": 23970
    },
    {
      "entropy": 0.4703827913152054,
      "epoch": 0.389465905491989,
      "grad_norm": 4.96875,
      "learning_rate": 3.424934167324579e-05,
      "loss": 0.464036750793457,
      "mean_token_accuracy": 0.8807435438036919,
      "num_tokens": 49036684.0,
      "step": 23980
    },
    {
      "entropy": 0.5090048952028156,
      "epoch": 0.3896283182966145,
      "grad_norm": 6.4375,
      "learning_rate": 3.423729413656905e-05,
      "loss": 0.5437698364257812,
      "mean_token_accuracy": 0.8648832157254219,
      "num_tokens": 49057995.0,
      "step": 23990
    },
    {
      "epoch": 0.38979073110124,
      "eval_entropy": 0.5637254813313484,
      "eval_loss": 0.5675074458122253,
      "eval_mean_token_accuracy": 0.8551123147010803,
      "eval_num_tokens": 49078774.0,
      "eval_runtime": 40.294,
      "eval_samples_per_second": 49.635,
      "eval_steps_per_second": 6.204,
      "step": 24000
    },
    {
      "entropy": 0.5470815454609692,
      "epoch": 0.38995314390586555,
      "grad_norm": 5.375,
      "learning_rate": 3.4213191611898405e-05,
      "loss": 0.5593103885650634,
      "mean_token_accuracy": 0.8582735929638148,
      "num_tokens": 49099171.0,
      "step": 24010
    },
    {
      "entropy": 0.5273791500367224,
      "epoch": 0.39011555671049103,
      "grad_norm": 5.9375,
      "learning_rate": 3.420113663038814e-05,
      "loss": 0.5481484413146973,
      "mean_token_accuracy": 0.85798534527421,
      "num_tokens": 49119274.0,
      "step": 24020
    },
    {
      "entropy": 0.5714621988590807,
      "epoch": 0.39027796951511656,
      "grad_norm": 8.0625,
      "learning_rate": 3.418907917375144e-05,
      "loss": 0.5809464931488038,
      "mean_token_accuracy": 0.859968900680542,
      "num_tokens": 49139083.0,
      "step": 24030
    },
    {
      "entropy": 0.5914743417873979,
      "epoch": 0.3904403823197421,
      "grad_norm": 8.9375,
      "learning_rate": 3.4177019245231786e-05,
      "loss": 0.5923002243041993,
      "mean_token_accuracy": 0.8478016547858716,
      "num_tokens": 49159312.0,
      "step": 24040
    },
    {
      "entropy": 0.5385115864686668,
      "epoch": 0.39060279512436763,
      "grad_norm": 7.15625,
      "learning_rate": 3.4164956848073336e-05,
      "loss": 0.5466501235961914,
      "mean_token_accuracy": 0.8565024055540562,
      "num_tokens": 49180250.0,
      "step": 24050
    },
    {
      "entropy": 0.5059388472698629,
      "epoch": 0.3907652079289931,
      "grad_norm": 6.75,
      "learning_rate": 3.4152891985520894e-05,
      "loss": 0.5066898345947266,
      "mean_token_accuracy": 0.8683594718575478,
      "num_tokens": 49201019.0,
      "step": 24060
    },
    {
      "entropy": 0.5544144594110548,
      "epoch": 0.39092762073361864,
      "grad_norm": 9.75,
      "learning_rate": 3.4140824660819946e-05,
      "loss": 0.5690405368804932,
      "mean_token_accuracy": 0.8509377285838127,
      "num_tokens": 49220995.0,
      "step": 24070
    },
    {
      "entropy": 0.5950842717662453,
      "epoch": 0.3910900335382442,
      "grad_norm": 5.0625,
      "learning_rate": 3.4128754877216606e-05,
      "loss": 0.5678687572479248,
      "mean_token_accuracy": 0.8584673322737217,
      "num_tokens": 49241743.0,
      "step": 24080
    },
    {
      "entropy": 0.544534741831012,
      "epoch": 0.39125244634286965,
      "grad_norm": 9.25,
      "learning_rate": 3.4116682637957707e-05,
      "loss": 0.5424981117248535,
      "mean_token_accuracy": 0.8638590648770332,
      "num_tokens": 49262114.0,
      "step": 24090
    },
    {
      "entropy": 0.5847654730081558,
      "epoch": 0.3914148591474952,
      "grad_norm": 8.8125,
      "learning_rate": 3.410460794629068e-05,
      "loss": 0.6122738361358643,
      "mean_token_accuracy": 0.8516114003956318,
      "num_tokens": 49283411.0,
      "step": 24100
    },
    {
      "entropy": 0.5177680325694383,
      "epoch": 0.3915772719521207,
      "grad_norm": 9.75,
      "learning_rate": 3.4092530805463655e-05,
      "loss": 0.53689546585083,
      "mean_token_accuracy": 0.8627045162022113,
      "num_tokens": 49303443.0,
      "step": 24110
    },
    {
      "entropy": 0.5199734096881002,
      "epoch": 0.3917396847567462,
      "grad_norm": 7.53125,
      "learning_rate": 3.408045121872541e-05,
      "loss": 0.4856442928314209,
      "mean_token_accuracy": 0.881899195164442,
      "num_tokens": 49323782.0,
      "step": 24120
    },
    {
      "entropy": 0.513450882351026,
      "epoch": 0.39190209756137173,
      "grad_norm": 8.1875,
      "learning_rate": 3.406836918932538e-05,
      "loss": 0.5112613201141357,
      "mean_token_accuracy": 0.8693561427295208,
      "num_tokens": 49343626.0,
      "step": 24130
    },
    {
      "entropy": 0.5693189200013876,
      "epoch": 0.39206451036599727,
      "grad_norm": 8.625,
      "learning_rate": 3.405628472051365e-05,
      "loss": 0.591523265838623,
      "mean_token_accuracy": 0.8597451001405716,
      "num_tokens": 49364281.0,
      "step": 24140
    },
    {
      "entropy": 0.5522109396755696,
      "epoch": 0.3922269231706228,
      "grad_norm": 7.75,
      "learning_rate": 3.404419781554099e-05,
      "loss": 0.5487237930297851,
      "mean_token_accuracy": 0.8625265426933766,
      "num_tokens": 49384550.0,
      "step": 24150
    },
    {
      "entropy": 0.593453058693558,
      "epoch": 0.3923893359752483,
      "grad_norm": 7.3125,
      "learning_rate": 3.403210847765878e-05,
      "loss": 0.6052311897277832,
      "mean_token_accuracy": 0.8471031688153744,
      "num_tokens": 49404580.0,
      "step": 24160
    },
    {
      "entropy": 0.5515667457133532,
      "epoch": 0.3925517487798738,
      "grad_norm": 11.75,
      "learning_rate": 3.4020016710119104e-05,
      "loss": 0.5541321277618408,
      "mean_token_accuracy": 0.8659717224538326,
      "num_tokens": 49425312.0,
      "step": 24170
    },
    {
      "entropy": 0.5102088353596628,
      "epoch": 0.39271416158449934,
      "grad_norm": 11.125,
      "learning_rate": 3.4007922516174665e-05,
      "loss": 0.5203387260437011,
      "mean_token_accuracy": 0.8678213790059089,
      "num_tokens": 49445943.0,
      "step": 24180
    },
    {
      "entropy": 0.521721200272441,
      "epoch": 0.3928765743891248,
      "grad_norm": 14.0625,
      "learning_rate": 3.399582589907881e-05,
      "loss": 0.5335447311401367,
      "mean_token_accuracy": 0.8675961889326572,
      "num_tokens": 49465942.0,
      "step": 24190
    },
    {
      "entropy": 0.5248227306175977,
      "epoch": 0.39303898719375036,
      "grad_norm": 6.8125,
      "learning_rate": 3.398372686208557e-05,
      "loss": 0.4982480049133301,
      "mean_token_accuracy": 0.875780388712883,
      "num_tokens": 49485004.0,
      "step": 24200
    },
    {
      "entropy": 0.5399140848778188,
      "epoch": 0.3932013999983759,
      "grad_norm": 7.96875,
      "learning_rate": 3.397162540844963e-05,
      "loss": 0.5400734424591065,
      "mean_token_accuracy": 0.8679162740707398,
      "num_tokens": 49504741.0,
      "step": 24210
    },
    {
      "entropy": 0.6539133792743087,
      "epoch": 0.39336381280300137,
      "grad_norm": 9.1875,
      "learning_rate": 3.395952154142628e-05,
      "loss": 0.7067049026489258,
      "mean_token_accuracy": 0.8358887210488319,
      "num_tokens": 49525284.0,
      "step": 24220
    },
    {
      "entropy": 0.5917888195253909,
      "epoch": 0.3935262256076269,
      "grad_norm": 10.1875,
      "learning_rate": 3.39474152642715e-05,
      "loss": 0.6000495433807373,
      "mean_token_accuracy": 0.8480644874274731,
      "num_tokens": 49544947.0,
      "step": 24230
    },
    {
      "entropy": 0.4984207819681615,
      "epoch": 0.39368863841225243,
      "grad_norm": 8.625,
      "learning_rate": 3.393530658024191e-05,
      "loss": 0.47518553733825686,
      "mean_token_accuracy": 0.8829759016633034,
      "num_tokens": 49565151.0,
      "step": 24240
    },
    {
      "entropy": 0.5419592611724511,
      "epoch": 0.3938510512168779,
      "grad_norm": 5.3125,
      "learning_rate": 3.3923195492594776e-05,
      "loss": 0.5468581676483154,
      "mean_token_accuracy": 0.8680461443960666,
      "num_tokens": 49586642.0,
      "step": 24250
    },
    {
      "entropy": 0.5260915311053396,
      "epoch": 0.39401346402150345,
      "grad_norm": 5.96875,
      "learning_rate": 3.3911082004588e-05,
      "loss": 0.5469674110412598,
      "mean_token_accuracy": 0.8541105460375548,
      "num_tokens": 49606186.0,
      "step": 24260
    },
    {
      "entropy": 0.47945679258555174,
      "epoch": 0.394175876826129,
      "grad_norm": 8.375,
      "learning_rate": 3.3898966119480125e-05,
      "loss": 0.4717095375061035,
      "mean_token_accuracy": 0.8782904975116252,
      "num_tokens": 49625578.0,
      "step": 24270
    },
    {
      "entropy": 0.4894809193909168,
      "epoch": 0.3943382896307545,
      "grad_norm": 8.375,
      "learning_rate": 3.3886847840530374e-05,
      "loss": 0.5080100536346436,
      "mean_token_accuracy": 0.8695916786789895,
      "num_tokens": 49645466.0,
      "step": 24280
    },
    {
      "entropy": 0.5191498143598438,
      "epoch": 0.39450070243538,
      "grad_norm": 8.6875,
      "learning_rate": 3.3874727170998574e-05,
      "loss": 0.5614278793334961,
      "mean_token_accuracy": 0.8652841538190842,
      "num_tokens": 49666219.0,
      "step": 24290
    },
    {
      "entropy": 0.556421373737976,
      "epoch": 0.3946631152400055,
      "grad_norm": 7.15625,
      "learning_rate": 3.386260411414523e-05,
      "loss": 0.5604465007781982,
      "mean_token_accuracy": 0.8581210598349571,
      "num_tokens": 49685223.0,
      "step": 24300
    },
    {
      "entropy": 0.48089061630889773,
      "epoch": 0.39482552804463106,
      "grad_norm": 7.8125,
      "learning_rate": 3.385047867323145e-05,
      "loss": 0.44666128158569335,
      "mean_token_accuracy": 0.8811507910490036,
      "num_tokens": 49705750.0,
      "step": 24310
    },
    {
      "entropy": 0.5026311038527638,
      "epoch": 0.39498794084925654,
      "grad_norm": 8.0,
      "learning_rate": 3.383835085151904e-05,
      "loss": 0.4860729694366455,
      "mean_token_accuracy": 0.8721625335514546,
      "num_tokens": 49726102.0,
      "step": 24320
    },
    {
      "entropy": 0.6206942440476269,
      "epoch": 0.39515035365388207,
      "grad_norm": 7.59375,
      "learning_rate": 3.3826220652270365e-05,
      "loss": 0.6403229236602783,
      "mean_token_accuracy": 0.8495640516281128,
      "num_tokens": 49746757.0,
      "step": 24330
    },
    {
      "entropy": 0.6080902880057693,
      "epoch": 0.3953127664585076,
      "grad_norm": 8.6875,
      "learning_rate": 3.381408807874852e-05,
      "loss": 0.6647547245025635,
      "mean_token_accuracy": 0.8369383759796619,
      "num_tokens": 49767947.0,
      "step": 24340
    },
    {
      "entropy": 0.562875070096925,
      "epoch": 0.3954751792631331,
      "grad_norm": 10.375,
      "learning_rate": 3.380195313421715e-05,
      "loss": 0.5710540294647217,
      "mean_token_accuracy": 0.8555803343653678,
      "num_tokens": 49788844.0,
      "step": 24350
    },
    {
      "entropy": 0.5438224884681404,
      "epoch": 0.3956375920677586,
      "grad_norm": 7.46875,
      "learning_rate": 3.378981582194062e-05,
      "loss": 0.5352099895477295,
      "mean_token_accuracy": 0.8640500456094742,
      "num_tokens": 49809020.0,
      "step": 24360
    },
    {
      "entropy": 0.5487508226651698,
      "epoch": 0.39580000487238415,
      "grad_norm": 6.53125,
      "learning_rate": 3.377767614518388e-05,
      "loss": 0.5067363739013672,
      "mean_token_accuracy": 0.8645423963665962,
      "num_tokens": 49830165.0,
      "step": 24370
    },
    {
      "entropy": 0.45935055054724216,
      "epoch": 0.3959624176770097,
      "grad_norm": 6.40625,
      "learning_rate": 3.376553410721252e-05,
      "loss": 0.44804673194885253,
      "mean_token_accuracy": 0.8846891708672047,
      "num_tokens": 49849847.0,
      "step": 24380
    },
    {
      "entropy": 0.5758420048980042,
      "epoch": 0.39612483048163516,
      "grad_norm": 6.78125,
      "learning_rate": 3.375338971129279e-05,
      "loss": 0.5575816631317139,
      "mean_token_accuracy": 0.8618034847080708,
      "num_tokens": 49870067.0,
      "step": 24390
    },
    {
      "entropy": 0.45775031694211066,
      "epoch": 0.3962872432862607,
      "grad_norm": 6.5,
      "learning_rate": 3.374124296069156e-05,
      "loss": 0.4685710906982422,
      "mean_token_accuracy": 0.878773408383131,
      "num_tokens": 49890641.0,
      "step": 24400
    },
    {
      "entropy": 0.4571589208673686,
      "epoch": 0.3964496560908862,
      "grad_norm": 6.0625,
      "learning_rate": 3.372909385867633e-05,
      "loss": 0.4606015205383301,
      "mean_token_accuracy": 0.8764919295907021,
      "num_tokens": 49910677.0,
      "step": 24410
    },
    {
      "entropy": 0.5417944345623255,
      "epoch": 0.3966120688955117,
      "grad_norm": 8.625,
      "learning_rate": 3.3716942408515226e-05,
      "loss": 0.5845203399658203,
      "mean_token_accuracy": 0.8530892439186573,
      "num_tokens": 49930442.0,
      "step": 24420
    },
    {
      "entropy": 0.583834565198049,
      "epoch": 0.39677448170013724,
      "grad_norm": 7.125,
      "learning_rate": 3.370478861347703e-05,
      "loss": 0.6038324356079101,
      "mean_token_accuracy": 0.8622652448713779,
      "num_tokens": 49952232.0,
      "step": 24430
    },
    {
      "entropy": 0.5695758547633887,
      "epoch": 0.39693689450476277,
      "grad_norm": 7.8125,
      "learning_rate": 3.369263247683114e-05,
      "loss": 0.605810546875,
      "mean_token_accuracy": 0.8507026106119155,
      "num_tokens": 49973416.0,
      "step": 24440
    },
    {
      "entropy": 0.5338407902047038,
      "epoch": 0.39709930730938825,
      "grad_norm": 8.5,
      "learning_rate": 3.368047400184757e-05,
      "loss": 0.5314123630523682,
      "mean_token_accuracy": 0.8644703276455402,
      "num_tokens": 49993341.0,
      "step": 24450
    },
    {
      "entropy": 0.587674260744825,
      "epoch": 0.3972617201140138,
      "grad_norm": 6.90625,
      "learning_rate": 3.3668313191796984e-05,
      "loss": 0.5830754280090332,
      "mean_token_accuracy": 0.849614730104804,
      "num_tokens": 50012579.0,
      "step": 24460
    },
    {
      "entropy": 0.5092047724872828,
      "epoch": 0.3974241329186393,
      "grad_norm": 8.875,
      "learning_rate": 3.365615004995067e-05,
      "loss": 0.521315050125122,
      "mean_token_accuracy": 0.8619146205484867,
      "num_tokens": 50033066.0,
      "step": 24470
    },
    {
      "entropy": 0.5565886603202671,
      "epoch": 0.3975865457232648,
      "grad_norm": 6.5,
      "learning_rate": 3.364398457958055e-05,
      "loss": 0.5631035804748535,
      "mean_token_accuracy": 0.8540438883006573,
      "num_tokens": 50054324.0,
      "step": 24480
    },
    {
      "entropy": 0.5142906873952597,
      "epoch": 0.39774895852789033,
      "grad_norm": 14.375,
      "learning_rate": 3.3631816783959136e-05,
      "loss": 0.5352102756500244,
      "mean_token_accuracy": 0.8597650971263647,
      "num_tokens": 50073873.0,
      "step": 24490
    },
    {
      "entropy": 0.564775818400085,
      "epoch": 0.39791137133251586,
      "grad_norm": 8.3125,
      "learning_rate": 3.361964666635962e-05,
      "loss": 0.5884331703186035,
      "mean_token_accuracy": 0.8570561379194259,
      "num_tokens": 50093299.0,
      "step": 24500
    },
    {
      "entropy": 0.5846339884679764,
      "epoch": 0.3980737841371414,
      "grad_norm": 6.1875,
      "learning_rate": 3.360747423005577e-05,
      "loss": 0.6248421192169189,
      "mean_token_accuracy": 0.8512643650174141,
      "num_tokens": 50113942.0,
      "step": 24510
    },
    {
      "entropy": 0.5850527752190828,
      "epoch": 0.3982361969417669,
      "grad_norm": 10.8125,
      "learning_rate": 3.3595299478322017e-05,
      "loss": 0.5873960018157959,
      "mean_token_accuracy": 0.8515430822968483,
      "num_tokens": 50134492.0,
      "step": 24520
    },
    {
      "entropy": 0.5614682870917023,
      "epoch": 0.3983986097463924,
      "grad_norm": 8.5,
      "learning_rate": 3.358312241443338e-05,
      "loss": 0.5744276523590088,
      "mean_token_accuracy": 0.8537463754415512,
      "num_tokens": 50155004.0,
      "step": 24530
    },
    {
      "entropy": 0.5558976626023651,
      "epoch": 0.39856102255101794,
      "grad_norm": 8.4375,
      "learning_rate": 3.357094304166553e-05,
      "loss": 0.5068215370178223,
      "mean_token_accuracy": 0.8669219389557838,
      "num_tokens": 50174133.0,
      "step": 24540
    },
    {
      "entropy": 0.6058720836881548,
      "epoch": 0.3987234353556434,
      "grad_norm": 8.6875,
      "learning_rate": 3.355876136329473e-05,
      "loss": 0.6238439559936524,
      "mean_token_accuracy": 0.8591556489467621,
      "num_tokens": 50195298.0,
      "step": 24550
    },
    {
      "entropy": 0.5775792581262067,
      "epoch": 0.39888584816026895,
      "grad_norm": 8.75,
      "learning_rate": 3.3546577382597896e-05,
      "loss": 0.6167443275451661,
      "mean_token_accuracy": 0.8506046891212463,
      "num_tokens": 50216301.0,
      "step": 24560
    },
    {
      "entropy": 0.5926234765443951,
      "epoch": 0.3990482609648945,
      "grad_norm": 6.46875,
      "learning_rate": 3.353439110285254e-05,
      "loss": 0.571668291091919,
      "mean_token_accuracy": 0.8554886281490326,
      "num_tokens": 50236744.0,
      "step": 24570
    },
    {
      "entropy": 0.5382444981485606,
      "epoch": 0.39921067376951996,
      "grad_norm": 6.3125,
      "learning_rate": 3.35222025273368e-05,
      "loss": 0.5097333908081054,
      "mean_token_accuracy": 0.8638208977878093,
      "num_tokens": 50257176.0,
      "step": 24580
    },
    {
      "entropy": 0.5809641570784152,
      "epoch": 0.3993730865741455,
      "grad_norm": 6.84375,
      "learning_rate": 3.351001165932942e-05,
      "loss": 0.5785306453704834,
      "mean_token_accuracy": 0.8542505957186222,
      "num_tokens": 50277880.0,
      "step": 24590
    },
    {
      "entropy": 0.5576749980449677,
      "epoch": 0.39953549937877103,
      "grad_norm": 9.4375,
      "learning_rate": 3.3497818502109787e-05,
      "loss": 0.5372408390045166,
      "mean_token_accuracy": 0.8587405133992434,
      "num_tokens": 50298402.0,
      "step": 24600
    },
    {
      "entropy": 0.5698979392182082,
      "epoch": 0.39969791218339656,
      "grad_norm": 6.71875,
      "learning_rate": 3.348562305895787e-05,
      "loss": 0.5533380508422852,
      "mean_token_accuracy": 0.8619211509823799,
      "num_tokens": 50318902.0,
      "step": 24610
    },
    {
      "entropy": 0.5035459279548377,
      "epoch": 0.39986032498802204,
      "grad_norm": 8.5,
      "learning_rate": 3.3473425333154286e-05,
      "loss": 0.48060007095336915,
      "mean_token_accuracy": 0.8768001101911068,
      "num_tokens": 50339801.0,
      "step": 24620
    },
    {
      "entropy": 0.4768885377794504,
      "epoch": 0.4000227377926476,
      "grad_norm": 7.03125,
      "learning_rate": 3.3461225327980236e-05,
      "loss": 0.4980560302734375,
      "mean_token_accuracy": 0.8777055807411671,
      "num_tokens": 50359692.0,
      "step": 24630
    },
    {
      "entropy": 0.5606810877099633,
      "epoch": 0.4001851505972731,
      "grad_norm": 8.625,
      "learning_rate": 3.3449023046717574e-05,
      "loss": 0.6263270854949952,
      "mean_token_accuracy": 0.8446399018168449,
      "num_tokens": 50380089.0,
      "step": 24640
    },
    {
      "entropy": 0.567219059728086,
      "epoch": 0.4003475634018986,
      "grad_norm": 7.6875,
      "learning_rate": 3.343681849264871e-05,
      "loss": 0.5682858943939209,
      "mean_token_accuracy": 0.8557911373674869,
      "num_tokens": 50401013.0,
      "step": 24650
    },
    {
      "entropy": 0.5201390456175432,
      "epoch": 0.4005099762065241,
      "grad_norm": 6.90625,
      "learning_rate": 3.3424611669056724e-05,
      "loss": 0.5405611515045166,
      "mean_token_accuracy": 0.865534345805645,
      "num_tokens": 50421210.0,
      "step": 24660
    },
    {
      "entropy": 0.5210041671991348,
      "epoch": 0.40067238901114965,
      "grad_norm": 7.78125,
      "learning_rate": 3.341240257922526e-05,
      "loss": 0.5044244289398193,
      "mean_token_accuracy": 0.8687691271305085,
      "num_tokens": 50441565.0,
      "step": 24670
    },
    {
      "entropy": 0.5202719015069306,
      "epoch": 0.40083480181577513,
      "grad_norm": 5.40625,
      "learning_rate": 3.340019122643859e-05,
      "loss": 0.48703689575195314,
      "mean_token_accuracy": 0.8775925043970346,
      "num_tokens": 50461385.0,
      "step": 24680
    },
    {
      "entropy": 0.424806459993124,
      "epoch": 0.40099721462040067,
      "grad_norm": 6.125,
      "learning_rate": 3.33879776139816e-05,
      "loss": 0.4296908378601074,
      "mean_token_accuracy": 0.8890633910894394,
      "num_tokens": 50481043.0,
      "step": 24690
    },
    {
      "entropy": 0.5221292734146118,
      "epoch": 0.4011596274250262,
      "grad_norm": 7.78125,
      "learning_rate": 3.3375761745139796e-05,
      "loss": 0.5456648349761963,
      "mean_token_accuracy": 0.8620258621871472,
      "num_tokens": 50501461.0,
      "step": 24700
    },
    {
      "entropy": 0.5230364803224802,
      "epoch": 0.40132204022965173,
      "grad_norm": 9.0,
      "learning_rate": 3.3363543623199244e-05,
      "loss": 0.5583702087402344,
      "mean_token_accuracy": 0.8629333063960075,
      "num_tokens": 50524402.0,
      "step": 24710
    },
    {
      "entropy": 0.5414495103526861,
      "epoch": 0.4014844530342772,
      "grad_norm": 6.28125,
      "learning_rate": 3.335132325144667e-05,
      "loss": 0.5327662944793701,
      "mean_token_accuracy": 0.8654169566929341,
      "num_tokens": 50545037.0,
      "step": 24720
    },
    {
      "entropy": 0.5475381553173065,
      "epoch": 0.40164686583890274,
      "grad_norm": 8.9375,
      "learning_rate": 3.333910063316937e-05,
      "loss": 0.4954291820526123,
      "mean_token_accuracy": 0.8769735060632229,
      "num_tokens": 50564929.0,
      "step": 24730
    },
    {
      "entropy": 0.5223632203880697,
      "epoch": 0.4018092786435283,
      "grad_norm": 12.375,
      "learning_rate": 3.332687577165526e-05,
      "loss": 0.5556440830230713,
      "mean_token_accuracy": 0.862972405552864,
      "num_tokens": 50585969.0,
      "step": 24740
    },
    {
      "entropy": 0.536286451946944,
      "epoch": 0.40197169144815376,
      "grad_norm": 10.375,
      "learning_rate": 3.331464867019286e-05,
      "loss": 0.5522490501403808,
      "mean_token_accuracy": 0.8656185768544674,
      "num_tokens": 50606386.0,
      "step": 24750
    },
    {
      "entropy": 0.4679999236017466,
      "epoch": 0.4021341042527793,
      "grad_norm": 7.0625,
      "learning_rate": 3.3302419332071264e-05,
      "loss": 0.5227734565734863,
      "mean_token_accuracy": 0.8635454043745995,
      "num_tokens": 50625576.0,
      "step": 24760
    },
    {
      "entropy": 0.588271027058363,
      "epoch": 0.4022965170574048,
      "grad_norm": 8.875,
      "learning_rate": 3.3290187760580224e-05,
      "loss": 0.6271987915039062,
      "mean_token_accuracy": 0.8493224211037159,
      "num_tokens": 50645384.0,
      "step": 24770
    },
    {
      "entropy": 0.5425276891328394,
      "epoch": 0.4024589298620303,
      "grad_norm": 7.0625,
      "learning_rate": 3.327795395901004e-05,
      "loss": 0.509632158279419,
      "mean_token_accuracy": 0.8600460760295391,
      "num_tokens": 50667221.0,
      "step": 24780
    },
    {
      "entropy": 0.5516342369839549,
      "epoch": 0.40262134266665583,
      "grad_norm": 7.0,
      "learning_rate": 3.326571793065164e-05,
      "loss": 0.5292848110198974,
      "mean_token_accuracy": 0.8611587196588516,
      "num_tokens": 50688163.0,
      "step": 24790
    },
    {
      "entropy": 0.5037799165584147,
      "epoch": 0.40278375547128137,
      "grad_norm": 7.59375,
      "learning_rate": 3.325347967879654e-05,
      "loss": 0.5035677433013916,
      "mean_token_accuracy": 0.8791987538337708,
      "num_tokens": 50709200.0,
      "step": 24800
    },
    {
      "entropy": 0.5215102039743215,
      "epoch": 0.40294616827590685,
      "grad_norm": 7.0,
      "learning_rate": 3.324123920673685e-05,
      "loss": 0.549717903137207,
      "mean_token_accuracy": 0.8651593759655952,
      "num_tokens": 50729448.0,
      "step": 24810
    },
    {
      "entropy": 0.5296195364557207,
      "epoch": 0.4031085810805324,
      "grad_norm": 7.53125,
      "learning_rate": 3.322899651776531e-05,
      "loss": 0.5221252918243409,
      "mean_token_accuracy": 0.8729255452752114,
      "num_tokens": 50750465.0,
      "step": 24820
    },
    {
      "entropy": 0.5017266682349145,
      "epoch": 0.4032709938851579,
      "grad_norm": 10.375,
      "learning_rate": 3.3216751615175204e-05,
      "loss": 0.5189737796783447,
      "mean_token_accuracy": 0.8803450286388397,
      "num_tokens": 50770167.0,
      "step": 24830
    },
    {
      "entropy": 0.5359395843930542,
      "epoch": 0.40343340668978345,
      "grad_norm": 9.25,
      "learning_rate": 3.320450450226044e-05,
      "loss": 0.534287977218628,
      "mean_token_accuracy": 0.8644127279520035,
      "num_tokens": 50791229.0,
      "step": 24840
    },
    {
      "entropy": 0.5798644173890353,
      "epoch": 0.4035958194944089,
      "grad_norm": 12.875,
      "learning_rate": 3.3192255182315535e-05,
      "loss": 0.596154260635376,
      "mean_token_accuracy": 0.8500757347792387,
      "num_tokens": 50813451.0,
      "step": 24850
    },
    {
      "entropy": 0.5313674295321107,
      "epoch": 0.40375823229903446,
      "grad_norm": 8.0,
      "learning_rate": 3.318000365863558e-05,
      "loss": 0.5799715518951416,
      "mean_token_accuracy": 0.8544572860002517,
      "num_tokens": 50832349.0,
      "step": 24860
    },
    {
      "entropy": 0.5053093465976417,
      "epoch": 0.40392064510366,
      "grad_norm": 7.25,
      "learning_rate": 3.316774993451624e-05,
      "loss": 0.5047046661376953,
      "mean_token_accuracy": 0.8757122941315174,
      "num_tokens": 50852593.0,
      "step": 24870
    },
    {
      "entropy": 0.480148171633482,
      "epoch": 0.40408305790828547,
      "grad_norm": 8.1875,
      "learning_rate": 3.315549401325383e-05,
      "loss": 0.4358085632324219,
      "mean_token_accuracy": 0.8822451248764992,
      "num_tokens": 50872351.0,
      "step": 24880
    },
    {
      "entropy": 0.5351269998587668,
      "epoch": 0.404245470712911,
      "grad_norm": 8.875,
      "learning_rate": 3.314323589814519e-05,
      "loss": 0.5243041515350342,
      "mean_token_accuracy": 0.87459683008492,
      "num_tokens": 50892056.0,
      "step": 24890
    },
    {
      "entropy": 0.5375764241674915,
      "epoch": 0.40440788351753654,
      "grad_norm": 10.5,
      "learning_rate": 3.313097559248779e-05,
      "loss": 0.5837990760803222,
      "mean_token_accuracy": 0.862605344504118,
      "num_tokens": 50911746.0,
      "step": 24900
    },
    {
      "entropy": 0.5436143177561462,
      "epoch": 0.404570296322162,
      "grad_norm": 11.375,
      "learning_rate": 3.3118713099579695e-05,
      "loss": 0.5762386798858643,
      "mean_token_accuracy": 0.8500058472156524,
      "num_tokens": 50931745.0,
      "step": 24910
    },
    {
      "entropy": 0.5300541246077046,
      "epoch": 0.40473270912678755,
      "grad_norm": 6.3125,
      "learning_rate": 3.310644842271953e-05,
      "loss": 0.5195698261260986,
      "mean_token_accuracy": 0.8673166073858738,
      "num_tokens": 50952171.0,
      "step": 24920
    },
    {
      "entropy": 0.5580717776436359,
      "epoch": 0.4048951219314131,
      "grad_norm": 12.5,
      "learning_rate": 3.309418156520652e-05,
      "loss": 0.5701028823852539,
      "mean_token_accuracy": 0.8580047309398651,
      "num_tokens": 50972923.0,
      "step": 24930
    },
    {
      "entropy": 0.4800017481669784,
      "epoch": 0.4050575347360386,
      "grad_norm": 8.9375,
      "learning_rate": 3.308191253034047e-05,
      "loss": 0.4500971794128418,
      "mean_token_accuracy": 0.8837778083980083,
      "num_tokens": 50993112.0,
      "step": 24940
    },
    {
      "entropy": 0.4880117943510413,
      "epoch": 0.4052199475406641,
      "grad_norm": 7.3125,
      "learning_rate": 3.306964132142179e-05,
      "loss": 0.4946317195892334,
      "mean_token_accuracy": 0.8702634982764721,
      "num_tokens": 51012658.0,
      "step": 24950
    },
    {
      "entropy": 0.5226860797964037,
      "epoch": 0.4053823603452896,
      "grad_norm": 8.4375,
      "learning_rate": 3.305736794175146e-05,
      "loss": 0.5503151893615723,
      "mean_token_accuracy": 0.8650079503655433,
      "num_tokens": 51033008.0,
      "step": 24960
    },
    {
      "entropy": 0.5558324968907982,
      "epoch": 0.40554477314991516,
      "grad_norm": 6.65625,
      "learning_rate": 3.304509239463104e-05,
      "loss": 0.566602087020874,
      "mean_token_accuracy": 0.8595930002629757,
      "num_tokens": 51054426.0,
      "step": 24970
    },
    {
      "entropy": 0.5375002248678357,
      "epoch": 0.40570718595454064,
      "grad_norm": 12.875,
      "learning_rate": 3.30328146833627e-05,
      "loss": 0.5405380725860596,
      "mean_token_accuracy": 0.8717418611049652,
      "num_tokens": 51074224.0,
      "step": 24980
    },
    {
      "entropy": 0.5348454692400992,
      "epoch": 0.4058695987591662,
      "grad_norm": 5.59375,
      "learning_rate": 3.302053481124914e-05,
      "loss": 0.5977858543395996,
      "mean_token_accuracy": 0.855983504280448,
      "num_tokens": 51096757.0,
      "step": 24990
    },
    {
      "entropy": 0.5838438208214939,
      "epoch": 0.4060320115637917,
      "grad_norm": 14.875,
      "learning_rate": 3.3008252781593705e-05,
      "loss": 0.6205543518066406,
      "mean_token_accuracy": 0.8500596769154072,
      "num_tokens": 51119258.0,
      "step": 25000
    },
    {
      "entropy": 0.5193705357611179,
      "epoch": 0.4061944243684172,
      "grad_norm": 9.3125,
      "learning_rate": 3.299596859770027e-05,
      "loss": 0.5671008586883545,
      "mean_token_accuracy": 0.8613362364470959,
      "num_tokens": 51138247.0,
      "step": 25010
    },
    {
      "entropy": 0.5452842829748988,
      "epoch": 0.4063568371730427,
      "grad_norm": 8.6875,
      "learning_rate": 3.298368226287332e-05,
      "loss": 0.5821037292480469,
      "mean_token_accuracy": 0.8549463383853435,
      "num_tokens": 51157375.0,
      "step": 25020
    },
    {
      "entropy": 0.5559474711306394,
      "epoch": 0.40651924997766825,
      "grad_norm": 8.625,
      "learning_rate": 3.297139378041788e-05,
      "loss": 0.5703703403472901,
      "mean_token_accuracy": 0.8571316428482533,
      "num_tokens": 51177682.0,
      "step": 25030
    },
    {
      "entropy": 0.5868739212397486,
      "epoch": 0.40668166278229373,
      "grad_norm": 8.375,
      "learning_rate": 3.2959103153639616e-05,
      "loss": 0.5792486667633057,
      "mean_token_accuracy": 0.8579266123473644,
      "num_tokens": 51198079.0,
      "step": 25040
    },
    {
      "entropy": 0.4383959240280092,
      "epoch": 0.40684407558691926,
      "grad_norm": 6.78125,
      "learning_rate": 3.2946810385844705e-05,
      "loss": 0.44569907188415525,
      "mean_token_accuracy": 0.8822499930858612,
      "num_tokens": 51218068.0,
      "step": 25050
    },
    {
      "entropy": 0.5247417388949543,
      "epoch": 0.4070064883915448,
      "grad_norm": 10.3125,
      "learning_rate": 3.2934515480339945e-05,
      "loss": 0.4918965339660645,
      "mean_token_accuracy": 0.8747635371983051,
      "num_tokens": 51237900.0,
      "step": 25060
    },
    {
      "entropy": 0.4831520746462047,
      "epoch": 0.40716890119617033,
      "grad_norm": 11.6875,
      "learning_rate": 3.292221844043269e-05,
      "loss": 0.5054845809936523,
      "mean_token_accuracy": 0.8761356875300408,
      "num_tokens": 51257170.0,
      "step": 25070
    },
    {
      "entropy": 0.5313878068700433,
      "epoch": 0.4073313140007958,
      "grad_norm": 6.6875,
      "learning_rate": 3.290991926943087e-05,
      "loss": 0.508045244216919,
      "mean_token_accuracy": 0.871259118616581,
      "num_tokens": 51276177.0,
      "step": 25080
    },
    {
      "entropy": 0.5120634741149843,
      "epoch": 0.40749372680542134,
      "grad_norm": 6.0,
      "learning_rate": 3.2897617970642994e-05,
      "loss": 0.5466601371765136,
      "mean_token_accuracy": 0.8665314018726349,
      "num_tokens": 51296054.0,
      "step": 25090
    },
    {
      "entropy": 0.6233436285983771,
      "epoch": 0.4076561396100469,
      "grad_norm": 7.90625,
      "learning_rate": 3.288531454737812e-05,
      "loss": 0.6286232471466064,
      "mean_token_accuracy": 0.8417238719761372,
      "num_tokens": 51316457.0,
      "step": 25100
    },
    {
      "entropy": 0.5086704493500293,
      "epoch": 0.40781855241467235,
      "grad_norm": 7.34375,
      "learning_rate": 3.287300900294593e-05,
      "loss": 0.48212876319885256,
      "mean_token_accuracy": 0.8741145327687263,
      "num_tokens": 51337529.0,
      "step": 25110
    },
    {
      "entropy": 0.5538184839067981,
      "epoch": 0.4079809652192979,
      "grad_norm": 7.78125,
      "learning_rate": 3.286070134065662e-05,
      "loss": 0.569442892074585,
      "mean_token_accuracy": 0.8591975264251233,
      "num_tokens": 51357111.0,
      "step": 25120
    },
    {
      "entropy": 0.5504454242996871,
      "epoch": 0.4081433780239234,
      "grad_norm": 6.65625,
      "learning_rate": 3.284839156382098e-05,
      "loss": 0.5442337989807129,
      "mean_token_accuracy": 0.8666186206042766,
      "num_tokens": 51378062.0,
      "step": 25130
    },
    {
      "entropy": 0.5188783780671656,
      "epoch": 0.4083057908285489,
      "grad_norm": 7.5625,
      "learning_rate": 3.283607967575037e-05,
      "loss": 0.5225536823272705,
      "mean_token_accuracy": 0.8741969250142574,
      "num_tokens": 51397852.0,
      "step": 25140
    },
    {
      "entropy": 0.5153032909147441,
      "epoch": 0.40846820363317443,
      "grad_norm": 5.59375,
      "learning_rate": 3.282376567975672e-05,
      "loss": 0.5229459762573242,
      "mean_token_accuracy": 0.8715800285339356,
      "num_tokens": 51417793.0,
      "step": 25150
    },
    {
      "entropy": 0.5368939979467541,
      "epoch": 0.40863061643779997,
      "grad_norm": 9.1875,
      "learning_rate": 3.281144957915252e-05,
      "loss": 0.5754826068878174,
      "mean_token_accuracy": 0.8616510044783354,
      "num_tokens": 51440667.0,
      "step": 25160
    },
    {
      "entropy": 0.49383406941778957,
      "epoch": 0.4087930292424255,
      "grad_norm": 7.75,
      "learning_rate": 3.279913137725082e-05,
      "loss": 0.47629785537719727,
      "mean_token_accuracy": 0.8758905433118344,
      "num_tokens": 51461719.0,
      "step": 25170
    },
    {
      "entropy": 0.5647580330260098,
      "epoch": 0.408955442047051,
      "grad_norm": 6.09375,
      "learning_rate": 3.2786811077365265e-05,
      "loss": 0.5508701801300049,
      "mean_token_accuracy": 0.8634980488568544,
      "num_tokens": 51482187.0,
      "step": 25180
    },
    {
      "entropy": 0.5274709517136216,
      "epoch": 0.4091178548516765,
      "grad_norm": 5.21875,
      "learning_rate": 3.277448868281002e-05,
      "loss": 0.5422244548797608,
      "mean_token_accuracy": 0.8664735473692418,
      "num_tokens": 51503218.0,
      "step": 25190
    },
    {
      "entropy": 0.4733066062442958,
      "epoch": 0.40928026765630204,
      "grad_norm": 8.75,
      "learning_rate": 3.276216419689986e-05,
      "loss": 0.4914716720581055,
      "mean_token_accuracy": 0.8818962126970291,
      "num_tokens": 51522181.0,
      "step": 25200
    },
    {
      "entropy": 0.5192671744152904,
      "epoch": 0.4094426804609275,
      "grad_norm": 6.65625,
      "learning_rate": 3.2749837622950067e-05,
      "loss": 0.4869736671447754,
      "mean_token_accuracy": 0.8693846091628075,
      "num_tokens": 51542289.0,
      "step": 25210
    },
    {
      "entropy": 0.590949431899935,
      "epoch": 0.40960509326555306,
      "grad_norm": 7.375,
      "learning_rate": 3.2737508964276556e-05,
      "loss": 0.6090972423553467,
      "mean_token_accuracy": 0.8470668002963067,
      "num_tokens": 51564324.0,
      "step": 25220
    },
    {
      "entropy": 0.4933886076323688,
      "epoch": 0.4097675060701786,
      "grad_norm": 10.1875,
      "learning_rate": 3.272517822419573e-05,
      "loss": 0.5148904323577881,
      "mean_token_accuracy": 0.8677328310906887,
      "num_tokens": 51585350.0,
      "step": 25230
    },
    {
      "entropy": 0.5094031355343759,
      "epoch": 0.40992991887480407,
      "grad_norm": 4.9375,
      "learning_rate": 3.271284540602461e-05,
      "loss": 0.5282598018646241,
      "mean_token_accuracy": 0.8695204310119152,
      "num_tokens": 51605613.0,
      "step": 25240
    },
    {
      "entropy": 0.5088961606845259,
      "epoch": 0.4100923316794296,
      "grad_norm": 6.03125,
      "learning_rate": 3.270051051308074e-05,
      "loss": 0.5168599605560302,
      "mean_token_accuracy": 0.8719507984817028,
      "num_tokens": 51627366.0,
      "step": 25250
    },
    {
      "entropy": 0.5920502545777708,
      "epoch": 0.41025474448405513,
      "grad_norm": 7.53125,
      "learning_rate": 3.268817354868223e-05,
      "loss": 0.5765781402587891,
      "mean_token_accuracy": 0.8519936833530665,
      "num_tokens": 51648361.0,
      "step": 25260
    },
    {
      "entropy": 0.5519197563640773,
      "epoch": 0.4104171572886806,
      "grad_norm": 8.75,
      "learning_rate": 3.267583451614777e-05,
      "loss": 0.5451009750366211,
      "mean_token_accuracy": 0.8552171923220158,
      "num_tokens": 51668594.0,
      "step": 25270
    },
    {
      "entropy": 0.5631777686998248,
      "epoch": 0.41057957009330615,
      "grad_norm": 6.3125,
      "learning_rate": 3.2663493418796566e-05,
      "loss": 0.5508751392364502,
      "mean_token_accuracy": 0.8606623455882072,
      "num_tokens": 51689915.0,
      "step": 25280
    },
    {
      "entropy": 0.4951519231311977,
      "epoch": 0.4107419828979317,
      "grad_norm": 9.375,
      "learning_rate": 3.265115025994841e-05,
      "loss": 0.5097496032714843,
      "mean_token_accuracy": 0.8752794310450553,
      "num_tokens": 51709410.0,
      "step": 25290
    },
    {
      "entropy": 0.60302872704342,
      "epoch": 0.4109043957025572,
      "grad_norm": 7.96875,
      "learning_rate": 3.2638805042923635e-05,
      "loss": 0.5931108474731446,
      "mean_token_accuracy": 0.857911828905344,
      "num_tokens": 51730534.0,
      "step": 25300
    },
    {
      "entropy": 0.5163030728232115,
      "epoch": 0.4110668085071827,
      "grad_norm": 8.5,
      "learning_rate": 3.262645777104315e-05,
      "loss": 0.504947566986084,
      "mean_token_accuracy": 0.8620601341128349,
      "num_tokens": 51750419.0,
      "step": 25310
    },
    {
      "entropy": 0.6103019653353841,
      "epoch": 0.4112292213118082,
      "grad_norm": 7.46875,
      "learning_rate": 3.2614108447628376e-05,
      "loss": 0.6235942840576172,
      "mean_token_accuracy": 0.843829907476902,
      "num_tokens": 51772364.0,
      "step": 25320
    },
    {
      "entropy": 0.5975800118874759,
      "epoch": 0.41139163411643376,
      "grad_norm": 7.3125,
      "learning_rate": 3.260175707600132e-05,
      "loss": 0.6280171871185303,
      "mean_token_accuracy": 0.845548202842474,
      "num_tokens": 51793452.0,
      "step": 25330
    },
    {
      "entropy": 0.4531507635489106,
      "epoch": 0.41155404692105924,
      "grad_norm": 5.40625,
      "learning_rate": 3.258940365948452e-05,
      "loss": 0.4364427089691162,
      "mean_token_accuracy": 0.8826680421829224,
      "num_tokens": 51814768.0,
      "step": 25340
    },
    {
      "entropy": 0.5348030606284737,
      "epoch": 0.41171645972568477,
      "grad_norm": 9.1875,
      "learning_rate": 3.2577048201401085e-05,
      "loss": 0.5560841083526611,
      "mean_token_accuracy": 0.8594465799629688,
      "num_tokens": 51835517.0,
      "step": 25350
    },
    {
      "entropy": 0.5656724194064736,
      "epoch": 0.4118788725303103,
      "grad_norm": 7.25,
      "learning_rate": 3.256469070507466e-05,
      "loss": 0.5837292194366455,
      "mean_token_accuracy": 0.8510655704885721,
      "num_tokens": 51856206.0,
      "step": 25360
    },
    {
      "entropy": 0.5455660265404731,
      "epoch": 0.4120412853349358,
      "grad_norm": 8.5625,
      "learning_rate": 3.2552331173829424e-05,
      "loss": 0.5626893520355225,
      "mean_token_accuracy": 0.8596199974417686,
      "num_tokens": 51875750.0,
      "step": 25370
    },
    {
      "entropy": 0.5807529872749001,
      "epoch": 0.4122036981395613,
      "grad_norm": 7.15625,
      "learning_rate": 3.253996961099014e-05,
      "loss": 0.5728814601898193,
      "mean_token_accuracy": 0.8529332995414733,
      "num_tokens": 51897242.0,
      "step": 25380
    },
    {
      "entropy": 0.4849427772220224,
      "epoch": 0.41236611094418685,
      "grad_norm": 7.15625,
      "learning_rate": 3.2527606019882086e-05,
      "loss": 0.48970746994018555,
      "mean_token_accuracy": 0.8802861593663692,
      "num_tokens": 51916240.0,
      "step": 25390
    },
    {
      "entropy": 0.5293574736453592,
      "epoch": 0.4125285237488124,
      "grad_norm": 7.53125,
      "learning_rate": 3.251524040383108e-05,
      "loss": 0.5553319454193115,
      "mean_token_accuracy": 0.8593196079134942,
      "num_tokens": 51936543.0,
      "step": 25400
    },
    {
      "entropy": 0.5929553141817451,
      "epoch": 0.41269093655343786,
      "grad_norm": 8.4375,
      "learning_rate": 3.250287276616353e-05,
      "loss": 0.6112259864807129,
      "mean_token_accuracy": 0.8469288147985935,
      "num_tokens": 51955387.0,
      "step": 25410
    },
    {
      "entropy": 0.540330576710403,
      "epoch": 0.4128533493580634,
      "grad_norm": 6.5,
      "learning_rate": 3.249050311020634e-05,
      "loss": 0.5391010761260986,
      "mean_token_accuracy": 0.8628093712031841,
      "num_tokens": 51975647.0,
      "step": 25420
    },
    {
      "entropy": 0.520256845606491,
      "epoch": 0.4130157621626889,
      "grad_norm": 5.71875,
      "learning_rate": 3.247813143928698e-05,
      "loss": 0.5635770320892334,
      "mean_token_accuracy": 0.865212307870388,
      "num_tokens": 51996035.0,
      "step": 25430
    },
    {
      "entropy": 0.5693472034297884,
      "epoch": 0.4131781749673144,
      "grad_norm": 8.1875,
      "learning_rate": 3.2465757756733454e-05,
      "loss": 0.5446886539459228,
      "mean_token_accuracy": 0.8683592341840267,
      "num_tokens": 52016722.0,
      "step": 25440
    },
    {
      "entropy": 0.5002526983618736,
      "epoch": 0.41334058777193994,
      "grad_norm": 7.21875,
      "learning_rate": 3.245338206587432e-05,
      "loss": 0.47034468650817873,
      "mean_token_accuracy": 0.8788267187774181,
      "num_tokens": 52037205.0,
      "step": 25450
    },
    {
      "entropy": 0.5580893530510366,
      "epoch": 0.41350300057656547,
      "grad_norm": 6.65625,
      "learning_rate": 3.2441004370038656e-05,
      "loss": 0.5750881195068359,
      "mean_token_accuracy": 0.8579423114657402,
      "num_tokens": 52057299.0,
      "step": 25460
    },
    {
      "entropy": 0.5026374509092421,
      "epoch": 0.41366541338119095,
      "grad_norm": 6.3125,
      "learning_rate": 3.2428624672556095e-05,
      "loss": 0.5181890487670898,
      "mean_token_accuracy": 0.8667039625346661,
      "num_tokens": 52077676.0,
      "step": 25470
    },
    {
      "entropy": 0.4719900216907263,
      "epoch": 0.4138278261858165,
      "grad_norm": 9.4375,
      "learning_rate": 3.24162429767568e-05,
      "loss": 0.46311087608337403,
      "mean_token_accuracy": 0.8847668409347534,
      "num_tokens": 52096380.0,
      "step": 25480
    },
    {
      "entropy": 0.49754854626953604,
      "epoch": 0.413990238990442,
      "grad_norm": 7.3125,
      "learning_rate": 3.240385928597148e-05,
      "loss": 0.49266839027404785,
      "mean_token_accuracy": 0.8770412042737007,
      "num_tokens": 52116400.0,
      "step": 25490
    },
    {
      "entropy": 0.4713777111377567,
      "epoch": 0.41415265179506755,
      "grad_norm": 7.71875,
      "learning_rate": 3.2391473603531374e-05,
      "loss": 0.44735121726989746,
      "mean_token_accuracy": 0.8792950481176376,
      "num_tokens": 52136643.0,
      "step": 25500
    },
    {
      "entropy": 0.5501636788481846,
      "epoch": 0.41431506459969303,
      "grad_norm": 8.5,
      "learning_rate": 3.237908593276825e-05,
      "loss": 0.5728245258331299,
      "mean_token_accuracy": 0.8529164776206016,
      "num_tokens": 52156536.0,
      "step": 25510
    },
    {
      "entropy": 0.5381138104945421,
      "epoch": 0.41447747740431856,
      "grad_norm": 6.84375,
      "learning_rate": 3.236669627701443e-05,
      "loss": 0.5558248519897461,
      "mean_token_accuracy": 0.8595959283411503,
      "num_tokens": 52176535.0,
      "step": 25520
    },
    {
      "entropy": 0.5629498870112002,
      "epoch": 0.4146398902089441,
      "grad_norm": 8.0625,
      "learning_rate": 3.2354304639602765e-05,
      "loss": 0.5796255111694336,
      "mean_token_accuracy": 0.8553454361855983,
      "num_tokens": 52197749.0,
      "step": 25530
    },
    {
      "entropy": 0.5713775873184204,
      "epoch": 0.4148023030135696,
      "grad_norm": 8.75,
      "learning_rate": 3.2341911023866617e-05,
      "loss": 0.5597432613372803,
      "mean_token_accuracy": 0.8614897020161152,
      "num_tokens": 52218289.0,
      "step": 25540
    },
    {
      "entropy": 0.5614398820325732,
      "epoch": 0.4149647158181951,
      "grad_norm": 9.8125,
      "learning_rate": 3.2329515433139897e-05,
      "loss": 0.6000655651092529,
      "mean_token_accuracy": 0.8536012053489686,
      "num_tokens": 52240413.0,
      "step": 25550
    },
    {
      "entropy": 0.6442205234896392,
      "epoch": 0.41512712862282064,
      "grad_norm": 6.46875,
      "learning_rate": 3.231711787075706e-05,
      "loss": 0.6977475643157959,
      "mean_token_accuracy": 0.8337103236466646,
      "num_tokens": 52261170.0,
      "step": 25560
    },
    {
      "entropy": 0.5533858236391097,
      "epoch": 0.4152895414274461,
      "grad_norm": 7.34375,
      "learning_rate": 3.230471834005308e-05,
      "loss": 0.5420894622802734,
      "mean_token_accuracy": 0.8603348501026631,
      "num_tokens": 52282515.0,
      "step": 25570
    },
    {
      "entropy": 0.5799517730250955,
      "epoch": 0.41545195423207165,
      "grad_norm": 10.0625,
      "learning_rate": 3.229231684436344e-05,
      "loss": 0.5732659816741943,
      "mean_token_accuracy": 0.84960383400321,
      "num_tokens": 52302630.0,
      "step": 25580
    },
    {
      "entropy": 0.4566038035438396,
      "epoch": 0.4156143670366972,
      "grad_norm": 5.09375,
      "learning_rate": 3.227991338702417e-05,
      "loss": 0.4110555648803711,
      "mean_token_accuracy": 0.8857577197253704,
      "num_tokens": 52323175.0,
      "step": 25590
    },
    {
      "entropy": 0.4929009010549635,
      "epoch": 0.41577677984132266,
      "grad_norm": 6.625,
      "learning_rate": 3.226750797137185e-05,
      "loss": 0.46628379821777344,
      "mean_token_accuracy": 0.8678748615086078,
      "num_tokens": 52344101.0,
      "step": 25600
    },
    {
      "entropy": 0.47562762089073657,
      "epoch": 0.4159391926459482,
      "grad_norm": 5.4375,
      "learning_rate": 3.225510060074356e-05,
      "loss": 0.49250307083129885,
      "mean_token_accuracy": 0.8703025840222836,
      "num_tokens": 52364114.0,
      "step": 25610
    },
    {
      "entropy": 0.5088329302147031,
      "epoch": 0.41610160545057373,
      "grad_norm": 6.6875,
      "learning_rate": 3.22426912784769e-05,
      "loss": 0.4973293304443359,
      "mean_token_accuracy": 0.8665605306625366,
      "num_tokens": 52385463.0,
      "step": 25620
    },
    {
      "entropy": 0.5468743508681655,
      "epoch": 0.41626401825519926,
      "grad_norm": 10.5625,
      "learning_rate": 3.223028000791e-05,
      "loss": 0.5036086559295654,
      "mean_token_accuracy": 0.868207111954689,
      "num_tokens": 52406049.0,
      "step": 25630
    },
    {
      "entropy": 0.44798689559102056,
      "epoch": 0.41642643105982474,
      "grad_norm": 5.53125,
      "learning_rate": 3.221786679238154e-05,
      "loss": 0.44849424362182616,
      "mean_token_accuracy": 0.8869002774357796,
      "num_tokens": 52426784.0,
      "step": 25640
    },
    {
      "entropy": 0.42774236453697084,
      "epoch": 0.4165888438644503,
      "grad_norm": 11.375,
      "learning_rate": 3.220545163523068e-05,
      "loss": 0.4663724899291992,
      "mean_token_accuracy": 0.8823844365775585,
      "num_tokens": 52447285.0,
      "step": 25650
    },
    {
      "entropy": 0.5102943607140332,
      "epoch": 0.4167512566690758,
      "grad_norm": 6.375,
      "learning_rate": 3.219303453979715e-05,
      "loss": 0.503302526473999,
      "mean_token_accuracy": 0.8683681294322014,
      "num_tokens": 52467046.0,
      "step": 25660
    },
    {
      "entropy": 0.5356067596003413,
      "epoch": 0.4169136694737013,
      "grad_norm": 9.875,
      "learning_rate": 3.218061550942115e-05,
      "loss": 0.597426462173462,
      "mean_token_accuracy": 0.84616807885468,
      "num_tokens": 52488399.0,
      "step": 25670
    },
    {
      "entropy": 0.5494789345189929,
      "epoch": 0.4170760822783268,
      "grad_norm": 8.0625,
      "learning_rate": 3.216819454744345e-05,
      "loss": 0.5711367130279541,
      "mean_token_accuracy": 0.8596862509846688,
      "num_tokens": 52508870.0,
      "step": 25680
    },
    {
      "entropy": 0.5665111511480063,
      "epoch": 0.41723849508295235,
      "grad_norm": 9.875,
      "learning_rate": 3.2155771657205304e-05,
      "loss": 0.5598531723022461,
      "mean_token_accuracy": 0.8618370190262794,
      "num_tokens": 52530077.0,
      "step": 25690
    },
    {
      "entropy": 0.5230646975338459,
      "epoch": 0.41740090788757783,
      "grad_norm": 8.3125,
      "learning_rate": 3.21433468420485e-05,
      "loss": 0.5428545475006104,
      "mean_token_accuracy": 0.8602124769240618,
      "num_tokens": 52549291.0,
      "step": 25700
    },
    {
      "entropy": 0.49194653034210206,
      "epoch": 0.41756332069220337,
      "grad_norm": 5.46875,
      "learning_rate": 3.213092010531534e-05,
      "loss": 0.46704850196838377,
      "mean_token_accuracy": 0.8761406935751438,
      "num_tokens": 52569178.0,
      "step": 25710
    },
    {
      "entropy": 0.518027339456603,
      "epoch": 0.4177257334968289,
      "grad_norm": 7.5625,
      "learning_rate": 3.211849145034865e-05,
      "loss": 0.508852767944336,
      "mean_token_accuracy": 0.8703674167394638,
      "num_tokens": 52589296.0,
      "step": 25720
    },
    {
      "entropy": 0.5370525011792779,
      "epoch": 0.41788814630145443,
      "grad_norm": 8.4375,
      "learning_rate": 3.210606088049176e-05,
      "loss": 0.5458516120910645,
      "mean_token_accuracy": 0.8614116158336401,
      "num_tokens": 52610241.0,
      "step": 25730
    },
    {
      "entropy": 0.5422722487710416,
      "epoch": 0.4180505591060799,
      "grad_norm": 10.625,
      "learning_rate": 3.209362839908853e-05,
      "loss": 0.6065820693969727,
      "mean_token_accuracy": 0.8461984626948833,
      "num_tokens": 52629335.0,
      "step": 25740
    },
    {
      "entropy": 0.5094201260246336,
      "epoch": 0.41821297191070544,
      "grad_norm": 7.78125,
      "learning_rate": 3.208119400948331e-05,
      "loss": 0.48579044342041017,
      "mean_token_accuracy": 0.8735195487737656,
      "num_tokens": 52650933.0,
      "step": 25750
    },
    {
      "entropy": 0.516224688757211,
      "epoch": 0.418375384715331,
      "grad_norm": 6.8125,
      "learning_rate": 3.2068757715020995e-05,
      "loss": 0.483828067779541,
      "mean_token_accuracy": 0.8724147580564022,
      "num_tokens": 52671127.0,
      "step": 25760
    },
    {
      "entropy": 0.4885833171196282,
      "epoch": 0.41853779751995646,
      "grad_norm": 6.71875,
      "learning_rate": 3.205631951904696e-05,
      "loss": 0.4948446750640869,
      "mean_token_accuracy": 0.8720498990267516,
      "num_tokens": 52692165.0,
      "step": 25770
    },
    {
      "entropy": 0.5013317247852683,
      "epoch": 0.418700210324582,
      "grad_norm": 7.9375,
      "learning_rate": 3.204387942490712e-05,
      "loss": 0.5151903629302979,
      "mean_token_accuracy": 0.874787350744009,
      "num_tokens": 52712219.0,
      "step": 25780
    },
    {
      "entropy": 0.5527803116478026,
      "epoch": 0.4188626231292075,
      "grad_norm": 7.6875,
      "learning_rate": 3.203143743594789e-05,
      "loss": 0.5959082126617432,
      "mean_token_accuracy": 0.8578294094651937,
      "num_tokens": 52733150.0,
      "step": 25790
    },
    {
      "entropy": 0.48373413900844753,
      "epoch": 0.419025035933833,
      "grad_norm": 6.9375,
      "learning_rate": 3.201899355551618e-05,
      "loss": 0.4935576915740967,
      "mean_token_accuracy": 0.871484537422657,
      "num_tokens": 52753274.0,
      "step": 25800
    },
    {
      "entropy": 0.5884950891137123,
      "epoch": 0.41918744873845853,
      "grad_norm": 7.71875,
      "learning_rate": 3.200654778695943e-05,
      "loss": 0.6307910442352295,
      "mean_token_accuracy": 0.8478124819695949,
      "num_tokens": 52773925.0,
      "step": 25810
    },
    {
      "entropy": 0.47720295009203256,
      "epoch": 0.41934986154308407,
      "grad_norm": 5.75,
      "learning_rate": 3.199410013362558e-05,
      "loss": 0.44926652908325193,
      "mean_token_accuracy": 0.8743626110255718,
      "num_tokens": 52793907.0,
      "step": 25820
    },
    {
      "entropy": 0.5504545530304312,
      "epoch": 0.41951227434770955,
      "grad_norm": 7.90625,
      "learning_rate": 3.198165059886307e-05,
      "loss": 0.5332975387573242,
      "mean_token_accuracy": 0.863922755792737,
      "num_tokens": 52814112.0,
      "step": 25830
    },
    {
      "entropy": 0.5630201070569456,
      "epoch": 0.4196746871523351,
      "grad_norm": 7.3125,
      "learning_rate": 3.196919918602087e-05,
      "loss": 0.5809246063232422,
      "mean_token_accuracy": 0.8578296817839146,
      "num_tokens": 52834995.0,
      "step": 25840
    },
    {
      "entropy": 0.5794985789339989,
      "epoch": 0.4198370999569606,
      "grad_norm": 10.9375,
      "learning_rate": 3.195674589844842e-05,
      "loss": 0.6090047359466553,
      "mean_token_accuracy": 0.8550545275211334,
      "num_tokens": 52854951.0,
      "step": 25850
    },
    {
      "entropy": 0.4636266532354057,
      "epoch": 0.41999951276158615,
      "grad_norm": 8.5,
      "learning_rate": 3.1944290739495695e-05,
      "loss": 0.4624065399169922,
      "mean_token_accuracy": 0.8780642818659544,
      "num_tokens": 52875530.0,
      "step": 25860
    },
    {
      "entropy": 0.4846934449742548,
      "epoch": 0.4201619255662116,
      "grad_norm": 8.5625,
      "learning_rate": 3.1931833712513146e-05,
      "loss": 0.4879265308380127,
      "mean_token_accuracy": 0.8752511214464903,
      "num_tokens": 52895664.0,
      "step": 25870
    },
    {
      "entropy": 0.5569766034139321,
      "epoch": 0.42032433837083716,
      "grad_norm": 6.3125,
      "learning_rate": 3.191937482085176e-05,
      "loss": 0.5507461071014405,
      "mean_token_accuracy": 0.8651439655572176,
      "num_tokens": 52917394.0,
      "step": 25880
    },
    {
      "entropy": 0.5131411178037524,
      "epoch": 0.4204867511754627,
      "grad_norm": 7.125,
      "learning_rate": 3.190691406786299e-05,
      "loss": 0.47635536193847655,
      "mean_token_accuracy": 0.8790477592498064,
      "num_tokens": 52937302.0,
      "step": 25890
    },
    {
      "entropy": 0.516530344914645,
      "epoch": 0.42064916398008817,
      "grad_norm": 6.875,
      "learning_rate": 3.189445145689882e-05,
      "loss": 0.5579797267913819,
      "mean_token_accuracy": 0.8636225178837776,
      "num_tokens": 52957890.0,
      "step": 25900
    },
    {
      "entropy": 0.5130226615816355,
      "epoch": 0.4208115767847137,
      "grad_norm": 5.875,
      "learning_rate": 3.188198699131171e-05,
      "loss": 0.5303301334381103,
      "mean_token_accuracy": 0.869339008629322,
      "num_tokens": 52977769.0,
      "step": 25910
    },
    {
      "entropy": 0.5385393127333373,
      "epoch": 0.42097398958933924,
      "grad_norm": 8.75,
      "learning_rate": 3.1869520674454624e-05,
      "loss": 0.530564546585083,
      "mean_token_accuracy": 0.8683712385594845,
      "num_tokens": 52998757.0,
      "step": 25920
    },
    {
      "entropy": 0.5261168368626385,
      "epoch": 0.4211364023939647,
      "grad_norm": 7.75,
      "learning_rate": 3.185705250968105e-05,
      "loss": 0.47595987319946287,
      "mean_token_accuracy": 0.883159664273262,
      "num_tokens": 53018613.0,
      "step": 25930
    },
    {
      "entropy": 0.529576008557342,
      "epoch": 0.42129881519859025,
      "grad_norm": 7.28125,
      "learning_rate": 3.184458250034493e-05,
      "loss": 0.5502344608306885,
      "mean_token_accuracy": 0.8604525402188301,
      "num_tokens": 53039739.0,
      "step": 25940
    },
    {
      "entropy": 0.5339185941964388,
      "epoch": 0.4214612280032158,
      "grad_norm": 9.5625,
      "learning_rate": 3.183211064980075e-05,
      "loss": 0.578618335723877,
      "mean_token_accuracy": 0.8596599020063878,
      "num_tokens": 53060331.0,
      "step": 25950
    },
    {
      "entropy": 0.5812776100821793,
      "epoch": 0.4216236408078413,
      "grad_norm": 13.0,
      "learning_rate": 3.1819636961403443e-05,
      "loss": 0.5753157615661622,
      "mean_token_accuracy": 0.8559711903333664,
      "num_tokens": 53081190.0,
      "step": 25960
    },
    {
      "entropy": 0.5282956954091788,
      "epoch": 0.4217860536124668,
      "grad_norm": 7.84375,
      "learning_rate": 3.1807161438508465e-05,
      "loss": 0.5454727649688721,
      "mean_token_accuracy": 0.86385711543262,
      "num_tokens": 53102555.0,
      "step": 25970
    },
    {
      "entropy": 0.495172376325354,
      "epoch": 0.4219484664170923,
      "grad_norm": 8.5,
      "learning_rate": 3.179468408447176e-05,
      "loss": 0.5281413555145263,
      "mean_token_accuracy": 0.8639149211347104,
      "num_tokens": 53122194.0,
      "step": 25980
    },
    {
      "entropy": 0.5663021387532353,
      "epoch": 0.42211087922171786,
      "grad_norm": 6.8125,
      "learning_rate": 3.178220490264977e-05,
      "loss": 0.6023267269134521,
      "mean_token_accuracy": 0.850651728734374,
      "num_tokens": 53143523.0,
      "step": 25990
    },
    {
      "entropy": 0.46800175299867985,
      "epoch": 0.42227329202634334,
      "grad_norm": 7.84375,
      "learning_rate": 3.1769723896399415e-05,
      "loss": 0.4666964054107666,
      "mean_token_accuracy": 0.8823631525039672,
      "num_tokens": 53165978.0,
      "step": 26000
    },
    {
      "entropy": 0.5163110538385809,
      "epoch": 0.4224357048309689,
      "grad_norm": 5.28125,
      "learning_rate": 3.1757241069078126e-05,
      "loss": 0.5024280071258544,
      "mean_token_accuracy": 0.8668322116136551,
      "num_tokens": 53187506.0,
      "step": 26010
    },
    {
      "entropy": 0.48313176329247653,
      "epoch": 0.4225981176355944,
      "grad_norm": 5.75,
      "learning_rate": 3.1744756424043803e-05,
      "loss": 0.5220269203186035,
      "mean_token_accuracy": 0.8713211715221405,
      "num_tokens": 53207209.0,
      "step": 26020
    },
    {
      "entropy": 0.5362095199991017,
      "epoch": 0.4227605304402199,
      "grad_norm": 8.0,
      "learning_rate": 3.173226996465484e-05,
      "loss": 0.5468008995056153,
      "mean_token_accuracy": 0.8625670209527015,
      "num_tokens": 53226835.0,
      "step": 26030
    },
    {
      "entropy": 0.5095091768074781,
      "epoch": 0.4229229432448454,
      "grad_norm": 6.75,
      "learning_rate": 3.1719781694270124e-05,
      "loss": 0.49617924690246584,
      "mean_token_accuracy": 0.8644309170544148,
      "num_tokens": 53247078.0,
      "step": 26040
    },
    {
      "entropy": 0.4972918125800788,
      "epoch": 0.42308535604947095,
      "grad_norm": 8.625,
      "learning_rate": 3.170729161624904e-05,
      "loss": 0.5204285144805908,
      "mean_token_accuracy": 0.8703681707382203,
      "num_tokens": 53267859.0,
      "step": 26050
    },
    {
      "entropy": 0.5583084385376423,
      "epoch": 0.4232477688540965,
      "grad_norm": 8.6875,
      "learning_rate": 3.169479973395143e-05,
      "loss": 0.5382142066955566,
      "mean_token_accuracy": 0.861668660491705,
      "num_tokens": 53288153.0,
      "step": 26060
    },
    {
      "entropy": 0.5674798764288426,
      "epoch": 0.42341018165872196,
      "grad_norm": 9.25,
      "learning_rate": 3.168230605073765e-05,
      "loss": 0.5714748859405517,
      "mean_token_accuracy": 0.8572727739810944,
      "num_tokens": 53308704.0,
      "step": 26070
    },
    {
      "entropy": 0.5764185952953994,
      "epoch": 0.4235725944633475,
      "grad_norm": 7.3125,
      "learning_rate": 3.166981056996852e-05,
      "loss": 0.5755550384521484,
      "mean_token_accuracy": 0.8588298425078392,
      "num_tokens": 53329202.0,
      "step": 26080
    },
    {
      "entropy": 0.4808438041713089,
      "epoch": 0.42373500726797303,
      "grad_norm": 6.125,
      "learning_rate": 3.165731329500536e-05,
      "loss": 0.4663823127746582,
      "mean_token_accuracy": 0.8789558589458466,
      "num_tokens": 53348725.0,
      "step": 26090
    },
    {
      "entropy": 0.48298922828398644,
      "epoch": 0.4238974200725985,
      "grad_norm": 7.9375,
      "learning_rate": 3.164481422920996e-05,
      "loss": 0.47258687019348145,
      "mean_token_accuracy": 0.8739002346992493,
      "num_tokens": 53368321.0,
      "step": 26100
    },
    {
      "entropy": 0.4612211043480784,
      "epoch": 0.42405983287722404,
      "grad_norm": 11.375,
      "learning_rate": 3.16323133759446e-05,
      "loss": 0.429952335357666,
      "mean_token_accuracy": 0.8857774078845978,
      "num_tokens": 53388986.0,
      "step": 26110
    },
    {
      "entropy": 0.5063950899988413,
      "epoch": 0.4242222456818496,
      "grad_norm": 6.5,
      "learning_rate": 3.161981073857203e-05,
      "loss": 0.516152811050415,
      "mean_token_accuracy": 0.8707935132086277,
      "num_tokens": 53409712.0,
      "step": 26120
    },
    {
      "entropy": 0.5596029178239406,
      "epoch": 0.42438465848647505,
      "grad_norm": 5.625,
      "learning_rate": 3.160730632045551e-05,
      "loss": 0.5536952495574952,
      "mean_token_accuracy": 0.859734732657671,
      "num_tokens": 53430995.0,
      "step": 26130
    },
    {
      "entropy": 0.46577100141439587,
      "epoch": 0.4245470712911006,
      "grad_norm": 6.6875,
      "learning_rate": 3.159480012495873e-05,
      "loss": 0.47863178253173827,
      "mean_token_accuracy": 0.8788735270500183,
      "num_tokens": 53451531.0,
      "step": 26140
    },
    {
      "entropy": 0.4203683691797778,
      "epoch": 0.4247094840957261,
      "grad_norm": 8.375,
      "learning_rate": 3.1582292155445896e-05,
      "loss": 0.4302022457122803,
      "mean_token_accuracy": 0.8908935882151127,
      "num_tokens": 53470653.0,
      "step": 26150
    },
    {
      "entropy": 0.5022329020313918,
      "epoch": 0.4248718969003516,
      "grad_norm": 6.65625,
      "learning_rate": 3.156978241528168e-05,
      "loss": 0.518714714050293,
      "mean_token_accuracy": 0.8695842556655407,
      "num_tokens": 53491423.0,
      "step": 26160
    },
    {
      "entropy": 0.5512746045831591,
      "epoch": 0.42503430970497713,
      "grad_norm": 9.75,
      "learning_rate": 3.155727090783123e-05,
      "loss": 0.5485617637634277,
      "mean_token_accuracy": 0.8620844349265099,
      "num_tokens": 53511769.0,
      "step": 26170
    },
    {
      "entropy": 0.5347740378696472,
      "epoch": 0.42519672250960266,
      "grad_norm": 5.125,
      "learning_rate": 3.1544757636460176e-05,
      "loss": 0.5319727420806885,
      "mean_token_accuracy": 0.8667205452919007,
      "num_tokens": 53532387.0,
      "step": 26180
    },
    {
      "entropy": 0.5781703687272965,
      "epoch": 0.4253591353142282,
      "grad_norm": 10.3125,
      "learning_rate": 3.15322426045346e-05,
      "loss": 0.6596169471740723,
      "mean_token_accuracy": 0.8383496649563312,
      "num_tokens": 53554291.0,
      "step": 26190
    },
    {
      "entropy": 0.5523478910792619,
      "epoch": 0.4255215481188537,
      "grad_norm": 5.1875,
      "learning_rate": 3.1519725815421086e-05,
      "loss": 0.5360152721405029,
      "mean_token_accuracy": 0.8639686457812786,
      "num_tokens": 53574168.0,
      "step": 26200
    },
    {
      "entropy": 0.5805981829296798,
      "epoch": 0.4256839609234792,
      "grad_norm": 8.4375,
      "learning_rate": 3.150720727248668e-05,
      "loss": 0.5832035064697265,
      "mean_token_accuracy": 0.8591724395751953,
      "num_tokens": 53595180.0,
      "step": 26210
    },
    {
      "entropy": 0.4912083494476974,
      "epoch": 0.42584637372810474,
      "grad_norm": 5.125,
      "learning_rate": 3.149468697909889e-05,
      "loss": 0.4988560199737549,
      "mean_token_accuracy": 0.8784642018377781,
      "num_tokens": 53615193.0,
      "step": 26220
    },
    {
      "entropy": 0.5159098997013644,
      "epoch": 0.4260087865327302,
      "grad_norm": 6.28125,
      "learning_rate": 3.14821649386257e-05,
      "loss": 0.5352456092834472,
      "mean_token_accuracy": 0.8633601270616055,
      "num_tokens": 53636589.0,
      "step": 26230
    },
    {
      "entropy": 0.49961121287196875,
      "epoch": 0.42617119933735575,
      "grad_norm": 12.4375,
      "learning_rate": 3.146964115443559e-05,
      "loss": 0.4994646549224854,
      "mean_token_accuracy": 0.8702877499163151,
      "num_tokens": 53656808.0,
      "step": 26240
    },
    {
      "entropy": 0.5086843435186893,
      "epoch": 0.4263336121419813,
      "grad_norm": 7.53125,
      "learning_rate": 3.1457115629897456e-05,
      "loss": 0.5040114879608154,
      "mean_token_accuracy": 0.8692098781466484,
      "num_tokens": 53676697.0,
      "step": 26250
    },
    {
      "entropy": 0.5707472438924015,
      "epoch": 0.42649602494660677,
      "grad_norm": 7.09375,
      "learning_rate": 3.1444588368380715e-05,
      "loss": 0.5983742237091064,
      "mean_token_accuracy": 0.8518050938844681,
      "num_tokens": 53696767.0,
      "step": 26260
    },
    {
      "entropy": 0.4743402528576553,
      "epoch": 0.4266584377512323,
      "grad_norm": 5.59375,
      "learning_rate": 3.143205937325521e-05,
      "loss": 0.4283727169036865,
      "mean_token_accuracy": 0.8885200098156929,
      "num_tokens": 53717109.0,
      "step": 26270
    },
    {
      "entropy": 0.4650275648571551,
      "epoch": 0.42682085055585783,
      "grad_norm": 7.71875,
      "learning_rate": 3.141952864789128e-05,
      "loss": 0.45522541999816896,
      "mean_token_accuracy": 0.8747443877160549,
      "num_tokens": 53736952.0,
      "step": 26280
    },
    {
      "entropy": 0.520511639630422,
      "epoch": 0.42698326336048337,
      "grad_norm": 6.625,
      "learning_rate": 3.140699619565972e-05,
      "loss": 0.5124590873718262,
      "mean_token_accuracy": 0.8666912041604519,
      "num_tokens": 53756854.0,
      "step": 26290
    },
    {
      "entropy": 0.5814689291641116,
      "epoch": 0.42714567616510885,
      "grad_norm": 10.8125,
      "learning_rate": 3.1394462019931785e-05,
      "loss": 0.5890640258789063,
      "mean_token_accuracy": 0.8591370917856693,
      "num_tokens": 53778286.0,
      "step": 26300
    },
    {
      "entropy": 0.48275368600152435,
      "epoch": 0.4273080889697344,
      "grad_norm": 8.9375,
      "learning_rate": 3.1381926124079196e-05,
      "loss": 0.47987875938415525,
      "mean_token_accuracy": 0.8782180935144425,
      "num_tokens": 53798203.0,
      "step": 26310
    },
    {
      "entropy": 0.5148432906717062,
      "epoch": 0.4274705017743599,
      "grad_norm": 5.84375,
      "learning_rate": 3.136938851147414e-05,
      "loss": 0.5323846817016602,
      "mean_token_accuracy": 0.8673043251037598,
      "num_tokens": 53818641.0,
      "step": 26320
    },
    {
      "entropy": 0.4841863142326474,
      "epoch": 0.4276329145789854,
      "grad_norm": 6.59375,
      "learning_rate": 3.135684918548925e-05,
      "loss": 0.4680328845977783,
      "mean_token_accuracy": 0.8780590943992138,
      "num_tokens": 53840093.0,
      "step": 26330
    },
    {
      "entropy": 0.5157970013096929,
      "epoch": 0.4277953273836109,
      "grad_norm": 10.25,
      "learning_rate": 3.1344308149497655e-05,
      "loss": 0.5487277030944824,
      "mean_token_accuracy": 0.8679249748587609,
      "num_tokens": 53859973.0,
      "step": 26340
    },
    {
      "entropy": 0.5025227293372154,
      "epoch": 0.42795774018823646,
      "grad_norm": 9.25,
      "learning_rate": 3.13317654068729e-05,
      "loss": 0.4844356536865234,
      "mean_token_accuracy": 0.8813383288681507,
      "num_tokens": 53879408.0,
      "step": 26350
    },
    {
      "entropy": 0.45562847042456267,
      "epoch": 0.42812015299286194,
      "grad_norm": 4.71875,
      "learning_rate": 3.1319220960989024e-05,
      "loss": 0.5080379962921142,
      "mean_token_accuracy": 0.8688146375119686,
      "num_tokens": 53898922.0,
      "step": 26360
    },
    {
      "entropy": 0.5546802303986624,
      "epoch": 0.42828256579748747,
      "grad_norm": 7.3125,
      "learning_rate": 3.1306674815220516e-05,
      "loss": 0.5671001434326172,
      "mean_token_accuracy": 0.8508445955812931,
      "num_tokens": 53920089.0,
      "step": 26370
    },
    {
      "entropy": 0.49256305387243626,
      "epoch": 0.428444978602113,
      "grad_norm": 6.0625,
      "learning_rate": 3.12941269729423e-05,
      "loss": 0.47075700759887695,
      "mean_token_accuracy": 0.8754027597606182,
      "num_tokens": 53939617.0,
      "step": 26380
    },
    {
      "entropy": 0.5048099600942806,
      "epoch": 0.4286073914067385,
      "grad_norm": 10.875,
      "learning_rate": 3.1281577437529786e-05,
      "loss": 0.50218825340271,
      "mean_token_accuracy": 0.8753619566559792,
      "num_tokens": 53960825.0,
      "step": 26390
    },
    {
      "entropy": 0.5073998478706926,
      "epoch": 0.428769804211364,
      "grad_norm": 8.6875,
      "learning_rate": 3.126902621235882e-05,
      "loss": 0.5164266109466553,
      "mean_token_accuracy": 0.8696310162544251,
      "num_tokens": 53981707.0,
      "step": 26400
    },
    {
      "entropy": 0.5174917437601835,
      "epoch": 0.42893221701598955,
      "grad_norm": 7.9375,
      "learning_rate": 3.125647330080572e-05,
      "loss": 0.5196106433868408,
      "mean_token_accuracy": 0.8668940372765064,
      "num_tokens": 54001638.0,
      "step": 26410
    },
    {
      "entropy": 0.5059850884601473,
      "epoch": 0.4290946298206151,
      "grad_norm": 7.78125,
      "learning_rate": 3.1243918706247245e-05,
      "loss": 0.5082090854644775,
      "mean_token_accuracy": 0.8626607395708561,
      "num_tokens": 54021523.0,
      "step": 26420
    },
    {
      "entropy": 0.5528865655418486,
      "epoch": 0.42925704262524056,
      "grad_norm": 7.125,
      "learning_rate": 3.123136243206059e-05,
      "loss": 0.54991455078125,
      "mean_token_accuracy": 0.8627529352903366,
      "num_tokens": 54042675.0,
      "step": 26430
    },
    {
      "entropy": 0.5532626616535709,
      "epoch": 0.4294194554298661,
      "grad_norm": 6.125,
      "learning_rate": 3.121880448162345e-05,
      "loss": 0.5937977313995362,
      "mean_token_accuracy": 0.8589829917997122,
      "num_tokens": 54062895.0,
      "step": 26440
    },
    {
      "entropy": 0.5254064027685672,
      "epoch": 0.4295818682344916,
      "grad_norm": 7.28125,
      "learning_rate": 3.1206244858313935e-05,
      "loss": 0.4956506729125977,
      "mean_token_accuracy": 0.8662436306476593,
      "num_tokens": 54084011.0,
      "step": 26450
    },
    {
      "entropy": 0.6057888876646758,
      "epoch": 0.4297442810391171,
      "grad_norm": 8.1875,
      "learning_rate": 3.119368356551059e-05,
      "loss": 0.5886111736297608,
      "mean_token_accuracy": 0.8451783813536167,
      "num_tokens": 54105189.0,
      "step": 26460
    },
    {
      "entropy": 0.5290971436537802,
      "epoch": 0.42990669384374264,
      "grad_norm": 6.84375,
      "learning_rate": 3.1181120606592454e-05,
      "loss": 0.5410046100616455,
      "mean_token_accuracy": 0.860647463798523,
      "num_tokens": 54124676.0,
      "step": 26470
    },
    {
      "entropy": 0.48946149055846033,
      "epoch": 0.43006910664836817,
      "grad_norm": 7.875,
      "learning_rate": 3.1168555984938986e-05,
      "loss": 0.4987939357757568,
      "mean_token_accuracy": 0.8819429256021977,
      "num_tokens": 54145848.0,
      "step": 26480
    },
    {
      "entropy": 0.5439423554576933,
      "epoch": 0.43023151945299365,
      "grad_norm": 7.59375,
      "learning_rate": 3.1155989703930086e-05,
      "loss": 0.5868628025054932,
      "mean_token_accuracy": 0.8605824045836925,
      "num_tokens": 54165748.0,
      "step": 26490
    },
    {
      "entropy": 0.5332016918575391,
      "epoch": 0.4303939322576192,
      "grad_norm": 14.25,
      "learning_rate": 3.1143421766946126e-05,
      "loss": 0.5294876575469971,
      "mean_token_accuracy": 0.8552493035793305,
      "num_tokens": 54185891.0,
      "step": 26500
    },
    {
      "entropy": 0.5729429565835744,
      "epoch": 0.4305563450622447,
      "grad_norm": 11.5625,
      "learning_rate": 3.1130852177367905e-05,
      "loss": 0.5412538051605225,
      "mean_token_accuracy": 0.8650202982127666,
      "num_tokens": 54206353.0,
      "step": 26510
    },
    {
      "entropy": 0.5026615750975907,
      "epoch": 0.43071875786687025,
      "grad_norm": 6.625,
      "learning_rate": 3.111828093857667e-05,
      "loss": 0.5295844554901123,
      "mean_token_accuracy": 0.8712313890457153,
      "num_tokens": 54226437.0,
      "step": 26520
    },
    {
      "entropy": 0.4892513652332127,
      "epoch": 0.43088117067149573,
      "grad_norm": 9.875,
      "learning_rate": 3.1105708053954106e-05,
      "loss": 0.5077725887298584,
      "mean_token_accuracy": 0.876500590890646,
      "num_tokens": 54246896.0,
      "step": 26530
    },
    {
      "entropy": 0.5436348012648523,
      "epoch": 0.43104358347612126,
      "grad_norm": 9.0625,
      "learning_rate": 3.109313352688237e-05,
      "loss": 0.5929782390594482,
      "mean_token_accuracy": 0.852525944635272,
      "num_tokens": 54267094.0,
      "step": 26540
    },
    {
      "entropy": 0.5234177305363119,
      "epoch": 0.4312059962807468,
      "grad_norm": 8.9375,
      "learning_rate": 3.1080557360744005e-05,
      "loss": 0.5364881038665772,
      "mean_token_accuracy": 0.8646044284105301,
      "num_tokens": 54286352.0,
      "step": 26550
    },
    {
      "entropy": 0.5222048241645098,
      "epoch": 0.4313684090853723,
      "grad_norm": 9.0625,
      "learning_rate": 3.106797955892205e-05,
      "loss": 0.5703032493591309,
      "mean_token_accuracy": 0.8520666599273682,
      "num_tokens": 54307009.0,
      "step": 26560
    },
    {
      "entropy": 0.5417647248599679,
      "epoch": 0.4315308218899978,
      "grad_norm": 8.0625,
      "learning_rate": 3.105540012479996e-05,
      "loss": 0.5160929203033447,
      "mean_token_accuracy": 0.8734475433826446,
      "num_tokens": 54326766.0,
      "step": 26570
    },
    {
      "entropy": 0.5138268779963255,
      "epoch": 0.43169323469462334,
      "grad_norm": 8.125,
      "learning_rate": 3.104281906176163e-05,
      "loss": 0.48660926818847655,
      "mean_token_accuracy": 0.8752601712942123,
      "num_tokens": 54347130.0,
      "step": 26580
    },
    {
      "entropy": 0.5025562706636265,
      "epoch": 0.4318556474992488,
      "grad_norm": 7.5625,
      "learning_rate": 3.1030236373191385e-05,
      "loss": 0.5202531337738037,
      "mean_token_accuracy": 0.8699207060039044,
      "num_tokens": 54367048.0,
      "step": 26590
    },
    {
      "entropy": 0.5354378638789058,
      "epoch": 0.43201806030387435,
      "grad_norm": 8.5625,
      "learning_rate": 3.101765206247401e-05,
      "loss": 0.5609186172485352,
      "mean_token_accuracy": 0.8593226723372936,
      "num_tokens": 54386962.0,
      "step": 26600
    },
    {
      "entropy": 0.4143619808135554,
      "epoch": 0.4321804731084999,
      "grad_norm": 7.34375,
      "learning_rate": 3.1005066132994696e-05,
      "loss": 0.4217390537261963,
      "mean_token_accuracy": 0.8915362112224102,
      "num_tokens": 54407075.0,
      "step": 26610
    },
    {
      "entropy": 0.4838870458304882,
      "epoch": 0.43234288591312536,
      "grad_norm": 8.875,
      "learning_rate": 3.099247858813911e-05,
      "loss": 0.5110352993011474,
      "mean_token_accuracy": 0.8693156875669956,
      "num_tokens": 54427925.0,
      "step": 26620
    },
    {
      "entropy": 0.5089731890708208,
      "epoch": 0.4325052987177509,
      "grad_norm": 5.75,
      "learning_rate": 3.097988943129331e-05,
      "loss": 0.5225047588348388,
      "mean_token_accuracy": 0.860106123238802,
      "num_tokens": 54448530.0,
      "step": 26630
    },
    {
      "entropy": 0.5895540602505207,
      "epoch": 0.43266771152237643,
      "grad_norm": 8.5625,
      "learning_rate": 3.096729866584381e-05,
      "loss": 0.6228594779968262,
      "mean_token_accuracy": 0.8524701084941626,
      "num_tokens": 54469003.0,
      "step": 26640
    },
    {
      "entropy": 0.514575906842947,
      "epoch": 0.43283012432700196,
      "grad_norm": 11.25,
      "learning_rate": 3.095470629517756e-05,
      "loss": 0.5403034687042236,
      "mean_token_accuracy": 0.8595052137970924,
      "num_tokens": 54488798.0,
      "step": 26650
    },
    {
      "entropy": 0.5221792967990041,
      "epoch": 0.43299253713162744,
      "grad_norm": 6.96875,
      "learning_rate": 3.094211232268194e-05,
      "loss": 0.5204185485839844,
      "mean_token_accuracy": 0.8598261848092079,
      "num_tokens": 54508761.0,
      "step": 26660
    },
    {
      "entropy": 0.5681801720522344,
      "epoch": 0.433154949936253,
      "grad_norm": 7.46875,
      "learning_rate": 3.092951675174475e-05,
      "loss": 0.58130784034729,
      "mean_token_accuracy": 0.8572632074356079,
      "num_tokens": 54529524.0,
      "step": 26670
    },
    {
      "entropy": 0.48390958122909067,
      "epoch": 0.4333173627408785,
      "grad_norm": 8.4375,
      "learning_rate": 3.091691958575422e-05,
      "loss": 0.4913118839263916,
      "mean_token_accuracy": 0.8714043438434601,
      "num_tokens": 54550424.0,
      "step": 26680
    },
    {
      "entropy": 0.5250225475057959,
      "epoch": 0.433479775545504,
      "grad_norm": 6.90625,
      "learning_rate": 3.0904320828099026e-05,
      "loss": 0.5219854354858399,
      "mean_token_accuracy": 0.8722835328429938,
      "num_tokens": 54570742.0,
      "step": 26690
    },
    {
      "entropy": 0.4988068257458508,
      "epoch": 0.4336421883501295,
      "grad_norm": 6.59375,
      "learning_rate": 3.089172048216827e-05,
      "loss": 0.4991009712219238,
      "mean_token_accuracy": 0.8717376034706831,
      "num_tokens": 54590738.0,
      "step": 26700
    },
    {
      "entropy": 0.5385075101163238,
      "epoch": 0.43380460115475505,
      "grad_norm": 8.0,
      "learning_rate": 3.0879118551351455e-05,
      "loss": 0.5113287448883057,
      "mean_token_accuracy": 0.8682886637747288,
      "num_tokens": 54611425.0,
      "step": 26710
    },
    {
      "entropy": 0.49123778035864235,
      "epoch": 0.43396701395938053,
      "grad_norm": 8.125,
      "learning_rate": 3.0866515039038543e-05,
      "loss": 0.43929152488708495,
      "mean_token_accuracy": 0.8864768534898758,
      "num_tokens": 54631298.0,
      "step": 26720
    },
    {
      "entropy": 0.4656914257444441,
      "epoch": 0.43412942676400607,
      "grad_norm": 7.375,
      "learning_rate": 3.085390994861989e-05,
      "loss": 0.4675886154174805,
      "mean_token_accuracy": 0.8772530876100063,
      "num_tokens": 54651744.0,
      "step": 26730
    },
    {
      "entropy": 0.5404176487587392,
      "epoch": 0.4342918395686316,
      "grad_norm": 8.125,
      "learning_rate": 3.084130328348632e-05,
      "loss": 0.5728754997253418,
      "mean_token_accuracy": 0.858978233486414,
      "num_tokens": 54672689.0,
      "step": 26740
    },
    {
      "entropy": 0.4994218789041042,
      "epoch": 0.43445425237325713,
      "grad_norm": 5.5625,
      "learning_rate": 3.0828695047029036e-05,
      "loss": 0.5237968444824219,
      "mean_token_accuracy": 0.8697491139173508,
      "num_tokens": 54692767.0,
      "step": 26750
    },
    {
      "entropy": 0.4951876137405634,
      "epoch": 0.4346166651778826,
      "grad_norm": 8.25,
      "learning_rate": 3.081608524263967e-05,
      "loss": 0.49139742851257323,
      "mean_token_accuracy": 0.8708185113966465,
      "num_tokens": 54713406.0,
      "step": 26760
    },
    {
      "entropy": 0.5145915291272104,
      "epoch": 0.43477907798250814,
      "grad_norm": 8.625,
      "learning_rate": 3.080347387371032e-05,
      "loss": 0.5114877700805665,
      "mean_token_accuracy": 0.8670777320861817,
      "num_tokens": 54734301.0,
      "step": 26770
    },
    {
      "entropy": 0.5081913491245359,
      "epoch": 0.4349414907871337,
      "grad_norm": 8.6875,
      "learning_rate": 3.079086094363345e-05,
      "loss": 0.5182557106018066,
      "mean_token_accuracy": 0.8704255856573582,
      "num_tokens": 54755207.0,
      "step": 26780
    },
    {
      "entropy": 0.48840864514932036,
      "epoch": 0.43510390359175916,
      "grad_norm": 7.03125,
      "learning_rate": 3.077824645580198e-05,
      "loss": 0.5048822879791259,
      "mean_token_accuracy": 0.8727167975157499,
      "num_tokens": 54773829.0,
      "step": 26790
    },
    {
      "entropy": 0.5090606729499996,
      "epoch": 0.4352663163963847,
      "grad_norm": 8.9375,
      "learning_rate": 3.07656304136092e-05,
      "loss": 0.5203195571899414,
      "mean_token_accuracy": 0.8654130011796951,
      "num_tokens": 54794210.0,
      "step": 26800
    },
    {
      "entropy": 0.4721269342582673,
      "epoch": 0.4354287292010102,
      "grad_norm": 8.8125,
      "learning_rate": 3.0753012820448914e-05,
      "loss": 0.46477470397949217,
      "mean_token_accuracy": 0.8802007094025612,
      "num_tokens": 54814036.0,
      "step": 26810
    },
    {
      "entropy": 0.5001543298945762,
      "epoch": 0.4355911420056357,
      "grad_norm": 6.125,
      "learning_rate": 3.074039367971523e-05,
      "loss": 0.48691325187683104,
      "mean_token_accuracy": 0.8793206624686718,
      "num_tokens": 54835281.0,
      "step": 26820
    },
    {
      "entropy": 0.5345731656532735,
      "epoch": 0.43575355481026123,
      "grad_norm": 7.5625,
      "learning_rate": 3.072777299480275e-05,
      "loss": 0.5122151851654053,
      "mean_token_accuracy": 0.866938441246748,
      "num_tokens": 54855534.0,
      "step": 26830
    },
    {
      "entropy": 0.5127610915806144,
      "epoch": 0.43591596761488677,
      "grad_norm": 6.125,
      "learning_rate": 3.071515076910647e-05,
      "loss": 0.5024378776550293,
      "mean_token_accuracy": 0.8693154700100422,
      "num_tokens": 54875642.0,
      "step": 26840
    },
    {
      "entropy": 0.5013603906612843,
      "epoch": 0.4360783804195123,
      "grad_norm": 10.9375,
      "learning_rate": 3.0702527006021774e-05,
      "loss": 0.5494681835174561,
      "mean_token_accuracy": 0.8603448450565339,
      "num_tokens": 54894335.0,
      "step": 26850
    },
    {
      "entropy": 0.5531791246728972,
      "epoch": 0.4362407932241378,
      "grad_norm": 8.6875,
      "learning_rate": 3.0689901708944515e-05,
      "loss": 0.6018117427825928,
      "mean_token_accuracy": 0.85612678155303,
      "num_tokens": 54915442.0,
      "step": 26860
    },
    {
      "entropy": 0.557512700278312,
      "epoch": 0.4364032060287633,
      "grad_norm": 8.4375,
      "learning_rate": 3.0677274881270894e-05,
      "loss": 0.523082685470581,
      "mean_token_accuracy": 0.8692874372005462,
      "num_tokens": 54935811.0,
      "step": 26870
    },
    {
      "entropy": 0.49158645793795586,
      "epoch": 0.43656561883338885,
      "grad_norm": 7.8125,
      "learning_rate": 3.066464652639758e-05,
      "loss": 0.5066465854644775,
      "mean_token_accuracy": 0.8714259028434753,
      "num_tokens": 54954733.0,
      "step": 26880
    },
    {
      "entropy": 0.5155477747321129,
      "epoch": 0.4367280316380143,
      "grad_norm": 9.3125,
      "learning_rate": 3.0652016647721623e-05,
      "loss": 0.5126651287078857,
      "mean_token_accuracy": 0.8676353961229324,
      "num_tokens": 54975094.0,
      "step": 26890
    },
    {
      "entropy": 0.5137879548594355,
      "epoch": 0.43689044444263986,
      "grad_norm": 5.59375,
      "learning_rate": 3.063938524864048e-05,
      "loss": 0.49982824325561526,
      "mean_token_accuracy": 0.8678163841366768,
      "num_tokens": 54997122.0,
      "step": 26900
    },
    {
      "entropy": 0.5218381367623806,
      "epoch": 0.4370528572472654,
      "grad_norm": 6.375,
      "learning_rate": 3.0626752332552025e-05,
      "loss": 0.568983507156372,
      "mean_token_accuracy": 0.8636690985411406,
      "num_tokens": 55017057.0,
      "step": 26910
    },
    {
      "entropy": 0.5401337065733969,
      "epoch": 0.43721527005189087,
      "grad_norm": 6.84375,
      "learning_rate": 3.061411790285456e-05,
      "loss": 0.5535965919494629,
      "mean_token_accuracy": 0.8644356988370419,
      "num_tokens": 55037012.0,
      "step": 26920
    },
    {
      "entropy": 0.5869622863829136,
      "epoch": 0.4373776828565164,
      "grad_norm": 10.875,
      "learning_rate": 3.060148196294675e-05,
      "loss": 0.5416614532470703,
      "mean_token_accuracy": 0.8606539323925972,
      "num_tokens": 55056798.0,
      "step": 26930
    },
    {
      "entropy": 0.5363588805310429,
      "epoch": 0.43754009566114194,
      "grad_norm": 7.3125,
      "learning_rate": 3.0588844516227714e-05,
      "loss": 0.49663877487182617,
      "mean_token_accuracy": 0.8696520425379276,
      "num_tokens": 55077220.0,
      "step": 26940
    },
    {
      "entropy": 0.5014222712256015,
      "epoch": 0.4377025084657674,
      "grad_norm": 5.71875,
      "learning_rate": 3.057620556609694e-05,
      "loss": 0.5350190162658691,
      "mean_token_accuracy": 0.8645533107221126,
      "num_tokens": 55097829.0,
      "step": 26950
    },
    {
      "entropy": 0.5029997228644788,
      "epoch": 0.43786492127039295,
      "grad_norm": 11.875,
      "learning_rate": 3.056356511595433e-05,
      "loss": 0.5386734008789062,
      "mean_token_accuracy": 0.869632513821125,
      "num_tokens": 55117005.0,
      "step": 26960
    },
    {
      "entropy": 0.575042939465493,
      "epoch": 0.4380273340750185,
      "grad_norm": 9.3125,
      "learning_rate": 3.0550923169200214e-05,
      "loss": 0.595947265625,
      "mean_token_accuracy": 0.8545689858496189,
      "num_tokens": 55137849.0,
      "step": 26970
    },
    {
      "entropy": 0.6491784100653604,
      "epoch": 0.438189746879644,
      "grad_norm": 8.625,
      "learning_rate": 3.053827972923528e-05,
      "loss": 0.6883252143859864,
      "mean_token_accuracy": 0.8324871927499771,
      "num_tokens": 55159729.0,
      "step": 26980
    },
    {
      "entropy": 0.5628588520921767,
      "epoch": 0.4383521596842695,
      "grad_norm": 6.65625,
      "learning_rate": 3.0525634799460654e-05,
      "loss": 0.5863224506378174,
      "mean_token_accuracy": 0.8498340670019389,
      "num_tokens": 55179567.0,
      "step": 26990
    },
    {
      "entropy": 0.5358456050977111,
      "epoch": 0.438514572488895,
      "grad_norm": 8.3125,
      "learning_rate": 3.0512988383277853e-05,
      "loss": 0.5146778106689454,
      "mean_token_accuracy": 0.8667303539812565,
      "num_tokens": 55199502.0,
      "step": 27000
    },
    {
      "entropy": 0.5640515458770097,
      "epoch": 0.43867698529352056,
      "grad_norm": 8.75,
      "learning_rate": 3.0500340484088778e-05,
      "loss": 0.5817858695983886,
      "mean_token_accuracy": 0.8544801115989685,
      "num_tokens": 55219042.0,
      "step": 27010
    },
    {
      "entropy": 0.5506502404343336,
      "epoch": 0.43883939809814604,
      "grad_norm": 7.21875,
      "learning_rate": 3.0487691105295758e-05,
      "loss": 0.548121976852417,
      "mean_token_accuracy": 0.850363128632307,
      "num_tokens": 55238104.0,
      "step": 27020
    },
    {
      "entropy": 0.4861097265034914,
      "epoch": 0.43900181090277157,
      "grad_norm": 7.09375,
      "learning_rate": 3.0475040250301497e-05,
      "loss": 0.4275508880615234,
      "mean_token_accuracy": 0.8820276372134686,
      "num_tokens": 55258317.0,
      "step": 27030
    },
    {
      "entropy": 0.481755858566612,
      "epoch": 0.4391642237073971,
      "grad_norm": 8.4375,
      "learning_rate": 3.0462387922509105e-05,
      "loss": 0.45993905067443847,
      "mean_token_accuracy": 0.8868843279778957,
      "num_tokens": 55278323.0,
      "step": 27040
    },
    {
      "entropy": 0.4703269128454849,
      "epoch": 0.4393266365120226,
      "grad_norm": 4.71875,
      "learning_rate": 3.044973412532209e-05,
      "loss": 0.4828486442565918,
      "mean_token_accuracy": 0.8776221714913846,
      "num_tokens": 55297420.0,
      "step": 27050
    },
    {
      "entropy": 0.5157658666837961,
      "epoch": 0.4394890493166481,
      "grad_norm": 7.875,
      "learning_rate": 3.043707886214434e-05,
      "loss": 0.510365629196167,
      "mean_token_accuracy": 0.8775257594883442,
      "num_tokens": 55318038.0,
      "step": 27060
    },
    {
      "entropy": 0.5182433150708675,
      "epoch": 0.43965146212127365,
      "grad_norm": 10.0625,
      "learning_rate": 3.0424422136380158e-05,
      "loss": 0.5361198902130127,
      "mean_token_accuracy": 0.8708617962896824,
      "num_tokens": 55339045.0,
      "step": 27070
    },
    {
      "entropy": 0.47627518344670533,
      "epoch": 0.4398138749258992,
      "grad_norm": 9.1875,
      "learning_rate": 3.0411763951434242e-05,
      "loss": 0.46596250534057615,
      "mean_token_accuracy": 0.881960965692997,
      "num_tokens": 55359496.0,
      "step": 27080
    },
    {
      "entropy": 0.48303752364590763,
      "epoch": 0.43997628773052466,
      "grad_norm": 7.40625,
      "learning_rate": 3.039910431071165e-05,
      "loss": 0.5369595527648926,
      "mean_token_accuracy": 0.8641105860471725,
      "num_tokens": 55380095.0,
      "step": 27090
    },
    {
      "entropy": 0.5046060610096902,
      "epoch": 0.4401387005351502,
      "grad_norm": 7.0,
      "learning_rate": 3.0386443217617867e-05,
      "loss": 0.5194183826446533,
      "mean_token_accuracy": 0.8675149701535702,
      "num_tokens": 55400296.0,
      "step": 27100
    },
    {
      "entropy": 0.5057170979678631,
      "epoch": 0.44030111333977573,
      "grad_norm": 8.5,
      "learning_rate": 3.037378067555875e-05,
      "loss": 0.5203274726867676,
      "mean_token_accuracy": 0.8720692843198776,
      "num_tokens": 55421395.0,
      "step": 27110
    },
    {
      "entropy": 0.6131428938359023,
      "epoch": 0.4404635261444012,
      "grad_norm": 6.375,
      "learning_rate": 3.0361116687940556e-05,
      "loss": 0.6149850368499756,
      "mean_token_accuracy": 0.8417527608573436,
      "num_tokens": 55442167.0,
      "step": 27120
    },
    {
      "entropy": 0.5441217758692801,
      "epoch": 0.44062593894902674,
      "grad_norm": 9.3125,
      "learning_rate": 3.0348451258169925e-05,
      "loss": 0.5217996120452881,
      "mean_token_accuracy": 0.8671445786952973,
      "num_tokens": 55462832.0,
      "step": 27130
    },
    {
      "entropy": 0.4976544429548085,
      "epoch": 0.4407883517536523,
      "grad_norm": 7.59375,
      "learning_rate": 3.0335784389653887e-05,
      "loss": 0.4890916347503662,
      "mean_token_accuracy": 0.8734675623476506,
      "num_tokens": 55483892.0,
      "step": 27140
    },
    {
      "entropy": 0.5568031307775527,
      "epoch": 0.44095076455827775,
      "grad_norm": 6.21875,
      "learning_rate": 3.032311608579986e-05,
      "loss": 0.5358655452728271,
      "mean_token_accuracy": 0.8669014245271682,
      "num_tokens": 55504320.0,
      "step": 27150
    },
    {
      "entropy": 0.49871573350392284,
      "epoch": 0.4411131773629033,
      "grad_norm": 6.21875,
      "learning_rate": 3.0310446350015636e-05,
      "loss": 0.49774761199951173,
      "mean_token_accuracy": 0.8651471607387066,
      "num_tokens": 55526353.0,
      "step": 27160
    },
    {
      "entropy": 0.5685152924619615,
      "epoch": 0.4412755901675288,
      "grad_norm": 9.1875,
      "learning_rate": 3.0297775185709414e-05,
      "loss": 0.5961421012878418,
      "mean_token_accuracy": 0.8542404055595398,
      "num_tokens": 55547541.0,
      "step": 27170
    },
    {
      "entropy": 0.49678133903071287,
      "epoch": 0.4414380029721543,
      "grad_norm": 4.8125,
      "learning_rate": 3.028510259628976e-05,
      "loss": 0.4837193012237549,
      "mean_token_accuracy": 0.8756991527974606,
      "num_tokens": 55568647.0,
      "step": 27180
    },
    {
      "entropy": 0.5176765469834208,
      "epoch": 0.44160041577677983,
      "grad_norm": 5.21875,
      "learning_rate": 3.027242858516563e-05,
      "loss": 0.539128303527832,
      "mean_token_accuracy": 0.8731474369764328,
      "num_tokens": 55588744.0,
      "step": 27190
    },
    {
      "entropy": 0.550701069040224,
      "epoch": 0.44176282858140536,
      "grad_norm": 6.71875,
      "learning_rate": 3.0259753155746363e-05,
      "loss": 0.5892716884613037,
      "mean_token_accuracy": 0.8614097677171231,
      "num_tokens": 55608248.0,
      "step": 27200
    },
    {
      "entropy": 0.5029158903285861,
      "epoch": 0.4419252413860309,
      "grad_norm": 5.4375,
      "learning_rate": 3.024707631144167e-05,
      "loss": 0.5024892330169678,
      "mean_token_accuracy": 0.8727006711065769,
      "num_tokens": 55628516.0,
      "step": 27210
    },
    {
      "entropy": 0.4638066618703306,
      "epoch": 0.4420876541906564,
      "grad_norm": 8.4375,
      "learning_rate": 3.0234398055661667e-05,
      "loss": 0.48854885101318357,
      "mean_token_accuracy": 0.873700724542141,
      "num_tokens": 55649273.0,
      "step": 27220
    },
    {
      "entropy": 0.47275223713368175,
      "epoch": 0.4422500669952819,
      "grad_norm": 9.3125,
      "learning_rate": 3.0221718391816816e-05,
      "loss": 0.4710687160491943,
      "mean_token_accuracy": 0.8771762073040008,
      "num_tokens": 55669210.0,
      "step": 27230
    },
    {
      "entropy": 0.5611609892453998,
      "epoch": 0.44241247979990744,
      "grad_norm": 9.125,
      "learning_rate": 3.020903732331799e-05,
      "loss": 0.5483808517456055,
      "mean_token_accuracy": 0.8577362351119519,
      "num_tokens": 55690165.0,
      "step": 27240
    },
    {
      "entropy": 0.5019849573262036,
      "epoch": 0.4425748926045329,
      "grad_norm": 11.75,
      "learning_rate": 3.0196354853576404e-05,
      "loss": 0.47623019218444823,
      "mean_token_accuracy": 0.8878120966255665,
      "num_tokens": 55709978.0,
      "step": 27250
    },
    {
      "entropy": 0.5766485585831106,
      "epoch": 0.44273730540915845,
      "grad_norm": 9.0625,
      "learning_rate": 3.01836709860037e-05,
      "loss": 0.56859130859375,
      "mean_token_accuracy": 0.8532906875014306,
      "num_tokens": 55731107.0,
      "step": 27260
    },
    {
      "entropy": 0.46530290339142083,
      "epoch": 0.442899718213784,
      "grad_norm": 11.0,
      "learning_rate": 3.0170985724011842e-05,
      "loss": 0.4409780979156494,
      "mean_token_accuracy": 0.8799609616398811,
      "num_tokens": 55751181.0,
      "step": 27270
    },
    {
      "entropy": 0.5087373574264348,
      "epoch": 0.44306213101840947,
      "grad_norm": 6.40625,
      "learning_rate": 3.0158299071013202e-05,
      "loss": 0.49903407096862795,
      "mean_token_accuracy": 0.8628100242465735,
      "num_tokens": 55770590.0,
      "step": 27280
    },
    {
      "entropy": 0.5108993250876666,
      "epoch": 0.443224543823035,
      "grad_norm": 6.40625,
      "learning_rate": 3.0145611030420522e-05,
      "loss": 0.5626221656799316,
      "mean_token_accuracy": 0.8675781033933163,
      "num_tokens": 55791618.0,
      "step": 27290
    },
    {
      "entropy": 0.4797052348963916,
      "epoch": 0.44338695662766053,
      "grad_norm": 7.4375,
      "learning_rate": 3.0132921605646912e-05,
      "loss": 0.5313747882843017,
      "mean_token_accuracy": 0.8695507697761059,
      "num_tokens": 55811544.0,
      "step": 27300
    },
    {
      "entropy": 0.5900356122292578,
      "epoch": 0.44354936943228607,
      "grad_norm": 8.4375,
      "learning_rate": 3.0120230800105857e-05,
      "loss": 0.6271039485931397,
      "mean_token_accuracy": 0.857377254217863,
      "num_tokens": 55832286.0,
      "step": 27310
    },
    {
      "entropy": 0.49382907184772196,
      "epoch": 0.44371178223691154,
      "grad_norm": 7.375,
      "learning_rate": 3.01075386172112e-05,
      "loss": 0.5063539981842041,
      "mean_token_accuracy": 0.8756177663803101,
      "num_tokens": 55852017.0,
      "step": 27320
    },
    {
      "entropy": 0.5082624503411353,
      "epoch": 0.4438741950415371,
      "grad_norm": 9.0,
      "learning_rate": 3.009484506037719e-05,
      "loss": 0.5057880401611328,
      "mean_token_accuracy": 0.8652433432638645,
      "num_tokens": 55873681.0,
      "step": 27330
    },
    {
      "entropy": 0.5111311489716173,
      "epoch": 0.4440366078461626,
      "grad_norm": 6.03125,
      "learning_rate": 3.00821501330184e-05,
      "loss": 0.5077598094940186,
      "mean_token_accuracy": 0.8754937760531902,
      "num_tokens": 55894903.0,
      "step": 27340
    },
    {
      "entropy": 0.4634359278716147,
      "epoch": 0.4441990206507881,
      "grad_norm": 9.0,
      "learning_rate": 3.0069453838549806e-05,
      "loss": 0.47876534461975095,
      "mean_token_accuracy": 0.8788744933903218,
      "num_tokens": 55915418.0,
      "step": 27350
    },
    {
      "entropy": 0.4893330675549805,
      "epoch": 0.4443614334554136,
      "grad_norm": 7.34375,
      "learning_rate": 3.005675618038673e-05,
      "loss": 0.44977564811706544,
      "mean_token_accuracy": 0.87604189068079,
      "num_tokens": 55935275.0,
      "step": 27360
    },
    {
      "entropy": 0.5168673981912434,
      "epoch": 0.44452384626003916,
      "grad_norm": 11.1875,
      "learning_rate": 3.004405716194488e-05,
      "loss": 0.5626243114471435,
      "mean_token_accuracy": 0.8583581402897835,
      "num_tokens": 55954690.0,
      "step": 27370
    },
    {
      "entropy": 0.4817249019630253,
      "epoch": 0.44468625906466464,
      "grad_norm": 10.75,
      "learning_rate": 3.003135678664033e-05,
      "loss": 0.529530382156372,
      "mean_token_accuracy": 0.8730773575603962,
      "num_tokens": 55974465.0,
      "step": 27380
    },
    {
      "entropy": 0.5289487045258283,
      "epoch": 0.44484867186929017,
      "grad_norm": 5.5,
      "learning_rate": 3.0018655057889482e-05,
      "loss": 0.5650976181030274,
      "mean_token_accuracy": 0.8635201297700406,
      "num_tokens": 55994092.0,
      "step": 27390
    },
    {
      "entropy": 0.5725059876218438,
      "epoch": 0.4450110846739157,
      "grad_norm": 8.5625,
      "learning_rate": 3.0005951979109155e-05,
      "loss": 0.6047881126403809,
      "mean_token_accuracy": 0.8567812725901603,
      "num_tokens": 56014419.0,
      "step": 27400
    },
    {
      "entropy": 0.5280761973001062,
      "epoch": 0.44517349747854124,
      "grad_norm": 7.28125,
      "learning_rate": 2.9993247553716484e-05,
      "loss": 0.508406114578247,
      "mean_token_accuracy": 0.8674389019608497,
      "num_tokens": 56035522.0,
      "step": 27410
    },
    {
      "entropy": 0.5342308286111802,
      "epoch": 0.4453359102831667,
      "grad_norm": 5.90625,
      "learning_rate": 2.9980541785129012e-05,
      "loss": 0.5391977310180665,
      "mean_token_accuracy": 0.8635302193462848,
      "num_tokens": 56055465.0,
      "step": 27420
    },
    {
      "entropy": 0.487649381486699,
      "epoch": 0.44549832308779225,
      "grad_norm": 7.4375,
      "learning_rate": 2.99678346767646e-05,
      "loss": 0.4839106559753418,
      "mean_token_accuracy": 0.8780412517488003,
      "num_tokens": 56075077.0,
      "step": 27430
    },
    {
      "entropy": 0.47437108277808876,
      "epoch": 0.4456607358924178,
      "grad_norm": 7.5625,
      "learning_rate": 2.9955126232041496e-05,
      "loss": 0.46820816993713377,
      "mean_token_accuracy": 0.8784167058765888,
      "num_tokens": 56095576.0,
      "step": 27440
    },
    {
      "entropy": 0.5081771889701485,
      "epoch": 0.44582314869704326,
      "grad_norm": 9.1875,
      "learning_rate": 2.9942416454378296e-05,
      "loss": 0.5080283164978028,
      "mean_token_accuracy": 0.8738152541220188,
      "num_tokens": 56114594.0,
      "step": 27450
    },
    {
      "entropy": 0.5496200503315777,
      "epoch": 0.4459855615016688,
      "grad_norm": 6.0,
      "learning_rate": 2.9929705347193966e-05,
      "loss": 0.5539786338806152,
      "mean_token_accuracy": 0.8692952737212181,
      "num_tokens": 56135148.0,
      "step": 27460
    },
    {
      "entropy": 0.45578996525146065,
      "epoch": 0.4461479743062943,
      "grad_norm": 5.34375,
      "learning_rate": 2.9916992913907814e-05,
      "loss": 0.47879323959350584,
      "mean_token_accuracy": 0.8847390830516815,
      "num_tokens": 56154659.0,
      "step": 27470
    },
    {
      "entropy": 0.5028160045389086,
      "epoch": 0.4463103871109198,
      "grad_norm": 8.6875,
      "learning_rate": 2.9904279157939525e-05,
      "loss": 0.5629745483398437,
      "mean_token_accuracy": 0.860260084271431,
      "num_tokens": 56175762.0,
      "step": 27480
    },
    {
      "entropy": 0.42568334874231367,
      "epoch": 0.44647279991554534,
      "grad_norm": 8.9375,
      "learning_rate": 2.9891564082709123e-05,
      "loss": 0.4326143264770508,
      "mean_token_accuracy": 0.8924735195934772,
      "num_tokens": 56195953.0,
      "step": 27490
    },
    {
      "entropy": 0.4986663500778377,
      "epoch": 0.44663521272017087,
      "grad_norm": 5.15625,
      "learning_rate": 2.9878847691636985e-05,
      "loss": 0.4687326908111572,
      "mean_token_accuracy": 0.8736995488405228,
      "num_tokens": 56216604.0,
      "step": 27500
    },
    {
      "entropy": 0.47569279102608564,
      "epoch": 0.44679762552479635,
      "grad_norm": 6.8125,
      "learning_rate": 2.9866129988143853e-05,
      "loss": 0.4851207733154297,
      "mean_token_accuracy": 0.8794075772166252,
      "num_tokens": 56237849.0,
      "step": 27510
    },
    {
      "entropy": 0.48038350851275025,
      "epoch": 0.4469600383294219,
      "grad_norm": 7.375,
      "learning_rate": 2.985341097565082e-05,
      "loss": 0.4679347038269043,
      "mean_token_accuracy": 0.8777019068598747,
      "num_tokens": 56258647.0,
      "step": 27520
    },
    {
      "entropy": 0.5460156882880256,
      "epoch": 0.4471224511340474,
      "grad_norm": 7.34375,
      "learning_rate": 2.9840690657579324e-05,
      "loss": 0.4965061187744141,
      "mean_token_accuracy": 0.8728871699422598,
      "num_tokens": 56279234.0,
      "step": 27530
    },
    {
      "entropy": 0.5350760090630502,
      "epoch": 0.44728486393867295,
      "grad_norm": 9.4375,
      "learning_rate": 2.9827969037351166e-05,
      "loss": 0.5187799453735351,
      "mean_token_accuracy": 0.8699410758912564,
      "num_tokens": 56300189.0,
      "step": 27540
    },
    {
      "entropy": 0.5151982740266249,
      "epoch": 0.4474472767432984,
      "grad_norm": 5.71875,
      "learning_rate": 2.981524611838848e-05,
      "loss": 0.45757360458374025,
      "mean_token_accuracy": 0.8829277276992797,
      "num_tokens": 56321698.0,
      "step": 27550
    },
    {
      "entropy": 0.4480190242640674,
      "epoch": 0.44760968954792396,
      "grad_norm": 9.1875,
      "learning_rate": 2.9802521904113778e-05,
      "loss": 0.5054731845855713,
      "mean_token_accuracy": 0.8751645080745221,
      "num_tokens": 56342065.0,
      "step": 27560
    },
    {
      "entropy": 0.5197412563953548,
      "epoch": 0.4477721023525495,
      "grad_norm": 9.5,
      "learning_rate": 2.978979639794988e-05,
      "loss": 0.561060094833374,
      "mean_token_accuracy": 0.8642024844884872,
      "num_tokens": 56362415.0,
      "step": 27570
    },
    {
      "entropy": 0.46230391040444374,
      "epoch": 0.447934515157175,
      "grad_norm": 8.375,
      "learning_rate": 2.9777069603319994e-05,
      "loss": 0.49129185676574705,
      "mean_token_accuracy": 0.8724095463752747,
      "num_tokens": 56382271.0,
      "step": 27580
    },
    {
      "entropy": 0.5471215604804456,
      "epoch": 0.4480969279618005,
      "grad_norm": 7.1875,
      "learning_rate": 2.9764341523647645e-05,
      "loss": 0.6013251304626465,
      "mean_token_accuracy": 0.8551538266241551,
      "num_tokens": 56402687.0,
      "step": 27590
    },
    {
      "entropy": 0.5851567406672984,
      "epoch": 0.44825934076642604,
      "grad_norm": 8.75,
      "learning_rate": 2.9751612162356725e-05,
      "loss": 0.5470786094665527,
      "mean_token_accuracy": 0.8647656932473182,
      "num_tokens": 56423274.0,
      "step": 27600
    },
    {
      "entropy": 0.5084506246494129,
      "epoch": 0.4484217535710515,
      "grad_norm": 8.3125,
      "learning_rate": 2.973888152287144e-05,
      "loss": 0.5340044975280762,
      "mean_token_accuracy": 0.8664151422679425,
      "num_tokens": 56444441.0,
      "step": 27610
    },
    {
      "entropy": 0.5077030701562762,
      "epoch": 0.44858416637567705,
      "grad_norm": 8.0625,
      "learning_rate": 2.972614960861639e-05,
      "loss": 0.5306628227233887,
      "mean_token_accuracy": 0.8696650039404631,
      "num_tokens": 56463507.0,
      "step": 27620
    },
    {
      "entropy": 0.5160880568902939,
      "epoch": 0.4487465791803026,
      "grad_norm": 6.65625,
      "learning_rate": 2.9713416423016466e-05,
      "loss": 0.4837037086486816,
      "mean_token_accuracy": 0.8810078144073487,
      "num_tokens": 56482722.0,
      "step": 27630
    },
    {
      "entropy": 0.5227928462438285,
      "epoch": 0.4489089919849281,
      "grad_norm": 8.9375,
      "learning_rate": 2.9700681969496937e-05,
      "loss": 0.5156248092651368,
      "mean_token_accuracy": 0.8673313677310943,
      "num_tokens": 56504096.0,
      "step": 27640
    },
    {
      "entropy": 0.5935252169147134,
      "epoch": 0.4490714047895536,
      "grad_norm": 6.96875,
      "learning_rate": 2.9687946251483394e-05,
      "loss": 0.5783194065093994,
      "mean_token_accuracy": 0.8519051276147366,
      "num_tokens": 56524768.0,
      "step": 27650
    },
    {
      "entropy": 0.6159563443623484,
      "epoch": 0.44923381759417913,
      "grad_norm": 6.875,
      "learning_rate": 2.9675209272401765e-05,
      "loss": 0.6225059986114502,
      "mean_token_accuracy": 0.8387514986097813,
      "num_tokens": 56544927.0,
      "step": 27660
    },
    {
      "entropy": 0.5161966452840716,
      "epoch": 0.44939623039880466,
      "grad_norm": 6.0625,
      "learning_rate": 2.9662471035678346e-05,
      "loss": 0.5211599349975586,
      "mean_token_accuracy": 0.861641599982977,
      "num_tokens": 56565486.0,
      "step": 27670
    },
    {
      "entropy": 0.4803266603965312,
      "epoch": 0.44955864320343014,
      "grad_norm": 5.40625,
      "learning_rate": 2.9649731544739733e-05,
      "loss": 0.49917969703674314,
      "mean_token_accuracy": 0.8728828027844429,
      "num_tokens": 56586134.0,
      "step": 27680
    },
    {
      "entropy": 0.47412106483243405,
      "epoch": 0.4497210560080557,
      "grad_norm": 9.1875,
      "learning_rate": 2.9636990803012893e-05,
      "loss": 0.4657122611999512,
      "mean_token_accuracy": 0.8911771476268768,
      "num_tokens": 56606440.0,
      "step": 27690
    },
    {
      "entropy": 0.5424385031219572,
      "epoch": 0.4498834688126812,
      "grad_norm": 6.8125,
      "learning_rate": 2.9624248813925094e-05,
      "loss": 0.567103385925293,
      "mean_token_accuracy": 0.8623527109622955,
      "num_tokens": 56626049.0,
      "step": 27700
    },
    {
      "entropy": 0.4593048396985978,
      "epoch": 0.4500458816173067,
      "grad_norm": 9.75,
      "learning_rate": 2.9611505580903977e-05,
      "loss": 0.47066497802734375,
      "mean_token_accuracy": 0.8742348283529282,
      "num_tokens": 56645533.0,
      "step": 27710
    },
    {
      "entropy": 0.48223865386098624,
      "epoch": 0.4502082944219322,
      "grad_norm": 7.9375,
      "learning_rate": 2.9598761107377498e-05,
      "loss": 0.47821626663208006,
      "mean_token_accuracy": 0.881100806593895,
      "num_tokens": 56665584.0,
      "step": 27720
    },
    {
      "entropy": 0.5023659316357225,
      "epoch": 0.45037070722655775,
      "grad_norm": 7.09375,
      "learning_rate": 2.958601539677394e-05,
      "loss": 0.48133211135864257,
      "mean_token_accuracy": 0.8764194183051586,
      "num_tokens": 56687023.0,
      "step": 27730
    },
    {
      "entropy": 0.5197680281009525,
      "epoch": 0.45053312003118323,
      "grad_norm": 9.4375,
      "learning_rate": 2.957326845252194e-05,
      "loss": 0.5338258266448974,
      "mean_token_accuracy": 0.8639291673898697,
      "num_tokens": 56707177.0,
      "step": 27740
    },
    {
      "entropy": 0.5231988635845483,
      "epoch": 0.45069553283580877,
      "grad_norm": 8.8125,
      "learning_rate": 2.9560520278050448e-05,
      "loss": 0.5094373226165771,
      "mean_token_accuracy": 0.8687764957547188,
      "num_tokens": 56728713.0,
      "step": 27750
    },
    {
      "entropy": 0.47884522438980637,
      "epoch": 0.4508579456404343,
      "grad_norm": 8.125,
      "learning_rate": 2.954777087678875e-05,
      "loss": 0.4620362281799316,
      "mean_token_accuracy": 0.8785881973803044,
      "num_tokens": 56748557.0,
      "step": 27760
    },
    {
      "entropy": 0.49885212825611236,
      "epoch": 0.45102035844505983,
      "grad_norm": 9.625,
      "learning_rate": 2.953502025216646e-05,
      "loss": 0.503712511062622,
      "mean_token_accuracy": 0.8713699534535408,
      "num_tokens": 56769356.0,
      "step": 27770
    },
    {
      "entropy": 0.46684436984360217,
      "epoch": 0.4511827712496853,
      "grad_norm": 9.9375,
      "learning_rate": 2.952226840761354e-05,
      "loss": 0.4527806282043457,
      "mean_token_accuracy": 0.8795828327536583,
      "num_tokens": 56789479.0,
      "step": 27780
    },
    {
      "entropy": 0.48974854149855673,
      "epoch": 0.45134518405431084,
      "grad_norm": 7.0,
      "learning_rate": 2.950951534656025e-05,
      "loss": 0.46338319778442383,
      "mean_token_accuracy": 0.8781877338886261,
      "num_tokens": 56810659.0,
      "step": 27790
    },
    {
      "entropy": 0.49248120239935816,
      "epoch": 0.4515075968589364,
      "grad_norm": 8.0625,
      "learning_rate": 2.9496761072437202e-05,
      "loss": 0.4911224842071533,
      "mean_token_accuracy": 0.8764248505234719,
      "num_tokens": 56831701.0,
      "step": 27800
    },
    {
      "entropy": 0.49278630600310863,
      "epoch": 0.45167000966356186,
      "grad_norm": 7.03125,
      "learning_rate": 2.9484005588675313e-05,
      "loss": 0.47745490074157715,
      "mean_token_accuracy": 0.8704216599464416,
      "num_tokens": 56850634.0,
      "step": 27810
    },
    {
      "entropy": 0.47929648105055095,
      "epoch": 0.4518324224681874,
      "grad_norm": 7.40625,
      "learning_rate": 2.947124889870585e-05,
      "loss": 0.4843841552734375,
      "mean_token_accuracy": 0.8734408773481845,
      "num_tokens": 56872522.0,
      "step": 27820
    },
    {
      "entropy": 0.47668720008805393,
      "epoch": 0.4519948352728129,
      "grad_norm": 10.375,
      "learning_rate": 2.9458491005960386e-05,
      "loss": 0.5225075244903564,
      "mean_token_accuracy": 0.8707329206168651,
      "num_tokens": 56892175.0,
      "step": 27830
    },
    {
      "entropy": 0.49260620479471984,
      "epoch": 0.4521572480774384,
      "grad_norm": 5.75,
      "learning_rate": 2.9445731913870816e-05,
      "loss": 0.4771301746368408,
      "mean_token_accuracy": 0.8791799023747444,
      "num_tokens": 56912585.0,
      "step": 27840
    },
    {
      "entropy": 0.4822661988670006,
      "epoch": 0.45231966088206393,
      "grad_norm": 10.0625,
      "learning_rate": 2.9432971625869367e-05,
      "loss": 0.5201068878173828,
      "mean_token_accuracy": 0.8631275564432144,
      "num_tokens": 56933266.0,
      "step": 27850
    },
    {
      "entropy": 0.51104031347204,
      "epoch": 0.45248207368668947,
      "grad_norm": 12.1875,
      "learning_rate": 2.9420210145388595e-05,
      "loss": 0.5032454490661621,
      "mean_token_accuracy": 0.8668810598552227,
      "num_tokens": 56953909.0,
      "step": 27860
    },
    {
      "entropy": 0.6333530560135842,
      "epoch": 0.452644486491315,
      "grad_norm": 11.5,
      "learning_rate": 2.9407447475861356e-05,
      "loss": 0.6678153991699218,
      "mean_token_accuracy": 0.8415273122489453,
      "num_tokens": 56974936.0,
      "step": 27870
    },
    {
      "entropy": 0.533442331571132,
      "epoch": 0.4528068992959405,
      "grad_norm": 8.625,
      "learning_rate": 2.9394683620720846e-05,
      "loss": 0.5945525646209717,
      "mean_token_accuracy": 0.8582693263888359,
      "num_tokens": 56995058.0,
      "step": 27880
    },
    {
      "entropy": 0.46883983693551273,
      "epoch": 0.452969312100566,
      "grad_norm": 8.4375,
      "learning_rate": 2.9381918583400553e-05,
      "loss": 0.4554112434387207,
      "mean_token_accuracy": 0.8908465370535851,
      "num_tokens": 57015191.0,
      "step": 27890
    },
    {
      "entropy": 0.577692665765062,
      "epoch": 0.45313172490519155,
      "grad_norm": 10.5625,
      "learning_rate": 2.936915236733433e-05,
      "loss": 0.549635362625122,
      "mean_token_accuracy": 0.8569009244441986,
      "num_tokens": 57034717.0,
      "step": 27900
    },
    {
      "entropy": 0.4780918015167117,
      "epoch": 0.453294137709817,
      "grad_norm": 9.0625,
      "learning_rate": 2.935638497595629e-05,
      "loss": 0.5154565334320068,
      "mean_token_accuracy": 0.8733496107161045,
      "num_tokens": 57055057.0,
      "step": 27910
    },
    {
      "entropy": 0.5785620251204818,
      "epoch": 0.45345655051444256,
      "grad_norm": 10.125,
      "learning_rate": 2.9343616412700908e-05,
      "loss": 0.5828054428100586,
      "mean_token_accuracy": 0.8555382870137691,
      "num_tokens": 57076239.0,
      "step": 27920
    },
    {
      "entropy": 0.5751172923715785,
      "epoch": 0.4536189633190681,
      "grad_norm": 8.6875,
      "learning_rate": 2.9330846681002942e-05,
      "loss": 0.6005470752716064,
      "mean_token_accuracy": 0.8553027667105197,
      "num_tokens": 57098929.0,
      "step": 27930
    },
    {
      "entropy": 0.5203103151172399,
      "epoch": 0.45378137612369357,
      "grad_norm": 8.625,
      "learning_rate": 2.9318075784297494e-05,
      "loss": 0.5234109401702881,
      "mean_token_accuracy": 0.8663516607135534,
      "num_tokens": 57120099.0,
      "step": 27940
    },
    {
      "entropy": 0.5321633242070675,
      "epoch": 0.4539437889283191,
      "grad_norm": 11.0625,
      "learning_rate": 2.930530372601995e-05,
      "loss": 0.5512485504150391,
      "mean_token_accuracy": 0.8660279542207718,
      "num_tokens": 57139795.0,
      "step": 27950
    },
    {
      "entropy": 0.5003785068169236,
      "epoch": 0.45410620173294464,
      "grad_norm": 8.0625,
      "learning_rate": 2.9292530509606033e-05,
      "loss": 0.4947153091430664,
      "mean_token_accuracy": 0.8785861395299435,
      "num_tokens": 57160397.0,
      "step": 27960
    },
    {
      "entropy": 0.4445462642703205,
      "epoch": 0.4542686145375701,
      "grad_norm": 9.4375,
      "learning_rate": 2.9279756138491754e-05,
      "loss": 0.4654655456542969,
      "mean_token_accuracy": 0.8850097946822644,
      "num_tokens": 57181727.0,
      "step": 27970
    },
    {
      "entropy": 0.5683333247900009,
      "epoch": 0.45443102734219565,
      "grad_norm": 10.25,
      "learning_rate": 2.926698061611346e-05,
      "loss": 0.5688388824462891,
      "mean_token_accuracy": 0.860228106752038,
      "num_tokens": 57202140.0,
      "step": 27980
    },
    {
      "entropy": 0.48379323692061005,
      "epoch": 0.4545934401468212,
      "grad_norm": 8.5,
      "learning_rate": 2.9254203945907794e-05,
      "loss": 0.47608389854431155,
      "mean_token_accuracy": 0.879845168441534,
      "num_tokens": 57223058.0,
      "step": 27990
    },
    {
      "epoch": 0.4547558529514467,
      "eval_entropy": 0.5441888830661774,
      "eval_loss": 0.5518063902854919,
      "eval_mean_token_accuracy": 0.8600729928016663,
      "eval_num_tokens": 57242778.0,
      "eval_runtime": 40.2499,
      "eval_samples_per_second": 49.69,
      "eval_steps_per_second": 6.211,
      "step": 28000
    },
    {
      "entropy": 0.48952402198920025,
      "epoch": 0.4549182657560722,
      "grad_norm": 7.28125,
      "learning_rate": 2.9228647175762445e-05,
      "loss": 0.5143888473510743,
      "mean_token_accuracy": 0.867078723013401,
      "num_tokens": 57263299.0,
      "step": 28010
    },
    {
      "entropy": 0.5396759864408522,
      "epoch": 0.4550806785606977,
      "grad_norm": 6.15625,
      "learning_rate": 2.92158670826976e-05,
      "loss": 0.5393098831176758,
      "mean_token_accuracy": 0.8725659303367138,
      "num_tokens": 57284598.0,
      "step": 28020
    },
    {
      "entropy": 0.517999223107472,
      "epoch": 0.45524309136532326,
      "grad_norm": 8.8125,
      "learning_rate": 2.920308585555504e-05,
      "loss": 0.5075906276702881,
      "mean_token_accuracy": 0.8702138021588326,
      "num_tokens": 57305160.0,
      "step": 28030
    },
    {
      "entropy": 0.5549280947539955,
      "epoch": 0.45540550416994874,
      "grad_norm": 6.15625,
      "learning_rate": 2.919030349777293e-05,
      "loss": 0.5839489936828614,
      "mean_token_accuracy": 0.8565610155463219,
      "num_tokens": 57325441.0,
      "step": 28040
    },
    {
      "entropy": 0.5128659505862743,
      "epoch": 0.45556791697457427,
      "grad_norm": 6.3125,
      "learning_rate": 2.917752001278976e-05,
      "loss": 0.4966397762298584,
      "mean_token_accuracy": 0.8731828995049,
      "num_tokens": 57345297.0,
      "step": 28050
    },
    {
      "entropy": 0.5385017185704782,
      "epoch": 0.4557303297791998,
      "grad_norm": 8.0,
      "learning_rate": 2.916473540404433e-05,
      "loss": 0.49066753387451173,
      "mean_token_accuracy": 0.8759444318711758,
      "num_tokens": 57366497.0,
      "step": 28060
    },
    {
      "entropy": 0.529580403631553,
      "epoch": 0.4558927425838253,
      "grad_norm": 9.0625,
      "learning_rate": 2.915194967497571e-05,
      "loss": 0.5299291133880615,
      "mean_token_accuracy": 0.8664688996970653,
      "num_tokens": 57386949.0,
      "step": 28070
    },
    {
      "entropy": 0.5750752655323594,
      "epoch": 0.4560551553884508,
      "grad_norm": 6.125,
      "learning_rate": 2.9139162829023292e-05,
      "loss": 0.5896079540252686,
      "mean_token_accuracy": 0.8533597849309444,
      "num_tokens": 57408045.0,
      "step": 28080
    },
    {
      "entropy": 0.4991929487325251,
      "epoch": 0.45621756819307635,
      "grad_norm": 9.125,
      "learning_rate": 2.9126374869626777e-05,
      "loss": 0.5069374561309814,
      "mean_token_accuracy": 0.8679572634398938,
      "num_tokens": 57427807.0,
      "step": 28090
    },
    {
      "entropy": 0.46944609307684004,
      "epoch": 0.4563799809977019,
      "grad_norm": 7.1875,
      "learning_rate": 2.911358580022615e-05,
      "loss": 0.4558931827545166,
      "mean_token_accuracy": 0.8824787326157093,
      "num_tokens": 57447866.0,
      "step": 28100
    },
    {
      "entropy": 0.43078374401666225,
      "epoch": 0.45654239380232736,
      "grad_norm": 8.8125,
      "learning_rate": 2.910079562426169e-05,
      "loss": 0.4022679805755615,
      "mean_token_accuracy": 0.8916513793170452,
      "num_tokens": 57466679.0,
      "step": 28110
    },
    {
      "entropy": 0.5269088610075414,
      "epoch": 0.4567048066069529,
      "grad_norm": 7.46875,
      "learning_rate": 2.9088004345174e-05,
      "loss": 0.5316169261932373,
      "mean_token_accuracy": 0.8533508881926537,
      "num_tokens": 57486343.0,
      "step": 28120
    },
    {
      "entropy": 0.4818964248523116,
      "epoch": 0.45686721941157843,
      "grad_norm": 7.28125,
      "learning_rate": 2.9075211966403948e-05,
      "loss": 0.4827631950378418,
      "mean_token_accuracy": 0.8727773398160934,
      "num_tokens": 57506978.0,
      "step": 28130
    },
    {
      "entropy": 0.4838607113342732,
      "epoch": 0.4570296322162039,
      "grad_norm": 6.59375,
      "learning_rate": 2.9062418491392718e-05,
      "loss": 0.4867546558380127,
      "mean_token_accuracy": 0.8754599928855896,
      "num_tokens": 57526599.0,
      "step": 28140
    },
    {
      "entropy": 0.5435472069308162,
      "epoch": 0.45719204502082944,
      "grad_norm": 11.3125,
      "learning_rate": 2.9049623923581788e-05,
      "loss": 0.6364171981811524,
      "mean_token_accuracy": 0.8509659618139267,
      "num_tokens": 57547188.0,
      "step": 28150
    },
    {
      "entropy": 0.49644737197086214,
      "epoch": 0.457354457825455,
      "grad_norm": 7.3125,
      "learning_rate": 2.903682826641292e-05,
      "loss": 0.515169906616211,
      "mean_token_accuracy": 0.8726398512721062,
      "num_tokens": 57569570.0,
      "step": 28160
    },
    {
      "entropy": 0.48946651970036326,
      "epoch": 0.45751687063008045,
      "grad_norm": 12.75,
      "learning_rate": 2.9024031523328183e-05,
      "loss": 0.4747940540313721,
      "mean_token_accuracy": 0.8815559990704059,
      "num_tokens": 57589610.0,
      "step": 28170
    },
    {
      "entropy": 0.5725207781419158,
      "epoch": 0.457679283434706,
      "grad_norm": 6.5625,
      "learning_rate": 2.9011233697769914e-05,
      "loss": 0.5641712665557861,
      "mean_token_accuracy": 0.8569286536425352,
      "num_tokens": 57609812.0,
      "step": 28180
    },
    {
      "entropy": 0.533456351654604,
      "epoch": 0.4578416962393315,
      "grad_norm": 8.1875,
      "learning_rate": 2.8998434793180767e-05,
      "loss": 0.5626682281494141,
      "mean_token_accuracy": 0.861941971629858,
      "num_tokens": 57629214.0,
      "step": 28190
    },
    {
      "entropy": 0.49125935332849624,
      "epoch": 0.45800410904395705,
      "grad_norm": 12.125,
      "learning_rate": 2.8985634813003672e-05,
      "loss": 0.4998145580291748,
      "mean_token_accuracy": 0.8718176349997521,
      "num_tokens": 57650587.0,
      "step": 28200
    },
    {
      "entropy": 0.5387174333445728,
      "epoch": 0.45816652184858253,
      "grad_norm": 7.40625,
      "learning_rate": 2.8972833760681867e-05,
      "loss": 0.534962797164917,
      "mean_token_accuracy": 0.869414459913969,
      "num_tokens": 57671898.0,
      "step": 28210
    },
    {
      "entropy": 0.4518477004021406,
      "epoch": 0.45832893465320806,
      "grad_norm": 7.03125,
      "learning_rate": 2.896003163965884e-05,
      "loss": 0.4726675510406494,
      "mean_token_accuracy": 0.8794634446501732,
      "num_tokens": 57692822.0,
      "step": 28220
    },
    {
      "entropy": 0.5405137675348669,
      "epoch": 0.4584913474578336,
      "grad_norm": 8.6875,
      "learning_rate": 2.89472284533784e-05,
      "loss": 0.5719991683959961,
      "mean_token_accuracy": 0.858391311392188,
      "num_tokens": 57713030.0,
      "step": 28230
    },
    {
      "entropy": 0.5613314555026591,
      "epoch": 0.4586537602624591,
      "grad_norm": 6.25,
      "learning_rate": 2.8934424205284643e-05,
      "loss": 0.5509742736816406,
      "mean_token_accuracy": 0.866412004083395,
      "num_tokens": 57734798.0,
      "step": 28240
    },
    {
      "entropy": 0.5574503833195195,
      "epoch": 0.4588161730670846,
      "grad_norm": 8.75,
      "learning_rate": 2.8921618898821934e-05,
      "loss": 0.5286017894744873,
      "mean_token_accuracy": 0.8658315978944302,
      "num_tokens": 57756250.0,
      "step": 28250
    },
    {
      "entropy": 0.5151393934618682,
      "epoch": 0.45897858587171014,
      "grad_norm": 10.3125,
      "learning_rate": 2.890881253743492e-05,
      "loss": 0.5381489753723144,
      "mean_token_accuracy": 0.8738049898296595,
      "num_tokens": 57776265.0,
      "step": 28260
    },
    {
      "entropy": 0.46738332798704507,
      "epoch": 0.4591409986763356,
      "grad_norm": 11.5,
      "learning_rate": 2.8896005124568554e-05,
      "loss": 0.4683993339538574,
      "mean_token_accuracy": 0.8755813539028168,
      "num_tokens": 57796790.0,
      "step": 28270
    },
    {
      "entropy": 0.5273630083538592,
      "epoch": 0.45930341148096115,
      "grad_norm": 9.6875,
      "learning_rate": 2.8883196663668062e-05,
      "loss": 0.5507990837097168,
      "mean_token_accuracy": 0.8600769765675068,
      "num_tokens": 57817484.0,
      "step": 28280
    },
    {
      "entropy": 0.4415387250483036,
      "epoch": 0.4594658242855867,
      "grad_norm": 6.03125,
      "learning_rate": 2.887038715817893e-05,
      "loss": 0.4413275718688965,
      "mean_token_accuracy": 0.8893660947680473,
      "num_tokens": 57837740.0,
      "step": 28290
    },
    {
      "entropy": 0.5317445192020387,
      "epoch": 0.45962823709021217,
      "grad_norm": 7.0,
      "learning_rate": 2.8857576611546953e-05,
      "loss": 0.559794807434082,
      "mean_token_accuracy": 0.8656041763722897,
      "num_tokens": 57857361.0,
      "step": 28300
    },
    {
      "entropy": 0.502260118466802,
      "epoch": 0.4597906498948377,
      "grad_norm": 6.4375,
      "learning_rate": 2.88447650272182e-05,
      "loss": 0.5337009429931641,
      "mean_token_accuracy": 0.8711214151233435,
      "num_tokens": 57876673.0,
      "step": 28310
    },
    {
      "entropy": 0.5326050275471061,
      "epoch": 0.45995306269946323,
      "grad_norm": 9.9375,
      "learning_rate": 2.8831952408639017e-05,
      "loss": 0.5522456169128418,
      "mean_token_accuracy": 0.8649732425808907,
      "num_tokens": 57897760.0,
      "step": 28320
    },
    {
      "entropy": 0.512941948743537,
      "epoch": 0.46011547550408877,
      "grad_norm": 11.1875,
      "learning_rate": 2.8819138759256025e-05,
      "loss": 0.5521686553955079,
      "mean_token_accuracy": 0.8665618255734444,
      "num_tokens": 57917713.0,
      "step": 28330
    },
    {
      "entropy": 0.4751192986033857,
      "epoch": 0.46027788830871424,
      "grad_norm": 6.375,
      "learning_rate": 2.8806324082516118e-05,
      "loss": 0.4707766532897949,
      "mean_token_accuracy": 0.8736432760953903,
      "num_tokens": 57938881.0,
      "step": 28340
    },
    {
      "entropy": 0.5120963967172429,
      "epoch": 0.4604403011133398,
      "grad_norm": 9.5,
      "learning_rate": 2.8793508381866485e-05,
      "loss": 0.4795109272003174,
      "mean_token_accuracy": 0.8774783879518508,
      "num_tokens": 57959295.0,
      "step": 28350
    },
    {
      "entropy": 0.5658920949324966,
      "epoch": 0.4606027139179653,
      "grad_norm": 13.6875,
      "learning_rate": 2.8780691660754565e-05,
      "loss": 0.6059384346008301,
      "mean_token_accuracy": 0.8520291686058045,
      "num_tokens": 57979009.0,
      "step": 28360
    },
    {
      "entropy": 0.4981432196684182,
      "epoch": 0.4607651267225908,
      "grad_norm": 8.5625,
      "learning_rate": 2.8767873922628102e-05,
      "loss": 0.43939967155456544,
      "mean_token_accuracy": 0.8817379660904408,
      "num_tokens": 57999339.0,
      "step": 28370
    },
    {
      "entropy": 0.5212194129824639,
      "epoch": 0.4609275395272163,
      "grad_norm": 6.0625,
      "learning_rate": 2.8755055170935076e-05,
      "loss": 0.5792467594146729,
      "mean_token_accuracy": 0.8642659697681665,
      "num_tokens": 58019417.0,
      "step": 28380
    },
    {
      "entropy": 0.44297404419630765,
      "epoch": 0.46108995233184186,
      "grad_norm": 7.125,
      "learning_rate": 2.8742235409123774e-05,
      "loss": 0.4357861042022705,
      "mean_token_accuracy": 0.8881089258939028,
      "num_tokens": 58038736.0,
      "step": 28390
    },
    {
      "entropy": 0.5363168688025326,
      "epoch": 0.46125236513646733,
      "grad_norm": 8.3125,
      "learning_rate": 2.872941464064274e-05,
      "loss": 0.5447350978851319,
      "mean_token_accuracy": 0.8607410121709108,
      "num_tokens": 58059315.0,
      "step": 28400
    },
    {
      "entropy": 0.5475703152827919,
      "epoch": 0.46141477794109287,
      "grad_norm": 8.375,
      "learning_rate": 2.8716592868940777e-05,
      "loss": 0.5251903533935547,
      "mean_token_accuracy": 0.861194783449173,
      "num_tokens": 58080513.0,
      "step": 28410
    },
    {
      "entropy": 0.4576488234102726,
      "epoch": 0.4615771907457184,
      "grad_norm": 6.25,
      "learning_rate": 2.8703770097466986e-05,
      "loss": 0.4897459030151367,
      "mean_token_accuracy": 0.8768083900213242,
      "num_tokens": 58100753.0,
      "step": 28420
    },
    {
      "entropy": 0.5401749092619866,
      "epoch": 0.46173960355034394,
      "grad_norm": 7.53125,
      "learning_rate": 2.869094632967072e-05,
      "loss": 0.5256563663482666,
      "mean_token_accuracy": 0.8685476124286652,
      "num_tokens": 58120992.0,
      "step": 28430
    },
    {
      "entropy": 0.5994820178952068,
      "epoch": 0.4619020163549694,
      "grad_norm": 7.03125,
      "learning_rate": 2.8678121569001592e-05,
      "loss": 0.5905335903167724,
      "mean_token_accuracy": 0.8479108743369579,
      "num_tokens": 58141673.0,
      "step": 28440
    },
    {
      "entropy": 0.5807786268182099,
      "epoch": 0.46206442915959495,
      "grad_norm": 7.125,
      "learning_rate": 2.866529581890949e-05,
      "loss": 0.5700356960296631,
      "mean_token_accuracy": 0.8585917323827743,
      "num_tokens": 58162745.0,
      "step": 28450
    },
    {
      "entropy": 0.49119173910003155,
      "epoch": 0.4622268419642205,
      "grad_norm": 7.84375,
      "learning_rate": 2.8652469082844585e-05,
      "loss": 0.4945389270782471,
      "mean_token_accuracy": 0.8758151896297932,
      "num_tokens": 58183459.0,
      "step": 28460
    },
    {
      "entropy": 0.4786935143638402,
      "epoch": 0.46238925476884596,
      "grad_norm": 6.71875,
      "learning_rate": 2.8639641364257284e-05,
      "loss": 0.5056787014007569,
      "mean_token_accuracy": 0.8769511818885803,
      "num_tokens": 58204210.0,
      "step": 28470
    },
    {
      "entropy": 0.4530389823485166,
      "epoch": 0.4625516675734715,
      "grad_norm": 12.4375,
      "learning_rate": 2.8626812666598275e-05,
      "loss": 0.45955824851989746,
      "mean_token_accuracy": 0.8805523261427879,
      "num_tokens": 58224947.0,
      "step": 28480
    },
    {
      "entropy": 0.5226126336492598,
      "epoch": 0.462714080378097,
      "grad_norm": 7.875,
      "learning_rate": 2.861398299331851e-05,
      "loss": 0.5214745044708252,
      "mean_token_accuracy": 0.866798072308302,
      "num_tokens": 58246051.0,
      "step": 28490
    },
    {
      "entropy": 0.4578794526401907,
      "epoch": 0.4628764931827225,
      "grad_norm": 5.125,
      "learning_rate": 2.8601152347869193e-05,
      "loss": 0.4578597068786621,
      "mean_token_accuracy": 0.8783680595457554,
      "num_tokens": 58266863.0,
      "step": 28500
    },
    {
      "entropy": 0.4489408145658672,
      "epoch": 0.46303890598734804,
      "grad_norm": 7.46875,
      "learning_rate": 2.8588320733701813e-05,
      "loss": 0.4768265724182129,
      "mean_token_accuracy": 0.8775707565248012,
      "num_tokens": 58285919.0,
      "step": 28510
    },
    {
      "entropy": 0.49431963795796036,
      "epoch": 0.46320131879197357,
      "grad_norm": 5.03125,
      "learning_rate": 2.8575488154268083e-05,
      "loss": 0.5043485164642334,
      "mean_token_accuracy": 0.8650501355528831,
      "num_tokens": 58306424.0,
      "step": 28520
    },
    {
      "entropy": 0.5237580328248441,
      "epoch": 0.46336373159659905,
      "grad_norm": 8.4375,
      "learning_rate": 2.856265461302e-05,
      "loss": 0.5038021564483642,
      "mean_token_accuracy": 0.8681864835321903,
      "num_tokens": 58326799.0,
      "step": 28530
    },
    {
      "entropy": 0.49096395750530064,
      "epoch": 0.4635261444012246,
      "grad_norm": 5.3125,
      "learning_rate": 2.8549820113409824e-05,
      "loss": 0.5077643394470215,
      "mean_token_accuracy": 0.8642884336411953,
      "num_tokens": 58348097.0,
      "step": 28540
    },
    {
      "entropy": 0.5835176645312459,
      "epoch": 0.4636885572058501,
      "grad_norm": 8.25,
      "learning_rate": 2.8536984658890064e-05,
      "loss": 0.6336695194244385,
      "mean_token_accuracy": 0.84471683986485,
      "num_tokens": 58367712.0,
      "step": 28550
    },
    {
      "entropy": 0.5091809654608369,
      "epoch": 0.46385097001047565,
      "grad_norm": 8.4375,
      "learning_rate": 2.852414825291348e-05,
      "loss": 0.5190454483032226,
      "mean_token_accuracy": 0.8638127416372299,
      "num_tokens": 58387056.0,
      "step": 28560
    },
    {
      "entropy": 0.46120275696739554,
      "epoch": 0.4640133828151011,
      "grad_norm": 4.96875,
      "learning_rate": 2.8511310898933085e-05,
      "loss": 0.44015069007873536,
      "mean_token_accuracy": 0.8889320299029351,
      "num_tokens": 58407491.0,
      "step": 28570
    },
    {
      "entropy": 0.6074223978444934,
      "epoch": 0.46417579561972666,
      "grad_norm": 9.5,
      "learning_rate": 2.8498472600402183e-05,
      "loss": 0.6582001686096192,
      "mean_token_accuracy": 0.8509986132383347,
      "num_tokens": 58428341.0,
      "step": 28580
    },
    {
      "entropy": 0.5233629588969052,
      "epoch": 0.4643382084243522,
      "grad_norm": 7.9375,
      "learning_rate": 2.8485633360774284e-05,
      "loss": 0.5261407852172851,
      "mean_token_accuracy": 0.8707919381558895,
      "num_tokens": 58448294.0,
      "step": 28590
    },
    {
      "entropy": 0.5530929511412979,
      "epoch": 0.4645006212289777,
      "grad_norm": 8.4375,
      "learning_rate": 2.8472793183503173e-05,
      "loss": 0.5527029514312745,
      "mean_token_accuracy": 0.8626071806997061,
      "num_tokens": 58469783.0,
      "step": 28600
    },
    {
      "entropy": 0.5605999634368345,
      "epoch": 0.4646630340336032,
      "grad_norm": 7.71875,
      "learning_rate": 2.8459952072042894e-05,
      "loss": 0.5172740459442139,
      "mean_token_accuracy": 0.8667139351367951,
      "num_tokens": 58490031.0,
      "step": 28610
    },
    {
      "entropy": 0.5199708007276058,
      "epoch": 0.46482544683822874,
      "grad_norm": 8.75,
      "learning_rate": 2.8447110029847735e-05,
      "loss": 0.5435936927795411,
      "mean_token_accuracy": 0.8610306687653064,
      "num_tokens": 58511167.0,
      "step": 28620
    },
    {
      "entropy": 0.5285219707060606,
      "epoch": 0.4649878596428542,
      "grad_norm": 9.0,
      "learning_rate": 2.843426706037223e-05,
      "loss": 0.5509239673614502,
      "mean_token_accuracy": 0.8529490254819393,
      "num_tokens": 58531846.0,
      "step": 28630
    },
    {
      "entropy": 0.4617819300852716,
      "epoch": 0.46515027244747975,
      "grad_norm": 8.625,
      "learning_rate": 2.842142316707116e-05,
      "loss": 0.4571998119354248,
      "mean_token_accuracy": 0.8861551232635975,
      "num_tokens": 58551960.0,
      "step": 28640
    },
    {
      "entropy": 0.5548846970312298,
      "epoch": 0.4653126852521053,
      "grad_norm": 7.28125,
      "learning_rate": 2.840857835339957e-05,
      "loss": 0.5538515090942383,
      "mean_token_accuracy": 0.8615544237196445,
      "num_tokens": 58571994.0,
      "step": 28650
    },
    {
      "entropy": 0.5113757980056107,
      "epoch": 0.4654750980567308,
      "grad_norm": 9.5,
      "learning_rate": 2.8395732622812753e-05,
      "loss": 0.4978204250335693,
      "mean_token_accuracy": 0.8723662622272969,
      "num_tokens": 58592663.0,
      "step": 28660
    },
    {
      "entropy": 0.5653941868338734,
      "epoch": 0.4656375108613563,
      "grad_norm": 7.75,
      "learning_rate": 2.8382885978766226e-05,
      "loss": 0.6409220218658447,
      "mean_token_accuracy": 0.8487643830478191,
      "num_tokens": 58613682.0,
      "step": 28670
    },
    {
      "entropy": 0.5341850281460211,
      "epoch": 0.46579992366598183,
      "grad_norm": 12.0,
      "learning_rate": 2.8370038424715762e-05,
      "loss": 0.5393381118774414,
      "mean_token_accuracy": 0.8591853231191635,
      "num_tokens": 58634313.0,
      "step": 28680
    },
    {
      "entropy": 0.5011114836204797,
      "epoch": 0.46596233647060736,
      "grad_norm": 6.59375,
      "learning_rate": 2.8357189964117393e-05,
      "loss": 0.5079142093658447,
      "mean_token_accuracy": 0.8740311890840531,
      "num_tokens": 58654664.0,
      "step": 28690
    },
    {
      "entropy": 0.6170508376322686,
      "epoch": 0.46612474927523284,
      "grad_norm": 9.0,
      "learning_rate": 2.8344340600427377e-05,
      "loss": 0.6418595314025879,
      "mean_token_accuracy": 0.8460695169866085,
      "num_tokens": 58675988.0,
      "step": 28700
    },
    {
      "entropy": 0.5121612361632287,
      "epoch": 0.4662871620798584,
      "grad_norm": 6.90625,
      "learning_rate": 2.8331490337102227e-05,
      "loss": 0.5404640674591065,
      "mean_token_accuracy": 0.8714973345398903,
      "num_tokens": 58696159.0,
      "step": 28710
    },
    {
      "entropy": 0.4462528958567418,
      "epoch": 0.4664495748844839,
      "grad_norm": 8.9375,
      "learning_rate": 2.8318639177598676e-05,
      "loss": 0.48027911186218264,
      "mean_token_accuracy": 0.8837172731757164,
      "num_tokens": 58716041.0,
      "step": 28720
    },
    {
      "entropy": 0.44541050363332035,
      "epoch": 0.4666119876891094,
      "grad_norm": 5.53125,
      "learning_rate": 2.8305787125373745e-05,
      "loss": 0.4326921939849854,
      "mean_token_accuracy": 0.8939703673124313,
      "num_tokens": 58736381.0,
      "step": 28730
    },
    {
      "entropy": 0.5445876111276448,
      "epoch": 0.4667744004937349,
      "grad_norm": 5.84375,
      "learning_rate": 2.829293418388464e-05,
      "loss": 0.5757375717163086,
      "mean_token_accuracy": 0.86097097620368,
      "num_tokens": 58756944.0,
      "step": 28740
    },
    {
      "entropy": 0.4825132834259421,
      "epoch": 0.46693681329836045,
      "grad_norm": 8.25,
      "learning_rate": 2.8280080356588835e-05,
      "loss": 0.47270898818969725,
      "mean_token_accuracy": 0.8781957753002644,
      "num_tokens": 58775845.0,
      "step": 28750
    },
    {
      "entropy": 0.5101991995703429,
      "epoch": 0.467099226102986,
      "grad_norm": 7.59375,
      "learning_rate": 2.826722564694405e-05,
      "loss": 0.5021873950958252,
      "mean_token_accuracy": 0.8747146472334861,
      "num_tokens": 58795264.0,
      "step": 28760
    },
    {
      "entropy": 0.4870492432266474,
      "epoch": 0.46726163890761147,
      "grad_norm": 7.8125,
      "learning_rate": 2.825437005840822e-05,
      "loss": 0.4967355728149414,
      "mean_token_accuracy": 0.8747383810579776,
      "num_tokens": 58815089.0,
      "step": 28770
    },
    {
      "entropy": 0.6020934372674673,
      "epoch": 0.467424051712237,
      "grad_norm": 7.0625,
      "learning_rate": 2.824151359443954e-05,
      "loss": 0.5662275314331054,
      "mean_token_accuracy": 0.8609477512538433,
      "num_tokens": 58836348.0,
      "step": 28780
    },
    {
      "entropy": 0.5522577329073102,
      "epoch": 0.46758646451686253,
      "grad_norm": 6.1875,
      "learning_rate": 2.82286562584964e-05,
      "loss": 0.5624095439910889,
      "mean_token_accuracy": 0.8500551402568817,
      "num_tokens": 58856950.0,
      "step": 28790
    },
    {
      "entropy": 0.5230826886137947,
      "epoch": 0.467748877321488,
      "grad_norm": 7.59375,
      "learning_rate": 2.8215798054037483e-05,
      "loss": 0.4668538570404053,
      "mean_token_accuracy": 0.8786075439304113,
      "num_tokens": 58877433.0,
      "step": 28800
    },
    {
      "entropy": 0.46910892869345844,
      "epoch": 0.46791129012611354,
      "grad_norm": 8.4375,
      "learning_rate": 2.820293898452166e-05,
      "loss": 0.45151596069335936,
      "mean_token_accuracy": 0.8821590825915336,
      "num_tokens": 58897446.0,
      "step": 28810
    },
    {
      "entropy": 0.5674657616298646,
      "epoch": 0.4680737029307391,
      "grad_norm": 8.4375,
      "learning_rate": 2.8190079053408052e-05,
      "loss": 0.6006256103515625,
      "mean_token_accuracy": 0.8582635506987571,
      "num_tokens": 58917580.0,
      "step": 28820
    },
    {
      "entropy": 0.5289128978736699,
      "epoch": 0.46823611573536456,
      "grad_norm": 8.3125,
      "learning_rate": 2.8177218264156008e-05,
      "loss": 0.5238483428955079,
      "mean_token_accuracy": 0.8720385029911994,
      "num_tokens": 58937111.0,
      "step": 28830
    },
    {
      "entropy": 0.4805091550806537,
      "epoch": 0.4683985285399901,
      "grad_norm": 8.875,
      "learning_rate": 2.8164356620225114e-05,
      "loss": 0.4741964817047119,
      "mean_token_accuracy": 0.8736087627708912,
      "num_tokens": 58957083.0,
      "step": 28840
    },
    {
      "entropy": 0.4591565925627947,
      "epoch": 0.4685609413446156,
      "grad_norm": 8.75,
      "learning_rate": 2.815149412507518e-05,
      "loss": 0.4875450134277344,
      "mean_token_accuracy": 0.8846498399972915,
      "num_tokens": 58976646.0,
      "step": 28850
    },
    {
      "entropy": 0.5367325347382575,
      "epoch": 0.4687233541492411,
      "grad_norm": 8.5625,
      "learning_rate": 2.8138630782166236e-05,
      "loss": 0.5388582229614258,
      "mean_token_accuracy": 0.8628377974033355,
      "num_tokens": 58998093.0,
      "step": 28860
    },
    {
      "entropy": 0.5019594680983573,
      "epoch": 0.46888576695386663,
      "grad_norm": 10.3125,
      "learning_rate": 2.812576659495856e-05,
      "loss": 0.5273561000823974,
      "mean_token_accuracy": 0.8658452503383159,
      "num_tokens": 59018151.0,
      "step": 28870
    },
    {
      "entropy": 0.5993008362944238,
      "epoch": 0.46904817975849217,
      "grad_norm": 9.0625,
      "learning_rate": 2.8112901566912646e-05,
      "loss": 0.583241319656372,
      "mean_token_accuracy": 0.8501493118703365,
      "num_tokens": 59038409.0,
      "step": 28880
    },
    {
      "entropy": 0.6123212797334417,
      "epoch": 0.4692105925631177,
      "grad_norm": 9.5,
      "learning_rate": 2.810003570148922e-05,
      "loss": 0.6621934413909912,
      "mean_token_accuracy": 0.8378510192036629,
      "num_tokens": 59058351.0,
      "step": 28890
    },
    {
      "entropy": 0.5347143078222871,
      "epoch": 0.4693730053677432,
      "grad_norm": 7.65625,
      "learning_rate": 2.808716900214921e-05,
      "loss": 0.557252311706543,
      "mean_token_accuracy": 0.8570691607892513,
      "num_tokens": 59080849.0,
      "step": 28900
    },
    {
      "entropy": 0.5318448594072833,
      "epoch": 0.4695354181723687,
      "grad_norm": 7.53125,
      "learning_rate": 2.8074301472353807e-05,
      "loss": 0.5008243560791016,
      "mean_token_accuracy": 0.8729621157050133,
      "num_tokens": 59101671.0,
      "step": 28910
    },
    {
      "entropy": 0.5269583736546337,
      "epoch": 0.46969783097699425,
      "grad_norm": 8.875,
      "learning_rate": 2.8061433115564407e-05,
      "loss": 0.5474551677703857,
      "mean_token_accuracy": 0.8615919455885888,
      "num_tokens": 59122218.0,
      "step": 28920
    },
    {
      "entropy": 0.5491807425860316,
      "epoch": 0.4698602437816197,
      "grad_norm": 9.25,
      "learning_rate": 2.8048563935242613e-05,
      "loss": 0.5691217899322509,
      "mean_token_accuracy": 0.8603252366185188,
      "num_tokens": 59143448.0,
      "step": 28930
    },
    {
      "entropy": 0.4917367325630039,
      "epoch": 0.47002265658624526,
      "grad_norm": 8.75,
      "learning_rate": 2.8035693934850265e-05,
      "loss": 0.4639258861541748,
      "mean_token_accuracy": 0.8873022764921188,
      "num_tokens": 59164432.0,
      "step": 28940
    },
    {
      "entropy": 0.48173368512652814,
      "epoch": 0.4701850693908708,
      "grad_norm": 7.09375,
      "learning_rate": 2.8022823117849434e-05,
      "loss": 0.4740555763244629,
      "mean_token_accuracy": 0.8831047616899014,
      "num_tokens": 59184924.0,
      "step": 28950
    },
    {
      "entropy": 0.4607706246431917,
      "epoch": 0.47034748219549627,
      "grad_norm": 9.5625,
      "learning_rate": 2.800995148770239e-05,
      "loss": 0.4716629505157471,
      "mean_token_accuracy": 0.8890001893043518,
      "num_tokens": 59204387.0,
      "step": 28960
    },
    {
      "entropy": 0.41450703288428487,
      "epoch": 0.4705098950001218,
      "grad_norm": 5.3125,
      "learning_rate": 2.7997079047871628e-05,
      "loss": 0.38582582473754884,
      "mean_token_accuracy": 0.9042799971997738,
      "num_tokens": 59224540.0,
      "step": 28970
    },
    {
      "entropy": 0.498604900855571,
      "epoch": 0.47067230780474734,
      "grad_norm": 7.625,
      "learning_rate": 2.7984205801819863e-05,
      "loss": 0.4882101535797119,
      "mean_token_accuracy": 0.8776780173182488,
      "num_tokens": 59245730.0,
      "step": 28980
    },
    {
      "entropy": 0.4720327663002536,
      "epoch": 0.47083472060937287,
      "grad_norm": 9.875,
      "learning_rate": 2.7971331753010033e-05,
      "loss": 0.4667468547821045,
      "mean_token_accuracy": 0.8862298645079136,
      "num_tokens": 59265028.0,
      "step": 28990
    },
    {
      "entropy": 0.4733597475569695,
      "epoch": 0.47099713341399835,
      "grad_norm": 7.28125,
      "learning_rate": 2.7958456904905284e-05,
      "loss": 0.49491591453552247,
      "mean_token_accuracy": 0.8707137577235698,
      "num_tokens": 59285453.0,
      "step": 29000
    },
    {
      "entropy": 0.507611849764362,
      "epoch": 0.4711595462186239,
      "grad_norm": 7.8125,
      "learning_rate": 2.7945581260968988e-05,
      "loss": 0.4839050769805908,
      "mean_token_accuracy": 0.867521571740508,
      "num_tokens": 59305589.0,
      "step": 29010
    },
    {
      "entropy": 0.5105169345159084,
      "epoch": 0.4713219590232494,
      "grad_norm": 7.6875,
      "learning_rate": 2.7932704824664696e-05,
      "loss": 0.4917591094970703,
      "mean_token_accuracy": 0.8640990361571312,
      "num_tokens": 59325640.0,
      "step": 29020
    },
    {
      "entropy": 0.4640126708894968,
      "epoch": 0.4714843718278749,
      "grad_norm": 8.25,
      "learning_rate": 2.7919827599456232e-05,
      "loss": 0.5124655723571777,
      "mean_token_accuracy": 0.8804964117705822,
      "num_tokens": 59345425.0,
      "step": 29030
    },
    {
      "entropy": 0.5048807728104293,
      "epoch": 0.4716467846325004,
      "grad_norm": 5.90625,
      "learning_rate": 2.7906949588807573e-05,
      "loss": 0.49563260078430177,
      "mean_token_accuracy": 0.8725179046392441,
      "num_tokens": 59366680.0,
      "step": 29040
    },
    {
      "entropy": 0.4982226439286023,
      "epoch": 0.47180919743712596,
      "grad_norm": 9.6875,
      "learning_rate": 2.7894070796182946e-05,
      "loss": 0.5183835506439209,
      "mean_token_accuracy": 0.8684291578829288,
      "num_tokens": 59386179.0,
      "step": 29050
    },
    {
      "entropy": 0.49047725009731946,
      "epoch": 0.47197161024175144,
      "grad_norm": 11.75,
      "learning_rate": 2.7881191225046766e-05,
      "loss": 0.5048642158508301,
      "mean_token_accuracy": 0.8687760289758444,
      "num_tokens": 59406008.0,
      "step": 29060
    },
    {
      "entropy": 0.49520368240773677,
      "epoch": 0.47213402304637697,
      "grad_norm": 5.5,
      "learning_rate": 2.7868310878863672e-05,
      "loss": 0.46657638549804686,
      "mean_token_accuracy": 0.8850192427635193,
      "num_tokens": 59425854.0,
      "step": 29070
    },
    {
      "entropy": 0.5125074000097811,
      "epoch": 0.4722964358510025,
      "grad_norm": 10.5625,
      "learning_rate": 2.785542976109851e-05,
      "loss": 0.5276947498321534,
      "mean_token_accuracy": 0.8698401913046837,
      "num_tokens": 59445909.0,
      "step": 29080
    },
    {
      "entropy": 0.6231117453426123,
      "epoch": 0.472458848655628,
      "grad_norm": 9.8125,
      "learning_rate": 2.7842547875216314e-05,
      "loss": 0.700288724899292,
      "mean_token_accuracy": 0.8300345376133919,
      "num_tokens": 59468174.0,
      "step": 29090
    },
    {
      "entropy": 0.4805035094730556,
      "epoch": 0.4726212614602535,
      "grad_norm": 6.59375,
      "learning_rate": 2.7829665224682367e-05,
      "loss": 0.46932177543640136,
      "mean_token_accuracy": 0.8789479933679104,
      "num_tokens": 59488572.0,
      "step": 29100
    },
    {
      "entropy": 0.4812570647103712,
      "epoch": 0.47278367426487905,
      "grad_norm": 10.25,
      "learning_rate": 2.7816781812962106e-05,
      "loss": 0.4593153953552246,
      "mean_token_accuracy": 0.8801739007234574,
      "num_tokens": 59508626.0,
      "step": 29110
    },
    {
      "entropy": 0.5619163550203666,
      "epoch": 0.4729460870695046,
      "grad_norm": 7.4375,
      "learning_rate": 2.7803897643521208e-05,
      "loss": 0.6146229267120361,
      "mean_token_accuracy": 0.840416020154953,
      "num_tokens": 59529377.0,
      "step": 29120
    },
    {
      "entropy": 0.4514863653574139,
      "epoch": 0.47310849987413006,
      "grad_norm": 7.96875,
      "learning_rate": 2.7791012719825537e-05,
      "loss": 0.4712690830230713,
      "mean_token_accuracy": 0.8814768403768539,
      "num_tokens": 59549754.0,
      "step": 29130
    },
    {
      "entropy": 0.5462221841327846,
      "epoch": 0.4732709126787556,
      "grad_norm": 7.125,
      "learning_rate": 2.7778127045341184e-05,
      "loss": 0.5591458797454834,
      "mean_token_accuracy": 0.8581615753471852,
      "num_tokens": 59569285.0,
      "step": 29140
    },
    {
      "entropy": 0.5204723903210834,
      "epoch": 0.47343332548338113,
      "grad_norm": 10.9375,
      "learning_rate": 2.7765240623534406e-05,
      "loss": 0.5422287940979004,
      "mean_token_accuracy": 0.8627918712794781,
      "num_tokens": 59590648.0,
      "step": 29150
    },
    {
      "entropy": 0.5784099218435585,
      "epoch": 0.4735957382880066,
      "grad_norm": 8.875,
      "learning_rate": 2.775235345787169e-05,
      "loss": 0.5869207859039307,
      "mean_token_accuracy": 0.8526345402002334,
      "num_tokens": 59610828.0,
      "step": 29160
    },
    {
      "entropy": 0.48779931450262665,
      "epoch": 0.47375815109263214,
      "grad_norm": 5.75,
      "learning_rate": 2.77394655518197e-05,
      "loss": 0.4360316276550293,
      "mean_token_accuracy": 0.8829068727791309,
      "num_tokens": 59632275.0,
      "step": 29170
    },
    {
      "entropy": 0.5344619671814144,
      "epoch": 0.4739205638972577,
      "grad_norm": 8.625,
      "learning_rate": 2.7726576908845326e-05,
      "loss": 0.5644847393035889,
      "mean_token_accuracy": 0.8634223960340023,
      "num_tokens": 59652355.0,
      "step": 29180
    },
    {
      "entropy": 0.5304775146767498,
      "epoch": 0.47408297670188315,
      "grad_norm": 9.6875,
      "learning_rate": 2.7713687532415634e-05,
      "loss": 0.5688430309295655,
      "mean_token_accuracy": 0.8622008301317692,
      "num_tokens": 59673353.0,
      "step": 29190
    },
    {
      "entropy": 0.4949146155733615,
      "epoch": 0.4742453895065087,
      "grad_norm": 9.25,
      "learning_rate": 2.77007974259979e-05,
      "loss": 0.5196728229522705,
      "mean_token_accuracy": 0.8749715499579906,
      "num_tokens": 59693373.0,
      "step": 29200
    },
    {
      "entropy": 0.5511972334701568,
      "epoch": 0.4744078023111342,
      "grad_norm": 5.90625,
      "learning_rate": 2.768790659305958e-05,
      "loss": 0.5451839447021485,
      "mean_token_accuracy": 0.8599359475076198,
      "num_tokens": 59714269.0,
      "step": 29210
    },
    {
      "entropy": 0.5794609194621444,
      "epoch": 0.47457021511575975,
      "grad_norm": 6.9375,
      "learning_rate": 2.7675015037068353e-05,
      "loss": 0.6084275245666504,
      "mean_token_accuracy": 0.856339019536972,
      "num_tokens": 59735049.0,
      "step": 29220
    },
    {
      "entropy": 0.46456572310999034,
      "epoch": 0.47473262792038523,
      "grad_norm": 5.59375,
      "learning_rate": 2.7662122761492077e-05,
      "loss": 0.4709340572357178,
      "mean_token_accuracy": 0.8785843297839164,
      "num_tokens": 59755248.0,
      "step": 29230
    },
    {
      "entropy": 0.49348001040052625,
      "epoch": 0.47489504072501076,
      "grad_norm": 9.875,
      "learning_rate": 2.7649229769798778e-05,
      "loss": 0.4944154739379883,
      "mean_token_accuracy": 0.8823047116398811,
      "num_tokens": 59777055.0,
      "step": 29240
    },
    {
      "entropy": 0.5304674033541232,
      "epoch": 0.4750574535296363,
      "grad_norm": 9.9375,
      "learning_rate": 2.763633606545673e-05,
      "loss": 0.5062361240386963,
      "mean_token_accuracy": 0.8706802144646645,
      "num_tokens": 59797661.0,
      "step": 29250
    },
    {
      "entropy": 0.5543469178490341,
      "epoch": 0.4752198663342618,
      "grad_norm": 6.4375,
      "learning_rate": 2.762344165193435e-05,
      "loss": 0.5556524753570556,
      "mean_token_accuracy": 0.8573232799768448,
      "num_tokens": 59817646.0,
      "step": 29260
    },
    {
      "entropy": 0.5326763639226556,
      "epoch": 0.4753822791388873,
      "grad_norm": 7.8125,
      "learning_rate": 2.7610546532700273e-05,
      "loss": 0.4878638744354248,
      "mean_token_accuracy": 0.8750081643462181,
      "num_tokens": 59838036.0,
      "step": 29270
    },
    {
      "entropy": 0.5074692221358419,
      "epoch": 0.47554469194351284,
      "grad_norm": 7.375,
      "learning_rate": 2.7597650711223305e-05,
      "loss": 0.5110429286956787,
      "mean_token_accuracy": 0.8702082328498364,
      "num_tokens": 59858754.0,
      "step": 29280
    },
    {
      "entropy": 0.47091527259908617,
      "epoch": 0.4757071047481383,
      "grad_norm": 9.75,
      "learning_rate": 2.7584754190972463e-05,
      "loss": 0.508536958694458,
      "mean_token_accuracy": 0.8694832641631365,
      "num_tokens": 59880655.0,
      "step": 29290
    },
    {
      "entropy": 0.4723741309950128,
      "epoch": 0.47586951755276385,
      "grad_norm": 7.03125,
      "learning_rate": 2.7571856975416944e-05,
      "loss": 0.4935288906097412,
      "mean_token_accuracy": 0.8806958936154843,
      "num_tokens": 59900987.0,
      "step": 29300
    },
    {
      "entropy": 0.49176813787780704,
      "epoch": 0.4760319303573894,
      "grad_norm": 6.625,
      "learning_rate": 2.7558959068026106e-05,
      "loss": 0.4761043071746826,
      "mean_token_accuracy": 0.8785827994346619,
      "num_tokens": 59921704.0,
      "step": 29310
    },
    {
      "entropy": 0.4842653120867908,
      "epoch": 0.47619434316201487,
      "grad_norm": 9.875,
      "learning_rate": 2.754606047226953e-05,
      "loss": 0.47976126670837405,
      "mean_token_accuracy": 0.8742985285818576,
      "num_tokens": 59942440.0,
      "step": 29320
    },
    {
      "entropy": 0.45057608569040897,
      "epoch": 0.4763567559666404,
      "grad_norm": 8.3125,
      "learning_rate": 2.753316119161696e-05,
      "loss": 0.4638981342315674,
      "mean_token_accuracy": 0.8838927946984768,
      "num_tokens": 59963542.0,
      "step": 29330
    },
    {
      "entropy": 0.48528456832282246,
      "epoch": 0.47651916877126593,
      "grad_norm": 6.53125,
      "learning_rate": 2.7520261229538347e-05,
      "loss": 0.4632599353790283,
      "mean_token_accuracy": 0.8749178193509579,
      "num_tokens": 59984405.0,
      "step": 29340
    },
    {
      "entropy": 0.5292535034939647,
      "epoch": 0.47668158157589147,
      "grad_norm": 7.625,
      "learning_rate": 2.75073605895038e-05,
      "loss": 0.5723861217498779,
      "mean_token_accuracy": 0.8543614126741886,
      "num_tokens": 60006474.0,
      "step": 29350
    },
    {
      "entropy": 0.56714669726789,
      "epoch": 0.47684399438051694,
      "grad_norm": 6.8125,
      "learning_rate": 2.749445927498361e-05,
      "loss": 0.5599283218383789,
      "mean_token_accuracy": 0.8447634615004063,
      "num_tokens": 60027765.0,
      "step": 29360
    },
    {
      "entropy": 0.5392264002468437,
      "epoch": 0.4770064071851425,
      "grad_norm": 9.8125,
      "learning_rate": 2.748155728944828e-05,
      "loss": 0.5512895107269287,
      "mean_token_accuracy": 0.8631967589259147,
      "num_tokens": 60047664.0,
      "step": 29370
    },
    {
      "entropy": 0.46034361869096757,
      "epoch": 0.477168819989768,
      "grad_norm": 6.34375,
      "learning_rate": 2.746865463636845e-05,
      "loss": 0.4695130348205566,
      "mean_token_accuracy": 0.8772197350859642,
      "num_tokens": 60068312.0,
      "step": 29380
    },
    {
      "entropy": 0.5103020229493268,
      "epoch": 0.4773312327943935,
      "grad_norm": 11.6875,
      "learning_rate": 2.745575131921498e-05,
      "loss": 0.543897008895874,
      "mean_token_accuracy": 0.8584680661559105,
      "num_tokens": 60088003.0,
      "step": 29390
    },
    {
      "entropy": 0.5412410012446344,
      "epoch": 0.477493645599019,
      "grad_norm": 6.90625,
      "learning_rate": 2.744284734145889e-05,
      "loss": 0.5487354755401611,
      "mean_token_accuracy": 0.8645126767456531,
      "num_tokens": 60108639.0,
      "step": 29400
    },
    {
      "entropy": 0.5813955397112295,
      "epoch": 0.47765605840364456,
      "grad_norm": 7.40625,
      "learning_rate": 2.7429942706571376e-05,
      "loss": 0.628877830505371,
      "mean_token_accuracy": 0.8525042816996574,
      "num_tokens": 60129699.0,
      "step": 29410
    },
    {
      "entropy": 0.45172173441387714,
      "epoch": 0.47781847120827003,
      "grad_norm": 7.75,
      "learning_rate": 2.7417037418023807e-05,
      "loss": 0.4584473133087158,
      "mean_token_accuracy": 0.8905608534812928,
      "num_tokens": 60148885.0,
      "step": 29420
    },
    {
      "entropy": 0.5677950949408114,
      "epoch": 0.47798088401289557,
      "grad_norm": 7.6875,
      "learning_rate": 2.7404131479287737e-05,
      "loss": 0.5595879077911377,
      "mean_token_accuracy": 0.8744869694113732,
      "num_tokens": 60167978.0,
      "step": 29430
    },
    {
      "entropy": 0.48404775145463647,
      "epoch": 0.4781432968175211,
      "grad_norm": 9.125,
      "learning_rate": 2.7391224893834904e-05,
      "loss": 0.5031661510467529,
      "mean_token_accuracy": 0.8685784190893173,
      "num_tokens": 60187473.0,
      "step": 29440
    },
    {
      "entropy": 0.5043280594749376,
      "epoch": 0.47830570962214664,
      "grad_norm": 7.875,
      "learning_rate": 2.7378317665137197e-05,
      "loss": 0.5054443359375,
      "mean_token_accuracy": 0.8712377823889256,
      "num_tokens": 60208791.0,
      "step": 29450
    },
    {
      "entropy": 0.6021320603787899,
      "epoch": 0.4784681224267721,
      "grad_norm": 6.25,
      "learning_rate": 2.7365409796666703e-05,
      "loss": 0.6106886863708496,
      "mean_token_accuracy": 0.8551202431321144,
      "num_tokens": 60229743.0,
      "step": 29460
    },
    {
      "entropy": 0.5146654705982655,
      "epoch": 0.47863053523139765,
      "grad_norm": 5.9375,
      "learning_rate": 2.7352501291895637e-05,
      "loss": 0.5126700878143311,
      "mean_token_accuracy": 0.8685191497206688,
      "num_tokens": 60250674.0,
      "step": 29470
    },
    {
      "entropy": 0.632385182287544,
      "epoch": 0.4787929480360232,
      "grad_norm": 7.5,
      "learning_rate": 2.7339592154296457e-05,
      "loss": 0.6202254295349121,
      "mean_token_accuracy": 0.8507012009620667,
      "num_tokens": 60271039.0,
      "step": 29480
    },
    {
      "entropy": 0.5470962394028902,
      "epoch": 0.47895536084064866,
      "grad_norm": 4.53125,
      "learning_rate": 2.7326682387341717e-05,
      "loss": 0.5675343036651611,
      "mean_token_accuracy": 0.8566773720085621,
      "num_tokens": 60292537.0,
      "step": 29490
    },
    {
      "entropy": 0.5249214741867035,
      "epoch": 0.4791177736452742,
      "grad_norm": 6.9375,
      "learning_rate": 2.7313771994504195e-05,
      "loss": 0.5128068923950195,
      "mean_token_accuracy": 0.8744842410087585,
      "num_tokens": 60312647.0,
      "step": 29500
    },
    {
      "entropy": 0.5530703178141266,
      "epoch": 0.4792801864498997,
      "grad_norm": 7.90625,
      "learning_rate": 2.7300860979256798e-05,
      "loss": 0.5479613780975342,
      "mean_token_accuracy": 0.8654669560492039,
      "num_tokens": 60332689.0,
      "step": 29510
    },
    {
      "entropy": 0.4852082619909197,
      "epoch": 0.4794425992545252,
      "grad_norm": 11.3125,
      "learning_rate": 2.7287949345072635e-05,
      "loss": 0.47780413627624513,
      "mean_token_accuracy": 0.8818675145506859,
      "num_tokens": 60351996.0,
      "step": 29520
    },
    {
      "entropy": 0.5326118987752124,
      "epoch": 0.47960501205915074,
      "grad_norm": 7.5625,
      "learning_rate": 2.7275037095424953e-05,
      "loss": 0.5676815986633301,
      "mean_token_accuracy": 0.8570737045258283,
      "num_tokens": 60372274.0,
      "step": 29530
    },
    {
      "entropy": 0.4631880828179419,
      "epoch": 0.47976742486377627,
      "grad_norm": 7.25,
      "learning_rate": 2.7262124233787172e-05,
      "loss": 0.4440745830535889,
      "mean_token_accuracy": 0.8860004730522633,
      "num_tokens": 60392906.0,
      "step": 29540
    },
    {
      "entropy": 0.5011167306452989,
      "epoch": 0.4799298376684018,
      "grad_norm": 6.8125,
      "learning_rate": 2.7249210763632898e-05,
      "loss": 0.5023388862609863,
      "mean_token_accuracy": 0.8715539656579494,
      "num_tokens": 60413309.0,
      "step": 29550
    },
    {
      "entropy": 0.4597462934907526,
      "epoch": 0.4800922504730273,
      "grad_norm": 11.75,
      "learning_rate": 2.7236296688435868e-05,
      "loss": 0.4675177574157715,
      "mean_token_accuracy": 0.8796481259167195,
      "num_tokens": 60434245.0,
      "step": 29560
    },
    {
      "entropy": 0.5782324160449207,
      "epoch": 0.4802546632776528,
      "grad_norm": 6.65625,
      "learning_rate": 2.7223382011670012e-05,
      "loss": 0.591136884689331,
      "mean_token_accuracy": 0.8440025478601456,
      "num_tokens": 60454589.0,
      "step": 29570
    },
    {
      "entropy": 0.4709653503727168,
      "epoch": 0.48041707608227835,
      "grad_norm": 7.84375,
      "learning_rate": 2.7210466736809387e-05,
      "loss": 0.47792782783508303,
      "mean_token_accuracy": 0.8835742913186551,
      "num_tokens": 60475205.0,
      "step": 29580
    },
    {
      "entropy": 0.5613249571062624,
      "epoch": 0.4805794888869038,
      "grad_norm": 8.4375,
      "learning_rate": 2.719755086732825e-05,
      "loss": 0.5236572742462158,
      "mean_token_accuracy": 0.8690123841166496,
      "num_tokens": 60495646.0,
      "step": 29590
    },
    {
      "entropy": 0.5990663065109402,
      "epoch": 0.48074190169152936,
      "grad_norm": 6.65625,
      "learning_rate": 2.7184634406700986e-05,
      "loss": 0.6056470394134521,
      "mean_token_accuracy": 0.8554785706102848,
      "num_tokens": 60516240.0,
      "step": 29600
    },
    {
      "entropy": 0.4954465480521321,
      "epoch": 0.4809043144961549,
      "grad_norm": 7.59375,
      "learning_rate": 2.717171735840216e-05,
      "loss": 0.4734655857086182,
      "mean_token_accuracy": 0.8827458553016185,
      "num_tokens": 60535498.0,
      "step": 29610
    },
    {
      "entropy": 0.41187417223118244,
      "epoch": 0.4810667273007804,
      "grad_norm": 11.875,
      "learning_rate": 2.7158799725906482e-05,
      "loss": 0.4087969303131104,
      "mean_token_accuracy": 0.8972854971885681,
      "num_tokens": 60556508.0,
      "step": 29620
    },
    {
      "entropy": 0.5187026370549574,
      "epoch": 0.4812291401054059,
      "grad_norm": 5.53125,
      "learning_rate": 2.7145881512688824e-05,
      "loss": 0.5632044792175293,
      "mean_token_accuracy": 0.8694371372461319,
      "num_tokens": 60576456.0,
      "step": 29630
    },
    {
      "entropy": 0.49102704410906883,
      "epoch": 0.48139155291003144,
      "grad_norm": 7.09375,
      "learning_rate": 2.7132962722224227e-05,
      "loss": 0.498667573928833,
      "mean_token_accuracy": 0.8760030813515186,
      "num_tokens": 60597658.0,
      "step": 29640
    },
    {
      "entropy": 0.4475715471664444,
      "epoch": 0.4815539657146569,
      "grad_norm": 7.5,
      "learning_rate": 2.7120043357987856e-05,
      "loss": 0.4275506019592285,
      "mean_token_accuracy": 0.8866591908037662,
      "num_tokens": 60617752.0,
      "step": 29650
    },
    {
      "entropy": 0.5014886402990669,
      "epoch": 0.48171637851928245,
      "grad_norm": 8.0,
      "learning_rate": 2.7107123423455057e-05,
      "loss": 0.5038645267486572,
      "mean_token_accuracy": 0.86918326318264,
      "num_tokens": 60638325.0,
      "step": 29660
    },
    {
      "entropy": 0.519360332749784,
      "epoch": 0.481878791323908,
      "grad_norm": 10.875,
      "learning_rate": 2.7094202922101318e-05,
      "loss": 0.5435659408569335,
      "mean_token_accuracy": 0.8648259781301022,
      "num_tokens": 60658792.0,
      "step": 29670
    },
    {
      "entropy": 0.6445486386306584,
      "epoch": 0.4820412041285335,
      "grad_norm": 8.75,
      "learning_rate": 2.708128185740229e-05,
      "loss": 0.6282527446746826,
      "mean_token_accuracy": 0.8412114650011062,
      "num_tokens": 60682503.0,
      "step": 29680
    },
    {
      "entropy": 0.5103652227902785,
      "epoch": 0.482203616933159,
      "grad_norm": 10.6875,
      "learning_rate": 2.7068360232833772e-05,
      "loss": 0.5392825126647949,
      "mean_token_accuracy": 0.8645791403949261,
      "num_tokens": 60703347.0,
      "step": 29690
    },
    {
      "entropy": 0.4788790072314441,
      "epoch": 0.48236602973778453,
      "grad_norm": 6.3125,
      "learning_rate": 2.705543805187169e-05,
      "loss": 0.4928687572479248,
      "mean_token_accuracy": 0.8770027764141559,
      "num_tokens": 60723680.0,
      "step": 29700
    },
    {
      "entropy": 0.5170168743468821,
      "epoch": 0.48252844254241006,
      "grad_norm": 11.9375,
      "learning_rate": 2.704251531799217e-05,
      "loss": 0.4810340881347656,
      "mean_token_accuracy": 0.8725479528307915,
      "num_tokens": 60745029.0,
      "step": 29710
    },
    {
      "entropy": 0.5183397804386913,
      "epoch": 0.48269085534703554,
      "grad_norm": 8.25,
      "learning_rate": 2.7029592034671424e-05,
      "loss": 0.5425444126129151,
      "mean_token_accuracy": 0.8662149779498577,
      "num_tokens": 60766015.0,
      "step": 29720
    },
    {
      "entropy": 0.5395031598396599,
      "epoch": 0.4828532681516611,
      "grad_norm": 11.1875,
      "learning_rate": 2.7016668205385863e-05,
      "loss": 0.5130730628967285,
      "mean_token_accuracy": 0.8666693948209285,
      "num_tokens": 60786405.0,
      "step": 29730
    },
    {
      "entropy": 0.5517880907282233,
      "epoch": 0.4830156809562866,
      "grad_norm": 7.15625,
      "learning_rate": 2.7003743833612023e-05,
      "loss": 0.5733532428741455,
      "mean_token_accuracy": 0.8584562078118324,
      "num_tokens": 60806698.0,
      "step": 29740
    },
    {
      "entropy": 0.5730183784849942,
      "epoch": 0.4831780937609121,
      "grad_norm": 8.3125,
      "learning_rate": 2.6990818922826596e-05,
      "loss": 0.57197585105896,
      "mean_token_accuracy": 0.8476547561585903,
      "num_tokens": 60827537.0,
      "step": 29750
    },
    {
      "entropy": 0.5533103136345744,
      "epoch": 0.4833405065655376,
      "grad_norm": 7.5,
      "learning_rate": 2.6977893476506393e-05,
      "loss": 0.577481460571289,
      "mean_token_accuracy": 0.860558132827282,
      "num_tokens": 60848066.0,
      "step": 29760
    },
    {
      "entropy": 0.45370709961280226,
      "epoch": 0.48350291937016315,
      "grad_norm": 7.96875,
      "learning_rate": 2.6964967498128407e-05,
      "loss": 0.4289956569671631,
      "mean_token_accuracy": 0.892080084234476,
      "num_tokens": 60868493.0,
      "step": 29770
    },
    {
      "entropy": 0.4949894939083606,
      "epoch": 0.4836653321747887,
      "grad_norm": 6.78125,
      "learning_rate": 2.6952040991169747e-05,
      "loss": 0.4752922534942627,
      "mean_token_accuracy": 0.8784846656024456,
      "num_tokens": 60889076.0,
      "step": 29780
    },
    {
      "entropy": 0.48860285901464523,
      "epoch": 0.48382774497941416,
      "grad_norm": 8.375,
      "learning_rate": 2.693911395910767e-05,
      "loss": 0.5030402660369873,
      "mean_token_accuracy": 0.867689611017704,
      "num_tokens": 60909607.0,
      "step": 29790
    },
    {
      "entropy": 0.481493732961826,
      "epoch": 0.4839901577840397,
      "grad_norm": 4.71875,
      "learning_rate": 2.692618640541959e-05,
      "loss": 0.46890997886657715,
      "mean_token_accuracy": 0.8801246553659439,
      "num_tokens": 60930179.0,
      "step": 29800
    },
    {
      "entropy": 0.4926418269518763,
      "epoch": 0.48415257058866523,
      "grad_norm": 10.9375,
      "learning_rate": 2.6913258333583025e-05,
      "loss": 0.5208756446838378,
      "mean_token_accuracy": 0.8652375198900699,
      "num_tokens": 60949757.0,
      "step": 29810
    },
    {
      "entropy": 0.5450511113740504,
      "epoch": 0.4843149833932907,
      "grad_norm": 9.125,
      "learning_rate": 2.6900329747075677e-05,
      "loss": 0.5472739219665528,
      "mean_token_accuracy": 0.8598699398338795,
      "num_tokens": 60970071.0,
      "step": 29820
    },
    {
      "entropy": 0.5112455891678109,
      "epoch": 0.48447739619791624,
      "grad_norm": 8.0625,
      "learning_rate": 2.688740064937535e-05,
      "loss": 0.5195834159851074,
      "mean_token_accuracy": 0.8668784521520138,
      "num_tokens": 60990112.0,
      "step": 29830
    },
    {
      "entropy": 0.517966901184991,
      "epoch": 0.4846398090025418,
      "grad_norm": 10.125,
      "learning_rate": 2.687447104396001e-05,
      "loss": 0.49804072380065917,
      "mean_token_accuracy": 0.8773394979536533,
      "num_tokens": 61012147.0,
      "step": 29840
    },
    {
      "entropy": 0.5584859501570463,
      "epoch": 0.48480222180716726,
      "grad_norm": 5.9375,
      "learning_rate": 2.6861540934307746e-05,
      "loss": 0.5686408519744873,
      "mean_token_accuracy": 0.8544702749699354,
      "num_tokens": 61032762.0,
      "step": 29850
    },
    {
      "entropy": 0.4955867235548794,
      "epoch": 0.4849646346117928,
      "grad_norm": 8.625,
      "learning_rate": 2.684861032389679e-05,
      "loss": 0.49292740821838377,
      "mean_token_accuracy": 0.8721981771290302,
      "num_tokens": 61053004.0,
      "step": 29860
    },
    {
      "entropy": 0.649409282905981,
      "epoch": 0.4851270474164183,
      "grad_norm": 8.5,
      "learning_rate": 2.68356792162055e-05,
      "loss": 0.6697579860687256,
      "mean_token_accuracy": 0.8376047853380442,
      "num_tokens": 61074051.0,
      "step": 29870
    },
    {
      "entropy": 0.487559350207448,
      "epoch": 0.4852894602210438,
      "grad_norm": 6.1875,
      "learning_rate": 2.6822747614712374e-05,
      "loss": 0.4505396842956543,
      "mean_token_accuracy": 0.8798145987093449,
      "num_tokens": 61094196.0,
      "step": 29880
    },
    {
      "entropy": 0.4557596504222602,
      "epoch": 0.48545187302566933,
      "grad_norm": 9.5,
      "learning_rate": 2.6809815522896047e-05,
      "loss": 0.44684624671936035,
      "mean_token_accuracy": 0.8865221582353116,
      "num_tokens": 61114216.0,
      "step": 29890
    },
    {
      "entropy": 0.49129079645499585,
      "epoch": 0.48561428583029487,
      "grad_norm": 6.0,
      "learning_rate": 2.6796882944235275e-05,
      "loss": 0.4906649112701416,
      "mean_token_accuracy": 0.8728243939578533,
      "num_tokens": 61134573.0,
      "step": 29900
    },
    {
      "entropy": 0.5212351752445101,
      "epoch": 0.4857766986349204,
      "grad_norm": 7.15625,
      "learning_rate": 2.678394988220896e-05,
      "loss": 0.5722396373748779,
      "mean_token_accuracy": 0.8581037942320109,
      "num_tokens": 61154788.0,
      "step": 29910
    },
    {
      "entropy": 0.498322302242741,
      "epoch": 0.4859391114395459,
      "grad_norm": 7.0625,
      "learning_rate": 2.6771016340296102e-05,
      "loss": 0.4833559036254883,
      "mean_token_accuracy": 0.8750042654573917,
      "num_tokens": 61176820.0,
      "step": 29920
    },
    {
      "entropy": 0.5365836209617555,
      "epoch": 0.4861015242441714,
      "grad_norm": 7.9375,
      "learning_rate": 2.675808232197588e-05,
      "loss": 0.5318466663360596,
      "mean_token_accuracy": 0.8658442594110966,
      "num_tokens": 61197244.0,
      "step": 29930
    },
    {
      "entropy": 0.5024980888934806,
      "epoch": 0.48626393704879695,
      "grad_norm": 6.8125,
      "learning_rate": 2.6745147830727557e-05,
      "loss": 0.50679612159729,
      "mean_token_accuracy": 0.8752139590680599,
      "num_tokens": 61217182.0,
      "step": 29940
    },
    {
      "entropy": 0.4963731272146106,
      "epoch": 0.4864263498534224,
      "grad_norm": 6.0,
      "learning_rate": 2.673221287003055e-05,
      "loss": 0.5420843124389648,
      "mean_token_accuracy": 0.8680335745215416,
      "num_tokens": 61237226.0,
      "step": 29950
    },
    {
      "entropy": 0.4804054842446931,
      "epoch": 0.48658876265804796,
      "grad_norm": 7.5625,
      "learning_rate": 2.671927744336438e-05,
      "loss": 0.5258739471435547,
      "mean_token_accuracy": 0.8703325003385544,
      "num_tokens": 61257079.0,
      "step": 29960
    },
    {
      "entropy": 0.47754820762202144,
      "epoch": 0.4867511754626735,
      "grad_norm": 8.5,
      "learning_rate": 2.670634155420872e-05,
      "loss": 0.4926462650299072,
      "mean_token_accuracy": 0.869492755830288,
      "num_tokens": 61277613.0,
      "step": 29970
    },
    {
      "entropy": 0.5729579982114956,
      "epoch": 0.48691358826729897,
      "grad_norm": 6.6875,
      "learning_rate": 2.6693405206043354e-05,
      "loss": 0.5480486869812011,
      "mean_token_accuracy": 0.86570425927639,
      "num_tokens": 61298213.0,
      "step": 29980
    },
    {
      "entropy": 0.49753935243934394,
      "epoch": 0.4870760010719245,
      "grad_norm": 8.9375,
      "learning_rate": 2.668046840234817e-05,
      "loss": 0.463856840133667,
      "mean_token_accuracy": 0.8777034610509873,
      "num_tokens": 61318897.0,
      "step": 29990
    },
    {
      "entropy": 0.4641304031945765,
      "epoch": 0.48723841387655004,
      "grad_norm": 10.5,
      "learning_rate": 2.6667531146603216e-05,
      "loss": 0.43685169219970704,
      "mean_token_accuracy": 0.8847464174032211,
      "num_tokens": 61338608.0,
      "step": 30000
    },
    {
      "entropy": 0.43955771944019945,
      "epoch": 0.48740082668117557,
      "grad_norm": 5.78125,
      "learning_rate": 2.665459344228863e-05,
      "loss": 0.4642635345458984,
      "mean_token_accuracy": 0.8810644365847111,
      "num_tokens": 61358800.0,
      "step": 30010
    },
    {
      "entropy": 0.4733219462446868,
      "epoch": 0.48756323948580105,
      "grad_norm": 8.125,
      "learning_rate": 2.6641655292884694e-05,
      "loss": 0.47419443130493166,
      "mean_token_accuracy": 0.8825159944593907,
      "num_tokens": 61379146.0,
      "step": 30020
    },
    {
      "entropy": 0.5096843206323683,
      "epoch": 0.4877256522904266,
      "grad_norm": 8.0625,
      "learning_rate": 2.6628716701871793e-05,
      "loss": 0.527325439453125,
      "mean_token_accuracy": 0.8592043444514275,
      "num_tokens": 61398997.0,
      "step": 30030
    },
    {
      "entropy": 0.535060270037502,
      "epoch": 0.4878880650950521,
      "grad_norm": 7.5625,
      "learning_rate": 2.6615777672730437e-05,
      "loss": 0.5630969047546387,
      "mean_token_accuracy": 0.8506309509277343,
      "num_tokens": 61419827.0,
      "step": 30040
    },
    {
      "entropy": 0.4415871178731322,
      "epoch": 0.4880504778996776,
      "grad_norm": 7.46875,
      "learning_rate": 2.6602838208941266e-05,
      "loss": 0.4716456890106201,
      "mean_token_accuracy": 0.8887098826467991,
      "num_tokens": 61440173.0,
      "step": 30050
    },
    {
      "entropy": 0.45274502753745766,
      "epoch": 0.4882128907043031,
      "grad_norm": 6.65625,
      "learning_rate": 2.6589898313985006e-05,
      "loss": 0.4595757007598877,
      "mean_token_accuracy": 0.8788973763585091,
      "num_tokens": 61460553.0,
      "step": 30060
    },
    {
      "entropy": 0.4252335591241717,
      "epoch": 0.48837530350892866,
      "grad_norm": 9.125,
      "learning_rate": 2.657695799134253e-05,
      "loss": 0.42985143661499026,
      "mean_token_accuracy": 0.8924656808376312,
      "num_tokens": 61481079.0,
      "step": 30070
    },
    {
      "entropy": 0.44788708586711434,
      "epoch": 0.48853771631355414,
      "grad_norm": 7.4375,
      "learning_rate": 2.6564017244494805e-05,
      "loss": 0.42978811264038086,
      "mean_token_accuracy": 0.8829262234270573,
      "num_tokens": 61500762.0,
      "step": 30080
    },
    {
      "entropy": 0.5700707169249654,
      "epoch": 0.48870012911817967,
      "grad_norm": 7.25,
      "learning_rate": 2.6551076076922933e-05,
      "loss": 0.5816641807556152,
      "mean_token_accuracy": 0.857812387496233,
      "num_tokens": 61520628.0,
      "step": 30090
    },
    {
      "entropy": 0.5153474589809776,
      "epoch": 0.4888625419228052,
      "grad_norm": 6.875,
      "learning_rate": 2.6538134492108103e-05,
      "loss": 0.5383924007415771,
      "mean_token_accuracy": 0.8654929660260677,
      "num_tokens": 61541351.0,
      "step": 30100
    },
    {
      "entropy": 0.5095272520091385,
      "epoch": 0.4890249547274307,
      "grad_norm": 7.28125,
      "learning_rate": 2.652519249353164e-05,
      "loss": 0.5477429389953613,
      "mean_token_accuracy": 0.8602794758975506,
      "num_tokens": 61561791.0,
      "step": 30110
    },
    {
      "entropy": 0.4994806745555252,
      "epoch": 0.4891873675320562,
      "grad_norm": 8.75,
      "learning_rate": 2.6512250084674973e-05,
      "loss": 0.4817654132843018,
      "mean_token_accuracy": 0.8751686945557594,
      "num_tokens": 61582246.0,
      "step": 30120
    },
    {
      "entropy": 0.5301328948931768,
      "epoch": 0.48934978033668175,
      "grad_norm": 7.03125,
      "learning_rate": 2.649930726901963e-05,
      "loss": 0.5003461837768555,
      "mean_token_accuracy": 0.8695394821465016,
      "num_tokens": 61604283.0,
      "step": 30130
    },
    {
      "entropy": 0.5047398585826158,
      "epoch": 0.4895121931413073,
      "grad_norm": 5.21875,
      "learning_rate": 2.648636405004727e-05,
      "loss": 0.4982298851013184,
      "mean_token_accuracy": 0.8746279820799827,
      "num_tokens": 61623753.0,
      "step": 30140
    },
    {
      "entropy": 0.4831450752448291,
      "epoch": 0.48967460594593276,
      "grad_norm": 9.4375,
      "learning_rate": 2.6473420431239626e-05,
      "loss": 0.5067430019378663,
      "mean_token_accuracy": 0.8715280137956143,
      "num_tokens": 61644275.0,
      "step": 30150
    },
    {
      "entropy": 0.5027948100585491,
      "epoch": 0.4898370187505583,
      "grad_norm": 9.5625,
      "learning_rate": 2.646047641607859e-05,
      "loss": 0.5231240749359131,
      "mean_token_accuracy": 0.8735895469784737,
      "num_tokens": 61663702.0,
      "step": 30160
    },
    {
      "entropy": 0.5178452826105058,
      "epoch": 0.48999943155518383,
      "grad_norm": 10.0625,
      "learning_rate": 2.6447532008046106e-05,
      "loss": 0.5340951919555664,
      "mean_token_accuracy": 0.8653065502643585,
      "num_tokens": 61684386.0,
      "step": 30170
    },
    {
      "entropy": 0.521889213193208,
      "epoch": 0.4901618443598093,
      "grad_norm": 10.1875,
      "learning_rate": 2.6434587210624263e-05,
      "loss": 0.5686691284179688,
      "mean_token_accuracy": 0.8573576383292675,
      "num_tokens": 61704673.0,
      "step": 30180
    },
    {
      "entropy": 0.5135290110483766,
      "epoch": 0.49032425716443484,
      "grad_norm": 9.875,
      "learning_rate": 2.6421642027295236e-05,
      "loss": 0.5327964305877686,
      "mean_token_accuracy": 0.8700820945203305,
      "num_tokens": 61725753.0,
      "step": 30190
    },
    {
      "entropy": 0.4099705256987363,
      "epoch": 0.4904866699690604,
      "grad_norm": 6.125,
      "learning_rate": 2.6408696461541303e-05,
      "loss": 0.4193225383758545,
      "mean_token_accuracy": 0.897350399941206,
      "num_tokens": 61747581.0,
      "step": 30200
    },
    {
      "entropy": 0.4896069903392345,
      "epoch": 0.49064908277368585,
      "grad_norm": 7.6875,
      "learning_rate": 2.6395750516844863e-05,
      "loss": 0.47652711868286135,
      "mean_token_accuracy": 0.8685698784887791,
      "num_tokens": 61768406.0,
      "step": 30210
    },
    {
      "entropy": 0.4952986028511077,
      "epoch": 0.4908114955783114,
      "grad_norm": 6.34375,
      "learning_rate": 2.6382804196688394e-05,
      "loss": 0.4614094257354736,
      "mean_token_accuracy": 0.8882255688309669,
      "num_tokens": 61789082.0,
      "step": 30220
    },
    {
      "entropy": 0.4760035391896963,
      "epoch": 0.4909739083829369,
      "grad_norm": 6.90625,
      "learning_rate": 2.636985750455449e-05,
      "loss": 0.4864801406860352,
      "mean_token_accuracy": 0.8773373775184155,
      "num_tokens": 61809169.0,
      "step": 30230
    },
    {
      "entropy": 0.5891461042687297,
      "epoch": 0.49113632118756245,
      "grad_norm": 6.75,
      "learning_rate": 2.6356910443925837e-05,
      "loss": 0.5992877006530761,
      "mean_token_accuracy": 0.8500635012984276,
      "num_tokens": 61830123.0,
      "step": 30240
    },
    {
      "entropy": 0.501731023634784,
      "epoch": 0.49129873399218793,
      "grad_norm": 10.25,
      "learning_rate": 2.6343963018285235e-05,
      "loss": 0.5126551628112793,
      "mean_token_accuracy": 0.8749392934143543,
      "num_tokens": 61852243.0,
      "step": 30250
    },
    {
      "entropy": 0.49700449258089063,
      "epoch": 0.49146114679681346,
      "grad_norm": 9.5625,
      "learning_rate": 2.6331015231115542e-05,
      "loss": 0.513553524017334,
      "mean_token_accuracy": 0.8654677368700504,
      "num_tokens": 61872540.0,
      "step": 30260
    },
    {
      "entropy": 0.4762661008164287,
      "epoch": 0.491623559601439,
      "grad_norm": 9.4375,
      "learning_rate": 2.6318067085899777e-05,
      "loss": 0.4792464256286621,
      "mean_token_accuracy": 0.8799596525728702,
      "num_tokens": 61893221.0,
      "step": 30270
    },
    {
      "entropy": 0.5100882638711483,
      "epoch": 0.4917859724060645,
      "grad_norm": 8.5625,
      "learning_rate": 2.6305118586120998e-05,
      "loss": 0.5225683212280273,
      "mean_token_accuracy": 0.8678221181035042,
      "num_tokens": 61913529.0,
      "step": 30280
    },
    {
      "entropy": 0.4800251376349479,
      "epoch": 0.49194838521069,
      "grad_norm": 7.21875,
      "learning_rate": 2.629216973526238e-05,
      "loss": 0.4860246181488037,
      "mean_token_accuracy": 0.8772603366523981,
      "num_tokens": 61933456.0,
      "step": 30290
    },
    {
      "entropy": 0.5236603042110801,
      "epoch": 0.49211079801531554,
      "grad_norm": 7.875,
      "learning_rate": 2.6279220536807204e-05,
      "loss": 0.5455100059509277,
      "mean_token_accuracy": 0.8637721229344606,
      "num_tokens": 61954176.0,
      "step": 30300
    },
    {
      "entropy": 0.5543377702590078,
      "epoch": 0.492273210819941,
      "grad_norm": 7.40625,
      "learning_rate": 2.6266270994238828e-05,
      "loss": 0.5468708992004394,
      "mean_token_accuracy": 0.8660032354295254,
      "num_tokens": 61974333.0,
      "step": 30310
    },
    {
      "entropy": 0.4878885676851496,
      "epoch": 0.49243562362456655,
      "grad_norm": 7.5,
      "learning_rate": 2.625332111104072e-05,
      "loss": 0.5142910003662109,
      "mean_token_accuracy": 0.8704902812838554,
      "num_tokens": 61996078.0,
      "step": 30320
    },
    {
      "entropy": 0.5132138452958316,
      "epoch": 0.4925980364291921,
      "grad_norm": 6.46875,
      "learning_rate": 2.6240370890696403e-05,
      "loss": 0.5062345027923584,
      "mean_token_accuracy": 0.8702041000127793,
      "num_tokens": 62014909.0,
      "step": 30330
    },
    {
      "entropy": 0.5446961538866162,
      "epoch": 0.4927604492338176,
      "grad_norm": 10.375,
      "learning_rate": 2.622742033668953e-05,
      "loss": 0.5493142604827881,
      "mean_token_accuracy": 0.8677078902721405,
      "num_tokens": 62035563.0,
      "step": 30340
    },
    {
      "entropy": 0.5544630941469222,
      "epoch": 0.4929228620384431,
      "grad_norm": 10.9375,
      "learning_rate": 2.621446945250383e-05,
      "loss": 0.570211124420166,
      "mean_token_accuracy": 0.8567478392273188,
      "num_tokens": 62056184.0,
      "step": 30350
    },
    {
      "entropy": 0.4904654514975846,
      "epoch": 0.49308527484306863,
      "grad_norm": 9.25,
      "learning_rate": 2.620151824162312e-05,
      "loss": 0.4716200351715088,
      "mean_token_accuracy": 0.8723306380212307,
      "num_tokens": 62076062.0,
      "step": 30360
    },
    {
      "entropy": 0.49410928809084,
      "epoch": 0.49324768764769417,
      "grad_norm": 9.25,
      "learning_rate": 2.61885667075313e-05,
      "loss": 0.49779596328735354,
      "mean_token_accuracy": 0.8639520108699799,
      "num_tokens": 62095762.0,
      "step": 30370
    },
    {
      "entropy": 0.5638323184568435,
      "epoch": 0.49341010045231964,
      "grad_norm": 8.9375,
      "learning_rate": 2.617561485371236e-05,
      "loss": 0.5657856464385986,
      "mean_token_accuracy": 0.8515051126480102,
      "num_tokens": 62117100.0,
      "step": 30380
    },
    {
      "entropy": 0.5236329250968993,
      "epoch": 0.4935725132569452,
      "grad_norm": 5.25,
      "learning_rate": 2.6162662683650397e-05,
      "loss": 0.4660347461700439,
      "mean_token_accuracy": 0.8810234762728214,
      "num_tokens": 62137127.0,
      "step": 30390
    },
    {
      "entropy": 0.47539623114280405,
      "epoch": 0.4937349260615707,
      "grad_norm": 7.375,
      "learning_rate": 2.6149710200829545e-05,
      "loss": 0.47725510597229004,
      "mean_token_accuracy": 0.8792604945600033,
      "num_tokens": 62156995.0,
      "step": 30400
    },
    {
      "entropy": 0.5026486547663808,
      "epoch": 0.4938973388661962,
      "grad_norm": 9.75,
      "learning_rate": 2.613675740873407e-05,
      "loss": 0.504782772064209,
      "mean_token_accuracy": 0.8726330913603306,
      "num_tokens": 62176071.0,
      "step": 30410
    },
    {
      "entropy": 0.46541387708857657,
      "epoch": 0.4940597516708217,
      "grad_norm": 8.5625,
      "learning_rate": 2.6123804310848298e-05,
      "loss": 0.47075252532958983,
      "mean_token_accuracy": 0.8812093429267407,
      "num_tokens": 62194942.0,
      "step": 30420
    },
    {
      "entropy": 0.4851918583270162,
      "epoch": 0.49422216447544726,
      "grad_norm": 10.6875,
      "learning_rate": 2.611085091065665e-05,
      "loss": 0.5061672687530517,
      "mean_token_accuracy": 0.8788064964115619,
      "num_tokens": 62215478.0,
      "step": 30430
    },
    {
      "entropy": 0.5110025990288705,
      "epoch": 0.49438457728007273,
      "grad_norm": 13.3125,
      "learning_rate": 2.6097897211643597e-05,
      "loss": 0.5120035171508789,
      "mean_token_accuracy": 0.8699561581015587,
      "num_tokens": 62235550.0,
      "step": 30440
    },
    {
      "entropy": 0.5003997090272605,
      "epoch": 0.49454699008469827,
      "grad_norm": 8.625,
      "learning_rate": 2.608494321729373e-05,
      "loss": 0.49400973320007324,
      "mean_token_accuracy": 0.8729841209948063,
      "num_tokens": 62255440.0,
      "step": 30450
    },
    {
      "entropy": 0.4641249504871666,
      "epoch": 0.4947094028893238,
      "grad_norm": 5.34375,
      "learning_rate": 2.6071988931091695e-05,
      "loss": 0.4550004959106445,
      "mean_token_accuracy": 0.8866696208715439,
      "num_tokens": 62275552.0,
      "step": 30460
    },
    {
      "entropy": 0.5039339892799035,
      "epoch": 0.49487181569394934,
      "grad_norm": 7.625,
      "learning_rate": 2.6059034356522234e-05,
      "loss": 0.5211454868316651,
      "mean_token_accuracy": 0.8695507273077965,
      "num_tokens": 62296710.0,
      "step": 30470
    },
    {
      "entropy": 0.46454361209180206,
      "epoch": 0.4950342284985748,
      "grad_norm": 5.6875,
      "learning_rate": 2.604607949707016e-05,
      "loss": 0.46201448440551757,
      "mean_token_accuracy": 0.882608300447464,
      "num_tokens": 62317282.0,
      "step": 30480
    },
    {
      "entropy": 0.4845169786014594,
      "epoch": 0.49519664130320035,
      "grad_norm": 7.40625,
      "learning_rate": 2.6033124356220328e-05,
      "loss": 0.4958409309387207,
      "mean_token_accuracy": 0.873705892264843,
      "num_tokens": 62337931.0,
      "step": 30490
    },
    {
      "entropy": 0.4919731826521456,
      "epoch": 0.4953590541078259,
      "grad_norm": 10.125,
      "learning_rate": 2.6020168937457744e-05,
      "loss": 0.5272331714630127,
      "mean_token_accuracy": 0.8704997308552265,
      "num_tokens": 62357696.0,
      "step": 30500
    },
    {
      "entropy": 0.4228378449101001,
      "epoch": 0.49552146691245136,
      "grad_norm": 8.75,
      "learning_rate": 2.6007213244267415e-05,
      "loss": 0.4252029895782471,
      "mean_token_accuracy": 0.8881394982337951,
      "num_tokens": 62377249.0,
      "step": 30510
    },
    {
      "entropy": 0.4533903363160789,
      "epoch": 0.4956838797170769,
      "grad_norm": 9.25,
      "learning_rate": 2.5994257280134455e-05,
      "loss": 0.45027670860290525,
      "mean_token_accuracy": 0.8905290596187114,
      "num_tokens": 62398389.0,
      "step": 30520
    },
    {
      "entropy": 0.4873589690774679,
      "epoch": 0.4958462925217024,
      "grad_norm": 6.75,
      "learning_rate": 2.5981301048544055e-05,
      "loss": 0.5163526058197021,
      "mean_token_accuracy": 0.871699383109808,
      "num_tokens": 62417621.0,
      "step": 30530
    },
    {
      "entropy": 0.46431480599567293,
      "epoch": 0.4960087053263279,
      "grad_norm": 10.3125,
      "learning_rate": 2.5968344552981465e-05,
      "loss": 0.4917804718017578,
      "mean_token_accuracy": 0.8729633808135986,
      "num_tokens": 62437578.0,
      "step": 30540
    },
    {
      "entropy": 0.5074082003440707,
      "epoch": 0.49617111813095344,
      "grad_norm": 6.5625,
      "learning_rate": 2.5955387796932022e-05,
      "loss": 0.47925896644592286,
      "mean_token_accuracy": 0.8755578853189945,
      "num_tokens": 62458133.0,
      "step": 30550
    },
    {
      "entropy": 0.4348384527023882,
      "epoch": 0.49633353093557897,
      "grad_norm": 8.875,
      "learning_rate": 2.5942430783881104e-05,
      "loss": 0.445039701461792,
      "mean_token_accuracy": 0.8834964036941528,
      "num_tokens": 62478307.0,
      "step": 30560
    },
    {
      "entropy": 0.47746557565405967,
      "epoch": 0.4964959437402045,
      "grad_norm": 7.875,
      "learning_rate": 2.59294735173142e-05,
      "loss": 0.44989709854125975,
      "mean_token_accuracy": 0.8790071040391922,
      "num_tokens": 62499555.0,
      "step": 30570
    },
    {
      "entropy": 0.5187020897865295,
      "epoch": 0.49665835654483,
      "grad_norm": 6.09375,
      "learning_rate": 2.5916516000716822e-05,
      "loss": 0.5426670551300049,
      "mean_token_accuracy": 0.8655966587364674,
      "num_tokens": 62520494.0,
      "step": 30580
    },
    {
      "entropy": 0.4227032272145152,
      "epoch": 0.4968207693494555,
      "grad_norm": 8.0625,
      "learning_rate": 2.5903558237574594e-05,
      "loss": 0.41092934608459475,
      "mean_token_accuracy": 0.8920796245336533,
      "num_tokens": 62540511.0,
      "step": 30590
    },
    {
      "entropy": 0.5064334815368057,
      "epoch": 0.49698318215408105,
      "grad_norm": 10.125,
      "learning_rate": 2.5890600231373147e-05,
      "loss": 0.5478888988494873,
      "mean_token_accuracy": 0.8684739585965872,
      "num_tokens": 62559843.0,
      "step": 30600
    },
    {
      "entropy": 0.4649269653949887,
      "epoch": 0.4971455949587065,
      "grad_norm": 11.0,
      "learning_rate": 2.5877641985598254e-05,
      "loss": 0.46546039581298826,
      "mean_token_accuracy": 0.8835222393274307,
      "num_tokens": 62578645.0,
      "step": 30610
    },
    {
      "entropy": 0.4410695023369044,
      "epoch": 0.49730800776333206,
      "grad_norm": 7.03125,
      "learning_rate": 2.5864683503735688e-05,
      "loss": 0.4625762939453125,
      "mean_token_accuracy": 0.8809315547347069,
      "num_tokens": 62598959.0,
      "step": 30620
    },
    {
      "entropy": 0.5100359302014112,
      "epoch": 0.4974704205679576,
      "grad_norm": 11.5625,
      "learning_rate": 2.5851724789271315e-05,
      "loss": 0.5133030414581299,
      "mean_token_accuracy": 0.8688674956560135,
      "num_tokens": 62619936.0,
      "step": 30630
    },
    {
      "entropy": 0.4976271322928369,
      "epoch": 0.49763283337258307,
      "grad_norm": 6.96875,
      "learning_rate": 2.583876584569106e-05,
      "loss": 0.4955767631530762,
      "mean_token_accuracy": 0.8729942969977855,
      "num_tokens": 62640492.0,
      "step": 30640
    },
    {
      "entropy": 0.43746135646943,
      "epoch": 0.4977952461772086,
      "grad_norm": 7.65625,
      "learning_rate": 2.5825806676480907e-05,
      "loss": 0.4749462127685547,
      "mean_token_accuracy": 0.8774041071534157,
      "num_tokens": 62659339.0,
      "step": 30650
    },
    {
      "entropy": 0.4396726899314672,
      "epoch": 0.49795765898183414,
      "grad_norm": 5.84375,
      "learning_rate": 2.5812847285126906e-05,
      "loss": 0.4430410861968994,
      "mean_token_accuracy": 0.8917746014893055,
      "num_tokens": 62678698.0,
      "step": 30660
    },
    {
      "entropy": 0.480262959189713,
      "epoch": 0.4981200717864596,
      "grad_norm": 9.9375,
      "learning_rate": 2.5799887675115147e-05,
      "loss": 0.4709348201751709,
      "mean_token_accuracy": 0.8819316975772381,
      "num_tokens": 62699193.0,
      "step": 30670
    },
    {
      "entropy": 0.5463363481219858,
      "epoch": 0.49828248459108515,
      "grad_norm": 8.4375,
      "learning_rate": 2.5786927849931825e-05,
      "loss": 0.5104465961456299,
      "mean_token_accuracy": 0.8709375485777855,
      "num_tokens": 62719791.0,
      "step": 30680
    },
    {
      "entropy": 0.5104331964626908,
      "epoch": 0.4984448973957107,
      "grad_norm": 6.125,
      "learning_rate": 2.5773967813063125e-05,
      "loss": 0.5357676029205323,
      "mean_token_accuracy": 0.8564973168075085,
      "num_tokens": 62739928.0,
      "step": 30690
    },
    {
      "entropy": 0.48209630395285785,
      "epoch": 0.4986073102003362,
      "grad_norm": 8.625,
      "learning_rate": 2.5761007567995353e-05,
      "loss": 0.45403451919555665,
      "mean_token_accuracy": 0.8767270036041737,
      "num_tokens": 62760333.0,
      "step": 30700
    },
    {
      "entropy": 0.5153867456130683,
      "epoch": 0.4987697230049617,
      "grad_norm": 13.1875,
      "learning_rate": 2.5748047118214835e-05,
      "loss": 0.5748306274414062,
      "mean_token_accuracy": 0.855607844889164,
      "num_tokens": 62780264.0,
      "step": 30710
    },
    {
      "entropy": 0.5527590315788984,
      "epoch": 0.49893213580958723,
      "grad_norm": 13.25,
      "learning_rate": 2.573508646720796e-05,
      "loss": 0.5738747119903564,
      "mean_token_accuracy": 0.8579062327742577,
      "num_tokens": 62800186.0,
      "step": 30720
    },
    {
      "entropy": 0.4992442848160863,
      "epoch": 0.49909454861421276,
      "grad_norm": 10.0625,
      "learning_rate": 2.5722125618461175e-05,
      "loss": 0.4925652027130127,
      "mean_token_accuracy": 0.8746158614754677,
      "num_tokens": 62822757.0,
      "step": 30730
    },
    {
      "entropy": 0.5208871163893491,
      "epoch": 0.49925696141883824,
      "grad_norm": 8.5625,
      "learning_rate": 2.570916457546097e-05,
      "loss": 0.5250263214111328,
      "mean_token_accuracy": 0.8665183283388614,
      "num_tokens": 62843430.0,
      "step": 30740
    },
    {
      "entropy": 0.4987202765420079,
      "epoch": 0.4994193742234638,
      "grad_norm": 10.6875,
      "learning_rate": 2.5696203341693902e-05,
      "loss": 0.5430459022521973,
      "mean_token_accuracy": 0.8693216603249312,
      "num_tokens": 62863427.0,
      "step": 30750
    },
    {
      "entropy": 0.5829684255179017,
      "epoch": 0.4995817870280893,
      "grad_norm": 7.15625,
      "learning_rate": 2.5683241920646572e-05,
      "loss": 0.601649808883667,
      "mean_token_accuracy": 0.8550139106810093,
      "num_tokens": 62883846.0,
      "step": 30760
    },
    {
      "entropy": 0.5745293566025793,
      "epoch": 0.4997441998327148,
      "grad_norm": 7.96875,
      "learning_rate": 2.567028031580564e-05,
      "loss": 0.6170824527740478,
      "mean_token_accuracy": 0.8481758117675782,
      "num_tokens": 62904156.0,
      "step": 30770
    },
    {
      "entropy": 0.5470445358194411,
      "epoch": 0.4999066126373403,
      "grad_norm": 8.75,
      "learning_rate": 2.565731853065778e-05,
      "loss": 0.5533169269561767,
      "mean_token_accuracy": 0.8700152479112149,
      "num_tokens": 62925795.0,
      "step": 30780
    },
    {
      "entropy": 0.5426269664429129,
      "epoch": 0.5000690254419659,
      "grad_norm": 4.21875,
      "learning_rate": 2.5644356568689763e-05,
      "loss": 0.5475377559661865,
      "mean_token_accuracy": 0.8683408368378878,
      "num_tokens": 62946833.0,
      "step": 30790
    },
    {
      "entropy": 0.532044337131083,
      "epoch": 0.5002314382465913,
      "grad_norm": 8.8125,
      "learning_rate": 2.5631394433388385e-05,
      "loss": 0.490903377532959,
      "mean_token_accuracy": 0.8670671299099922,
      "num_tokens": 62967548.0,
      "step": 30800
    },
    {
      "entropy": 0.4674599841004238,
      "epoch": 0.5003938510512169,
      "grad_norm": 11.6875,
      "learning_rate": 2.5618432128240483e-05,
      "loss": 0.4400550365447998,
      "mean_token_accuracy": 0.8835840925574303,
      "num_tokens": 62987169.0,
      "step": 30810
    },
    {
      "entropy": 0.45229013967327775,
      "epoch": 0.5005562638558424,
      "grad_norm": 6.0,
      "learning_rate": 2.5605469656732954e-05,
      "loss": 0.4245868682861328,
      "mean_token_accuracy": 0.8871519461274147,
      "num_tokens": 63007221.0,
      "step": 30820
    },
    {
      "entropy": 0.49253651523031294,
      "epoch": 0.5007186766604679,
      "grad_norm": 16.75,
      "learning_rate": 2.5592507022352713e-05,
      "loss": 0.48865370750427245,
      "mean_token_accuracy": 0.8773426309227943,
      "num_tokens": 63028681.0,
      "step": 30830
    },
    {
      "entropy": 0.4299283406231552,
      "epoch": 0.5008810894650935,
      "grad_norm": 6.4375,
      "learning_rate": 2.5579544228586767e-05,
      "loss": 0.43002886772155763,
      "mean_token_accuracy": 0.8833753518760205,
      "num_tokens": 63049481.0,
      "step": 30840
    },
    {
      "entropy": 0.5228127751033753,
      "epoch": 0.5010435022697189,
      "grad_norm": 8.8125,
      "learning_rate": 2.5566581278922113e-05,
      "loss": 0.5524937152862549,
      "mean_token_accuracy": 0.8659674070775509,
      "num_tokens": 63071360.0,
      "step": 30850
    },
    {
      "entropy": 0.5086765705840662,
      "epoch": 0.5012059150743444,
      "grad_norm": 7.59375,
      "learning_rate": 2.5553618176845824e-05,
      "loss": 0.5200385093688965,
      "mean_token_accuracy": 0.8698768258094788,
      "num_tokens": 63092473.0,
      "step": 30860
    },
    {
      "entropy": 0.4547089242376387,
      "epoch": 0.50136832787897,
      "grad_norm": 6.34375,
      "learning_rate": 2.5540654925844997e-05,
      "loss": 0.4663630485534668,
      "mean_token_accuracy": 0.8817520800977945,
      "num_tokens": 63111948.0,
      "step": 30870
    },
    {
      "entropy": 0.4980192950926721,
      "epoch": 0.5015307406835955,
      "grad_norm": 8.625,
      "learning_rate": 2.5527691529406785e-05,
      "loss": 0.5350961208343505,
      "mean_token_accuracy": 0.863745704293251,
      "num_tokens": 63132471.0,
      "step": 30880
    },
    {
      "entropy": 0.45823753252625465,
      "epoch": 0.501693153488221,
      "grad_norm": 6.71875,
      "learning_rate": 2.5514727991018366e-05,
      "loss": 0.45459766387939454,
      "mean_token_accuracy": 0.8867032773792743,
      "num_tokens": 63152876.0,
      "step": 30890
    },
    {
      "entropy": 0.5019978042691946,
      "epoch": 0.5018555662928466,
      "grad_norm": 6.53125,
      "learning_rate": 2.5501764314166948e-05,
      "loss": 0.49185924530029296,
      "mean_token_accuracy": 0.876114359870553,
      "num_tokens": 63173810.0,
      "step": 30900
    },
    {
      "entropy": 0.46008493879344314,
      "epoch": 0.502017979097472,
      "grad_norm": 7.8125,
      "learning_rate": 2.548880050233982e-05,
      "loss": 0.4573627471923828,
      "mean_token_accuracy": 0.8838027566671371,
      "num_tokens": 63193253.0,
      "step": 30910
    },
    {
      "entropy": 0.4920669169630855,
      "epoch": 0.5021803919020975,
      "grad_norm": 8.5,
      "learning_rate": 2.5475836559024245e-05,
      "loss": 0.5201551914215088,
      "mean_token_accuracy": 0.8709423765540123,
      "num_tokens": 63213476.0,
      "step": 30920
    },
    {
      "entropy": 0.5327320579439402,
      "epoch": 0.5023428047067231,
      "grad_norm": 6.5,
      "learning_rate": 2.5462872487707578e-05,
      "loss": 0.5551751136779786,
      "mean_token_accuracy": 0.8605267081409693,
      "num_tokens": 63234608.0,
      "step": 30930
    },
    {
      "entropy": 0.4709260633215308,
      "epoch": 0.5025052175113486,
      "grad_norm": 10.5,
      "learning_rate": 2.5449908291877163e-05,
      "loss": 0.4904721736907959,
      "mean_token_accuracy": 0.874935220927,
      "num_tokens": 63254706.0,
      "step": 30940
    },
    {
      "entropy": 0.49455576394684614,
      "epoch": 0.5026676303159741,
      "grad_norm": 7.1875,
      "learning_rate": 2.543694397502041e-05,
      "loss": 0.5085570812225342,
      "mean_token_accuracy": 0.8718381352722645,
      "num_tokens": 63275045.0,
      "step": 30950
    },
    {
      "entropy": 0.5725363776087761,
      "epoch": 0.5028300431205996,
      "grad_norm": 7.125,
      "learning_rate": 2.5423979540624743e-05,
      "loss": 0.5924774169921875,
      "mean_token_accuracy": 0.8532403245568275,
      "num_tokens": 63295595.0,
      "step": 30960
    },
    {
      "entropy": 0.5804558306001126,
      "epoch": 0.5029924559252251,
      "grad_norm": 9.9375,
      "learning_rate": 2.541101499217763e-05,
      "loss": 0.5645534992218018,
      "mean_token_accuracy": 0.8602942600846291,
      "num_tokens": 63314737.0,
      "step": 30970
    },
    {
      "entropy": 0.5781186825595797,
      "epoch": 0.5031548687298507,
      "grad_norm": 8.375,
      "learning_rate": 2.5398050333166556e-05,
      "loss": 0.6102898597717286,
      "mean_token_accuracy": 0.8500448048114777,
      "num_tokens": 63336456.0,
      "step": 30980
    },
    {
      "entropy": 0.5248013722710312,
      "epoch": 0.5033172815344762,
      "grad_norm": 8.625,
      "learning_rate": 2.538508556707905e-05,
      "loss": 0.5418924808502197,
      "mean_token_accuracy": 0.8645344406366349,
      "num_tokens": 63357546.0,
      "step": 30990
    },
    {
      "entropy": 0.5431352938525379,
      "epoch": 0.5034796943391017,
      "grad_norm": 6.9375,
      "learning_rate": 2.5372120697402668e-05,
      "loss": 0.545280933380127,
      "mean_token_accuracy": 0.869857943803072,
      "num_tokens": 63377496.0,
      "step": 31000
    },
    {
      "entropy": 0.5434500108938665,
      "epoch": 0.5036421071437273,
      "grad_norm": 6.03125,
      "learning_rate": 2.5359155727624967e-05,
      "loss": 0.5474634647369385,
      "mean_token_accuracy": 0.8586061395704746,
      "num_tokens": 63397760.0,
      "step": 31010
    },
    {
      "entropy": 0.5027830949053168,
      "epoch": 0.5038045199483527,
      "grad_norm": 8.1875,
      "learning_rate": 2.534619066123358e-05,
      "loss": 0.4951018810272217,
      "mean_token_accuracy": 0.8699811305850744,
      "num_tokens": 63417909.0,
      "step": 31020
    },
    {
      "entropy": 0.48969662794843316,
      "epoch": 0.5039669327529782,
      "grad_norm": 8.4375,
      "learning_rate": 2.5333225501716122e-05,
      "loss": 0.4860862731933594,
      "mean_token_accuracy": 0.8704923883080482,
      "num_tokens": 63438182.0,
      "step": 31030
    },
    {
      "entropy": 0.4967236865311861,
      "epoch": 0.5041293455576038,
      "grad_norm": 6.8125,
      "learning_rate": 2.5320260252560257e-05,
      "loss": 0.5233484268188476,
      "mean_token_accuracy": 0.870791421085596,
      "num_tokens": 63458102.0,
      "step": 31040
    },
    {
      "entropy": 0.4368639747845009,
      "epoch": 0.5042917583622293,
      "grad_norm": 5.65625,
      "learning_rate": 2.5307294917253667e-05,
      "loss": 0.41996021270751954,
      "mean_token_accuracy": 0.8908371031284332,
      "num_tokens": 63478270.0,
      "step": 31050
    },
    {
      "entropy": 0.5225863297469914,
      "epoch": 0.5044541711668548,
      "grad_norm": 6.21875,
      "learning_rate": 2.5294329499284047e-05,
      "loss": 0.5451902866363525,
      "mean_token_accuracy": 0.8659935861825943,
      "num_tokens": 63499471.0,
      "step": 31060
    },
    {
      "entropy": 0.5028060196666047,
      "epoch": 0.5046165839714803,
      "grad_norm": 9.0,
      "learning_rate": 2.528136400213914e-05,
      "loss": 0.5199184894561768,
      "mean_token_accuracy": 0.8694649368524552,
      "num_tokens": 63519591.0,
      "step": 31070
    },
    {
      "entropy": 0.5243601767346263,
      "epoch": 0.5047789967761058,
      "grad_norm": 8.5625,
      "learning_rate": 2.5268398429306677e-05,
      "loss": 0.4969650745391846,
      "mean_token_accuracy": 0.8783114463090896,
      "num_tokens": 63539936.0,
      "step": 31080
    },
    {
      "entropy": 0.5960184166207909,
      "epoch": 0.5049414095807313,
      "grad_norm": 8.25,
      "learning_rate": 2.525543278427443e-05,
      "loss": 0.592011547088623,
      "mean_token_accuracy": 0.8490048408508301,
      "num_tokens": 63560489.0,
      "step": 31090
    },
    {
      "entropy": 0.580484717618674,
      "epoch": 0.5051038223853569,
      "grad_norm": 11.5,
      "learning_rate": 2.5242467070530195e-05,
      "loss": 0.5620294094085694,
      "mean_token_accuracy": 0.8598048843443393,
      "num_tokens": 63581026.0,
      "step": 31100
    },
    {
      "entropy": 0.5274917416274547,
      "epoch": 0.5052662351899824,
      "grad_norm": 9.3125,
      "learning_rate": 2.522950129156178e-05,
      "loss": 0.5459061622619629,
      "mean_token_accuracy": 0.8565484628081321,
      "num_tokens": 63601413.0,
      "step": 31110
    },
    {
      "entropy": 0.47519553927704694,
      "epoch": 0.5054286479946078,
      "grad_norm": 6.0625,
      "learning_rate": 2.5216535450856994e-05,
      "loss": 0.4461238384246826,
      "mean_token_accuracy": 0.8849045678973197,
      "num_tokens": 63621908.0,
      "step": 31120
    },
    {
      "entropy": 0.5771022098138928,
      "epoch": 0.5055910607992334,
      "grad_norm": 6.84375,
      "learning_rate": 2.5203569551903682e-05,
      "loss": 0.591285514831543,
      "mean_token_accuracy": 0.8545915607362986,
      "num_tokens": 63643462.0,
      "step": 31130
    },
    {
      "entropy": 0.5222721460275352,
      "epoch": 0.5057534736038589,
      "grad_norm": 7.5625,
      "learning_rate": 2.5190603598189704e-05,
      "loss": 0.565674877166748,
      "mean_token_accuracy": 0.8669797025620938,
      "num_tokens": 63664113.0,
      "step": 31140
    },
    {
      "entropy": 0.5848565527237952,
      "epoch": 0.5059158864084844,
      "grad_norm": 6.34375,
      "learning_rate": 2.5177637593202935e-05,
      "loss": 0.5782303810119629,
      "mean_token_accuracy": 0.8618705078959465,
      "num_tokens": 63685544.0,
      "step": 31150
    },
    {
      "entropy": 0.5247034734115005,
      "epoch": 0.50607829921311,
      "grad_norm": 6.78125,
      "learning_rate": 2.516467154043125e-05,
      "loss": 0.5311241149902344,
      "mean_token_accuracy": 0.8705053735524416,
      "num_tokens": 63705578.0,
      "step": 31160
    },
    {
      "entropy": 0.508901415579021,
      "epoch": 0.5062407120177355,
      "grad_norm": 5.53125,
      "learning_rate": 2.515170544336255e-05,
      "loss": 0.48520455360412595,
      "mean_token_accuracy": 0.8807548485696316,
      "num_tokens": 63725717.0,
      "step": 31170
    },
    {
      "entropy": 0.5674579416401684,
      "epoch": 0.506403124822361,
      "grad_norm": 9.3125,
      "learning_rate": 2.513873930548475e-05,
      "loss": 0.5618189811706543,
      "mean_token_accuracy": 0.8561816483736038,
      "num_tokens": 63746195.0,
      "step": 31180
    },
    {
      "entropy": 0.5311344156973064,
      "epoch": 0.5065655376269865,
      "grad_norm": 7.5625,
      "learning_rate": 2.512577313028576e-05,
      "loss": 0.5378102779388427,
      "mean_token_accuracy": 0.8680837921798229,
      "num_tokens": 63767194.0,
      "step": 31190
    },
    {
      "entropy": 0.47301930682733656,
      "epoch": 0.506727950431612,
      "grad_norm": 9.1875,
      "learning_rate": 2.511280692125352e-05,
      "loss": 0.4473231315612793,
      "mean_token_accuracy": 0.881066520512104,
      "num_tokens": 63786718.0,
      "step": 31200
    },
    {
      "entropy": 0.5166785879526288,
      "epoch": 0.5068903632362376,
      "grad_norm": 12.5,
      "learning_rate": 2.509984068187596e-05,
      "loss": 0.5207250118255615,
      "mean_token_accuracy": 0.8656733512878418,
      "num_tokens": 63807722.0,
      "step": 31210
    },
    {
      "entropy": 0.5005590539425612,
      "epoch": 0.5070527760408631,
      "grad_norm": 7.25,
      "learning_rate": 2.5086874415641033e-05,
      "loss": 0.5609787464141845,
      "mean_token_accuracy": 0.8576747052371502,
      "num_tokens": 63827510.0,
      "step": 31220
    },
    {
      "entropy": 0.5807335109682754,
      "epoch": 0.5072151888454886,
      "grad_norm": 9.6875,
      "learning_rate": 2.50739081260367e-05,
      "loss": 0.6269726753234863,
      "mean_token_accuracy": 0.8458425734192133,
      "num_tokens": 63846283.0,
      "step": 31230
    },
    {
      "entropy": 0.45723277390934525,
      "epoch": 0.5073776016501141,
      "grad_norm": 8.3125,
      "learning_rate": 2.50609418165509e-05,
      "loss": 0.44083271026611326,
      "mean_token_accuracy": 0.8905237339437008,
      "num_tokens": 63865923.0,
      "step": 31240
    },
    {
      "entropy": 0.5534209537319839,
      "epoch": 0.5075400144547396,
      "grad_norm": 6.4375,
      "learning_rate": 2.5047975490671627e-05,
      "loss": 0.523739242553711,
      "mean_token_accuracy": 0.8712362129241228,
      "num_tokens": 63886664.0,
      "step": 31250
    },
    {
      "entropy": 0.47124530901201067,
      "epoch": 0.5077024272593651,
      "grad_norm": 6.34375,
      "learning_rate": 2.5035009151886836e-05,
      "loss": 0.4780773162841797,
      "mean_token_accuracy": 0.872762282192707,
      "num_tokens": 63906926.0,
      "step": 31260
    },
    {
      "entropy": 0.5377285491907969,
      "epoch": 0.5078648400639907,
      "grad_norm": 13.0625,
      "learning_rate": 2.502204280368451e-05,
      "loss": 0.5421071529388428,
      "mean_token_accuracy": 0.8675767444074154,
      "num_tokens": 63926573.0,
      "step": 31270
    },
    {
      "entropy": 0.5739461981225759,
      "epoch": 0.5080272528686162,
      "grad_norm": 10.5625,
      "learning_rate": 2.5009076449552605e-05,
      "loss": 0.5823330879211426,
      "mean_token_accuracy": 0.8554668717086316,
      "num_tokens": 63948250.0,
      "step": 31280
    },
    {
      "entropy": 0.4982933820458129,
      "epoch": 0.5081896656732416,
      "grad_norm": 9.9375,
      "learning_rate": 2.4996110092979132e-05,
      "loss": 0.5096698760986328,
      "mean_token_accuracy": 0.8730714991688728,
      "num_tokens": 63968639.0,
      "step": 31290
    },
    {
      "entropy": 0.5116185680963099,
      "epoch": 0.5083520784778672,
      "grad_norm": 7.1875,
      "learning_rate": 2.4983143737452036e-05,
      "loss": 0.5201897144317627,
      "mean_token_accuracy": 0.8637016639113426,
      "num_tokens": 63989879.0,
      "step": 31300
    },
    {
      "entropy": 0.5345291260629892,
      "epoch": 0.5085144912824927,
      "grad_norm": 8.9375,
      "learning_rate": 2.4970177386459325e-05,
      "loss": 0.5226802349090576,
      "mean_token_accuracy": 0.8631611682474614,
      "num_tokens": 64010612.0,
      "step": 31310
    },
    {
      "entropy": 0.5360772182233632,
      "epoch": 0.5086769040871182,
      "grad_norm": 6.53125,
      "learning_rate": 2.495721104348896e-05,
      "loss": 0.5293684959411621,
      "mean_token_accuracy": 0.876145800948143,
      "num_tokens": 64030120.0,
      "step": 31320
    },
    {
      "entropy": 0.5112530216341838,
      "epoch": 0.5088393168917438,
      "grad_norm": 8.4375,
      "learning_rate": 2.494424471202892e-05,
      "loss": 0.5075409889221192,
      "mean_token_accuracy": 0.8725536562502384,
      "num_tokens": 64050784.0,
      "step": 31330
    },
    {
      "entropy": 0.48728248351253567,
      "epoch": 0.5090017296963693,
      "grad_norm": 10.8125,
      "learning_rate": 2.493127839556718e-05,
      "loss": 0.5027622699737548,
      "mean_token_accuracy": 0.8670290302485227,
      "num_tokens": 64070428.0,
      "step": 31340
    },
    {
      "entropy": 0.4942189686000347,
      "epoch": 0.5091641425009947,
      "grad_norm": 7.5625,
      "learning_rate": 2.4918312097591702e-05,
      "loss": 0.49208431243896483,
      "mean_token_accuracy": 0.8711544938385487,
      "num_tokens": 64090718.0,
      "step": 31350
    },
    {
      "entropy": 0.49619610672816633,
      "epoch": 0.5093265553056203,
      "grad_norm": 7.0,
      "learning_rate": 2.490534582159046e-05,
      "loss": 0.500821304321289,
      "mean_token_accuracy": 0.868914358317852,
      "num_tokens": 64112181.0,
      "step": 31360
    },
    {
      "entropy": 0.4967054320499301,
      "epoch": 0.5094889681102458,
      "grad_norm": 7.15625,
      "learning_rate": 2.48923795710514e-05,
      "loss": 0.48534693717956545,
      "mean_token_accuracy": 0.8766528733074666,
      "num_tokens": 64131941.0,
      "step": 31370
    },
    {
      "entropy": 0.5339177510235459,
      "epoch": 0.5096513809148713,
      "grad_norm": 7.125,
      "learning_rate": 2.4879413349462473e-05,
      "loss": 0.5206130981445313,
      "mean_token_accuracy": 0.8724849846214056,
      "num_tokens": 64153536.0,
      "step": 31380
    },
    {
      "entropy": 0.4945803613867611,
      "epoch": 0.5098137937194969,
      "grad_norm": 10.5,
      "learning_rate": 2.4866447160311636e-05,
      "loss": 0.5093672275543213,
      "mean_token_accuracy": 0.8763057559728622,
      "num_tokens": 64173712.0,
      "step": 31390
    },
    {
      "entropy": 0.5137289888691157,
      "epoch": 0.5099762065241223,
      "grad_norm": 7.1875,
      "learning_rate": 2.4853481007086803e-05,
      "loss": 0.5466082572937012,
      "mean_token_accuracy": 0.8600450672209263,
      "num_tokens": 64192958.0,
      "step": 31400
    },
    {
      "entropy": 0.4881234313827008,
      "epoch": 0.5101386193287479,
      "grad_norm": 7.125,
      "learning_rate": 2.4840514893275914e-05,
      "loss": 0.46816329956054686,
      "mean_token_accuracy": 0.8737611286342144,
      "num_tokens": 64212687.0,
      "step": 31410
    },
    {
      "entropy": 0.43072693292051556,
      "epoch": 0.5103010321333734,
      "grad_norm": 6.59375,
      "learning_rate": 2.4827548822366877e-05,
      "loss": 0.42693161964416504,
      "mean_token_accuracy": 0.8838998459279537,
      "num_tokens": 64232216.0,
      "step": 31420
    },
    {
      "entropy": 0.45978101273067296,
      "epoch": 0.5104634449379989,
      "grad_norm": 8.125,
      "learning_rate": 2.4814582797847593e-05,
      "loss": 0.46312775611877444,
      "mean_token_accuracy": 0.8838886834681035,
      "num_tokens": 64251446.0,
      "step": 31430
    },
    {
      "entropy": 0.5035333069041371,
      "epoch": 0.5106258577426245,
      "grad_norm": 7.4375,
      "learning_rate": 2.480161682320596e-05,
      "loss": 0.5440796852111817,
      "mean_token_accuracy": 0.8676906295120717,
      "num_tokens": 64272414.0,
      "step": 31440
    },
    {
      "entropy": 0.5059620055370033,
      "epoch": 0.51078827054725,
      "grad_norm": 10.5,
      "learning_rate": 2.4788650901929842e-05,
      "loss": 0.5036671161651611,
      "mean_token_accuracy": 0.8809638246893883,
      "num_tokens": 64292301.0,
      "step": 31450
    },
    {
      "entropy": 0.4990936552174389,
      "epoch": 0.5109506833518754,
      "grad_norm": 10.0625,
      "learning_rate": 2.47756850375071e-05,
      "loss": 0.5343925952911377,
      "mean_token_accuracy": 0.8702289573848248,
      "num_tokens": 64311661.0,
      "step": 31460
    },
    {
      "entropy": 0.5140576610341668,
      "epoch": 0.511113096156501,
      "grad_norm": 8.3125,
      "learning_rate": 2.4762719233425595e-05,
      "loss": 0.5367635250091553,
      "mean_token_accuracy": 0.8706135220825673,
      "num_tokens": 64331457.0,
      "step": 31470
    },
    {
      "entropy": 0.564902906329371,
      "epoch": 0.5112755089611265,
      "grad_norm": 9.0625,
      "learning_rate": 2.4749753493173146e-05,
      "loss": 0.5571750640869141,
      "mean_token_accuracy": 0.8642473790794611,
      "num_tokens": 64352127.0,
      "step": 31480
    },
    {
      "entropy": 0.5148959412472323,
      "epoch": 0.511437921765752,
      "grad_norm": 7.125,
      "learning_rate": 2.473678782023757e-05,
      "loss": 0.522284460067749,
      "mean_token_accuracy": 0.8729870699346065,
      "num_tokens": 64373480.0,
      "step": 31490
    },
    {
      "entropy": 0.4708524478599429,
      "epoch": 0.5116003345703776,
      "grad_norm": 8.5625,
      "learning_rate": 2.4723822218106666e-05,
      "loss": 0.4871383190155029,
      "mean_token_accuracy": 0.8796190045773983,
      "num_tokens": 64394266.0,
      "step": 31500
    },
    {
      "entropy": 0.5530656065791846,
      "epoch": 0.511762747375003,
      "grad_norm": 6.9375,
      "learning_rate": 2.4710856690268196e-05,
      "loss": 0.5431325435638428,
      "mean_token_accuracy": 0.8641827218234539,
      "num_tokens": 64415238.0,
      "step": 31510
    },
    {
      "entropy": 0.5239378379890696,
      "epoch": 0.5119251601796285,
      "grad_norm": 8.0625,
      "learning_rate": 2.4697891240209933e-05,
      "loss": 0.49851303100585936,
      "mean_token_accuracy": 0.872267734259367,
      "num_tokens": 64436652.0,
      "step": 31520
    },
    {
      "entropy": 0.5107231616042555,
      "epoch": 0.5120875729842541,
      "grad_norm": 6.96875,
      "learning_rate": 2.468492587141961e-05,
      "loss": 0.4844343185424805,
      "mean_token_accuracy": 0.8753704614937305,
      "num_tokens": 64458019.0,
      "step": 31530
    },
    {
      "entropy": 0.44907164145261047,
      "epoch": 0.5122499857888796,
      "grad_norm": 7.8125,
      "learning_rate": 2.467196058738493e-05,
      "loss": 0.4586945533752441,
      "mean_token_accuracy": 0.8821403846144676,
      "num_tokens": 64478005.0,
      "step": 31540
    },
    {
      "entropy": 0.528500972664915,
      "epoch": 0.5124123985935051,
      "grad_norm": 5.84375,
      "learning_rate": 2.4658995391593593e-05,
      "loss": 0.5393474102020264,
      "mean_token_accuracy": 0.8670339226722718,
      "num_tokens": 64498309.0,
      "step": 31550
    },
    {
      "entropy": 0.4714937834069133,
      "epoch": 0.5125748113981307,
      "grad_norm": 6.4375,
      "learning_rate": 2.464603028753325e-05,
      "loss": 0.514289140701294,
      "mean_token_accuracy": 0.8751979980617761,
      "num_tokens": 64518926.0,
      "step": 31560
    },
    {
      "entropy": 0.4604245376540348,
      "epoch": 0.5127372242027561,
      "grad_norm": 9.0625,
      "learning_rate": 2.4633065278691567e-05,
      "loss": 0.4549668312072754,
      "mean_token_accuracy": 0.8814656987786293,
      "num_tokens": 64539533.0,
      "step": 31570
    },
    {
      "entropy": 0.48911209269426764,
      "epoch": 0.5128996370073816,
      "grad_norm": 7.28125,
      "learning_rate": 2.462010036855615e-05,
      "loss": 0.4882661819458008,
      "mean_token_accuracy": 0.8762936003506183,
      "num_tokens": 64560603.0,
      "step": 31580
    },
    {
      "entropy": 0.5008675747551024,
      "epoch": 0.5130620498120072,
      "grad_norm": 7.75,
      "learning_rate": 2.4607135560614583e-05,
      "loss": 0.510643196105957,
      "mean_token_accuracy": 0.8800863225013018,
      "num_tokens": 64579709.0,
      "step": 31590
    },
    {
      "entropy": 0.40986216969322414,
      "epoch": 0.5132244626166327,
      "grad_norm": 10.5,
      "learning_rate": 2.4594170858354427e-05,
      "loss": 0.3996280670166016,
      "mean_token_accuracy": 0.8924286879599095,
      "num_tokens": 64599982.0,
      "step": 31600
    },
    {
      "entropy": 0.5349804350640625,
      "epoch": 0.5133868754212582,
      "grad_norm": 8.625,
      "learning_rate": 2.458120626526324e-05,
      "loss": 0.5295713424682618,
      "mean_token_accuracy": 0.8756115455180407,
      "num_tokens": 64620576.0,
      "step": 31610
    },
    {
      "entropy": 0.45686810282059015,
      "epoch": 0.5135492882258837,
      "grad_norm": 11.75,
      "learning_rate": 2.456824178482849e-05,
      "loss": 0.4990899085998535,
      "mean_token_accuracy": 0.8711816787719726,
      "num_tokens": 64640474.0,
      "step": 31620
    },
    {
      "entropy": 0.5743285831529648,
      "epoch": 0.5137117010305092,
      "grad_norm": 5.625,
      "learning_rate": 2.4555277420537675e-05,
      "loss": 0.5815680503845215,
      "mean_token_accuracy": 0.8535645857453347,
      "num_tokens": 64660601.0,
      "step": 31630
    },
    {
      "entropy": 0.46380920028313993,
      "epoch": 0.5138741138351348,
      "grad_norm": 9.125,
      "learning_rate": 2.454231317587824e-05,
      "loss": 0.4839338302612305,
      "mean_token_accuracy": 0.8802500292658806,
      "num_tokens": 64679925.0,
      "step": 31640
    },
    {
      "entropy": 0.530510958051309,
      "epoch": 0.5140365266397603,
      "grad_norm": 6.4375,
      "learning_rate": 2.4529349054337587e-05,
      "loss": 0.5314265251159668,
      "mean_token_accuracy": 0.8676993541419507,
      "num_tokens": 64699973.0,
      "step": 31650
    },
    {
      "entropy": 0.5221024069003761,
      "epoch": 0.5141989394443858,
      "grad_norm": 7.90625,
      "learning_rate": 2.4516385059403095e-05,
      "loss": 0.5489985942840576,
      "mean_token_accuracy": 0.8641322743147611,
      "num_tokens": 64719177.0,
      "step": 31660
    },
    {
      "entropy": 0.45507272984832525,
      "epoch": 0.5143613522490114,
      "grad_norm": 8.125,
      "learning_rate": 2.4503421194562103e-05,
      "loss": 0.45491909980773926,
      "mean_token_accuracy": 0.8864268340170384,
      "num_tokens": 64739546.0,
      "step": 31670
    },
    {
      "entropy": 0.48251351070357484,
      "epoch": 0.5145237650536368,
      "grad_norm": 8.75,
      "learning_rate": 2.449045746330193e-05,
      "loss": 0.5065985202789307,
      "mean_token_accuracy": 0.8716546334326267,
      "num_tokens": 64759150.0,
      "step": 31680
    },
    {
      "entropy": 0.4678797002416104,
      "epoch": 0.5146861778582623,
      "grad_norm": 6.34375,
      "learning_rate": 2.4477493869109853e-05,
      "loss": 0.4818274974822998,
      "mean_token_accuracy": 0.8805451069027186,
      "num_tokens": 64780408.0,
      "step": 31690
    },
    {
      "entropy": 0.5049496863503009,
      "epoch": 0.5148485906628879,
      "grad_norm": 7.78125,
      "learning_rate": 2.4464530415473087e-05,
      "loss": 0.5188385486602783,
      "mean_token_accuracy": 0.8720504656434059,
      "num_tokens": 64801061.0,
      "step": 31700
    },
    {
      "entropy": 0.5490647946717218,
      "epoch": 0.5150110034675134,
      "grad_norm": 8.25,
      "learning_rate": 2.4451567105878842e-05,
      "loss": 0.532465124130249,
      "mean_token_accuracy": 0.8642654992640019,
      "num_tokens": 64822404.0,
      "step": 31710
    },
    {
      "entropy": 0.5230568445753306,
      "epoch": 0.5151734162721389,
      "grad_norm": 11.625,
      "learning_rate": 2.443860394381429e-05,
      "loss": 0.4831412315368652,
      "mean_token_accuracy": 0.881611493229866,
      "num_tokens": 64841067.0,
      "step": 31720
    },
    {
      "entropy": 0.5389516274444759,
      "epoch": 0.5153358290767645,
      "grad_norm": 7.71875,
      "learning_rate": 2.442564093276653e-05,
      "loss": 0.5225939750671387,
      "mean_token_accuracy": 0.8703916773200036,
      "num_tokens": 64861714.0,
      "step": 31730
    },
    {
      "entropy": 0.5079233683412895,
      "epoch": 0.5154982418813899,
      "grad_norm": 6.5,
      "learning_rate": 2.441267807622266e-05,
      "loss": 0.49684629440307615,
      "mean_token_accuracy": 0.8753706328570843,
      "num_tokens": 64881656.0,
      "step": 31740
    },
    {
      "entropy": 0.5793736977502704,
      "epoch": 0.5156606546860154,
      "grad_norm": 10.9375,
      "learning_rate": 2.4399715377669698e-05,
      "loss": 0.6116445541381836,
      "mean_token_accuracy": 0.8517182812094688,
      "num_tokens": 64901869.0,
      "step": 31750
    },
    {
      "entropy": 0.5432271152269095,
      "epoch": 0.515823067490641,
      "grad_norm": 8.9375,
      "learning_rate": 2.4386752840594643e-05,
      "loss": 0.5649291038513183,
      "mean_token_accuracy": 0.8622763954102993,
      "num_tokens": 64922658.0,
      "step": 31760
    },
    {
      "entropy": 0.5102298968471587,
      "epoch": 0.5159854802952665,
      "grad_norm": 7.4375,
      "learning_rate": 2.4373790468484466e-05,
      "loss": 0.49146018028259275,
      "mean_token_accuracy": 0.8747186385095119,
      "num_tokens": 64943489.0,
      "step": 31770
    },
    {
      "entropy": 0.5531921558082104,
      "epoch": 0.516147893099892,
      "grad_norm": 8.8125,
      "learning_rate": 2.436082826482605e-05,
      "loss": 0.5821649551391601,
      "mean_token_accuracy": 0.8579642526805401,
      "num_tokens": 64964866.0,
      "step": 31780
    },
    {
      "entropy": 0.44418733306229113,
      "epoch": 0.5163103059045175,
      "grad_norm": 6.8125,
      "learning_rate": 2.434786623310627e-05,
      "loss": 0.43855991363525393,
      "mean_token_accuracy": 0.8791685063391924,
      "num_tokens": 64985152.0,
      "step": 31790
    },
    {
      "entropy": 0.4625800663838163,
      "epoch": 0.516472718709143,
      "grad_norm": 7.875,
      "learning_rate": 2.433490437681194e-05,
      "loss": 0.47405142784118653,
      "mean_token_accuracy": 0.8854402892291546,
      "num_tokens": 65004500.0,
      "step": 31800
    },
    {
      "entropy": 0.5459943887079135,
      "epoch": 0.5166351315137685,
      "grad_norm": 6.6875,
      "learning_rate": 2.432194269942983e-05,
      "loss": 0.5206427097320556,
      "mean_token_accuracy": 0.870201250910759,
      "num_tokens": 65024830.0,
      "step": 31810
    },
    {
      "entropy": 0.5844416637904942,
      "epoch": 0.5167975443183941,
      "grad_norm": 9.75,
      "learning_rate": 2.4308981204446647e-05,
      "loss": 0.5958810329437256,
      "mean_token_accuracy": 0.8488966196775436,
      "num_tokens": 65044785.0,
      "step": 31820
    },
    {
      "entropy": 0.4825925830751657,
      "epoch": 0.5169599571230196,
      "grad_norm": 10.6875,
      "learning_rate": 2.4296019895349083e-05,
      "loss": 0.5136378288269043,
      "mean_token_accuracy": 0.8730720303952694,
      "num_tokens": 65065003.0,
      "step": 31830
    },
    {
      "entropy": 0.5214458112604916,
      "epoch": 0.517122369927645,
      "grad_norm": 6.03125,
      "learning_rate": 2.428305877562375e-05,
      "loss": 0.5323106288909912,
      "mean_token_accuracy": 0.8703538931906223,
      "num_tokens": 65085528.0,
      "step": 31840
    },
    {
      "entropy": 0.49421712281182406,
      "epoch": 0.5172847827322706,
      "grad_norm": 7.5,
      "learning_rate": 2.4270097848757217e-05,
      "loss": 0.4815335273742676,
      "mean_token_accuracy": 0.8732552267611027,
      "num_tokens": 65106591.0,
      "step": 31850
    },
    {
      "entropy": 0.4858395718038082,
      "epoch": 0.5174471955368961,
      "grad_norm": 6.15625,
      "learning_rate": 2.4257137118236e-05,
      "loss": 0.4936209201812744,
      "mean_token_accuracy": 0.8734801806509495,
      "num_tokens": 65127690.0,
      "step": 31860
    },
    {
      "entropy": 0.4252887805458158,
      "epoch": 0.5176096083415217,
      "grad_norm": 6.71875,
      "learning_rate": 2.424417658754656e-05,
      "loss": 0.4176274299621582,
      "mean_token_accuracy": 0.8891563273966312,
      "num_tokens": 65148254.0,
      "step": 31870
    },
    {
      "entropy": 0.5193524038186297,
      "epoch": 0.5177720211461472,
      "grad_norm": 9.0625,
      "learning_rate": 2.423121626017533e-05,
      "loss": 0.5388798713684082,
      "mean_token_accuracy": 0.8683447428047657,
      "num_tokens": 65169269.0,
      "step": 31880
    },
    {
      "entropy": 0.5068851973395795,
      "epoch": 0.5179344339507727,
      "grad_norm": 6.5625,
      "learning_rate": 2.4218256139608646e-05,
      "loss": 0.49756393432617185,
      "mean_token_accuracy": 0.8802759349346161,
      "num_tokens": 65188452.0,
      "step": 31890
    },
    {
      "entropy": 0.4957798327319324,
      "epoch": 0.5180968467553982,
      "grad_norm": 9.1875,
      "learning_rate": 2.420529622933282e-05,
      "loss": 0.4983179569244385,
      "mean_token_accuracy": 0.8740420959889889,
      "num_tokens": 65209357.0,
      "step": 31900
    },
    {
      "entropy": 0.5331252472475171,
      "epoch": 0.5182592595600237,
      "grad_norm": 6.0,
      "learning_rate": 2.41923365328341e-05,
      "loss": 0.5302881240844727,
      "mean_token_accuracy": 0.8621334679424763,
      "num_tokens": 65230310.0,
      "step": 31910
    },
    {
      "entropy": 0.42941749540623275,
      "epoch": 0.5184216723646492,
      "grad_norm": 8.6875,
      "learning_rate": 2.4179377053598654e-05,
      "loss": 0.43677868843078616,
      "mean_token_accuracy": 0.8920993342995643,
      "num_tokens": 65251135.0,
      "step": 31920
    },
    {
      "entropy": 0.500172538345214,
      "epoch": 0.5185840851692748,
      "grad_norm": 5.1875,
      "learning_rate": 2.4166417795112646e-05,
      "loss": 0.4706252098083496,
      "mean_token_accuracy": 0.8766123689711094,
      "num_tokens": 65271783.0,
      "step": 31930
    },
    {
      "entropy": 0.6117228420451284,
      "epoch": 0.5187464979739003,
      "grad_norm": 6.5,
      "learning_rate": 2.4153458760862104e-05,
      "loss": 0.635905122756958,
      "mean_token_accuracy": 0.8449424345046281,
      "num_tokens": 65292520.0,
      "step": 31940
    },
    {
      "entropy": 0.5325430804863572,
      "epoch": 0.5189089107785257,
      "grad_norm": 6.90625,
      "learning_rate": 2.4140499954333063e-05,
      "loss": 0.5645117282867431,
      "mean_token_accuracy": 0.8663497149944306,
      "num_tokens": 65313614.0,
      "step": 31950
    },
    {
      "entropy": 0.4516424720175564,
      "epoch": 0.5190713235831513,
      "grad_norm": 7.65625,
      "learning_rate": 2.412754137901147e-05,
      "loss": 0.470012903213501,
      "mean_token_accuracy": 0.881708151102066,
      "num_tokens": 65334038.0,
      "step": 31960
    },
    {
      "entropy": 0.5022185544949025,
      "epoch": 0.5192337363877768,
      "grad_norm": 10.8125,
      "learning_rate": 2.4114583038383204e-05,
      "loss": 0.5040915489196778,
      "mean_token_accuracy": 0.874006712436676,
      "num_tokens": 65355467.0,
      "step": 31970
    },
    {
      "entropy": 0.5295085556805134,
      "epoch": 0.5193961491924023,
      "grad_norm": 7.40625,
      "learning_rate": 2.4101624935934085e-05,
      "loss": 0.5014502048492432,
      "mean_token_accuracy": 0.8715894818305969,
      "num_tokens": 65376116.0,
      "step": 31980
    },
    {
      "entropy": 0.47011025808751583,
      "epoch": 0.5195585619970279,
      "grad_norm": 11.5625,
      "learning_rate": 2.4088667075149886e-05,
      "loss": 0.5108585834503174,
      "mean_token_accuracy": 0.8754473805427552,
      "num_tokens": 65396866.0,
      "step": 31990
    },
    {
      "epoch": 0.5197209748016534,
      "eval_entropy": 0.5435711845159531,
      "eval_loss": 0.5334779620170593,
      "eval_mean_token_accuracy": 0.8646890954971314,
      "eval_num_tokens": 65417249.0,
      "eval_runtime": 40.3049,
      "eval_samples_per_second": 49.622,
      "eval_steps_per_second": 6.203,
      "step": 32000
    },
    {
      "entropy": 0.4463274016743526,
      "epoch": 0.5198833876062788,
      "grad_norm": 6.71875,
      "learning_rate": 2.4062752092518927e-05,
      "loss": 0.4454516887664795,
      "mean_token_accuracy": 0.8882415909320116,
      "num_tokens": 65437831.0,
      "step": 32010
    },
    {
      "entropy": 0.5641618329100311,
      "epoch": 0.5200458004109044,
      "grad_norm": 8.1875,
      "learning_rate": 2.404979497764336e-05,
      "loss": 0.5450178146362304,
      "mean_token_accuracy": 0.8561033755540848,
      "num_tokens": 65458779.0,
      "step": 32020
    },
    {
      "entropy": 0.5010769475251436,
      "epoch": 0.5202082132155299,
      "grad_norm": 8.25,
      "learning_rate": 2.4036838118375067e-05,
      "loss": 0.4833087921142578,
      "mean_token_accuracy": 0.879993773251772,
      "num_tokens": 65478715.0,
      "step": 32030
    },
    {
      "entropy": 0.5423857097513973,
      "epoch": 0.5203706260201554,
      "grad_norm": 8.75,
      "learning_rate": 2.4023881518199497e-05,
      "loss": 0.5583311557769776,
      "mean_token_accuracy": 0.8702725648880005,
      "num_tokens": 65498467.0,
      "step": 32040
    },
    {
      "entropy": 0.4668980533955619,
      "epoch": 0.520533038824781,
      "grad_norm": 5.15625,
      "learning_rate": 2.401092518060199e-05,
      "loss": 0.4571049690246582,
      "mean_token_accuracy": 0.8776378765702247,
      "num_tokens": 65518112.0,
      "step": 32050
    },
    {
      "entropy": 0.4601549171842635,
      "epoch": 0.5206954516294064,
      "grad_norm": 7.84375,
      "learning_rate": 2.399796910906783e-05,
      "loss": 0.4762396335601807,
      "mean_token_accuracy": 0.8714640635997057,
      "num_tokens": 65538076.0,
      "step": 32060
    },
    {
      "entropy": 0.45349740230012686,
      "epoch": 0.5208578644340319,
      "grad_norm": 4.71875,
      "learning_rate": 2.3985013307082238e-05,
      "loss": 0.4853828907012939,
      "mean_token_accuracy": 0.8840164136141538,
      "num_tokens": 65558116.0,
      "step": 32070
    },
    {
      "entropy": 0.5141229214146733,
      "epoch": 0.5210202772386575,
      "grad_norm": 6.25,
      "learning_rate": 2.3972057778130337e-05,
      "loss": 0.5007837295532227,
      "mean_token_accuracy": 0.8701067987829447,
      "num_tokens": 65578251.0,
      "step": 32080
    },
    {
      "entropy": 0.4534931580070406,
      "epoch": 0.521182690043283,
      "grad_norm": 6.53125,
      "learning_rate": 2.3959102525697207e-05,
      "loss": 0.45786185264587403,
      "mean_token_accuracy": 0.8853714436292648,
      "num_tokens": 65599426.0,
      "step": 32090
    },
    {
      "entropy": 0.4763696023263037,
      "epoch": 0.5213451028479086,
      "grad_norm": 7.84375,
      "learning_rate": 2.3946147553267843e-05,
      "loss": 0.46722073554992677,
      "mean_token_accuracy": 0.8808918125927448,
      "num_tokens": 65620349.0,
      "step": 32100
    },
    {
      "entropy": 0.5246857958380133,
      "epoch": 0.5215075156525341,
      "grad_norm": 7.40625,
      "learning_rate": 2.393319286432715e-05,
      "loss": 0.5786768436431885,
      "mean_token_accuracy": 0.8668648295104504,
      "num_tokens": 65640301.0,
      "step": 32110
    },
    {
      "entropy": 0.5341463503893464,
      "epoch": 0.5216699284571595,
      "grad_norm": 9.875,
      "learning_rate": 2.3920238462359967e-05,
      "loss": 0.5437932014465332,
      "mean_token_accuracy": 0.8627737656235694,
      "num_tokens": 65660582.0,
      "step": 32120
    },
    {
      "entropy": 0.5422573374584317,
      "epoch": 0.5218323412617851,
      "grad_norm": 5.65625,
      "learning_rate": 2.3907284350851083e-05,
      "loss": 0.5173760414123535,
      "mean_token_accuracy": 0.8667926982045173,
      "num_tokens": 65680419.0,
      "step": 32130
    },
    {
      "entropy": 0.4727510468568653,
      "epoch": 0.5219947540664106,
      "grad_norm": 11.9375,
      "learning_rate": 2.389433053328515e-05,
      "loss": 0.4518135547637939,
      "mean_token_accuracy": 0.8855110205709934,
      "num_tokens": 65701564.0,
      "step": 32140
    },
    {
      "entropy": 0.4822386335581541,
      "epoch": 0.5221571668710361,
      "grad_norm": 10.625,
      "learning_rate": 2.3881377013146793e-05,
      "loss": 0.47096867561340333,
      "mean_token_accuracy": 0.8798721224069596,
      "num_tokens": 65721192.0,
      "step": 32150
    },
    {
      "entropy": 0.4734748762333766,
      "epoch": 0.5223195796756617,
      "grad_norm": 7.375,
      "learning_rate": 2.3868423793920532e-05,
      "loss": 0.44486727714538576,
      "mean_token_accuracy": 0.8806769132614136,
      "num_tokens": 65741762.0,
      "step": 32160
    },
    {
      "entropy": 0.4480273856315762,
      "epoch": 0.5224819924802871,
      "grad_norm": 5.65625,
      "learning_rate": 2.3855470879090813e-05,
      "loss": 0.43381896018981936,
      "mean_token_accuracy": 0.8878937222063541,
      "num_tokens": 65762385.0,
      "step": 32170
    },
    {
      "entropy": 0.4639682239620015,
      "epoch": 0.5226444052849126,
      "grad_norm": 10.0,
      "learning_rate": 2.3842518272142006e-05,
      "loss": 0.5031438350677491,
      "mean_token_accuracy": 0.8769525602459908,
      "num_tokens": 65782484.0,
      "step": 32180
    },
    {
      "entropy": 0.5111422085901722,
      "epoch": 0.5228068180895382,
      "grad_norm": 5.625,
      "learning_rate": 2.3829565976558373e-05,
      "loss": 0.5638608455657959,
      "mean_token_accuracy": 0.8621011957526207,
      "num_tokens": 65802628.0,
      "step": 32190
    },
    {
      "entropy": 0.5033729222603143,
      "epoch": 0.5229692308941637,
      "grad_norm": 7.15625,
      "learning_rate": 2.381661399582413e-05,
      "loss": 0.5432794094085693,
      "mean_token_accuracy": 0.8604748986661435,
      "num_tokens": 65822636.0,
      "step": 32200
    },
    {
      "entropy": 0.5175486814230681,
      "epoch": 0.5231316436987892,
      "grad_norm": 9.5,
      "learning_rate": 2.3803662333423386e-05,
      "loss": 0.5050787925720215,
      "mean_token_accuracy": 0.8639198429882526,
      "num_tokens": 65843033.0,
      "step": 32210
    },
    {
      "entropy": 0.5353206885512918,
      "epoch": 0.5232940565034148,
      "grad_norm": 7.96875,
      "learning_rate": 2.3790710992840155e-05,
      "loss": 0.5240823745727539,
      "mean_token_accuracy": 0.8657391272485256,
      "num_tokens": 65864420.0,
      "step": 32220
    },
    {
      "entropy": 0.4753953646635637,
      "epoch": 0.5234564693080402,
      "grad_norm": 6.03125,
      "learning_rate": 2.3777759977558392e-05,
      "loss": 0.47089762687683107,
      "mean_token_accuracy": 0.8850571863353253,
      "num_tokens": 65884671.0,
      "step": 32230
    },
    {
      "entropy": 0.42645889101549983,
      "epoch": 0.5236188821126657,
      "grad_norm": 8.9375,
      "learning_rate": 2.376480929106193e-05,
      "loss": 0.4254753589630127,
      "mean_token_accuracy": 0.8885055363178254,
      "num_tokens": 65905856.0,
      "step": 32240
    },
    {
      "entropy": 0.4537255208473653,
      "epoch": 0.5237812949172913,
      "grad_norm": 7.5625,
      "learning_rate": 2.3751858936834554e-05,
      "loss": 0.4510823726654053,
      "mean_token_accuracy": 0.883241830766201,
      "num_tokens": 65925331.0,
      "step": 32250
    },
    {
      "entropy": 0.5443071022164077,
      "epoch": 0.5239437077219168,
      "grad_norm": 10.5625,
      "learning_rate": 2.373890891835993e-05,
      "loss": 0.5385807991027832,
      "mean_token_accuracy": 0.8695555370301008,
      "num_tokens": 65946632.0,
      "step": 32260
    },
    {
      "entropy": 0.5175013852771372,
      "epoch": 0.5241061205265423,
      "grad_norm": 13.125,
      "learning_rate": 2.372595923912164e-05,
      "loss": 0.5482803344726562,
      "mean_token_accuracy": 0.8657186351716518,
      "num_tokens": 65967565.0,
      "step": 32270
    },
    {
      "entropy": 0.5001365350093693,
      "epoch": 0.5242685333311679,
      "grad_norm": 6.65625,
      "learning_rate": 2.3713009902603167e-05,
      "loss": 0.5293495655059814,
      "mean_token_accuracy": 0.8662685737013817,
      "num_tokens": 65988889.0,
      "step": 32280
    },
    {
      "entropy": 0.5404625029768795,
      "epoch": 0.5244309461357933,
      "grad_norm": 6.875,
      "learning_rate": 2.3700060912287943e-05,
      "loss": 0.5786478519439697,
      "mean_token_accuracy": 0.8626107566058636,
      "num_tokens": 66009279.0,
      "step": 32290
    },
    {
      "entropy": 0.48064499637112024,
      "epoch": 0.5245933589404188,
      "grad_norm": 5.5,
      "learning_rate": 2.3687112271659236e-05,
      "loss": 0.4971757411956787,
      "mean_token_accuracy": 0.8816525593400002,
      "num_tokens": 66029684.0,
      "step": 32300
    },
    {
      "entropy": 0.49382392605766656,
      "epoch": 0.5247557717450444,
      "grad_norm": 15.6875,
      "learning_rate": 2.3674163984200287e-05,
      "loss": 0.5158262729644776,
      "mean_token_accuracy": 0.8745053738355637,
      "num_tokens": 66048892.0,
      "step": 32310
    },
    {
      "entropy": 0.540617058495991,
      "epoch": 0.5249181845496699,
      "grad_norm": 7.5,
      "learning_rate": 2.366121605339421e-05,
      "loss": 0.5566507816314697,
      "mean_token_accuracy": 0.8635015293955803,
      "num_tokens": 66070299.0,
      "step": 32320
    },
    {
      "entropy": 0.48672333364374937,
      "epoch": 0.5250805973542955,
      "grad_norm": 9.5625,
      "learning_rate": 2.3648268482724017e-05,
      "loss": 0.4804685592651367,
      "mean_token_accuracy": 0.8746005900204181,
      "num_tokens": 66090539.0,
      "step": 32330
    },
    {
      "entropy": 0.5427389019168913,
      "epoch": 0.5252430101589209,
      "grad_norm": 7.1875,
      "learning_rate": 2.3635321275672644e-05,
      "loss": 0.545236349105835,
      "mean_token_accuracy": 0.8599127843976021,
      "num_tokens": 66111077.0,
      "step": 32340
    },
    {
      "entropy": 0.5261860045138747,
      "epoch": 0.5254054229635464,
      "grad_norm": 8.625,
      "learning_rate": 2.3622374435722908e-05,
      "loss": 0.5440515041351318,
      "mean_token_accuracy": 0.8716329850256443,
      "num_tokens": 66131500.0,
      "step": 32350
    },
    {
      "entropy": 0.4874029252678156,
      "epoch": 0.525567835768172,
      "grad_norm": 8.125,
      "learning_rate": 2.3609427966357547e-05,
      "loss": 0.5141664028167725,
      "mean_token_accuracy": 0.872572049498558,
      "num_tokens": 66151210.0,
      "step": 32360
    },
    {
      "entropy": 0.5343034153804183,
      "epoch": 0.5257302485727975,
      "grad_norm": 9.1875,
      "learning_rate": 2.3596481871059194e-05,
      "loss": 0.5240890502929687,
      "mean_token_accuracy": 0.870276103168726,
      "num_tokens": 66171569.0,
      "step": 32370
    },
    {
      "entropy": 0.47301261839456854,
      "epoch": 0.525892661377423,
      "grad_norm": 8.625,
      "learning_rate": 2.358353615331037e-05,
      "loss": 0.4896370887756348,
      "mean_token_accuracy": 0.8805594921112061,
      "num_tokens": 66193658.0,
      "step": 32380
    },
    {
      "entropy": 0.5156035942956805,
      "epoch": 0.5260550741820486,
      "grad_norm": 7.75,
      "learning_rate": 2.3570590816593494e-05,
      "loss": 0.49200100898742677,
      "mean_token_accuracy": 0.8764811046421528,
      "num_tokens": 66214280.0,
      "step": 32390
    },
    {
      "entropy": 0.518183997599408,
      "epoch": 0.526217486986674,
      "grad_norm": 9.9375,
      "learning_rate": 2.355764586439091e-05,
      "loss": 0.49608497619628905,
      "mean_token_accuracy": 0.8673455268144608,
      "num_tokens": 66234867.0,
      "step": 32400
    },
    {
      "entropy": 0.5413148131221532,
      "epoch": 0.5263798997912995,
      "grad_norm": 5.6875,
      "learning_rate": 2.354470130018483e-05,
      "loss": 0.5391173362731934,
      "mean_token_accuracy": 0.8639550812542438,
      "num_tokens": 66256453.0,
      "step": 32410
    },
    {
      "entropy": 0.5153211651369929,
      "epoch": 0.5265423125959251,
      "grad_norm": 9.8125,
      "learning_rate": 2.3531757127457374e-05,
      "loss": 0.48618292808532715,
      "mean_token_accuracy": 0.8799338944256305,
      "num_tokens": 66277425.0,
      "step": 32420
    },
    {
      "entropy": 0.46757388330297545,
      "epoch": 0.5267047254005506,
      "grad_norm": 10.8125,
      "learning_rate": 2.3518813349690547e-05,
      "loss": 0.47550210952758787,
      "mean_token_accuracy": 0.8820863835513592,
      "num_tokens": 66297358.0,
      "step": 32430
    },
    {
      "entropy": 0.4802468687528744,
      "epoch": 0.526867138205176,
      "grad_norm": 8.25,
      "learning_rate": 2.3505869970366247e-05,
      "loss": 0.4909496307373047,
      "mean_token_accuracy": 0.8810056120157241,
      "num_tokens": 66317323.0,
      "step": 32440
    },
    {
      "entropy": 0.5636429914273322,
      "epoch": 0.5270295510098016,
      "grad_norm": 8.875,
      "learning_rate": 2.3492926992966303e-05,
      "loss": 0.5893250465393066,
      "mean_token_accuracy": 0.8638626627624035,
      "num_tokens": 66337681.0,
      "step": 32450
    },
    {
      "entropy": 0.4628427307587117,
      "epoch": 0.5271919638144271,
      "grad_norm": 6.40625,
      "learning_rate": 2.347998442097237e-05,
      "loss": 0.4464442729949951,
      "mean_token_accuracy": 0.8814917452633381,
      "num_tokens": 66358081.0,
      "step": 32460
    },
    {
      "entropy": 0.47034822651185093,
      "epoch": 0.5273543766190526,
      "grad_norm": 7.1875,
      "learning_rate": 2.3467042257866044e-05,
      "loss": 0.46502079963684084,
      "mean_token_accuracy": 0.8840247943997384,
      "num_tokens": 66379646.0,
      "step": 32470
    },
    {
      "entropy": 0.5026594616589136,
      "epoch": 0.5275167894236782,
      "grad_norm": 7.3125,
      "learning_rate": 2.3454100507128806e-05,
      "loss": 0.5435523509979248,
      "mean_token_accuracy": 0.8607279852032661,
      "num_tokens": 66400110.0,
      "step": 32480
    },
    {
      "entropy": 0.6047575959935785,
      "epoch": 0.5276792022283037,
      "grad_norm": 8.5625,
      "learning_rate": 2.3441159172242e-05,
      "loss": 0.6161753177642822,
      "mean_token_accuracy": 0.8454909287393093,
      "num_tokens": 66420762.0,
      "step": 32490
    },
    {
      "entropy": 0.46706012487411497,
      "epoch": 0.5278416150329291,
      "grad_norm": 10.0,
      "learning_rate": 2.3428218256686873e-05,
      "loss": 0.47799954414367674,
      "mean_token_accuracy": 0.8782646067440509,
      "num_tokens": 66441315.0,
      "step": 32500
    },
    {
      "entropy": 0.5199498703121208,
      "epoch": 0.5280040278375547,
      "grad_norm": 11.1875,
      "learning_rate": 2.3415277763944578e-05,
      "loss": 0.5070500373840332,
      "mean_token_accuracy": 0.867745628207922,
      "num_tokens": 66460962.0,
      "step": 32510
    },
    {
      "entropy": 0.4659481802256778,
      "epoch": 0.5281664406421802,
      "grad_norm": 6.5,
      "learning_rate": 2.340233769749612e-05,
      "loss": 0.4874136447906494,
      "mean_token_accuracy": 0.879593276232481,
      "num_tokens": 66480823.0,
      "step": 32520
    },
    {
      "entropy": 0.4709164334461093,
      "epoch": 0.5283288534468057,
      "grad_norm": 9.375,
      "learning_rate": 2.3389398060822422e-05,
      "loss": 0.4702265739440918,
      "mean_token_accuracy": 0.8867265716195106,
      "num_tokens": 66501840.0,
      "step": 32530
    },
    {
      "entropy": 0.5098220149986445,
      "epoch": 0.5284912662514313,
      "grad_norm": 8.1875,
      "learning_rate": 2.3376458857404253e-05,
      "loss": 0.5513124465942383,
      "mean_token_accuracy": 0.8693359225988389,
      "num_tokens": 66524094.0,
      "step": 32540
    },
    {
      "entropy": 0.48883825172670187,
      "epoch": 0.5286536790560568,
      "grad_norm": 7.375,
      "learning_rate": 2.33635200907223e-05,
      "loss": 0.5303766250610351,
      "mean_token_accuracy": 0.8605763830244542,
      "num_tokens": 66543805.0,
      "step": 32550
    },
    {
      "entropy": 0.5398591757752002,
      "epoch": 0.5288160918606823,
      "grad_norm": 10.6875,
      "learning_rate": 2.3350581764257124e-05,
      "loss": 0.5733153820037842,
      "mean_token_accuracy": 0.8481892824172974,
      "num_tokens": 66564588.0,
      "step": 32560
    },
    {
      "entropy": 0.5216809779405593,
      "epoch": 0.5289785046653078,
      "grad_norm": 6.5625,
      "learning_rate": 2.3337643881489163e-05,
      "loss": 0.4994957447052002,
      "mean_token_accuracy": 0.875600402802229,
      "num_tokens": 66585677.0,
      "step": 32570
    },
    {
      "entropy": 0.4802111144643277,
      "epoch": 0.5291409174699333,
      "grad_norm": 5.6875,
      "learning_rate": 2.332470644589873e-05,
      "loss": 0.4922330379486084,
      "mean_token_accuracy": 0.8726156920194625,
      "num_tokens": 66606189.0,
      "step": 32580
    },
    {
      "entropy": 0.49278554660268126,
      "epoch": 0.5293033302745589,
      "grad_norm": 9.25,
      "learning_rate": 2.3311769460966034e-05,
      "loss": 0.4768968105316162,
      "mean_token_accuracy": 0.8778623759746551,
      "num_tokens": 66626925.0,
      "step": 32590
    },
    {
      "entropy": 0.5064407004509122,
      "epoch": 0.5294657430791844,
      "grad_norm": 8.125,
      "learning_rate": 2.3298832930171134e-05,
      "loss": 0.5814823150634766,
      "mean_token_accuracy": 0.8563963498920202,
      "num_tokens": 66646830.0,
      "step": 32600
    },
    {
      "entropy": 0.5224996482720599,
      "epoch": 0.5296281558838098,
      "grad_norm": 6.75,
      "learning_rate": 2.3285896856994006e-05,
      "loss": 0.5544665336608887,
      "mean_token_accuracy": 0.8646035399287939,
      "num_tokens": 66667856.0,
      "step": 32610
    },
    {
      "entropy": 0.528996616275981,
      "epoch": 0.5297905686884354,
      "grad_norm": 11.1875,
      "learning_rate": 2.3272961244914483e-05,
      "loss": 0.4975640296936035,
      "mean_token_accuracy": 0.8739037767052651,
      "num_tokens": 66687976.0,
      "step": 32620
    },
    {
      "entropy": 0.533230008976534,
      "epoch": 0.5299529814930609,
      "grad_norm": 5.78125,
      "learning_rate": 2.326002609741226e-05,
      "loss": 0.52183518409729,
      "mean_token_accuracy": 0.8701559405773878,
      "num_tokens": 66708620.0,
      "step": 32630
    },
    {
      "entropy": 0.5010134418960661,
      "epoch": 0.5301153942976864,
      "grad_norm": 7.0625,
      "learning_rate": 2.3247091417966932e-05,
      "loss": 0.4694943904876709,
      "mean_token_accuracy": 0.8811018671840429,
      "num_tokens": 66728854.0,
      "step": 32640
    },
    {
      "entropy": 0.4928140091709793,
      "epoch": 0.530277807102312,
      "grad_norm": 6.5,
      "learning_rate": 2.3234157210057948e-05,
      "loss": 0.46771931648254395,
      "mean_token_accuracy": 0.8821816585958004,
      "num_tokens": 66749262.0,
      "step": 32650
    },
    {
      "entropy": 0.5664487617090345,
      "epoch": 0.5304402199069375,
      "grad_norm": 7.59375,
      "learning_rate": 2.3221223477164637e-05,
      "loss": 0.6064012050628662,
      "mean_token_accuracy": 0.8469003453850746,
      "num_tokens": 66770290.0,
      "step": 32660
    },
    {
      "entropy": 0.4953888944815844,
      "epoch": 0.5306026327115629,
      "grad_norm": 6.03125,
      "learning_rate": 2.3208290222766215e-05,
      "loss": 0.4613226890563965,
      "mean_token_accuracy": 0.8770062454044819,
      "num_tokens": 66792253.0,
      "step": 32670
    },
    {
      "entropy": 0.4760645147413015,
      "epoch": 0.5307650455161885,
      "grad_norm": 6.375,
      "learning_rate": 2.3195357450341747e-05,
      "loss": 0.4352574348449707,
      "mean_token_accuracy": 0.8907764203846454,
      "num_tokens": 66813127.0,
      "step": 32680
    },
    {
      "entropy": 0.49212415223009887,
      "epoch": 0.530927458320814,
      "grad_norm": 8.6875,
      "learning_rate": 2.3182425163370184e-05,
      "loss": 0.5220905303955078,
      "mean_token_accuracy": 0.8718458153307438,
      "num_tokens": 66833914.0,
      "step": 32690
    },
    {
      "entropy": 0.476950765773654,
      "epoch": 0.5310898711254395,
      "grad_norm": 5.28125,
      "learning_rate": 2.3169493365330338e-05,
      "loss": 0.4816749572753906,
      "mean_token_accuracy": 0.8748804412782192,
      "num_tokens": 66854306.0,
      "step": 32700
    },
    {
      "entropy": 0.4427878403570503,
      "epoch": 0.5312522839300651,
      "grad_norm": 6.53125,
      "learning_rate": 2.315656205970088e-05,
      "loss": 0.4731252670288086,
      "mean_token_accuracy": 0.8855576887726784,
      "num_tokens": 66873903.0,
      "step": 32710
    },
    {
      "entropy": 0.498014506441541,
      "epoch": 0.5314146967346906,
      "grad_norm": 8.125,
      "learning_rate": 2.3143631249960386e-05,
      "loss": 0.4536584854125977,
      "mean_token_accuracy": 0.8867255732417106,
      "num_tokens": 66894100.0,
      "step": 32720
    },
    {
      "entropy": 0.477185506792739,
      "epoch": 0.531577109539316,
      "grad_norm": 5.75,
      "learning_rate": 2.3130700939587247e-05,
      "loss": 0.5139286041259765,
      "mean_token_accuracy": 0.8775724165141583,
      "num_tokens": 66914360.0,
      "step": 32730
    },
    {
      "entropy": 0.5347449270077049,
      "epoch": 0.5317395223439416,
      "grad_norm": 10.75,
      "learning_rate": 2.311777113205976e-05,
      "loss": 0.5629463195800781,
      "mean_token_accuracy": 0.8536611661314965,
      "num_tokens": 66934832.0,
      "step": 32740
    },
    {
      "entropy": 0.552800724375993,
      "epoch": 0.5319019351485671,
      "grad_norm": 8.75,
      "learning_rate": 2.3104841830856073e-05,
      "loss": 0.546446418762207,
      "mean_token_accuracy": 0.8640116959810257,
      "num_tokens": 66954493.0,
      "step": 32750
    },
    {
      "entropy": 0.48592754974961283,
      "epoch": 0.5320643479531927,
      "grad_norm": 9.125,
      "learning_rate": 2.309191303945418e-05,
      "loss": 0.4768341541290283,
      "mean_token_accuracy": 0.8734835833311081,
      "num_tokens": 66974922.0,
      "step": 32760
    },
    {
      "entropy": 0.5000361226731911,
      "epoch": 0.5322267607578182,
      "grad_norm": 5.71875,
      "learning_rate": 2.3078984761331973e-05,
      "loss": 0.4872912406921387,
      "mean_token_accuracy": 0.8747485183179379,
      "num_tokens": 66996536.0,
      "step": 32770
    },
    {
      "entropy": 0.4710672049317509,
      "epoch": 0.5323891735624436,
      "grad_norm": 7.875,
      "learning_rate": 2.306605699996719e-05,
      "loss": 0.48632097244262695,
      "mean_token_accuracy": 0.8779416307806969,
      "num_tokens": 67016313.0,
      "step": 32780
    },
    {
      "entropy": 0.5291117968503386,
      "epoch": 0.5325515863670692,
      "grad_norm": 8.3125,
      "learning_rate": 2.305312975883741e-05,
      "loss": 0.5384933948516846,
      "mean_token_accuracy": 0.8686050578951836,
      "num_tokens": 67037028.0,
      "step": 32790
    },
    {
      "entropy": 0.4670780538814142,
      "epoch": 0.5327139991716947,
      "grad_norm": 5.90625,
      "learning_rate": 2.30402030414201e-05,
      "loss": 0.48860921859741213,
      "mean_token_accuracy": 0.8706880435347557,
      "num_tokens": 67057534.0,
      "step": 32800
    },
    {
      "entropy": 0.4696903506293893,
      "epoch": 0.5328764119763202,
      "grad_norm": 7.65625,
      "learning_rate": 2.302727685119259e-05,
      "loss": 0.5334705352783203,
      "mean_token_accuracy": 0.8719631433486938,
      "num_tokens": 67077047.0,
      "step": 32810
    },
    {
      "entropy": 0.44733525696210563,
      "epoch": 0.5330388247809458,
      "grad_norm": 7.90625,
      "learning_rate": 2.3014351191632026e-05,
      "loss": 0.4464900016784668,
      "mean_token_accuracy": 0.886394377052784,
      "num_tokens": 67096524.0,
      "step": 32820
    },
    {
      "entropy": 0.5560731885954737,
      "epoch": 0.5332012375855713,
      "grad_norm": 7.875,
      "learning_rate": 2.3001426066215463e-05,
      "loss": 0.5633484363555908,
      "mean_token_accuracy": 0.8631261631846427,
      "num_tokens": 67118200.0,
      "step": 32830
    },
    {
      "entropy": 0.5145490691997111,
      "epoch": 0.5333636503901967,
      "grad_norm": 7.71875,
      "learning_rate": 2.2988501478419775e-05,
      "loss": 0.4940930366516113,
      "mean_token_accuracy": 0.8738794352859258,
      "num_tokens": 67138452.0,
      "step": 32840
    },
    {
      "entropy": 0.5695694834925235,
      "epoch": 0.5335260631948223,
      "grad_norm": 7.21875,
      "learning_rate": 2.297557743172171e-05,
      "loss": 0.5532413959503174,
      "mean_token_accuracy": 0.8605085492134095,
      "num_tokens": 67160100.0,
      "step": 32850
    },
    {
      "entropy": 0.5626272991299629,
      "epoch": 0.5336884759994478,
      "grad_norm": 6.1875,
      "learning_rate": 2.296265392959787e-05,
      "loss": 0.5708122253417969,
      "mean_token_accuracy": 0.861725141108036,
      "num_tokens": 67180290.0,
      "step": 32860
    },
    {
      "entropy": 0.538824743963778,
      "epoch": 0.5338508888040733,
      "grad_norm": 6.8125,
      "learning_rate": 2.294973097552469e-05,
      "loss": 0.5272441864013672,
      "mean_token_accuracy": 0.86572335511446,
      "num_tokens": 67201461.0,
      "step": 32870
    },
    {
      "entropy": 0.514518678188324,
      "epoch": 0.5340133016086989,
      "grad_norm": 6.3125,
      "learning_rate": 2.2936808572978492e-05,
      "loss": 0.4908015251159668,
      "mean_token_accuracy": 0.8665127895772458,
      "num_tokens": 67222450.0,
      "step": 32880
    },
    {
      "entropy": 0.5197007052600384,
      "epoch": 0.5341757144133243,
      "grad_norm": 7.15625,
      "learning_rate": 2.2923886725435433e-05,
      "loss": 0.5046964645385742,
      "mean_token_accuracy": 0.8710944265127182,
      "num_tokens": 67241686.0,
      "step": 32890
    },
    {
      "entropy": 0.5433647131547332,
      "epoch": 0.5343381272179498,
      "grad_norm": 9.875,
      "learning_rate": 2.2910965436371502e-05,
      "loss": 0.5563732147216797,
      "mean_token_accuracy": 0.861873684078455,
      "num_tokens": 67262403.0,
      "step": 32900
    },
    {
      "entropy": 0.5445385276339948,
      "epoch": 0.5345005400225754,
      "grad_norm": 11.5625,
      "learning_rate": 2.289804470926257e-05,
      "loss": 0.59857177734375,
      "mean_token_accuracy": 0.8586912170052529,
      "num_tokens": 67283407.0,
      "step": 32910
    },
    {
      "entropy": 0.5489644240122289,
      "epoch": 0.5346629528272009,
      "grad_norm": 9.25,
      "learning_rate": 2.288512454758432e-05,
      "loss": 0.5666347980499268,
      "mean_token_accuracy": 0.8678211003541947,
      "num_tokens": 67302719.0,
      "step": 32920
    },
    {
      "entropy": 0.5456996961496771,
      "epoch": 0.5348253656318264,
      "grad_norm": 8.875,
      "learning_rate": 2.2872204954812332e-05,
      "loss": 0.5822239398956299,
      "mean_token_accuracy": 0.8573476634919643,
      "num_tokens": 67323279.0,
      "step": 32930
    },
    {
      "entropy": 0.5652053585799877,
      "epoch": 0.534987778436452,
      "grad_norm": 7.03125,
      "learning_rate": 2.2859285934421996e-05,
      "loss": 0.5865334033966064,
      "mean_token_accuracy": 0.8632900461554527,
      "num_tokens": 67344229.0,
      "step": 32940
    },
    {
      "entropy": 0.46165919820778073,
      "epoch": 0.5351501912410774,
      "grad_norm": 9.125,
      "learning_rate": 2.2846367489888544e-05,
      "loss": 0.44862751960754393,
      "mean_token_accuracy": 0.8846436686813831,
      "num_tokens": 67365004.0,
      "step": 32950
    },
    {
      "entropy": 0.5541324409656226,
      "epoch": 0.5353126040457029,
      "grad_norm": 12.125,
      "learning_rate": 2.2833449624687078e-05,
      "loss": 0.5242507934570313,
      "mean_token_accuracy": 0.8652288518846035,
      "num_tokens": 67386889.0,
      "step": 32960
    },
    {
      "entropy": 0.5944210954941809,
      "epoch": 0.5354750168503285,
      "grad_norm": 9.4375,
      "learning_rate": 2.2820532342292547e-05,
      "loss": 0.6086171627044678,
      "mean_token_accuracy": 0.8493060059845448,
      "num_tokens": 67407297.0,
      "step": 32970
    },
    {
      "entropy": 0.5019004900241271,
      "epoch": 0.535637429654954,
      "grad_norm": 9.5625,
      "learning_rate": 2.2807615646179697e-05,
      "loss": 0.46617608070373534,
      "mean_token_accuracy": 0.8732406802475452,
      "num_tokens": 67427997.0,
      "step": 32980
    },
    {
      "entropy": 0.5707862267270685,
      "epoch": 0.5357998424595796,
      "grad_norm": 6.65625,
      "learning_rate": 2.279469953982317e-05,
      "loss": 0.6045993804931641,
      "mean_token_accuracy": 0.8481650006026029,
      "num_tokens": 67450231.0,
      "step": 32990
    },
    {
      "entropy": 0.4815137625904754,
      "epoch": 0.535962255264205,
      "grad_norm": 8.8125,
      "learning_rate": 2.2781784026697427e-05,
      "loss": 0.49737749099731443,
      "mean_token_accuracy": 0.8748418495059014,
      "num_tokens": 67472258.0,
      "step": 33000
    },
    {
      "entropy": 0.4480097529478371,
      "epoch": 0.5361246680688305,
      "grad_norm": 5.28125,
      "learning_rate": 2.2768869110276764e-05,
      "loss": 0.3788824319839478,
      "mean_token_accuracy": 0.8982210613787174,
      "num_tokens": 67492073.0,
      "step": 33010
    },
    {
      "entropy": 0.45280841409694406,
      "epoch": 0.5362870808734561,
      "grad_norm": 7.28125,
      "learning_rate": 2.2755954794035327e-05,
      "loss": 0.46574993133544923,
      "mean_token_accuracy": 0.8809876352548599,
      "num_tokens": 67512778.0,
      "step": 33020
    },
    {
      "entropy": 0.517536657396704,
      "epoch": 0.5364494936780816,
      "grad_norm": 6.90625,
      "learning_rate": 2.2743041081447088e-05,
      "loss": 0.5407838821411133,
      "mean_token_accuracy": 0.8746400058269501,
      "num_tokens": 67533897.0,
      "step": 33030
    },
    {
      "entropy": 0.580740409810096,
      "epoch": 0.5366119064827071,
      "grad_norm": 9.4375,
      "learning_rate": 2.2730127975985868e-05,
      "loss": 0.5848737716674804,
      "mean_token_accuracy": 0.8543719828128815,
      "num_tokens": 67554717.0,
      "step": 33040
    },
    {
      "entropy": 0.5211548042949289,
      "epoch": 0.5367743192873327,
      "grad_norm": 4.34375,
      "learning_rate": 2.2717215481125333e-05,
      "loss": 0.5052491188049316,
      "mean_token_accuracy": 0.8700866349041462,
      "num_tokens": 67574665.0,
      "step": 33050
    },
    {
      "entropy": 0.49489269563928245,
      "epoch": 0.5369367320919581,
      "grad_norm": 14.0625,
      "learning_rate": 2.2704303600338962e-05,
      "loss": 0.49632997512817384,
      "mean_token_accuracy": 0.8713954783976078,
      "num_tokens": 67596493.0,
      "step": 33060
    },
    {
      "entropy": 0.4595108923036605,
      "epoch": 0.5370991448965836,
      "grad_norm": 9.125,
      "learning_rate": 2.269139233710007e-05,
      "loss": 0.4537147045135498,
      "mean_token_accuracy": 0.8804702654480934,
      "num_tokens": 67616283.0,
      "step": 33070
    },
    {
      "entropy": 0.5467124318587594,
      "epoch": 0.5372615577012092,
      "grad_norm": 7.21875,
      "learning_rate": 2.2678481694881843e-05,
      "loss": 0.5416676998138428,
      "mean_token_accuracy": 0.8634032782167196,
      "num_tokens": 67637740.0,
      "step": 33080
    },
    {
      "entropy": 0.4828449568711221,
      "epoch": 0.5374239705058347,
      "grad_norm": 8.8125,
      "learning_rate": 2.266557167715725e-05,
      "loss": 0.4983325958251953,
      "mean_token_accuracy": 0.8741713687777519,
      "num_tokens": 67659010.0,
      "step": 33090
    },
    {
      "entropy": 0.5056463782209903,
      "epoch": 0.5375863833104602,
      "grad_norm": 7.46875,
      "learning_rate": 2.265266228739912e-05,
      "loss": 0.5212023258209229,
      "mean_token_accuracy": 0.8703475572168827,
      "num_tokens": 67679648.0,
      "step": 33100
    },
    {
      "entropy": 0.5141458952799439,
      "epoch": 0.5377487961150857,
      "grad_norm": 8.1875,
      "learning_rate": 2.2639753529080117e-05,
      "loss": 0.5216042518615722,
      "mean_token_accuracy": 0.8696788311004638,
      "num_tokens": 67700601.0,
      "step": 33110
    },
    {
      "entropy": 0.5577704962808638,
      "epoch": 0.5379112089197112,
      "grad_norm": 7.6875,
      "learning_rate": 2.262684540567271e-05,
      "loss": 0.5485371589660645,
      "mean_token_accuracy": 0.8682889953255654,
      "num_tokens": 67722446.0,
      "step": 33120
    },
    {
      "entropy": 0.4419416519580409,
      "epoch": 0.5380736217243367,
      "grad_norm": 6.90625,
      "learning_rate": 2.2613937920649236e-05,
      "loss": 0.5068124771118164,
      "mean_token_accuracy": 0.873594519495964,
      "num_tokens": 67741980.0,
      "step": 33130
    },
    {
      "entropy": 0.47142268856987357,
      "epoch": 0.5382360345289623,
      "grad_norm": 8.5625,
      "learning_rate": 2.260103107748181e-05,
      "loss": 0.4757381916046143,
      "mean_token_accuracy": 0.8767985232174397,
      "num_tokens": 67762889.0,
      "step": 33140
    },
    {
      "entropy": 0.5141531300963834,
      "epoch": 0.5383984473335878,
      "grad_norm": 6.53125,
      "learning_rate": 2.258812487964242e-05,
      "loss": 0.5164625644683838,
      "mean_token_accuracy": 0.8741278141736984,
      "num_tokens": 67782869.0,
      "step": 33150
    },
    {
      "entropy": 0.4289492588490248,
      "epoch": 0.5385608601382132,
      "grad_norm": 11.25,
      "learning_rate": 2.257521933060286e-05,
      "loss": 0.4134061336517334,
      "mean_token_accuracy": 0.8899612694978714,
      "num_tokens": 67803858.0,
      "step": 33160
    },
    {
      "entropy": 0.47381586651317775,
      "epoch": 0.5387232729428388,
      "grad_norm": 13.5625,
      "learning_rate": 2.256231443383475e-05,
      "loss": 0.4945474147796631,
      "mean_token_accuracy": 0.8788340352475643,
      "num_tokens": 67824984.0,
      "step": 33170
    },
    {
      "entropy": 0.4264356805477291,
      "epoch": 0.5388856857474643,
      "grad_norm": 12.3125,
      "learning_rate": 2.2549410192809526e-05,
      "loss": 0.44061965942382814,
      "mean_token_accuracy": 0.8830974742770195,
      "num_tokens": 67843941.0,
      "step": 33180
    },
    {
      "entropy": 0.4913007877767086,
      "epoch": 0.5390480985520898,
      "grad_norm": 5.78125,
      "learning_rate": 2.253650661099848e-05,
      "loss": 0.4912854194641113,
      "mean_token_accuracy": 0.8743470579385757,
      "num_tokens": 67866012.0,
      "step": 33190
    },
    {
      "entropy": 0.5307337297592312,
      "epoch": 0.5392105113567154,
      "grad_norm": 10.125,
      "learning_rate": 2.2523603691872683e-05,
      "loss": 0.5261868476867676,
      "mean_token_accuracy": 0.8672286428511142,
      "num_tokens": 67885633.0,
      "step": 33200
    },
    {
      "entropy": 0.4887002401985228,
      "epoch": 0.5393729241613409,
      "grad_norm": 9.625,
      "learning_rate": 2.2510701438903068e-05,
      "loss": 0.503861379623413,
      "mean_token_accuracy": 0.8765039809048176,
      "num_tokens": 67905336.0,
      "step": 33210
    },
    {
      "entropy": 0.5051737172994762,
      "epoch": 0.5395353369659664,
      "grad_norm": 8.3125,
      "learning_rate": 2.2497799855560354e-05,
      "loss": 0.5371496677398682,
      "mean_token_accuracy": 0.8708290956914425,
      "num_tokens": 67926501.0,
      "step": 33220
    },
    {
      "entropy": 0.45476394777651874,
      "epoch": 0.5396977497705919,
      "grad_norm": 6.1875,
      "learning_rate": 2.2484898945315095e-05,
      "loss": 0.44231410026550294,
      "mean_token_accuracy": 0.8890948362648488,
      "num_tokens": 67947136.0,
      "step": 33230
    },
    {
      "entropy": 0.44398899832740424,
      "epoch": 0.5398601625752174,
      "grad_norm": 9.1875,
      "learning_rate": 2.2471998711637684e-05,
      "loss": 0.4634708881378174,
      "mean_token_accuracy": 0.8752835385501385,
      "num_tokens": 67966198.0,
      "step": 33240
    },
    {
      "entropy": 0.4963882165029645,
      "epoch": 0.540022575379843,
      "grad_norm": 8.3125,
      "learning_rate": 2.2459099157998296e-05,
      "loss": 0.5084300518035889,
      "mean_token_accuracy": 0.8769156686961651,
      "num_tokens": 67986407.0,
      "step": 33250
    },
    {
      "entropy": 0.47158871341962366,
      "epoch": 0.5401849881844685,
      "grad_norm": 8.3125,
      "learning_rate": 2.2446200287866944e-05,
      "loss": 0.46752400398254396,
      "mean_token_accuracy": 0.8747147388756276,
      "num_tokens": 68005869.0,
      "step": 33260
    },
    {
      "entropy": 0.49041389748454095,
      "epoch": 0.540347400989094,
      "grad_norm": 10.8125,
      "learning_rate": 2.2433302104713462e-05,
      "loss": 0.5097508907318116,
      "mean_token_accuracy": 0.873805795609951,
      "num_tokens": 68026202.0,
      "step": 33270
    },
    {
      "entropy": 0.474134402256459,
      "epoch": 0.5405098137937195,
      "grad_norm": 9.625,
      "learning_rate": 2.2420404612007468e-05,
      "loss": 0.4815934181213379,
      "mean_token_accuracy": 0.8758148342370987,
      "num_tokens": 68045893.0,
      "step": 33280
    },
    {
      "entropy": 0.49069400699809196,
      "epoch": 0.540672226598345,
      "grad_norm": 8.3125,
      "learning_rate": 2.2407507813218434e-05,
      "loss": 0.48001389503479003,
      "mean_token_accuracy": 0.8661003768444061,
      "num_tokens": 68066068.0,
      "step": 33290
    },
    {
      "entropy": 0.5716423511505127,
      "epoch": 0.5408346394029705,
      "grad_norm": 7.25,
      "learning_rate": 2.2394611711815634e-05,
      "loss": 0.5595417022705078,
      "mean_token_accuracy": 0.8686589658260345,
      "num_tokens": 68086944.0,
      "step": 33300
    },
    {
      "entropy": 0.5035804902901873,
      "epoch": 0.5409970522075961,
      "grad_norm": 7.0625,
      "learning_rate": 2.238171631126813e-05,
      "loss": 0.4869039535522461,
      "mean_token_accuracy": 0.8801019251346588,
      "num_tokens": 68107169.0,
      "step": 33310
    },
    {
      "entropy": 0.5052243522368371,
      "epoch": 0.5411594650122216,
      "grad_norm": 6.125,
      "learning_rate": 2.2368821615044832e-05,
      "loss": 0.5146710395812988,
      "mean_token_accuracy": 0.8662437550723553,
      "num_tokens": 68127610.0,
      "step": 33320
    },
    {
      "entropy": 0.41654260715004054,
      "epoch": 0.541321877816847,
      "grad_norm": 5.78125,
      "learning_rate": 2.2355927626614426e-05,
      "loss": 0.43950557708740234,
      "mean_token_accuracy": 0.8930185467004776,
      "num_tokens": 68147415.0,
      "step": 33330
    },
    {
      "entropy": 0.5206709966994822,
      "epoch": 0.5414842906214726,
      "grad_norm": 7.125,
      "learning_rate": 2.2343034349445423e-05,
      "loss": 0.5341089248657227,
      "mean_token_accuracy": 0.8702145893126726,
      "num_tokens": 68168517.0,
      "step": 33340
    },
    {
      "entropy": 0.5909528512042016,
      "epoch": 0.5416467034260981,
      "grad_norm": 9.1875,
      "learning_rate": 2.2330141787006165e-05,
      "loss": 0.6003854274749756,
      "mean_token_accuracy": 0.8529349848628044,
      "num_tokens": 68191697.0,
      "step": 33350
    },
    {
      "entropy": 0.5397138382773846,
      "epoch": 0.5418091162307236,
      "grad_norm": 7.25,
      "learning_rate": 2.2317249942764758e-05,
      "loss": 0.5493967533111572,
      "mean_token_accuracy": 0.8606333419680595,
      "num_tokens": 68212250.0,
      "step": 33360
    },
    {
      "entropy": 0.47205501173157244,
      "epoch": 0.5419715290353492,
      "grad_norm": 8.75,
      "learning_rate": 2.2304358820189146e-05,
      "loss": 0.4590886116027832,
      "mean_token_accuracy": 0.8820004306733609,
      "num_tokens": 68234249.0,
      "step": 33370
    },
    {
      "entropy": 0.48822519700042905,
      "epoch": 0.5421339418399747,
      "grad_norm": 9.9375,
      "learning_rate": 2.2291468422747077e-05,
      "loss": 0.4978174686431885,
      "mean_token_accuracy": 0.8755841057747602,
      "num_tokens": 68255569.0,
      "step": 33380
    },
    {
      "entropy": 0.46375337410718204,
      "epoch": 0.5422963546446001,
      "grad_norm": 10.5625,
      "learning_rate": 2.227857875390608e-05,
      "loss": 0.44492011070251464,
      "mean_token_accuracy": 0.8772645220160484,
      "num_tokens": 68276870.0,
      "step": 33390
    },
    {
      "entropy": 0.5155336276628077,
      "epoch": 0.5424587674492257,
      "grad_norm": 5.28125,
      "learning_rate": 2.2265689817133517e-05,
      "loss": 0.5314523220062256,
      "mean_token_accuracy": 0.8737161293625831,
      "num_tokens": 68297073.0,
      "step": 33400
    },
    {
      "entropy": 0.5359852105844766,
      "epoch": 0.5426211802538512,
      "grad_norm": 10.0625,
      "learning_rate": 2.2252801615896542e-05,
      "loss": 0.5427725315093994,
      "mean_token_accuracy": 0.8727028504014015,
      "num_tokens": 68317043.0,
      "step": 33410
    },
    {
      "entropy": 0.5307764233555645,
      "epoch": 0.5427835930584767,
      "grad_norm": 6.6875,
      "learning_rate": 2.2239914153662102e-05,
      "loss": 0.5087746620178223,
      "mean_token_accuracy": 0.8781487956643105,
      "num_tokens": 68337049.0,
      "step": 33420
    },
    {
      "entropy": 0.4663276055827737,
      "epoch": 0.5429460058631023,
      "grad_norm": 11.625,
      "learning_rate": 2.2227027433896967e-05,
      "loss": 0.5130875110626221,
      "mean_token_accuracy": 0.8678036965429783,
      "num_tokens": 68357111.0,
      "step": 33430
    },
    {
      "entropy": 0.4962276759557426,
      "epoch": 0.5431084186677277,
      "grad_norm": 6.5625,
      "learning_rate": 2.2214141460067667e-05,
      "loss": 0.5082298278808594,
      "mean_token_accuracy": 0.8749689199030399,
      "num_tokens": 68377983.0,
      "step": 33440
    },
    {
      "entropy": 0.44234788888134063,
      "epoch": 0.5432708314723533,
      "grad_norm": 9.25,
      "learning_rate": 2.220125623564058e-05,
      "loss": 0.43184924125671387,
      "mean_token_accuracy": 0.8864670112729073,
      "num_tokens": 68398507.0,
      "step": 33450
    },
    {
      "entropy": 0.4878842759877443,
      "epoch": 0.5434332442769788,
      "grad_norm": 8.9375,
      "learning_rate": 2.218837176408187e-05,
      "loss": 0.4968769073486328,
      "mean_token_accuracy": 0.8653305366635322,
      "num_tokens": 68418531.0,
      "step": 33460
    },
    {
      "entropy": 0.5374108772724867,
      "epoch": 0.5435956570816043,
      "grad_norm": 7.84375,
      "learning_rate": 2.2175488048857462e-05,
      "loss": 0.5138206958770752,
      "mean_token_accuracy": 0.8728295043110847,
      "num_tokens": 68439131.0,
      "step": 33470
    },
    {
      "entropy": 0.4325492143630981,
      "epoch": 0.5437580698862299,
      "grad_norm": 7.625,
      "learning_rate": 2.216260509343311e-05,
      "loss": 0.4399118423461914,
      "mean_token_accuracy": 0.8919104568660259,
      "num_tokens": 68458918.0,
      "step": 33480
    },
    {
      "entropy": 0.49625964760780333,
      "epoch": 0.5439204826908554,
      "grad_norm": 9.0,
      "learning_rate": 2.2149722901274382e-05,
      "loss": 0.5188203811645508,
      "mean_token_accuracy": 0.8696750447154045,
      "num_tokens": 68478584.0,
      "step": 33490
    },
    {
      "entropy": 0.46760306269861757,
      "epoch": 0.5440828954954808,
      "grad_norm": 5.9375,
      "learning_rate": 2.2136841475846586e-05,
      "loss": 0.4818405628204346,
      "mean_token_accuracy": 0.8756172627210617,
      "num_tokens": 68499434.0,
      "step": 33500
    },
    {
      "entropy": 0.5448808514280244,
      "epoch": 0.5442453083001064,
      "grad_norm": 7.65625,
      "learning_rate": 2.2123960820614875e-05,
      "loss": 0.5551241397857666,
      "mean_token_accuracy": 0.8585790652781725,
      "num_tokens": 68520045.0,
      "step": 33510
    },
    {
      "entropy": 0.48159954412840306,
      "epoch": 0.5444077211047319,
      "grad_norm": 12.25,
      "learning_rate": 2.2111080939044166e-05,
      "loss": 0.4752756118774414,
      "mean_token_accuracy": 0.8761787042021751,
      "num_tokens": 68541304.0,
      "step": 33520
    },
    {
      "entropy": 0.4943916528020054,
      "epoch": 0.5445701339093574,
      "grad_norm": 7.0,
      "learning_rate": 2.2098201834599172e-05,
      "loss": 0.5155853271484375,
      "mean_token_accuracy": 0.86746766269207,
      "num_tokens": 68561698.0,
      "step": 33530
    },
    {
      "entropy": 0.5340118791442364,
      "epoch": 0.544732546713983,
      "grad_norm": 13.1875,
      "learning_rate": 2.2085323510744414e-05,
      "loss": 0.5341507911682128,
      "mean_token_accuracy": 0.865185060352087,
      "num_tokens": 68581862.0,
      "step": 33540
    },
    {
      "entropy": 0.5048008145415224,
      "epoch": 0.5448949595186084,
      "grad_norm": 7.15625,
      "learning_rate": 2.2072445970944166e-05,
      "loss": 0.5080407142639161,
      "mean_token_accuracy": 0.8736014045774937,
      "num_tokens": 68603229.0,
      "step": 33550
    },
    {
      "entropy": 0.4910031044855714,
      "epoch": 0.5450573723232339,
      "grad_norm": 6.46875,
      "learning_rate": 2.205956921866254e-05,
      "loss": 0.4944744110107422,
      "mean_token_accuracy": 0.8780560486018658,
      "num_tokens": 68624069.0,
      "step": 33560
    },
    {
      "entropy": 0.5115315243136138,
      "epoch": 0.5452197851278595,
      "grad_norm": 6.90625,
      "learning_rate": 2.204669325736341e-05,
      "loss": 0.5151458263397217,
      "mean_token_accuracy": 0.8617636248469353,
      "num_tokens": 68644276.0,
      "step": 33570
    },
    {
      "entropy": 0.48581056750845164,
      "epoch": 0.545382197932485,
      "grad_norm": 7.09375,
      "learning_rate": 2.2033818090510422e-05,
      "loss": 0.5089701175689697,
      "mean_token_accuracy": 0.8723434753715992,
      "num_tokens": 68665579.0,
      "step": 33580
    },
    {
      "entropy": 0.4839793669525534,
      "epoch": 0.5455446107371105,
      "grad_norm": 4.75,
      "learning_rate": 2.202094372156703e-05,
      "loss": 0.5076016426086426,
      "mean_token_accuracy": 0.8720679119229316,
      "num_tokens": 68685170.0,
      "step": 33590
    },
    {
      "entropy": 0.5663524148985744,
      "epoch": 0.5457070235417361,
      "grad_norm": 6.96875,
      "learning_rate": 2.2008070153996483e-05,
      "loss": 0.5927314758300781,
      "mean_token_accuracy": 0.859750097990036,
      "num_tokens": 68705699.0,
      "step": 33600
    },
    {
      "entropy": 0.5428223856724799,
      "epoch": 0.5458694363463615,
      "grad_norm": 6.3125,
      "learning_rate": 2.1995197391261782e-05,
      "loss": 0.5632279872894287,
      "mean_token_accuracy": 0.8607417088001966,
      "num_tokens": 68726554.0,
      "step": 33610
    },
    {
      "entropy": 0.5287891311571002,
      "epoch": 0.546031849150987,
      "grad_norm": 7.40625,
      "learning_rate": 2.1982325436825743e-05,
      "loss": 0.5187876701354981,
      "mean_token_accuracy": 0.8670607775449752,
      "num_tokens": 68748173.0,
      "step": 33620
    },
    {
      "entropy": 0.5366115199401975,
      "epoch": 0.5461942619556126,
      "grad_norm": 7.09375,
      "learning_rate": 2.196945429415094e-05,
      "loss": 0.5440903663635254,
      "mean_token_accuracy": 0.8679592341184617,
      "num_tokens": 68770135.0,
      "step": 33630
    },
    {
      "entropy": 0.5490993307903409,
      "epoch": 0.5463566747602381,
      "grad_norm": 10.25,
      "learning_rate": 2.195658396669973e-05,
      "loss": 0.5862680435180664,
      "mean_token_accuracy": 0.8561960939317942,
      "num_tokens": 68790143.0,
      "step": 33640
    },
    {
      "entropy": 0.4596920091426,
      "epoch": 0.5465190875648636,
      "grad_norm": 7.84375,
      "learning_rate": 2.194371445793429e-05,
      "loss": 0.4475874900817871,
      "mean_token_accuracy": 0.8831777170300483,
      "num_tokens": 68809910.0,
      "step": 33650
    },
    {
      "entropy": 0.5676915866322816,
      "epoch": 0.5466815003694891,
      "grad_norm": 7.28125,
      "learning_rate": 2.193084577131652e-05,
      "loss": 0.5510447978973388,
      "mean_token_accuracy": 0.8659780111163855,
      "num_tokens": 68831340.0,
      "step": 33660
    },
    {
      "entropy": 0.5031451881397515,
      "epoch": 0.5468439131741146,
      "grad_norm": 13.125,
      "learning_rate": 2.1917977910308134e-05,
      "loss": 0.4976961612701416,
      "mean_token_accuracy": 0.8731383413076401,
      "num_tokens": 68852521.0,
      "step": 33670
    },
    {
      "entropy": 0.5082315989769995,
      "epoch": 0.5470063259787402,
      "grad_norm": 6.90625,
      "learning_rate": 2.1905110878370624e-05,
      "loss": 0.5394809722900391,
      "mean_token_accuracy": 0.881472185254097,
      "num_tokens": 68871990.0,
      "step": 33680
    },
    {
      "entropy": 0.5268838602583855,
      "epoch": 0.5471687387833657,
      "grad_norm": 8.0,
      "learning_rate": 2.1892244678965236e-05,
      "loss": 0.5325503349304199,
      "mean_token_accuracy": 0.8647225953638553,
      "num_tokens": 68891570.0,
      "step": 33690
    },
    {
      "entropy": 0.5486882321536541,
      "epoch": 0.5473311515879912,
      "grad_norm": 6.5,
      "learning_rate": 2.1879379315553022e-05,
      "loss": 0.5232158660888672,
      "mean_token_accuracy": 0.8633068993687629,
      "num_tokens": 68912236.0,
      "step": 33700
    },
    {
      "entropy": 0.5080272956751287,
      "epoch": 0.5474935643926168,
      "grad_norm": 11.5625,
      "learning_rate": 2.1866514791594776e-05,
      "loss": 0.5114031791687011,
      "mean_token_accuracy": 0.8697652652859688,
      "num_tokens": 68933295.0,
      "step": 33710
    },
    {
      "entropy": 0.520780862821266,
      "epoch": 0.5476559771972422,
      "grad_norm": 10.0625,
      "learning_rate": 2.1853651110551094e-05,
      "loss": 0.5365932941436767,
      "mean_token_accuracy": 0.8683645501732826,
      "num_tokens": 68954075.0,
      "step": 33720
    },
    {
      "entropy": 0.5485076494049281,
      "epoch": 0.5478183900018677,
      "grad_norm": 7.5625,
      "learning_rate": 2.184078827588234e-05,
      "loss": 0.5130879402160644,
      "mean_token_accuracy": 0.8693035677075386,
      "num_tokens": 68973666.0,
      "step": 33730
    },
    {
      "entropy": 0.40100018768571316,
      "epoch": 0.5479808028064933,
      "grad_norm": 7.40625,
      "learning_rate": 2.1827926291048638e-05,
      "loss": 0.3670509338378906,
      "mean_token_accuracy": 0.9048851899802685,
      "num_tokens": 68993260.0,
      "step": 33740
    },
    {
      "entropy": 0.5185030865948648,
      "epoch": 0.5481432156111188,
      "grad_norm": 8.25,
      "learning_rate": 2.1815065159509885e-05,
      "loss": 0.5452479362487793,
      "mean_token_accuracy": 0.8645625326782465,
      "num_tokens": 69014320.0,
      "step": 33750
    },
    {
      "entropy": 0.4520452327094972,
      "epoch": 0.5483056284157443,
      "grad_norm": 6.6875,
      "learning_rate": 2.180220488472577e-05,
      "loss": 0.4525897979736328,
      "mean_token_accuracy": 0.8847043149173259,
      "num_tokens": 69034069.0,
      "step": 33760
    },
    {
      "entropy": 0.4792000159155577,
      "epoch": 0.5484680412203699,
      "grad_norm": 5.8125,
      "learning_rate": 2.1789345470155725e-05,
      "loss": 0.47873744964599607,
      "mean_token_accuracy": 0.870918034017086,
      "num_tokens": 69054041.0,
      "step": 33770
    },
    {
      "entropy": 0.4692484918050468,
      "epoch": 0.5486304540249953,
      "grad_norm": 8.0,
      "learning_rate": 2.1776486919258965e-05,
      "loss": 0.4822674751281738,
      "mean_token_accuracy": 0.8736551456153393,
      "num_tokens": 69075049.0,
      "step": 33780
    },
    {
      "entropy": 0.5533133916556835,
      "epoch": 0.5487928668296208,
      "grad_norm": 9.75,
      "learning_rate": 2.176362923549447e-05,
      "loss": 0.599947452545166,
      "mean_token_accuracy": 0.8578931078314781,
      "num_tokens": 69095224.0,
      "step": 33790
    },
    {
      "entropy": 0.47489711875095963,
      "epoch": 0.5489552796342464,
      "grad_norm": 8.1875,
      "learning_rate": 2.1750772422320977e-05,
      "loss": 0.5112692832946777,
      "mean_token_accuracy": 0.8759594984352589,
      "num_tokens": 69115669.0,
      "step": 33800
    },
    {
      "entropy": 0.5247186324093491,
      "epoch": 0.5491176924388719,
      "grad_norm": 6.90625,
      "learning_rate": 2.1737916483197017e-05,
      "loss": 0.5539400100708007,
      "mean_token_accuracy": 0.860515919700265,
      "num_tokens": 69137870.0,
      "step": 33810
    },
    {
      "entropy": 0.45998743185773494,
      "epoch": 0.5492801052434974,
      "grad_norm": 9.25,
      "learning_rate": 2.172506142158084e-05,
      "loss": 0.44403390884399413,
      "mean_token_accuracy": 0.8863782770931721,
      "num_tokens": 69159243.0,
      "step": 33820
    },
    {
      "entropy": 0.4849973525851965,
      "epoch": 0.5494425180481229,
      "grad_norm": 6.8125,
      "learning_rate": 2.1712207240930507e-05,
      "loss": 0.4921281814575195,
      "mean_token_accuracy": 0.8784383341670037,
      "num_tokens": 69178783.0,
      "step": 33830
    },
    {
      "entropy": 0.459086738480255,
      "epoch": 0.5496049308527484,
      "grad_norm": 7.6875,
      "learning_rate": 2.169935394470382e-05,
      "loss": 0.4879415512084961,
      "mean_token_accuracy": 0.8805905781686306,
      "num_tokens": 69199069.0,
      "step": 33840
    },
    {
      "entropy": 0.5058258512523025,
      "epoch": 0.5497673436573739,
      "grad_norm": 6.96875,
      "learning_rate": 2.1686501536358334e-05,
      "loss": 0.5280792713165283,
      "mean_token_accuracy": 0.8809782069176435,
      "num_tokens": 69219119.0,
      "step": 33850
    },
    {
      "entropy": 0.507034054864198,
      "epoch": 0.5499297564619995,
      "grad_norm": 6.09375,
      "learning_rate": 2.1673650019351374e-05,
      "loss": 0.5133120059967041,
      "mean_token_accuracy": 0.8727388210594654,
      "num_tokens": 69239714.0,
      "step": 33860
    },
    {
      "entropy": 0.4960570607334375,
      "epoch": 0.550092169266625,
      "grad_norm": 5.46875,
      "learning_rate": 2.1660799397140045e-05,
      "loss": 0.4704521656036377,
      "mean_token_accuracy": 0.8823664113879204,
      "num_tokens": 69260074.0,
      "step": 33870
    },
    {
      "entropy": 0.523968976829201,
      "epoch": 0.5502545820712504,
      "grad_norm": 7.4375,
      "learning_rate": 2.164794967318118e-05,
      "loss": 0.5161651134490967,
      "mean_token_accuracy": 0.8688462004065514,
      "num_tokens": 69280491.0,
      "step": 33880
    },
    {
      "entropy": 0.4810294737573713,
      "epoch": 0.550416994875876,
      "grad_norm": 5.28125,
      "learning_rate": 2.1635100850931385e-05,
      "loss": 0.5014121055603027,
      "mean_token_accuracy": 0.870734129846096,
      "num_tokens": 69300322.0,
      "step": 33890
    },
    {
      "entropy": 0.4866638079751283,
      "epoch": 0.5505794076805015,
      "grad_norm": 7.28125,
      "learning_rate": 2.162225293384703e-05,
      "loss": 0.516301441192627,
      "mean_token_accuracy": 0.8712336659431458,
      "num_tokens": 69321225.0,
      "step": 33900
    },
    {
      "entropy": 0.5268201882019639,
      "epoch": 0.5507418204851271,
      "grad_norm": 8.5,
      "learning_rate": 2.1609405925384214e-05,
      "loss": 0.520289945602417,
      "mean_token_accuracy": 0.8740871854126453,
      "num_tokens": 69340836.0,
      "step": 33910
    },
    {
      "entropy": 0.47803408522158863,
      "epoch": 0.5509042332897526,
      "grad_norm": 6.84375,
      "learning_rate": 2.1596559828998837e-05,
      "loss": 0.45181875228881835,
      "mean_token_accuracy": 0.882061368227005,
      "num_tokens": 69361302.0,
      "step": 33920
    },
    {
      "entropy": 0.4712110889144242,
      "epoch": 0.551066646094378,
      "grad_norm": 5.9375,
      "learning_rate": 2.1583714648146504e-05,
      "loss": 0.44107418060302733,
      "mean_token_accuracy": 0.8868093475699425,
      "num_tokens": 69380602.0,
      "step": 33930
    },
    {
      "entropy": 0.502981933276169,
      "epoch": 0.5512290588990036,
      "grad_norm": 13.6875,
      "learning_rate": 2.1570870386282615e-05,
      "loss": 0.5006446361541748,
      "mean_token_accuracy": 0.8774623736739159,
      "num_tokens": 69399570.0,
      "step": 33940
    },
    {
      "entropy": 0.4921667458023876,
      "epoch": 0.5513914717036291,
      "grad_norm": 8.25,
      "learning_rate": 2.15580270468623e-05,
      "loss": 0.5033316135406494,
      "mean_token_accuracy": 0.8668560326099396,
      "num_tokens": 69418998.0,
      "step": 33950
    },
    {
      "entropy": 0.47964778216555715,
      "epoch": 0.5515538845082546,
      "grad_norm": 7.59375,
      "learning_rate": 2.1545184633340434e-05,
      "loss": 0.4808070659637451,
      "mean_token_accuracy": 0.882119120657444,
      "num_tokens": 69438662.0,
      "step": 33960
    },
    {
      "entropy": 0.43518576892092825,
      "epoch": 0.5517162973128802,
      "grad_norm": 11.1875,
      "learning_rate": 2.1532343149171675e-05,
      "loss": 0.4336407661437988,
      "mean_token_accuracy": 0.889134718477726,
      "num_tokens": 69458909.0,
      "step": 33970
    },
    {
      "entropy": 0.4329780153464526,
      "epoch": 0.5518787101175057,
      "grad_norm": 12.4375,
      "learning_rate": 2.1519502597810404e-05,
      "loss": 0.4519818782806396,
      "mean_token_accuracy": 0.8897483959794045,
      "num_tokens": 69479312.0,
      "step": 33980
    },
    {
      "entropy": 0.5104085814207793,
      "epoch": 0.5520411229221311,
      "grad_norm": 8.25,
      "learning_rate": 2.1506662982710747e-05,
      "loss": 0.5020429134368897,
      "mean_token_accuracy": 0.8746480740606785,
      "num_tokens": 69500181.0,
      "step": 33990
    },
    {
      "entropy": 0.47903393004089595,
      "epoch": 0.5522035357267567,
      "grad_norm": 7.9375,
      "learning_rate": 2.149382430732661e-05,
      "loss": 0.46356797218322754,
      "mean_token_accuracy": 0.8766411602497101,
      "num_tokens": 69520889.0,
      "step": 34000
    },
    {
      "entropy": 0.4973522378131747,
      "epoch": 0.5523659485313822,
      "grad_norm": 8.625,
      "learning_rate": 2.1480986575111605e-05,
      "loss": 0.49213075637817383,
      "mean_token_accuracy": 0.8718222737312317,
      "num_tokens": 69541367.0,
      "step": 34010
    },
    {
      "entropy": 0.4829199007712305,
      "epoch": 0.5525283613360077,
      "grad_norm": 11.4375,
      "learning_rate": 2.1468149789519118e-05,
      "loss": 0.45187668800354003,
      "mean_token_accuracy": 0.8809490002691746,
      "num_tokens": 69561008.0,
      "step": 34020
    },
    {
      "entropy": 0.425837971130386,
      "epoch": 0.5526907741406333,
      "grad_norm": 9.0,
      "learning_rate": 2.1455313954002285e-05,
      "loss": 0.44309382438659667,
      "mean_token_accuracy": 0.888883751630783,
      "num_tokens": 69581900.0,
      "step": 34030
    },
    {
      "entropy": 0.4946030925028026,
      "epoch": 0.5528531869452588,
      "grad_norm": 9.75,
      "learning_rate": 2.1442479072013958e-05,
      "loss": 0.5483765125274658,
      "mean_token_accuracy": 0.8593298442661762,
      "num_tokens": 69602478.0,
      "step": 34040
    },
    {
      "entropy": 0.5531061929184944,
      "epoch": 0.5530155997498842,
      "grad_norm": 9.0,
      "learning_rate": 2.1429645147006754e-05,
      "loss": 0.5781039714813232,
      "mean_token_accuracy": 0.8600262388586998,
      "num_tokens": 69623401.0,
      "step": 34050
    },
    {
      "entropy": 0.4687629329040647,
      "epoch": 0.5531780125545098,
      "grad_norm": 6.5625,
      "learning_rate": 2.1416812182433035e-05,
      "loss": 0.44787936210632323,
      "mean_token_accuracy": 0.8818682730197906,
      "num_tokens": 69642663.0,
      "step": 34060
    },
    {
      "entropy": 0.501962796645239,
      "epoch": 0.5533404253591353,
      "grad_norm": 6.09375,
      "learning_rate": 2.140398018174488e-05,
      "loss": 0.49068689346313477,
      "mean_token_accuracy": 0.8727966424077749,
      "num_tokens": 69661937.0,
      "step": 34070
    },
    {
      "entropy": 0.44825147041119634,
      "epoch": 0.5535028381637608,
      "grad_norm": 6.84375,
      "learning_rate": 2.1391149148394148e-05,
      "loss": 0.46203007698059084,
      "mean_token_accuracy": 0.8813515804708004,
      "num_tokens": 69683161.0,
      "step": 34080
    },
    {
      "entropy": 0.45110149746760725,
      "epoch": 0.5536652509683864,
      "grad_norm": 4.0,
      "learning_rate": 2.137831908583241e-05,
      "loss": 0.4494613170623779,
      "mean_token_accuracy": 0.8835438802838326,
      "num_tokens": 69703800.0,
      "step": 34090
    },
    {
      "entropy": 0.4750289857387543,
      "epoch": 0.5538276637730118,
      "grad_norm": 7.78125,
      "learning_rate": 2.1365489997510967e-05,
      "loss": 0.44250850677490233,
      "mean_token_accuracy": 0.8751864179968833,
      "num_tokens": 69723797.0,
      "step": 34100
    },
    {
      "entropy": 0.4391176711767912,
      "epoch": 0.5539900765776374,
      "grad_norm": 9.5625,
      "learning_rate": 2.1352661886880894e-05,
      "loss": 0.43527789115905763,
      "mean_token_accuracy": 0.8833812512457371,
      "num_tokens": 69743336.0,
      "step": 34110
    },
    {
      "entropy": 0.5042473228182643,
      "epoch": 0.5541524893822629,
      "grad_norm": 6.15625,
      "learning_rate": 2.1339834757392953e-05,
      "loss": 0.5285406112670898,
      "mean_token_accuracy": 0.8727432809770107,
      "num_tokens": 69763830.0,
      "step": 34120
    },
    {
      "entropy": 0.4043325950391591,
      "epoch": 0.5543149021868884,
      "grad_norm": 9.1875,
      "learning_rate": 2.13270086124977e-05,
      "loss": 0.3670841455459595,
      "mean_token_accuracy": 0.901578602194786,
      "num_tokens": 69783525.0,
      "step": 34130
    },
    {
      "entropy": 0.5282034748466685,
      "epoch": 0.554477314991514,
      "grad_norm": 7.28125,
      "learning_rate": 2.1314183455645386e-05,
      "loss": 0.5173911094665528,
      "mean_token_accuracy": 0.8635507769882679,
      "num_tokens": 69804212.0,
      "step": 34140
    },
    {
      "entropy": 0.4405926693230867,
      "epoch": 0.5546397277961395,
      "grad_norm": 9.4375,
      "learning_rate": 2.1301359290286e-05,
      "loss": 0.4537059783935547,
      "mean_token_accuracy": 0.8873001031577588,
      "num_tokens": 69823011.0,
      "step": 34150
    },
    {
      "entropy": 0.533292758371681,
      "epoch": 0.5548021406007649,
      "grad_norm": 8.625,
      "learning_rate": 2.1288536119869273e-05,
      "loss": 0.5358974933624268,
      "mean_token_accuracy": 0.8623500406742096,
      "num_tokens": 69844624.0,
      "step": 34160
    },
    {
      "entropy": 0.5525579012464732,
      "epoch": 0.5549645534053905,
      "grad_norm": 7.09375,
      "learning_rate": 2.127571394784469e-05,
      "loss": 0.5271025657653808,
      "mean_token_accuracy": 0.8684543140232563,
      "num_tokens": 69864762.0,
      "step": 34170
    },
    {
      "entropy": 0.5145470515359193,
      "epoch": 0.555126966210016,
      "grad_norm": 13.6875,
      "learning_rate": 2.1262892777661404e-05,
      "loss": 0.5839977264404297,
      "mean_token_accuracy": 0.8599419817328453,
      "num_tokens": 69885383.0,
      "step": 34180
    },
    {
      "entropy": 0.45715162423439326,
      "epoch": 0.5552893790146415,
      "grad_norm": 12.0,
      "learning_rate": 2.125007261276837e-05,
      "loss": 0.4776647090911865,
      "mean_token_accuracy": 0.8766120292246342,
      "num_tokens": 69904977.0,
      "step": 34190
    },
    {
      "entropy": 0.5160606379155069,
      "epoch": 0.5554517918192671,
      "grad_norm": 12.375,
      "learning_rate": 2.1237253456614227e-05,
      "loss": 0.5432309627532959,
      "mean_token_accuracy": 0.8706870548427105,
      "num_tokens": 69925515.0,
      "step": 34200
    },
    {
      "entropy": 0.4255206938600168,
      "epoch": 0.5556142046238925,
      "grad_norm": 6.90625,
      "learning_rate": 2.122443531264736e-05,
      "loss": 0.45685949325561526,
      "mean_token_accuracy": 0.883085048943758,
      "num_tokens": 69946226.0,
      "step": 34210
    },
    {
      "entropy": 0.4560190875083208,
      "epoch": 0.555776617428518,
      "grad_norm": 5.5625,
      "learning_rate": 2.1211618184315885e-05,
      "loss": 0.4478724479675293,
      "mean_token_accuracy": 0.8852220684289932,
      "num_tokens": 69966852.0,
      "step": 34220
    },
    {
      "entropy": 0.44994520633481444,
      "epoch": 0.5559390302331436,
      "grad_norm": 9.375,
      "learning_rate": 2.1198802075067624e-05,
      "loss": 0.4342986583709717,
      "mean_token_accuracy": 0.8882243476808072,
      "num_tokens": 69987177.0,
      "step": 34230
    },
    {
      "entropy": 0.49745861971750854,
      "epoch": 0.5561014430377691,
      "grad_norm": 7.625,
      "learning_rate": 2.1185986988350147e-05,
      "loss": 0.5224419116973877,
      "mean_token_accuracy": 0.8665642373263835,
      "num_tokens": 70007261.0,
      "step": 34240
    },
    {
      "entropy": 0.5065399416722357,
      "epoch": 0.5562638558423946,
      "grad_norm": 8.8125,
      "learning_rate": 2.117317292761075e-05,
      "loss": 0.5284903049468994,
      "mean_token_accuracy": 0.8648033283650876,
      "num_tokens": 70026609.0,
      "step": 34250
    },
    {
      "entropy": 0.4760451527312398,
      "epoch": 0.5564262686470202,
      "grad_norm": 8.6875,
      "learning_rate": 2.116035989629643e-05,
      "loss": 0.5072333335876464,
      "mean_token_accuracy": 0.8727751493453979,
      "num_tokens": 70047018.0,
      "step": 34260
    },
    {
      "entropy": 0.482893331348896,
      "epoch": 0.5565886814516456,
      "grad_norm": 6.90625,
      "learning_rate": 2.1147547897853924e-05,
      "loss": 0.5377142429351807,
      "mean_token_accuracy": 0.8681183196604252,
      "num_tokens": 70067879.0,
      "step": 34270
    },
    {
      "entropy": 0.5108528202166781,
      "epoch": 0.5567510942562711,
      "grad_norm": 9.3125,
      "learning_rate": 2.11347369357297e-05,
      "loss": 0.4921539306640625,
      "mean_token_accuracy": 0.8748038411140442,
      "num_tokens": 70088165.0,
      "step": 34280
    },
    {
      "entropy": 0.5154920598259196,
      "epoch": 0.5569135070608967,
      "grad_norm": 6.6875,
      "learning_rate": 2.1121927013369923e-05,
      "loss": 0.5216649055480957,
      "mean_token_accuracy": 0.8720419008284808,
      "num_tokens": 70109038.0,
      "step": 34290
    },
    {
      "entropy": 0.5074167178012431,
      "epoch": 0.5570759198655222,
      "grad_norm": 7.9375,
      "learning_rate": 2.1109118134220495e-05,
      "loss": 0.5181425094604493,
      "mean_token_accuracy": 0.8697020083665847,
      "num_tokens": 70129164.0,
      "step": 34300
    },
    {
      "entropy": 0.4884337204508483,
      "epoch": 0.5572383326701477,
      "grad_norm": 9.4375,
      "learning_rate": 2.1096310301727035e-05,
      "loss": 0.46446948051452636,
      "mean_token_accuracy": 0.8758760750293731,
      "num_tokens": 70149233.0,
      "step": 34310
    },
    {
      "entropy": 0.5107352879829705,
      "epoch": 0.5574007454747733,
      "grad_norm": 4.84375,
      "learning_rate": 2.1083503519334867e-05,
      "loss": 0.48760204315185546,
      "mean_token_accuracy": 0.8766903355717659,
      "num_tokens": 70169796.0,
      "step": 34320
    },
    {
      "entropy": 0.47839214936830105,
      "epoch": 0.5575631582793987,
      "grad_norm": 14.9375,
      "learning_rate": 2.107069779048907e-05,
      "loss": 0.4767155647277832,
      "mean_token_accuracy": 0.8806215837597847,
      "num_tokens": 70191282.0,
      "step": 34330
    },
    {
      "entropy": 0.46691877890843897,
      "epoch": 0.5577255710840243,
      "grad_norm": 7.28125,
      "learning_rate": 2.1057893118634382e-05,
      "loss": 0.48032350540161134,
      "mean_token_accuracy": 0.8726262830197811,
      "num_tokens": 70212136.0,
      "step": 34340
    },
    {
      "entropy": 0.46525451326742767,
      "epoch": 0.5578879838886498,
      "grad_norm": 5.5625,
      "learning_rate": 2.1045089507215312e-05,
      "loss": 0.4670886993408203,
      "mean_token_accuracy": 0.8801200285553932,
      "num_tokens": 70232730.0,
      "step": 34350
    },
    {
      "entropy": 0.515624888241291,
      "epoch": 0.5580503966932753,
      "grad_norm": 9.125,
      "learning_rate": 2.1032286959676058e-05,
      "loss": 0.5364349842071533,
      "mean_token_accuracy": 0.865925046056509,
      "num_tokens": 70253482.0,
      "step": 34360
    },
    {
      "entropy": 0.5046710526105016,
      "epoch": 0.5582128094979009,
      "grad_norm": 7.5,
      "learning_rate": 2.101948547946052e-05,
      "loss": 0.5447606086730957,
      "mean_token_accuracy": 0.8708166740834713,
      "num_tokens": 70273913.0,
      "step": 34370
    },
    {
      "entropy": 0.4832702114712447,
      "epoch": 0.5583752223025263,
      "grad_norm": 7.78125,
      "learning_rate": 2.1006685070012334e-05,
      "loss": 0.4952534198760986,
      "mean_token_accuracy": 0.8793291226029396,
      "num_tokens": 70294916.0,
      "step": 34380
    },
    {
      "entropy": 0.447965563647449,
      "epoch": 0.5585376351071518,
      "grad_norm": 10.1875,
      "learning_rate": 2.0993885734774842e-05,
      "loss": 0.45066261291503906,
      "mean_token_accuracy": 0.8823843881487846,
      "num_tokens": 70315875.0,
      "step": 34390
    },
    {
      "entropy": 0.4531406541354954,
      "epoch": 0.5587000479117774,
      "grad_norm": 8.5625,
      "learning_rate": 2.098108747719109e-05,
      "loss": 0.43594064712524416,
      "mean_token_accuracy": 0.8798953130841255,
      "num_tokens": 70335888.0,
      "step": 34400
    },
    {
      "entropy": 0.49902948264498265,
      "epoch": 0.5588624607164029,
      "grad_norm": 7.84375,
      "learning_rate": 2.0968290300703845e-05,
      "loss": 0.49384756088256837,
      "mean_token_accuracy": 0.8716082014143467,
      "num_tokens": 70358155.0,
      "step": 34410
    },
    {
      "entropy": 0.5495085548842326,
      "epoch": 0.5590248735210284,
      "grad_norm": 10.1875,
      "learning_rate": 2.095549420875556e-05,
      "loss": 0.5554338455200195,
      "mean_token_accuracy": 0.8570383638143539,
      "num_tokens": 70377752.0,
      "step": 34420
    },
    {
      "entropy": 0.5181454066187143,
      "epoch": 0.559187286325654,
      "grad_norm": 8.875,
      "learning_rate": 2.094269920478842e-05,
      "loss": 0.494932222366333,
      "mean_token_accuracy": 0.8733607940375805,
      "num_tokens": 70398505.0,
      "step": 34430
    },
    {
      "entropy": 0.4953248830744997,
      "epoch": 0.5593496991302794,
      "grad_norm": 4.90625,
      "learning_rate": 2.0929905292244322e-05,
      "loss": 0.4984746932983398,
      "mean_token_accuracy": 0.8765162169933319,
      "num_tokens": 70418657.0,
      "step": 34440
    },
    {
      "entropy": 0.5314223393099382,
      "epoch": 0.5595121119349049,
      "grad_norm": 7.8125,
      "learning_rate": 2.091711247456484e-05,
      "loss": 0.5134416103363038,
      "mean_token_accuracy": 0.8746513843536377,
      "num_tokens": 70438831.0,
      "step": 34450
    },
    {
      "entropy": 0.5021224085241556,
      "epoch": 0.5596745247395305,
      "grad_norm": 6.65625,
      "learning_rate": 2.0904320755191283e-05,
      "loss": 0.50058012008667,
      "mean_token_accuracy": 0.869370809942484,
      "num_tokens": 70458054.0,
      "step": 34460
    },
    {
      "entropy": 0.48685859008692206,
      "epoch": 0.559836937544156,
      "grad_norm": 13.6875,
      "learning_rate": 2.089153013756465e-05,
      "loss": 0.5046381473541259,
      "mean_token_accuracy": 0.8752521492540837,
      "num_tokens": 70478965.0,
      "step": 34470
    },
    {
      "entropy": 0.4195751552004367,
      "epoch": 0.5599993503487815,
      "grad_norm": 7.625,
      "learning_rate": 2.0878740625125633e-05,
      "loss": 0.4254859447479248,
      "mean_token_accuracy": 0.8863438948988914,
      "num_tokens": 70499725.0,
      "step": 34480
    },
    {
      "entropy": 0.4875357083976269,
      "epoch": 0.560161763153407,
      "grad_norm": 6.40625,
      "learning_rate": 2.0865952221314664e-05,
      "loss": 0.49010448455810546,
      "mean_token_accuracy": 0.871698647737503,
      "num_tokens": 70521428.0,
      "step": 34490
    },
    {
      "entropy": 0.448056256538257,
      "epoch": 0.5603241759580325,
      "grad_norm": 8.4375,
      "learning_rate": 2.0853164929571822e-05,
      "loss": 0.4018381118774414,
      "mean_token_accuracy": 0.8886393658816815,
      "num_tokens": 70541418.0,
      "step": 34500
    },
    {
      "entropy": 0.5430981740355492,
      "epoch": 0.560486588762658,
      "grad_norm": 5.78125,
      "learning_rate": 2.0840378753336942e-05,
      "loss": 0.5877717494964599,
      "mean_token_accuracy": 0.8484640121459961,
      "num_tokens": 70562553.0,
      "step": 34510
    },
    {
      "entropy": 0.4600142488256097,
      "epoch": 0.5606490015672836,
      "grad_norm": 6.28125,
      "learning_rate": 2.0827593696049525e-05,
      "loss": 0.4282061100006104,
      "mean_token_accuracy": 0.8892242413014173,
      "num_tokens": 70582558.0,
      "step": 34520
    },
    {
      "entropy": 0.4776402392424643,
      "epoch": 0.5608114143719091,
      "grad_norm": 6.25,
      "learning_rate": 2.081480976114878e-05,
      "loss": 0.486096715927124,
      "mean_token_accuracy": 0.8741221770644187,
      "num_tokens": 70602900.0,
      "step": 34530
    },
    {
      "entropy": 0.48262185615021735,
      "epoch": 0.5609738271765345,
      "grad_norm": 12.4375,
      "learning_rate": 2.080202695207361e-05,
      "loss": 0.5496110916137695,
      "mean_token_accuracy": 0.8730777487158775,
      "num_tokens": 70623249.0,
      "step": 34540
    },
    {
      "entropy": 0.39689771803095936,
      "epoch": 0.5611362399811601,
      "grad_norm": 7.625,
      "learning_rate": 2.0789245272262638e-05,
      "loss": 0.3777339458465576,
      "mean_token_accuracy": 0.8984143860638142,
      "num_tokens": 70642381.0,
      "step": 34550
    },
    {
      "entropy": 0.49803549514617773,
      "epoch": 0.5612986527857856,
      "grad_norm": 7.875,
      "learning_rate": 2.0776464725154137e-05,
      "loss": 0.5172844409942627,
      "mean_token_accuracy": 0.8698162406682968,
      "num_tokens": 70663173.0,
      "step": 34560
    },
    {
      "entropy": 0.5323108837008477,
      "epoch": 0.5614610655904112,
      "grad_norm": 8.4375,
      "learning_rate": 2.076368531418612e-05,
      "loss": 0.5813912868499755,
      "mean_token_accuracy": 0.8548152774572373,
      "num_tokens": 70683330.0,
      "step": 34570
    },
    {
      "entropy": 0.5067643852438778,
      "epoch": 0.5616234783950367,
      "grad_norm": 6.09375,
      "learning_rate": 2.075090704279628e-05,
      "loss": 0.4940641403198242,
      "mean_token_accuracy": 0.8793788313865661,
      "num_tokens": 70703283.0,
      "step": 34580
    },
    {
      "entropy": 0.5791491781594231,
      "epoch": 0.5617858911996622,
      "grad_norm": 10.125,
      "learning_rate": 2.0738129914421987e-05,
      "loss": 0.5809484958648682,
      "mean_token_accuracy": 0.858603087067604,
      "num_tokens": 70723836.0,
      "step": 34590
    },
    {
      "entropy": 0.4569374238606542,
      "epoch": 0.5619483040042877,
      "grad_norm": 8.4375,
      "learning_rate": 2.072535393250033e-05,
      "loss": 0.4381058692932129,
      "mean_token_accuracy": 0.8907965552061796,
      "num_tokens": 70743963.0,
      "step": 34600
    },
    {
      "entropy": 0.5010301585076377,
      "epoch": 0.5621107168089132,
      "grad_norm": 8.8125,
      "learning_rate": 2.071257910046807e-05,
      "loss": 0.48470497131347656,
      "mean_token_accuracy": 0.8748321052640676,
      "num_tokens": 70763876.0,
      "step": 34610
    },
    {
      "entropy": 0.5651604687795043,
      "epoch": 0.5622731296135387,
      "grad_norm": 6.875,
      "learning_rate": 2.069980542176167e-05,
      "loss": 0.5963554859161377,
      "mean_token_accuracy": 0.8573240779340268,
      "num_tokens": 70785318.0,
      "step": 34620
    },
    {
      "entropy": 0.47854647524654864,
      "epoch": 0.5624355424181643,
      "grad_norm": 6.0,
      "learning_rate": 2.0687032899817277e-05,
      "loss": 0.49874148368835447,
      "mean_token_accuracy": 0.881337983906269,
      "num_tokens": 70806107.0,
      "step": 34630
    },
    {
      "entropy": 0.4789683695882559,
      "epoch": 0.5625979552227898,
      "grad_norm": 9.375,
      "learning_rate": 2.0674261538070717e-05,
      "loss": 0.5007937908172607,
      "mean_token_accuracy": 0.8765693627297878,
      "num_tokens": 70826906.0,
      "step": 34640
    },
    {
      "entropy": 0.4771538968663663,
      "epoch": 0.5627603680274152,
      "grad_norm": 3.78125,
      "learning_rate": 2.0661491339957528e-05,
      "loss": 0.4639010429382324,
      "mean_token_accuracy": 0.8857994571328163,
      "num_tokens": 70847464.0,
      "step": 34650
    },
    {
      "entropy": 0.499520261446014,
      "epoch": 0.5629227808320408,
      "grad_norm": 9.875,
      "learning_rate": 2.0648722308912927e-05,
      "loss": 0.4947873592376709,
      "mean_token_accuracy": 0.8731232985854149,
      "num_tokens": 70867012.0,
      "step": 34660
    },
    {
      "entropy": 0.49294444895349443,
      "epoch": 0.5630851936366663,
      "grad_norm": 8.125,
      "learning_rate": 2.0635954448371796e-05,
      "loss": 0.4743643760681152,
      "mean_token_accuracy": 0.8843513838946819,
      "num_tokens": 70886604.0,
      "step": 34670
    },
    {
      "entropy": 0.4815678860293701,
      "epoch": 0.5632476064412918,
      "grad_norm": 13.6875,
      "learning_rate": 2.0623187761768734e-05,
      "loss": 0.49619135856628416,
      "mean_token_accuracy": 0.8789019376039505,
      "num_tokens": 70907758.0,
      "step": 34680
    },
    {
      "entropy": 0.42690492630936205,
      "epoch": 0.5634100192459174,
      "grad_norm": 9.25,
      "learning_rate": 2.0610422252537996e-05,
      "loss": 0.42781615257263184,
      "mean_token_accuracy": 0.8851472303271294,
      "num_tokens": 70927292.0,
      "step": 34690
    },
    {
      "entropy": 0.4761223338311538,
      "epoch": 0.5635724320505429,
      "grad_norm": 9.6875,
      "learning_rate": 2.0597657924113528e-05,
      "loss": 0.4773038387298584,
      "mean_token_accuracy": 0.8756697084754705,
      "num_tokens": 70947417.0,
      "step": 34700
    },
    {
      "entropy": 0.6108558603562415,
      "epoch": 0.5637348448551683,
      "grad_norm": 6.1875,
      "learning_rate": 2.0584894779928985e-05,
      "loss": 0.6215550422668457,
      "mean_token_accuracy": 0.8545157637447118,
      "num_tokens": 70969063.0,
      "step": 34710
    },
    {
      "entropy": 0.45702446419745685,
      "epoch": 0.5638972576597939,
      "grad_norm": 7.96875,
      "learning_rate": 2.0572132823417662e-05,
      "loss": 0.46111421585083007,
      "mean_token_accuracy": 0.8747239202260971,
      "num_tokens": 70988585.0,
      "step": 34720
    },
    {
      "entropy": 0.4605277282302268,
      "epoch": 0.5640596704644194,
      "grad_norm": 8.25,
      "learning_rate": 2.055937205801256e-05,
      "loss": 0.4342050552368164,
      "mean_token_accuracy": 0.8870020620524883,
      "num_tokens": 71008208.0,
      "step": 34730
    },
    {
      "entropy": 0.4896334719378501,
      "epoch": 0.5642220832690449,
      "grad_norm": 10.6875,
      "learning_rate": 2.054661248714636e-05,
      "loss": 0.45511322021484374,
      "mean_token_accuracy": 0.8830964401364326,
      "num_tokens": 71028189.0,
      "step": 34740
    },
    {
      "entropy": 0.4962799789849669,
      "epoch": 0.5643844960736705,
      "grad_norm": 9.5625,
      "learning_rate": 2.05338541142514e-05,
      "loss": 0.5084120273590088,
      "mean_token_accuracy": 0.8721415407955646,
      "num_tokens": 71047824.0,
      "step": 34750
    },
    {
      "entropy": 0.45862949253059926,
      "epoch": 0.564546908878296,
      "grad_norm": 5.84375,
      "learning_rate": 2.0521096942759723e-05,
      "loss": 0.4405822277069092,
      "mean_token_accuracy": 0.8863287039101124,
      "num_tokens": 71068283.0,
      "step": 34760
    },
    {
      "entropy": 0.49449151852168144,
      "epoch": 0.5647093216829214,
      "grad_norm": 6.6875,
      "learning_rate": 2.0508340976103043e-05,
      "loss": 0.5077680110931396,
      "mean_token_accuracy": 0.8650071255862712,
      "num_tokens": 71088106.0,
      "step": 34770
    },
    {
      "entropy": 0.4664253987139091,
      "epoch": 0.564871734487547,
      "grad_norm": 6.875,
      "learning_rate": 2.0495586217712725e-05,
      "loss": 0.4763810634613037,
      "mean_token_accuracy": 0.8787409633398056,
      "num_tokens": 71107801.0,
      "step": 34780
    },
    {
      "entropy": 0.5265321785816923,
      "epoch": 0.5650341472921725,
      "grad_norm": 6.28125,
      "learning_rate": 2.0482832671019847e-05,
      "loss": 0.5626398086547851,
      "mean_token_accuracy": 0.8643428958952427,
      "num_tokens": 71128909.0,
      "step": 34790
    },
    {
      "entropy": 0.5055589692550712,
      "epoch": 0.5651965600967981,
      "grad_norm": 8.0625,
      "learning_rate": 2.047008033945512e-05,
      "loss": 0.502316665649414,
      "mean_token_accuracy": 0.8745052278041839,
      "num_tokens": 71149971.0,
      "step": 34800
    },
    {
      "entropy": 0.47698192845564336,
      "epoch": 0.5653589729014236,
      "grad_norm": 9.5,
      "learning_rate": 2.0457329226448967e-05,
      "loss": 0.5061979293823242,
      "mean_token_accuracy": 0.8723123468458652,
      "num_tokens": 71169525.0,
      "step": 34810
    },
    {
      "entropy": 0.4782237640582025,
      "epoch": 0.565521385706049,
      "grad_norm": 6.78125,
      "learning_rate": 2.0444579335431464e-05,
      "loss": 0.49452881813049315,
      "mean_token_accuracy": 0.8736629325896501,
      "num_tokens": 71189826.0,
      "step": 34820
    },
    {
      "entropy": 0.5603102063760161,
      "epoch": 0.5656837985106746,
      "grad_norm": 8.3125,
      "learning_rate": 2.043183066983236e-05,
      "loss": 0.5362167835235596,
      "mean_token_accuracy": 0.8641893390566111,
      "num_tokens": 71212809.0,
      "step": 34830
    },
    {
      "entropy": 0.47653927877545355,
      "epoch": 0.5658462113153001,
      "grad_norm": 7.78125,
      "learning_rate": 2.041908323308106e-05,
      "loss": 0.516403341293335,
      "mean_token_accuracy": 0.8692436680197716,
      "num_tokens": 71232661.0,
      "step": 34840
    },
    {
      "entropy": 0.4793540730141103,
      "epoch": 0.5660086241199256,
      "grad_norm": 6.625,
      "learning_rate": 2.040633702860668e-05,
      "loss": 0.454522705078125,
      "mean_token_accuracy": 0.8846742898225785,
      "num_tokens": 71253827.0,
      "step": 34850
    },
    {
      "entropy": 0.48299675539601594,
      "epoch": 0.5661710369245512,
      "grad_norm": 7.75,
      "learning_rate": 2.039359205983795e-05,
      "loss": 0.5054035186767578,
      "mean_token_accuracy": 0.8699159834533929,
      "num_tokens": 71274366.0,
      "step": 34860
    },
    {
      "entropy": 0.45336519037373363,
      "epoch": 0.5663334497291767,
      "grad_norm": 8.0,
      "learning_rate": 2.0380848330203314e-05,
      "loss": 0.3939234495162964,
      "mean_token_accuracy": 0.8970321334898472,
      "num_tokens": 71294116.0,
      "step": 34870
    },
    {
      "entropy": 0.5371098374482244,
      "epoch": 0.5664958625338021,
      "grad_norm": 11.75,
      "learning_rate": 2.0368105843130862e-05,
      "loss": 0.5254314422607422,
      "mean_token_accuracy": 0.8690562717616558,
      "num_tokens": 71314566.0,
      "step": 34880
    },
    {
      "entropy": 0.48186783902347086,
      "epoch": 0.5666582753384277,
      "grad_norm": 10.5625,
      "learning_rate": 2.0355364602048342e-05,
      "loss": 0.4568496227264404,
      "mean_token_accuracy": 0.8833194449543953,
      "num_tokens": 71334539.0,
      "step": 34890
    },
    {
      "entropy": 0.4095654758391902,
      "epoch": 0.5668206881430532,
      "grad_norm": 10.1875,
      "learning_rate": 2.034262461038319e-05,
      "loss": 0.40563116073608396,
      "mean_token_accuracy": 0.8931937098503113,
      "num_tokens": 71354976.0,
      "step": 34900
    },
    {
      "entropy": 0.44991155792959037,
      "epoch": 0.5669831009476787,
      "grad_norm": 8.0,
      "learning_rate": 2.0329885871562475e-05,
      "loss": 0.48073339462280273,
      "mean_token_accuracy": 0.8749138675630093,
      "num_tokens": 71377026.0,
      "step": 34910
    },
    {
      "entropy": 0.5136917638126761,
      "epoch": 0.5671455137523043,
      "grad_norm": 7.25,
      "learning_rate": 2.0317148389012956e-05,
      "loss": 0.5577004432678223,
      "mean_token_accuracy": 0.8620994292199612,
      "num_tokens": 71396577.0,
      "step": 34920
    },
    {
      "entropy": 0.4831500891596079,
      "epoch": 0.5673079265569297,
      "grad_norm": 5.25,
      "learning_rate": 2.0304412166161052e-05,
      "loss": 0.4902970790863037,
      "mean_token_accuracy": 0.8798348784446717,
      "num_tokens": 71417750.0,
      "step": 34930
    },
    {
      "entropy": 0.47617808654904364,
      "epoch": 0.5674703393615552,
      "grad_norm": 10.5625,
      "learning_rate": 2.0291677206432823e-05,
      "loss": 0.4725971698760986,
      "mean_token_accuracy": 0.8770605310797691,
      "num_tokens": 71439085.0,
      "step": 34940
    },
    {
      "entropy": 0.4150344125460833,
      "epoch": 0.5676327521661808,
      "grad_norm": 6.96875,
      "learning_rate": 2.0278943513254005e-05,
      "loss": 0.40584430694580076,
      "mean_token_accuracy": 0.8975309379398823,
      "num_tokens": 71457851.0,
      "step": 34950
    },
    {
      "entropy": 0.5292735514231026,
      "epoch": 0.5677951649708063,
      "grad_norm": 8.125,
      "learning_rate": 2.0266211090049996e-05,
      "loss": 0.5454298973083496,
      "mean_token_accuracy": 0.8667261973023415,
      "num_tokens": 71477803.0,
      "step": 34960
    },
    {
      "entropy": 0.5285453464370221,
      "epoch": 0.5679575777754318,
      "grad_norm": 8.5625,
      "learning_rate": 2.025347994024584e-05,
      "loss": 0.5311562538146972,
      "mean_token_accuracy": 0.867181496322155,
      "num_tokens": 71498260.0,
      "step": 34970
    },
    {
      "entropy": 0.43806261613499375,
      "epoch": 0.5681199905800574,
      "grad_norm": 5.9375,
      "learning_rate": 2.0240750067266255e-05,
      "loss": 0.448976993560791,
      "mean_token_accuracy": 0.8785400088876486,
      "num_tokens": 71519345.0,
      "step": 34980
    },
    {
      "entropy": 0.5054113526362926,
      "epoch": 0.5682824033846828,
      "grad_norm": 8.25,
      "learning_rate": 2.022802147453559e-05,
      "loss": 0.5149348258972168,
      "mean_token_accuracy": 0.8674313440918923,
      "num_tokens": 71540488.0,
      "step": 34990
    },
    {
      "entropy": 0.5302013392560184,
      "epoch": 0.5684448161893083,
      "grad_norm": 5.40625,
      "learning_rate": 2.0215294165477867e-05,
      "loss": 0.5469722747802734,
      "mean_token_accuracy": 0.8635726720094681,
      "num_tokens": 71563109.0,
      "step": 35000
    },
    {
      "entropy": 0.4694493652903475,
      "epoch": 0.5686072289939339,
      "grad_norm": 5.90625,
      "learning_rate": 2.020256814351678e-05,
      "loss": 0.4577460765838623,
      "mean_token_accuracy": 0.8836593680083752,
      "num_tokens": 71583247.0,
      "step": 35010
    },
    {
      "entropy": 0.4979432867374271,
      "epoch": 0.5687696417985594,
      "grad_norm": 9.5,
      "learning_rate": 2.0189843412075632e-05,
      "loss": 0.5117532253265381,
      "mean_token_accuracy": 0.8673950858414173,
      "num_tokens": 71603861.0,
      "step": 35020
    },
    {
      "entropy": 0.5268384081777185,
      "epoch": 0.568932054603185,
      "grad_norm": 8.5,
      "learning_rate": 2.0177119974577423e-05,
      "loss": 0.5456787586212158,
      "mean_token_accuracy": 0.8603182222694159,
      "num_tokens": 71624383.0,
      "step": 35030
    },
    {
      "entropy": 0.5225970756262541,
      "epoch": 0.5690944674078104,
      "grad_norm": 10.5625,
      "learning_rate": 2.0164397834444782e-05,
      "loss": 0.5266626834869385,
      "mean_token_accuracy": 0.868008466809988,
      "num_tokens": 71644111.0,
      "step": 35040
    },
    {
      "entropy": 0.48486509285867213,
      "epoch": 0.5692568802124359,
      "grad_norm": 6.84375,
      "learning_rate": 2.015167699509998e-05,
      "loss": 0.49770417213439944,
      "mean_token_accuracy": 0.8824911795556545,
      "num_tokens": 71664275.0,
      "step": 35050
    },
    {
      "entropy": 0.5637024248950183,
      "epoch": 0.5694192930170615,
      "grad_norm": 9.3125,
      "learning_rate": 2.0138957459964966e-05,
      "loss": 0.5851668357849121,
      "mean_token_accuracy": 0.8578349262475967,
      "num_tokens": 71685989.0,
      "step": 35060
    },
    {
      "entropy": 0.5482973799109458,
      "epoch": 0.569581705821687,
      "grad_norm": 9.0,
      "learning_rate": 2.012623923246133e-05,
      "loss": 0.5288400173187255,
      "mean_token_accuracy": 0.8677950225770473,
      "num_tokens": 71706896.0,
      "step": 35070
    },
    {
      "entropy": 0.5307497866917401,
      "epoch": 0.5697441186263125,
      "grad_norm": 7.625,
      "learning_rate": 2.0113522316010283e-05,
      "loss": 0.5533623695373535,
      "mean_token_accuracy": 0.8618308059871197,
      "num_tokens": 71727167.0,
      "step": 35080
    },
    {
      "entropy": 0.5067502087913454,
      "epoch": 0.5699065314309381,
      "grad_norm": 8.1875,
      "learning_rate": 2.010080671403272e-05,
      "loss": 0.49077458381652833,
      "mean_token_accuracy": 0.8750096470117569,
      "num_tokens": 71747153.0,
      "step": 35090
    },
    {
      "entropy": 0.508166627259925,
      "epoch": 0.5700689442355635,
      "grad_norm": 7.8125,
      "learning_rate": 2.0088092429949163e-05,
      "loss": 0.4806680679321289,
      "mean_token_accuracy": 0.8751934170722961,
      "num_tokens": 71768260.0,
      "step": 35100
    },
    {
      "entropy": 0.46800614716485145,
      "epoch": 0.570231357040189,
      "grad_norm": 9.1875,
      "learning_rate": 2.007537946717977e-05,
      "loss": 0.5247076034545899,
      "mean_token_accuracy": 0.8686347384005785,
      "num_tokens": 71789366.0,
      "step": 35110
    },
    {
      "entropy": 0.4788985221646726,
      "epoch": 0.5703937698448146,
      "grad_norm": 5.65625,
      "learning_rate": 2.0062667829144377e-05,
      "loss": 0.4638078689575195,
      "mean_token_accuracy": 0.8813885979354381,
      "num_tokens": 71810142.0,
      "step": 35120
    },
    {
      "entropy": 0.45482188570313153,
      "epoch": 0.5705561826494401,
      "grad_norm": 4.34375,
      "learning_rate": 2.0049957519262434e-05,
      "loss": 0.4294625759124756,
      "mean_token_accuracy": 0.8898667350411416,
      "num_tokens": 71830390.0,
      "step": 35130
    },
    {
      "entropy": 0.522005493263714,
      "epoch": 0.5707185954540656,
      "grad_norm": 10.0,
      "learning_rate": 2.0037248540953038e-05,
      "loss": 0.5120690345764161,
      "mean_token_accuracy": 0.8703705713152885,
      "num_tokens": 71852728.0,
      "step": 35140
    },
    {
      "entropy": 0.43475394051056354,
      "epoch": 0.5708810082586911,
      "grad_norm": 5.5625,
      "learning_rate": 2.0024540897634947e-05,
      "loss": 0.4568382740020752,
      "mean_token_accuracy": 0.8903486330062151,
      "num_tokens": 71872652.0,
      "step": 35150
    },
    {
      "entropy": 0.5192686026683078,
      "epoch": 0.5710434210633166,
      "grad_norm": 5.125,
      "learning_rate": 2.001183459272652e-05,
      "loss": 0.5284985542297364,
      "mean_token_accuracy": 0.8661310613155365,
      "num_tokens": 71893520.0,
      "step": 35160
    },
    {
      "entropy": 0.4849213314475492,
      "epoch": 0.5712058338679421,
      "grad_norm": 11.4375,
      "learning_rate": 1.9999129629645806e-05,
      "loss": 0.4959848880767822,
      "mean_token_accuracy": 0.8776094879955053,
      "num_tokens": 71913553.0,
      "step": 35170
    },
    {
      "entropy": 0.5057167560094967,
      "epoch": 0.5713682466725677,
      "grad_norm": 10.5625,
      "learning_rate": 1.998642601181047e-05,
      "loss": 0.5124267101287842,
      "mean_token_accuracy": 0.8666249215602875,
      "num_tokens": 71934676.0,
      "step": 35180
    },
    {
      "entropy": 0.5166067194426432,
      "epoch": 0.5715306594771932,
      "grad_norm": 11.8125,
      "learning_rate": 1.997372374263779e-05,
      "loss": 0.5389686584472656,
      "mean_token_accuracy": 0.865152994915843,
      "num_tokens": 71954271.0,
      "step": 35190
    },
    {
      "entropy": 0.4393458048463799,
      "epoch": 0.5716930722818186,
      "grad_norm": 8.0,
      "learning_rate": 1.9961022825544735e-05,
      "loss": 0.45366339683532714,
      "mean_token_accuracy": 0.8860217802226543,
      "num_tokens": 71973412.0,
      "step": 35200
    },
    {
      "entropy": 0.4900574030121788,
      "epoch": 0.5718554850864442,
      "grad_norm": 8.125,
      "learning_rate": 1.9948323263947853e-05,
      "loss": 0.48357267379760743,
      "mean_token_accuracy": 0.8728819444775582,
      "num_tokens": 71994229.0,
      "step": 35210
    },
    {
      "entropy": 0.4179455946199596,
      "epoch": 0.5720178978910697,
      "grad_norm": 5.28125,
      "learning_rate": 1.9935625061263356e-05,
      "loss": 0.44046435356140134,
      "mean_token_accuracy": 0.890372796356678,
      "num_tokens": 72013770.0,
      "step": 35220
    },
    {
      "entropy": 0.49527778872288764,
      "epoch": 0.5721803106956952,
      "grad_norm": 15.8125,
      "learning_rate": 1.9922928220907116e-05,
      "loss": 0.5227673530578614,
      "mean_token_accuracy": 0.8691906869411469,
      "num_tokens": 72034092.0,
      "step": 35230
    },
    {
      "entropy": 0.5062321400269866,
      "epoch": 0.5723427235003208,
      "grad_norm": 7.78125,
      "learning_rate": 1.991023274629459e-05,
      "loss": 0.5155374526977539,
      "mean_token_accuracy": 0.8683955699205399,
      "num_tokens": 72053333.0,
      "step": 35240
    },
    {
      "entropy": 0.4899992852238938,
      "epoch": 0.5725051363049463,
      "grad_norm": 8.5,
      "learning_rate": 1.989753864084089e-05,
      "loss": 0.5187952995300293,
      "mean_token_accuracy": 0.8693899199366569,
      "num_tokens": 72075098.0,
      "step": 35250
    },
    {
      "entropy": 0.4520259898621589,
      "epoch": 0.5726675491095718,
      "grad_norm": 11.0625,
      "learning_rate": 1.988484590796077e-05,
      "loss": 0.4368502140045166,
      "mean_token_accuracy": 0.8857277475297451,
      "num_tokens": 72094792.0,
      "step": 35260
    },
    {
      "entropy": 0.5363670466933399,
      "epoch": 0.5728299619141973,
      "grad_norm": 5.90625,
      "learning_rate": 1.9872154551068587e-05,
      "loss": 0.565367317199707,
      "mean_token_accuracy": 0.8561111085116864,
      "num_tokens": 72115364.0,
      "step": 35270
    },
    {
      "entropy": 0.47905132616870105,
      "epoch": 0.5729923747188228,
      "grad_norm": 9.3125,
      "learning_rate": 1.985946457357837e-05,
      "loss": 0.4833566665649414,
      "mean_token_accuracy": 0.8734516359865665,
      "num_tokens": 72135474.0,
      "step": 35280
    },
    {
      "entropy": 0.5327798760961742,
      "epoch": 0.5731547875234484,
      "grad_norm": 8.1875,
      "learning_rate": 1.9846775978903715e-05,
      "loss": 0.5180971145629882,
      "mean_token_accuracy": 0.8669820465147495,
      "num_tokens": 72154897.0,
      "step": 35290
    },
    {
      "entropy": 0.5559980712831021,
      "epoch": 0.5733172003280739,
      "grad_norm": 7.03125,
      "learning_rate": 1.983408877045791e-05,
      "loss": 0.5702974319458007,
      "mean_token_accuracy": 0.8580952174961567,
      "num_tokens": 72175256.0,
      "step": 35300
    },
    {
      "entropy": 0.5266974648460746,
      "epoch": 0.5734796131326994,
      "grad_norm": 8.625,
      "learning_rate": 1.982140295165384e-05,
      "loss": 0.5337141036987305,
      "mean_token_accuracy": 0.8667919367551804,
      "num_tokens": 72195687.0,
      "step": 35310
    },
    {
      "entropy": 0.5135111227631569,
      "epoch": 0.5736420259373249,
      "grad_norm": 6.8125,
      "learning_rate": 1.980871852590401e-05,
      "loss": 0.5278810501098633,
      "mean_token_accuracy": 0.8711406491696835,
      "num_tokens": 72216419.0,
      "step": 35320
    },
    {
      "entropy": 0.4720841435249895,
      "epoch": 0.5738044387419504,
      "grad_norm": 6.84375,
      "learning_rate": 1.979603549662056e-05,
      "loss": 0.4846165180206299,
      "mean_token_accuracy": 0.8822045281529427,
      "num_tokens": 72237783.0,
      "step": 35330
    },
    {
      "entropy": 0.44560605542501436,
      "epoch": 0.5739668515465759,
      "grad_norm": 7.6875,
      "learning_rate": 1.9783353867215265e-05,
      "loss": 0.43014960289001464,
      "mean_token_accuracy": 0.8870272479951382,
      "num_tokens": 72258499.0,
      "step": 35340
    },
    {
      "entropy": 0.545111465966329,
      "epoch": 0.5741292643512015,
      "grad_norm": 8.25,
      "learning_rate": 1.9770673641099498e-05,
      "loss": 0.5253788471221924,
      "mean_token_accuracy": 0.8628722935914993,
      "num_tokens": 72279692.0,
      "step": 35350
    },
    {
      "entropy": 0.4645933476276696,
      "epoch": 0.574291677155827,
      "grad_norm": 6.25,
      "learning_rate": 1.975799482168427e-05,
      "loss": 0.4438615322113037,
      "mean_token_accuracy": 0.8875696711242199,
      "num_tokens": 72299471.0,
      "step": 35360
    },
    {
      "entropy": 0.4763076829724014,
      "epoch": 0.5744540899604524,
      "grad_norm": 6.9375,
      "learning_rate": 1.974531741238023e-05,
      "loss": 0.48317227363586424,
      "mean_token_accuracy": 0.8727102398872375,
      "num_tokens": 72319585.0,
      "step": 35370
    },
    {
      "entropy": 0.4537927029421553,
      "epoch": 0.574616502765078,
      "grad_norm": 5.8125,
      "learning_rate": 1.97326414165976e-05,
      "loss": 0.4454843044281006,
      "mean_token_accuracy": 0.8842326655983925,
      "num_tokens": 72339531.0,
      "step": 35380
    },
    {
      "entropy": 0.518293599714525,
      "epoch": 0.5747789155697035,
      "grad_norm": 10.1875,
      "learning_rate": 1.9719966837746276e-05,
      "loss": 0.5239322662353516,
      "mean_token_accuracy": 0.8714905016124248,
      "num_tokens": 72359162.0,
      "step": 35390
    },
    {
      "entropy": 0.5017513856291771,
      "epoch": 0.574941328374329,
      "grad_norm": 5.625,
      "learning_rate": 1.970729367923573e-05,
      "loss": 0.4994038581848145,
      "mean_token_accuracy": 0.877447097748518,
      "num_tokens": 72380379.0,
      "step": 35400
    },
    {
      "entropy": 0.47544234625529497,
      "epoch": 0.5751037411789546,
      "grad_norm": 8.1875,
      "learning_rate": 1.969462194447509e-05,
      "loss": 0.49881682395935056,
      "mean_token_accuracy": 0.8745658576488495,
      "num_tokens": 72400727.0,
      "step": 35410
    },
    {
      "entropy": 0.394749592943117,
      "epoch": 0.57526615398358,
      "grad_norm": 10.25,
      "learning_rate": 1.9681951636873067e-05,
      "loss": 0.3541309118270874,
      "mean_token_accuracy": 0.9039084300398826,
      "num_tokens": 72420304.0,
      "step": 35420
    },
    {
      "entropy": 0.4846257843775675,
      "epoch": 0.5754285667882055,
      "grad_norm": 6.0625,
      "learning_rate": 1.9669282759837996e-05,
      "loss": 0.46775059700012206,
      "mean_token_accuracy": 0.878752438351512,
      "num_tokens": 72441385.0,
      "step": 35430
    },
    {
      "entropy": 0.512512823799625,
      "epoch": 0.5755909795928311,
      "grad_norm": 6.75,
      "learning_rate": 1.965661531677785e-05,
      "loss": 0.511643362045288,
      "mean_token_accuracy": 0.8738086454570293,
      "num_tokens": 72461314.0,
      "step": 35440
    },
    {
      "entropy": 0.47105136755853894,
      "epoch": 0.5757533923974566,
      "grad_norm": 13.0625,
      "learning_rate": 1.9643949311100193e-05,
      "loss": 0.49589967727661133,
      "mean_token_accuracy": 0.8749259889125824,
      "num_tokens": 72482582.0,
      "step": 35450
    },
    {
      "entropy": 0.5483934213174507,
      "epoch": 0.5759158052020822,
      "grad_norm": 7.59375,
      "learning_rate": 1.9631284746212207e-05,
      "loss": 0.5773508548736572,
      "mean_token_accuracy": 0.8603455200791359,
      "num_tokens": 72504052.0,
      "step": 35460
    },
    {
      "entropy": 0.4333242515102029,
      "epoch": 0.5760782180067077,
      "grad_norm": 7.6875,
      "learning_rate": 1.9618621625520695e-05,
      "loss": 0.4417896270751953,
      "mean_token_accuracy": 0.8854976907372475,
      "num_tokens": 72523657.0,
      "step": 35470
    },
    {
      "entropy": 0.5490050204098225,
      "epoch": 0.5762406308113331,
      "grad_norm": 8.875,
      "learning_rate": 1.9605959952432045e-05,
      "loss": 0.5026678085327149,
      "mean_token_accuracy": 0.8730398714542389,
      "num_tokens": 72545133.0,
      "step": 35480
    },
    {
      "entropy": 0.4752177944406867,
      "epoch": 0.5764030436159587,
      "grad_norm": 8.0625,
      "learning_rate": 1.95932997303523e-05,
      "loss": 0.4895893096923828,
      "mean_token_accuracy": 0.8736348383128643,
      "num_tokens": 72566157.0,
      "step": 35490
    },
    {
      "entropy": 0.47617207895964386,
      "epoch": 0.5765654564205842,
      "grad_norm": 8.875,
      "learning_rate": 1.958064096268708e-05,
      "loss": 0.5119708061218262,
      "mean_token_accuracy": 0.8702007882297039,
      "num_tokens": 72586335.0,
      "step": 35500
    },
    {
      "entropy": 0.485381368547678,
      "epoch": 0.5767278692252097,
      "grad_norm": 6.875,
      "learning_rate": 1.9567983652841618e-05,
      "loss": 0.5113878726959229,
      "mean_token_accuracy": 0.8741186879575252,
      "num_tokens": 72606460.0,
      "step": 35510
    },
    {
      "entropy": 0.46341732763685284,
      "epoch": 0.5768902820298353,
      "grad_norm": 8.375,
      "learning_rate": 1.9555327804220754e-05,
      "loss": 0.45360622406005857,
      "mean_token_accuracy": 0.8794531367719174,
      "num_tokens": 72627322.0,
      "step": 35520
    },
    {
      "entropy": 0.41934563680551945,
      "epoch": 0.5770526948344608,
      "grad_norm": 7.78125,
      "learning_rate": 1.9542673420228964e-05,
      "loss": 0.3949080228805542,
      "mean_token_accuracy": 0.8971073001623153,
      "num_tokens": 72646666.0,
      "step": 35530
    },
    {
      "entropy": 0.4801793386694044,
      "epoch": 0.5772151076390862,
      "grad_norm": 10.75,
      "learning_rate": 1.9530020504270273e-05,
      "loss": 0.48345298767089845,
      "mean_token_accuracy": 0.8773797497153282,
      "num_tokens": 72667594.0,
      "step": 35540
    },
    {
      "entropy": 0.49723189789801836,
      "epoch": 0.5773775204437118,
      "grad_norm": 6.375,
      "learning_rate": 1.9517369059748365e-05,
      "loss": 0.4977464199066162,
      "mean_token_accuracy": 0.8691658906638622,
      "num_tokens": 72688273.0,
      "step": 35550
    },
    {
      "entropy": 0.44376993768382816,
      "epoch": 0.5775399332483373,
      "grad_norm": 7.4375,
      "learning_rate": 1.9504719090066502e-05,
      "loss": 0.4893683910369873,
      "mean_token_accuracy": 0.8861746147274971,
      "num_tokens": 72708834.0,
      "step": 35560
    },
    {
      "entropy": 0.4419066194910556,
      "epoch": 0.5777023460529628,
      "grad_norm": 8.625,
      "learning_rate": 1.9492070598627553e-05,
      "loss": 0.4583296775817871,
      "mean_token_accuracy": 0.8798258394002915,
      "num_tokens": 72730356.0,
      "step": 35570
    },
    {
      "entropy": 0.5207978910300881,
      "epoch": 0.5778647588575884,
      "grad_norm": 6.59375,
      "learning_rate": 1.9479423588833996e-05,
      "loss": 0.5213712692260742,
      "mean_token_accuracy": 0.8730552222579717,
      "num_tokens": 72751759.0,
      "step": 35580
    },
    {
      "entropy": 0.47828973256982865,
      "epoch": 0.5780271716622138,
      "grad_norm": 11.125,
      "learning_rate": 1.9466778064087886e-05,
      "loss": 0.4969003200531006,
      "mean_token_accuracy": 0.8758047930896282,
      "num_tokens": 72773300.0,
      "step": 35590
    },
    {
      "entropy": 0.48286981508135796,
      "epoch": 0.5781895844668393,
      "grad_norm": 6.78125,
      "learning_rate": 1.945413402779092e-05,
      "loss": 0.4850332736968994,
      "mean_token_accuracy": 0.8762808255851269,
      "num_tokens": 72795365.0,
      "step": 35600
    },
    {
      "entropy": 0.5551985192578286,
      "epoch": 0.5783519972714649,
      "grad_norm": 8.5,
      "learning_rate": 1.944149148334437e-05,
      "loss": 0.5783005714416504,
      "mean_token_accuracy": 0.8668156832456588,
      "num_tokens": 72817101.0,
      "step": 35610
    },
    {
      "entropy": 0.47192288199439647,
      "epoch": 0.5785144100760904,
      "grad_norm": 6.9375,
      "learning_rate": 1.9428850434149098e-05,
      "loss": 0.43670010566711426,
      "mean_token_accuracy": 0.884991142898798,
      "num_tokens": 72838431.0,
      "step": 35620
    },
    {
      "entropy": 0.5473644613754004,
      "epoch": 0.5786768228807159,
      "grad_norm": 7.46875,
      "learning_rate": 1.9416210883605573e-05,
      "loss": 0.5788946151733398,
      "mean_token_accuracy": 0.8434464104473591,
      "num_tokens": 72859123.0,
      "step": 35630
    },
    {
      "entropy": 0.478962611919269,
      "epoch": 0.5788392356853415,
      "grad_norm": 9.125,
      "learning_rate": 1.940357283511388e-05,
      "loss": 0.5026509761810303,
      "mean_token_accuracy": 0.8742302693426609,
      "num_tokens": 72878970.0,
      "step": 35640
    },
    {
      "entropy": 0.5686134153744206,
      "epoch": 0.5790016484899669,
      "grad_norm": 8.3125,
      "learning_rate": 1.939093629207367e-05,
      "loss": 0.6010343551635742,
      "mean_token_accuracy": 0.8534506179392338,
      "num_tokens": 72898490.0,
      "step": 35650
    },
    {
      "entropy": 0.4544618075247854,
      "epoch": 0.5791640612945924,
      "grad_norm": 5.5625,
      "learning_rate": 1.9378301257884195e-05,
      "loss": 0.445522403717041,
      "mean_token_accuracy": 0.8852240912616253,
      "num_tokens": 72918567.0,
      "step": 35660
    },
    {
      "entropy": 0.4762196332740132,
      "epoch": 0.579326474099218,
      "grad_norm": 7.03125,
      "learning_rate": 1.936566773594433e-05,
      "loss": 0.46373400688171384,
      "mean_token_accuracy": 0.8773585550487042,
      "num_tokens": 72939852.0,
      "step": 35670
    },
    {
      "entropy": 0.5091265562921763,
      "epoch": 0.5794888869038435,
      "grad_norm": 7.71875,
      "learning_rate": 1.935303572965249e-05,
      "loss": 0.4906982421875,
      "mean_token_accuracy": 0.8700317494571209,
      "num_tokens": 72959350.0,
      "step": 35680
    },
    {
      "entropy": 0.4609913203399628,
      "epoch": 0.5796512997084691,
      "grad_norm": 8.0625,
      "learning_rate": 1.9340405242406744e-05,
      "loss": 0.48447117805480955,
      "mean_token_accuracy": 0.8790274687111378,
      "num_tokens": 72979708.0,
      "step": 35690
    },
    {
      "entropy": 0.4752211211249232,
      "epoch": 0.5798137125130945,
      "grad_norm": 8.3125,
      "learning_rate": 1.9327776277604694e-05,
      "loss": 0.5047016143798828,
      "mean_token_accuracy": 0.8818192154169082,
      "num_tokens": 73001174.0,
      "step": 35700
    },
    {
      "entropy": 0.5462072098860518,
      "epoch": 0.57997612531772,
      "grad_norm": 7.21875,
      "learning_rate": 1.931514883864358e-05,
      "loss": 0.5003799438476563,
      "mean_token_accuracy": 0.8733808405697345,
      "num_tokens": 73022151.0,
      "step": 35710
    },
    {
      "entropy": 0.4847595132421702,
      "epoch": 0.5801385381223456,
      "grad_norm": 8.75,
      "learning_rate": 1.9302522928920207e-05,
      "loss": 0.519412612915039,
      "mean_token_accuracy": 0.8738117277622223,
      "num_tokens": 73041597.0,
      "step": 35720
    },
    {
      "entropy": 0.42400085178669544,
      "epoch": 0.5803009509269711,
      "grad_norm": 6.21875,
      "learning_rate": 1.9289898551830965e-05,
      "loss": 0.4307117938995361,
      "mean_token_accuracy": 0.8873087473213672,
      "num_tokens": 73061173.0,
      "step": 35730
    },
    {
      "entropy": 0.46986113467719404,
      "epoch": 0.5804633637315966,
      "grad_norm": 11.0,
      "learning_rate": 1.927727571077184e-05,
      "loss": 0.5222992897033691,
      "mean_token_accuracy": 0.8686073735356331,
      "num_tokens": 73080682.0,
      "step": 35740
    },
    {
      "entropy": 0.5160985419293865,
      "epoch": 0.5806257765362222,
      "grad_norm": 7.1875,
      "learning_rate": 1.926465440913842e-05,
      "loss": 0.5411452293395996,
      "mean_token_accuracy": 0.8689249783754349,
      "num_tokens": 73101638.0,
      "step": 35750
    },
    {
      "entropy": 0.48366700068581847,
      "epoch": 0.5807881893408476,
      "grad_norm": 7.84375,
      "learning_rate": 1.925203465032585e-05,
      "loss": 0.48059887886047364,
      "mean_token_accuracy": 0.8789863005280495,
      "num_tokens": 73122320.0,
      "step": 35760
    },
    {
      "entropy": 0.48181718395790085,
      "epoch": 0.5809506021454731,
      "grad_norm": 11.3125,
      "learning_rate": 1.923941643772888e-05,
      "loss": 0.49228410720825194,
      "mean_token_accuracy": 0.8737196993082762,
      "num_tokens": 73142969.0,
      "step": 35770
    },
    {
      "entropy": 0.47220885392744094,
      "epoch": 0.5811130149500987,
      "grad_norm": 11.3125,
      "learning_rate": 1.922679977474183e-05,
      "loss": 0.49103217124938964,
      "mean_token_accuracy": 0.8732566699385643,
      "num_tokens": 73163397.0,
      "step": 35780
    },
    {
      "entropy": 0.49429815537296234,
      "epoch": 0.5812754277547242,
      "grad_norm": 7.4375,
      "learning_rate": 1.9214184664758612e-05,
      "loss": 0.5304154872894287,
      "mean_token_accuracy": 0.8679283633828163,
      "num_tokens": 73183166.0,
      "step": 35790
    },
    {
      "entropy": 0.49542769910767676,
      "epoch": 0.5814378405593497,
      "grad_norm": 7.6875,
      "learning_rate": 1.9201571111172732e-05,
      "loss": 0.5472351551055908,
      "mean_token_accuracy": 0.8648921579122544,
      "num_tokens": 73202664.0,
      "step": 35800
    },
    {
      "entropy": 0.41655356702394786,
      "epoch": 0.5816002533639753,
      "grad_norm": 4.875,
      "learning_rate": 1.918895911737724e-05,
      "loss": 0.387218976020813,
      "mean_token_accuracy": 0.8966086618602276,
      "num_tokens": 73221834.0,
      "step": 35810
    },
    {
      "entropy": 0.5333805223926902,
      "epoch": 0.5817626661686007,
      "grad_norm": 4.84375,
      "learning_rate": 1.9176348686764816e-05,
      "loss": 0.5333885669708252,
      "mean_token_accuracy": 0.8728076040744781,
      "num_tokens": 73242720.0,
      "step": 35820
    },
    {
      "entropy": 0.47388181819114833,
      "epoch": 0.5819250789732262,
      "grad_norm": 7.65625,
      "learning_rate": 1.9163739822727683e-05,
      "loss": 0.5109090805053711,
      "mean_token_accuracy": 0.8860897071659565,
      "num_tokens": 73263361.0,
      "step": 35830
    },
    {
      "entropy": 0.5140248330309987,
      "epoch": 0.5820874917778518,
      "grad_norm": 9.0,
      "learning_rate": 1.9151132528657645e-05,
      "loss": 0.4885869026184082,
      "mean_token_accuracy": 0.8726772665977478,
      "num_tokens": 73283738.0,
      "step": 35840
    },
    {
      "entropy": 0.45320728844963015,
      "epoch": 0.5822499045824773,
      "grad_norm": 8.4375,
      "learning_rate": 1.9138526807946105e-05,
      "loss": 0.4760264873504639,
      "mean_token_accuracy": 0.8808666363358497,
      "num_tokens": 73304230.0,
      "step": 35850
    },
    {
      "entropy": 0.4778599556419067,
      "epoch": 0.5824123173871028,
      "grad_norm": 11.1875,
      "learning_rate": 1.9125922663984034e-05,
      "loss": 0.4544682502746582,
      "mean_token_accuracy": 0.8863369505852461,
      "num_tokens": 73324292.0,
      "step": 35860
    },
    {
      "entropy": 0.4511441068025306,
      "epoch": 0.5825747301917283,
      "grad_norm": 7.9375,
      "learning_rate": 1.9113320100161965e-05,
      "loss": 0.463943338394165,
      "mean_token_accuracy": 0.8822577059268951,
      "num_tokens": 73344905.0,
      "step": 35870
    },
    {
      "entropy": 0.5073189568473027,
      "epoch": 0.5827371429963538,
      "grad_norm": 7.875,
      "learning_rate": 1.910071911987002e-05,
      "loss": 0.5422058582305909,
      "mean_token_accuracy": 0.8703247912228107,
      "num_tokens": 73365568.0,
      "step": 35880
    },
    {
      "entropy": 0.5112473953515291,
      "epoch": 0.5828995558009793,
      "grad_norm": 7.875,
      "learning_rate": 1.9088119726497882e-05,
      "loss": 0.5209843635559082,
      "mean_token_accuracy": 0.8778479151427746,
      "num_tokens": 73386129.0,
      "step": 35890
    },
    {
      "entropy": 0.4623198303859681,
      "epoch": 0.5830619686056049,
      "grad_norm": 8.6875,
      "learning_rate": 1.9075521923434822e-05,
      "loss": 0.42987847328186035,
      "mean_token_accuracy": 0.8828278444707394,
      "num_tokens": 73405946.0,
      "step": 35900
    },
    {
      "entropy": 0.4787001106888056,
      "epoch": 0.5832243814102304,
      "grad_norm": 6.53125,
      "learning_rate": 1.9062925714069688e-05,
      "loss": 0.4607100486755371,
      "mean_token_accuracy": 0.8803922191262246,
      "num_tokens": 73425701.0,
      "step": 35910
    },
    {
      "entropy": 0.4909517400432378,
      "epoch": 0.583386794214856,
      "grad_norm": 5.90625,
      "learning_rate": 1.9050331101790874e-05,
      "loss": 0.49463305473327634,
      "mean_token_accuracy": 0.8764684133231639,
      "num_tokens": 73446062.0,
      "step": 35920
    },
    {
      "entropy": 0.4801846581976861,
      "epoch": 0.5835492070194814,
      "grad_norm": 13.875,
      "learning_rate": 1.9037738089986368e-05,
      "loss": 0.48386492729187014,
      "mean_token_accuracy": 0.8831513248383999,
      "num_tokens": 73465983.0,
      "step": 35930
    },
    {
      "entropy": 0.44464865741319953,
      "epoch": 0.5837116198241069,
      "grad_norm": 10.25,
      "learning_rate": 1.9025146682043714e-05,
      "loss": 0.4053050994873047,
      "mean_token_accuracy": 0.8936971254646778,
      "num_tokens": 73486804.0,
      "step": 35940
    },
    {
      "entropy": 0.4627448882907629,
      "epoch": 0.5838740326287325,
      "grad_norm": 6.0625,
      "learning_rate": 1.9012556881350028e-05,
      "loss": 0.4583892822265625,
      "mean_token_accuracy": 0.8845570560544729,
      "num_tokens": 73508053.0,
      "step": 35950
    },
    {
      "entropy": 0.43732062744093125,
      "epoch": 0.584036445433358,
      "grad_norm": 4.59375,
      "learning_rate": 1.8999968691292008e-05,
      "loss": 0.418790864944458,
      "mean_token_accuracy": 0.8850777491927146,
      "num_tokens": 73526769.0,
      "step": 35960
    },
    {
      "entropy": 0.5147913709515706,
      "epoch": 0.5841988582379835,
      "grad_norm": 9.8125,
      "learning_rate": 1.898738211525588e-05,
      "loss": 0.49486594200134276,
      "mean_token_accuracy": 0.8699352584779263,
      "num_tokens": 73548090.0,
      "step": 35970
    },
    {
      "entropy": 0.4678251658566296,
      "epoch": 0.584361271042609,
      "grad_norm": 6.15625,
      "learning_rate": 1.8974797156627476e-05,
      "loss": 0.4651028633117676,
      "mean_token_accuracy": 0.8825579956173897,
      "num_tokens": 73566792.0,
      "step": 35980
    },
    {
      "entropy": 0.5122425441164523,
      "epoch": 0.5845236838472345,
      "grad_norm": 7.71875,
      "learning_rate": 1.8962213818792186e-05,
      "loss": 0.5353062152862549,
      "mean_token_accuracy": 0.8677972353994846,
      "num_tokens": 73588499.0,
      "step": 35990
    },
    {
      "epoch": 0.58468609665186,
      "eval_entropy": 0.5086739993989468,
      "eval_loss": 0.5173910856246948,
      "eval_mean_token_accuracy": 0.8680103075504303,
      "eval_num_tokens": 73608229.0,
      "eval_runtime": 40.4342,
      "eval_samples_per_second": 49.463,
      "eval_steps_per_second": 6.183,
      "step": 36000
    }
  ],
  "logging_steps": 10,
  "max_steps": 61572,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 4000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.88763742720516e+18,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}