{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.0095,
  "eval_steps": 500,
  "global_step": 9500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "aux_loss": 8.092667961120606,
      "entropy": 0.2859670266509056,
      "epoch": 5e-06,
      "grad_norm": 2.4607067108154297,
      "learning_rate": 8.000000000000001e-07,
      "loss": 1.162985134124756,
      "mean_token_accuracy": 0.8323121547698975,
      "num_tokens": 81020.0,
      "step": 5
    },
    {
      "aux_loss": 8.10206651687622,
      "entropy": 0.3462228536605835,
      "epoch": 1e-05,
      "grad_norm": 2.7678725719451904,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 1.4827698707580566,
      "mean_token_accuracy": 0.7950647979974746,
      "num_tokens": 162205.0,
      "step": 10
    },
    {
      "aux_loss": 8.108795881271362,
      "entropy": 0.38629237115383147,
      "epoch": 1.5e-05,
      "grad_norm": 3.27398681640625,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 1.6987945556640625,
      "mean_token_accuracy": 0.7703906655311584,
      "num_tokens": 243180.0,
      "step": 15
    },
    {
      "aux_loss": 8.126577138900757,
      "entropy": 0.44169870764017105,
      "epoch": 2e-05,
      "grad_norm": 3.469611644744873,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 1.8818172454833983,
      "mean_token_accuracy": 0.745928606390953,
      "num_tokens": 323659.0,
      "step": 20
    },
    {
      "aux_loss": 8.139866304397582,
      "entropy": 0.4981244429945946,
      "epoch": 2.5e-05,
      "grad_norm": 3.9781763553619385,
      "learning_rate": 4.800000000000001e-06,
      "loss": 1.963961410522461,
      "mean_token_accuracy": 0.7302836388349533,
      "num_tokens": 404012.0,
      "step": 25
    },
    {
      "aux_loss": 8.1594473361969,
      "entropy": 0.612455278635025,
      "epoch": 3e-05,
      "grad_norm": 4.717221736907959,
      "learning_rate": 5.8e-06,
      "loss": 2.052127647399902,
      "mean_token_accuracy": 0.702911964058876,
      "num_tokens": 484061.0,
      "step": 30
    },
    {
      "aux_loss": 8.10626163482666,
      "entropy": 0.43849745094776155,
      "epoch": 3.5e-05,
      "grad_norm": 1.3155338764190674,
      "learning_rate": 6.800000000000001e-06,
      "loss": 1.0553879737854004,
      "mean_token_accuracy": 0.8176445186138153,
      "num_tokens": 561596.0,
      "step": 35
    },
    {
      "aux_loss": 8.10414137840271,
      "entropy": 0.5332785502076149,
      "epoch": 4e-05,
      "grad_norm": 1.514621615409851,
      "learning_rate": 7.800000000000002e-06,
      "loss": 1.0922231674194336,
      "mean_token_accuracy": 0.7994577556848526,
      "num_tokens": 643068.0,
      "step": 40
    },
    {
      "aux_loss": 8.110194063186645,
      "entropy": 0.6411030411720275,
      "epoch": 4.5e-05,
      "grad_norm": 1.3585166931152344,
      "learning_rate": 8.8e-06,
      "loss": 1.1537811279296875,
      "mean_token_accuracy": 0.7801551461219788,
      "num_tokens": 723967.0,
      "step": 45
    },
    {
      "aux_loss": 8.122820091247558,
      "entropy": 0.762116265296936,
      "epoch": 5e-05,
      "grad_norm": 1.0342539548873901,
      "learning_rate": 9.800000000000001e-06,
      "loss": 1.17335844039917,
      "mean_token_accuracy": 0.7652206391096115,
      "num_tokens": 804455.0,
      "step": 50
    },
    {
      "aux_loss": 8.138481664657593,
      "entropy": 0.8767565250396728,
      "epoch": 5.5e-05,
      "grad_norm": 0.8618419766426086,
      "learning_rate": 1.0800000000000002e-05,
      "loss": 1.1562211990356446,
      "mean_token_accuracy": 0.7581683337688446,
      "num_tokens": 885084.0,
      "step": 55
    },
    {
      "aux_loss": 8.161038303375244,
      "entropy": 0.9870363444089889,
      "epoch": 6e-05,
      "grad_norm": 0.9211753010749817,
      "learning_rate": 1.18e-05,
      "loss": 1.1537209510803224,
      "mean_token_accuracy": 0.7541291087865829,
      "num_tokens": 965552.0,
      "step": 60
    },
    {
      "aux_loss": 8.108438205718993,
      "entropy": 0.6005895540118218,
      "epoch": 6.5e-05,
      "grad_norm": 0.5573590397834778,
      "learning_rate": 1.2800000000000001e-05,
      "loss": 0.6857633113861084,
      "mean_token_accuracy": 0.8443360924720764,
      "num_tokens": 1045440.0,
      "step": 65
    },
    {
      "aux_loss": 8.097761487960815,
      "entropy": 0.5134793236851692,
      "epoch": 7e-05,
      "grad_norm": 0.8289216756820679,
      "learning_rate": 1.38e-05,
      "loss": 0.6059335708618164,
      "mean_token_accuracy": 0.8547728210687637,
      "num_tokens": 1126198.0,
      "step": 70
    },
    {
      "aux_loss": 8.109341192245484,
      "entropy": 0.49506455659866333,
      "epoch": 7.5e-05,
      "grad_norm": 0.7367376089096069,
      "learning_rate": 1.48e-05,
      "loss": 0.5785280227661133,
      "mean_token_accuracy": 0.8590773671865464,
      "num_tokens": 1206891.0,
      "step": 75
    },
    {
      "aux_loss": 8.126173639297486,
      "entropy": 0.47545491009950636,
      "epoch": 8e-05,
      "grad_norm": 0.5796106457710266,
      "learning_rate": 1.58e-05,
      "loss": 0.5134618282318115,
      "mean_token_accuracy": 0.8742193549871444,
      "num_tokens": 1287594.0,
      "step": 80
    },
    {
      "aux_loss": 8.144159412384033,
      "entropy": 0.4454115808010101,
      "epoch": 8.5e-05,
      "grad_norm": 0.37113332748413086,
      "learning_rate": 1.6800000000000002e-05,
      "loss": 0.47019548416137696,
      "mean_token_accuracy": 0.8867231249809265,
      "num_tokens": 1368588.0,
      "step": 85
    },
    {
      "aux_loss": 8.16485824584961,
      "entropy": 0.42590916603803636,
      "epoch": 9e-05,
      "grad_norm": 0.37275785207748413,
      "learning_rate": 1.7800000000000002e-05,
      "loss": 0.4420657157897949,
      "mean_token_accuracy": 0.8914682984352111,
      "num_tokens": 1448962.0,
      "step": 90
    },
    {
      "aux_loss": 8.138917446136475,
      "entropy": 0.32791209444403646,
      "epoch": 9.5e-05,
      "grad_norm": 0.181448832154274,
      "learning_rate": 1.88e-05,
      "loss": 0.34783947467803955,
      "mean_token_accuracy": 0.911352476477623,
      "num_tokens": 1527530.0,
      "step": 95
    },
    {
      "aux_loss": 8.104297828674316,
      "entropy": 0.2813743956387043,
      "epoch": 0.0001,
      "grad_norm": 0.22963790595531464,
      "learning_rate": 1.98e-05,
      "loss": 0.31741673946380616,
      "mean_token_accuracy": 0.9149413764476776,
      "num_tokens": 1608627.0,
      "step": 100
    },
    {
      "aux_loss": 8.11442060470581,
      "entropy": 0.2937369041144848,
      "epoch": 0.000105,
      "grad_norm": 0.24627141654491425,
      "learning_rate": 2e-05,
      "loss": 0.32099955081939696,
      "mean_token_accuracy": 0.9122195959091186,
      "num_tokens": 1689020.0,
      "step": 105
    },
    {
      "aux_loss": 8.135867309570312,
      "entropy": 0.2759621672332287,
      "epoch": 0.00011,
      "grad_norm": 0.2679576873779297,
      "learning_rate": 2e-05,
      "loss": 0.3021723508834839,
      "mean_token_accuracy": 0.9157016634941101,
      "num_tokens": 1769747.0,
      "step": 110
    },
    {
      "aux_loss": 8.152368021011352,
      "entropy": 0.2531339399516582,
      "epoch": 0.000115,
      "grad_norm": 0.14514628052711487,
      "learning_rate": 2e-05,
      "loss": 0.2765481948852539,
      "mean_token_accuracy": 0.9218116343021393,
      "num_tokens": 1850938.0,
      "step": 115
    },
    {
      "aux_loss": 8.1763521194458,
      "entropy": 0.2417703628540039,
      "epoch": 0.00012,
      "grad_norm": 0.13516604900360107,
      "learning_rate": 2e-05,
      "loss": 0.26600430011749265,
      "mean_token_accuracy": 0.9240559995174408,
      "num_tokens": 1931354.0,
      "step": 120
    },
    {
      "aux_loss": 8.152435255050658,
      "entropy": 0.22490062788128853,
      "epoch": 0.000125,
      "grad_norm": 0.13898135721683502,
      "learning_rate": 2e-05,
      "loss": 0.2723809242248535,
      "mean_token_accuracy": 0.9253725618124008,
      "num_tokens": 2011951.0,
      "step": 125
    },
    {
      "aux_loss": 8.103985023498534,
      "entropy": 0.21289734095335006,
      "epoch": 0.00013,
      "grad_norm": 0.11599434912204742,
      "learning_rate": 2e-05,
      "loss": 0.2714021444320679,
      "mean_token_accuracy": 0.925862830877304,
      "num_tokens": 2093277.0,
      "step": 130
    },
    {
      "aux_loss": 8.117699193954468,
      "entropy": 0.23202661350369452,
      "epoch": 0.000135,
      "grad_norm": 0.11791031807661057,
      "learning_rate": 2e-05,
      "loss": 0.2785206317901611,
      "mean_token_accuracy": 0.9215139657258987,
      "num_tokens": 2174577.0,
      "step": 135
    },
    {
      "aux_loss": 8.131106090545654,
      "entropy": 0.22132659405469896,
      "epoch": 0.00014,
      "grad_norm": 0.09535504132509232,
      "learning_rate": 2e-05,
      "loss": 0.2573246002197266,
      "mean_token_accuracy": 0.925444370508194,
      "num_tokens": 2255453.0,
      "step": 140
    },
    {
      "aux_loss": 8.153035163879395,
      "entropy": 0.2323688253760338,
      "epoch": 0.000145,
      "grad_norm": 0.11295939981937408,
      "learning_rate": 2e-05,
      "loss": 0.2621062517166138,
      "mean_token_accuracy": 0.9242130070924759,
      "num_tokens": 2335773.0,
      "step": 145
    },
    {
      "aux_loss": 8.169135093688965,
      "entropy": 0.23263223320245743,
      "epoch": 0.00015,
      "grad_norm": 0.10522045195102692,
      "learning_rate": 2e-05,
      "loss": 0.267002010345459,
      "mean_token_accuracy": 0.9240644991397857,
      "num_tokens": 2416318.0,
      "step": 150
    },
    {
      "aux_loss": 8.183323335647582,
      "entropy": 0.2133930116891861,
      "epoch": 0.000155,
      "grad_norm": 0.11165736615657806,
      "learning_rate": 2e-05,
      "loss": 0.24688491821289063,
      "mean_token_accuracy": 0.9294364064931869,
      "num_tokens": 2496506.0,
      "step": 155
    },
    {
      "aux_loss": 8.095394563674926,
      "entropy": 0.19811301529407502,
      "epoch": 0.00016,
      "grad_norm": 0.12067431211471558,
      "learning_rate": 2e-05,
      "loss": 0.2500493288040161,
      "mean_token_accuracy": 0.9307984799146652,
      "num_tokens": 2577798.0,
      "step": 160
    },
    {
      "aux_loss": 8.118247032165527,
      "entropy": 0.2177561827003956,
      "epoch": 0.000165,
      "grad_norm": 0.11767563968896866,
      "learning_rate": 2e-05,
      "loss": 0.2695132255554199,
      "mean_token_accuracy": 0.9233422338962555,
      "num_tokens": 2659301.0,
      "step": 165
    },
    {
      "aux_loss": 8.132621622085571,
      "entropy": 0.22528393045067788,
      "epoch": 0.00017,
      "grad_norm": 0.1076337918639183,
      "learning_rate": 2e-05,
      "loss": 0.2669530391693115,
      "mean_token_accuracy": 0.9225011795759201,
      "num_tokens": 2740256.0,
      "step": 170
    },
    {
      "aux_loss": 8.148080205917358,
      "entropy": 0.22090085670351983,
      "epoch": 0.000175,
      "grad_norm": 0.10490445792675018,
      "learning_rate": 2e-05,
      "loss": 0.2574422836303711,
      "mean_token_accuracy": 0.9257927656173706,
      "num_tokens": 2820851.0,
      "step": 175
    },
    {
      "aux_loss": 8.163924217224121,
      "entropy": 0.22146919295191764,
      "epoch": 0.00018,
      "grad_norm": 0.1082926020026207,
      "learning_rate": 2e-05,
      "loss": 0.250842022895813,
      "mean_token_accuracy": 0.9289938122034073,
      "num_tokens": 2901081.0,
      "step": 180
    },
    {
      "aux_loss": 8.191630840301514,
      "entropy": 0.21994876563549043,
      "epoch": 0.000185,
      "grad_norm": 0.13312949240207672,
      "learning_rate": 2e-05,
      "loss": 0.24324214458465576,
      "mean_token_accuracy": 0.9298162490129471,
      "num_tokens": 2981067.0,
      "step": 185
    },
    {
      "aux_loss": 8.111298513412475,
      "entropy": 0.19961925223469734,
      "epoch": 0.00019,
      "grad_norm": 0.14753255248069763,
      "learning_rate": 2e-05,
      "loss": 0.25237910747528075,
      "mean_token_accuracy": 0.9300692468881607,
      "num_tokens": 3059730.0,
      "step": 190
    },
    {
      "aux_loss": 8.112194967269897,
      "entropy": 0.21024976968765258,
      "epoch": 0.000195,
      "grad_norm": 0.10278994590044022,
      "learning_rate": 2e-05,
      "loss": 0.2596331357955933,
      "mean_token_accuracy": 0.9263428509235382,
      "num_tokens": 3140878.0,
      "step": 195
    },
    {
      "aux_loss": 8.126857614517212,
      "entropy": 0.2193007118999958,
      "epoch": 0.0002,
      "grad_norm": 0.11093388497829437,
      "learning_rate": 2e-05,
      "loss": 0.2619086742401123,
      "mean_token_accuracy": 0.9256764084100724,
      "num_tokens": 3221992.0,
      "step": 200
    },
    {
      "aux_loss": 8.144866943359375,
      "entropy": 0.21325586959719658,
      "epoch": 0.000205,
      "grad_norm": 0.10353685915470123,
      "learning_rate": 2e-05,
      "loss": 0.2519110441207886,
      "mean_token_accuracy": 0.9273806542158127,
      "num_tokens": 3302738.0,
      "step": 205
    },
    {
      "aux_loss": 8.165263128280639,
      "entropy": 0.21110628470778464,
      "epoch": 0.00021,
      "grad_norm": 0.09743088483810425,
      "learning_rate": 2e-05,
      "loss": 0.23808212280273439,
      "mean_token_accuracy": 0.9318860292434692,
      "num_tokens": 3383596.0,
      "step": 210
    },
    {
      "aux_loss": 8.185856103897095,
      "entropy": 0.21532292664051056,
      "epoch": 0.000215,
      "grad_norm": 0.11625949293375015,
      "learning_rate": 2e-05,
      "loss": 0.23955197334289552,
      "mean_token_accuracy": 0.9299207180738449,
      "num_tokens": 3463857.0,
      "step": 215
    },
    {
      "aux_loss": 8.122901678085327,
      "entropy": 0.19247540533542634,
      "epoch": 0.00022,
      "grad_norm": 0.12606287002563477,
      "learning_rate": 2e-05,
      "loss": 0.23701121807098388,
      "mean_token_accuracy": 0.9341312944889069,
      "num_tokens": 3544682.0,
      "step": 220
    },
    {
      "aux_loss": 8.113785791397095,
      "entropy": 0.20620187148451805,
      "epoch": 0.000225,
      "grad_norm": 0.10787980258464813,
      "learning_rate": 2e-05,
      "loss": 0.2567843198776245,
      "mean_token_accuracy": 0.9279820889234542,
      "num_tokens": 3625248.0,
      "step": 225
    },
    {
      "aux_loss": 8.12384238243103,
      "entropy": 0.21107216626405717,
      "epoch": 0.00023,
      "grad_norm": 0.10271225869655609,
      "learning_rate": 2e-05,
      "loss": 0.2494570255279541,
      "mean_token_accuracy": 0.9288909673690796,
      "num_tokens": 3706159.0,
      "step": 230
    },
    {
      "aux_loss": 8.142868852615356,
      "entropy": 0.21522793173789978,
      "epoch": 0.000235,
      "grad_norm": 0.10796686261892319,
      "learning_rate": 2e-05,
      "loss": 0.24782259464263917,
      "mean_token_accuracy": 0.9284009218215943,
      "num_tokens": 3786934.0,
      "step": 235
    },
    {
      "aux_loss": 8.161288928985595,
      "entropy": 0.21401349380612372,
      "epoch": 0.00024,
      "grad_norm": 0.10282823443412781,
      "learning_rate": 2e-05,
      "loss": 0.23868253231048583,
      "mean_token_accuracy": 0.9310090273618699,
      "num_tokens": 3867737.0,
      "step": 240
    },
    {
      "aux_loss": 8.182607841491699,
      "entropy": 0.20493827313184737,
      "epoch": 0.000245,
      "grad_norm": 0.11914316564798355,
      "learning_rate": 2e-05,
      "loss": 0.2305370330810547,
      "mean_token_accuracy": 0.933690196275711,
      "num_tokens": 3948051.0,
      "step": 245
    },
    {
      "aux_loss": 8.1415030002594,
      "entropy": 0.1825829602777958,
      "epoch": 0.00025,
      "grad_norm": 0.11836111545562744,
      "learning_rate": 2e-05,
      "loss": 0.22231316566467285,
      "mean_token_accuracy": 0.9372205168008805,
      "num_tokens": 4027476.0,
      "step": 250
    },
    {
      "aux_loss": 8.11691393852234,
      "entropy": 0.19727027863264085,
      "epoch": 0.000255,
      "grad_norm": 0.12442458420991898,
      "learning_rate": 2e-05,
      "loss": 0.2484445333480835,
      "mean_token_accuracy": 0.9305040508508682,
      "num_tokens": 4107946.0,
      "step": 255
    },
    {
      "aux_loss": 8.125603771209716,
      "entropy": 0.20766830518841745,
      "epoch": 0.00026,
      "grad_norm": 0.09718529134988785,
      "learning_rate": 2e-05,
      "loss": 0.25551691055297854,
      "mean_token_accuracy": 0.9271604478359222,
      "num_tokens": 4188909.0,
      "step": 260
    },
    {
      "aux_loss": 8.1395188331604,
      "entropy": 0.20430102646350862,
      "epoch": 0.000265,
      "grad_norm": 0.10315221548080444,
      "learning_rate": 2e-05,
      "loss": 0.23392271995544434,
      "mean_token_accuracy": 0.9329472333192825,
      "num_tokens": 4269781.0,
      "step": 265
    },
    {
      "aux_loss": 8.156323575973511,
      "entropy": 0.20779709219932557,
      "epoch": 0.00027,
      "grad_norm": 0.10915031284093857,
      "learning_rate": 2e-05,
      "loss": 0.23439457416534423,
      "mean_token_accuracy": 0.9330718845129014,
      "num_tokens": 4350638.0,
      "step": 270
    },
    {
      "aux_loss": 8.177580308914184,
      "entropy": 0.20390753895044328,
      "epoch": 0.000275,
      "grad_norm": 0.11818801611661911,
      "learning_rate": 2e-05,
      "loss": 0.22784197330474854,
      "mean_token_accuracy": 0.9339498370885849,
      "num_tokens": 4430597.0,
      "step": 275
    },
    {
      "aux_loss": 8.151418304443359,
      "entropy": 0.18322188295423986,
      "epoch": 0.00028,
      "grad_norm": 0.12752026319503784,
      "learning_rate": 2e-05,
      "loss": 0.2191540241241455,
      "mean_token_accuracy": 0.9375271767377853,
      "num_tokens": 4509115.0,
      "step": 280
    },
    {
      "aux_loss": 8.10854730606079,
      "entropy": 0.2067515507340431,
      "epoch": 0.000285,
      "grad_norm": 0.1464259922504425,
      "learning_rate": 2e-05,
      "loss": 0.26910512447357177,
      "mean_token_accuracy": 0.9251548409461975,
      "num_tokens": 4589888.0,
      "step": 285
    },
    {
      "aux_loss": 8.123592424392701,
      "entropy": 0.21190579310059549,
      "epoch": 0.00029,
      "grad_norm": 0.12418748438358307,
      "learning_rate": 2e-05,
      "loss": 0.24806389808654786,
      "mean_token_accuracy": 0.9289951413869858,
      "num_tokens": 4670831.0,
      "step": 290
    },
    {
      "aux_loss": 8.134210300445556,
      "entropy": 0.2096714936196804,
      "epoch": 0.000295,
      "grad_norm": 0.09857596457004547,
      "learning_rate": 2e-05,
      "loss": 0.24017581939697266,
      "mean_token_accuracy": 0.931833091378212,
      "num_tokens": 4751506.0,
      "step": 295
    },
    {
      "aux_loss": 8.1527024269104,
      "entropy": 0.2133574441075325,
      "epoch": 0.0003,
      "grad_norm": 0.12382720410823822,
      "learning_rate": 2e-05,
      "loss": 0.23982751369476318,
      "mean_token_accuracy": 0.9317940264940262,
      "num_tokens": 4832077.0,
      "step": 300
    },
    {
      "aux_loss": 8.172873163223267,
      "entropy": 0.2010139726102352,
      "epoch": 0.000305,
      "grad_norm": 0.1136835366487503,
      "learning_rate": 2e-05,
      "loss": 0.2233577013015747,
      "mean_token_accuracy": 0.9361098736524582,
      "num_tokens": 4913061.0,
      "step": 305
    },
    {
      "aux_loss": 8.17594952583313,
      "entropy": 0.1964607909321785,
      "epoch": 0.00031,
      "grad_norm": 0.14149561524391174,
      "learning_rate": 2e-05,
      "loss": 0.2333085536956787,
      "mean_token_accuracy": 0.9334494590759277,
      "num_tokens": 4989644.0,
      "step": 310
    },
    {
      "aux_loss": 8.103254461288453,
      "entropy": 0.1957407720386982,
      "epoch": 0.000315,
      "grad_norm": 0.13898971676826477,
      "learning_rate": 2e-05,
      "loss": 0.2496751070022583,
      "mean_token_accuracy": 0.9303697794675827,
      "num_tokens": 5070985.0,
      "step": 315
    },
    {
      "aux_loss": 8.121169662475586,
      "entropy": 0.21436288058757783,
      "epoch": 0.00032,
      "grad_norm": 0.10650767385959625,
      "learning_rate": 2e-05,
      "loss": 0.2578300476074219,
      "mean_token_accuracy": 0.9276322662830353,
      "num_tokens": 5151799.0,
      "step": 320
    },
    {
      "aux_loss": 8.133362054824829,
      "entropy": 0.20707877054810525,
      "epoch": 0.000325,
      "grad_norm": 0.10141773521900177,
      "learning_rate": 2e-05,
      "loss": 0.2367786169052124,
      "mean_token_accuracy": 0.9322249799966812,
      "num_tokens": 5232787.0,
      "step": 325
    },
    {
      "aux_loss": 8.152495861053467,
      "entropy": 0.20323231145739556,
      "epoch": 0.00033,
      "grad_norm": 0.1185888797044754,
      "learning_rate": 2e-05,
      "loss": 0.231177020072937,
      "mean_token_accuracy": 0.9329827636480331,
      "num_tokens": 5313384.0,
      "step": 330
    },
    {
      "aux_loss": 8.174307060241699,
      "entropy": 0.19869768768548965,
      "epoch": 0.000335,
      "grad_norm": 0.11668054759502411,
      "learning_rate": 2e-05,
      "loss": 0.2284449815750122,
      "mean_token_accuracy": 0.934393760561943,
      "num_tokens": 5394275.0,
      "step": 335
    },
    {
      "aux_loss": 8.171902513504028,
      "entropy": 0.19377838149666787,
      "epoch": 0.00034,
      "grad_norm": 0.12018025666475296,
      "learning_rate": 2e-05,
      "loss": 0.22451364994049072,
      "mean_token_accuracy": 0.9363054871559143,
      "num_tokens": 5472436.0,
      "step": 340
    },
    {
      "aux_loss": 8.103836250305175,
      "entropy": 0.2022476501762867,
      "epoch": 0.000345,
      "grad_norm": 0.12494606524705887,
      "learning_rate": 2e-05,
      "loss": 0.2549619436264038,
      "mean_token_accuracy": 0.9293957650661469,
      "num_tokens": 5553488.0,
      "step": 345
    },
    {
      "aux_loss": 8.11767029762268,
      "entropy": 0.20732319429516793,
      "epoch": 0.00035,
      "grad_norm": 0.10775358974933624,
      "learning_rate": 2e-05,
      "loss": 0.2460080623626709,
      "mean_token_accuracy": 0.9307799786329269,
      "num_tokens": 5634296.0,
      "step": 350
    },
    {
      "aux_loss": 8.13014645576477,
      "entropy": 0.20114240497350694,
      "epoch": 0.000355,
      "grad_norm": 0.10871438682079315,
      "learning_rate": 2e-05,
      "loss": 0.23498148918151857,
      "mean_token_accuracy": 0.9330670058727264,
      "num_tokens": 5714763.0,
      "step": 355
    },
    {
      "aux_loss": 8.151902627944946,
      "entropy": 0.19216700047254562,
      "epoch": 0.00036,
      "grad_norm": 0.1080172136425972,
      "learning_rate": 2e-05,
      "loss": 0.21803610324859618,
      "mean_token_accuracy": 0.9378397405147553,
      "num_tokens": 5795149.0,
      "step": 360
    },
    {
      "aux_loss": 8.171459436416626,
      "entropy": 0.20071458742022513,
      "epoch": 0.000365,
      "grad_norm": 0.11842406541109085,
      "learning_rate": 2e-05,
      "loss": 0.23490922451019286,
      "mean_token_accuracy": 0.9331449836492538,
      "num_tokens": 5875713.0,
      "step": 365
    },
    {
      "aux_loss": 8.188628768920898,
      "entropy": 0.18819867298007012,
      "epoch": 0.00037,
      "grad_norm": 0.11708404123783112,
      "learning_rate": 2e-05,
      "loss": 0.21983647346496582,
      "mean_token_accuracy": 0.9381567299365997,
      "num_tokens": 5955091.0,
      "step": 370
    },
    {
      "aux_loss": 8.103630065917969,
      "entropy": 0.17806926667690276,
      "epoch": 0.000375,
      "grad_norm": 0.11385245621204376,
      "learning_rate": 2e-05,
      "loss": 0.2229466438293457,
      "mean_token_accuracy": 0.9370234042406083,
      "num_tokens": 6036204.0,
      "step": 375
    },
    {
      "aux_loss": 8.118957471847533,
      "entropy": 0.19929006695747375,
      "epoch": 0.00038,
      "grad_norm": 0.11459151655435562,
      "learning_rate": 2e-05,
      "loss": 0.24367763996124267,
      "mean_token_accuracy": 0.9314134418964386,
      "num_tokens": 6117453.0,
      "step": 380
    },
    {
      "aux_loss": 8.131189393997193,
      "entropy": 0.19944887161254882,
      "epoch": 0.000385,
      "grad_norm": 0.11293925344944,
      "learning_rate": 2e-05,
      "loss": 0.23520164489746093,
      "mean_token_accuracy": 0.932736599445343,
      "num_tokens": 6198381.0,
      "step": 385
    },
    {
      "aux_loss": 8.145823335647583,
      "entropy": 0.20357693061232568,
      "epoch": 0.00039,
      "grad_norm": 0.1089092567563057,
      "learning_rate": 2e-05,
      "loss": 0.22967128753662108,
      "mean_token_accuracy": 0.9340643346309662,
      "num_tokens": 6279079.0,
      "step": 390
    },
    {
      "aux_loss": 8.167986011505127,
      "entropy": 0.1950391486287117,
      "epoch": 0.000395,
      "grad_norm": 0.11145143955945969,
      "learning_rate": 2e-05,
      "loss": 0.22251276969909667,
      "mean_token_accuracy": 0.9362853616476059,
      "num_tokens": 6359105.0,
      "step": 395
    },
    {
      "aux_loss": 8.193504858016968,
      "entropy": 0.19332419708371162,
      "epoch": 0.0004,
      "grad_norm": 0.1461859792470932,
      "learning_rate": 2e-05,
      "loss": 0.22106502056121827,
      "mean_token_accuracy": 0.9367363721132278,
      "num_tokens": 6439228.0,
      "step": 400
    },
    {
      "aux_loss": 8.104427671432495,
      "entropy": 0.18344127163290977,
      "epoch": 0.000405,
      "grad_norm": 0.1249011904001236,
      "learning_rate": 2e-05,
      "loss": 0.22748680114746095,
      "mean_token_accuracy": 0.9364313662052155,
      "num_tokens": 6519620.0,
      "step": 405
    },
    {
      "aux_loss": 8.115176773071289,
      "entropy": 0.1931718721985817,
      "epoch": 0.00041,
      "grad_norm": 0.11378726363182068,
      "learning_rate": 2e-05,
      "loss": 0.23756885528564453,
      "mean_token_accuracy": 0.9335996091365815,
      "num_tokens": 6600777.0,
      "step": 410
    },
    {
      "aux_loss": 8.128407907485961,
      "entropy": 0.2001208856701851,
      "epoch": 0.000415,
      "grad_norm": 0.11233960092067719,
      "learning_rate": 2e-05,
      "loss": 0.2344135284423828,
      "mean_token_accuracy": 0.9312853068113327,
      "num_tokens": 6681796.0,
      "step": 415
    },
    {
      "aux_loss": 8.14580340385437,
      "entropy": 0.19529841765761374,
      "epoch": 0.00042,
      "grad_norm": 0.12032654881477356,
      "learning_rate": 2e-05,
      "loss": 0.22535357475280762,
      "mean_token_accuracy": 0.9345941811800003,
      "num_tokens": 6762546.0,
      "step": 420
    },
    {
      "aux_loss": 8.167212009429932,
      "entropy": 0.20289452597498894,
      "epoch": 0.000425,
      "grad_norm": 0.11672218888998032,
      "learning_rate": 2e-05,
      "loss": 0.22811779975891114,
      "mean_token_accuracy": 0.9348306357860565,
      "num_tokens": 6843084.0,
      "step": 425
    },
    {
      "aux_loss": 8.18767786026001,
      "entropy": 0.18519736900925637,
      "epoch": 0.00043,
      "grad_norm": 0.12846212089061737,
      "learning_rate": 2e-05,
      "loss": 0.21719591617584227,
      "mean_token_accuracy": 0.9371245980262757,
      "num_tokens": 6923337.0,
      "step": 430
    },
    {
      "aux_loss": 8.120493125915527,
      "entropy": 0.1844749353826046,
      "epoch": 0.000435,
      "grad_norm": 0.14547668397426605,
      "learning_rate": 2e-05,
      "loss": 0.23188719749450684,
      "mean_token_accuracy": 0.9346409678459168,
      "num_tokens": 7002600.0,
      "step": 435
    },
    {
      "aux_loss": 8.116820049285888,
      "entropy": 0.20214365869760514,
      "epoch": 0.00044,
      "grad_norm": 0.1273147016763687,
      "learning_rate": 2e-05,
      "loss": 0.24130103588104249,
      "mean_token_accuracy": 0.9309691965579987,
      "num_tokens": 7083762.0,
      "step": 440
    },
    {
      "aux_loss": 8.130047225952149,
      "entropy": 0.21163361221551896,
      "epoch": 0.000445,
      "grad_norm": 0.11172276735305786,
      "learning_rate": 2e-05,
      "loss": 0.24803059101104735,
      "mean_token_accuracy": 0.9303495198488235,
      "num_tokens": 7164904.0,
      "step": 445
    },
    {
      "aux_loss": 8.146463441848756,
      "entropy": 0.20141847357153891,
      "epoch": 0.00045,
      "grad_norm": 0.11067729443311691,
      "learning_rate": 2e-05,
      "loss": 0.2322296380996704,
      "mean_token_accuracy": 0.934284684062004,
      "num_tokens": 7245505.0,
      "step": 450
    },
    {
      "aux_loss": 8.159644889831544,
      "entropy": 0.19601927921175957,
      "epoch": 0.000455,
      "grad_norm": 0.11938250809907913,
      "learning_rate": 2e-05,
      "loss": 0.22233147621154786,
      "mean_token_accuracy": 0.9349738001823426,
      "num_tokens": 7325901.0,
      "step": 455
    },
    {
      "aux_loss": 8.181401491165161,
      "entropy": 0.188271477073431,
      "epoch": 0.00046,
      "grad_norm": 0.13487257063388824,
      "learning_rate": 2e-05,
      "loss": 0.213454270362854,
      "mean_token_accuracy": 0.9380605906248093,
      "num_tokens": 7405893.0,
      "step": 460
    },
    {
      "aux_loss": 8.13626651763916,
      "entropy": 0.17855463474988936,
      "epoch": 0.000465,
      "grad_norm": 0.14394786953926086,
      "learning_rate": 2e-05,
      "loss": 0.22088232040405273,
      "mean_token_accuracy": 0.9383244127035141,
      "num_tokens": 7484471.0,
      "step": 465
    },
    {
      "aux_loss": 8.11563949584961,
      "entropy": 0.19367720708251,
      "epoch": 0.00047,
      "grad_norm": 0.12225597351789474,
      "learning_rate": 2e-05,
      "loss": 0.2356712579727173,
      "mean_token_accuracy": 0.9334611773490906,
      "num_tokens": 7565565.0,
      "step": 470
    },
    {
      "aux_loss": 8.126933383941651,
      "entropy": 0.19567672535777092,
      "epoch": 0.000475,
      "grad_norm": 0.10960134863853455,
      "learning_rate": 2e-05,
      "loss": 0.2261427640914917,
      "mean_token_accuracy": 0.9353327989578247,
      "num_tokens": 7646830.0,
      "step": 475
    },
    {
      "aux_loss": 8.141116952896118,
      "entropy": 0.19941989183425904,
      "epoch": 0.00048,
      "grad_norm": 0.1227649673819542,
      "learning_rate": 2e-05,
      "loss": 0.228026819229126,
      "mean_token_accuracy": 0.935494351387024,
      "num_tokens": 7727875.0,
      "step": 480
    },
    {
      "aux_loss": 8.162264680862426,
      "entropy": 0.19571123644709587,
      "epoch": 0.000485,
      "grad_norm": 0.12896135449409485,
      "learning_rate": 2e-05,
      "loss": 0.229170560836792,
      "mean_token_accuracy": 0.9355918109416962,
      "num_tokens": 7808288.0,
      "step": 485
    },
    {
      "aux_loss": 8.186082220077514,
      "entropy": 0.18303756937384605,
      "epoch": 0.00049,
      "grad_norm": 0.12394686043262482,
      "learning_rate": 2e-05,
      "loss": 0.20925889015197754,
      "mean_token_accuracy": 0.9398917376995086,
      "num_tokens": 7888831.0,
      "step": 490
    },
    {
      "aux_loss": 8.130529880523682,
      "entropy": 0.17214761935174466,
      "epoch": 0.000495,
      "grad_norm": 0.12712115049362183,
      "learning_rate": 2e-05,
      "loss": 0.21500439643859864,
      "mean_token_accuracy": 0.9404178261756897,
      "num_tokens": 7967750.0,
      "step": 495
    },
    {
      "aux_loss": 8.108060312271117,
      "entropy": 0.21055722162127494,
      "epoch": 0.0005,
      "grad_norm": 0.13972988724708557,
      "learning_rate": 2e-05,
      "loss": 0.2587005615234375,
      "mean_token_accuracy": 0.9271103173494339,
      "num_tokens": 8049056.0,
      "step": 500
    },
    {
      "aux_loss": 8.123171758651733,
      "entropy": 0.2044980011880398,
      "epoch": 0.000505,
      "grad_norm": 0.11161034554243088,
      "learning_rate": 2e-05,
      "loss": 0.23147244453430177,
      "mean_token_accuracy": 0.9337287068367004,
      "num_tokens": 8130294.0,
      "step": 505
    },
    {
      "aux_loss": 8.136773443222046,
      "entropy": 0.19301708787679672,
      "epoch": 0.00051,
      "grad_norm": 0.11636194586753845,
      "learning_rate": 2e-05,
      "loss": 0.2207319736480713,
      "mean_token_accuracy": 0.9375490337610245,
      "num_tokens": 8210854.0,
      "step": 510
    },
    {
      "aux_loss": 8.149327182769776,
      "entropy": 0.18693748190999032,
      "epoch": 0.000515,
      "grad_norm": 0.13859185576438904,
      "learning_rate": 2e-05,
      "loss": 0.22074565887451172,
      "mean_token_accuracy": 0.9367436707019806,
      "num_tokens": 8291598.0,
      "step": 515
    },
    {
      "aux_loss": 8.175821590423585,
      "entropy": 0.1823076017200947,
      "epoch": 0.00052,
      "grad_norm": 0.14429710805416107,
      "learning_rate": 2e-05,
      "loss": 0.21230535507202147,
      "mean_token_accuracy": 0.9388055384159089,
      "num_tokens": 8371590.0,
      "step": 520
    },
    {
      "aux_loss": 8.148676204681397,
      "entropy": 0.18088727369904517,
      "epoch": 0.000525,
      "grad_norm": 0.1233874037861824,
      "learning_rate": 2e-05,
      "loss": 0.2183612823486328,
      "mean_token_accuracy": 0.9389139652252197,
      "num_tokens": 8448788.0,
      "step": 525
    },
    {
      "aux_loss": 8.11525845527649,
      "entropy": 0.1864571325480938,
      "epoch": 0.00053,
      "grad_norm": 0.12095758318901062,
      "learning_rate": 2e-05,
      "loss": 0.22606940269470216,
      "mean_token_accuracy": 0.9359039962291718,
      "num_tokens": 8530095.0,
      "step": 530
    },
    {
      "aux_loss": 8.119484186172485,
      "entropy": 0.2005111776292324,
      "epoch": 0.000535,
      "grad_norm": 0.12065824866294861,
      "learning_rate": 2e-05,
      "loss": 0.23620870113372802,
      "mean_token_accuracy": 0.9338431239128113,
      "num_tokens": 8610731.0,
      "step": 535
    },
    {
      "aux_loss": 8.133606958389283,
      "entropy": 0.19584577158093452,
      "epoch": 0.00054,
      "grad_norm": 0.12134315818548203,
      "learning_rate": 2e-05,
      "loss": 0.22811131477355956,
      "mean_token_accuracy": 0.93468978703022,
      "num_tokens": 8691103.0,
      "step": 540
    },
    {
      "aux_loss": 8.153080463409424,
      "entropy": 0.19136801436543466,
      "epoch": 0.000545,
      "grad_norm": 0.13523323833942413,
      "learning_rate": 2e-05,
      "loss": 0.22138473987579346,
      "mean_token_accuracy": 0.9364001035690308,
      "num_tokens": 8771839.0,
      "step": 545
    },
    {
      "aux_loss": 8.168652248382568,
      "entropy": 0.1879940316081047,
      "epoch": 0.00055,
      "grad_norm": 0.13683131337165833,
      "learning_rate": 2e-05,
      "loss": 0.21518070697784425,
      "mean_token_accuracy": 0.9380461663007736,
      "num_tokens": 8852541.0,
      "step": 550
    },
    {
      "aux_loss": 8.163324642181397,
      "entropy": 0.1806576546281576,
      "epoch": 0.000555,
      "grad_norm": 0.12295546382665634,
      "learning_rate": 2e-05,
      "loss": 0.21502649784088135,
      "mean_token_accuracy": 0.9389211475849152,
      "num_tokens": 8930934.0,
      "step": 555
    },
    {
      "aux_loss": 8.106051921844482,
      "entropy": 0.17880526185035706,
      "epoch": 0.00056,
      "grad_norm": 0.1354513019323349,
      "learning_rate": 2e-05,
      "loss": 0.2253559112548828,
      "mean_token_accuracy": 0.9372707158327103,
      "num_tokens": 9012314.0,
      "step": 560
    },
    {
      "aux_loss": 8.11518898010254,
      "entropy": 0.20097925364971161,
      "epoch": 0.000565,
      "grad_norm": 0.1232772096991539,
      "learning_rate": 2e-05,
      "loss": 0.24141240119934082,
      "mean_token_accuracy": 0.9333308070898056,
      "num_tokens": 9093472.0,
      "step": 565
    },
    {
      "aux_loss": 8.129575252532959,
      "entropy": 0.2135834276676178,
      "epoch": 0.00057,
      "grad_norm": 0.12575826048851013,
      "learning_rate": 2e-05,
      "loss": 0.2407468318939209,
      "mean_token_accuracy": 0.9321556568145752,
      "num_tokens": 9174479.0,
      "step": 570
    },
    {
      "aux_loss": 8.141605377197266,
      "entropy": 0.1949714571237564,
      "epoch": 0.000575,
      "grad_norm": 0.12399072200059891,
      "learning_rate": 2e-05,
      "loss": 0.22630493640899657,
      "mean_token_accuracy": 0.9361388981342316,
      "num_tokens": 9255274.0,
      "step": 575
    },
    {
      "aux_loss": 8.161956882476806,
      "entropy": 0.18306322023272514,
      "epoch": 0.00058,
      "grad_norm": 0.1358226090669632,
      "learning_rate": 2e-05,
      "loss": 0.21267693042755126,
      "mean_token_accuracy": 0.9395399630069733,
      "num_tokens": 9336087.0,
      "step": 580
    },
    {
      "aux_loss": 8.178706502914428,
      "entropy": 0.17525342367589475,
      "epoch": 0.000585,
      "grad_norm": 0.13497048616409302,
      "learning_rate": 2e-05,
      "loss": 0.20523464679718018,
      "mean_token_accuracy": 0.9408984541893005,
      "num_tokens": 9415415.0,
      "step": 585
    },
    {
      "aux_loss": 8.097288703918457,
      "entropy": 0.17932368852198124,
      "epoch": 0.00059,
      "grad_norm": 0.13442210853099823,
      "learning_rate": 2e-05,
      "loss": 0.22426512241363525,
      "mean_token_accuracy": 0.9379236102104187,
      "num_tokens": 9496545.0,
      "step": 590
    },
    {
      "aux_loss": 8.115752935409546,
      "entropy": 0.1957818239927292,
      "epoch": 0.000595,
      "grad_norm": 0.1318245828151703,
      "learning_rate": 2e-05,
      "loss": 0.23539261817932128,
      "mean_token_accuracy": 0.9342283546924591,
      "num_tokens": 9577562.0,
      "step": 595
    },
    {
      "aux_loss": 8.121268844604492,
      "entropy": 0.20243508517742156,
      "epoch": 0.0006,
      "grad_norm": 0.12045804411172867,
      "learning_rate": 2e-05,
      "loss": 0.2283019781112671,
      "mean_token_accuracy": 0.9349256962537765,
      "num_tokens": 9658735.0,
      "step": 600
    },
    {
      "aux_loss": 8.138788747787476,
      "entropy": 0.19625149294734,
      "epoch": 0.000605,
      "grad_norm": 0.1283811777830124,
      "learning_rate": 2e-05,
      "loss": 0.22196848392486573,
      "mean_token_accuracy": 0.936600586771965,
      "num_tokens": 9739325.0,
      "step": 605
    },
    {
      "aux_loss": 8.153925466537476,
      "entropy": 0.1859537072479725,
      "epoch": 0.00061,
      "grad_norm": 0.1319660097360611,
      "learning_rate": 2e-05,
      "loss": 0.21823840141296386,
      "mean_token_accuracy": 0.9378139019012451,
      "num_tokens": 9819316.0,
      "step": 610
    },
    {
      "aux_loss": 8.181534004211425,
      "entropy": 0.1830529935657978,
      "epoch": 0.000615,
      "grad_norm": 0.1498909741640091,
      "learning_rate": 2e-05,
      "loss": 0.21233899593353273,
      "mean_token_accuracy": 0.9398808598518371,
      "num_tokens": 9899847.0,
      "step": 615
    },
    {
      "aux_loss": 8.115496301651001,
      "entropy": 0.1680178303271532,
      "epoch": 0.00062,
      "grad_norm": 0.13132670521736145,
      "learning_rate": 2e-05,
      "loss": 0.2065460205078125,
      "mean_token_accuracy": 0.9419705748558045,
      "num_tokens": 9979438.0,
      "step": 620
    },
    {
      "aux_loss": 8.112633323669433,
      "entropy": 0.19637811034917832,
      "epoch": 0.000625,
      "grad_norm": 0.11578872054815292,
      "learning_rate": 2e-05,
      "loss": 0.23919968605041503,
      "mean_token_accuracy": 0.9333801180124283,
      "num_tokens": 10060958.0,
      "step": 625
    },
    {
      "aux_loss": 8.12286148071289,
      "entropy": 0.20346488654613495,
      "epoch": 0.00063,
      "grad_norm": 0.11164384335279465,
      "learning_rate": 2e-05,
      "loss": 0.23783347606658936,
      "mean_token_accuracy": 0.9328191936016083,
      "num_tokens": 10141992.0,
      "step": 630
    },
    {
      "aux_loss": 8.13811559677124,
      "entropy": 0.19954773262143136,
      "epoch": 0.000635,
      "grad_norm": 0.12549450993537903,
      "learning_rate": 2e-05,
      "loss": 0.22470993995666505,
      "mean_token_accuracy": 0.9348821938037872,
      "num_tokens": 10222969.0,
      "step": 635
    },
    {
      "aux_loss": 8.154943752288819,
      "entropy": 0.1856667123734951,
      "epoch": 0.00064,
      "grad_norm": 0.1315392553806305,
      "learning_rate": 2e-05,
      "loss": 0.21411411762237548,
      "mean_token_accuracy": 0.9383790761232376,
      "num_tokens": 10303283.0,
      "step": 640
    },
    {
      "aux_loss": 8.175086259841919,
      "entropy": 0.18237411677837373,
      "epoch": 0.000645,
      "grad_norm": 0.13341780006885529,
      "learning_rate": 2e-05,
      "loss": 0.21301708221435547,
      "mean_token_accuracy": 0.9392763644456863,
      "num_tokens": 10383851.0,
      "step": 645
    },
    {
      "aux_loss": 8.12028365135193,
      "entropy": 0.17168023064732552,
      "epoch": 0.00065,
      "grad_norm": 0.13317014276981354,
      "learning_rate": 2e-05,
      "loss": 0.21729826927185059,
      "mean_token_accuracy": 0.9406512379646301,
      "num_tokens": 10461204.0,
      "step": 650
    },
    {
      "aux_loss": 8.111561059951782,
      "entropy": 0.2008303001523018,
      "epoch": 0.000655,
      "grad_norm": 0.13740909099578857,
      "learning_rate": 2e-05,
      "loss": 0.24498250484466552,
      "mean_token_accuracy": 0.9311482429504394,
      "num_tokens": 10542394.0,
      "step": 655
    },
    {
      "aux_loss": 8.121409940719605,
      "entropy": 0.2036907948553562,
      "epoch": 0.00066,
      "grad_norm": 0.11630047112703323,
      "learning_rate": 2e-05,
      "loss": 0.22815239429473877,
      "mean_token_accuracy": 0.9339557975530625,
      "num_tokens": 10623242.0,
      "step": 660
    },
    {
      "aux_loss": 8.135656070709228,
      "entropy": 0.1957025058567524,
      "epoch": 0.000665,
      "grad_norm": 0.11823663860559464,
      "learning_rate": 2e-05,
      "loss": 0.22408857345581054,
      "mean_token_accuracy": 0.9361691862344742,
      "num_tokens": 10703802.0,
      "step": 665
    },
    {
      "aux_loss": 8.155420827865601,
      "entropy": 0.19255246967077255,
      "epoch": 0.00067,
      "grad_norm": 0.1313610076904297,
      "learning_rate": 2e-05,
      "loss": 0.2256295919418335,
      "mean_token_accuracy": 0.9346400886774063,
      "num_tokens": 10784336.0,
      "step": 670
    },
    {
      "aux_loss": 8.172291088104249,
      "entropy": 0.1758721239864826,
      "epoch": 0.000675,
      "grad_norm": 0.12752971053123474,
      "learning_rate": 2e-05,
      "loss": 0.20256171226501465,
      "mean_token_accuracy": 0.9408331990242005,
      "num_tokens": 10864297.0,
      "step": 675
    },
    {
      "aux_loss": 8.12908205986023,
      "entropy": 0.1669745821505785,
      "epoch": 0.00068,
      "grad_norm": 0.1285611391067505,
      "learning_rate": 2e-05,
      "loss": 0.20708107948303223,
      "mean_token_accuracy": 0.9414781242609024,
      "num_tokens": 10944480.0,
      "step": 680
    },
    {
      "aux_loss": 8.110326957702636,
      "entropy": 0.19593036100268363,
      "epoch": 0.000685,
      "grad_norm": 0.1288689821958542,
      "learning_rate": 2e-05,
      "loss": 0.2419433116912842,
      "mean_token_accuracy": 0.9323927283287048,
      "num_tokens": 11025645.0,
      "step": 685
    },
    {
      "aux_loss": 8.118507814407348,
      "entropy": 0.20737494602799417,
      "epoch": 0.00069,
      "grad_norm": 0.12270528078079224,
      "learning_rate": 2e-05,
      "loss": 0.24025404453277588,
      "mean_token_accuracy": 0.9311066091060638,
      "num_tokens": 11106719.0,
      "step": 690
    },
    {
      "aux_loss": 8.131222677230834,
      "entropy": 0.19384507685899735,
      "epoch": 0.000695,
      "grad_norm": 0.11885134130716324,
      "learning_rate": 2e-05,
      "loss": 0.22150125503540039,
      "mean_token_accuracy": 0.9368506163358689,
      "num_tokens": 11187483.0,
      "step": 695
    },
    {
      "aux_loss": 8.152690219879151,
      "entropy": 0.1862498939037323,
      "epoch": 0.0007,
      "grad_norm": 0.13085342943668365,
      "learning_rate": 2e-05,
      "loss": 0.21512579917907715,
      "mean_token_accuracy": 0.9381736785173416,
      "num_tokens": 11267923.0,
      "step": 700
    },
    {
      "aux_loss": 8.175735139846802,
      "entropy": 0.18144834637641907,
      "epoch": 0.000705,
      "grad_norm": 0.157309427857399,
      "learning_rate": 2e-05,
      "loss": 0.21283621788024903,
      "mean_token_accuracy": 0.9390768080949783,
      "num_tokens": 11348330.0,
      "step": 705
    },
    {
      "aux_loss": 8.131483697891236,
      "entropy": 0.169484831020236,
      "epoch": 0.00071,
      "grad_norm": 0.12376774102449417,
      "learning_rate": 2e-05,
      "loss": 0.20716946125030516,
      "mean_token_accuracy": 0.942305651307106,
      "num_tokens": 11427342.0,
      "step": 710
    },
    {
      "aux_loss": 8.102163362503052,
      "entropy": 0.20284102559089662,
      "epoch": 0.000715,
      "grad_norm": 0.1368631273508072,
      "learning_rate": 2e-05,
      "loss": 0.2438335657119751,
      "mean_token_accuracy": 0.9310838371515274,
      "num_tokens": 11508525.0,
      "step": 715
    },
    {
      "aux_loss": 8.116524744033814,
      "entropy": 0.1933087445795536,
      "epoch": 0.00072,
      "grad_norm": 0.11646907031536102,
      "learning_rate": 2e-05,
      "loss": 0.22562990188598633,
      "mean_token_accuracy": 0.9347972393035888,
      "num_tokens": 11589651.0,
      "step": 720
    },
    {
      "aux_loss": 8.130318403244019,
      "entropy": 0.20118043795228005,
      "epoch": 0.000725,
      "grad_norm": 0.12211769074201584,
      "learning_rate": 2e-05,
      "loss": 0.22835359573364258,
      "mean_token_accuracy": 0.9346433848142623,
      "num_tokens": 11670537.0,
      "step": 725
    },
    {
      "aux_loss": 8.145296144485474,
      "entropy": 0.18327361345291138,
      "epoch": 0.00073,
      "grad_norm": 0.14909130334854126,
      "learning_rate": 2e-05,
      "loss": 0.22180075645446778,
      "mean_token_accuracy": 0.9366574764251709,
      "num_tokens": 11751080.0,
      "step": 730
    },
    {
      "aux_loss": 8.16052017211914,
      "entropy": 0.17084143832325935,
      "epoch": 0.000735,
      "grad_norm": 0.14019134640693665,
      "learning_rate": 2e-05,
      "loss": 0.2031708002090454,
      "mean_token_accuracy": 0.9429173946380616,
      "num_tokens": 11831584.0,
      "step": 735
    },
    {
      "aux_loss": 8.145300006866455,
      "entropy": 0.16553889848291875,
      "epoch": 0.00074,
      "grad_norm": 0.12312433123588562,
      "learning_rate": 2e-05,
      "loss": 0.1950891613960266,
      "mean_token_accuracy": 0.945001271367073,
      "num_tokens": 11912161.0,
      "step": 740
    },
    {
      "aux_loss": 8.100783348083496,
      "entropy": 0.18602003231644632,
      "epoch": 0.000745,
      "grad_norm": 0.16615012288093567,
      "learning_rate": 2e-05,
      "loss": 0.23647782802581788,
      "mean_token_accuracy": 0.9346523225307465,
      "num_tokens": 11993125.0,
      "step": 745
    },
    {
      "aux_loss": 8.119870710372926,
      "entropy": 0.20997865498065948,
      "epoch": 0.00075,
      "grad_norm": 0.12530550360679626,
      "learning_rate": 2e-05,
      "loss": 0.2479936122894287,
      "mean_token_accuracy": 0.9292480409145355,
      "num_tokens": 12074082.0,
      "step": 750
    },
    {
      "aux_loss": 8.132441186904908,
      "entropy": 0.20384614467620848,
      "epoch": 0.000755,
      "grad_norm": 0.14014708995819092,
      "learning_rate": 2e-05,
      "loss": 0.2232731342315674,
      "mean_token_accuracy": 0.9353633552789689,
      "num_tokens": 12154977.0,
      "step": 755
    },
    {
      "aux_loss": 8.147329187393188,
      "entropy": 0.1960611179471016,
      "epoch": 0.00076,
      "grad_norm": 0.141123428940773,
      "learning_rate": 2e-05,
      "loss": 0.2233285427093506,
      "mean_token_accuracy": 0.9361371099948883,
      "num_tokens": 12235552.0,
      "step": 760
    },
    {
      "aux_loss": 8.164418029785157,
      "entropy": 0.177785038203001,
      "epoch": 0.000765,
      "grad_norm": 0.13308079540729523,
      "learning_rate": 2e-05,
      "loss": 0.21046712398529052,
      "mean_token_accuracy": 0.9409039944410325,
      "num_tokens": 12316296.0,
      "step": 765
    },
    {
      "aux_loss": 8.148503684997559,
      "entropy": 0.17363472767174243,
      "epoch": 0.00077,
      "grad_norm": 0.14745160937309265,
      "learning_rate": 2e-05,
      "loss": 0.20947306156158446,
      "mean_token_accuracy": 0.9408982396125793,
      "num_tokens": 12396200.0,
      "step": 770
    },
    {
      "aux_loss": 8.100437784194947,
      "entropy": 0.18707914352416993,
      "epoch": 0.000775,
      "grad_norm": 0.13240523636341095,
      "learning_rate": 2e-05,
      "loss": 0.23113799095153809,
      "mean_token_accuracy": 0.9352542102336884,
      "num_tokens": 12477581.0,
      "step": 775
    },
    {
      "aux_loss": 8.109345197677612,
      "entropy": 0.20186578631401061,
      "epoch": 0.00078,
      "grad_norm": 0.1274145543575287,
      "learning_rate": 2e-05,
      "loss": 0.238570499420166,
      "mean_token_accuracy": 0.9327393889427185,
      "num_tokens": 12558633.0,
      "step": 780
    },
    {
      "aux_loss": 8.122262763977051,
      "entropy": 0.19397404193878173,
      "epoch": 0.000785,
      "grad_norm": 0.1275504231452942,
      "learning_rate": 2e-05,
      "loss": 0.21904168128967286,
      "mean_token_accuracy": 0.9374355524778366,
      "num_tokens": 12639315.0,
      "step": 785
    },
    {
      "aux_loss": 8.143395376205444,
      "entropy": 0.1881797805428505,
      "epoch": 0.00079,
      "grad_norm": 0.133524551987648,
      "learning_rate": 2e-05,
      "loss": 0.2175664186477661,
      "mean_token_accuracy": 0.9382498949766159,
      "num_tokens": 12719815.0,
      "step": 790
    },
    {
      "aux_loss": 8.156324195861817,
      "entropy": 0.17355761900544167,
      "epoch": 0.000795,
      "grad_norm": 0.13200172781944275,
      "learning_rate": 2e-05,
      "loss": 0.20816025733947754,
      "mean_token_accuracy": 0.9410201340913773,
      "num_tokens": 12800294.0,
      "step": 795
    },
    {
      "aux_loss": 8.171008968353272,
      "entropy": 0.17404237911105155,
      "epoch": 0.0008,
      "grad_norm": 0.14056964218616486,
      "learning_rate": 2e-05,
      "loss": 0.20798285007476808,
      "mean_token_accuracy": 0.9413327813148499,
      "num_tokens": 12878773.0,
      "step": 800
    },
    {
      "aux_loss": 8.093837356567382,
      "entropy": 0.1741633154451847,
      "epoch": 0.000805,
      "grad_norm": 0.13568736612796783,
      "learning_rate": 2e-05,
      "loss": 0.2179255723953247,
      "mean_token_accuracy": 0.9385777384042739,
      "num_tokens": 12959682.0,
      "step": 805
    },
    {
      "aux_loss": 8.111591148376466,
      "entropy": 0.19361605867743492,
      "epoch": 0.00081,
      "grad_norm": 0.12338358908891678,
      "learning_rate": 2e-05,
      "loss": 0.22942121028900148,
      "mean_token_accuracy": 0.9353196829557419,
      "num_tokens": 13041028.0,
      "step": 810
    },
    {
      "aux_loss": 8.12508053779602,
      "entropy": 0.19188614636659623,
      "epoch": 0.000815,
      "grad_norm": 0.12431506812572479,
      "learning_rate": 2e-05,
      "loss": 0.21545052528381348,
      "mean_token_accuracy": 0.9393882095813751,
      "num_tokens": 13121701.0,
      "step": 815
    },
    {
      "aux_loss": 8.13652834892273,
      "entropy": 0.19292113482952117,
      "epoch": 0.00082,
      "grad_norm": 0.1260223686695099,
      "learning_rate": 2e-05,
      "loss": 0.22584521770477295,
      "mean_token_accuracy": 0.9350881546735763,
      "num_tokens": 13202292.0,
      "step": 820
    },
    {
      "aux_loss": 8.149944734573364,
      "entropy": 0.17534302845597266,
      "epoch": 0.000825,
      "grad_norm": 0.13426516950130463,
      "learning_rate": 2e-05,
      "loss": 0.20456910133361816,
      "mean_token_accuracy": 0.941878753900528,
      "num_tokens": 13282210.0,
      "step": 825
    },
    {
      "aux_loss": 8.178964471817016,
      "entropy": 0.1746487282216549,
      "epoch": 0.00083,
      "grad_norm": 0.15093551576137543,
      "learning_rate": 2e-05,
      "loss": 0.20533223152160646,
      "mean_token_accuracy": 0.9409088015556335,
      "num_tokens": 13362632.0,
      "step": 830
    },
    {
      "aux_loss": 8.11063618659973,
      "entropy": 0.18316879943013192,
      "epoch": 0.000835,
      "grad_norm": 0.14530125260353088,
      "learning_rate": 2e-05,
      "loss": 0.22833695411682128,
      "mean_token_accuracy": 0.9373126685619354,
      "num_tokens": 13439754.0,
      "step": 835
    },
    {
      "aux_loss": 8.109144926071167,
      "entropy": 0.19589049592614174,
      "epoch": 0.00084,
      "grad_norm": 0.1471167951822281,
      "learning_rate": 2e-05,
      "loss": 0.2323617696762085,
      "mean_token_accuracy": 0.935021448135376,
      "num_tokens": 13521162.0,
      "step": 840
    },
    {
      "aux_loss": 8.11714415550232,
      "entropy": 0.20246537178754806,
      "epoch": 0.000845,
      "grad_norm": 0.12362229079008102,
      "learning_rate": 2e-05,
      "loss": 0.23810834884643556,
      "mean_token_accuracy": 0.9331830322742463,
      "num_tokens": 13601917.0,
      "step": 845
    },
    {
      "aux_loss": 8.133970737457275,
      "entropy": 0.1931244671344757,
      "epoch": 0.00085,
      "grad_norm": 0.13273629546165466,
      "learning_rate": 2e-05,
      "loss": 0.22032694816589354,
      "mean_token_accuracy": 0.936900520324707,
      "num_tokens": 13682685.0,
      "step": 850
    },
    {
      "aux_loss": 8.143550634384155,
      "entropy": 0.18127841055393218,
      "epoch": 0.000855,
      "grad_norm": 0.13045166432857513,
      "learning_rate": 2e-05,
      "loss": 0.2100224494934082,
      "mean_token_accuracy": 0.9403670132160187,
      "num_tokens": 13763113.0,
      "step": 855
    },
    {
      "aux_loss": 8.167861986160279,
      "entropy": 0.1719556473195553,
      "epoch": 0.00086,
      "grad_norm": 0.1533770114183426,
      "learning_rate": 2e-05,
      "loss": 0.2010368824005127,
      "mean_token_accuracy": 0.9427653759717941,
      "num_tokens": 13843381.0,
      "step": 860
    },
    {
      "aux_loss": 8.11334171295166,
      "entropy": 0.17256785109639167,
      "epoch": 0.000865,
      "grad_norm": 0.1296316236257553,
      "learning_rate": 2e-05,
      "loss": 0.22134318351745605,
      "mean_token_accuracy": 0.939113312959671,
      "num_tokens": 13922305.0,
      "step": 865
    },
    {
      "aux_loss": 8.108716487884521,
      "entropy": 0.1922435976564884,
      "epoch": 0.00087,
      "grad_norm": 0.1125635877251625,
      "learning_rate": 2e-05,
      "loss": 0.2310950994491577,
      "mean_token_accuracy": 0.9351043969392776,
      "num_tokens": 14003490.0,
      "step": 870
    },
    {
      "aux_loss": 8.115301895141602,
      "entropy": 0.2005027063190937,
      "epoch": 0.000875,
      "grad_norm": 0.12468750029802322,
      "learning_rate": 2e-05,
      "loss": 0.22837119102478026,
      "mean_token_accuracy": 0.9355538696050644,
      "num_tokens": 14084716.0,
      "step": 875
    },
    {
      "aux_loss": 8.126211452484132,
      "entropy": 0.19354628697037696,
      "epoch": 0.00088,
      "grad_norm": 0.13211360573768616,
      "learning_rate": 2e-05,
      "loss": 0.2266221523284912,
      "mean_token_accuracy": 0.9371717393398284,
      "num_tokens": 14165484.0,
      "step": 880
    },
    {
      "aux_loss": 8.14353904724121,
      "entropy": 0.18160487785935403,
      "epoch": 0.000885,
      "grad_norm": 0.14947129786014557,
      "learning_rate": 2e-05,
      "loss": 0.2137932538986206,
      "mean_token_accuracy": 0.9408360600471497,
      "num_tokens": 14245987.0,
      "step": 885
    },
    {
      "aux_loss": 8.163965654373168,
      "entropy": 0.17387786507606506,
      "epoch": 0.00089,
      "grad_norm": 0.15340450406074524,
      "learning_rate": 2e-05,
      "loss": 0.2016824245452881,
      "mean_token_accuracy": 0.9424707919359208,
      "num_tokens": 14326178.0,
      "step": 890
    },
    {
      "aux_loss": 8.119355916976929,
      "entropy": 0.17598350942134858,
      "epoch": 0.000895,
      "grad_norm": 0.14417360723018646,
      "learning_rate": 2e-05,
      "loss": 0.21732549667358397,
      "mean_token_accuracy": 0.9407132625579834,
      "num_tokens": 14405313.0,
      "step": 895
    },
    {
      "aux_loss": 8.109018659591674,
      "entropy": 0.20914908573031427,
      "epoch": 0.0009,
      "grad_norm": 0.12596318125724792,
      "learning_rate": 2e-05,
      "loss": 0.24976136684417724,
      "mean_token_accuracy": 0.9293664872646332,
      "num_tokens": 14486566.0,
      "step": 900
    },
    {
      "aux_loss": 8.115027236938477,
      "entropy": 0.20074048563838004,
      "epoch": 0.000905,
      "grad_norm": 0.12749867141246796,
      "learning_rate": 2e-05,
      "loss": 0.2284632921218872,
      "mean_token_accuracy": 0.9368485689163208,
      "num_tokens": 14567772.0,
      "step": 905
    },
    {
      "aux_loss": 8.127698040008545,
      "entropy": 0.19091494828462602,
      "epoch": 0.00091,
      "grad_norm": 0.1260538548231125,
      "learning_rate": 2e-05,
      "loss": 0.22049989700317382,
      "mean_token_accuracy": 0.9376890927553176,
      "num_tokens": 14648419.0,
      "step": 910
    },
    {
      "aux_loss": 8.13921823501587,
      "entropy": 0.1774464137852192,
      "epoch": 0.000915,
      "grad_norm": 0.12893904745578766,
      "learning_rate": 2e-05,
      "loss": 0.20832493305206298,
      "mean_token_accuracy": 0.9396044939756394,
      "num_tokens": 14729366.0,
      "step": 915
    },
    {
      "aux_loss": 8.16345157623291,
      "entropy": 0.17128808721899985,
      "epoch": 0.00092,
      "grad_norm": 0.14869266748428345,
      "learning_rate": 2e-05,
      "loss": 0.20440850257873536,
      "mean_token_accuracy": 0.9419215887784957,
      "num_tokens": 14809681.0,
      "step": 920
    },
    {
      "aux_loss": 8.12254056930542,
      "entropy": 0.17211619690060614,
      "epoch": 0.000925,
      "grad_norm": 0.14314009249210358,
      "learning_rate": 2e-05,
      "loss": 0.20708854198455812,
      "mean_token_accuracy": 0.943778908252716,
      "num_tokens": 14888981.0,
      "step": 925
    },
    {
      "aux_loss": 8.102337169647218,
      "entropy": 0.1863011859357357,
      "epoch": 0.00093,
      "grad_norm": 0.12896594405174255,
      "learning_rate": 2e-05,
      "loss": 0.22866301536560057,
      "mean_token_accuracy": 0.9369279652833938,
      "num_tokens": 14970153.0,
      "step": 930
    },
    {
      "aux_loss": 8.112103414535522,
      "entropy": 0.20329497009515762,
      "epoch": 0.000935,
      "grad_norm": 0.13223311305046082,
      "learning_rate": 2e-05,
      "loss": 0.23256449699401854,
      "mean_token_accuracy": 0.934928897023201,
      "num_tokens": 15051345.0,
      "step": 935
    },
    {
      "aux_loss": 8.12607283592224,
      "entropy": 0.18796944245696068,
      "epoch": 0.00094,
      "grad_norm": 0.14386539161205292,
      "learning_rate": 2e-05,
      "loss": 0.21558585166931152,
      "mean_token_accuracy": 0.9385993003845214,
      "num_tokens": 15131948.0,
      "step": 940
    },
    {
      "aux_loss": 8.143296909332275,
      "entropy": 0.18020025864243508,
      "epoch": 0.000945,
      "grad_norm": 0.1396578699350357,
      "learning_rate": 2e-05,
      "loss": 0.21119444370269774,
      "mean_token_accuracy": 0.9392522275447845,
      "num_tokens": 15211936.0,
      "step": 945
    },
    {
      "aux_loss": 8.155585670471192,
      "entropy": 0.17179183289408684,
      "epoch": 0.00095,
      "grad_norm": 0.1429106742143631,
      "learning_rate": 2e-05,
      "loss": 0.20406196117401124,
      "mean_token_accuracy": 0.9423564583063125,
      "num_tokens": 15291992.0,
      "step": 950
    },
    {
      "aux_loss": 8.129588842391968,
      "entropy": 0.17628065794706343,
      "epoch": 0.000955,
      "grad_norm": 0.13858522474765778,
      "learning_rate": 2e-05,
      "loss": 0.21669340133666992,
      "mean_token_accuracy": 0.9396115422248841,
      "num_tokens": 15372441.0,
      "step": 955
    },
    {
      "aux_loss": 8.100811052322388,
      "entropy": 0.1907297946512699,
      "epoch": 0.00096,
      "grad_norm": 0.14532077312469482,
      "learning_rate": 2e-05,
      "loss": 0.2299985408782959,
      "mean_token_accuracy": 0.9367765963077546,
      "num_tokens": 15453322.0,
      "step": 960
    },
    {
      "aux_loss": 8.109273386001586,
      "entropy": 0.20441971346735954,
      "epoch": 0.000965,
      "grad_norm": 0.12614189088344574,
      "learning_rate": 2e-05,
      "loss": 0.23587870597839355,
      "mean_token_accuracy": 0.9339224755764007,
      "num_tokens": 15534402.0,
      "step": 965
    },
    {
      "aux_loss": 8.121965646743774,
      "entropy": 0.1921239286661148,
      "epoch": 0.00097,
      "grad_norm": 0.14098156988620758,
      "learning_rate": 2e-05,
      "loss": 0.21978702545166015,
      "mean_token_accuracy": 0.937722310423851,
      "num_tokens": 15614958.0,
      "step": 970
    },
    {
      "aux_loss": 8.135463666915893,
      "entropy": 0.181341952085495,
      "epoch": 0.000975,
      "grad_norm": 0.13644331693649292,
      "learning_rate": 2e-05,
      "loss": 0.21015706062316894,
      "mean_token_accuracy": 0.9399736613035202,
      "num_tokens": 15695868.0,
      "step": 975
    },
    {
      "aux_loss": 8.154515409469605,
      "entropy": 0.17385695725679398,
      "epoch": 0.00098,
      "grad_norm": 0.1496136635541916,
      "learning_rate": 2e-05,
      "loss": 0.20413517951965332,
      "mean_token_accuracy": 0.9421140640974045,
      "num_tokens": 15776187.0,
      "step": 980
    },
    {
      "aux_loss": 8.142510080337525,
      "entropy": 0.1882740218192339,
      "epoch": 0.000985,
      "grad_norm": 0.13687245547771454,
      "learning_rate": 2e-05,
      "loss": 0.2333993673324585,
      "mean_token_accuracy": 0.9334321290254592,
      "num_tokens": 15854768.0,
      "step": 985
    },
    {
      "aux_loss": 8.09783320426941,
      "entropy": 0.181090035289526,
      "epoch": 0.00099,
      "grad_norm": 0.1412193477153778,
      "learning_rate": 2e-05,
      "loss": 0.2248934507369995,
      "mean_token_accuracy": 0.9371724635362625,
      "num_tokens": 15936013.0,
      "step": 990
    },
    {
      "aux_loss": 8.108890724182128,
      "entropy": 0.1935890458524227,
      "epoch": 0.000995,
      "grad_norm": 0.1459343284368515,
      "learning_rate": 2e-05,
      "loss": 0.21834125518798828,
      "mean_token_accuracy": 0.9386760383844376,
      "num_tokens": 16016975.0,
      "step": 995
    },
    {
      "aux_loss": 8.11813826560974,
      "entropy": 0.18227510675787925,
      "epoch": 0.001,
      "grad_norm": 0.13530665636062622,
      "learning_rate": 2e-05,
      "loss": 0.2158998966217041,
      "mean_token_accuracy": 0.9394854247570038,
      "num_tokens": 16097623.0,
      "step": 1000
    },
    {
      "aux_loss": 8.132675504684448,
      "entropy": 0.18303612917661666,
      "epoch": 0.001005,
      "grad_norm": 0.1295279860496521,
      "learning_rate": 2e-05,
      "loss": 0.21372718811035157,
      "mean_token_accuracy": 0.9388361543416976,
      "num_tokens": 16178571.0,
      "step": 1005
    },
    {
      "aux_loss": 8.148015975952148,
      "entropy": 0.1663380205631256,
      "epoch": 0.00101,
      "grad_norm": 0.1477927565574646,
      "learning_rate": 2e-05,
      "loss": 0.19427905082702637,
      "mean_token_accuracy": 0.944655567407608,
      "num_tokens": 16259014.0,
      "step": 1010
    },
    {
      "aux_loss": 8.15143895149231,
      "entropy": 0.1644660137593746,
      "epoch": 0.001015,
      "grad_norm": 0.14161117374897003,
      "learning_rate": 2e-05,
      "loss": 0.19924207925796508,
      "mean_token_accuracy": 0.9437342047691345,
      "num_tokens": 16335868.0,
      "step": 1015
    },
    {
      "aux_loss": 8.0966055393219,
      "entropy": 0.18626073226332665,
      "epoch": 0.00102,
      "grad_norm": 0.14405575394630432,
      "learning_rate": 2e-05,
      "loss": 0.2318559169769287,
      "mean_token_accuracy": 0.9360893279314041,
      "num_tokens": 16416964.0,
      "step": 1020
    },
    {
      "aux_loss": 8.109320878982544,
      "entropy": 0.20917314514517785,
      "epoch": 0.001025,
      "grad_norm": 0.13582730293273926,
      "learning_rate": 2e-05,
      "loss": 0.24025216102600097,
      "mean_token_accuracy": 0.9321926653385162,
      "num_tokens": 16497864.0,
      "step": 1025
    },
    {
      "aux_loss": 8.120928478240966,
      "entropy": 0.19562898352742195,
      "epoch": 0.00103,
      "grad_norm": 0.14005805552005768,
      "learning_rate": 2e-05,
      "loss": 0.22124676704406737,
      "mean_token_accuracy": 0.9378227382898331,
      "num_tokens": 16578957.0,
      "step": 1030
    },
    {
      "aux_loss": 8.131659936904907,
      "entropy": 0.1797639399766922,
      "epoch": 0.001035,
      "grad_norm": 0.1292354166507721,
      "learning_rate": 2e-05,
      "loss": 0.20929946899414062,
      "mean_token_accuracy": 0.9413349151611328,
      "num_tokens": 16659600.0,
      "step": 1035
    },
    {
      "aux_loss": 8.150179290771485,
      "entropy": 0.17218540385365486,
      "epoch": 0.00104,
      "grad_norm": 0.14203843474388123,
      "learning_rate": 2e-05,
      "loss": 0.2047133207321167,
      "mean_token_accuracy": 0.9421777725219727,
      "num_tokens": 16740165.0,
      "step": 1040
    },
    {
      "aux_loss": 8.14911766052246,
      "entropy": 0.17258057817816735,
      "epoch": 0.001045,
      "grad_norm": 0.15955905616283417,
      "learning_rate": 2e-05,
      "loss": 0.20898404121398925,
      "mean_token_accuracy": 0.9419188648462296,
      "num_tokens": 16818770.0,
      "step": 1045
    },
    {
      "aux_loss": 8.100459671020507,
      "entropy": 0.18287332057952882,
      "epoch": 0.00105,
      "grad_norm": 0.13039886951446533,
      "learning_rate": 2e-05,
      "loss": 0.2245183229446411,
      "mean_token_accuracy": 0.9368477880954742,
      "num_tokens": 16899620.0,
      "step": 1050
    },
    {
      "aux_loss": 8.104825401306153,
      "entropy": 0.20207613930106164,
      "epoch": 0.001055,
      "grad_norm": 0.1333446055650711,
      "learning_rate": 2e-05,
      "loss": 0.2343914747238159,
      "mean_token_accuracy": 0.9339895784854889,
      "num_tokens": 16980428.0,
      "step": 1055
    },
    {
      "aux_loss": 8.121085977554321,
      "entropy": 0.18368163779377938,
      "epoch": 0.00106,
      "grad_norm": 0.14122816920280457,
      "learning_rate": 2e-05,
      "loss": 0.20634570121765136,
      "mean_token_accuracy": 0.9415034621953964,
      "num_tokens": 17061163.0,
      "step": 1060
    },
    {
      "aux_loss": 8.13550944328308,
      "entropy": 0.17058610990643502,
      "epoch": 0.001065,
      "grad_norm": 0.13418757915496826,
      "learning_rate": 2e-05,
      "loss": 0.20201709270477294,
      "mean_token_accuracy": 0.942056930065155,
      "num_tokens": 17142015.0,
      "step": 1065
    },
    {
      "aux_loss": 8.153172969818115,
      "entropy": 0.1640140824019909,
      "epoch": 0.00107,
      "grad_norm": 0.1374797224998474,
      "learning_rate": 2e-05,
      "loss": 0.1974026918411255,
      "mean_token_accuracy": 0.9445619136095047,
      "num_tokens": 17222876.0,
      "step": 1070
    },
    {
      "aux_loss": 8.147079563140869,
      "entropy": 0.16375502794981003,
      "epoch": 0.001075,
      "grad_norm": 0.149629145860672,
      "learning_rate": 2e-05,
      "loss": 0.19806833267211915,
      "mean_token_accuracy": 0.9455973088741303,
      "num_tokens": 17299616.0,
      "step": 1075
    },
    {
      "aux_loss": 8.09938554763794,
      "entropy": 0.18405816331505775,
      "epoch": 0.00108,
      "grad_norm": 0.1336614340543747,
      "learning_rate": 2e-05,
      "loss": 0.22501940727233888,
      "mean_token_accuracy": 0.9368825078010559,
      "num_tokens": 17381036.0,
      "step": 1080
    },
    {
      "aux_loss": 8.105483961105346,
      "entropy": 0.20805612877011298,
      "epoch": 0.001085,
      "grad_norm": 0.14559486508369446,
      "learning_rate": 2e-05,
      "loss": 0.23807644844055176,
      "mean_token_accuracy": 0.932142361998558,
      "num_tokens": 17461835.0,
      "step": 1085
    },
    {
      "aux_loss": 8.114073085784913,
      "entropy": 0.19431799054145812,
      "epoch": 0.00109,
      "grad_norm": 0.1483253389596939,
      "learning_rate": 2e-05,
      "loss": 0.224235463142395,
      "mean_token_accuracy": 0.9369283556938172,
      "num_tokens": 17542394.0,
      "step": 1090
    },
    {
      "aux_loss": 8.13317642211914,
      "entropy": 0.18543633222579955,
      "epoch": 0.001095,
      "grad_norm": 0.13417592644691467,
      "learning_rate": 2e-05,
      "loss": 0.21352875232696533,
      "mean_token_accuracy": 0.9396424561738967,
      "num_tokens": 17623111.0,
      "step": 1095
    },
    {
      "aux_loss": 8.143168830871582,
      "entropy": 0.16634368896484375,
      "epoch": 0.0011,
      "grad_norm": 0.15244615077972412,
      "learning_rate": 2e-05,
      "loss": 0.19832861423492432,
      "mean_token_accuracy": 0.9432316929101944,
      "num_tokens": 17703783.0,
      "step": 1100
    },
    {
      "aux_loss": 8.154546213150024,
      "entropy": 0.1635368935763836,
      "epoch": 0.001105,
      "grad_norm": 0.12932510673999786,
      "learning_rate": 2e-05,
      "loss": 0.19644242525100708,
      "mean_token_accuracy": 0.9460396349430085,
      "num_tokens": 17780357.0,
      "step": 1105
    },
    {
      "aux_loss": 8.091581392288209,
      "entropy": 0.16594813540577888,
      "epoch": 0.00111,
      "grad_norm": 0.1556437462568283,
      "learning_rate": 2e-05,
      "loss": 0.2127126932144165,
      "mean_token_accuracy": 0.9413120537996292,
      "num_tokens": 17861703.0,
      "step": 1110
    },
    {
      "aux_loss": 8.103971719741821,
      "entropy": 0.18995240107178687,
      "epoch": 0.001115,
      "grad_norm": 0.13781099021434784,
      "learning_rate": 2e-05,
      "loss": 0.22258846759796141,
      "mean_token_accuracy": 0.9372970789670945,
      "num_tokens": 17942701.0,
      "step": 1115
    },
    {
      "aux_loss": 8.116778898239136,
      "entropy": 0.18525892719626427,
      "epoch": 0.00112,
      "grad_norm": 0.13170120120048523,
      "learning_rate": 2e-05,
      "loss": 0.21453588008880614,
      "mean_token_accuracy": 0.9394149154424667,
      "num_tokens": 18023228.0,
      "step": 1120
    },
    {
      "aux_loss": 8.129170846939086,
      "entropy": 0.18442671671509742,
      "epoch": 0.001125,
      "grad_norm": 0.13744796812534332,
      "learning_rate": 2e-05,
      "loss": 0.2131723165512085,
      "mean_token_accuracy": 0.9389948576688767,
      "num_tokens": 18104109.0,
      "step": 1125
    },
    {
      "aux_loss": 8.143761157989502,
      "entropy": 0.17146908715367318,
      "epoch": 0.00113,
      "grad_norm": 0.1306254118680954,
      "learning_rate": 2e-05,
      "loss": 0.20081534385681152,
      "mean_token_accuracy": 0.9426989138126374,
      "num_tokens": 18185131.0,
      "step": 1130
    },
    {
      "aux_loss": 8.155526971817016,
      "entropy": 0.17061771005392073,
      "epoch": 0.001135,
      "grad_norm": 0.15940193831920624,
      "learning_rate": 2e-05,
      "loss": 0.20533814430236816,
      "mean_token_accuracy": 0.9431990951299667,
      "num_tokens": 18262808.0,
      "step": 1135
    },
    {
      "aux_loss": 8.088939094543457,
      "entropy": 0.162516987323761,
      "epoch": 0.00114,
      "grad_norm": 0.13790178298950195,
      "learning_rate": 2e-05,
      "loss": 0.2050339937210083,
      "mean_token_accuracy": 0.9429171621799469,
      "num_tokens": 18343894.0,
      "step": 1140
    },
    {
      "aux_loss": 8.099242973327637,
      "entropy": 0.1909376263618469,
      "epoch": 0.001145,
      "grad_norm": 0.13040708005428314,
      "learning_rate": 2e-05,
      "loss": 0.23065586090087892,
      "mean_token_accuracy": 0.9361128509044647,
      "num_tokens": 18425420.0,
      "step": 1145
    },
    {
      "aux_loss": 8.109739398956298,
      "entropy": 0.19257219433784484,
      "epoch": 0.00115,
      "grad_norm": 0.13617382943630219,
      "learning_rate": 2e-05,
      "loss": 0.2169320821762085,
      "mean_token_accuracy": 0.9386547476053237,
      "num_tokens": 18506048.0,
      "step": 1150
    },
    {
      "aux_loss": 8.119318962097168,
      "entropy": 0.17985256910324096,
      "epoch": 0.001155,
      "grad_norm": 0.1464129090309143,
      "learning_rate": 2e-05,
      "loss": 0.20620300769805908,
      "mean_token_accuracy": 0.9404237031936645,
      "num_tokens": 18586609.0,
      "step": 1155
    },
    {
      "aux_loss": 8.136944580078126,
      "entropy": 0.17993081212043763,
      "epoch": 0.00116,
      "grad_norm": 0.1385568082332611,
      "learning_rate": 2e-05,
      "loss": 0.21773624420166016,
      "mean_token_accuracy": 0.9385370492935181,
      "num_tokens": 18666884.0,
      "step": 1160
    },
    {
      "aux_loss": 8.16204137802124,
      "entropy": 0.1734834022819996,
      "epoch": 0.001165,
      "grad_norm": 0.16297543048858643,
      "learning_rate": 2e-05,
      "loss": 0.19739640951156617,
      "mean_token_accuracy": 0.9443773657083512,
      "num_tokens": 18747693.0,
      "step": 1165
    },
    {
      "aux_loss": 8.095090389251709,
      "entropy": 0.17245537899434565,
      "epoch": 0.00117,
      "grad_norm": 0.1412307322025299,
      "learning_rate": 2e-05,
      "loss": 0.2139646053314209,
      "mean_token_accuracy": 0.9405614197254181,
      "num_tokens": 18826580.0,
      "step": 1170
    },
    {
      "aux_loss": 8.09668779373169,
      "entropy": 0.1829852543771267,
      "epoch": 0.001175,
      "grad_norm": 0.1572728455066681,
      "learning_rate": 2e-05,
      "loss": 0.22257506847381592,
      "mean_token_accuracy": 0.9376355350017548,
      "num_tokens": 18907868.0,
      "step": 1175
    },
    {
      "aux_loss": 8.104577207565308,
      "entropy": 0.18904553726315498,
      "epoch": 0.00118,
      "grad_norm": 0.13721878826618195,
      "learning_rate": 2e-05,
      "loss": 0.2151425838470459,
      "mean_token_accuracy": 0.9397189229726791,
      "num_tokens": 18988636.0,
      "step": 1180
    },
    {
      "aux_loss": 8.118225622177125,
      "entropy": 0.18333376720547676,
      "epoch": 0.001185,
      "grad_norm": 0.14412416517734528,
      "learning_rate": 2e-05,
      "loss": 0.21259968280792235,
      "mean_token_accuracy": 0.9398420035839081,
      "num_tokens": 19069174.0,
      "step": 1185
    },
    {
      "aux_loss": 8.134906196594239,
      "entropy": 0.17814260721206665,
      "epoch": 0.00119,
      "grad_norm": 0.13498298823833466,
      "learning_rate": 2e-05,
      "loss": 0.2124964714050293,
      "mean_token_accuracy": 0.9405255049467087,
      "num_tokens": 19149371.0,
      "step": 1190
    },
    {
      "aux_loss": 8.156946563720703,
      "entropy": 0.1696353040635586,
      "epoch": 0.001195,
      "grad_norm": 0.16295582056045532,
      "learning_rate": 2e-05,
      "loss": 0.19403015375137328,
      "mean_token_accuracy": 0.9445232212543487,
      "num_tokens": 19229674.0,
      "step": 1195
    },
    {
      "aux_loss": 8.112064218521118,
      "entropy": 0.1671122845262289,
      "epoch": 0.0012,
      "grad_norm": 0.14335326850414276,
      "learning_rate": 2e-05,
      "loss": 0.2131701946258545,
      "mean_token_accuracy": 0.9426359623670578,
      "num_tokens": 19306860.0,
      "step": 1200
    },
    {
      "aux_loss": 8.102148056030273,
      "entropy": 0.17936039455235003,
      "epoch": 0.001205,
      "grad_norm": 0.138744056224823,
      "learning_rate": 2e-05,
      "loss": 0.2263488292694092,
      "mean_token_accuracy": 0.9367114216089248,
      "num_tokens": 19388241.0,
      "step": 1205
    },
    {
      "aux_loss": 8.107982635498047,
      "entropy": 0.18988944292068483,
      "epoch": 0.00121,
      "grad_norm": 0.1490761637687683,
      "learning_rate": 2e-05,
      "loss": 0.211828875541687,
      "mean_token_accuracy": 0.9403735935688019,
      "num_tokens": 19469572.0,
      "step": 1210
    },
    {
      "aux_loss": 8.120537662506104,
      "entropy": 0.18598054051399232,
      "epoch": 0.001215,
      "grad_norm": 0.13718858361244202,
      "learning_rate": 2e-05,
      "loss": 0.21541576385498046,
      "mean_token_accuracy": 0.9391979366540909,
      "num_tokens": 19550158.0,
      "step": 1215
    },
    {
      "aux_loss": 8.129882097244263,
      "entropy": 0.1692873366177082,
      "epoch": 0.00122,
      "grad_norm": 0.1462230086326599,
      "learning_rate": 2e-05,
      "loss": 0.19773582220077515,
      "mean_token_accuracy": 0.9440513670444488,
      "num_tokens": 19630593.0,
      "step": 1220
    },
    {
      "aux_loss": 8.150879144668579,
      "entropy": 0.1672566942870617,
      "epoch": 0.001225,
      "grad_norm": 0.1628735363483429,
      "learning_rate": 2e-05,
      "loss": 0.20410618782043458,
      "mean_token_accuracy": 0.9435446441173554,
      "num_tokens": 19710578.0,
      "step": 1225
    },
    {
      "aux_loss": 8.112903356552124,
      "entropy": 0.16016247272491455,
      "epoch": 0.00123,
      "grad_norm": 0.13584807515144348,
      "learning_rate": 2e-05,
      "loss": 0.19806501865386963,
      "mean_token_accuracy": 0.9454169005155564,
      "num_tokens": 19790178.0,
      "step": 1230
    },
    {
      "aux_loss": 8.097063112258912,
      "entropy": 0.1896512873470783,
      "epoch": 0.001235,
      "grad_norm": 0.13905711472034454,
      "learning_rate": 2e-05,
      "loss": 0.22745237350463868,
      "mean_token_accuracy": 0.9353796660900116,
      "num_tokens": 19871450.0,
      "step": 1235
    },
    {
      "aux_loss": 8.103591251373292,
      "entropy": 0.18758923336863517,
      "epoch": 0.00124,
      "grad_norm": 0.14005883038043976,
      "learning_rate": 2e-05,
      "loss": 0.21161456108093263,
      "mean_token_accuracy": 0.9404743552207947,
      "num_tokens": 19952818.0,
      "step": 1240
    },
    {
      "aux_loss": 8.119750118255615,
      "entropy": 0.17323974296450614,
      "epoch": 0.001245,
      "grad_norm": 0.14812205731868744,
      "learning_rate": 2e-05,
      "loss": 0.20181419849395751,
      "mean_token_accuracy": 0.943002137541771,
      "num_tokens": 20033581.0,
      "step": 1245
    },
    {
      "aux_loss": 8.13001561164856,
      "entropy": 0.17530782744288445,
      "epoch": 0.00125,
      "grad_norm": 0.13251017034053802,
      "learning_rate": 2e-05,
      "loss": 0.21283555030822754,
      "mean_token_accuracy": 0.9399103164672852,
      "num_tokens": 20114398.0,
      "step": 1250
    },
    {
      "aux_loss": 8.147759580612183,
      "entropy": 0.16442596018314362,
      "epoch": 0.001255,
      "grad_norm": 0.15714676678180695,
      "learning_rate": 2e-05,
      "loss": 0.18702534437179566,
      "mean_token_accuracy": 0.9467514336109162,
      "num_tokens": 20194434.0,
      "step": 1255
    },
    {
      "aux_loss": 8.109157800674438,
      "entropy": 0.1626373641192913,
      "epoch": 0.00126,
      "grad_norm": 0.1565043330192566,
      "learning_rate": 2e-05,
      "loss": 0.21089496612548828,
      "mean_token_accuracy": 0.942788028717041,
      "num_tokens": 20275698.0,
      "step": 1260
    },
    {
      "aux_loss": 8.093947696685792,
      "entropy": 0.19617915898561478,
      "epoch": 0.001265,
      "grad_norm": 0.1432366520166397,
      "learning_rate": 2e-05,
      "loss": 0.24150006771087645,
      "mean_token_accuracy": 0.9325732469558716,
      "num_tokens": 20356606.0,
      "step": 1265
    },
    {
      "aux_loss": 8.100923013687133,
      "entropy": 0.20489457100629807,
      "epoch": 0.00127,
      "grad_norm": 0.142571359872818,
      "learning_rate": 2e-05,
      "loss": 0.22859885692596435,
      "mean_token_accuracy": 0.9360031843185425,
      "num_tokens": 20437492.0,
      "step": 1270
    },
    {
      "aux_loss": 8.113832139968872,
      "entropy": 0.18682107999920844,
      "epoch": 0.001275,
      "grad_norm": 0.15194223821163177,
      "learning_rate": 2e-05,
      "loss": 0.21301717758178712,
      "mean_token_accuracy": 0.93876693546772,
      "num_tokens": 20518204.0,
      "step": 1275
    },
    {
      "aux_loss": 8.12835898399353,
      "entropy": 0.172673948854208,
      "epoch": 0.00128,
      "grad_norm": 0.14351797103881836,
      "learning_rate": 2e-05,
      "loss": 0.20544931888580323,
      "mean_token_accuracy": 0.9409793645143509,
      "num_tokens": 20598785.0,
      "step": 1280
    },
    {
      "aux_loss": 8.146044445037841,
      "entropy": 0.1725030966103077,
      "epoch": 0.001285,
      "grad_norm": 0.14401860535144806,
      "learning_rate": 2e-05,
      "loss": 0.20696258544921875,
      "mean_token_accuracy": 0.9420131802558899,
      "num_tokens": 20678911.0,
      "step": 1285
    },
    {
      "aux_loss": 8.123119592666626,
      "entropy": 0.17102996557950972,
      "epoch": 0.00129,
      "grad_norm": 0.1375003606081009,
      "learning_rate": 2e-05,
      "loss": 0.20492122173309327,
      "mean_token_accuracy": 0.9413111358880997,
      "num_tokens": 20755503.0,
      "step": 1290
    },
    {
      "aux_loss": 8.0912504196167,
      "entropy": 0.17658306807279586,
      "epoch": 0.001295,
      "grad_norm": 0.13257145881652832,
      "learning_rate": 2e-05,
      "loss": 0.20998308658599854,
      "mean_token_accuracy": 0.940663531422615,
      "num_tokens": 20836245.0,
      "step": 1295
    },
    {
      "aux_loss": 8.099381446838379,
      "entropy": 0.1868967518210411,
      "epoch": 0.0013,
      "grad_norm": 0.14198297262191772,
      "learning_rate": 2e-05,
      "loss": 0.21974670886993408,
      "mean_token_accuracy": 0.9376268059015274,
      "num_tokens": 20917294.0,
      "step": 1300
    },
    {
      "aux_loss": 8.108508634567261,
      "entropy": 0.18228753432631492,
      "epoch": 0.001305,
      "grad_norm": 0.13187234103679657,
      "learning_rate": 2e-05,
      "loss": 0.21817548274993898,
      "mean_token_accuracy": 0.9392153471708298,
      "num_tokens": 20997680.0,
      "step": 1305
    },
    {
      "aux_loss": 8.122104930877686,
      "entropy": 0.18339259847998618,
      "epoch": 0.00131,
      "grad_norm": 0.1490059345960617,
      "learning_rate": 2e-05,
      "loss": 0.21026885509490967,
      "mean_token_accuracy": 0.9402256816625595,
      "num_tokens": 21078472.0,
      "step": 1310
    },
    {
      "aux_loss": 8.134609413146972,
      "entropy": 0.17264059111475943,
      "epoch": 0.001315,
      "grad_norm": 0.1463150978088379,
      "learning_rate": 2e-05,
      "loss": 0.20235130786895753,
      "mean_token_accuracy": 0.9432796269655228,
      "num_tokens": 21158851.0,
      "step": 1315
    },
    {
      "aux_loss": 8.132504081726074,
      "entropy": 0.15988932885229587,
      "epoch": 0.00132,
      "grad_norm": 0.14908353984355927,
      "learning_rate": 2e-05,
      "loss": 0.1955566883087158,
      "mean_token_accuracy": 0.9455756425857544,
      "num_tokens": 21236649.0,
      "step": 1320
    },
    {
      "aux_loss": 8.091651821136475,
      "entropy": 0.16718392595648765,
      "epoch": 0.001325,
      "grad_norm": 0.15275311470031738,
      "learning_rate": 2e-05,
      "loss": 0.21028823852539064,
      "mean_token_accuracy": 0.9421744287014008,
      "num_tokens": 21317997.0,
      "step": 1325
    },
    {
      "aux_loss": 8.095997381210328,
      "entropy": 0.1901338778436184,
      "epoch": 0.00133,
      "grad_norm": 0.14271008968353271,
      "learning_rate": 2e-05,
      "loss": 0.2200601577758789,
      "mean_token_accuracy": 0.9385367631912231,
      "num_tokens": 21399182.0,
      "step": 1330
    },
    {
      "aux_loss": 8.110585117340088,
      "entropy": 0.20013106390833854,
      "epoch": 0.001335,
      "grad_norm": 0.13731908798217773,
      "learning_rate": 2e-05,
      "loss": 0.22662696838378907,
      "mean_token_accuracy": 0.9368092805147171,
      "num_tokens": 21480251.0,
      "step": 1335
    },
    {
      "aux_loss": 8.120805788040162,
      "entropy": 0.1727154590189457,
      "epoch": 0.00134,
      "grad_norm": 0.1378210335969925,
      "learning_rate": 2e-05,
      "loss": 0.20113024711608887,
      "mean_token_accuracy": 0.9439943104982376,
      "num_tokens": 21561156.0,
      "step": 1340
    },
    {
      "aux_loss": 8.133132648468017,
      "entropy": 0.1674143709242344,
      "epoch": 0.001345,
      "grad_norm": 0.1513887643814087,
      "learning_rate": 2e-05,
      "loss": 0.20577425956726075,
      "mean_token_accuracy": 0.942402470111847,
      "num_tokens": 21642185.0,
      "step": 1345
    },
    {
      "aux_loss": 8.139795017242431,
      "entropy": 0.17304934822022916,
      "epoch": 0.00135,
      "grad_norm": 0.12145114690065384,
      "learning_rate": 2e-05,
      "loss": 0.20785124301910402,
      "mean_token_accuracy": 0.942343357205391,
      "num_tokens": 21719162.0,
      "step": 1350
    },
    {
      "aux_loss": 8.08781909942627,
      "entropy": 0.1742361642420292,
      "epoch": 0.001355,
      "grad_norm": 0.14099696278572083,
      "learning_rate": 2e-05,
      "loss": 0.2182718276977539,
      "mean_token_accuracy": 0.940433731675148,
      "num_tokens": 21800626.0,
      "step": 1355
    },
    {
      "aux_loss": 8.098934936523438,
      "entropy": 0.20285899192094803,
      "epoch": 0.00136,
      "grad_norm": 0.15401683747768402,
      "learning_rate": 2e-05,
      "loss": 0.23468153476715087,
      "mean_token_accuracy": 0.9335884660482406,
      "num_tokens": 21881696.0,
      "step": 1360
    },
    {
      "aux_loss": 8.106411504745484,
      "entropy": 0.18526409342885017,
      "epoch": 0.001365,
      "grad_norm": 0.1306249499320984,
      "learning_rate": 2e-05,
      "loss": 0.21111090183258058,
      "mean_token_accuracy": 0.9408644020557404,
      "num_tokens": 21962450.0,
      "step": 1365
    },
    {
      "aux_loss": 8.121489858627319,
      "entropy": 0.18036528900265694,
      "epoch": 0.00137,
      "grad_norm": 0.14501120150089264,
      "learning_rate": 2e-05,
      "loss": 0.21063520908355712,
      "mean_token_accuracy": 0.9411169201135635,
      "num_tokens": 22042655.0,
      "step": 1370
    },
    {
      "aux_loss": 8.132449626922607,
      "entropy": 0.16052772253751754,
      "epoch": 0.001375,
      "grad_norm": 0.142227441072464,
      "learning_rate": 2e-05,
      "loss": 0.19645824432373046,
      "mean_token_accuracy": 0.94445960521698,
      "num_tokens": 22123291.0,
      "step": 1375
    },
    {
      "aux_loss": 8.15138282775879,
      "entropy": 0.16159301958978176,
      "epoch": 0.00138,
      "grad_norm": 0.13989758491516113,
      "learning_rate": 2e-05,
      "loss": 0.1932361364364624,
      "mean_token_accuracy": 0.9453170746564865,
      "num_tokens": 22202462.0,
      "step": 1380
    },
    {
      "aux_loss": 8.08789939880371,
      "entropy": 0.17011022567749023,
      "epoch": 0.001385,
      "grad_norm": 0.14079993963241577,
      "learning_rate": 2e-05,
      "loss": 0.20923681259155275,
      "mean_token_accuracy": 0.9421351909637451,
      "num_tokens": 22283958.0,
      "step": 1385
    },
    {
      "aux_loss": 8.095738554000855,
      "entropy": 0.18748774603009224,
      "epoch": 0.00139,
      "grad_norm": 0.1360437422990799,
      "learning_rate": 2e-05,
      "loss": 0.2199646234512329,
      "mean_token_accuracy": 0.9384114235639572,
      "num_tokens": 22365454.0,
      "step": 1390
    },
    {
      "aux_loss": 8.10390763282776,
      "entropy": 0.1814654491841793,
      "epoch": 0.001395,
      "grad_norm": 0.12834835052490234,
      "learning_rate": 2e-05,
      "loss": 0.21202080249786376,
      "mean_token_accuracy": 0.9412189185619354,
      "num_tokens": 22446565.0,
      "step": 1395
    },
    {
      "aux_loss": 8.116164302825927,
      "entropy": 0.17915418669581412,
      "epoch": 0.0014,
      "grad_norm": 0.1470014899969101,
      "learning_rate": 2e-05,
      "loss": 0.20995094776153564,
      "mean_token_accuracy": 0.9401764899492264,
      "num_tokens": 22527025.0,
      "step": 1400
    },
    {
      "aux_loss": 8.131404733657837,
      "entropy": 0.17990972548723222,
      "epoch": 0.001405,
      "grad_norm": 0.14340564608573914,
      "learning_rate": 2e-05,
      "loss": 0.2145388126373291,
      "mean_token_accuracy": 0.9400021612644196,
      "num_tokens": 22607222.0,
      "step": 1405
    },
    {
      "aux_loss": 8.152696418762208,
      "entropy": 0.1697441078722477,
      "epoch": 0.00141,
      "grad_norm": 0.1787918657064438,
      "learning_rate": 2e-05,
      "loss": 0.19432330131530762,
      "mean_token_accuracy": 0.9452683657407761,
      "num_tokens": 22687723.0,
      "step": 1410
    },
    {
      "aux_loss": 8.087120485305785,
      "entropy": 0.17364331483840942,
      "epoch": 0.001415,
      "grad_norm": 0.14804264903068542,
      "learning_rate": 2e-05,
      "loss": 0.22220132350921631,
      "mean_token_accuracy": 0.9390887200832367,
      "num_tokens": 22766548.0,
      "step": 1415
    },
    {
      "aux_loss": 8.093894338607788,
      "entropy": 0.18652292788028718,
      "epoch": 0.00142,
      "grad_norm": 0.14664113521575928,
      "learning_rate": 2e-05,
      "loss": 0.2241745948791504,
      "mean_token_accuracy": 0.9366730749607086,
      "num_tokens": 22847865.0,
      "step": 1420
    },
    {
      "aux_loss": 8.102896785736084,
      "entropy": 0.19468777775764465,
      "epoch": 0.001425,
      "grad_norm": 0.1493048369884491,
      "learning_rate": 2e-05,
      "loss": 0.21715831756591797,
      "mean_token_accuracy": 0.9390844434499741,
      "num_tokens": 22928761.0,
      "step": 1425
    },
    {
      "aux_loss": 8.116300249099732,
      "entropy": 0.18231945857405663,
      "epoch": 0.00143,
      "grad_norm": 0.13884210586547852,
      "learning_rate": 2e-05,
      "loss": 0.2111462116241455,
      "mean_token_accuracy": 0.939801424741745,
      "num_tokens": 23009438.0,
      "step": 1430
    },
    {
      "aux_loss": 8.132174634933472,
      "entropy": 0.16873669773340225,
      "epoch": 0.001435,
      "grad_norm": 0.16086970269680023,
      "learning_rate": 2e-05,
      "loss": 0.1996437430381775,
      "mean_token_accuracy": 0.944518530368805,
      "num_tokens": 23089375.0,
      "step": 1435
    },
    {
      "aux_loss": 8.149925804138183,
      "entropy": 0.16066225245594978,
      "epoch": 0.00144,
      "grad_norm": 0.16884085536003113,
      "learning_rate": 2e-05,
      "loss": 0.19438048601150512,
      "mean_token_accuracy": 0.9448157876729966,
      "num_tokens": 23170066.0,
      "step": 1440
    },
    {
      "aux_loss": 8.09245195388794,
      "entropy": 0.15585733875632285,
      "epoch": 0.001445,
      "grad_norm": 0.14111007750034332,
      "learning_rate": 2e-05,
      "loss": 0.19613447189331054,
      "mean_token_accuracy": 0.9457630664110184,
      "num_tokens": 23250511.0,
      "step": 1445
    },
    {
      "aux_loss": 8.091105413436889,
      "entropy": 0.1920392207801342,
      "epoch": 0.00145,
      "grad_norm": 0.15117305517196655,
      "learning_rate": 2e-05,
      "loss": 0.2302844524383545,
      "mean_token_accuracy": 0.9350064992904663,
      "num_tokens": 23332036.0,
      "step": 1450
    },
    {
      "aux_loss": 8.103524017333985,
      "entropy": 0.17666609063744546,
      "epoch": 0.001455,
      "grad_norm": 0.15425564348697662,
      "learning_rate": 2e-05,
      "loss": 0.20472273826599122,
      "mean_token_accuracy": 0.9426793485879899,
      "num_tokens": 23412945.0,
      "step": 1455
    },
    {
      "aux_loss": 8.113294744491578,
      "entropy": 0.1808626964688301,
      "epoch": 0.00146,
      "grad_norm": 0.14899621903896332,
      "learning_rate": 2e-05,
      "loss": 0.21448283195495604,
      "mean_token_accuracy": 0.9400674223899841,
      "num_tokens": 23493932.0,
      "step": 1460
    },
    {
      "aux_loss": 8.12929720878601,
      "entropy": 0.1730622112751007,
      "epoch": 0.001465,
      "grad_norm": 0.15310509502887726,
      "learning_rate": 2e-05,
      "loss": 0.1996227025985718,
      "mean_token_accuracy": 0.943952077627182,
      "num_tokens": 23574061.0,
      "step": 1465
    },
    {
      "aux_loss": 8.151718235015869,
      "entropy": 0.16063716560602187,
      "epoch": 0.00147,
      "grad_norm": 0.16773603856563568,
      "learning_rate": 2e-05,
      "loss": 0.189278507232666,
      "mean_token_accuracy": 0.9467889577150345,
      "num_tokens": 23654574.0,
      "step": 1470
    },
    {
      "aux_loss": 8.094487380981445,
      "entropy": 0.15401179529726505,
      "epoch": 0.001475,
      "grad_norm": 0.14586107432842255,
      "learning_rate": 2e-05,
      "loss": 0.20016000270843506,
      "mean_token_accuracy": 0.9449825167655945,
      "num_tokens": 23733817.0,
      "step": 1475
    },
    {
      "aux_loss": 8.091149282455444,
      "entropy": 0.19360100850462914,
      "epoch": 0.00148,
      "grad_norm": 0.14854799211025238,
      "learning_rate": 2e-05,
      "loss": 0.23483653068542482,
      "mean_token_accuracy": 0.9342969328165054,
      "num_tokens": 23815269.0,
      "step": 1480
    },
    {
      "aux_loss": 8.098963975906372,
      "entropy": 0.1929106667637825,
      "epoch": 0.001485,
      "grad_norm": 0.13886724412441254,
      "learning_rate": 2e-05,
      "loss": 0.21658732891082763,
      "mean_token_accuracy": 0.939777034521103,
      "num_tokens": 23896528.0,
      "step": 1485
    },
    {
      "aux_loss": 8.108890295028687,
      "entropy": 0.18312879726290704,
      "epoch": 0.00149,
      "grad_norm": 0.14512109756469727,
      "learning_rate": 2e-05,
      "loss": 0.2098008394241333,
      "mean_token_accuracy": 0.9403581470251083,
      "num_tokens": 23977386.0,
      "step": 1490
    },
    {
      "aux_loss": 8.12533917427063,
      "entropy": 0.16724784225225447,
      "epoch": 0.001495,
      "grad_norm": 0.14004266262054443,
      "learning_rate": 2e-05,
      "loss": 0.2040783405303955,
      "mean_token_accuracy": 0.9425091087818146,
      "num_tokens": 24057921.0,
      "step": 1495
    },
    {
      "aux_loss": 8.14401888847351,
      "entropy": 0.16301089823246,
      "epoch": 0.0015,
      "grad_norm": 0.16092567145824432,
      "learning_rate": 2e-05,
      "loss": 0.19011870622634888,
      "mean_token_accuracy": 0.9468242824077606,
      "num_tokens": 24138053.0,
      "step": 1500
    },
    {
      "aux_loss": 8.108096361160278,
      "entropy": 0.1561386350542307,
      "epoch": 0.001505,
      "grad_norm": 0.15491639077663422,
      "learning_rate": 2e-05,
      "loss": 0.19484918117523192,
      "mean_token_accuracy": 0.9462997406721115,
      "num_tokens": 24215221.0,
      "step": 1505
    },
    {
      "aux_loss": 8.088674736022949,
      "entropy": 0.17986297979950905,
      "epoch": 0.00151,
      "grad_norm": 0.14700467884540558,
      "learning_rate": 2e-05,
      "loss": 0.2231841802597046,
      "mean_token_accuracy": 0.9368600070476532,
      "num_tokens": 24296079.0,
      "step": 1510
    },
    {
      "aux_loss": 8.095492744445801,
      "entropy": 0.18229425847530364,
      "epoch": 0.001515,
      "grad_norm": 0.14003045856952667,
      "learning_rate": 2e-05,
      "loss": 0.20808348655700684,
      "mean_token_accuracy": 0.9413382917642593,
      "num_tokens": 24377526.0,
      "step": 1515
    },
    {
      "aux_loss": 8.10906138420105,
      "entropy": 0.18038261011242868,
      "epoch": 0.00152,
      "grad_norm": 0.15010018646717072,
      "learning_rate": 2e-05,
      "loss": 0.20852205753326417,
      "mean_token_accuracy": 0.9413996934890747,
      "num_tokens": 24458627.0,
      "step": 1520
    },
    {
      "aux_loss": 8.123844289779663,
      "entropy": 0.17206328213214875,
      "epoch": 0.001525,
      "grad_norm": 0.14683103561401367,
      "learning_rate": 2e-05,
      "loss": 0.20467498302459716,
      "mean_token_accuracy": 0.9411923170089722,
      "num_tokens": 24539422.0,
      "step": 1525
    },
    {
      "aux_loss": 8.140221214294433,
      "entropy": 0.1623208202421665,
      "epoch": 0.00153,
      "grad_norm": 0.16020053625106812,
      "learning_rate": 2e-05,
      "loss": 0.1922840356826782,
      "mean_token_accuracy": 0.945524924993515,
      "num_tokens": 24619704.0,
      "step": 1530
    },
    {
      "aux_loss": 8.112646102905273,
      "entropy": 0.15188264660537243,
      "epoch": 0.001535,
      "grad_norm": 0.1351615935564041,
      "learning_rate": 2e-05,
      "loss": 0.18567148447036744,
      "mean_token_accuracy": 0.9497504830360413,
      "num_tokens": 24700558.0,
      "step": 1535
    },
    {
      "aux_loss": 8.093827390670777,
      "entropy": 0.18044227585196496,
      "epoch": 0.00154,
      "grad_norm": 0.1527668535709381,
      "learning_rate": 2e-05,
      "loss": 0.22727639675140382,
      "mean_token_accuracy": 0.9360405266284942,
      "num_tokens": 24781672.0,
      "step": 1540
    },
    {
      "aux_loss": 8.094546461105347,
      "entropy": 0.20025886446237565,
      "epoch": 0.001545,
      "grad_norm": 0.14517077803611755,
      "learning_rate": 2e-05,
      "loss": 0.2319253444671631,
      "mean_token_accuracy": 0.9350155264139175,
      "num_tokens": 24862726.0,
      "step": 1545
    },
    {
      "aux_loss": 8.110199642181396,
      "entropy": 0.1861427828669548,
      "epoch": 0.00155,
      "grad_norm": 0.15331503748893738,
      "learning_rate": 2e-05,
      "loss": 0.21457738876342775,
      "mean_token_accuracy": 0.9389505892992019,
      "num_tokens": 24943775.0,
      "step": 1550
    },
    {
      "aux_loss": 8.120828199386597,
      "entropy": 0.1685674399137497,
      "epoch": 0.001555,
      "grad_norm": 0.1569015085697174,
      "learning_rate": 2e-05,
      "loss": 0.19857349395751953,
      "mean_token_accuracy": 0.9437560766935349,
      "num_tokens": 25024772.0,
      "step": 1555
    },
    {
      "aux_loss": 8.13948163986206,
      "entropy": 0.15821726694703103,
      "epoch": 0.00156,
      "grad_norm": 0.16434122622013092,
      "learning_rate": 2e-05,
      "loss": 0.1915326476097107,
      "mean_token_accuracy": 0.9459531784057618,
      "num_tokens": 25105051.0,
      "step": 1560
    },
    {
      "aux_loss": 8.127675104141236,
      "entropy": 0.1545293591916561,
      "epoch": 0.001565,
      "grad_norm": 0.15886199474334717,
      "learning_rate": 2e-05,
      "loss": 0.19543263912200928,
      "mean_token_accuracy": 0.9462945342063904,
      "num_tokens": 25184642.0,
      "step": 1565
    },
    {
      "aux_loss": 8.08992862701416,
      "entropy": 0.17979532293975353,
      "epoch": 0.00157,
      "grad_norm": 0.13161753118038177,
      "learning_rate": 2e-05,
      "loss": 0.21586837768554687,
      "mean_token_accuracy": 0.9389263749122619,
      "num_tokens": 25266000.0,
      "step": 1570
    },
    {
      "aux_loss": 8.09758381843567,
      "entropy": 0.20703106448054315,
      "epoch": 0.001575,
      "grad_norm": 0.13990826904773712,
      "learning_rate": 2e-05,
      "loss": 0.23759994506835938,
      "mean_token_accuracy": 0.9331592172384262,
      "num_tokens": 25346583.0,
      "step": 1575
    },
    {
      "aux_loss": 8.106312370300293,
      "entropy": 0.1796684518456459,
      "epoch": 0.00158,
      "grad_norm": 0.14089328050613403,
      "learning_rate": 2e-05,
      "loss": 0.2043560266494751,
      "mean_token_accuracy": 0.942157757282257,
      "num_tokens": 25427514.0,
      "step": 1580
    },
    {
      "aux_loss": 8.124406957626343,
      "entropy": 0.17517981231212615,
      "epoch": 0.001585,
      "grad_norm": 0.14138734340667725,
      "learning_rate": 2e-05,
      "loss": 0.20602011680603027,
      "mean_token_accuracy": 0.9417855650186538,
      "num_tokens": 25508402.0,
      "step": 1585
    },
    {
      "aux_loss": 8.14156994819641,
      "entropy": 0.159897642955184,
      "epoch": 0.00159,
      "grad_norm": 0.14980235695838928,
      "learning_rate": 2e-05,
      "loss": 0.18918629884719848,
      "mean_token_accuracy": 0.9464654803276062,
      "num_tokens": 25588248.0,
      "step": 1590
    },
    {
      "aux_loss": 8.132224559783936,
      "entropy": 0.15646722465753554,
      "epoch": 0.001595,
      "grad_norm": 0.14346034824848175,
      "learning_rate": 2e-05,
      "loss": 0.19590044021606445,
      "mean_token_accuracy": 0.9455728471279145,
      "num_tokens": 25668812.0,
      "step": 1595
    },
    {
      "aux_loss": 8.091664552688599,
      "entropy": 0.16691523045301437,
      "epoch": 0.0016,
      "grad_norm": 0.14989890158176422,
      "learning_rate": 2e-05,
      "loss": 0.20818891525268554,
      "mean_token_accuracy": 0.9413408249616623,
      "num_tokens": 25750022.0,
      "step": 1600
    },
    {
      "aux_loss": 8.093881845474243,
      "entropy": 0.19223699048161508,
      "epoch": 0.001605,
      "grad_norm": 0.15126468241214752,
      "learning_rate": 2e-05,
      "loss": 0.22417640686035156,
      "mean_token_accuracy": 0.9378127545118332,
      "num_tokens": 25830960.0,
      "step": 1605
    },
    {
      "aux_loss": 8.106926345825196,
      "entropy": 0.18031547740101814,
      "epoch": 0.00161,
      "grad_norm": 0.14342601597309113,
      "learning_rate": 2e-05,
      "loss": 0.20777347087860107,
      "mean_token_accuracy": 0.9411538213491439,
      "num_tokens": 25912001.0,
      "step": 1610
    },
    {
      "aux_loss": 8.118624687194824,
      "entropy": 0.17412861213088035,
      "epoch": 0.001615,
      "grad_norm": 0.14276643097400665,
      "learning_rate": 2e-05,
      "loss": 0.20309576988220215,
      "mean_token_accuracy": 0.9427552461624146,
      "num_tokens": 25992996.0,
      "step": 1615
    },
    {
      "aux_loss": 8.134443330764771,
      "entropy": 0.1625348597764969,
      "epoch": 0.00162,
      "grad_norm": 0.16904227435588837,
      "learning_rate": 2e-05,
      "loss": 0.19113484621047974,
      "mean_token_accuracy": 0.946293231844902,
      "num_tokens": 26073465.0,
      "step": 1620
    },
    {
      "aux_loss": 8.128686761856079,
      "entropy": 0.14892494566738607,
      "epoch": 0.001625,
      "grad_norm": 0.14239241182804108,
      "learning_rate": 2e-05,
      "loss": 0.18040697574615477,
      "mean_token_accuracy": 0.9493981957435608,
      "num_tokens": 26152691.0,
      "step": 1625
    },
    {
      "aux_loss": 8.086278486251832,
      "entropy": 0.17021566741168498,
      "epoch": 0.00163,
      "grad_norm": 0.14199447631835938,
      "learning_rate": 2e-05,
      "loss": 0.22021899223327637,
      "mean_token_accuracy": 0.9390756160020828,
      "num_tokens": 26233800.0,
      "step": 1630
    },
    {
      "aux_loss": 8.091834163665771,
      "entropy": 0.19363221898674965,
      "epoch": 0.001635,
      "grad_norm": 0.141326442360878,
      "learning_rate": 2e-05,
      "loss": 0.2295623540878296,
      "mean_token_accuracy": 0.9363398462533951,
      "num_tokens": 26314237.0,
      "step": 1635
    },
    {
      "aux_loss": 8.099656438827514,
      "entropy": 0.1877741038799286,
      "epoch": 0.00164,
      "grad_norm": 0.13307121396064758,
      "learning_rate": 2e-05,
      "loss": 0.21384093761444092,
      "mean_token_accuracy": 0.9402267515659333,
      "num_tokens": 26395203.0,
      "step": 1640
    },
    {
      "aux_loss": 8.112567138671874,
      "entropy": 0.17270937636494638,
      "epoch": 0.001645,
      "grad_norm": 0.1454281508922577,
      "learning_rate": 2e-05,
      "loss": 0.19741926193237305,
      "mean_token_accuracy": 0.9443966925144196,
      "num_tokens": 26475783.0,
      "step": 1645
    },
    {
      "aux_loss": 8.130879163742065,
      "entropy": 0.16381614953279494,
      "epoch": 0.00165,
      "grad_norm": 0.15070879459381104,
      "learning_rate": 2e-05,
      "loss": 0.19763474464416503,
      "mean_token_accuracy": 0.9448977142572403,
      "num_tokens": 26556693.0,
      "step": 1650
    },
    {
      "aux_loss": 8.136204147338868,
      "entropy": 0.15377900041639805,
      "epoch": 0.001655,
      "grad_norm": 0.12270507216453552,
      "learning_rate": 2e-05,
      "loss": 0.18577005863189697,
      "mean_token_accuracy": 0.9487336248159408,
      "num_tokens": 26636170.0,
      "step": 1655
    },
    {
      "aux_loss": 8.082316684722901,
      "entropy": 0.1666357085108757,
      "epoch": 0.00166,
      "grad_norm": 0.1402491331100464,
      "learning_rate": 2e-05,
      "loss": 0.2062994956970215,
      "mean_token_accuracy": 0.9427699983119965,
      "num_tokens": 26717546.0,
      "step": 1660
    },
    {
      "aux_loss": 8.094042301177979,
      "entropy": 0.18402569144964218,
      "epoch": 0.001665,
      "grad_norm": 0.1399538666009903,
      "learning_rate": 2e-05,
      "loss": 0.22184386253356933,
      "mean_token_accuracy": 0.9360486805438996,
      "num_tokens": 26798880.0,
      "step": 1665
    },
    {
      "aux_loss": 8.096484851837157,
      "entropy": 0.17516063004732133,
      "epoch": 0.00167,
      "grad_norm": 0.1460048258304596,
      "learning_rate": 2e-05,
      "loss": 0.20700514316558838,
      "mean_token_accuracy": 0.9405145049095154,
      "num_tokens": 26879876.0,
      "step": 1670
    },
    {
      "aux_loss": 8.108837652206422,
      "entropy": 0.17554794177412986,
      "epoch": 0.001675,
      "grad_norm": 0.14934812486171722,
      "learning_rate": 2e-05,
      "loss": 0.1992682099342346,
      "mean_token_accuracy": 0.9441122353076935,
      "num_tokens": 26960459.0,
      "step": 1675
    },
    {
      "aux_loss": 8.122825431823731,
      "entropy": 0.16358745619654655,
      "epoch": 0.00168,
      "grad_norm": 0.1521233320236206,
      "learning_rate": 2e-05,
      "loss": 0.1915867567062378,
      "mean_token_accuracy": 0.945333406329155,
      "num_tokens": 27041160.0,
      "step": 1680
    },
    {
      "aux_loss": 8.144129180908203,
      "entropy": 0.156498297303915,
      "epoch": 0.001685,
      "grad_norm": 0.1905045360326767,
      "learning_rate": 2e-05,
      "loss": 0.19240328073501586,
      "mean_token_accuracy": 0.9464808136224747,
      "num_tokens": 27120205.0,
      "step": 1685
    },
    {
      "aux_loss": 8.077440357208252,
      "entropy": 0.16339176818728446,
      "epoch": 0.00169,
      "grad_norm": 0.15257541835308075,
      "learning_rate": 2e-05,
      "loss": 0.21015667915344238,
      "mean_token_accuracy": 0.9428196430206299,
      "num_tokens": 27201641.0,
      "step": 1690
    },
    {
      "aux_loss": 8.091274213790893,
      "entropy": 0.1816990975290537,
      "epoch": 0.001695,
      "grad_norm": 0.15236543118953705,
      "learning_rate": 2e-05,
      "loss": 0.21230292320251465,
      "mean_token_accuracy": 0.9404520213603973,
      "num_tokens": 27282918.0,
      "step": 1695
    },
    {
      "aux_loss": 8.102615594863892,
      "entropy": 0.1775689408183098,
      "epoch": 0.0017,
      "grad_norm": 0.14349766075611115,
      "learning_rate": 2e-05,
      "loss": 0.20254044532775878,
      "mean_token_accuracy": 0.9426239758729935,
      "num_tokens": 27363680.0,
      "step": 1700
    },
    {
      "aux_loss": 8.108851909637451,
      "entropy": 0.16214391365647315,
      "epoch": 0.001705,
      "grad_norm": 0.15616129338741302,
      "learning_rate": 2e-05,
      "loss": 0.1956779718399048,
      "mean_token_accuracy": 0.9449942111968994,
      "num_tokens": 27444588.0,
      "step": 1705
    },
    {
      "aux_loss": 8.122281932830811,
      "entropy": 0.16567001715302468,
      "epoch": 0.00171,
      "grad_norm": 0.15975213050842285,
      "learning_rate": 2e-05,
      "loss": 0.20318365097045898,
      "mean_token_accuracy": 0.9430002510547638,
      "num_tokens": 27524930.0,
      "step": 1710
    },
    {
      "aux_loss": 8.146432733535766,
      "entropy": 0.16198174953460692,
      "epoch": 0.001715,
      "grad_norm": 0.21213486790657043,
      "learning_rate": 2e-05,
      "loss": 0.19222015142440796,
      "mean_token_accuracy": 0.9461041837930679,
      "num_tokens": 27605483.0,
      "step": 1715
    },
    {
      "aux_loss": 8.08912272453308,
      "entropy": 0.16188026033341885,
      "epoch": 0.00172,
      "grad_norm": 0.1400371640920639,
      "learning_rate": 2e-05,
      "loss": 0.19799292087554932,
      "mean_token_accuracy": 0.9432717949151993,
      "num_tokens": 27682781.0,
      "step": 1720
    },
    {
      "aux_loss": 8.090563344955445,
      "entropy": 0.19338417798280716,
      "epoch": 0.001725,
      "grad_norm": 0.13796886801719666,
      "learning_rate": 2e-05,
      "loss": 0.22614567279815673,
      "mean_token_accuracy": 0.935646939277649,
      "num_tokens": 27763895.0,
      "step": 1725
    },
    {
      "aux_loss": 8.098656415939331,
      "entropy": 0.18461048305034639,
      "epoch": 0.00173,
      "grad_norm": 0.14468136429786682,
      "learning_rate": 2e-05,
      "loss": 0.213616943359375,
      "mean_token_accuracy": 0.9396229416131974,
      "num_tokens": 27844834.0,
      "step": 1730
    },
    {
      "aux_loss": 8.107629537582397,
      "entropy": 0.17966359108686447,
      "epoch": 0.001735,
      "grad_norm": 0.1427018791437149,
      "learning_rate": 2e-05,
      "loss": 0.21277756690979005,
      "mean_token_accuracy": 0.9404060482978821,
      "num_tokens": 27925652.0,
      "step": 1735
    },
    {
      "aux_loss": 8.12034010887146,
      "entropy": 0.17336866483092309,
      "epoch": 0.00174,
      "grad_norm": 0.1495121866464615,
      "learning_rate": 2e-05,
      "loss": 0.20275940895080566,
      "mean_token_accuracy": 0.9420665025711059,
      "num_tokens": 28005982.0,
      "step": 1740
    },
    {
      "aux_loss": 8.13797106742859,
      "entropy": 0.16741465479135514,
      "epoch": 0.001745,
      "grad_norm": 0.15425525605678558,
      "learning_rate": 2e-05,
      "loss": 0.19253748655319214,
      "mean_token_accuracy": 0.9453565120697022,
      "num_tokens": 28086419.0,
      "step": 1745
    },
    {
      "aux_loss": 8.09626030921936,
      "entropy": 0.15883845277130604,
      "epoch": 0.00175,
      "grad_norm": 0.14984235167503357,
      "learning_rate": 2e-05,
      "loss": 0.20106861591339112,
      "mean_token_accuracy": 0.9447483986616134,
      "num_tokens": 28166183.0,
      "step": 1750
    },
    {
      "aux_loss": 8.087971305847168,
      "entropy": 0.17699833139777182,
      "epoch": 0.001755,
      "grad_norm": 0.14571627974510193,
      "learning_rate": 2e-05,
      "loss": 0.22462019920349122,
      "mean_token_accuracy": 0.937322124838829,
      "num_tokens": 28247260.0,
      "step": 1755
    },
    {
      "aux_loss": 8.096086835861206,
      "entropy": 0.18670467361807824,
      "epoch": 0.00176,
      "grad_norm": 0.13775542378425598,
      "learning_rate": 2e-05,
      "loss": 0.21440637111663818,
      "mean_token_accuracy": 0.9395901650190354,
      "num_tokens": 28328086.0,
      "step": 1760
    },
    {
      "aux_loss": 8.105480766296386,
      "entropy": 0.18005181178450586,
      "epoch": 0.001765,
      "grad_norm": 0.13863572478294373,
      "learning_rate": 2e-05,
      "loss": 0.20243821144104004,
      "mean_token_accuracy": 0.9433570712804794,
      "num_tokens": 28409087.0,
      "step": 1765
    },
    {
      "aux_loss": 8.116328716278076,
      "entropy": 0.17669251263141633,
      "epoch": 0.00177,
      "grad_norm": 0.1435736119747162,
      "learning_rate": 2e-05,
      "loss": 0.2058091640472412,
      "mean_token_accuracy": 0.9422786742448807,
      "num_tokens": 28489547.0,
      "step": 1770
    },
    {
      "aux_loss": 8.138267469406127,
      "entropy": 0.16357129886746408,
      "epoch": 0.001775,
      "grad_norm": 0.1697431057691574,
      "learning_rate": 2e-05,
      "loss": 0.1961698889732361,
      "mean_token_accuracy": 0.9445254772901535,
      "num_tokens": 28569719.0,
      "step": 1775
    },
    {
      "aux_loss": 8.095318269729614,
      "entropy": 0.15334185361862182,
      "epoch": 0.00178,
      "grad_norm": 0.14353254437446594,
      "learning_rate": 2e-05,
      "loss": 0.19230631589889527,
      "mean_token_accuracy": 0.9462363421916962,
      "num_tokens": 28649541.0,
      "step": 1780
    },
    {
      "aux_loss": 8.084946012496948,
      "entropy": 0.18362632989883423,
      "epoch": 0.001785,
      "grad_norm": 0.13997416198253632,
      "learning_rate": 2e-05,
      "loss": 0.22572917938232423,
      "mean_token_accuracy": 0.9371131926774978,
      "num_tokens": 28730801.0,
      "step": 1785
    },
    {
      "aux_loss": 8.094417142868043,
      "entropy": 0.1935125134885311,
      "epoch": 0.00179,
      "grad_norm": 0.14033225178718567,
      "learning_rate": 2e-05,
      "loss": 0.22036070823669435,
      "mean_token_accuracy": 0.9385657787322998,
      "num_tokens": 28812114.0,
      "step": 1790
    },
    {
      "aux_loss": 8.10587649345398,
      "entropy": 0.17012702152132989,
      "epoch": 0.001795,
      "grad_norm": 0.1463741660118103,
      "learning_rate": 2e-05,
      "loss": 0.19968732595443725,
      "mean_token_accuracy": 0.9432043999433517,
      "num_tokens": 28893187.0,
      "step": 1795
    },
    {
      "aux_loss": 8.123018264770508,
      "entropy": 0.1629740722477436,
      "epoch": 0.0018,
      "grad_norm": 0.15700559318065643,
      "learning_rate": 2e-05,
      "loss": 0.19563982486724854,
      "mean_token_accuracy": 0.9459733098745347,
      "num_tokens": 28973708.0,
      "step": 1800
    },
    {
      "aux_loss": 8.141767835617065,
      "entropy": 0.1542024113237858,
      "epoch": 0.001805,
      "grad_norm": 0.1794741153717041,
      "learning_rate": 2e-05,
      "loss": 0.18798991441726684,
      "mean_token_accuracy": 0.9461712419986725,
      "num_tokens": 29054450.0,
      "step": 1805
    },
    {
      "aux_loss": 8.097171592712403,
      "entropy": 0.14869276769459247,
      "epoch": 0.00181,
      "grad_norm": 0.15480752289295197,
      "learning_rate": 2e-05,
      "loss": 0.18727688789367675,
      "mean_token_accuracy": 0.9484751909971237,
      "num_tokens": 29135641.0,
      "step": 1810
    },
    {
      "aux_loss": 8.085370969772338,
      "entropy": 0.18647595047950744,
      "epoch": 0.001815,
      "grad_norm": 0.14736691117286682,
      "learning_rate": 2e-05,
      "loss": 0.22024307250976563,
      "mean_token_accuracy": 0.9377060443162918,
      "num_tokens": 29216504.0,
      "step": 1815
    },
    {
      "aux_loss": 8.093893194198609,
      "entropy": 0.19387242645025254,
      "epoch": 0.00182,
      "grad_norm": 0.14280001819133759,
      "learning_rate": 2e-05,
      "loss": 0.22777719497680665,
      "mean_token_accuracy": 0.937128022313118,
      "num_tokens": 29297660.0,
      "step": 1820
    },
    {
      "aux_loss": 8.103721952438354,
      "entropy": 0.17211199477314948,
      "epoch": 0.001825,
      "grad_norm": 0.153641477227211,
      "learning_rate": 2e-05,
      "loss": 0.19921499490737915,
      "mean_token_accuracy": 0.9442539513111115,
      "num_tokens": 29378679.0,
      "step": 1825
    },
    {
      "aux_loss": 8.121135091781616,
      "entropy": 0.17612731233239173,
      "epoch": 0.00183,
      "grad_norm": 0.16757187247276306,
      "learning_rate": 2e-05,
      "loss": 0.20550386905670165,
      "mean_token_accuracy": 0.9418410420417785,
      "num_tokens": 29459520.0,
      "step": 1830
    },
    {
      "aux_loss": 8.134055376052856,
      "entropy": 0.15651994571089745,
      "epoch": 0.001835,
      "grad_norm": 0.16769260168075562,
      "learning_rate": 2e-05,
      "loss": 0.1892096519470215,
      "mean_token_accuracy": 0.9467572629451751,
      "num_tokens": 29539625.0,
      "step": 1835
    },
    {
      "aux_loss": 8.108575201034546,
      "entropy": 0.15799305215477943,
      "epoch": 0.00184,
      "grad_norm": 0.14953041076660156,
      "learning_rate": 2e-05,
      "loss": 0.20081510543823242,
      "mean_token_accuracy": 0.9454599142074585,
      "num_tokens": 29620351.0,
      "step": 1840
    },
    {
      "aux_loss": 8.090334177017212,
      "entropy": 0.18302324786782265,
      "epoch": 0.001845,
      "grad_norm": 0.14703971147537231,
      "learning_rate": 2e-05,
      "loss": 0.22612268924713136,
      "mean_token_accuracy": 0.9370312541723251,
      "num_tokens": 29701395.0,
      "step": 1845
    },
    {
      "aux_loss": 8.092384195327758,
      "entropy": 0.19149321541190148,
      "epoch": 0.00185,
      "grad_norm": 0.14011870324611664,
      "learning_rate": 2e-05,
      "loss": 0.21809930801391603,
      "mean_token_accuracy": 0.93871491253376,
      "num_tokens": 29782863.0,
      "step": 1850
    },
    {
      "aux_loss": 8.104417562484741,
      "entropy": 0.17967173233628272,
      "epoch": 0.001855,
      "grad_norm": 0.1392931342124939,
      "learning_rate": 2e-05,
      "loss": 0.2052596092224121,
      "mean_token_accuracy": 0.9424006640911102,
      "num_tokens": 29863853.0,
      "step": 1855
    },
    {
      "aux_loss": 8.115437746047974,
      "entropy": 0.17473116293549537,
      "epoch": 0.00186,
      "grad_norm": 0.15456846356391907,
      "learning_rate": 2e-05,
      "loss": 0.20883481502532958,
      "mean_token_accuracy": 0.9422131240367889,
      "num_tokens": 29944494.0,
      "step": 1860
    },
    {
      "aux_loss": 8.130228281021118,
      "entropy": 0.1598011538386345,
      "epoch": 0.001865,
      "grad_norm": 0.15138360857963562,
      "learning_rate": 2e-05,
      "loss": 0.19179458618164064,
      "mean_token_accuracy": 0.9459455698728562,
      "num_tokens": 30024871.0,
      "step": 1865
    },
    {
      "aux_loss": 8.114875078201294,
      "entropy": 0.15188466049730778,
      "epoch": 0.00187,
      "grad_norm": 0.1505664438009262,
      "learning_rate": 2e-05,
      "loss": 0.1842581272125244,
      "mean_token_accuracy": 0.9483438968658447,
      "num_tokens": 30103012.0,
      "step": 1870
    },
    {
      "aux_loss": 8.087232494354248,
      "entropy": 0.16997748129069806,
      "epoch": 0.001875,
      "grad_norm": 0.14826248586177826,
      "learning_rate": 2e-05,
      "loss": 0.2133951187133789,
      "mean_token_accuracy": 0.9409424215555191,
      "num_tokens": 30184174.0,
      "step": 1875
    },
    {
      "aux_loss": 8.091681003570557,
      "entropy": 0.18082854598760606,
      "epoch": 0.00188,
      "grad_norm": 0.1385442465543747,
      "learning_rate": 2e-05,
      "loss": 0.21204240322113038,
      "mean_token_accuracy": 0.9416619211435318,
      "num_tokens": 30264898.0,
      "step": 1880
    },
    {
      "aux_loss": 8.100720834732055,
      "entropy": 0.1882795512676239,
      "epoch": 0.001885,
      "grad_norm": 0.14320985972881317,
      "learning_rate": 2e-05,
      "loss": 0.21383564472198485,
      "mean_token_accuracy": 0.9398417532444,
      "num_tokens": 30345836.0,
      "step": 1885
    },
    {
      "aux_loss": 8.11526861190796,
      "entropy": 0.17213281691074372,
      "epoch": 0.00189,
      "grad_norm": 0.143971249461174,
      "learning_rate": 2e-05,
      "loss": 0.20359275341033936,
      "mean_token_accuracy": 0.94215886592865,
      "num_tokens": 30426658.0,
      "step": 1890
    },
    {
      "aux_loss": 8.128854751586914,
      "entropy": 0.15709721222519873,
      "epoch": 0.001895,
      "grad_norm": 0.169982448220253,
      "learning_rate": 2e-05,
      "loss": 0.18121217489242553,
      "mean_token_accuracy": 0.948867654800415,
      "num_tokens": 30506978.0,
      "step": 1895
    },
    {
      "aux_loss": 8.122401666641235,
      "entropy": 0.15198111161589622,
      "epoch": 0.0019,
      "grad_norm": 0.2494894117116928,
      "learning_rate": 2e-05,
      "loss": 0.18944740295410156,
      "mean_token_accuracy": 0.9484657168388366,
      "num_tokens": 30584707.0,
      "step": 1900
    },
    {
      "aux_loss": 8.08495569229126,
      "entropy": 0.16648186147212982,
      "epoch": 0.001905,
      "grad_norm": 0.14557261765003204,
      "learning_rate": 2e-05,
      "loss": 0.21277298927307128,
      "mean_token_accuracy": 0.9397358149290085,
      "num_tokens": 30665855.0,
      "step": 1905
    },
    {
      "aux_loss": 8.089331197738648,
      "entropy": 0.18394614458084108,
      "epoch": 0.00191,
      "grad_norm": 0.14418573677539825,
      "learning_rate": 2e-05,
      "loss": 0.22145624160766603,
      "mean_token_accuracy": 0.9384402185678482,
      "num_tokens": 30746741.0,
      "step": 1910
    },
    {
      "aux_loss": 8.099579572677612,
      "entropy": 0.18058552592992783,
      "epoch": 0.001915,
      "grad_norm": 0.14464372396469116,
      "learning_rate": 2e-05,
      "loss": 0.19679237604141236,
      "mean_token_accuracy": 0.9442365169525146,
      "num_tokens": 30827519.0,
      "step": 1915
    },
    {
      "aux_loss": 8.111036348342896,
      "entropy": 0.16745228171348572,
      "epoch": 0.00192,
      "grad_norm": 0.1453181803226471,
      "learning_rate": 2e-05,
      "loss": 0.19425774812698365,
      "mean_token_accuracy": 0.9451937824487686,
      "num_tokens": 30908391.0,
      "step": 1920
    },
    {
      "aux_loss": 8.12428812980652,
      "entropy": 0.15277336463332175,
      "epoch": 0.001925,
      "grad_norm": 0.17287026345729828,
      "learning_rate": 2e-05,
      "loss": 0.19231228828430175,
      "mean_token_accuracy": 0.9467857658863068,
      "num_tokens": 30988775.0,
      "step": 1925
    },
    {
      "aux_loss": 8.128771543502808,
      "entropy": 0.14848360568284988,
      "epoch": 0.00193,
      "grad_norm": 0.12818454205989838,
      "learning_rate": 2e-05,
      "loss": 0.17717504501342773,
      "mean_token_accuracy": 0.9506623685359955,
      "num_tokens": 31066517.0,
      "step": 1930
    },
    {
      "aux_loss": 8.07816481590271,
      "entropy": 0.17916654124855996,
      "epoch": 0.001935,
      "grad_norm": 0.1318196952342987,
      "learning_rate": 2e-05,
      "loss": 0.22017006874084472,
      "mean_token_accuracy": 0.9393600344657898,
      "num_tokens": 31148198.0,
      "step": 1935
    },
    {
      "aux_loss": 8.085155296325684,
      "entropy": 0.18373383060097695,
      "epoch": 0.00194,
      "grad_norm": 0.13922302424907684,
      "learning_rate": 2e-05,
      "loss": 0.2204740285873413,
      "mean_token_accuracy": 0.9385798037052154,
      "num_tokens": 31228701.0,
      "step": 1940
    },
    {
      "aux_loss": 8.093752336502074,
      "entropy": 0.18153189793229102,
      "epoch": 0.001945,
      "grad_norm": 0.14095920324325562,
      "learning_rate": 2e-05,
      "loss": 0.20607366561889648,
      "mean_token_accuracy": 0.9425692021846771,
      "num_tokens": 31309643.0,
      "step": 1945
    },
    {
      "aux_loss": 8.109092903137206,
      "entropy": 0.17216213867068292,
      "epoch": 0.00195,
      "grad_norm": 0.1501334309577942,
      "learning_rate": 2e-05,
      "loss": 0.20081067085266113,
      "mean_token_accuracy": 0.9431844651699066,
      "num_tokens": 31390333.0,
      "step": 1950
    },
    {
      "aux_loss": 8.121693420410157,
      "entropy": 0.16054480373859406,
      "epoch": 0.001955,
      "grad_norm": 0.14425206184387207,
      "learning_rate": 2e-05,
      "loss": 0.19371211528778076,
      "mean_token_accuracy": 0.9458245187997818,
      "num_tokens": 31471108.0,
      "step": 1955
    },
    {
      "aux_loss": 8.131749486923217,
      "entropy": 0.1476591568440199,
      "epoch": 0.00196,
      "grad_norm": 0.13990935683250427,
      "learning_rate": 2e-05,
      "loss": 0.18084168434143066,
      "mean_token_accuracy": 0.9503275483846665,
      "num_tokens": 31551156.0,
      "step": 1960
    },
    {
      "aux_loss": 8.079078769683838,
      "entropy": 0.15950408726930618,
      "epoch": 0.001965,
      "grad_norm": 0.1457347720861435,
      "learning_rate": 2e-05,
      "loss": 0.20008344650268556,
      "mean_token_accuracy": 0.9451485365629196,
      "num_tokens": 31632531.0,
      "step": 1965
    },
    {
      "aux_loss": 8.090669679641724,
      "entropy": 0.17626864910125734,
      "epoch": 0.00197,
      "grad_norm": 0.14180968701839447,
      "learning_rate": 2e-05,
      "loss": 0.21406903266906738,
      "mean_token_accuracy": 0.9399964392185212,
      "num_tokens": 31713571.0,
      "step": 1970
    },
    {
      "aux_loss": 8.094075250625611,
      "entropy": 0.17952204570174218,
      "epoch": 0.001975,
      "grad_norm": 0.14584553241729736,
      "learning_rate": 2e-05,
      "loss": 0.2037555456161499,
      "mean_token_accuracy": 0.9423517286777496,
      "num_tokens": 31794181.0,
      "step": 1975
    },
    {
      "aux_loss": 8.108042097091674,
      "entropy": 0.17397107556462288,
      "epoch": 0.00198,
      "grad_norm": 0.13913865387439728,
      "learning_rate": 2e-05,
      "loss": 0.20009984970092773,
      "mean_token_accuracy": 0.9435578137636185,
      "num_tokens": 31874688.0,
      "step": 1980
    },
    {
      "aux_loss": 8.124657011032104,
      "entropy": 0.16394528411328793,
      "epoch": 0.001985,
      "grad_norm": 0.16476790606975555,
      "learning_rate": 2e-05,
      "loss": 0.19735250473022461,
      "mean_token_accuracy": 0.9441554218530654,
      "num_tokens": 31955614.0,
      "step": 1985
    },
    {
      "aux_loss": 8.139951753616334,
      "entropy": 0.16006705798208715,
      "epoch": 0.00199,
      "grad_norm": 0.16014958918094635,
      "learning_rate": 2e-05,
      "loss": 0.193767249584198,
      "mean_token_accuracy": 0.9465360701084137,
      "num_tokens": 32033705.0,
      "step": 1990
    },
    {
      "aux_loss": 8.076089763641358,
      "entropy": 0.16580818742513656,
      "epoch": 0.001995,
      "grad_norm": 0.14132636785507202,
      "learning_rate": 2e-05,
      "loss": 0.20737054347991943,
      "mean_token_accuracy": 0.9432087421417237,
      "num_tokens": 32115159.0,
      "step": 1995
    },
    {
      "aux_loss": 8.089388179779053,
      "entropy": 0.17119593992829324,
      "epoch": 0.002,
      "grad_norm": 0.13946352899074554,
      "learning_rate": 2e-05,
      "loss": 0.2031093120574951,
      "mean_token_accuracy": 0.9424938827753067,
      "num_tokens": 32195875.0,
      "step": 2000
    },
    {
      "aux_loss": 8.099317407608032,
      "entropy": 0.17812084034085274,
      "epoch": 0.002005,
      "grad_norm": 0.13404801487922668,
      "learning_rate": 2e-05,
      "loss": 0.2054826498031616,
      "mean_token_accuracy": 0.9427031606435776,
      "num_tokens": 32276639.0,
      "step": 2005
    },
    {
      "aux_loss": 8.11078701019287,
      "entropy": 0.16636469662189485,
      "epoch": 0.00201,
      "grad_norm": 0.1405758410692215,
      "learning_rate": 2e-05,
      "loss": 0.1961885690689087,
      "mean_token_accuracy": 0.9454802840948104,
      "num_tokens": 32357510.0,
      "step": 2010
    },
    {
      "aux_loss": 8.124329996109008,
      "entropy": 0.16409692093729972,
      "epoch": 0.002015,
      "grad_norm": 0.1530579924583435,
      "learning_rate": 2e-05,
      "loss": 0.19761037826538086,
      "mean_token_accuracy": 0.9441994547843933,
      "num_tokens": 32438571.0,
      "step": 2015
    },
    {
      "aux_loss": 8.134271717071533,
      "entropy": 0.15539358295500277,
      "epoch": 0.00202,
      "grad_norm": 0.1314486712217331,
      "learning_rate": 2e-05,
      "loss": 0.18882243633270263,
      "mean_token_accuracy": 0.9479888468980789,
      "num_tokens": 32516933.0,
      "step": 2020
    },
    {
      "aux_loss": 8.076760530471802,
      "entropy": 0.160697041451931,
      "epoch": 0.002025,
      "grad_norm": 0.1575176864862442,
      "learning_rate": 2e-05,
      "loss": 0.20129084587097168,
      "mean_token_accuracy": 0.9434007734060288,
      "num_tokens": 32598260.0,
      "step": 2025
    },
    {
      "aux_loss": 8.088034677505494,
      "entropy": 0.18619557917118074,
      "epoch": 0.00203,
      "grad_norm": 0.1533048450946808,
      "learning_rate": 2e-05,
      "loss": 0.22235991954803466,
      "mean_token_accuracy": 0.9385304391384125,
      "num_tokens": 32679421.0,
      "step": 2030
    },
    {
      "aux_loss": 8.09389386177063,
      "entropy": 0.1984347604215145,
      "epoch": 0.002035,
      "grad_norm": 0.13730265200138092,
      "learning_rate": 2e-05,
      "loss": 0.22851157188415527,
      "mean_token_accuracy": 0.9353628069162369,
      "num_tokens": 32760105.0,
      "step": 2035
    },
    {
      "aux_loss": 8.10667643547058,
      "entropy": 0.1851852796971798,
      "epoch": 0.00204,
      "grad_norm": 0.136952742934227,
      "learning_rate": 2e-05,
      "loss": 0.20700442790985107,
      "mean_token_accuracy": 0.9416122794151306,
      "num_tokens": 32840448.0,
      "step": 2040
    },
    {
      "aux_loss": 8.119586324691772,
      "entropy": 0.1611733615398407,
      "epoch": 0.002045,
      "grad_norm": 0.15282323956489563,
      "learning_rate": 2e-05,
      "loss": 0.19286082983016967,
      "mean_token_accuracy": 0.9459655225276947,
      "num_tokens": 32920963.0,
      "step": 2045
    },
    {
      "aux_loss": 8.142139720916749,
      "entropy": 0.1511260338127613,
      "epoch": 0.00205,
      "grad_norm": 0.20718199014663696,
      "learning_rate": 2e-05,
      "loss": 0.18599768877029418,
      "mean_token_accuracy": 0.9473131716251373,
      "num_tokens": 32999796.0,
      "step": 2050
    },
    {
      "aux_loss": 8.073504304885864,
      "entropy": 0.15352291576564311,
      "epoch": 0.002055,
      "grad_norm": 0.1586863100528717,
      "learning_rate": 2e-05,
      "loss": 0.20044825077056885,
      "mean_token_accuracy": 0.9450691044330597,
      "num_tokens": 33081090.0,
      "step": 2055
    },
    {
      "aux_loss": 8.085515308380128,
      "entropy": 0.19169245585799216,
      "epoch": 0.00206,
      "grad_norm": 0.14856073260307312,
      "learning_rate": 2e-05,
      "loss": 0.2272887945175171,
      "mean_token_accuracy": 0.9367268264293671,
      "num_tokens": 33162496.0,
      "step": 2060
    },
    {
      "aux_loss": 8.090879011154176,
      "entropy": 0.1914783239364624,
      "epoch": 0.002065,
      "grad_norm": 0.13699573278427124,
      "learning_rate": 2e-05,
      "loss": 0.21631879806518556,
      "mean_token_accuracy": 0.9397626191377639,
      "num_tokens": 33243363.0,
      "step": 2065
    },
    {
      "aux_loss": 8.102555894851685,
      "entropy": 0.17537787184119225,
      "epoch": 0.00207,
      "grad_norm": 0.14345507323741913,
      "learning_rate": 2e-05,
      "loss": 0.20378334522247316,
      "mean_token_accuracy": 0.9429430037736892,
      "num_tokens": 33324009.0,
      "step": 2070
    },
    {
      "aux_loss": 8.116124439239503,
      "entropy": 0.16216097101569177,
      "epoch": 0.002075,
      "grad_norm": 0.15001976490020752,
      "learning_rate": 2e-05,
      "loss": 0.1955229640007019,
      "mean_token_accuracy": 0.945282730460167,
      "num_tokens": 33404354.0,
      "step": 2075
    },
    {
      "aux_loss": 8.13131160736084,
      "entropy": 0.15755471885204314,
      "epoch": 0.00208,
      "grad_norm": 0.1630067080259323,
      "learning_rate": 2e-05,
      "loss": 0.18815797567367554,
      "mean_token_accuracy": 0.9474914819002151,
      "num_tokens": 33485370.0,
      "step": 2080
    },
    {
      "aux_loss": 8.085240173339844,
      "entropy": 0.1570974349975586,
      "epoch": 0.002085,
      "grad_norm": 0.16560609638690948,
      "learning_rate": 2e-05,
      "loss": 0.19717142581939698,
      "mean_token_accuracy": 0.9450071066617965,
      "num_tokens": 33563829.0,
      "step": 2085
    },
    {
      "aux_loss": 8.079175758361817,
      "entropy": 0.1813222225755453,
      "epoch": 0.00209,
      "grad_norm": 0.15199394524097443,
      "learning_rate": 2e-05,
      "loss": 0.22196767330169678,
      "mean_token_accuracy": 0.9381584227085114,
      "num_tokens": 33645265.0,
      "step": 2090
    },
    {
      "aux_loss": 8.089537858963013,
      "entropy": 0.1975115083158016,
      "epoch": 0.002095,
      "grad_norm": 0.14169766008853912,
      "learning_rate": 2e-05,
      "loss": 0.2236358642578125,
      "mean_token_accuracy": 0.9374071359634399,
      "num_tokens": 33726036.0,
      "step": 2095
    },
    {
      "aux_loss": 8.101109313964844,
      "entropy": 0.1791003830730915,
      "epoch": 0.0021,
      "grad_norm": 0.1525634527206421,
      "learning_rate": 2e-05,
      "loss": 0.20314218997955322,
      "mean_token_accuracy": 0.9433953523635864,
      "num_tokens": 33806473.0,
      "step": 2100
    },
    {
      "aux_loss": 8.115503787994385,
      "entropy": 0.15914352983236313,
      "epoch": 0.002105,
      "grad_norm": 0.15232878923416138,
      "learning_rate": 2e-05,
      "loss": 0.1909801483154297,
      "mean_token_accuracy": 0.9463894724845886,
      "num_tokens": 33886645.0,
      "step": 2105
    },
    {
      "aux_loss": 8.136928987503051,
      "entropy": 0.1533144772052765,
      "epoch": 0.00211,
      "grad_norm": 0.16553930938243866,
      "learning_rate": 2e-05,
      "loss": 0.187789249420166,
      "mean_token_accuracy": 0.9473093211650848,
      "num_tokens": 33967099.0,
      "step": 2110
    },
    {
      "aux_loss": 8.085592317581177,
      "entropy": 0.15226773768663407,
      "epoch": 0.002115,
      "grad_norm": 0.1545340120792389,
      "learning_rate": 2e-05,
      "loss": 0.19702117443084716,
      "mean_token_accuracy": 0.9462773114442825,
      "num_tokens": 34047867.0,
      "step": 2115
    },
    {
      "aux_loss": 8.084630918502807,
      "entropy": 0.17928903549909592,
      "epoch": 0.00212,
      "grad_norm": 0.14029167592525482,
      "learning_rate": 2e-05,
      "loss": 0.21474015712738037,
      "mean_token_accuracy": 0.9403113484382629,
      "num_tokens": 34129348.0,
      "step": 2120
    },
    {
      "aux_loss": 8.092728471755981,
      "entropy": 0.18752039447426797,
      "epoch": 0.002125,
      "grad_norm": 0.16090719401836395,
      "learning_rate": 2e-05,
      "loss": 0.21388823986053468,
      "mean_token_accuracy": 0.9389720141887665,
      "num_tokens": 34209888.0,
      "step": 2125
    },
    {
      "aux_loss": 8.105613899230956,
      "entropy": 0.17475795224308968,
      "epoch": 0.00213,
      "grad_norm": 0.15447764098644257,
      "learning_rate": 2e-05,
      "loss": 0.20451686382293702,
      "mean_token_accuracy": 0.9426469892263413,
      "num_tokens": 34290780.0,
      "step": 2130
    },
    {
      "aux_loss": 8.117050743103027,
      "entropy": 0.1618360310792923,
      "epoch": 0.002135,
      "grad_norm": 0.1479986011981964,
      "learning_rate": 2e-05,
      "loss": 0.1921260714530945,
      "mean_token_accuracy": 0.9462061405181885,
      "num_tokens": 34371120.0,
      "step": 2135
    },
    {
      "aux_loss": 8.137471532821655,
      "entropy": 0.15500931218266487,
      "epoch": 0.00214,
      "grad_norm": 0.1681438386440277,
      "learning_rate": 2e-05,
      "loss": 0.18469098806381226,
      "mean_token_accuracy": 0.9486718863248825,
      "num_tokens": 34451650.0,
      "step": 2140
    },
    {
      "aux_loss": 8.082095575332641,
      "entropy": 0.14619569703936577,
      "epoch": 0.002145,
      "grad_norm": 0.15538665652275085,
      "learning_rate": 2e-05,
      "loss": 0.19378514289855958,
      "mean_token_accuracy": 0.9471194446086884,
      "num_tokens": 34532208.0,
      "step": 2145
    },
    {
      "aux_loss": 8.085667753219605,
      "entropy": 0.17372070737183093,
      "epoch": 0.00215,
      "grad_norm": 0.14733856916427612,
      "learning_rate": 2e-05,
      "loss": 0.21246514320373536,
      "mean_token_accuracy": 0.9406029522418976,
      "num_tokens": 34613186.0,
      "step": 2150
    },
    {
      "aux_loss": 8.090953254699707,
      "entropy": 0.18648151084780692,
      "epoch": 0.002155,
      "grad_norm": 0.13900934159755707,
      "learning_rate": 2e-05,
      "loss": 0.21289048194885254,
      "mean_token_accuracy": 0.9395630806684494,
      "num_tokens": 34694000.0,
      "step": 2155
    },
    {
      "aux_loss": 8.099458074569702,
      "entropy": 0.179243303835392,
      "epoch": 0.00216,
      "grad_norm": 0.1422865390777588,
      "learning_rate": 2e-05,
      "loss": 0.20033249855041504,
      "mean_token_accuracy": 0.9438653618097306,
      "num_tokens": 34774874.0,
      "step": 2160
    },
    {
      "aux_loss": 8.11012306213379,
      "entropy": 0.16056850105524062,
      "epoch": 0.002165,
      "grad_norm": 0.1500975340604782,
      "learning_rate": 2e-05,
      "loss": 0.1950578808784485,
      "mean_token_accuracy": 0.9458033829927445,
      "num_tokens": 34855714.0,
      "step": 2165
    },
    {
      "aux_loss": 8.128828239440917,
      "entropy": 0.15131179094314576,
      "epoch": 0.00217,
      "grad_norm": 0.17130112648010254,
      "learning_rate": 2e-05,
      "loss": 0.1841041088104248,
      "mean_token_accuracy": 0.9479575306177139,
      "num_tokens": 34936115.0,
      "step": 2170
    },
    {
      "aux_loss": 8.096973991394043,
      "entropy": 0.1353026580065489,
      "epoch": 0.002175,
      "grad_norm": 0.14092183113098145,
      "learning_rate": 2e-05,
      "loss": 0.1694318413734436,
      "mean_token_accuracy": 0.953260064125061,
      "num_tokens": 35016252.0,
      "step": 2175
    },
    {
      "aux_loss": 8.082904958724976,
      "entropy": 0.18240192383527756,
      "epoch": 0.00218,
      "grad_norm": 0.14787666499614716,
      "learning_rate": 2e-05,
      "loss": 0.2253516674041748,
      "mean_token_accuracy": 0.9372398585081101,
      "num_tokens": 35097431.0,
      "step": 2180
    },
    {
      "aux_loss": 8.090396499633789,
      "entropy": 0.188712240755558,
      "epoch": 0.002185,
      "grad_norm": 0.17046400904655457,
      "learning_rate": 2e-05,
      "loss": 0.2209940195083618,
      "mean_token_accuracy": 0.9375521659851074,
      "num_tokens": 35178778.0,
      "step": 2185
    },
    {
      "aux_loss": 8.098586225509644,
      "entropy": 0.17917247265577316,
      "epoch": 0.00219,
      "grad_norm": 0.13856396079063416,
      "learning_rate": 2e-05,
      "loss": 0.20676376819610595,
      "mean_token_accuracy": 0.9419531643390655,
      "num_tokens": 35259574.0,
      "step": 2190
    },
    {
      "aux_loss": 8.111574029922485,
      "entropy": 0.17405206337571144,
      "epoch": 0.002195,
      "grad_norm": 0.14484310150146484,
      "learning_rate": 2e-05,
      "loss": 0.20087049007415772,
      "mean_token_accuracy": 0.9425794541835785,
      "num_tokens": 35340171.0,
      "step": 2195
    },
    {
      "aux_loss": 8.12577714920044,
      "entropy": 0.15521480180323124,
      "epoch": 0.0022,
      "grad_norm": 0.18309779465198517,
      "learning_rate": 2e-05,
      "loss": 0.1860737681388855,
      "mean_token_accuracy": 0.9473440408706665,
      "num_tokens": 35420499.0,
      "step": 2200
    },
    {
      "aux_loss": 8.106541919708253,
      "entropy": 0.15410314984619616,
      "epoch": 0.002205,
      "grad_norm": 0.16265305876731873,
      "learning_rate": 2e-05,
      "loss": 0.19503383636474608,
      "mean_token_accuracy": 0.9455761820077896,
      "num_tokens": 35499651.0,
      "step": 2205
    },
    {
      "aux_loss": 8.081146049499512,
      "entropy": 0.17743488028645515,
      "epoch": 0.00221,
      "grad_norm": 0.15264394879341125,
      "learning_rate": 2e-05,
      "loss": 0.22179040908813477,
      "mean_token_accuracy": 0.9375330954790115,
      "num_tokens": 35581141.0,
      "step": 2210
    },
    {
      "aux_loss": 8.089364290237427,
      "entropy": 0.19082226306200029,
      "epoch": 0.002215,
      "grad_norm": 0.13382941484451294,
      "learning_rate": 2e-05,
      "loss": 0.22009677886962892,
      "mean_token_accuracy": 0.9386865109205246,
      "num_tokens": 35661933.0,
      "step": 2215
    },
    {
      "aux_loss": 8.09914779663086,
      "entropy": 0.18173426166176795,
      "epoch": 0.00222,
      "grad_norm": 0.13940785825252533,
      "learning_rate": 2e-05,
      "loss": 0.20357420444488525,
      "mean_token_accuracy": 0.9428389191627502,
      "num_tokens": 35742852.0,
      "step": 2220
    },
    {
      "aux_loss": 8.109477663040161,
      "entropy": 0.16557887718081474,
      "epoch": 0.002225,
      "grad_norm": 0.1517818123102188,
      "learning_rate": 2e-05,
      "loss": 0.19582066535949708,
      "mean_token_accuracy": 0.9441197693347931,
      "num_tokens": 35823590.0,
      "step": 2225
    },
    {
      "aux_loss": 8.124861478805542,
      "entropy": 0.15990642234683036,
      "epoch": 0.00223,
      "grad_norm": 0.16026835143566132,
      "learning_rate": 2e-05,
      "loss": 0.19691332578659057,
      "mean_token_accuracy": 0.9450530737638474,
      "num_tokens": 35903659.0,
      "step": 2230
    },
    {
      "aux_loss": 8.110028553009034,
      "entropy": 0.14880426675081254,
      "epoch": 0.002235,
      "grad_norm": 0.14345495402812958,
      "learning_rate": 2e-05,
      "loss": 0.17638733386993408,
      "mean_token_accuracy": 0.951531520485878,
      "num_tokens": 35980742.0,
      "step": 2235
    },
    {
      "aux_loss": 8.076312446594239,
      "entropy": 0.1716868907213211,
      "epoch": 0.00224,
      "grad_norm": 0.15603093802928925,
      "learning_rate": 2e-05,
      "loss": 0.21976280212402344,
      "mean_token_accuracy": 0.939437958598137,
      "num_tokens": 36061594.0,
      "step": 2240
    },
    {
      "aux_loss": 8.084784698486327,
      "entropy": 0.17795654386281967,
      "epoch": 0.002245,
      "grad_norm": 0.13866472244262695,
      "learning_rate": 2e-05,
      "loss": 0.20380654335021972,
      "mean_token_accuracy": 0.9432132840156555,
      "num_tokens": 36142538.0,
      "step": 2245
    },
    {
      "aux_loss": 8.096397113800048,
      "entropy": 0.1781972512602806,
      "epoch": 0.00225,
      "grad_norm": 0.1484403908252716,
      "learning_rate": 2e-05,
      "loss": 0.20842454433441163,
      "mean_token_accuracy": 0.9416657865047455,
      "num_tokens": 36223141.0,
      "step": 2250
    },
    {
      "aux_loss": 8.106841039657592,
      "entropy": 0.17267693802714348,
      "epoch": 0.002255,
      "grad_norm": 0.1455325186252594,
      "learning_rate": 2e-05,
      "loss": 0.20392305850982667,
      "mean_token_accuracy": 0.94244185090065,
      "num_tokens": 36303819.0,
      "step": 2255
    },
    {
      "aux_loss": 8.123301219940185,
      "entropy": 0.1638832077383995,
      "epoch": 0.00226,
      "grad_norm": 0.15240441262722015,
      "learning_rate": 2e-05,
      "loss": 0.19455764293670655,
      "mean_token_accuracy": 0.9457614630460739,
      "num_tokens": 36383712.0,
      "step": 2260
    },
    {
      "aux_loss": 8.109774494171143,
      "entropy": 0.15629686638712884,
      "epoch": 0.002265,
      "grad_norm": 0.14580827951431274,
      "learning_rate": 2e-05,
      "loss": 0.19248292446136475,
      "mean_token_accuracy": 0.9473178863525391,
      "num_tokens": 36462413.0,
      "step": 2265
    },
    {
      "aux_loss": 8.080343246459961,
      "entropy": 0.17776158899068834,
      "epoch": 0.00227,
      "grad_norm": 0.15754243731498718,
      "learning_rate": 2e-05,
      "loss": 0.21568129062652588,
      "mean_token_accuracy": 0.9400018811225891,
      "num_tokens": 36543608.0,
      "step": 2270
    },
    {
      "aux_loss": 8.084667730331422,
      "entropy": 0.18670802041888238,
      "epoch": 0.002275,
      "grad_norm": 0.13676391541957855,
      "learning_rate": 2e-05,
      "loss": 0.21331896781921386,
      "mean_token_accuracy": 0.9408599764108658,
      "num_tokens": 36624156.0,
      "step": 2275
    },
    {
      "aux_loss": 8.094764137268067,
      "entropy": 0.1731523796916008,
      "epoch": 0.00228,
      "grad_norm": 0.16389347612857819,
      "learning_rate": 2e-05,
      "loss": 0.20079305171966552,
      "mean_token_accuracy": 0.9436883717775345,
      "num_tokens": 36705126.0,
      "step": 2280
    },
    {
      "aux_loss": 8.10275354385376,
      "entropy": 0.16792446225881577,
      "epoch": 0.002285,
      "grad_norm": 0.14619581401348114,
      "learning_rate": 2e-05,
      "loss": 0.19990408420562744,
      "mean_token_accuracy": 0.9430625677108765,
      "num_tokens": 36786106.0,
      "step": 2285
    },
    {
      "aux_loss": 8.118331623077392,
      "entropy": 0.16288357749581336,
      "epoch": 0.00229,
      "grad_norm": 0.16086938977241516,
      "learning_rate": 2e-05,
      "loss": 0.19735727310180665,
      "mean_token_accuracy": 0.9458436697721482,
      "num_tokens": 36867047.0,
      "step": 2290
    },
    {
      "aux_loss": 8.118902730941773,
      "entropy": 0.15250914059579374,
      "epoch": 0.002295,
      "grad_norm": 0.14330445230007172,
      "learning_rate": 2e-05,
      "loss": 0.18190444707870485,
      "mean_token_accuracy": 0.9490704566240311,
      "num_tokens": 36944943.0,
      "step": 2295
    },
    {
      "aux_loss": 8.075653839111329,
      "entropy": 0.1593605875968933,
      "epoch": 0.0023,
      "grad_norm": 0.13995373249053955,
      "learning_rate": 2e-05,
      "loss": 0.2039646625518799,
      "mean_token_accuracy": 0.9433200567960739,
      "num_tokens": 37025984.0,
      "step": 2300
    },
    {
      "aux_loss": 8.087512254714966,
      "entropy": 0.1860686607658863,
      "epoch": 0.002305,
      "grad_norm": 0.13759204745292664,
      "learning_rate": 2e-05,
      "loss": 0.22247304916381835,
      "mean_token_accuracy": 0.9367720305919647,
      "num_tokens": 37106861.0,
      "step": 2305
    },
    {
      "aux_loss": 8.095287132263184,
      "entropy": 0.18836626335978507,
      "epoch": 0.00231,
      "grad_norm": 0.15044982731342316,
      "learning_rate": 2e-05,
      "loss": 0.21037421226501465,
      "mean_token_accuracy": 0.9411572962999344,
      "num_tokens": 37187899.0,
      "step": 2310
    },
    {
      "aux_loss": 8.108015441894532,
      "entropy": 0.18079396188259125,
      "epoch": 0.002315,
      "grad_norm": 0.14435166120529175,
      "learning_rate": 2e-05,
      "loss": 0.21002936363220215,
      "mean_token_accuracy": 0.9402021259069443,
      "num_tokens": 37268517.0,
      "step": 2315
    },
    {
      "aux_loss": 8.118687677383424,
      "entropy": 0.1508309967815876,
      "epoch": 0.00232,
      "grad_norm": 0.1588558554649353,
      "learning_rate": 2e-05,
      "loss": 0.18262405395507814,
      "mean_token_accuracy": 0.9487547636032104,
      "num_tokens": 37348643.0,
      "step": 2320
    },
    {
      "aux_loss": 8.11876459121704,
      "entropy": 0.15192754790186883,
      "epoch": 0.002325,
      "grad_norm": 0.14980925619602203,
      "learning_rate": 2e-05,
      "loss": 0.19291154146194459,
      "mean_token_accuracy": 0.9454974114894867,
      "num_tokens": 37426287.0,
      "step": 2325
    },
    {
      "aux_loss": 8.072935342788696,
      "entropy": 0.15959618985652924,
      "epoch": 0.00233,
      "grad_norm": 0.14502452313899994,
      "learning_rate": 2e-05,
      "loss": 0.20303568840026856,
      "mean_token_accuracy": 0.9447141468524933,
      "num_tokens": 37507440.0,
      "step": 2330
    },
    {
      "aux_loss": 8.083228874206544,
      "entropy": 0.18669020012021065,
      "epoch": 0.002335,
      "grad_norm": 0.14419959485530853,
      "learning_rate": 2e-05,
      "loss": 0.2143770456314087,
      "mean_token_accuracy": 0.939791139960289,
      "num_tokens": 37588258.0,
      "step": 2335
    },
    {
      "aux_loss": 8.092568063735962,
      "entropy": 0.18322967886924743,
      "epoch": 0.00234,
      "grad_norm": 0.1554855853319168,
      "learning_rate": 2e-05,
      "loss": 0.2063164472579956,
      "mean_token_accuracy": 0.9414166659116745,
      "num_tokens": 37668728.0,
      "step": 2340
    },
    {
      "aux_loss": 8.102509689331054,
      "entropy": 0.16345815286040305,
      "epoch": 0.002345,
      "grad_norm": 0.1634492725133896,
      "learning_rate": 2e-05,
      "loss": 0.20011322498321532,
      "mean_token_accuracy": 0.9438972026109695,
      "num_tokens": 37748838.0,
      "step": 2345
    },
    {
      "aux_loss": 8.11418375968933,
      "entropy": 0.16603684350848197,
      "epoch": 0.00235,
      "grad_norm": 0.14245444536209106,
      "learning_rate": 2e-05,
      "loss": 0.1968133568763733,
      "mean_token_accuracy": 0.9455876886844635,
      "num_tokens": 37829396.0,
      "step": 2350
    },
    {
      "aux_loss": 8.131700086593629,
      "entropy": 0.15448009595274925,
      "epoch": 0.002355,
      "grad_norm": 0.1676913946866989,
      "learning_rate": 2e-05,
      "loss": 0.1844778060913086,
      "mean_token_accuracy": 0.9480048835277557,
      "num_tokens": 37906525.0,
      "step": 2355
    },
    {
      "aux_loss": 8.073725318908691,
      "entropy": 0.16473549902439116,
      "epoch": 0.00236,
      "grad_norm": 0.14581012725830078,
      "learning_rate": 2e-05,
      "loss": 0.2133479118347168,
      "mean_token_accuracy": 0.9407844752073288,
      "num_tokens": 37987845.0,
      "step": 2360
    },
    {
      "aux_loss": 8.079457998275757,
      "entropy": 0.18857136219739914,
      "epoch": 0.002365,
      "grad_norm": 0.14681008458137512,
      "learning_rate": 2e-05,
      "loss": 0.22654881477355956,
      "mean_token_accuracy": 0.93775754570961,
      "num_tokens": 38068960.0,
      "step": 2365
    },
    {
      "aux_loss": 8.087099742889404,
      "entropy": 0.17707170620560647,
      "epoch": 0.00237,
      "grad_norm": 0.1334906816482544,
      "learning_rate": 2e-05,
      "loss": 0.20186762809753417,
      "mean_token_accuracy": 0.9438526868820191,
      "num_tokens": 38149738.0,
      "step": 2370
    },
    {
      "aux_loss": 8.09603304862976,
      "entropy": 0.1817078672349453,
      "epoch": 0.002375,
      "grad_norm": 0.15234078466892242,
      "learning_rate": 2e-05,
      "loss": 0.20077755451202392,
      "mean_token_accuracy": 0.9430094093084336,
      "num_tokens": 38229880.0,
      "step": 2375
    },
    {
      "aux_loss": 8.11332926750183,
      "entropy": 0.15103389993309974,
      "epoch": 0.00238,
      "grad_norm": 0.16679836809635162,
      "learning_rate": 2e-05,
      "loss": 0.1859709620475769,
      "mean_token_accuracy": 0.9472784757614136,
      "num_tokens": 38310817.0,
      "step": 2380
    },
    {
      "aux_loss": 8.128608703613281,
      "entropy": 0.14925035387277602,
      "epoch": 0.002385,
      "grad_norm": 0.14778488874435425,
      "learning_rate": 2e-05,
      "loss": 0.18576887845993043,
      "mean_token_accuracy": 0.9485006332397461,
      "num_tokens": 38391282.0,
      "step": 2385
    },
    {
      "aux_loss": 8.078663730621338,
      "entropy": 0.1730654701590538,
      "epoch": 0.00239,
      "grad_norm": 0.15327969193458557,
      "learning_rate": 2e-05,
      "loss": 0.21485276222229005,
      "mean_token_accuracy": 0.9404427438974381,
      "num_tokens": 38472686.0,
      "step": 2390
    },
    {
      "aux_loss": 8.07740821838379,
      "entropy": 0.19329795911908149,
      "epoch": 0.002395,
      "grad_norm": 0.145245760679245,
      "learning_rate": 2e-05,
      "loss": 0.23386235237121583,
      "mean_token_accuracy": 0.9349213540554047,
      "num_tokens": 38553848.0,
      "step": 2395
    },
    {
      "aux_loss": 8.083292770385743,
      "entropy": 0.19434554129838943,
      "epoch": 0.0024,
      "grad_norm": 0.13770611584186554,
      "learning_rate": 2e-05,
      "loss": 0.22028076648712158,
      "mean_token_accuracy": 0.9392184466123581,
      "num_tokens": 38634744.0,
      "step": 2400
    },
    {
      "aux_loss": 8.097694253921508,
      "entropy": 0.18237790688872338,
      "epoch": 0.002405,
      "grad_norm": 0.1487669199705124,
      "learning_rate": 2e-05,
      "loss": 0.2100670576095581,
      "mean_token_accuracy": 0.9413860589265823,
      "num_tokens": 38715496.0,
      "step": 2405
    },
    {
      "aux_loss": 8.10801305770874,
      "entropy": 0.16891746148467063,
      "epoch": 0.00241,
      "grad_norm": 0.14763256907463074,
      "learning_rate": 2e-05,
      "loss": 0.20151259899139404,
      "mean_token_accuracy": 0.9435642391443253,
      "num_tokens": 38796096.0,
      "step": 2410
    },
    {
      "aux_loss": 8.12585744857788,
      "entropy": 0.15316050089895725,
      "epoch": 0.002415,
      "grad_norm": 0.1634683907032013,
      "learning_rate": 2e-05,
      "loss": 0.18161994218826294,
      "mean_token_accuracy": 0.949803501367569,
      "num_tokens": 38876276.0,
      "step": 2415
    },
    {
      "aux_loss": 8.077015113830566,
      "entropy": 0.1525297023355961,
      "epoch": 0.00242,
      "grad_norm": 0.1803746074438095,
      "learning_rate": 2e-05,
      "loss": 0.20174922943115234,
      "mean_token_accuracy": 0.9449002683162689,
      "num_tokens": 38957001.0,
      "step": 2420
    },
    {
      "aux_loss": 8.078759098052979,
      "entropy": 0.1808013513684273,
      "epoch": 0.002425,
      "grad_norm": 0.1409953236579895,
      "learning_rate": 2e-05,
      "loss": 0.22260196208953859,
      "mean_token_accuracy": 0.9379542499780655,
      "num_tokens": 39038111.0,
      "step": 2425
    },
    {
      "aux_loss": 8.088305902481078,
      "entropy": 0.1743409626185894,
      "epoch": 0.00243,
      "grad_norm": 0.14868317544460297,
      "learning_rate": 2e-05,
      "loss": 0.1922954797744751,
      "mean_token_accuracy": 0.94639051258564,
      "num_tokens": 39119062.0,
      "step": 2430
    },
    {
      "aux_loss": 8.099035120010376,
      "entropy": 0.1727360025048256,
      "epoch": 0.002435,
      "grad_norm": 0.15608380734920502,
      "learning_rate": 2e-05,
      "loss": 0.19858545064926147,
      "mean_token_accuracy": 0.9447236627340316,
      "num_tokens": 39199781.0,
      "step": 2435
    },
    {
      "aux_loss": 8.109528255462646,
      "entropy": 0.16413209736347198,
      "epoch": 0.00244,
      "grad_norm": 0.1472751945257187,
      "learning_rate": 2e-05,
      "loss": 0.20274732112884522,
      "mean_token_accuracy": 0.9434846758842468,
      "num_tokens": 39280279.0,
      "step": 2440
    },
    {
      "aux_loss": 8.127269220352172,
      "entropy": 0.1620781935751438,
      "epoch": 0.002445,
      "grad_norm": 0.14663583040237427,
      "learning_rate": 2e-05,
      "loss": 0.1924894332885742,
      "mean_token_accuracy": 0.9457176208496094,
      "num_tokens": 39360400.0,
      "step": 2445
    },
    {
      "aux_loss": 8.089293003082275,
      "entropy": 0.18135993257164956,
      "epoch": 0.00245,
      "grad_norm": 0.15306329727172852,
      "learning_rate": 2e-05,
      "loss": 0.21985092163085937,
      "mean_token_accuracy": 0.9400442719459534,
      "num_tokens": 39440866.0,
      "step": 2450
    },
    {
      "aux_loss": 8.076449537277222,
      "entropy": 0.1844419553875923,
      "epoch": 0.002455,
      "grad_norm": 0.13918258249759674,
      "learning_rate": 2e-05,
      "loss": 0.22127432823181153,
      "mean_token_accuracy": 0.9395702898502349,
      "num_tokens": 39522161.0,
      "step": 2455
    },
    {
      "aux_loss": 8.084309816360474,
      "entropy": 0.18715021312236785,
      "epoch": 0.00246,
      "grad_norm": 0.14264462888240814,
      "learning_rate": 2e-05,
      "loss": 0.21688578128814698,
      "mean_token_accuracy": 0.9393721938133239,
      "num_tokens": 39603268.0,
      "step": 2460
    },
    {
      "aux_loss": 8.09578332901001,
      "entropy": 0.1833598129451275,
      "epoch": 0.002465,
      "grad_norm": 0.13851940631866455,
      "learning_rate": 2e-05,
      "loss": 0.20649237632751466,
      "mean_token_accuracy": 0.9424688786268234,
      "num_tokens": 39683887.0,
      "step": 2465
    },
    {
      "aux_loss": 8.112656593322754,
      "entropy": 0.16445145159959793,
      "epoch": 0.00247,
      "grad_norm": 0.14665396511554718,
      "learning_rate": 2e-05,
      "loss": 0.19205303192138673,
      "mean_token_accuracy": 0.945525485277176,
      "num_tokens": 39764133.0,
      "step": 2470
    },
    {
      "aux_loss": 8.122511672973634,
      "entropy": 0.1466979943215847,
      "epoch": 0.002475,
      "grad_norm": 0.14952561259269714,
      "learning_rate": 2e-05,
      "loss": 0.1791680335998535,
      "mean_token_accuracy": 0.9493878811597825,
      "num_tokens": 39844487.0,
      "step": 2475
    },
    {
      "aux_loss": 8.092697048187256,
      "entropy": 0.13941412158310412,
      "epoch": 0.00248,
      "grad_norm": 0.1423688679933548,
      "learning_rate": 2e-05,
      "loss": 0.18052189350128173,
      "mean_token_accuracy": 0.949979031085968,
      "num_tokens": 39922970.0,
      "step": 2480
    },
    {
      "aux_loss": 8.077152633666993,
      "entropy": 0.17317311018705367,
      "epoch": 0.002485,
      "grad_norm": 0.1295267641544342,
      "learning_rate": 2e-05,
      "loss": 0.21593210697174073,
      "mean_token_accuracy": 0.9398978143930435,
      "num_tokens": 40004037.0,
      "step": 2485
    },
    {
      "aux_loss": 8.079712390899658,
      "entropy": 0.17868908829987049,
      "epoch": 0.00249,
      "grad_norm": 0.14083313941955566,
      "learning_rate": 2e-05,
      "loss": 0.2062427282333374,
      "mean_token_accuracy": 0.9426745235919952,
      "num_tokens": 40084736.0,
      "step": 2490
    },
    {
      "aux_loss": 8.089857292175292,
      "entropy": 0.1889190748333931,
      "epoch": 0.002495,
      "grad_norm": 0.1387125551700592,
      "learning_rate": 2e-05,
      "loss": 0.21451432704925538,
      "mean_token_accuracy": 0.9401138484477997,
      "num_tokens": 40165274.0,
      "step": 2495
    },
    {
      "aux_loss": 8.10280442237854,
      "entropy": 0.17657697722315788,
      "epoch": 0.0025,
      "grad_norm": 0.1414203643798828,
      "learning_rate": 2e-05,
      "loss": 0.20892417430877686,
      "mean_token_accuracy": 0.941176226735115,
      "num_tokens": 40245693.0,
      "step": 2500
    },
    {
      "aux_loss": 8.118335151672364,
      "entropy": 0.1604831449687481,
      "epoch": 0.002505,
      "grad_norm": 0.1640281230211258,
      "learning_rate": 2e-05,
      "loss": 0.1901395082473755,
      "mean_token_accuracy": 0.9472380697727203,
      "num_tokens": 40326091.0,
      "step": 2505
    },
    {
      "aux_loss": 8.095034790039062,
      "entropy": 0.15698655024170877,
      "epoch": 0.00251,
      "grad_norm": 0.15623047947883606,
      "learning_rate": 2e-05,
      "loss": 0.20149469375610352,
      "mean_token_accuracy": 0.9449017673730851,
      "num_tokens": 40407165.0,
      "step": 2510
    },
    {
      "aux_loss": 8.076070404052734,
      "entropy": 0.18253343403339387,
      "epoch": 0.002515,
      "grad_norm": 0.14557181298732758,
      "learning_rate": 2e-05,
      "loss": 0.22407917976379393,
      "mean_token_accuracy": 0.9371753901243209,
      "num_tokens": 40488218.0,
      "step": 2515
    },
    {
      "aux_loss": 8.079090929031372,
      "entropy": 0.1965438723564148,
      "epoch": 0.00252,
      "grad_norm": 0.15903256833553314,
      "learning_rate": 2e-05,
      "loss": 0.22432379722595214,
      "mean_token_accuracy": 0.9371412605047226,
      "num_tokens": 40569357.0,
      "step": 2520
    },
    {
      "aux_loss": 8.092989253997803,
      "entropy": 0.1891261577606201,
      "epoch": 0.002525,
      "grad_norm": 0.13752049207687378,
      "learning_rate": 2e-05,
      "loss": 0.20952942371368408,
      "mean_token_accuracy": 0.9412952989339829,
      "num_tokens": 40650473.0,
      "step": 2525
    },
    {
      "aux_loss": 8.10186824798584,
      "entropy": 0.16528766751289367,
      "epoch": 0.00253,
      "grad_norm": 0.1446021944284439,
      "learning_rate": 2e-05,
      "loss": 0.1990690588951111,
      "mean_token_accuracy": 0.9436188697814941,
      "num_tokens": 40731040.0,
      "step": 2530
    },
    {
      "aux_loss": 8.115973663330077,
      "entropy": 0.15740383863449098,
      "epoch": 0.002535,
      "grad_norm": 0.14759008586406708,
      "learning_rate": 2e-05,
      "loss": 0.18687137365341186,
      "mean_token_accuracy": 0.9475764125585556,
      "num_tokens": 40811409.0,
      "step": 2535
    },
    {
      "aux_loss": 8.113844347000121,
      "entropy": 0.15397238209843636,
      "epoch": 0.00254,
      "grad_norm": 0.15065361559391022,
      "learning_rate": 2e-05,
      "loss": 0.1924121141433716,
      "mean_token_accuracy": 0.9462917983531952,
      "num_tokens": 40889180.0,
      "step": 2540
    },
    {
      "aux_loss": 8.073805284500121,
      "entropy": 0.18016892522573472,
      "epoch": 0.002545,
      "grad_norm": 0.15082381665706635,
      "learning_rate": 2e-05,
      "loss": 0.2252495765686035,
      "mean_token_accuracy": 0.9375465750694275,
      "num_tokens": 40969964.0,
      "step": 2545
    },
    {
      "aux_loss": 8.079817867279052,
      "entropy": 0.19116854891180993,
      "epoch": 0.00255,
      "grad_norm": 0.13598640263080597,
      "learning_rate": 2e-05,
      "loss": 0.22265090942382812,
      "mean_token_accuracy": 0.9392884552478791,
      "num_tokens": 41050990.0,
      "step": 2550
    },
    {
      "aux_loss": 8.091651058197021,
      "entropy": 0.19393812492489815,
      "epoch": 0.002555,
      "grad_norm": 0.1380283236503601,
      "learning_rate": 2e-05,
      "loss": 0.22394993305206298,
      "mean_token_accuracy": 0.9375712990760803,
      "num_tokens": 41131628.0,
      "step": 2555
    },
    {
      "aux_loss": 8.099888706207276,
      "entropy": 0.17336956933140754,
      "epoch": 0.00256,
      "grad_norm": 0.13645745813846588,
      "learning_rate": 2e-05,
      "loss": 0.2004483938217163,
      "mean_token_accuracy": 0.9430395931005477,
      "num_tokens": 41212285.0,
      "step": 2560
    },
    {
      "aux_loss": 8.111267566680908,
      "entropy": 0.16155279204249381,
      "epoch": 0.002565,
      "grad_norm": 0.14703181385993958,
      "learning_rate": 2e-05,
      "loss": 0.189046311378479,
      "mean_token_accuracy": 0.9468112826347351,
      "num_tokens": 41292841.0,
      "step": 2565
    },
    {
      "aux_loss": 8.109141540527343,
      "entropy": 0.1456486813724041,
      "epoch": 0.00257,
      "grad_norm": 0.14463473856449127,
      "learning_rate": 2e-05,
      "loss": 0.18097591400146484,
      "mean_token_accuracy": 0.9507924824953079,
      "num_tokens": 41372172.0,
      "step": 2570
    },
    {
      "aux_loss": 8.073219919204712,
      "entropy": 0.18117170445621014,
      "epoch": 0.002575,
      "grad_norm": 0.1482602208852768,
      "learning_rate": 2e-05,
      "loss": 0.22767367362976074,
      "mean_token_accuracy": 0.935913696885109,
      "num_tokens": 41453307.0,
      "step": 2575
    },
    {
      "aux_loss": 8.075421333312988,
      "entropy": 0.19263429790735245,
      "epoch": 0.00258,
      "grad_norm": 0.1455337107181549,
      "learning_rate": 2e-05,
      "loss": 0.22617831230163574,
      "mean_token_accuracy": 0.9362822055816651,
      "num_tokens": 41534026.0,
      "step": 2580
    },
    {
      "aux_loss": 8.087105178833008,
      "entropy": 0.17997874394059182,
      "epoch": 0.002585,
      "grad_norm": 0.14783543348312378,
      "learning_rate": 2e-05,
      "loss": 0.20190536975860596,
      "mean_token_accuracy": 0.9435222715139389,
      "num_tokens": 41614804.0,
      "step": 2585
    },
    {
      "aux_loss": 8.100873613357544,
      "entropy": 0.17839294150471688,
      "epoch": 0.00259,
      "grad_norm": 0.15630635619163513,
      "learning_rate": 2e-05,
      "loss": 0.21114187240600585,
      "mean_token_accuracy": 0.9409147024154663,
      "num_tokens": 41695832.0,
      "step": 2590
    },
    {
      "aux_loss": 8.111894226074218,
      "entropy": 0.16124511361122132,
      "epoch": 0.002595,
      "grad_norm": 0.15398986637592316,
      "learning_rate": 2e-05,
      "loss": 0.19127076864242554,
      "mean_token_accuracy": 0.9467630207538604,
      "num_tokens": 41776558.0,
      "step": 2595
    },
    {
      "aux_loss": 8.11488184928894,
      "entropy": 0.14952887110412122,
      "epoch": 0.0026,
      "grad_norm": 0.12254941463470459,
      "learning_rate": 2e-05,
      "loss": 0.18442397117614745,
      "mean_token_accuracy": 0.9494154214859009,
      "num_tokens": 41856347.0,
      "step": 2600
    },
    {
      "aux_loss": 8.070796585083007,
      "entropy": 0.17625901177525521,
      "epoch": 0.002605,
      "grad_norm": 0.15750719606876373,
      "learning_rate": 2e-05,
      "loss": 0.2269012451171875,
      "mean_token_accuracy": 0.9383298277854919,
      "num_tokens": 41937637.0,
      "step": 2605
    },
    {
      "aux_loss": 8.078232192993164,
      "entropy": 0.20892537087202073,
      "epoch": 0.00261,
      "grad_norm": 0.14309179782867432,
      "learning_rate": 2e-05,
      "loss": 0.2362462043762207,
      "mean_token_accuracy": 0.933705860376358,
      "num_tokens": 42018224.0,
      "step": 2610
    },
    {
      "aux_loss": 8.088038635253906,
      "entropy": 0.18252842277288436,
      "epoch": 0.002615,
      "grad_norm": 0.13642433285713196,
      "learning_rate": 2e-05,
      "loss": 0.2050029754638672,
      "mean_token_accuracy": 0.9437324970960617,
      "num_tokens": 42098789.0,
      "step": 2615
    },
    {
      "aux_loss": 8.101947784423828,
      "entropy": 0.18167223706841468,
      "epoch": 0.00262,
      "grad_norm": 0.14701588451862335,
      "learning_rate": 2e-05,
      "loss": 0.2142505407333374,
      "mean_token_accuracy": 0.939968267083168,
      "num_tokens": 42179697.0,
      "step": 2620
    },
    {
      "aux_loss": 8.112688302993774,
      "entropy": 0.1608296036720276,
      "epoch": 0.002625,
      "grad_norm": 0.15386417508125305,
      "learning_rate": 2e-05,
      "loss": 0.1939125895500183,
      "mean_token_accuracy": 0.9460368245840073,
      "num_tokens": 42260272.0,
      "step": 2625
    },
    {
      "aux_loss": 8.120966529846191,
      "entropy": 0.15800306014716625,
      "epoch": 0.00263,
      "grad_norm": 0.14041326940059662,
      "learning_rate": 2e-05,
      "loss": 0.19136632680892945,
      "mean_token_accuracy": 0.9471804887056351,
      "num_tokens": 42339732.0,
      "step": 2630
    },
    {
      "aux_loss": 8.071980142593384,
      "entropy": 0.18256490118801594,
      "epoch": 0.002635,
      "grad_norm": 0.154375359416008,
      "learning_rate": 2e-05,
      "loss": 0.2265719175338745,
      "mean_token_accuracy": 0.9373707473278046,
      "num_tokens": 42421141.0,
      "step": 2635
    },
    {
      "aux_loss": 8.075599956512452,
      "entropy": 0.19487906619906425,
      "epoch": 0.00264,
      "grad_norm": 0.1399112492799759,
      "learning_rate": 2e-05,
      "loss": 0.2332533836364746,
      "mean_token_accuracy": 0.9351333647966384,
      "num_tokens": 42502206.0,
      "step": 2640
    },
    {
      "aux_loss": 8.084503602981567,
      "entropy": 0.19292669147253036,
      "epoch": 0.002645,
      "grad_norm": 0.1477968543767929,
      "learning_rate": 2e-05,
      "loss": 0.2195763349533081,
      "mean_token_accuracy": 0.9394425630569458,
      "num_tokens": 42582876.0,
      "step": 2645
    },
    {
      "aux_loss": 8.09777274131775,
      "entropy": 0.17275489643216133,
      "epoch": 0.00265,
      "grad_norm": 0.14520341157913208,
      "learning_rate": 2e-05,
      "loss": 0.19572608470916747,
      "mean_token_accuracy": 0.9456899732351303,
      "num_tokens": 42663373.0,
      "step": 2650
    },
    {
      "aux_loss": 8.110652685165405,
      "entropy": 0.16136607751250268,
      "epoch": 0.002655,
      "grad_norm": 0.1658102571964264,
      "learning_rate": 2e-05,
      "loss": 0.19246256351470947,
      "mean_token_accuracy": 0.945327153801918,
      "num_tokens": 42744168.0,
      "step": 2655
    },
    {
      "aux_loss": 8.12573184967041,
      "entropy": 0.14542376026511192,
      "epoch": 0.00266,
      "grad_norm": 0.1605246365070343,
      "learning_rate": 2e-05,
      "loss": 0.17110786437988282,
      "mean_token_accuracy": 0.9521415412425995,
      "num_tokens": 42822630.0,
      "step": 2660
    },
    {
      "aux_loss": 8.069727516174316,
      "entropy": 0.17552723549306393,
      "epoch": 0.002665,
      "grad_norm": 0.2469029575586319,
      "learning_rate": 2e-05,
      "loss": 0.2253035306930542,
      "mean_token_accuracy": 0.9392019242048264,
      "num_tokens": 42903872.0,
      "step": 2665
    },
    {
      "aux_loss": 8.07782793045044,
      "entropy": 0.19894684851169586,
      "epoch": 0.00267,
      "grad_norm": 0.13871465623378754,
      "learning_rate": 2e-05,
      "loss": 0.23623056411743165,
      "mean_token_accuracy": 0.9340197622776032,
      "num_tokens": 42985349.0,
      "step": 2670
    },
    {
      "aux_loss": 8.085161781311035,
      "entropy": 0.19190286472439766,
      "epoch": 0.002675,
      "grad_norm": 0.14929887652397156,
      "learning_rate": 2e-05,
      "loss": 0.21718759536743165,
      "mean_token_accuracy": 0.9392900198698044,
      "num_tokens": 43066375.0,
      "step": 2675
    },
    {
      "aux_loss": 8.096126270294189,
      "entropy": 0.19029331654310228,
      "epoch": 0.00268,
      "grad_norm": 0.1458783745765686,
      "learning_rate": 2e-05,
      "loss": 0.21430518627166747,
      "mean_token_accuracy": 0.9394995540380477,
      "num_tokens": 43147219.0,
      "step": 2680
    },
    {
      "aux_loss": 8.107530117034912,
      "entropy": 0.15586493760347367,
      "epoch": 0.002685,
      "grad_norm": 0.1702526956796646,
      "learning_rate": 2e-05,
      "loss": 0.1878519892692566,
      "mean_token_accuracy": 0.9479204326868057,
      "num_tokens": 43227344.0,
      "step": 2685
    },
    {
      "aux_loss": 8.130838680267335,
      "entropy": 0.15244097486138344,
      "epoch": 0.00269,
      "grad_norm": 0.16751837730407715,
      "learning_rate": 2e-05,
      "loss": 0.1854215145111084,
      "mean_token_accuracy": 0.9483686149120331,
      "num_tokens": 43307625.0,
      "step": 2690
    },
    {
      "aux_loss": 8.078331851959229,
      "entropy": 0.1527701273560524,
      "epoch": 0.002695,
      "grad_norm": 0.14889752864837646,
      "learning_rate": 2e-05,
      "loss": 0.19740285873413085,
      "mean_token_accuracy": 0.9465331047773361,
      "num_tokens": 43388342.0,
      "step": 2695
    },
    {
      "aux_loss": 8.078295803070068,
      "entropy": 0.18069629594683648,
      "epoch": 0.0027,
      "grad_norm": 0.14572887122631073,
      "learning_rate": 2e-05,
      "loss": 0.21931033134460448,
      "mean_token_accuracy": 0.9392302930355072,
      "num_tokens": 43469497.0,
      "step": 2700
    },
    {
      "aux_loss": 8.083542346954346,
      "entropy": 0.18787576258182526,
      "epoch": 0.002705,
      "grad_norm": 0.13534726202487946,
      "learning_rate": 2e-05,
      "loss": 0.2157909631729126,
      "mean_token_accuracy": 0.9400046467781067,
      "num_tokens": 43550520.0,
      "step": 2705
    },
    {
      "aux_loss": 8.098087263107299,
      "entropy": 0.18121880441904067,
      "epoch": 0.00271,
      "grad_norm": 0.14448904991149902,
      "learning_rate": 2e-05,
      "loss": 0.20662832260131836,
      "mean_token_accuracy": 0.941738149523735,
      "num_tokens": 43631338.0,
      "step": 2710
    },
    {
      "aux_loss": 8.112855195999146,
      "entropy": 0.1707596518099308,
      "epoch": 0.002715,
      "grad_norm": 0.1639878749847412,
      "learning_rate": 2e-05,
      "loss": 0.2031870126724243,
      "mean_token_accuracy": 0.9440413445234299,
      "num_tokens": 43711619.0,
      "step": 2715
    },
    {
      "aux_loss": 8.126987075805664,
      "entropy": 0.15146464705467225,
      "epoch": 0.00272,
      "grad_norm": 0.15374919772148132,
      "learning_rate": 2e-05,
      "loss": 0.17934576272964478,
      "mean_token_accuracy": 0.9499849855899811,
      "num_tokens": 43792031.0,
      "step": 2720
    },
    {
      "aux_loss": 8.08648157119751,
      "entropy": 0.16811617240309715,
      "epoch": 0.002725,
      "grad_norm": 0.16933873295783997,
      "learning_rate": 2e-05,
      "loss": 0.20919368267059327,
      "mean_token_accuracy": 0.9438339740037918,
      "num_tokens": 43871936.0,
      "step": 2725
    },
    {
      "aux_loss": 8.075786447525024,
      "entropy": 0.19027187675237656,
      "epoch": 0.00273,
      "grad_norm": 0.15974991023540497,
      "learning_rate": 2e-05,
      "loss": 0.2363669157028198,
      "mean_token_accuracy": 0.9353165745735168,
      "num_tokens": 43952962.0,
      "step": 2730
    },
    {
      "aux_loss": 8.080311822891236,
      "entropy": 0.198441019654274,
      "epoch": 0.002735,
      "grad_norm": 0.15618927776813507,
      "learning_rate": 2e-05,
      "loss": 0.2246633768081665,
      "mean_token_accuracy": 0.9376106888055802,
      "num_tokens": 44034116.0,
      "step": 2735
    },
    {
      "aux_loss": 8.0945152759552,
      "entropy": 0.19126661866903305,
      "epoch": 0.00274,
      "grad_norm": 0.14107875525951385,
      "learning_rate": 2e-05,
      "loss": 0.21505730152130126,
      "mean_token_accuracy": 0.9397924184799195,
      "num_tokens": 44115130.0,
      "step": 2740
    },
    {
      "aux_loss": 8.10862979888916,
      "entropy": 0.1707291953265667,
      "epoch": 0.002745,
      "grad_norm": 0.15384399890899658,
      "learning_rate": 2e-05,
      "loss": 0.19822795391082765,
      "mean_token_accuracy": 0.9442929923534393,
      "num_tokens": 44195401.0,
      "step": 2745
    },
    {
      "aux_loss": 8.125518846511842,
      "entropy": 0.14779275730252267,
      "epoch": 0.00275,
      "grad_norm": 0.16695097088813782,
      "learning_rate": 2e-05,
      "loss": 0.1799459218978882,
      "mean_token_accuracy": 0.9499589592218399,
      "num_tokens": 44275896.0,
      "step": 2750
    },
    {
      "aux_loss": 8.091786193847657,
      "entropy": 0.1664949256926775,
      "epoch": 0.002755,
      "grad_norm": 0.16463322937488556,
      "learning_rate": 2e-05,
      "loss": 0.21682701110839844,
      "mean_token_accuracy": 0.940763384103775,
      "num_tokens": 44354850.0,
      "step": 2755
    },
    {
      "aux_loss": 8.078452444076538,
      "entropy": 0.19329728484153746,
      "epoch": 0.00276,
      "grad_norm": 0.13854876160621643,
      "learning_rate": 2e-05,
      "loss": 0.23405907154083253,
      "mean_token_accuracy": 0.9349021703004837,
      "num_tokens": 44435935.0,
      "step": 2760
    },
    {
      "aux_loss": 8.081658744812012,
      "entropy": 0.19469547793269157,
      "epoch": 0.002765,
      "grad_norm": 0.13840121030807495,
      "learning_rate": 2e-05,
      "loss": 0.21995303630828858,
      "mean_token_accuracy": 0.9390068531036377,
      "num_tokens": 44516956.0,
      "step": 2765
    },
    {
      "aux_loss": 8.094853734970092,
      "entropy": 0.18673215955495834,
      "epoch": 0.00277,
      "grad_norm": 0.14616280794143677,
      "learning_rate": 2e-05,
      "loss": 0.21082022190093994,
      "mean_token_accuracy": 0.9406619310379029,
      "num_tokens": 44597778.0,
      "step": 2770
    },
    {
      "aux_loss": 8.106273174285889,
      "entropy": 0.16729313060641288,
      "epoch": 0.002775,
      "grad_norm": 0.16103743016719818,
      "learning_rate": 2e-05,
      "loss": 0.19992878437042236,
      "mean_token_accuracy": 0.944680991768837,
      "num_tokens": 44678107.0,
      "step": 2775
    },
    {
      "aux_loss": 8.1201500415802,
      "entropy": 0.1520601712167263,
      "epoch": 0.00278,
      "grad_norm": 0.15789227187633514,
      "learning_rate": 2e-05,
      "loss": 0.1843225598335266,
      "mean_token_accuracy": 0.9481644630432129,
      "num_tokens": 44757973.0,
      "step": 2780
    },
    {
      "aux_loss": 8.096934032440185,
      "entropy": 0.14989367127418518,
      "epoch": 0.002785,
      "grad_norm": 0.15040884912014008,
      "learning_rate": 2e-05,
      "loss": 0.18910366296768188,
      "mean_token_accuracy": 0.9482103943824768,
      "num_tokens": 44836566.0,
      "step": 2785
    },
    {
      "aux_loss": 8.076406717300415,
      "entropy": 0.19880288168787957,
      "epoch": 0.00279,
      "grad_norm": 0.14457979798316956,
      "learning_rate": 2e-05,
      "loss": 0.24200451374053955,
      "mean_token_accuracy": 0.9328301399946213,
      "num_tokens": 44917514.0,
      "step": 2790
    },
    {
      "aux_loss": 8.084968185424804,
      "entropy": 0.18911403678357602,
      "epoch": 0.002795,
      "grad_norm": 0.14090609550476074,
      "learning_rate": 2e-05,
      "loss": 0.21615207195281982,
      "mean_token_accuracy": 0.939749613404274,
      "num_tokens": 44998413.0,
      "step": 2795
    },
    {
      "aux_loss": 8.091570806503295,
      "entropy": 0.18450974449515342,
      "epoch": 0.0028,
      "grad_norm": 0.14722010493278503,
      "learning_rate": 2e-05,
      "loss": 0.2080524444580078,
      "mean_token_accuracy": 0.9409097254276275,
      "num_tokens": 45079215.0,
      "step": 2800
    },
    {
      "aux_loss": 8.100802278518676,
      "entropy": 0.1651144064962864,
      "epoch": 0.002805,
      "grad_norm": 0.14703086018562317,
      "learning_rate": 2e-05,
      "loss": 0.1942728877067566,
      "mean_token_accuracy": 0.9462512761354447,
      "num_tokens": 45159778.0,
      "step": 2805
    },
    {
      "aux_loss": 8.116804504394532,
      "entropy": 0.15706228241324424,
      "epoch": 0.00281,
      "grad_norm": 0.16164728999137878,
      "learning_rate": 2e-05,
      "loss": 0.19007450342178345,
      "mean_token_accuracy": 0.9475653886795044,
      "num_tokens": 45239925.0,
      "step": 2810
    },
    {
      "aux_loss": 8.105684089660645,
      "entropy": 0.15516949817538261,
      "epoch": 0.002815,
      "grad_norm": 0.1553998589515686,
      "learning_rate": 2e-05,
      "loss": 0.19682148694992066,
      "mean_token_accuracy": 0.9462418019771576,
      "num_tokens": 45318777.0,
      "step": 2815
    },
    {
      "aux_loss": 8.06877179145813,
      "entropy": 0.17810099869966506,
      "epoch": 0.00282,
      "grad_norm": 0.15135815739631653,
      "learning_rate": 2e-05,
      "loss": 0.22194890975952147,
      "mean_token_accuracy": 0.9384380996227264,
      "num_tokens": 45400058.0,
      "step": 2820
    },
    {
      "aux_loss": 8.078272819519043,
      "entropy": 0.19753627777099608,
      "epoch": 0.002825,
      "grad_norm": 0.13577862083911896,
      "learning_rate": 2e-05,
      "loss": 0.23035862445831298,
      "mean_token_accuracy": 0.935256740450859,
      "num_tokens": 45480492.0,
      "step": 2825
    },
    {
      "aux_loss": 8.088609981536866,
      "entropy": 0.17886079102754593,
      "epoch": 0.00283,
      "grad_norm": 0.1576220989227295,
      "learning_rate": 2e-05,
      "loss": 0.19549431800842285,
      "mean_token_accuracy": 0.9454530507326127,
      "num_tokens": 45561258.0,
      "step": 2830
    },
    {
      "aux_loss": 8.098433589935302,
      "entropy": 0.17717600762844085,
      "epoch": 0.002835,
      "grad_norm": 0.14695903658866882,
      "learning_rate": 2e-05,
      "loss": 0.2073725700378418,
      "mean_token_accuracy": 0.9411997050046921,
      "num_tokens": 45641626.0,
      "step": 2835
    },
    {
      "aux_loss": 8.112347507476807,
      "entropy": 0.15916404463350772,
      "epoch": 0.00284,
      "grad_norm": 0.16437484323978424,
      "learning_rate": 2e-05,
      "loss": 0.19357267618179322,
      "mean_token_accuracy": 0.9453988254070282,
      "num_tokens": 45722304.0,
      "step": 2840
    },
    {
      "aux_loss": 8.113294410705567,
      "entropy": 0.14726108722388745,
      "epoch": 0.002845,
      "grad_norm": 0.13739712536334991,
      "learning_rate": 2e-05,
      "loss": 0.18099336624145507,
      "mean_token_accuracy": 0.950264859199524,
      "num_tokens": 45802174.0,
      "step": 2845
    },
    {
      "aux_loss": 8.067266464233398,
      "entropy": 0.16752013862133025,
      "epoch": 0.00285,
      "grad_norm": 0.1620785892009735,
      "learning_rate": 2e-05,
      "loss": 0.21951792240142823,
      "mean_token_accuracy": 0.9401003777980804,
      "num_tokens": 45883444.0,
      "step": 2850
    },
    {
      "aux_loss": 8.073601818084716,
      "entropy": 0.18826278373599054,
      "epoch": 0.002855,
      "grad_norm": 0.14571647346019745,
      "learning_rate": 2e-05,
      "loss": 0.21801767349243165,
      "mean_token_accuracy": 0.9387796074151993,
      "num_tokens": 45964043.0,
      "step": 2855
    },
    {
      "aux_loss": 8.087750053405761,
      "entropy": 0.19014779180288316,
      "epoch": 0.00286,
      "grad_norm": 0.14232376217842102,
      "learning_rate": 2e-05,
      "loss": 0.21073970794677735,
      "mean_token_accuracy": 0.9416055470705033,
      "num_tokens": 46044722.0,
      "step": 2860
    },
    {
      "aux_loss": 8.097754430770873,
      "entropy": 0.1776500701904297,
      "epoch": 0.002865,
      "grad_norm": 0.17109878361225128,
      "learning_rate": 2e-05,
      "loss": 0.20462331771850586,
      "mean_token_accuracy": 0.9429490447044373,
      "num_tokens": 46125391.0,
      "step": 2865
    },
    {
      "aux_loss": 8.109619522094727,
      "entropy": 0.1521120510995388,
      "epoch": 0.00287,
      "grad_norm": 0.15355637669563293,
      "learning_rate": 2e-05,
      "loss": 0.1853906273841858,
      "mean_token_accuracy": 0.9488934278488159,
      "num_tokens": 46205778.0,
      "step": 2870
    },
    {
      "aux_loss": 8.12909483909607,
      "entropy": 0.15182966738939285,
      "epoch": 0.002875,
      "grad_norm": 0.15022306144237518,
      "learning_rate": 2e-05,
      "loss": 0.17824344635009765,
      "mean_token_accuracy": 0.9494897872209549,
      "num_tokens": 46282933.0,
      "step": 2875
    },
    {
      "aux_loss": 8.072309017181396,
      "entropy": 0.16922202073037623,
      "epoch": 0.00288,
      "grad_norm": 0.1451803594827652,
      "learning_rate": 2e-05,
      "loss": 0.21520595550537108,
      "mean_token_accuracy": 0.9406840562820434,
      "num_tokens": 46364260.0,
      "step": 2880
    },
    {
      "aux_loss": 8.07764573097229,
      "entropy": 0.1865469954907894,
      "epoch": 0.002885,
      "grad_norm": 0.1395547091960907,
      "learning_rate": 2e-05,
      "loss": 0.22133069038391112,
      "mean_token_accuracy": 0.9387134283781051,
      "num_tokens": 46444829.0,
      "step": 2885
    },
    {
      "aux_loss": 8.087185287475586,
      "entropy": 0.18829575702548026,
      "epoch": 0.00289,
      "grad_norm": 0.15173520147800446,
      "learning_rate": 2e-05,
      "loss": 0.21082444190979005,
      "mean_token_accuracy": 0.9413496106863022,
      "num_tokens": 46525629.0,
      "step": 2890
    },
    {
      "aux_loss": 8.098417282104492,
      "entropy": 0.17964600697159766,
      "epoch": 0.002895,
      "grad_norm": 0.1490006297826767,
      "learning_rate": 2e-05,
      "loss": 0.20410447120666503,
      "mean_token_accuracy": 0.9433809965848923,
      "num_tokens": 46606117.0,
      "step": 2895
    },
    {
      "aux_loss": 8.108289337158203,
      "entropy": 0.15851739272475243,
      "epoch": 0.0029,
      "grad_norm": 0.16662052273750305,
      "learning_rate": 2e-05,
      "loss": 0.19273626804351807,
      "mean_token_accuracy": 0.9464959114789963,
      "num_tokens": 46685972.0,
      "step": 2900
    },
    {
      "aux_loss": 8.125025749206543,
      "entropy": 0.1530994538217783,
      "epoch": 0.002905,
      "grad_norm": 0.15623949468135834,
      "learning_rate": 2e-05,
      "loss": 0.1868635892868042,
      "mean_token_accuracy": 0.9484824895858764,
      "num_tokens": 46766456.0,
      "step": 2905
    },
    {
      "aux_loss": 8.075625801086426,
      "entropy": 0.16139253973960876,
      "epoch": 0.00291,
      "grad_norm": 0.14356064796447754,
      "learning_rate": 2e-05,
      "loss": 0.20534379482269288,
      "mean_token_accuracy": 0.9439013361930847,
      "num_tokens": 46847537.0,
      "step": 2910
    },
    {
      "aux_loss": 8.076686811447143,
      "entropy": 0.18802704960107802,
      "epoch": 0.002915,
      "grad_norm": 0.1410534828901291,
      "learning_rate": 2e-05,
      "loss": 0.22646405696868896,
      "mean_token_accuracy": 0.9367367178201675,
      "num_tokens": 46928461.0,
      "step": 2915
    },
    {
      "aux_loss": 8.08428888320923,
      "entropy": 0.1758427046239376,
      "epoch": 0.00292,
      "grad_norm": 0.14738109707832336,
      "learning_rate": 2e-05,
      "loss": 0.20100624561309816,
      "mean_token_accuracy": 0.9441363543272019,
      "num_tokens": 47009106.0,
      "step": 2920
    },
    {
      "aux_loss": 8.095292329788208,
      "entropy": 0.18274475634098053,
      "epoch": 0.002925,
      "grad_norm": 0.15297847986221313,
      "learning_rate": 2e-05,
      "loss": 0.2086167335510254,
      "mean_token_accuracy": 0.9418325036764145,
      "num_tokens": 47089327.0,
      "step": 2925
    },
    {
      "aux_loss": 8.107821607589722,
      "entropy": 0.15642909482121467,
      "epoch": 0.00293,
      "grad_norm": 0.14811629056930542,
      "learning_rate": 2e-05,
      "loss": 0.1830051898956299,
      "mean_token_accuracy": 0.9496629565954209,
      "num_tokens": 47170193.0,
      "step": 2930
    },
    {
      "aux_loss": 8.131646299362183,
      "entropy": 0.15163895227015017,
      "epoch": 0.002935,
      "grad_norm": 0.22720278799533844,
      "learning_rate": 2e-05,
      "loss": 0.19013959169387817,
      "mean_token_accuracy": 0.9475005149841309,
      "num_tokens": 47247385.0,
      "step": 2935
    },
    {
      "aux_loss": 8.076407861709594,
      "entropy": 0.16601541228592395,
      "epoch": 0.00294,
      "grad_norm": 0.16172954440116882,
      "learning_rate": 2e-05,
      "loss": 0.2110731601715088,
      "mean_token_accuracy": 0.9428413033485412,
      "num_tokens": 47328897.0,
      "step": 2940
    },
    {
      "aux_loss": 8.080685424804688,
      "entropy": 0.16975508704781533,
      "epoch": 0.002945,
      "grad_norm": 0.13648611307144165,
      "learning_rate": 2e-05,
      "loss": 0.20457279682159424,
      "mean_token_accuracy": 0.9430114537477493,
      "num_tokens": 47409920.0,
      "step": 2945
    },
    {
      "aux_loss": 8.08769178390503,
      "entropy": 0.17321218699216842,
      "epoch": 0.00295,
      "grad_norm": 0.1371011734008789,
      "learning_rate": 2e-05,
      "loss": 0.20059237480163575,
      "mean_token_accuracy": 0.944344025850296,
      "num_tokens": 47490216.0,
      "step": 2950
    },
    {
      "aux_loss": 8.09835376739502,
      "entropy": 0.17534421011805534,
      "epoch": 0.002955,
      "grad_norm": 0.14919205009937286,
      "learning_rate": 2e-05,
      "loss": 0.20029969215393068,
      "mean_token_accuracy": 0.9436137437820434,
      "num_tokens": 47570540.0,
      "step": 2955
    },
    {
      "aux_loss": 8.108619117736817,
      "entropy": 0.15632903352379798,
      "epoch": 0.00296,
      "grad_norm": 0.17632590234279633,
      "learning_rate": 2e-05,
      "loss": 0.1897244095802307,
      "mean_token_accuracy": 0.9480899810791016,
      "num_tokens": 47651290.0,
      "step": 2960
    },
    {
      "aux_loss": 8.126777124404907,
      "entropy": 0.15064034573733806,
      "epoch": 0.002965,
      "grad_norm": 0.16173529624938965,
      "learning_rate": 2e-05,
      "loss": 0.18522897958755494,
      "mean_token_accuracy": 0.9485573321580887,
      "num_tokens": 47731384.0,
      "step": 2965
    },
    {
      "aux_loss": 8.069377374649047,
      "entropy": 0.17484852522611619,
      "epoch": 0.00297,
      "grad_norm": 0.14227241277694702,
      "learning_rate": 2e-05,
      "loss": 0.22020425796508789,
      "mean_token_accuracy": 0.9394288778305053,
      "num_tokens": 47812737.0,
      "step": 2970
    },
    {
      "aux_loss": 8.074917316436768,
      "entropy": 0.1867200769484043,
      "epoch": 0.002975,
      "grad_norm": 0.1428276002407074,
      "learning_rate": 2e-05,
      "loss": 0.2199643850326538,
      "mean_token_accuracy": 0.9371882557868958,
      "num_tokens": 47893883.0,
      "step": 2975
    },
    {
      "aux_loss": 8.08695330619812,
      "entropy": 0.185770183801651,
      "epoch": 0.00298,
      "grad_norm": 0.1431232988834381,
      "learning_rate": 2e-05,
      "loss": 0.21032371520996093,
      "mean_token_accuracy": 0.9413160055875778,
      "num_tokens": 47974773.0,
      "step": 2980
    },
    {
      "aux_loss": 8.093664646148682,
      "entropy": 0.17664483636617662,
      "epoch": 0.002985,
      "grad_norm": 0.14912278950214386,
      "learning_rate": 2e-05,
      "loss": 0.20335035324096679,
      "mean_token_accuracy": 0.9418219566345215,
      "num_tokens": 48055120.0,
      "step": 2985
    },
    {
      "aux_loss": 8.10689287185669,
      "entropy": 0.15533287599682807,
      "epoch": 0.00299,
      "grad_norm": 0.14180661737918854,
      "learning_rate": 2e-05,
      "loss": 0.18206703662872314,
      "mean_token_accuracy": 0.9497777760028839,
      "num_tokens": 48135634.0,
      "step": 2990
    },
    {
      "aux_loss": 8.128514432907105,
      "entropy": 0.15512944869697093,
      "epoch": 0.002995,
      "grad_norm": 0.23194196820259094,
      "learning_rate": 2e-05,
      "loss": 0.1939287304878235,
      "mean_token_accuracy": 0.9470187544822692,
      "num_tokens": 48212781.0,
      "step": 2995
    },
    {
      "aux_loss": 8.074347972869873,
      "entropy": 0.14761393442749976,
      "epoch": 0.003,
      "grad_norm": 0.16289988160133362,
      "learning_rate": 2e-05,
      "loss": 0.19531513452529908,
      "mean_token_accuracy": 0.9461183100938797,
      "num_tokens": 48293983.0,
      "step": 3000
    },
    {
      "aux_loss": 8.074123859405518,
      "entropy": 0.17624220326542855,
      "epoch": 0.003005,
      "grad_norm": 0.1353483945131302,
      "learning_rate": 2e-05,
      "loss": 0.20937368869781495,
      "mean_token_accuracy": 0.941815647482872,
      "num_tokens": 48374979.0,
      "step": 3005
    },
    {
      "aux_loss": 8.084510660171508,
      "entropy": 0.18216314911842346,
      "epoch": 0.00301,
      "grad_norm": 0.14181570708751678,
      "learning_rate": 2e-05,
      "loss": 0.20339524745941162,
      "mean_token_accuracy": 0.9437639027833938,
      "num_tokens": 48455692.0,
      "step": 3010
    },
    {
      "aux_loss": 8.094678926467896,
      "entropy": 0.1740648839622736,
      "epoch": 0.003015,
      "grad_norm": 0.15326496958732605,
      "learning_rate": 2e-05,
      "loss": 0.2038062334060669,
      "mean_token_accuracy": 0.9432501882314682,
      "num_tokens": 48536353.0,
      "step": 3015
    },
    {
      "aux_loss": 8.106809997558594,
      "entropy": 0.16099482402205467,
      "epoch": 0.00302,
      "grad_norm": 0.17165681719779968,
      "learning_rate": 2e-05,
      "loss": 0.19483649730682373,
      "mean_token_accuracy": 0.946119824051857,
      "num_tokens": 48616858.0,
      "step": 3020
    },
    {
      "aux_loss": 8.123516130447388,
      "entropy": 0.14078146517276763,
      "epoch": 0.003025,
      "grad_norm": 0.17677168548107147,
      "learning_rate": 2e-05,
      "loss": 0.1703662395477295,
      "mean_token_accuracy": 0.9534485638141632,
      "num_tokens": 48697699.0,
      "step": 3025
    },
    {
      "aux_loss": 8.073600101470948,
      "entropy": 0.1642740063369274,
      "epoch": 0.00303,
      "grad_norm": 0.15738898515701294,
      "learning_rate": 2e-05,
      "loss": 0.20547797679901122,
      "mean_token_accuracy": 0.9430563062429428,
      "num_tokens": 48776328.0,
      "step": 3030
    },
    {
      "aux_loss": 8.074311685562133,
      "entropy": 0.1766006827354431,
      "epoch": 0.003035,
      "grad_norm": 0.13500206172466278,
      "learning_rate": 2e-05,
      "loss": 0.21654367446899414,
      "mean_token_accuracy": 0.939488896727562,
      "num_tokens": 48857588.0,
      "step": 3035
    },
    {
      "aux_loss": 8.08253984451294,
      "entropy": 0.1705415889620781,
      "epoch": 0.00304,
      "grad_norm": 0.13991662859916687,
      "learning_rate": 2e-05,
      "loss": 0.19832667112350463,
      "mean_token_accuracy": 0.9440321087837219,
      "num_tokens": 48938218.0,
      "step": 3040
    },
    {
      "aux_loss": 8.09462513923645,
      "entropy": 0.1893142431974411,
      "epoch": 0.003045,
      "grad_norm": 0.1491444706916809,
      "learning_rate": 2e-05,
      "loss": 0.21176254749298096,
      "mean_token_accuracy": 0.9404118835926056,
      "num_tokens": 49018629.0,
      "step": 3045
    },
    {
      "aux_loss": 8.106306886672973,
      "entropy": 0.16372528150677682,
      "epoch": 0.00305,
      "grad_norm": 0.14525392651557922,
      "learning_rate": 2e-05,
      "loss": 0.18929215669631957,
      "mean_token_accuracy": 0.9470975756645202,
      "num_tokens": 49099554.0,
      "step": 3050
    },
    {
      "aux_loss": 8.121476459503175,
      "entropy": 0.14607863649725913,
      "epoch": 0.003055,
      "grad_norm": 0.172555610537529,
      "learning_rate": 2e-05,
      "loss": 0.17976579666137696,
      "mean_token_accuracy": 0.9501303941011429,
      "num_tokens": 49179679.0,
      "step": 3055
    },
    {
      "aux_loss": 8.070800018310546,
      "entropy": 0.1742791451513767,
      "epoch": 0.00306,
      "grad_norm": 0.16902503371238708,
      "learning_rate": 2e-05,
      "loss": 0.22889356613159179,
      "mean_token_accuracy": 0.9380523175001144,
      "num_tokens": 49259841.0,
      "step": 3060
    },
    {
      "aux_loss": 8.075080823898315,
      "entropy": 0.18140368089079856,
      "epoch": 0.003065,
      "grad_norm": 0.14427874982357025,
      "learning_rate": 2e-05,
      "loss": 0.21710011959075928,
      "mean_token_accuracy": 0.9395097434520722,
      "num_tokens": 49340761.0,
      "step": 3065
    },
    {
      "aux_loss": 8.080142736434937,
      "entropy": 0.17514398097991943,
      "epoch": 0.00307,
      "grad_norm": 0.1458827555179596,
      "learning_rate": 2e-05,
      "loss": 0.20087404251098634,
      "mean_token_accuracy": 0.9440953344106674,
      "num_tokens": 49421716.0,
      "step": 3070
    },
    {
      "aux_loss": 8.094359970092773,
      "entropy": 0.18348851203918456,
      "epoch": 0.003075,
      "grad_norm": 0.1496962159872055,
      "learning_rate": 2e-05,
      "loss": 0.2093740224838257,
      "mean_token_accuracy": 0.9411183804273605,
      "num_tokens": 49502346.0,
      "step": 3075
    },
    {
      "aux_loss": 8.107255792617797,
      "entropy": 0.1585214577615261,
      "epoch": 0.00308,
      "grad_norm": 0.14572924375534058,
      "learning_rate": 2e-05,
      "loss": 0.18582788705825806,
      "mean_token_accuracy": 0.9484485566616059,
      "num_tokens": 49582931.0,
      "step": 3080
    },
    {
      "aux_loss": 8.125768804550171,
      "entropy": 0.1524410367012024,
      "epoch": 0.003085,
      "grad_norm": 0.19554057717323303,
      "learning_rate": 2e-05,
      "loss": 0.18749512434005738,
      "mean_token_accuracy": 0.9487124681472778,
      "num_tokens": 49662038.0,
      "step": 3085
    },
    {
      "aux_loss": 8.065309143066406,
      "entropy": 0.1476170975714922,
      "epoch": 0.00309,
      "grad_norm": 0.1500067263841629,
      "learning_rate": 2e-05,
      "loss": 0.1956866502761841,
      "mean_token_accuracy": 0.9471260637044907,
      "num_tokens": 49743454.0,
      "step": 3090
    },
    {
      "aux_loss": 8.070446252822876,
      "entropy": 0.18643608838319778,
      "epoch": 0.003095,
      "grad_norm": 0.12826579809188843,
      "learning_rate": 2e-05,
      "loss": 0.22123312950134277,
      "mean_token_accuracy": 0.9385463774204255,
      "num_tokens": 49824814.0,
      "step": 3095
    },
    {
      "aux_loss": 8.083160877227783,
      "entropy": 0.18798451498150826,
      "epoch": 0.0031,
      "grad_norm": 0.14452606439590454,
      "learning_rate": 2e-05,
      "loss": 0.21503002643585206,
      "mean_token_accuracy": 0.9404649019241333,
      "num_tokens": 49905807.0,
      "step": 3100
    },
    {
      "aux_loss": 8.092270660400391,
      "entropy": 0.18032723516225815,
      "epoch": 0.003105,
      "grad_norm": 0.1419580727815628,
      "learning_rate": 2e-05,
      "loss": 0.2056110382080078,
      "mean_token_accuracy": 0.9420620232820511,
      "num_tokens": 49986673.0,
      "step": 3105
    },
    {
      "aux_loss": 8.102667379379273,
      "entropy": 0.16459283158183097,
      "epoch": 0.00311,
      "grad_norm": 0.14356949925422668,
      "learning_rate": 2e-05,
      "loss": 0.19574639797210694,
      "mean_token_accuracy": 0.9458548665046692,
      "num_tokens": 50067358.0,
      "step": 3110
    },
    {
      "aux_loss": 8.119340229034425,
      "entropy": 0.15142956078052522,
      "epoch": 0.003115,
      "grad_norm": 0.16615331172943115,
      "learning_rate": 2e-05,
      "loss": 0.17887283563613893,
      "mean_token_accuracy": 0.9498388588428497,
      "num_tokens": 50147498.0,
      "step": 3115
    },
    {
      "aux_loss": 8.083586502075196,
      "entropy": 0.15965948812663555,
      "epoch": 0.00312,
      "grad_norm": 0.15895089507102966,
      "learning_rate": 2e-05,
      "loss": 0.20383920669555664,
      "mean_token_accuracy": 0.943869486451149,
      "num_tokens": 50228432.0,
      "step": 3120
    },
    {
      "aux_loss": 8.079378700256347,
      "entropy": 0.17228425592184066,
      "epoch": 0.003125,
      "grad_norm": 0.14265304803848267,
      "learning_rate": 2e-05,
      "loss": 0.21231956481933595,
      "mean_token_accuracy": 0.9405779361724853,
      "num_tokens": 50309758.0,
      "step": 3125
    },
    {
      "aux_loss": 8.081025695800781,
      "entropy": 0.18083779141306877,
      "epoch": 0.00313,
      "grad_norm": 0.149937704205513,
      "learning_rate": 2e-05,
      "loss": 0.2152872323989868,
      "mean_token_accuracy": 0.9383395344018937,
      "num_tokens": 50390948.0,
      "step": 3130
    },
    {
      "aux_loss": 8.091249132156372,
      "entropy": 0.17624640986323356,
      "epoch": 0.003135,
      "grad_norm": 0.1447724848985672,
      "learning_rate": 2e-05,
      "loss": 0.1994763970375061,
      "mean_token_accuracy": 0.9429386019706726,
      "num_tokens": 50472129.0,
      "step": 3135
    },
    {
      "aux_loss": 8.101275825500489,
      "entropy": 0.1689361333847046,
      "epoch": 0.00314,
      "grad_norm": 0.1408073753118515,
      "learning_rate": 2e-05,
      "loss": 0.19626353979110717,
      "mean_token_accuracy": 0.9448360979557038,
      "num_tokens": 50553021.0,
      "step": 3140
    },
    {
      "aux_loss": 8.11643452644348,
      "entropy": 0.14498640932142734,
      "epoch": 0.003145,
      "grad_norm": 0.15813784301280975,
      "learning_rate": 2e-05,
      "loss": 0.171815824508667,
      "mean_token_accuracy": 0.951966792345047,
      "num_tokens": 50633262.0,
      "step": 3145
    },
    {
      "aux_loss": 8.091406774520873,
      "entropy": 0.1517676327377558,
      "epoch": 0.00315,
      "grad_norm": 0.1492190659046173,
      "learning_rate": 2e-05,
      "loss": 0.1973658561706543,
      "mean_token_accuracy": 0.9443111538887023,
      "num_tokens": 50710671.0,
      "step": 3150
    },
    {
      "aux_loss": 8.072928285598755,
      "entropy": 0.17895912900567054,
      "epoch": 0.003155,
      "grad_norm": 0.13991013169288635,
      "learning_rate": 2e-05,
      "loss": 0.22549657821655272,
      "mean_token_accuracy": 0.9366275131702423,
      "num_tokens": 50791741.0,
      "step": 3155
    },
    {
      "aux_loss": 8.078258371353149,
      "entropy": 0.1867855153977871,
      "epoch": 0.00316,
      "grad_norm": 0.13619062304496765,
      "learning_rate": 2e-05,
      "loss": 0.21175603866577147,
      "mean_token_accuracy": 0.9412827521562577,
      "num_tokens": 50872809.0,
      "step": 3160
    },
    {
      "aux_loss": 8.088674831390382,
      "entropy": 0.17993197366595268,
      "epoch": 0.003165,
      "grad_norm": 0.13888010382652283,
      "learning_rate": 2e-05,
      "loss": 0.2040722131729126,
      "mean_token_accuracy": 0.9433172762393951,
      "num_tokens": 50953775.0,
      "step": 3165
    },
    {
      "aux_loss": 8.101012372970581,
      "entropy": 0.17506337612867356,
      "epoch": 0.00317,
      "grad_norm": 0.15246519446372986,
      "learning_rate": 2e-05,
      "loss": 0.20098731517791749,
      "mean_token_accuracy": 0.9440871596336364,
      "num_tokens": 51034235.0,
      "step": 3170
    },
    {
      "aux_loss": 8.11236925125122,
      "entropy": 0.1487548641860485,
      "epoch": 0.003175,
      "grad_norm": 0.162531778216362,
      "learning_rate": 2e-05,
      "loss": 0.17819650173187257,
      "mean_token_accuracy": 0.9502140015363694,
      "num_tokens": 51114583.0,
      "step": 3175
    },
    {
      "aux_loss": 8.091565465927124,
      "entropy": 0.1420391496270895,
      "epoch": 0.00318,
      "grad_norm": 0.14105230569839478,
      "learning_rate": 2e-05,
      "loss": 0.1813441276550293,
      "mean_token_accuracy": 0.9506178289651871,
      "num_tokens": 51193495.0,
      "step": 3180
    },
    {
      "aux_loss": 8.07268500328064,
      "entropy": 0.17803226746618747,
      "epoch": 0.003185,
      "grad_norm": 0.15835879743099213,
      "learning_rate": 2e-05,
      "loss": 0.23109543323516846,
      "mean_token_accuracy": 0.9360948204994202,
      "num_tokens": 51274914.0,
      "step": 3185
    },
    {
      "aux_loss": 8.079576635360718,
      "entropy": 0.16796971559524537,
      "epoch": 0.00319,
      "grad_norm": 0.14183682203292847,
      "learning_rate": 2e-05,
      "loss": 0.1991797685623169,
      "mean_token_accuracy": 0.9444876432418823,
      "num_tokens": 51356123.0,
      "step": 3190
    },
    {
      "aux_loss": 8.0917893409729,
      "entropy": 0.19391123130917548,
      "epoch": 0.003195,
      "grad_norm": 0.14483098685741425,
      "learning_rate": 2e-05,
      "loss": 0.21371831893920898,
      "mean_token_accuracy": 0.9389345049858093,
      "num_tokens": 51436905.0,
      "step": 3195
    },
    {
      "aux_loss": 8.096118450164795,
      "entropy": 0.1660480782389641,
      "epoch": 0.0032,
      "grad_norm": 0.13997644186019897,
      "learning_rate": 2e-05,
      "loss": 0.18765841722488402,
      "mean_token_accuracy": 0.9476926326751709,
      "num_tokens": 51517505.0,
      "step": 3200
    },
    {
      "aux_loss": 8.113101434707641,
      "entropy": 0.14779697582125664,
      "epoch": 0.003205,
      "grad_norm": 0.1604810208082199,
      "learning_rate": 2e-05,
      "loss": 0.18181984424591063,
      "mean_token_accuracy": 0.9491806507110596,
      "num_tokens": 51597899.0,
      "step": 3205
    },
    {
      "aux_loss": 8.091599655151366,
      "entropy": 0.1628108948469162,
      "epoch": 0.00321,
      "grad_norm": 0.15442314743995667,
      "learning_rate": 2e-05,
      "loss": 0.2045971155166626,
      "mean_token_accuracy": 0.9435444265604019,
      "num_tokens": 51674792.0,
      "step": 3210
    },
    {
      "aux_loss": 8.073364400863648,
      "entropy": 0.16871255934238433,
      "epoch": 0.003215,
      "grad_norm": 0.1368168443441391,
      "learning_rate": 2e-05,
      "loss": 0.21433780193328858,
      "mean_token_accuracy": 0.9394196689128875,
      "num_tokens": 51756132.0,
      "step": 3215
    },
    {
      "aux_loss": 8.079710388183594,
      "entropy": 0.18508594259619712,
      "epoch": 0.00322,
      "grad_norm": 0.1367693841457367,
      "learning_rate": 2e-05,
      "loss": 0.21680617332458496,
      "mean_token_accuracy": 0.9394427359104156,
      "num_tokens": 51837220.0,
      "step": 3220
    },
    {
      "aux_loss": 8.088467407226563,
      "entropy": 0.18197764903306962,
      "epoch": 0.003225,
      "grad_norm": 0.1577661782503128,
      "learning_rate": 2e-05,
      "loss": 0.20405569076538085,
      "mean_token_accuracy": 0.9435792595148087,
      "num_tokens": 51917893.0,
      "step": 3225
    },
    {
      "aux_loss": 8.099485826492309,
      "entropy": 0.15860071107745172,
      "epoch": 0.00323,
      "grad_norm": 0.14509201049804688,
      "learning_rate": 2e-05,
      "loss": 0.1842892050743103,
      "mean_token_accuracy": 0.9479869216680527,
      "num_tokens": 51998477.0,
      "step": 3230
    },
    {
      "aux_loss": 8.114618968963622,
      "entropy": 0.15259107425808907,
      "epoch": 0.003235,
      "grad_norm": 0.1713511049747467,
      "learning_rate": 2e-05,
      "loss": 0.18523633480072021,
      "mean_token_accuracy": 0.9479565292596817,
      "num_tokens": 52079119.0,
      "step": 3235
    },
    {
      "aux_loss": 8.089430856704713,
      "entropy": 0.14369250796735286,
      "epoch": 0.00324,
      "grad_norm": 0.15875156223773956,
      "learning_rate": 2e-05,
      "loss": 0.18524037599563598,
      "mean_token_accuracy": 0.9484268635511398,
      "num_tokens": 52159396.0,
      "step": 3240
    },
    {
      "aux_loss": 8.072626447677612,
      "entropy": 0.1836976647377014,
      "epoch": 0.003245,
      "grad_norm": 0.14659079909324646,
      "learning_rate": 2e-05,
      "loss": 0.22915871143341066,
      "mean_token_accuracy": 0.9364413321018219,
      "num_tokens": 52240466.0,
      "step": 3245
    },
    {
      "aux_loss": 8.078272390365601,
      "entropy": 0.1815402589738369,
      "epoch": 0.00325,
      "grad_norm": 0.1364625096321106,
      "learning_rate": 2e-05,
      "loss": 0.20854880809783935,
      "mean_token_accuracy": 0.9424823135137558,
      "num_tokens": 52321440.0,
      "step": 3250
    },
    {
      "aux_loss": 8.0862726688385,
      "entropy": 0.17093037217855453,
      "epoch": 0.003255,
      "grad_norm": 0.146819069981575,
      "learning_rate": 2e-05,
      "loss": 0.19406936168670655,
      "mean_token_accuracy": 0.9462387025356293,
      "num_tokens": 52402267.0,
      "step": 3255
    },
    {
      "aux_loss": 8.100034379959107,
      "entropy": 0.1684187613427639,
      "epoch": 0.00326,
      "grad_norm": 0.1474911868572235,
      "learning_rate": 2e-05,
      "loss": 0.19510631561279296,
      "mean_token_accuracy": 0.9447874993085861,
      "num_tokens": 52482738.0,
      "step": 3260
    },
    {
      "aux_loss": 8.113502073287965,
      "entropy": 0.14618208706378938,
      "epoch": 0.003265,
      "grad_norm": 0.14918093383312225,
      "learning_rate": 2e-05,
      "loss": 0.17784619331359863,
      "mean_token_accuracy": 0.9503018856048584,
      "num_tokens": 52563073.0,
      "step": 3265
    },
    {
      "aux_loss": 8.086842918395996,
      "entropy": 0.1440690655261278,
      "epoch": 0.00327,
      "grad_norm": 0.14992688596248627,
      "learning_rate": 2e-05,
      "loss": 0.18557535409927367,
      "mean_token_accuracy": 0.9500605285167694,
      "num_tokens": 52642342.0,
      "step": 3270
    },
    {
      "aux_loss": 8.069199800491333,
      "entropy": 0.16331303752958776,
      "epoch": 0.003275,
      "grad_norm": 0.15484365820884705,
      "learning_rate": 2e-05,
      "loss": 0.20972805023193358,
      "mean_token_accuracy": 0.9413518786430359,
      "num_tokens": 52723373.0,
      "step": 3275
    },
    {
      "aux_loss": 8.07596459388733,
      "entropy": 0.17387017607688904,
      "epoch": 0.00328,
      "grad_norm": 0.13718627393245697,
      "learning_rate": 2e-05,
      "loss": 0.2032240152359009,
      "mean_token_accuracy": 0.9430759161710739,
      "num_tokens": 52804217.0,
      "step": 3280
    },
    {
      "aux_loss": 8.083146572113037,
      "entropy": 0.17956893593072892,
      "epoch": 0.003285,
      "grad_norm": 0.1390460878610611,
      "learning_rate": 2e-05,
      "loss": 0.20335347652435304,
      "mean_token_accuracy": 0.9434962630271911,
      "num_tokens": 52884823.0,
      "step": 3285
    },
    {
      "aux_loss": 8.096046590805054,
      "entropy": 0.1636742062866688,
      "epoch": 0.00329,
      "grad_norm": 0.15083882212638855,
      "learning_rate": 2e-05,
      "loss": 0.18862991333007811,
      "mean_token_accuracy": 0.9463885605335236,
      "num_tokens": 52965238.0,
      "step": 3290
    },
    {
      "aux_loss": 8.109536933898926,
      "entropy": 0.14898678436875343,
      "epoch": 0.003295,
      "grad_norm": 0.1536809504032135,
      "learning_rate": 2e-05,
      "loss": 0.17851805686950684,
      "mean_token_accuracy": 0.9510477185249329,
      "num_tokens": 53045705.0,
      "step": 3295
    },
    {
      "aux_loss": 8.088180685043335,
      "entropy": 0.16010745838284493,
      "epoch": 0.0033,
      "grad_norm": 0.14198707044124603,
      "learning_rate": 2e-05,
      "loss": 0.19941191673278807,
      "mean_token_accuracy": 0.9460611671209336,
      "num_tokens": 53126029.0,
      "step": 3300
    },
    {
      "aux_loss": 8.072169733047485,
      "entropy": 0.17422449737787246,
      "epoch": 0.003305,
      "grad_norm": 0.1427670568227768,
      "learning_rate": 2e-05,
      "loss": 0.21760165691375732,
      "mean_token_accuracy": 0.9400460362434387,
      "num_tokens": 53207154.0,
      "step": 3305
    },
    {
      "aux_loss": 8.077490329742432,
      "entropy": 0.1694101832807064,
      "epoch": 0.00331,
      "grad_norm": 0.1344936639070511,
      "learning_rate": 2e-05,
      "loss": 0.20693485736846923,
      "mean_token_accuracy": 0.9421370506286622,
      "num_tokens": 53288152.0,
      "step": 3310
    },
    {
      "aux_loss": 8.08360242843628,
      "entropy": 0.1677086018025875,
      "epoch": 0.003315,
      "grad_norm": 0.13482902944087982,
      "learning_rate": 2e-05,
      "loss": 0.19454469680786132,
      "mean_token_accuracy": 0.9452098190784455,
      "num_tokens": 53369094.0,
      "step": 3315
    },
    {
      "aux_loss": 8.094222927093506,
      "entropy": 0.17268427312374116,
      "epoch": 0.00332,
      "grad_norm": 0.1527445763349533,
      "learning_rate": 2e-05,
      "loss": 0.19566521644592286,
      "mean_token_accuracy": 0.945321798324585,
      "num_tokens": 53449882.0,
      "step": 3320
    },
    {
      "aux_loss": 8.110411405563354,
      "entropy": 0.15340660512447357,
      "epoch": 0.003325,
      "grad_norm": 0.163923978805542,
      "learning_rate": 2e-05,
      "loss": 0.18321604728698732,
      "mean_token_accuracy": 0.9484834492206573,
      "num_tokens": 53530134.0,
      "step": 3325
    },
    {
      "aux_loss": 8.09237003326416,
      "entropy": 0.15732431635260583,
      "epoch": 0.00333,
      "grad_norm": 0.13197200000286102,
      "learning_rate": 2e-05,
      "loss": 0.19181399345397948,
      "mean_token_accuracy": 0.948565936088562,
      "num_tokens": 53607669.0,
      "step": 3330
    },
    {
      "aux_loss": 8.073115730285645,
      "entropy": 0.18500021658837795,
      "epoch": 0.003335,
      "grad_norm": 0.14987976849079132,
      "learning_rate": 2e-05,
      "loss": 0.2311016082763672,
      "mean_token_accuracy": 0.9357433319091797,
      "num_tokens": 53689066.0,
      "step": 3335
    },
    {
      "aux_loss": 8.077186584472656,
      "entropy": 0.1729499664157629,
      "epoch": 0.00334,
      "grad_norm": 0.13073405623435974,
      "learning_rate": 2e-05,
      "loss": 0.20392858982086182,
      "mean_token_accuracy": 0.9428721487522125,
      "num_tokens": 53770472.0,
      "step": 3340
    },
    {
      "aux_loss": 8.086996364593507,
      "entropy": 0.1811000734567642,
      "epoch": 0.003345,
      "grad_norm": 0.13385631144046783,
      "learning_rate": 2e-05,
      "loss": 0.20783238410949706,
      "mean_token_accuracy": 0.9418109148740769,
      "num_tokens": 53851149.0,
      "step": 3345
    },
    {
      "aux_loss": 8.096043014526368,
      "entropy": 0.167522644251585,
      "epoch": 0.00335,
      "grad_norm": 0.14239554107189178,
      "learning_rate": 2e-05,
      "loss": 0.18875793218612671,
      "mean_token_accuracy": 0.9475765883922577,
      "num_tokens": 53931702.0,
      "step": 3350
    },
    {
      "aux_loss": 8.111851119995118,
      "entropy": 0.1472678080201149,
      "epoch": 0.003355,
      "grad_norm": 0.1470818966627121,
      "learning_rate": 2e-05,
      "loss": 0.1803680419921875,
      "mean_token_accuracy": 0.949494594335556,
      "num_tokens": 54011870.0,
      "step": 3355
    },
    {
      "aux_loss": 8.096990203857422,
      "entropy": 0.13597517758607863,
      "epoch": 0.00336,
      "grad_norm": 0.14512929320335388,
      "learning_rate": 2e-05,
      "loss": 0.177366304397583,
      "mean_token_accuracy": 0.9522892445325851,
      "num_tokens": 54091666.0,
      "step": 3360
    },
    {
      "aux_loss": 8.069016027450562,
      "entropy": 0.16606580428779125,
      "epoch": 0.003365,
      "grad_norm": 0.1469336450099945,
      "learning_rate": 2e-05,
      "loss": 0.2099947452545166,
      "mean_token_accuracy": 0.9417406618595123,
      "num_tokens": 54172689.0,
      "step": 3365
    },
    {
      "aux_loss": 8.073838710784912,
      "entropy": 0.1717648908495903,
      "epoch": 0.00337,
      "grad_norm": 0.13721416890621185,
      "learning_rate": 2e-05,
      "loss": 0.2064809560775757,
      "mean_token_accuracy": 0.9425877064466477,
      "num_tokens": 54253733.0,
      "step": 3370
    },
    {
      "aux_loss": 8.08399519920349,
      "entropy": 0.17570768669247627,
      "epoch": 0.003375,
      "grad_norm": 0.1387743204832077,
      "learning_rate": 2e-05,
      "loss": 0.2002772331237793,
      "mean_token_accuracy": 0.9443808704614639,
      "num_tokens": 54334687.0,
      "step": 3375
    },
    {
      "aux_loss": 8.093376922607423,
      "entropy": 0.17167234867811204,
      "epoch": 0.00338,
      "grad_norm": 0.14760684967041016,
      "learning_rate": 2e-05,
      "loss": 0.19618290662765503,
      "mean_token_accuracy": 0.9448229610919953,
      "num_tokens": 54415453.0,
      "step": 3380
    },
    {
      "aux_loss": 8.107440710067749,
      "entropy": 0.15284837558865547,
      "epoch": 0.003385,
      "grad_norm": 0.14630548655986786,
      "learning_rate": 2e-05,
      "loss": 0.1791166067123413,
      "mean_token_accuracy": 0.9502524435520172,
      "num_tokens": 54495604.0,
      "step": 3385
    },
    {
      "aux_loss": 8.091902589797973,
      "entropy": 0.14478393606841564,
      "epoch": 0.00339,
      "grad_norm": 0.16304811835289001,
      "learning_rate": 2e-05,
      "loss": 0.18242955207824707,
      "mean_token_accuracy": 0.9496553808450698,
      "num_tokens": 54576298.0,
      "step": 3390
    },
    {
      "aux_loss": 8.067150974273682,
      "entropy": 0.1716588746756315,
      "epoch": 0.003395,
      "grad_norm": 0.16855032742023468,
      "learning_rate": 2e-05,
      "loss": 0.2224963426589966,
      "mean_token_accuracy": 0.9379703015089035,
      "num_tokens": 54657649.0,
      "step": 3395
    },
    {
      "aux_loss": 8.074144315719604,
      "entropy": 0.17227399796247483,
      "epoch": 0.0034,
      "grad_norm": 0.14275826513767242,
      "learning_rate": 2e-05,
      "loss": 0.2041255474090576,
      "mean_token_accuracy": 0.9419350147247314,
      "num_tokens": 54738703.0,
      "step": 3400
    },
    {
      "aux_loss": 8.084643983840943,
      "entropy": 0.1803322732448578,
      "epoch": 0.003405,
      "grad_norm": 0.1536228209733963,
      "learning_rate": 2e-05,
      "loss": 0.20024464130401612,
      "mean_token_accuracy": 0.9435871928930283,
      "num_tokens": 54819670.0,
      "step": 3405
    },
    {
      "aux_loss": 8.096064472198487,
      "entropy": 0.16384155228734015,
      "epoch": 0.00341,
      "grad_norm": 0.1480627804994583,
      "learning_rate": 2e-05,
      "loss": 0.18860028982162474,
      "mean_token_accuracy": 0.9474248379468918,
      "num_tokens": 54900088.0,
      "step": 3410
    },
    {
      "aux_loss": 8.109060859680175,
      "entropy": 0.14899858981370925,
      "epoch": 0.003415,
      "grad_norm": 0.15656854212284088,
      "learning_rate": 2e-05,
      "loss": 0.18060693740844727,
      "mean_token_accuracy": 0.9499949842691422,
      "num_tokens": 54980631.0,
      "step": 3415
    },
    {
      "aux_loss": 8.095376110076904,
      "entropy": 0.13952750600874425,
      "epoch": 0.00342,
      "grad_norm": 0.14643265306949615,
      "learning_rate": 2e-05,
      "loss": 0.17561327219009398,
      "mean_token_accuracy": 0.9521472573280334,
      "num_tokens": 55059774.0,
      "step": 3420
    },
    {
      "aux_loss": 8.070878791809083,
      "entropy": 0.1734847042709589,
      "epoch": 0.003425,
      "grad_norm": 0.14761346578598022,
      "learning_rate": 2e-05,
      "loss": 0.22200329303741456,
      "mean_token_accuracy": 0.9386822372674942,
      "num_tokens": 55140940.0,
      "step": 3425
    },
    {
      "aux_loss": 8.071926259994507,
      "entropy": 0.18360228911042215,
      "epoch": 0.00343,
      "grad_norm": 0.1406898945569992,
      "learning_rate": 2e-05,
      "loss": 0.21827876567840576,
      "mean_token_accuracy": 0.9393490523099899,
      "num_tokens": 55221990.0,
      "step": 3430
    },
    {
      "aux_loss": 8.083148193359374,
      "entropy": 0.16763183027505874,
      "epoch": 0.003435,
      "grad_norm": 0.14203479886054993,
      "learning_rate": 2e-05,
      "loss": 0.1955203652381897,
      "mean_token_accuracy": 0.9449512451887131,
      "num_tokens": 55302530.0,
      "step": 3435
    },
    {
      "aux_loss": 8.091049718856812,
      "entropy": 0.16986002177000045,
      "epoch": 0.00344,
      "grad_norm": 0.14416195452213287,
      "learning_rate": 2e-05,
      "loss": 0.19449214935302733,
      "mean_token_accuracy": 0.9441195905208588,
      "num_tokens": 55383244.0,
      "step": 3440
    },
    {
      "aux_loss": 8.105107021331786,
      "entropy": 0.15056025087833405,
      "epoch": 0.003445,
      "grad_norm": 0.15131866931915283,
      "learning_rate": 2e-05,
      "loss": 0.17404098510742189,
      "mean_token_accuracy": 0.9523547291755676,
      "num_tokens": 55463258.0,
      "step": 3445
    },
    {
      "aux_loss": 8.098848581314087,
      "entropy": 0.1485459551215172,
      "epoch": 0.00345,
      "grad_norm": 0.1486630141735077,
      "learning_rate": 2e-05,
      "loss": 0.18512657880783082,
      "mean_token_accuracy": 0.9497396767139434,
      "num_tokens": 55542988.0,
      "step": 3450
    },
    {
      "aux_loss": 8.071794891357422,
      "entropy": 0.1657637305557728,
      "epoch": 0.003455,
      "grad_norm": 0.1553289145231247,
      "learning_rate": 2e-05,
      "loss": 0.21695096492767335,
      "mean_token_accuracy": 0.9409345209598541,
      "num_tokens": 55624496.0,
      "step": 3455
    },
    {
      "aux_loss": 8.07252402305603,
      "entropy": 0.16912850216031075,
      "epoch": 0.00346,
      "grad_norm": 0.14527633786201477,
      "learning_rate": 2e-05,
      "loss": 0.20679042339324952,
      "mean_token_accuracy": 0.9425709515810012,
      "num_tokens": 55705532.0,
      "step": 3460
    },
    {
      "aux_loss": 8.080512189865113,
      "entropy": 0.17733409106731415,
      "epoch": 0.003465,
      "grad_norm": 0.14004933834075928,
      "learning_rate": 2e-05,
      "loss": 0.20295884609222412,
      "mean_token_accuracy": 0.9436537146568298,
      "num_tokens": 55786302.0,
      "step": 3465
    },
    {
      "aux_loss": 8.09162197113037,
      "entropy": 0.17465185448527337,
      "epoch": 0.00347,
      "grad_norm": 0.14595374464988708,
      "learning_rate": 2e-05,
      "loss": 0.19612003564834596,
      "mean_token_accuracy": 0.9450876235961914,
      "num_tokens": 55867205.0,
      "step": 3470
    },
    {
      "aux_loss": 8.104484224319458,
      "entropy": 0.15324413031339645,
      "epoch": 0.003475,
      "grad_norm": 0.14978156983852386,
      "learning_rate": 2e-05,
      "loss": 0.18199578523635865,
      "mean_token_accuracy": 0.9488621115684509,
      "num_tokens": 55947604.0,
      "step": 3475
    },
    {
      "aux_loss": 8.105511283874511,
      "entropy": 0.14812915921211242,
      "epoch": 0.00348,
      "grad_norm": 0.13192299008369446,
      "learning_rate": 2e-05,
      "loss": 0.17839200496673585,
      "mean_token_accuracy": 0.9508434116840363,
      "num_tokens": 56024999.0,
      "step": 3480
    },
    {
      "aux_loss": 8.066703128814698,
      "entropy": 0.1732874248176813,
      "epoch": 0.003485,
      "grad_norm": 0.1783316731452942,
      "learning_rate": 2e-05,
      "loss": 0.22596149444580077,
      "mean_token_accuracy": 0.9383738577365875,
      "num_tokens": 56106248.0,
      "step": 3485
    },
    {
      "aux_loss": 8.074977588653564,
      "entropy": 0.1672011435031891,
      "epoch": 0.00349,
      "grad_norm": 0.12999005615711212,
      "learning_rate": 2e-05,
      "loss": 0.20716416835784912,
      "mean_token_accuracy": 0.9420315384864807,
      "num_tokens": 56187356.0,
      "step": 3490
    },
    {
      "aux_loss": 8.084104919433594,
      "entropy": 0.17579501047730445,
      "epoch": 0.003495,
      "grad_norm": 0.14199404418468475,
      "learning_rate": 2e-05,
      "loss": 0.19671993255615233,
      "mean_token_accuracy": 0.946016663312912,
      "num_tokens": 56268163.0,
      "step": 3495
    },
    {
      "aux_loss": 8.095600652694703,
      "entropy": 0.1678522802889347,
      "epoch": 0.0035,
      "grad_norm": 0.14334514737129211,
      "learning_rate": 2e-05,
      "loss": 0.1883704900741577,
      "mean_token_accuracy": 0.9474934220314026,
      "num_tokens": 56348970.0,
      "step": 3500
    },
    {
      "aux_loss": 8.109538269042968,
      "entropy": 0.14645420126616954,
      "epoch": 0.003505,
      "grad_norm": 0.15478180348873138,
      "learning_rate": 2e-05,
      "loss": 0.17634756565093995,
      "mean_token_accuracy": 0.9506915777921676,
      "num_tokens": 56429279.0,
      "step": 3505
    },
    {
      "aux_loss": 8.102639722824097,
      "entropy": 0.1441028606146574,
      "epoch": 0.00351,
      "grad_norm": 0.13839173316955566,
      "learning_rate": 2e-05,
      "loss": 0.1813260793685913,
      "mean_token_accuracy": 0.9499836087226867,
      "num_tokens": 56509051.0,
      "step": 3510
    },
    {
      "aux_loss": 8.074317789077758,
      "entropy": 0.17779307663440705,
      "epoch": 0.003515,
      "grad_norm": 0.14687684178352356,
      "learning_rate": 2e-05,
      "loss": 0.22249352931976318,
      "mean_token_accuracy": 0.9384042322635651,
      "num_tokens": 56590270.0,
      "step": 3515
    },
    {
      "aux_loss": 8.078415966033935,
      "entropy": 0.1729664221405983,
      "epoch": 0.00352,
      "grad_norm": 0.13611707091331482,
      "learning_rate": 2e-05,
      "loss": 0.20582201480865478,
      "mean_token_accuracy": 0.9415476769208908,
      "num_tokens": 56671323.0,
      "step": 3520
    },
    {
      "aux_loss": 8.087087154388428,
      "entropy": 0.1833976648747921,
      "epoch": 0.003525,
      "grad_norm": 0.1398952603340149,
      "learning_rate": 2e-05,
      "loss": 0.2122587203979492,
      "mean_token_accuracy": 0.9403575837612153,
      "num_tokens": 56752246.0,
      "step": 3525
    },
    {
      "aux_loss": 8.096710443496704,
      "entropy": 0.17299836575984956,
      "epoch": 0.00353,
      "grad_norm": 0.13910555839538574,
      "learning_rate": 2e-05,
      "loss": 0.19738600254058838,
      "mean_token_accuracy": 0.9448669463396072,
      "num_tokens": 56832808.0,
      "step": 3530
    },
    {
      "aux_loss": 8.10430006980896,
      "entropy": 0.15525901839137077,
      "epoch": 0.003535,
      "grad_norm": 0.15603163838386536,
      "learning_rate": 2e-05,
      "loss": 0.1801411509513855,
      "mean_token_accuracy": 0.9500389933586121,
      "num_tokens": 56913267.0,
      "step": 3535
    },
    {
      "aux_loss": 8.091502237319947,
      "entropy": 0.14775969237089157,
      "epoch": 0.00354,
      "grad_norm": 0.13365484774112701,
      "learning_rate": 2e-05,
      "loss": 0.18674581050872802,
      "mean_token_accuracy": 0.9489086121320724,
      "num_tokens": 56994050.0,
      "step": 3540
    },
    {
      "aux_loss": 8.068204164505005,
      "entropy": 0.1635946251451969,
      "epoch": 0.003545,
      "grad_norm": 0.15203210711479187,
      "learning_rate": 2e-05,
      "loss": 0.21183254718780517,
      "mean_token_accuracy": 0.9427683711051941,
      "num_tokens": 57075181.0,
      "step": 3545
    },
    {
      "aux_loss": 8.076211643218993,
      "entropy": 0.17809797935187816,
      "epoch": 0.00355,
      "grad_norm": 0.1417941451072693,
      "learning_rate": 2e-05,
      "loss": 0.21140050888061523,
      "mean_token_accuracy": 0.9404971837997437,
      "num_tokens": 57156292.0,
      "step": 3550
    },
    {
      "aux_loss": 8.084080123901368,
      "entropy": 0.17739937528967858,
      "epoch": 0.003555,
      "grad_norm": 0.1385902315378189,
      "learning_rate": 2e-05,
      "loss": 0.20577735900878907,
      "mean_token_accuracy": 0.9431796312332154,
      "num_tokens": 57237218.0,
      "step": 3555
    },
    {
      "aux_loss": 8.094055414199829,
      "entropy": 0.1631332315504551,
      "epoch": 0.00356,
      "grad_norm": 0.14270524680614471,
      "learning_rate": 2e-05,
      "loss": 0.18669048547744752,
      "mean_token_accuracy": 0.9465743929147721,
      "num_tokens": 57318047.0,
      "step": 3560
    },
    {
      "aux_loss": 8.10590057373047,
      "entropy": 0.1550621699541807,
      "epoch": 0.003565,
      "grad_norm": 0.14862309396266937,
      "learning_rate": 2e-05,
      "loss": 0.18273617029190065,
      "mean_token_accuracy": 0.9494030088186264,
      "num_tokens": 57398501.0,
      "step": 3565
    },
    {
      "aux_loss": 8.1013503074646,
      "entropy": 0.15278596021234989,
      "epoch": 0.00357,
      "grad_norm": 0.1578340083360672,
      "learning_rate": 2e-05,
      "loss": 0.19054425954818727,
      "mean_token_accuracy": 0.9464311182498932,
      "num_tokens": 57476607.0,
      "step": 3570
    },
    {
      "aux_loss": 8.06883101463318,
      "entropy": 0.16016253717243673,
      "epoch": 0.003575,
      "grad_norm": 0.13430093228816986,
      "learning_rate": 2e-05,
      "loss": 0.20853614807128906,
      "mean_token_accuracy": 0.9426379173994064,
      "num_tokens": 57557852.0,
      "step": 3575
    },
    {
      "aux_loss": 8.073803567886353,
      "entropy": 0.16622515581548214,
      "epoch": 0.00358,
      "grad_norm": 0.1397142857313156,
      "learning_rate": 2e-05,
      "loss": 0.2018742084503174,
      "mean_token_accuracy": 0.9440779864788056,
      "num_tokens": 57638613.0,
      "step": 3580
    },
    {
      "aux_loss": 8.085051679611206,
      "entropy": 0.16953913792967795,
      "epoch": 0.003585,
      "grad_norm": 0.13832564651966095,
      "learning_rate": 2e-05,
      "loss": 0.20050559043884278,
      "mean_token_accuracy": 0.9444626539945602,
      "num_tokens": 57719544.0,
      "step": 3585
    },
    {
      "aux_loss": 8.095213031768798,
      "entropy": 0.17049626186490058,
      "epoch": 0.00359,
      "grad_norm": 0.1392274647951126,
      "learning_rate": 2e-05,
      "loss": 0.19836210012435912,
      "mean_token_accuracy": 0.9441302388906478,
      "num_tokens": 57800028.0,
      "step": 3590
    },
    {
      "aux_loss": 8.1052903175354,
      "entropy": 0.15605995878577233,
      "epoch": 0.003595,
      "grad_norm": 0.14226338267326355,
      "learning_rate": 2e-05,
      "loss": 0.18160698413848878,
      "mean_token_accuracy": 0.9485923439264298,
      "num_tokens": 57880482.0,
      "step": 3595
    },
    {
      "aux_loss": 8.103016996383667,
      "entropy": 0.1513067062944174,
      "epoch": 0.0036,
      "grad_norm": 0.13649244606494904,
      "learning_rate": 2e-05,
      "loss": 0.1775503635406494,
      "mean_token_accuracy": 0.9513054430484772,
      "num_tokens": 57958701.0,
      "step": 3600
    },
    {
      "aux_loss": 8.068546104431153,
      "entropy": 0.1584773451089859,
      "epoch": 0.003605,
      "grad_norm": 0.14773598313331604,
      "learning_rate": 2e-05,
      "loss": 0.20502164363861083,
      "mean_token_accuracy": 0.9424826383590699,
      "num_tokens": 58039900.0,
      "step": 3605
    },
    {
      "aux_loss": 8.07251763343811,
      "entropy": 0.16773774065077304,
      "epoch": 0.00361,
      "grad_norm": 0.13793109357357025,
      "learning_rate": 2e-05,
      "loss": 0.21010069847106932,
      "mean_token_accuracy": 0.9423064380884171,
      "num_tokens": 58120837.0,
      "step": 3610
    },
    {
      "aux_loss": 8.081832790374756,
      "entropy": 0.16447483748197556,
      "epoch": 0.003615,
      "grad_norm": 0.12927091121673584,
      "learning_rate": 2e-05,
      "loss": 0.19288171529769899,
      "mean_token_accuracy": 0.9468847870826721,
      "num_tokens": 58201567.0,
      "step": 3615
    },
    {
      "aux_loss": 8.092286443710327,
      "entropy": 0.1687350258231163,
      "epoch": 0.00362,
      "grad_norm": 0.14320580661296844,
      "learning_rate": 2e-05,
      "loss": 0.1897109031677246,
      "mean_token_accuracy": 0.9469810336828232,
      "num_tokens": 58282395.0,
      "step": 3620
    },
    {
      "aux_loss": 8.10400595664978,
      "entropy": 0.15696103647351264,
      "epoch": 0.003625,
      "grad_norm": 0.1543847620487213,
      "learning_rate": 2e-05,
      "loss": 0.1850444793701172,
      "mean_token_accuracy": 0.9474167734384537,
      "num_tokens": 58362730.0,
      "step": 3625
    },
    {
      "aux_loss": 8.10280966758728,
      "entropy": 0.1554012343287468,
      "epoch": 0.00363,
      "grad_norm": 0.1395951807498932,
      "learning_rate": 2e-05,
      "loss": 0.1888120412826538,
      "mean_token_accuracy": 0.947870847582817,
      "num_tokens": 58439441.0,
      "step": 3630
    },
    {
      "aux_loss": 8.07128839492798,
      "entropy": 0.16419425942003726,
      "epoch": 0.003635,
      "grad_norm": 0.15131539106369019,
      "learning_rate": 2e-05,
      "loss": 0.20569977760314942,
      "mean_token_accuracy": 0.94364333152771,
      "num_tokens": 58520811.0,
      "step": 3635
    },
    {
      "aux_loss": 8.074514245986938,
      "entropy": 0.1789974257349968,
      "epoch": 0.00364,
      "grad_norm": 0.1431204080581665,
      "learning_rate": 2e-05,
      "loss": 0.22197389602661133,
      "mean_token_accuracy": 0.9386385768651963,
      "num_tokens": 58601757.0,
      "step": 3640
    },
    {
      "aux_loss": 8.079972505569458,
      "entropy": 0.17981437221169472,
      "epoch": 0.003645,
      "grad_norm": 0.13867253065109253,
      "learning_rate": 2e-05,
      "loss": 0.2087554931640625,
      "mean_token_accuracy": 0.9421653658151626,
      "num_tokens": 58682738.0,
      "step": 3645
    },
    {
      "aux_loss": 8.09368281364441,
      "entropy": 0.17152813225984573,
      "epoch": 0.00365,
      "grad_norm": 0.14049142599105835,
      "learning_rate": 2e-05,
      "loss": 0.1897148847579956,
      "mean_token_accuracy": 0.9465822011232377,
      "num_tokens": 58763483.0,
      "step": 3650
    },
    {
      "aux_loss": 8.101623916625977,
      "entropy": 0.1555980086326599,
      "epoch": 0.003655,
      "grad_norm": 0.1593800038099289,
      "learning_rate": 2e-05,
      "loss": 0.18208436965942382,
      "mean_token_accuracy": 0.9497519314289093,
      "num_tokens": 58844426.0,
      "step": 3655
    },
    {
      "aux_loss": 8.104884815216064,
      "entropy": 0.14434664137661457,
      "epoch": 0.00366,
      "grad_norm": 0.12744976580142975,
      "learning_rate": 2e-05,
      "loss": 0.17775565385818481,
      "mean_token_accuracy": 0.9501483708620071,
      "num_tokens": 58921220.0,
      "step": 3660
    },
    {
      "aux_loss": 8.065509510040282,
      "entropy": 0.17364537231624128,
      "epoch": 0.003665,
      "grad_norm": 0.14052312076091766,
      "learning_rate": 2e-05,
      "loss": 0.21997876167297364,
      "mean_token_accuracy": 0.9393134415149689,
      "num_tokens": 59002459.0,
      "step": 3665
    },
    {
      "aux_loss": 8.074805450439452,
      "entropy": 0.16845594197511674,
      "epoch": 0.00367,
      "grad_norm": 0.13931669294834137,
      "learning_rate": 2e-05,
      "loss": 0.20685987472534179,
      "mean_token_accuracy": 0.9421530127525329,
      "num_tokens": 59083041.0,
      "step": 3670
    },
    {
      "aux_loss": 8.080971336364746,
      "entropy": 0.17533733174204827,
      "epoch": 0.003675,
      "grad_norm": 0.136427640914917,
      "learning_rate": 2e-05,
      "loss": 0.2057826042175293,
      "mean_token_accuracy": 0.9421284943819046,
      "num_tokens": 59164100.0,
      "step": 3675
    },
    {
      "aux_loss": 8.092462491989135,
      "entropy": 0.16860179007053375,
      "epoch": 0.00368,
      "grad_norm": 0.13490749895572662,
      "learning_rate": 2e-05,
      "loss": 0.19366698265075682,
      "mean_token_accuracy": 0.9455673635005951,
      "num_tokens": 59244953.0,
      "step": 3680
    },
    {
      "aux_loss": 8.102824401855468,
      "entropy": 0.14722870886325837,
      "epoch": 0.003685,
      "grad_norm": 0.14679794013500214,
      "learning_rate": 2e-05,
      "loss": 0.17461202144622803,
      "mean_token_accuracy": 0.9518853396177291,
      "num_tokens": 59325485.0,
      "step": 3685
    },
    {
      "aux_loss": 8.103275966644286,
      "entropy": 0.13416015766561032,
      "epoch": 0.00369,
      "grad_norm": 0.12525036931037903,
      "learning_rate": 2e-05,
      "loss": 0.1626958131790161,
      "mean_token_accuracy": 0.9550653249025345,
      "num_tokens": 59405516.0,
      "step": 3690
    },
    {
      "aux_loss": 8.067550325393677,
      "entropy": 0.1654042899608612,
      "epoch": 0.003695,
      "grad_norm": 0.13868121802806854,
      "learning_rate": 2e-05,
      "loss": 0.20632545948028563,
      "mean_token_accuracy": 0.9427901446819306,
      "num_tokens": 59486998.0,
      "step": 3695
    },
    {
      "aux_loss": 8.075068044662476,
      "entropy": 0.16679728627204896,
      "epoch": 0.0037,
      "grad_norm": 0.15251266956329346,
      "learning_rate": 2e-05,
      "loss": 0.21037230491638184,
      "mean_token_accuracy": 0.941617614030838,
      "num_tokens": 59567565.0,
      "step": 3700
    },
    {
      "aux_loss": 8.079337167739869,
      "entropy": 0.17589445263147355,
      "epoch": 0.003705,
      "grad_norm": 0.139875590801239,
      "learning_rate": 2e-05,
      "loss": 0.2073918104171753,
      "mean_token_accuracy": 0.9415038168430329,
      "num_tokens": 59648671.0,
      "step": 3705
    },
    {
      "aux_loss": 8.089571523666383,
      "entropy": 0.17697745338082313,
      "epoch": 0.00371,
      "grad_norm": 0.14507681131362915,
      "learning_rate": 2e-05,
      "loss": 0.2054443597793579,
      "mean_token_accuracy": 0.9419445693492889,
      "num_tokens": 59729407.0,
      "step": 3710
    },
    {
      "aux_loss": 8.099748992919922,
      "entropy": 0.16234676837921141,
      "epoch": 0.003715,
      "grad_norm": 0.14972840249538422,
      "learning_rate": 2e-05,
      "loss": 0.18711310625076294,
      "mean_token_accuracy": 0.9482452005147934,
      "num_tokens": 59810060.0,
      "step": 3715
    },
    {
      "aux_loss": 8.108324575424195,
      "entropy": 0.14043421484529972,
      "epoch": 0.00372,
      "grad_norm": 0.12313748896121979,
      "learning_rate": 2e-05,
      "loss": 0.16772583723068238,
      "mean_token_accuracy": 0.9538856446743011,
      "num_tokens": 59886481.0,
      "step": 3720
    },
    {
      "aux_loss": 8.066444635391235,
      "entropy": 0.16332019604742526,
      "epoch": 0.003725,
      "grad_norm": 0.1655113846063614,
      "learning_rate": 2e-05,
      "loss": 0.21046555042266846,
      "mean_token_accuracy": 0.9426675111055374,
      "num_tokens": 59968013.0,
      "step": 3725
    },
    {
      "aux_loss": 8.071371078491211,
      "entropy": 0.17669796496629714,
      "epoch": 0.00373,
      "grad_norm": 0.14448584616184235,
      "learning_rate": 2e-05,
      "loss": 0.22329461574554443,
      "mean_token_accuracy": 0.9376091629266738,
      "num_tokens": 60048893.0,
      "step": 3730
    },
    {
      "aux_loss": 8.07760648727417,
      "entropy": 0.1781152181327343,
      "epoch": 0.003735,
      "grad_norm": 0.13516509532928467,
      "learning_rate": 2e-05,
      "loss": 0.20711853504180908,
      "mean_token_accuracy": 0.9421719670295715,
      "num_tokens": 60129799.0,
      "step": 3735
    },
    {
      "aux_loss": 8.092101621627808,
      "entropy": 0.17709542885422708,
      "epoch": 0.00374,
      "grad_norm": 0.14650335907936096,
      "learning_rate": 2e-05,
      "loss": 0.19788501262664795,
      "mean_token_accuracy": 0.9451971143484116,
      "num_tokens": 60210518.0,
      "step": 3740
    },
    {
      "aux_loss": 8.099432039260865,
      "entropy": 0.15634507611393927,
      "epoch": 0.003745,
      "grad_norm": 0.13907134532928467,
      "learning_rate": 2e-05,
      "loss": 0.17914718389511108,
      "mean_token_accuracy": 0.9502253204584121,
      "num_tokens": 60291392.0,
      "step": 3745
    },
    {
      "aux_loss": 8.110324239730835,
      "entropy": 0.13975763842463493,
      "epoch": 0.00375,
      "grad_norm": 0.13600419461727142,
      "learning_rate": 2e-05,
      "loss": 0.16946213245391845,
      "mean_token_accuracy": 0.9522726953029632,
      "num_tokens": 60368839.0,
      "step": 3750
    },
    {
      "aux_loss": 8.067803287506104,
      "entropy": 0.17291736453771592,
      "epoch": 0.003755,
      "grad_norm": 0.19572490453720093,
      "learning_rate": 2e-05,
      "loss": 0.22342123985290527,
      "mean_token_accuracy": 0.9385370045900345,
      "num_tokens": 60449738.0,
      "step": 3755
    },
    {
      "aux_loss": 8.075034904479981,
      "entropy": 0.16723310016095638,
      "epoch": 0.00376,
      "grad_norm": 0.15154753625392914,
      "learning_rate": 2e-05,
      "loss": 0.20492987632751464,
      "mean_token_accuracy": 0.9426275312900543,
      "num_tokens": 60530692.0,
      "step": 3760
    },
    {
      "aux_loss": 8.079753303527832,
      "entropy": 0.17016384825110437,
      "epoch": 0.003765,
      "grad_norm": 0.1361101269721985,
      "learning_rate": 2e-05,
      "loss": 0.1980854630470276,
      "mean_token_accuracy": 0.9451626509428024,
      "num_tokens": 60611814.0,
      "step": 3765
    },
    {
      "aux_loss": 8.090311813354493,
      "entropy": 0.18353770449757575,
      "epoch": 0.00377,
      "grad_norm": 0.1370791345834732,
      "learning_rate": 2e-05,
      "loss": 0.20895819664001464,
      "mean_token_accuracy": 0.9414872795343399,
      "num_tokens": 60692659.0,
      "step": 3770
    },
    {
      "aux_loss": 8.101877975463868,
      "entropy": 0.15832723006606103,
      "epoch": 0.003775,
      "grad_norm": 0.1525694727897644,
      "learning_rate": 2e-05,
      "loss": 0.18391293287277222,
      "mean_token_accuracy": 0.9479856729507447,
      "num_tokens": 60773046.0,
      "step": 3775
    },
    {
      "aux_loss": 8.103589153289795,
      "entropy": 0.14184964299201966,
      "epoch": 0.00378,
      "grad_norm": 0.12755383551120758,
      "learning_rate": 2e-05,
      "loss": 0.16912662982940674,
      "mean_token_accuracy": 0.9529174774885177,
      "num_tokens": 60850842.0,
      "step": 3780
    },
    {
      "aux_loss": 8.067220973968507,
      "entropy": 0.17528922557830812,
      "epoch": 0.003785,
      "grad_norm": 0.14482465386390686,
      "learning_rate": 2e-05,
      "loss": 0.22777693271636962,
      "mean_token_accuracy": 0.9383696466684341,
      "num_tokens": 60932278.0,
      "step": 3785
    },
    {
      "aux_loss": 8.072800445556641,
      "entropy": 0.17815810069441795,
      "epoch": 0.00379,
      "grad_norm": 0.14308401942253113,
      "learning_rate": 2e-05,
      "loss": 0.22371666431427,
      "mean_token_accuracy": 0.9380281120538712,
      "num_tokens": 61012759.0,
      "step": 3790
    },
    {
      "aux_loss": 8.076924514770507,
      "entropy": 0.17671712972223758,
      "epoch": 0.003795,
      "grad_norm": 0.1398933380842209,
      "learning_rate": 2e-05,
      "loss": 0.2038175106048584,
      "mean_token_accuracy": 0.9430381059646606,
      "num_tokens": 61093403.0,
      "step": 3795
    },
    {
      "aux_loss": 8.087810611724853,
      "entropy": 0.18333133086562156,
      "epoch": 0.0038,
      "grad_norm": 0.14304576814174652,
      "learning_rate": 2e-05,
      "loss": 0.20525126457214354,
      "mean_token_accuracy": 0.9424682438373566,
      "num_tokens": 61173942.0,
      "step": 3800
    },
    {
      "aux_loss": 8.097368049621583,
      "entropy": 0.1570002555847168,
      "epoch": 0.003805,
      "grad_norm": 0.1716916561126709,
      "learning_rate": 2e-05,
      "loss": 0.18336018323898315,
      "mean_token_accuracy": 0.948911139369011,
      "num_tokens": 61254576.0,
      "step": 3805
    },
    {
      "aux_loss": 8.106603574752807,
      "entropy": 0.14477055370807648,
      "epoch": 0.00381,
      "grad_norm": 0.13840284943580627,
      "learning_rate": 2e-05,
      "loss": 0.17411131858825685,
      "mean_token_accuracy": 0.9516070276498795,
      "num_tokens": 61330978.0,
      "step": 3810
    },
    {
      "aux_loss": 8.06417441368103,
      "entropy": 0.18448591083288193,
      "epoch": 0.003815,
      "grad_norm": 0.1575627475976944,
      "learning_rate": 2e-05,
      "loss": 0.22809510231018065,
      "mean_token_accuracy": 0.9370896995067597,
      "num_tokens": 61412024.0,
      "step": 3815
    },
    {
      "aux_loss": 8.066959285736084,
      "entropy": 0.18426196910440923,
      "epoch": 0.00382,
      "grad_norm": 0.1520664542913437,
      "learning_rate": 2e-05,
      "loss": 0.22588849067687988,
      "mean_token_accuracy": 0.937090840935707,
      "num_tokens": 61492663.0,
      "step": 3820
    },
    {
      "aux_loss": 8.077086305618286,
      "entropy": 0.18012150302529334,
      "epoch": 0.003825,
      "grad_norm": 0.13913986086845398,
      "learning_rate": 2e-05,
      "loss": 0.21055953502655028,
      "mean_token_accuracy": 0.9416161686182022,
      "num_tokens": 61573614.0,
      "step": 3825
    },
    {
      "aux_loss": 8.084502267837525,
      "entropy": 0.17572448030114174,
      "epoch": 0.00383,
      "grad_norm": 0.1395370364189148,
      "learning_rate": 2e-05,
      "loss": 0.20020182132720948,
      "mean_token_accuracy": 0.9457064777612686,
      "num_tokens": 61654548.0,
      "step": 3830
    },
    {
      "aux_loss": 8.09701132774353,
      "entropy": 0.160580874979496,
      "epoch": 0.003835,
      "grad_norm": 0.14620131254196167,
      "learning_rate": 2e-05,
      "loss": 0.1879928946495056,
      "mean_token_accuracy": 0.9474564075469971,
      "num_tokens": 61734815.0,
      "step": 3835
    },
    {
      "aux_loss": 8.113717746734618,
      "entropy": 0.1552987787872553,
      "epoch": 0.00384,
      "grad_norm": 0.16251729428768158,
      "learning_rate": 2e-05,
      "loss": 0.18154183626174927,
      "mean_token_accuracy": 0.950005042552948,
      "num_tokens": 61813093.0,
      "step": 3840
    },
    {
      "aux_loss": 8.068704652786256,
      "entropy": 0.17081396877765656,
      "epoch": 0.003845,
      "grad_norm": 0.13388794660568237,
      "learning_rate": 2e-05,
      "loss": 0.21023707389831542,
      "mean_token_accuracy": 0.9419908076524734,
      "num_tokens": 61894066.0,
      "step": 3845
    },
    {
      "aux_loss": 8.07093186378479,
      "entropy": 0.17962607443332673,
      "epoch": 0.00385,
      "grad_norm": 0.14779174327850342,
      "learning_rate": 2e-05,
      "loss": 0.22672898769378663,
      "mean_token_accuracy": 0.9358665734529495,
      "num_tokens": 61975158.0,
      "step": 3850
    },
    {
      "aux_loss": 8.076707983016968,
      "entropy": 0.17310950197279454,
      "epoch": 0.003855,
      "grad_norm": 0.140579491853714,
      "learning_rate": 2e-05,
      "loss": 0.20651617050170898,
      "mean_token_accuracy": 0.9428789287805557,
      "num_tokens": 62055831.0,
      "step": 3855
    },
    {
      "aux_loss": 8.086752367019653,
      "entropy": 0.18669826090335845,
      "epoch": 0.00386,
      "grad_norm": 0.13994470238685608,
      "learning_rate": 2e-05,
      "loss": 0.2107757329940796,
      "mean_token_accuracy": 0.9407466769218444,
      "num_tokens": 62136632.0,
      "step": 3860
    },
    {
      "aux_loss": 8.092320489883424,
      "entropy": 0.16844304874539376,
      "epoch": 0.003865,
      "grad_norm": 0.1472175419330597,
      "learning_rate": 2e-05,
      "loss": 0.19181419610977174,
      "mean_token_accuracy": 0.9457267761230469,
      "num_tokens": 62216811.0,
      "step": 3865
    },
    {
      "aux_loss": 8.111100244522095,
      "entropy": 0.15630946606397628,
      "epoch": 0.00387,
      "grad_norm": 0.1743493378162384,
      "learning_rate": 2e-05,
      "loss": 0.18492677211761474,
      "mean_token_accuracy": 0.9484911233186721,
      "num_tokens": 62297082.0,
      "step": 3870
    },
    {
      "aux_loss": 8.071560907363892,
      "entropy": 0.17498248666524888,
      "epoch": 0.003875,
      "grad_norm": 0.14743833243846893,
      "learning_rate": 2e-05,
      "loss": 0.21595487594604493,
      "mean_token_accuracy": 0.9405759006738663,
      "num_tokens": 62375241.0,
      "step": 3875
    },
    {
      "aux_loss": 8.06474723815918,
      "entropy": 0.201045473664999,
      "epoch": 0.00388,
      "grad_norm": 0.14226999878883362,
      "learning_rate": 2e-05,
      "loss": 0.2451887845993042,
      "mean_token_accuracy": 0.9316706866025924,
      "num_tokens": 62456545.0,
      "step": 3880
    },
    {
      "aux_loss": 8.077294397354127,
      "entropy": 0.1726114943623543,
      "epoch": 0.003885,
      "grad_norm": 0.13420595228672028,
      "learning_rate": 2e-05,
      "loss": 0.20281989574432374,
      "mean_token_accuracy": 0.943801012635231,
      "num_tokens": 62537361.0,
      "step": 3885
    },
    {
      "aux_loss": 8.084831619262696,
      "entropy": 0.17151982635259627,
      "epoch": 0.00389,
      "grad_norm": 0.14446011185646057,
      "learning_rate": 2e-05,
      "loss": 0.19781169891357422,
      "mean_token_accuracy": 0.9445989429950714,
      "num_tokens": 62618051.0,
      "step": 3890
    },
    {
      "aux_loss": 8.095459461212158,
      "entropy": 0.16679829508066177,
      "epoch": 0.003895,
      "grad_norm": 0.14256325364112854,
      "learning_rate": 2e-05,
      "loss": 0.18840779066085817,
      "mean_token_accuracy": 0.9474774956703186,
      "num_tokens": 62698436.0,
      "step": 3895
    },
    {
      "aux_loss": 8.112070798873901,
      "entropy": 0.14683061018586158,
      "epoch": 0.0039,
      "grad_norm": 0.16215519607067108,
      "learning_rate": 2e-05,
      "loss": 0.17416359186172486,
      "mean_token_accuracy": 0.9515390157699585,
      "num_tokens": 62779041.0,
      "step": 3900
    },
    {
      "aux_loss": 8.068126726150513,
      "entropy": 0.16317534297704697,
      "epoch": 0.003905,
      "grad_norm": 0.1673935055732727,
      "learning_rate": 2e-05,
      "loss": 0.20950429439544677,
      "mean_token_accuracy": 0.9425682723522186,
      "num_tokens": 62858638.0,
      "step": 3905
    },
    {
      "aux_loss": 8.067454957962036,
      "entropy": 0.17523740008473396,
      "epoch": 0.00391,
      "grad_norm": 0.1550503373146057,
      "learning_rate": 2e-05,
      "loss": 0.22011191844940187,
      "mean_token_accuracy": 0.938545498251915,
      "num_tokens": 62939766.0,
      "step": 3910
    },
    {
      "aux_loss": 8.07533793449402,
      "entropy": 0.18380124494433403,
      "epoch": 0.003915,
      "grad_norm": 0.13432689011096954,
      "learning_rate": 2e-05,
      "loss": 0.21724610328674315,
      "mean_token_accuracy": 0.9401372313499451,
      "num_tokens": 63020707.0,
      "step": 3915
    },
    {
      "aux_loss": 8.086825227737426,
      "entropy": 0.18134923428297042,
      "epoch": 0.00392,
      "grad_norm": 0.13876844942569733,
      "learning_rate": 2e-05,
      "loss": 0.20158858299255372,
      "mean_token_accuracy": 0.943289241194725,
      "num_tokens": 63101008.0,
      "step": 3920
    },
    {
      "aux_loss": 8.093204641342163,
      "entropy": 0.17062823176383973,
      "epoch": 0.003925,
      "grad_norm": 0.15416906774044037,
      "learning_rate": 2e-05,
      "loss": 0.19223999977111816,
      "mean_token_accuracy": 0.9459930211305618,
      "num_tokens": 63180814.0,
      "step": 3925
    },
    {
      "aux_loss": 8.106504344940186,
      "entropy": 0.15065026395022868,
      "epoch": 0.00393,
      "grad_norm": 0.13469359278678894,
      "learning_rate": 2e-05,
      "loss": 0.18335039615631105,
      "mean_token_accuracy": 0.9487786680459976,
      "num_tokens": 63260847.0,
      "step": 3930
    },
    {
      "aux_loss": 8.07407112121582,
      "entropy": 0.17284853234887124,
      "epoch": 0.003935,
      "grad_norm": 0.155502051115036,
      "learning_rate": 2e-05,
      "loss": 0.21779682636260986,
      "mean_token_accuracy": 0.9408769845962525,
      "num_tokens": 63338365.0,
      "step": 3935
    },
    {
      "aux_loss": 8.064867067337037,
      "entropy": 0.17973878383636474,
      "epoch": 0.00394,
      "grad_norm": 0.1558724343776703,
      "learning_rate": 2e-05,
      "loss": 0.22478768825531006,
      "mean_token_accuracy": 0.9373753875494003,
      "num_tokens": 63420001.0,
      "step": 3940
    },
    {
      "aux_loss": 8.07354006767273,
      "entropy": 0.18883172050118446,
      "epoch": 0.003945,
      "grad_norm": 0.1374303549528122,
      "learning_rate": 2e-05,
      "loss": 0.21975882053375245,
      "mean_token_accuracy": 0.9390784353017807,
      "num_tokens": 63500945.0,
      "step": 3945
    },
    {
      "aux_loss": 8.081611967086792,
      "entropy": 0.1774898014962673,
      "epoch": 0.00395,
      "grad_norm": 0.13622717559337616,
      "learning_rate": 2e-05,
      "loss": 0.2004019498825073,
      "mean_token_accuracy": 0.9445203125476838,
      "num_tokens": 63581595.0,
      "step": 3950
    },
    {
      "aux_loss": 8.094461631774902,
      "entropy": 0.1645282708108425,
      "epoch": 0.003955,
      "grad_norm": 0.14535793662071228,
      "learning_rate": 2e-05,
      "loss": 0.19013161659240724,
      "mean_token_accuracy": 0.9472324132919312,
      "num_tokens": 63661915.0,
      "step": 3955
    },
    {
      "aux_loss": 8.110036945343017,
      "entropy": 0.15100119560956954,
      "epoch": 0.00396,
      "grad_norm": 0.15752434730529785,
      "learning_rate": 2e-05,
      "loss": 0.18374007940292358,
      "mean_token_accuracy": 0.9492660373449325,
      "num_tokens": 63742004.0,
      "step": 3960
    },
    {
      "aux_loss": 8.072049474716186,
      "entropy": 0.1649086859077215,
      "epoch": 0.003965,
      "grad_norm": 0.1616693139076233,
      "learning_rate": 2e-05,
      "loss": 0.20715844631195068,
      "mean_token_accuracy": 0.9433943182229996,
      "num_tokens": 63822259.0,
      "step": 3965
    },
    {
      "aux_loss": 8.065539264678955,
      "entropy": 0.1898988239467144,
      "epoch": 0.00397,
      "grad_norm": 0.14644716680049896,
      "learning_rate": 2e-05,
      "loss": 0.23113880157470704,
      "mean_token_accuracy": 0.9356475621461868,
      "num_tokens": 63903211.0,
      "step": 3970
    },
    {
      "aux_loss": 8.072787141799926,
      "entropy": 0.1908182680606842,
      "epoch": 0.003975,
      "grad_norm": 0.13250933587551117,
      "learning_rate": 2e-05,
      "loss": 0.22197015285491944,
      "mean_token_accuracy": 0.939109081029892,
      "num_tokens": 63984548.0,
      "step": 3975
    },
    {
      "aux_loss": 8.082742023468018,
      "entropy": 0.17957219183444978,
      "epoch": 0.00398,
      "grad_norm": 0.13384675979614258,
      "learning_rate": 2e-05,
      "loss": 0.20482177734375,
      "mean_token_accuracy": 0.9431985884904861,
      "num_tokens": 64065346.0,
      "step": 3980
    },
    {
      "aux_loss": 8.090649700164795,
      "entropy": 0.16390545591711997,
      "epoch": 0.003985,
      "grad_norm": 0.13855285942554474,
      "learning_rate": 2e-05,
      "loss": 0.19011374711990356,
      "mean_token_accuracy": 0.9466467708349228,
      "num_tokens": 64145867.0,
      "step": 3985
    },
    {
      "aux_loss": 8.106703615188598,
      "entropy": 0.16054214015603066,
      "epoch": 0.00399,
      "grad_norm": 0.15388259291648865,
      "learning_rate": 2e-05,
      "loss": 0.18943867683410645,
      "mean_token_accuracy": 0.9464708685874939,
      "num_tokens": 64226088.0,
      "step": 3990
    },
    {
      "aux_loss": 8.081350135803223,
      "entropy": 0.1579239446669817,
      "epoch": 0.003995,
      "grad_norm": 0.14588339626789093,
      "learning_rate": 2e-05,
      "loss": 0.19378492832183838,
      "mean_token_accuracy": 0.9476122468709945,
      "num_tokens": 64303351.0,
      "step": 3995
    },
    {
      "aux_loss": 8.06923942565918,
      "entropy": 0.17941427305340768,
      "epoch": 0.004,
      "grad_norm": 0.14514535665512085,
      "learning_rate": 2e-05,
      "loss": 0.22122440338134766,
      "mean_token_accuracy": 0.9387730300426483,
      "num_tokens": 64384402.0,
      "step": 4000
    },
    {
      "aux_loss": 8.069340658187866,
      "entropy": 0.18155089989304543,
      "epoch": 0.004005,
      "grad_norm": 0.14123430848121643,
      "learning_rate": 2e-05,
      "loss": 0.21901488304138184,
      "mean_token_accuracy": 0.9397453844547272,
      "num_tokens": 64465427.0,
      "step": 4005
    },
    {
      "aux_loss": 8.08221082687378,
      "entropy": 0.1821060985326767,
      "epoch": 0.00401,
      "grad_norm": 0.13074013590812683,
      "learning_rate": 2e-05,
      "loss": 0.2080132007598877,
      "mean_token_accuracy": 0.9409109026193618,
      "num_tokens": 64546356.0,
      "step": 4010
    },
    {
      "aux_loss": 8.090443134307861,
      "entropy": 0.16765516772866249,
      "epoch": 0.004015,
      "grad_norm": 0.1483263075351715,
      "learning_rate": 2e-05,
      "loss": 0.1874303102493286,
      "mean_token_accuracy": 0.9475386768579483,
      "num_tokens": 64626873.0,
      "step": 4015
    },
    {
      "aux_loss": 8.10508303642273,
      "entropy": 0.15679123848676682,
      "epoch": 0.00402,
      "grad_norm": 0.1835867166519165,
      "learning_rate": 2e-05,
      "loss": 0.19097511768341063,
      "mean_token_accuracy": 0.9467210471630096,
      "num_tokens": 64707233.0,
      "step": 4020
    },
    {
      "aux_loss": 8.083331775665282,
      "entropy": 0.15017431192100048,
      "epoch": 0.004025,
      "grad_norm": 0.13776715099811554,
      "learning_rate": 2e-05,
      "loss": 0.18937180042266846,
      "mean_token_accuracy": 0.948642173409462,
      "num_tokens": 64788031.0,
      "step": 4025
    },
    {
      "aux_loss": 8.070507955551147,
      "entropy": 0.19104250706732273,
      "epoch": 0.00403,
      "grad_norm": 0.14764906466007233,
      "learning_rate": 2e-05,
      "loss": 0.23659148216247558,
      "mean_token_accuracy": 0.9353424161672592,
      "num_tokens": 64869070.0,
      "step": 4030
    },
    {
      "aux_loss": 8.069948101043702,
      "entropy": 0.17851735912263395,
      "epoch": 0.004035,
      "grad_norm": 0.13509583473205566,
      "learning_rate": 2e-05,
      "loss": 0.21018548011779786,
      "mean_token_accuracy": 0.9413766711950302,
      "num_tokens": 64949884.0,
      "step": 4035
    },
    {
      "aux_loss": 8.081086015701294,
      "entropy": 0.184680961817503,
      "epoch": 0.00404,
      "grad_norm": 0.13973838090896606,
      "learning_rate": 2e-05,
      "loss": 0.21257233619689941,
      "mean_token_accuracy": 0.9408496230840683,
      "num_tokens": 65030661.0,
      "step": 4040
    },
    {
      "aux_loss": 8.087134504318238,
      "entropy": 0.17623230963945388,
      "epoch": 0.004045,
      "grad_norm": 0.1628706008195877,
      "learning_rate": 2e-05,
      "loss": 0.20290110111236573,
      "mean_token_accuracy": 0.9433736443519593,
      "num_tokens": 65111662.0,
      "step": 4045
    },
    {
      "aux_loss": 8.10088005065918,
      "entropy": 0.159989033639431,
      "epoch": 0.00405,
      "grad_norm": 0.14700080454349518,
      "learning_rate": 2e-05,
      "loss": 0.18420116901397704,
      "mean_token_accuracy": 0.9484766960144043,
      "num_tokens": 65191660.0,
      "step": 4050
    },
    {
      "aux_loss": 8.09033432006836,
      "entropy": 0.174098851531744,
      "epoch": 0.004055,
      "grad_norm": 0.152372345328331,
      "learning_rate": 2e-05,
      "loss": 0.20993974208831787,
      "mean_token_accuracy": 0.9438207447528839,
      "num_tokens": 65269913.0,
      "step": 4055
    },
    {
      "aux_loss": 8.06638798713684,
      "entropy": 0.19519088454544545,
      "epoch": 0.00406,
      "grad_norm": 0.1604108065366745,
      "learning_rate": 2e-05,
      "loss": 0.2431988000869751,
      "mean_token_accuracy": 0.9322654813528061,
      "num_tokens": 65350862.0,
      "step": 4060
    },
    {
      "aux_loss": 8.070633697509766,
      "entropy": 0.1889781817793846,
      "epoch": 0.004065,
      "grad_norm": 0.12792906165122986,
      "learning_rate": 2e-05,
      "loss": 0.22586872577667236,
      "mean_token_accuracy": 0.9373369872570038,
      "num_tokens": 65432074.0,
      "step": 4065
    },
    {
      "aux_loss": 8.07981243133545,
      "entropy": 0.1819232426583767,
      "epoch": 0.00407,
      "grad_norm": 0.136160746216774,
      "learning_rate": 2e-05,
      "loss": 0.20806806087493895,
      "mean_token_accuracy": 0.9421541363000869,
      "num_tokens": 65512746.0,
      "step": 4070
    },
    {
      "aux_loss": 8.08708896636963,
      "entropy": 0.1701486200094223,
      "epoch": 0.004075,
      "grad_norm": 0.1507285088300705,
      "learning_rate": 2e-05,
      "loss": 0.19780858755111694,
      "mean_token_accuracy": 0.9454937070608139,
      "num_tokens": 65593234.0,
      "step": 4075
    },
    {
      "aux_loss": 8.099062633514404,
      "entropy": 0.1580248475074768,
      "epoch": 0.00408,
      "grad_norm": 0.15038086473941803,
      "learning_rate": 2e-05,
      "loss": 0.18610332012176514,
      "mean_token_accuracy": 0.9480824947357178,
      "num_tokens": 65673733.0,
      "step": 4080
    },
    {
      "aux_loss": 8.091101217269898,
      "entropy": 0.15355209447443485,
      "epoch": 0.004085,
      "grad_norm": 0.1268012821674347,
      "learning_rate": 2e-05,
      "loss": 0.18365161418914794,
      "mean_token_accuracy": 0.9506066590547562,
      "num_tokens": 65750893.0,
      "step": 4085
    },
    {
      "aux_loss": 8.064303636550903,
      "entropy": 0.18075645230710508,
      "epoch": 0.00409,
      "grad_norm": 0.14601679146289825,
      "learning_rate": 2e-05,
      "loss": 0.22333426475524903,
      "mean_token_accuracy": 0.9376881331205368,
      "num_tokens": 65832039.0,
      "step": 4090
    },
    {
      "aux_loss": 8.068262529373168,
      "entropy": 0.19310965687036513,
      "epoch": 0.004095,
      "grad_norm": 0.13745450973510742,
      "learning_rate": 2e-05,
      "loss": 0.23399546146392822,
      "mean_token_accuracy": 0.9342675775289535,
      "num_tokens": 65913070.0,
      "step": 4095
    },
    {
      "aux_loss": 8.074833726882934,
      "entropy": 0.17377459220588207,
      "epoch": 0.0041,
      "grad_norm": 0.13420403003692627,
      "learning_rate": 2e-05,
      "loss": 0.20697827339172364,
      "mean_token_accuracy": 0.9427221775054931,
      "num_tokens": 65993809.0,
      "step": 4100
    },
    {
      "aux_loss": 8.08429675102234,
      "entropy": 0.1748849056661129,
      "epoch": 0.004105,
      "grad_norm": 0.13983213901519775,
      "learning_rate": 2e-05,
      "loss": 0.19947532415390015,
      "mean_token_accuracy": 0.943372043967247,
      "num_tokens": 66074604.0,
      "step": 4105
    },
    {
      "aux_loss": 8.094364404678345,
      "entropy": 0.15670294016599656,
      "epoch": 0.00411,
      "grad_norm": 0.1450595110654831,
      "learning_rate": 2e-05,
      "loss": 0.17926582098007202,
      "mean_token_accuracy": 0.9512201547622681,
      "num_tokens": 66155539.0,
      "step": 4110
    },
    {
      "aux_loss": 8.101194286346436,
      "entropy": 0.1474872872233391,
      "epoch": 0.004115,
      "grad_norm": 0.13709039986133575,
      "learning_rate": 2e-05,
      "loss": 0.1780955672264099,
      "mean_token_accuracy": 0.9507912933826447,
      "num_tokens": 66234447.0,
      "step": 4115
    },
    {
      "aux_loss": 8.057816457748412,
      "entropy": 0.1664051443338394,
      "epoch": 0.00412,
      "grad_norm": 0.1568625420331955,
      "learning_rate": 2e-05,
      "loss": 0.2139223575592041,
      "mean_token_accuracy": 0.9409090608358384,
      "num_tokens": 66315669.0,
      "step": 4120
    },
    {
      "aux_loss": 8.069647598266602,
      "entropy": 0.19159175381064414,
      "epoch": 0.004125,
      "grad_norm": 0.1470678448677063,
      "learning_rate": 2e-05,
      "loss": 0.23636178970336913,
      "mean_token_accuracy": 0.9345293343067169,
      "num_tokens": 66396554.0,
      "step": 4125
    },
    {
      "aux_loss": 8.072778749465943,
      "entropy": 0.18799845278263091,
      "epoch": 0.00413,
      "grad_norm": 0.13189157843589783,
      "learning_rate": 2e-05,
      "loss": 0.21511499881744384,
      "mean_token_accuracy": 0.9406137645244599,
      "num_tokens": 66477671.0,
      "step": 4130
    },
    {
      "aux_loss": 8.08385615348816,
      "entropy": 0.1776302933692932,
      "epoch": 0.004135,
      "grad_norm": 0.13385000824928284,
      "learning_rate": 2e-05,
      "loss": 0.19926952123641967,
      "mean_token_accuracy": 0.9443440586328506,
      "num_tokens": 66558415.0,
      "step": 4135
    },
    {
      "aux_loss": 8.092693519592284,
      "entropy": 0.165148364007473,
      "epoch": 0.00414,
      "grad_norm": 0.14408499002456665,
      "learning_rate": 2e-05,
      "loss": 0.19295555353164673,
      "mean_token_accuracy": 0.9458460479974746,
      "num_tokens": 66639392.0,
      "step": 4140
    },
    {
      "aux_loss": 8.10185284614563,
      "entropy": 0.15802374631166458,
      "epoch": 0.004145,
      "grad_norm": 0.12628768384456635,
      "learning_rate": 2e-05,
      "loss": 0.18807635307312012,
      "mean_token_accuracy": 0.9486256182193756,
      "num_tokens": 66717706.0,
      "step": 4145
    },
    {
      "aux_loss": 8.061801862716674,
      "entropy": 0.17386738657951356,
      "epoch": 0.00415,
      "grad_norm": 0.14630699157714844,
      "learning_rate": 2e-05,
      "loss": 0.21836941242218016,
      "mean_token_accuracy": 0.940554091334343,
      "num_tokens": 66798955.0,
      "step": 4150
    },
    {
      "aux_loss": 8.070080280303955,
      "entropy": 0.16589830182492732,
      "epoch": 0.004155,
      "grad_norm": 0.15465304255485535,
      "learning_rate": 2e-05,
      "loss": 0.20467777252197267,
      "mean_token_accuracy": 0.9426725238561631,
      "num_tokens": 66879375.0,
      "step": 4155
    },
    {
      "aux_loss": 8.07814974784851,
      "entropy": 0.17901213020086287,
      "epoch": 0.00416,
      "grad_norm": 0.13934604823589325,
      "learning_rate": 2e-05,
      "loss": 0.21030039787292482,
      "mean_token_accuracy": 0.9412366211414337,
      "num_tokens": 66960058.0,
      "step": 4160
    },
    {
      "aux_loss": 8.088335704803466,
      "entropy": 0.1755394533276558,
      "epoch": 0.004165,
      "grad_norm": 0.14618510007858276,
      "learning_rate": 2e-05,
      "loss": 0.20091817378997803,
      "mean_token_accuracy": 0.9444429874420166,
      "num_tokens": 67041056.0,
      "step": 4165
    },
    {
      "aux_loss": 8.093344831466675,
      "entropy": 0.15623823255300523,
      "epoch": 0.00417,
      "grad_norm": 0.14904572069644928,
      "learning_rate": 2e-05,
      "loss": 0.18160520792007445,
      "mean_token_accuracy": 0.9495443314313888,
      "num_tokens": 67122012.0,
      "step": 4170
    },
    {
      "aux_loss": 8.102756452560424,
      "entropy": 0.1416645497083664,
      "epoch": 0.004175,
      "grad_norm": 0.12421419471502304,
      "learning_rate": 2e-05,
      "loss": 0.17385153770446776,
      "mean_token_accuracy": 0.9521195501089096,
      "num_tokens": 67200616.0,
      "step": 4175
    },
    {
      "aux_loss": 8.062928533554077,
      "entropy": 0.1789213791489601,
      "epoch": 0.00418,
      "grad_norm": 0.13390590250492096,
      "learning_rate": 2e-05,
      "loss": 0.22279748916625977,
      "mean_token_accuracy": 0.9387729048728943,
      "num_tokens": 67281934.0,
      "step": 4180
    },
    {
      "aux_loss": 8.067456197738647,
      "entropy": 0.19295660853385926,
      "epoch": 0.004185,
      "grad_norm": 0.15121915936470032,
      "learning_rate": 2e-05,
      "loss": 0.2366931676864624,
      "mean_token_accuracy": 0.9342221319675446,
      "num_tokens": 67362931.0,
      "step": 4185
    },
    {
      "aux_loss": 8.07349090576172,
      "entropy": 0.19296474382281303,
      "epoch": 0.00419,
      "grad_norm": 0.13033528625965118,
      "learning_rate": 2e-05,
      "loss": 0.22286410331726075,
      "mean_token_accuracy": 0.9372146010398865,
      "num_tokens": 67443902.0,
      "step": 4190
    },
    {
      "aux_loss": 8.078996515274047,
      "entropy": 0.17323909923434258,
      "epoch": 0.004195,
      "grad_norm": 0.1359061300754547,
      "learning_rate": 2e-05,
      "loss": 0.19408833980560303,
      "mean_token_accuracy": 0.9463406056165695,
      "num_tokens": 67524610.0,
      "step": 4195
    },
    {
      "aux_loss": 8.091639518737793,
      "entropy": 0.16558289527893066,
      "epoch": 0.0042,
      "grad_norm": 0.14134995639324188,
      "learning_rate": 2e-05,
      "loss": 0.18870824575424194,
      "mean_token_accuracy": 0.9481595456600189,
      "num_tokens": 67605394.0,
      "step": 4200
    },
    {
      "aux_loss": 8.11300106048584,
      "entropy": 0.156542756408453,
      "epoch": 0.004205,
      "grad_norm": 0.1970505714416504,
      "learning_rate": 2e-05,
      "loss": 0.18325804471969603,
      "mean_token_accuracy": 0.9480713158845901,
      "num_tokens": 67682390.0,
      "step": 4205
    },
    {
      "aux_loss": 8.062799263000489,
      "entropy": 0.18372222259640694,
      "epoch": 0.00421,
      "grad_norm": 0.16438975930213928,
      "learning_rate": 2e-05,
      "loss": 0.23308656215667725,
      "mean_token_accuracy": 0.936217337846756,
      "num_tokens": 67763766.0,
      "step": 4210
    },
    {
      "aux_loss": 8.065517807006836,
      "entropy": 0.2014029435813427,
      "epoch": 0.004215,
      "grad_norm": 0.13802672922611237,
      "learning_rate": 2e-05,
      "loss": 0.24253413677215577,
      "mean_token_accuracy": 0.9319497972726822,
      "num_tokens": 67845050.0,
      "step": 4215
    },
    {
      "aux_loss": 8.071827507019043,
      "entropy": 0.19400066658854484,
      "epoch": 0.00422,
      "grad_norm": 0.13618911802768707,
      "learning_rate": 2e-05,
      "loss": 0.2283384084701538,
      "mean_token_accuracy": 0.9372434586286544,
      "num_tokens": 67926012.0,
      "step": 4220
    },
    {
      "aux_loss": 8.079135036468506,
      "entropy": 0.17796122208237647,
      "epoch": 0.004225,
      "grad_norm": 0.134163960814476,
      "learning_rate": 2e-05,
      "loss": 0.20136067867279053,
      "mean_token_accuracy": 0.9438400983810424,
      "num_tokens": 68006576.0,
      "step": 4225
    },
    {
      "aux_loss": 8.089920711517333,
      "entropy": 0.16169902980327605,
      "epoch": 0.00423,
      "grad_norm": 0.1498737633228302,
      "learning_rate": 2e-05,
      "loss": 0.1858538269996643,
      "mean_token_accuracy": 0.9484745532274246,
      "num_tokens": 68086579.0,
      "step": 4230
    },
    {
      "aux_loss": 8.105748748779297,
      "entropy": 0.15697465613484382,
      "epoch": 0.004235,
      "grad_norm": 0.15838921070098877,
      "learning_rate": 2e-05,
      "loss": 0.18612390756607056,
      "mean_token_accuracy": 0.9480006545782089,
      "num_tokens": 68167179.0,
      "step": 4235
    },
    {
      "aux_loss": 8.064301204681396,
      "entropy": 0.1899186883121729,
      "epoch": 0.00424,
      "grad_norm": 0.1430978775024414,
      "learning_rate": 2e-05,
      "loss": 0.2342383623123169,
      "mean_token_accuracy": 0.9365141242742538,
      "num_tokens": 68247740.0,
      "step": 4240
    },
    {
      "aux_loss": 8.071315956115722,
      "entropy": 0.19949283003807067,
      "epoch": 0.004245,
      "grad_norm": 0.14402714371681213,
      "learning_rate": 2e-05,
      "loss": 0.24164390563964844,
      "mean_token_accuracy": 0.9334080964326859,
      "num_tokens": 68328781.0,
      "step": 4245
    },
    {
      "aux_loss": 8.07277536392212,
      "entropy": 0.1642709244042635,
      "epoch": 0.00425,
      "grad_norm": 0.13262556493282318,
      "learning_rate": 2e-05,
      "loss": 0.1938338041305542,
      "mean_token_accuracy": 0.9459922581911087,
      "num_tokens": 68409641.0,
      "step": 4250
    },
    {
      "aux_loss": 8.081375312805175,
      "entropy": 0.17299794778227806,
      "epoch": 0.004255,
      "grad_norm": 0.14518195390701294,
      "learning_rate": 2e-05,
      "loss": 0.20028214454650878,
      "mean_token_accuracy": 0.9443636804819107,
      "num_tokens": 68490108.0,
      "step": 4255
    },
    {
      "aux_loss": 8.088916969299316,
      "entropy": 0.1673464857041836,
      "epoch": 0.00426,
      "grad_norm": 0.1402788609266281,
      "learning_rate": 2e-05,
      "loss": 0.19545431137084962,
      "mean_token_accuracy": 0.945445927977562,
      "num_tokens": 68570407.0,
      "step": 4260
    },
    {
      "aux_loss": 8.104100513458253,
      "entropy": 0.1506233938038349,
      "epoch": 0.004265,
      "grad_norm": 0.16530829668045044,
      "learning_rate": 2e-05,
      "loss": 0.17668191194534302,
      "mean_token_accuracy": 0.9513014763593673,
      "num_tokens": 68650835.0,
      "step": 4265
    },
    {
      "aux_loss": 8.062661266326904,
      "entropy": 0.16357909217476846,
      "epoch": 0.00427,
      "grad_norm": 0.14734461903572083,
      "learning_rate": 2e-05,
      "loss": 0.20471456050872802,
      "mean_token_accuracy": 0.9446876019239425,
      "num_tokens": 68731722.0,
      "step": 4270
    },
    {
      "aux_loss": 8.064127683639526,
      "entropy": 0.1830916054546833,
      "epoch": 0.004275,
      "grad_norm": 0.13832776248455048,
      "learning_rate": 2e-05,
      "loss": 0.22243731021881102,
      "mean_token_accuracy": 0.9390664726495743,
      "num_tokens": 68813208.0,
      "step": 4275
    },
    {
      "aux_loss": 8.069669198989867,
      "entropy": 0.17019328102469444,
      "epoch": 0.00428,
      "grad_norm": 0.13978317379951477,
      "learning_rate": 2e-05,
      "loss": 0.20442538261413573,
      "mean_token_accuracy": 0.9432233452796936,
      "num_tokens": 68894144.0,
      "step": 4280
    },
    {
      "aux_loss": 8.080967378616332,
      "entropy": 0.17591089904308319,
      "epoch": 0.004285,
      "grad_norm": 0.13872495293617249,
      "learning_rate": 2e-05,
      "loss": 0.2043544292449951,
      "mean_token_accuracy": 0.9426625460386276,
      "num_tokens": 68974842.0,
      "step": 4285
    },
    {
      "aux_loss": 8.088066577911377,
      "entropy": 0.15964294224977493,
      "epoch": 0.00429,
      "grad_norm": 0.14347712695598602,
      "learning_rate": 2e-05,
      "loss": 0.18445612192153932,
      "mean_token_accuracy": 0.9482874900102616,
      "num_tokens": 69054956.0,
      "step": 4290
    },
    {
      "aux_loss": 8.10286865234375,
      "entropy": 0.1529075898230076,
      "epoch": 0.004295,
      "grad_norm": 0.16215908527374268,
      "learning_rate": 2e-05,
      "loss": 0.17725415229797364,
      "mean_token_accuracy": 0.9505029529333114,
      "num_tokens": 69135438.0,
      "step": 4295
    },
    {
      "aux_loss": 8.064771461486817,
      "entropy": 0.1863943636417389,
      "epoch": 0.0043,
      "grad_norm": 0.14068421721458435,
      "learning_rate": 2e-05,
      "loss": 0.22935824394226073,
      "mean_token_accuracy": 0.9385807782411575,
      "num_tokens": 69215103.0,
      "step": 4300
    },
    {
      "aux_loss": 8.070622444152832,
      "entropy": 0.1977604880928993,
      "epoch": 0.004305,
      "grad_norm": 0.1485324501991272,
      "learning_rate": 2e-05,
      "loss": 0.24191112518310548,
      "mean_token_accuracy": 0.9326848983764648,
      "num_tokens": 69296223.0,
      "step": 4305
    },
    {
      "aux_loss": 8.068218088150024,
      "entropy": 0.17825442291796206,
      "epoch": 0.00431,
      "grad_norm": 0.13282841444015503,
      "learning_rate": 2e-05,
      "loss": 0.21320128440856934,
      "mean_token_accuracy": 0.9405946582555771,
      "num_tokens": 69377141.0,
      "step": 4310
    },
    {
      "aux_loss": 8.079132413864135,
      "entropy": 0.176223087310791,
      "epoch": 0.004315,
      "grad_norm": 0.1335696429014206,
      "learning_rate": 2e-05,
      "loss": 0.20367257595062255,
      "mean_token_accuracy": 0.9429424643516541,
      "num_tokens": 69457684.0,
      "step": 4315
    },
    {
      "aux_loss": 8.088479137420654,
      "entropy": 0.16690649688243867,
      "epoch": 0.00432,
      "grad_norm": 0.1405257135629654,
      "learning_rate": 2e-05,
      "loss": 0.19091715812683105,
      "mean_token_accuracy": 0.9464767873287201,
      "num_tokens": 69538378.0,
      "step": 4320
    },
    {
      "aux_loss": 8.09993510246277,
      "entropy": 0.1499143972992897,
      "epoch": 0.004325,
      "grad_norm": 0.15144208073616028,
      "learning_rate": 2e-05,
      "loss": 0.17648680210113527,
      "mean_token_accuracy": 0.9506325542926788,
      "num_tokens": 69618500.0,
      "step": 4325
    },
    {
      "aux_loss": 8.075860214233398,
      "entropy": 0.17051323652267455,
      "epoch": 0.00433,
      "grad_norm": 0.16274310648441315,
      "learning_rate": 2e-05,
      "loss": 0.20888113975524902,
      "mean_token_accuracy": 0.9421725541353225,
      "num_tokens": 69698610.0,
      "step": 4330
    },
    {
      "aux_loss": 8.067101573944091,
      "entropy": 0.18359835296869279,
      "epoch": 0.004335,
      "grad_norm": 0.14738775789737701,
      "learning_rate": 2e-05,
      "loss": 0.22505500316619872,
      "mean_token_accuracy": 0.9374961674213409,
      "num_tokens": 69779681.0,
      "step": 4335
    },
    {
      "aux_loss": 8.069784450531007,
      "entropy": 0.18074525594711305,
      "epoch": 0.00434,
      "grad_norm": 0.13869988918304443,
      "learning_rate": 2e-05,
      "loss": 0.2161696195602417,
      "mean_token_accuracy": 0.9398662030696869,
      "num_tokens": 69860607.0,
      "step": 4340
    },
    {
      "aux_loss": 8.079578924179078,
      "entropy": 0.1802871987223625,
      "epoch": 0.004345,
      "grad_norm": 0.13982544839382172,
      "learning_rate": 2e-05,
      "loss": 0.20732157230377196,
      "mean_token_accuracy": 0.9423238128423691,
      "num_tokens": 69941509.0,
      "step": 4345
    },
    {
      "aux_loss": 8.08824462890625,
      "entropy": 0.16554139778017998,
      "epoch": 0.00435,
      "grad_norm": 0.1411389410495758,
      "learning_rate": 2e-05,
      "loss": 0.18847463130950928,
      "mean_token_accuracy": 0.9479102939367294,
      "num_tokens": 70021906.0,
      "step": 4350
    },
    {
      "aux_loss": 8.102502489089966,
      "entropy": 0.15165012702345848,
      "epoch": 0.004355,
      "grad_norm": 0.17146232724189758,
      "learning_rate": 2e-05,
      "loss": 0.18178340196609497,
      "mean_token_accuracy": 0.9504520446062088,
      "num_tokens": 70102174.0,
      "step": 4355
    },
    {
      "aux_loss": 8.080121850967407,
      "entropy": 0.17862374149262905,
      "epoch": 0.00436,
      "grad_norm": 0.154321551322937,
      "learning_rate": 2e-05,
      "loss": 0.22329416275024414,
      "mean_token_accuracy": 0.9388250648975373,
      "num_tokens": 70179755.0,
      "step": 4360
    },
    {
      "aux_loss": 8.065065860748291,
      "entropy": 0.19679731577634813,
      "epoch": 0.004365,
      "grad_norm": 0.1517145186662674,
      "learning_rate": 2e-05,
      "loss": 0.2374509811401367,
      "mean_token_accuracy": 0.9333733379840851,
      "num_tokens": 70261012.0,
      "step": 4365
    },
    {
      "aux_loss": 8.071874523162842,
      "entropy": 0.1774231106042862,
      "epoch": 0.00437,
      "grad_norm": 0.13734932243824005,
      "learning_rate": 2e-05,
      "loss": 0.20597360134124756,
      "mean_token_accuracy": 0.9424983054399491,
      "num_tokens": 70342063.0,
      "step": 4370
    },
    {
      "aux_loss": 8.075448513031006,
      "entropy": 0.17250633724033831,
      "epoch": 0.004375,
      "grad_norm": 0.13557101786136627,
      "learning_rate": 2e-05,
      "loss": 0.19838166236877441,
      "mean_token_accuracy": 0.9442262798547745,
      "num_tokens": 70422480.0,
      "step": 4375
    },
    {
      "aux_loss": 8.085636949539184,
      "entropy": 0.1633521683514118,
      "epoch": 0.00438,
      "grad_norm": 0.14496545493602753,
      "learning_rate": 2e-05,
      "loss": 0.19115960597991943,
      "mean_token_accuracy": 0.9469558268785476,
      "num_tokens": 70502787.0,
      "step": 4380
    },
    {
      "aux_loss": 8.098505687713622,
      "entropy": 0.14625002555549144,
      "epoch": 0.004385,
      "grad_norm": 0.15544292330741882,
      "learning_rate": 2e-05,
      "loss": 0.17514681816101074,
      "mean_token_accuracy": 0.9512954622507095,
      "num_tokens": 70583182.0,
      "step": 4385
    },
    {
      "aux_loss": 8.079198598861694,
      "entropy": 0.16400582976639272,
      "epoch": 0.00439,
      "grad_norm": 0.15685784816741943,
      "learning_rate": 2e-05,
      "loss": 0.2038726329803467,
      "mean_token_accuracy": 0.944197016954422,
      "num_tokens": 70661098.0,
      "step": 4390
    },
    {
      "aux_loss": 8.063077163696288,
      "entropy": 0.18995024338364602,
      "epoch": 0.004395,
      "grad_norm": 0.1617853045463562,
      "learning_rate": 2e-05,
      "loss": 0.23596880435943604,
      "mean_token_accuracy": 0.9353494733572006,
      "num_tokens": 70742372.0,
      "step": 4395
    },
    {
      "aux_loss": 8.068993473052979,
      "entropy": 0.18315186426043512,
      "epoch": 0.0044,
      "grad_norm": 0.14959554374217987,
      "learning_rate": 2e-05,
      "loss": 0.21350929737091065,
      "mean_token_accuracy": 0.9409860402345658,
      "num_tokens": 70823521.0,
      "step": 4400
    },
    {
      "aux_loss": 8.076430606842042,
      "entropy": 0.1799841694533825,
      "epoch": 0.004405,
      "grad_norm": 0.13897870481014252,
      "learning_rate": 2e-05,
      "loss": 0.20462429523468018,
      "mean_token_accuracy": 0.9430956065654754,
      "num_tokens": 70904491.0,
      "step": 4405
    },
    {
      "aux_loss": 8.087015199661256,
      "entropy": 0.16791487634181976,
      "epoch": 0.00441,
      "grad_norm": 0.1454988569021225,
      "learning_rate": 2e-05,
      "loss": 0.19633032083511354,
      "mean_token_accuracy": 0.9444218277931213,
      "num_tokens": 70985017.0,
      "step": 4410
    },
    {
      "aux_loss": 8.099424028396607,
      "entropy": 0.15029532685875893,
      "epoch": 0.004415,
      "grad_norm": 0.15232595801353455,
      "learning_rate": 2e-05,
      "loss": 0.17951246500015258,
      "mean_token_accuracy": 0.9501245677471161,
      "num_tokens": 71065525.0,
      "step": 4415
    },
    {
      "aux_loss": 8.077301216125488,
      "entropy": 0.17666647508740424,
      "epoch": 0.00442,
      "grad_norm": 0.14066213369369507,
      "learning_rate": 2e-05,
      "loss": 0.21807899475097656,
      "mean_token_accuracy": 0.9400651812553406,
      "num_tokens": 71144378.0,
      "step": 4420
    },
    {
      "aux_loss": 8.062133598327637,
      "entropy": 0.19724079072475434,
      "epoch": 0.004425,
      "grad_norm": 0.13821446895599365,
      "learning_rate": 2e-05,
      "loss": 0.24060750007629395,
      "mean_token_accuracy": 0.9340013116598129,
      "num_tokens": 71225128.0,
      "step": 4425
    },
    {
      "aux_loss": 8.067744398117066,
      "entropy": 0.18977774679660797,
      "epoch": 0.00443,
      "grad_norm": 0.13182908296585083,
      "learning_rate": 2e-05,
      "loss": 0.2200251579284668,
      "mean_token_accuracy": 0.9383378446102142,
      "num_tokens": 71305827.0,
      "step": 4430
    },
    {
      "aux_loss": 8.074538040161134,
      "entropy": 0.17881755754351616,
      "epoch": 0.004435,
      "grad_norm": 0.13338777422904968,
      "learning_rate": 2e-05,
      "loss": 0.20556602478027344,
      "mean_token_accuracy": 0.9426519364118576,
      "num_tokens": 71386866.0,
      "step": 4435
    },
    {
      "aux_loss": 8.084324836730957,
      "entropy": 0.17350154370069504,
      "epoch": 0.00444,
      "grad_norm": 0.13931666314601898,
      "learning_rate": 2e-05,
      "loss": 0.20212831497192382,
      "mean_token_accuracy": 0.9437998682260513,
      "num_tokens": 71467809.0,
      "step": 4440
    },
    {
      "aux_loss": 8.09551944732666,
      "entropy": 0.15271651223301888,
      "epoch": 0.004445,
      "grad_norm": 0.14519935846328735,
      "learning_rate": 2e-05,
      "loss": 0.17921783924102783,
      "mean_token_accuracy": 0.949969157576561,
      "num_tokens": 71547991.0,
      "step": 4445
    },
    {
      "aux_loss": 8.085372018814088,
      "entropy": 0.16865719333291054,
      "epoch": 0.00445,
      "grad_norm": 0.1393231302499771,
      "learning_rate": 2e-05,
      "loss": 0.2130376100540161,
      "mean_token_accuracy": 0.941288697719574,
      "num_tokens": 71626670.0,
      "step": 4450
    },
    {
      "aux_loss": 8.066629123687743,
      "entropy": 0.21009229570627214,
      "epoch": 0.004455,
      "grad_norm": 0.14402218163013458,
      "learning_rate": 2e-05,
      "loss": 0.2533489465713501,
      "mean_token_accuracy": 0.9295083075761795,
      "num_tokens": 71707870.0,
      "step": 4455
    },
    {
      "aux_loss": 8.070453786849976,
      "entropy": 0.1969396859407425,
      "epoch": 0.00446,
      "grad_norm": 0.14154215157032013,
      "learning_rate": 2e-05,
      "loss": 0.22565038204193116,
      "mean_token_accuracy": 0.9378685653209686,
      "num_tokens": 71788873.0,
      "step": 4460
    },
    {
      "aux_loss": 8.076510000228883,
      "entropy": 0.18202112466096879,
      "epoch": 0.004465,
      "grad_norm": 0.14607785642147064,
      "learning_rate": 2e-05,
      "loss": 0.2105297327041626,
      "mean_token_accuracy": 0.9412449181079865,
      "num_tokens": 71869739.0,
      "step": 4465
    },
    {
      "aux_loss": 8.084033250808716,
      "entropy": 0.16639509126544,
      "epoch": 0.00447,
      "grad_norm": 0.13450272381305695,
      "learning_rate": 2e-05,
      "loss": 0.18952665328979493,
      "mean_token_accuracy": 0.9482693105936051,
      "num_tokens": 71950652.0,
      "step": 4470
    },
    {
      "aux_loss": 8.096046972274781,
      "entropy": 0.1618128690868616,
      "epoch": 0.004475,
      "grad_norm": 0.14755448698997498,
      "learning_rate": 2e-05,
      "loss": 0.1920398950576782,
      "mean_token_accuracy": 0.9457824021577835,
      "num_tokens": 72030610.0,
      "step": 4475
    },
    {
      "aux_loss": 8.083650875091553,
      "entropy": 0.1595099586993456,
      "epoch": 0.00448,
      "grad_norm": 0.1382610946893692,
      "learning_rate": 2e-05,
      "loss": 0.19593231678009032,
      "mean_token_accuracy": 0.9457450628280639,
      "num_tokens": 72107579.0,
      "step": 4480
    },
    {
      "aux_loss": 8.062685012817383,
      "entropy": 0.19694138392806054,
      "epoch": 0.004485,
      "grad_norm": 0.15121030807495117,
      "learning_rate": 2e-05,
      "loss": 0.24766852855682372,
      "mean_token_accuracy": 0.932568433880806,
      "num_tokens": 72188409.0,
      "step": 4485
    },
    {
      "aux_loss": 8.066452789306641,
      "entropy": 0.1974887542426586,
      "epoch": 0.00449,
      "grad_norm": 0.14627841114997864,
      "learning_rate": 2e-05,
      "loss": 0.22699663639068604,
      "mean_token_accuracy": 0.9374784499406814,
      "num_tokens": 72269638.0,
      "step": 4490
    },
    {
      "aux_loss": 8.072482252120972,
      "entropy": 0.19781241416931153,
      "epoch": 0.004495,
      "grad_norm": 0.1398935765028,
      "learning_rate": 2e-05,
      "loss": 0.22060873508453369,
      "mean_token_accuracy": 0.9392734676599502,
      "num_tokens": 72350725.0,
      "step": 4495
    },
    {
      "aux_loss": 8.082854652404786,
      "entropy": 0.1693793423473835,
      "epoch": 0.0045,
      "grad_norm": 0.14737220108509064,
      "learning_rate": 2e-05,
      "loss": 0.19731720685958862,
      "mean_token_accuracy": 0.9453026652336121,
      "num_tokens": 72431768.0,
      "step": 4500
    },
    {
      "aux_loss": 8.091933298110963,
      "entropy": 0.15824628472328187,
      "epoch": 0.004505,
      "grad_norm": 0.15237073600292206,
      "learning_rate": 2e-05,
      "loss": 0.1889147400856018,
      "mean_token_accuracy": 0.9477914452552796,
      "num_tokens": 72512635.0,
      "step": 4505
    },
    {
      "aux_loss": 8.095367622375488,
      "entropy": 0.1589428626000881,
      "epoch": 0.00451,
      "grad_norm": 0.14845195412635803,
      "learning_rate": 2e-05,
      "loss": 0.194513738155365,
      "mean_token_accuracy": 0.9472707033157348,
      "num_tokens": 72589799.0,
      "step": 4510
    },
    {
      "aux_loss": 8.063193702697754,
      "entropy": 0.19799885898828506,
      "epoch": 0.004515,
      "grad_norm": 0.15138953924179077,
      "learning_rate": 2e-05,
      "loss": 0.24119656085968016,
      "mean_token_accuracy": 0.9317484855651855,
      "num_tokens": 72670741.0,
      "step": 4515
    },
    {
      "aux_loss": 8.06618161201477,
      "entropy": 0.18955951258540155,
      "epoch": 0.00452,
      "grad_norm": 0.13983069360256195,
      "learning_rate": 2e-05,
      "loss": 0.2227219820022583,
      "mean_token_accuracy": 0.9374795973300933,
      "num_tokens": 72751380.0,
      "step": 4520
    },
    {
      "aux_loss": 8.07211389541626,
      "entropy": 0.19891044422984122,
      "epoch": 0.004525,
      "grad_norm": 0.1337806135416031,
      "learning_rate": 2e-05,
      "loss": 0.22794482707977295,
      "mean_token_accuracy": 0.9362816601991654,
      "num_tokens": 72832575.0,
      "step": 4525
    },
    {
      "aux_loss": 8.081197166442871,
      "entropy": 0.1760702535510063,
      "epoch": 0.00453,
      "grad_norm": 0.1429797112941742,
      "learning_rate": 2e-05,
      "loss": 0.20033011436462403,
      "mean_token_accuracy": 0.9451403826475143,
      "num_tokens": 72913402.0,
      "step": 4530
    },
    {
      "aux_loss": 8.092799615859985,
      "entropy": 0.15617439076304435,
      "epoch": 0.004535,
      "grad_norm": 0.15613104403018951,
      "learning_rate": 2e-05,
      "loss": 0.18931454420089722,
      "mean_token_accuracy": 0.9478604257106781,
      "num_tokens": 72994260.0,
      "step": 4535
    },
    {
      "aux_loss": 8.090886211395263,
      "entropy": 0.15939540602266788,
      "epoch": 0.00454,
      "grad_norm": 0.1329685002565384,
      "learning_rate": 2e-05,
      "loss": 0.19394994974136354,
      "mean_token_accuracy": 0.9464447289705277,
      "num_tokens": 73073334.0,
      "step": 4540
    },
    {
      "aux_loss": 8.056928014755249,
      "entropy": 0.17319882661104202,
      "epoch": 0.004545,
      "grad_norm": 0.14538465440273285,
      "learning_rate": 2e-05,
      "loss": 0.21885814666748046,
      "mean_token_accuracy": 0.9401337534189225,
      "num_tokens": 73154244.0,
      "step": 4545
    },
    {
      "aux_loss": 8.064096784591674,
      "entropy": 0.18502655103802682,
      "epoch": 0.00455,
      "grad_norm": 0.13656294345855713,
      "learning_rate": 2e-05,
      "loss": 0.22055163383483886,
      "mean_token_accuracy": 0.9385868400335312,
      "num_tokens": 73235183.0,
      "step": 4550
    },
    {
      "aux_loss": 8.072021579742431,
      "entropy": 0.18931702598929406,
      "epoch": 0.004555,
      "grad_norm": 0.13566571474075317,
      "learning_rate": 2e-05,
      "loss": 0.2179636240005493,
      "mean_token_accuracy": 0.9405626714229584,
      "num_tokens": 73316031.0,
      "step": 4555
    },
    {
      "aux_loss": 8.079217147827148,
      "entropy": 0.18022140115499496,
      "epoch": 0.00456,
      "grad_norm": 0.14091147482395172,
      "learning_rate": 2e-05,
      "loss": 0.20822691917419434,
      "mean_token_accuracy": 0.9418215036392212,
      "num_tokens": 73396740.0,
      "step": 4560
    },
    {
      "aux_loss": 8.089881706237794,
      "entropy": 0.1692284107208252,
      "epoch": 0.004565,
      "grad_norm": 0.13896194100379944,
      "learning_rate": 2e-05,
      "loss": 0.1954060673713684,
      "mean_token_accuracy": 0.9457089990377426,
      "num_tokens": 73477466.0,
      "step": 4565
    },
    {
      "aux_loss": 8.096963691711426,
      "entropy": 0.15040182694792747,
      "epoch": 0.00457,
      "grad_norm": 0.13605912029743195,
      "learning_rate": 2e-05,
      "loss": 0.17977070808410645,
      "mean_token_accuracy": 0.950329777598381,
      "num_tokens": 73556185.0,
      "step": 4570
    },
    {
      "aux_loss": 8.062443923950195,
      "entropy": 0.19017603173851966,
      "epoch": 0.004575,
      "grad_norm": 0.14468735456466675,
      "learning_rate": 2e-05,
      "loss": 0.23940153121948243,
      "mean_token_accuracy": 0.9329633384943008,
      "num_tokens": 73637641.0,
      "step": 4575
    },
    {
      "aux_loss": 8.066125631332397,
      "entropy": 0.19175320342183114,
      "epoch": 0.00458,
      "grad_norm": 0.13012750446796417,
      "learning_rate": 2e-05,
      "loss": 0.23033366203308106,
      "mean_token_accuracy": 0.9355287790298462,
      "num_tokens": 73718521.0,
      "step": 4580
    },
    {
      "aux_loss": 8.071811771392822,
      "entropy": 0.1796797424554825,
      "epoch": 0.004585,
      "grad_norm": 0.13333414494991302,
      "learning_rate": 2e-05,
      "loss": 0.20581986904144287,
      "mean_token_accuracy": 0.9426154911518096,
      "num_tokens": 73799328.0,
      "step": 4585
    },
    {
      "aux_loss": 8.080306148529052,
      "entropy": 0.19109393954277037,
      "epoch": 0.00459,
      "grad_norm": 0.1370532512664795,
      "learning_rate": 2e-05,
      "loss": 0.21568603515625,
      "mean_token_accuracy": 0.9396095305681229,
      "num_tokens": 73879633.0,
      "step": 4590
    },
    {
      "aux_loss": 8.088784456253052,
      "entropy": 0.16321615129709244,
      "epoch": 0.004595,
      "grad_norm": 0.13728748261928558,
      "learning_rate": 2e-05,
      "loss": 0.18967210054397582,
      "mean_token_accuracy": 0.9470471858978271,
      "num_tokens": 73959933.0,
      "step": 4595
    },
    {
      "aux_loss": 8.108656072616578,
      "entropy": 0.14840723127126693,
      "epoch": 0.0046,
      "grad_norm": 0.17282789945602417,
      "learning_rate": 2e-05,
      "loss": 0.17800523042678834,
      "mean_token_accuracy": 0.9506584674119949,
      "num_tokens": 74036968.0,
      "step": 4600
    },
    {
      "aux_loss": 8.06267852783203,
      "entropy": 0.1675733968615532,
      "epoch": 0.004605,
      "grad_norm": 0.14014120399951935,
      "learning_rate": 2e-05,
      "loss": 0.2125887393951416,
      "mean_token_accuracy": 0.9414013952016831,
      "num_tokens": 74118273.0,
      "step": 4605
    },
    {
      "aux_loss": 8.063804531097412,
      "entropy": 0.2013581320643425,
      "epoch": 0.00461,
      "grad_norm": 0.14335060119628906,
      "learning_rate": 2e-05,
      "loss": 0.2425384283065796,
      "mean_token_accuracy": 0.9319942027330399,
      "num_tokens": 74199507.0,
      "step": 4610
    },
    {
      "aux_loss": 8.071268701553345,
      "entropy": 0.1975974678993225,
      "epoch": 0.004615,
      "grad_norm": 0.13530243933200836,
      "learning_rate": 2e-05,
      "loss": 0.2292879819869995,
      "mean_token_accuracy": 0.9366386622190476,
      "num_tokens": 74280396.0,
      "step": 4615
    },
    {
      "aux_loss": 8.076024723052978,
      "entropy": 0.18943899720907212,
      "epoch": 0.00462,
      "grad_norm": 0.1392260193824768,
      "learning_rate": 2e-05,
      "loss": 0.21794638633728028,
      "mean_token_accuracy": 0.9395203143358231,
      "num_tokens": 74361468.0,
      "step": 4620
    },
    {
      "aux_loss": 8.085432767868042,
      "entropy": 0.174921815097332,
      "epoch": 0.004625,
      "grad_norm": 0.1396854817867279,
      "learning_rate": 2e-05,
      "loss": 0.19981579780578612,
      "mean_token_accuracy": 0.944375878572464,
      "num_tokens": 74442097.0,
      "step": 4625
    },
    {
      "aux_loss": 8.096465492248536,
      "entropy": 0.15287185683846474,
      "epoch": 0.00463,
      "grad_norm": 0.14518404006958008,
      "learning_rate": 2e-05,
      "loss": 0.17887674570083617,
      "mean_token_accuracy": 0.9504520565271377,
      "num_tokens": 74522340.0,
      "step": 4630
    },
    {
      "aux_loss": 8.077925872802734,
      "entropy": 0.15998136401176452,
      "epoch": 0.004635,
      "grad_norm": 0.14441661536693573,
      "learning_rate": 2e-05,
      "loss": 0.194937002658844,
      "mean_token_accuracy": 0.9472737938165665,
      "num_tokens": 74600424.0,
      "step": 4635
    },
    {
      "aux_loss": 8.063422870635986,
      "entropy": 0.18733035810291768,
      "epoch": 0.00464,
      "grad_norm": 0.14657369256019592,
      "learning_rate": 2e-05,
      "loss": 0.23482840061187743,
      "mean_token_accuracy": 0.9350446581840515,
      "num_tokens": 74681412.0,
      "step": 4640
    },
    {
      "aux_loss": 8.065973854064941,
      "entropy": 0.17998111471533776,
      "epoch": 0.004645,
      "grad_norm": 0.13485363125801086,
      "learning_rate": 2e-05,
      "loss": 0.2157827854156494,
      "mean_token_accuracy": 0.940943318605423,
      "num_tokens": 74762360.0,
      "step": 4645
    },
    {
      "aux_loss": 8.074100160598755,
      "entropy": 0.17944046780467032,
      "epoch": 0.00465,
      "grad_norm": 0.13718393445014954,
      "learning_rate": 2e-05,
      "loss": 0.20702033042907714,
      "mean_token_accuracy": 0.9419519454240799,
      "num_tokens": 74843171.0,
      "step": 4650
    },
    {
      "aux_loss": 8.086026668548584,
      "entropy": 0.18462952673435212,
      "epoch": 0.004655,
      "grad_norm": 0.1488821804523468,
      "learning_rate": 2e-05,
      "loss": 0.2104006290435791,
      "mean_token_accuracy": 0.9410360902547836,
      "num_tokens": 74924206.0,
      "step": 4655
    },
    {
      "aux_loss": 8.096301364898682,
      "entropy": 0.15810611769556998,
      "epoch": 0.00466,
      "grad_norm": 0.14179250597953796,
      "learning_rate": 2e-05,
      "loss": 0.18454512357711791,
      "mean_token_accuracy": 0.9477861642837524,
      "num_tokens": 75004486.0,
      "step": 4660
    },
    {
      "aux_loss": 8.07915906906128,
      "entropy": 0.15486926287412645,
      "epoch": 0.004665,
      "grad_norm": 0.13104258477687836,
      "learning_rate": 2e-05,
      "loss": 0.18867655992507934,
      "mean_token_accuracy": 0.9478426486253738,
      "num_tokens": 75083410.0,
      "step": 4665
    },
    {
      "aux_loss": 8.063070440292359,
      "entropy": 0.18661090694367885,
      "epoch": 0.00467,
      "grad_norm": 0.1346316635608673,
      "learning_rate": 2e-05,
      "loss": 0.23294384479522706,
      "mean_token_accuracy": 0.9356189996004105,
      "num_tokens": 75164368.0,
      "step": 4670
    },
    {
      "aux_loss": 8.06877965927124,
      "entropy": 0.19713190719485282,
      "epoch": 0.004675,
      "grad_norm": 0.1399718075990677,
      "learning_rate": 2e-05,
      "loss": 0.2288973331451416,
      "mean_token_accuracy": 0.9362384736537933,
      "num_tokens": 75245552.0,
      "step": 4675
    },
    {
      "aux_loss": 8.072309732437134,
      "entropy": 0.1855604901909828,
      "epoch": 0.00468,
      "grad_norm": 0.1335286647081375,
      "learning_rate": 2e-05,
      "loss": 0.21088075637817383,
      "mean_token_accuracy": 0.9411711722612381,
      "num_tokens": 75326295.0,
      "step": 4680
    },
    {
      "aux_loss": 8.081314134597779,
      "entropy": 0.17673607021570206,
      "epoch": 0.004685,
      "grad_norm": 0.13869231939315796,
      "learning_rate": 2e-05,
      "loss": 0.20768117904663086,
      "mean_token_accuracy": 0.9430624306201935,
      "num_tokens": 75406987.0,
      "step": 4685
    },
    {
      "aux_loss": 8.094422912597656,
      "entropy": 0.15325603298842908,
      "epoch": 0.00469,
      "grad_norm": 0.15016576647758484,
      "learning_rate": 2e-05,
      "loss": 0.17913633584976196,
      "mean_token_accuracy": 0.9492972373962403,
      "num_tokens": 75486875.0,
      "step": 4690
    },
    {
      "aux_loss": 8.085892486572266,
      "entropy": 0.15288294851779938,
      "epoch": 0.004695,
      "grad_norm": 0.13211746513843536,
      "learning_rate": 2e-05,
      "loss": 0.1904144287109375,
      "mean_token_accuracy": 0.9474628537893295,
      "num_tokens": 75567676.0,
      "step": 4695
    },
    {
      "aux_loss": 8.05788812637329,
      "entropy": 0.18550010696053504,
      "epoch": 0.0047,
      "grad_norm": 0.1705792248249054,
      "learning_rate": 2e-05,
      "loss": 0.2348736047744751,
      "mean_token_accuracy": 0.9357887744903565,
      "num_tokens": 75648822.0,
      "step": 4700
    },
    {
      "aux_loss": 8.064849948883056,
      "entropy": 0.19827280640602113,
      "epoch": 0.004705,
      "grad_norm": 0.13893216848373413,
      "learning_rate": 2e-05,
      "loss": 0.23682036399841308,
      "mean_token_accuracy": 0.934507954120636,
      "num_tokens": 75729826.0,
      "step": 4705
    },
    {
      "aux_loss": 8.06988878250122,
      "entropy": 0.18553285971283912,
      "epoch": 0.00471,
      "grad_norm": 0.14590932428836823,
      "learning_rate": 2e-05,
      "loss": 0.21061787605285645,
      "mean_token_accuracy": 0.9408619731664658,
      "num_tokens": 75810376.0,
      "step": 4710
    },
    {
      "aux_loss": 8.079116249084473,
      "entropy": 0.1813324846327305,
      "epoch": 0.004715,
      "grad_norm": 0.15733754634857178,
      "learning_rate": 2e-05,
      "loss": 0.20513761043548584,
      "mean_token_accuracy": 0.9425485312938691,
      "num_tokens": 75891195.0,
      "step": 4715
    },
    {
      "aux_loss": 8.090752458572387,
      "entropy": 0.15269932709634304,
      "epoch": 0.00472,
      "grad_norm": 0.1568419337272644,
      "learning_rate": 2e-05,
      "loss": 0.17674895524978637,
      "mean_token_accuracy": 0.9509547144174576,
      "num_tokens": 75971569.0,
      "step": 4720
    },
    {
      "aux_loss": 8.096448373794555,
      "entropy": 0.14826879277825356,
      "epoch": 0.004725,
      "grad_norm": 0.12922972440719604,
      "learning_rate": 2e-05,
      "loss": 0.18003151416778565,
      "mean_token_accuracy": 0.9500920981168747,
      "num_tokens": 76048280.0,
      "step": 4725
    },
    {
      "aux_loss": 8.062775564193725,
      "entropy": 0.17861160933971404,
      "epoch": 0.00473,
      "grad_norm": 0.15754659473896027,
      "learning_rate": 2e-05,
      "loss": 0.2280363082885742,
      "mean_token_accuracy": 0.9365868926048279,
      "num_tokens": 76129158.0,
      "step": 4730
    },
    {
      "aux_loss": 8.066216230392456,
      "entropy": 0.18023304902017118,
      "epoch": 0.004735,
      "grad_norm": 0.1284816414117813,
      "learning_rate": 2e-05,
      "loss": 0.21792955398559571,
      "mean_token_accuracy": 0.9398057073354721,
      "num_tokens": 76210467.0,
      "step": 4735
    },
    {
      "aux_loss": 8.074613666534423,
      "entropy": 0.18975592032074928,
      "epoch": 0.00474,
      "grad_norm": 0.13700903952121735,
      "learning_rate": 2e-05,
      "loss": 0.2202444076538086,
      "mean_token_accuracy": 0.9387628763914109,
      "num_tokens": 76291112.0,
      "step": 4740
    },
    {
      "aux_loss": 8.078449916839599,
      "entropy": 0.17308646738529204,
      "epoch": 0.004745,
      "grad_norm": 0.13061454892158508,
      "learning_rate": 2e-05,
      "loss": 0.19888010025024414,
      "mean_token_accuracy": 0.944050857424736,
      "num_tokens": 76371992.0,
      "step": 4745
    },
    {
      "aux_loss": 8.090685415267945,
      "entropy": 0.15347178801894187,
      "epoch": 0.00475,
      "grad_norm": 0.15530313551425934,
      "learning_rate": 2e-05,
      "loss": 0.17448859214782714,
      "mean_token_accuracy": 0.9517248034477234,
      "num_tokens": 76452652.0,
      "step": 4750
    },
    {
      "aux_loss": 8.095803833007812,
      "entropy": 0.1440520141273737,
      "epoch": 0.004755,
      "grad_norm": 0.1254265159368515,
      "learning_rate": 2e-05,
      "loss": 0.17638027667999268,
      "mean_token_accuracy": 0.9516173630952836,
      "num_tokens": 76530326.0,
      "step": 4755
    },
    {
      "aux_loss": 8.061766147613525,
      "entropy": 0.1593510240316391,
      "epoch": 0.00476,
      "grad_norm": 0.1422550082206726,
      "learning_rate": 2e-05,
      "loss": 0.20485897064208985,
      "mean_token_accuracy": 0.9428268611431122,
      "num_tokens": 76611731.0,
      "step": 4760
    },
    {
      "aux_loss": 8.065559148788452,
      "entropy": 0.18134529888629913,
      "epoch": 0.004765,
      "grad_norm": 0.14673522114753723,
      "learning_rate": 2e-05,
      "loss": 0.22229044437408446,
      "mean_token_accuracy": 0.937855190038681,
      "num_tokens": 76692544.0,
      "step": 4765
    },
    {
      "aux_loss": 8.069455003738403,
      "entropy": 0.19137707054615022,
      "epoch": 0.00477,
      "grad_norm": 0.13313166797161102,
      "learning_rate": 2e-05,
      "loss": 0.22386796474456788,
      "mean_token_accuracy": 0.9381320953369141,
      "num_tokens": 76773321.0,
      "step": 4770
    },
    {
      "aux_loss": 8.08010835647583,
      "entropy": 0.18328376412391661,
      "epoch": 0.004775,
      "grad_norm": 0.13736172020435333,
      "learning_rate": 2e-05,
      "loss": 0.20520222187042236,
      "mean_token_accuracy": 0.9428687363862991,
      "num_tokens": 76854322.0,
      "step": 4775
    },
    {
      "aux_loss": 8.086656522750854,
      "entropy": 0.16473961025476455,
      "epoch": 0.00478,
      "grad_norm": 0.14471887052059174,
      "learning_rate": 2e-05,
      "loss": 0.19400408267974853,
      "mean_token_accuracy": 0.9451741725206375,
      "num_tokens": 76935072.0,
      "step": 4780
    },
    {
      "aux_loss": 8.094878768920898,
      "entropy": 0.14804205670952797,
      "epoch": 0.004785,
      "grad_norm": 0.1242397353053093,
      "learning_rate": 2e-05,
      "loss": 0.17491521835327148,
      "mean_token_accuracy": 0.9517376780509949,
      "num_tokens": 77014990.0,
      "step": 4785
    },
    {
      "aux_loss": 8.05673713684082,
      "entropy": 0.17769003957509993,
      "epoch": 0.00479,
      "grad_norm": 0.13810604810714722,
      "learning_rate": 2e-05,
      "loss": 0.22061588764190673,
      "mean_token_accuracy": 0.9399046152830124,
      "num_tokens": 77096204.0,
      "step": 4790
    },
    {
      "aux_loss": 8.063803243637086,
      "entropy": 0.1688243556767702,
      "epoch": 0.004795,
      "grad_norm": 0.14647211134433746,
      "learning_rate": 2e-05,
      "loss": 0.2114875316619873,
      "mean_token_accuracy": 0.9427297741174698,
      "num_tokens": 77177244.0,
      "step": 4795
    },
    {
      "aux_loss": 8.071053886413575,
      "entropy": 0.1860219307243824,
      "epoch": 0.0048,
      "grad_norm": 0.1336272656917572,
      "learning_rate": 2e-05,
      "loss": 0.2182929754257202,
      "mean_token_accuracy": 0.9387612760066986,
      "num_tokens": 77258308.0,
      "step": 4800
    },
    {
      "aux_loss": 8.077130365371705,
      "entropy": 0.17437779307365417,
      "epoch": 0.004805,
      "grad_norm": 0.14074093103408813,
      "learning_rate": 2e-05,
      "loss": 0.2036818027496338,
      "mean_token_accuracy": 0.9434036195278168,
      "num_tokens": 77338936.0,
      "step": 4805
    },
    {
      "aux_loss": 8.08580207824707,
      "entropy": 0.16227206736803054,
      "epoch": 0.00481,
      "grad_norm": 0.1438792198896408,
      "learning_rate": 2e-05,
      "loss": 0.187655246257782,
      "mean_token_accuracy": 0.9477988064289093,
      "num_tokens": 77419517.0,
      "step": 4810
    },
    {
      "aux_loss": 8.100102949142457,
      "entropy": 0.14868246652185918,
      "epoch": 0.004815,
      "grad_norm": 0.14531210064888,
      "learning_rate": 2e-05,
      "loss": 0.17584103345870972,
      "mean_token_accuracy": 0.9517533600330352,
      "num_tokens": 77498691.0,
      "step": 4815
    },
    {
      "aux_loss": 8.06145749092102,
      "entropy": 0.17999815195798874,
      "epoch": 0.00482,
      "grad_norm": 0.15414494276046753,
      "learning_rate": 2e-05,
      "loss": 0.22422258853912352,
      "mean_token_accuracy": 0.9383078992366791,
      "num_tokens": 77580251.0,
      "step": 4820
    },
    {
      "aux_loss": 8.062994575500488,
      "entropy": 0.19598802849650382,
      "epoch": 0.004825,
      "grad_norm": 0.13477541506290436,
      "learning_rate": 2e-05,
      "loss": 0.23447961807250978,
      "mean_token_accuracy": 0.9342438578605652,
      "num_tokens": 77661720.0,
      "step": 4825
    },
    {
      "aux_loss": 8.06834659576416,
      "entropy": 0.19157903753221034,
      "epoch": 0.00483,
      "grad_norm": 0.13233202695846558,
      "learning_rate": 2e-05,
      "loss": 0.2236931562423706,
      "mean_token_accuracy": 0.9371049910783767,
      "num_tokens": 77742821.0,
      "step": 4830
    },
    {
      "aux_loss": 8.079009389877319,
      "entropy": 0.19248158112168312,
      "epoch": 0.004835,
      "grad_norm": 0.14351986348628998,
      "learning_rate": 2e-05,
      "loss": 0.21961729526519774,
      "mean_token_accuracy": 0.9385960161685943,
      "num_tokens": 77823431.0,
      "step": 4835
    },
    {
      "aux_loss": 8.0873046875,
      "entropy": 0.16416383236646653,
      "epoch": 0.00484,
      "grad_norm": 0.1315767765045166,
      "learning_rate": 2e-05,
      "loss": 0.18500778675079346,
      "mean_token_accuracy": 0.9487603038549424,
      "num_tokens": 77904075.0,
      "step": 4840
    },
    {
      "aux_loss": 8.105883741378785,
      "entropy": 0.15333222597837448,
      "epoch": 0.004845,
      "grad_norm": 0.1693802773952484,
      "learning_rate": 2e-05,
      "loss": 0.1815592885017395,
      "mean_token_accuracy": 0.9489593893289566,
      "num_tokens": 77984572.0,
      "step": 4845
    },
    {
      "aux_loss": 8.063505125045776,
      "entropy": 0.1847516719251871,
      "epoch": 0.00485,
      "grad_norm": 0.14307129383087158,
      "learning_rate": 2e-05,
      "loss": 0.22595858573913574,
      "mean_token_accuracy": 0.938094162940979,
      "num_tokens": 78062556.0,
      "step": 4850
    },
    {
      "aux_loss": 8.06514253616333,
      "entropy": 0.18040362000465393,
      "epoch": 0.004855,
      "grad_norm": 0.15260963141918182,
      "learning_rate": 2e-05,
      "loss": 0.22822363376617433,
      "mean_token_accuracy": 0.9371621757745743,
      "num_tokens": 78143960.0,
      "step": 4855
    },
    {
      "aux_loss": 8.069723987579346,
      "entropy": 0.16827748566865922,
      "epoch": 0.00486,
      "grad_norm": 0.13363219797611237,
      "learning_rate": 2e-05,
      "loss": 0.20340406894683838,
      "mean_token_accuracy": 0.9436208039522171,
      "num_tokens": 78225141.0,
      "step": 4860
    },
    {
      "aux_loss": 8.07733221054077,
      "entropy": 0.18799750804901122,
      "epoch": 0.004865,
      "grad_norm": 0.13352744281291962,
      "learning_rate": 2e-05,
      "loss": 0.21118266582489015,
      "mean_token_accuracy": 0.9427341043949127,
      "num_tokens": 78305728.0,
      "step": 4865
    },
    {
      "aux_loss": 8.08578381538391,
      "entropy": 0.17517005577683448,
      "epoch": 0.00487,
      "grad_norm": 0.14015154540538788,
      "learning_rate": 2e-05,
      "loss": 0.1990761399269104,
      "mean_token_accuracy": 0.9440957307815552,
      "num_tokens": 78386350.0,
      "step": 4870
    },
    {
      "aux_loss": 8.0989990234375,
      "entropy": 0.1563259456306696,
      "epoch": 0.004875,
      "grad_norm": 0.15986883640289307,
      "learning_rate": 2e-05,
      "loss": 0.18830337524414062,
      "mean_token_accuracy": 0.948091197013855,
      "num_tokens": 78466611.0,
      "step": 4875
    },
    {
      "aux_loss": 8.068572950363158,
      "entropy": 0.16635826528072356,
      "epoch": 0.00488,
      "grad_norm": 0.1429399698972702,
      "learning_rate": 2e-05,
      "loss": 0.20412166118621827,
      "mean_token_accuracy": 0.9434427320957184,
      "num_tokens": 78545911.0,
      "step": 4880
    },
    {
      "aux_loss": 8.063074445724487,
      "entropy": 0.19215986505150795,
      "epoch": 0.004885,
      "grad_norm": 0.14082340896129608,
      "learning_rate": 2e-05,
      "loss": 0.2396674394607544,
      "mean_token_accuracy": 0.9329192250967026,
      "num_tokens": 78627110.0,
      "step": 4885
    },
    {
      "aux_loss": 8.069276094436646,
      "entropy": 0.18511325493454933,
      "epoch": 0.00489,
      "grad_norm": 0.12917843461036682,
      "learning_rate": 2e-05,
      "loss": 0.2201209306716919,
      "mean_token_accuracy": 0.9387602388858796,
      "num_tokens": 78708416.0,
      "step": 4890
    },
    {
      "aux_loss": 8.074924802780151,
      "entropy": 0.19096734598279,
      "epoch": 0.004895,
      "grad_norm": 0.13760483264923096,
      "learning_rate": 2e-05,
      "loss": 0.21770994663238524,
      "mean_token_accuracy": 0.9395693838596344,
      "num_tokens": 78789288.0,
      "step": 4895
    },
    {
      "aux_loss": 8.084355926513672,
      "entropy": 0.1727454349398613,
      "epoch": 0.0049,
      "grad_norm": 0.13361328840255737,
      "learning_rate": 2e-05,
      "loss": 0.195175838470459,
      "mean_token_accuracy": 0.9459042578935624,
      "num_tokens": 78870329.0,
      "step": 4900
    },
    {
      "aux_loss": 8.09723253250122,
      "entropy": 0.152843826264143,
      "epoch": 0.004905,
      "grad_norm": 0.1537516564130783,
      "learning_rate": 2e-05,
      "loss": 0.17511388063430786,
      "mean_token_accuracy": 0.9508573204278946,
      "num_tokens": 78950772.0,
      "step": 4905
    },
    {
      "aux_loss": 8.075333881378175,
      "entropy": 0.15962165966629982,
      "epoch": 0.00491,
      "grad_norm": 0.15547269582748413,
      "learning_rate": 2e-05,
      "loss": 0.20460450649261475,
      "mean_token_accuracy": 0.9436504900455475,
      "num_tokens": 79028899.0,
      "step": 4910
    },
    {
      "aux_loss": 8.065459394454956,
      "entropy": 0.17354790680110455,
      "epoch": 0.004915,
      "grad_norm": 0.15293730795383453,
      "learning_rate": 2e-05,
      "loss": 0.22032437324523926,
      "mean_token_accuracy": 0.9393253415822983,
      "num_tokens": 79110116.0,
      "step": 4915
    },
    {
      "aux_loss": 8.070332384109497,
      "entropy": 0.18256853595376016,
      "epoch": 0.00492,
      "grad_norm": 0.128585085272789,
      "learning_rate": 2e-05,
      "loss": 0.2146277904510498,
      "mean_token_accuracy": 0.9398369520902634,
      "num_tokens": 79191137.0,
      "step": 4920
    },
    {
      "aux_loss": 8.07633810043335,
      "entropy": 0.1892526388168335,
      "epoch": 0.004925,
      "grad_norm": 0.1317615956068039,
      "learning_rate": 2e-05,
      "loss": 0.2122637987136841,
      "mean_token_accuracy": 0.9413473963737488,
      "num_tokens": 79272063.0,
      "step": 4925
    },
    {
      "aux_loss": 8.083935403823853,
      "entropy": 0.17361845299601555,
      "epoch": 0.00493,
      "grad_norm": 0.14347580075263977,
      "learning_rate": 2e-05,
      "loss": 0.20168979167938234,
      "mean_token_accuracy": 0.9447771817445755,
      "num_tokens": 79352811.0,
      "step": 4930
    },
    {
      "aux_loss": 8.092362785339356,
      "entropy": 0.15533897280693054,
      "epoch": 0.004935,
      "grad_norm": 0.15174630284309387,
      "learning_rate": 2e-05,
      "loss": 0.18947105407714843,
      "mean_token_accuracy": 0.9474773168563843,
      "num_tokens": 79432879.0,
      "step": 4935
    },
    {
      "aux_loss": 8.081296300888061,
      "entropy": 0.16013988479971886,
      "epoch": 0.00494,
      "grad_norm": 0.1388075351715088,
      "learning_rate": 2e-05,
      "loss": 0.1940844774246216,
      "mean_token_accuracy": 0.947017589211464,
      "num_tokens": 79513506.0,
      "step": 4940
    },
    {
      "aux_loss": 8.064167737960815,
      "entropy": 0.18577812500298024,
      "epoch": 0.004945,
      "grad_norm": 0.1492556631565094,
      "learning_rate": 2e-05,
      "loss": 0.22762320041656495,
      "mean_token_accuracy": 0.9366672039031982,
      "num_tokens": 79594247.0,
      "step": 4945
    },
    {
      "aux_loss": 8.068421649932862,
      "entropy": 0.19005856439471244,
      "epoch": 0.00495,
      "grad_norm": 0.13822248578071594,
      "learning_rate": 2e-05,
      "loss": 0.22751774787902831,
      "mean_token_accuracy": 0.9371238201856613,
      "num_tokens": 79675230.0,
      "step": 4950
    },
    {
      "aux_loss": 8.072866487503052,
      "entropy": 0.18283550590276718,
      "epoch": 0.004955,
      "grad_norm": 0.13207697868347168,
      "learning_rate": 2e-05,
      "loss": 0.20969955921173095,
      "mean_token_accuracy": 0.9412348508834839,
      "num_tokens": 79755840.0,
      "step": 4955
    },
    {
      "aux_loss": 8.082518911361694,
      "entropy": 0.1677012600004673,
      "epoch": 0.00496,
      "grad_norm": 0.13073845207691193,
      "learning_rate": 2e-05,
      "loss": 0.18848557472229005,
      "mean_token_accuracy": 0.947378334403038,
      "num_tokens": 79836457.0,
      "step": 4960
    },
    {
      "aux_loss": 8.093623876571655,
      "entropy": 0.14853578731417655,
      "epoch": 0.004965,
      "grad_norm": 0.15753896534442902,
      "learning_rate": 2e-05,
      "loss": 0.17656341791152955,
      "mean_token_accuracy": 0.9510572373867034,
      "num_tokens": 79917044.0,
      "step": 4965
    },
    {
      "aux_loss": 8.076998138427735,
      "entropy": 0.1635641474276781,
      "epoch": 0.00497,
      "grad_norm": 0.14596302807331085,
      "learning_rate": 2e-05,
      "loss": 0.20475146770477295,
      "mean_token_accuracy": 0.9442510575056076,
      "num_tokens": 79997860.0,
      "step": 4970
    },
    {
      "aux_loss": 8.05983076095581,
      "entropy": 0.18420847207307817,
      "epoch": 0.004975,
      "grad_norm": 0.1406341940164566,
      "learning_rate": 2e-05,
      "loss": 0.22545886039733887,
      "mean_token_accuracy": 0.938109427690506,
      "num_tokens": 80078637.0,
      "step": 4975
    },
    {
      "aux_loss": 8.065752983093262,
      "entropy": 0.20092643201351165,
      "epoch": 0.00498,
      "grad_norm": 0.13014748692512512,
      "learning_rate": 2e-05,
      "loss": 0.24155800342559813,
      "mean_token_accuracy": 0.9334942549467087,
      "num_tokens": 80159684.0,
      "step": 4980
    },
    {
      "aux_loss": 8.069738149642944,
      "entropy": 0.18131208866834642,
      "epoch": 0.004985,
      "grad_norm": 0.128971129655838,
      "learning_rate": 2e-05,
      "loss": 0.206261944770813,
      "mean_token_accuracy": 0.9428774952888489,
      "num_tokens": 80240406.0,
      "step": 4985
    },
    {
      "aux_loss": 8.080774974822997,
      "entropy": 0.17356322705745697,
      "epoch": 0.00499,
      "grad_norm": 0.13693855702877045,
      "learning_rate": 2e-05,
      "loss": 0.19833028316497803,
      "mean_token_accuracy": 0.9450587689876556,
      "num_tokens": 80321090.0,
      "step": 4990
    },
    {
      "aux_loss": 8.088388538360595,
      "entropy": 0.1589390866458416,
      "epoch": 0.004995,
      "grad_norm": 0.144724503159523,
      "learning_rate": 2e-05,
      "loss": 0.1885025382041931,
      "mean_token_accuracy": 0.9478643119335175,
      "num_tokens": 80401591.0,
      "step": 4995
    },
    {
      "aux_loss": 8.086220598220825,
      "entropy": 0.1696808345615864,
      "epoch": 0.005,
      "grad_norm": 0.14380449056625366,
      "learning_rate": 2e-05,
      "loss": 0.20389807224273682,
      "mean_token_accuracy": 0.9446290701627731,
      "num_tokens": 80481663.0,
      "step": 5000
    },
    {
      "aux_loss": 8.061460494995117,
      "entropy": 0.17909952998161316,
      "epoch": 0.005005,
      "grad_norm": 0.1437544822692871,
      "learning_rate": 2e-05,
      "loss": 0.22281675338745116,
      "mean_token_accuracy": 0.937743729352951,
      "num_tokens": 80562998.0,
      "step": 5005
    },
    {
      "aux_loss": 8.062940502166748,
      "entropy": 0.17133663594722748,
      "epoch": 0.00501,
      "grad_norm": 0.13513287901878357,
      "learning_rate": 2e-05,
      "loss": 0.20670864582061768,
      "mean_token_accuracy": 0.942376172542572,
      "num_tokens": 80643879.0,
      "step": 5010
    },
    {
      "aux_loss": 8.07051615715027,
      "entropy": 0.18950156792998313,
      "epoch": 0.005015,
      "grad_norm": 0.1351640522480011,
      "learning_rate": 2e-05,
      "loss": 0.2214902639389038,
      "mean_token_accuracy": 0.9381137132644654,
      "num_tokens": 80724362.0,
      "step": 5015
    },
    {
      "aux_loss": 8.077671670913697,
      "entropy": 0.1622502975165844,
      "epoch": 0.00502,
      "grad_norm": 0.1385820209980011,
      "learning_rate": 2e-05,
      "loss": 0.1887761116027832,
      "mean_token_accuracy": 0.9463288217782975,
      "num_tokens": 80804977.0,
      "step": 5020
    },
    {
      "aux_loss": 8.087444496154784,
      "entropy": 0.164411199092865,
      "epoch": 0.005025,
      "grad_norm": 0.14165249466896057,
      "learning_rate": 2e-05,
      "loss": 0.1889095425605774,
      "mean_token_accuracy": 0.9469385474920273,
      "num_tokens": 80885739.0,
      "step": 5025
    },
    {
      "aux_loss": 8.089562845230102,
      "entropy": 0.14132896289229394,
      "epoch": 0.00503,
      "grad_norm": 0.12848259508609772,
      "learning_rate": 2e-05,
      "loss": 0.17200437784194947,
      "mean_token_accuracy": 0.9527578234672547,
      "num_tokens": 80965365.0,
      "step": 5030
    },
    {
      "aux_loss": 8.059129858016968,
      "entropy": 0.18484954573214055,
      "epoch": 0.005035,
      "grad_norm": 0.15159401297569275,
      "learning_rate": 2e-05,
      "loss": 0.23313052654266359,
      "mean_token_accuracy": 0.9361837476491928,
      "num_tokens": 81046759.0,
      "step": 5035
    },
    {
      "aux_loss": 8.065659332275391,
      "entropy": 0.18609695509076118,
      "epoch": 0.00504,
      "grad_norm": 0.13380348682403564,
      "learning_rate": 2e-05,
      "loss": 0.22719082832336426,
      "mean_token_accuracy": 0.9366167664527894,
      "num_tokens": 81127339.0,
      "step": 5040
    },
    {
      "aux_loss": 8.071785831451416,
      "entropy": 0.18999304734170436,
      "epoch": 0.005045,
      "grad_norm": 0.12963952124118805,
      "learning_rate": 2e-05,
      "loss": 0.2202707290649414,
      "mean_token_accuracy": 0.9385194927453995,
      "num_tokens": 81208150.0,
      "step": 5045
    },
    {
      "aux_loss": 8.075503396987916,
      "entropy": 0.17498328685760497,
      "epoch": 0.00505,
      "grad_norm": 0.13342489302158356,
      "learning_rate": 2e-05,
      "loss": 0.20067176818847657,
      "mean_token_accuracy": 0.944417205452919,
      "num_tokens": 81288788.0,
      "step": 5050
    },
    {
      "aux_loss": 8.085416507720947,
      "entropy": 0.1612071469426155,
      "epoch": 0.005055,
      "grad_norm": 0.13029880821704865,
      "learning_rate": 2e-05,
      "loss": 0.18787113428115845,
      "mean_token_accuracy": 0.9476497143507003,
      "num_tokens": 81369821.0,
      "step": 5055
    },
    {
      "aux_loss": 8.094400930404664,
      "entropy": 0.14648021943867207,
      "epoch": 0.00506,
      "grad_norm": 0.12763769924640656,
      "learning_rate": 2e-05,
      "loss": 0.17607131004333496,
      "mean_token_accuracy": 0.9516687840223312,
      "num_tokens": 81447752.0,
      "step": 5060
    },
    {
      "aux_loss": 8.056461143493653,
      "entropy": 0.1861129380762577,
      "epoch": 0.005065,
      "grad_norm": 0.14374208450317383,
      "learning_rate": 2e-05,
      "loss": 0.23259367942810058,
      "mean_token_accuracy": 0.9371941596269607,
      "num_tokens": 81529164.0,
      "step": 5065
    },
    {
      "aux_loss": 8.06020212173462,
      "entropy": 0.1942598707973957,
      "epoch": 0.00507,
      "grad_norm": 0.13377441465854645,
      "learning_rate": 2e-05,
      "loss": 0.23065507411956787,
      "mean_token_accuracy": 0.9355861335992813,
      "num_tokens": 81609913.0,
      "step": 5070
    },
    {
      "aux_loss": 8.068313694000244,
      "entropy": 0.19032789394259453,
      "epoch": 0.005075,
      "grad_norm": 0.13738398253917694,
      "learning_rate": 2e-05,
      "loss": 0.2128575325012207,
      "mean_token_accuracy": 0.9400048702955246,
      "num_tokens": 81690581.0,
      "step": 5075
    },
    {
      "aux_loss": 8.078317451477051,
      "entropy": 0.18149514123797417,
      "epoch": 0.00508,
      "grad_norm": 0.14576472342014313,
      "learning_rate": 2e-05,
      "loss": 0.21017444133758545,
      "mean_token_accuracy": 0.9409641116857529,
      "num_tokens": 81770764.0,
      "step": 5080
    },
    {
      "aux_loss": 8.087953996658324,
      "entropy": 0.16609180569648743,
      "epoch": 0.005085,
      "grad_norm": 0.14495544135570526,
      "learning_rate": 2e-05,
      "loss": 0.19417307376861573,
      "mean_token_accuracy": 0.9456107228994369,
      "num_tokens": 81851547.0,
      "step": 5085
    },
    {
      "aux_loss": 8.098352146148681,
      "entropy": 0.14779425226151943,
      "epoch": 0.00509,
      "grad_norm": 0.1287878304719925,
      "learning_rate": 2e-05,
      "loss": 0.17721613645553588,
      "mean_token_accuracy": 0.9516434460878372,
      "num_tokens": 81930829.0,
      "step": 5090
    },
    {
      "aux_loss": 8.057374095916748,
      "entropy": 0.17751976512372494,
      "epoch": 0.005095,
      "grad_norm": 0.14047560095787048,
      "learning_rate": 2e-05,
      "loss": 0.22454257011413575,
      "mean_token_accuracy": 0.9388383090496063,
      "num_tokens": 82012400.0,
      "step": 5095
    },
    {
      "aux_loss": 8.061855792999268,
      "entropy": 0.1945471465587616,
      "epoch": 0.0051,
      "grad_norm": 0.14557085931301117,
      "learning_rate": 2e-05,
      "loss": 0.2323601245880127,
      "mean_token_accuracy": 0.9356596738100051,
      "num_tokens": 82093115.0,
      "step": 5100
    },
    {
      "aux_loss": 8.068995046615601,
      "entropy": 0.18055291585624217,
      "epoch": 0.005105,
      "grad_norm": 0.1313052773475647,
      "learning_rate": 2e-05,
      "loss": 0.21209371089935303,
      "mean_token_accuracy": 0.9422195374965667,
      "num_tokens": 82173958.0,
      "step": 5105
    },
    {
      "aux_loss": 8.08045072555542,
      "entropy": 0.1748869799077511,
      "epoch": 0.00511,
      "grad_norm": 0.1337028592824936,
      "learning_rate": 2e-05,
      "loss": 0.1990726947784424,
      "mean_token_accuracy": 0.9446351289749145,
      "num_tokens": 82254669.0,
      "step": 5110
    },
    {
      "aux_loss": 8.086453580856324,
      "entropy": 0.15801681503653525,
      "epoch": 0.005115,
      "grad_norm": 0.1370856761932373,
      "learning_rate": 2e-05,
      "loss": 0.18352211713790895,
      "mean_token_accuracy": 0.9489396572113037,
      "num_tokens": 82335264.0,
      "step": 5115
    },
    {
      "aux_loss": 8.097521829605103,
      "entropy": 0.1501435101032257,
      "epoch": 0.00512,
      "grad_norm": 0.1464599072933197,
      "learning_rate": 2e-05,
      "loss": 0.18293447494506837,
      "mean_token_accuracy": 0.9493194729089737,
      "num_tokens": 82412840.0,
      "step": 5120
    },
    {
      "aux_loss": 8.0606849193573,
      "entropy": 0.18269062973558903,
      "epoch": 0.005125,
      "grad_norm": 0.14097903668880463,
      "learning_rate": 2e-05,
      "loss": 0.22738711833953856,
      "mean_token_accuracy": 0.9382804304361343,
      "num_tokens": 82494410.0,
      "step": 5125
    },
    {
      "aux_loss": 8.05986065864563,
      "entropy": 0.20089913681149482,
      "epoch": 0.00513,
      "grad_norm": 0.13728372752666473,
      "learning_rate": 2e-05,
      "loss": 0.24180009365081787,
      "mean_token_accuracy": 0.9336772918701172,
      "num_tokens": 82575568.0,
      "step": 5130
    },
    {
      "aux_loss": 8.06652398109436,
      "entropy": 0.19021136164665223,
      "epoch": 0.005135,
      "grad_norm": 0.13539330661296844,
      "learning_rate": 2e-05,
      "loss": 0.22068052291870116,
      "mean_token_accuracy": 0.93862384557724,
      "num_tokens": 82656732.0,
      "step": 5135
    },
    {
      "aux_loss": 8.076739740371703,
      "entropy": 0.18896769285202025,
      "epoch": 0.00514,
      "grad_norm": 0.14026601612567902,
      "learning_rate": 2e-05,
      "loss": 0.21176626682281494,
      "mean_token_accuracy": 0.9415340900421143,
      "num_tokens": 82737361.0,
      "step": 5140
    },
    {
      "aux_loss": 8.085612916946411,
      "entropy": 0.16142391487956048,
      "epoch": 0.005145,
      "grad_norm": 0.14975136518478394,
      "learning_rate": 2e-05,
      "loss": 0.18413556814193727,
      "mean_token_accuracy": 0.9488928973674774,
      "num_tokens": 82817843.0,
      "step": 5145
    },
    {
      "aux_loss": 8.102252101898193,
      "entropy": 0.14715617932379246,
      "epoch": 0.00515,
      "grad_norm": 0.18000397086143494,
      "learning_rate": 2e-05,
      "loss": 0.17973248958587645,
      "mean_token_accuracy": 0.9500347584486007,
      "num_tokens": 82895353.0,
      "step": 5150
    },
    {
      "aux_loss": 8.057420682907104,
      "entropy": 0.1684798751026392,
      "epoch": 0.005155,
      "grad_norm": 0.15003588795661926,
      "learning_rate": 2e-05,
      "loss": 0.2151956081390381,
      "mean_token_accuracy": 0.9406437695026397,
      "num_tokens": 82977015.0,
      "step": 5155
    },
    {
      "aux_loss": 8.05977005958557,
      "entropy": 0.19859195798635482,
      "epoch": 0.00516,
      "grad_norm": 0.13261450827121735,
      "learning_rate": 2e-05,
      "loss": 0.24169466495513917,
      "mean_token_accuracy": 0.9339892268180847,
      "num_tokens": 83058388.0,
      "step": 5160
    },
    {
      "aux_loss": 8.066183423995971,
      "entropy": 0.20120759382843972,
      "epoch": 0.005165,
      "grad_norm": 0.1291927993297577,
      "learning_rate": 2e-05,
      "loss": 0.2316272497177124,
      "mean_token_accuracy": 0.9348942846059799,
      "num_tokens": 83139193.0,
      "step": 5165
    },
    {
      "aux_loss": 8.07408447265625,
      "entropy": 0.1641852840781212,
      "epoch": 0.00517,
      "grad_norm": 0.12719838321208954,
      "learning_rate": 2e-05,
      "loss": 0.18633654117584228,
      "mean_token_accuracy": 0.9486995756626129,
      "num_tokens": 83219768.0,
      "step": 5170
    },
    {
      "aux_loss": 8.085832214355468,
      "entropy": 0.16438565030694008,
      "epoch": 0.005175,
      "grad_norm": 0.14058950543403625,
      "learning_rate": 2e-05,
      "loss": 0.19205108880996705,
      "mean_token_accuracy": 0.9463698744773865,
      "num_tokens": 83300341.0,
      "step": 5175
    },
    {
      "aux_loss": 8.092932653427123,
      "entropy": 0.14931646287441253,
      "epoch": 0.00518,
      "grad_norm": 0.14637906849384308,
      "learning_rate": 2e-05,
      "loss": 0.17636841535568237,
      "mean_token_accuracy": 0.9511190354824066,
      "num_tokens": 83380113.0,
      "step": 5180
    },
    {
      "aux_loss": 8.071480417251587,
      "entropy": 0.16499637849628926,
      "epoch": 0.005185,
      "grad_norm": 0.14911912381649017,
      "learning_rate": 2e-05,
      "loss": 0.20396580696105956,
      "mean_token_accuracy": 0.9451363801956176,
      "num_tokens": 83461026.0,
      "step": 5185
    },
    {
      "aux_loss": 8.063604736328125,
      "entropy": 0.18016661144793034,
      "epoch": 0.00519,
      "grad_norm": 0.15766309201717377,
      "learning_rate": 2e-05,
      "loss": 0.22609267234802247,
      "mean_token_accuracy": 0.9369242399930954,
      "num_tokens": 83541945.0,
      "step": 5190
    },
    {
      "aux_loss": 8.064621305465698,
      "entropy": 0.1844883807003498,
      "epoch": 0.005195,
      "grad_norm": 0.13181759417057037,
      "learning_rate": 2e-05,
      "loss": 0.217124342918396,
      "mean_token_accuracy": 0.9403899610042572,
      "num_tokens": 83622962.0,
      "step": 5195
    },
    {
      "aux_loss": 8.071509265899659,
      "entropy": 0.19331589862704276,
      "epoch": 0.0052,
      "grad_norm": 0.1289542019367218,
      "learning_rate": 2e-05,
      "loss": 0.21748409271240235,
      "mean_token_accuracy": 0.9402452617883682,
      "num_tokens": 83703623.0,
      "step": 5200
    },
    {
      "aux_loss": 8.077685022354126,
      "entropy": 0.17031075283885003,
      "epoch": 0.005205,
      "grad_norm": 0.1487719863653183,
      "learning_rate": 2e-05,
      "loss": 0.19834980964660645,
      "mean_token_accuracy": 0.9444656133651733,
      "num_tokens": 83784228.0,
      "step": 5205
    },
    {
      "aux_loss": 8.086824178695679,
      "entropy": 0.16505861431360244,
      "epoch": 0.00521,
      "grad_norm": 0.1417459398508072,
      "learning_rate": 2e-05,
      "loss": 0.1932835340499878,
      "mean_token_accuracy": 0.9459480851888656,
      "num_tokens": 83864661.0,
      "step": 5210
    },
    {
      "aux_loss": 8.082101011276245,
      "entropy": 0.1893757451325655,
      "epoch": 0.005215,
      "grad_norm": 0.1346246898174286,
      "learning_rate": 2e-05,
      "loss": 0.22587037086486816,
      "mean_token_accuracy": 0.9373359382152557,
      "num_tokens": 83945040.0,
      "step": 5215
    },
    {
      "aux_loss": 8.060264587402344,
      "entropy": 0.18647052422165872,
      "epoch": 0.00522,
      "grad_norm": 0.13630251586437225,
      "learning_rate": 2e-05,
      "loss": 0.229781174659729,
      "mean_token_accuracy": 0.9363776117563247,
      "num_tokens": 84025943.0,
      "step": 5220
    },
    {
      "aux_loss": 8.063036727905274,
      "entropy": 0.18653476648032666,
      "epoch": 0.005225,
      "grad_norm": 0.13629674911499023,
      "learning_rate": 2e-05,
      "loss": 0.22703688144683837,
      "mean_token_accuracy": 0.9371970355510711,
      "num_tokens": 84107044.0,
      "step": 5225
    },
    {
      "aux_loss": 8.070794439315796,
      "entropy": 0.18378162607550622,
      "epoch": 0.00523,
      "grad_norm": 0.13028231263160706,
      "learning_rate": 2e-05,
      "loss": 0.2100696563720703,
      "mean_token_accuracy": 0.941874486207962,
      "num_tokens": 84188012.0,
      "step": 5230
    },
    {
      "aux_loss": 8.077144384384155,
      "entropy": 0.17059426307678222,
      "epoch": 0.005235,
      "grad_norm": 0.12864133715629578,
      "learning_rate": 2e-05,
      "loss": 0.1947857141494751,
      "mean_token_accuracy": 0.9454405397176743,
      "num_tokens": 84268484.0,
      "step": 5235
    },
    {
      "aux_loss": 8.088841867446899,
      "entropy": 0.1648114152252674,
      "epoch": 0.00524,
      "grad_norm": 0.14165791869163513,
      "learning_rate": 2e-05,
      "loss": 0.19274320602416992,
      "mean_token_accuracy": 0.9462246894836426,
      "num_tokens": 84349041.0,
      "step": 5240
    },
    {
      "aux_loss": 8.083983564376831,
      "entropy": 0.15550471656024456,
      "epoch": 0.005245,
      "grad_norm": 0.12493729591369629,
      "learning_rate": 2e-05,
      "loss": 0.18871302604675294,
      "mean_token_accuracy": 0.9482835561037064,
      "num_tokens": 84428624.0,
      "step": 5245
    },
    {
      "aux_loss": 8.058908605575562,
      "entropy": 0.2004340335726738,
      "epoch": 0.00525,
      "grad_norm": 0.13948389887809753,
      "learning_rate": 2e-05,
      "loss": 0.24442296028137206,
      "mean_token_accuracy": 0.93235542178154,
      "num_tokens": 84509816.0,
      "step": 5250
    },
    {
      "aux_loss": 8.064950942993164,
      "entropy": 0.19418388232588768,
      "epoch": 0.005255,
      "grad_norm": 0.13848178088665009,
      "learning_rate": 2e-05,
      "loss": 0.23547282218933105,
      "mean_token_accuracy": 0.9355094879865646,
      "num_tokens": 84591063.0,
      "step": 5255
    },
    {
      "aux_loss": 8.071905946731567,
      "entropy": 0.188873939961195,
      "epoch": 0.00526,
      "grad_norm": 0.13631193339824677,
      "learning_rate": 2e-05,
      "loss": 0.21399741172790526,
      "mean_token_accuracy": 0.9406247943639755,
      "num_tokens": 84672185.0,
      "step": 5260
    },
    {
      "aux_loss": 8.078228664398193,
      "entropy": 0.17337775602936745,
      "epoch": 0.005265,
      "grad_norm": 0.1347685158252716,
      "learning_rate": 2e-05,
      "loss": 0.19666157960891723,
      "mean_token_accuracy": 0.9445564061403274,
      "num_tokens": 84753131.0,
      "step": 5265
    },
    {
      "aux_loss": 8.090518760681153,
      "entropy": 0.15564620941877366,
      "epoch": 0.00527,
      "grad_norm": 0.14129430055618286,
      "learning_rate": 2e-05,
      "loss": 0.18422791957855225,
      "mean_token_accuracy": 0.9499352663755417,
      "num_tokens": 84833759.0,
      "step": 5270
    },
    {
      "aux_loss": 8.08458776473999,
      "entropy": 0.1570093970745802,
      "epoch": 0.005275,
      "grad_norm": 0.16710075736045837,
      "learning_rate": 2e-05,
      "loss": 0.1880483865737915,
      "mean_token_accuracy": 0.9488036781549454,
      "num_tokens": 84913344.0,
      "step": 5275
    },
    {
      "aux_loss": 8.063005447387695,
      "entropy": 0.183569348603487,
      "epoch": 0.00528,
      "grad_norm": 0.14091730117797852,
      "learning_rate": 2e-05,
      "loss": 0.22685303688049316,
      "mean_token_accuracy": 0.9369056910276413,
      "num_tokens": 84994646.0,
      "step": 5280
    },
    {
      "aux_loss": 8.064584589004516,
      "entropy": 0.17309212908148766,
      "epoch": 0.005285,
      "grad_norm": 0.1365366280078888,
      "learning_rate": 2e-05,
      "loss": 0.21304657459259033,
      "mean_token_accuracy": 0.9410705208778382,
      "num_tokens": 85075695.0,
      "step": 5285
    },
    {
      "aux_loss": 8.074301385879517,
      "entropy": 0.1822326511144638,
      "epoch": 0.00529,
      "grad_norm": 0.13143017888069153,
      "learning_rate": 2e-05,
      "loss": 0.21302130222320556,
      "mean_token_accuracy": 0.9400135189294815,
      "num_tokens": 85156515.0,
      "step": 5290
    },
    {
      "aux_loss": 8.07977910041809,
      "entropy": 0.17323933616280557,
      "epoch": 0.005295,
      "grad_norm": 0.13945405185222626,
      "learning_rate": 2e-05,
      "loss": 0.19553561210632325,
      "mean_token_accuracy": 0.9454155951738358,
      "num_tokens": 85237456.0,
      "step": 5295
    },
    {
      "aux_loss": 8.089977407455445,
      "entropy": 0.159903197363019,
      "epoch": 0.0053,
      "grad_norm": 0.14117389917373657,
      "learning_rate": 2e-05,
      "loss": 0.1822801947593689,
      "mean_token_accuracy": 0.9490360617637634,
      "num_tokens": 85317735.0,
      "step": 5300
    },
    {
      "aux_loss": 8.084252834320068,
      "entropy": 0.16214874982833863,
      "epoch": 0.005305,
      "grad_norm": 0.1376124620437622,
      "learning_rate": 2e-05,
      "loss": 0.1983538508415222,
      "mean_token_accuracy": 0.9464780122041703,
      "num_tokens": 85396313.0,
      "step": 5305
    },
    {
      "aux_loss": 8.060075807571412,
      "entropy": 0.18716093078255652,
      "epoch": 0.00531,
      "grad_norm": 0.14581257104873657,
      "learning_rate": 2e-05,
      "loss": 0.23828446865081787,
      "mean_token_accuracy": 0.934212327003479,
      "num_tokens": 85477074.0,
      "step": 5310
    },
    {
      "aux_loss": 8.066683912277222,
      "entropy": 0.18466987274587154,
      "epoch": 0.005315,
      "grad_norm": 0.12460780888795853,
      "learning_rate": 2e-05,
      "loss": 0.22361156940460206,
      "mean_token_accuracy": 0.9388301581144333,
      "num_tokens": 85557935.0,
      "step": 5315
    },
    {
      "aux_loss": 8.070351076126098,
      "entropy": 0.18355950862169265,
      "epoch": 0.00532,
      "grad_norm": 0.13842551410198212,
      "learning_rate": 2e-05,
      "loss": 0.20714399814605713,
      "mean_token_accuracy": 0.9423571646213531,
      "num_tokens": 85638600.0,
      "step": 5320
    },
    {
      "aux_loss": 8.080270481109618,
      "entropy": 0.16325409561395646,
      "epoch": 0.005325,
      "grad_norm": 0.13353529572486877,
      "learning_rate": 2e-05,
      "loss": 0.18565356731414795,
      "mean_token_accuracy": 0.9481933623552322,
      "num_tokens": 85719369.0,
      "step": 5325
    },
    {
      "aux_loss": 8.085492944717407,
      "entropy": 0.16201958134770394,
      "epoch": 0.00533,
      "grad_norm": 0.137704998254776,
      "learning_rate": 2e-05,
      "loss": 0.19290823936462403,
      "mean_token_accuracy": 0.9460921108722686,
      "num_tokens": 85799996.0,
      "step": 5330
    },
    {
      "aux_loss": 8.087014818191529,
      "entropy": 0.14468047209084034,
      "epoch": 0.005335,
      "grad_norm": 0.12653417885303497,
      "learning_rate": 2e-05,
      "loss": 0.17807286977767944,
      "mean_token_accuracy": 0.9512067526578903,
      "num_tokens": 85880492.0,
      "step": 5335
    },
    {
      "aux_loss": 8.060007762908935,
      "entropy": 0.18660878837108613,
      "epoch": 0.00534,
      "grad_norm": 0.13945764303207397,
      "learning_rate": 2e-05,
      "loss": 0.23587532043457032,
      "mean_token_accuracy": 0.9343234062194824,
      "num_tokens": 85961912.0,
      "step": 5340
    },
    {
      "aux_loss": 8.062325048446656,
      "entropy": 0.1941206455230713,
      "epoch": 0.005345,
      "grad_norm": 0.13073861598968506,
      "learning_rate": 2e-05,
      "loss": 0.2310464859008789,
      "mean_token_accuracy": 0.936710524559021,
      "num_tokens": 86042465.0,
      "step": 5345
    },
    {
      "aux_loss": 8.07067289352417,
      "entropy": 0.18799100294709206,
      "epoch": 0.00535,
      "grad_norm": 0.13589370250701904,
      "learning_rate": 2e-05,
      "loss": 0.21719734668731688,
      "mean_token_accuracy": 0.9395054072141648,
      "num_tokens": 86123253.0,
      "step": 5350
    },
    {
      "aux_loss": 8.077297639846801,
      "entropy": 0.17483488768339156,
      "epoch": 0.005355,
      "grad_norm": 0.13774454593658447,
      "learning_rate": 2e-05,
      "loss": 0.197356116771698,
      "mean_token_accuracy": 0.9447283208370209,
      "num_tokens": 86204153.0,
      "step": 5355
    },
    {
      "aux_loss": 8.085671281814575,
      "entropy": 0.15701987519860267,
      "epoch": 0.00536,
      "grad_norm": 0.1307866871356964,
      "learning_rate": 2e-05,
      "loss": 0.1806653618812561,
      "mean_token_accuracy": 0.9495237529277801,
      "num_tokens": 86285077.0,
      "step": 5360
    },
    {
      "aux_loss": 8.090608692169189,
      "entropy": 0.14949651174247264,
      "epoch": 0.005365,
      "grad_norm": 0.13744370639324188,
      "learning_rate": 2e-05,
      "loss": 0.18164455890655518,
      "mean_token_accuracy": 0.9505838185548783,
      "num_tokens": 86362910.0,
      "step": 5365
    },
    {
      "aux_loss": 8.059525680541991,
      "entropy": 0.1972865153104067,
      "epoch": 0.00537,
      "grad_norm": 0.144741028547287,
      "learning_rate": 2e-05,
      "loss": 0.2484058618545532,
      "mean_token_accuracy": 0.9318668723106385,
      "num_tokens": 86444017.0,
      "step": 5370
    },
    {
      "aux_loss": 8.06361231803894,
      "entropy": 0.18635189421474935,
      "epoch": 0.005375,
      "grad_norm": 0.1357920914888382,
      "learning_rate": 2e-05,
      "loss": 0.22629961967468262,
      "mean_token_accuracy": 0.9371060013771058,
      "num_tokens": 86524770.0,
      "step": 5375
    },
    {
      "aux_loss": 8.07099962234497,
      "entropy": 0.18715228661894798,
      "epoch": 0.00538,
      "grad_norm": 0.1337023675441742,
      "learning_rate": 2e-05,
      "loss": 0.21162376403808594,
      "mean_token_accuracy": 0.9418139934539795,
      "num_tokens": 86605479.0,
      "step": 5380
    },
    {
      "aux_loss": 8.07848834991455,
      "entropy": 0.17354466021060944,
      "epoch": 0.005385,
      "grad_norm": 0.13858440518379211,
      "learning_rate": 2e-05,
      "loss": 0.19837032556533812,
      "mean_token_accuracy": 0.9447529882192611,
      "num_tokens": 86686227.0,
      "step": 5385
    },
    {
      "aux_loss": 8.087332487106323,
      "entropy": 0.16211861222982407,
      "epoch": 0.00539,
      "grad_norm": 0.14103250205516815,
      "learning_rate": 2e-05,
      "loss": 0.19104729890823363,
      "mean_token_accuracy": 0.9458187043666839,
      "num_tokens": 86766849.0,
      "step": 5390
    },
    {
      "aux_loss": 8.094975423812866,
      "entropy": 0.14273872263729573,
      "epoch": 0.005395,
      "grad_norm": 0.12239940464496613,
      "learning_rate": 2e-05,
      "loss": 0.16792721748352052,
      "mean_token_accuracy": 0.9539982408285141,
      "num_tokens": 86845433.0,
      "step": 5395
    },
    {
      "aux_loss": 8.056214094161987,
      "entropy": 0.1847220130264759,
      "epoch": 0.0054,
      "grad_norm": 0.1403694897890091,
      "learning_rate": 2e-05,
      "loss": 0.22664308547973633,
      "mean_token_accuracy": 0.9382877796888351,
      "num_tokens": 86926653.0,
      "step": 5400
    },
    {
      "aux_loss": 8.065641975402832,
      "entropy": 0.17645796462893487,
      "epoch": 0.005405,
      "grad_norm": 0.14281390607357025,
      "learning_rate": 2e-05,
      "loss": 0.21644015312194825,
      "mean_token_accuracy": 0.939283487200737,
      "num_tokens": 87007776.0,
      "step": 5405
    },
    {
      "aux_loss": 8.066926908493041,
      "entropy": 0.18893500715494155,
      "epoch": 0.00541,
      "grad_norm": 0.1328655183315277,
      "learning_rate": 2e-05,
      "loss": 0.22411150932312013,
      "mean_token_accuracy": 0.9372498899698257,
      "num_tokens": 87088607.0,
      "step": 5410
    },
    {
      "aux_loss": 8.073578310012817,
      "entropy": 0.18239122703671456,
      "epoch": 0.005415,
      "grad_norm": 0.1392800509929657,
      "learning_rate": 2e-05,
      "loss": 0.2100072145462036,
      "mean_token_accuracy": 0.9420152723789215,
      "num_tokens": 87169734.0,
      "step": 5415
    },
    {
      "aux_loss": 8.084091281890869,
      "entropy": 0.16939120516180992,
      "epoch": 0.00542,
      "grad_norm": 0.13912147283554077,
      "learning_rate": 2e-05,
      "loss": 0.18994629383087158,
      "mean_token_accuracy": 0.9470632135868072,
      "num_tokens": 87250032.0,
      "step": 5420
    },
    {
      "aux_loss": 8.100236654281616,
      "entropy": 0.14679208323359488,
      "epoch": 0.005425,
      "grad_norm": 0.1625506728887558,
      "learning_rate": 2e-05,
      "loss": 0.17082153558731078,
      "mean_token_accuracy": 0.9530383318662643,
      "num_tokens": 87330203.0,
      "step": 5425
    },
    {
      "aux_loss": 8.065205430984497,
      "entropy": 0.16032120436429978,
      "epoch": 0.00543,
      "grad_norm": 0.14751924574375153,
      "learning_rate": 2e-05,
      "loss": 0.2046891927719116,
      "mean_token_accuracy": 0.9443695902824402,
      "num_tokens": 87408106.0,
      "step": 5430
    },
    {
      "aux_loss": 8.061290168762207,
      "entropy": 0.18574356958270072,
      "epoch": 0.005435,
      "grad_norm": 0.15134435892105103,
      "learning_rate": 2e-05,
      "loss": 0.23141441345214844,
      "mean_token_accuracy": 0.9357331395149231,
      "num_tokens": 87489430.0,
      "step": 5435
    },
    {
      "aux_loss": 8.065538454055787,
      "entropy": 0.18175650313496589,
      "epoch": 0.00544,
      "grad_norm": 0.13342812657356262,
      "learning_rate": 2e-05,
      "loss": 0.21436078548431398,
      "mean_token_accuracy": 0.939489334821701,
      "num_tokens": 87570553.0,
      "step": 5440
    },
    {
      "aux_loss": 8.074884796142578,
      "entropy": 0.1787301439791918,
      "epoch": 0.005445,
      "grad_norm": 0.1371317058801651,
      "learning_rate": 2e-05,
      "loss": 0.20495855808258057,
      "mean_token_accuracy": 0.9420325130224227,
      "num_tokens": 87651757.0,
      "step": 5445
    },
    {
      "aux_loss": 8.08654351234436,
      "entropy": 0.1611557759344578,
      "epoch": 0.00545,
      "grad_norm": 0.14200739562511444,
      "learning_rate": 2e-05,
      "loss": 0.1845611810684204,
      "mean_token_accuracy": 0.9484672844409943,
      "num_tokens": 87731849.0,
      "step": 5450
    },
    {
      "aux_loss": 8.096723937988282,
      "entropy": 0.1464724324643612,
      "epoch": 0.005455,
      "grad_norm": 0.1508539766073227,
      "learning_rate": 2e-05,
      "loss": 0.1742676854133606,
      "mean_token_accuracy": 0.951809149980545,
      "num_tokens": 87812673.0,
      "step": 5455
    },
    {
      "aux_loss": 8.066505670547485,
      "entropy": 0.15454460829496383,
      "epoch": 0.00546,
      "grad_norm": 0.13381434977054596,
      "learning_rate": 2e-05,
      "loss": 0.19344613552093506,
      "mean_token_accuracy": 0.9480178236961365,
      "num_tokens": 87891413.0,
      "step": 5460
    },
    {
      "aux_loss": 8.062973499298096,
      "entropy": 0.17748164683580397,
      "epoch": 0.005465,
      "grad_norm": 0.1435805857181549,
      "learning_rate": 2e-05,
      "loss": 0.2242818832397461,
      "mean_token_accuracy": 0.9377706199884415,
      "num_tokens": 87972408.0,
      "step": 5465
    },
    {
      "aux_loss": 8.064704418182373,
      "entropy": 0.1739771731197834,
      "epoch": 0.00547,
      "grad_norm": 0.1375543177127838,
      "learning_rate": 2e-05,
      "loss": 0.2104339599609375,
      "mean_token_accuracy": 0.9420272171497345,
      "num_tokens": 88053615.0,
      "step": 5470
    },
    {
      "aux_loss": 8.07089900970459,
      "entropy": 0.16684116907417773,
      "epoch": 0.005475,
      "grad_norm": 0.1268276870250702,
      "learning_rate": 2e-05,
      "loss": 0.18953068256378175,
      "mean_token_accuracy": 0.9469234615564346,
      "num_tokens": 88134384.0,
      "step": 5475
    },
    {
      "aux_loss": 8.080683755874634,
      "entropy": 0.17217904850840568,
      "epoch": 0.00548,
      "grad_norm": 0.13503451645374298,
      "learning_rate": 2e-05,
      "loss": 0.19765374660491944,
      "mean_token_accuracy": 0.9450942516326905,
      "num_tokens": 88215085.0,
      "step": 5480
    },
    {
      "aux_loss": 8.092326545715332,
      "entropy": 0.15209664180874824,
      "epoch": 0.005485,
      "grad_norm": 0.1457759141921997,
      "learning_rate": 2e-05,
      "loss": 0.180171000957489,
      "mean_token_accuracy": 0.9496074229478836,
      "num_tokens": 88295455.0,
      "step": 5485
    },
    {
      "aux_loss": 8.071015930175781,
      "entropy": 0.16407919004559518,
      "epoch": 0.00549,
      "grad_norm": 0.1388535499572754,
      "learning_rate": 2e-05,
      "loss": 0.19918595552444457,
      "mean_token_accuracy": 0.9456561833620072,
      "num_tokens": 88373371.0,
      "step": 5490
    },
    {
      "aux_loss": 8.058560371398926,
      "entropy": 0.20277629159390925,
      "epoch": 0.005495,
      "grad_norm": 0.14603985846042633,
      "learning_rate": 2e-05,
      "loss": 0.2442614793777466,
      "mean_token_accuracy": 0.9321949660778046,
      "num_tokens": 88454172.0,
      "step": 5495
    },
    {
      "aux_loss": 8.060793733596801,
      "entropy": 0.17903067097067832,
      "epoch": 0.0055,
      "grad_norm": 0.14053769409656525,
      "learning_rate": 2e-05,
      "loss": 0.216259765625,
      "mean_token_accuracy": 0.9401715993881226,
      "num_tokens": 88535183.0,
      "step": 5500
    },
    {
      "aux_loss": 8.071125936508178,
      "entropy": 0.18102438598871232,
      "epoch": 0.005505,
      "grad_norm": 0.13056859374046326,
      "learning_rate": 2e-05,
      "loss": 0.2098301887512207,
      "mean_token_accuracy": 0.9413354009389877,
      "num_tokens": 88615999.0,
      "step": 5505
    },
    {
      "aux_loss": 8.078808498382568,
      "entropy": 0.1832273855805397,
      "epoch": 0.00551,
      "grad_norm": 0.14732852578163147,
      "learning_rate": 2e-05,
      "loss": 0.20819444656372071,
      "mean_token_accuracy": 0.9419064611196518,
      "num_tokens": 88696463.0,
      "step": 5510
    },
    {
      "aux_loss": 8.087247800827026,
      "entropy": 0.1631258063018322,
      "epoch": 0.005515,
      "grad_norm": 0.14680707454681396,
      "learning_rate": 2e-05,
      "loss": 0.1912326216697693,
      "mean_token_accuracy": 0.9472010612487793,
      "num_tokens": 88776951.0,
      "step": 5515
    },
    {
      "aux_loss": 8.078309249877929,
      "entropy": 0.15659815147519113,
      "epoch": 0.00552,
      "grad_norm": 0.15139469504356384,
      "learning_rate": 2e-05,
      "loss": 0.19072411060333253,
      "mean_token_accuracy": 0.9476257890462876,
      "num_tokens": 88856044.0,
      "step": 5520
    },
    {
      "aux_loss": 8.062022686004639,
      "entropy": 0.17389807254076003,
      "epoch": 0.005525,
      "grad_norm": 0.1451343446969986,
      "learning_rate": 2e-05,
      "loss": 0.220885705947876,
      "mean_token_accuracy": 0.9391647338867187,
      "num_tokens": 88937315.0,
      "step": 5525
    },
    {
      "aux_loss": 8.06127314567566,
      "entropy": 0.17330899611115455,
      "epoch": 0.00553,
      "grad_norm": 0.1304226815700531,
      "learning_rate": 2e-05,
      "loss": 0.21205427646636962,
      "mean_token_accuracy": 0.941293078660965,
      "num_tokens": 89018380.0,
      "step": 5530
    },
    {
      "aux_loss": 8.066385984420776,
      "entropy": 0.17221021950244902,
      "epoch": 0.005535,
      "grad_norm": 0.12877008318901062,
      "learning_rate": 2e-05,
      "loss": 0.2028797149658203,
      "mean_token_accuracy": 0.9435760766267777,
      "num_tokens": 89099215.0,
      "step": 5535
    },
    {
      "aux_loss": 8.074336624145507,
      "entropy": 0.17517672404646872,
      "epoch": 0.00554,
      "grad_norm": 0.13778316974639893,
      "learning_rate": 2e-05,
      "loss": 0.19934819936752318,
      "mean_token_accuracy": 0.9456212908029556,
      "num_tokens": 89180428.0,
      "step": 5540
    },
    {
      "aux_loss": 8.086762952804566,
      "entropy": 0.16597272530198098,
      "epoch": 0.005545,
      "grad_norm": 0.14513112604618073,
      "learning_rate": 2e-05,
      "loss": 0.18906669616699218,
      "mean_token_accuracy": 0.9467913210391998,
      "num_tokens": 89260792.0,
      "step": 5545
    },
    {
      "aux_loss": 8.084904623031616,
      "entropy": 0.15381703563034535,
      "epoch": 0.00555,
      "grad_norm": 0.13529625535011292,
      "learning_rate": 2e-05,
      "loss": 0.1874503254890442,
      "mean_token_accuracy": 0.948190101981163,
      "num_tokens": 89337942.0,
      "step": 5550
    },
    {
      "aux_loss": 8.063801765441895,
      "entropy": 0.20141154229640962,
      "epoch": 0.005555,
      "grad_norm": 0.14477373659610748,
      "learning_rate": 2e-05,
      "loss": 0.24707629680633544,
      "mean_token_accuracy": 0.9326157689094543,
      "num_tokens": 89418892.0,
      "step": 5555
    },
    {
      "aux_loss": 8.062500095367431,
      "entropy": 0.1710236206650734,
      "epoch": 0.00556,
      "grad_norm": 0.14716032147407532,
      "learning_rate": 2e-05,
      "loss": 0.21237616539001464,
      "mean_token_accuracy": 0.9414898663759231,
      "num_tokens": 89499950.0,
      "step": 5560
    },
    {
      "aux_loss": 8.071409225463867,
      "entropy": 0.18398527428507805,
      "epoch": 0.005565,
      "grad_norm": 0.12915635108947754,
      "learning_rate": 2e-05,
      "loss": 0.2138502597808838,
      "mean_token_accuracy": 0.9398818194866181,
      "num_tokens": 89580930.0,
      "step": 5565
    },
    {
      "aux_loss": 8.08089838027954,
      "entropy": 0.1699889160692692,
      "epoch": 0.00557,
      "grad_norm": 0.13302767276763916,
      "learning_rate": 2e-05,
      "loss": 0.19393314123153688,
      "mean_token_accuracy": 0.9458663403987885,
      "num_tokens": 89661544.0,
      "step": 5570
    },
    {
      "aux_loss": 8.087613248825074,
      "entropy": 0.15916759930551053,
      "epoch": 0.005575,
      "grad_norm": 0.15169700980186462,
      "learning_rate": 2e-05,
      "loss": 0.18777084350585938,
      "mean_token_accuracy": 0.9470723569393158,
      "num_tokens": 89741791.0,
      "step": 5575
    },
    {
      "aux_loss": 8.083750867843628,
      "entropy": 0.15333622619509696,
      "epoch": 0.00558,
      "grad_norm": 0.13448326289653778,
      "learning_rate": 2e-05,
      "loss": 0.18871968984603882,
      "mean_token_accuracy": 0.9483475059270858,
      "num_tokens": 89821499.0,
      "step": 5580
    },
    {
      "aux_loss": 8.062133312225342,
      "entropy": 0.18548016324639321,
      "epoch": 0.005585,
      "grad_norm": 0.1459362506866455,
      "learning_rate": 2e-05,
      "loss": 0.22562904357910157,
      "mean_token_accuracy": 0.9373085409402847,
      "num_tokens": 89902521.0,
      "step": 5585
    },
    {
      "aux_loss": 8.060822057724,
      "entropy": 0.1857463113963604,
      "epoch": 0.00559,
      "grad_norm": 0.1406942456960678,
      "learning_rate": 2e-05,
      "loss": 0.21963930130004883,
      "mean_token_accuracy": 0.9383396595716477,
      "num_tokens": 89982753.0,
      "step": 5590
    },
    {
      "aux_loss": 8.068700551986694,
      "entropy": 0.17866287454962732,
      "epoch": 0.005595,
      "grad_norm": 0.14183767139911652,
      "learning_rate": 2e-05,
      "loss": 0.2091233253479004,
      "mean_token_accuracy": 0.9427858710289001,
      "num_tokens": 90063839.0,
      "step": 5595
    },
    {
      "aux_loss": 8.073639488220214,
      "entropy": 0.1639278568327427,
      "epoch": 0.0056,
      "grad_norm": 0.13105636835098267,
      "learning_rate": 2e-05,
      "loss": 0.19065216779708863,
      "mean_token_accuracy": 0.9459737181663513,
      "num_tokens": 90144830.0,
      "step": 5600
    },
    {
      "aux_loss": 8.083618307113648,
      "entropy": 0.15334885641932489,
      "epoch": 0.005605,
      "grad_norm": 0.1375649869441986,
      "learning_rate": 2e-05,
      "loss": 0.17763084173202515,
      "mean_token_accuracy": 0.9501822859048843,
      "num_tokens": 90225731.0,
      "step": 5605
    },
    {
      "aux_loss": 8.09196228981018,
      "entropy": 0.14317901767790317,
      "epoch": 0.00561,
      "grad_norm": 0.12069131433963776,
      "learning_rate": 2e-05,
      "loss": 0.1715116024017334,
      "mean_token_accuracy": 0.952927190065384,
      "num_tokens": 90303347.0,
      "step": 5610
    },
    {
      "aux_loss": 8.061731386184693,
      "entropy": 0.17649685703217982,
      "epoch": 0.005615,
      "grad_norm": 0.1547907143831253,
      "learning_rate": 2e-05,
      "loss": 0.2208183526992798,
      "mean_token_accuracy": 0.9393579095602036,
      "num_tokens": 90384904.0,
      "step": 5615
    },
    {
      "aux_loss": 8.0631507396698,
      "entropy": 0.1872194517403841,
      "epoch": 0.00562,
      "grad_norm": 0.1439310759305954,
      "learning_rate": 2e-05,
      "loss": 0.23037052154541016,
      "mean_token_accuracy": 0.9356203705072403,
      "num_tokens": 90465586.0,
      "step": 5620
    },
    {
      "aux_loss": 8.067120122909547,
      "entropy": 0.18145886287093163,
      "epoch": 0.005625,
      "grad_norm": 0.1358509361743927,
      "learning_rate": 2e-05,
      "loss": 0.21520378589630126,
      "mean_token_accuracy": 0.9384305089712143,
      "num_tokens": 90546069.0,
      "step": 5625
    },
    {
      "aux_loss": 8.075404167175293,
      "entropy": 0.17724177688360215,
      "epoch": 0.00563,
      "grad_norm": 0.13912159204483032,
      "learning_rate": 2e-05,
      "loss": 0.19740923643112182,
      "mean_token_accuracy": 0.9450008779764175,
      "num_tokens": 90626910.0,
      "step": 5630
    },
    {
      "aux_loss": 8.084629440307618,
      "entropy": 0.1531679518520832,
      "epoch": 0.005635,
      "grad_norm": 0.13781334459781647,
      "learning_rate": 2e-05,
      "loss": 0.17628220319747925,
      "mean_token_accuracy": 0.9514947712421418,
      "num_tokens": 90707845.0,
      "step": 5635
    },
    {
      "aux_loss": 8.095353937149047,
      "entropy": 0.14077298939228058,
      "epoch": 0.00564,
      "grad_norm": 0.331421434879303,
      "learning_rate": 2e-05,
      "loss": 0.17239704132080078,
      "mean_token_accuracy": 0.9524759769439697,
      "num_tokens": 90786711.0,
      "step": 5640
    },
    {
      "aux_loss": 8.05805230140686,
      "entropy": 0.17784008271992208,
      "epoch": 0.005645,
      "grad_norm": 0.13844691216945648,
      "learning_rate": 2e-05,
      "loss": 0.22189898490905763,
      "mean_token_accuracy": 0.937601524591446,
      "num_tokens": 90868301.0,
      "step": 5645
    },
    {
      "aux_loss": 8.064334726333618,
      "entropy": 0.1882697395980358,
      "epoch": 0.00565,
      "grad_norm": 0.13702885806560516,
      "learning_rate": 2e-05,
      "loss": 0.23438878059387208,
      "mean_token_accuracy": 0.9356648206710816,
      "num_tokens": 90949176.0,
      "step": 5650
    },
    {
      "aux_loss": 8.067155647277833,
      "entropy": 0.18150743171572686,
      "epoch": 0.005655,
      "grad_norm": 0.13390353322029114,
      "learning_rate": 2e-05,
      "loss": 0.20990867614746095,
      "mean_token_accuracy": 0.9414841532707214,
      "num_tokens": 91029644.0,
      "step": 5655
    },
    {
      "aux_loss": 8.074916362762451,
      "entropy": 0.1758174940943718,
      "epoch": 0.00566,
      "grad_norm": 0.13451914489269257,
      "learning_rate": 2e-05,
      "loss": 0.2010176420211792,
      "mean_token_accuracy": 0.9433561027050018,
      "num_tokens": 91110401.0,
      "step": 5660
    },
    {
      "aux_loss": 8.082951211929322,
      "entropy": 0.16548511534929275,
      "epoch": 0.005665,
      "grad_norm": 0.14417754113674164,
      "learning_rate": 2e-05,
      "loss": 0.18805559873580932,
      "mean_token_accuracy": 0.9471803724765777,
      "num_tokens": 91191163.0,
      "step": 5665
    },
    {
      "aux_loss": 8.094138383865356,
      "entropy": 0.1513815179467201,
      "epoch": 0.00567,
      "grad_norm": 0.1468878984451294,
      "learning_rate": 2e-05,
      "loss": 0.18069925308227539,
      "mean_token_accuracy": 0.9499895632266998,
      "num_tokens": 91269061.0,
      "step": 5670
    },
    {
      "aux_loss": 8.056322288513183,
      "entropy": 0.17581670098006724,
      "epoch": 0.005675,
      "grad_norm": 0.14119933545589447,
      "learning_rate": 2e-05,
      "loss": 0.2202775478363037,
      "mean_token_accuracy": 0.93985455930233,
      "num_tokens": 91350350.0,
      "step": 5675
    },
    {
      "aux_loss": 8.059354782104492,
      "entropy": 0.18535528592765332,
      "epoch": 0.00568,
      "grad_norm": 0.13150428235530853,
      "learning_rate": 2e-05,
      "loss": 0.22885429859161377,
      "mean_token_accuracy": 0.9363972008228302,
      "num_tokens": 91431249.0,
      "step": 5680
    },
    {
      "aux_loss": 8.068913793563842,
      "entropy": 0.18798594176769257,
      "epoch": 0.005685,
      "grad_norm": 0.13334661722183228,
      "learning_rate": 2e-05,
      "loss": 0.2183297395706177,
      "mean_token_accuracy": 0.9391392350196839,
      "num_tokens": 91511768.0,
      "step": 5685
    },
    {
      "aux_loss": 8.073178625106811,
      "entropy": 0.17933037132024765,
      "epoch": 0.00569,
      "grad_norm": 0.1362065076828003,
      "learning_rate": 2e-05,
      "loss": 0.2034616470336914,
      "mean_token_accuracy": 0.9438274651765823,
      "num_tokens": 91592991.0,
      "step": 5690
    },
    {
      "aux_loss": 8.080089426040649,
      "entropy": 0.1637831013649702,
      "epoch": 0.005695,
      "grad_norm": 0.14167571067810059,
      "learning_rate": 2e-05,
      "loss": 0.1888332724571228,
      "mean_token_accuracy": 0.9474999457597733,
      "num_tokens": 91673276.0,
      "step": 5695
    },
    {
      "aux_loss": 8.09348554611206,
      "entropy": 0.14930419959127902,
      "epoch": 0.0057,
      "grad_norm": 0.1610335260629654,
      "learning_rate": 2e-05,
      "loss": 0.18102054595947265,
      "mean_token_accuracy": 0.9492920756340026,
      "num_tokens": 91753593.0,
      "step": 5700
    },
    {
      "aux_loss": 8.054687738418579,
      "entropy": 0.17351944260299207,
      "epoch": 0.005705,
      "grad_norm": 0.15405237674713135,
      "learning_rate": 2e-05,
      "loss": 0.2203193187713623,
      "mean_token_accuracy": 0.9391676157712936,
      "num_tokens": 91835086.0,
      "step": 5705
    },
    {
      "aux_loss": 8.061115980148315,
      "entropy": 0.18555759713053704,
      "epoch": 0.00571,
      "grad_norm": 0.1384715735912323,
      "learning_rate": 2e-05,
      "loss": 0.22710886001586914,
      "mean_token_accuracy": 0.9366005331277847,
      "num_tokens": 91916369.0,
      "step": 5710
    },
    {
      "aux_loss": 8.066951417922974,
      "entropy": 0.19230927899479866,
      "epoch": 0.005715,
      "grad_norm": 0.12675443291664124,
      "learning_rate": 2e-05,
      "loss": 0.21959903240203857,
      "mean_token_accuracy": 0.9389223873615264,
      "num_tokens": 91997062.0,
      "step": 5715
    },
    {
      "aux_loss": 8.07173776626587,
      "entropy": 0.1709363006055355,
      "epoch": 0.00572,
      "grad_norm": 0.13290657103061676,
      "learning_rate": 2e-05,
      "loss": 0.19522292613983155,
      "mean_token_accuracy": 0.9457512766122818,
      "num_tokens": 92078160.0,
      "step": 5720
    },
    {
      "aux_loss": 8.081733465194702,
      "entropy": 0.15775464847683907,
      "epoch": 0.005725,
      "grad_norm": 0.13461284339427948,
      "learning_rate": 2e-05,
      "loss": 0.18419874906539918,
      "mean_token_accuracy": 0.9491376131772995,
      "num_tokens": 92158400.0,
      "step": 5725
    },
    {
      "aux_loss": 8.09744577407837,
      "entropy": 0.15009067244827748,
      "epoch": 0.00573,
      "grad_norm": 0.1676807850599289,
      "learning_rate": 2e-05,
      "loss": 0.1772789478302002,
      "mean_token_accuracy": 0.9510623186826705,
      "num_tokens": 92239268.0,
      "step": 5730
    },
    {
      "aux_loss": 8.061673212051392,
      "entropy": 0.15475926585495473,
      "epoch": 0.005735,
      "grad_norm": 0.1452476531267166,
      "learning_rate": 2e-05,
      "loss": 0.193684720993042,
      "mean_token_accuracy": 0.946747025847435,
      "num_tokens": 92320198.0,
      "step": 5735
    },
    {
      "aux_loss": 8.060331678390503,
      "entropy": 0.17048021629452706,
      "epoch": 0.00574,
      "grad_norm": 0.1505119800567627,
      "learning_rate": 2e-05,
      "loss": 0.21258339881896973,
      "mean_token_accuracy": 0.942168366909027,
      "num_tokens": 92401220.0,
      "step": 5740
    },
    {
      "aux_loss": 8.064356708526612,
      "entropy": 0.17428167089819907,
      "epoch": 0.005745,
      "grad_norm": 0.13418808579444885,
      "learning_rate": 2e-05,
      "loss": 0.21015555858612062,
      "mean_token_accuracy": 0.9408140420913697,
      "num_tokens": 92482191.0,
      "step": 5745
    },
    {
      "aux_loss": 8.073054027557372,
      "entropy": 0.1741316296160221,
      "epoch": 0.00575,
      "grad_norm": 0.1389142870903015,
      "learning_rate": 2e-05,
      "loss": 0.20202784538269042,
      "mean_token_accuracy": 0.9437748968601227,
      "num_tokens": 92562836.0,
      "step": 5750
    },
    {
      "aux_loss": 8.080522537231445,
      "entropy": 0.1635540872812271,
      "epoch": 0.005755,
      "grad_norm": 0.14265964925289154,
      "learning_rate": 2e-05,
      "loss": 0.18728913068771363,
      "mean_token_accuracy": 0.9472408413887023,
      "num_tokens": 92643436.0,
      "step": 5755
    },
    {
      "aux_loss": 8.092672872543336,
      "entropy": 0.14982328303158282,
      "epoch": 0.00576,
      "grad_norm": 0.14974676072597504,
      "learning_rate": 2e-05,
      "loss": 0.173487389087677,
      "mean_token_accuracy": 0.9515981703996659,
      "num_tokens": 92723306.0,
      "step": 5760
    },
    {
      "aux_loss": 8.063954257965088,
      "entropy": 0.16390583664178848,
      "epoch": 0.005765,
      "grad_norm": 0.1341383159160614,
      "learning_rate": 2e-05,
      "loss": 0.2032714605331421,
      "mean_token_accuracy": 0.9443684726953506,
      "num_tokens": 92802842.0,
      "step": 5765
    },
    {
      "aux_loss": 8.06337742805481,
      "entropy": 0.1663944326341152,
      "epoch": 0.00577,
      "grad_norm": 0.14594364166259766,
      "learning_rate": 2e-05,
      "loss": 0.20972006320953368,
      "mean_token_accuracy": 0.9422449350357056,
      "num_tokens": 92883987.0,
      "step": 5770
    },
    {
      "aux_loss": 8.06315712928772,
      "entropy": 0.17991503998637198,
      "epoch": 0.005775,
      "grad_norm": 0.15171870589256287,
      "learning_rate": 2e-05,
      "loss": 0.21673328876495362,
      "mean_token_accuracy": 0.9400419503450393,
      "num_tokens": 92965089.0,
      "step": 5775
    },
    {
      "aux_loss": 8.068653440475464,
      "entropy": 0.1760217808187008,
      "epoch": 0.00578,
      "grad_norm": 0.13597488403320312,
      "learning_rate": 2e-05,
      "loss": 0.20551862716674804,
      "mean_token_accuracy": 0.9422009438276291,
      "num_tokens": 93045989.0,
      "step": 5780
    },
    {
      "aux_loss": 8.07674880027771,
      "entropy": 0.1621110685169697,
      "epoch": 0.005785,
      "grad_norm": 0.13650666177272797,
      "learning_rate": 2e-05,
      "loss": 0.19024025201797484,
      "mean_token_accuracy": 0.9472737580537796,
      "num_tokens": 93127019.0,
      "step": 5785
    },
    {
      "aux_loss": 8.089992189407349,
      "entropy": 0.15284370705485345,
      "epoch": 0.00579,
      "grad_norm": 0.14131860435009003,
      "learning_rate": 2e-05,
      "loss": 0.17998089790344238,
      "mean_token_accuracy": 0.9499067604541779,
      "num_tokens": 93207117.0,
      "step": 5790
    },
    {
      "aux_loss": 8.072627401351928,
      "entropy": 0.16572054252028465,
      "epoch": 0.005795,
      "grad_norm": 0.13508719205856323,
      "learning_rate": 2e-05,
      "loss": 0.197525954246521,
      "mean_token_accuracy": 0.9458871275186539,
      "num_tokens": 93287901.0,
      "step": 5795
    },
    {
      "aux_loss": 8.063549137115478,
      "entropy": 0.16287589445710182,
      "epoch": 0.0058,
      "grad_norm": 0.1373695582151413,
      "learning_rate": 2e-05,
      "loss": 0.204803991317749,
      "mean_token_accuracy": 0.9441636741161347,
      "num_tokens": 93368746.0,
      "step": 5800
    },
    {
      "aux_loss": 8.06470890045166,
      "entropy": 0.17902035936713218,
      "epoch": 0.005805,
      "grad_norm": 0.14767873287200928,
      "learning_rate": 2e-05,
      "loss": 0.22090959548950195,
      "mean_token_accuracy": 0.9376872271299362,
      "num_tokens": 93449568.0,
      "step": 5805
    },
    {
      "aux_loss": 8.069039487838745,
      "entropy": 0.1771714173257351,
      "epoch": 0.00581,
      "grad_norm": 0.13491499423980713,
      "learning_rate": 2e-05,
      "loss": 0.20453858375549316,
      "mean_token_accuracy": 0.9425706744194031,
      "num_tokens": 93530412.0,
      "step": 5810
    },
    {
      "aux_loss": 8.078837060928345,
      "entropy": 0.17171581760048865,
      "epoch": 0.005815,
      "grad_norm": 0.14088688790798187,
      "learning_rate": 2e-05,
      "loss": 0.19795360565185546,
      "mean_token_accuracy": 0.9439451515674591,
      "num_tokens": 93611214.0,
      "step": 5815
    },
    {
      "aux_loss": 8.086272764205933,
      "entropy": 0.15185713097453118,
      "epoch": 0.00582,
      "grad_norm": 0.14482736587524414,
      "learning_rate": 2e-05,
      "loss": 0.1787477970123291,
      "mean_token_accuracy": 0.9499908089637756,
      "num_tokens": 93691899.0,
      "step": 5820
    },
    {
      "aux_loss": 8.085255575180053,
      "entropy": 0.1480364143848419,
      "epoch": 0.005825,
      "grad_norm": 0.1324480175971985,
      "learning_rate": 2e-05,
      "loss": 0.18043410778045654,
      "mean_token_accuracy": 0.9507531315088272,
      "num_tokens": 93770107.0,
      "step": 5825
    },
    {
      "aux_loss": 8.055831813812256,
      "entropy": 0.183095084130764,
      "epoch": 0.00583,
      "grad_norm": 0.15179389715194702,
      "learning_rate": 2e-05,
      "loss": 0.2280038833618164,
      "mean_token_accuracy": 0.9371652454137802,
      "num_tokens": 93851643.0,
      "step": 5830
    },
    {
      "aux_loss": 8.063481044769286,
      "entropy": 0.18543468043208122,
      "epoch": 0.005835,
      "grad_norm": 0.12911377847194672,
      "learning_rate": 2e-05,
      "loss": 0.21943447589874268,
      "mean_token_accuracy": 0.938288950920105,
      "num_tokens": 93932478.0,
      "step": 5835
    },
    {
      "aux_loss": 8.065958642959595,
      "entropy": 0.17744500562548637,
      "epoch": 0.00584,
      "grad_norm": 0.1379033476114273,
      "learning_rate": 2e-05,
      "loss": 0.2041388988494873,
      "mean_token_accuracy": 0.9421236902475357,
      "num_tokens": 94013130.0,
      "step": 5840
    },
    {
      "aux_loss": 8.072461032867432,
      "entropy": 0.1693294659256935,
      "epoch": 0.005845,
      "grad_norm": 0.14475242793560028,
      "learning_rate": 2e-05,
      "loss": 0.19758639335632325,
      "mean_token_accuracy": 0.9444235742092133,
      "num_tokens": 94093276.0,
      "step": 5845
    },
    {
      "aux_loss": 8.082824802398681,
      "entropy": 0.15156857818365096,
      "epoch": 0.00585,
      "grad_norm": 0.14044874906539917,
      "learning_rate": 2e-05,
      "loss": 0.17861695289611818,
      "mean_token_accuracy": 0.9495580226182938,
      "num_tokens": 94173881.0,
      "step": 5850
    },
    {
      "aux_loss": 8.091250610351562,
      "entropy": 0.14027350395917892,
      "epoch": 0.005855,
      "grad_norm": 0.11501962691545486,
      "learning_rate": 2e-05,
      "loss": 0.1679123282432556,
      "mean_token_accuracy": 0.9536539733409881,
      "num_tokens": 94252994.0,
      "step": 5855
    },
    {
      "aux_loss": 8.055892133712769,
      "entropy": 0.17486198730766772,
      "epoch": 0.00586,
      "grad_norm": 0.14729391038417816,
      "learning_rate": 2e-05,
      "loss": 0.21825242042541504,
      "mean_token_accuracy": 0.938835597038269,
      "num_tokens": 94334247.0,
      "step": 5860
    },
    {
      "aux_loss": 8.06143307685852,
      "entropy": 0.1837987508624792,
      "epoch": 0.005865,
      "grad_norm": 0.12966404855251312,
      "learning_rate": 2e-05,
      "loss": 0.22734589576721193,
      "mean_token_accuracy": 0.9369234770536423,
      "num_tokens": 94415628.0,
      "step": 5865
    },
    {
      "aux_loss": 8.063664722442628,
      "entropy": 0.1713401522487402,
      "epoch": 0.00587,
      "grad_norm": 0.13503432273864746,
      "learning_rate": 2e-05,
      "loss": 0.20385169982910156,
      "mean_token_accuracy": 0.9436463266611099,
      "num_tokens": 94496544.0,
      "step": 5870
    },
    {
      "aux_loss": 8.070697546005249,
      "entropy": 0.1775601290166378,
      "epoch": 0.005875,
      "grad_norm": 0.12884454429149628,
      "learning_rate": 2e-05,
      "loss": 0.20616636276245118,
      "mean_token_accuracy": 0.9421791166067124,
      "num_tokens": 94577478.0,
      "step": 5875
    },
    {
      "aux_loss": 8.081576299667358,
      "entropy": 0.1703704960644245,
      "epoch": 0.00588,
      "grad_norm": 0.14574073255062103,
      "learning_rate": 2e-05,
      "loss": 0.19270837306976318,
      "mean_token_accuracy": 0.9460537075996399,
      "num_tokens": 94657487.0,
      "step": 5880
    },
    {
      "aux_loss": 8.097192907333374,
      "entropy": 0.1473420437425375,
      "epoch": 0.005885,
      "grad_norm": 0.17822988331317902,
      "learning_rate": 2e-05,
      "loss": 0.17440905570983886,
      "mean_token_accuracy": 0.9511974841356278,
      "num_tokens": 94735178.0,
      "step": 5885
    },
    {
      "aux_loss": 8.058767366409302,
      "entropy": 0.1708623431622982,
      "epoch": 0.00589,
      "grad_norm": 0.13845281302928925,
      "learning_rate": 2e-05,
      "loss": 0.21285016536712648,
      "mean_token_accuracy": 0.9420379728078843,
      "num_tokens": 94816378.0,
      "step": 5890
    },
    {
      "aux_loss": 8.061467742919922,
      "entropy": 0.18202443197369575,
      "epoch": 0.005895,
      "grad_norm": 0.13410122692584991,
      "learning_rate": 2e-05,
      "loss": 0.21937117576599122,
      "mean_token_accuracy": 0.9389735221862793,
      "num_tokens": 94897959.0,
      "step": 5895
    },
    {
      "aux_loss": 8.063257217407227,
      "entropy": 0.17064057104289532,
      "epoch": 0.0059,
      "grad_norm": 0.1352560818195343,
      "learning_rate": 2e-05,
      "loss": 0.20185585021972657,
      "mean_token_accuracy": 0.9441125452518463,
      "num_tokens": 94979042.0,
      "step": 5900
    },
    {
      "aux_loss": 8.072481775283814,
      "entropy": 0.16953162774443625,
      "epoch": 0.005905,
      "grad_norm": 0.13705940544605255,
      "learning_rate": 2e-05,
      "loss": 0.19920376539230347,
      "mean_token_accuracy": 0.9433676153421402,
      "num_tokens": 95059868.0,
      "step": 5905
    },
    {
      "aux_loss": 8.078711700439452,
      "entropy": 0.15986860617995263,
      "epoch": 0.00591,
      "grad_norm": 0.14904817938804626,
      "learning_rate": 2e-05,
      "loss": 0.19004034996032715,
      "mean_token_accuracy": 0.9468602746725082,
      "num_tokens": 95140070.0,
      "step": 5910
    },
    {
      "aux_loss": 8.092806434631347,
      "entropy": 0.1493261508643627,
      "epoch": 0.005915,
      "grad_norm": 0.14852458238601685,
      "learning_rate": 2e-05,
      "loss": 0.17741813659667968,
      "mean_token_accuracy": 0.9513360589742661,
      "num_tokens": 95220308.0,
      "step": 5915
    },
    {
      "aux_loss": 8.065175008773803,
      "entropy": 0.17146045416593553,
      "epoch": 0.00592,
      "grad_norm": 0.1456707864999771,
      "learning_rate": 2e-05,
      "loss": 0.2067434310913086,
      "mean_token_accuracy": 0.9436594158411026,
      "num_tokens": 95299270.0,
      "step": 5920
    },
    {
      "aux_loss": 8.06174168586731,
      "entropy": 0.1672132160514593,
      "epoch": 0.005925,
      "grad_norm": 0.1400015652179718,
      "learning_rate": 2e-05,
      "loss": 0.20917375087738038,
      "mean_token_accuracy": 0.9424566030502319,
      "num_tokens": 95380537.0,
      "step": 5925
    },
    {
      "aux_loss": 8.066592931747437,
      "entropy": 0.17035450227558613,
      "epoch": 0.00593,
      "grad_norm": 0.13348719477653503,
      "learning_rate": 2e-05,
      "loss": 0.20101003646850585,
      "mean_token_accuracy": 0.9435734242200852,
      "num_tokens": 95461363.0,
      "step": 5930
    },
    {
      "aux_loss": 8.072800540924073,
      "entropy": 0.17302406653761865,
      "epoch": 0.005935,
      "grad_norm": 0.1358296424150467,
      "learning_rate": 2e-05,
      "loss": 0.2002587080001831,
      "mean_token_accuracy": 0.9450927078723907,
      "num_tokens": 95542166.0,
      "step": 5935
    },
    {
      "aux_loss": 8.080530643463135,
      "entropy": 0.1578762486577034,
      "epoch": 0.00594,
      "grad_norm": 0.13561034202575684,
      "learning_rate": 2e-05,
      "loss": 0.18470389842987062,
      "mean_token_accuracy": 0.9497940570116044,
      "num_tokens": 95622653.0,
      "step": 5940
    },
    {
      "aux_loss": 8.091456985473632,
      "entropy": 0.15263076946139337,
      "epoch": 0.005945,
      "grad_norm": 0.1546875685453415,
      "learning_rate": 2e-05,
      "loss": 0.18000965118408202,
      "mean_token_accuracy": 0.9487313568592072,
      "num_tokens": 95703334.0,
      "step": 5945
    },
    {
      "aux_loss": 8.067452478408814,
      "entropy": 0.16000871136784553,
      "epoch": 0.00595,
      "grad_norm": 0.14300614595413208,
      "learning_rate": 2e-05,
      "loss": 0.1960815668106079,
      "mean_token_accuracy": 0.9453659981489182,
      "num_tokens": 95780602.0,
      "step": 5950
    },
    {
      "aux_loss": 8.062563467025758,
      "entropy": 0.18366865627467632,
      "epoch": 0.005955,
      "grad_norm": 0.1349765509366989,
      "learning_rate": 2e-05,
      "loss": 0.224930477142334,
      "mean_token_accuracy": 0.9381143927574158,
      "num_tokens": 95861860.0,
      "step": 5955
    },
    {
      "aux_loss": 8.061886882781982,
      "entropy": 0.17511266022920607,
      "epoch": 0.00596,
      "grad_norm": 0.12589815258979797,
      "learning_rate": 2e-05,
      "loss": 0.20954318046569825,
      "mean_token_accuracy": 0.9415994316339493,
      "num_tokens": 95943070.0,
      "step": 5960
    },
    {
      "aux_loss": 8.068562984466553,
      "entropy": 0.16178049817681311,
      "epoch": 0.005965,
      "grad_norm": 0.12789005041122437,
      "learning_rate": 2e-05,
      "loss": 0.1920637607574463,
      "mean_token_accuracy": 0.9466611444950104,
      "num_tokens": 96024218.0,
      "step": 5965
    },
    {
      "aux_loss": 8.07816834449768,
      "entropy": 0.15705137699842453,
      "epoch": 0.00597,
      "grad_norm": 0.13880205154418945,
      "learning_rate": 2e-05,
      "loss": 0.1792244791984558,
      "mean_token_accuracy": 0.9492392539978027,
      "num_tokens": 96104912.0,
      "step": 5970
    },
    {
      "aux_loss": 8.090532302856445,
      "entropy": 0.1515044815838337,
      "epoch": 0.005975,
      "grad_norm": 0.14545196294784546,
      "learning_rate": 2e-05,
      "loss": 0.17885712385177613,
      "mean_token_accuracy": 0.9505871325731278,
      "num_tokens": 96185410.0,
      "step": 5975
    },
    {
      "aux_loss": 8.072143363952637,
      "entropy": 0.14746857099235058,
      "epoch": 0.00598,
      "grad_norm": 0.12406041473150253,
      "learning_rate": 2e-05,
      "loss": 0.18974361419677735,
      "mean_token_accuracy": 0.9486129403114318,
      "num_tokens": 96262450.0,
      "step": 5980
    },
    {
      "aux_loss": 8.053602933883667,
      "entropy": 0.18247710466384887,
      "epoch": 0.005985,
      "grad_norm": 0.12539565563201904,
      "learning_rate": 2e-05,
      "loss": 0.22521460056304932,
      "mean_token_accuracy": 0.9376156717538834,
      "num_tokens": 96342965.0,
      "step": 5985
    },
    {
      "aux_loss": 8.061638069152831,
      "entropy": 0.1799737609922886,
      "epoch": 0.00599,
      "grad_norm": 0.13107004761695862,
      "learning_rate": 2e-05,
      "loss": 0.2145768880844116,
      "mean_token_accuracy": 0.9398253291845322,
      "num_tokens": 96424174.0,
      "step": 5990
    },
    {
      "aux_loss": 8.06684627532959,
      "entropy": 0.17635875716805458,
      "epoch": 0.005995,
      "grad_norm": 0.13319900631904602,
      "learning_rate": 2e-05,
      "loss": 0.20265119075775145,
      "mean_token_accuracy": 0.9436583459377289,
      "num_tokens": 96505120.0,
      "step": 5995
    },
    {
      "aux_loss": 8.07758674621582,
      "entropy": 0.17258130609989167,
      "epoch": 0.006,
      "grad_norm": 0.12928390502929688,
      "learning_rate": 2e-05,
      "loss": 0.19766595363616943,
      "mean_token_accuracy": 0.9446023225784301,
      "num_tokens": 96586037.0,
      "step": 6000
    },
    {
      "aux_loss": 8.08571252822876,
      "entropy": 0.15081175677478315,
      "epoch": 0.006005,
      "grad_norm": 0.1503567397594452,
      "learning_rate": 2e-05,
      "loss": 0.17959368228912354,
      "mean_token_accuracy": 0.9497321128845215,
      "num_tokens": 96666252.0,
      "step": 6005
    },
    {
      "aux_loss": 8.07593240737915,
      "entropy": 0.14692696668207644,
      "epoch": 0.00601,
      "grad_norm": 0.1549401730298996,
      "learning_rate": 2e-05,
      "loss": 0.18516803979873658,
      "mean_token_accuracy": 0.9497349351644516,
      "num_tokens": 96746963.0,
      "step": 6010
    },
    {
      "aux_loss": 8.060930681228637,
      "entropy": 0.1806217808276415,
      "epoch": 0.006015,
      "grad_norm": 0.13608407974243164,
      "learning_rate": 2e-05,
      "loss": 0.2211301803588867,
      "mean_token_accuracy": 0.937987893819809,
      "num_tokens": 96827662.0,
      "step": 6015
    },
    {
      "aux_loss": 8.063177013397217,
      "entropy": 0.17464426308870315,
      "epoch": 0.00602,
      "grad_norm": 0.13371874392032623,
      "learning_rate": 2e-05,
      "loss": 0.20847699642181397,
      "mean_token_accuracy": 0.9415706694126129,
      "num_tokens": 96908701.0,
      "step": 6020
    },
    {
      "aux_loss": 8.068975162506103,
      "entropy": 0.1864542119204998,
      "epoch": 0.006025,
      "grad_norm": 0.1279734969139099,
      "learning_rate": 2e-05,
      "loss": 0.2206197500228882,
      "mean_token_accuracy": 0.9386757463216782,
      "num_tokens": 96989269.0,
      "step": 6025
    },
    {
      "aux_loss": 8.07884931564331,
      "entropy": 0.17818933054804803,
      "epoch": 0.00603,
      "grad_norm": 0.13667216897010803,
      "learning_rate": 2e-05,
      "loss": 0.20026471614837646,
      "mean_token_accuracy": 0.9444550067186356,
      "num_tokens": 97069757.0,
      "step": 6030
    },
    {
      "aux_loss": 8.084789896011353,
      "entropy": 0.15738881714642047,
      "epoch": 0.006035,
      "grad_norm": 0.13657960295677185,
      "learning_rate": 2e-05,
      "loss": 0.1845165014266968,
      "mean_token_accuracy": 0.9477501511573792,
      "num_tokens": 97149843.0,
      "step": 6035
    },
    {
      "aux_loss": 8.077795934677123,
      "entropy": 0.1544258899986744,
      "epoch": 0.00604,
      "grad_norm": 0.14422611892223358,
      "learning_rate": 2e-05,
      "loss": 0.18736767768859863,
      "mean_token_accuracy": 0.9476096302270889,
      "num_tokens": 97227475.0,
      "step": 6040
    },
    {
      "aux_loss": 8.064987087249756,
      "entropy": 0.16860076040029526,
      "epoch": 0.006045,
      "grad_norm": 0.15342532098293304,
      "learning_rate": 2e-05,
      "loss": 0.2130617618560791,
      "mean_token_accuracy": 0.9399748772382737,
      "num_tokens": 97308306.0,
      "step": 6045
    },
    {
      "aux_loss": 8.060701751708985,
      "entropy": 0.1847734682261944,
      "epoch": 0.00605,
      "grad_norm": 0.1469544619321823,
      "learning_rate": 2e-05,
      "loss": 0.21930246353149413,
      "mean_token_accuracy": 0.9377886474132537,
      "num_tokens": 97389210.0,
      "step": 6050
    },
    {
      "aux_loss": 8.066556262969971,
      "entropy": 0.1651120275259018,
      "epoch": 0.006055,
      "grad_norm": 0.14201335608959198,
      "learning_rate": 2e-05,
      "loss": 0.19273279905319213,
      "mean_token_accuracy": 0.9464361488819122,
      "num_tokens": 97470207.0,
      "step": 6055
    },
    {
      "aux_loss": 8.075764894485474,
      "entropy": 0.1742984712123871,
      "epoch": 0.00606,
      "grad_norm": 0.14166109263896942,
      "learning_rate": 2e-05,
      "loss": 0.20348267555236815,
      "mean_token_accuracy": 0.9441876828670501,
      "num_tokens": 97550879.0,
      "step": 6060
    },
    {
      "aux_loss": 8.080997848510743,
      "entropy": 0.1588553436100483,
      "epoch": 0.006065,
      "grad_norm": 0.13352887332439423,
      "learning_rate": 2e-05,
      "loss": 0.18576077222824097,
      "mean_token_accuracy": 0.9482071310281753,
      "num_tokens": 97631362.0,
      "step": 6065
    },
    {
      "aux_loss": 8.08425498008728,
      "entropy": 0.14988943040370942,
      "epoch": 0.00607,
      "grad_norm": 0.132766455411911,
      "learning_rate": 2e-05,
      "loss": 0.1825609564781189,
      "mean_token_accuracy": 0.9494982123374939,
      "num_tokens": 97709729.0,
      "step": 6070
    },
    {
      "aux_loss": 8.055536222457885,
      "entropy": 0.1706943243741989,
      "epoch": 0.006075,
      "grad_norm": 0.14236465096473694,
      "learning_rate": 2e-05,
      "loss": 0.21214184761047364,
      "mean_token_accuracy": 0.9419941782951355,
      "num_tokens": 97791098.0,
      "step": 6075
    },
    {
      "aux_loss": 8.058393096923828,
      "entropy": 0.1728857483714819,
      "epoch": 0.00608,
      "grad_norm": 0.12840798497200012,
      "learning_rate": 2e-05,
      "loss": 0.2110762357711792,
      "mean_token_accuracy": 0.9423777967691421,
      "num_tokens": 97871896.0,
      "step": 6080
    },
    {
      "aux_loss": 8.065342998504638,
      "entropy": 0.17374960109591484,
      "epoch": 0.006085,
      "grad_norm": 0.13968020677566528,
      "learning_rate": 2e-05,
      "loss": 0.20638532638549806,
      "mean_token_accuracy": 0.9422299891710282,
      "num_tokens": 97952649.0,
      "step": 6085
    },
    {
      "aux_loss": 8.07047872543335,
      "entropy": 0.17087947130203246,
      "epoch": 0.00609,
      "grad_norm": 0.13061648607254028,
      "learning_rate": 2e-05,
      "loss": 0.19820499420166016,
      "mean_token_accuracy": 0.9450456500053406,
      "num_tokens": 98033617.0,
      "step": 6090
    },
    {
      "aux_loss": 8.081618547439575,
      "entropy": 0.16440374180674552,
      "epoch": 0.006095,
      "grad_norm": 0.13969780504703522,
      "learning_rate": 2e-05,
      "loss": 0.1900782108306885,
      "mean_token_accuracy": 0.9466186791658402,
      "num_tokens": 98113780.0,
      "step": 6095
    },
    {
      "aux_loss": 8.088514423370361,
      "entropy": 0.14573372192680836,
      "epoch": 0.0061,
      "grad_norm": 0.13671861588954926,
      "learning_rate": 2e-05,
      "loss": 0.17280199527740478,
      "mean_token_accuracy": 0.9539888024330139,
      "num_tokens": 98190540.0,
      "step": 6100
    },
    {
      "aux_loss": 8.058065843582153,
      "entropy": 0.17734894305467605,
      "epoch": 0.006105,
      "grad_norm": 0.14421439170837402,
      "learning_rate": 2e-05,
      "loss": 0.22198522090911865,
      "mean_token_accuracy": 0.9395839691162109,
      "num_tokens": 98271598.0,
      "step": 6105
    },
    {
      "aux_loss": 8.06196870803833,
      "entropy": 0.17928044497966766,
      "epoch": 0.00611,
      "grad_norm": 0.13943056762218475,
      "learning_rate": 2e-05,
      "loss": 0.22034146785736083,
      "mean_token_accuracy": 0.9395325124263764,
      "num_tokens": 98352898.0,
      "step": 6110
    },
    {
      "aux_loss": 8.065202140808106,
      "entropy": 0.17982555776834488,
      "epoch": 0.006115,
      "grad_norm": 0.14371375739574432,
      "learning_rate": 2e-05,
      "loss": 0.21403896808624268,
      "mean_token_accuracy": 0.9406698137521744,
      "num_tokens": 98433784.0,
      "step": 6115
    },
    {
      "aux_loss": 8.067698335647583,
      "entropy": 0.173493380099535,
      "epoch": 0.00612,
      "grad_norm": 0.14029116928577423,
      "learning_rate": 2e-05,
      "loss": 0.19781092405319214,
      "mean_token_accuracy": 0.9440387845039367,
      "num_tokens": 98514535.0,
      "step": 6120
    },
    {
      "aux_loss": 8.077132511138917,
      "entropy": 0.16261643767356873,
      "epoch": 0.006125,
      "grad_norm": 0.1420287787914276,
      "learning_rate": 2e-05,
      "loss": 0.18982197046279908,
      "mean_token_accuracy": 0.947833177447319,
      "num_tokens": 98594371.0,
      "step": 6125
    },
    {
      "aux_loss": 8.093416452407837,
      "entropy": 0.15497832894325256,
      "epoch": 0.00613,
      "grad_norm": 0.16165927052497864,
      "learning_rate": 2e-05,
      "loss": 0.17958693504333495,
      "mean_token_accuracy": 0.9492352396249771,
      "num_tokens": 98673792.0,
      "step": 6130
    },
    {
      "aux_loss": 8.055329036712646,
      "entropy": 0.1695503953844309,
      "epoch": 0.006135,
      "grad_norm": 0.1473519653081894,
      "learning_rate": 2e-05,
      "loss": 0.2120358943939209,
      "mean_token_accuracy": 0.9426074802875519,
      "num_tokens": 98754618.0,
      "step": 6135
    },
    {
      "aux_loss": 8.061479568481445,
      "entropy": 0.1997602991759777,
      "epoch": 0.00614,
      "grad_norm": 0.13871872425079346,
      "learning_rate": 2e-05,
      "loss": 0.24574472904205322,
      "mean_token_accuracy": 0.9321439087390899,
      "num_tokens": 98835844.0,
      "step": 6140
    },
    {
      "aux_loss": 8.059854507446289,
      "entropy": 0.19934454634785653,
      "epoch": 0.006145,
      "grad_norm": 0.13629339635372162,
      "learning_rate": 2e-05,
      "loss": 0.22751383781433104,
      "mean_token_accuracy": 0.936464735865593,
      "num_tokens": 98916650.0,
      "step": 6145
    },
    {
      "aux_loss": 8.067006635665894,
      "entropy": 0.18546929582953453,
      "epoch": 0.00615,
      "grad_norm": 0.1268135905265808,
      "learning_rate": 2e-05,
      "loss": 0.2111495018005371,
      "mean_token_accuracy": 0.9414886891841888,
      "num_tokens": 98997540.0,
      "step": 6150
    },
    {
      "aux_loss": 8.078183698654176,
      "entropy": 0.16131289191544057,
      "epoch": 0.006155,
      "grad_norm": 0.1337651014328003,
      "learning_rate": 2e-05,
      "loss": 0.18758971691131593,
      "mean_token_accuracy": 0.9483745336532593,
      "num_tokens": 99077962.0,
      "step": 6155
    },
    {
      "aux_loss": 8.089561986923218,
      "entropy": 0.1473751597106457,
      "epoch": 0.00616,
      "grad_norm": 0.146542489528656,
      "learning_rate": 2e-05,
      "loss": 0.17467809915542604,
      "mean_token_accuracy": 0.9505683511495591,
      "num_tokens": 99158018.0,
      "step": 6160
    },
    {
      "aux_loss": 8.063340997695922,
      "entropy": 0.16819728650152682,
      "epoch": 0.006165,
      "grad_norm": 0.13880492746829987,
      "learning_rate": 2e-05,
      "loss": 0.20851426124572753,
      "mean_token_accuracy": 0.9430805325508118,
      "num_tokens": 99238930.0,
      "step": 6165
    },
    {
      "aux_loss": 8.057694911956787,
      "entropy": 0.18912734240293502,
      "epoch": 0.00617,
      "grad_norm": 0.134108304977417,
      "learning_rate": 2e-05,
      "loss": 0.23240768909454346,
      "mean_token_accuracy": 0.9356161594390869,
      "num_tokens": 99319755.0,
      "step": 6170
    },
    {
      "aux_loss": 8.05859932899475,
      "entropy": 0.18864412009716033,
      "epoch": 0.006175,
      "grad_norm": 0.12981119751930237,
      "learning_rate": 2e-05,
      "loss": 0.2230241060256958,
      "mean_token_accuracy": 0.9389087319374084,
      "num_tokens": 99400974.0,
      "step": 6175
    },
    {
      "aux_loss": 8.06623592376709,
      "entropy": 0.1835307776927948,
      "epoch": 0.00618,
      "grad_norm": 0.14064079523086548,
      "learning_rate": 2e-05,
      "loss": 0.212884521484375,
      "mean_token_accuracy": 0.9408647507429123,
      "num_tokens": 99481987.0,
      "step": 6180
    },
    {
      "aux_loss": 8.072245693206787,
      "entropy": 0.15626964047551156,
      "epoch": 0.006185,
      "grad_norm": 0.13499991595745087,
      "learning_rate": 2e-05,
      "loss": 0.18445948362350464,
      "mean_token_accuracy": 0.9477288573980331,
      "num_tokens": 99562468.0,
      "step": 6185
    },
    {
      "aux_loss": 8.084248924255371,
      "entropy": 0.15240877121686935,
      "epoch": 0.00619,
      "grad_norm": 0.13890454173088074,
      "learning_rate": 2e-05,
      "loss": 0.1771385431289673,
      "mean_token_accuracy": 0.9499257773160934,
      "num_tokens": 99642977.0,
      "step": 6190
    },
    {
      "aux_loss": 8.074431467056275,
      "entropy": 0.167390775680542,
      "epoch": 0.006195,
      "grad_norm": 0.12660688161849976,
      "learning_rate": 2e-05,
      "loss": 0.19968509674072266,
      "mean_token_accuracy": 0.9451132565736771,
      "num_tokens": 99721646.0,
      "step": 6195
    },
    {
      "aux_loss": 8.059612607955932,
      "entropy": 0.18690274953842162,
      "epoch": 0.0062,
      "grad_norm": 0.13673487305641174,
      "learning_rate": 2e-05,
      "loss": 0.2261310338973999,
      "mean_token_accuracy": 0.9381637156009675,
      "num_tokens": 99802848.0,
      "step": 6200
    },
    {
      "aux_loss": 8.060397386550903,
      "entropy": 0.1755303718149662,
      "epoch": 0.006205,
      "grad_norm": 0.14175184071063995,
      "learning_rate": 2e-05,
      "loss": 0.216101336479187,
      "mean_token_accuracy": 0.9398821413516998,
      "num_tokens": 99883821.0,
      "step": 6205
    },
    {
      "aux_loss": 8.070406627655029,
      "entropy": 0.17731958478689194,
      "epoch": 0.00621,
      "grad_norm": 0.13555993139743805,
      "learning_rate": 2e-05,
      "loss": 0.20861904621124266,
      "mean_token_accuracy": 0.9407115608453751,
      "num_tokens": 99964762.0,
      "step": 6210
    },
    {
      "aux_loss": 8.078168630599976,
      "entropy": 0.16126366034150125,
      "epoch": 0.006215,
      "grad_norm": 0.12997405230998993,
      "learning_rate": 2e-05,
      "loss": 0.18080706596374513,
      "mean_token_accuracy": 0.9492483675479889,
      "num_tokens": 100045203.0,
      "step": 6215
    },
    {
      "aux_loss": 8.088132619857788,
      "entropy": 0.14728614166378975,
      "epoch": 0.00622,
      "grad_norm": 0.14376845955848694,
      "learning_rate": 2e-05,
      "loss": 0.17079941034317017,
      "mean_token_accuracy": 0.9525896340608597,
      "num_tokens": 100125157.0,
      "step": 6220
    },
    {
      "aux_loss": 8.072662687301635,
      "entropy": 0.15450909920036793,
      "epoch": 0.006225,
      "grad_norm": 0.1307871788740158,
      "learning_rate": 2e-05,
      "loss": 0.19495983123779298,
      "mean_token_accuracy": 0.9467739045619965,
      "num_tokens": 100206074.0,
      "step": 6225
    },
    {
      "aux_loss": 8.061028575897216,
      "entropy": 0.1787221372127533,
      "epoch": 0.00623,
      "grad_norm": 0.14120079576969147,
      "learning_rate": 2e-05,
      "loss": 0.22198593616485596,
      "mean_token_accuracy": 0.9384955644607544,
      "num_tokens": 100286699.0,
      "step": 6230
    },
    {
      "aux_loss": 8.06080503463745,
      "entropy": 0.1750737138092518,
      "epoch": 0.006235,
      "grad_norm": 0.13577698171138763,
      "learning_rate": 2e-05,
      "loss": 0.21695075035095215,
      "mean_token_accuracy": 0.9401729851961136,
      "num_tokens": 100367634.0,
      "step": 6235
    },
    {
      "aux_loss": 8.067031860351562,
      "entropy": 0.16431664153933526,
      "epoch": 0.00624,
      "grad_norm": 0.12588439881801605,
      "learning_rate": 2e-05,
      "loss": 0.19172556400299073,
      "mean_token_accuracy": 0.9468090206384658,
      "num_tokens": 100448579.0,
      "step": 6240
    },
    {
      "aux_loss": 8.074049615859986,
      "entropy": 0.1671641945838928,
      "epoch": 0.006245,
      "grad_norm": 0.1321130394935608,
      "learning_rate": 2e-05,
      "loss": 0.19521715641021728,
      "mean_token_accuracy": 0.9454416334629059,
      "num_tokens": 100529476.0,
      "step": 6245
    },
    {
      "aux_loss": 8.084184312820435,
      "entropy": 0.15946810208261014,
      "epoch": 0.00625,
      "grad_norm": 0.1479744017124176,
      "learning_rate": 2e-05,
      "loss": 0.1840827465057373,
      "mean_token_accuracy": 0.9490292429924011,
      "num_tokens": 100610102.0,
      "step": 6250
    },
    {
      "aux_loss": 8.077231216430665,
      "entropy": 0.14674910046160222,
      "epoch": 0.006255,
      "grad_norm": 0.12755580246448517,
      "learning_rate": 2e-05,
      "loss": 0.17703778743743898,
      "mean_token_accuracy": 0.9516955465078354,
      "num_tokens": 100688691.0,
      "step": 6255
    },
    {
      "aux_loss": 8.059980821609496,
      "entropy": 0.17733871527016162,
      "epoch": 0.00626,
      "grad_norm": 0.14013931155204773,
      "learning_rate": 2e-05,
      "loss": 0.2219689130783081,
      "mean_token_accuracy": 0.938349649310112,
      "num_tokens": 100769628.0,
      "step": 6260
    },
    {
      "aux_loss": 8.063177871704102,
      "entropy": 0.1751440867781639,
      "epoch": 0.006265,
      "grad_norm": 0.14352500438690186,
      "learning_rate": 2e-05,
      "loss": 0.21170456409454347,
      "mean_token_accuracy": 0.9407127380371094,
      "num_tokens": 100850505.0,
      "step": 6265
    },
    {
      "aux_loss": 8.065592050552368,
      "entropy": 0.18136587142944335,
      "epoch": 0.00627,
      "grad_norm": 0.13016898930072784,
      "learning_rate": 2e-05,
      "loss": 0.21362464427947997,
      "mean_token_accuracy": 0.9404471904039383,
      "num_tokens": 100931559.0,
      "step": 6270
    },
    {
      "aux_loss": 8.071725225448608,
      "entropy": 0.1680338393896818,
      "epoch": 0.006275,
      "grad_norm": 0.13893038034439087,
      "learning_rate": 2e-05,
      "loss": 0.19443680047988893,
      "mean_token_accuracy": 0.9450006574392319,
      "num_tokens": 101012279.0,
      "step": 6275
    },
    {
      "aux_loss": 8.082179164886474,
      "entropy": 0.15971801355481147,
      "epoch": 0.00628,
      "grad_norm": 0.1487090140581131,
      "learning_rate": 2e-05,
      "loss": 0.18496278524398804,
      "mean_token_accuracy": 0.9478775680065155,
      "num_tokens": 101093008.0,
      "step": 6280
    },
    {
      "aux_loss": 8.083986186981202,
      "entropy": 0.16268995590507984,
      "epoch": 0.006285,
      "grad_norm": 0.13213466107845306,
      "learning_rate": 2e-05,
      "loss": 0.1961376667022705,
      "mean_token_accuracy": 0.9464788019657135,
      "num_tokens": 101170521.0,
      "step": 6285
    },
    {
      "aux_loss": 8.059894895553589,
      "entropy": 0.17033642530441284,
      "epoch": 0.00629,
      "grad_norm": 0.13566872477531433,
      "learning_rate": 2e-05,
      "loss": 0.21158432960510254,
      "mean_token_accuracy": 0.9407631993293762,
      "num_tokens": 101251626.0,
      "step": 6290
    },
    {
      "aux_loss": 8.059821033477784,
      "entropy": 0.17324242703616619,
      "epoch": 0.006295,
      "grad_norm": 0.14442917704582214,
      "learning_rate": 2e-05,
      "loss": 0.21338980197906493,
      "mean_token_accuracy": 0.9399789214134217,
      "num_tokens": 101332185.0,
      "step": 6295
    },
    {
      "aux_loss": 8.064768934249878,
      "entropy": 0.18178710490465164,
      "epoch": 0.0063,
      "grad_norm": 0.14033494889736176,
      "learning_rate": 2e-05,
      "loss": 0.21303117275238037,
      "mean_token_accuracy": 0.9392391294240952,
      "num_tokens": 101413044.0,
      "step": 6300
    },
    {
      "aux_loss": 8.07487931251526,
      "entropy": 0.1670963678508997,
      "epoch": 0.006305,
      "grad_norm": 0.13624851405620575,
      "learning_rate": 2e-05,
      "loss": 0.19358983039855956,
      "mean_token_accuracy": 0.9451935350894928,
      "num_tokens": 101493841.0,
      "step": 6305
    },
    {
      "aux_loss": 8.080358409881592,
      "entropy": 0.15010572001338005,
      "epoch": 0.00631,
      "grad_norm": 0.13583829998970032,
      "learning_rate": 2e-05,
      "loss": 0.17737114429473877,
      "mean_token_accuracy": 0.9494899690151215,
      "num_tokens": 101574315.0,
      "step": 6310
    },
    {
      "aux_loss": 8.08941583633423,
      "entropy": 0.15250032395124435,
      "epoch": 0.006315,
      "grad_norm": 0.14238521456718445,
      "learning_rate": 2e-05,
      "loss": 0.18297399282455445,
      "mean_token_accuracy": 0.949578019976616,
      "num_tokens": 101650808.0,
      "step": 6315
    },
    {
      "aux_loss": 8.05661997795105,
      "entropy": 0.1723372034728527,
      "epoch": 0.00632,
      "grad_norm": 0.13784413039684296,
      "learning_rate": 2e-05,
      "loss": 0.2155602216720581,
      "mean_token_accuracy": 0.9406784355640412,
      "num_tokens": 101732175.0,
      "step": 6320
    },
    {
      "aux_loss": 8.056038904190064,
      "entropy": 0.17770125046372415,
      "epoch": 0.006325,
      "grad_norm": 0.1384645253419876,
      "learning_rate": 2e-05,
      "loss": 0.2155226230621338,
      "mean_token_accuracy": 0.940444940328598,
      "num_tokens": 101813586.0,
      "step": 6325
    },
    {
      "aux_loss": 8.061874866485596,
      "entropy": 0.17923748418688773,
      "epoch": 0.00633,
      "grad_norm": 0.14557155966758728,
      "learning_rate": 2e-05,
      "loss": 0.2121119499206543,
      "mean_token_accuracy": 0.9412698000669479,
      "num_tokens": 101894258.0,
      "step": 6330
    },
    {
      "aux_loss": 8.067919588088989,
      "entropy": 0.16947432681918145,
      "epoch": 0.006335,
      "grad_norm": 0.1293715536594391,
      "learning_rate": 2e-05,
      "loss": 0.1959705114364624,
      "mean_token_accuracy": 0.9455020755529404,
      "num_tokens": 101975137.0,
      "step": 6335
    },
    {
      "aux_loss": 8.075682258605957,
      "entropy": 0.16904918551445008,
      "epoch": 0.00634,
      "grad_norm": 0.1394459307193756,
      "learning_rate": 2e-05,
      "loss": 0.19398807287216185,
      "mean_token_accuracy": 0.9463216930627822,
      "num_tokens": 102055325.0,
      "step": 6340
    },
    {
      "aux_loss": 8.088437271118163,
      "entropy": 0.14480783976614475,
      "epoch": 0.006345,
      "grad_norm": 0.1445549726486206,
      "learning_rate": 2e-05,
      "loss": 0.17504599094390869,
      "mean_token_accuracy": 0.9519671112298965,
      "num_tokens": 102133231.0,
      "step": 6345
    },
    {
      "aux_loss": 8.053894186019898,
      "entropy": 0.17075674533843993,
      "epoch": 0.00635,
      "grad_norm": 0.13737104833126068,
      "learning_rate": 2e-05,
      "loss": 0.21471831798553467,
      "mean_token_accuracy": 0.9409557968378067,
      "num_tokens": 102214602.0,
      "step": 6350
    },
    {
      "aux_loss": 8.059332466125488,
      "entropy": 0.18495972119271756,
      "epoch": 0.006355,
      "grad_norm": 0.13046260178089142,
      "learning_rate": 2e-05,
      "loss": 0.22336864471435547,
      "mean_token_accuracy": 0.9380752146244049,
      "num_tokens": 102296034.0,
      "step": 6355
    },
    {
      "aux_loss": 8.062976837158203,
      "entropy": 0.18716762810945511,
      "epoch": 0.00636,
      "grad_norm": 0.12837707996368408,
      "learning_rate": 2e-05,
      "loss": 0.21661136150360108,
      "mean_token_accuracy": 0.93944251537323,
      "num_tokens": 102377116.0,
      "step": 6360
    },
    {
      "aux_loss": 8.069791984558105,
      "entropy": 0.18775770887732507,
      "epoch": 0.006365,
      "grad_norm": 0.13980534672737122,
      "learning_rate": 2e-05,
      "loss": 0.21184928417205812,
      "mean_token_accuracy": 0.9406641364097595,
      "num_tokens": 102457851.0,
      "step": 6365
    },
    {
      "aux_loss": 8.076382970809936,
      "entropy": 0.15570278875529767,
      "epoch": 0.00637,
      "grad_norm": 0.1451767235994339,
      "learning_rate": 2e-05,
      "loss": 0.1806432008743286,
      "mean_token_accuracy": 0.9495420396327973,
      "num_tokens": 102538278.0,
      "step": 6370
    },
    {
      "aux_loss": 8.093045139312744,
      "entropy": 0.14868444874882697,
      "epoch": 0.006375,
      "grad_norm": 0.1694030910730362,
      "learning_rate": 2e-05,
      "loss": 0.17895843982696533,
      "mean_token_accuracy": 0.950489941239357,
      "num_tokens": 102618109.0,
      "step": 6375
    },
    {
      "aux_loss": 8.05609827041626,
      "entropy": 0.15887579768896104,
      "epoch": 0.00638,
      "grad_norm": 0.13222035765647888,
      "learning_rate": 2e-05,
      "loss": 0.19839054346084595,
      "mean_token_accuracy": 0.9463102877140045,
      "num_tokens": 102699390.0,
      "step": 6380
    },
    {
      "aux_loss": 8.058115196228027,
      "entropy": 0.1806506983935833,
      "epoch": 0.006385,
      "grad_norm": 0.13981851935386658,
      "learning_rate": 2e-05,
      "loss": 0.22575092315673828,
      "mean_token_accuracy": 0.9366847574710846,
      "num_tokens": 102780414.0,
      "step": 6385
    },
    {
      "aux_loss": 8.060818958282471,
      "entropy": 0.17934260740876198,
      "epoch": 0.00639,
      "grad_norm": 0.1297139823436737,
      "learning_rate": 2e-05,
      "loss": 0.21333825588226318,
      "mean_token_accuracy": 0.9409090429544449,
      "num_tokens": 102860940.0,
      "step": 6390
    },
    {
      "aux_loss": 8.06777834892273,
      "entropy": 0.1966196946799755,
      "epoch": 0.006395,
      "grad_norm": 0.1342628300189972,
      "learning_rate": 2e-05,
      "loss": 0.22602782249450684,
      "mean_token_accuracy": 0.9370049566030503,
      "num_tokens": 102941531.0,
      "step": 6395
    },
    {
      "aux_loss": 8.073350095748902,
      "entropy": 0.16740402728319168,
      "epoch": 0.0064,
      "grad_norm": 0.1337825208902359,
      "learning_rate": 2e-05,
      "loss": 0.19232274293899537,
      "mean_token_accuracy": 0.9470282524824143,
      "num_tokens": 103022099.0,
      "step": 6400
    },
    {
      "aux_loss": 8.085779571533203,
      "entropy": 0.15219030603766442,
      "epoch": 0.006405,
      "grad_norm": 0.1452208161354065,
      "learning_rate": 2e-05,
      "loss": 0.1763453245162964,
      "mean_token_accuracy": 0.9512188225984574,
      "num_tokens": 103102334.0,
      "step": 6405
    },
    {
      "aux_loss": 8.065481805801392,
      "entropy": 0.1620359502732754,
      "epoch": 0.00641,
      "grad_norm": 0.1380251795053482,
      "learning_rate": 2e-05,
      "loss": 0.20248186588287354,
      "mean_token_accuracy": 0.9455227285623551,
      "num_tokens": 103183019.0,
      "step": 6410
    },
    {
      "aux_loss": 8.060709190368652,
      "entropy": 0.1893113970756531,
      "epoch": 0.006415,
      "grad_norm": 0.14346621930599213,
      "learning_rate": 2e-05,
      "loss": 0.2336345672607422,
      "mean_token_accuracy": 0.9357190012931824,
      "num_tokens": 103263812.0,
      "step": 6415
    },
    {
      "aux_loss": 8.060740518569947,
      "entropy": 0.20097941234707833,
      "epoch": 0.00642,
      "grad_norm": 0.1351483017206192,
      "learning_rate": 2e-05,
      "loss": 0.2385958194732666,
      "mean_token_accuracy": 0.9341014057397843,
      "num_tokens": 103344593.0,
      "step": 6420
    },
    {
      "aux_loss": 8.06397876739502,
      "entropy": 0.1782874498516321,
      "epoch": 0.006425,
      "grad_norm": 0.12901869416236877,
      "learning_rate": 2e-05,
      "loss": 0.20582175254821777,
      "mean_token_accuracy": 0.9428041458129883,
      "num_tokens": 103425352.0,
      "step": 6425
    },
    {
      "aux_loss": 8.073601055145264,
      "entropy": 0.17462822794914246,
      "epoch": 0.00643,
      "grad_norm": 0.1316116601228714,
      "learning_rate": 2e-05,
      "loss": 0.20616085529327394,
      "mean_token_accuracy": 0.9420482069253922,
      "num_tokens": 103505951.0,
      "step": 6430
    },
    {
      "aux_loss": 8.080321311950684,
      "entropy": 0.1603805750608444,
      "epoch": 0.006435,
      "grad_norm": 0.13217490911483765,
      "learning_rate": 2e-05,
      "loss": 0.18733654022216797,
      "mean_token_accuracy": 0.9473046660423279,
      "num_tokens": 103586403.0,
      "step": 6435
    },
    {
      "aux_loss": 8.074872922897338,
      "entropy": 0.15219646506011486,
      "epoch": 0.00644,
      "grad_norm": 0.1349012553691864,
      "learning_rate": 2e-05,
      "loss": 0.1850438594818115,
      "mean_token_accuracy": 0.9502466320991516,
      "num_tokens": 103663739.0,
      "step": 6440
    },
    {
      "aux_loss": 8.055966472625732,
      "entropy": 0.20168813094496726,
      "epoch": 0.006445,
      "grad_norm": 0.13414886593818665,
      "learning_rate": 2e-05,
      "loss": 0.24502916336059571,
      "mean_token_accuracy": 0.9330766826868058,
      "num_tokens": 103744688.0,
      "step": 6445
    },
    {
      "aux_loss": 8.059682655334473,
      "entropy": 0.1899139355868101,
      "epoch": 0.00645,
      "grad_norm": 0.13743431866168976,
      "learning_rate": 2e-05,
      "loss": 0.22691383361816406,
      "mean_token_accuracy": 0.9376573890447617,
      "num_tokens": 103825465.0,
      "step": 6450
    },
    {
      "aux_loss": 8.066794633865356,
      "entropy": 0.17871159315109253,
      "epoch": 0.006455,
      "grad_norm": 0.13151606917381287,
      "learning_rate": 2e-05,
      "loss": 0.2089684247970581,
      "mean_token_accuracy": 0.9403085857629776,
      "num_tokens": 103906392.0,
      "step": 6455
    },
    {
      "aux_loss": 8.07235927581787,
      "entropy": 0.17127743512392044,
      "epoch": 0.00646,
      "grad_norm": 0.12921243906021118,
      "learning_rate": 2e-05,
      "loss": 0.1983320951461792,
      "mean_token_accuracy": 0.9451368868350982,
      "num_tokens": 103987013.0,
      "step": 6460
    },
    {
      "aux_loss": 8.079568862915039,
      "entropy": 0.15652903616428376,
      "epoch": 0.006465,
      "grad_norm": 0.13730360567569733,
      "learning_rate": 2e-05,
      "loss": 0.18192354440689087,
      "mean_token_accuracy": 0.9490347474813461,
      "num_tokens": 104067556.0,
      "step": 6465
    },
    {
      "aux_loss": 8.08714051246643,
      "entropy": 0.14996255822479726,
      "epoch": 0.00647,
      "grad_norm": 0.12541280686855316,
      "learning_rate": 2e-05,
      "loss": 0.17527691125869752,
      "mean_token_accuracy": 0.9513368755578995,
      "num_tokens": 104144236.0,
      "step": 6470
    },
    {
      "aux_loss": 8.059282493591308,
      "entropy": 0.17776259817183018,
      "epoch": 0.006475,
      "grad_norm": 0.13069605827331543,
      "learning_rate": 2e-05,
      "loss": 0.2222141981124878,
      "mean_token_accuracy": 0.9384141057729721,
      "num_tokens": 104225446.0,
      "step": 6475
    },
    {
      "aux_loss": 8.054843664169312,
      "entropy": 0.1879831813275814,
      "epoch": 0.00648,
      "grad_norm": 0.1447993963956833,
      "learning_rate": 2e-05,
      "loss": 0.2318018913269043,
      "mean_token_accuracy": 0.9365807831287384,
      "num_tokens": 104306350.0,
      "step": 6480
    },
    {
      "aux_loss": 8.058440113067627,
      "entropy": 0.1796765249222517,
      "epoch": 0.006485,
      "grad_norm": 0.13044868409633636,
      "learning_rate": 2e-05,
      "loss": 0.21003224849700927,
      "mean_token_accuracy": 0.9415479063987732,
      "num_tokens": 104387085.0,
      "step": 6485
    },
    {
      "aux_loss": 8.067186880111695,
      "entropy": 0.16888436824083328,
      "epoch": 0.00649,
      "grad_norm": 0.1329645812511444,
      "learning_rate": 2e-05,
      "loss": 0.19239667654037476,
      "mean_token_accuracy": 0.9472585588693618,
      "num_tokens": 104467319.0,
      "step": 6490
    },
    {
      "aux_loss": 8.074488878250122,
      "entropy": 0.15988884083926677,
      "epoch": 0.006495,
      "grad_norm": 0.1404368132352829,
      "learning_rate": 2e-05,
      "loss": 0.18557705879211425,
      "mean_token_accuracy": 0.9484501391649246,
      "num_tokens": 104546959.0,
      "step": 6495
    },
    {
      "aux_loss": 8.08824725151062,
      "entropy": 0.1398022759705782,
      "epoch": 0.0065,
      "grad_norm": 0.12826049327850342,
      "learning_rate": 2e-05,
      "loss": 0.1691373825073242,
      "mean_token_accuracy": 0.9537356555461883,
      "num_tokens": 104627237.0,
      "step": 6500
    },
    {
      "aux_loss": 8.052957391738891,
      "entropy": 0.18530977740883828,
      "epoch": 0.006505,
      "grad_norm": 0.1500970721244812,
      "learning_rate": 2e-05,
      "loss": 0.23158650398254393,
      "mean_token_accuracy": 0.9359501153230667,
      "num_tokens": 104708198.0,
      "step": 6505
    },
    {
      "aux_loss": 8.056018400192261,
      "entropy": 0.1863008677959442,
      "epoch": 0.00651,
      "grad_norm": 0.1350659728050232,
      "learning_rate": 2e-05,
      "loss": 0.22994539737701417,
      "mean_token_accuracy": 0.9373692750930787,
      "num_tokens": 104789530.0,
      "step": 6510
    },
    {
      "aux_loss": 8.06305227279663,
      "entropy": 0.19534257426857948,
      "epoch": 0.006515,
      "grad_norm": 0.12743796408176422,
      "learning_rate": 2e-05,
      "loss": 0.22684235572814943,
      "mean_token_accuracy": 0.9365925937891006,
      "num_tokens": 104870440.0,
      "step": 6515
    },
    {
      "aux_loss": 8.070610857009887,
      "entropy": 0.1902066335082054,
      "epoch": 0.00652,
      "grad_norm": 0.13365595042705536,
      "learning_rate": 2e-05,
      "loss": 0.2153306007385254,
      "mean_token_accuracy": 0.939258161187172,
      "num_tokens": 104950859.0,
      "step": 6520
    },
    {
      "aux_loss": 8.073332023620605,
      "entropy": 0.17365288510918617,
      "epoch": 0.006525,
      "grad_norm": 0.13237911462783813,
      "learning_rate": 2e-05,
      "loss": 0.1991954565048218,
      "mean_token_accuracy": 0.944195294380188,
      "num_tokens": 105030900.0,
      "step": 6525
    },
    {
      "aux_loss": 8.086445665359497,
      "entropy": 0.15020677708089353,
      "epoch": 0.00653,
      "grad_norm": 0.14996060729026794,
      "learning_rate": 2e-05,
      "loss": 0.17386245727539062,
      "mean_token_accuracy": 0.9523511111736298,
      "num_tokens": 105110880.0,
      "step": 6530
    },
    {
      "aux_loss": 8.054518985748292,
      "entropy": 0.1636942494660616,
      "epoch": 0.006535,
      "grad_norm": 0.139048233628273,
      "learning_rate": 2e-05,
      "loss": 0.20500919818878174,
      "mean_token_accuracy": 0.9441078394651413,
      "num_tokens": 105188071.0,
      "step": 6535
    },
    {
      "aux_loss": 8.058283233642578,
      "entropy": 0.18873534649610518,
      "epoch": 0.00654,
      "grad_norm": 0.1392601728439331,
      "learning_rate": 2e-05,
      "loss": 0.23121559619903564,
      "mean_token_accuracy": 0.9366243630647659,
      "num_tokens": 105269529.0,
      "step": 6540
    },
    {
      "aux_loss": 8.058865451812744,
      "entropy": 0.16948034428060055,
      "epoch": 0.006545,
      "grad_norm": 0.12779845297336578,
      "learning_rate": 2e-05,
      "loss": 0.2034911870956421,
      "mean_token_accuracy": 0.9437923997640609,
      "num_tokens": 105350826.0,
      "step": 6545
    },
    {
      "aux_loss": 8.067020893096924,
      "entropy": 0.1822333849966526,
      "epoch": 0.00655,
      "grad_norm": 0.1375105381011963,
      "learning_rate": 2e-05,
      "loss": 0.20614545345306395,
      "mean_token_accuracy": 0.9421275317668915,
      "num_tokens": 105431431.0,
      "step": 6550
    },
    {
      "aux_loss": 8.07479681968689,
      "entropy": 0.17016190513968468,
      "epoch": 0.006555,
      "grad_norm": 0.13541489839553833,
      "learning_rate": 2e-05,
      "loss": 0.19411683082580566,
      "mean_token_accuracy": 0.9465625077486038,
      "num_tokens": 105511533.0,
      "step": 6555
    },
    {
      "aux_loss": 8.08745527267456,
      "entropy": 0.15058867894113065,
      "epoch": 0.00656,
      "grad_norm": 0.15061993896961212,
      "learning_rate": 2e-05,
      "loss": 0.1770238995552063,
      "mean_token_accuracy": 0.9511159539222718,
      "num_tokens": 105591922.0,
      "step": 6560
    },
    {
      "aux_loss": 8.066929578781128,
      "entropy": 0.17327417582273483,
      "epoch": 0.006565,
      "grad_norm": 0.13963890075683594,
      "learning_rate": 2e-05,
      "loss": 0.2097869873046875,
      "mean_token_accuracy": 0.9418965160846711,
      "num_tokens": 105669729.0,
      "step": 6565
    },
    {
      "aux_loss": 8.059988069534302,
      "entropy": 0.1767417922616005,
      "epoch": 0.00657,
      "grad_norm": 0.13648806512355804,
      "learning_rate": 2e-05,
      "loss": 0.2210536479949951,
      "mean_token_accuracy": 0.939487236738205,
      "num_tokens": 105751040.0,
      "step": 6570
    },
    {
      "aux_loss": 8.060567808151244,
      "entropy": 0.18365311175584792,
      "epoch": 0.006575,
      "grad_norm": 0.13076962530612946,
      "learning_rate": 2e-05,
      "loss": 0.22124106884002687,
      "mean_token_accuracy": 0.938056617975235,
      "num_tokens": 105832440.0,
      "step": 6575
    },
    {
      "aux_loss": 8.06791181564331,
      "entropy": 0.18670534677803516,
      "epoch": 0.00658,
      "grad_norm": 0.12424053251743317,
      "learning_rate": 2e-05,
      "loss": 0.212841796875,
      "mean_token_accuracy": 0.9406958937644958,
      "num_tokens": 105913085.0,
      "step": 6580
    },
    {
      "aux_loss": 8.074774360656738,
      "entropy": 0.1809455320239067,
      "epoch": 0.006585,
      "grad_norm": 0.13560666143894196,
      "learning_rate": 2e-05,
      "loss": 0.2035428285598755,
      "mean_token_accuracy": 0.9440999895334243,
      "num_tokens": 105993790.0,
      "step": 6585
    },
    {
      "aux_loss": 8.08399920463562,
      "entropy": 0.15627671107649804,
      "epoch": 0.00659,
      "grad_norm": 0.15000271797180176,
      "learning_rate": 2e-05,
      "loss": 0.18291122913360597,
      "mean_token_accuracy": 0.9494013279676438,
      "num_tokens": 106073949.0,
      "step": 6590
    },
    {
      "aux_loss": 8.072545909881592,
      "entropy": 0.1805926974862814,
      "epoch": 0.006595,
      "grad_norm": 0.1438024938106537,
      "learning_rate": 2e-05,
      "loss": 0.21861820220947265,
      "mean_token_accuracy": 0.9400606960058212,
      "num_tokens": 106152757.0,
      "step": 6595
    },
    {
      "aux_loss": 8.057387018203736,
      "entropy": 0.19217270687222482,
      "epoch": 0.0066,
      "grad_norm": 0.1483086347579956,
      "learning_rate": 2e-05,
      "loss": 0.23820018768310547,
      "mean_token_accuracy": 0.9352487921714783,
      "num_tokens": 106233759.0,
      "step": 6600
    },
    {
      "aux_loss": 8.061714553833008,
      "entropy": 0.19212173148989678,
      "epoch": 0.006605,
      "grad_norm": 0.12826994061470032,
      "learning_rate": 2e-05,
      "loss": 0.22636728286743163,
      "mean_token_accuracy": 0.9375009626150131,
      "num_tokens": 106314761.0,
      "step": 6605
    },
    {
      "aux_loss": 8.063360214233398,
      "entropy": 0.19681259840726853,
      "epoch": 0.00661,
      "grad_norm": 0.12922720611095428,
      "learning_rate": 2e-05,
      "loss": 0.22415735721588134,
      "mean_token_accuracy": 0.9371958076953888,
      "num_tokens": 106395504.0,
      "step": 6610
    },
    {
      "aux_loss": 8.067996263504028,
      "entropy": 0.18302029445767404,
      "epoch": 0.006615,
      "grad_norm": 0.1312997043132782,
      "learning_rate": 2e-05,
      "loss": 0.2081216096878052,
      "mean_token_accuracy": 0.9416180312633514,
      "num_tokens": 106476323.0,
      "step": 6615
    },
    {
      "aux_loss": 8.075970935821534,
      "entropy": 0.1569758526980877,
      "epoch": 0.00662,
      "grad_norm": 0.15728217363357544,
      "learning_rate": 2e-05,
      "loss": 0.18248648643493653,
      "mean_token_accuracy": 0.9502352803945542,
      "num_tokens": 106556954.0,
      "step": 6620
    },
    {
      "aux_loss": 8.081261444091798,
      "entropy": 0.15056833885610105,
      "epoch": 0.006625,
      "grad_norm": 0.13029922544956207,
      "learning_rate": 2e-05,
      "loss": 0.17897058725357057,
      "mean_token_accuracy": 0.9469171613454819,
      "num_tokens": 106633460.0,
      "step": 6625
    },
    {
      "aux_loss": 8.056975650787354,
      "entropy": 0.20196614861488343,
      "epoch": 0.00663,
      "grad_norm": 0.14535565674304962,
      "learning_rate": 2e-05,
      "loss": 0.2484807014465332,
      "mean_token_accuracy": 0.9315470159053802,
      "num_tokens": 106714714.0,
      "step": 6630
    },
    {
      "aux_loss": 8.059915542602539,
      "entropy": 0.1933729939162731,
      "epoch": 0.006635,
      "grad_norm": 0.13469654321670532,
      "learning_rate": 2e-05,
      "loss": 0.22994585037231446,
      "mean_token_accuracy": 0.9362727642059326,
      "num_tokens": 106796062.0,
      "step": 6635
    },
    {
      "aux_loss": 8.064970064163209,
      "entropy": 0.1833428829908371,
      "epoch": 0.00664,
      "grad_norm": 0.1303439885377884,
      "learning_rate": 2e-05,
      "loss": 0.213004469871521,
      "mean_token_accuracy": 0.9411578148603439,
      "num_tokens": 106877213.0,
      "step": 6640
    },
    {
      "aux_loss": 8.073453855514526,
      "entropy": 0.18130681216716765,
      "epoch": 0.006645,
      "grad_norm": 0.1338777095079422,
      "learning_rate": 2e-05,
      "loss": 0.20527844429016112,
      "mean_token_accuracy": 0.94286450445652,
      "num_tokens": 106958021.0,
      "step": 6645
    },
    {
      "aux_loss": 8.081600046157837,
      "entropy": 0.1550985224545002,
      "epoch": 0.00665,
      "grad_norm": 0.14633561670780182,
      "learning_rate": 2e-05,
      "loss": 0.17817087173461915,
      "mean_token_accuracy": 0.9505529850721359,
      "num_tokens": 107038583.0,
      "step": 6650
    },
    {
      "aux_loss": 8.08010368347168,
      "entropy": 0.14544684253633022,
      "epoch": 0.006655,
      "grad_norm": 0.12753990292549133,
      "learning_rate": 2e-05,
      "loss": 0.17821503877639772,
      "mean_token_accuracy": 0.9524681568145752,
      "num_tokens": 107115835.0,
      "step": 6655
    },
    {
      "aux_loss": 8.058242654800415,
      "entropy": 0.19374196976423264,
      "epoch": 0.00666,
      "grad_norm": 0.14286461472511292,
      "learning_rate": 2e-05,
      "loss": 0.2412653923034668,
      "mean_token_accuracy": 0.9333578646183014,
      "num_tokens": 107197217.0,
      "step": 6660
    },
    {
      "aux_loss": 8.058524560928344,
      "entropy": 0.193814517557621,
      "epoch": 0.006665,
      "grad_norm": 0.14072448015213013,
      "learning_rate": 2e-05,
      "loss": 0.23591794967651367,
      "mean_token_accuracy": 0.9340570271015167,
      "num_tokens": 107277838.0,
      "step": 6665
    },
    {
      "aux_loss": 8.060419511795043,
      "entropy": 0.19526009038090705,
      "epoch": 0.00667,
      "grad_norm": 0.1275109201669693,
      "learning_rate": 2e-05,
      "loss": 0.2266437292098999,
      "mean_token_accuracy": 0.9376020550727844,
      "num_tokens": 107358411.0,
      "step": 6670
    },
    {
      "aux_loss": 8.068568563461303,
      "entropy": 0.178493682295084,
      "epoch": 0.006675,
      "grad_norm": 0.13420528173446655,
      "learning_rate": 2e-05,
      "loss": 0.20035793781280517,
      "mean_token_accuracy": 0.9438051968812943,
      "num_tokens": 107439121.0,
      "step": 6675
    },
    {
      "aux_loss": 8.077151107788087,
      "entropy": 0.17396341487765313,
      "epoch": 0.00668,
      "grad_norm": 0.1457800567150116,
      "learning_rate": 2e-05,
      "loss": 0.2007878303527832,
      "mean_token_accuracy": 0.9430379956960678,
      "num_tokens": 107520306.0,
      "step": 6680
    },
    {
      "aux_loss": 8.083810710906983,
      "entropy": 0.14405076950788498,
      "epoch": 0.006685,
      "grad_norm": 0.12061306834220886,
      "learning_rate": 2e-05,
      "loss": 0.1787621021270752,
      "mean_token_accuracy": 0.9509928166866303,
      "num_tokens": 107597366.0,
      "step": 6685
    },
    {
      "aux_loss": 8.056791305541992,
      "entropy": 0.21013906523585318,
      "epoch": 0.00669,
      "grad_norm": 0.14182820916175842,
      "learning_rate": 2e-05,
      "loss": 0.25118963718414306,
      "mean_token_accuracy": 0.9320299386978149,
      "num_tokens": 107678724.0,
      "step": 6690
    },
    {
      "aux_loss": 8.059663486480712,
      "entropy": 0.20437658838927747,
      "epoch": 0.006695,
      "grad_norm": 0.12801802158355713,
      "learning_rate": 2e-05,
      "loss": 0.23997869491577148,
      "mean_token_accuracy": 0.9333798080682755,
      "num_tokens": 107759944.0,
      "step": 6695
    },
    {
      "aux_loss": 8.060472249984741,
      "entropy": 0.1899130627512932,
      "epoch": 0.0067,
      "grad_norm": 0.13734589517116547,
      "learning_rate": 2e-05,
      "loss": 0.2205308198928833,
      "mean_token_accuracy": 0.9388278216123581,
      "num_tokens": 107840832.0,
      "step": 6700
    },
    {
      "aux_loss": 8.068573808670044,
      "entropy": 0.18932122439146043,
      "epoch": 0.006705,
      "grad_norm": 0.14030024409294128,
      "learning_rate": 2e-05,
      "loss": 0.2195117712020874,
      "mean_token_accuracy": 0.9391480773687363,
      "num_tokens": 107921401.0,
      "step": 6705
    },
    {
      "aux_loss": 8.076237630844116,
      "entropy": 0.16646337546408177,
      "epoch": 0.00671,
      "grad_norm": 0.12890629470348358,
      "learning_rate": 2e-05,
      "loss": 0.19483060836791993,
      "mean_token_accuracy": 0.9460500359535218,
      "num_tokens": 108001687.0,
      "step": 6710
    },
    {
      "aux_loss": 8.089707183837891,
      "entropy": 0.1615411765873432,
      "epoch": 0.006715,
      "grad_norm": 0.21329468488693237,
      "learning_rate": 2e-05,
      "loss": 0.1896498680114746,
      "mean_token_accuracy": 0.9470278322696686,
      "num_tokens": 108080166.0,
      "step": 6715
    },
    {
      "aux_loss": 8.053832149505615,
      "entropy": 0.17431953698396682,
      "epoch": 0.00672,
      "grad_norm": 0.13431163132190704,
      "learning_rate": 2e-05,
      "loss": 0.21453843116760254,
      "mean_token_accuracy": 0.9422074437141419,
      "num_tokens": 108161359.0,
      "step": 6720
    },
    {
      "aux_loss": 8.056922817230225,
      "entropy": 0.1907948836684227,
      "epoch": 0.006725,
      "grad_norm": 0.13440436124801636,
      "learning_rate": 2e-05,
      "loss": 0.2308185577392578,
      "mean_token_accuracy": 0.9372014433145524,
      "num_tokens": 108242649.0,
      "step": 6725
    },
    {
      "aux_loss": 8.060697078704834,
      "entropy": 0.1860272318124771,
      "epoch": 0.00673,
      "grad_norm": 0.12962786853313446,
      "learning_rate": 2e-05,
      "loss": 0.21740972995758057,
      "mean_token_accuracy": 0.939772167801857,
      "num_tokens": 108323689.0,
      "step": 6730
    },
    {
      "aux_loss": 8.067785120010376,
      "entropy": 0.20486112534999848,
      "epoch": 0.006735,
      "grad_norm": 0.14315727353096008,
      "learning_rate": 2e-05,
      "loss": 0.2384364128112793,
      "mean_token_accuracy": 0.9337804347276688,
      "num_tokens": 108404556.0,
      "step": 6735
    },
    {
      "aux_loss": 8.070926666259766,
      "entropy": 0.16807297691702844,
      "epoch": 0.00674,
      "grad_norm": 0.1316627562046051,
      "learning_rate": 2e-05,
      "loss": 0.1896495819091797,
      "mean_token_accuracy": 0.9470909118652344,
      "num_tokens": 108485280.0,
      "step": 6740
    },
    {
      "aux_loss": 8.083463144302367,
      "entropy": 0.15742594823241235,
      "epoch": 0.006745,
      "grad_norm": 0.1355815976858139,
      "learning_rate": 2e-05,
      "loss": 0.18207114934921265,
      "mean_token_accuracy": 0.9488089293241501,
      "num_tokens": 108565343.0,
      "step": 6745
    },
    {
      "aux_loss": 8.065484142303466,
      "entropy": 0.17837621867656708,
      "epoch": 0.00675,
      "grad_norm": 0.13580040633678436,
      "learning_rate": 2e-05,
      "loss": 0.2193075656890869,
      "mean_token_accuracy": 0.9402038544416428,
      "num_tokens": 108644820.0,
      "step": 6750
    },
    {
      "aux_loss": 8.059473514556885,
      "entropy": 0.19034894444048406,
      "epoch": 0.006755,
      "grad_norm": 0.14436422288417816,
      "learning_rate": 2e-05,
      "loss": 0.23359949588775636,
      "mean_token_accuracy": 0.935259535908699,
      "num_tokens": 108726002.0,
      "step": 6755
    },
    {
      "aux_loss": 8.057626676559448,
      "entropy": 0.18970154076814652,
      "epoch": 0.00676,
      "grad_norm": 0.13981929421424866,
      "learning_rate": 2e-05,
      "loss": 0.22816820144653321,
      "mean_token_accuracy": 0.9363198816776276,
      "num_tokens": 108806816.0,
      "step": 6760
    },
    {
      "aux_loss": 8.063619422912598,
      "entropy": 0.1876445196568966,
      "epoch": 0.006765,
      "grad_norm": 0.12872658669948578,
      "learning_rate": 2e-05,
      "loss": 0.2164677858352661,
      "mean_token_accuracy": 0.9406172186136246,
      "num_tokens": 108887661.0,
      "step": 6765
    },
    {
      "aux_loss": 8.07038254737854,
      "entropy": 0.1758854180574417,
      "epoch": 0.00677,
      "grad_norm": 0.12422648072242737,
      "learning_rate": 2e-05,
      "loss": 0.19471333026885987,
      "mean_token_accuracy": 0.9462656289339065,
      "num_tokens": 108968469.0,
      "step": 6770
    },
    {
      "aux_loss": 8.080739402770996,
      "entropy": 0.14386206939816476,
      "epoch": 0.006775,
      "grad_norm": 0.13551288843154907,
      "learning_rate": 2e-05,
      "loss": 0.1684157967567444,
      "mean_token_accuracy": 0.9528355211019516,
      "num_tokens": 109048940.0,
      "step": 6775
    },
    {
      "aux_loss": 8.07333745956421,
      "entropy": 0.16987128220498562,
      "epoch": 0.00678,
      "grad_norm": 0.12569330632686615,
      "learning_rate": 2e-05,
      "loss": 0.21025750637054444,
      "mean_token_accuracy": 0.9418981492519378,
      "num_tokens": 109128048.0,
      "step": 6780
    },
    {
      "aux_loss": 8.057387399673463,
      "entropy": 0.18619118109345437,
      "epoch": 0.006785,
      "grad_norm": 0.13500316441059113,
      "learning_rate": 2e-05,
      "loss": 0.22751965522766113,
      "mean_token_accuracy": 0.9370787292718887,
      "num_tokens": 109209432.0,
      "step": 6785
    },
    {
      "aux_loss": 8.060741424560547,
      "entropy": 0.19404347911477088,
      "epoch": 0.00679,
      "grad_norm": 0.13726806640625,
      "learning_rate": 2e-05,
      "loss": 0.23631765842437744,
      "mean_token_accuracy": 0.9343833029270172,
      "num_tokens": 109290434.0,
      "step": 6790
    },
    {
      "aux_loss": 8.065060949325561,
      "entropy": 0.18128382861614228,
      "epoch": 0.006795,
      "grad_norm": 0.12944266200065613,
      "learning_rate": 2e-05,
      "loss": 0.2106311321258545,
      "mean_token_accuracy": 0.9406280428171158,
      "num_tokens": 109371345.0,
      "step": 6795
    },
    {
      "aux_loss": 8.067907905578613,
      "entropy": 0.17607021257281302,
      "epoch": 0.0068,
      "grad_norm": 0.1292889565229416,
      "learning_rate": 2e-05,
      "loss": 0.20047850608825685,
      "mean_token_accuracy": 0.943754893541336,
      "num_tokens": 109451740.0,
      "step": 6800
    },
    {
      "aux_loss": 8.07772216796875,
      "entropy": 0.15810563787817955,
      "epoch": 0.006805,
      "grad_norm": 0.13309086859226227,
      "learning_rate": 2e-05,
      "loss": 0.18045300245285034,
      "mean_token_accuracy": 0.9496598452329635,
      "num_tokens": 109531905.0,
      "step": 6805
    },
    {
      "aux_loss": 8.070142364501953,
      "entropy": 0.16251008845865728,
      "epoch": 0.00681,
      "grad_norm": 0.1362977772951126,
      "learning_rate": 2e-05,
      "loss": 0.1953115940093994,
      "mean_token_accuracy": 0.9467148929834366,
      "num_tokens": 109610227.0,
      "step": 6810
    },
    {
      "aux_loss": 8.05427656173706,
      "entropy": 0.20478485673666,
      "epoch": 0.006815,
      "grad_norm": 0.1396312266588211,
      "learning_rate": 2e-05,
      "loss": 0.25188241004943845,
      "mean_token_accuracy": 0.9313136786222458,
      "num_tokens": 109691387.0,
      "step": 6815
    },
    {
      "aux_loss": 8.058727979660034,
      "entropy": 0.2000688396394253,
      "epoch": 0.00682,
      "grad_norm": 0.12259568274021149,
      "learning_rate": 2e-05,
      "loss": 0.2394339323043823,
      "mean_token_accuracy": 0.9337880730628967,
      "num_tokens": 109772332.0,
      "step": 6820
    },
    {
      "aux_loss": 8.06331491470337,
      "entropy": 0.20067718476057053,
      "epoch": 0.006825,
      "grad_norm": 0.13288083672523499,
      "learning_rate": 2e-05,
      "loss": 0.22876579761505128,
      "mean_token_accuracy": 0.9371572852134704,
      "num_tokens": 109853329.0,
      "step": 6825
    },
    {
      "aux_loss": 8.068853902816773,
      "entropy": 0.18307801187038422,
      "epoch": 0.00683,
      "grad_norm": 0.13197623193264008,
      "learning_rate": 2e-05,
      "loss": 0.21064381599426268,
      "mean_token_accuracy": 0.9424758553504944,
      "num_tokens": 109934003.0,
      "step": 6830
    },
    {
      "aux_loss": 8.080074024200439,
      "entropy": 0.1613745927810669,
      "epoch": 0.006835,
      "grad_norm": 0.1528049111366272,
      "learning_rate": 2e-05,
      "loss": 0.18981387615203857,
      "mean_token_accuracy": 0.9484105855226517,
      "num_tokens": 110013900.0,
      "step": 6835
    },
    {
      "aux_loss": 8.078043985366822,
      "entropy": 0.14951721727848052,
      "epoch": 0.00684,
      "grad_norm": 0.13994275033473969,
      "learning_rate": 2e-05,
      "loss": 0.18397578001022338,
      "mean_token_accuracy": 0.9501771062612534,
      "num_tokens": 110092248.0,
      "step": 6840
    },
    {
      "aux_loss": 8.055387353897094,
      "entropy": 0.19120416045188904,
      "epoch": 0.006845,
      "grad_norm": 0.14257748425006866,
      "learning_rate": 2e-05,
      "loss": 0.23187901973724365,
      "mean_token_accuracy": 0.9358863234519958,
      "num_tokens": 110173065.0,
      "step": 6845
    },
    {
      "aux_loss": 8.056835651397705,
      "entropy": 0.18925052024424077,
      "epoch": 0.00685,
      "grad_norm": 0.13090164959430695,
      "learning_rate": 2e-05,
      "loss": 0.22437198162078859,
      "mean_token_accuracy": 0.9382225751876831,
      "num_tokens": 110254084.0,
      "step": 6850
    },
    {
      "aux_loss": 8.06151032447815,
      "entropy": 0.1916178446263075,
      "epoch": 0.006855,
      "grad_norm": 0.13993686437606812,
      "learning_rate": 2e-05,
      "loss": 0.223028826713562,
      "mean_token_accuracy": 0.9386214166879654,
      "num_tokens": 110334504.0,
      "step": 6855
    },
    {
      "aux_loss": 8.065795183181763,
      "entropy": 0.18025753945112227,
      "epoch": 0.00686,
      "grad_norm": 0.1293414980173111,
      "learning_rate": 2e-05,
      "loss": 0.2055180311203003,
      "mean_token_accuracy": 0.9433819651603699,
      "num_tokens": 110415448.0,
      "step": 6860
    },
    {
      "aux_loss": 8.072910642623901,
      "entropy": 0.16579332202672958,
      "epoch": 0.006865,
      "grad_norm": 0.15642303228378296,
      "learning_rate": 2e-05,
      "loss": 0.192667818069458,
      "mean_token_accuracy": 0.9462066143751144,
      "num_tokens": 110496544.0,
      "step": 6865
    },
    {
      "aux_loss": 8.079344987869263,
      "entropy": 0.14516615755856038,
      "epoch": 0.00687,
      "grad_norm": 0.13828928768634796,
      "learning_rate": 2e-05,
      "loss": 0.1700227975845337,
      "mean_token_accuracy": 0.9530481725931168,
      "num_tokens": 110573773.0,
      "step": 6870
    },
    {
      "aux_loss": 8.050804615020752,
      "entropy": 0.1830222938209772,
      "epoch": 0.006875,
      "grad_norm": 0.13731269538402557,
      "learning_rate": 2e-05,
      "loss": 0.22500884532928467,
      "mean_token_accuracy": 0.9397697508335113,
      "num_tokens": 110655033.0,
      "step": 6875
    },
    {
      "aux_loss": 8.053100633621217,
      "entropy": 0.19751237854361534,
      "epoch": 0.00688,
      "grad_norm": 0.13944876194000244,
      "learning_rate": 2e-05,
      "loss": 0.2381190538406372,
      "mean_token_accuracy": 0.9329667806625366,
      "num_tokens": 110736212.0,
      "step": 6880
    },
    {
      "aux_loss": 8.058423900604248,
      "entropy": 0.18282825984060763,
      "epoch": 0.006885,
      "grad_norm": 0.12871089577674866,
      "learning_rate": 2e-05,
      "loss": 0.2117380142211914,
      "mean_token_accuracy": 0.9409862369298935,
      "num_tokens": 110817125.0,
      "step": 6885
    },
    {
      "aux_loss": 8.063421773910523,
      "entropy": 0.17388938404619694,
      "epoch": 0.00689,
      "grad_norm": 0.13460378348827362,
      "learning_rate": 2e-05,
      "loss": 0.19637068510055541,
      "mean_token_accuracy": 0.9451661705970764,
      "num_tokens": 110897779.0,
      "step": 6890
    },
    {
      "aux_loss": 8.071660709381103,
      "entropy": 0.1595969472080469,
      "epoch": 0.006895,
      "grad_norm": 0.1393076479434967,
      "learning_rate": 2e-05,
      "loss": 0.18909016847610474,
      "mean_token_accuracy": 0.9462069272994995,
      "num_tokens": 110978080.0,
      "step": 6895
    },
    {
      "aux_loss": 8.086488819122314,
      "entropy": 0.15020788870751858,
      "epoch": 0.0069,
      "grad_norm": 0.15149039030075073,
      "learning_rate": 2e-05,
      "loss": 0.1846014976501465,
      "mean_token_accuracy": 0.9486435681581498,
      "num_tokens": 111057346.0,
      "step": 6900
    },
    {
      "aux_loss": 8.055602884292602,
      "entropy": 0.1961719837039709,
      "epoch": 0.006905,
      "grad_norm": 0.14332440495491028,
      "learning_rate": 2e-05,
      "loss": 0.24031534194946289,
      "mean_token_accuracy": 0.9335872083902359,
      "num_tokens": 111138608.0,
      "step": 6905
    },
    {
      "aux_loss": 8.059308862686157,
      "entropy": 0.20703301578760147,
      "epoch": 0.00691,
      "grad_norm": 0.12914623320102692,
      "learning_rate": 2e-05,
      "loss": 0.24857690334320068,
      "mean_token_accuracy": 0.931088387966156,
      "num_tokens": 111219408.0,
      "step": 6910
    },
    {
      "aux_loss": 8.059395360946656,
      "entropy": 0.202732365578413,
      "epoch": 0.006915,
      "grad_norm": 0.1352163404226303,
      "learning_rate": 2e-05,
      "loss": 0.2315981149673462,
      "mean_token_accuracy": 0.9357743471860885,
      "num_tokens": 111300285.0,
      "step": 6915
    },
    {
      "aux_loss": 8.062900590896607,
      "entropy": 0.18121012523770333,
      "epoch": 0.00692,
      "grad_norm": 0.13391900062561035,
      "learning_rate": 2e-05,
      "loss": 0.2070383310317993,
      "mean_token_accuracy": 0.943183696269989,
      "num_tokens": 111381188.0,
      "step": 6920
    },
    {
      "aux_loss": 8.06819167137146,
      "entropy": 0.16773489862680435,
      "epoch": 0.006925,
      "grad_norm": 0.1380244940519333,
      "learning_rate": 2e-05,
      "loss": 0.19404807090759277,
      "mean_token_accuracy": 0.9467169791460037,
      "num_tokens": 111461254.0,
      "step": 6925
    },
    {
      "aux_loss": 8.083512306213379,
      "entropy": 0.14787791818380355,
      "epoch": 0.00693,
      "grad_norm": 0.15291079878807068,
      "learning_rate": 2e-05,
      "loss": 0.17515230178833008,
      "mean_token_accuracy": 0.951891052722931,
      "num_tokens": 111541466.0,
      "step": 6930
    },
    {
      "aux_loss": 8.063453865051269,
      "entropy": 0.15990073122084142,
      "epoch": 0.006935,
      "grad_norm": 0.13127268850803375,
      "learning_rate": 2e-05,
      "loss": 0.19819576740264894,
      "mean_token_accuracy": 0.9464002907276153,
      "num_tokens": 111621110.0,
      "step": 6935
    },
    {
      "aux_loss": 8.056852865219117,
      "entropy": 0.19791870787739754,
      "epoch": 0.00694,
      "grad_norm": 0.14063888788223267,
      "learning_rate": 2e-05,
      "loss": 0.24480123519897462,
      "mean_token_accuracy": 0.9323440104722976,
      "num_tokens": 111702240.0,
      "step": 6940
    },
    {
      "aux_loss": 8.058293867111207,
      "entropy": 0.19194333180785178,
      "epoch": 0.006945,
      "grad_norm": 0.13223256170749664,
      "learning_rate": 2e-05,
      "loss": 0.2275765895843506,
      "mean_token_accuracy": 0.937348023056984,
      "num_tokens": 111783161.0,
      "step": 6945
    },
    {
      "aux_loss": 8.064247035980225,
      "entropy": 0.18550173714756965,
      "epoch": 0.00695,
      "grad_norm": 0.12810027599334717,
      "learning_rate": 2e-05,
      "loss": 0.20763940811157228,
      "mean_token_accuracy": 0.9421748757362366,
      "num_tokens": 111864063.0,
      "step": 6950
    },
    {
      "aux_loss": 8.070037651062012,
      "entropy": 0.17414481341838836,
      "epoch": 0.006955,
      "grad_norm": 0.14044809341430664,
      "learning_rate": 2e-05,
      "loss": 0.20043206214904785,
      "mean_token_accuracy": 0.9433826714754104,
      "num_tokens": 111944599.0,
      "step": 6955
    },
    {
      "aux_loss": 8.079628562927246,
      "entropy": 0.15028354302048683,
      "epoch": 0.00696,
      "grad_norm": 0.153854638338089,
      "learning_rate": 2e-05,
      "loss": 0.18153855800628663,
      "mean_token_accuracy": 0.9484508752822876,
      "num_tokens": 112024764.0,
      "step": 6960
    },
    {
      "aux_loss": 8.061726951599121,
      "entropy": 0.16690603084862232,
      "epoch": 0.006965,
      "grad_norm": 0.13620921969413757,
      "learning_rate": 2e-05,
      "loss": 0.20388710498809814,
      "mean_token_accuracy": 0.9443705022335053,
      "num_tokens": 112103878.0,
      "step": 6965
    },
    {
      "aux_loss": 8.058962440490722,
      "entropy": 0.19669478423893452,
      "epoch": 0.00697,
      "grad_norm": 0.13166803121566772,
      "learning_rate": 2e-05,
      "loss": 0.23782813549041748,
      "mean_token_accuracy": 0.9345280230045319,
      "num_tokens": 112185391.0,
      "step": 6970
    },
    {
      "aux_loss": 8.057507133483886,
      "entropy": 0.1985299989581108,
      "epoch": 0.006975,
      "grad_norm": 0.12910211086273193,
      "learning_rate": 2e-05,
      "loss": 0.23177578449249267,
      "mean_token_accuracy": 0.9366781413555145,
      "num_tokens": 112266553.0,
      "step": 6975
    },
    {
      "aux_loss": 8.063517522811889,
      "entropy": 0.18326300829648973,
      "epoch": 0.00698,
      "grad_norm": 0.1300823986530304,
      "learning_rate": 2e-05,
      "loss": 0.21248078346252441,
      "mean_token_accuracy": 0.9411573708057404,
      "num_tokens": 112347294.0,
      "step": 6980
    },
    {
      "aux_loss": 8.07069435119629,
      "entropy": 0.1764574684202671,
      "epoch": 0.006985,
      "grad_norm": 0.13278327882289886,
      "learning_rate": 2e-05,
      "loss": 0.20272331237792968,
      "mean_token_accuracy": 0.9434526592493058,
      "num_tokens": 112428013.0,
      "step": 6985
    },
    {
      "aux_loss": 8.081476879119872,
      "entropy": 0.15756783448159695,
      "epoch": 0.00699,
      "grad_norm": 0.1429092139005661,
      "learning_rate": 2e-05,
      "loss": 0.18209214210510255,
      "mean_token_accuracy": 0.9492216110229492,
      "num_tokens": 112508231.0,
      "step": 6990
    },
    {
      "aux_loss": 8.06993646621704,
      "entropy": 0.16216953173279763,
      "epoch": 0.006995,
      "grad_norm": 0.12714068591594696,
      "learning_rate": 2e-05,
      "loss": 0.1991791009902954,
      "mean_token_accuracy": 0.9440642416477203,
      "num_tokens": 112586809.0,
      "step": 6995
    },
    {
      "aux_loss": 8.052938604354859,
      "entropy": 0.20342568233609198,
      "epoch": 0.007,
      "grad_norm": 0.1341390162706375,
      "learning_rate": 2e-05,
      "loss": 0.2487004280090332,
      "mean_token_accuracy": 0.9307210355997085,
      "num_tokens": 112668141.0,
      "step": 7000
    },
    {
      "aux_loss": 8.057039022445679,
      "entropy": 0.1822150766849518,
      "epoch": 0.007005,
      "grad_norm": 0.1320662498474121,
      "learning_rate": 2e-05,
      "loss": 0.21836693286895753,
      "mean_token_accuracy": 0.939070850610733,
      "num_tokens": 112749233.0,
      "step": 7005
    },
    {
      "aux_loss": 8.060214424133301,
      "entropy": 0.17933097407221793,
      "epoch": 0.00701,
      "grad_norm": 0.1267113983631134,
      "learning_rate": 2e-05,
      "loss": 0.20422806739807128,
      "mean_token_accuracy": 0.9432374060153961,
      "num_tokens": 112830244.0,
      "step": 7010
    },
    {
      "aux_loss": 8.067915773391723,
      "entropy": 0.17348047643899916,
      "epoch": 0.007015,
      "grad_norm": 0.1358555406332016,
      "learning_rate": 2e-05,
      "loss": 0.19757884740829468,
      "mean_token_accuracy": 0.9455861330032349,
      "num_tokens": 112911185.0,
      "step": 7015
    },
    {
      "aux_loss": 8.07640118598938,
      "entropy": 0.14997454285621642,
      "epoch": 0.00702,
      "grad_norm": 0.14231212437152863,
      "learning_rate": 2e-05,
      "loss": 0.1767973780632019,
      "mean_token_accuracy": 0.9509914219379425,
      "num_tokens": 112991404.0,
      "step": 7020
    },
    {
      "aux_loss": 8.070006895065308,
      "entropy": 0.1480850413441658,
      "epoch": 0.007025,
      "grad_norm": 0.13705086708068848,
      "learning_rate": 2e-05,
      "loss": 0.18057994842529296,
      "mean_token_accuracy": 0.9501135468482971,
      "num_tokens": 113069054.0,
      "step": 7025
    },
    {
      "aux_loss": 8.054139804840087,
      "entropy": 0.17381192483007907,
      "epoch": 0.00703,
      "grad_norm": 0.14269228279590607,
      "learning_rate": 2e-05,
      "loss": 0.21657919883728027,
      "mean_token_accuracy": 0.9390967428684235,
      "num_tokens": 113150161.0,
      "step": 7030
    },
    {
      "aux_loss": 8.055738592147828,
      "entropy": 0.19614654183387756,
      "epoch": 0.007035,
      "grad_norm": 0.12956149876117706,
      "learning_rate": 2e-05,
      "loss": 0.23727197647094728,
      "mean_token_accuracy": 0.9344140857458114,
      "num_tokens": 113230726.0,
      "step": 7035
    },
    {
      "aux_loss": 8.058074474334717,
      "entropy": 0.18538858070969583,
      "epoch": 0.00704,
      "grad_norm": 0.13408729434013367,
      "learning_rate": 2e-05,
      "loss": 0.21406421661376954,
      "mean_token_accuracy": 0.940402963757515,
      "num_tokens": 113311589.0,
      "step": 7040
    },
    {
      "aux_loss": 8.068364429473878,
      "entropy": 0.1827181063592434,
      "epoch": 0.007045,
      "grad_norm": 0.14206305146217346,
      "learning_rate": 2e-05,
      "loss": 0.20695745944976807,
      "mean_token_accuracy": 0.941447052359581,
      "num_tokens": 113392472.0,
      "step": 7045
    },
    {
      "aux_loss": 8.073644208908082,
      "entropy": 0.16167523339390755,
      "epoch": 0.00705,
      "grad_norm": 0.1369583010673523,
      "learning_rate": 2e-05,
      "loss": 0.18708597421646117,
      "mean_token_accuracy": 0.9479620009660721,
      "num_tokens": 113473166.0,
      "step": 7050
    },
    {
      "aux_loss": 8.079384756088256,
      "entropy": 0.15518963374197484,
      "epoch": 0.007055,
      "grad_norm": 0.126423180103302,
      "learning_rate": 2e-05,
      "loss": 0.18250396251678466,
      "mean_token_accuracy": 0.9494188785552978,
      "num_tokens": 113551458.0,
      "step": 7055
    },
    {
      "aux_loss": 8.052441120147705,
      "entropy": 0.20016354694962502,
      "epoch": 0.00706,
      "grad_norm": 0.1303359568119049,
      "learning_rate": 2e-05,
      "loss": 0.24556608200073243,
      "mean_token_accuracy": 0.9331448823213577,
      "num_tokens": 113633087.0,
      "step": 7060
    },
    {
      "aux_loss": 8.056761503219604,
      "entropy": 0.19472678676247596,
      "epoch": 0.007065,
      "grad_norm": 0.1315099149942398,
      "learning_rate": 2e-05,
      "loss": 0.2353368043899536,
      "mean_token_accuracy": 0.9340995758771896,
      "num_tokens": 113714117.0,
      "step": 7065
    },
    {
      "aux_loss": 8.060474061965943,
      "entropy": 0.17509630098938941,
      "epoch": 0.00707,
      "grad_norm": 0.12468890845775604,
      "learning_rate": 2e-05,
      "loss": 0.20644593238830566,
      "mean_token_accuracy": 0.9434526115655899,
      "num_tokens": 113794646.0,
      "step": 7070
    },
    {
      "aux_loss": 8.068728876113891,
      "entropy": 0.1740468643605709,
      "epoch": 0.007075,
      "grad_norm": 0.13405142724514008,
      "learning_rate": 2e-05,
      "loss": 0.1977548360824585,
      "mean_token_accuracy": 0.9444215714931488,
      "num_tokens": 113875078.0,
      "step": 7075
    },
    {
      "aux_loss": 8.074010181427003,
      "entropy": 0.15743556693196298,
      "epoch": 0.00708,
      "grad_norm": 0.1447540819644928,
      "learning_rate": 2e-05,
      "loss": 0.18424359560012818,
      "mean_token_accuracy": 0.9496011435985565,
      "num_tokens": 113955611.0,
      "step": 7080
    },
    {
      "aux_loss": 8.083940553665162,
      "entropy": 0.14909599982202054,
      "epoch": 0.007085,
      "grad_norm": 0.12713557481765747,
      "learning_rate": 2e-05,
      "loss": 0.1818614959716797,
      "mean_token_accuracy": 0.9497849643230438,
      "num_tokens": 114034251.0,
      "step": 7085
    },
    {
      "aux_loss": 8.049956130981446,
      "entropy": 0.18191827796399593,
      "epoch": 0.00709,
      "grad_norm": 0.13210304081439972,
      "learning_rate": 2e-05,
      "loss": 0.2215649127960205,
      "mean_token_accuracy": 0.9386815667152405,
      "num_tokens": 114115770.0,
      "step": 7090
    },
    {
      "aux_loss": 8.056961631774902,
      "entropy": 0.20289888903498648,
      "epoch": 0.007095,
      "grad_norm": 0.13716211915016174,
      "learning_rate": 2e-05,
      "loss": 0.24096074104309081,
      "mean_token_accuracy": 0.9343296110630035,
      "num_tokens": 114196903.0,
      "step": 7095
    },
    {
      "aux_loss": 8.057987880706786,
      "entropy": 0.18323367536067964,
      "epoch": 0.0071,
      "grad_norm": 0.12688027322292328,
      "learning_rate": 2e-05,
      "loss": 0.21829211711883545,
      "mean_token_accuracy": 0.9390118360519409,
      "num_tokens": 114277890.0,
      "step": 7100
    },
    {
      "aux_loss": 8.062729549407958,
      "entropy": 0.1816129706799984,
      "epoch": 0.007105,
      "grad_norm": 0.1281527578830719,
      "learning_rate": 2e-05,
      "loss": 0.21081395149230958,
      "mean_token_accuracy": 0.9405385702848434,
      "num_tokens": 114358553.0,
      "step": 7105
    },
    {
      "aux_loss": 8.070510625839233,
      "entropy": 0.1619888201355934,
      "epoch": 0.00711,
      "grad_norm": 0.1359483003616333,
      "learning_rate": 2e-05,
      "loss": 0.18314563035964965,
      "mean_token_accuracy": 0.9491298496723175,
      "num_tokens": 114439175.0,
      "step": 7110
    },
    {
      "aux_loss": 8.089284420013428,
      "entropy": 0.15384032763540745,
      "epoch": 0.007115,
      "grad_norm": 0.15055975317955017,
      "learning_rate": 2e-05,
      "loss": 0.1828370451927185,
      "mean_token_accuracy": 0.9485554873943329,
      "num_tokens": 114519315.0,
      "step": 7115
    },
    {
      "aux_loss": 8.050078392028809,
      "entropy": 0.181279044598341,
      "epoch": 0.00712,
      "grad_norm": 0.13102883100509644,
      "learning_rate": 2e-05,
      "loss": 0.22407925128936768,
      "mean_token_accuracy": 0.9366942703723907,
      "num_tokens": 114600595.0,
      "step": 7120
    },
    {
      "aux_loss": 8.05292468070984,
      "entropy": 0.19039197117090226,
      "epoch": 0.007125,
      "grad_norm": 0.14553800225257874,
      "learning_rate": 2e-05,
      "loss": 0.2279196262359619,
      "mean_token_accuracy": 0.9349955052137375,
      "num_tokens": 114681650.0,
      "step": 7125
    },
    {
      "aux_loss": 8.056312847137452,
      "entropy": 0.1868167079985142,
      "epoch": 0.00713,
      "grad_norm": 0.12976329028606415,
      "learning_rate": 2e-05,
      "loss": 0.22218914031982423,
      "mean_token_accuracy": 0.9386069148778915,
      "num_tokens": 114762443.0,
      "step": 7130
    },
    {
      "aux_loss": 8.064145469665528,
      "entropy": 0.17997071743011475,
      "epoch": 0.007135,
      "grad_norm": 0.13100044429302216,
      "learning_rate": 2e-05,
      "loss": 0.20837724208831787,
      "mean_token_accuracy": 0.9420854121446609,
      "num_tokens": 114843149.0,
      "step": 7135
    },
    {
      "aux_loss": 8.072184228897095,
      "entropy": 0.15913472957909108,
      "epoch": 0.00714,
      "grad_norm": 0.1354636549949646,
      "learning_rate": 2e-05,
      "loss": 0.18547892570495605,
      "mean_token_accuracy": 0.9483176797628403,
      "num_tokens": 114923306.0,
      "step": 7140
    },
    {
      "aux_loss": 8.08354058265686,
      "entropy": 0.14665511325001718,
      "epoch": 0.007145,
      "grad_norm": 0.1437242180109024,
      "learning_rate": 2e-05,
      "loss": 0.1727093815803528,
      "mean_token_accuracy": 0.9517208278179169,
      "num_tokens": 115003325.0,
      "step": 7145
    },
    {
      "aux_loss": 8.05896463394165,
      "entropy": 0.1686363026499748,
      "epoch": 0.00715,
      "grad_norm": 0.15330128371715546,
      "learning_rate": 2e-05,
      "loss": 0.21119279861450196,
      "mean_token_accuracy": 0.9434751361608505,
      "num_tokens": 115081108.0,
      "step": 7150
    },
    {
      "aux_loss": 8.054275941848754,
      "entropy": 0.17525300458073617,
      "epoch": 0.007155,
      "grad_norm": 0.14003179967403412,
      "learning_rate": 2e-05,
      "loss": 0.21937448978424073,
      "mean_token_accuracy": 0.9394598424434661,
      "num_tokens": 115162274.0,
      "step": 7155
    },
    {
      "aux_loss": 8.056509160995484,
      "entropy": 0.1968074031174183,
      "epoch": 0.00716,
      "grad_norm": 0.1276974380016327,
      "learning_rate": 2e-05,
      "loss": 0.23514800071716307,
      "mean_token_accuracy": 0.9346697181463242,
      "num_tokens": 115243231.0,
      "step": 7160
    },
    {
      "aux_loss": 8.062007665634155,
      "entropy": 0.17011548653244973,
      "epoch": 0.007165,
      "grad_norm": 0.1272108405828476,
      "learning_rate": 2e-05,
      "loss": 0.19616918563842772,
      "mean_token_accuracy": 0.9454666316509247,
      "num_tokens": 115324054.0,
      "step": 7165
    },
    {
      "aux_loss": 8.068539762496949,
      "entropy": 0.16745559722185135,
      "epoch": 0.00717,
      "grad_norm": 0.12922310829162598,
      "learning_rate": 2e-05,
      "loss": 0.18886666297912597,
      "mean_token_accuracy": 0.9473625183105469,
      "num_tokens": 115404540.0,
      "step": 7170
    },
    {
      "aux_loss": 8.078390312194824,
      "entropy": 0.15166331827640533,
      "epoch": 0.007175,
      "grad_norm": 0.1345159262418747,
      "learning_rate": 2e-05,
      "loss": 0.17675564289093018,
      "mean_token_accuracy": 0.9505210280418396,
      "num_tokens": 115484724.0,
      "step": 7175
    },
    {
      "aux_loss": 8.06920928955078,
      "entropy": 0.16094010695815086,
      "epoch": 0.00718,
      "grad_norm": 0.13655336201190948,
      "learning_rate": 2e-05,
      "loss": 0.1940350890159607,
      "mean_token_accuracy": 0.9466844081878663,
      "num_tokens": 115562335.0,
      "step": 7180
    },
    {
      "aux_loss": 8.054381799697875,
      "entropy": 0.17820226773619652,
      "epoch": 0.007185,
      "grad_norm": 0.14267687499523163,
      "learning_rate": 2e-05,
      "loss": 0.22416069507598876,
      "mean_token_accuracy": 0.9382634341716767,
      "num_tokens": 115643339.0,
      "step": 7185
    },
    {
      "aux_loss": 8.057310247421265,
      "entropy": 0.19102592542767524,
      "epoch": 0.00719,
      "grad_norm": 0.14033392071723938,
      "learning_rate": 2e-05,
      "loss": 0.23336381912231446,
      "mean_token_accuracy": 0.9354314297437668,
      "num_tokens": 115724316.0,
      "step": 7190
    },
    {
      "aux_loss": 8.059385013580322,
      "entropy": 0.19008596166968345,
      "epoch": 0.007195,
      "grad_norm": 0.12919111549854279,
      "learning_rate": 2e-05,
      "loss": 0.21567316055297853,
      "mean_token_accuracy": 0.9399946600198745,
      "num_tokens": 115805291.0,
      "step": 7195
    },
    {
      "aux_loss": 8.062178421020509,
      "entropy": 0.1773022785782814,
      "epoch": 0.0072,
      "grad_norm": 0.1276605874300003,
      "learning_rate": 2e-05,
      "loss": 0.20061724185943602,
      "mean_token_accuracy": 0.9441809803247452,
      "num_tokens": 115886098.0,
      "step": 7200
    },
    {
      "aux_loss": 8.072524499893188,
      "entropy": 0.15648553147912025,
      "epoch": 0.007205,
      "grad_norm": 0.13309496641159058,
      "learning_rate": 2e-05,
      "loss": 0.1798306107521057,
      "mean_token_accuracy": 0.9506972104310989,
      "num_tokens": 115966782.0,
      "step": 7205
    },
    {
      "aux_loss": 8.072466135025024,
      "entropy": 0.1448201522231102,
      "epoch": 0.00721,
      "grad_norm": 0.13043193519115448,
      "learning_rate": 2e-05,
      "loss": 0.17662678956985473,
      "mean_token_accuracy": 0.9528814524412155,
      "num_tokens": 116046137.0,
      "step": 7210
    },
    {
      "aux_loss": 8.054290866851806,
      "entropy": 0.1953566338866949,
      "epoch": 0.007215,
      "grad_norm": 0.15037333965301514,
      "learning_rate": 2e-05,
      "loss": 0.24480528831481935,
      "mean_token_accuracy": 0.9315298944711685,
      "num_tokens": 116127333.0,
      "step": 7215
    },
    {
      "aux_loss": 8.055727910995483,
      "entropy": 0.18882788643240928,
      "epoch": 0.00722,
      "grad_norm": 0.13716623187065125,
      "learning_rate": 2e-05,
      "loss": 0.23057069778442382,
      "mean_token_accuracy": 0.937417060136795,
      "num_tokens": 116208330.0,
      "step": 7220
    },
    {
      "aux_loss": 8.059576272964478,
      "entropy": 0.1771923318505287,
      "epoch": 0.007225,
      "grad_norm": 0.1354140043258667,
      "learning_rate": 2e-05,
      "loss": 0.20474693775177003,
      "mean_token_accuracy": 0.9419525802135468,
      "num_tokens": 116289165.0,
      "step": 7225
    },
    {
      "aux_loss": 8.065571212768555,
      "entropy": 0.18347064033150673,
      "epoch": 0.00723,
      "grad_norm": 0.13612456619739532,
      "learning_rate": 2e-05,
      "loss": 0.2070516586303711,
      "mean_token_accuracy": 0.94222591817379,
      "num_tokens": 116369854.0,
      "step": 7230
    },
    {
      "aux_loss": 8.074532508850098,
      "entropy": 0.1525070745497942,
      "epoch": 0.007235,
      "grad_norm": 0.13450807332992554,
      "learning_rate": 2e-05,
      "loss": 0.17552847862243653,
      "mean_token_accuracy": 0.9517581731081008,
      "num_tokens": 116450659.0,
      "step": 7235
    },
    {
      "aux_loss": 8.07882218360901,
      "entropy": 0.14312020763754846,
      "epoch": 0.00724,
      "grad_norm": 0.1265956610441208,
      "learning_rate": 2e-05,
      "loss": 0.17811346054077148,
      "mean_token_accuracy": 0.9513128101825714,
      "num_tokens": 116529257.0,
      "step": 7240
    },
    {
      "aux_loss": 8.059095239639282,
      "entropy": 0.19069296047091483,
      "epoch": 0.007245,
      "grad_norm": 0.14973682165145874,
      "learning_rate": 2e-05,
      "loss": 0.2341454267501831,
      "mean_token_accuracy": 0.9352933436632156,
      "num_tokens": 116610855.0,
      "step": 7245
    },
    {
      "aux_loss": 8.055321168899535,
      "entropy": 0.18375627025961877,
      "epoch": 0.00725,
      "grad_norm": 0.12588109076023102,
      "learning_rate": 2e-05,
      "loss": 0.2217010259628296,
      "mean_token_accuracy": 0.93807832300663,
      "num_tokens": 116691708.0,
      "step": 7250
    },
    {
      "aux_loss": 8.059062385559082,
      "entropy": 0.17112191617488862,
      "epoch": 0.007255,
      "grad_norm": 0.12933211028575897,
      "learning_rate": 2e-05,
      "loss": 0.20111255645751952,
      "mean_token_accuracy": 0.9439534664154052,
      "num_tokens": 116772246.0,
      "step": 7255
    },
    {
      "aux_loss": 8.064316415786744,
      "entropy": 0.1701140858232975,
      "epoch": 0.00726,
      "grad_norm": 0.13094095885753632,
      "learning_rate": 2e-05,
      "loss": 0.19298509359359742,
      "mean_token_accuracy": 0.9466363191604614,
      "num_tokens": 116852645.0,
      "step": 7260
    },
    {
      "aux_loss": 8.070601797103881,
      "entropy": 0.16739392802119255,
      "epoch": 0.007265,
      "grad_norm": 0.14695289731025696,
      "learning_rate": 2e-05,
      "loss": 0.1961276888847351,
      "mean_token_accuracy": 0.945297583937645,
      "num_tokens": 116933229.0,
      "step": 7265
    },
    {
      "aux_loss": 8.08184199333191,
      "entropy": 0.1523813996464014,
      "epoch": 0.00727,
      "grad_norm": 0.13648247718811035,
      "learning_rate": 2e-05,
      "loss": 0.18260360956192018,
      "mean_token_accuracy": 0.9500061243772506,
      "num_tokens": 117011076.0,
      "step": 7270
    },
    {
      "aux_loss": 8.056529569625855,
      "entropy": 0.18706797063350677,
      "epoch": 0.007275,
      "grad_norm": 0.13587868213653564,
      "learning_rate": 2e-05,
      "loss": 0.23394424915313722,
      "mean_token_accuracy": 0.9364314496517181,
      "num_tokens": 117092640.0,
      "step": 7275
    },
    {
      "aux_loss": 8.05380277633667,
      "entropy": 0.18629618287086486,
      "epoch": 0.00728,
      "grad_norm": 0.17872829735279083,
      "learning_rate": 2e-05,
      "loss": 0.2321547508239746,
      "mean_token_accuracy": 0.936711460351944,
      "num_tokens": 117173746.0,
      "step": 7280
    },
    {
      "aux_loss": 8.05741114616394,
      "entropy": 0.1780652090907097,
      "epoch": 0.007285,
      "grad_norm": 0.12533755600452423,
      "learning_rate": 2e-05,
      "loss": 0.20582013130187987,
      "mean_token_accuracy": 0.9429244548082352,
      "num_tokens": 117254458.0,
      "step": 7285
    },
    {
      "aux_loss": 8.065899991989136,
      "entropy": 0.18077191561460496,
      "epoch": 0.00729,
      "grad_norm": 0.13414017856121063,
      "learning_rate": 2e-05,
      "loss": 0.20282549858093263,
      "mean_token_accuracy": 0.9424093931913375,
      "num_tokens": 117335127.0,
      "step": 7290
    },
    {
      "aux_loss": 8.071962690353393,
      "entropy": 0.16169586554169654,
      "epoch": 0.007295,
      "grad_norm": 0.1344667673110962,
      "learning_rate": 2e-05,
      "loss": 0.18243684768676757,
      "mean_token_accuracy": 0.9489847272634506,
      "num_tokens": 117416024.0,
      "step": 7295
    },
    {
      "aux_loss": 8.083722734451294,
      "entropy": 0.14176189824938773,
      "epoch": 0.0073,
      "grad_norm": 0.13662175834178925,
      "learning_rate": 2e-05,
      "loss": 0.17092795372009278,
      "mean_token_accuracy": 0.9530090302228927,
      "num_tokens": 117496453.0,
      "step": 7300
    },
    {
      "aux_loss": 8.057459402084351,
      "entropy": 0.17937030084431171,
      "epoch": 0.007305,
      "grad_norm": 0.13250751793384552,
      "learning_rate": 2e-05,
      "loss": 0.22172956466674804,
      "mean_token_accuracy": 0.9390877991914749,
      "num_tokens": 117577913.0,
      "step": 7305
    },
    {
      "aux_loss": 8.057992649078368,
      "entropy": 0.1838606558740139,
      "epoch": 0.00731,
      "grad_norm": 0.13557037711143494,
      "learning_rate": 2e-05,
      "loss": 0.23216652870178223,
      "mean_token_accuracy": 0.9359318941831589,
      "num_tokens": 117658940.0,
      "step": 7310
    },
    {
      "aux_loss": 8.056634569168091,
      "entropy": 0.17050421684980394,
      "epoch": 0.007315,
      "grad_norm": 0.127749502658844,
      "learning_rate": 2e-05,
      "loss": 0.20294065475463868,
      "mean_token_accuracy": 0.9430437445640564,
      "num_tokens": 117739805.0,
      "step": 7315
    },
    {
      "aux_loss": 8.063086032867432,
      "entropy": 0.17265646159648895,
      "epoch": 0.00732,
      "grad_norm": 0.12932780385017395,
      "learning_rate": 2e-05,
      "loss": 0.20049242973327636,
      "mean_token_accuracy": 0.9436680555343628,
      "num_tokens": 117819931.0,
      "step": 7320
    },
    {
      "aux_loss": 8.070355701446534,
      "entropy": 0.1700688511133194,
      "epoch": 0.007325,
      "grad_norm": 0.13648325204849243,
      "learning_rate": 2e-05,
      "loss": 0.19605551958084105,
      "mean_token_accuracy": 0.9448462516069412,
      "num_tokens": 117900606.0,
      "step": 7325
    },
    {
      "aux_loss": 8.08477554321289,
      "entropy": 0.151838706061244,
      "epoch": 0.00733,
      "grad_norm": 0.1555393785238266,
      "learning_rate": 2e-05,
      "loss": 0.17659342288970947,
      "mean_token_accuracy": 0.950711703300476,
      "num_tokens": 117981034.0,
      "step": 7330
    },
    {
      "aux_loss": 8.053568506240845,
      "entropy": 0.17783264219760894,
      "epoch": 0.007335,
      "grad_norm": 0.13863061368465424,
      "learning_rate": 2e-05,
      "loss": 0.21795213222503662,
      "mean_token_accuracy": 0.94056396484375,
      "num_tokens": 118058681.0,
      "step": 7335
    },
    {
      "aux_loss": 8.056577110290528,
      "entropy": 0.18549672774970533,
      "epoch": 0.00734,
      "grad_norm": 0.13383427262306213,
      "learning_rate": 2e-05,
      "loss": 0.22846200466156005,
      "mean_token_accuracy": 0.9356410890817642,
      "num_tokens": 118139985.0,
      "step": 7340
    },
    {
      "aux_loss": 8.055075359344482,
      "entropy": 0.18474764078855516,
      "epoch": 0.007345,
      "grad_norm": 0.1304062455892563,
      "learning_rate": 2e-05,
      "loss": 0.2212611675262451,
      "mean_token_accuracy": 0.9390469253063202,
      "num_tokens": 118220716.0,
      "step": 7345
    },
    {
      "aux_loss": 8.062121868133545,
      "entropy": 0.17709342762827873,
      "epoch": 0.00735,
      "grad_norm": 0.13126607239246368,
      "learning_rate": 2e-05,
      "loss": 0.20466187000274658,
      "mean_token_accuracy": 0.9431264877319336,
      "num_tokens": 118301339.0,
      "step": 7350
    },
    {
      "aux_loss": 8.069013690948486,
      "entropy": 0.16827764362096786,
      "epoch": 0.007355,
      "grad_norm": 0.13136102259159088,
      "learning_rate": 2e-05,
      "loss": 0.19352166652679442,
      "mean_token_accuracy": 0.9460424661636353,
      "num_tokens": 118382054.0,
      "step": 7355
    },
    {
      "aux_loss": 8.079519033432007,
      "entropy": 0.14911518841981888,
      "epoch": 0.00736,
      "grad_norm": 0.14757278561592102,
      "learning_rate": 2e-05,
      "loss": 0.17423219680786134,
      "mean_token_accuracy": 0.9520506620407104,
      "num_tokens": 118462143.0,
      "step": 7360
    },
    {
      "aux_loss": 8.061421775817871,
      "entropy": 0.15900407992303373,
      "epoch": 0.007365,
      "grad_norm": 0.1237441748380661,
      "learning_rate": 2e-05,
      "loss": 0.19232628345489503,
      "mean_token_accuracy": 0.9466270864009857,
      "num_tokens": 118541586.0,
      "step": 7365
    },
    {
      "aux_loss": 8.051958465576172,
      "entropy": 0.18195629641413688,
      "epoch": 0.00737,
      "grad_norm": 0.1475217193365097,
      "learning_rate": 2e-05,
      "loss": 0.2241210699081421,
      "mean_token_accuracy": 0.9381725996732712,
      "num_tokens": 118622519.0,
      "step": 7370
    },
    {
      "aux_loss": 8.054046058654786,
      "entropy": 0.17723673954606056,
      "epoch": 0.007375,
      "grad_norm": 0.127950981259346,
      "learning_rate": 2e-05,
      "loss": 0.21344656944274903,
      "mean_token_accuracy": 0.9404371291399002,
      "num_tokens": 118703590.0,
      "step": 7375
    },
    {
      "aux_loss": 8.062593698501587,
      "entropy": 0.17338769212365152,
      "epoch": 0.00738,
      "grad_norm": 0.13531026244163513,
      "learning_rate": 2e-05,
      "loss": 0.20329663753509522,
      "mean_token_accuracy": 0.9436062783002853,
      "num_tokens": 118784188.0,
      "step": 7380
    },
    {
      "aux_loss": 8.068348217010499,
      "entropy": 0.16674674972891806,
      "epoch": 0.007385,
      "grad_norm": 0.1276444047689438,
      "learning_rate": 2e-05,
      "loss": 0.18860353231430055,
      "mean_token_accuracy": 0.9473716080188751,
      "num_tokens": 118865016.0,
      "step": 7385
    },
    {
      "aux_loss": 8.079969835281371,
      "entropy": 0.15172515511512757,
      "epoch": 0.00739,
      "grad_norm": 0.15022414922714233,
      "learning_rate": 2e-05,
      "loss": 0.17702598571777345,
      "mean_token_accuracy": 0.9499790847301484,
      "num_tokens": 118944694.0,
      "step": 7390
    },
    {
      "aux_loss": 8.066068696975709,
      "entropy": 0.16314495652914046,
      "epoch": 0.007395,
      "grad_norm": 0.1417473405599594,
      "learning_rate": 2e-05,
      "loss": 0.20094895362854004,
      "mean_token_accuracy": 0.9441868036985397,
      "num_tokens": 119022092.0,
      "step": 7395
    },
    {
      "aux_loss": 8.052758121490479,
      "entropy": 0.1985016480088234,
      "epoch": 0.0074,
      "grad_norm": 0.13020913302898407,
      "learning_rate": 2e-05,
      "loss": 0.2348928451538086,
      "mean_token_accuracy": 0.9354265242815017,
      "num_tokens": 119103277.0,
      "step": 7400
    },
    {
      "aux_loss": 8.056073951721192,
      "entropy": 0.18207105696201326,
      "epoch": 0.007405,
      "grad_norm": 0.12403420358896255,
      "learning_rate": 2e-05,
      "loss": 0.22015740871429443,
      "mean_token_accuracy": 0.938755315542221,
      "num_tokens": 119184370.0,
      "step": 7405
    },
    {
      "aux_loss": 8.05927095413208,
      "entropy": 0.17326905950903893,
      "epoch": 0.00741,
      "grad_norm": 0.12449563294649124,
      "learning_rate": 2e-05,
      "loss": 0.2053593158721924,
      "mean_token_accuracy": 0.9431898385286331,
      "num_tokens": 119265148.0,
      "step": 7410
    },
    {
      "aux_loss": 8.068750143051147,
      "entropy": 0.17745597884058953,
      "epoch": 0.007415,
      "grad_norm": 0.13368143141269684,
      "learning_rate": 2e-05,
      "loss": 0.20241856575012207,
      "mean_token_accuracy": 0.9431264072656631,
      "num_tokens": 119346114.0,
      "step": 7415
    },
    {
      "aux_loss": 8.076086902618409,
      "entropy": 0.15370094850659372,
      "epoch": 0.00742,
      "grad_norm": 0.13952258229255676,
      "learning_rate": 2e-05,
      "loss": 0.18110870122909545,
      "mean_token_accuracy": 0.9495438784360886,
      "num_tokens": 119426614.0,
      "step": 7420
    },
    {
      "aux_loss": 8.067528915405273,
      "entropy": 0.15559654757380487,
      "epoch": 0.007425,
      "grad_norm": 0.14000247418880463,
      "learning_rate": 2e-05,
      "loss": 0.18898707628250122,
      "mean_token_accuracy": 0.9487629324197769,
      "num_tokens": 119506486.0,
      "step": 7425
    },
    {
      "aux_loss": 8.05845079421997,
      "entropy": 0.1831148162484169,
      "epoch": 0.00743,
      "grad_norm": 0.13033998012542725,
      "learning_rate": 2e-05,
      "loss": 0.22222542762756348,
      "mean_token_accuracy": 0.9379464477300644,
      "num_tokens": 119587735.0,
      "step": 7430
    },
    {
      "aux_loss": 8.05538592338562,
      "entropy": 0.2005470596253872,
      "epoch": 0.007435,
      "grad_norm": 0.14411033689975739,
      "learning_rate": 2e-05,
      "loss": 0.2414017915725708,
      "mean_token_accuracy": 0.9326531320810318,
      "num_tokens": 119668293.0,
      "step": 7435
    },
    {
      "aux_loss": 8.058018922805786,
      "entropy": 0.1771654523909092,
      "epoch": 0.00744,
      "grad_norm": 0.12871888279914856,
      "learning_rate": 2e-05,
      "loss": 0.20639445781707763,
      "mean_token_accuracy": 0.9422460258007049,
      "num_tokens": 119749331.0,
      "step": 7440
    },
    {
      "aux_loss": 8.064875555038451,
      "entropy": 0.17126045003533363,
      "epoch": 0.007445,
      "grad_norm": 0.13269393146038055,
      "learning_rate": 2e-05,
      "loss": 0.1968636155128479,
      "mean_token_accuracy": 0.9442877054214478,
      "num_tokens": 119829926.0,
      "step": 7445
    },
    {
      "aux_loss": 8.0713294506073,
      "entropy": 0.15380880497395993,
      "epoch": 0.00745,
      "grad_norm": 0.13929136097431183,
      "learning_rate": 2e-05,
      "loss": 0.18095406293869018,
      "mean_token_accuracy": 0.9498785108327865,
      "num_tokens": 119910651.0,
      "step": 7450
    },
    {
      "aux_loss": 8.074373769760133,
      "entropy": 0.16200978085398673,
      "epoch": 0.007455,
      "grad_norm": 0.12612785398960114,
      "learning_rate": 2e-05,
      "loss": 0.1933169960975647,
      "mean_token_accuracy": 0.9484478265047074,
      "num_tokens": 119988727.0,
      "step": 7455
    },
    {
      "aux_loss": 8.051265430450439,
      "entropy": 0.17319760657846928,
      "epoch": 0.00746,
      "grad_norm": 0.13716913759708405,
      "learning_rate": 2e-05,
      "loss": 0.21310434341430665,
      "mean_token_accuracy": 0.9411862343549728,
      "num_tokens": 120070092.0,
      "step": 7460
    },
    {
      "aux_loss": 8.053029489517211,
      "entropy": 0.19725120663642884,
      "epoch": 0.007465,
      "grad_norm": 0.13327398896217346,
      "learning_rate": 2e-05,
      "loss": 0.2380702018737793,
      "mean_token_accuracy": 0.9335362702608109,
      "num_tokens": 120151478.0,
      "step": 7465
    },
    {
      "aux_loss": 8.057619667053222,
      "entropy": 0.17242650762200357,
      "epoch": 0.00747,
      "grad_norm": 0.13533847033977509,
      "learning_rate": 2e-05,
      "loss": 0.20545673370361328,
      "mean_token_accuracy": 0.9422155261039734,
      "num_tokens": 120232252.0,
      "step": 7470
    },
    {
      "aux_loss": 8.064508438110352,
      "entropy": 0.17887631431221962,
      "epoch": 0.007475,
      "grad_norm": 0.13431933522224426,
      "learning_rate": 2e-05,
      "loss": 0.20763728618621827,
      "mean_token_accuracy": 0.941964665055275,
      "num_tokens": 120312653.0,
      "step": 7475
    },
    {
      "aux_loss": 8.071344709396362,
      "entropy": 0.16794106587767602,
      "epoch": 0.00748,
      "grad_norm": 0.15036065876483917,
      "learning_rate": 2e-05,
      "loss": 0.19158920049667358,
      "mean_token_accuracy": 0.9471013605594635,
      "num_tokens": 120392736.0,
      "step": 7480
    },
    {
      "aux_loss": 8.079354572296143,
      "entropy": 0.1548157688230276,
      "epoch": 0.007485,
      "grad_norm": 0.12153560668230057,
      "learning_rate": 2e-05,
      "loss": 0.18247909545898439,
      "mean_token_accuracy": 0.9490600466728211,
      "num_tokens": 120473263.0,
      "step": 7485
    },
    {
      "aux_loss": 8.051816368103028,
      "entropy": 0.176940343901515,
      "epoch": 0.00749,
      "grad_norm": 0.14291629195213318,
      "learning_rate": 2e-05,
      "loss": 0.21626551151275636,
      "mean_token_accuracy": 0.9405858308076859,
      "num_tokens": 120554642.0,
      "step": 7490
    },
    {
      "aux_loss": 8.054406261444091,
      "entropy": 0.1862452931702137,
      "epoch": 0.007495,
      "grad_norm": 0.1422058492898941,
      "learning_rate": 2e-05,
      "loss": 0.23038835525512696,
      "mean_token_accuracy": 0.9352040529251099,
      "num_tokens": 120635636.0,
      "step": 7495
    },
    {
      "aux_loss": 8.056235361099244,
      "entropy": 0.1768251147121191,
      "epoch": 0.0075,
      "grad_norm": 0.12467633932828903,
      "learning_rate": 2e-05,
      "loss": 0.21371710300445557,
      "mean_token_accuracy": 0.940923273563385,
      "num_tokens": 120716640.0,
      "step": 7500
    },
    {
      "aux_loss": 8.064361953735352,
      "entropy": 0.17170063331723212,
      "epoch": 0.007505,
      "grad_norm": 0.1362258344888687,
      "learning_rate": 2e-05,
      "loss": 0.19958794116973877,
      "mean_token_accuracy": 0.9444036543369293,
      "num_tokens": 120797337.0,
      "step": 7505
    },
    {
      "aux_loss": 8.071210527420044,
      "entropy": 0.16280821934342385,
      "epoch": 0.00751,
      "grad_norm": 0.1373574137687683,
      "learning_rate": 2e-05,
      "loss": 0.1837894558906555,
      "mean_token_accuracy": 0.9482373565435409,
      "num_tokens": 120877833.0,
      "step": 7510
    },
    {
      "aux_loss": 8.082080364227295,
      "entropy": 0.14672463983297349,
      "epoch": 0.007515,
      "grad_norm": 0.14578260481357574,
      "learning_rate": 2e-05,
      "loss": 0.16725674867630005,
      "mean_token_accuracy": 0.9539514601230621,
      "num_tokens": 120956989.0,
      "step": 7515
    },
    {
      "aux_loss": 8.047899627685547,
      "entropy": 0.1632910158485174,
      "epoch": 0.00752,
      "grad_norm": 0.14295019209384918,
      "learning_rate": 2e-05,
      "loss": 0.20444486141204835,
      "mean_token_accuracy": 0.9441313356161117,
      "num_tokens": 121038244.0,
      "step": 7520
    },
    {
      "aux_loss": 8.051692342758178,
      "entropy": 0.18788138255476952,
      "epoch": 0.007525,
      "grad_norm": 0.1524629294872284,
      "learning_rate": 2e-05,
      "loss": 0.23267843723297119,
      "mean_token_accuracy": 0.9353946089744568,
      "num_tokens": 121119392.0,
      "step": 7525
    },
    {
      "aux_loss": 8.056363153457642,
      "entropy": 0.18160314559936525,
      "epoch": 0.00753,
      "grad_norm": 0.13505490124225616,
      "learning_rate": 2e-05,
      "loss": 0.21701672077178955,
      "mean_token_accuracy": 0.9390973597764969,
      "num_tokens": 121200362.0,
      "step": 7530
    },
    {
      "aux_loss": 8.060441398620606,
      "entropy": 0.16757689155638217,
      "epoch": 0.007535,
      "grad_norm": 0.1300678253173828,
      "learning_rate": 2e-05,
      "loss": 0.19440244436264037,
      "mean_token_accuracy": 0.9457934498786926,
      "num_tokens": 121281028.0,
      "step": 7535
    },
    {
      "aux_loss": 8.068178510665893,
      "entropy": 0.17026686370372773,
      "epoch": 0.00754,
      "grad_norm": 0.13784855604171753,
      "learning_rate": 2e-05,
      "loss": 0.199432373046875,
      "mean_token_accuracy": 0.9441833674907685,
      "num_tokens": 121361445.0,
      "step": 7540
    },
    {
      "aux_loss": 8.075921297073364,
      "entropy": 0.14733664877712727,
      "epoch": 0.007545,
      "grad_norm": 0.14461226761341095,
      "learning_rate": 2e-05,
      "loss": 0.1730791449546814,
      "mean_token_accuracy": 0.9519186556339264,
      "num_tokens": 121441239.0,
      "step": 7545
    },
    {
      "aux_loss": 8.067001104354858,
      "entropy": 0.1626121636480093,
      "epoch": 0.00755,
      "grad_norm": 0.14616137742996216,
      "learning_rate": 2e-05,
      "loss": 0.20098023414611815,
      "mean_token_accuracy": 0.944268774986267,
      "num_tokens": 121518539.0,
      "step": 7550
    },
    {
      "aux_loss": 8.054377555847168,
      "entropy": 0.18031487762928008,
      "epoch": 0.007555,
      "grad_norm": 0.14393800497055054,
      "learning_rate": 2e-05,
      "loss": 0.22135207653045655,
      "mean_token_accuracy": 0.9373565286397934,
      "num_tokens": 121599628.0,
      "step": 7555
    },
    {
      "aux_loss": 8.05363154411316,
      "entropy": 0.17373102754354477,
      "epoch": 0.00756,
      "grad_norm": 0.13420702517032623,
      "learning_rate": 2e-05,
      "loss": 0.2052135467529297,
      "mean_token_accuracy": 0.9428716838359833,
      "num_tokens": 121680685.0,
      "step": 7560
    },
    {
      "aux_loss": 8.063855361938476,
      "entropy": 0.16535890772938727,
      "epoch": 0.007565,
      "grad_norm": 0.13558802008628845,
      "learning_rate": 2e-05,
      "loss": 0.19632102251052858,
      "mean_token_accuracy": 0.9453599065542221,
      "num_tokens": 121761568.0,
      "step": 7565
    },
    {
      "aux_loss": 8.070244646072387,
      "entropy": 0.16432089656591414,
      "epoch": 0.00757,
      "grad_norm": 0.13941632211208344,
      "learning_rate": 2e-05,
      "loss": 0.19304463863372803,
      "mean_token_accuracy": 0.9463385283946991,
      "num_tokens": 121842146.0,
      "step": 7570
    },
    {
      "aux_loss": 8.07728853225708,
      "entropy": 0.14477277621626855,
      "epoch": 0.007575,
      "grad_norm": 0.1449158489704132,
      "learning_rate": 2e-05,
      "loss": 0.16983110904693605,
      "mean_token_accuracy": 0.9528191447257995,
      "num_tokens": 121922136.0,
      "step": 7575
    },
    {
      "aux_loss": 8.065555381774903,
      "entropy": 0.15253821276128293,
      "epoch": 0.00758,
      "grad_norm": 0.12851905822753906,
      "learning_rate": 2e-05,
      "loss": 0.18775851726531984,
      "mean_token_accuracy": 0.9485505431890487,
      "num_tokens": 122002609.0,
      "step": 7580
    },
    {
      "aux_loss": 8.053164148330689,
      "entropy": 0.18466558493673801,
      "epoch": 0.007585,
      "grad_norm": 0.1372760385274887,
      "learning_rate": 2e-05,
      "loss": 0.22767891883850097,
      "mean_token_accuracy": 0.9371002525091171,
      "num_tokens": 122083285.0,
      "step": 7585
    },
    {
      "aux_loss": 8.055616188049317,
      "entropy": 0.1753864422440529,
      "epoch": 0.00759,
      "grad_norm": 0.1325058937072754,
      "learning_rate": 2e-05,
      "loss": 0.2120638370513916,
      "mean_token_accuracy": 0.940838199853897,
      "num_tokens": 122164375.0,
      "step": 7590
    },
    {
      "aux_loss": 8.060683679580688,
      "entropy": 0.169303497672081,
      "epoch": 0.007595,
      "grad_norm": 0.1259465217590332,
      "learning_rate": 2e-05,
      "loss": 0.19654493331909179,
      "mean_token_accuracy": 0.9455428004264832,
      "num_tokens": 122244917.0,
      "step": 7595
    },
    {
      "aux_loss": 8.065549325942992,
      "entropy": 0.15612652525305748,
      "epoch": 0.0076,
      "grad_norm": 0.12974722683429718,
      "learning_rate": 2e-05,
      "loss": 0.18145638704299927,
      "mean_token_accuracy": 0.9491685390472412,
      "num_tokens": 122325668.0,
      "step": 7600
    },
    {
      "aux_loss": 8.072974014282227,
      "entropy": 0.15129366442561148,
      "epoch": 0.007605,
      "grad_norm": 0.13044056296348572,
      "learning_rate": 2e-05,
      "loss": 0.18290570974349976,
      "mean_token_accuracy": 0.9489581197500229,
      "num_tokens": 122405878.0,
      "step": 7605
    },
    {
      "aux_loss": 8.066445779800414,
      "entropy": 0.16404522247612477,
      "epoch": 0.00761,
      "grad_norm": 0.12882989645004272,
      "learning_rate": 2e-05,
      "loss": 0.1982866883277893,
      "mean_token_accuracy": 0.9456588923931122,
      "num_tokens": 122484204.0,
      "step": 7610
    },
    {
      "aux_loss": 8.056326627731323,
      "entropy": 0.21864307895302773,
      "epoch": 0.007615,
      "grad_norm": 0.13727900385856628,
      "learning_rate": 2e-05,
      "loss": 0.2567129373550415,
      "mean_token_accuracy": 0.9283473819494248,
      "num_tokens": 122565009.0,
      "step": 7615
    },
    {
      "aux_loss": 8.05384349822998,
      "entropy": 0.1832655794918537,
      "epoch": 0.00762,
      "grad_norm": 0.13248056173324585,
      "learning_rate": 2e-05,
      "loss": 0.21849074363708496,
      "mean_token_accuracy": 0.9386864900588989,
      "num_tokens": 122645726.0,
      "step": 7620
    },
    {
      "aux_loss": 8.05781512260437,
      "entropy": 0.16786095947027208,
      "epoch": 0.007625,
      "grad_norm": 0.12765242159366608,
      "learning_rate": 2e-05,
      "loss": 0.19740880727767945,
      "mean_token_accuracy": 0.9447900116443634,
      "num_tokens": 122726776.0,
      "step": 7625
    },
    {
      "aux_loss": 8.065768098831176,
      "entropy": 0.1635085940361023,
      "epoch": 0.00763,
      "grad_norm": 0.1307392120361328,
      "learning_rate": 2e-05,
      "loss": 0.19003418684005738,
      "mean_token_accuracy": 0.9469250172376633,
      "num_tokens": 122807539.0,
      "step": 7630
    },
    {
      "aux_loss": 8.073894548416138,
      "entropy": 0.1470636773854494,
      "epoch": 0.007635,
      "grad_norm": 0.14527620375156403,
      "learning_rate": 2e-05,
      "loss": 0.17303662300109862,
      "mean_token_accuracy": 0.9513357639312744,
      "num_tokens": 122888085.0,
      "step": 7635
    },
    {
      "aux_loss": 8.072463321685792,
      "entropy": 0.14796018600463867,
      "epoch": 0.00764,
      "grad_norm": 0.15510867536067963,
      "learning_rate": 2e-05,
      "loss": 0.1794682741165161,
      "mean_token_accuracy": 0.9507189899682998,
      "num_tokens": 122966503.0,
      "step": 7640
    },
    {
      "aux_loss": 8.052925539016723,
      "entropy": 0.16258029043674468,
      "epoch": 0.007645,
      "grad_norm": 0.13088901340961456,
      "learning_rate": 2e-05,
      "loss": 0.2010512113571167,
      "mean_token_accuracy": 0.9435966789722443,
      "num_tokens": 123047905.0,
      "step": 7645
    },
    {
      "aux_loss": 8.054289102554321,
      "entropy": 0.18538018241524695,
      "epoch": 0.00765,
      "grad_norm": 0.15072676539421082,
      "learning_rate": 2e-05,
      "loss": 0.22367489337921143,
      "mean_token_accuracy": 0.9375688791275024,
      "num_tokens": 123129027.0,
      "step": 7650
    },
    {
      "aux_loss": 8.057654905319215,
      "entropy": 0.17644346468150615,
      "epoch": 0.007655,
      "grad_norm": 0.12897905707359314,
      "learning_rate": 2e-05,
      "loss": 0.21030006408691407,
      "mean_token_accuracy": 0.9423573017120361,
      "num_tokens": 123209763.0,
      "step": 7655
    },
    {
      "aux_loss": 8.067772102355956,
      "entropy": 0.17088019102811813,
      "epoch": 0.00766,
      "grad_norm": 0.1275871992111206,
      "learning_rate": 2e-05,
      "loss": 0.19940216541290284,
      "mean_token_accuracy": 0.944449645280838,
      "num_tokens": 123290106.0,
      "step": 7660
    },
    {
      "aux_loss": 8.07044243812561,
      "entropy": 0.15822128541767597,
      "epoch": 0.007665,
      "grad_norm": 0.1283460557460785,
      "learning_rate": 2e-05,
      "loss": 0.18370022773742675,
      "mean_token_accuracy": 0.9491430759429932,
      "num_tokens": 123370909.0,
      "step": 7665
    },
    {
      "aux_loss": 8.078686094284057,
      "entropy": 0.14930059984326363,
      "epoch": 0.00767,
      "grad_norm": 0.12206239998340607,
      "learning_rate": 2e-05,
      "loss": 0.17540135383605956,
      "mean_token_accuracy": 0.9517033100128174,
      "num_tokens": 123451278.0,
      "step": 7670
    },
    {
      "aux_loss": 8.051547527313232,
      "entropy": 0.17841746509075165,
      "epoch": 0.007675,
      "grad_norm": 0.12721508741378784,
      "learning_rate": 2e-05,
      "loss": 0.21916866302490234,
      "mean_token_accuracy": 0.9399241238832474,
      "num_tokens": 123532686.0,
      "step": 7675
    },
    {
      "aux_loss": 8.054313230514527,
      "entropy": 0.1667678453028202,
      "epoch": 0.00768,
      "grad_norm": 0.1395113617181778,
      "learning_rate": 2e-05,
      "loss": 0.20610229969024657,
      "mean_token_accuracy": 0.9425065070390701,
      "num_tokens": 123613871.0,
      "step": 7680
    },
    {
      "aux_loss": 8.05824065208435,
      "entropy": 0.1794706739485264,
      "epoch": 0.007685,
      "grad_norm": 0.13812685012817383,
      "learning_rate": 2e-05,
      "loss": 0.21608576774597169,
      "mean_token_accuracy": 0.9394313514232635,
      "num_tokens": 123694867.0,
      "step": 7685
    },
    {
      "aux_loss": 8.061620092391967,
      "entropy": 0.15949661284685135,
      "epoch": 0.00769,
      "grad_norm": 0.1312873512506485,
      "learning_rate": 2e-05,
      "loss": 0.18671815395355223,
      "mean_token_accuracy": 0.9473142325878143,
      "num_tokens": 123775333.0,
      "step": 7690
    },
    {
      "aux_loss": 8.069132089614868,
      "entropy": 0.16831732913851738,
      "epoch": 0.007695,
      "grad_norm": 0.1347435712814331,
      "learning_rate": 2e-05,
      "loss": 0.19773918390274048,
      "mean_token_accuracy": 0.9447491824626922,
      "num_tokens": 123855686.0,
      "step": 7695
    },
    {
      "aux_loss": 8.081768035888672,
      "entropy": 0.1415973886847496,
      "epoch": 0.0077,
      "grad_norm": 0.15629427134990692,
      "learning_rate": 2e-05,
      "loss": 0.16426509618759155,
      "mean_token_accuracy": 0.9545311212539673,
      "num_tokens": 123933075.0,
      "step": 7700
    },
    {
      "aux_loss": 8.051624774932861,
      "entropy": 0.17327583506703376,
      "epoch": 0.007705,
      "grad_norm": 0.12539736926555634,
      "learning_rate": 2e-05,
      "loss": 0.21044507026672363,
      "mean_token_accuracy": 0.9425005942583085,
      "num_tokens": 124014183.0,
      "step": 7705
    },
    {
      "aux_loss": 8.053359985351562,
      "entropy": 0.19304492510855198,
      "epoch": 0.00771,
      "grad_norm": 0.14401879906654358,
      "learning_rate": 2e-05,
      "loss": 0.2403738260269165,
      "mean_token_accuracy": 0.9334207534790039,
      "num_tokens": 124095638.0,
      "step": 7710
    },
    {
      "aux_loss": 8.053791809082032,
      "entropy": 0.1789645031094551,
      "epoch": 0.007715,
      "grad_norm": 0.12512809038162231,
      "learning_rate": 2e-05,
      "loss": 0.2131068229675293,
      "mean_token_accuracy": 0.9400732308626175,
      "num_tokens": 124176537.0,
      "step": 7715
    },
    {
      "aux_loss": 8.059452676773072,
      "entropy": 0.173047836124897,
      "epoch": 0.00772,
      "grad_norm": 0.12605106830596924,
      "learning_rate": 2e-05,
      "loss": 0.20206212997436523,
      "mean_token_accuracy": 0.9426600307226181,
      "num_tokens": 124257414.0,
      "step": 7720
    },
    {
      "aux_loss": 8.065120697021484,
      "entropy": 0.16913856267929078,
      "epoch": 0.007725,
      "grad_norm": 0.1334429383277893,
      "learning_rate": 2e-05,
      "loss": 0.19155629873275756,
      "mean_token_accuracy": 0.946245226264,
      "num_tokens": 124337538.0,
      "step": 7725
    },
    {
      "aux_loss": 8.078563451766968,
      "entropy": 0.15281990244984628,
      "epoch": 0.00773,
      "grad_norm": 0.16049043834209442,
      "learning_rate": 2e-05,
      "loss": 0.17544158697128295,
      "mean_token_accuracy": 0.9512608706951141,
      "num_tokens": 124417699.0,
      "step": 7730
    },
    {
      "aux_loss": 8.057663822174073,
      "entropy": 0.15443732179701328,
      "epoch": 0.007735,
      "grad_norm": 0.1292727142572403,
      "learning_rate": 2e-05,
      "loss": 0.19397801160812378,
      "mean_token_accuracy": 0.9491463840007782,
      "num_tokens": 124494889.0,
      "step": 7735
    },
    {
      "aux_loss": 8.052012920379639,
      "entropy": 0.17123932167887687,
      "epoch": 0.00774,
      "grad_norm": 0.1302178055047989,
      "learning_rate": 2e-05,
      "loss": 0.2160214900970459,
      "mean_token_accuracy": 0.9405649453401566,
      "num_tokens": 124576113.0,
      "step": 7740
    },
    {
      "aux_loss": 8.058313179016114,
      "entropy": 0.1709307014942169,
      "epoch": 0.007745,
      "grad_norm": 0.13111530244350433,
      "learning_rate": 2e-05,
      "loss": 0.20689597129821777,
      "mean_token_accuracy": 0.9426312148571014,
      "num_tokens": 124656747.0,
      "step": 7745
    },
    {
      "aux_loss": 8.057962608337402,
      "entropy": 0.17740309238433838,
      "epoch": 0.00775,
      "grad_norm": 0.1303430199623108,
      "learning_rate": 2e-05,
      "loss": 0.20656147003173828,
      "mean_token_accuracy": 0.9424556732177735,
      "num_tokens": 124737374.0,
      "step": 7750
    },
    {
      "aux_loss": 8.064849615097046,
      "entropy": 0.15833495184779167,
      "epoch": 0.007755,
      "grad_norm": 0.124373659491539,
      "learning_rate": 2e-05,
      "loss": 0.18394724130630494,
      "mean_token_accuracy": 0.9486096560955047,
      "num_tokens": 124817976.0,
      "step": 7755
    },
    {
      "aux_loss": 8.072395133972169,
      "entropy": 0.14352682642638684,
      "epoch": 0.00776,
      "grad_norm": 0.13774539530277252,
      "learning_rate": 2e-05,
      "loss": 0.1703740954399109,
      "mean_token_accuracy": 0.9524211406707763,
      "num_tokens": 124897912.0,
      "step": 7760
    },
    {
      "aux_loss": 8.066678285598755,
      "entropy": 0.16153050549328327,
      "epoch": 0.007765,
      "grad_norm": 0.13381630182266235,
      "learning_rate": 2e-05,
      "loss": 0.19915086030960083,
      "mean_token_accuracy": 0.9459080785512924,
      "num_tokens": 124978157.0,
      "step": 7765
    },
    {
      "aux_loss": 8.053351259231567,
      "entropy": 0.17324205189943315,
      "epoch": 0.00777,
      "grad_norm": 0.1400909423828125,
      "learning_rate": 2e-05,
      "loss": 0.215484094619751,
      "mean_token_accuracy": 0.9400692701339721,
      "num_tokens": 125058705.0,
      "step": 7770
    },
    {
      "aux_loss": 8.05493893623352,
      "entropy": 0.1917740911245346,
      "epoch": 0.007775,
      "grad_norm": 0.12173166126012802,
      "learning_rate": 2e-05,
      "loss": 0.22488539218902587,
      "mean_token_accuracy": 0.9369278907775879,
      "num_tokens": 125139991.0,
      "step": 7775
    },
    {
      "aux_loss": 8.058575105667114,
      "entropy": 0.1666221559047699,
      "epoch": 0.00778,
      "grad_norm": 0.13220971822738647,
      "learning_rate": 2e-05,
      "loss": 0.196097469329834,
      "mean_token_accuracy": 0.9448536068201066,
      "num_tokens": 125220978.0,
      "step": 7780
    },
    {
      "aux_loss": 8.065988111495972,
      "entropy": 0.16583029180765152,
      "epoch": 0.007785,
      "grad_norm": 0.1359696388244629,
      "learning_rate": 2e-05,
      "loss": 0.1912519931793213,
      "mean_token_accuracy": 0.946163147687912,
      "num_tokens": 125302057.0,
      "step": 7785
    },
    {
      "aux_loss": 8.07320785522461,
      "entropy": 0.15409406274557114,
      "epoch": 0.00779,
      "grad_norm": 0.1364138424396515,
      "learning_rate": 2e-05,
      "loss": 0.1849233865737915,
      "mean_token_accuracy": 0.9479338139295578,
      "num_tokens": 125382267.0,
      "step": 7790
    },
    {
      "aux_loss": 8.069149160385132,
      "entropy": 0.14468498453497886,
      "epoch": 0.007795,
      "grad_norm": 0.13282909989356995,
      "learning_rate": 2e-05,
      "loss": 0.1789170742034912,
      "mean_token_accuracy": 0.9510094195604324,
      "num_tokens": 125460965.0,
      "step": 7795
    },
    {
      "aux_loss": 8.052641677856446,
      "entropy": 0.1654026973992586,
      "epoch": 0.0078,
      "grad_norm": 0.13480037450790405,
      "learning_rate": 2e-05,
      "loss": 0.2051823377609253,
      "mean_token_accuracy": 0.9445670366287231,
      "num_tokens": 125542449.0,
      "step": 7800
    },
    {
      "aux_loss": 8.055175399780273,
      "entropy": 0.18958356529474257,
      "epoch": 0.007805,
      "grad_norm": 0.13398399949073792,
      "learning_rate": 2e-05,
      "loss": 0.22677903175354003,
      "mean_token_accuracy": 0.9368277102708816,
      "num_tokens": 125623439.0,
      "step": 7805
    },
    {
      "aux_loss": 8.056657552719116,
      "entropy": 0.18027772307395934,
      "epoch": 0.00781,
      "grad_norm": 0.12332537770271301,
      "learning_rate": 2e-05,
      "loss": 0.21051394939422607,
      "mean_token_accuracy": 0.9407501101493836,
      "num_tokens": 125704018.0,
      "step": 7810
    },
    {
      "aux_loss": 8.064786958694459,
      "entropy": 0.16789413020014762,
      "epoch": 0.007815,
      "grad_norm": 0.12678587436676025,
      "learning_rate": 2e-05,
      "loss": 0.19255117177963257,
      "mean_token_accuracy": 0.9461271107196808,
      "num_tokens": 125784608.0,
      "step": 7815
    },
    {
      "aux_loss": 8.070134830474853,
      "entropy": 0.1539592742919922,
      "epoch": 0.00782,
      "grad_norm": 0.1389128416776657,
      "learning_rate": 2e-05,
      "loss": 0.17781422138214112,
      "mean_token_accuracy": 0.9500072836875916,
      "num_tokens": 125865019.0,
      "step": 7820
    },
    {
      "aux_loss": 8.074520206451416,
      "entropy": 0.14029172472655774,
      "epoch": 0.007825,
      "grad_norm": 0.1327447146177292,
      "learning_rate": 2e-05,
      "loss": 0.17502390146255492,
      "mean_token_accuracy": 0.9525757849216461,
      "num_tokens": 125943171.0,
      "step": 7825
    },
    {
      "aux_loss": 8.052855587005615,
      "entropy": 0.1754423726350069,
      "epoch": 0.00783,
      "grad_norm": 0.1373121738433838,
      "learning_rate": 2e-05,
      "loss": 0.22010231018066406,
      "mean_token_accuracy": 0.9398065447807312,
      "num_tokens": 126024596.0,
      "step": 7830
    },
    {
      "aux_loss": 8.049565553665161,
      "entropy": 0.1772006146609783,
      "epoch": 0.007835,
      "grad_norm": 0.12699949741363525,
      "learning_rate": 2e-05,
      "loss": 0.21540539264678954,
      "mean_token_accuracy": 0.9408989787101746,
      "num_tokens": 126105305.0,
      "step": 7835
    },
    {
      "aux_loss": 8.05538649559021,
      "entropy": 0.1763097494840622,
      "epoch": 0.00784,
      "grad_norm": 0.1293422281742096,
      "learning_rate": 2e-05,
      "loss": 0.20866866111755372,
      "mean_token_accuracy": 0.9412191480398178,
      "num_tokens": 126185985.0,
      "step": 7840
    },
    {
      "aux_loss": 8.062546110153198,
      "entropy": 0.17701613903045654,
      "epoch": 0.007845,
      "grad_norm": 0.12797299027442932,
      "learning_rate": 2e-05,
      "loss": 0.20301802158355714,
      "mean_token_accuracy": 0.9436441868543625,
      "num_tokens": 126266332.0,
      "step": 7845
    },
    {
      "aux_loss": 8.068743419647216,
      "entropy": 0.15298430100083352,
      "epoch": 0.00785,
      "grad_norm": 0.13106869161128998,
      "learning_rate": 2e-05,
      "loss": 0.17925429344177246,
      "mean_token_accuracy": 0.9499890148639679,
      "num_tokens": 126346300.0,
      "step": 7850
    },
    {
      "aux_loss": 8.075567436218261,
      "entropy": 0.14626502506434919,
      "epoch": 0.007855,
      "grad_norm": 0.12238899618387222,
      "learning_rate": 2e-05,
      "loss": 0.17421101331710814,
      "mean_token_accuracy": 0.9522601574659347,
      "num_tokens": 126425009.0,
      "step": 7855
    },
    {
      "aux_loss": 8.054703283309937,
      "entropy": 0.1732068732380867,
      "epoch": 0.00786,
      "grad_norm": 0.13118509948253632,
      "learning_rate": 2e-05,
      "loss": 0.21082921028137208,
      "mean_token_accuracy": 0.9412829577922821,
      "num_tokens": 126506479.0,
      "step": 7860
    },
    {
      "aux_loss": 8.0575541973114,
      "entropy": 0.183889240026474,
      "epoch": 0.007865,
      "grad_norm": 0.16069471836090088,
      "learning_rate": 2e-05,
      "loss": 0.2293248653411865,
      "mean_token_accuracy": 0.9364004522562027,
      "num_tokens": 126587524.0,
      "step": 7865
    },
    {
      "aux_loss": 8.056588840484618,
      "entropy": 0.17077396139502526,
      "epoch": 0.00787,
      "grad_norm": 0.13866876065731049,
      "learning_rate": 2e-05,
      "loss": 0.20553219318389893,
      "mean_token_accuracy": 0.9434812992811203,
      "num_tokens": 126668412.0,
      "step": 7870
    },
    {
      "aux_loss": 8.06143045425415,
      "entropy": 0.17998394519090652,
      "epoch": 0.007875,
      "grad_norm": 0.13197578489780426,
      "learning_rate": 2e-05,
      "loss": 0.20621256828308104,
      "mean_token_accuracy": 0.9409545570611954,
      "num_tokens": 126749257.0,
      "step": 7875
    },
    {
      "aux_loss": 8.06725378036499,
      "entropy": 0.15875052027404307,
      "epoch": 0.00788,
      "grad_norm": 0.1366870254278183,
      "learning_rate": 2e-05,
      "loss": 0.18191620111465454,
      "mean_token_accuracy": 0.9497178018093109,
      "num_tokens": 126829294.0,
      "step": 7880
    },
    {
      "aux_loss": 8.076392650604248,
      "entropy": 0.14412251934409143,
      "epoch": 0.007885,
      "grad_norm": 0.1481265127658844,
      "learning_rate": 2e-05,
      "loss": 0.1692441701889038,
      "mean_token_accuracy": 0.9525374323129654,
      "num_tokens": 126909412.0,
      "step": 7885
    },
    {
      "aux_loss": 8.052158451080322,
      "entropy": 0.16255763173103333,
      "epoch": 0.00789,
      "grad_norm": 0.1406784951686859,
      "learning_rate": 2e-05,
      "loss": 0.20298330783843993,
      "mean_token_accuracy": 0.9450664490461349,
      "num_tokens": 126990584.0,
      "step": 7890
    },
    {
      "aux_loss": 8.050649499893188,
      "entropy": 0.18288838416337966,
      "epoch": 0.007895,
      "grad_norm": 0.1396569311618805,
      "learning_rate": 2e-05,
      "loss": 0.2286553144454956,
      "mean_token_accuracy": 0.9361393839120865,
      "num_tokens": 127071982.0,
      "step": 7895
    },
    {
      "aux_loss": 8.05613989830017,
      "entropy": 0.1704040512442589,
      "epoch": 0.0079,
      "grad_norm": 0.13003630936145782,
      "learning_rate": 2e-05,
      "loss": 0.202703857421875,
      "mean_token_accuracy": 0.9429990768432617,
      "num_tokens": 127152972.0,
      "step": 7900
    },
    {
      "aux_loss": 8.060097646713256,
      "entropy": 0.16928130351006984,
      "epoch": 0.007905,
      "grad_norm": 0.12603473663330078,
      "learning_rate": 2e-05,
      "loss": 0.19504098892211913,
      "mean_token_accuracy": 0.9456968605518341,
      "num_tokens": 127233770.0,
      "step": 7905
    },
    {
      "aux_loss": 8.06521725654602,
      "entropy": 0.15720516219735145,
      "epoch": 0.00791,
      "grad_norm": 0.13574708998203278,
      "learning_rate": 2e-05,
      "loss": 0.18702263832092286,
      "mean_token_accuracy": 0.9477150440216064,
      "num_tokens": 127313889.0,
      "step": 7910
    },
    {
      "aux_loss": 8.076163482666015,
      "entropy": 0.1441133599728346,
      "epoch": 0.007915,
      "grad_norm": 0.1436004936695099,
      "learning_rate": 2e-05,
      "loss": 0.172634494304657,
      "mean_token_accuracy": 0.9515066981315613,
      "num_tokens": 127394272.0,
      "step": 7915
    },
    {
      "aux_loss": 8.056488227844238,
      "entropy": 0.14625754654407502,
      "epoch": 0.00792,
      "grad_norm": 0.1260545551776886,
      "learning_rate": 2e-05,
      "loss": 0.1745867133140564,
      "mean_token_accuracy": 0.9520705491304398,
      "num_tokens": 127472076.0,
      "step": 7920
    },
    {
      "aux_loss": 8.053277969360352,
      "entropy": 0.18621368780732156,
      "epoch": 0.007925,
      "grad_norm": 0.12595397233963013,
      "learning_rate": 2e-05,
      "loss": 0.22350747585296632,
      "mean_token_accuracy": 0.9376633852720261,
      "num_tokens": 127552974.0,
      "step": 7925
    },
    {
      "aux_loss": 8.052712059020996,
      "entropy": 0.1803848709911108,
      "epoch": 0.00793,
      "grad_norm": 0.1303734928369522,
      "learning_rate": 2e-05,
      "loss": 0.2149219512939453,
      "mean_token_accuracy": 0.9405826300382614,
      "num_tokens": 127633891.0,
      "step": 7930
    },
    {
      "aux_loss": 8.056089830398559,
      "entropy": 0.1527490597218275,
      "epoch": 0.007935,
      "grad_norm": 0.13994599878787994,
      "learning_rate": 2e-05,
      "loss": 0.18676354885101318,
      "mean_token_accuracy": 0.9470791161060333,
      "num_tokens": 127714652.0,
      "step": 7935
    },
    {
      "aux_loss": 8.060953140258789,
      "entropy": 0.16185988783836364,
      "epoch": 0.00794,
      "grad_norm": 0.13381613790988922,
      "learning_rate": 2e-05,
      "loss": 0.19044151306152343,
      "mean_token_accuracy": 0.9470368057489396,
      "num_tokens": 127795393.0,
      "step": 7940
    },
    {
      "aux_loss": 8.070991706848144,
      "entropy": 0.15282550193369387,
      "epoch": 0.007945,
      "grad_norm": 0.1395682543516159,
      "learning_rate": 2e-05,
      "loss": 0.1772878050804138,
      "mean_token_accuracy": 0.9517181694507599,
      "num_tokens": 127875299.0,
      "step": 7945
    },
    {
      "aux_loss": 8.065530252456664,
      "entropy": 0.149153945595026,
      "epoch": 0.00795,
      "grad_norm": 0.13513027131557465,
      "learning_rate": 2e-05,
      "loss": 0.18359854221343994,
      "mean_token_accuracy": 0.9488667905330658,
      "num_tokens": 127956214.0,
      "step": 7950
    },
    {
      "aux_loss": 8.051602363586426,
      "entropy": 0.17141111344099044,
      "epoch": 0.007955,
      "grad_norm": 0.13007114827632904,
      "learning_rate": 2e-05,
      "loss": 0.21193037033081055,
      "mean_token_accuracy": 0.941623193025589,
      "num_tokens": 128037138.0,
      "step": 7955
    },
    {
      "aux_loss": 8.055979585647583,
      "entropy": 0.16566436626017095,
      "epoch": 0.00796,
      "grad_norm": 0.12722444534301758,
      "learning_rate": 2e-05,
      "loss": 0.19971907138824463,
      "mean_token_accuracy": 0.9445472031831741,
      "num_tokens": 128118108.0,
      "step": 7960
    },
    {
      "aux_loss": 8.057784366607667,
      "entropy": 0.1565207026898861,
      "epoch": 0.007965,
      "grad_norm": 0.12743087112903595,
      "learning_rate": 2e-05,
      "loss": 0.18870077133178711,
      "mean_token_accuracy": 0.947584468126297,
      "num_tokens": 128198916.0,
      "step": 7965
    },
    {
      "aux_loss": 8.065590524673462,
      "entropy": 0.1583856925368309,
      "epoch": 0.00797,
      "grad_norm": 0.1353423297405243,
      "learning_rate": 2e-05,
      "loss": 0.18316779136657715,
      "mean_token_accuracy": 0.9478757351636886,
      "num_tokens": 128279609.0,
      "step": 7970
    },
    {
      "aux_loss": 8.072882652282715,
      "entropy": 0.14844279438257219,
      "epoch": 0.007975,
      "grad_norm": 0.14464455842971802,
      "learning_rate": 2e-05,
      "loss": 0.17363022565841674,
      "mean_token_accuracy": 0.9515461266040802,
      "num_tokens": 128359551.0,
      "step": 7975
    },
    {
      "aux_loss": 8.0706711769104,
      "entropy": 0.143844673037529,
      "epoch": 0.00798,
      "grad_norm": 0.1265324205160141,
      "learning_rate": 2e-05,
      "loss": 0.1778017044067383,
      "mean_token_accuracy": 0.9520799815654755,
      "num_tokens": 128437018.0,
      "step": 7980
    },
    {
      "aux_loss": 8.053995752334595,
      "entropy": 0.16581640467047692,
      "epoch": 0.007985,
      "grad_norm": 0.14176373183727264,
      "learning_rate": 2e-05,
      "loss": 0.21161484718322754,
      "mean_token_accuracy": 0.9414434880018234,
      "num_tokens": 128518006.0,
      "step": 7985
    },
    {
      "aux_loss": 8.051170873641968,
      "entropy": 0.17790948115289212,
      "epoch": 0.00799,
      "grad_norm": 0.1317165195941925,
      "learning_rate": 2e-05,
      "loss": 0.21526544094085692,
      "mean_token_accuracy": 0.9398356437683105,
      "num_tokens": 128598913.0,
      "step": 7990
    },
    {
      "aux_loss": 8.054621934890747,
      "entropy": 0.16488940343260766,
      "epoch": 0.007995,
      "grad_norm": 0.12735776603221893,
      "learning_rate": 2e-05,
      "loss": 0.19196200370788574,
      "mean_token_accuracy": 0.9462250411510468,
      "num_tokens": 128679788.0,
      "step": 7995
    },
    {
      "aux_loss": 8.060392284393311,
      "entropy": 0.1674186483025551,
      "epoch": 0.008,
      "grad_norm": 0.13624000549316406,
      "learning_rate": 2e-05,
      "loss": 0.1934329390525818,
      "mean_token_accuracy": 0.9455029308795929,
      "num_tokens": 128760490.0,
      "step": 8000
    },
    {
      "aux_loss": 8.070809412002564,
      "entropy": 0.1470926396548748,
      "epoch": 0.008005,
      "grad_norm": 0.13767880201339722,
      "learning_rate": 2e-05,
      "loss": 0.17155596017837524,
      "mean_token_accuracy": 0.9519717872142792,
      "num_tokens": 128840501.0,
      "step": 8005
    },
    {
      "aux_loss": 8.070143985748292,
      "entropy": 0.14362324588000774,
      "epoch": 0.00801,
      "grad_norm": 0.11937398463487625,
      "learning_rate": 2e-05,
      "loss": 0.17357660531997682,
      "mean_token_accuracy": 0.9517420142889023,
      "num_tokens": 128919701.0,
      "step": 8010
    },
    {
      "aux_loss": 8.05161256790161,
      "entropy": 0.1736305795609951,
      "epoch": 0.008015,
      "grad_norm": 0.14193867146968842,
      "learning_rate": 2e-05,
      "loss": 0.21575624942779542,
      "mean_token_accuracy": 0.9399879932403564,
      "num_tokens": 129001254.0,
      "step": 8015
    },
    {
      "aux_loss": 8.052344560623169,
      "entropy": 0.17643555104732514,
      "epoch": 0.00802,
      "grad_norm": 0.1344258338212967,
      "learning_rate": 2e-05,
      "loss": 0.21650593280792235,
      "mean_token_accuracy": 0.9389111161231994,
      "num_tokens": 129082013.0,
      "step": 8020
    },
    {
      "aux_loss": 8.055804872512818,
      "entropy": 0.16876336000859737,
      "epoch": 0.008025,
      "grad_norm": 0.1343068927526474,
      "learning_rate": 2e-05,
      "loss": 0.20226428508758545,
      "mean_token_accuracy": 0.9438631623983383,
      "num_tokens": 129163035.0,
      "step": 8025
    },
    {
      "aux_loss": 8.059794282913208,
      "entropy": 0.16408565491437913,
      "epoch": 0.00803,
      "grad_norm": 0.13268572092056274,
      "learning_rate": 2e-05,
      "loss": 0.1947793483734131,
      "mean_token_accuracy": 0.9456487149000168,
      "num_tokens": 129243298.0,
      "step": 8030
    },
    {
      "aux_loss": 8.068254852294922,
      "entropy": 0.15262793451547624,
      "epoch": 0.008035,
      "grad_norm": 0.134384423494339,
      "learning_rate": 2e-05,
      "loss": 0.17796486616134644,
      "mean_token_accuracy": 0.9506339967250824,
      "num_tokens": 129323881.0,
      "step": 8035
    },
    {
      "aux_loss": 8.0743483543396,
      "entropy": 0.14847229681909085,
      "epoch": 0.00804,
      "grad_norm": 0.12602965533733368,
      "learning_rate": 2e-05,
      "loss": 0.17821271419525148,
      "mean_token_accuracy": 0.9505399107933045,
      "num_tokens": 129403401.0,
      "step": 8040
    },
    {
      "aux_loss": 8.048740434646607,
      "entropy": 0.1746581345796585,
      "epoch": 0.008045,
      "grad_norm": 0.13220183551311493,
      "learning_rate": 2e-05,
      "loss": 0.21524398326873778,
      "mean_token_accuracy": 0.9402400761842727,
      "num_tokens": 129484815.0,
      "step": 8045
    },
    {
      "aux_loss": 8.051606273651123,
      "entropy": 0.19396723434329033,
      "epoch": 0.00805,
      "grad_norm": 0.15582771599292755,
      "learning_rate": 2e-05,
      "loss": 0.23267974853515624,
      "mean_token_accuracy": 0.9352190732955933,
      "num_tokens": 129566009.0,
      "step": 8050
    },
    {
      "aux_loss": 8.053307962417602,
      "entropy": 0.1655742671340704,
      "epoch": 0.008055,
      "grad_norm": 0.12590418756008148,
      "learning_rate": 2e-05,
      "loss": 0.1977403163909912,
      "mean_token_accuracy": 0.9455384075641632,
      "num_tokens": 129647119.0,
      "step": 8055
    },
    {
      "aux_loss": 8.056657075881958,
      "entropy": 0.16041493639349938,
      "epoch": 0.00806,
      "grad_norm": 0.12716123461723328,
      "learning_rate": 2e-05,
      "loss": 0.1873391389846802,
      "mean_token_accuracy": 0.9478627860546112,
      "num_tokens": 129727915.0,
      "step": 8060
    },
    {
      "aux_loss": 8.06365156173706,
      "entropy": 0.1458283070474863,
      "epoch": 0.008065,
      "grad_norm": 0.13084760308265686,
      "learning_rate": 2e-05,
      "loss": 0.1737341284751892,
      "mean_token_accuracy": 0.9514607548713684,
      "num_tokens": 129808142.0,
      "step": 8065
    },
    {
      "aux_loss": 8.072865676879882,
      "entropy": 0.1436923887580633,
      "epoch": 0.00807,
      "grad_norm": 0.12995615601539612,
      "learning_rate": 2e-05,
      "loss": 0.17381809949874877,
      "mean_token_accuracy": 0.9522612005472183,
      "num_tokens": 129888034.0,
      "step": 8070
    },
    {
      "aux_loss": 8.050793600082397,
      "entropy": 0.16842188835144042,
      "epoch": 0.008075,
      "grad_norm": 0.13960705697536469,
      "learning_rate": 2e-05,
      "loss": 0.2096801519393921,
      "mean_token_accuracy": 0.9427917838096619,
      "num_tokens": 129969336.0,
      "step": 8075
    },
    {
      "aux_loss": 8.051363229751587,
      "entropy": 0.16905580945312976,
      "epoch": 0.00808,
      "grad_norm": 0.13340699672698975,
      "learning_rate": 2e-05,
      "loss": 0.20787465572357178,
      "mean_token_accuracy": 0.9414985030889511,
      "num_tokens": 130050704.0,
      "step": 8080
    },
    {
      "aux_loss": 8.05513892173767,
      "entropy": 0.17234189957380294,
      "epoch": 0.008085,
      "grad_norm": 0.12723803520202637,
      "learning_rate": 2e-05,
      "loss": 0.205952787399292,
      "mean_token_accuracy": 0.9429928481578826,
      "num_tokens": 130131685.0,
      "step": 8085
    },
    {
      "aux_loss": 8.060874843597412,
      "entropy": 0.16077012233436108,
      "epoch": 0.00809,
      "grad_norm": 0.12754301726818085,
      "learning_rate": 2e-05,
      "loss": 0.18734116554260255,
      "mean_token_accuracy": 0.947236305475235,
      "num_tokens": 130212333.0,
      "step": 8090
    },
    {
      "aux_loss": 8.067550802230835,
      "entropy": 0.1531975518912077,
      "epoch": 0.008095,
      "grad_norm": 0.13674671947956085,
      "learning_rate": 2e-05,
      "loss": 0.17868854999542236,
      "mean_token_accuracy": 0.950370478630066,
      "num_tokens": 130292662.0,
      "step": 8095
    },
    {
      "aux_loss": 8.078342294692993,
      "entropy": 0.13891354911029338,
      "epoch": 0.0081,
      "grad_norm": 0.1541026383638382,
      "learning_rate": 2e-05,
      "loss": 0.16874922513961793,
      "mean_token_accuracy": 0.9530214101076127,
      "num_tokens": 130372981.0,
      "step": 8100
    },
    {
      "aux_loss": 8.051136112213134,
      "entropy": 0.1572195865213871,
      "epoch": 0.008105,
      "grad_norm": 0.14138002693653107,
      "learning_rate": 2e-05,
      "loss": 0.19828492403030396,
      "mean_token_accuracy": 0.9464273273944854,
      "num_tokens": 130454458.0,
      "step": 8105
    },
    {
      "aux_loss": 8.054611682891846,
      "entropy": 0.1773559059947729,
      "epoch": 0.00811,
      "grad_norm": 0.13690850138664246,
      "learning_rate": 2e-05,
      "loss": 0.21153836250305175,
      "mean_token_accuracy": 0.9408694803714752,
      "num_tokens": 130535536.0,
      "step": 8110
    },
    {
      "aux_loss": 8.0561372756958,
      "entropy": 0.18155391663312911,
      "epoch": 0.008115,
      "grad_norm": 0.13156554102897644,
      "learning_rate": 2e-05,
      "loss": 0.21547298431396483,
      "mean_token_accuracy": 0.940059581398964,
      "num_tokens": 130616282.0,
      "step": 8115
    },
    {
      "aux_loss": 8.061676025390625,
      "entropy": 0.15969658866524697,
      "epoch": 0.00812,
      "grad_norm": 0.12900713086128235,
      "learning_rate": 2e-05,
      "loss": 0.18574862480163573,
      "mean_token_accuracy": 0.948108771443367,
      "num_tokens": 130697326.0,
      "step": 8120
    },
    {
      "aux_loss": 8.068650913238525,
      "entropy": 0.15999038219451905,
      "epoch": 0.008125,
      "grad_norm": 0.14218170940876007,
      "learning_rate": 2e-05,
      "loss": 0.18584848642349244,
      "mean_token_accuracy": 0.9481202960014343,
      "num_tokens": 130777453.0,
      "step": 8125
    },
    {
      "aux_loss": 8.077187442779541,
      "entropy": 0.14273985028266906,
      "epoch": 0.00813,
      "grad_norm": 0.13563385605812073,
      "learning_rate": 2e-05,
      "loss": 0.17152726650238037,
      "mean_token_accuracy": 0.9528078377246857,
      "num_tokens": 130857778.0,
      "step": 8130
    },
    {
      "aux_loss": 8.053657245635986,
      "entropy": 0.14082985408604146,
      "epoch": 0.008135,
      "grad_norm": 0.14208555221557617,
      "learning_rate": 2e-05,
      "loss": 0.1782923698425293,
      "mean_token_accuracy": 0.9514079868793488,
      "num_tokens": 130936964.0,
      "step": 8135
    },
    {
      "aux_loss": 8.055822563171386,
      "entropy": 0.17075503170490264,
      "epoch": 0.00814,
      "grad_norm": 0.1350121647119522,
      "learning_rate": 2e-05,
      "loss": 0.20721492767333985,
      "mean_token_accuracy": 0.9425066232681274,
      "num_tokens": 131018374.0,
      "step": 8140
    },
    {
      "aux_loss": 8.053418064117432,
      "entropy": 0.15526145249605178,
      "epoch": 0.008145,
      "grad_norm": 0.13038256764411926,
      "learning_rate": 2e-05,
      "loss": 0.18912750482559204,
      "mean_token_accuracy": 0.9473474532365799,
      "num_tokens": 131099411.0,
      "step": 8145
    },
    {
      "aux_loss": 8.057427453994752,
      "entropy": 0.15396433770656587,
      "epoch": 0.00815,
      "grad_norm": 0.1316199153661728,
      "learning_rate": 2e-05,
      "loss": 0.18397905826568603,
      "mean_token_accuracy": 0.9487013101577759,
      "num_tokens": 131179905.0,
      "step": 8150
    },
    {
      "aux_loss": 8.063847064971924,
      "entropy": 0.1535062536597252,
      "epoch": 0.008155,
      "grad_norm": 0.1289980262517929,
      "learning_rate": 2e-05,
      "loss": 0.184206759929657,
      "mean_token_accuracy": 0.9487711012363433,
      "num_tokens": 131260460.0,
      "step": 8155
    },
    {
      "aux_loss": 8.072390508651733,
      "entropy": 0.14371876530349253,
      "epoch": 0.00816,
      "grad_norm": 0.1354750245809555,
      "learning_rate": 2e-05,
      "loss": 0.16831915378570556,
      "mean_token_accuracy": 0.9534879624843597,
      "num_tokens": 131340247.0,
      "step": 8160
    },
    {
      "aux_loss": 8.063643741607667,
      "entropy": 0.16286289542913437,
      "epoch": 0.008165,
      "grad_norm": 0.13989847898483276,
      "learning_rate": 2e-05,
      "loss": 0.19906128644943238,
      "mean_token_accuracy": 0.9450865864753724,
      "num_tokens": 131419822.0,
      "step": 8165
    },
    {
      "aux_loss": 8.052569580078124,
      "entropy": 0.17371668070554733,
      "epoch": 0.00817,
      "grad_norm": 0.13841158151626587,
      "learning_rate": 2e-05,
      "loss": 0.21525461673736573,
      "mean_token_accuracy": 0.9402058362960816,
      "num_tokens": 131501314.0,
      "step": 8170
    },
    {
      "aux_loss": 8.05364236831665,
      "entropy": 0.16507818698883056,
      "epoch": 0.008175,
      "grad_norm": 0.12713953852653503,
      "learning_rate": 2e-05,
      "loss": 0.19896907806396485,
      "mean_token_accuracy": 0.9446077913045883,
      "num_tokens": 131582432.0,
      "step": 8175
    },
    {
      "aux_loss": 8.05719633102417,
      "entropy": 0.15961042642593384,
      "epoch": 0.00818,
      "grad_norm": 0.12847581505775452,
      "learning_rate": 2e-05,
      "loss": 0.18849375247955322,
      "mean_token_accuracy": 0.9474297881126403,
      "num_tokens": 131662760.0,
      "step": 8180
    },
    {
      "aux_loss": 8.065714836120605,
      "entropy": 0.15520738139748574,
      "epoch": 0.008185,
      "grad_norm": 0.128628671169281,
      "learning_rate": 2e-05,
      "loss": 0.18253229856491088,
      "mean_token_accuracy": 0.9487132281064987,
      "num_tokens": 131743360.0,
      "step": 8185
    },
    {
      "aux_loss": 8.073528146743774,
      "entropy": 0.14960686787962912,
      "epoch": 0.00819,
      "grad_norm": 0.13791678845882416,
      "learning_rate": 2e-05,
      "loss": 0.1772822380065918,
      "mean_token_accuracy": 0.9508191347122192,
      "num_tokens": 131823293.0,
      "step": 8190
    },
    {
      "aux_loss": 8.069000148773194,
      "entropy": 0.14849095791578293,
      "epoch": 0.008195,
      "grad_norm": 0.1389141082763672,
      "learning_rate": 2e-05,
      "loss": 0.17870173454284669,
      "mean_token_accuracy": 0.95055011510849,
      "num_tokens": 131900144.0,
      "step": 8195
    },
    {
      "aux_loss": 8.05020546913147,
      "entropy": 0.1647246763110161,
      "epoch": 0.0082,
      "grad_norm": 0.13321228325366974,
      "learning_rate": 2e-05,
      "loss": 0.2091508388519287,
      "mean_token_accuracy": 0.9420828849077225,
      "num_tokens": 131980926.0,
      "step": 8200
    },
    {
      "aux_loss": 8.053854846954346,
      "entropy": 0.15634859763085843,
      "epoch": 0.008205,
      "grad_norm": 0.13224320113658905,
      "learning_rate": 2e-05,
      "loss": 0.19497339725494384,
      "mean_token_accuracy": 0.9453929275274277,
      "num_tokens": 132062090.0,
      "step": 8205
    },
    {
      "aux_loss": 8.054818630218506,
      "entropy": 0.1541820175945759,
      "epoch": 0.00821,
      "grad_norm": 0.12972988188266754,
      "learning_rate": 2e-05,
      "loss": 0.18355741500854492,
      "mean_token_accuracy": 0.9486854493618011,
      "num_tokens": 132142939.0,
      "step": 8210
    },
    {
      "aux_loss": 8.062036609649658,
      "entropy": 0.15288175493478776,
      "epoch": 0.008215,
      "grad_norm": 0.11998964101076126,
      "learning_rate": 2e-05,
      "loss": 0.1793956995010376,
      "mean_token_accuracy": 0.9495918989181519,
      "num_tokens": 132223672.0,
      "step": 8215
    },
    {
      "aux_loss": 8.073719263076782,
      "entropy": 0.14733343049883843,
      "epoch": 0.00822,
      "grad_norm": 0.14416344463825226,
      "learning_rate": 2e-05,
      "loss": 0.17541441917419434,
      "mean_token_accuracy": 0.950934499502182,
      "num_tokens": 132303650.0,
      "step": 8220
    },
    {
      "aux_loss": 8.063137292861938,
      "entropy": 0.14231124371290207,
      "epoch": 0.008225,
      "grad_norm": 0.12300004810094833,
      "learning_rate": 2e-05,
      "loss": 0.17548067569732667,
      "mean_token_accuracy": 0.9528339743614197,
      "num_tokens": 132383299.0,
      "step": 8225
    },
    {
      "aux_loss": 8.054024934768677,
      "entropy": 0.1865336313843727,
      "epoch": 0.00823,
      "grad_norm": 0.12891030311584473,
      "learning_rate": 2e-05,
      "loss": 0.2253323793411255,
      "mean_token_accuracy": 0.936671644449234,
      "num_tokens": 132464255.0,
      "step": 8230
    },
    {
      "aux_loss": 8.052585649490357,
      "entropy": 0.1659695204347372,
      "epoch": 0.008235,
      "grad_norm": 0.1280297040939331,
      "learning_rate": 2e-05,
      "loss": 0.20176198482513427,
      "mean_token_accuracy": 0.9440831243991852,
      "num_tokens": 132545373.0,
      "step": 8235
    },
    {
      "aux_loss": 8.057102346420288,
      "entropy": 0.15998641178011894,
      "epoch": 0.00824,
      "grad_norm": 0.12822078168392181,
      "learning_rate": 2e-05,
      "loss": 0.18847664594650268,
      "mean_token_accuracy": 0.9475738674402236,
      "num_tokens": 132625981.0,
      "step": 8240
    },
    {
      "aux_loss": 8.060812997817994,
      "entropy": 0.16090417951345443,
      "epoch": 0.008245,
      "grad_norm": 0.12101796269416809,
      "learning_rate": 2e-05,
      "loss": 0.1924592971801758,
      "mean_token_accuracy": 0.9462516725063324,
      "num_tokens": 132706743.0,
      "step": 8245
    },
    {
      "aux_loss": 8.070701217651367,
      "entropy": 0.145225390791893,
      "epoch": 0.00825,
      "grad_norm": 0.14184513688087463,
      "learning_rate": 2e-05,
      "loss": 0.17299001216888427,
      "mean_token_accuracy": 0.9512601137161255,
      "num_tokens": 132787061.0,
      "step": 8250
    },
    {
      "aux_loss": 8.070589399337768,
      "entropy": 0.13938662335276603,
      "epoch": 0.008255,
      "grad_norm": 0.12141731381416321,
      "learning_rate": 2e-05,
      "loss": 0.1703873634338379,
      "mean_token_accuracy": 0.9535843312740326,
      "num_tokens": 132866564.0,
      "step": 8255
    },
    {
      "aux_loss": 8.052244663238525,
      "entropy": 0.17886133380234243,
      "epoch": 0.00826,
      "grad_norm": 0.1335018426179886,
      "learning_rate": 2e-05,
      "loss": 0.21689348220825194,
      "mean_token_accuracy": 0.9410188883543015,
      "num_tokens": 132947556.0,
      "step": 8260
    },
    {
      "aux_loss": 8.050800132751466,
      "entropy": 0.1688988868147135,
      "epoch": 0.008265,
      "grad_norm": 0.1476362943649292,
      "learning_rate": 2e-05,
      "loss": 0.20815763473510743,
      "mean_token_accuracy": 0.942469248175621,
      "num_tokens": 133028724.0,
      "step": 8265
    },
    {
      "aux_loss": 8.057167053222656,
      "entropy": 0.16154177039861678,
      "epoch": 0.00827,
      "grad_norm": 0.13477906584739685,
      "learning_rate": 2e-05,
      "loss": 0.1949950337409973,
      "mean_token_accuracy": 0.9448888748884201,
      "num_tokens": 133109722.0,
      "step": 8270
    },
    {
      "aux_loss": 8.06676697731018,
      "entropy": 0.15860776528716086,
      "epoch": 0.008275,
      "grad_norm": 0.13367654383182526,
      "learning_rate": 2e-05,
      "loss": 0.18534401655197144,
      "mean_token_accuracy": 0.9481903582811355,
      "num_tokens": 133190601.0,
      "step": 8275
    },
    {
      "aux_loss": 8.070173358917236,
      "entropy": 0.14624718800187111,
      "epoch": 0.00828,
      "grad_norm": 0.13547837734222412,
      "learning_rate": 2e-05,
      "loss": 0.1699145793914795,
      "mean_token_accuracy": 0.9528974741697311,
      "num_tokens": 133270725.0,
      "step": 8280
    },
    {
      "aux_loss": 8.068696594238281,
      "entropy": 0.14652694389224052,
      "epoch": 0.008285,
      "grad_norm": 0.12477568536996841,
      "learning_rate": 2e-05,
      "loss": 0.17745449542999267,
      "mean_token_accuracy": 0.9509728610515594,
      "num_tokens": 133350981.0,
      "step": 8285
    },
    {
      "aux_loss": 8.051817417144775,
      "entropy": 0.16215978749096394,
      "epoch": 0.00829,
      "grad_norm": 0.13373258709907532,
      "learning_rate": 2e-05,
      "loss": 0.20403625965118408,
      "mean_token_accuracy": 0.9429250538349152,
      "num_tokens": 133432006.0,
      "step": 8290
    },
    {
      "aux_loss": 8.052235221862793,
      "entropy": 0.16459218487143518,
      "epoch": 0.008295,
      "grad_norm": 0.13437971472740173,
      "learning_rate": 2e-05,
      "loss": 0.20010752677917482,
      "mean_token_accuracy": 0.9441726863384247,
      "num_tokens": 133513057.0,
      "step": 8295
    },
    {
      "aux_loss": 8.056406164169312,
      "entropy": 0.16486706621944905,
      "epoch": 0.0083,
      "grad_norm": 0.13076648116111755,
      "learning_rate": 2e-05,
      "loss": 0.19752124547958375,
      "mean_token_accuracy": 0.9446148037910461,
      "num_tokens": 133593871.0,
      "step": 8300
    },
    {
      "aux_loss": 8.061829566955566,
      "entropy": 0.15138532780110836,
      "epoch": 0.008305,
      "grad_norm": 0.1317295879125595,
      "learning_rate": 2e-05,
      "loss": 0.1803811311721802,
      "mean_token_accuracy": 0.9497204631567001,
      "num_tokens": 133674497.0,
      "step": 8305
    },
    {
      "aux_loss": 8.069890785217286,
      "entropy": 0.14223296083509923,
      "epoch": 0.00831,
      "grad_norm": 0.13315069675445557,
      "learning_rate": 2e-05,
      "loss": 0.16793813705444335,
      "mean_token_accuracy": 0.9531345009803772,
      "num_tokens": 133754710.0,
      "step": 8310
    },
    {
      "aux_loss": 8.069337129592896,
      "entropy": 0.14927798248827456,
      "epoch": 0.008315,
      "grad_norm": 0.1374337375164032,
      "learning_rate": 2e-05,
      "loss": 0.1796567440032959,
      "mean_token_accuracy": 0.9504870891571044,
      "num_tokens": 133834987.0,
      "step": 8315
    },
    {
      "aux_loss": 8.053461647033691,
      "entropy": 0.16906164549291133,
      "epoch": 0.00832,
      "grad_norm": 0.13867270946502686,
      "learning_rate": 2e-05,
      "loss": 0.21054699420928955,
      "mean_token_accuracy": 0.9407571345567703,
      "num_tokens": 133916106.0,
      "step": 8320
    },
    {
      "aux_loss": 8.0525053024292,
      "entropy": 0.16746328137814998,
      "epoch": 0.008325,
      "grad_norm": 0.13184724748134613,
      "learning_rate": 2e-05,
      "loss": 0.20621328353881835,
      "mean_token_accuracy": 0.9427377462387085,
      "num_tokens": 133996984.0,
      "step": 8325
    },
    {
      "aux_loss": 8.055532121658326,
      "entropy": 0.1552527479827404,
      "epoch": 0.00833,
      "grad_norm": 0.12123899161815643,
      "learning_rate": 2e-05,
      "loss": 0.1862831234931946,
      "mean_token_accuracy": 0.9481147855520249,
      "num_tokens": 134077744.0,
      "step": 8330
    },
    {
      "aux_loss": 8.060292482376099,
      "entropy": 0.15543766990303992,
      "epoch": 0.008335,
      "grad_norm": 0.13376720249652863,
      "learning_rate": 2e-05,
      "loss": 0.18655211925506593,
      "mean_token_accuracy": 0.948227059841156,
      "num_tokens": 134158621.0,
      "step": 8335
    },
    {
      "aux_loss": 8.065773534774781,
      "entropy": 0.14543639309704304,
      "epoch": 0.00834,
      "grad_norm": 0.1333523690700531,
      "learning_rate": 2e-05,
      "loss": 0.17411571741104126,
      "mean_token_accuracy": 0.9517302125692367,
      "num_tokens": 134239428.0,
      "step": 8340
    },
    {
      "aux_loss": 8.07011046409607,
      "entropy": 0.13980746679008008,
      "epoch": 0.008345,
      "grad_norm": 0.13500382006168365,
      "learning_rate": 2e-05,
      "loss": 0.16771507263183594,
      "mean_token_accuracy": 0.9539576590061187,
      "num_tokens": 134318586.0,
      "step": 8345
    },
    {
      "aux_loss": 8.050806093215943,
      "entropy": 0.1748618222773075,
      "epoch": 0.00835,
      "grad_norm": 0.12813574075698853,
      "learning_rate": 2e-05,
      "loss": 0.21088182926177979,
      "mean_token_accuracy": 0.9425250440835953,
      "num_tokens": 134399942.0,
      "step": 8350
    },
    {
      "aux_loss": 8.050401067733764,
      "entropy": 0.16043828427791595,
      "epoch": 0.008355,
      "grad_norm": 0.1389087438583374,
      "learning_rate": 2e-05,
      "loss": 0.20061631202697755,
      "mean_token_accuracy": 0.944438835978508,
      "num_tokens": 134480226.0,
      "step": 8355
    },
    {
      "aux_loss": 8.055367946624756,
      "entropy": 0.16035865880548955,
      "epoch": 0.00836,
      "grad_norm": 0.13739383220672607,
      "learning_rate": 2e-05,
      "loss": 0.19648418426513672,
      "mean_token_accuracy": 0.9449416488409043,
      "num_tokens": 134560975.0,
      "step": 8360
    },
    {
      "aux_loss": 8.060657596588134,
      "entropy": 0.15242384225130082,
      "epoch": 0.008365,
      "grad_norm": 0.12526613473892212,
      "learning_rate": 2e-05,
      "loss": 0.17871508598327637,
      "mean_token_accuracy": 0.9498950958251953,
      "num_tokens": 134641902.0,
      "step": 8365
    },
    {
      "aux_loss": 8.066985416412354,
      "entropy": 0.14445665068924426,
      "epoch": 0.00837,
      "grad_norm": 0.12973423302173615,
      "learning_rate": 2e-05,
      "loss": 0.16594820022583007,
      "mean_token_accuracy": 0.9540929168462753,
      "num_tokens": 134722740.0,
      "step": 8370
    },
    {
      "aux_loss": 8.07032914161682,
      "entropy": 0.13631007261574268,
      "epoch": 0.008375,
      "grad_norm": 0.1262480467557907,
      "learning_rate": 2e-05,
      "loss": 0.1668606996536255,
      "mean_token_accuracy": 0.9549294799566269,
      "num_tokens": 134802594.0,
      "step": 8375
    },
    {
      "aux_loss": 8.051563024520874,
      "entropy": 0.15825699903070928,
      "epoch": 0.00838,
      "grad_norm": 0.14672711491584778,
      "learning_rate": 2e-05,
      "loss": 0.1998654007911682,
      "mean_token_accuracy": 0.9451659232378006,
      "num_tokens": 134884114.0,
      "step": 8380
    },
    {
      "aux_loss": 8.054048776626587,
      "entropy": 0.15660641267895697,
      "epoch": 0.008385,
      "grad_norm": 0.12988872826099396,
      "learning_rate": 2e-05,
      "loss": 0.19593580961227416,
      "mean_token_accuracy": 0.9451469659805298,
      "num_tokens": 134965204.0,
      "step": 8385
    },
    {
      "aux_loss": 8.05440993309021,
      "entropy": 0.15677015855908394,
      "epoch": 0.00839,
      "grad_norm": 0.1295607089996338,
      "learning_rate": 2e-05,
      "loss": 0.1928365111351013,
      "mean_token_accuracy": 0.9462762326002121,
      "num_tokens": 135045536.0,
      "step": 8390
    },
    {
      "aux_loss": 8.062279844284058,
      "entropy": 0.1617270179092884,
      "epoch": 0.008395,
      "grad_norm": 0.12894052267074585,
      "learning_rate": 2e-05,
      "loss": 0.1873775005340576,
      "mean_token_accuracy": 0.9466507911682129,
      "num_tokens": 135126285.0,
      "step": 8395
    },
    {
      "aux_loss": 8.067923831939698,
      "entropy": 0.14252947755157946,
      "epoch": 0.0084,
      "grad_norm": 0.13394547998905182,
      "learning_rate": 2e-05,
      "loss": 0.1701805353164673,
      "mean_token_accuracy": 0.9530052065849304,
      "num_tokens": 135206887.0,
      "step": 8400
    },
    {
      "aux_loss": 8.075802040100097,
      "entropy": 0.13822566233575345,
      "epoch": 0.008405,
      "grad_norm": 0.13011668622493744,
      "learning_rate": 2e-05,
      "loss": 0.1682661294937134,
      "mean_token_accuracy": 0.9539941906929016,
      "num_tokens": 135286660.0,
      "step": 8405
    },
    {
      "aux_loss": 8.051136493682861,
      "entropy": 0.17228760942816734,
      "epoch": 0.00841,
      "grad_norm": 0.1299499124288559,
      "learning_rate": 2e-05,
      "loss": 0.21304645538330078,
      "mean_token_accuracy": 0.9413011193275451,
      "num_tokens": 135367984.0,
      "step": 8410
    },
    {
      "aux_loss": 8.050592994689941,
      "entropy": 0.16765848957002163,
      "epoch": 0.008415,
      "grad_norm": 0.13136902451515198,
      "learning_rate": 2e-05,
      "loss": 0.2054699420928955,
      "mean_token_accuracy": 0.9424240410327911,
      "num_tokens": 135448603.0,
      "step": 8415
    },
    {
      "aux_loss": 8.052868938446045,
      "entropy": 0.1590926930308342,
      "epoch": 0.00842,
      "grad_norm": 0.1242709681391716,
      "learning_rate": 2e-05,
      "loss": 0.1873373031616211,
      "mean_token_accuracy": 0.9476733565330505,
      "num_tokens": 135529829.0,
      "step": 8420
    },
    {
      "aux_loss": 8.061859369277954,
      "entropy": 0.15883804187178613,
      "epoch": 0.008425,
      "grad_norm": 0.136344775557518,
      "learning_rate": 2e-05,
      "loss": 0.1884535789489746,
      "mean_token_accuracy": 0.9476337909698487,
      "num_tokens": 135610568.0,
      "step": 8425
    },
    {
      "aux_loss": 8.067611932754517,
      "entropy": 0.14792343154549598,
      "epoch": 0.00843,
      "grad_norm": 0.1428966373205185,
      "learning_rate": 2e-05,
      "loss": 0.17854593992233275,
      "mean_token_accuracy": 0.9499095141887665,
      "num_tokens": 135691584.0,
      "step": 8430
    },
    {
      "aux_loss": 8.077659225463867,
      "entropy": 0.136220546066761,
      "epoch": 0.008435,
      "grad_norm": 0.13112741708755493,
      "learning_rate": 2e-05,
      "loss": 0.16483110189437866,
      "mean_token_accuracy": 0.9541334748268128,
      "num_tokens": 135770218.0,
      "step": 8435
    },
    {
      "aux_loss": 8.04805555343628,
      "entropy": 0.17007865346968173,
      "epoch": 0.00844,
      "grad_norm": 0.12503069639205933,
      "learning_rate": 2e-05,
      "loss": 0.20843098163604737,
      "mean_token_accuracy": 0.941634801030159,
      "num_tokens": 135851612.0,
      "step": 8440
    },
    {
      "aux_loss": 8.049964237213135,
      "entropy": 0.17655111774802207,
      "epoch": 0.008445,
      "grad_norm": 0.1310964673757553,
      "learning_rate": 2e-05,
      "loss": 0.21329019069671631,
      "mean_token_accuracy": 0.9415773063898086,
      "num_tokens": 135932161.0,
      "step": 8445
    },
    {
      "aux_loss": 8.052983236312866,
      "entropy": 0.16025624498724939,
      "epoch": 0.00845,
      "grad_norm": 0.13463611900806427,
      "learning_rate": 2e-05,
      "loss": 0.1935208797454834,
      "mean_token_accuracy": 0.945856460928917,
      "num_tokens": 136013260.0,
      "step": 8450
    },
    {
      "aux_loss": 8.061661100387573,
      "entropy": 0.15678115971386433,
      "epoch": 0.008455,
      "grad_norm": 0.12999224662780762,
      "learning_rate": 2e-05,
      "loss": 0.18524272441864015,
      "mean_token_accuracy": 0.9482613444328308,
      "num_tokens": 136093767.0,
      "step": 8455
    },
    {
      "aux_loss": 8.066637802124024,
      "entropy": 0.13855443336069584,
      "epoch": 0.00846,
      "grad_norm": 0.13254959881305695,
      "learning_rate": 2e-05,
      "loss": 0.16542829275131227,
      "mean_token_accuracy": 0.9536194920539856,
      "num_tokens": 136174562.0,
      "step": 8460
    },
    {
      "aux_loss": 8.08025507926941,
      "entropy": 0.1366057839244604,
      "epoch": 0.008465,
      "grad_norm": 0.1474137157201767,
      "learning_rate": 2e-05,
      "loss": 0.1648195743560791,
      "mean_token_accuracy": 0.9535883605480194,
      "num_tokens": 136251997.0,
      "step": 8465
    },
    {
      "aux_loss": 8.050443744659423,
      "entropy": 0.1504580620676279,
      "epoch": 0.00847,
      "grad_norm": 0.12698541581630707,
      "learning_rate": 2e-05,
      "loss": 0.19243426322937013,
      "mean_token_accuracy": 0.9472835689783097,
      "num_tokens": 136333090.0,
      "step": 8470
    },
    {
      "aux_loss": 8.051616382598876,
      "entropy": 0.1817252304404974,
      "epoch": 0.008475,
      "grad_norm": 0.13129396736621857,
      "learning_rate": 2e-05,
      "loss": 0.22213666439056395,
      "mean_token_accuracy": 0.9386663466691971,
      "num_tokens": 136414015.0,
      "step": 8475
    },
    {
      "aux_loss": 8.05089349746704,
      "entropy": 0.16867434233427048,
      "epoch": 0.00848,
      "grad_norm": 0.13452528417110443,
      "learning_rate": 2e-05,
      "loss": 0.1958939552307129,
      "mean_token_accuracy": 0.9444753050804138,
      "num_tokens": 136495101.0,
      "step": 8480
    },
    {
      "aux_loss": 8.053499364852906,
      "entropy": 0.1629290573298931,
      "epoch": 0.008485,
      "grad_norm": 0.1275554746389389,
      "learning_rate": 2e-05,
      "loss": 0.19197280406951905,
      "mean_token_accuracy": 0.9465134739875793,
      "num_tokens": 136575540.0,
      "step": 8485
    },
    {
      "aux_loss": 8.064253425598144,
      "entropy": 0.1574890699237585,
      "epoch": 0.00849,
      "grad_norm": 0.1433914452791214,
      "learning_rate": 2e-05,
      "loss": 0.1865792989730835,
      "mean_token_accuracy": 0.9476440668106079,
      "num_tokens": 136656189.0,
      "step": 8490
    },
    {
      "aux_loss": 8.074000263214112,
      "entropy": 0.1440015934407711,
      "epoch": 0.008495,
      "grad_norm": 0.13831086456775665,
      "learning_rate": 2e-05,
      "loss": 0.17209904193878173,
      "mean_token_accuracy": 0.9527287960052491,
      "num_tokens": 136736909.0,
      "step": 8495
    },
    {
      "aux_loss": 8.062353897094727,
      "entropy": 0.15502053014934064,
      "epoch": 0.0085,
      "grad_norm": 0.12684522569179535,
      "learning_rate": 2e-05,
      "loss": 0.18892711400985718,
      "mean_token_accuracy": 0.9479603826999664,
      "num_tokens": 136815132.0,
      "step": 8500
    },
    {
      "aux_loss": 8.05538215637207,
      "entropy": 0.16450981870293618,
      "epoch": 0.008505,
      "grad_norm": 0.1259032040834427,
      "learning_rate": 2e-05,
      "loss": 0.19836450815200807,
      "mean_token_accuracy": 0.9436654269695282,
      "num_tokens": 136896292.0,
      "step": 8505
    },
    {
      "aux_loss": 8.050623846054076,
      "entropy": 0.16529802270233632,
      "epoch": 0.00851,
      "grad_norm": 0.13030876219272614,
      "learning_rate": 2e-05,
      "loss": 0.20448694229125977,
      "mean_token_accuracy": 0.9423883527517318,
      "num_tokens": 136977300.0,
      "step": 8510
    },
    {
      "aux_loss": 8.054156970977782,
      "entropy": 0.16047285869717598,
      "epoch": 0.008515,
      "grad_norm": 0.1274743676185608,
      "learning_rate": 2e-05,
      "loss": 0.1904818296432495,
      "mean_token_accuracy": 0.9465765982866288,
      "num_tokens": 137058037.0,
      "step": 8515
    },
    {
      "aux_loss": 8.063240242004394,
      "entropy": 0.16352380365133284,
      "epoch": 0.00852,
      "grad_norm": 0.12635862827301025,
      "learning_rate": 2e-05,
      "loss": 0.1896214723587036,
      "mean_token_accuracy": 0.9474231600761414,
      "num_tokens": 137138516.0,
      "step": 8520
    },
    {
      "aux_loss": 8.07029881477356,
      "entropy": 0.1384718779474497,
      "epoch": 0.008525,
      "grad_norm": 0.1427355408668518,
      "learning_rate": 2e-05,
      "loss": 0.16541393995285034,
      "mean_token_accuracy": 0.9543010950088501,
      "num_tokens": 137218756.0,
      "step": 8525
    },
    {
      "aux_loss": 8.065940237045288,
      "entropy": 0.15036954395473004,
      "epoch": 0.00853,
      "grad_norm": 0.1259373277425766,
      "learning_rate": 2e-05,
      "loss": 0.18651095628738404,
      "mean_token_accuracy": 0.9489887952804565,
      "num_tokens": 137297311.0,
      "step": 8530
    },
    {
      "aux_loss": 8.050931978225709,
      "entropy": 0.17691235840320588,
      "epoch": 0.008535,
      "grad_norm": 0.13844381272792816,
      "learning_rate": 2e-05,
      "loss": 0.216611909866333,
      "mean_token_accuracy": 0.9383084505796433,
      "num_tokens": 137378163.0,
      "step": 8535
    },
    {
      "aux_loss": 8.052419710159302,
      "entropy": 0.16057549603283405,
      "epoch": 0.00854,
      "grad_norm": 0.12409794330596924,
      "learning_rate": 2e-05,
      "loss": 0.1939744234085083,
      "mean_token_accuracy": 0.9457886725664139,
      "num_tokens": 137459014.0,
      "step": 8540
    },
    {
      "aux_loss": 8.05696759223938,
      "entropy": 0.15709731429815293,
      "epoch": 0.008545,
      "grad_norm": 0.12415891885757446,
      "learning_rate": 2e-05,
      "loss": 0.18806880712509155,
      "mean_token_accuracy": 0.9476287961006165,
      "num_tokens": 137539713.0,
      "step": 8545
    },
    {
      "aux_loss": 8.06216320991516,
      "entropy": 0.1605144962668419,
      "epoch": 0.00855,
      "grad_norm": 0.1295788288116455,
      "learning_rate": 2e-05,
      "loss": 0.18548530340194702,
      "mean_token_accuracy": 0.948658874630928,
      "num_tokens": 137620469.0,
      "step": 8550
    },
    {
      "aux_loss": 8.07037444114685,
      "entropy": 0.14625077769160272,
      "epoch": 0.008555,
      "grad_norm": 0.1383758783340454,
      "learning_rate": 2e-05,
      "loss": 0.1735527515411377,
      "mean_token_accuracy": 0.9521961241960526,
      "num_tokens": 137700728.0,
      "step": 8555
    },
    {
      "aux_loss": 8.070580625534058,
      "entropy": 0.12732557877898215,
      "epoch": 0.00856,
      "grad_norm": 0.1281554400920868,
      "learning_rate": 2e-05,
      "loss": 0.15901296138763427,
      "mean_token_accuracy": 0.9563904017210006,
      "num_tokens": 137780886.0,
      "step": 8560
    },
    {
      "aux_loss": 8.053347492218018,
      "entropy": 0.15632971115410327,
      "epoch": 0.008565,
      "grad_norm": 0.1276077926158905,
      "learning_rate": 2e-05,
      "loss": 0.20193641185760497,
      "mean_token_accuracy": 0.9426657378673553,
      "num_tokens": 137862342.0,
      "step": 8565
    },
    {
      "aux_loss": 8.056013822555542,
      "entropy": 0.15691122002899646,
      "epoch": 0.00857,
      "grad_norm": 0.13058046996593475,
      "learning_rate": 2e-05,
      "loss": 0.19218190908432006,
      "mean_token_accuracy": 0.9463227838277817,
      "num_tokens": 137943351.0,
      "step": 8570
    },
    {
      "aux_loss": 8.054302597045899,
      "entropy": 0.16809199415147305,
      "epoch": 0.008575,
      "grad_norm": 0.12302262336015701,
      "learning_rate": 2e-05,
      "loss": 0.20412819385528563,
      "mean_token_accuracy": 0.9438839465379715,
      "num_tokens": 138023937.0,
      "step": 8575
    },
    {
      "aux_loss": 8.059054136276245,
      "entropy": 0.15220602191984653,
      "epoch": 0.00858,
      "grad_norm": 0.1288856863975525,
      "learning_rate": 2e-05,
      "loss": 0.18049123287200927,
      "mean_token_accuracy": 0.9497587084770203,
      "num_tokens": 138104329.0,
      "step": 8580
    },
    {
      "aux_loss": 8.066953325271607,
      "entropy": 0.14887753091752529,
      "epoch": 0.008585,
      "grad_norm": 0.13680119812488556,
      "learning_rate": 2e-05,
      "loss": 0.17565748691558838,
      "mean_token_accuracy": 0.9511402636766434,
      "num_tokens": 138185089.0,
      "step": 8585
    },
    {
      "aux_loss": 8.07802700996399,
      "entropy": 0.14214944653213024,
      "epoch": 0.00859,
      "grad_norm": 0.12327562272548676,
      "learning_rate": 2e-05,
      "loss": 0.16872937679290773,
      "mean_token_accuracy": 0.952862823009491,
      "num_tokens": 138263094.0,
      "step": 8590
    },
    {
      "aux_loss": 8.053894662857056,
      "entropy": 0.17544727213680744,
      "epoch": 0.008595,
      "grad_norm": 0.12496910989284515,
      "learning_rate": 2e-05,
      "loss": 0.21347546577453613,
      "mean_token_accuracy": 0.9417501091957092,
      "num_tokens": 138344538.0,
      "step": 8595
    },
    {
      "aux_loss": 8.054403495788574,
      "entropy": 0.16302029825747014,
      "epoch": 0.0086,
      "grad_norm": 0.12998966872692108,
      "learning_rate": 2e-05,
      "loss": 0.19678072929382323,
      "mean_token_accuracy": 0.9450888693332672,
      "num_tokens": 138425256.0,
      "step": 8600
    },
    {
      "aux_loss": 8.053652095794678,
      "entropy": 0.15328414253890515,
      "epoch": 0.008605,
      "grad_norm": 0.1274077296257019,
      "learning_rate": 2e-05,
      "loss": 0.18891308307647706,
      "mean_token_accuracy": 0.9469194948673249,
      "num_tokens": 138505465.0,
      "step": 8605
    },
    {
      "aux_loss": 8.064128017425537,
      "entropy": 0.15721628814935684,
      "epoch": 0.00861,
      "grad_norm": 0.1264345645904541,
      "learning_rate": 2e-05,
      "loss": 0.1871870279312134,
      "mean_token_accuracy": 0.9478241115808487,
      "num_tokens": 138586320.0,
      "step": 8610
    },
    {
      "aux_loss": 8.068868494033813,
      "entropy": 0.14275341480970383,
      "epoch": 0.008615,
      "grad_norm": 0.13732150197029114,
      "learning_rate": 2e-05,
      "loss": 0.16966525316238404,
      "mean_token_accuracy": 0.9530564188957215,
      "num_tokens": 138667068.0,
      "step": 8615
    },
    {
      "aux_loss": 8.075186586380005,
      "entropy": 0.142410034686327,
      "epoch": 0.00862,
      "grad_norm": 0.11204396188259125,
      "learning_rate": 2e-05,
      "loss": 0.16834025382995604,
      "mean_token_accuracy": 0.9537007331848144,
      "num_tokens": 138743979.0,
      "step": 8620
    },
    {
      "aux_loss": 8.050409317016602,
      "entropy": 0.17698446996510028,
      "epoch": 0.008625,
      "grad_norm": 0.1299743801355362,
      "learning_rate": 2e-05,
      "loss": 0.21800837516784669,
      "mean_token_accuracy": 0.9389644712209702,
      "num_tokens": 138825070.0,
      "step": 8625
    },
    {
      "aux_loss": 8.051132202148438,
      "entropy": 0.1597856990993023,
      "epoch": 0.00863,
      "grad_norm": 0.13730962574481964,
      "learning_rate": 2e-05,
      "loss": 0.19670662879943848,
      "mean_token_accuracy": 0.9442116737365722,
      "num_tokens": 138905312.0,
      "step": 8630
    },
    {
      "aux_loss": 8.058001804351807,
      "entropy": 0.15900781080126764,
      "epoch": 0.008635,
      "grad_norm": 0.14008685946464539,
      "learning_rate": 2e-05,
      "loss": 0.19398657083511353,
      "mean_token_accuracy": 0.9463704049587249,
      "num_tokens": 138986333.0,
      "step": 8635
    },
    {
      "aux_loss": 8.05928988456726,
      "entropy": 0.16415026746690273,
      "epoch": 0.00864,
      "grad_norm": 0.12896999716758728,
      "learning_rate": 2e-05,
      "loss": 0.192586088180542,
      "mean_token_accuracy": 0.9446914702653885,
      "num_tokens": 139066909.0,
      "step": 8640
    },
    {
      "aux_loss": 8.068218898773193,
      "entropy": 0.14910345450043677,
      "epoch": 0.008645,
      "grad_norm": 0.13533496856689453,
      "learning_rate": 2e-05,
      "loss": 0.1701783299446106,
      "mean_token_accuracy": 0.9525663197040558,
      "num_tokens": 139147854.0,
      "step": 8645
    },
    {
      "aux_loss": 8.07570676803589,
      "entropy": 0.13731190077960492,
      "epoch": 0.00865,
      "grad_norm": 0.12068255245685577,
      "learning_rate": 2e-05,
      "loss": 0.16553701162338258,
      "mean_token_accuracy": 0.9541501373052597,
      "num_tokens": 139225902.0,
      "step": 8650
    },
    {
      "aux_loss": 8.05265121459961,
      "entropy": 0.16186390668153763,
      "epoch": 0.008655,
      "grad_norm": 0.1321835219860077,
      "learning_rate": 2e-05,
      "loss": 0.20611445903778075,
      "mean_token_accuracy": 0.9439223557710648,
      "num_tokens": 139307121.0,
      "step": 8655
    },
    {
      "aux_loss": 8.048187494277954,
      "entropy": 0.16961295604705812,
      "epoch": 0.00866,
      "grad_norm": 0.12428125739097595,
      "learning_rate": 2e-05,
      "loss": 0.21001293659210205,
      "mean_token_accuracy": 0.9411908566951752,
      "num_tokens": 139388288.0,
      "step": 8660
    },
    {
      "aux_loss": 8.054498624801635,
      "entropy": 0.16876161135733128,
      "epoch": 0.008665,
      "grad_norm": 0.1274934709072113,
      "learning_rate": 2e-05,
      "loss": 0.20555145740509034,
      "mean_token_accuracy": 0.9421246469020843,
      "num_tokens": 139468718.0,
      "step": 8665
    },
    {
      "aux_loss": 8.059780836105347,
      "entropy": 0.16027447320520877,
      "epoch": 0.00867,
      "grad_norm": 0.12797242403030396,
      "learning_rate": 2e-05,
      "loss": 0.1871555209159851,
      "mean_token_accuracy": 0.9473997086286545,
      "num_tokens": 139549670.0,
      "step": 8670
    },
    {
      "aux_loss": 8.064860773086547,
      "entropy": 0.14896807856857777,
      "epoch": 0.008675,
      "grad_norm": 0.13439537584781647,
      "learning_rate": 2e-05,
      "loss": 0.17393001317977905,
      "mean_token_accuracy": 0.9516429126262664,
      "num_tokens": 139629546.0,
      "step": 8675
    },
    {
      "aux_loss": 8.07807092666626,
      "entropy": 0.1481859538704157,
      "epoch": 0.00868,
      "grad_norm": 0.14601923525333405,
      "learning_rate": 2e-05,
      "loss": 0.1798541784286499,
      "mean_token_accuracy": 0.9498418360948563,
      "num_tokens": 139707364.0,
      "step": 8680
    },
    {
      "aux_loss": 8.05081205368042,
      "entropy": 0.1466105841100216,
      "epoch": 0.008685,
      "grad_norm": 0.1306084543466568,
      "learning_rate": 2e-05,
      "loss": 0.17915236949920654,
      "mean_token_accuracy": 0.95189129114151,
      "num_tokens": 139788612.0,
      "step": 8685
    },
    {
      "aux_loss": 8.052356147766114,
      "entropy": 0.16529877707362176,
      "epoch": 0.00869,
      "grad_norm": 0.13377101719379425,
      "learning_rate": 2e-05,
      "loss": 0.20282554626464844,
      "mean_token_accuracy": 0.9433024913072586,
      "num_tokens": 139869885.0,
      "step": 8690
    },
    {
      "aux_loss": 8.050832414627076,
      "entropy": 0.16516531445086002,
      "epoch": 0.008695,
      "grad_norm": 0.12875154614448547,
      "learning_rate": 2e-05,
      "loss": 0.20171666145324707,
      "mean_token_accuracy": 0.9431208789348602,
      "num_tokens": 139950740.0,
      "step": 8695
    },
    {
      "aux_loss": 8.056227970123292,
      "entropy": 0.1561167225241661,
      "epoch": 0.0087,
      "grad_norm": 0.13064496219158173,
      "learning_rate": 2e-05,
      "loss": 0.18403198719024658,
      "mean_token_accuracy": 0.9483417361974716,
      "num_tokens": 140031451.0,
      "step": 8700
    },
    {
      "aux_loss": 8.063406038284302,
      "entropy": 0.14633053094148635,
      "epoch": 0.008705,
      "grad_norm": 0.12414082139730453,
      "learning_rate": 2e-05,
      "loss": 0.1717940092086792,
      "mean_token_accuracy": 0.9514490932226181,
      "num_tokens": 140112273.0,
      "step": 8705
    },
    {
      "aux_loss": 8.07530369758606,
      "entropy": 0.14537588655948638,
      "epoch": 0.00871,
      "grad_norm": 0.14689278602600098,
      "learning_rate": 2e-05,
      "loss": 0.17688130140304564,
      "mean_token_accuracy": 0.9504638940095902,
      "num_tokens": 140192712.0,
      "step": 8710
    },
    {
      "aux_loss": 8.053073024749756,
      "entropy": 0.1428489603102207,
      "epoch": 0.008715,
      "grad_norm": 0.12161094695329666,
      "learning_rate": 2e-05,
      "loss": 0.1810328483581543,
      "mean_token_accuracy": 0.9500407069921494,
      "num_tokens": 140270661.0,
      "step": 8715
    },
    {
      "aux_loss": 8.051481294631959,
      "entropy": 0.16562087461352348,
      "epoch": 0.00872,
      "grad_norm": 0.1386113166809082,
      "learning_rate": 2e-05,
      "loss": 0.20505356788635254,
      "mean_token_accuracy": 0.9423316895961762,
      "num_tokens": 140352134.0,
      "step": 8720
    },
    {
      "aux_loss": 8.049740409851074,
      "entropy": 0.1607420913875103,
      "epoch": 0.008725,
      "grad_norm": 0.1264772266149521,
      "learning_rate": 2e-05,
      "loss": 0.19718654155731202,
      "mean_token_accuracy": 0.9457148790359498,
      "num_tokens": 140432667.0,
      "step": 8725
    },
    {
      "aux_loss": 8.056254959106445,
      "entropy": 0.15717368833720685,
      "epoch": 0.00873,
      "grad_norm": 0.13263815641403198,
      "learning_rate": 2e-05,
      "loss": 0.18402861356735228,
      "mean_token_accuracy": 0.9484142869710922,
      "num_tokens": 140513406.0,
      "step": 8730
    },
    {
      "aux_loss": 8.060372877120972,
      "entropy": 0.15108941197395326,
      "epoch": 0.008735,
      "grad_norm": 0.13129562139511108,
      "learning_rate": 2e-05,
      "loss": 0.17795922756195068,
      "mean_token_accuracy": 0.951002436876297,
      "num_tokens": 140593343.0,
      "step": 8735
    },
    {
      "aux_loss": 8.068865633010864,
      "entropy": 0.13550023287534713,
      "epoch": 0.00874,
      "grad_norm": 0.14219635725021362,
      "learning_rate": 2e-05,
      "loss": 0.16220284700393678,
      "mean_token_accuracy": 0.9552569538354874,
      "num_tokens": 140673906.0,
      "step": 8740
    },
    {
      "aux_loss": 8.055937242507934,
      "entropy": 0.1420279834419489,
      "epoch": 0.008745,
      "grad_norm": 0.12697762250900269,
      "learning_rate": 2e-05,
      "loss": 0.18226102590560914,
      "mean_token_accuracy": 0.9506237059831619,
      "num_tokens": 140754779.0,
      "step": 8745
    },
    {
      "aux_loss": 8.046888780593871,
      "entropy": 0.16474676579236985,
      "epoch": 0.00875,
      "grad_norm": 0.1317567527294159,
      "learning_rate": 2e-05,
      "loss": 0.2031479597091675,
      "mean_token_accuracy": 0.9435187935829162,
      "num_tokens": 140836322.0,
      "step": 8750
    },
    {
      "aux_loss": 8.049375629425048,
      "entropy": 0.1640807218849659,
      "epoch": 0.008755,
      "grad_norm": 0.12557213008403778,
      "learning_rate": 2e-05,
      "loss": 0.19852058887481688,
      "mean_token_accuracy": 0.9439573079347611,
      "num_tokens": 140917371.0,
      "step": 8755
    },
    {
      "aux_loss": 8.052918243408204,
      "entropy": 0.16071385629475116,
      "epoch": 0.00876,
      "grad_norm": 0.12216376513242722,
      "learning_rate": 2e-05,
      "loss": 0.18771981000900267,
      "mean_token_accuracy": 0.9486925154924393,
      "num_tokens": 140998020.0,
      "step": 8760
    },
    {
      "aux_loss": 8.059798336029052,
      "entropy": 0.13861278481781483,
      "epoch": 0.008765,
      "grad_norm": 0.1295582801103592,
      "learning_rate": 2e-05,
      "loss": 0.16302660703659058,
      "mean_token_accuracy": 0.9547215700149536,
      "num_tokens": 141078039.0,
      "step": 8765
    },
    {
      "aux_loss": 8.069599533081055,
      "entropy": 0.13799573630094528,
      "epoch": 0.00877,
      "grad_norm": 0.1403513103723526,
      "learning_rate": 2e-05,
      "loss": 0.16963460445404052,
      "mean_token_accuracy": 0.9537664920091629,
      "num_tokens": 141158150.0,
      "step": 8770
    },
    {
      "aux_loss": 8.063237476348878,
      "entropy": 0.15479829013347626,
      "epoch": 0.008775,
      "grad_norm": 0.13407620787620544,
      "learning_rate": 2e-05,
      "loss": 0.18906344175338746,
      "mean_token_accuracy": 0.946771365404129,
      "num_tokens": 141236398.0,
      "step": 8775
    },
    {
      "aux_loss": 8.04686884880066,
      "entropy": 0.17361192852258683,
      "epoch": 0.00878,
      "grad_norm": 0.1317320466041565,
      "learning_rate": 2e-05,
      "loss": 0.21579275131225586,
      "mean_token_accuracy": 0.9400768518447876,
      "num_tokens": 141317447.0,
      "step": 8780
    },
    {
      "aux_loss": 8.05041151046753,
      "entropy": 0.16427178904414177,
      "epoch": 0.008785,
      "grad_norm": 0.12869320809841156,
      "learning_rate": 2e-05,
      "loss": 0.19875767230987548,
      "mean_token_accuracy": 0.9443625956773758,
      "num_tokens": 141398194.0,
      "step": 8785
    },
    {
      "aux_loss": 8.056263542175293,
      "entropy": 0.16415370851755143,
      "epoch": 0.00879,
      "grad_norm": 0.13013140857219696,
      "learning_rate": 2e-05,
      "loss": 0.19186978340148925,
      "mean_token_accuracy": 0.9460722267627716,
      "num_tokens": 141479065.0,
      "step": 8790
    },
    {
      "aux_loss": 8.059855604171753,
      "entropy": 0.15008809715509414,
      "epoch": 0.008795,
      "grad_norm": 0.13298286497592926,
      "learning_rate": 2e-05,
      "loss": 0.17762097120285034,
      "mean_token_accuracy": 0.9506973773241043,
      "num_tokens": 141559578.0,
      "step": 8795
    },
    {
      "aux_loss": 8.070141315460205,
      "entropy": 0.1445841770619154,
      "epoch": 0.0088,
      "grad_norm": 0.13302505016326904,
      "learning_rate": 2e-05,
      "loss": 0.1716432213783264,
      "mean_token_accuracy": 0.9523935735225677,
      "num_tokens": 141639537.0,
      "step": 8800
    },
    {
      "aux_loss": 8.062156009674073,
      "entropy": 0.1437410395592451,
      "epoch": 0.008805,
      "grad_norm": 0.12990376353263855,
      "learning_rate": 2e-05,
      "loss": 0.1756006360054016,
      "mean_token_accuracy": 0.9523844093084335,
      "num_tokens": 141719918.0,
      "step": 8805
    },
    {
      "aux_loss": 8.050819587707519,
      "entropy": 0.16217534951865673,
      "epoch": 0.00881,
      "grad_norm": 0.14529691636562347,
      "learning_rate": 2e-05,
      "loss": 0.20822794437408448,
      "mean_token_accuracy": 0.9423745334148407,
      "num_tokens": 141800620.0,
      "step": 8810
    },
    {
      "aux_loss": 8.051668262481689,
      "entropy": 0.15166816152632237,
      "epoch": 0.008815,
      "grad_norm": 0.12711620330810547,
      "learning_rate": 2e-05,
      "loss": 0.1852765917778015,
      "mean_token_accuracy": 0.9483495861291885,
      "num_tokens": 141881777.0,
      "step": 8815
    },
    {
      "aux_loss": 8.053888511657714,
      "entropy": 0.1557618010789156,
      "epoch": 0.00882,
      "grad_norm": 0.1266649067401886,
      "learning_rate": 2e-05,
      "loss": 0.18496228456497193,
      "mean_token_accuracy": 0.9490999162197113,
      "num_tokens": 141962642.0,
      "step": 8820
    },
    {
      "aux_loss": 8.06244354248047,
      "entropy": 0.15281456038355828,
      "epoch": 0.008825,
      "grad_norm": 0.1309521198272705,
      "learning_rate": 2e-05,
      "loss": 0.18130017518997193,
      "mean_token_accuracy": 0.9499205708503723,
      "num_tokens": 142043191.0,
      "step": 8825
    },
    {
      "aux_loss": 8.070529413223266,
      "entropy": 0.14602159895002842,
      "epoch": 0.00883,
      "grad_norm": 0.13235162198543549,
      "learning_rate": 2e-05,
      "loss": 0.17312501668930053,
      "mean_token_accuracy": 0.9522285014390945,
      "num_tokens": 142123628.0,
      "step": 8830
    },
    {
      "aux_loss": 8.065625190734863,
      "entropy": 0.15181401334702968,
      "epoch": 0.008835,
      "grad_norm": 0.13476738333702087,
      "learning_rate": 2e-05,
      "loss": 0.18659589290618897,
      "mean_token_accuracy": 0.9489993035793305,
      "num_tokens": 142203232.0,
      "step": 8835
    },
    {
      "aux_loss": 8.051854848861694,
      "entropy": 0.15467614158988,
      "epoch": 0.00884,
      "grad_norm": 0.1443680375814438,
      "learning_rate": 2e-05,
      "loss": 0.19742339849472046,
      "mean_token_accuracy": 0.9460978031158447,
      "num_tokens": 142283945.0,
      "step": 8840
    },
    {
      "aux_loss": 8.050033140182496,
      "entropy": 0.1654368791729212,
      "epoch": 0.008845,
      "grad_norm": 0.12915019690990448,
      "learning_rate": 2e-05,
      "loss": 0.20108115673065186,
      "mean_token_accuracy": 0.9435195207595826,
      "num_tokens": 142365144.0,
      "step": 8845
    },
    {
      "aux_loss": 8.05650281906128,
      "entropy": 0.15751716382801534,
      "epoch": 0.00885,
      "grad_norm": 0.12440650165081024,
      "learning_rate": 2e-05,
      "loss": 0.1870269298553467,
      "mean_token_accuracy": 0.9477470695972443,
      "num_tokens": 142445969.0,
      "step": 8850
    },
    {
      "aux_loss": 8.05877652168274,
      "entropy": 0.15325161516666413,
      "epoch": 0.008855,
      "grad_norm": 0.12807123363018036,
      "learning_rate": 2e-05,
      "loss": 0.18139255046844482,
      "mean_token_accuracy": 0.9493809968233109,
      "num_tokens": 142526769.0,
      "step": 8855
    },
    {
      "aux_loss": 8.068867444992065,
      "entropy": 0.14166127592325212,
      "epoch": 0.00886,
      "grad_norm": 0.1370983123779297,
      "learning_rate": 2e-05,
      "loss": 0.16508629322052001,
      "mean_token_accuracy": 0.9536879241466523,
      "num_tokens": 142606917.0,
      "step": 8860
    },
    {
      "aux_loss": 8.068143367767334,
      "entropy": 0.13985994793474674,
      "epoch": 0.008865,
      "grad_norm": 0.1331673115491867,
      "learning_rate": 2e-05,
      "loss": 0.17156301736831664,
      "mean_token_accuracy": 0.9526797980070114,
      "num_tokens": 142686149.0,
      "step": 8865
    },
    {
      "aux_loss": 8.051418209075928,
      "entropy": 0.16913191936910152,
      "epoch": 0.00887,
      "grad_norm": 0.13198046386241913,
      "learning_rate": 2e-05,
      "loss": 0.2088552474975586,
      "mean_token_accuracy": 0.9411212980747223,
      "num_tokens": 142767589.0,
      "step": 8870
    },
    {
      "aux_loss": 8.047991275787354,
      "entropy": 0.17505197674036027,
      "epoch": 0.008875,
      "grad_norm": 0.13860420882701874,
      "learning_rate": 2e-05,
      "loss": 0.21608192920684816,
      "mean_token_accuracy": 0.9397012293338776,
      "num_tokens": 142848136.0,
      "step": 8875
    },
    {
      "aux_loss": 8.051774311065675,
      "entropy": 0.16211238801479338,
      "epoch": 0.00888,
      "grad_norm": 0.1306205838918686,
      "learning_rate": 2e-05,
      "loss": 0.19351627826690673,
      "mean_token_accuracy": 0.9458301037549972,
      "num_tokens": 142929218.0,
      "step": 8880
    },
    {
      "aux_loss": 8.060734033584595,
      "entropy": 0.15532847866415977,
      "epoch": 0.008885,
      "grad_norm": 0.13024985790252686,
      "learning_rate": 2e-05,
      "loss": 0.18095829486846923,
      "mean_token_accuracy": 0.9493039816617965,
      "num_tokens": 143010029.0,
      "step": 8885
    },
    {
      "aux_loss": 8.066394472122193,
      "entropy": 0.13877962827682494,
      "epoch": 0.00889,
      "grad_norm": 0.13255876302719116,
      "learning_rate": 2e-05,
      "loss": 0.1647440552711487,
      "mean_token_accuracy": 0.9544006258249282,
      "num_tokens": 143090694.0,
      "step": 8890
    },
    {
      "aux_loss": 8.071954727172852,
      "entropy": 0.1325362291187048,
      "epoch": 0.008895,
      "grad_norm": 0.12084479629993439,
      "learning_rate": 2e-05,
      "loss": 0.15787626504898072,
      "mean_token_accuracy": 0.9556850016117096,
      "num_tokens": 143168492.0,
      "step": 8895
    },
    {
      "aux_loss": 8.04944043159485,
      "entropy": 0.1741876568645239,
      "epoch": 0.0089,
      "grad_norm": 0.13142059743404388,
      "learning_rate": 2e-05,
      "loss": 0.22046067714691162,
      "mean_token_accuracy": 0.9394972771406174,
      "num_tokens": 143249591.0,
      "step": 8900
    },
    {
      "aux_loss": 8.0515212059021,
      "entropy": 0.151621338352561,
      "epoch": 0.008905,
      "grad_norm": 0.13156643509864807,
      "learning_rate": 2e-05,
      "loss": 0.1894966959953308,
      "mean_token_accuracy": 0.9466851204633713,
      "num_tokens": 143330373.0,
      "step": 8905
    },
    {
      "aux_loss": 8.052351999282838,
      "entropy": 0.15956550613045692,
      "epoch": 0.00891,
      "grad_norm": 0.12576551735401154,
      "learning_rate": 2e-05,
      "loss": 0.19171146154403687,
      "mean_token_accuracy": 0.9459272176027298,
      "num_tokens": 143411203.0,
      "step": 8910
    },
    {
      "aux_loss": 8.055167388916015,
      "entropy": 0.16308680921792984,
      "epoch": 0.008915,
      "grad_norm": 0.13080812990665436,
      "learning_rate": 2e-05,
      "loss": 0.19286198616027833,
      "mean_token_accuracy": 0.9454808354377746,
      "num_tokens": 143491925.0,
      "step": 8915
    },
    {
      "aux_loss": 8.064648008346557,
      "entropy": 0.1466072913259268,
      "epoch": 0.00892,
      "grad_norm": 0.13442130386829376,
      "learning_rate": 2e-05,
      "loss": 0.17161482572555542,
      "mean_token_accuracy": 0.9522063434123993,
      "num_tokens": 143572940.0,
      "step": 8920
    },
    {
      "aux_loss": 8.07026448249817,
      "entropy": 0.140037352591753,
      "epoch": 0.008925,
      "grad_norm": 0.11821462213993073,
      "learning_rate": 2e-05,
      "loss": 0.16644347906112672,
      "mean_token_accuracy": 0.9547454357147217,
      "num_tokens": 143651762.0,
      "step": 8925
    },
    {
      "aux_loss": 8.049960374832153,
      "entropy": 0.168941380828619,
      "epoch": 0.00893,
      "grad_norm": 0.13707335293293,
      "learning_rate": 2e-05,
      "loss": 0.21055347919464112,
      "mean_token_accuracy": 0.9414200127124787,
      "num_tokens": 143733269.0,
      "step": 8930
    },
    {
      "aux_loss": 8.047146701812744,
      "entropy": 0.1807384926825762,
      "epoch": 0.008935,
      "grad_norm": 0.17462655901908875,
      "learning_rate": 2e-05,
      "loss": 0.2216662883758545,
      "mean_token_accuracy": 0.9369146227836609,
      "num_tokens": 143813814.0,
      "step": 8935
    },
    {
      "aux_loss": 8.053944206237793,
      "entropy": 0.15766459032893182,
      "epoch": 0.00894,
      "grad_norm": 0.13653764128684998,
      "learning_rate": 2e-05,
      "loss": 0.19147311449050902,
      "mean_token_accuracy": 0.946722275018692,
      "num_tokens": 143894598.0,
      "step": 8940
    },
    {
      "aux_loss": 8.056559181213379,
      "entropy": 0.1534650482237339,
      "epoch": 0.008945,
      "grad_norm": 0.12604378163814545,
      "learning_rate": 2e-05,
      "loss": 0.18191812038421631,
      "mean_token_accuracy": 0.9490950018167496,
      "num_tokens": 143975134.0,
      "step": 8945
    },
    {
      "aux_loss": 8.062952136993408,
      "entropy": 0.1457659013569355,
      "epoch": 0.00895,
      "grad_norm": 0.12857455015182495,
      "learning_rate": 2e-05,
      "loss": 0.16952617168426515,
      "mean_token_accuracy": 0.9529273092746735,
      "num_tokens": 144055863.0,
      "step": 8950
    },
    {
      "aux_loss": 8.073321533203124,
      "entropy": 0.13516576513648032,
      "epoch": 0.008955,
      "grad_norm": 0.12914860248565674,
      "learning_rate": 2e-05,
      "loss": 0.16308298110961914,
      "mean_token_accuracy": 0.9545054614543915,
      "num_tokens": 144133736.0,
      "step": 8955
    },
    {
      "aux_loss": 8.047145748138428,
      "entropy": 0.1553576286882162,
      "epoch": 0.00896,
      "grad_norm": 0.13375549018383026,
      "learning_rate": 2e-05,
      "loss": 0.19422754049301147,
      "mean_token_accuracy": 0.9465140253305435,
      "num_tokens": 144215187.0,
      "step": 8960
    },
    {
      "aux_loss": 8.048073196411133,
      "entropy": 0.1687317706644535,
      "epoch": 0.008965,
      "grad_norm": 0.1284179538488388,
      "learning_rate": 2e-05,
      "loss": 0.20868115425109862,
      "mean_token_accuracy": 0.9416483998298645,
      "num_tokens": 144296078.0,
      "step": 8965
    },
    {
      "aux_loss": 8.051653337478637,
      "entropy": 0.17172398492693902,
      "epoch": 0.00897,
      "grad_norm": 0.13168437778949738,
      "learning_rate": 2e-05,
      "loss": 0.2059194564819336,
      "mean_token_accuracy": 0.942954096198082,
      "num_tokens": 144376988.0,
      "step": 8970
    },
    {
      "aux_loss": 8.056134557723999,
      "entropy": 0.16314014792442322,
      "epoch": 0.008975,
      "grad_norm": 0.1363796591758728,
      "learning_rate": 2e-05,
      "loss": 0.19433751106262206,
      "mean_token_accuracy": 0.9456538796424866,
      "num_tokens": 144457297.0,
      "step": 8975
    },
    {
      "aux_loss": 8.064738988876343,
      "entropy": 0.15046918019652367,
      "epoch": 0.00898,
      "grad_norm": 0.12787559628486633,
      "learning_rate": 2e-05,
      "loss": 0.1770918607711792,
      "mean_token_accuracy": 0.9500804215669632,
      "num_tokens": 144538120.0,
      "step": 8980
    },
    {
      "aux_loss": 8.078823328018188,
      "entropy": 0.14470162987709045,
      "epoch": 0.008985,
      "grad_norm": 0.15308259427547455,
      "learning_rate": 2e-05,
      "loss": 0.17147524356842042,
      "mean_token_accuracy": 0.9527350008487702,
      "num_tokens": 144615756.0,
      "step": 8985
    },
    {
      "aux_loss": 8.053917407989502,
      "entropy": 0.16611836478114128,
      "epoch": 0.00899,
      "grad_norm": 0.14119122922420502,
      "learning_rate": 2e-05,
      "loss": 0.20463342666625978,
      "mean_token_accuracy": 0.9434159636497498,
      "num_tokens": 144696882.0,
      "step": 8990
    },
    {
      "aux_loss": 8.05145330429077,
      "entropy": 0.1684564959257841,
      "epoch": 0.008995,
      "grad_norm": 0.1359567940235138,
      "learning_rate": 2e-05,
      "loss": 0.20629723072052003,
      "mean_token_accuracy": 0.9411316961050034,
      "num_tokens": 144778403.0,
      "step": 8995
    },
    {
      "aux_loss": 8.051396226882934,
      "entropy": 0.16917419098317624,
      "epoch": 0.009,
      "grad_norm": 0.12857814133167267,
      "learning_rate": 2e-05,
      "loss": 0.20049822330474854,
      "mean_token_accuracy": 0.9434998363256455,
      "num_tokens": 144859171.0,
      "step": 9000
    },
    {
      "aux_loss": 8.055720281600951,
      "entropy": 0.16034072861075402,
      "epoch": 0.009005,
      "grad_norm": 0.13157115876674652,
      "learning_rate": 2e-05,
      "loss": 0.1890867233276367,
      "mean_token_accuracy": 0.9466980814933776,
      "num_tokens": 144939851.0,
      "step": 9005
    },
    {
      "aux_loss": 8.061790704727173,
      "entropy": 0.1485297095030546,
      "epoch": 0.00901,
      "grad_norm": 0.12871001660823822,
      "learning_rate": 2e-05,
      "loss": 0.17670243978500366,
      "mean_token_accuracy": 0.9509409308433533,
      "num_tokens": 145020422.0,
      "step": 9010
    },
    {
      "aux_loss": 8.071242952346802,
      "entropy": 0.14230466708540918,
      "epoch": 0.009015,
      "grad_norm": 0.13953736424446106,
      "learning_rate": 2e-05,
      "loss": 0.17031856775283813,
      "mean_token_accuracy": 0.9534539580345154,
      "num_tokens": 145100765.0,
      "step": 9015
    },
    {
      "aux_loss": 8.053312683105469,
      "entropy": 0.14188128374516965,
      "epoch": 0.00902,
      "grad_norm": 0.1283850222826004,
      "learning_rate": 2e-05,
      "loss": 0.18063976764678955,
      "mean_token_accuracy": 0.9500974535942077,
      "num_tokens": 145181867.0,
      "step": 9020
    },
    {
      "aux_loss": 8.050068616867065,
      "entropy": 0.15850713439285755,
      "epoch": 0.009025,
      "grad_norm": 0.1412983387708664,
      "learning_rate": 2e-05,
      "loss": 0.19685086011886596,
      "mean_token_accuracy": 0.9451803863048553,
      "num_tokens": 145263150.0,
      "step": 9025
    },
    {
      "aux_loss": 8.046318244934081,
      "entropy": 0.1529967050999403,
      "epoch": 0.00903,
      "grad_norm": 0.1289587765932083,
      "learning_rate": 2e-05,
      "loss": 0.1920097589492798,
      "mean_token_accuracy": 0.9464766293764114,
      "num_tokens": 145344167.0,
      "step": 9030
    },
    {
      "aux_loss": 8.05451159477234,
      "entropy": 0.1611484669148922,
      "epoch": 0.009035,
      "grad_norm": 0.12812398374080658,
      "learning_rate": 2e-05,
      "loss": 0.19337185621261596,
      "mean_token_accuracy": 0.9459359347820282,
      "num_tokens": 145424886.0,
      "step": 9035
    },
    {
      "aux_loss": 8.063358736038207,
      "entropy": 0.1576306439936161,
      "epoch": 0.00904,
      "grad_norm": 0.12644045054912567,
      "learning_rate": 2e-05,
      "loss": 0.1833519458770752,
      "mean_token_accuracy": 0.9484833687543869,
      "num_tokens": 145505588.0,
      "step": 9040
    },
    {
      "aux_loss": 8.069869709014892,
      "entropy": 0.13947059065103531,
      "epoch": 0.009045,
      "grad_norm": 0.136777862906456,
      "learning_rate": 2e-05,
      "loss": 0.16213153600692748,
      "mean_token_accuracy": 0.954826545715332,
      "num_tokens": 145585934.0,
      "step": 9045
    },
    {
      "aux_loss": 8.054084968566894,
      "entropy": 0.15392575040459633,
      "epoch": 0.00905,
      "grad_norm": 0.1301695853471756,
      "learning_rate": 2e-05,
      "loss": 0.19046540260314943,
      "mean_token_accuracy": 0.9471210896968841,
      "num_tokens": 145666589.0,
      "step": 9050
    },
    {
      "aux_loss": 8.04963812828064,
      "entropy": 0.1462317556142807,
      "epoch": 0.009055,
      "grad_norm": 0.14226828515529633,
      "learning_rate": 2e-05,
      "loss": 0.18507881164550782,
      "mean_token_accuracy": 0.9491311490535737,
      "num_tokens": 145747944.0,
      "step": 9055
    },
    {
      "aux_loss": 8.049897718429566,
      "entropy": 0.1655551202595234,
      "epoch": 0.00906,
      "grad_norm": 0.13290181756019592,
      "learning_rate": 2e-05,
      "loss": 0.20115478038787843,
      "mean_token_accuracy": 0.9442602813243866,
      "num_tokens": 145828834.0,
      "step": 9060
    },
    {
      "aux_loss": 8.053511095046996,
      "entropy": 0.15808416530489922,
      "epoch": 0.009065,
      "grad_norm": 0.125838503241539,
      "learning_rate": 2e-05,
      "loss": 0.1863319158554077,
      "mean_token_accuracy": 0.948188191652298,
      "num_tokens": 145909366.0,
      "step": 9065
    },
    {
      "aux_loss": 8.058942651748657,
      "entropy": 0.1557976320385933,
      "epoch": 0.00907,
      "grad_norm": 0.1331806480884552,
      "learning_rate": 2e-05,
      "loss": 0.1798400640487671,
      "mean_token_accuracy": 0.9495758682489395,
      "num_tokens": 145989592.0,
      "step": 9070
    },
    {
      "aux_loss": 8.070401811599732,
      "entropy": 0.14249608814716339,
      "epoch": 0.009075,
      "grad_norm": 0.1402416229248047,
      "learning_rate": 2e-05,
      "loss": 0.17083990573883057,
      "mean_token_accuracy": 0.9521846741437912,
      "num_tokens": 146070084.0,
      "step": 9075
    },
    {
      "aux_loss": 8.052979326248169,
      "entropy": 0.1582282304763794,
      "epoch": 0.00908,
      "grad_norm": 0.12918207049369812,
      "learning_rate": 2e-05,
      "loss": 0.19609121084213257,
      "mean_token_accuracy": 0.9468253374099731,
      "num_tokens": 146150901.0,
      "step": 9080
    },
    {
      "aux_loss": 8.046459197998047,
      "entropy": 0.15578912794589997,
      "epoch": 0.009085,
      "grad_norm": 0.13867315649986267,
      "learning_rate": 2e-05,
      "loss": 0.19550650119781493,
      "mean_token_accuracy": 0.9448387354612351,
      "num_tokens": 146232225.0,
      "step": 9085
    },
    {
      "aux_loss": 8.048712778091431,
      "entropy": 0.1605131860822439,
      "epoch": 0.00909,
      "grad_norm": 0.12908585369586945,
      "learning_rate": 2e-05,
      "loss": 0.19709932804107666,
      "mean_token_accuracy": 0.9448313176631927,
      "num_tokens": 146312867.0,
      "step": 9090
    },
    {
      "aux_loss": 8.05649242401123,
      "entropy": 0.16510931104421617,
      "epoch": 0.009095,
      "grad_norm": 0.12659822404384613,
      "learning_rate": 2e-05,
      "loss": 0.19316548109054565,
      "mean_token_accuracy": 0.9458854496479034,
      "num_tokens": 146393782.0,
      "step": 9095
    },
    {
      "aux_loss": 8.060756492614747,
      "entropy": 0.15448057651519775,
      "epoch": 0.0091,
      "grad_norm": 0.13716357946395874,
      "learning_rate": 2e-05,
      "loss": 0.18324564695358275,
      "mean_token_accuracy": 0.9493474274873733,
      "num_tokens": 146474610.0,
      "step": 9100
    },
    {
      "aux_loss": 8.069791078567505,
      "entropy": 0.14029395207762718,
      "epoch": 0.009105,
      "grad_norm": 0.15031740069389343,
      "learning_rate": 2e-05,
      "loss": 0.16438581943511962,
      "mean_token_accuracy": 0.9537094175815582,
      "num_tokens": 146554908.0,
      "step": 9105
    },
    {
      "aux_loss": 8.051681613922119,
      "entropy": 0.15240961983799933,
      "epoch": 0.00911,
      "grad_norm": 0.1311596930027008,
      "learning_rate": 2e-05,
      "loss": 0.19222482442855834,
      "mean_token_accuracy": 0.9482855081558228,
      "num_tokens": 146635858.0,
      "step": 9110
    },
    {
      "aux_loss": 8.046053886413574,
      "entropy": 0.17746356800198554,
      "epoch": 0.009115,
      "grad_norm": 0.13681170344352722,
      "learning_rate": 2e-05,
      "loss": 0.21867761611938477,
      "mean_token_accuracy": 0.9398280680179596,
      "num_tokens": 146717222.0,
      "step": 9115
    },
    {
      "aux_loss": 8.047878742218018,
      "entropy": 0.14980819076299667,
      "epoch": 0.00912,
      "grad_norm": 0.12490449845790863,
      "learning_rate": 2e-05,
      "loss": 0.1871268033981323,
      "mean_token_accuracy": 0.9484641134738923,
      "num_tokens": 146798068.0,
      "step": 9120
    },
    {
      "aux_loss": 8.057613420486451,
      "entropy": 0.16235484033823014,
      "epoch": 0.009125,
      "grad_norm": 0.13405996561050415,
      "learning_rate": 2e-05,
      "loss": 0.19359736442565917,
      "mean_token_accuracy": 0.9450442284345627,
      "num_tokens": 146878773.0,
      "step": 9125
    },
    {
      "aux_loss": 8.060540294647216,
      "entropy": 0.1512028731405735,
      "epoch": 0.00913,
      "grad_norm": 0.13461905717849731,
      "learning_rate": 2e-05,
      "loss": 0.17625993490219116,
      "mean_token_accuracy": 0.9509659796953202,
      "num_tokens": 146959463.0,
      "step": 9130
    },
    {
      "aux_loss": 8.071784734725952,
      "entropy": 0.15256218314170839,
      "epoch": 0.009135,
      "grad_norm": 0.1549593210220337,
      "learning_rate": 2e-05,
      "loss": 0.1816193699836731,
      "mean_token_accuracy": 0.9490591704845428,
      "num_tokens": 147040075.0,
      "step": 9135
    },
    {
      "aux_loss": 8.054101610183716,
      "entropy": 0.1635272864252329,
      "epoch": 0.00914,
      "grad_norm": 0.1223096251487732,
      "learning_rate": 2e-05,
      "loss": 0.200970458984375,
      "mean_token_accuracy": 0.9452550083398819,
      "num_tokens": 147120148.0,
      "step": 9140
    },
    {
      "aux_loss": 8.047113227844239,
      "entropy": 0.1671906251460314,
      "epoch": 0.009145,
      "grad_norm": 0.1311800479888916,
      "learning_rate": 2e-05,
      "loss": 0.202933931350708,
      "mean_token_accuracy": 0.9446672409772873,
      "num_tokens": 147201725.0,
      "step": 9145
    },
    {
      "aux_loss": 8.051253843307496,
      "entropy": 0.17207262180745603,
      "epoch": 0.00915,
      "grad_norm": 0.13126836717128754,
      "learning_rate": 2e-05,
      "loss": 0.20531783103942872,
      "mean_token_accuracy": 0.9422431230545044,
      "num_tokens": 147282800.0,
      "step": 9150
    },
    {
      "aux_loss": 8.05524377822876,
      "entropy": 0.15392919145524503,
      "epoch": 0.009155,
      "grad_norm": 0.1381392627954483,
      "learning_rate": 2e-05,
      "loss": 0.18467797040939332,
      "mean_token_accuracy": 0.9486317545175552,
      "num_tokens": 147363409.0,
      "step": 9155
    },
    {
      "aux_loss": 8.060203075408936,
      "entropy": 0.14846806265413762,
      "epoch": 0.00916,
      "grad_norm": 0.1275726705789566,
      "learning_rate": 2e-05,
      "loss": 0.17700949907302857,
      "mean_token_accuracy": 0.9502640396356583,
      "num_tokens": 147443900.0,
      "step": 9160
    },
    {
      "aux_loss": 8.066796827316285,
      "entropy": 0.14137993082404138,
      "epoch": 0.009165,
      "grad_norm": 0.1385326385498047,
      "learning_rate": 2e-05,
      "loss": 0.16897222995758057,
      "mean_token_accuracy": 0.9529264569282532,
      "num_tokens": 147524202.0,
      "step": 9165
    },
    {
      "aux_loss": 8.055108118057252,
      "entropy": 0.17844887264072895,
      "epoch": 0.00917,
      "grad_norm": 0.1283971071243286,
      "learning_rate": 2e-05,
      "loss": 0.21684696674346923,
      "mean_token_accuracy": 0.9413057148456574,
      "num_tokens": 147604113.0,
      "step": 9170
    },
    {
      "aux_loss": 8.048505544662476,
      "entropy": 0.16603213101625441,
      "epoch": 0.009175,
      "grad_norm": 0.12222763895988464,
      "learning_rate": 2e-05,
      "loss": 0.20498316287994384,
      "mean_token_accuracy": 0.9432746320962906,
      "num_tokens": 147685017.0,
      "step": 9175
    },
    {
      "aux_loss": 8.048001146316528,
      "entropy": 0.1670322071760893,
      "epoch": 0.00918,
      "grad_norm": 0.1308424472808838,
      "learning_rate": 2e-05,
      "loss": 0.20279350280761718,
      "mean_token_accuracy": 0.9437865585088729,
      "num_tokens": 147766105.0,
      "step": 9180
    },
    {
      "aux_loss": 8.054187297821045,
      "entropy": 0.1686132237315178,
      "epoch": 0.009185,
      "grad_norm": 0.1288473904132843,
      "learning_rate": 2e-05,
      "loss": 0.19990822076797485,
      "mean_token_accuracy": 0.9437294691801071,
      "num_tokens": 147847262.0,
      "step": 9185
    },
    {
      "aux_loss": 8.059163904190063,
      "entropy": 0.15092508792877196,
      "epoch": 0.00919,
      "grad_norm": 0.1380072981119156,
      "learning_rate": 2e-05,
      "loss": 0.17768815755844117,
      "mean_token_accuracy": 0.9510469049215317,
      "num_tokens": 147927576.0,
      "step": 9190
    },
    {
      "aux_loss": 8.065919303894043,
      "entropy": 0.14715459160506725,
      "epoch": 0.009195,
      "grad_norm": 0.14201250672340393,
      "learning_rate": 2e-05,
      "loss": 0.17172386646270751,
      "mean_token_accuracy": 0.9520582288503647,
      "num_tokens": 148007903.0,
      "step": 9195
    },
    {
      "aux_loss": 8.05676827430725,
      "entropy": 0.15409750305116177,
      "epoch": 0.0092,
      "grad_norm": 0.12849180400371552,
      "learning_rate": 2e-05,
      "loss": 0.19143229722976685,
      "mean_token_accuracy": 0.9480628252029419,
      "num_tokens": 148088823.0,
      "step": 9200
    },
    {
      "aux_loss": 8.05044469833374,
      "entropy": 0.19112828448414804,
      "epoch": 0.009205,
      "grad_norm": 0.13316133618354797,
      "learning_rate": 2e-05,
      "loss": 0.23454911708831788,
      "mean_token_accuracy": 0.9347047060728073,
      "num_tokens": 148170251.0,
      "step": 9205
    },
    {
      "aux_loss": 8.049229526519776,
      "entropy": 0.16306887194514275,
      "epoch": 0.00921,
      "grad_norm": 0.12697415053844452,
      "learning_rate": 2e-05,
      "loss": 0.19756876230239867,
      "mean_token_accuracy": 0.9448862552642823,
      "num_tokens": 148250965.0,
      "step": 9210
    },
    {
      "aux_loss": 8.05115294456482,
      "entropy": 0.15382013358175756,
      "epoch": 0.009215,
      "grad_norm": 0.1288222223520279,
      "learning_rate": 2e-05,
      "loss": 0.18505736589431762,
      "mean_token_accuracy": 0.9490427553653717,
      "num_tokens": 148332150.0,
      "step": 9215
    },
    {
      "aux_loss": 8.057027864456177,
      "entropy": 0.1658315747976303,
      "epoch": 0.00922,
      "grad_norm": 0.12217347323894501,
      "learning_rate": 2e-05,
      "loss": 0.189755117893219,
      "mean_token_accuracy": 0.9472373455762864,
      "num_tokens": 148412683.0,
      "step": 9220
    },
    {
      "aux_loss": 8.065322017669677,
      "entropy": 0.1417682718485594,
      "epoch": 0.009225,
      "grad_norm": 0.14299456775188446,
      "learning_rate": 2e-05,
      "loss": 0.1710767388343811,
      "mean_token_accuracy": 0.9516008555889129,
      "num_tokens": 148493362.0,
      "step": 9225
    },
    {
      "aux_loss": 8.064607858657837,
      "entropy": 0.13697895631194115,
      "epoch": 0.00923,
      "grad_norm": 0.1288357973098755,
      "learning_rate": 2e-05,
      "loss": 0.16802023649215697,
      "mean_token_accuracy": 0.954392597079277,
      "num_tokens": 148573368.0,
      "step": 9230
    },
    {
      "aux_loss": 8.051961183547974,
      "entropy": 0.16174781173467637,
      "epoch": 0.009235,
      "grad_norm": 0.12534639239311218,
      "learning_rate": 2e-05,
      "loss": 0.19824026823043822,
      "mean_token_accuracy": 0.945355799794197,
      "num_tokens": 148654770.0,
      "step": 9235
    },
    {
      "aux_loss": 8.047585153579712,
      "entropy": 0.18336153365671634,
      "epoch": 0.00924,
      "grad_norm": 0.12391963601112366,
      "learning_rate": 2e-05,
      "loss": 0.22359731197357177,
      "mean_token_accuracy": 0.9378106683492661,
      "num_tokens": 148735336.0,
      "step": 9240
    },
    {
      "aux_loss": 8.050562191009522,
      "entropy": 0.16212284602224827,
      "epoch": 0.009245,
      "grad_norm": 0.12470485270023346,
      "learning_rate": 2e-05,
      "loss": 0.1927903413772583,
      "mean_token_accuracy": 0.9465355992317199,
      "num_tokens": 148816387.0,
      "step": 9245
    },
    {
      "aux_loss": 8.057068967819214,
      "entropy": 0.15883617624640464,
      "epoch": 0.00925,
      "grad_norm": 0.13021220266819,
      "learning_rate": 2e-05,
      "loss": 0.1838752031326294,
      "mean_token_accuracy": 0.94894839823246,
      "num_tokens": 148897298.0,
      "step": 9250
    },
    {
      "aux_loss": 8.062517213821412,
      "entropy": 0.14380777589976787,
      "epoch": 0.009255,
      "grad_norm": 0.13678789138793945,
      "learning_rate": 2e-05,
      "loss": 0.17271702289581298,
      "mean_token_accuracy": 0.9515684425830842,
      "num_tokens": 148978453.0,
      "step": 9255
    },
    {
      "aux_loss": 8.0665846824646,
      "entropy": 0.14376781955361367,
      "epoch": 0.00926,
      "grad_norm": 0.12270132452249527,
      "learning_rate": 2e-05,
      "loss": 0.1778918147087097,
      "mean_token_accuracy": 0.9517032414674759,
      "num_tokens": 149057082.0,
      "step": 9260
    },
    {
      "aux_loss": 8.05322332382202,
      "entropy": 0.16869215443730354,
      "epoch": 0.009265,
      "grad_norm": 0.14173255860805511,
      "learning_rate": 2e-05,
      "loss": 0.20956835746765137,
      "mean_token_accuracy": 0.9413413912057876,
      "num_tokens": 149138132.0,
      "step": 9265
    },
    {
      "aux_loss": 8.046040391921997,
      "entropy": 0.16795603334903716,
      "epoch": 0.00927,
      "grad_norm": 0.12431542575359344,
      "learning_rate": 2e-05,
      "loss": 0.20726256370544432,
      "mean_token_accuracy": 0.9420913815498352,
      "num_tokens": 149219229.0,
      "step": 9270
    },
    {
      "aux_loss": 8.052736711502074,
      "entropy": 0.1785639014095068,
      "epoch": 0.009275,
      "grad_norm": 0.13082771003246307,
      "learning_rate": 2e-05,
      "loss": 0.20987801551818847,
      "mean_token_accuracy": 0.9412637978792191,
      "num_tokens": 149300103.0,
      "step": 9275
    },
    {
      "aux_loss": 8.055554676055909,
      "entropy": 0.1684865839779377,
      "epoch": 0.00928,
      "grad_norm": 0.12718334794044495,
      "learning_rate": 2e-05,
      "loss": 0.19301304817199708,
      "mean_token_accuracy": 0.9460320919752121,
      "num_tokens": 149380463.0,
      "step": 9280
    },
    {
      "aux_loss": 8.060051774978637,
      "entropy": 0.16233376264572144,
      "epoch": 0.009285,
      "grad_norm": 0.13252019882202148,
      "learning_rate": 2e-05,
      "loss": 0.18949836492538452,
      "mean_token_accuracy": 0.9469490438699723,
      "num_tokens": 149460212.0,
      "step": 9285
    },
    {
      "aux_loss": 8.073284721374511,
      "entropy": 0.1458605956286192,
      "epoch": 0.00929,
      "grad_norm": 0.15022385120391846,
      "learning_rate": 2e-05,
      "loss": 0.17203078269958497,
      "mean_token_accuracy": 0.951647412776947,
      "num_tokens": 149538025.0,
      "step": 9290
    },
    {
      "aux_loss": 8.04990553855896,
      "entropy": 0.1576648410409689,
      "epoch": 0.009295,
      "grad_norm": 0.13055454194545746,
      "learning_rate": 2e-05,
      "loss": 0.19502304792404174,
      "mean_token_accuracy": 0.9459043979644776,
      "num_tokens": 149619267.0,
      "step": 9295
    },
    {
      "aux_loss": 8.051372957229614,
      "entropy": 0.18589525669813156,
      "epoch": 0.0093,
      "grad_norm": 0.12683400511741638,
      "learning_rate": 2e-05,
      "loss": 0.22710530757904052,
      "mean_token_accuracy": 0.9372732311487197,
      "num_tokens": 149700369.0,
      "step": 9300
    },
    {
      "aux_loss": 8.049410343170166,
      "entropy": 0.16699933186173438,
      "epoch": 0.009305,
      "grad_norm": 0.13137462735176086,
      "learning_rate": 2e-05,
      "loss": 0.2062304735183716,
      "mean_token_accuracy": 0.9432043164968491,
      "num_tokens": 149781352.0,
      "step": 9305
    },
    {
      "aux_loss": 8.055490255355835,
      "entropy": 0.15752097703516482,
      "epoch": 0.00931,
      "grad_norm": 0.13234452903270721,
      "learning_rate": 2e-05,
      "loss": 0.18671668767929078,
      "mean_token_accuracy": 0.9475500673055649,
      "num_tokens": 149862178.0,
      "step": 9310
    },
    {
      "aux_loss": 8.061890697479248,
      "entropy": 0.16386120915412902,
      "epoch": 0.009315,
      "grad_norm": 0.1288726031780243,
      "learning_rate": 2e-05,
      "loss": 0.18764032125473024,
      "mean_token_accuracy": 0.9478613406419754,
      "num_tokens": 149942677.0,
      "step": 9315
    },
    {
      "aux_loss": 8.06913390159607,
      "entropy": 0.1462075099349022,
      "epoch": 0.00932,
      "grad_norm": 0.14157095551490784,
      "learning_rate": 2e-05,
      "loss": 0.16920729875564575,
      "mean_token_accuracy": 0.9530714064836502,
      "num_tokens": 150023246.0,
      "step": 9320
    },
    {
      "aux_loss": 8.057884168624877,
      "entropy": 0.161289606615901,
      "epoch": 0.009325,
      "grad_norm": 0.15061193704605103,
      "learning_rate": 2e-05,
      "loss": 0.1942088484764099,
      "mean_token_accuracy": 0.9444399297237396,
      "num_tokens": 150100414.0,
      "step": 9325
    },
    {
      "aux_loss": 8.047045135498047,
      "entropy": 0.1670971255749464,
      "epoch": 0.00933,
      "grad_norm": 0.14655718207359314,
      "learning_rate": 2e-05,
      "loss": 0.20956218242645264,
      "mean_token_accuracy": 0.9412062048912049,
      "num_tokens": 150181739.0,
      "step": 9330
    },
    {
      "aux_loss": 8.0474187374115,
      "entropy": 0.17152599543333052,
      "epoch": 0.009335,
      "grad_norm": 0.13407909870147705,
      "learning_rate": 2e-05,
      "loss": 0.20760955810546874,
      "mean_token_accuracy": 0.9430797725915909,
      "num_tokens": 150262474.0,
      "step": 9335
    },
    {
      "aux_loss": 8.052974796295166,
      "entropy": 0.16562261804938316,
      "epoch": 0.00934,
      "grad_norm": 0.12974247336387634,
      "learning_rate": 2e-05,
      "loss": 0.1965591073036194,
      "mean_token_accuracy": 0.944324043393135,
      "num_tokens": 150343156.0,
      "step": 9340
    },
    {
      "aux_loss": 8.059053182601929,
      "entropy": 0.16600474789738656,
      "epoch": 0.009345,
      "grad_norm": 0.13165518641471863,
      "learning_rate": 2e-05,
      "loss": 0.18874115943908693,
      "mean_token_accuracy": 0.9474812120199203,
      "num_tokens": 150423862.0,
      "step": 9345
    },
    {
      "aux_loss": 8.068104648590088,
      "entropy": 0.14390633776783943,
      "epoch": 0.00935,
      "grad_norm": 0.1445103883743286,
      "learning_rate": 2e-05,
      "loss": 0.17139971256256104,
      "mean_token_accuracy": 0.951501852273941,
      "num_tokens": 150503705.0,
      "step": 9350
    },
    {
      "aux_loss": 8.054257488250732,
      "entropy": 0.17811726741492748,
      "epoch": 0.009355,
      "grad_norm": 0.1314753293991089,
      "learning_rate": 2e-05,
      "loss": 0.21900644302368164,
      "mean_token_accuracy": 0.9397265940904618,
      "num_tokens": 150584678.0,
      "step": 9355
    },
    {
      "aux_loss": 8.048040676116944,
      "entropy": 0.17390961721539497,
      "epoch": 0.00936,
      "grad_norm": 0.12444669008255005,
      "learning_rate": 2e-05,
      "loss": 0.20949745178222656,
      "mean_token_accuracy": 0.9416788220405579,
      "num_tokens": 150665669.0,
      "step": 9360
    },
    {
      "aux_loss": 8.051009035110473,
      "entropy": 0.1653123665601015,
      "epoch": 0.009365,
      "grad_norm": 0.12438629567623138,
      "learning_rate": 2e-05,
      "loss": 0.2017763614654541,
      "mean_token_accuracy": 0.9444821387529373,
      "num_tokens": 150746692.0,
      "step": 9365
    },
    {
      "aux_loss": 8.056476259231568,
      "entropy": 0.16538954228162767,
      "epoch": 0.00937,
      "grad_norm": 0.1241748109459877,
      "learning_rate": 2e-05,
      "loss": 0.1942046880722046,
      "mean_token_accuracy": 0.9446871757507325,
      "num_tokens": 150827509.0,
      "step": 9370
    },
    {
      "aux_loss": 8.059260368347168,
      "entropy": 0.1478948902338743,
      "epoch": 0.009375,
      "grad_norm": 0.1315631866455078,
      "learning_rate": 2e-05,
      "loss": 0.1773110508918762,
      "mean_token_accuracy": 0.9499401599168777,
      "num_tokens": 150908359.0,
      "step": 9375
    },
    {
      "aux_loss": 8.067628479003906,
      "entropy": 0.14016885533928872,
      "epoch": 0.00938,
      "grad_norm": 0.1329381763935089,
      "learning_rate": 2e-05,
      "loss": 0.1666790008544922,
      "mean_token_accuracy": 0.9528997659683227,
      "num_tokens": 150988255.0,
      "step": 9380
    },
    {
      "aux_loss": 8.064125585556031,
      "entropy": 0.16074576787650585,
      "epoch": 0.009385,
      "grad_norm": 0.12546353042125702,
      "learning_rate": 2e-05,
      "loss": 0.1952120542526245,
      "mean_token_accuracy": 0.9466593503952027,
      "num_tokens": 151067831.0,
      "step": 9385
    },
    {
      "aux_loss": 8.045863056182862,
      "entropy": 0.16810012757778167,
      "epoch": 0.00939,
      "grad_norm": 0.13055840134620667,
      "learning_rate": 2e-05,
      "loss": 0.2092730760574341,
      "mean_token_accuracy": 0.9432838469743728,
      "num_tokens": 151148899.0,
      "step": 9390
    },
    {
      "aux_loss": 8.04754114151001,
      "entropy": 0.18048907145857812,
      "epoch": 0.009395,
      "grad_norm": 0.12861719727516174,
      "learning_rate": 2e-05,
      "loss": 0.21951625347137452,
      "mean_token_accuracy": 0.9389490485191345,
      "num_tokens": 151229743.0,
      "step": 9395
    },
    {
      "aux_loss": 8.051449251174926,
      "entropy": 0.16877450346946715,
      "epoch": 0.0094,
      "grad_norm": 0.12695060670375824,
      "learning_rate": 2e-05,
      "loss": 0.19677878618240358,
      "mean_token_accuracy": 0.9440664023160934,
      "num_tokens": 151310549.0,
      "step": 9400
    },
    {
      "aux_loss": 8.055356740951538,
      "entropy": 0.149481039121747,
      "epoch": 0.009405,
      "grad_norm": 0.12921400368213654,
      "learning_rate": 2e-05,
      "loss": 0.1782172441482544,
      "mean_token_accuracy": 0.9512036591768265,
      "num_tokens": 151391373.0,
      "step": 9405
    },
    {
      "aux_loss": 8.061852836608887,
      "entropy": 0.143968391045928,
      "epoch": 0.00941,
      "grad_norm": 0.13102933764457703,
      "learning_rate": 2e-05,
      "loss": 0.17220641374588014,
      "mean_token_accuracy": 0.9524245947599411,
      "num_tokens": 151472193.0,
      "step": 9410
    },
    {
      "aux_loss": 8.061843824386596,
      "entropy": 0.14332425817847252,
      "epoch": 0.009415,
      "grad_norm": 0.12662266194820404,
      "learning_rate": 2e-05,
      "loss": 0.1767818808555603,
      "mean_token_accuracy": 0.9512009918689728,
      "num_tokens": 151552749.0,
      "step": 9415
    },
    {
      "aux_loss": 8.047246408462524,
      "entropy": 0.1688066367059946,
      "epoch": 0.00942,
      "grad_norm": 0.12467551231384277,
      "learning_rate": 2e-05,
      "loss": 0.20529661178588868,
      "mean_token_accuracy": 0.9429667621850968,
      "num_tokens": 151633940.0,
      "step": 9420
    },
    {
      "aux_loss": 8.046308708190917,
      "entropy": 0.1844594843685627,
      "epoch": 0.009425,
      "grad_norm": 0.12745779752731323,
      "learning_rate": 2e-05,
      "loss": 0.22105760574340821,
      "mean_token_accuracy": 0.9385595500469208,
      "num_tokens": 151714556.0,
      "step": 9425
    },
    {
      "aux_loss": 8.04946813583374,
      "entropy": 0.16270146295428276,
      "epoch": 0.00943,
      "grad_norm": 0.12307646870613098,
      "learning_rate": 2e-05,
      "loss": 0.1920972943305969,
      "mean_token_accuracy": 0.9450286716222763,
      "num_tokens": 151795279.0,
      "step": 9430
    },
    {
      "aux_loss": 8.057387161254884,
      "entropy": 0.1639138750731945,
      "epoch": 0.009435,
      "grad_norm": 0.12869898974895477,
      "learning_rate": 2e-05,
      "loss": 0.1896457076072693,
      "mean_token_accuracy": 0.9474249124526978,
      "num_tokens": 151876130.0,
      "step": 9435
    },
    {
      "aux_loss": 8.06136450767517,
      "entropy": 0.13895124793052674,
      "epoch": 0.00944,
      "grad_norm": 0.13967812061309814,
      "learning_rate": 2e-05,
      "loss": 0.1674266815185547,
      "mean_token_accuracy": 0.9536738961935043,
      "num_tokens": 151956699.0,
      "step": 9440
    },
    {
      "aux_loss": 8.068367719650269,
      "entropy": 0.13663856834173202,
      "epoch": 0.009445,
      "grad_norm": 0.1271567940711975,
      "learning_rate": 2e-05,
      "loss": 0.16691681146621704,
      "mean_token_accuracy": 0.9556272029876709,
      "num_tokens": 152033952.0,
      "step": 9445
    },
    {
      "aux_loss": 8.050022172927857,
      "entropy": 0.16802804507315158,
      "epoch": 0.00945,
      "grad_norm": 0.13574174046516418,
      "learning_rate": 2e-05,
      "loss": 0.21252655982971191,
      "mean_token_accuracy": 0.9412559688091278,
      "num_tokens": 152115171.0,
      "step": 9450
    },
    {
      "aux_loss": 8.04716248512268,
      "entropy": 0.166149490699172,
      "epoch": 0.009455,
      "grad_norm": 0.13551300764083862,
      "learning_rate": 2e-05,
      "loss": 0.20771293640136718,
      "mean_token_accuracy": 0.9421769857406617,
      "num_tokens": 152196226.0,
      "step": 9455
    },
    {
      "aux_loss": 8.050882148742676,
      "entropy": 0.17923564054071903,
      "epoch": 0.00946,
      "grad_norm": 0.12311870604753494,
      "learning_rate": 2e-05,
      "loss": 0.2105072021484375,
      "mean_token_accuracy": 0.9422770380973816,
      "num_tokens": 152277324.0,
      "step": 9460
    },
    {
      "aux_loss": 8.055430030822754,
      "entropy": 0.17071404084563255,
      "epoch": 0.009465,
      "grad_norm": 0.1360500305891037,
      "learning_rate": 2e-05,
      "loss": 0.19662503004074097,
      "mean_token_accuracy": 0.9438480734825134,
      "num_tokens": 152357789.0,
      "step": 9465
    },
    {
      "aux_loss": 8.060851955413819,
      "entropy": 0.15003802850842476,
      "epoch": 0.00947,
      "grad_norm": 0.13290728628635406,
      "learning_rate": 2e-05,
      "loss": 0.1774725317955017,
      "mean_token_accuracy": 0.9489647716283798,
      "num_tokens": 152438001.0,
      "step": 9470
    },
    {
      "aux_loss": 8.069437170028687,
      "entropy": 0.1424240183085203,
      "epoch": 0.009475,
      "grad_norm": 0.1440083533525467,
      "learning_rate": 2e-05,
      "loss": 0.1703614592552185,
      "mean_token_accuracy": 0.9517173022031784,
      "num_tokens": 152517099.0,
      "step": 9475
    },
    {
      "aux_loss": 8.049043893814087,
      "entropy": 0.17002014629542828,
      "epoch": 0.00948,
      "grad_norm": 0.12988366186618805,
      "learning_rate": 2e-05,
      "loss": 0.20711610317230225,
      "mean_token_accuracy": 0.9426903814077378,
      "num_tokens": 152597797.0,
      "step": 9480
    },
    {
      "aux_loss": 8.04736876487732,
      "entropy": 0.16849479153752328,
      "epoch": 0.009485,
      "grad_norm": 0.12628424167633057,
      "learning_rate": 2e-05,
      "loss": 0.2082036018371582,
      "mean_token_accuracy": 0.9422867804765701,
      "num_tokens": 152679166.0,
      "step": 9485
    },
    {
      "aux_loss": 8.050393438339233,
      "entropy": 0.16659751236438752,
      "epoch": 0.00949,
      "grad_norm": 0.13501644134521484,
      "learning_rate": 2e-05,
      "loss": 0.2020411729812622,
      "mean_token_accuracy": 0.944596815109253,
      "num_tokens": 152759962.0,
      "step": 9490
    },
    {
      "aux_loss": 8.052613925933837,
      "entropy": 0.1595939502120018,
      "epoch": 0.009495,
      "grad_norm": 0.11969030648469925,
      "learning_rate": 2e-05,
      "loss": 0.18960103988647461,
      "mean_token_accuracy": 0.9479102104902267,
      "num_tokens": 152840713.0,
      "step": 9495
    },
    {
      "aux_loss": 8.058992767333985,
      "entropy": 0.14904571026563646,
      "epoch": 0.0095,
      "grad_norm": 0.12419265508651733,
      "learning_rate": 2e-05,
      "loss": 0.17722588777542114,
      "mean_token_accuracy": 0.9507673263549805,
      "num_tokens": 152921053.0,
      "step": 9500
    }
  ],
  "logging_steps": 5,
  "max_steps": 1000000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 250,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.7777590941142225e+19,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}