{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.4670588763641263,
  "eval_steps": 1000,
  "global_step": 26000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.2285142108798027,
      "epoch": 0.0001796380293708178,
      "grad_norm": 338.0,
      "learning_rate": 4.5e-07,
      "loss": 4.8883522033691404,
      "mean_token_accuracy": 0.614159119874239,
      "num_tokens": 19881.0,
      "step": 10
    },
    {
      "entropy": 0.2592433709651232,
      "epoch": 0.0003592760587416356,
      "grad_norm": 225.0,
      "learning_rate": 9.5e-07,
      "loss": 4.066074371337891,
      "mean_token_accuracy": 0.615167886018753,
      "num_tokens": 40073.0,
      "step": 20
    },
    {
      "entropy": 0.5413158625364304,
      "epoch": 0.0005389140881124534,
      "grad_norm": 78.0,
      "learning_rate": 1.45e-06,
      "loss": 2.8527931213378905,
      "mean_token_accuracy": 0.6295808032155037,
      "num_tokens": 59507.0,
      "step": 30
    },
    {
      "entropy": 0.8526842221617699,
      "epoch": 0.0007185521174832712,
      "grad_norm": 40.0,
      "learning_rate": 1.95e-06,
      "loss": 1.788459014892578,
      "mean_token_accuracy": 0.6713394083082675,
      "num_tokens": 78952.0,
      "step": 40
    },
    {
      "entropy": 1.3804137662053109,
      "epoch": 0.000898190146854089,
      "grad_norm": 34.25,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 1.77496395111084,
      "mean_token_accuracy": 0.6312253788113594,
      "num_tokens": 99568.0,
      "step": 50
    },
    {
      "entropy": 1.1601247757673263,
      "epoch": 0.0010778281762249067,
      "grad_norm": 21.625,
      "learning_rate": 2.95e-06,
      "loss": 1.3090765953063965,
      "mean_token_accuracy": 0.7052405722439289,
      "num_tokens": 120527.0,
      "step": 60
    },
    {
      "entropy": 1.253069232404232,
      "epoch": 0.0012574662055957245,
      "grad_norm": 27.5,
      "learning_rate": 3.4500000000000004e-06,
      "loss": 1.3177606582641601,
      "mean_token_accuracy": 0.7069540455937385,
      "num_tokens": 140444.0,
      "step": 70
    },
    {
      "entropy": 1.3081351161003112,
      "epoch": 0.0014371042349665423,
      "grad_norm": 33.75,
      "learning_rate": 3.95e-06,
      "loss": 1.38143949508667,
      "mean_token_accuracy": 0.6947338208556175,
      "num_tokens": 160447.0,
      "step": 80
    },
    {
      "entropy": 1.2931678161025046,
      "epoch": 0.0016167422643373602,
      "grad_norm": 28.625,
      "learning_rate": 4.45e-06,
      "loss": 1.321197509765625,
      "mean_token_accuracy": 0.696695514768362,
      "num_tokens": 179917.0,
      "step": 90
    },
    {
      "entropy": 1.3114918872714043,
      "epoch": 0.001796380293708178,
      "grad_norm": 26.25,
      "learning_rate": 4.950000000000001e-06,
      "loss": 1.4064659118652343,
      "mean_token_accuracy": 0.6969998598098754,
      "num_tokens": 198929.0,
      "step": 100
    },
    {
      "entropy": 1.2356058835983277,
      "epoch": 0.001976018323078996,
      "grad_norm": 23.125,
      "learning_rate": 5.45e-06,
      "loss": 1.3247491836547851,
      "mean_token_accuracy": 0.6962062671780587,
      "num_tokens": 220465.0,
      "step": 110
    },
    {
      "entropy": 1.1230886340141297,
      "epoch": 0.0021556563524498134,
      "grad_norm": 23.5,
      "learning_rate": 5.95e-06,
      "loss": 1.176513957977295,
      "mean_token_accuracy": 0.7297163873910903,
      "num_tokens": 239249.0,
      "step": 120
    },
    {
      "entropy": 1.313824026286602,
      "epoch": 0.0023352943818206314,
      "grad_norm": 33.5,
      "learning_rate": 6.45e-06,
      "loss": 1.3627755165100097,
      "mean_token_accuracy": 0.7020026236772537,
      "num_tokens": 258824.0,
      "step": 130
    },
    {
      "entropy": 1.3096917569637299,
      "epoch": 0.002514932411191449,
      "grad_norm": 37.0,
      "learning_rate": 6.950000000000001e-06,
      "loss": 1.4027366638183594,
      "mean_token_accuracy": 0.689975680410862,
      "num_tokens": 278392.0,
      "step": 140
    },
    {
      "entropy": 1.229283632338047,
      "epoch": 0.002694570440562267,
      "grad_norm": 27.875,
      "learning_rate": 7.45e-06,
      "loss": 1.193472194671631,
      "mean_token_accuracy": 0.7135655298829079,
      "num_tokens": 298114.0,
      "step": 150
    },
    {
      "entropy": 1.129289910197258,
      "epoch": 0.0028742084699330847,
      "grad_norm": 28.25,
      "learning_rate": 7.95e-06,
      "loss": 1.2268380165100097,
      "mean_token_accuracy": 0.7122256726026535,
      "num_tokens": 317682.0,
      "step": 160
    },
    {
      "entropy": 1.3514228284358978,
      "epoch": 0.0030538464993039027,
      "grad_norm": 25.75,
      "learning_rate": 8.45e-06,
      "loss": 1.3145798683166503,
      "mean_token_accuracy": 0.6925913117825985,
      "num_tokens": 337243.0,
      "step": 170
    },
    {
      "entropy": 1.1698046237230302,
      "epoch": 0.0032334845286747203,
      "grad_norm": 23.0,
      "learning_rate": 8.95e-06,
      "loss": 1.202991771697998,
      "mean_token_accuracy": 0.7119750328361988,
      "num_tokens": 357044.0,
      "step": 180
    },
    {
      "entropy": 1.1908135280013084,
      "epoch": 0.0034131225580455384,
      "grad_norm": 30.0,
      "learning_rate": 9.450000000000001e-06,
      "loss": 1.2616134643554688,
      "mean_token_accuracy": 0.7143504373729229,
      "num_tokens": 376996.0,
      "step": 190
    },
    {
      "entropy": 1.1140592023730278,
      "epoch": 0.003592760587416356,
      "grad_norm": 20.625,
      "learning_rate": 9.950000000000001e-06,
      "loss": 1.1481157302856446,
      "mean_token_accuracy": 0.7259704366326332,
      "num_tokens": 395940.0,
      "step": 200
    },
    {
      "entropy": 1.185268647968769,
      "epoch": 0.003772398616787174,
      "grad_norm": 22.0,
      "learning_rate": 1.045e-05,
      "loss": 1.1380942344665528,
      "mean_token_accuracy": 0.7433141075074673,
      "num_tokens": 414953.0,
      "step": 210
    },
    {
      "entropy": 1.1160883173346519,
      "epoch": 0.003952036646157992,
      "grad_norm": 20.125,
      "learning_rate": 1.095e-05,
      "loss": 1.234296703338623,
      "mean_token_accuracy": 0.7160610310733319,
      "num_tokens": 435066.0,
      "step": 220
    },
    {
      "entropy": 1.3014020532369615,
      "epoch": 0.004131674675528809,
      "grad_norm": 17.75,
      "learning_rate": 1.145e-05,
      "loss": 1.2594083786010741,
      "mean_token_accuracy": 0.7140978299081325,
      "num_tokens": 455791.0,
      "step": 230
    },
    {
      "entropy": 1.186601100116968,
      "epoch": 0.004311312704899627,
      "grad_norm": 35.0,
      "learning_rate": 1.195e-05,
      "loss": 1.2977076530456544,
      "mean_token_accuracy": 0.7061133705079555,
      "num_tokens": 474818.0,
      "step": 240
    },
    {
      "entropy": 1.2088973760604858,
      "epoch": 0.004490950734270445,
      "grad_norm": 22.125,
      "learning_rate": 1.2450000000000001e-05,
      "loss": 1.1662664413452148,
      "mean_token_accuracy": 0.7203448742628098,
      "num_tokens": 496453.0,
      "step": 250
    },
    {
      "entropy": 1.2616489186882973,
      "epoch": 0.004670588763641263,
      "grad_norm": 28.0,
      "learning_rate": 1.2950000000000001e-05,
      "loss": 1.298599624633789,
      "mean_token_accuracy": 0.7160171762108802,
      "num_tokens": 516400.0,
      "step": 260
    },
    {
      "entropy": 1.126600243151188,
      "epoch": 0.0048502267930120805,
      "grad_norm": 19.75,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 1.1665369987487793,
      "mean_token_accuracy": 0.7411642163991928,
      "num_tokens": 535475.0,
      "step": 270
    },
    {
      "entropy": 1.1618767216801644,
      "epoch": 0.005029864822382898,
      "grad_norm": 27.875,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 1.2132204055786133,
      "mean_token_accuracy": 0.7233127027750015,
      "num_tokens": 554868.0,
      "step": 280
    },
    {
      "entropy": 1.3146504908800125,
      "epoch": 0.005209502851753717,
      "grad_norm": 22.0,
      "learning_rate": 1.4449999999999999e-05,
      "loss": 1.3070467948913573,
      "mean_token_accuracy": 0.7006118722259999,
      "num_tokens": 574711.0,
      "step": 290
    },
    {
      "entropy": 1.1791881158947946,
      "epoch": 0.005389140881124534,
      "grad_norm": 22.5,
      "learning_rate": 1.4950000000000001e-05,
      "loss": 1.2319019317626954,
      "mean_token_accuracy": 0.7152372360229492,
      "num_tokens": 595068.0,
      "step": 300
    },
    {
      "entropy": 1.1496956929564477,
      "epoch": 0.005568778910495352,
      "grad_norm": 23.25,
      "learning_rate": 1.545e-05,
      "loss": 1.2204404830932618,
      "mean_token_accuracy": 0.720104095339775,
      "num_tokens": 615186.0,
      "step": 310
    },
    {
      "entropy": 1.1919636234641076,
      "epoch": 0.005748416939866169,
      "grad_norm": 22.25,
      "learning_rate": 1.595e-05,
      "loss": 1.2102306365966797,
      "mean_token_accuracy": 0.7176152899861336,
      "num_tokens": 636045.0,
      "step": 320
    },
    {
      "entropy": 1.2924261584877967,
      "epoch": 0.005928054969236988,
      "grad_norm": 24.0,
      "learning_rate": 1.645e-05,
      "loss": 1.256188201904297,
      "mean_token_accuracy": 0.698620592802763,
      "num_tokens": 656037.0,
      "step": 330
    },
    {
      "entropy": 1.2066523447632789,
      "epoch": 0.0061076929986078055,
      "grad_norm": 23.875,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 1.300409984588623,
      "mean_token_accuracy": 0.6900593861937523,
      "num_tokens": 674706.0,
      "step": 340
    },
    {
      "entropy": 1.1717823699116707,
      "epoch": 0.006287331027978623,
      "grad_norm": 21.75,
      "learning_rate": 1.745e-05,
      "loss": 1.1818981170654297,
      "mean_token_accuracy": 0.7271141156554222,
      "num_tokens": 697034.0,
      "step": 350
    },
    {
      "entropy": 1.347169689834118,
      "epoch": 0.006466969057349441,
      "grad_norm": 20.5,
      "learning_rate": 1.795e-05,
      "loss": 1.386529541015625,
      "mean_token_accuracy": 0.6903549000620842,
      "num_tokens": 716356.0,
      "step": 360
    },
    {
      "entropy": 1.1157519415020942,
      "epoch": 0.006646607086720258,
      "grad_norm": 19.875,
      "learning_rate": 1.845e-05,
      "loss": 1.1746755599975587,
      "mean_token_accuracy": 0.7316143766045571,
      "num_tokens": 736010.0,
      "step": 370
    },
    {
      "entropy": 1.2086873427033424,
      "epoch": 0.006826245116091077,
      "grad_norm": 21.0,
      "learning_rate": 1.895e-05,
      "loss": 1.1709842681884766,
      "mean_token_accuracy": 0.721261890232563,
      "num_tokens": 755905.0,
      "step": 380
    },
    {
      "entropy": 1.0470090851187706,
      "epoch": 0.007005883145461894,
      "grad_norm": 15.875,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 1.1902419090270997,
      "mean_token_accuracy": 0.730524119734764,
      "num_tokens": 776342.0,
      "step": 390
    },
    {
      "entropy": 1.3544242575764656,
      "epoch": 0.007185521174832712,
      "grad_norm": 16.5,
      "learning_rate": 1.995e-05,
      "loss": 1.3294052124023437,
      "mean_token_accuracy": 0.6911105871200561,
      "num_tokens": 796912.0,
      "step": 400
    },
    {
      "entropy": 1.294105887413025,
      "epoch": 0.0073651592042035295,
      "grad_norm": 26.0,
      "learning_rate": 2.045e-05,
      "loss": 1.3535008430480957,
      "mean_token_accuracy": 0.6951926596462726,
      "num_tokens": 816253.0,
      "step": 410
    },
    {
      "entropy": 1.3261266335844994,
      "epoch": 0.007544797233574348,
      "grad_norm": 20.125,
      "learning_rate": 2.095e-05,
      "loss": 1.3798662185668946,
      "mean_token_accuracy": 0.6866254404187202,
      "num_tokens": 835424.0,
      "step": 420
    },
    {
      "entropy": 1.3382393702864648,
      "epoch": 0.007724435262945166,
      "grad_norm": 22.125,
      "learning_rate": 2.145e-05,
      "loss": 1.3444744110107423,
      "mean_token_accuracy": 0.6885690718889237,
      "num_tokens": 853906.0,
      "step": 430
    },
    {
      "entropy": 1.190830898284912,
      "epoch": 0.007904073292315983,
      "grad_norm": 27.0,
      "learning_rate": 2.195e-05,
      "loss": 1.2053414344787599,
      "mean_token_accuracy": 0.7217165172100067,
      "num_tokens": 873159.0,
      "step": 440
    },
    {
      "entropy": 1.1526850268244744,
      "epoch": 0.0080837113216868,
      "grad_norm": 20.125,
      "learning_rate": 2.245e-05,
      "loss": 1.3208661079406738,
      "mean_token_accuracy": 0.6984982281923294,
      "num_tokens": 893835.0,
      "step": 450
    },
    {
      "entropy": 1.3733031839132308,
      "epoch": 0.008263349351057618,
      "grad_norm": 19.875,
      "learning_rate": 2.2950000000000002e-05,
      "loss": 1.2875539779663085,
      "mean_token_accuracy": 0.70067323371768,
      "num_tokens": 914236.0,
      "step": 460
    },
    {
      "entropy": 1.1398809850215912,
      "epoch": 0.008442987380428436,
      "grad_norm": 17.5,
      "learning_rate": 2.345e-05,
      "loss": 1.2422693252563477,
      "mean_token_accuracy": 0.7127251163125038,
      "num_tokens": 933711.0,
      "step": 470
    },
    {
      "entropy": 1.3505023419857025,
      "epoch": 0.008622625409799254,
      "grad_norm": 19.25,
      "learning_rate": 2.395e-05,
      "loss": 1.3500717163085938,
      "mean_token_accuracy": 0.7032387189567089,
      "num_tokens": 952965.0,
      "step": 480
    },
    {
      "entropy": 1.2616789430379867,
      "epoch": 0.008802263439170073,
      "grad_norm": 23.75,
      "learning_rate": 2.445e-05,
      "loss": 1.3580240249633788,
      "mean_token_accuracy": 0.6813065484166145,
      "num_tokens": 973043.0,
      "step": 490
    },
    {
      "entropy": 1.1928757190704347,
      "epoch": 0.00898190146854089,
      "grad_norm": 18.75,
      "learning_rate": 2.495e-05,
      "loss": 1.2279540061950684,
      "mean_token_accuracy": 0.7097863912582397,
      "num_tokens": 992534.0,
      "step": 500
    },
    {
      "entropy": 1.3313418433070183,
      "epoch": 0.009161539497911708,
      "grad_norm": 25.625,
      "learning_rate": 2.5450000000000002e-05,
      "loss": 1.392829418182373,
      "mean_token_accuracy": 0.687772911787033,
      "num_tokens": 1011503.0,
      "step": 510
    },
    {
      "entropy": 1.2510645657777786,
      "epoch": 0.009341177527282526,
      "grad_norm": 17.375,
      "learning_rate": 2.595e-05,
      "loss": 1.2840056419372559,
      "mean_token_accuracy": 0.7028377979993821,
      "num_tokens": 1031078.0,
      "step": 520
    },
    {
      "entropy": 1.3048915401101113,
      "epoch": 0.009520815556653343,
      "grad_norm": 16.75,
      "learning_rate": 2.6450000000000003e-05,
      "loss": 1.3614519119262696,
      "mean_token_accuracy": 0.6907784312963485,
      "num_tokens": 1051001.0,
      "step": 530
    },
    {
      "entropy": 1.2187796041369439,
      "epoch": 0.009700453586024161,
      "grad_norm": 20.875,
      "learning_rate": 2.6950000000000005e-05,
      "loss": 1.2859848976135253,
      "mean_token_accuracy": 0.7090503506362438,
      "num_tokens": 1070569.0,
      "step": 540
    },
    {
      "entropy": 1.3603180184960366,
      "epoch": 0.009880091615394979,
      "grad_norm": 18.25,
      "learning_rate": 2.7450000000000003e-05,
      "loss": 1.3390862464904785,
      "mean_token_accuracy": 0.6932703338563442,
      "num_tokens": 1089715.0,
      "step": 550
    },
    {
      "entropy": 1.229197269678116,
      "epoch": 0.010059729644765796,
      "grad_norm": 18.375,
      "learning_rate": 2.7950000000000005e-05,
      "loss": 1.3082941055297852,
      "mean_token_accuracy": 0.7056461185216903,
      "num_tokens": 1110066.0,
      "step": 560
    },
    {
      "entropy": 1.3566848933696747,
      "epoch": 0.010239367674136616,
      "grad_norm": 15.375,
      "learning_rate": 2.845e-05,
      "loss": 1.3479373931884766,
      "mean_token_accuracy": 0.6874948844313622,
      "num_tokens": 1129467.0,
      "step": 570
    },
    {
      "entropy": 1.165881448984146,
      "epoch": 0.010419005703507433,
      "grad_norm": 15.25,
      "learning_rate": 2.895e-05,
      "loss": 1.2552820205688477,
      "mean_token_accuracy": 0.7130140498280525,
      "num_tokens": 1150425.0,
      "step": 580
    },
    {
      "entropy": 1.2559717029333115,
      "epoch": 0.01059864373287825,
      "grad_norm": 14.875,
      "learning_rate": 2.945e-05,
      "loss": 1.3594337463378907,
      "mean_token_accuracy": 0.6969697877764702,
      "num_tokens": 1169390.0,
      "step": 590
    },
    {
      "entropy": 1.527521088719368,
      "epoch": 0.010778281762249068,
      "grad_norm": 18.125,
      "learning_rate": 2.995e-05,
      "loss": 1.5145947456359863,
      "mean_token_accuracy": 0.6706581458449363,
      "num_tokens": 1188454.0,
      "step": 600
    },
    {
      "entropy": 1.3165057599544525,
      "epoch": 0.010957919791619886,
      "grad_norm": 22.75,
      "learning_rate": 3.045e-05,
      "loss": 1.3955662727355957,
      "mean_token_accuracy": 0.6902709446847439,
      "num_tokens": 1208049.0,
      "step": 610
    },
    {
      "entropy": 1.2833266004920005,
      "epoch": 0.011137557820990704,
      "grad_norm": 30.125,
      "learning_rate": 3.095e-05,
      "loss": 1.3738471031188966,
      "mean_token_accuracy": 0.695693901181221,
      "num_tokens": 1227680.0,
      "step": 620
    },
    {
      "entropy": 1.3729368403553963,
      "epoch": 0.011317195850361521,
      "grad_norm": 27.75,
      "learning_rate": 3.145e-05,
      "loss": 1.3908536911010743,
      "mean_token_accuracy": 0.6883269481360912,
      "num_tokens": 1248247.0,
      "step": 630
    },
    {
      "entropy": 1.321514193713665,
      "epoch": 0.011496833879732339,
      "grad_norm": 15.75,
      "learning_rate": 3.1950000000000004e-05,
      "loss": 1.4234540939331055,
      "mean_token_accuracy": 0.6745201140642166,
      "num_tokens": 1268043.0,
      "step": 640
    },
    {
      "entropy": 1.4137424260377884,
      "epoch": 0.011676471909103156,
      "grad_norm": 16.375,
      "learning_rate": 3.245e-05,
      "loss": 1.465153121948242,
      "mean_token_accuracy": 0.6703482449054718,
      "num_tokens": 1287035.0,
      "step": 650
    },
    {
      "entropy": 1.4118265092372895,
      "epoch": 0.011856109938473976,
      "grad_norm": 16.625,
      "learning_rate": 3.295e-05,
      "loss": 1.4218225479125977,
      "mean_token_accuracy": 0.6761282578110694,
      "num_tokens": 1307713.0,
      "step": 660
    },
    {
      "entropy": 1.3070572301745416,
      "epoch": 0.012035747967844793,
      "grad_norm": 12.4375,
      "learning_rate": 3.345000000000001e-05,
      "loss": 1.4144351005554199,
      "mean_token_accuracy": 0.6854814760386944,
      "num_tokens": 1328401.0,
      "step": 670
    },
    {
      "entropy": 1.4844876140356065,
      "epoch": 0.012215385997215611,
      "grad_norm": 14.0,
      "learning_rate": 3.3950000000000005e-05,
      "loss": 1.5067203521728516,
      "mean_token_accuracy": 0.6696059130132198,
      "num_tokens": 1348081.0,
      "step": 680
    },
    {
      "entropy": 1.5267313480377198,
      "epoch": 0.012395024026586429,
      "grad_norm": 22.125,
      "learning_rate": 3.445e-05,
      "loss": 1.547717571258545,
      "mean_token_accuracy": 0.6525343872606755,
      "num_tokens": 1366620.0,
      "step": 690
    },
    {
      "entropy": 1.409482868015766,
      "epoch": 0.012574662055957246,
      "grad_norm": 15.25,
      "learning_rate": 3.495e-05,
      "loss": 1.5240630149841308,
      "mean_token_accuracy": 0.6619769029319287,
      "num_tokens": 1387634.0,
      "step": 700
    },
    {
      "entropy": 1.385390780866146,
      "epoch": 0.012754300085328064,
      "grad_norm": 17.875,
      "learning_rate": 3.545e-05,
      "loss": 1.4197970390319825,
      "mean_token_accuracy": 0.6821596398949623,
      "num_tokens": 1407691.0,
      "step": 710
    },
    {
      "entropy": 1.4515428811311721,
      "epoch": 0.012933938114698881,
      "grad_norm": 13.3125,
      "learning_rate": 3.595e-05,
      "loss": 1.5343506813049317,
      "mean_token_accuracy": 0.6703817009925842,
      "num_tokens": 1428549.0,
      "step": 720
    },
    {
      "entropy": 1.422910524904728,
      "epoch": 0.013113576144069699,
      "grad_norm": 16.25,
      "learning_rate": 3.645e-05,
      "loss": 1.4436484336853028,
      "mean_token_accuracy": 0.679908835887909,
      "num_tokens": 1448696.0,
      "step": 730
    },
    {
      "entropy": 1.3541560366749763,
      "epoch": 0.013293214173440517,
      "grad_norm": 13.5625,
      "learning_rate": 3.6950000000000004e-05,
      "loss": 1.4838427543640136,
      "mean_token_accuracy": 0.6745669797062874,
      "num_tokens": 1467982.0,
      "step": 740
    },
    {
      "entropy": 1.4543387055397035,
      "epoch": 0.013472852202811336,
      "grad_norm": 24.375,
      "learning_rate": 3.745e-05,
      "loss": 1.5257116317749024,
      "mean_token_accuracy": 0.6583885066211224,
      "num_tokens": 1486504.0,
      "step": 750
    },
    {
      "entropy": 1.4233024775981904,
      "epoch": 0.013652490232182153,
      "grad_norm": 14.3125,
      "learning_rate": 3.795e-05,
      "loss": 1.5261417388916017,
      "mean_token_accuracy": 0.6611535474658012,
      "num_tokens": 1507944.0,
      "step": 760
    },
    {
      "entropy": 1.5868904277682305,
      "epoch": 0.013832128261552971,
      "grad_norm": 15.4375,
      "learning_rate": 3.845e-05,
      "loss": 1.589180088043213,
      "mean_token_accuracy": 0.652680542320013,
      "num_tokens": 1527573.0,
      "step": 770
    },
    {
      "entropy": 1.4296333938837051,
      "epoch": 0.014011766290923789,
      "grad_norm": 15.875,
      "learning_rate": 3.8950000000000005e-05,
      "loss": 1.4833106994628906,
      "mean_token_accuracy": 0.664696491509676,
      "num_tokens": 1547688.0,
      "step": 780
    },
    {
      "entropy": 1.4370834082365036,
      "epoch": 0.014191404320294606,
      "grad_norm": 19.0,
      "learning_rate": 3.9450000000000003e-05,
      "loss": 1.5261101722717285,
      "mean_token_accuracy": 0.6580417729914189,
      "num_tokens": 1567143.0,
      "step": 790
    },
    {
      "entropy": 1.5416127756237983,
      "epoch": 0.014371042349665424,
      "grad_norm": 11.3125,
      "learning_rate": 3.995e-05,
      "loss": 1.5723227500915526,
      "mean_token_accuracy": 0.6579069018363952,
      "num_tokens": 1587504.0,
      "step": 800
    },
    {
      "entropy": 1.4236389979720117,
      "epoch": 0.014550680379036241,
      "grad_norm": 15.5625,
      "learning_rate": 4.045000000000001e-05,
      "loss": 1.5273307800292968,
      "mean_token_accuracy": 0.6582598991692066,
      "num_tokens": 1606283.0,
      "step": 810
    },
    {
      "entropy": 1.478298343718052,
      "epoch": 0.014730318408407059,
      "grad_norm": 20.5,
      "learning_rate": 4.095e-05,
      "loss": 1.5232603073120117,
      "mean_token_accuracy": 0.6629001379013062,
      "num_tokens": 1625648.0,
      "step": 820
    },
    {
      "entropy": 1.5232951745390892,
      "epoch": 0.014909956437777878,
      "grad_norm": 14.5,
      "learning_rate": 4.145e-05,
      "loss": 1.5676536560058594,
      "mean_token_accuracy": 0.6548260487616062,
      "num_tokens": 1645079.0,
      "step": 830
    },
    {
      "entropy": 1.3451756298542024,
      "epoch": 0.015089594467148696,
      "grad_norm": 17.25,
      "learning_rate": 4.195e-05,
      "loss": 1.4040653228759765,
      "mean_token_accuracy": 0.6850749261677265,
      "num_tokens": 1665644.0,
      "step": 840
    },
    {
      "entropy": 1.553303211927414,
      "epoch": 0.015269232496519514,
      "grad_norm": 15.5625,
      "learning_rate": 4.245e-05,
      "loss": 1.6475976943969726,
      "mean_token_accuracy": 0.6442298710346221,
      "num_tokens": 1684318.0,
      "step": 850
    },
    {
      "entropy": 1.4236965268850326,
      "epoch": 0.015448870525890331,
      "grad_norm": 14.1875,
      "learning_rate": 4.295e-05,
      "loss": 1.513192844390869,
      "mean_token_accuracy": 0.6647441633045673,
      "num_tokens": 1705464.0,
      "step": 860
    },
    {
      "entropy": 1.6874301448464393,
      "epoch": 0.01562850855526115,
      "grad_norm": 14.1875,
      "learning_rate": 4.345e-05,
      "loss": 1.6767539978027344,
      "mean_token_accuracy": 0.6399445720016956,
      "num_tokens": 1724729.0,
      "step": 870
    },
    {
      "entropy": 1.455209381878376,
      "epoch": 0.015808146584631966,
      "grad_norm": 15.8125,
      "learning_rate": 4.3950000000000004e-05,
      "loss": 1.6115209579467773,
      "mean_token_accuracy": 0.6471515670418739,
      "num_tokens": 1744724.0,
      "step": 880
    },
    {
      "entropy": 1.4922844484448432,
      "epoch": 0.015987784614002784,
      "grad_norm": 14.6875,
      "learning_rate": 4.445e-05,
      "loss": 1.572502613067627,
      "mean_token_accuracy": 0.6561724334955216,
      "num_tokens": 1764414.0,
      "step": 890
    },
    {
      "entropy": 1.7285323530435561,
      "epoch": 0.0161674226433736,
      "grad_norm": 16.0,
      "learning_rate": 4.495e-05,
      "loss": 1.8349027633666992,
      "mean_token_accuracy": 0.6137065529823303,
      "num_tokens": 1783351.0,
      "step": 900
    },
    {
      "entropy": 1.510180902481079,
      "epoch": 0.01634706067274442,
      "grad_norm": 20.375,
      "learning_rate": 4.545000000000001e-05,
      "loss": 1.5554212570190429,
      "mean_token_accuracy": 0.6492202296853066,
      "num_tokens": 1802521.0,
      "step": 910
    },
    {
      "entropy": 1.6031754165887833,
      "epoch": 0.016526698702115237,
      "grad_norm": 13.375,
      "learning_rate": 4.5950000000000006e-05,
      "loss": 1.7184165954589843,
      "mean_token_accuracy": 0.6307573899626732,
      "num_tokens": 1821027.0,
      "step": 920
    },
    {
      "entropy": 1.5202497854828834,
      "epoch": 0.016706336731486054,
      "grad_norm": 13.0625,
      "learning_rate": 4.6450000000000004e-05,
      "loss": 1.674123191833496,
      "mean_token_accuracy": 0.6374556072056293,
      "num_tokens": 1840799.0,
      "step": 930
    },
    {
      "entropy": 1.577104239165783,
      "epoch": 0.016885974760856872,
      "grad_norm": 17.125,
      "learning_rate": 4.695e-05,
      "loss": 1.6293512344360352,
      "mean_token_accuracy": 0.662019107490778,
      "num_tokens": 1860158.0,
      "step": 940
    },
    {
      "entropy": 1.474810865521431,
      "epoch": 0.01706561279022769,
      "grad_norm": 13.25,
      "learning_rate": 4.745e-05,
      "loss": 1.5657214164733886,
      "mean_token_accuracy": 0.6575634054839611,
      "num_tokens": 1879251.0,
      "step": 950
    },
    {
      "entropy": 1.376376987993717,
      "epoch": 0.017245250819598507,
      "grad_norm": 22.25,
      "learning_rate": 4.795e-05,
      "loss": 1.505562686920166,
      "mean_token_accuracy": 0.6564014129340648,
      "num_tokens": 1898911.0,
      "step": 960
    },
    {
      "entropy": 1.5325671911239624,
      "epoch": 0.01742488884896933,
      "grad_norm": 14.875,
      "learning_rate": 4.845e-05,
      "loss": 1.6113946914672852,
      "mean_token_accuracy": 0.6567016690969467,
      "num_tokens": 1918843.0,
      "step": 970
    },
    {
      "entropy": 1.6021498128771783,
      "epoch": 0.017604526878340146,
      "grad_norm": 18.625,
      "learning_rate": 4.8950000000000004e-05,
      "loss": 1.7239124298095703,
      "mean_token_accuracy": 0.6299502395093441,
      "num_tokens": 1939280.0,
      "step": 980
    },
    {
      "entropy": 1.5801695257425308,
      "epoch": 0.017784164907710964,
      "grad_norm": 12.5,
      "learning_rate": 4.945e-05,
      "loss": 1.6025472640991212,
      "mean_token_accuracy": 0.6534447938203811,
      "num_tokens": 1959517.0,
      "step": 990
    },
    {
      "epoch": 0.01796380293708178,
      "eval_entropy": 1.6879122581481933,
      "eval_loss": 1.6562248468399048,
      "eval_mean_token_accuracy": 0.6401306872367859,
      "eval_num_tokens": 1977717.0,
      "eval_runtime": 40.5307,
      "eval_samples_per_second": 49.345,
      "eval_steps_per_second": 6.168,
      "step": 1000
    },
    {
      "entropy": 1.5339436739683152,
      "epoch": 0.0181434409664526,
      "grad_norm": 12.875,
      "learning_rate": 4.99999966562916e-05,
      "loss": 1.6479965209960938,
      "mean_token_accuracy": 0.6450760822743178,
      "num_tokens": 1997777.0,
      "step": 1010
    },
    {
      "entropy": 1.481242610514164,
      "epoch": 0.018323078995823416,
      "grad_norm": 11.9375,
      "learning_rate": 4.999998509779458e-05,
      "loss": 1.5723793983459473,
      "mean_token_accuracy": 0.6561716787517071,
      "num_tokens": 2017891.0,
      "step": 1020
    },
    {
      "entropy": 1.3482632651925086,
      "epoch": 0.018502717025194234,
      "grad_norm": 14.375,
      "learning_rate": 4.99999652832324e-05,
      "loss": 1.5132084846496583,
      "mean_token_accuracy": 0.6613622292876243,
      "num_tokens": 2037900.0,
      "step": 1030
    },
    {
      "entropy": 1.7140410661697387,
      "epoch": 0.01868235505456505,
      "grad_norm": 13.6875,
      "learning_rate": 4.9999937212611616e-05,
      "loss": 1.825594711303711,
      "mean_token_accuracy": 0.6063587903976441,
      "num_tokens": 2058263.0,
      "step": 1040
    },
    {
      "entropy": 1.69788548797369,
      "epoch": 0.01886199308393587,
      "grad_norm": 14.4375,
      "learning_rate": 4.9999900885941476e-05,
      "loss": 1.7420467376708983,
      "mean_token_accuracy": 0.6239760592579842,
      "num_tokens": 2078942.0,
      "step": 1050
    },
    {
      "entropy": 1.521554161608219,
      "epoch": 0.019041631113306687,
      "grad_norm": 18.0,
      "learning_rate": 4.9999856303234e-05,
      "loss": 1.7311466217041016,
      "mean_token_accuracy": 0.6197005368769168,
      "num_tokens": 2097309.0,
      "step": 1060
    },
    {
      "entropy": 1.5711013540625571,
      "epoch": 0.019221269142677504,
      "grad_norm": 16.375,
      "learning_rate": 4.99998034645039e-05,
      "loss": 1.6851322174072265,
      "mean_token_accuracy": 0.6406420513987541,
      "num_tokens": 2116331.0,
      "step": 1070
    },
    {
      "entropy": 1.517425474524498,
      "epoch": 0.019400907172048322,
      "grad_norm": 14.6875,
      "learning_rate": 4.999974236976862e-05,
      "loss": 1.5062079429626465,
      "mean_token_accuracy": 0.6739716276526451,
      "num_tokens": 2135284.0,
      "step": 1080
    },
    {
      "entropy": 1.402970278263092,
      "epoch": 0.01958054520141914,
      "grad_norm": 22.625,
      "learning_rate": 4.999967301904834e-05,
      "loss": 1.614433479309082,
      "mean_token_accuracy": 0.6460897564888001,
      "num_tokens": 2154764.0,
      "step": 1090
    },
    {
      "entropy": 1.4869570776820182,
      "epoch": 0.019760183230789957,
      "grad_norm": 15.3125,
      "learning_rate": 4.999959541236597e-05,
      "loss": 1.5126214981079102,
      "mean_token_accuracy": 0.6601682506501675,
      "num_tokens": 2175073.0,
      "step": 1100
    },
    {
      "entropy": 1.4203249007463454,
      "epoch": 0.019939821260160775,
      "grad_norm": 13.125,
      "learning_rate": 4.9999509549747143e-05,
      "loss": 1.6122428894042968,
      "mean_token_accuracy": 0.6398775935173034,
      "num_tokens": 2195533.0,
      "step": 1110
    },
    {
      "entropy": 1.7457217395305633,
      "epoch": 0.020119459289531592,
      "grad_norm": 14.25,
      "learning_rate": 4.99994154312202e-05,
      "loss": 1.7978387832641602,
      "mean_token_accuracy": 0.6236461505293847,
      "num_tokens": 2215392.0,
      "step": 1120
    },
    {
      "entropy": 1.4203085482120514,
      "epoch": 0.02029909731890241,
      "grad_norm": 13.3125,
      "learning_rate": 4.999931305681622e-05,
      "loss": 1.535835361480713,
      "mean_token_accuracy": 0.66737802028656,
      "num_tokens": 2234775.0,
      "step": 1130
    },
    {
      "entropy": 1.6592785954475402,
      "epoch": 0.02047873534827323,
      "grad_norm": 13.6875,
      "learning_rate": 4.9999202426569024e-05,
      "loss": 1.717864418029785,
      "mean_token_accuracy": 0.6368574008345604,
      "num_tokens": 2253731.0,
      "step": 1140
    },
    {
      "entropy": 1.5235640779137611,
      "epoch": 0.02065837337764405,
      "grad_norm": 11.3125,
      "learning_rate": 4.999908354051515e-05,
      "loss": 1.6711483001708984,
      "mean_token_accuracy": 0.6409434221684933,
      "num_tokens": 2273704.0,
      "step": 1150
    },
    {
      "entropy": 1.5913785725831986,
      "epoch": 0.020838011407014866,
      "grad_norm": 10.75,
      "learning_rate": 4.999895639869385e-05,
      "loss": 1.6213735580444335,
      "mean_token_accuracy": 0.6473274938762188,
      "num_tokens": 2294883.0,
      "step": 1160
    },
    {
      "entropy": 1.4578886479139328,
      "epoch": 0.021017649436385684,
      "grad_norm": 17.375,
      "learning_rate": 4.999882100114711e-05,
      "loss": 1.5965503692626952,
      "mean_token_accuracy": 0.6535281084477902,
      "num_tokens": 2315060.0,
      "step": 1170
    },
    {
      "entropy": 1.703429938852787,
      "epoch": 0.0211972874657565,
      "grad_norm": 11.5,
      "learning_rate": 4.999867734791965e-05,
      "loss": 1.7444644927978517,
      "mean_token_accuracy": 0.6244046315550804,
      "num_tokens": 2335097.0,
      "step": 1180
    },
    {
      "entropy": 1.4013918712735176,
      "epoch": 0.02137692549512732,
      "grad_norm": 10.375,
      "learning_rate": 4.9998525439058906e-05,
      "loss": 1.4863811492919923,
      "mean_token_accuracy": 0.6667556263506412,
      "num_tokens": 2355374.0,
      "step": 1190
    },
    {
      "entropy": 1.5714426934719086,
      "epoch": 0.021556563524498137,
      "grad_norm": 10.875,
      "learning_rate": 4.999836527461505e-05,
      "loss": 1.6421844482421875,
      "mean_token_accuracy": 0.6483815163373947,
      "num_tokens": 2375236.0,
      "step": 1200
    },
    {
      "entropy": 1.48934256285429,
      "epoch": 0.021736201553868954,
      "grad_norm": 12.6875,
      "learning_rate": 4.999819685464097e-05,
      "loss": 1.635277557373047,
      "mean_token_accuracy": 0.6506827063858509,
      "num_tokens": 2395068.0,
      "step": 1210
    },
    {
      "entropy": 1.6177606523036956,
      "epoch": 0.021915839583239772,
      "grad_norm": 13.75,
      "learning_rate": 4.999802017919229e-05,
      "loss": 1.7069984436035157,
      "mean_token_accuracy": 0.6416367538273334,
      "num_tokens": 2414218.0,
      "step": 1220
    },
    {
      "entropy": 1.6434327408671379,
      "epoch": 0.02209547761261059,
      "grad_norm": 12.6875,
      "learning_rate": 4.999783524832735e-05,
      "loss": 1.7221960067749023,
      "mean_token_accuracy": 0.6327094666659832,
      "num_tokens": 2433952.0,
      "step": 1230
    },
    {
      "entropy": 1.5641241088509559,
      "epoch": 0.022275115641981407,
      "grad_norm": 13.5,
      "learning_rate": 4.999764206210723e-05,
      "loss": 1.5754161834716798,
      "mean_token_accuracy": 0.65119748711586,
      "num_tokens": 2453433.0,
      "step": 1240
    },
    {
      "entropy": 1.5548364117741584,
      "epoch": 0.022454753671352225,
      "grad_norm": 10.4375,
      "learning_rate": 4.9997440620595724e-05,
      "loss": 1.745047378540039,
      "mean_token_accuracy": 0.6357388451695443,
      "num_tokens": 2473929.0,
      "step": 1250
    },
    {
      "entropy": 1.5932861298322678,
      "epoch": 0.022634391700723042,
      "grad_norm": 10.75,
      "learning_rate": 4.999723092385935e-05,
      "loss": 1.6144153594970703,
      "mean_token_accuracy": 0.6489748314023018,
      "num_tokens": 2493661.0,
      "step": 1260
    },
    {
      "entropy": 1.5058840662240982,
      "epoch": 0.02281402973009386,
      "grad_norm": 13.5625,
      "learning_rate": 4.9997012971967375e-05,
      "loss": 1.6505386352539062,
      "mean_token_accuracy": 0.6422648474574089,
      "num_tokens": 2513309.0,
      "step": 1270
    },
    {
      "entropy": 1.517166194319725,
      "epoch": 0.022993667759464678,
      "grad_norm": 11.5,
      "learning_rate": 4.9996786764991766e-05,
      "loss": 1.576535701751709,
      "mean_token_accuracy": 0.6516160413622856,
      "num_tokens": 2533388.0,
      "step": 1280
    },
    {
      "entropy": 1.4109832972288132,
      "epoch": 0.023173305788835495,
      "grad_norm": 17.625,
      "learning_rate": 4.9996552303007216e-05,
      "loss": 1.5971531867980957,
      "mean_token_accuracy": 0.6632041566073894,
      "num_tokens": 2552530.0,
      "step": 1290
    },
    {
      "entropy": 1.5902836352586747,
      "epoch": 0.023352943818206313,
      "grad_norm": 11.5,
      "learning_rate": 4.999630958609117e-05,
      "loss": 1.5698233604431153,
      "mean_token_accuracy": 0.6543730370700359,
      "num_tokens": 2572748.0,
      "step": 1300
    },
    {
      "entropy": 1.4055397614836693,
      "epoch": 0.023532581847577134,
      "grad_norm": 13.0,
      "learning_rate": 4.999605861432378e-05,
      "loss": 1.6371458053588868,
      "mean_token_accuracy": 0.6445710822939873,
      "num_tokens": 2592065.0,
      "step": 1310
    },
    {
      "entropy": 1.6797343373298645,
      "epoch": 0.02371221987694795,
      "grad_norm": 12.3125,
      "learning_rate": 4.999579938778794e-05,
      "loss": 1.6646812438964844,
      "mean_token_accuracy": 0.6332690112292767,
      "num_tokens": 2611257.0,
      "step": 1320
    },
    {
      "entropy": 1.4525077223777771,
      "epoch": 0.02389185790631877,
      "grad_norm": 13.25,
      "learning_rate": 4.9995531906569224e-05,
      "loss": 1.5657713890075684,
      "mean_token_accuracy": 0.6668823085725307,
      "num_tokens": 2630888.0,
      "step": 1330
    },
    {
      "entropy": 1.4237727612257003,
      "epoch": 0.024071495935689587,
      "grad_norm": 11.75,
      "learning_rate": 4.9995256170755986e-05,
      "loss": 1.5716552734375,
      "mean_token_accuracy": 0.6479331746697425,
      "num_tokens": 2651428.0,
      "step": 1340
    },
    {
      "entropy": 1.4339660495519637,
      "epoch": 0.024251133965060404,
      "grad_norm": 9.375,
      "learning_rate": 4.99949721804393e-05,
      "loss": 1.4079453468322753,
      "mean_token_accuracy": 0.6819032713770866,
      "num_tokens": 2671526.0,
      "step": 1350
    },
    {
      "entropy": 1.6262118130922318,
      "epoch": 0.024430771994431222,
      "grad_norm": 14.1875,
      "learning_rate": 4.9994679935712926e-05,
      "loss": 1.8461585998535157,
      "mean_token_accuracy": 0.6129544958472252,
      "num_tokens": 2691008.0,
      "step": 1360
    },
    {
      "entropy": 1.732082599401474,
      "epoch": 0.02461041002380204,
      "grad_norm": 12.625,
      "learning_rate": 4.999437943667339e-05,
      "loss": 1.7744600296020507,
      "mean_token_accuracy": 0.6243151530623436,
      "num_tokens": 2710558.0,
      "step": 1370
    },
    {
      "entropy": 1.5513766437768937,
      "epoch": 0.024790048053172857,
      "grad_norm": 10.3125,
      "learning_rate": 4.999407068341992e-05,
      "loss": 1.7053058624267579,
      "mean_token_accuracy": 0.6329125322401523,
      "num_tokens": 2731179.0,
      "step": 1380
    },
    {
      "entropy": 1.5373933896422387,
      "epoch": 0.024969686082543675,
      "grad_norm": 13.9375,
      "learning_rate": 4.99937536760545e-05,
      "loss": 1.616126823425293,
      "mean_token_accuracy": 0.6419561401009559,
      "num_tokens": 2750381.0,
      "step": 1390
    },
    {
      "entropy": 1.4567629858851432,
      "epoch": 0.025149324111914492,
      "grad_norm": 8.75,
      "learning_rate": 4.9993428414681796e-05,
      "loss": 1.5642082214355468,
      "mean_token_accuracy": 0.6601189814507962,
      "num_tokens": 2769782.0,
      "step": 1400
    },
    {
      "entropy": 1.507647468149662,
      "epoch": 0.02532896214128531,
      "grad_norm": 16.375,
      "learning_rate": 4.999309489940924e-05,
      "loss": 1.5823881149291992,
      "mean_token_accuracy": 0.650723684579134,
      "num_tokens": 2789949.0,
      "step": 1410
    },
    {
      "entropy": 1.468251334130764,
      "epoch": 0.025508600170656127,
      "grad_norm": 10.4375,
      "learning_rate": 4.999275313034696e-05,
      "loss": 1.5939903259277344,
      "mean_token_accuracy": 0.6539834812283516,
      "num_tokens": 2810085.0,
      "step": 1420
    },
    {
      "entropy": 1.573106411099434,
      "epoch": 0.025688238200026945,
      "grad_norm": 11.0625,
      "learning_rate": 4.999240310760783e-05,
      "loss": 1.6474645614624024,
      "mean_token_accuracy": 0.6424420893192291,
      "num_tokens": 2828371.0,
      "step": 1430
    },
    {
      "entropy": 1.5729375794529914,
      "epoch": 0.025867876229397763,
      "grad_norm": 10.5,
      "learning_rate": 4.999204483130744e-05,
      "loss": 1.6772174835205078,
      "mean_token_accuracy": 0.6429912939667701,
      "num_tokens": 2848250.0,
      "step": 1440
    },
    {
      "entropy": 1.592067340016365,
      "epoch": 0.02604751425876858,
      "grad_norm": 12.0,
      "learning_rate": 4.999167830156411e-05,
      "loss": 1.6595354080200195,
      "mean_token_accuracy": 0.6351709939539433,
      "num_tokens": 2867489.0,
      "step": 1450
    },
    {
      "entropy": 1.377654667198658,
      "epoch": 0.026227152288139398,
      "grad_norm": 10.625,
      "learning_rate": 4.9991303518498884e-05,
      "loss": 1.5140768051147462,
      "mean_token_accuracy": 0.6695582553744316,
      "num_tokens": 2887394.0,
      "step": 1460
    },
    {
      "entropy": 1.5429529666900634,
      "epoch": 0.026406790317510215,
      "grad_norm": 11.0,
      "learning_rate": 4.9990920482235526e-05,
      "loss": 1.669426155090332,
      "mean_token_accuracy": 0.6400448359549046,
      "num_tokens": 2906814.0,
      "step": 1470
    },
    {
      "entropy": 1.5632092237472535,
      "epoch": 0.026586428346881033,
      "grad_norm": 10.0,
      "learning_rate": 4.999052919290054e-05,
      "loss": 1.6766624450683594,
      "mean_token_accuracy": 0.6423386096954345,
      "num_tokens": 2927477.0,
      "step": 1480
    },
    {
      "entropy": 1.4513872116804123,
      "epoch": 0.026766066376251854,
      "grad_norm": 9.9375,
      "learning_rate": 4.9990129650623144e-05,
      "loss": 1.5277776718139648,
      "mean_token_accuracy": 0.6600641779601574,
      "num_tokens": 2946858.0,
      "step": 1490
    },
    {
      "entropy": 1.4783962324261666,
      "epoch": 0.026945704405622672,
      "grad_norm": 12.25,
      "learning_rate": 4.998972185553528e-05,
      "loss": 1.5461612701416017,
      "mean_token_accuracy": 0.6608979612588882,
      "num_tokens": 2966658.0,
      "step": 1500
    },
    {
      "entropy": 1.4329083040356636,
      "epoch": 0.02712534243499349,
      "grad_norm": 9.125,
      "learning_rate": 4.998930580777162e-05,
      "loss": 1.5002537727355958,
      "mean_token_accuracy": 0.6745608203113079,
      "num_tokens": 2985784.0,
      "step": 1510
    },
    {
      "entropy": 1.5033440753817557,
      "epoch": 0.027304980464364307,
      "grad_norm": 14.5,
      "learning_rate": 4.998888150746957e-05,
      "loss": 1.591783046722412,
      "mean_token_accuracy": 0.6536013603210449,
      "num_tokens": 3005268.0,
      "step": 1520
    },
    {
      "entropy": 1.5715586751699449,
      "epoch": 0.027484618493735125,
      "grad_norm": 11.125,
      "learning_rate": 4.998844895476923e-05,
      "loss": 1.5852946281433105,
      "mean_token_accuracy": 0.6514325991272927,
      "num_tokens": 3024640.0,
      "step": 1530
    },
    {
      "entropy": 1.4802327036857605,
      "epoch": 0.027664256523105942,
      "grad_norm": 10.0,
      "learning_rate": 4.998800814981346e-05,
      "loss": 1.568634605407715,
      "mean_token_accuracy": 0.6610087245702744,
      "num_tokens": 3044144.0,
      "step": 1540
    },
    {
      "entropy": 1.552686908841133,
      "epoch": 0.02784389455247676,
      "grad_norm": 9.5,
      "learning_rate": 4.998755909274785e-05,
      "loss": 1.6198490142822266,
      "mean_token_accuracy": 0.6509311467409133,
      "num_tokens": 3064442.0,
      "step": 1550
    },
    {
      "entropy": 1.4523080557584762,
      "epoch": 0.028023532581847577,
      "grad_norm": 15.75,
      "learning_rate": 4.998710178372067e-05,
      "loss": 1.621797752380371,
      "mean_token_accuracy": 0.6542368687689304,
      "num_tokens": 3083210.0,
      "step": 1560
    },
    {
      "entropy": 1.6241023704409598,
      "epoch": 0.028203170611218395,
      "grad_norm": 8.9375,
      "learning_rate": 4.998663622288296e-05,
      "loss": 1.707360076904297,
      "mean_token_accuracy": 0.6304206505417824,
      "num_tokens": 3103130.0,
      "step": 1570
    },
    {
      "entropy": 1.459706774353981,
      "epoch": 0.028382808640589213,
      "grad_norm": 9.375,
      "learning_rate": 4.998616241038846e-05,
      "loss": 1.4722458839416503,
      "mean_token_accuracy": 0.6724917612969875,
      "num_tokens": 3122604.0,
      "step": 1580
    },
    {
      "entropy": 1.3426085755228996,
      "epoch": 0.02856244666996003,
      "grad_norm": 37.25,
      "learning_rate": 4.9985680346393646e-05,
      "loss": 1.6074480056762694,
      "mean_token_accuracy": 0.6560868576169014,
      "num_tokens": 3142076.0,
      "step": 1590
    },
    {
      "entropy": 1.366212722659111,
      "epoch": 0.028742084699330848,
      "grad_norm": 9.4375,
      "learning_rate": 4.9985190031057727e-05,
      "loss": 1.3982316970825195,
      "mean_token_accuracy": 0.6974723257124424,
      "num_tokens": 3162989.0,
      "step": 1600
    },
    {
      "entropy": 1.321020522713661,
      "epoch": 0.028921722728701665,
      "grad_norm": 12.1875,
      "learning_rate": 4.998469146454261e-05,
      "loss": 1.444266414642334,
      "mean_token_accuracy": 0.6938928335905075,
      "num_tokens": 3181845.0,
      "step": 1610
    },
    {
      "entropy": 1.4611617252230644,
      "epoch": 0.029101360758072483,
      "grad_norm": 11.375,
      "learning_rate": 4.9984184647012945e-05,
      "loss": 1.5311830520629883,
      "mean_token_accuracy": 0.6752639405429364,
      "num_tokens": 3201278.0,
      "step": 1620
    },
    {
      "entropy": 1.4955737888813019,
      "epoch": 0.0292809987874433,
      "grad_norm": 9.5625,
      "learning_rate": 4.998366957863611e-05,
      "loss": 1.493691349029541,
      "mean_token_accuracy": 0.6672499142587185,
      "num_tokens": 3221704.0,
      "step": 1630
    },
    {
      "entropy": 1.3047339126467705,
      "epoch": 0.029460636816814118,
      "grad_norm": 11.375,
      "learning_rate": 4.99831462595822e-05,
      "loss": 1.4614850997924804,
      "mean_token_accuracy": 0.6840211108326912,
      "num_tokens": 3242445.0,
      "step": 1640
    },
    {
      "entropy": 1.3840580880641937,
      "epoch": 0.029640274846184936,
      "grad_norm": 14.875,
      "learning_rate": 4.9982614690024043e-05,
      "loss": 1.4759600639343262,
      "mean_token_accuracy": 0.6760420441627503,
      "num_tokens": 3262220.0,
      "step": 1650
    },
    {
      "entropy": 1.5510875165462494,
      "epoch": 0.029819912875555757,
      "grad_norm": 12.0,
      "learning_rate": 4.998207487013717e-05,
      "loss": 1.5977466583251954,
      "mean_token_accuracy": 0.6495107904076576,
      "num_tokens": 3281865.0,
      "step": 1660
    },
    {
      "entropy": 1.4049832925200463,
      "epoch": 0.029999550904926574,
      "grad_norm": 8.5,
      "learning_rate": 4.9981526800099875e-05,
      "loss": 1.4581579208374023,
      "mean_token_accuracy": 0.6835083514451981,
      "num_tokens": 3301008.0,
      "step": 1670
    },
    {
      "entropy": 1.5528059363365174,
      "epoch": 0.030179188934297392,
      "grad_norm": 11.5,
      "learning_rate": 4.998097048009315e-05,
      "loss": 1.6444576263427735,
      "mean_token_accuracy": 0.6547182284295558,
      "num_tokens": 3319908.0,
      "step": 1680
    },
    {
      "entropy": 1.2988245815038681,
      "epoch": 0.03035882696366821,
      "grad_norm": 10.5625,
      "learning_rate": 4.99804059103007e-05,
      "loss": 1.3999847412109374,
      "mean_token_accuracy": 0.6804589122533798,
      "num_tokens": 3340185.0,
      "step": 1690
    },
    {
      "entropy": 1.4380019143223763,
      "epoch": 0.030538464993039027,
      "grad_norm": 11.25,
      "learning_rate": 4.997983309090898e-05,
      "loss": 1.4962980270385742,
      "mean_token_accuracy": 0.6764723815023899,
      "num_tokens": 3359706.0,
      "step": 1700
    },
    {
      "entropy": 1.4233854606747627,
      "epoch": 0.030718103022409845,
      "grad_norm": 9.9375,
      "learning_rate": 4.997925202210715e-05,
      "loss": 1.539259910583496,
      "mean_token_accuracy": 0.6645733818411828,
      "num_tokens": 3379611.0,
      "step": 1710
    },
    {
      "entropy": 1.4388819113373756,
      "epoch": 0.030897741051780662,
      "grad_norm": 10.5625,
      "learning_rate": 4.997866270408712e-05,
      "loss": 1.5832198143005372,
      "mean_token_accuracy": 0.6639614477753639,
      "num_tokens": 3400687.0,
      "step": 1720
    },
    {
      "entropy": 1.5089255303144455,
      "epoch": 0.03107737908115148,
      "grad_norm": 12.9375,
      "learning_rate": 4.9978065137043506e-05,
      "loss": 1.5903237342834473,
      "mean_token_accuracy": 0.6547822445631027,
      "num_tokens": 3419916.0,
      "step": 1730
    },
    {
      "entropy": 1.4004312366247178,
      "epoch": 0.0312570171105223,
      "grad_norm": 10.5,
      "learning_rate": 4.997745932117364e-05,
      "loss": 1.4962909698486329,
      "mean_token_accuracy": 0.6745987348258495,
      "num_tokens": 3438709.0,
      "step": 1740
    },
    {
      "entropy": 1.4637741982936858,
      "epoch": 0.03143665513989312,
      "grad_norm": 9.9375,
      "learning_rate": 4.997684525667759e-05,
      "loss": 1.5000354766845703,
      "mean_token_accuracy": 0.6743131808936595,
      "num_tokens": 3458225.0,
      "step": 1750
    },
    {
      "entropy": 1.3026440516114235,
      "epoch": 0.03161629316926393,
      "grad_norm": 11.875,
      "learning_rate": 4.997622294375816e-05,
      "loss": 1.4002955436706543,
      "mean_token_accuracy": 0.692991704493761,
      "num_tokens": 3477608.0,
      "step": 1760
    },
    {
      "entropy": 1.443167109787464,
      "epoch": 0.031795931198634754,
      "grad_norm": 12.375,
      "learning_rate": 4.9975592382620836e-05,
      "loss": 1.4983144760131837,
      "mean_token_accuracy": 0.6700777739286423,
      "num_tokens": 3496123.0,
      "step": 1770
    },
    {
      "entropy": 1.459372314810753,
      "epoch": 0.03197556922800557,
      "grad_norm": 10.25,
      "learning_rate": 4.997495357347388e-05,
      "loss": 1.489708137512207,
      "mean_token_accuracy": 0.6677679911255836,
      "num_tokens": 3516787.0,
      "step": 1780
    },
    {
      "entropy": 1.333537481725216,
      "epoch": 0.03215520725737639,
      "grad_norm": 8.625,
      "learning_rate": 4.997430651652826e-05,
      "loss": 1.4847636222839355,
      "mean_token_accuracy": 0.6733242832124233,
      "num_tokens": 3536899.0,
      "step": 1790
    },
    {
      "entropy": 1.4999468088150025,
      "epoch": 0.0323348452867472,
      "grad_norm": 11.625,
      "learning_rate": 4.997365121199764e-05,
      "loss": 1.5769845008850099,
      "mean_token_accuracy": 0.6588590689003467,
      "num_tokens": 3557046.0,
      "step": 1800
    },
    {
      "entropy": 1.452918392419815,
      "epoch": 0.032514483316118024,
      "grad_norm": 10.0625,
      "learning_rate": 4.997298766009844e-05,
      "loss": 1.5399716377258301,
      "mean_token_accuracy": 0.6588214591145516,
      "num_tokens": 3577721.0,
      "step": 1810
    },
    {
      "entropy": 1.4200744450092315,
      "epoch": 0.03269412134548884,
      "grad_norm": 9.9375,
      "learning_rate": 4.9972315861049794e-05,
      "loss": 1.4729167938232421,
      "mean_token_accuracy": 0.6796995468437672,
      "num_tokens": 3597102.0,
      "step": 1820
    },
    {
      "entropy": 1.2982208222150802,
      "epoch": 0.03287375937485966,
      "grad_norm": 9.4375,
      "learning_rate": 4.997163581507356e-05,
      "loss": 1.3422197341918944,
      "mean_token_accuracy": 0.6908485218882561,
      "num_tokens": 3617565.0,
      "step": 1830
    },
    {
      "entropy": 1.3019727289676666,
      "epoch": 0.033053397404230474,
      "grad_norm": 11.5625,
      "learning_rate": 4.997094752239431e-05,
      "loss": 1.4144463539123535,
      "mean_token_accuracy": 0.6866289287805557,
      "num_tokens": 3638110.0,
      "step": 1840
    },
    {
      "entropy": 1.5307150796055793,
      "epoch": 0.033233035433601295,
      "grad_norm": 10.1875,
      "learning_rate": 4.997025098323936e-05,
      "loss": 1.5762164115905761,
      "mean_token_accuracy": 0.6582541570067406,
      "num_tokens": 3657763.0,
      "step": 1850
    },
    {
      "entropy": 1.3525944530963898,
      "epoch": 0.03341267346297211,
      "grad_norm": 9.4375,
      "learning_rate": 4.996954619783873e-05,
      "loss": 1.4310758590698243,
      "mean_token_accuracy": 0.6845531776547432,
      "num_tokens": 3676808.0,
      "step": 1860
    },
    {
      "entropy": 1.4061302855610847,
      "epoch": 0.03359231149234293,
      "grad_norm": 11.0625,
      "learning_rate": 4.996883316642517e-05,
      "loss": 1.454031753540039,
      "mean_token_accuracy": 0.6834939956665039,
      "num_tokens": 3696335.0,
      "step": 1870
    },
    {
      "entropy": 1.476578514277935,
      "epoch": 0.033771949521713744,
      "grad_norm": 10.8125,
      "learning_rate": 4.9968111889234145e-05,
      "loss": 1.58303804397583,
      "mean_token_accuracy": 0.6579067260026932,
      "num_tokens": 3716773.0,
      "step": 1880
    },
    {
      "entropy": 1.5915799140930176,
      "epoch": 0.033951587551084565,
      "grad_norm": 11.625,
      "learning_rate": 4.996738236650388e-05,
      "loss": 1.6760948181152344,
      "mean_token_accuracy": 0.6308161959052085,
      "num_tokens": 3736866.0,
      "step": 1890
    },
    {
      "entropy": 1.4185971394181252,
      "epoch": 0.03413122558045538,
      "grad_norm": 8.75,
      "learning_rate": 4.9966644598475266e-05,
      "loss": 1.4916116714477539,
      "mean_token_accuracy": 0.6840248756110668,
      "num_tokens": 3757687.0,
      "step": 1900
    },
    {
      "entropy": 1.4899285048246385,
      "epoch": 0.0343108636098262,
      "grad_norm": 10.875,
      "learning_rate": 4.9965898585391954e-05,
      "loss": 1.5667666435241698,
      "mean_token_accuracy": 0.6564998932182788,
      "num_tokens": 3777462.0,
      "step": 1910
    },
    {
      "entropy": 1.4775958225131034,
      "epoch": 0.034490501639197015,
      "grad_norm": 8.125,
      "learning_rate": 4.996514432750031e-05,
      "loss": 1.6225557327270508,
      "mean_token_accuracy": 0.6544752031564712,
      "num_tokens": 3796725.0,
      "step": 1920
    },
    {
      "entropy": 1.3850451871752738,
      "epoch": 0.034670139668567836,
      "grad_norm": 11.5625,
      "learning_rate": 4.996438182504942e-05,
      "loss": 1.4020709991455078,
      "mean_token_accuracy": 0.6894868232309819,
      "num_tokens": 3816310.0,
      "step": 1930
    },
    {
      "entropy": 1.4130028799176215,
      "epoch": 0.03484977769793866,
      "grad_norm": 9.4375,
      "learning_rate": 4.996361107829111e-05,
      "loss": 1.5707258224487304,
      "mean_token_accuracy": 0.6510094694793225,
      "num_tokens": 3835270.0,
      "step": 1940
    },
    {
      "entropy": 1.4181913778185844,
      "epoch": 0.03502941572730947,
      "grad_norm": 6.6875,
      "learning_rate": 4.996283208747989e-05,
      "loss": 1.4030420303344726,
      "mean_token_accuracy": 0.687630844861269,
      "num_tokens": 3855525.0,
      "step": 1950
    },
    {
      "entropy": 1.387476322054863,
      "epoch": 0.03520905375668029,
      "grad_norm": 9.5625,
      "learning_rate": 4.996204485287304e-05,
      "loss": 1.5385560989379883,
      "mean_token_accuracy": 0.6627428747713566,
      "num_tokens": 3876068.0,
      "step": 1960
    },
    {
      "entropy": 1.5595558032393455,
      "epoch": 0.035388691786051106,
      "grad_norm": 9.1875,
      "learning_rate": 4.996124937473051e-05,
      "loss": 1.5708704948425294,
      "mean_token_accuracy": 0.659713289141655,
      "num_tokens": 3896088.0,
      "step": 1970
    },
    {
      "entropy": 1.3917503014206887,
      "epoch": 0.03556832981542193,
      "grad_norm": 9.5,
      "learning_rate": 4.996044565331502e-05,
      "loss": 1.4669230461120606,
      "mean_token_accuracy": 0.672278282046318,
      "num_tokens": 3914554.0,
      "step": 1980
    },
    {
      "entropy": 1.5031164228916167,
      "epoch": 0.03574796784479274,
      "grad_norm": 9.625,
      "learning_rate": 4.9959633688892e-05,
      "loss": 1.6079071044921875,
      "mean_token_accuracy": 0.6591591216623783,
      "num_tokens": 3934276.0,
      "step": 1990
    },
    {
      "epoch": 0.03592760587416356,
      "eval_entropy": 1.386294443845749,
      "eval_loss": 1.4577912092208862,
      "eval_mean_token_accuracy": 0.6798514029979705,
      "eval_num_tokens": 3953916.0,
      "eval_runtime": 40.4522,
      "eval_samples_per_second": 49.441,
      "eval_steps_per_second": 6.18,
      "step": 2000
    },
    {
      "entropy": 1.3852115996181964,
      "epoch": 0.036107243903534376,
      "grad_norm": 8.5,
      "learning_rate": 4.9957985032098636e-05,
      "loss": 1.4510122299194337,
      "mean_token_accuracy": 0.6766671903431416,
      "num_tokens": 3976478.0,
      "step": 2010
    },
    {
      "entropy": 1.3892260804772376,
      "epoch": 0.0362868819329052,
      "grad_norm": 14.4375,
      "learning_rate": 4.995714834027275e-05,
      "loss": 1.4879559516906737,
      "mean_token_accuracy": 0.6728622771799564,
      "num_tokens": 3996357.0,
      "step": 2020
    },
    {
      "entropy": 1.2689381957054138,
      "epoch": 0.03646651996227601,
      "grad_norm": 11.375,
      "learning_rate": 4.995630340652825e-05,
      "loss": 1.3114850044250488,
      "mean_token_accuracy": 0.7042328543961048,
      "num_tokens": 4015938.0,
      "step": 2030
    },
    {
      "entropy": 1.346160574257374,
      "epoch": 0.03664615799164683,
      "grad_norm": 8.8125,
      "learning_rate": 4.995545023114414e-05,
      "loss": 1.4239927291870118,
      "mean_token_accuracy": 0.6828582987189293,
      "num_tokens": 4035180.0,
      "step": 2040
    },
    {
      "entropy": 1.3286015927791595,
      "epoch": 0.03682579602101765,
      "grad_norm": 8.5,
      "learning_rate": 4.99545888144022e-05,
      "loss": 1.3946846961975097,
      "mean_token_accuracy": 0.6950214438140392,
      "num_tokens": 4054300.0,
      "step": 2050
    },
    {
      "entropy": 1.4283461213111877,
      "epoch": 0.03700543405038847,
      "grad_norm": 11.25,
      "learning_rate": 4.995371915658691e-05,
      "loss": 1.5292243957519531,
      "mean_token_accuracy": 0.6782881774008274,
      "num_tokens": 4073937.0,
      "step": 2060
    },
    {
      "entropy": 1.5037285730242729,
      "epoch": 0.03718507207975928,
      "grad_norm": 10.6875,
      "learning_rate": 4.995284125798545e-05,
      "loss": 1.622320556640625,
      "mean_token_accuracy": 0.6501829117536545,
      "num_tokens": 4094074.0,
      "step": 2070
    },
    {
      "entropy": 1.4100628808140754,
      "epoch": 0.0373647101091301,
      "grad_norm": 10.4375,
      "learning_rate": 4.995195511888774e-05,
      "loss": 1.441728401184082,
      "mean_token_accuracy": 0.6748277321457863,
      "num_tokens": 4113754.0,
      "step": 2080
    },
    {
      "entropy": 1.2534951508045196,
      "epoch": 0.03754434813850092,
      "grad_norm": 8.875,
      "learning_rate": 4.995106073958644e-05,
      "loss": 1.3250720977783204,
      "mean_token_accuracy": 0.7045919358730316,
      "num_tokens": 4132875.0,
      "step": 2090
    },
    {
      "entropy": 1.3869805589318276,
      "epoch": 0.03772398616787174,
      "grad_norm": 9.5,
      "learning_rate": 4.995015812037689e-05,
      "loss": 1.500474452972412,
      "mean_token_accuracy": 0.6728356026113034,
      "num_tokens": 4152888.0,
      "step": 2100
    },
    {
      "entropy": 1.407675413787365,
      "epoch": 0.03790362419724256,
      "grad_norm": 12.5625,
      "learning_rate": 4.994924726155719e-05,
      "loss": 1.471339511871338,
      "mean_token_accuracy": 0.6686058312654495,
      "num_tokens": 4172954.0,
      "step": 2110
    },
    {
      "entropy": 1.3001487225294113,
      "epoch": 0.038083262226613374,
      "grad_norm": 7.65625,
      "learning_rate": 4.994832816342814e-05,
      "loss": 1.355784797668457,
      "mean_token_accuracy": 0.701902487128973,
      "num_tokens": 4192622.0,
      "step": 2120
    },
    {
      "entropy": 1.322553563117981,
      "epoch": 0.038262900255984195,
      "grad_norm": 9.125,
      "learning_rate": 4.994740082629325e-05,
      "loss": 1.3366957664489747,
      "mean_token_accuracy": 0.698298542201519,
      "num_tokens": 4212532.0,
      "step": 2130
    },
    {
      "entropy": 1.2555802136659622,
      "epoch": 0.03844253828535501,
      "grad_norm": 11.6875,
      "learning_rate": 4.99464652504588e-05,
      "loss": 1.3966214179992675,
      "mean_token_accuracy": 0.6893934711813927,
      "num_tokens": 4232317.0,
      "step": 2140
    },
    {
      "entropy": 1.3680005744099617,
      "epoch": 0.03862217631472583,
      "grad_norm": 9.375,
      "learning_rate": 4.9945521436233734e-05,
      "loss": 1.4108839988708497,
      "mean_token_accuracy": 0.676778818666935,
      "num_tokens": 4251961.0,
      "step": 2150
    },
    {
      "entropy": 1.3847421169281007,
      "epoch": 0.038801814344096644,
      "grad_norm": 8.5,
      "learning_rate": 4.9944569383929736e-05,
      "loss": 1.4746883392333985,
      "mean_token_accuracy": 0.6758048750460148,
      "num_tokens": 4271400.0,
      "step": 2160
    },
    {
      "entropy": 1.4548217207193375,
      "epoch": 0.038981452373467465,
      "grad_norm": 7.125,
      "learning_rate": 4.9943609093861224e-05,
      "loss": 1.5626228332519532,
      "mean_token_accuracy": 0.6512648619711399,
      "num_tokens": 4292287.0,
      "step": 2170
    },
    {
      "entropy": 1.4191979765892029,
      "epoch": 0.03916109040283828,
      "grad_norm": 9.25,
      "learning_rate": 4.9942640566345325e-05,
      "loss": 1.4875967025756835,
      "mean_token_accuracy": 0.6716679044067859,
      "num_tokens": 4311407.0,
      "step": 2180
    },
    {
      "entropy": 1.2986303925514222,
      "epoch": 0.0393407284322091,
      "grad_norm": 14.5625,
      "learning_rate": 4.994166380170189e-05,
      "loss": 1.344566249847412,
      "mean_token_accuracy": 0.7027465403079987,
      "num_tokens": 4331325.0,
      "step": 2190
    },
    {
      "entropy": 1.3378031596541404,
      "epoch": 0.039520366461579914,
      "grad_norm": 7.9375,
      "learning_rate": 4.994067880025349e-05,
      "loss": 1.4125244140625,
      "mean_token_accuracy": 0.6898518830537796,
      "num_tokens": 4351787.0,
      "step": 2200
    },
    {
      "entropy": 1.3462848573923112,
      "epoch": 0.039700004490950735,
      "grad_norm": 8.8125,
      "learning_rate": 4.9939685562325405e-05,
      "loss": 1.444535255432129,
      "mean_token_accuracy": 0.6816151268780232,
      "num_tokens": 4372412.0,
      "step": 2210
    },
    {
      "entropy": 1.4076301500201225,
      "epoch": 0.03987964252032155,
      "grad_norm": 8.625,
      "learning_rate": 4.993868408824566e-05,
      "loss": 1.440959644317627,
      "mean_token_accuracy": 0.6924315065145492,
      "num_tokens": 4392563.0,
      "step": 2220
    },
    {
      "entropy": 1.4046123221516609,
      "epoch": 0.04005928054969237,
      "grad_norm": 10.0,
      "learning_rate": 4.993767437834496e-05,
      "loss": 1.553008460998535,
      "mean_token_accuracy": 0.6718848347663879,
      "num_tokens": 4411924.0,
      "step": 2230
    },
    {
      "entropy": 1.2542652651667594,
      "epoch": 0.040238918579063185,
      "grad_norm": 7.4375,
      "learning_rate": 4.993665643295678e-05,
      "loss": 1.3188764572143554,
      "mean_token_accuracy": 0.6935629293322563,
      "num_tokens": 4431340.0,
      "step": 2240
    },
    {
      "entropy": 1.3614250645041466,
      "epoch": 0.040418556608434006,
      "grad_norm": 8.875,
      "learning_rate": 4.993563025241728e-05,
      "loss": 1.4111050605773925,
      "mean_token_accuracy": 0.689359862357378,
      "num_tokens": 4450515.0,
      "step": 2250
    },
    {
      "entropy": 1.310494926571846,
      "epoch": 0.04059819463780482,
      "grad_norm": 6.8125,
      "learning_rate": 4.993459583706534e-05,
      "loss": 1.3224611282348633,
      "mean_token_accuracy": 0.7060503855347633,
      "num_tokens": 4470031.0,
      "step": 2260
    },
    {
      "entropy": 1.3604912854731084,
      "epoch": 0.04077783266717564,
      "grad_norm": 9.5,
      "learning_rate": 4.9933553187242586e-05,
      "loss": 1.5020469665527343,
      "mean_token_accuracy": 0.6709420837461948,
      "num_tokens": 4490672.0,
      "step": 2270
    },
    {
      "entropy": 1.3310826301574707,
      "epoch": 0.04095747069654646,
      "grad_norm": 9.4375,
      "learning_rate": 4.9932502303293336e-05,
      "loss": 1.3717524528503418,
      "mean_token_accuracy": 0.6959741517901421,
      "num_tokens": 4509494.0,
      "step": 2280
    },
    {
      "entropy": 1.3372018203139304,
      "epoch": 0.041137108725917276,
      "grad_norm": 10.125,
      "learning_rate": 4.993144318556463e-05,
      "loss": 1.4499279975891113,
      "mean_token_accuracy": 0.678582813590765,
      "num_tokens": 4528424.0,
      "step": 2290
    },
    {
      "entropy": 1.4755570113658905,
      "epoch": 0.0413167467552881,
      "grad_norm": 8.1875,
      "learning_rate": 4.9930375834406245e-05,
      "loss": 1.5132333755493164,
      "mean_token_accuracy": 0.6683591969311238,
      "num_tokens": 4548511.0,
      "step": 2300
    },
    {
      "entropy": 1.2982542484998703,
      "epoch": 0.04149638478465891,
      "grad_norm": 7.03125,
      "learning_rate": 4.992930025017066e-05,
      "loss": 1.3485528945922851,
      "mean_token_accuracy": 0.6985410317778588,
      "num_tokens": 4568884.0,
      "step": 2310
    },
    {
      "entropy": 1.2956511452794075,
      "epoch": 0.04167602281402973,
      "grad_norm": 7.125,
      "learning_rate": 4.992821643321308e-05,
      "loss": 1.3949552536010743,
      "mean_token_accuracy": 0.6817234754562378,
      "num_tokens": 4589343.0,
      "step": 2320
    },
    {
      "entropy": 1.3517157122492791,
      "epoch": 0.04185566084340055,
      "grad_norm": 10.5625,
      "learning_rate": 4.992712438389143e-05,
      "loss": 1.423372459411621,
      "mean_token_accuracy": 0.6937249153852463,
      "num_tokens": 4608678.0,
      "step": 2330
    },
    {
      "entropy": 1.3493634089827538,
      "epoch": 0.04203529887277137,
      "grad_norm": 10.0625,
      "learning_rate": 4.992602410256635e-05,
      "loss": 1.3928610801696777,
      "mean_token_accuracy": 0.6890722371637821,
      "num_tokens": 4629429.0,
      "step": 2340
    },
    {
      "entropy": 1.2909897580742835,
      "epoch": 0.04221493690214218,
      "grad_norm": 9.25,
      "learning_rate": 4.99249155896012e-05,
      "loss": 1.306317138671875,
      "mean_token_accuracy": 0.7104037247598172,
      "num_tokens": 4649013.0,
      "step": 2350
    },
    {
      "entropy": 1.3496612012386322,
      "epoch": 0.042394574931513,
      "grad_norm": 7.875,
      "learning_rate": 4.9923798845362055e-05,
      "loss": 1.4225537300109863,
      "mean_token_accuracy": 0.6847415447235108,
      "num_tokens": 4667684.0,
      "step": 2360
    },
    {
      "entropy": 1.3888683900237084,
      "epoch": 0.04257421296088382,
      "grad_norm": 8.75,
      "learning_rate": 4.992267387021772e-05,
      "loss": 1.4505311965942382,
      "mean_token_accuracy": 0.6889954619109631,
      "num_tokens": 4688012.0,
      "step": 2370
    },
    {
      "entropy": 1.2660149484872818,
      "epoch": 0.04275385099025464,
      "grad_norm": 11.4375,
      "learning_rate": 4.9921540664539694e-05,
      "loss": 1.412290668487549,
      "mean_token_accuracy": 0.685637129098177,
      "num_tokens": 4706808.0,
      "step": 2380
    },
    {
      "entropy": 1.2213765501976013,
      "epoch": 0.04293348901962545,
      "grad_norm": 8.9375,
      "learning_rate": 4.992039922870223e-05,
      "loss": 1.3109620094299317,
      "mean_token_accuracy": 0.7055920973420143,
      "num_tokens": 4726385.0,
      "step": 2390
    },
    {
      "entropy": 1.4299337461590766,
      "epoch": 0.04311312704899627,
      "grad_norm": 9.0625,
      "learning_rate": 4.991924956308227e-05,
      "loss": 1.5099981307983399,
      "mean_token_accuracy": 0.6700854904949665,
      "num_tokens": 4745568.0,
      "step": 2400
    },
    {
      "entropy": 1.3655345007777213,
      "epoch": 0.04329276507836709,
      "grad_norm": 9.25,
      "learning_rate": 4.991809166805948e-05,
      "loss": 1.400827693939209,
      "mean_token_accuracy": 0.6930969566106796,
      "num_tokens": 4765320.0,
      "step": 2410
    },
    {
      "entropy": 1.3112824589014054,
      "epoch": 0.04347240310773791,
      "grad_norm": 9.0625,
      "learning_rate": 4.9916925544016246e-05,
      "loss": 1.3679177284240722,
      "mean_token_accuracy": 0.6980331242084503,
      "num_tokens": 4786248.0,
      "step": 2420
    },
    {
      "entropy": 1.2900094658136367,
      "epoch": 0.04365204113710872,
      "grad_norm": 7.0,
      "learning_rate": 4.991575119133768e-05,
      "loss": 1.3773300170898437,
      "mean_token_accuracy": 0.6965838596224785,
      "num_tokens": 4806161.0,
      "step": 2430
    },
    {
      "entropy": 1.3515348821878432,
      "epoch": 0.043831679166479544,
      "grad_norm": 8.375,
      "learning_rate": 4.991456861041161e-05,
      "loss": 1.4092613220214845,
      "mean_token_accuracy": 0.6918789118528366,
      "num_tokens": 4826047.0,
      "step": 2440
    },
    {
      "entropy": 1.4256507366895677,
      "epoch": 0.044011317195850365,
      "grad_norm": 14.875,
      "learning_rate": 4.991337780162854e-05,
      "loss": 1.3939717292785645,
      "mean_token_accuracy": 0.6889933459460735,
      "num_tokens": 4845778.0,
      "step": 2450
    },
    {
      "entropy": 1.3277788490056992,
      "epoch": 0.04419095522522118,
      "grad_norm": 9.25,
      "learning_rate": 4.991217876538177e-05,
      "loss": 1.388737964630127,
      "mean_token_accuracy": 0.6865644760429859,
      "num_tokens": 4865383.0,
      "step": 2460
    },
    {
      "entropy": 1.3833733096718788,
      "epoch": 0.044370593254592,
      "grad_norm": 9.0,
      "learning_rate": 4.991097150206724e-05,
      "loss": 1.4994051933288575,
      "mean_token_accuracy": 0.6717169411480427,
      "num_tokens": 4885618.0,
      "step": 2470
    },
    {
      "entropy": 1.2716542035341263,
      "epoch": 0.044550231283962814,
      "grad_norm": 8.4375,
      "learning_rate": 4.990975601208365e-05,
      "loss": 1.2832053184509278,
      "mean_token_accuracy": 0.7136360689997673,
      "num_tokens": 4906218.0,
      "step": 2480
    },
    {
      "entropy": 1.34033023416996,
      "epoch": 0.044729869313333635,
      "grad_norm": 9.375,
      "learning_rate": 4.990853229583242e-05,
      "loss": 1.436203956604004,
      "mean_token_accuracy": 0.6842044688761234,
      "num_tokens": 4927510.0,
      "step": 2490
    },
    {
      "entropy": 1.4591680765151978,
      "epoch": 0.04490950734270445,
      "grad_norm": 7.40625,
      "learning_rate": 4.9907300353717654e-05,
      "loss": 1.4958786964416504,
      "mean_token_accuracy": 0.6768533974885941,
      "num_tokens": 4947233.0,
      "step": 2500
    },
    {
      "entropy": 1.223151443898678,
      "epoch": 0.04508914537207527,
      "grad_norm": 16.75,
      "learning_rate": 4.990606018614621e-05,
      "loss": 1.2687625885009766,
      "mean_token_accuracy": 0.7144640415906907,
      "num_tokens": 4966394.0,
      "step": 2510
    },
    {
      "entropy": 1.321670588850975,
      "epoch": 0.045268783401446085,
      "grad_norm": 12.5,
      "learning_rate": 4.990481179352763e-05,
      "loss": 1.4832565307617187,
      "mean_token_accuracy": 0.6693767845630646,
      "num_tokens": 4985129.0,
      "step": 2520
    },
    {
      "entropy": 1.243193519115448,
      "epoch": 0.045448421430816906,
      "grad_norm": 8.6875,
      "learning_rate": 4.9903555176274195e-05,
      "loss": 1.259221649169922,
      "mean_token_accuracy": 0.7097675010561943,
      "num_tokens": 5005158.0,
      "step": 2530
    },
    {
      "entropy": 1.3275543317198752,
      "epoch": 0.04562805946018772,
      "grad_norm": 9.9375,
      "learning_rate": 4.9902290334800897e-05,
      "loss": 1.4005366325378419,
      "mean_token_accuracy": 0.6900383830070496,
      "num_tokens": 5025099.0,
      "step": 2540
    },
    {
      "entropy": 1.35167026668787,
      "epoch": 0.04580769748955854,
      "grad_norm": 9.4375,
      "learning_rate": 4.9901017269525425e-05,
      "loss": 1.3493748664855958,
      "mean_token_accuracy": 0.7060595467686653,
      "num_tokens": 5044943.0,
      "step": 2550
    },
    {
      "entropy": 1.1953320011496544,
      "epoch": 0.045987335518929355,
      "grad_norm": 9.1875,
      "learning_rate": 4.9899735980868214e-05,
      "loss": 1.279421615600586,
      "mean_token_accuracy": 0.7239278607070446,
      "num_tokens": 5064780.0,
      "step": 2560
    },
    {
      "entropy": 1.2834494024515153,
      "epoch": 0.046166973548300176,
      "grad_norm": 10.5,
      "learning_rate": 4.989844646925239e-05,
      "loss": 1.3923212051391602,
      "mean_token_accuracy": 0.6952004477381706,
      "num_tokens": 5084058.0,
      "step": 2570
    },
    {
      "entropy": 1.1845080688595773,
      "epoch": 0.04634661157767099,
      "grad_norm": 9.4375,
      "learning_rate": 4.9897148735103816e-05,
      "loss": 1.220262336730957,
      "mean_token_accuracy": 0.7271849617362023,
      "num_tokens": 5103486.0,
      "step": 2580
    },
    {
      "entropy": 1.4354370057582855,
      "epoch": 0.04652624960704181,
      "grad_norm": 9.5625,
      "learning_rate": 4.9895842778851056e-05,
      "loss": 1.5020055770874023,
      "mean_token_accuracy": 0.6726393468677998,
      "num_tokens": 5123695.0,
      "step": 2590
    },
    {
      "entropy": 1.3275725990533829,
      "epoch": 0.046705887636412625,
      "grad_norm": 7.65625,
      "learning_rate": 4.989452860092538e-05,
      "loss": 1.3116796493530274,
      "mean_token_accuracy": 0.7051602303981781,
      "num_tokens": 5142404.0,
      "step": 2600
    },
    {
      "entropy": 1.192495933175087,
      "epoch": 0.04688552566578345,
      "grad_norm": 10.0625,
      "learning_rate": 4.989320620176081e-05,
      "loss": 1.4025640487670898,
      "mean_token_accuracy": 0.688460073620081,
      "num_tokens": 5161318.0,
      "step": 2610
    },
    {
      "entropy": 1.2570716872811318,
      "epoch": 0.04706516369515427,
      "grad_norm": 7.0625,
      "learning_rate": 4.9891875581794033e-05,
      "loss": 1.3261414527893067,
      "mean_token_accuracy": 0.7105177581310272,
      "num_tokens": 5180927.0,
      "step": 2620
    },
    {
      "entropy": 1.337856189906597,
      "epoch": 0.04724480172452508,
      "grad_norm": 9.1875,
      "learning_rate": 4.98905367414645e-05,
      "loss": 1.3589582443237305,
      "mean_token_accuracy": 0.6933407261967659,
      "num_tokens": 5201323.0,
      "step": 2630
    },
    {
      "entropy": 1.3069262325763702,
      "epoch": 0.0474244397538959,
      "grad_norm": 9.5,
      "learning_rate": 4.9889189681214335e-05,
      "loss": 1.416452407836914,
      "mean_token_accuracy": 0.6833235710859299,
      "num_tokens": 5220469.0,
      "step": 2640
    },
    {
      "entropy": 1.422820433974266,
      "epoch": 0.04760407778326672,
      "grad_norm": 8.0625,
      "learning_rate": 4.9887834401488406e-05,
      "loss": 1.4213821411132812,
      "mean_token_accuracy": 0.683092050999403,
      "num_tokens": 5241302.0,
      "step": 2650
    },
    {
      "entropy": 1.2506478175520896,
      "epoch": 0.04778371581263754,
      "grad_norm": 10.125,
      "learning_rate": 4.988647090273428e-05,
      "loss": 1.396510696411133,
      "mean_token_accuracy": 0.6983902335166932,
      "num_tokens": 5261404.0,
      "step": 2660
    },
    {
      "entropy": 1.3648975372314454,
      "epoch": 0.04796335384200835,
      "grad_norm": 8.125,
      "learning_rate": 4.988509918540224e-05,
      "loss": 1.3885870933532716,
      "mean_token_accuracy": 0.6930775731801987,
      "num_tokens": 5281482.0,
      "step": 2670
    },
    {
      "entropy": 1.3311908036470412,
      "epoch": 0.04814299187137917,
      "grad_norm": 8.4375,
      "learning_rate": 4.988371924994529e-05,
      "loss": 1.3789891242980956,
      "mean_token_accuracy": 0.6992242433130741,
      "num_tokens": 5301408.0,
      "step": 2680
    },
    {
      "entropy": 1.3038343518972397,
      "epoch": 0.04832262990074999,
      "grad_norm": 7.6875,
      "learning_rate": 4.9882331096819145e-05,
      "loss": 1.3972562789916991,
      "mean_token_accuracy": 0.6883526988327503,
      "num_tokens": 5320443.0,
      "step": 2690
    },
    {
      "entropy": 1.311833919584751,
      "epoch": 0.04850226793012081,
      "grad_norm": 8.0625,
      "learning_rate": 4.9880934726482234e-05,
      "loss": 1.3173972129821778,
      "mean_token_accuracy": 0.7006198644638062,
      "num_tokens": 5340777.0,
      "step": 2700
    },
    {
      "entropy": 1.3987028837203979,
      "epoch": 0.04868190595949162,
      "grad_norm": 7.125,
      "learning_rate": 4.9879530139395695e-05,
      "loss": 1.5190353393554688,
      "mean_token_accuracy": 0.6705217584967613,
      "num_tokens": 5361079.0,
      "step": 2710
    },
    {
      "entropy": 1.300481103360653,
      "epoch": 0.048861543988862444,
      "grad_norm": 8.5,
      "learning_rate": 4.987811733602338e-05,
      "loss": 1.29676513671875,
      "mean_token_accuracy": 0.7194139674305916,
      "num_tokens": 5380992.0,
      "step": 2720
    },
    {
      "entropy": 1.294689479470253,
      "epoch": 0.04904118201823326,
      "grad_norm": 8.1875,
      "learning_rate": 4.987669631683185e-05,
      "loss": 1.4212841033935546,
      "mean_token_accuracy": 0.6948933452367783,
      "num_tokens": 5401446.0,
      "step": 2730
    },
    {
      "entropy": 1.26357045173645,
      "epoch": 0.04922082004760408,
      "grad_norm": 8.5625,
      "learning_rate": 4.987526708229041e-05,
      "loss": 1.331010627746582,
      "mean_token_accuracy": 0.6966155946254731,
      "num_tokens": 5420973.0,
      "step": 2740
    },
    {
      "entropy": 1.3671389058232308,
      "epoch": 0.04940045807697489,
      "grad_norm": 12.75,
      "learning_rate": 4.987382963287103e-05,
      "loss": 1.3746275901794434,
      "mean_token_accuracy": 0.6926651231944561,
      "num_tokens": 5441255.0,
      "step": 2750
    },
    {
      "entropy": 1.2061855614185333,
      "epoch": 0.049580096106345714,
      "grad_norm": 10.5,
      "learning_rate": 4.987238396904843e-05,
      "loss": 1.3396441459655761,
      "mean_token_accuracy": 0.7067731723189354,
      "num_tokens": 5461883.0,
      "step": 2760
    },
    {
      "entropy": 1.3966844588518144,
      "epoch": 0.04975973413571653,
      "grad_norm": 8.125,
      "learning_rate": 4.9870930091300034e-05,
      "loss": 1.4205124855041504,
      "mean_token_accuracy": 0.688123058527708,
      "num_tokens": 5481480.0,
      "step": 2770
    },
    {
      "entropy": 1.1706627994775771,
      "epoch": 0.04993937216508735,
      "grad_norm": 8.0625,
      "learning_rate": 4.986946800010597e-05,
      "loss": 1.2511099815368651,
      "mean_token_accuracy": 0.7223161771893502,
      "num_tokens": 5502785.0,
      "step": 2780
    },
    {
      "entropy": 1.281210759282112,
      "epoch": 0.05011901019445816,
      "grad_norm": 8.9375,
      "learning_rate": 4.986799769594907e-05,
      "loss": 1.387538242340088,
      "mean_token_accuracy": 0.6984438315033913,
      "num_tokens": 5524148.0,
      "step": 2790
    },
    {
      "entropy": 1.3482430651783943,
      "epoch": 0.050298648223828984,
      "grad_norm": 6.5,
      "learning_rate": 4.986651917931491e-05,
      "loss": 1.3431249618530274,
      "mean_token_accuracy": 0.704458849877119,
      "num_tokens": 5543845.0,
      "step": 2800
    },
    {
      "entropy": 1.2614577263593674,
      "epoch": 0.050478286253199806,
      "grad_norm": 6.46875,
      "learning_rate": 4.986503245069175e-05,
      "loss": 1.3213672637939453,
      "mean_token_accuracy": 0.7075468108057976,
      "num_tokens": 5564439.0,
      "step": 2810
    },
    {
      "entropy": 1.266054430603981,
      "epoch": 0.05065792428257062,
      "grad_norm": 9.4375,
      "learning_rate": 4.986353751057058e-05,
      "loss": 1.3275205612182617,
      "mean_token_accuracy": 0.7045779831707477,
      "num_tokens": 5583368.0,
      "step": 2820
    },
    {
      "entropy": 1.2890402510762216,
      "epoch": 0.05083756231194144,
      "grad_norm": 6.375,
      "learning_rate": 4.9862034359445076e-05,
      "loss": 1.3899471282958984,
      "mean_token_accuracy": 0.6825043484568596,
      "num_tokens": 5604202.0,
      "step": 2830
    },
    {
      "entropy": 1.3366580858826638,
      "epoch": 0.051017200341312255,
      "grad_norm": 8.1875,
      "learning_rate": 4.9860522997811664e-05,
      "loss": 1.3271031379699707,
      "mean_token_accuracy": 0.7065982103347779,
      "num_tokens": 5623725.0,
      "step": 2840
    },
    {
      "entropy": 1.2373004704713821,
      "epoch": 0.051196838370683076,
      "grad_norm": 8.75,
      "learning_rate": 4.985900342616945e-05,
      "loss": 1.3450727462768555,
      "mean_token_accuracy": 0.6976740017533303,
      "num_tokens": 5644627.0,
      "step": 2850
    },
    {
      "entropy": 1.369050419330597,
      "epoch": 0.05137647640005389,
      "grad_norm": 8.4375,
      "learning_rate": 4.985747564502026e-05,
      "loss": 1.4189981460571288,
      "mean_token_accuracy": 0.6970780231058598,
      "num_tokens": 5665570.0,
      "step": 2860
    },
    {
      "entropy": 1.1946534171700478,
      "epoch": 0.05155611442942471,
      "grad_norm": 8.0625,
      "learning_rate": 4.985593965486863e-05,
      "loss": 1.2828879356384277,
      "mean_token_accuracy": 0.7100204780697823,
      "num_tokens": 5685833.0,
      "step": 2870
    },
    {
      "entropy": 1.2458000168204308,
      "epoch": 0.051735752458795525,
      "grad_norm": 7.5,
      "learning_rate": 4.985439545622181e-05,
      "loss": 1.2325324058532714,
      "mean_token_accuracy": 0.7182733811438083,
      "num_tokens": 5706468.0,
      "step": 2880
    },
    {
      "entropy": 1.1770873591303825,
      "epoch": 0.051915390488166346,
      "grad_norm": 9.9375,
      "learning_rate": 4.985284304958978e-05,
      "loss": 1.2611387252807618,
      "mean_token_accuracy": 0.7225394740700721,
      "num_tokens": 5726485.0,
      "step": 2890
    },
    {
      "entropy": 1.279192292690277,
      "epoch": 0.05209502851753716,
      "grad_norm": 10.375,
      "learning_rate": 4.9851282435485184e-05,
      "loss": 1.3833558082580566,
      "mean_token_accuracy": 0.6912841096520423,
      "num_tokens": 5745620.0,
      "step": 2900
    },
    {
      "entropy": 1.2917813286185265,
      "epoch": 0.05227466654690798,
      "grad_norm": 7.78125,
      "learning_rate": 4.984971361442342e-05,
      "loss": 1.28564453125,
      "mean_token_accuracy": 0.7049105674028396,
      "num_tokens": 5764605.0,
      "step": 2910
    },
    {
      "entropy": 1.2992657899856568,
      "epoch": 0.052454304576278796,
      "grad_norm": 10.125,
      "learning_rate": 4.984813658692257e-05,
      "loss": 1.321458911895752,
      "mean_token_accuracy": 0.707523126900196,
      "num_tokens": 5784159.0,
      "step": 2920
    },
    {
      "entropy": 1.13544053286314,
      "epoch": 0.05263394260564962,
      "grad_norm": 5.84375,
      "learning_rate": 4.9846551353503445e-05,
      "loss": 1.2309046745300294,
      "mean_token_accuracy": 0.7213367313146591,
      "num_tokens": 5803812.0,
      "step": 2930
    },
    {
      "entropy": 1.3005485817790032,
      "epoch": 0.05281358063502043,
      "grad_norm": 10.25,
      "learning_rate": 4.984495791468956e-05,
      "loss": 1.4027809143066405,
      "mean_token_accuracy": 0.6931733332574368,
      "num_tokens": 5823166.0,
      "step": 2940
    },
    {
      "entropy": 1.254653562605381,
      "epoch": 0.05299321866439125,
      "grad_norm": 8.625,
      "learning_rate": 4.9843356271007125e-05,
      "loss": 1.300437068939209,
      "mean_token_accuracy": 0.705638687312603,
      "num_tokens": 5843386.0,
      "step": 2950
    },
    {
      "entropy": 1.3216019973158837,
      "epoch": 0.053172856693762066,
      "grad_norm": 7.25,
      "learning_rate": 4.984174642298507e-05,
      "loss": 1.359949493408203,
      "mean_token_accuracy": 0.7100481852889061,
      "num_tokens": 5862634.0,
      "step": 2960
    },
    {
      "entropy": 1.3008277341723442,
      "epoch": 0.05335249472313289,
      "grad_norm": 15.375,
      "learning_rate": 4.9840128371155046e-05,
      "loss": 1.368751621246338,
      "mean_token_accuracy": 0.6962978295981884,
      "num_tokens": 5881644.0,
      "step": 2970
    },
    {
      "entropy": 1.240821860730648,
      "epoch": 0.05353213275250371,
      "grad_norm": 8.5625,
      "learning_rate": 4.98385021160514e-05,
      "loss": 1.268119716644287,
      "mean_token_accuracy": 0.7124153539538384,
      "num_tokens": 5901449.0,
      "step": 2980
    },
    {
      "entropy": 1.2630082756280898,
      "epoch": 0.05371177078187452,
      "grad_norm": 7.3125,
      "learning_rate": 4.983686765821119e-05,
      "loss": 1.3535577774047851,
      "mean_token_accuracy": 0.7052506610751152,
      "num_tokens": 5920882.0,
      "step": 2990
    },
    {
      "epoch": 0.053891408811245343,
      "eval_entropy": 1.2305607693195344,
      "eval_loss": 1.2938153743743896,
      "eval_mean_token_accuracy": 0.7112859692573548,
      "eval_num_tokens": 5940705.0,
      "eval_runtime": 40.4865,
      "eval_samples_per_second": 49.399,
      "eval_steps_per_second": 6.175,
      "step": 3000
    },
    {
      "entropy": 1.212465363740921,
      "epoch": 0.05407104684061616,
      "grad_norm": 7.25,
      "learning_rate": 4.983357413648286e-05,
      "loss": 1.2685124397277832,
      "mean_token_accuracy": 0.7060285702347755,
      "num_tokens": 5959971.0,
      "step": 3010
    },
    {
      "entropy": 1.2705860674381255,
      "epoch": 0.05425068486998698,
      "grad_norm": 9.375,
      "learning_rate": 4.98319150736824e-05,
      "loss": 1.31303129196167,
      "mean_token_accuracy": 0.7058427125215531,
      "num_tokens": 5981168.0,
      "step": 3020
    },
    {
      "entropy": 1.2937784075737,
      "epoch": 0.05443032289935779,
      "grad_norm": 11.0625,
      "learning_rate": 4.9830247810320694e-05,
      "loss": 1.3791098594665527,
      "mean_token_accuracy": 0.6943519517779351,
      "num_tokens": 6001372.0,
      "step": 3030
    },
    {
      "entropy": 1.3318655461072921,
      "epoch": 0.054609960928728614,
      "grad_norm": 7.28125,
      "learning_rate": 4.982857234694836e-05,
      "loss": 1.3658021926879882,
      "mean_token_accuracy": 0.6942898020148277,
      "num_tokens": 6021178.0,
      "step": 3040
    },
    {
      "entropy": 1.3197183042764664,
      "epoch": 0.05478959895809943,
      "grad_norm": 9.125,
      "learning_rate": 4.982688868411869e-05,
      "loss": 1.404970359802246,
      "mean_token_accuracy": 0.7003599464893341,
      "num_tokens": 6041764.0,
      "step": 3050
    },
    {
      "entropy": 1.2844930931925773,
      "epoch": 0.05496923698747025,
      "grad_norm": 7.40625,
      "learning_rate": 4.9825196822387705e-05,
      "loss": 1.2841452598571776,
      "mean_token_accuracy": 0.7065952852368355,
      "num_tokens": 6062755.0,
      "step": 3060
    },
    {
      "entropy": 1.2265906885266304,
      "epoch": 0.05514887501684106,
      "grad_norm": 8.75,
      "learning_rate": 4.982349676231414e-05,
      "loss": 1.2465375900268554,
      "mean_token_accuracy": 0.7151035532355309,
      "num_tokens": 6082483.0,
      "step": 3070
    },
    {
      "entropy": 1.1339795097708703,
      "epoch": 0.055328513046211884,
      "grad_norm": 6.3125,
      "learning_rate": 4.982178850445941e-05,
      "loss": 1.2763813972473144,
      "mean_token_accuracy": 0.7256345644593238,
      "num_tokens": 6102407.0,
      "step": 3080
    },
    {
      "entropy": 1.270952996611595,
      "epoch": 0.0555081510755827,
      "grad_norm": 10.0625,
      "learning_rate": 4.982007204938768e-05,
      "loss": 1.3460318565368652,
      "mean_token_accuracy": 0.7137261286377907,
      "num_tokens": 6121728.0,
      "step": 3090
    },
    {
      "entropy": 1.2569975599646568,
      "epoch": 0.05568778910495352,
      "grad_norm": 7.09375,
      "learning_rate": 4.981834739766578e-05,
      "loss": 1.2724987030029298,
      "mean_token_accuracy": 0.7073746204376221,
      "num_tokens": 6143110.0,
      "step": 3100
    },
    {
      "entropy": 1.2817855432629586,
      "epoch": 0.055867427134324334,
      "grad_norm": 10.5625,
      "learning_rate": 4.981661454986326e-05,
      "loss": 1.3831307411193847,
      "mean_token_accuracy": 0.6954029813408852,
      "num_tokens": 6162077.0,
      "step": 3110
    },
    {
      "entropy": 1.2735674560070038,
      "epoch": 0.056047065163695155,
      "grad_norm": 8.75,
      "learning_rate": 4.981487350655239e-05,
      "loss": 1.2766504287719727,
      "mean_token_accuracy": 0.7066126257181168,
      "num_tokens": 6181108.0,
      "step": 3120
    },
    {
      "entropy": 1.1311923533678054,
      "epoch": 0.05622670319306597,
      "grad_norm": 7.28125,
      "learning_rate": 4.981312426830813e-05,
      "loss": 1.2055017471313476,
      "mean_token_accuracy": 0.7288956001400948,
      "num_tokens": 6200424.0,
      "step": 3130
    },
    {
      "entropy": 1.3666249245405198,
      "epoch": 0.05640634122243679,
      "grad_norm": 9.1875,
      "learning_rate": 4.981136683570816e-05,
      "loss": 1.4029760360717773,
      "mean_token_accuracy": 0.686952055990696,
      "num_tokens": 6220501.0,
      "step": 3140
    },
    {
      "entropy": 1.2993694990873337,
      "epoch": 0.05658597925180761,
      "grad_norm": 10.75,
      "learning_rate": 4.980960120933287e-05,
      "loss": 1.2893091201782227,
      "mean_token_accuracy": 0.7055289223790169,
      "num_tokens": 6240048.0,
      "step": 3150
    },
    {
      "entropy": 1.1119103491306306,
      "epoch": 0.056765617281178425,
      "grad_norm": 7.4375,
      "learning_rate": 4.9807827389765316e-05,
      "loss": 1.2223850250244142,
      "mean_token_accuracy": 0.7190957799553871,
      "num_tokens": 6260108.0,
      "step": 3160
    },
    {
      "entropy": 1.195087979733944,
      "epoch": 0.056945255310549246,
      "grad_norm": 8.375,
      "learning_rate": 4.9806045377591317e-05,
      "loss": 1.268524742126465,
      "mean_token_accuracy": 0.7110963732004165,
      "num_tokens": 6280073.0,
      "step": 3170
    },
    {
      "entropy": 1.2271991118788719,
      "epoch": 0.05712489333992006,
      "grad_norm": 7.3125,
      "learning_rate": 4.980425517339935e-05,
      "loss": 1.3134913444519043,
      "mean_token_accuracy": 0.7043784148991108,
      "num_tokens": 6299909.0,
      "step": 3180
    },
    {
      "entropy": 1.2218820795416832,
      "epoch": 0.05730453136929088,
      "grad_norm": 8.125,
      "learning_rate": 4.980245677778063e-05,
      "loss": 1.2430070877075194,
      "mean_token_accuracy": 0.7221039563417435,
      "num_tokens": 6319408.0,
      "step": 3190
    },
    {
      "entropy": 1.188573731482029,
      "epoch": 0.057484169398661696,
      "grad_norm": 9.1875,
      "learning_rate": 4.980065019132905e-05,
      "loss": 1.2444118499755858,
      "mean_token_accuracy": 0.7217368751764297,
      "num_tokens": 6339634.0,
      "step": 3200
    },
    {
      "entropy": 1.3362519547343255,
      "epoch": 0.05766380742803252,
      "grad_norm": 13.5625,
      "learning_rate": 4.979883541464124e-05,
      "loss": 1.3920759201049804,
      "mean_token_accuracy": 0.6976002052426338,
      "num_tokens": 6359387.0,
      "step": 3210
    },
    {
      "entropy": 1.1500658124685288,
      "epoch": 0.05784344545740333,
      "grad_norm": 6.6875,
      "learning_rate": 4.9797012448316505e-05,
      "loss": 1.1104830741882323,
      "mean_token_accuracy": 0.7497173845767975,
      "num_tokens": 6378775.0,
      "step": 3220
    },
    {
      "entropy": 1.206740863621235,
      "epoch": 0.05802308348677415,
      "grad_norm": 8.375,
      "learning_rate": 4.979518129295687e-05,
      "loss": 1.3307913780212401,
      "mean_token_accuracy": 0.7072183325886726,
      "num_tokens": 6399148.0,
      "step": 3230
    },
    {
      "entropy": 1.2591784119606018,
      "epoch": 0.058202721516144966,
      "grad_norm": 6.78125,
      "learning_rate": 4.9793341949167064e-05,
      "loss": 1.2734559059143067,
      "mean_token_accuracy": 0.7065724715590477,
      "num_tokens": 6419577.0,
      "step": 3240
    },
    {
      "entropy": 1.226381041109562,
      "epoch": 0.05838235954551579,
      "grad_norm": 7.21875,
      "learning_rate": 4.9791494417554517e-05,
      "loss": 1.2940117835998535,
      "mean_token_accuracy": 0.7089318469166755,
      "num_tokens": 6439454.0,
      "step": 3250
    },
    {
      "entropy": 1.2393014162778855,
      "epoch": 0.0585619975748866,
      "grad_norm": 8.0625,
      "learning_rate": 4.9789638698729356e-05,
      "loss": 1.29135160446167,
      "mean_token_accuracy": 0.7178616285324096,
      "num_tokens": 6458745.0,
      "step": 3260
    },
    {
      "entropy": 1.2197474554181098,
      "epoch": 0.05874163560425742,
      "grad_norm": 8.4375,
      "learning_rate": 4.978777479330442e-05,
      "loss": 1.2943720817565918,
      "mean_token_accuracy": 0.7084274925291538,
      "num_tokens": 6478206.0,
      "step": 3270
    },
    {
      "entropy": 1.2297852396965028,
      "epoch": 0.058921273633628236,
      "grad_norm": 7.65625,
      "learning_rate": 4.978590270189526e-05,
      "loss": 1.2721128463745117,
      "mean_token_accuracy": 0.7122110277414322,
      "num_tokens": 6498369.0,
      "step": 3280
    },
    {
      "entropy": 1.2309267163276671,
      "epoch": 0.05910091166299906,
      "grad_norm": 8.0625,
      "learning_rate": 4.978402242512011e-05,
      "loss": 1.3250860214233398,
      "mean_token_accuracy": 0.7046559602022171,
      "num_tokens": 6518459.0,
      "step": 3290
    },
    {
      "entropy": 1.268016543984413,
      "epoch": 0.05928054969236987,
      "grad_norm": 9.375,
      "learning_rate": 4.9782133963599916e-05,
      "loss": 1.3107852935791016,
      "mean_token_accuracy": 0.7096571244299412,
      "num_tokens": 6538537.0,
      "step": 3300
    },
    {
      "entropy": 1.2028040543198586,
      "epoch": 0.05946018772174069,
      "grad_norm": 8.375,
      "learning_rate": 4.978023731795834e-05,
      "loss": 1.2983840942382812,
      "mean_token_accuracy": 0.7168523222208023,
      "num_tokens": 6557820.0,
      "step": 3310
    },
    {
      "entropy": 1.3016474202275277,
      "epoch": 0.059639825751111514,
      "grad_norm": 9.125,
      "learning_rate": 4.977833248882173e-05,
      "loss": 1.307421875,
      "mean_token_accuracy": 0.7036651030182839,
      "num_tokens": 6578908.0,
      "step": 3320
    },
    {
      "entropy": 1.3456607446074487,
      "epoch": 0.05981946378048233,
      "grad_norm": 7.8125,
      "learning_rate": 4.9776419476819144e-05,
      "loss": 1.3548008918762207,
      "mean_token_accuracy": 0.7018690846860409,
      "num_tokens": 6597802.0,
      "step": 3330
    },
    {
      "entropy": 1.1719222888350487,
      "epoch": 0.05999910180985315,
      "grad_norm": 11.6875,
      "learning_rate": 4.9774498282582345e-05,
      "loss": 1.276802921295166,
      "mean_token_accuracy": 0.7090975195169449,
      "num_tokens": 6617697.0,
      "step": 3340
    },
    {
      "entropy": 1.2620284423232078,
      "epoch": 0.06017873983922396,
      "grad_norm": 5.71875,
      "learning_rate": 4.977256890674577e-05,
      "loss": 1.3803043365478516,
      "mean_token_accuracy": 0.7027500331401825,
      "num_tokens": 6636930.0,
      "step": 3350
    },
    {
      "entropy": 1.2817351326346398,
      "epoch": 0.060358377868594784,
      "grad_norm": 7.625,
      "learning_rate": 4.977063134994662e-05,
      "loss": 1.279933261871338,
      "mean_token_accuracy": 0.7034377135336399,
      "num_tokens": 6656377.0,
      "step": 3360
    },
    {
      "entropy": 1.167387916147709,
      "epoch": 0.0605380158979656,
      "grad_norm": 5.625,
      "learning_rate": 4.976868561282472e-05,
      "loss": 1.1824708938598634,
      "mean_token_accuracy": 0.7330655008554459,
      "num_tokens": 6677205.0,
      "step": 3370
    },
    {
      "entropy": 1.1286020129919052,
      "epoch": 0.06071765392733642,
      "grad_norm": 9.0,
      "learning_rate": 4.9766731696022664e-05,
      "loss": 1.232715129852295,
      "mean_token_accuracy": 0.7248808458447457,
      "num_tokens": 6698105.0,
      "step": 3380
    },
    {
      "entropy": 1.321954208612442,
      "epoch": 0.060897291956707233,
      "grad_norm": 8.5625,
      "learning_rate": 4.97647696001857e-05,
      "loss": 1.4307735443115235,
      "mean_token_accuracy": 0.69056516289711,
      "num_tokens": 6717826.0,
      "step": 3390
    },
    {
      "entropy": 1.2396196484565736,
      "epoch": 0.061076929986078055,
      "grad_norm": 8.25,
      "learning_rate": 4.976279932596182e-05,
      "loss": 1.1927993774414063,
      "mean_token_accuracy": 0.7270082898437977,
      "num_tokens": 6737263.0,
      "step": 3400
    },
    {
      "entropy": 1.1182163417339326,
      "epoch": 0.06125656801544887,
      "grad_norm": 6.34375,
      "learning_rate": 4.9760820874001664e-05,
      "loss": 1.1719481468200683,
      "mean_token_accuracy": 0.7272097051143647,
      "num_tokens": 6757784.0,
      "step": 3410
    },
    {
      "entropy": 1.2036172717809677,
      "epoch": 0.06143620604481969,
      "grad_norm": 6.34375,
      "learning_rate": 4.9758834244958616e-05,
      "loss": 1.2382091522216796,
      "mean_token_accuracy": 0.7252969890832901,
      "num_tokens": 6776728.0,
      "step": 3420
    },
    {
      "entropy": 1.2807849138975143,
      "epoch": 0.061615844074190504,
      "grad_norm": 5.25,
      "learning_rate": 4.975683943948875e-05,
      "loss": 1.352182388305664,
      "mean_token_accuracy": 0.7102423667907715,
      "num_tokens": 6796588.0,
      "step": 3430
    },
    {
      "entropy": 1.2119808204472065,
      "epoch": 0.061795482103561325,
      "grad_norm": 9.5625,
      "learning_rate": 4.975483645825083e-05,
      "loss": 1.2675718307495116,
      "mean_token_accuracy": 0.7183391459286212,
      "num_tokens": 6815557.0,
      "step": 3440
    },
    {
      "entropy": 1.2187032014131547,
      "epoch": 0.06197512013293214,
      "grad_norm": 12.0625,
      "learning_rate": 4.975282530190632e-05,
      "loss": 1.309666347503662,
      "mean_token_accuracy": 0.7133752577006817,
      "num_tokens": 6836335.0,
      "step": 3450
    },
    {
      "entropy": 1.2438659489154815,
      "epoch": 0.06215475816230296,
      "grad_norm": 7.9375,
      "learning_rate": 4.975080597111941e-05,
      "loss": 1.2932543754577637,
      "mean_token_accuracy": 0.7090259060263634,
      "num_tokens": 6856973.0,
      "step": 3460
    },
    {
      "entropy": 1.2497733801603317,
      "epoch": 0.062334396191673774,
      "grad_norm": 9.25,
      "learning_rate": 4.974877846655694e-05,
      "loss": 1.2450849533081054,
      "mean_token_accuracy": 0.7180824309587479,
      "num_tokens": 6876477.0,
      "step": 3470
    },
    {
      "entropy": 1.1382774874567985,
      "epoch": 0.0625140342210446,
      "grad_norm": 11.9375,
      "learning_rate": 4.974674278888851e-05,
      "loss": 1.2133026123046875,
      "mean_token_accuracy": 0.7236194267868996,
      "num_tokens": 6896275.0,
      "step": 3480
    },
    {
      "entropy": 1.235068529844284,
      "epoch": 0.06269367225041542,
      "grad_norm": 7.46875,
      "learning_rate": 4.974469893878637e-05,
      "loss": 1.3045209884643554,
      "mean_token_accuracy": 0.7078215204179287,
      "num_tokens": 6915823.0,
      "step": 3490
    },
    {
      "entropy": 1.2026451602578163,
      "epoch": 0.06287331027978624,
      "grad_norm": 8.5,
      "learning_rate": 4.9742646916925484e-05,
      "loss": 1.2315085411071778,
      "mean_token_accuracy": 0.7220328912138939,
      "num_tokens": 6934986.0,
      "step": 3500
    },
    {
      "entropy": 1.2231766253709793,
      "epoch": 0.06305294830915704,
      "grad_norm": 9.125,
      "learning_rate": 4.974058672398353e-05,
      "loss": 1.2991942405700683,
      "mean_token_accuracy": 0.7067679658532142,
      "num_tokens": 6954226.0,
      "step": 3510
    },
    {
      "entropy": 1.2861617535352707,
      "epoch": 0.06323258633852787,
      "grad_norm": 9.3125,
      "learning_rate": 4.973851836064086e-05,
      "loss": 1.332446002960205,
      "mean_token_accuracy": 0.6976198434829712,
      "num_tokens": 6972985.0,
      "step": 3520
    },
    {
      "entropy": 1.234875889122486,
      "epoch": 0.06341222436789869,
      "grad_norm": 6.1875,
      "learning_rate": 4.973644182758055e-05,
      "loss": 1.226964282989502,
      "mean_token_accuracy": 0.7273961171507836,
      "num_tokens": 6991688.0,
      "step": 3530
    },
    {
      "entropy": 1.223322905600071,
      "epoch": 0.06359186239726951,
      "grad_norm": 4.625,
      "learning_rate": 4.973435712548834e-05,
      "loss": 1.2936535835266114,
      "mean_token_accuracy": 0.7073257721960544,
      "num_tokens": 7012670.0,
      "step": 3540
    },
    {
      "entropy": 1.2125786438584327,
      "epoch": 0.06377150042664032,
      "grad_norm": 8.4375,
      "learning_rate": 4.9732264255052706e-05,
      "loss": 1.3368864059448242,
      "mean_token_accuracy": 0.7069816030561924,
      "num_tokens": 7031831.0,
      "step": 3550
    },
    {
      "entropy": 1.2602311804890634,
      "epoch": 0.06395113845601114,
      "grad_norm": 7.40625,
      "learning_rate": 4.97301632169648e-05,
      "loss": 1.3025297164916991,
      "mean_token_accuracy": 0.7098409436643124,
      "num_tokens": 7052248.0,
      "step": 3560
    },
    {
      "entropy": 1.364729504287243,
      "epoch": 0.06413077648538196,
      "grad_norm": 9.5,
      "learning_rate": 4.972805401191847e-05,
      "loss": 1.4347960472106933,
      "mean_token_accuracy": 0.688126140832901,
      "num_tokens": 7071284.0,
      "step": 3570
    },
    {
      "entropy": 1.1663824573159218,
      "epoch": 0.06431041451475278,
      "grad_norm": 7.3125,
      "learning_rate": 4.972593664061027e-05,
      "loss": 1.2339487075805664,
      "mean_token_accuracy": 0.7286138713359833,
      "num_tokens": 7090273.0,
      "step": 3580
    },
    {
      "entropy": 1.2514225229620934,
      "epoch": 0.06449005254412359,
      "grad_norm": 8.1875,
      "learning_rate": 4.9723811103739435e-05,
      "loss": 1.3342815399169923,
      "mean_token_accuracy": 0.7041991069912911,
      "num_tokens": 7110543.0,
      "step": 3590
    },
    {
      "entropy": 1.1802908971905708,
      "epoch": 0.0646696905734944,
      "grad_norm": 8.0,
      "learning_rate": 4.972167740200793e-05,
      "loss": 1.1685798645019532,
      "mean_token_accuracy": 0.7224047504365444,
      "num_tokens": 7131347.0,
      "step": 3600
    },
    {
      "entropy": 1.2886706575751306,
      "epoch": 0.06484932860286523,
      "grad_norm": 8.0625,
      "learning_rate": 4.971953553612036e-05,
      "loss": 1.4090861320495605,
      "mean_token_accuracy": 0.7011920288205147,
      "num_tokens": 7151376.0,
      "step": 3610
    },
    {
      "entropy": 1.347020211815834,
      "epoch": 0.06502896663223605,
      "grad_norm": 11.5625,
      "learning_rate": 4.971738550678411e-05,
      "loss": 1.3731415748596192,
      "mean_token_accuracy": 0.6966219522058964,
      "num_tokens": 7171119.0,
      "step": 3620
    },
    {
      "entropy": 1.2852986738085748,
      "epoch": 0.06520860466160686,
      "grad_norm": 9.1875,
      "learning_rate": 4.9715227314709164e-05,
      "loss": 1.2775721549987793,
      "mean_token_accuracy": 0.7031201519072056,
      "num_tokens": 7190493.0,
      "step": 3630
    },
    {
      "entropy": 1.2208514109253883,
      "epoch": 0.06538824269097768,
      "grad_norm": 8.5625,
      "learning_rate": 4.971306096060828e-05,
      "loss": 1.302250576019287,
      "mean_token_accuracy": 0.7101312294602394,
      "num_tokens": 7209508.0,
      "step": 3640
    },
    {
      "entropy": 1.245539802312851,
      "epoch": 0.0655678807203485,
      "grad_norm": 7.875,
      "learning_rate": 4.971088644519687e-05,
      "loss": 1.3021961212158204,
      "mean_token_accuracy": 0.7031107805669308,
      "num_tokens": 7229810.0,
      "step": 3650
    },
    {
      "entropy": 1.1274859696626662,
      "epoch": 0.06574751874971932,
      "grad_norm": 6.28125,
      "learning_rate": 4.970870376919305e-05,
      "loss": 1.1816731452941895,
      "mean_token_accuracy": 0.7423084408044816,
      "num_tokens": 7250518.0,
      "step": 3660
    },
    {
      "entropy": 1.0779212534427642,
      "epoch": 0.06592715677909014,
      "grad_norm": 6.125,
      "learning_rate": 4.970651293331764e-05,
      "loss": 1.1065179824829101,
      "mean_token_accuracy": 0.7425893932580948,
      "num_tokens": 7270177.0,
      "step": 3670
    },
    {
      "entropy": 1.1414249613881111,
      "epoch": 0.06610679480846095,
      "grad_norm": 5.96875,
      "learning_rate": 4.9704313938294134e-05,
      "loss": 1.2218378067016602,
      "mean_token_accuracy": 0.731036901473999,
      "num_tokens": 7289893.0,
      "step": 3680
    },
    {
      "entropy": 1.1677282959222794,
      "epoch": 0.06628643283783177,
      "grad_norm": 6.53125,
      "learning_rate": 4.970210678484876e-05,
      "loss": 1.1923569679260253,
      "mean_token_accuracy": 0.7306131199002266,
      "num_tokens": 7310224.0,
      "step": 3690
    },
    {
      "entropy": 1.1779899552464486,
      "epoch": 0.06646607086720259,
      "grad_norm": 7.1875,
      "learning_rate": 4.969989147371038e-05,
      "loss": 1.2312817573547363,
      "mean_token_accuracy": 0.7168428733944893,
      "num_tokens": 7330845.0,
      "step": 3700
    },
    {
      "entropy": 1.1872374355792998,
      "epoch": 0.06664570889657341,
      "grad_norm": 6.25,
      "learning_rate": 4.9697668005610613e-05,
      "loss": 1.2032230377197266,
      "mean_token_accuracy": 0.7239424347877502,
      "num_tokens": 7350255.0,
      "step": 3710
    },
    {
      "entropy": 1.2174224704504013,
      "epoch": 0.06682534692594422,
      "grad_norm": 6.25,
      "learning_rate": 4.969543638128373e-05,
      "loss": 1.3171369552612304,
      "mean_token_accuracy": 0.7075670704245567,
      "num_tokens": 7369666.0,
      "step": 3720
    },
    {
      "entropy": 1.2062487781047821,
      "epoch": 0.06700498495531504,
      "grad_norm": 8.0625,
      "learning_rate": 4.969319660146672e-05,
      "loss": 1.2684994697570802,
      "mean_token_accuracy": 0.7242234297096729,
      "num_tokens": 7389836.0,
      "step": 3730
    },
    {
      "entropy": 1.3156227365136146,
      "epoch": 0.06718462298468586,
      "grad_norm": 8.125,
      "learning_rate": 4.969094866689924e-05,
      "loss": 1.3338292121887207,
      "mean_token_accuracy": 0.7075921133160591,
      "num_tokens": 7409549.0,
      "step": 3740
    },
    {
      "entropy": 1.2621764317154884,
      "epoch": 0.06736426101405668,
      "grad_norm": 7.15625,
      "learning_rate": 4.968869257832366e-05,
      "loss": 1.3384495735168458,
      "mean_token_accuracy": 0.6944715850055218,
      "num_tokens": 7429142.0,
      "step": 3750
    },
    {
      "entropy": 1.2328008137643338,
      "epoch": 0.06754389904342749,
      "grad_norm": 9.4375,
      "learning_rate": 4.9686428336485033e-05,
      "loss": 1.3114291191101075,
      "mean_token_accuracy": 0.7137681439518928,
      "num_tokens": 7447923.0,
      "step": 3760
    },
    {
      "entropy": 1.221186363697052,
      "epoch": 0.06772353707279831,
      "grad_norm": 5.96875,
      "learning_rate": 4.968415594213113e-05,
      "loss": 1.2468070983886719,
      "mean_token_accuracy": 0.7111487925052643,
      "num_tokens": 7466912.0,
      "step": 3770
    },
    {
      "entropy": 1.2326500102877618,
      "epoch": 0.06790317510216913,
      "grad_norm": 7.09375,
      "learning_rate": 4.968187539601236e-05,
      "loss": 1.2358693122863769,
      "mean_token_accuracy": 0.7201651819050312,
      "num_tokens": 7486861.0,
      "step": 3780
    },
    {
      "entropy": 1.1974031433463097,
      "epoch": 0.06808281313153995,
      "grad_norm": 9.9375,
      "learning_rate": 4.967958669888188e-05,
      "loss": 1.281076431274414,
      "mean_token_accuracy": 0.7184802263975143,
      "num_tokens": 7506247.0,
      "step": 3790
    },
    {
      "entropy": 1.2204970180988313,
      "epoch": 0.06826245116091076,
      "grad_norm": 6.90625,
      "learning_rate": 4.96772898514955e-05,
      "loss": 1.285616970062256,
      "mean_token_accuracy": 0.7070894479751587,
      "num_tokens": 7527277.0,
      "step": 3800
    },
    {
      "entropy": 1.2314346462488175,
      "epoch": 0.06844208919028158,
      "grad_norm": 8.5625,
      "learning_rate": 4.967498485461176e-05,
      "loss": 1.2882765769958495,
      "mean_token_accuracy": 0.7152645960450172,
      "num_tokens": 7547363.0,
      "step": 3810
    },
    {
      "entropy": 1.2864285662770272,
      "epoch": 0.0686217272196524,
      "grad_norm": 5.53125,
      "learning_rate": 4.967267170899185e-05,
      "loss": 1.3151036262512208,
      "mean_token_accuracy": 0.6949542030692101,
      "num_tokens": 7568055.0,
      "step": 3820
    },
    {
      "entropy": 1.2726764366030694,
      "epoch": 0.06880136524902322,
      "grad_norm": 7.53125,
      "learning_rate": 4.967035041539967e-05,
      "loss": 1.303805446624756,
      "mean_token_accuracy": 0.7162467837333679,
      "num_tokens": 7589470.0,
      "step": 3830
    },
    {
      "entropy": 1.2356864973902701,
      "epoch": 0.06898100327839403,
      "grad_norm": 6.9375,
      "learning_rate": 4.9668020974601816e-05,
      "loss": 1.3131505966186523,
      "mean_token_accuracy": 0.7027673237025738,
      "num_tokens": 7608654.0,
      "step": 3840
    },
    {
      "entropy": 1.159817098081112,
      "epoch": 0.06916064130776485,
      "grad_norm": 6.25,
      "learning_rate": 4.966568338736757e-05,
      "loss": 1.1256381034851075,
      "mean_token_accuracy": 0.7296169057488442,
      "num_tokens": 7628615.0,
      "step": 3850
    },
    {
      "entropy": 1.3059486582875253,
      "epoch": 0.06934027933713567,
      "grad_norm": 7.1875,
      "learning_rate": 4.96633376544689e-05,
      "loss": 1.4329897880554199,
      "mean_token_accuracy": 0.6818833567202092,
      "num_tokens": 7648594.0,
      "step": 3860
    },
    {
      "entropy": 1.1422328650951385,
      "epoch": 0.06951991736650649,
      "grad_norm": 7.59375,
      "learning_rate": 4.966098377668047e-05,
      "loss": 1.181243896484375,
      "mean_token_accuracy": 0.7235778883099556,
      "num_tokens": 7668915.0,
      "step": 3870
    },
    {
      "entropy": 1.1333101496100426,
      "epoch": 0.06969955539587731,
      "grad_norm": 6.09375,
      "learning_rate": 4.9658621754779624e-05,
      "loss": 1.2051117897033692,
      "mean_token_accuracy": 0.7212189637124539,
      "num_tokens": 7688621.0,
      "step": 3880
    },
    {
      "entropy": 1.1232207968831063,
      "epoch": 0.06987919342524812,
      "grad_norm": 9.1875,
      "learning_rate": 4.965625158954642e-05,
      "loss": 1.1678098678588866,
      "mean_token_accuracy": 0.7345072537660599,
      "num_tokens": 7708625.0,
      "step": 3890
    },
    {
      "entropy": 1.1640655010938645,
      "epoch": 0.07005883145461894,
      "grad_norm": 7.78125,
      "learning_rate": 4.965387328176356e-05,
      "loss": 1.2449628829956054,
      "mean_token_accuracy": 0.725748098641634,
      "num_tokens": 7728005.0,
      "step": 3900
    },
    {
      "entropy": 1.3203783065080643,
      "epoch": 0.07023846948398976,
      "grad_norm": 6.75,
      "learning_rate": 4.96514868322165e-05,
      "loss": 1.3069683074951173,
      "mean_token_accuracy": 0.699828465282917,
      "num_tokens": 7748372.0,
      "step": 3910
    },
    {
      "entropy": 1.167919297516346,
      "epoch": 0.07041810751336058,
      "grad_norm": 5.84375,
      "learning_rate": 4.964909224169331e-05,
      "loss": 1.24308443069458,
      "mean_token_accuracy": 0.7124692603945733,
      "num_tokens": 7769053.0,
      "step": 3920
    },
    {
      "entropy": 1.2315762400627137,
      "epoch": 0.07059774554273139,
      "grad_norm": 8.3125,
      "learning_rate": 4.964668951098481e-05,
      "loss": 1.307978343963623,
      "mean_token_accuracy": 0.7066112831234932,
      "num_tokens": 7788054.0,
      "step": 3930
    },
    {
      "entropy": 1.1790028259158134,
      "epoch": 0.07077738357210221,
      "grad_norm": 7.09375,
      "learning_rate": 4.9644278640884476e-05,
      "loss": 1.1816357612609862,
      "mean_token_accuracy": 0.7243860140442848,
      "num_tokens": 7808464.0,
      "step": 3940
    },
    {
      "entropy": 1.1662120059132577,
      "epoch": 0.07095702160147303,
      "grad_norm": 6.46875,
      "learning_rate": 4.964185963218849e-05,
      "loss": 1.2439204216003419,
      "mean_token_accuracy": 0.7191622465848923,
      "num_tokens": 7827407.0,
      "step": 3950
    },
    {
      "entropy": 1.1534938678145408,
      "epoch": 0.07113665963084385,
      "grad_norm": 6.65625,
      "learning_rate": 4.9639432485695704e-05,
      "loss": 1.1548154830932618,
      "mean_token_accuracy": 0.7311466187238693,
      "num_tokens": 7846810.0,
      "step": 3960
    },
    {
      "entropy": 1.1068582475185393,
      "epoch": 0.07131629766021466,
      "grad_norm": 6.34375,
      "learning_rate": 4.963699720220766e-05,
      "loss": 1.1102201461791992,
      "mean_token_accuracy": 0.742414303123951,
      "num_tokens": 7866375.0,
      "step": 3970
    },
    {
      "entropy": 1.0965869441628455,
      "epoch": 0.07149593568958548,
      "grad_norm": 9.625,
      "learning_rate": 4.96345537825286e-05,
      "loss": 1.2301941871643067,
      "mean_token_accuracy": 0.724865297973156,
      "num_tokens": 7885541.0,
      "step": 3980
    },
    {
      "entropy": 1.1597801387310027,
      "epoch": 0.0716755737189563,
      "grad_norm": 8.3125,
      "learning_rate": 4.9632102227465454e-05,
      "loss": 1.2550856590270996,
      "mean_token_accuracy": 0.7191265910863877,
      "num_tokens": 7905579.0,
      "step": 3990
    },
    {
      "epoch": 0.07185521174832712,
      "eval_entropy": 1.1816870107650757,
      "eval_loss": 1.1931698322296143,
      "eval_mean_token_accuracy": 0.7292079088687897,
      "eval_num_tokens": 7926226.0,
      "eval_runtime": 40.4952,
      "eval_samples_per_second": 49.389,
      "eval_steps_per_second": 6.174,
      "step": 4000
    },
    {
      "entropy": 1.1713870141655207,
      "epoch": 0.07203484977769793,
      "grad_norm": 7.40625,
      "learning_rate": 4.962717471442799e-05,
      "loss": 1.1969499588012695,
      "mean_token_accuracy": 0.7299913689494133,
      "num_tokens": 7946346.0,
      "step": 4010
    },
    {
      "entropy": 1.1797849789261818,
      "epoch": 0.07221448780706875,
      "grad_norm": 7.1875,
      "learning_rate": 4.962469875808094e-05,
      "loss": 1.25166015625,
      "mean_token_accuracy": 0.7197384297847748,
      "num_tokens": 7966372.0,
      "step": 4020
    },
    {
      "entropy": 1.242465315759182,
      "epoch": 0.07239412583643957,
      "grad_norm": 7.5625,
      "learning_rate": 4.962221466960436e-05,
      "loss": 1.2804611206054688,
      "mean_token_accuracy": 0.7203630454838276,
      "num_tokens": 7985084.0,
      "step": 4030
    },
    {
      "entropy": 1.208653238415718,
      "epoch": 0.0725737638658104,
      "grad_norm": 9.6875,
      "learning_rate": 4.961972244981859e-05,
      "loss": 1.2687210083007812,
      "mean_token_accuracy": 0.7167617976665497,
      "num_tokens": 8004777.0,
      "step": 4040
    },
    {
      "entropy": 1.2139231905341148,
      "epoch": 0.07275340189518122,
      "grad_norm": 6.9375,
      "learning_rate": 4.9617222099546656e-05,
      "loss": 1.257332992553711,
      "mean_token_accuracy": 0.7208951704204083,
      "num_tokens": 8024146.0,
      "step": 4050
    },
    {
      "entropy": 1.152854146808386,
      "epoch": 0.07293303992455202,
      "grad_norm": 6.6875,
      "learning_rate": 4.96147136196143e-05,
      "loss": 1.215870475769043,
      "mean_token_accuracy": 0.717916489392519,
      "num_tokens": 8044151.0,
      "step": 4060
    },
    {
      "entropy": 1.074496054649353,
      "epoch": 0.07311267795392284,
      "grad_norm": 7.34375,
      "learning_rate": 4.9612197010849914e-05,
      "loss": 1.1245906829833985,
      "mean_token_accuracy": 0.7415511220693588,
      "num_tokens": 8063914.0,
      "step": 4070
    },
    {
      "entropy": 1.1561739817261696,
      "epoch": 0.07329231598329367,
      "grad_norm": 7.8125,
      "learning_rate": 4.96096722740846e-05,
      "loss": 1.2119399070739747,
      "mean_token_accuracy": 0.7264811292290687,
      "num_tokens": 8083712.0,
      "step": 4080
    },
    {
      "entropy": 1.1322413355112075,
      "epoch": 0.07347195401266449,
      "grad_norm": 7.1875,
      "learning_rate": 4.9607139410152125e-05,
      "loss": 1.1491485595703126,
      "mean_token_accuracy": 0.7246209889650345,
      "num_tokens": 8104099.0,
      "step": 4090
    },
    {
      "entropy": 1.1210455924272538,
      "epoch": 0.0736515920420353,
      "grad_norm": 8.8125,
      "learning_rate": 4.960459841988896e-05,
      "loss": 1.1640192985534668,
      "mean_token_accuracy": 0.7334762841463089,
      "num_tokens": 8124414.0,
      "step": 4100
    },
    {
      "entropy": 1.142209130525589,
      "epoch": 0.07383123007140611,
      "grad_norm": 9.375,
      "learning_rate": 4.9602049304134245e-05,
      "loss": 1.2349956512451172,
      "mean_token_accuracy": 0.7196511864662171,
      "num_tokens": 8144399.0,
      "step": 4110
    },
    {
      "entropy": 1.2242891952395438,
      "epoch": 0.07401086810077694,
      "grad_norm": 8.6875,
      "learning_rate": 4.95994920637298e-05,
      "loss": 1.2499198913574219,
      "mean_token_accuracy": 0.7192654147744179,
      "num_tokens": 8164376.0,
      "step": 4120
    },
    {
      "entropy": 1.218564397096634,
      "epoch": 0.07419050613014776,
      "grad_norm": 6.1875,
      "learning_rate": 4.9596926699520144e-05,
      "loss": 1.2232787132263183,
      "mean_token_accuracy": 0.7258368238806725,
      "num_tokens": 8184048.0,
      "step": 4130
    },
    {
      "entropy": 1.1712930262088777,
      "epoch": 0.07437014415951856,
      "grad_norm": 7.4375,
      "learning_rate": 4.959435321235246e-05,
      "loss": 1.2336976051330566,
      "mean_token_accuracy": 0.7239293172955513,
      "num_tokens": 8204084.0,
      "step": 4140
    },
    {
      "entropy": 1.2275750920176507,
      "epoch": 0.07454978218888939,
      "grad_norm": 8.6875,
      "learning_rate": 4.959177160307664e-05,
      "loss": 1.2959970474243163,
      "mean_token_accuracy": 0.7193170458078384,
      "num_tokens": 8223830.0,
      "step": 4150
    },
    {
      "entropy": 1.185786284506321,
      "epoch": 0.0747294202182602,
      "grad_norm": 8.375,
      "learning_rate": 4.9589181872545225e-05,
      "loss": 1.1565098762512207,
      "mean_token_accuracy": 0.7324448108673096,
      "num_tokens": 8243935.0,
      "step": 4160
    },
    {
      "entropy": 1.1236275151371955,
      "epoch": 0.07490905824763103,
      "grad_norm": 8.5,
      "learning_rate": 4.958658402161347e-05,
      "loss": 1.1814092636108398,
      "mean_token_accuracy": 0.7348843105137348,
      "num_tokens": 8263559.0,
      "step": 4170
    },
    {
      "entropy": 1.0817065611481667,
      "epoch": 0.07508869627700183,
      "grad_norm": 8.25,
      "learning_rate": 4.958397805113928e-05,
      "loss": 1.1719980239868164,
      "mean_token_accuracy": 0.737169261276722,
      "num_tokens": 8282921.0,
      "step": 4180
    },
    {
      "entropy": 1.2425470396876335,
      "epoch": 0.07526833430637266,
      "grad_norm": 6.28125,
      "learning_rate": 4.9581363961983265e-05,
      "loss": 1.3042529106140137,
      "mean_token_accuracy": 0.7026644073426723,
      "num_tokens": 8303879.0,
      "step": 4190
    },
    {
      "entropy": 1.2776909247040749,
      "epoch": 0.07544797233574348,
      "grad_norm": 7.1875,
      "learning_rate": 4.957874175500872e-05,
      "loss": 1.2881336212158203,
      "mean_token_accuracy": 0.716185811161995,
      "num_tokens": 8324551.0,
      "step": 4200
    },
    {
      "entropy": 1.1333178557455539,
      "epoch": 0.0756276103651143,
      "grad_norm": 7.875,
      "learning_rate": 4.9576111431081594e-05,
      "loss": 1.1647315979003907,
      "mean_token_accuracy": 0.7394245445728302,
      "num_tokens": 8343072.0,
      "step": 4210
    },
    {
      "entropy": 0.9866166695952415,
      "epoch": 0.07580724839448512,
      "grad_norm": 6.375,
      "learning_rate": 4.957347299107054e-05,
      "loss": 1.0758567810058595,
      "mean_token_accuracy": 0.7557776510715485,
      "num_tokens": 8362647.0,
      "step": 4220
    },
    {
      "entropy": 1.082628497481346,
      "epoch": 0.07598688642385593,
      "grad_norm": 7.03125,
      "learning_rate": 4.957082643584688e-05,
      "loss": 1.1338550567626953,
      "mean_token_accuracy": 0.7338819041848182,
      "num_tokens": 8382086.0,
      "step": 4230
    },
    {
      "entropy": 1.0527696318924427,
      "epoch": 0.07616652445322675,
      "grad_norm": 7.90625,
      "learning_rate": 4.956817176628463e-05,
      "loss": 1.0984151840209961,
      "mean_token_accuracy": 0.7494589716196061,
      "num_tokens": 8402690.0,
      "step": 4240
    },
    {
      "entropy": 1.185495026409626,
      "epoch": 0.07634616248259757,
      "grad_norm": 6.8125,
      "learning_rate": 4.9565508983260464e-05,
      "loss": 1.2181382179260254,
      "mean_token_accuracy": 0.7362824261188508,
      "num_tokens": 8423353.0,
      "step": 4250
    },
    {
      "entropy": 1.2501841343939304,
      "epoch": 0.07652580051196839,
      "grad_norm": 8.125,
      "learning_rate": 4.956283808765376e-05,
      "loss": 1.2641603469848632,
      "mean_token_accuracy": 0.7112665444612503,
      "num_tokens": 8443211.0,
      "step": 4260
    },
    {
      "entropy": 1.161339780688286,
      "epoch": 0.0767054385413392,
      "grad_norm": 10.4375,
      "learning_rate": 4.9560159080346545e-05,
      "loss": 1.232271385192871,
      "mean_token_accuracy": 0.7298238225281238,
      "num_tokens": 8462398.0,
      "step": 4270
    },
    {
      "entropy": 0.9935755603015423,
      "epoch": 0.07688507657071002,
      "grad_norm": 7.625,
      "learning_rate": 4.955747196222355e-05,
      "loss": 1.028653335571289,
      "mean_token_accuracy": 0.7640767946839333,
      "num_tokens": 8482730.0,
      "step": 4280
    },
    {
      "entropy": 1.0362499043345452,
      "epoch": 0.07706471460008084,
      "grad_norm": 7.375,
      "learning_rate": 4.955477673417218e-05,
      "loss": 1.1323720932006835,
      "mean_token_accuracy": 0.7439577117562294,
      "num_tokens": 8503077.0,
      "step": 4290
    },
    {
      "entropy": 1.1890631526708604,
      "epoch": 0.07724435262945166,
      "grad_norm": 6.15625,
      "learning_rate": 4.955207339708251e-05,
      "loss": 1.182280921936035,
      "mean_token_accuracy": 0.7326760351657867,
      "num_tokens": 8522290.0,
      "step": 4300
    },
    {
      "entropy": 1.1386354684829711,
      "epoch": 0.07742399065882247,
      "grad_norm": 6.375,
      "learning_rate": 4.9549361951847295e-05,
      "loss": 1.1753834724426269,
      "mean_token_accuracy": 0.7320934399962425,
      "num_tokens": 8542708.0,
      "step": 4310
    },
    {
      "entropy": 1.1046071857213975,
      "epoch": 0.07760362868819329,
      "grad_norm": 8.5,
      "learning_rate": 4.954664239936197e-05,
      "loss": 1.224449920654297,
      "mean_token_accuracy": 0.7294555336236954,
      "num_tokens": 8561394.0,
      "step": 4320
    },
    {
      "entropy": 1.1356689035892487,
      "epoch": 0.07778326671756411,
      "grad_norm": 7.59375,
      "learning_rate": 4.954391474052466e-05,
      "loss": 1.154538059234619,
      "mean_token_accuracy": 0.735795333981514,
      "num_tokens": 8581730.0,
      "step": 4330
    },
    {
      "entropy": 1.082607415318489,
      "epoch": 0.07796290474693493,
      "grad_norm": 6.875,
      "learning_rate": 4.954117897623616e-05,
      "loss": 1.0830055236816407,
      "mean_token_accuracy": 0.7537325888872146,
      "num_tokens": 8601341.0,
      "step": 4340
    },
    {
      "entropy": 1.1453176468610764,
      "epoch": 0.07814254277630574,
      "grad_norm": 6.71875,
      "learning_rate": 4.9538435107399904e-05,
      "loss": 1.2479177474975587,
      "mean_token_accuracy": 0.7211047016084194,
      "num_tokens": 8622368.0,
      "step": 4350
    },
    {
      "entropy": 1.1174189552664757,
      "epoch": 0.07832218080567656,
      "grad_norm": 7.9375,
      "learning_rate": 4.953568313492206e-05,
      "loss": 1.2095426559448241,
      "mean_token_accuracy": 0.7209716007113457,
      "num_tokens": 8641573.0,
      "step": 4360
    },
    {
      "entropy": 1.219351901113987,
      "epoch": 0.07850181883504738,
      "grad_norm": 7.5,
      "learning_rate": 4.9532923059711436e-05,
      "loss": 1.2631852149963378,
      "mean_token_accuracy": 0.7296150237321853,
      "num_tokens": 8663117.0,
      "step": 4370
    },
    {
      "entropy": 1.2396316826343536,
      "epoch": 0.0786814568644182,
      "grad_norm": 7.9375,
      "learning_rate": 4.953015488267954e-05,
      "loss": 1.2260595321655274,
      "mean_token_accuracy": 0.7207832835614681,
      "num_tokens": 8681971.0,
      "step": 4380
    },
    {
      "entropy": 1.1308637797832488,
      "epoch": 0.07886109489378902,
      "grad_norm": 7.125,
      "learning_rate": 4.9527378604740524e-05,
      "loss": 1.1758810043334962,
      "mean_token_accuracy": 0.7243160039186478,
      "num_tokens": 8701270.0,
      "step": 4390
    },
    {
      "entropy": 1.117917563021183,
      "epoch": 0.07904073292315983,
      "grad_norm": 8.375,
      "learning_rate": 4.952459422681126e-05,
      "loss": 1.1736659049987792,
      "mean_token_accuracy": 0.7362862914800644,
      "num_tokens": 8720977.0,
      "step": 4400
    },
    {
      "entropy": 1.1850198328495025,
      "epoch": 0.07922037095253065,
      "grad_norm": 8.0625,
      "learning_rate": 4.9521801749811244e-05,
      "loss": 1.229194736480713,
      "mean_token_accuracy": 0.7197527036070823,
      "num_tokens": 8741493.0,
      "step": 4410
    },
    {
      "entropy": 1.100655935704708,
      "epoch": 0.07940000898190147,
      "grad_norm": 8.3125,
      "learning_rate": 4.951900117466268e-05,
      "loss": 1.0920257568359375,
      "mean_token_accuracy": 0.7484180495142937,
      "num_tokens": 8761894.0,
      "step": 4420
    },
    {
      "entropy": 1.1196262806653976,
      "epoch": 0.07957964701127229,
      "grad_norm": 10.6875,
      "learning_rate": 4.951619250229044e-05,
      "loss": 1.2225539207458496,
      "mean_token_accuracy": 0.728130965679884,
      "num_tokens": 8782035.0,
      "step": 4430
    },
    {
      "entropy": 1.1870262697339058,
      "epoch": 0.0797592850406431,
      "grad_norm": 6.625,
      "learning_rate": 4.9513375733622064e-05,
      "loss": 1.2134931564331055,
      "mean_token_accuracy": 0.7289870753884315,
      "num_tokens": 8803256.0,
      "step": 4440
    },
    {
      "entropy": 1.1689264476299286,
      "epoch": 0.07993892307001392,
      "grad_norm": 9.6875,
      "learning_rate": 4.951055086958777e-05,
      "loss": 1.2450201988220215,
      "mean_token_accuracy": 0.7246900506317615,
      "num_tokens": 8822743.0,
      "step": 4450
    },
    {
      "entropy": 1.1491047069430351,
      "epoch": 0.08011856109938474,
      "grad_norm": 5.75,
      "learning_rate": 4.9507717911120456e-05,
      "loss": 1.2423948287963866,
      "mean_token_accuracy": 0.7236890837550163,
      "num_tokens": 8842515.0,
      "step": 4460
    },
    {
      "entropy": 1.1791701532900334,
      "epoch": 0.08029819912875556,
      "grad_norm": 9.75,
      "learning_rate": 4.950487685915568e-05,
      "loss": 1.2120024681091308,
      "mean_token_accuracy": 0.7256231009960175,
      "num_tokens": 8863647.0,
      "step": 4470
    },
    {
      "entropy": 1.1515010014176368,
      "epoch": 0.08047783715812637,
      "grad_norm": 7.1875,
      "learning_rate": 4.9502027714631684e-05,
      "loss": 1.207651138305664,
      "mean_token_accuracy": 0.7286435097455979,
      "num_tokens": 8884436.0,
      "step": 4480
    },
    {
      "entropy": 1.1190988317131996,
      "epoch": 0.08065747518749719,
      "grad_norm": 8.3125,
      "learning_rate": 4.9499170478489366e-05,
      "loss": 1.1839657783508302,
      "mean_token_accuracy": 0.734744630753994,
      "num_tokens": 8904065.0,
      "step": 4490
    },
    {
      "entropy": 1.1805353432893753,
      "epoch": 0.08083711321686801,
      "grad_norm": 11.75,
      "learning_rate": 4.949630515167232e-05,
      "loss": 1.2314042091369628,
      "mean_token_accuracy": 0.7194116234779357,
      "num_tokens": 8923449.0,
      "step": 4500
    },
    {
      "entropy": 1.2770184025168418,
      "epoch": 0.08101675124623883,
      "grad_norm": 7.75,
      "learning_rate": 4.949343173512679e-05,
      "loss": 1.259283447265625,
      "mean_token_accuracy": 0.7055680677294731,
      "num_tokens": 8941762.0,
      "step": 4510
    },
    {
      "entropy": 1.1458518490195275,
      "epoch": 0.08119638927560964,
      "grad_norm": 7.78125,
      "learning_rate": 4.9490550229801706e-05,
      "loss": 1.2183851242065429,
      "mean_token_accuracy": 0.7194598913192749,
      "num_tokens": 8962146.0,
      "step": 4520
    },
    {
      "entropy": 1.113838893175125,
      "epoch": 0.08137602730498046,
      "grad_norm": 10.0,
      "learning_rate": 4.948766063664867e-05,
      "loss": 1.1766281127929688,
      "mean_token_accuracy": 0.7418471738696099,
      "num_tokens": 8981459.0,
      "step": 4530
    },
    {
      "entropy": 1.0895928040146827,
      "epoch": 0.08155566533435128,
      "grad_norm": 5.8125,
      "learning_rate": 4.948476295662194e-05,
      "loss": 1.0964507102966308,
      "mean_token_accuracy": 0.7489591509103775,
      "num_tokens": 9000662.0,
      "step": 4540
    },
    {
      "entropy": 1.0721483424305915,
      "epoch": 0.0817353033637221,
      "grad_norm": 6.46875,
      "learning_rate": 4.948185719067846e-05,
      "loss": 1.0654159545898438,
      "mean_token_accuracy": 0.7556398212909698,
      "num_tokens": 9020337.0,
      "step": 4550
    },
    {
      "entropy": 1.1024317428469659,
      "epoch": 0.08191494139309292,
      "grad_norm": 6.25,
      "learning_rate": 4.947894333977783e-05,
      "loss": 1.1473363876342773,
      "mean_token_accuracy": 0.7449666991829872,
      "num_tokens": 9041389.0,
      "step": 4560
    },
    {
      "entropy": 1.0792812690138818,
      "epoch": 0.08209457942246373,
      "grad_norm": 6.59375,
      "learning_rate": 4.9476021404882334e-05,
      "loss": 1.1406134605407714,
      "mean_token_accuracy": 0.7428496971726417,
      "num_tokens": 9060752.0,
      "step": 4570
    },
    {
      "entropy": 1.1383121401071548,
      "epoch": 0.08227421745183455,
      "grad_norm": 7.34375,
      "learning_rate": 4.947309138695693e-05,
      "loss": 1.1429553985595704,
      "mean_token_accuracy": 0.7356810122728348,
      "num_tokens": 9081057.0,
      "step": 4580
    },
    {
      "entropy": 1.0782451003789901,
      "epoch": 0.08245385548120537,
      "grad_norm": 6.40625,
      "learning_rate": 4.947015328696921e-05,
      "loss": 1.0886828422546386,
      "mean_token_accuracy": 0.7469633638858795,
      "num_tokens": 9100489.0,
      "step": 4590
    },
    {
      "entropy": 1.095804797858,
      "epoch": 0.0826334935105762,
      "grad_norm": 6.03125,
      "learning_rate": 4.946720710588949e-05,
      "loss": 1.1337652206420898,
      "mean_token_accuracy": 0.735855221748352,
      "num_tokens": 9121072.0,
      "step": 4600
    },
    {
      "entropy": 1.0897636383771896,
      "epoch": 0.082813131539947,
      "grad_norm": 5.6875,
      "learning_rate": 4.946425284469071e-05,
      "loss": 1.1113767623901367,
      "mean_token_accuracy": 0.7458392336964608,
      "num_tokens": 9141351.0,
      "step": 4610
    },
    {
      "entropy": 1.1274853184819222,
      "epoch": 0.08299276956931782,
      "grad_norm": 8.625,
      "learning_rate": 4.946129050434848e-05,
      "loss": 1.239295482635498,
      "mean_token_accuracy": 0.7173780806362628,
      "num_tokens": 9160784.0,
      "step": 4620
    },
    {
      "entropy": 1.009946422278881,
      "epoch": 0.08317240759868864,
      "grad_norm": 6.9375,
      "learning_rate": 4.945832008584112e-05,
      "loss": 1.0282076835632323,
      "mean_token_accuracy": 0.759869922697544,
      "num_tokens": 9179862.0,
      "step": 4630
    },
    {
      "entropy": 1.1466631941497325,
      "epoch": 0.08335204562805947,
      "grad_norm": 8.1875,
      "learning_rate": 4.945534159014957e-05,
      "loss": 1.2382200241088868,
      "mean_token_accuracy": 0.7125390745699406,
      "num_tokens": 9199210.0,
      "step": 4640
    },
    {
      "entropy": 1.0501185223460197,
      "epoch": 0.08353168365743027,
      "grad_norm": 6.9375,
      "learning_rate": 4.945235501825747e-05,
      "loss": 1.0690990447998048,
      "mean_token_accuracy": 0.7579691767692566,
      "num_tokens": 9218872.0,
      "step": 4650
    },
    {
      "entropy": 1.1878760367631913,
      "epoch": 0.0837113216868011,
      "grad_norm": 9.25,
      "learning_rate": 4.944936037115109e-05,
      "loss": 1.2630474090576171,
      "mean_token_accuracy": 0.7189512997865677,
      "num_tokens": 9238842.0,
      "step": 4660
    },
    {
      "entropy": 1.2201133072376251,
      "epoch": 0.08389095971617191,
      "grad_norm": 7.09375,
      "learning_rate": 4.944635764981942e-05,
      "loss": 1.2702902793884276,
      "mean_token_accuracy": 0.7169070929288864,
      "num_tokens": 9257476.0,
      "step": 4670
    },
    {
      "entropy": 1.0432930171489716,
      "epoch": 0.08407059774554274,
      "grad_norm": 7.84375,
      "learning_rate": 4.944334685525407e-05,
      "loss": 1.0921135902404786,
      "mean_token_accuracy": 0.7399536401033402,
      "num_tokens": 9277743.0,
      "step": 4680
    },
    {
      "entropy": 1.078136332333088,
      "epoch": 0.08425023577491354,
      "grad_norm": 5.75,
      "learning_rate": 4.944032798844933e-05,
      "loss": 1.0558756828308105,
      "mean_token_accuracy": 0.7448537409305572,
      "num_tokens": 9297397.0,
      "step": 4690
    },
    {
      "entropy": 1.0190674111247062,
      "epoch": 0.08442987380428436,
      "grad_norm": 7.78125,
      "learning_rate": 4.943730105040217e-05,
      "loss": 1.1207489013671874,
      "mean_token_accuracy": 0.735479012131691,
      "num_tokens": 9317151.0,
      "step": 4700
    },
    {
      "entropy": 1.1430235594511031,
      "epoch": 0.08460951183365518,
      "grad_norm": 6.71875,
      "learning_rate": 4.943426604211222e-05,
      "loss": 1.1972759246826172,
      "mean_token_accuracy": 0.7309389486908913,
      "num_tokens": 9336304.0,
      "step": 4710
    },
    {
      "entropy": 1.2262649565935135,
      "epoch": 0.084789149863026,
      "grad_norm": 8.0,
      "learning_rate": 4.9431222964581756e-05,
      "loss": 1.2485018730163575,
      "mean_token_accuracy": 0.7141258925199508,
      "num_tokens": 9357327.0,
      "step": 4720
    },
    {
      "entropy": 1.1704344421625137,
      "epoch": 0.08496878789239683,
      "grad_norm": 8.3125,
      "learning_rate": 4.942817181881573e-05,
      "loss": 1.2148893356323243,
      "mean_token_accuracy": 0.725588996708393,
      "num_tokens": 9377018.0,
      "step": 4730
    },
    {
      "entropy": 1.0907916009426117,
      "epoch": 0.08514842592176763,
      "grad_norm": 7.09375,
      "learning_rate": 4.9425112605821764e-05,
      "loss": 1.0532269477844238,
      "mean_token_accuracy": 0.7582018673419952,
      "num_tokens": 9396611.0,
      "step": 4740
    },
    {
      "entropy": 1.085527414828539,
      "epoch": 0.08532806395113846,
      "grad_norm": 6.28125,
      "learning_rate": 4.942204532661016e-05,
      "loss": 1.1715558052062989,
      "mean_token_accuracy": 0.7360363766551018,
      "num_tokens": 9416791.0,
      "step": 4750
    },
    {
      "entropy": 1.0757691830396652,
      "epoch": 0.08550770198050928,
      "grad_norm": 5.71875,
      "learning_rate": 4.941896998219384e-05,
      "loss": 1.093851089477539,
      "mean_token_accuracy": 0.7434619531035424,
      "num_tokens": 9437765.0,
      "step": 4760
    },
    {
      "entropy": 1.170188531279564,
      "epoch": 0.0856873400098801,
      "grad_norm": 7.25,
      "learning_rate": 4.941588657358842e-05,
      "loss": 1.199110221862793,
      "mean_token_accuracy": 0.7326875366270542,
      "num_tokens": 9457939.0,
      "step": 4770
    },
    {
      "entropy": 1.1072071999311448,
      "epoch": 0.0858669780392509,
      "grad_norm": 6.65625,
      "learning_rate": 4.941279510181218e-05,
      "loss": 1.197795295715332,
      "mean_token_accuracy": 0.7166375622153283,
      "num_tokens": 9478026.0,
      "step": 4780
    },
    {
      "entropy": 1.1615648224949837,
      "epoch": 0.08604661606862173,
      "grad_norm": 5.375,
      "learning_rate": 4.940969556788606e-05,
      "loss": 1.1371450424194336,
      "mean_token_accuracy": 0.7349997401237488,
      "num_tokens": 9497918.0,
      "step": 4790
    },
    {
      "entropy": 1.1660807460546494,
      "epoch": 0.08622625409799255,
      "grad_norm": 9.0,
      "learning_rate": 4.940658797283365e-05,
      "loss": 1.2441143989562988,
      "mean_token_accuracy": 0.7208584748208523,
      "num_tokens": 9517487.0,
      "step": 4800
    },
    {
      "entropy": 1.0893693819642067,
      "epoch": 0.08640589212736337,
      "grad_norm": 6.125,
      "learning_rate": 4.940347231768121e-05,
      "loss": 1.127051830291748,
      "mean_token_accuracy": 0.7391569375991821,
      "num_tokens": 9536686.0,
      "step": 4810
    },
    {
      "entropy": 1.196195873618126,
      "epoch": 0.08658553015673418,
      "grad_norm": 7.4375,
      "learning_rate": 4.9400348603457666e-05,
      "loss": 1.2687178611755372,
      "mean_token_accuracy": 0.7227301627397538,
      "num_tokens": 9556431.0,
      "step": 4820
    },
    {
      "entropy": 1.0226236060261726,
      "epoch": 0.086765168186105,
      "grad_norm": 5.53125,
      "learning_rate": 4.93972168311946e-05,
      "loss": 1.0510937690734863,
      "mean_token_accuracy": 0.7562260389328003,
      "num_tokens": 9575900.0,
      "step": 4830
    },
    {
      "entropy": 1.249869167804718,
      "epoch": 0.08694480621547582,
      "grad_norm": 7.4375,
      "learning_rate": 4.939407700192627e-05,
      "loss": 1.254288101196289,
      "mean_token_accuracy": 0.714546873420477,
      "num_tokens": 9596978.0,
      "step": 4840
    },
    {
      "entropy": 1.0518839672207831,
      "epoch": 0.08712444424484664,
      "grad_norm": 9.1875,
      "learning_rate": 4.9390929116689564e-05,
      "loss": 1.0785776138305665,
      "mean_token_accuracy": 0.752020961791277,
      "num_tokens": 9616364.0,
      "step": 4850
    },
    {
      "entropy": 0.979648444801569,
      "epoch": 0.08730408227421745,
      "grad_norm": 7.6875,
      "learning_rate": 4.938777317652405e-05,
      "loss": 1.0633100509643554,
      "mean_token_accuracy": 0.7483566731214524,
      "num_tokens": 9637270.0,
      "step": 4860
    },
    {
      "entropy": 1.0675409823656081,
      "epoch": 0.08748372030358827,
      "grad_norm": 9.25,
      "learning_rate": 4.9384609182471964e-05,
      "loss": 1.1480443954467774,
      "mean_token_accuracy": 0.7428780734539032,
      "num_tokens": 9658063.0,
      "step": 4870
    },
    {
      "entropy": 1.1188408970832824,
      "epoch": 0.08766335833295909,
      "grad_norm": 6.71875,
      "learning_rate": 4.938143713557819e-05,
      "loss": 1.2195775985717774,
      "mean_token_accuracy": 0.721643791347742,
      "num_tokens": 9677401.0,
      "step": 4880
    },
    {
      "entropy": 1.0885242834687232,
      "epoch": 0.08784299636232991,
      "grad_norm": 8.125,
      "learning_rate": 4.9378257036890274e-05,
      "loss": 1.0725805282592773,
      "mean_token_accuracy": 0.7543836668133735,
      "num_tokens": 9696171.0,
      "step": 4890
    },
    {
      "entropy": 1.1103441014885902,
      "epoch": 0.08802263439170073,
      "grad_norm": 11.1875,
      "learning_rate": 4.937506888745842e-05,
      "loss": 1.1757125854492188,
      "mean_token_accuracy": 0.73391894698143,
      "num_tokens": 9715382.0,
      "step": 4900
    },
    {
      "entropy": 1.1567430004477501,
      "epoch": 0.08820227242107154,
      "grad_norm": 8.6875,
      "learning_rate": 4.937187268833549e-05,
      "loss": 1.1572379112243651,
      "mean_token_accuracy": 0.7308511160314083,
      "num_tokens": 9735614.0,
      "step": 4910
    },
    {
      "entropy": 1.0268317446112634,
      "epoch": 0.08838191045044236,
      "grad_norm": 8.1875,
      "learning_rate": 4.9368668440577004e-05,
      "loss": 1.1054457664489745,
      "mean_token_accuracy": 0.7446669086813926,
      "num_tokens": 9755779.0,
      "step": 4920
    },
    {
      "entropy": 1.0308701664209365,
      "epoch": 0.08856154847981318,
      "grad_norm": 6.84375,
      "learning_rate": 4.936545614524114e-05,
      "loss": 1.089671802520752,
      "mean_token_accuracy": 0.7485545501112938,
      "num_tokens": 9775888.0,
      "step": 4930
    },
    {
      "entropy": 1.1735311090946197,
      "epoch": 0.088741186509184,
      "grad_norm": 7.625,
      "learning_rate": 4.9362235803388743e-05,
      "loss": 1.204796028137207,
      "mean_token_accuracy": 0.7297091811895371,
      "num_tokens": 9795296.0,
      "step": 4940
    },
    {
      "entropy": 1.1966588482260705,
      "epoch": 0.08892082453855481,
      "grad_norm": 7.15625,
      "learning_rate": 4.935900741608331e-05,
      "loss": 1.1896883964538574,
      "mean_token_accuracy": 0.7273554235696793,
      "num_tokens": 9815947.0,
      "step": 4950
    },
    {
      "entropy": 1.1858868986368178,
      "epoch": 0.08910046256792563,
      "grad_norm": 7.09375,
      "learning_rate": 4.935577098439098e-05,
      "loss": 1.1679813385009765,
      "mean_token_accuracy": 0.7352235659956932,
      "num_tokens": 9834549.0,
      "step": 4960
    },
    {
      "entropy": 1.1753675110638142,
      "epoch": 0.08928010059729645,
      "grad_norm": 7.15625,
      "learning_rate": 4.935252650938057e-05,
      "loss": 1.2885382652282715,
      "mean_token_accuracy": 0.7026080086827278,
      "num_tokens": 9854159.0,
      "step": 4970
    },
    {
      "entropy": 1.0198906108736991,
      "epoch": 0.08945973862666727,
      "grad_norm": 4.75,
      "learning_rate": 4.934927399212354e-05,
      "loss": 1.0887751579284668,
      "mean_token_accuracy": 0.7540128245949745,
      "num_tokens": 9874278.0,
      "step": 4980
    },
    {
      "entropy": 1.2017800465226174,
      "epoch": 0.08963937665603808,
      "grad_norm": 7.53125,
      "learning_rate": 4.9346013433694015e-05,
      "loss": 1.2068198204040528,
      "mean_token_accuracy": 0.7233853235840797,
      "num_tokens": 9893070.0,
      "step": 4990
    },
    {
      "epoch": 0.0898190146854089,
      "eval_entropy": 1.0882056572437286,
      "eval_loss": 1.1298004388809204,
      "eval_mean_token_accuracy": 0.7420213644504547,
      "eval_num_tokens": 9912598.0,
      "eval_runtime": 40.4432,
      "eval_samples_per_second": 49.452,
      "eval_steps_per_second": 6.182,
      "step": 5000
    },
    {
      "entropy": 1.0914090752601624,
      "epoch": 0.08999865271477972,
      "grad_norm": 8.75,
      "learning_rate": 4.933946819762724e-05,
      "loss": 1.1383769035339355,
      "mean_token_accuracy": 0.7404438890516758,
      "num_tokens": 9932169.0,
      "step": 5010
    },
    {
      "entropy": 1.1380635425448418,
      "epoch": 0.09017829074415054,
      "grad_norm": 9.125,
      "learning_rate": 4.933618352215149e-05,
      "loss": 1.152083396911621,
      "mean_token_accuracy": 0.7271808430552482,
      "num_tokens": 9952227.0,
      "step": 5020
    },
    {
      "entropy": 1.0220148339867592,
      "epoch": 0.09035792877352135,
      "grad_norm": 7.125,
      "learning_rate": 4.9332890809826296e-05,
      "loss": 1.0692299842834472,
      "mean_token_accuracy": 0.7524761587381363,
      "num_tokens": 9971835.0,
      "step": 5030
    },
    {
      "entropy": 1.0026740372180938,
      "epoch": 0.09053756680289217,
      "grad_norm": 6.0,
      "learning_rate": 4.9329590061739014e-05,
      "loss": 1.0836249351501466,
      "mean_token_accuracy": 0.7460832521319389,
      "num_tokens": 9992157.0,
      "step": 5040
    },
    {
      "entropy": 1.178272594511509,
      "epoch": 0.09071720483226299,
      "grad_norm": 7.125,
      "learning_rate": 4.9326281278979724e-05,
      "loss": 1.2511772155761718,
      "mean_token_accuracy": 0.7174595117568969,
      "num_tokens": 10013018.0,
      "step": 5050
    },
    {
      "entropy": 1.2776155576109887,
      "epoch": 0.09089684286163381,
      "grad_norm": 6.8125,
      "learning_rate": 4.932296446264111e-05,
      "loss": 1.2887293815612793,
      "mean_token_accuracy": 0.7096857517957688,
      "num_tokens": 10033857.0,
      "step": 5060
    },
    {
      "entropy": 0.9546220138669014,
      "epoch": 0.09107648089100463,
      "grad_norm": 6.90625,
      "learning_rate": 4.9319639613818534e-05,
      "loss": 1.0112491607666017,
      "mean_token_accuracy": 0.7743618234992027,
      "num_tokens": 10053604.0,
      "step": 5070
    },
    {
      "entropy": 0.9962936744093895,
      "epoch": 0.09125611892037544,
      "grad_norm": 5.75,
      "learning_rate": 4.931630673361e-05,
      "loss": 1.0010974884033204,
      "mean_token_accuracy": 0.7660636007785797,
      "num_tokens": 10073538.0,
      "step": 5080
    },
    {
      "entropy": 1.0547944888472558,
      "epoch": 0.09143575694974626,
      "grad_norm": 6.4375,
      "learning_rate": 4.931296582311618e-05,
      "loss": 1.1485336303710938,
      "mean_token_accuracy": 0.7320487529039383,
      "num_tokens": 10093105.0,
      "step": 5090
    },
    {
      "entropy": 1.142357422411442,
      "epoch": 0.09161539497911708,
      "grad_norm": 8.25,
      "learning_rate": 4.930961688344036e-05,
      "loss": 1.2027498245239259,
      "mean_token_accuracy": 0.7288209445774555,
      "num_tokens": 10114480.0,
      "step": 5100
    },
    {
      "entropy": 1.0491223186254501,
      "epoch": 0.0917950330084879,
      "grad_norm": 7.03125,
      "learning_rate": 4.930625991568851e-05,
      "loss": 1.0472751617431642,
      "mean_token_accuracy": 0.7557002186775208,
      "num_tokens": 10134713.0,
      "step": 5110
    },
    {
      "entropy": 1.0559690818190575,
      "epoch": 0.09197467103785871,
      "grad_norm": 8.375,
      "learning_rate": 4.930289492096927e-05,
      "loss": 1.0869626998901367,
      "mean_token_accuracy": 0.7524943515658379,
      "num_tokens": 10154936.0,
      "step": 5120
    },
    {
      "entropy": 0.9721009105443954,
      "epoch": 0.09215430906722953,
      "grad_norm": 6.53125,
      "learning_rate": 4.929952190039388e-05,
      "loss": 1.018129253387451,
      "mean_token_accuracy": 0.7587607815861702,
      "num_tokens": 10174250.0,
      "step": 5130
    },
    {
      "entropy": 1.042512382566929,
      "epoch": 0.09233394709660035,
      "grad_norm": 5.71875,
      "learning_rate": 4.929614085507627e-05,
      "loss": 1.0626505851745605,
      "mean_token_accuracy": 0.7524214640259743,
      "num_tokens": 10193908.0,
      "step": 5140
    },
    {
      "entropy": 1.0346194952726364,
      "epoch": 0.09251358512597117,
      "grad_norm": 7.625,
      "learning_rate": 4.929275178613298e-05,
      "loss": 1.1145664215087892,
      "mean_token_accuracy": 0.7511129155755043,
      "num_tokens": 10214554.0,
      "step": 5150
    },
    {
      "entropy": 1.1123483985662461,
      "epoch": 0.09269322315534198,
      "grad_norm": 6.59375,
      "learning_rate": 4.928935469468327e-05,
      "loss": 1.1143043518066407,
      "mean_token_accuracy": 0.744641299545765,
      "num_tokens": 10236233.0,
      "step": 5160
    },
    {
      "entropy": 1.0230978533625603,
      "epoch": 0.0928728611847128,
      "grad_norm": 6.90625,
      "learning_rate": 4.9285949581848964e-05,
      "loss": 1.0614803314208985,
      "mean_token_accuracy": 0.7528670564293861,
      "num_tokens": 10256452.0,
      "step": 5170
    },
    {
      "entropy": 1.0922610640525818,
      "epoch": 0.09305249921408362,
      "grad_norm": 5.1875,
      "learning_rate": 4.928253644875459e-05,
      "loss": 1.1323586463928224,
      "mean_token_accuracy": 0.7398323759436607,
      "num_tokens": 10276611.0,
      "step": 5180
    },
    {
      "entropy": 1.1038916900753974,
      "epoch": 0.09323213724345444,
      "grad_norm": 7.09375,
      "learning_rate": 4.9279115296527324e-05,
      "loss": 1.1701594352722169,
      "mean_token_accuracy": 0.7353798605501651,
      "num_tokens": 10296993.0,
      "step": 5190
    },
    {
      "entropy": 1.1484145201742648,
      "epoch": 0.09341177527282525,
      "grad_norm": 6.375,
      "learning_rate": 4.927568612629696e-05,
      "loss": 1.1665028572082519,
      "mean_token_accuracy": 0.7322954878211021,
      "num_tokens": 10316448.0,
      "step": 5200
    },
    {
      "entropy": 1.012902107089758,
      "epoch": 0.09359141330219607,
      "grad_norm": 5.5,
      "learning_rate": 4.927224893919595e-05,
      "loss": 1.0114215850830077,
      "mean_token_accuracy": 0.765603469312191,
      "num_tokens": 10337065.0,
      "step": 5210
    },
    {
      "entropy": 1.0247754469513892,
      "epoch": 0.0937710513315669,
      "grad_norm": 5.0625,
      "learning_rate": 4.9268803736359426e-05,
      "loss": 1.1000990867614746,
      "mean_token_accuracy": 0.7513218760490418,
      "num_tokens": 10356953.0,
      "step": 5220
    },
    {
      "entropy": 1.1060189202427864,
      "epoch": 0.09395068936093771,
      "grad_norm": 6.59375,
      "learning_rate": 4.926535051892513e-05,
      "loss": 1.0986078262329102,
      "mean_token_accuracy": 0.7419395886361599,
      "num_tokens": 10376309.0,
      "step": 5230
    },
    {
      "entropy": 1.0925479769706725,
      "epoch": 0.09413032739030854,
      "grad_norm": 8.375,
      "learning_rate": 4.9261889288033445e-05,
      "loss": 1.1497454643249512,
      "mean_token_accuracy": 0.7397084981203079,
      "num_tokens": 10397081.0,
      "step": 5240
    },
    {
      "entropy": 1.1301081612706185,
      "epoch": 0.09430996541967934,
      "grad_norm": 10.0625,
      "learning_rate": 4.925842004482745e-05,
      "loss": 1.2125494956970215,
      "mean_token_accuracy": 0.7250224642455578,
      "num_tokens": 10416823.0,
      "step": 5250
    },
    {
      "entropy": 1.1606574088335038,
      "epoch": 0.09448960344905016,
      "grad_norm": 6.84375,
      "learning_rate": 4.925494279045281e-05,
      "loss": 1.2046463012695312,
      "mean_token_accuracy": 0.7321810111403465,
      "num_tokens": 10436705.0,
      "step": 5260
    },
    {
      "entropy": 1.162073915451765,
      "epoch": 0.09466924147842098,
      "grad_norm": 7.40625,
      "learning_rate": 4.925145752605788e-05,
      "loss": 1.1661548614501953,
      "mean_token_accuracy": 0.7304585590958595,
      "num_tokens": 10456339.0,
      "step": 5270
    },
    {
      "entropy": 1.136557611823082,
      "epoch": 0.0948488795077918,
      "grad_norm": 7.15625,
      "learning_rate": 4.9247964252793624e-05,
      "loss": 1.2424405097961426,
      "mean_token_accuracy": 0.7333075851202011,
      "num_tokens": 10477205.0,
      "step": 5280
    },
    {
      "entropy": 1.0063030883669852,
      "epoch": 0.09502851753716261,
      "grad_norm": 4.71875,
      "learning_rate": 4.9244462971813696e-05,
      "loss": 0.9620872497558594,
      "mean_token_accuracy": 0.7753238156437874,
      "num_tokens": 10497810.0,
      "step": 5290
    },
    {
      "entropy": 0.9935022726655006,
      "epoch": 0.09520815556653343,
      "grad_norm": 7.15625,
      "learning_rate": 4.924095368427435e-05,
      "loss": 1.0597262382507324,
      "mean_token_accuracy": 0.7567232459783554,
      "num_tokens": 10518412.0,
      "step": 5300
    },
    {
      "entropy": 1.1183653250336647,
      "epoch": 0.09538779359590425,
      "grad_norm": 7.03125,
      "learning_rate": 4.923743639133451e-05,
      "loss": 1.1989439010620118,
      "mean_token_accuracy": 0.7255274146795273,
      "num_tokens": 10538274.0,
      "step": 5310
    },
    {
      "entropy": 1.0052246682345867,
      "epoch": 0.09556743162527508,
      "grad_norm": 8.5625,
      "learning_rate": 4.9233911094155735e-05,
      "loss": 1.05171537399292,
      "mean_token_accuracy": 0.7563199654221535,
      "num_tokens": 10559604.0,
      "step": 5320
    },
    {
      "entropy": 1.1118532486259938,
      "epoch": 0.09574706965464588,
      "grad_norm": 7.09375,
      "learning_rate": 4.9230377793902236e-05,
      "loss": 1.1537075996398927,
      "mean_token_accuracy": 0.743959829211235,
      "num_tokens": 10579387.0,
      "step": 5330
    },
    {
      "entropy": 1.1166372366249562,
      "epoch": 0.0959267076840167,
      "grad_norm": 5.9375,
      "learning_rate": 4.9226836491740844e-05,
      "loss": 1.1717398643493653,
      "mean_token_accuracy": 0.733088281005621,
      "num_tokens": 10598453.0,
      "step": 5340
    },
    {
      "entropy": 1.1397918701171874,
      "epoch": 0.09610634571338753,
      "grad_norm": 5.4375,
      "learning_rate": 4.922328718884107e-05,
      "loss": 1.141579532623291,
      "mean_token_accuracy": 0.7417924642562866,
      "num_tokens": 10617952.0,
      "step": 5350
    },
    {
      "entropy": 1.047075329720974,
      "epoch": 0.09628598374275835,
      "grad_norm": 7.0625,
      "learning_rate": 4.9219729886375024e-05,
      "loss": 1.0157114028930665,
      "mean_token_accuracy": 0.765577121078968,
      "num_tokens": 10637374.0,
      "step": 5360
    },
    {
      "entropy": 1.020466460287571,
      "epoch": 0.09646562177212915,
      "grad_norm": 6.03125,
      "learning_rate": 4.9216164585517494e-05,
      "loss": 1.0876362800598145,
      "mean_token_accuracy": 0.7573550254106521,
      "num_tokens": 10657173.0,
      "step": 5370
    },
    {
      "entropy": 1.008359081298113,
      "epoch": 0.09664525980149997,
      "grad_norm": 6.9375,
      "learning_rate": 4.921259128744589e-05,
      "loss": 1.0307931900024414,
      "mean_token_accuracy": 0.7562927782535553,
      "num_tokens": 10677472.0,
      "step": 5380
    },
    {
      "entropy": 1.0701922655105591,
      "epoch": 0.0968248978308708,
      "grad_norm": 6.625,
      "learning_rate": 4.9209009993340264e-05,
      "loss": 1.1034717559814453,
      "mean_token_accuracy": 0.7494121044874191,
      "num_tokens": 10698124.0,
      "step": 5390
    },
    {
      "entropy": 1.1503942504525184,
      "epoch": 0.09700453586024162,
      "grad_norm": 7.5,
      "learning_rate": 4.920542070438332e-05,
      "loss": 1.2291152000427246,
      "mean_token_accuracy": 0.7066176936030388,
      "num_tokens": 10717994.0,
      "step": 5400
    },
    {
      "entropy": 1.035945226252079,
      "epoch": 0.09718417388961242,
      "grad_norm": 5.59375,
      "learning_rate": 4.9201823421760395e-05,
      "loss": 1.093758487701416,
      "mean_token_accuracy": 0.7496545135974884,
      "num_tokens": 10739824.0,
      "step": 5410
    },
    {
      "entropy": 1.155145500600338,
      "epoch": 0.09736381191898325,
      "grad_norm": 10.125,
      "learning_rate": 4.919821814665946e-05,
      "loss": 1.1667027473449707,
      "mean_token_accuracy": 0.732168273627758,
      "num_tokens": 10758786.0,
      "step": 5420
    },
    {
      "entropy": 1.170270571857691,
      "epoch": 0.09754344994835407,
      "grad_norm": 8.125,
      "learning_rate": 4.919460488027113e-05,
      "loss": 1.2354021072387695,
      "mean_token_accuracy": 0.7238908313214779,
      "num_tokens": 10780006.0,
      "step": 5430
    },
    {
      "entropy": 1.1628091730177403,
      "epoch": 0.09772308797772489,
      "grad_norm": 8.125,
      "learning_rate": 4.919098362378868e-05,
      "loss": 1.205706787109375,
      "mean_token_accuracy": 0.7184412762522697,
      "num_tokens": 10800674.0,
      "step": 5440
    },
    {
      "entropy": 1.185660894215107,
      "epoch": 0.09790272600709571,
      "grad_norm": 8.3125,
      "learning_rate": 4.918735437840797e-05,
      "loss": 1.1491731643676757,
      "mean_token_accuracy": 0.7371725812554359,
      "num_tokens": 10821089.0,
      "step": 5450
    },
    {
      "entropy": 1.1204293459653853,
      "epoch": 0.09808236403646652,
      "grad_norm": 5.8125,
      "learning_rate": 4.918371714532756e-05,
      "loss": 1.2428881645202636,
      "mean_token_accuracy": 0.7213858619332314,
      "num_tokens": 10842765.0,
      "step": 5460
    },
    {
      "entropy": 1.097996361553669,
      "epoch": 0.09826200206583734,
      "grad_norm": 6.28125,
      "learning_rate": 4.918007192574861e-05,
      "loss": 1.085082721710205,
      "mean_token_accuracy": 0.7507482290267944,
      "num_tokens": 10862834.0,
      "step": 5470
    },
    {
      "entropy": 1.0705626383423805,
      "epoch": 0.09844164009520816,
      "grad_norm": 6.53125,
      "learning_rate": 4.917641872087492e-05,
      "loss": 1.1050758361816406,
      "mean_token_accuracy": 0.745497590303421,
      "num_tokens": 10882137.0,
      "step": 5480
    },
    {
      "entropy": 1.1645829781889916,
      "epoch": 0.09862127812457898,
      "grad_norm": 6.59375,
      "learning_rate": 4.917275753191294e-05,
      "loss": 1.203923511505127,
      "mean_token_accuracy": 0.7298523351550102,
      "num_tokens": 10901620.0,
      "step": 5490
    },
    {
      "entropy": 1.1557079285383225,
      "epoch": 0.09880091615394979,
      "grad_norm": 6.0625,
      "learning_rate": 4.916908836007176e-05,
      "loss": 1.1813009262084961,
      "mean_token_accuracy": 0.7358975663781167,
      "num_tokens": 10921106.0,
      "step": 5500
    },
    {
      "entropy": 0.9995366096496582,
      "epoch": 0.09898055418332061,
      "grad_norm": 8.9375,
      "learning_rate": 4.91654112065631e-05,
      "loss": 1.0133164405822754,
      "mean_token_accuracy": 0.7629754096269608,
      "num_tokens": 10940701.0,
      "step": 5510
    },
    {
      "entropy": 1.0657085008919238,
      "epoch": 0.09916019221269143,
      "grad_norm": 6.9375,
      "learning_rate": 4.9161726072601286e-05,
      "loss": 1.174940299987793,
      "mean_token_accuracy": 0.7383551955223083,
      "num_tokens": 10961195.0,
      "step": 5520
    },
    {
      "entropy": 1.023510032892227,
      "epoch": 0.09933983024206225,
      "grad_norm": 6.1875,
      "learning_rate": 4.915803295940333e-05,
      "loss": 1.0425142288208007,
      "mean_token_accuracy": 0.7584219187498092,
      "num_tokens": 10979831.0,
      "step": 5530
    },
    {
      "entropy": 1.1189694494009017,
      "epoch": 0.09951946827143306,
      "grad_norm": 7.21875,
      "learning_rate": 4.9154331868188865e-05,
      "loss": 1.1405045509338378,
      "mean_token_accuracy": 0.7520097717642784,
      "num_tokens": 10998828.0,
      "step": 5540
    },
    {
      "entropy": 1.0844595924019813,
      "epoch": 0.09969910630080388,
      "grad_norm": 4.65625,
      "learning_rate": 4.915062280018012e-05,
      "loss": 1.0839847564697265,
      "mean_token_accuracy": 0.7443592548370361,
      "num_tokens": 11018857.0,
      "step": 5550
    },
    {
      "entropy": 1.1899598315358162,
      "epoch": 0.0998787443301747,
      "grad_norm": 5.5,
      "learning_rate": 4.914690575660202e-05,
      "loss": 1.165814208984375,
      "mean_token_accuracy": 0.7357245326042176,
      "num_tokens": 11037838.0,
      "step": 5560
    },
    {
      "entropy": 1.0412156522274016,
      "epoch": 0.10005838235954552,
      "grad_norm": 6.3125,
      "learning_rate": 4.914318073868207e-05,
      "loss": 1.1189002990722656,
      "mean_token_accuracy": 0.7408639818429947,
      "num_tokens": 11057599.0,
      "step": 5570
    },
    {
      "entropy": 1.0312755942344665,
      "epoch": 0.10023802038891633,
      "grad_norm": 9.125,
      "learning_rate": 4.913944774765044e-05,
      "loss": 1.061415672302246,
      "mean_token_accuracy": 0.7434593051671982,
      "num_tokens": 11077692.0,
      "step": 5580
    },
    {
      "entropy": 1.0085327550768852,
      "epoch": 0.10041765841828715,
      "grad_norm": 4.84375,
      "learning_rate": 4.913570678473991e-05,
      "loss": 1.036046314239502,
      "mean_token_accuracy": 0.7555720120668411,
      "num_tokens": 11098514.0,
      "step": 5590
    },
    {
      "entropy": 1.0291207820177077,
      "epoch": 0.10059729644765797,
      "grad_norm": 6.5,
      "learning_rate": 4.9131957851185915e-05,
      "loss": 1.128465175628662,
      "mean_token_accuracy": 0.7424156993627549,
      "num_tokens": 11118565.0,
      "step": 5600
    },
    {
      "entropy": 1.15163347274065,
      "epoch": 0.10077693447702879,
      "grad_norm": 7.4375,
      "learning_rate": 4.912820094822652e-05,
      "loss": 1.1871371269226074,
      "mean_token_accuracy": 0.7359251111745835,
      "num_tokens": 11138316.0,
      "step": 5610
    },
    {
      "entropy": 1.1466680020093918,
      "epoch": 0.10095657250639961,
      "grad_norm": 5.3125,
      "learning_rate": 4.912443607710241e-05,
      "loss": 1.148698616027832,
      "mean_token_accuracy": 0.7421729430556298,
      "num_tokens": 11158751.0,
      "step": 5620
    },
    {
      "entropy": 0.9591960638761521,
      "epoch": 0.10113621053577042,
      "grad_norm": 7.03125,
      "learning_rate": 4.9120663239056914e-05,
      "loss": 0.9600547790527344,
      "mean_token_accuracy": 0.7702839434146881,
      "num_tokens": 11178084.0,
      "step": 5630
    },
    {
      "entropy": 0.9736021757125854,
      "epoch": 0.10131584856514124,
      "grad_norm": 8.5625,
      "learning_rate": 4.9116882435335966e-05,
      "loss": 1.0426530838012695,
      "mean_token_accuracy": 0.7497874781489372,
      "num_tokens": 11198042.0,
      "step": 5640
    },
    {
      "entropy": 1.1649086952209473,
      "epoch": 0.10149548659451206,
      "grad_norm": 7.25,
      "learning_rate": 4.911309366718817e-05,
      "loss": 1.2064261436462402,
      "mean_token_accuracy": 0.7338306829333305,
      "num_tokens": 11217349.0,
      "step": 5650
    },
    {
      "entropy": 1.0595213398337364,
      "epoch": 0.10167512462388288,
      "grad_norm": 7.1875,
      "learning_rate": 4.910929693586472e-05,
      "loss": 1.0868009567260741,
      "mean_token_accuracy": 0.7516065120697022,
      "num_tokens": 11237284.0,
      "step": 5660
    },
    {
      "entropy": 1.1015644878149033,
      "epoch": 0.10185476265325369,
      "grad_norm": 7.75,
      "learning_rate": 4.910549224261947e-05,
      "loss": 1.1364654541015624,
      "mean_token_accuracy": 0.7441115751862526,
      "num_tokens": 11256958.0,
      "step": 5670
    },
    {
      "entropy": 1.0432453140616418,
      "epoch": 0.10203440068262451,
      "grad_norm": 6.5625,
      "learning_rate": 4.910167958870889e-05,
      "loss": 1.1159459114074708,
      "mean_token_accuracy": 0.7430467203259468,
      "num_tokens": 11276778.0,
      "step": 5680
    },
    {
      "entropy": 1.076836434006691,
      "epoch": 0.10221403871199533,
      "grad_norm": 6.5625,
      "learning_rate": 4.909785897539209e-05,
      "loss": 1.1307708740234375,
      "mean_token_accuracy": 0.7387826569378376,
      "num_tokens": 11297715.0,
      "step": 5690
    },
    {
      "entropy": 1.1495344325900079,
      "epoch": 0.10239367674136615,
      "grad_norm": 6.59375,
      "learning_rate": 4.909403040393079e-05,
      "loss": 1.192055606842041,
      "mean_token_accuracy": 0.72905587926507,
      "num_tokens": 11317785.0,
      "step": 5700
    },
    {
      "entropy": 1.0696873247623444,
      "epoch": 0.10257331477073696,
      "grad_norm": 7.8125,
      "learning_rate": 4.909019387558935e-05,
      "loss": 1.1156390190124512,
      "mean_token_accuracy": 0.7452369824051857,
      "num_tokens": 11338476.0,
      "step": 5710
    },
    {
      "entropy": 1.1068349450826644,
      "epoch": 0.10275295280010778,
      "grad_norm": 5.0625,
      "learning_rate": 4.9086349391634755e-05,
      "loss": 1.116074562072754,
      "mean_token_accuracy": 0.7409947916865349,
      "num_tokens": 11358803.0,
      "step": 5720
    },
    {
      "entropy": 1.0765526920557023,
      "epoch": 0.1029325908294786,
      "grad_norm": 6.15625,
      "learning_rate": 4.908249695333662e-05,
      "loss": 1.1018148422241212,
      "mean_token_accuracy": 0.7474873699247837,
      "num_tokens": 11379952.0,
      "step": 5730
    },
    {
      "entropy": 1.0575618125498294,
      "epoch": 0.10311222885884942,
      "grad_norm": 6.5,
      "learning_rate": 4.907863656196719e-05,
      "loss": 1.1194604873657226,
      "mean_token_accuracy": 0.7468373820185661,
      "num_tokens": 11399683.0,
      "step": 5740
    },
    {
      "entropy": 1.0685867860913276,
      "epoch": 0.10329186688822023,
      "grad_norm": 7.75,
      "learning_rate": 4.907476821880132e-05,
      "loss": 1.0872982025146485,
      "mean_token_accuracy": 0.748898808658123,
      "num_tokens": 11419410.0,
      "step": 5750
    },
    {
      "entropy": 1.0381993144750594,
      "epoch": 0.10347150491759105,
      "grad_norm": 7.1875,
      "learning_rate": 4.907089192511651e-05,
      "loss": 1.0862348556518555,
      "mean_token_accuracy": 0.746786642074585,
      "num_tokens": 11438772.0,
      "step": 5760
    },
    {
      "entropy": 1.0364542946219444,
      "epoch": 0.10365114294696187,
      "grad_norm": 4.9375,
      "learning_rate": 4.906700768219288e-05,
      "loss": 1.0907612800598145,
      "mean_token_accuracy": 0.7433740258216858,
      "num_tokens": 11459210.0,
      "step": 5770
    },
    {
      "entropy": 1.0087475448846817,
      "epoch": 0.10383078097633269,
      "grad_norm": 6.875,
      "learning_rate": 4.906311549131317e-05,
      "loss": 1.0176715850830078,
      "mean_token_accuracy": 0.7580270543694496,
      "num_tokens": 11478691.0,
      "step": 5780
    },
    {
      "entropy": 0.9338625639677047,
      "epoch": 0.10401041900570351,
      "grad_norm": 5.71875,
      "learning_rate": 4.905921535376275e-05,
      "loss": 0.9635801315307617,
      "mean_token_accuracy": 0.767186838388443,
      "num_tokens": 11498293.0,
      "step": 5790
    },
    {
      "entropy": 1.1373558983206749,
      "epoch": 0.10419005703507432,
      "grad_norm": 6.5625,
      "learning_rate": 4.905530727082961e-05,
      "loss": 1.1965977668762207,
      "mean_token_accuracy": 0.7290416166186333,
      "num_tokens": 11518715.0,
      "step": 5800
    },
    {
      "entropy": 1.1861480817198753,
      "epoch": 0.10436969506444514,
      "grad_norm": 6.5625,
      "learning_rate": 4.9051391243804364e-05,
      "loss": 1.1814220428466797,
      "mean_token_accuracy": 0.7295829258859158,
      "num_tokens": 11538859.0,
      "step": 5810
    },
    {
      "entropy": 1.0245934382081032,
      "epoch": 0.10454933309381596,
      "grad_norm": 5.6875,
      "learning_rate": 4.904746727398025e-05,
      "loss": 1.0298347473144531,
      "mean_token_accuracy": 0.7525382161140441,
      "num_tokens": 11557500.0,
      "step": 5820
    },
    {
      "entropy": 1.01600990742445,
      "epoch": 0.10472897112318678,
      "grad_norm": 7.6875,
      "learning_rate": 4.9043535362653146e-05,
      "loss": 1.0304162979125977,
      "mean_token_accuracy": 0.7633151859045029,
      "num_tokens": 11577984.0,
      "step": 5830
    },
    {
      "entropy": 1.0156742125749587,
      "epoch": 0.10490860915255759,
      "grad_norm": 7.34375,
      "learning_rate": 4.903959551112153e-05,
      "loss": 1.1048483848571777,
      "mean_token_accuracy": 0.7479206956923008,
      "num_tokens": 11598405.0,
      "step": 5840
    },
    {
      "entropy": 1.0360478840768337,
      "epoch": 0.10508824718192841,
      "grad_norm": 6.15625,
      "learning_rate": 4.9035647720686495e-05,
      "loss": 1.0700927734375,
      "mean_token_accuracy": 0.7488227054476738,
      "num_tokens": 11617654.0,
      "step": 5850
    },
    {
      "entropy": 1.009080171585083,
      "epoch": 0.10526788521129923,
      "grad_norm": 5.28125,
      "learning_rate": 4.9031691992651784e-05,
      "loss": 1.0658652305603027,
      "mean_token_accuracy": 0.7580617651343345,
      "num_tokens": 11637527.0,
      "step": 5860
    },
    {
      "entropy": 1.0495788708329201,
      "epoch": 0.10544752324067005,
      "grad_norm": 6.78125,
      "learning_rate": 4.9027728328323755e-05,
      "loss": 1.0697233200073242,
      "mean_token_accuracy": 0.7545733407139779,
      "num_tokens": 11657057.0,
      "step": 5870
    },
    {
      "entropy": 1.0322448551654815,
      "epoch": 0.10562716127004086,
      "grad_norm": 6.9375,
      "learning_rate": 4.9023756729011364e-05,
      "loss": 1.063349437713623,
      "mean_token_accuracy": 0.7505650848150254,
      "num_tokens": 11677237.0,
      "step": 5880
    },
    {
      "entropy": 1.0717719510197639,
      "epoch": 0.10580679929941168,
      "grad_norm": 6.34375,
      "learning_rate": 4.9019777196026204e-05,
      "loss": 1.0954110145568847,
      "mean_token_accuracy": 0.7442150920629501,
      "num_tokens": 11696801.0,
      "step": 5890
    },
    {
      "entropy": 1.0516410022974014,
      "epoch": 0.1059864373287825,
      "grad_norm": 8.25,
      "learning_rate": 4.901578973068249e-05,
      "loss": 1.123201847076416,
      "mean_token_accuracy": 0.7358275979757309,
      "num_tokens": 11716517.0,
      "step": 5900
    },
    {
      "entropy": 1.1413511157035827,
      "epoch": 0.10616607535815333,
      "grad_norm": 7.03125,
      "learning_rate": 4.9011794334297064e-05,
      "loss": 1.1876730918884277,
      "mean_token_accuracy": 0.7294520512223244,
      "num_tokens": 11735790.0,
      "step": 5910
    },
    {
      "entropy": 1.1290773004293442,
      "epoch": 0.10634571338752413,
      "grad_norm": 7.40625,
      "learning_rate": 4.9007791008189376e-05,
      "loss": 1.1772315979003907,
      "mean_token_accuracy": 0.7263943634927272,
      "num_tokens": 11755070.0,
      "step": 5920
    },
    {
      "entropy": 1.0778159394860267,
      "epoch": 0.10652535141689495,
      "grad_norm": 7.8125,
      "learning_rate": 4.9003779753681476e-05,
      "loss": 1.0553714752197265,
      "mean_token_accuracy": 0.7557582795619965,
      "num_tokens": 11775017.0,
      "step": 5930
    },
    {
      "entropy": 1.045047302544117,
      "epoch": 0.10670498944626577,
      "grad_norm": 8.0625,
      "learning_rate": 4.8999760572098076e-05,
      "loss": 1.1167244911193848,
      "mean_token_accuracy": 0.7462304264307023,
      "num_tokens": 11794892.0,
      "step": 5940
    },
    {
      "entropy": 1.0112846434116363,
      "epoch": 0.1068846274756366,
      "grad_norm": 6.6875,
      "learning_rate": 4.8995733464766456e-05,
      "loss": 0.9950170516967773,
      "mean_token_accuracy": 0.7710164472460747,
      "num_tokens": 11813821.0,
      "step": 5950
    },
    {
      "entropy": 0.9848114423453808,
      "epoch": 0.10706426550500742,
      "grad_norm": 5.5,
      "learning_rate": 4.8991698433016567e-05,
      "loss": 1.0798036575317382,
      "mean_token_accuracy": 0.7548213630914689,
      "num_tokens": 11834346.0,
      "step": 5960
    },
    {
      "entropy": 1.0721779204905033,
      "epoch": 0.10724390353437822,
      "grad_norm": 7.71875,
      "learning_rate": 4.8987655478180926e-05,
      "loss": 1.124882698059082,
      "mean_token_accuracy": 0.7332268178462982,
      "num_tokens": 11853516.0,
      "step": 5970
    },
    {
      "entropy": 1.143219293653965,
      "epoch": 0.10742354156374904,
      "grad_norm": 6.125,
      "learning_rate": 4.898360460159471e-05,
      "loss": 1.1659598350524902,
      "mean_token_accuracy": 0.7341182455420494,
      "num_tokens": 11873707.0,
      "step": 5980
    },
    {
      "entropy": 1.1560181632637978,
      "epoch": 0.10760317959311987,
      "grad_norm": 5.59375,
      "learning_rate": 4.897954580459567e-05,
      "loss": 1.1859731674194336,
      "mean_token_accuracy": 0.7260950461030007,
      "num_tokens": 11893813.0,
      "step": 5990
    },
    {
      "epoch": 0.10778281762249069,
      "eval_entropy": 1.0263335263729096,
      "eval_loss": 1.0623183250427246,
      "eval_mean_token_accuracy": 0.7533111233711243,
      "eval_num_tokens": 11913033.0,
      "eval_runtime": 40.5662,
      "eval_samples_per_second": 49.302,
      "eval_steps_per_second": 6.163,
      "step": 6000
    },
    {
      "entropy": 0.9956502024084329,
      "epoch": 0.1079624556518615,
      "grad_norm": 6.6875,
      "learning_rate": 4.897140445472334e-05,
      "loss": 1.023189640045166,
      "mean_token_accuracy": 0.7594320617616177,
      "num_tokens": 11933234.0,
      "step": 6010
    },
    {
      "entropy": 1.0894772738218308,
      "epoch": 0.10814209368123232,
      "grad_norm": 5.96875,
      "learning_rate": 4.896732190453866e-05,
      "loss": 1.1016148567199706,
      "mean_token_accuracy": 0.74625184237957,
      "num_tokens": 11954910.0,
      "step": 6020
    },
    {
      "entropy": 1.0253516979515553,
      "epoch": 0.10832173171060314,
      "grad_norm": 6.4375,
      "learning_rate": 4.896323143931841e-05,
      "loss": 1.107302761077881,
      "mean_token_accuracy": 0.7491895817220211,
      "num_tokens": 11975608.0,
      "step": 6030
    },
    {
      "entropy": 1.0746199190616608,
      "epoch": 0.10850136973997396,
      "grad_norm": 7.375,
      "learning_rate": 4.895913306041344e-05,
      "loss": 1.0935521125793457,
      "mean_token_accuracy": 0.7454762056469917,
      "num_tokens": 11995026.0,
      "step": 6040
    },
    {
      "entropy": 1.01798400208354,
      "epoch": 0.10868100776934476,
      "grad_norm": 7.90625,
      "learning_rate": 4.89550267691772e-05,
      "loss": 1.0459890365600586,
      "mean_token_accuracy": 0.7536133244633675,
      "num_tokens": 12015546.0,
      "step": 6050
    },
    {
      "entropy": 1.0714834108948708,
      "epoch": 0.10886064579871559,
      "grad_norm": 6.0625,
      "learning_rate": 4.895091256696578e-05,
      "loss": 1.118292999267578,
      "mean_token_accuracy": 0.7389300256967545,
      "num_tokens": 12035659.0,
      "step": 6060
    },
    {
      "entropy": 1.0880193039774895,
      "epoch": 0.1090402838280864,
      "grad_norm": 9.125,
      "learning_rate": 4.8946790455137856e-05,
      "loss": 1.0567352294921875,
      "mean_token_accuracy": 0.7475480064749718,
      "num_tokens": 12056542.0,
      "step": 6070
    },
    {
      "entropy": 1.0426205813884735,
      "epoch": 0.10921992185745723,
      "grad_norm": 6.5625,
      "learning_rate": 4.894266043505472e-05,
      "loss": 1.0677014350891114,
      "mean_token_accuracy": 0.7626322612166405,
      "num_tokens": 12077166.0,
      "step": 6080
    },
    {
      "entropy": 1.0352643243968487,
      "epoch": 0.10939955988682804,
      "grad_norm": 6.09375,
      "learning_rate": 4.89385225080803e-05,
      "loss": 1.0735333442687989,
      "mean_token_accuracy": 0.7500649943947792,
      "num_tokens": 12097940.0,
      "step": 6090
    },
    {
      "entropy": 1.0700816869735719,
      "epoch": 0.10957919791619886,
      "grad_norm": 8.5,
      "learning_rate": 4.8934376675581096e-05,
      "loss": 1.1178907394409179,
      "mean_token_accuracy": 0.743023757636547,
      "num_tokens": 12117716.0,
      "step": 6100
    },
    {
      "entropy": 0.9464042142033577,
      "epoch": 0.10975883594556968,
      "grad_norm": 6.53125,
      "learning_rate": 4.8930222938926245e-05,
      "loss": 0.9468300819396973,
      "mean_token_accuracy": 0.7686618760228157,
      "num_tokens": 12137933.0,
      "step": 6110
    },
    {
      "entropy": 1.091781882941723,
      "epoch": 0.1099384739749405,
      "grad_norm": 8.625,
      "learning_rate": 4.892606129948749e-05,
      "loss": 1.0873963356018066,
      "mean_token_accuracy": 0.7495502181351185,
      "num_tokens": 12158901.0,
      "step": 6120
    },
    {
      "entropy": 0.9942154839634896,
      "epoch": 0.11011811200431132,
      "grad_norm": 9.0,
      "learning_rate": 4.8921891758639196e-05,
      "loss": 1.0592898368835448,
      "mean_token_accuracy": 0.7540314614772796,
      "num_tokens": 12178934.0,
      "step": 6130
    },
    {
      "entropy": 0.9941693246364594,
      "epoch": 0.11029775003368213,
      "grad_norm": 6.90625,
      "learning_rate": 4.8917714317758304e-05,
      "loss": 1.0154906272888184,
      "mean_token_accuracy": 0.7608428508043289,
      "num_tokens": 12199661.0,
      "step": 6140
    },
    {
      "entropy": 1.033303937315941,
      "epoch": 0.11047738806305295,
      "grad_norm": 7.5625,
      "learning_rate": 4.891352897822439e-05,
      "loss": 1.0534424781799316,
      "mean_token_accuracy": 0.7513334073126317,
      "num_tokens": 12219554.0,
      "step": 6150
    },
    {
      "entropy": 1.0806662991642952,
      "epoch": 0.11065702609242377,
      "grad_norm": 6.90625,
      "learning_rate": 4.890933574141964e-05,
      "loss": 1.1575303077697754,
      "mean_token_accuracy": 0.7324025899171829,
      "num_tokens": 12240903.0,
      "step": 6160
    },
    {
      "entropy": 1.066120059788227,
      "epoch": 0.11083666412179459,
      "grad_norm": 6.96875,
      "learning_rate": 4.890513460872883e-05,
      "loss": 1.0571278572082519,
      "mean_token_accuracy": 0.7533085778355598,
      "num_tokens": 12260451.0,
      "step": 6170
    },
    {
      "entropy": 1.0459950476884843,
      "epoch": 0.1110163021511654,
      "grad_norm": 6.28125,
      "learning_rate": 4.890092558153935e-05,
      "loss": 1.1307196617126465,
      "mean_token_accuracy": 0.7439378961920738,
      "num_tokens": 12281836.0,
      "step": 6180
    },
    {
      "entropy": 0.9970750510692596,
      "epoch": 0.11119594018053622,
      "grad_norm": 11.375,
      "learning_rate": 4.889670866124122e-05,
      "loss": 1.059687614440918,
      "mean_token_accuracy": 0.74624994546175,
      "num_tokens": 12301409.0,
      "step": 6190
    },
    {
      "entropy": 1.0292142443358898,
      "epoch": 0.11137557820990704,
      "grad_norm": 7.15625,
      "learning_rate": 4.8892483849227034e-05,
      "loss": 1.0145362854003905,
      "mean_token_accuracy": 0.7649327725172043,
      "num_tokens": 12321304.0,
      "step": 6200
    },
    {
      "entropy": 1.0469703406095505,
      "epoch": 0.11155521623927786,
      "grad_norm": 7.78125,
      "learning_rate": 4.8888251146892e-05,
      "loss": 1.01773042678833,
      "mean_token_accuracy": 0.7551460549235344,
      "num_tokens": 12339859.0,
      "step": 6210
    },
    {
      "entropy": 0.9684609696269035,
      "epoch": 0.11173485426864867,
      "grad_norm": 5.65625,
      "learning_rate": 4.888401055563396e-05,
      "loss": 1.0565549850463867,
      "mean_token_accuracy": 0.7555677369236946,
      "num_tokens": 12359048.0,
      "step": 6220
    },
    {
      "entropy": 1.124122017621994,
      "epoch": 0.11191449229801949,
      "grad_norm": 6.375,
      "learning_rate": 4.8879762076853314e-05,
      "loss": 1.1532334327697753,
      "mean_token_accuracy": 0.740015585720539,
      "num_tokens": 12378388.0,
      "step": 6230
    },
    {
      "entropy": 1.0747182056307794,
      "epoch": 0.11209413032739031,
      "grad_norm": 7.4375,
      "learning_rate": 4.8875505711953103e-05,
      "loss": 1.108290958404541,
      "mean_token_accuracy": 0.7411955341696739,
      "num_tokens": 12397541.0,
      "step": 6240
    },
    {
      "entropy": 0.938264898955822,
      "epoch": 0.11227376835676113,
      "grad_norm": 5.28125,
      "learning_rate": 4.887124146233896e-05,
      "loss": 0.9431843757629395,
      "mean_token_accuracy": 0.7783607006072998,
      "num_tokens": 12417377.0,
      "step": 6250
    },
    {
      "entropy": 1.001605676114559,
      "epoch": 0.11245340638613194,
      "grad_norm": 6.40625,
      "learning_rate": 4.886696932941913e-05,
      "loss": 1.0423752784729003,
      "mean_token_accuracy": 0.7528838887810707,
      "num_tokens": 12437693.0,
      "step": 6260
    },
    {
      "entropy": 0.9650053367018699,
      "epoch": 0.11263304441550276,
      "grad_norm": 5.84375,
      "learning_rate": 4.8862689314604434e-05,
      "loss": 0.965366268157959,
      "mean_token_accuracy": 0.7738505646586418,
      "num_tokens": 12458209.0,
      "step": 6270
    },
    {
      "entropy": 0.9712089881300926,
      "epoch": 0.11281268244487358,
      "grad_norm": 7.96875,
      "learning_rate": 4.885840141930834e-05,
      "loss": 1.0391764640808105,
      "mean_token_accuracy": 0.7517055526375771,
      "num_tokens": 12478289.0,
      "step": 6280
    },
    {
      "entropy": 1.081689990311861,
      "epoch": 0.1129923204742444,
      "grad_norm": 7.84375,
      "learning_rate": 4.885410564494688e-05,
      "loss": 1.1368453979492188,
      "mean_token_accuracy": 0.7460439398884773,
      "num_tokens": 12498084.0,
      "step": 6290
    },
    {
      "entropy": 1.1053149342536925,
      "epoch": 0.11317195850361522,
      "grad_norm": 7.71875,
      "learning_rate": 4.8849801992938706e-05,
      "loss": 1.134336280822754,
      "mean_token_accuracy": 0.7440494827926158,
      "num_tokens": 12518286.0,
      "step": 6300
    },
    {
      "entropy": 1.0616450533270836,
      "epoch": 0.11335159653298603,
      "grad_norm": 6.59375,
      "learning_rate": 4.884549046470507e-05,
      "loss": 1.0437630653381347,
      "mean_token_accuracy": 0.752113826572895,
      "num_tokens": 12539095.0,
      "step": 6310
    },
    {
      "entropy": 1.03160240650177,
      "epoch": 0.11353123456235685,
      "grad_norm": 8.0625,
      "learning_rate": 4.884117106166981e-05,
      "loss": 1.0980774879455566,
      "mean_token_accuracy": 0.7493119806051254,
      "num_tokens": 12558359.0,
      "step": 6320
    },
    {
      "entropy": 1.05936993137002,
      "epoch": 0.11371087259172767,
      "grad_norm": 6.34375,
      "learning_rate": 4.88368437852594e-05,
      "loss": 1.0880445480346679,
      "mean_token_accuracy": 0.7531413897871971,
      "num_tokens": 12578082.0,
      "step": 6330
    },
    {
      "entropy": 0.9684530325233937,
      "epoch": 0.11389051062109849,
      "grad_norm": 6.9375,
      "learning_rate": 4.8832508636902876e-05,
      "loss": 0.9865660667419434,
      "mean_token_accuracy": 0.7738498479127884,
      "num_tokens": 12598376.0,
      "step": 6340
    },
    {
      "entropy": 1.0118293084204197,
      "epoch": 0.1140701486504693,
      "grad_norm": 7.6875,
      "learning_rate": 4.88281656180319e-05,
      "loss": 1.0487909317016602,
      "mean_token_accuracy": 0.7557375371456146,
      "num_tokens": 12617492.0,
      "step": 6350
    },
    {
      "entropy": 0.9186167255043983,
      "epoch": 0.11424978667984012,
      "grad_norm": 6.625,
      "learning_rate": 4.882381473008071e-05,
      "loss": 0.9821917533874511,
      "mean_token_accuracy": 0.7730137690901756,
      "num_tokens": 12636274.0,
      "step": 6360
    },
    {
      "entropy": 1.042784681916237,
      "epoch": 0.11442942470921094,
      "grad_norm": 7.46875,
      "learning_rate": 4.881945597448617e-05,
      "loss": 1.0676585197448731,
      "mean_token_accuracy": 0.7516802608966827,
      "num_tokens": 12655437.0,
      "step": 6370
    },
    {
      "entropy": 1.065847682952881,
      "epoch": 0.11460906273858176,
      "grad_norm": 6.625,
      "learning_rate": 4.881508935268772e-05,
      "loss": 1.055802822113037,
      "mean_token_accuracy": 0.7571970328688622,
      "num_tokens": 12675843.0,
      "step": 6380
    },
    {
      "entropy": 1.0163489624857902,
      "epoch": 0.11478870076795257,
      "grad_norm": 6.90625,
      "learning_rate": 4.8810714866127396e-05,
      "loss": 1.0799942016601562,
      "mean_token_accuracy": 0.7535074308514595,
      "num_tokens": 12696092.0,
      "step": 6390
    },
    {
      "entropy": 1.010008205473423,
      "epoch": 0.11496833879732339,
      "grad_norm": 5.5,
      "learning_rate": 4.880633251624986e-05,
      "loss": 1.0491214752197267,
      "mean_token_accuracy": 0.7569492131471633,
      "num_tokens": 12715034.0,
      "step": 6400
    },
    {
      "entropy": 1.0110047802329063,
      "epoch": 0.11514797682669421,
      "grad_norm": 6.8125,
      "learning_rate": 4.880194230450233e-05,
      "loss": 1.0598729133605957,
      "mean_token_accuracy": 0.7513307005167007,
      "num_tokens": 12735784.0,
      "step": 6410
    },
    {
      "entropy": 1.0210563033819198,
      "epoch": 0.11532761485606503,
      "grad_norm": 7.375,
      "learning_rate": 4.879754423233467e-05,
      "loss": 1.0464460372924804,
      "mean_token_accuracy": 0.7693593099713325,
      "num_tokens": 12754660.0,
      "step": 6420
    },
    {
      "entropy": 1.105432777106762,
      "epoch": 0.11550725288543584,
      "grad_norm": 8.8125,
      "learning_rate": 4.879313830119928e-05,
      "loss": 1.0998011589050294,
      "mean_token_accuracy": 0.746526052057743,
      "num_tokens": 12774946.0,
      "step": 6430
    },
    {
      "entropy": 0.9488676428794861,
      "epoch": 0.11568689091480666,
      "grad_norm": 6.8125,
      "learning_rate": 4.878872451255121e-05,
      "loss": 0.9547931671142578,
      "mean_token_accuracy": 0.7704014152288436,
      "num_tokens": 12794403.0,
      "step": 6440
    },
    {
      "entropy": 0.9039615742862225,
      "epoch": 0.11586652894417748,
      "grad_norm": 7.78125,
      "learning_rate": 4.8784302867848076e-05,
      "loss": 0.9366701126098633,
      "mean_token_accuracy": 0.7851650446653367,
      "num_tokens": 12815134.0,
      "step": 6450
    },
    {
      "entropy": 0.9854362159967422,
      "epoch": 0.1160461669735483,
      "grad_norm": 6.875,
      "learning_rate": 4.877987336855009e-05,
      "loss": 1.0387304306030274,
      "mean_token_accuracy": 0.7558018639683723,
      "num_tokens": 12835190.0,
      "step": 6460
    },
    {
      "entropy": 1.1121864467859268,
      "epoch": 0.11622580500291912,
      "grad_norm": 6.0625,
      "learning_rate": 4.8775436016120066e-05,
      "loss": 1.2135540008544923,
      "mean_token_accuracy": 0.7309125058352948,
      "num_tokens": 12854620.0,
      "step": 6470
    },
    {
      "entropy": 1.0853245466947556,
      "epoch": 0.11640544303228993,
      "grad_norm": 8.5625,
      "learning_rate": 4.877099081202341e-05,
      "loss": 1.0786243438720704,
      "mean_token_accuracy": 0.7556031823158265,
      "num_tokens": 12874317.0,
      "step": 6480
    },
    {
      "entropy": 0.9458428382873535,
      "epoch": 0.11658508106166075,
      "grad_norm": 5.21875,
      "learning_rate": 4.876653775772811e-05,
      "loss": 1.0157344818115235,
      "mean_token_accuracy": 0.7615962594747543,
      "num_tokens": 12894607.0,
      "step": 6490
    },
    {
      "entropy": 1.124277839064598,
      "epoch": 0.11676471909103157,
      "grad_norm": 6.4375,
      "learning_rate": 4.876207685470476e-05,
      "loss": 1.1769062995910644,
      "mean_token_accuracy": 0.7311655737459659,
      "num_tokens": 12915612.0,
      "step": 6500
    },
    {
      "entropy": 1.039088200777769,
      "epoch": 0.1169443571204024,
      "grad_norm": 7.53125,
      "learning_rate": 4.875760810442654e-05,
      "loss": 1.036177635192871,
      "mean_token_accuracy": 0.7529889523983002,
      "num_tokens": 12934003.0,
      "step": 6510
    },
    {
      "entropy": 0.9423630580306053,
      "epoch": 0.1171239951497732,
      "grad_norm": 7.78125,
      "learning_rate": 4.875313150836923e-05,
      "loss": 0.9441682815551757,
      "mean_token_accuracy": 0.7680635437369346,
      "num_tokens": 12953275.0,
      "step": 6520
    },
    {
      "entropy": 1.1086465388536453,
      "epoch": 0.11730363317914402,
      "grad_norm": 7.15625,
      "learning_rate": 4.874864706801119e-05,
      "loss": 1.1593046188354492,
      "mean_token_accuracy": 0.7296991139650345,
      "num_tokens": 12971672.0,
      "step": 6530
    },
    {
      "entropy": 1.0537022650241852,
      "epoch": 0.11748327120851484,
      "grad_norm": 7.5625,
      "learning_rate": 4.874415478483336e-05,
      "loss": 1.1074079513549804,
      "mean_token_accuracy": 0.7402363657951355,
      "num_tokens": 12992709.0,
      "step": 6540
    },
    {
      "entropy": 0.9904323205351829,
      "epoch": 0.11766290923788567,
      "grad_norm": 9.0625,
      "learning_rate": 4.873965466031931e-05,
      "loss": 1.0011497497558595,
      "mean_token_accuracy": 0.7709414944052696,
      "num_tokens": 13012268.0,
      "step": 6550
    },
    {
      "entropy": 1.0350611910223961,
      "epoch": 0.11784254726725647,
      "grad_norm": 8.25,
      "learning_rate": 4.8735146695955154e-05,
      "loss": 1.0971911430358887,
      "mean_token_accuracy": 0.7489110670983792,
      "num_tokens": 13032124.0,
      "step": 6560
    },
    {
      "entropy": 0.9896834671497345,
      "epoch": 0.1180221852966273,
      "grad_norm": 6.53125,
      "learning_rate": 4.873063089322962e-05,
      "loss": 0.9778719902038574,
      "mean_token_accuracy": 0.7633831724524498,
      "num_tokens": 13051789.0,
      "step": 6570
    },
    {
      "entropy": 1.069327637553215,
      "epoch": 0.11820182332599811,
      "grad_norm": 6.28125,
      "learning_rate": 4.872610725363402e-05,
      "loss": 1.0697262763977051,
      "mean_token_accuracy": 0.7499631449580193,
      "num_tokens": 13071140.0,
      "step": 6580
    },
    {
      "entropy": 0.9656549043953419,
      "epoch": 0.11838146135536894,
      "grad_norm": 6.84375,
      "learning_rate": 4.872157577866225e-05,
      "loss": 1.0009926795959472,
      "mean_token_accuracy": 0.7680483996868134,
      "num_tokens": 13091309.0,
      "step": 6590
    },
    {
      "entropy": 0.9845070227980613,
      "epoch": 0.11856109938473974,
      "grad_norm": 7.28125,
      "learning_rate": 4.871703646981081e-05,
      "loss": 1.0317476272583008,
      "mean_token_accuracy": 0.7533792406320572,
      "num_tokens": 13111371.0,
      "step": 6600
    },
    {
      "entropy": 1.0421454735100268,
      "epoch": 0.11874073741411056,
      "grad_norm": 6.3125,
      "learning_rate": 4.871248932857876e-05,
      "loss": 1.0890267372131348,
      "mean_token_accuracy": 0.7409923642873764,
      "num_tokens": 13130657.0,
      "step": 6610
    },
    {
      "entropy": 0.9024669304490089,
      "epoch": 0.11892037544348139,
      "grad_norm": 7.09375,
      "learning_rate": 4.870793435646776e-05,
      "loss": 0.9276371955871582,
      "mean_token_accuracy": 0.7823088929057121,
      "num_tokens": 13149764.0,
      "step": 6620
    },
    {
      "entropy": 0.9360634863376618,
      "epoch": 0.1191000134728522,
      "grad_norm": 5.8125,
      "learning_rate": 4.870337155498207e-05,
      "loss": 0.9467898368835449,
      "mean_token_accuracy": 0.7801152005791664,
      "num_tokens": 13169107.0,
      "step": 6630
    },
    {
      "entropy": 1.02987220287323,
      "epoch": 0.11927965150222303,
      "grad_norm": 6.375,
      "learning_rate": 4.8698800925628506e-05,
      "loss": 1.084320068359375,
      "mean_token_accuracy": 0.7482532218098641,
      "num_tokens": 13188581.0,
      "step": 6640
    },
    {
      "entropy": 1.029680258780718,
      "epoch": 0.11945928953159383,
      "grad_norm": 6.0,
      "learning_rate": 4.8694222469916504e-05,
      "loss": 1.0989627838134766,
      "mean_token_accuracy": 0.7460219338536263,
      "num_tokens": 13208346.0,
      "step": 6650
    },
    {
      "entropy": 1.0734458602964878,
      "epoch": 0.11963892756096466,
      "grad_norm": 8.1875,
      "learning_rate": 4.8689636189358045e-05,
      "loss": 1.0589815139770509,
      "mean_token_accuracy": 0.7575636148452759,
      "num_tokens": 13227406.0,
      "step": 6660
    },
    {
      "entropy": 0.9638423591852188,
      "epoch": 0.11981856559033548,
      "grad_norm": 5.6875,
      "learning_rate": 4.868504208546774e-05,
      "loss": 0.9934786796569824,
      "mean_token_accuracy": 0.7669479727745057,
      "num_tokens": 13247967.0,
      "step": 6670
    },
    {
      "entropy": 1.0991191014647483,
      "epoch": 0.1199982036197063,
      "grad_norm": 6.34375,
      "learning_rate": 4.868044015976273e-05,
      "loss": 1.1577638626098632,
      "mean_token_accuracy": 0.7318303808569908,
      "num_tokens": 13268459.0,
      "step": 6680
    },
    {
      "entropy": 0.9101608723402024,
      "epoch": 0.1201778416490771,
      "grad_norm": 7.40625,
      "learning_rate": 4.867583041376278e-05,
      "loss": 0.9195585250854492,
      "mean_token_accuracy": 0.7837798535823822,
      "num_tokens": 13288414.0,
      "step": 6690
    },
    {
      "entropy": 1.065487912297249,
      "epoch": 0.12035747967844793,
      "grad_norm": 6.1875,
      "learning_rate": 4.8671212848990246e-05,
      "loss": 1.1378360748291017,
      "mean_token_accuracy": 0.7402533307671547,
      "num_tokens": 13309189.0,
      "step": 6700
    },
    {
      "entropy": 1.0643005684018134,
      "epoch": 0.12053711770781875,
      "grad_norm": 7.96875,
      "learning_rate": 4.866658746697001e-05,
      "loss": 1.097043800354004,
      "mean_token_accuracy": 0.7491765782237053,
      "num_tokens": 13329882.0,
      "step": 6710
    },
    {
      "entropy": 0.9805642001330852,
      "epoch": 0.12071675573718957,
      "grad_norm": 7.34375,
      "learning_rate": 4.8661954269229595e-05,
      "loss": 0.9970211029052735,
      "mean_token_accuracy": 0.7660590589046479,
      "num_tokens": 13349183.0,
      "step": 6720
    },
    {
      "entropy": 0.9814873158931732,
      "epoch": 0.12089639376656038,
      "grad_norm": 6.78125,
      "learning_rate": 4.865731325729908e-05,
      "loss": 1.0121488571166992,
      "mean_token_accuracy": 0.7641021609306335,
      "num_tokens": 13368802.0,
      "step": 6730
    },
    {
      "entropy": 0.9894362859427929,
      "epoch": 0.1210760317959312,
      "grad_norm": 6.5625,
      "learning_rate": 4.8652664432711115e-05,
      "loss": 1.009317684173584,
      "mean_token_accuracy": 0.7588550344109535,
      "num_tokens": 13387113.0,
      "step": 6740
    },
    {
      "entropy": 1.0440159760415555,
      "epoch": 0.12125566982530202,
      "grad_norm": 6.09375,
      "learning_rate": 4.864800779700095e-05,
      "loss": 1.0911325454711913,
      "mean_token_accuracy": 0.7527034327387809,
      "num_tokens": 13406835.0,
      "step": 6750
    },
    {
      "entropy": 1.0030406430363654,
      "epoch": 0.12143530785467284,
      "grad_norm": 8.3125,
      "learning_rate": 4.86433433517064e-05,
      "loss": 0.9971686363220215,
      "mean_token_accuracy": 0.7668825313448906,
      "num_tokens": 13426869.0,
      "step": 6760
    },
    {
      "entropy": 0.9166123487055302,
      "epoch": 0.12161494588404365,
      "grad_norm": 7.03125,
      "learning_rate": 4.863867109836787e-05,
      "loss": 0.9854426383972168,
      "mean_token_accuracy": 0.7727517142891884,
      "num_tokens": 13446503.0,
      "step": 6770
    },
    {
      "entropy": 0.9390500813722611,
      "epoch": 0.12179458391341447,
      "grad_norm": 6.25,
      "learning_rate": 4.863399103852833e-05,
      "loss": 0.9305182456970215,
      "mean_token_accuracy": 0.7743746340274811,
      "num_tokens": 13465433.0,
      "step": 6780
    },
    {
      "entropy": 1.1078578799962997,
      "epoch": 0.12197422194278529,
      "grad_norm": 7.4375,
      "learning_rate": 4.862930317373334e-05,
      "loss": 1.1838531494140625,
      "mean_token_accuracy": 0.7291121244430542,
      "num_tokens": 13484665.0,
      "step": 6790
    },
    {
      "entropy": 0.9251707255840301,
      "epoch": 0.12215385997215611,
      "grad_norm": 6.4375,
      "learning_rate": 4.8624607505531036e-05,
      "loss": 0.9616168975830078,
      "mean_token_accuracy": 0.7723833799362183,
      "num_tokens": 13504329.0,
      "step": 6800
    },
    {
      "entropy": 0.9434578157961369,
      "epoch": 0.12233349800152692,
      "grad_norm": 5.9375,
      "learning_rate": 4.861990403547213e-05,
      "loss": 0.9922041893005371,
      "mean_token_accuracy": 0.7682307422161102,
      "num_tokens": 13525149.0,
      "step": 6810
    },
    {
      "entropy": 1.00031920671463,
      "epoch": 0.12251313603089774,
      "grad_norm": 7.1875,
      "learning_rate": 4.86151927651099e-05,
      "loss": 1.013861083984375,
      "mean_token_accuracy": 0.758407998085022,
      "num_tokens": 13544451.0,
      "step": 6820
    },
    {
      "entropy": 1.1202403873205184,
      "epoch": 0.12269277406026856,
      "grad_norm": 4.8125,
      "learning_rate": 4.8610473696000214e-05,
      "loss": 1.1820843696594239,
      "mean_token_accuracy": 0.7329499587416649,
      "num_tokens": 13564143.0,
      "step": 6830
    },
    {
      "entropy": 1.0417736902832986,
      "epoch": 0.12287241208963938,
      "grad_norm": 5.71875,
      "learning_rate": 4.8605746829701515e-05,
      "loss": 0.9679837226867676,
      "mean_token_accuracy": 0.7726904019713402,
      "num_tokens": 13584819.0,
      "step": 6840
    },
    {
      "entropy": 1.0067878305912017,
      "epoch": 0.1230520501190102,
      "grad_norm": 6.40625,
      "learning_rate": 4.8601012167774815e-05,
      "loss": 1.0070413589477538,
      "mean_token_accuracy": 0.7584892511367798,
      "num_tokens": 13603773.0,
      "step": 6850
    },
    {
      "entropy": 0.9952953092753887,
      "epoch": 0.12323168814838101,
      "grad_norm": 6.25,
      "learning_rate": 4.8596269711783696e-05,
      "loss": 1.1002662658691407,
      "mean_token_accuracy": 0.7561241887509823,
      "num_tokens": 13623456.0,
      "step": 6860
    },
    {
      "entropy": 1.003292979300022,
      "epoch": 0.12341132617775183,
      "grad_norm": 7.1875,
      "learning_rate": 4.8591519463294336e-05,
      "loss": 0.9929161071777344,
      "mean_token_accuracy": 0.7706490695476532,
      "num_tokens": 13643957.0,
      "step": 6870
    },
    {
      "entropy": 1.0403919465839864,
      "epoch": 0.12359096420712265,
      "grad_norm": 10.0,
      "learning_rate": 4.8586761423875447e-05,
      "loss": 1.1246644020080567,
      "mean_token_accuracy": 0.742063120007515,
      "num_tokens": 13663613.0,
      "step": 6880
    },
    {
      "entropy": 0.9565848641097545,
      "epoch": 0.12377060223649347,
      "grad_norm": 5.78125,
      "learning_rate": 4.858199559509835e-05,
      "loss": 0.9863670349121094,
      "mean_token_accuracy": 0.7635638147592545,
      "num_tokens": 13682920.0,
      "step": 6890
    },
    {
      "entropy": 0.9593957096338273,
      "epoch": 0.12395024026586428,
      "grad_norm": 7.25,
      "learning_rate": 4.8577221978536924e-05,
      "loss": 0.9902218818664551,
      "mean_token_accuracy": 0.7598926559090614,
      "num_tokens": 13701810.0,
      "step": 6900
    },
    {
      "entropy": 0.9058008939027786,
      "epoch": 0.1241298782952351,
      "grad_norm": 6.125,
      "learning_rate": 4.857244057576762e-05,
      "loss": 0.924420166015625,
      "mean_token_accuracy": 0.7777085155248642,
      "num_tokens": 13720738.0,
      "step": 6910
    },
    {
      "entropy": 0.9216203063726425,
      "epoch": 0.12430951632460592,
      "grad_norm": 6.5,
      "learning_rate": 4.856765138836947e-05,
      "loss": 0.9765249252319336,
      "mean_token_accuracy": 0.7725277796387673,
      "num_tokens": 13741224.0,
      "step": 6920
    },
    {
      "entropy": 0.9464600130915641,
      "epoch": 0.12448915435397674,
      "grad_norm": 5.4375,
      "learning_rate": 4.856285441792407e-05,
      "loss": 0.9995107650756836,
      "mean_token_accuracy": 0.7607283592224121,
      "num_tokens": 13760311.0,
      "step": 6930
    },
    {
      "entropy": 1.020284004509449,
      "epoch": 0.12466879238334755,
      "grad_norm": 4.40625,
      "learning_rate": 4.855804966601556e-05,
      "loss": 1.0106845855712892,
      "mean_token_accuracy": 0.7653924763202667,
      "num_tokens": 13780215.0,
      "step": 6940
    },
    {
      "entropy": 1.0780873231589794,
      "epoch": 0.12484843041271837,
      "grad_norm": 6.125,
      "learning_rate": 4.85532371342307e-05,
      "loss": 1.1189627647399902,
      "mean_token_accuracy": 0.7504855573177338,
      "num_tokens": 13800135.0,
      "step": 6950
    },
    {
      "entropy": 0.9893208250403405,
      "epoch": 0.1250280684420892,
      "grad_norm": 5.75,
      "learning_rate": 4.854841682415879e-05,
      "loss": 0.9784346580505371,
      "mean_token_accuracy": 0.7669732868671417,
      "num_tokens": 13820300.0,
      "step": 6960
    },
    {
      "entropy": 0.9498419493436814,
      "epoch": 0.12520770647146,
      "grad_norm": 5.8125,
      "learning_rate": 4.8543588737391696e-05,
      "loss": 0.9837165832519531,
      "mean_token_accuracy": 0.7657981343567372,
      "num_tokens": 13839598.0,
      "step": 6970
    },
    {
      "entropy": 0.8783250324428081,
      "epoch": 0.12538734450083083,
      "grad_norm": 5.65625,
      "learning_rate": 4.853875287552386e-05,
      "loss": 0.9952584266662597,
      "mean_token_accuracy": 0.774166339635849,
      "num_tokens": 13859574.0,
      "step": 6980
    },
    {
      "entropy": 1.0732742726802826,
      "epoch": 0.12556698253020165,
      "grad_norm": 8.5625,
      "learning_rate": 4.85339092401523e-05,
      "loss": 1.19972505569458,
      "mean_token_accuracy": 0.7316268265247345,
      "num_tokens": 13879974.0,
      "step": 6990
    },
    {
      "epoch": 0.12574662055957248,
      "eval_entropy": 1.0539333384037017,
      "eval_loss": 1.0188111066818237,
      "eval_mean_token_accuracy": 0.7641782150268555,
      "eval_num_tokens": 13900457.0,
      "eval_runtime": 40.4836,
      "eval_samples_per_second": 49.403,
      "eval_steps_per_second": 6.175,
      "step": 7000
    },
    {
      "entropy": 1.0704687289893626,
      "epoch": 0.12592625858894327,
      "grad_norm": 7.84375,
      "learning_rate": 4.8524198655298833e-05,
      "loss": 1.0548014640808105,
      "mean_token_accuracy": 0.7501308396458626,
      "num_tokens": 13919660.0,
      "step": 7010
    },
    {
      "entropy": 1.0547692447900772,
      "epoch": 0.1261058966183141,
      "grad_norm": 8.3125,
      "learning_rate": 4.8519331709023795e-05,
      "loss": 1.081246280670166,
      "mean_token_accuracy": 0.7473935708403587,
      "num_tokens": 13937628.0,
      "step": 7020
    },
    {
      "entropy": 0.9390965379774571,
      "epoch": 0.1262855346476849,
      "grad_norm": 8.4375,
      "learning_rate": 4.8514456995658726e-05,
      "loss": 0.9572797775268554,
      "mean_token_accuracy": 0.7749572098255157,
      "num_tokens": 13956825.0,
      "step": 7030
    },
    {
      "entropy": 0.9652213841676712,
      "epoch": 0.12646517267705573,
      "grad_norm": 7.28125,
      "learning_rate": 4.850957451681346e-05,
      "loss": 0.9720582962036133,
      "mean_token_accuracy": 0.7662696912884712,
      "num_tokens": 13976916.0,
      "step": 7040
    },
    {
      "entropy": 0.8637428626418113,
      "epoch": 0.12664481070642655,
      "grad_norm": 6.03125,
      "learning_rate": 4.850468427410041e-05,
      "loss": 0.9257391929626465,
      "mean_token_accuracy": 0.7808357328176498,
      "num_tokens": 13995934.0,
      "step": 7050
    },
    {
      "entropy": 0.9526255652308464,
      "epoch": 0.12682444873579737,
      "grad_norm": 7.40625,
      "learning_rate": 4.849978626913454e-05,
      "loss": 0.9906145095825195,
      "mean_token_accuracy": 0.7689179390668869,
      "num_tokens": 14015304.0,
      "step": 7060
    },
    {
      "entropy": 0.9305871874094009,
      "epoch": 0.1270040867651682,
      "grad_norm": 6.71875,
      "learning_rate": 4.849488050353337e-05,
      "loss": 0.9351966857910157,
      "mean_token_accuracy": 0.7824089989066124,
      "num_tokens": 14034097.0,
      "step": 7070
    },
    {
      "entropy": 0.9513274013996125,
      "epoch": 0.12718372479453902,
      "grad_norm": 4.65625,
      "learning_rate": 4.848996697891701e-05,
      "loss": 1.0010278701782227,
      "mean_token_accuracy": 0.7751732021570206,
      "num_tokens": 14054396.0,
      "step": 7080
    },
    {
      "entropy": 1.0039676330983638,
      "epoch": 0.1273633628239098,
      "grad_norm": 6.40625,
      "learning_rate": 4.848504569690812e-05,
      "loss": 1.0376245498657226,
      "mean_token_accuracy": 0.7558190628886223,
      "num_tokens": 14073554.0,
      "step": 7090
    },
    {
      "entropy": 1.0095471292734146,
      "epoch": 0.12754300085328063,
      "grad_norm": 7.5,
      "learning_rate": 4.848011665913189e-05,
      "loss": 1.0702735900878906,
      "mean_token_accuracy": 0.7530047968029976,
      "num_tokens": 14093419.0,
      "step": 7100
    },
    {
      "entropy": 0.9995038047432899,
      "epoch": 0.12772263888265145,
      "grad_norm": 6.0,
      "learning_rate": 4.8475179867216125e-05,
      "loss": 1.063162136077881,
      "mean_token_accuracy": 0.7515974521636963,
      "num_tokens": 14113055.0,
      "step": 7110
    },
    {
      "entropy": 1.0539027094841003,
      "epoch": 0.12790227691202227,
      "grad_norm": 6.0,
      "learning_rate": 4.8470235322791156e-05,
      "loss": 1.0371717453002929,
      "mean_token_accuracy": 0.7615969255566597,
      "num_tokens": 14132836.0,
      "step": 7120
    },
    {
      "entropy": 0.9602646186947823,
      "epoch": 0.1280819149413931,
      "grad_norm": 6.03125,
      "learning_rate": 4.8465283027489884e-05,
      "loss": 0.9974834442138671,
      "mean_token_accuracy": 0.7661203786730766,
      "num_tokens": 14152136.0,
      "step": 7130
    },
    {
      "entropy": 1.047901628166437,
      "epoch": 0.12826155297076391,
      "grad_norm": 7.6875,
      "learning_rate": 4.846032298294777e-05,
      "loss": 1.0882431030273438,
      "mean_token_accuracy": 0.7494831725955009,
      "num_tokens": 14171540.0,
      "step": 7140
    },
    {
      "entropy": 1.0146586880087853,
      "epoch": 0.12844119100013474,
      "grad_norm": 6.25,
      "learning_rate": 4.845535519080283e-05,
      "loss": 1.0395357131958007,
      "mean_token_accuracy": 0.753221869468689,
      "num_tokens": 14190585.0,
      "step": 7150
    },
    {
      "entropy": 0.9759780965745449,
      "epoch": 0.12862082902950556,
      "grad_norm": 6.375,
      "learning_rate": 4.8450379652695646e-05,
      "loss": 0.9921046257019043,
      "mean_token_accuracy": 0.7688451960682869,
      "num_tokens": 14209903.0,
      "step": 7160
    },
    {
      "entropy": 0.9458831459283829,
      "epoch": 0.12880046705887638,
      "grad_norm": 8.0625,
      "learning_rate": 4.844539637026935e-05,
      "loss": 1.0311005592346192,
      "mean_token_accuracy": 0.768575819581747,
      "num_tokens": 14229490.0,
      "step": 7170
    },
    {
      "entropy": 1.0068778857588767,
      "epoch": 0.12898010508824717,
      "grad_norm": 7.84375,
      "learning_rate": 4.844040534516964e-05,
      "loss": 1.0468737602233886,
      "mean_token_accuracy": 0.7575206249952317,
      "num_tokens": 14248452.0,
      "step": 7180
    },
    {
      "entropy": 1.0791752129793166,
      "epoch": 0.129159743117618,
      "grad_norm": 7.125,
      "learning_rate": 4.8435406579044764e-05,
      "loss": 1.075886344909668,
      "mean_token_accuracy": 0.7474431961774826,
      "num_tokens": 14268358.0,
      "step": 7190
    },
    {
      "entropy": 1.0529504910111427,
      "epoch": 0.1293393811469888,
      "grad_norm": 5.625,
      "learning_rate": 4.8430400073545516e-05,
      "loss": 1.0952861785888672,
      "mean_token_accuracy": 0.7439638927578927,
      "num_tokens": 14288465.0,
      "step": 7200
    },
    {
      "entropy": 1.0603987380862236,
      "epoch": 0.12951901917635963,
      "grad_norm": 5.40625,
      "learning_rate": 4.8425385830325274e-05,
      "loss": 1.056072998046875,
      "mean_token_accuracy": 0.7510070040822029,
      "num_tokens": 14308829.0,
      "step": 7210
    },
    {
      "entropy": 1.0066994294524192,
      "epoch": 0.12969865720573046,
      "grad_norm": 5.0,
      "learning_rate": 4.8420363851039955e-05,
      "loss": 1.0044511795043944,
      "mean_token_accuracy": 0.7635873422026634,
      "num_tokens": 14329682.0,
      "step": 7220
    },
    {
      "entropy": 1.0032708846032619,
      "epoch": 0.12987829523510128,
      "grad_norm": 7.15625,
      "learning_rate": 4.841533413734803e-05,
      "loss": 1.0734521865844726,
      "mean_token_accuracy": 0.7625238470733166,
      "num_tokens": 14348288.0,
      "step": 7230
    },
    {
      "entropy": 1.038369619846344,
      "epoch": 0.1300579332644721,
      "grad_norm": 5.9375,
      "learning_rate": 4.841029669091052e-05,
      "loss": 1.0857677459716797,
      "mean_token_accuracy": 0.7477707870304584,
      "num_tokens": 14368279.0,
      "step": 7240
    },
    {
      "entropy": 1.0475360125303268,
      "epoch": 0.13023757129384292,
      "grad_norm": 6.3125,
      "learning_rate": 4.8405251513391e-05,
      "loss": 1.1035821914672852,
      "mean_token_accuracy": 0.7405801579356194,
      "num_tokens": 14387685.0,
      "step": 7250
    },
    {
      "entropy": 1.0898923024535179,
      "epoch": 0.1304172093232137,
      "grad_norm": 5.0625,
      "learning_rate": 4.8400198606455626e-05,
      "loss": 1.085732364654541,
      "mean_token_accuracy": 0.7459083139896393,
      "num_tokens": 14407946.0,
      "step": 7260
    },
    {
      "entropy": 1.038026401400566,
      "epoch": 0.13059684735258453,
      "grad_norm": 4.75,
      "learning_rate": 4.839513797177307e-05,
      "loss": 1.048984909057617,
      "mean_token_accuracy": 0.7518763616681099,
      "num_tokens": 14427812.0,
      "step": 7270
    },
    {
      "entropy": 1.0761775322258473,
      "epoch": 0.13077648538195535,
      "grad_norm": 5.90625,
      "learning_rate": 4.839006961101457e-05,
      "loss": 1.0780851364135742,
      "mean_token_accuracy": 0.7360206946730614,
      "num_tokens": 14446940.0,
      "step": 7280
    },
    {
      "entropy": 0.9662975825369358,
      "epoch": 0.13095612341132618,
      "grad_norm": 8.3125,
      "learning_rate": 4.83849935258539e-05,
      "loss": 1.0302798271179199,
      "mean_token_accuracy": 0.764030534029007,
      "num_tokens": 14466639.0,
      "step": 7290
    },
    {
      "entropy": 1.0169863127171994,
      "epoch": 0.131135761440697,
      "grad_norm": 6.09375,
      "learning_rate": 4.837990971796744e-05,
      "loss": 1.031626319885254,
      "mean_token_accuracy": 0.7589772090315818,
      "num_tokens": 14486243.0,
      "step": 7300
    },
    {
      "entropy": 0.9524190604686738,
      "epoch": 0.13131539947006782,
      "grad_norm": 7.21875,
      "learning_rate": 4.837481818903404e-05,
      "loss": 1.017763328552246,
      "mean_token_accuracy": 0.7593987986445427,
      "num_tokens": 14506532.0,
      "step": 7310
    },
    {
      "entropy": 1.0019267067313193,
      "epoch": 0.13149503749943864,
      "grad_norm": 6.46875,
      "learning_rate": 4.8369718940735176e-05,
      "loss": 1.0181355476379395,
      "mean_token_accuracy": 0.7598321154713631,
      "num_tokens": 14527100.0,
      "step": 7320
    },
    {
      "entropy": 0.9895594835281372,
      "epoch": 0.13167467552880946,
      "grad_norm": 7.125,
      "learning_rate": 4.836461197475481e-05,
      "loss": 0.9963333129882812,
      "mean_token_accuracy": 0.7660549744963646,
      "num_tokens": 14546685.0,
      "step": 7330
    },
    {
      "entropy": 0.876677007228136,
      "epoch": 0.13185431355818028,
      "grad_norm": 5.875,
      "learning_rate": 4.8359497292779495e-05,
      "loss": 0.9290422439575196,
      "mean_token_accuracy": 0.7822725132107735,
      "num_tokens": 14566923.0,
      "step": 7340
    },
    {
      "entropy": 0.8709055915474891,
      "epoch": 0.13203395158755107,
      "grad_norm": 6.5,
      "learning_rate": 4.835437489649831e-05,
      "loss": 0.8853213310241699,
      "mean_token_accuracy": 0.7778545469045639,
      "num_tokens": 14586269.0,
      "step": 7350
    },
    {
      "entropy": 0.922557246685028,
      "epoch": 0.1322135896169219,
      "grad_norm": 7.03125,
      "learning_rate": 4.83492447876029e-05,
      "loss": 1.0160842895507813,
      "mean_token_accuracy": 0.769161856174469,
      "num_tokens": 14606334.0,
      "step": 7360
    },
    {
      "entropy": 1.0463509738445282,
      "epoch": 0.13239322764629272,
      "grad_norm": 4.8125,
      "learning_rate": 4.834410696778744e-05,
      "loss": 1.0613508224487305,
      "mean_token_accuracy": 0.7502421051263809,
      "num_tokens": 14626897.0,
      "step": 7370
    },
    {
      "entropy": 0.9834346681833267,
      "epoch": 0.13257286567566354,
      "grad_norm": 7.15625,
      "learning_rate": 4.833896143874865e-05,
      "loss": 0.9475133895874024,
      "mean_token_accuracy": 0.7763537585735321,
      "num_tokens": 14647276.0,
      "step": 7380
    },
    {
      "entropy": 0.956779134273529,
      "epoch": 0.13275250370503436,
      "grad_norm": 7.3125,
      "learning_rate": 4.8333808202185826e-05,
      "loss": 1.0467992782592774,
      "mean_token_accuracy": 0.7568529158830642,
      "num_tokens": 14666665.0,
      "step": 7390
    },
    {
      "entropy": 0.9405389867722989,
      "epoch": 0.13293214173440518,
      "grad_norm": 6.59375,
      "learning_rate": 4.8328647259800765e-05,
      "loss": 0.9760592460632325,
      "mean_token_accuracy": 0.7664414137601853,
      "num_tokens": 14686963.0,
      "step": 7400
    },
    {
      "entropy": 0.9767249166965485,
      "epoch": 0.133111779763776,
      "grad_norm": 6.21875,
      "learning_rate": 4.832347861329784e-05,
      "loss": 0.9937135696411132,
      "mean_token_accuracy": 0.7680794999003411,
      "num_tokens": 14705647.0,
      "step": 7410
    },
    {
      "entropy": 0.8972371503710747,
      "epoch": 0.13329141779314682,
      "grad_norm": 5.8125,
      "learning_rate": 4.831830226438396e-05,
      "loss": 0.9153555870056153,
      "mean_token_accuracy": 0.7840756922960281,
      "num_tokens": 14725781.0,
      "step": 7420
    },
    {
      "entropy": 1.0124459140002728,
      "epoch": 0.13347105582251761,
      "grad_norm": 5.65625,
      "learning_rate": 4.8313118214768584e-05,
      "loss": 1.027753734588623,
      "mean_token_accuracy": 0.7593206621706485,
      "num_tokens": 14745758.0,
      "step": 7430
    },
    {
      "entropy": 1.0883337348699569,
      "epoch": 0.13365069385188844,
      "grad_norm": 7.09375,
      "learning_rate": 4.830792646616369e-05,
      "loss": 1.1550491333007813,
      "mean_token_accuracy": 0.7440998949110508,
      "num_tokens": 14766717.0,
      "step": 7440
    },
    {
      "entropy": 0.9820266768336297,
      "epoch": 0.13383033188125926,
      "grad_norm": 10.5625,
      "learning_rate": 4.8302727020283825e-05,
      "loss": 1.018569564819336,
      "mean_token_accuracy": 0.7588348686695099,
      "num_tokens": 14786376.0,
      "step": 7450
    },
    {
      "entropy": 0.9752943694591523,
      "epoch": 0.13400996991063008,
      "grad_norm": 8.4375,
      "learning_rate": 4.8297519878846077e-05,
      "loss": 1.057596206665039,
      "mean_token_accuracy": 0.7617123052477837,
      "num_tokens": 14806150.0,
      "step": 7460
    },
    {
      "entropy": 0.9267949141561985,
      "epoch": 0.1341896079400009,
      "grad_norm": 6.53125,
      "learning_rate": 4.829230504357005e-05,
      "loss": 0.93712158203125,
      "mean_token_accuracy": 0.7776257634162903,
      "num_tokens": 14825824.0,
      "step": 7470
    },
    {
      "entropy": 1.0339096680283546,
      "epoch": 0.13436924596937172,
      "grad_norm": 5.96875,
      "learning_rate": 4.828708251617792e-05,
      "loss": 1.0638809204101562,
      "mean_token_accuracy": 0.7497575655579567,
      "num_tokens": 14846877.0,
      "step": 7480
    },
    {
      "entropy": 0.9317882686853409,
      "epoch": 0.13454888399874254,
      "grad_norm": 5.5625,
      "learning_rate": 4.8281852298394384e-05,
      "loss": 0.9406015396118164,
      "mean_token_accuracy": 0.7742516383528709,
      "num_tokens": 14866642.0,
      "step": 7490
    },
    {
      "entropy": 0.9686097353696823,
      "epoch": 0.13472852202811336,
      "grad_norm": 5.125,
      "learning_rate": 4.827661439194667e-05,
      "loss": 0.9904957771301269,
      "mean_token_accuracy": 0.7631622701883316,
      "num_tokens": 14886886.0,
      "step": 7500
    },
    {
      "entropy": 0.9410270944237709,
      "epoch": 0.13490816005748418,
      "grad_norm": 5.96875,
      "learning_rate": 4.827136879856459e-05,
      "loss": 0.9413405418395996,
      "mean_token_accuracy": 0.776588824391365,
      "num_tokens": 14905655.0,
      "step": 7510
    },
    {
      "entropy": 0.9666697688400745,
      "epoch": 0.13508779808685498,
      "grad_norm": 5.46875,
      "learning_rate": 4.826611551998044e-05,
      "loss": 1.0359756469726562,
      "mean_token_accuracy": 0.757803750783205,
      "num_tokens": 14926295.0,
      "step": 7520
    },
    {
      "entropy": 1.0256551466882229,
      "epoch": 0.1352674361162258,
      "grad_norm": 6.6875,
      "learning_rate": 4.8260854557929074e-05,
      "loss": 1.045433807373047,
      "mean_token_accuracy": 0.7525411635637284,
      "num_tokens": 14948090.0,
      "step": 7530
    },
    {
      "entropy": 0.9966421768069267,
      "epoch": 0.13544707414559662,
      "grad_norm": 5.75,
      "learning_rate": 4.8255585914147906e-05,
      "loss": 1.0045899391174316,
      "mean_token_accuracy": 0.7595632210373878,
      "num_tokens": 14968267.0,
      "step": 7540
    },
    {
      "entropy": 0.947544538974762,
      "epoch": 0.13562671217496744,
      "grad_norm": 6.21875,
      "learning_rate": 4.8250309590376854e-05,
      "loss": 0.9799577713012695,
      "mean_token_accuracy": 0.7670654103159904,
      "num_tokens": 14987214.0,
      "step": 7550
    },
    {
      "entropy": 1.0466882660984993,
      "epoch": 0.13580635020433826,
      "grad_norm": 6.03125,
      "learning_rate": 4.824502558835838e-05,
      "loss": 1.0922574996948242,
      "mean_token_accuracy": 0.7481233045458794,
      "num_tokens": 15006329.0,
      "step": 7560
    },
    {
      "entropy": 1.0878894224762916,
      "epoch": 0.13598598823370908,
      "grad_norm": 7.71875,
      "learning_rate": 4.8239733909837515e-05,
      "loss": 1.1031211853027343,
      "mean_token_accuracy": 0.7472087875008583,
      "num_tokens": 15027147.0,
      "step": 7570
    },
    {
      "entropy": 0.9644355952739716,
      "epoch": 0.1361656262630799,
      "grad_norm": 7.40625,
      "learning_rate": 4.823443455656177e-05,
      "loss": 1.001220989227295,
      "mean_token_accuracy": 0.7750869542360306,
      "num_tokens": 15047355.0,
      "step": 7580
    },
    {
      "entropy": 0.9090218998491764,
      "epoch": 0.13634526429245072,
      "grad_norm": 6.0625,
      "learning_rate": 4.8229127530281225e-05,
      "loss": 0.9393827438354492,
      "mean_token_accuracy": 0.7823689952492714,
      "num_tokens": 15066942.0,
      "step": 7590
    },
    {
      "entropy": 0.9752462819218636,
      "epoch": 0.13652490232182152,
      "grad_norm": 7.21875,
      "learning_rate": 4.822381283274849e-05,
      "loss": 1.00120849609375,
      "mean_token_accuracy": 0.7599927857518196,
      "num_tokens": 15086528.0,
      "step": 7600
    },
    {
      "entropy": 0.9841432318091392,
      "epoch": 0.13670454035119234,
      "grad_norm": 7.21875,
      "learning_rate": 4.8218490465718716e-05,
      "loss": 1.0175788879394532,
      "mean_token_accuracy": 0.7616926595568657,
      "num_tokens": 15106086.0,
      "step": 7610
    },
    {
      "entropy": 0.9625697776675224,
      "epoch": 0.13688417838056316,
      "grad_norm": 5.65625,
      "learning_rate": 4.8213160430949566e-05,
      "loss": 1.055706787109375,
      "mean_token_accuracy": 0.7496672600507737,
      "num_tokens": 15124972.0,
      "step": 7620
    },
    {
      "entropy": 1.0013868123292924,
      "epoch": 0.13706381640993398,
      "grad_norm": 6.5,
      "learning_rate": 4.8207822730201235e-05,
      "loss": 1.022496223449707,
      "mean_token_accuracy": 0.7650181606411934,
      "num_tokens": 15143851.0,
      "step": 7630
    },
    {
      "entropy": 1.1403584860265255,
      "epoch": 0.1372434544393048,
      "grad_norm": 7.25,
      "learning_rate": 4.820247736523648e-05,
      "loss": 1.117336368560791,
      "mean_token_accuracy": 0.7482697673141956,
      "num_tokens": 15164906.0,
      "step": 7640
    },
    {
      "entropy": 1.0574091345071792,
      "epoch": 0.13742309246867562,
      "grad_norm": 6.34375,
      "learning_rate": 4.819712433782056e-05,
      "loss": 1.0826851844787597,
      "mean_token_accuracy": 0.7504638671875,
      "num_tokens": 15184700.0,
      "step": 7650
    },
    {
      "entropy": 0.9792430564761162,
      "epoch": 0.13760273049804644,
      "grad_norm": 5.5625,
      "learning_rate": 4.819176364972127e-05,
      "loss": 0.9904280662536621,
      "mean_token_accuracy": 0.7649105042219162,
      "num_tokens": 15203812.0,
      "step": 7660
    },
    {
      "entropy": 0.9239024288952351,
      "epoch": 0.13778236852741726,
      "grad_norm": 5.65625,
      "learning_rate": 4.8186395302708945e-05,
      "loss": 0.9592381477355957,
      "mean_token_accuracy": 0.7689514562487603,
      "num_tokens": 15223923.0,
      "step": 7670
    },
    {
      "entropy": 0.9036650270223617,
      "epoch": 0.13796200655678806,
      "grad_norm": 5.90625,
      "learning_rate": 4.8181019298556444e-05,
      "loss": 0.9777517318725586,
      "mean_token_accuracy": 0.7790706425905227,
      "num_tokens": 15243500.0,
      "step": 7680
    },
    {
      "entropy": 1.0522332705557347,
      "epoch": 0.13814164458615888,
      "grad_norm": 7.28125,
      "learning_rate": 4.817563563903915e-05,
      "loss": 1.0460851669311524,
      "mean_token_accuracy": 0.7607108160853386,
      "num_tokens": 15262344.0,
      "step": 7690
    },
    {
      "entropy": 1.015552794188261,
      "epoch": 0.1383212826155297,
      "grad_norm": 6.03125,
      "learning_rate": 4.817024432593497e-05,
      "loss": 1.0146279335021973,
      "mean_token_accuracy": 0.7699651852250099,
      "num_tokens": 15281626.0,
      "step": 7700
    },
    {
      "entropy": 1.0855581656098365,
      "epoch": 0.13850092064490052,
      "grad_norm": 8.8125,
      "learning_rate": 4.816484536102437e-05,
      "loss": 1.1384079933166504,
      "mean_token_accuracy": 0.7484158635139465,
      "num_tokens": 15301492.0,
      "step": 7710
    },
    {
      "entropy": 0.9143221013247966,
      "epoch": 0.13868055867427134,
      "grad_norm": 6.5,
      "learning_rate": 4.815943874609029e-05,
      "loss": 0.9617671012878418,
      "mean_token_accuracy": 0.7745268329977989,
      "num_tokens": 15321187.0,
      "step": 7720
    },
    {
      "entropy": 0.9544545084238052,
      "epoch": 0.13886019670364216,
      "grad_norm": 7.125,
      "learning_rate": 4.815402448291825e-05,
      "loss": 0.97393798828125,
      "mean_token_accuracy": 0.7658943727612495,
      "num_tokens": 15339714.0,
      "step": 7730
    },
    {
      "entropy": 1.0051582425832748,
      "epoch": 0.13903983473301298,
      "grad_norm": 7.5625,
      "learning_rate": 4.814860257329625e-05,
      "loss": 1.000349998474121,
      "mean_token_accuracy": 0.7609384104609489,
      "num_tokens": 15358571.0,
      "step": 7740
    },
    {
      "entropy": 0.9305440656840801,
      "epoch": 0.1392194727623838,
      "grad_norm": 7.65625,
      "learning_rate": 4.814317301901486e-05,
      "loss": 0.958104133605957,
      "mean_token_accuracy": 0.770952707529068,
      "num_tokens": 15378411.0,
      "step": 7750
    },
    {
      "entropy": 0.9795505315065384,
      "epoch": 0.13939911079175463,
      "grad_norm": 6.21875,
      "learning_rate": 4.813773582186714e-05,
      "loss": 1.0074503898620606,
      "mean_token_accuracy": 0.7626443967223168,
      "num_tokens": 15401580.0,
      "step": 7760
    },
    {
      "entropy": 0.9503938175737858,
      "epoch": 0.13957874882112542,
      "grad_norm": 6.375,
      "learning_rate": 4.8132290983648675e-05,
      "loss": 1.0205866813659668,
      "mean_token_accuracy": 0.7669512376189231,
      "num_tokens": 15421437.0,
      "step": 7770
    },
    {
      "entropy": 1.0204586371779443,
      "epoch": 0.13975838685049624,
      "grad_norm": 8.3125,
      "learning_rate": 4.8126838506157594e-05,
      "loss": 1.0673606872558594,
      "mean_token_accuracy": 0.7425005987286568,
      "num_tokens": 15442504.0,
      "step": 7780
    },
    {
      "entropy": 0.9531779669225215,
      "epoch": 0.13993802487986706,
      "grad_norm": 7.90625,
      "learning_rate": 4.812137839119454e-05,
      "loss": 0.99929780960083,
      "mean_token_accuracy": 0.7683652810752392,
      "num_tokens": 15462444.0,
      "step": 7790
    },
    {
      "entropy": 0.989243657886982,
      "epoch": 0.14011766290923788,
      "grad_norm": 5.84375,
      "learning_rate": 4.8115910640562675e-05,
      "loss": 0.9879693984985352,
      "mean_token_accuracy": 0.7624568179249763,
      "num_tokens": 15481541.0,
      "step": 7800
    },
    {
      "entropy": 0.9361924007534981,
      "epoch": 0.1402973009386087,
      "grad_norm": 5.28125,
      "learning_rate": 4.8110435256067684e-05,
      "loss": 0.9400579452514648,
      "mean_token_accuracy": 0.7703248217701912,
      "num_tokens": 15501299.0,
      "step": 7810
    },
    {
      "entropy": 0.9134767703711987,
      "epoch": 0.14047693896797953,
      "grad_norm": 5.84375,
      "learning_rate": 4.810495223951777e-05,
      "loss": 0.9158266067504883,
      "mean_token_accuracy": 0.7841978080570697,
      "num_tokens": 15521099.0,
      "step": 7820
    },
    {
      "entropy": 1.0086168855428697,
      "epoch": 0.14065657699735035,
      "grad_norm": 6.125,
      "learning_rate": 4.809946159272366e-05,
      "loss": 1.1049494743347168,
      "mean_token_accuracy": 0.7437369585037231,
      "num_tokens": 15542244.0,
      "step": 7830
    },
    {
      "entropy": 0.9758957415819168,
      "epoch": 0.14083621502672117,
      "grad_norm": 6.40625,
      "learning_rate": 4.8093963317498614e-05,
      "loss": 0.9844758033752441,
      "mean_token_accuracy": 0.7701422110199928,
      "num_tokens": 15562892.0,
      "step": 7840
    },
    {
      "entropy": 0.975720651447773,
      "epoch": 0.14101585305609196,
      "grad_norm": 5.40625,
      "learning_rate": 4.8088457415658374e-05,
      "loss": 1.0480069160461425,
      "mean_token_accuracy": 0.7628588542342186,
      "num_tokens": 15582709.0,
      "step": 7850
    },
    {
      "entropy": 0.9315341129899025,
      "epoch": 0.14119549108546278,
      "grad_norm": 6.71875,
      "learning_rate": 4.808294388902125e-05,
      "loss": 0.9373558044433594,
      "mean_token_accuracy": 0.7789211988449096,
      "num_tokens": 15603694.0,
      "step": 7860
    },
    {
      "entropy": 0.9304000794887543,
      "epoch": 0.1413751291148336,
      "grad_norm": 7.53125,
      "learning_rate": 4.807742273940802e-05,
      "loss": 0.9249242782592774,
      "mean_token_accuracy": 0.7742099076509475,
      "num_tokens": 15624991.0,
      "step": 7870
    },
    {
      "entropy": 1.1340977996587753,
      "epoch": 0.14155476714420442,
      "grad_norm": 6.3125,
      "learning_rate": 4.807189396864202e-05,
      "loss": 1.1512449264526368,
      "mean_token_accuracy": 0.740560895204544,
      "num_tokens": 15645160.0,
      "step": 7880
    },
    {
      "entropy": 0.934343284368515,
      "epoch": 0.14173440517357525,
      "grad_norm": 6.96875,
      "learning_rate": 4.8066357578549084e-05,
      "loss": 0.9627437591552734,
      "mean_token_accuracy": 0.7725389644503593,
      "num_tokens": 15664325.0,
      "step": 7890
    },
    {
      "entropy": 0.9234327264130116,
      "epoch": 0.14191404320294607,
      "grad_norm": 6.6875,
      "learning_rate": 4.806081357095756e-05,
      "loss": 0.9461308479309082,
      "mean_token_accuracy": 0.7729640871286392,
      "num_tokens": 15683844.0,
      "step": 7900
    },
    {
      "entropy": 0.982674865424633,
      "epoch": 0.1420936812323169,
      "grad_norm": 5.21875,
      "learning_rate": 4.805526194769833e-05,
      "loss": 1.0271575927734375,
      "mean_token_accuracy": 0.7620200425386429,
      "num_tokens": 15703099.0,
      "step": 7910
    },
    {
      "entropy": 1.099584560841322,
      "epoch": 0.1422733192616877,
      "grad_norm": 5.96875,
      "learning_rate": 4.804970271060475e-05,
      "loss": 1.1175827980041504,
      "mean_token_accuracy": 0.7436011791229248,
      "num_tokens": 15722672.0,
      "step": 7920
    },
    {
      "entropy": 0.9634081177413464,
      "epoch": 0.14245295729105853,
      "grad_norm": 4.59375,
      "learning_rate": 4.804413586151275e-05,
      "loss": 0.9453004837036133,
      "mean_token_accuracy": 0.7736996680498123,
      "num_tokens": 15742433.0,
      "step": 7930
    },
    {
      "entropy": 0.9332099676132202,
      "epoch": 0.14263259532042932,
      "grad_norm": 5.71875,
      "learning_rate": 4.803856140226072e-05,
      "loss": 0.9822916984558105,
      "mean_token_accuracy": 0.7649647414684295,
      "num_tokens": 15761186.0,
      "step": 7940
    },
    {
      "entropy": 0.8899253077805043,
      "epoch": 0.14281223334980014,
      "grad_norm": 6.34375,
      "learning_rate": 4.80329793346896e-05,
      "loss": 0.9294010162353515,
      "mean_token_accuracy": 0.7811193868517876,
      "num_tokens": 15781247.0,
      "step": 7950
    },
    {
      "entropy": 0.9057660847902298,
      "epoch": 0.14299187137917097,
      "grad_norm": 6.9375,
      "learning_rate": 4.8027389660642815e-05,
      "loss": 0.8979418754577637,
      "mean_token_accuracy": 0.7901165187358856,
      "num_tokens": 15800944.0,
      "step": 7960
    },
    {
      "entropy": 0.8994721084833145,
      "epoch": 0.14317150940854179,
      "grad_norm": 6.15625,
      "learning_rate": 4.802179238196632e-05,
      "loss": 0.9426714897155761,
      "mean_token_accuracy": 0.7724975645542145,
      "num_tokens": 15821123.0,
      "step": 7970
    },
    {
      "entropy": 0.8700844459235668,
      "epoch": 0.1433511474379126,
      "grad_norm": 5.53125,
      "learning_rate": 4.801618750050858e-05,
      "loss": 0.9043972015380859,
      "mean_token_accuracy": 0.7886849567294121,
      "num_tokens": 15840633.0,
      "step": 7980
    },
    {
      "entropy": 0.9122070282697677,
      "epoch": 0.14353078546728343,
      "grad_norm": 5.46875,
      "learning_rate": 4.8010575018120565e-05,
      "loss": 0.940817928314209,
      "mean_token_accuracy": 0.783033049851656,
      "num_tokens": 15860763.0,
      "step": 7990
    },
    {
      "epoch": 0.14371042349665425,
      "eval_entropy": 0.9253975068330764,
      "eval_loss": 0.9760612845420837,
      "eval_mean_token_accuracy": 0.771456812620163,
      "eval_num_tokens": 15879974.0,
      "eval_runtime": 40.527,
      "eval_samples_per_second": 49.35,
      "eval_steps_per_second": 6.169,
      "step": 8000
    },
    {
      "entropy": 0.9213393233716488,
      "epoch": 0.14389006152602507,
      "grad_norm": 7.9375,
      "learning_rate": 4.7999327257970125e-05,
      "loss": 0.9661079406738281,
      "mean_token_accuracy": 0.7736393481492996,
      "num_tokens": 15899160.0,
      "step": 8010
    },
    {
      "entropy": 0.9077555313706398,
      "epoch": 0.14406969955539586,
      "grad_norm": 7.34375,
      "learning_rate": 4.799369198392221e-05,
      "loss": 0.8781991004943848,
      "mean_token_accuracy": 0.7922703772783279,
      "num_tokens": 15920544.0,
      "step": 8020
    },
    {
      "entropy": 0.8791537016630173,
      "epoch": 0.14424933758476668,
      "grad_norm": 5.84375,
      "learning_rate": 4.798804911637299e-05,
      "loss": 0.8872124671936035,
      "mean_token_accuracy": 0.7892002657055854,
      "num_tokens": 15939116.0,
      "step": 8030
    },
    {
      "entropy": 1.007106528431177,
      "epoch": 0.1444289756141375,
      "grad_norm": 6.34375,
      "learning_rate": 4.7982398657185994e-05,
      "loss": 1.0695843696594238,
      "mean_token_accuracy": 0.7552625820040703,
      "num_tokens": 15959916.0,
      "step": 8040
    },
    {
      "entropy": 0.9214366927742959,
      "epoch": 0.14460861364350833,
      "grad_norm": 6.34375,
      "learning_rate": 4.797674060822724e-05,
      "loss": 1.0163339614868163,
      "mean_token_accuracy": 0.766477893292904,
      "num_tokens": 15979421.0,
      "step": 8050
    },
    {
      "entropy": 1.0950290344655513,
      "epoch": 0.14478825167287915,
      "grad_norm": 6.28125,
      "learning_rate": 4.797107497136527e-05,
      "loss": 1.1203022956848145,
      "mean_token_accuracy": 0.7474785096943378,
      "num_tokens": 15999946.0,
      "step": 8060
    },
    {
      "entropy": 1.0755657061934472,
      "epoch": 0.14496788970224997,
      "grad_norm": 7.5625,
      "learning_rate": 4.7965401748471105e-05,
      "loss": 1.004679012298584,
      "mean_token_accuracy": 0.761923325061798,
      "num_tokens": 16020057.0,
      "step": 8070
    },
    {
      "entropy": 0.9756059482693672,
      "epoch": 0.1451475277316208,
      "grad_norm": 7.0,
      "learning_rate": 4.79597209414183e-05,
      "loss": 0.988649559020996,
      "mean_token_accuracy": 0.7638410523533821,
      "num_tokens": 16040036.0,
      "step": 8080
    },
    {
      "entropy": 1.0792164623737335,
      "epoch": 0.1453271657609916,
      "grad_norm": 6.28125,
      "learning_rate": 4.795403255208287e-05,
      "loss": 1.0785239219665528,
      "mean_token_accuracy": 0.7480552606284618,
      "num_tokens": 16060882.0,
      "step": 8090
    },
    {
      "entropy": 0.9377812132239342,
      "epoch": 0.14550680379036243,
      "grad_norm": 5.5,
      "learning_rate": 4.794833658234341e-05,
      "loss": 0.9462774276733399,
      "mean_token_accuracy": 0.7788886345922947,
      "num_tokens": 16079988.0,
      "step": 8100
    },
    {
      "entropy": 0.9954687006771564,
      "epoch": 0.14568644181973323,
      "grad_norm": 9.25,
      "learning_rate": 4.794263303408094e-05,
      "loss": 1.0717336654663085,
      "mean_token_accuracy": 0.7422982066869735,
      "num_tokens": 16099491.0,
      "step": 8110
    },
    {
      "entropy": 0.9266466960310936,
      "epoch": 0.14586607984910405,
      "grad_norm": 5.3125,
      "learning_rate": 4.7936921909179025e-05,
      "loss": 1.007299041748047,
      "mean_token_accuracy": 0.7658177345991135,
      "num_tokens": 16120315.0,
      "step": 8120
    },
    {
      "entropy": 0.918332064896822,
      "epoch": 0.14604571787847487,
      "grad_norm": 5.4375,
      "learning_rate": 4.793120320952373e-05,
      "loss": 0.9172868728637695,
      "mean_token_accuracy": 0.7765826389193535,
      "num_tokens": 16140305.0,
      "step": 8130
    },
    {
      "entropy": 0.9838981352746486,
      "epoch": 0.1462253559078457,
      "grad_norm": 6.25,
      "learning_rate": 4.79254769370036e-05,
      "loss": 1.0050840377807617,
      "mean_token_accuracy": 0.7727771311998367,
      "num_tokens": 16159957.0,
      "step": 8140
    },
    {
      "entropy": 0.9598341636359692,
      "epoch": 0.1464049939372165,
      "grad_norm": 7.53125,
      "learning_rate": 4.791974309350971e-05,
      "loss": 0.9321498870849609,
      "mean_token_accuracy": 0.775365024805069,
      "num_tokens": 16179409.0,
      "step": 8150
    },
    {
      "entropy": 0.8736059963703156,
      "epoch": 0.14658463196658733,
      "grad_norm": 5.8125,
      "learning_rate": 4.791400168093562e-05,
      "loss": 0.9543352127075195,
      "mean_token_accuracy": 0.7838498800992966,
      "num_tokens": 16198836.0,
      "step": 8160
    },
    {
      "entropy": 0.9364406079053879,
      "epoch": 0.14676426999595815,
      "grad_norm": 5.09375,
      "learning_rate": 4.790825270117737e-05,
      "loss": 0.9711810111999511,
      "mean_token_accuracy": 0.7722716882824898,
      "num_tokens": 16217946.0,
      "step": 8170
    },
    {
      "entropy": 1.050256796181202,
      "epoch": 0.14694390802532897,
      "grad_norm": 5.75,
      "learning_rate": 4.790249615613354e-05,
      "loss": 1.1144298553466796,
      "mean_token_accuracy": 0.7403399452567101,
      "num_tokens": 16239033.0,
      "step": 8180
    },
    {
      "entropy": 0.9918863236904144,
      "epoch": 0.14712354605469977,
      "grad_norm": 8.1875,
      "learning_rate": 4.7896732047705186e-05,
      "loss": 0.9685324668884278,
      "mean_token_accuracy": 0.7705891713500023,
      "num_tokens": 16257754.0,
      "step": 8190
    },
    {
      "entropy": 0.9363312378525734,
      "epoch": 0.1473031840840706,
      "grad_norm": 6.375,
      "learning_rate": 4.7890960377795854e-05,
      "loss": 0.9067818641662597,
      "mean_token_accuracy": 0.7761141359806061,
      "num_tokens": 16277770.0,
      "step": 8200
    },
    {
      "entropy": 1.0740804091095923,
      "epoch": 0.1474828221134414,
      "grad_norm": 7.28125,
      "learning_rate": 4.7885181148311595e-05,
      "loss": 1.175827407836914,
      "mean_token_accuracy": 0.7309317663311958,
      "num_tokens": 16297940.0,
      "step": 8210
    },
    {
      "entropy": 0.9732234723865986,
      "epoch": 0.14766246014281223,
      "grad_norm": 6.84375,
      "learning_rate": 4.787939436116097e-05,
      "loss": 1.0304366111755372,
      "mean_token_accuracy": 0.7548943281173706,
      "num_tokens": 16318456.0,
      "step": 8220
    },
    {
      "entropy": 1.0493127137422562,
      "epoch": 0.14784209817218305,
      "grad_norm": 6.46875,
      "learning_rate": 4.787360001825502e-05,
      "loss": 1.0820201873779296,
      "mean_token_accuracy": 0.7460463121533394,
      "num_tokens": 16338285.0,
      "step": 8230
    },
    {
      "entropy": 0.9108531437814236,
      "epoch": 0.14802173620155387,
      "grad_norm": 5.96875,
      "learning_rate": 4.786779812150727e-05,
      "loss": 0.8831881523132324,
      "mean_token_accuracy": 0.7933973103761673,
      "num_tokens": 16358087.0,
      "step": 8240
    },
    {
      "entropy": 0.9601771712303162,
      "epoch": 0.1482013742309247,
      "grad_norm": 6.15625,
      "learning_rate": 4.786198867283378e-05,
      "loss": 0.9855291366577148,
      "mean_token_accuracy": 0.7606870174407959,
      "num_tokens": 16378497.0,
      "step": 8250
    },
    {
      "entropy": 1.04051413834095,
      "epoch": 0.14838101226029551,
      "grad_norm": 7.0,
      "learning_rate": 4.785617167415306e-05,
      "loss": 1.095736312866211,
      "mean_token_accuracy": 0.7461801916360855,
      "num_tokens": 16398707.0,
      "step": 8260
    },
    {
      "entropy": 0.9777207329869271,
      "epoch": 0.14856065028966634,
      "grad_norm": 7.8125,
      "learning_rate": 4.785034712738614e-05,
      "loss": 0.9825257301330567,
      "mean_token_accuracy": 0.7654220774769783,
      "num_tokens": 16418593.0,
      "step": 8270
    },
    {
      "entropy": 0.9808871738612652,
      "epoch": 0.14874028831903713,
      "grad_norm": 7.6875,
      "learning_rate": 4.7844515034456535e-05,
      "loss": 1.0270327568054198,
      "mean_token_accuracy": 0.7590462177991867,
      "num_tokens": 16438576.0,
      "step": 8280
    },
    {
      "entropy": 1.08336161673069,
      "epoch": 0.14891992634840795,
      "grad_norm": 4.9375,
      "learning_rate": 4.7838675397290245e-05,
      "loss": 1.0961649894714356,
      "mean_token_accuracy": 0.7482455588877202,
      "num_tokens": 16459251.0,
      "step": 8290
    },
    {
      "entropy": 0.9407742746174336,
      "epoch": 0.14909956437777877,
      "grad_norm": 4.65625,
      "learning_rate": 4.783282821781577e-05,
      "loss": 0.9486184120178223,
      "mean_token_accuracy": 0.7776899576187134,
      "num_tokens": 16478530.0,
      "step": 8300
    },
    {
      "entropy": 0.9380235381424427,
      "epoch": 0.1492792024071496,
      "grad_norm": 6.09375,
      "learning_rate": 4.7826973497964104e-05,
      "loss": 0.9914567947387696,
      "mean_token_accuracy": 0.7693686172366142,
      "num_tokens": 16498815.0,
      "step": 8310
    },
    {
      "entropy": 0.9677539363503456,
      "epoch": 0.1494588404365204,
      "grad_norm": 6.3125,
      "learning_rate": 4.782111123966872e-05,
      "loss": 1.005695915222168,
      "mean_token_accuracy": 0.7632286712527275,
      "num_tokens": 16518149.0,
      "step": 8320
    },
    {
      "entropy": 0.9675017647445202,
      "epoch": 0.14963847846589123,
      "grad_norm": 6.34375,
      "learning_rate": 4.78152414448656e-05,
      "loss": 1.006098461151123,
      "mean_token_accuracy": 0.7647502481937408,
      "num_tokens": 16537357.0,
      "step": 8330
    },
    {
      "entropy": 0.9819768950343132,
      "epoch": 0.14981811649526205,
      "grad_norm": 6.0625,
      "learning_rate": 4.780936411549319e-05,
      "loss": 1.0014469146728515,
      "mean_token_accuracy": 0.7614009439945221,
      "num_tokens": 16557200.0,
      "step": 8340
    },
    {
      "entropy": 0.9535814724862576,
      "epoch": 0.14999775452463288,
      "grad_norm": 5.625,
      "learning_rate": 4.780347925349243e-05,
      "loss": 0.9643281936645508,
      "mean_token_accuracy": 0.7714976266026496,
      "num_tokens": 16576637.0,
      "step": 8350
    },
    {
      "entropy": 0.9912674829363823,
      "epoch": 0.15017739255400367,
      "grad_norm": 5.4375,
      "learning_rate": 4.7797586860806766e-05,
      "loss": 1.039186954498291,
      "mean_token_accuracy": 0.7689199656248092,
      "num_tokens": 16597417.0,
      "step": 8360
    },
    {
      "entropy": 1.0057676807045937,
      "epoch": 0.1503570305833745,
      "grad_norm": 7.59375,
      "learning_rate": 4.779168693938211e-05,
      "loss": 1.0194299697875977,
      "mean_token_accuracy": 0.763163010776043,
      "num_tokens": 16616941.0,
      "step": 8370
    },
    {
      "entropy": 0.9133186720311641,
      "epoch": 0.1505366686127453,
      "grad_norm": 6.90625,
      "learning_rate": 4.778577949116686e-05,
      "loss": 0.9048111915588379,
      "mean_token_accuracy": 0.7816048264503479,
      "num_tokens": 16636930.0,
      "step": 8380
    },
    {
      "entropy": 1.0196019619703294,
      "epoch": 0.15071630664211613,
      "grad_norm": 7.65625,
      "learning_rate": 4.7779864518111935e-05,
      "loss": 1.0508824348449708,
      "mean_token_accuracy": 0.7527576312422752,
      "num_tokens": 16656472.0,
      "step": 8390
    },
    {
      "entropy": 0.9308358244597912,
      "epoch": 0.15089594467148695,
      "grad_norm": 5.9375,
      "learning_rate": 4.777394202217069e-05,
      "loss": 0.9421045303344726,
      "mean_token_accuracy": 0.7774453893303871,
      "num_tokens": 16675576.0,
      "step": 8400
    },
    {
      "entropy": 1.064289902150631,
      "epoch": 0.15107558270085777,
      "grad_norm": 7.59375,
      "learning_rate": 4.7768012005299e-05,
      "loss": 1.1038736343383788,
      "mean_token_accuracy": 0.749541050195694,
      "num_tokens": 16694850.0,
      "step": 8410
    },
    {
      "entropy": 0.9885524310171604,
      "epoch": 0.1512552207302286,
      "grad_norm": 8.625,
      "learning_rate": 4.77620744694552e-05,
      "loss": 1.0435017585754394,
      "mean_token_accuracy": 0.7530829295516014,
      "num_tokens": 16714257.0,
      "step": 8420
    },
    {
      "entropy": 0.976657697558403,
      "epoch": 0.15143485875959942,
      "grad_norm": 5.59375,
      "learning_rate": 4.7756129416600125e-05,
      "loss": 0.9995203018188477,
      "mean_token_accuracy": 0.7667400926351547,
      "num_tokens": 16733991.0,
      "step": 8430
    },
    {
      "entropy": 0.9459144331514835,
      "epoch": 0.15161449678897024,
      "grad_norm": 7.3125,
      "learning_rate": 4.775017684869708e-05,
      "loss": 0.9626112937927246,
      "mean_token_accuracy": 0.7768433019518852,
      "num_tokens": 16753119.0,
      "step": 8440
    },
    {
      "entropy": 0.9674123585224151,
      "epoch": 0.15179413481834103,
      "grad_norm": 6.25,
      "learning_rate": 4.7744216767711856e-05,
      "loss": 0.9742136001586914,
      "mean_token_accuracy": 0.7641598656773567,
      "num_tokens": 16773481.0,
      "step": 8450
    },
    {
      "entropy": 0.947027649730444,
      "epoch": 0.15197377284771185,
      "grad_norm": 6.0,
      "learning_rate": 4.7738249175612736e-05,
      "loss": 1.0270215034484864,
      "mean_token_accuracy": 0.7674704849720001,
      "num_tokens": 16793170.0,
      "step": 8460
    },
    {
      "entropy": 0.899222207814455,
      "epoch": 0.15215341087708267,
      "grad_norm": 7.25,
      "learning_rate": 4.7732274074370475e-05,
      "loss": 0.9216394424438477,
      "mean_token_accuracy": 0.7811160445213318,
      "num_tokens": 16812152.0,
      "step": 8470
    },
    {
      "entropy": 0.9994746647775173,
      "epoch": 0.1523330489064535,
      "grad_norm": 6.125,
      "learning_rate": 4.77262914659583e-05,
      "loss": 1.0397178649902343,
      "mean_token_accuracy": 0.760996862500906,
      "num_tokens": 16832704.0,
      "step": 8480
    },
    {
      "entropy": 0.9703105866909028,
      "epoch": 0.15251268693582432,
      "grad_norm": 6.4375,
      "learning_rate": 4.7720301352351924e-05,
      "loss": 0.9766875267028808,
      "mean_token_accuracy": 0.7668307051062584,
      "num_tokens": 16853565.0,
      "step": 8490
    },
    {
      "entropy": 1.0501516416668892,
      "epoch": 0.15269232496519514,
      "grad_norm": 6.625,
      "learning_rate": 4.771430373552955e-05,
      "loss": 1.101221752166748,
      "mean_token_accuracy": 0.7473446115851402,
      "num_tokens": 16873410.0,
      "step": 8500
    },
    {
      "entropy": 1.0002856373786926,
      "epoch": 0.15287196299456596,
      "grad_norm": 6.625,
      "learning_rate": 4.7708298617471826e-05,
      "loss": 1.0198898315429688,
      "mean_token_accuracy": 0.7665119498968125,
      "num_tokens": 16892331.0,
      "step": 8510
    },
    {
      "entropy": 0.9968453854322433,
      "epoch": 0.15305160102393678,
      "grad_norm": 6.78125,
      "learning_rate": 4.770228600016192e-05,
      "loss": 0.9913382530212402,
      "mean_token_accuracy": 0.7704442054033279,
      "num_tokens": 16911188.0,
      "step": 8520
    },
    {
      "entropy": 0.9885257661342621,
      "epoch": 0.15323123905330757,
      "grad_norm": 6.125,
      "learning_rate": 4.769626588558545e-05,
      "loss": 0.9797019958496094,
      "mean_token_accuracy": 0.7724747464060784,
      "num_tokens": 16931509.0,
      "step": 8530
    },
    {
      "entropy": 0.9782149374485016,
      "epoch": 0.1534108770826784,
      "grad_norm": 8.125,
      "learning_rate": 4.769023827573051e-05,
      "loss": 1.0146746635437012,
      "mean_token_accuracy": 0.7616164013743401,
      "num_tokens": 16950743.0,
      "step": 8540
    },
    {
      "entropy": 0.8727515764534474,
      "epoch": 0.15359051511204921,
      "grad_norm": 6.0625,
      "learning_rate": 4.768420317258768e-05,
      "loss": 0.8935771942138672,
      "mean_token_accuracy": 0.7834074899554253,
      "num_tokens": 16970015.0,
      "step": 8550
    },
    {
      "entropy": 0.9122753046452999,
      "epoch": 0.15377015314142004,
      "grad_norm": 5.90625,
      "learning_rate": 4.767816057815001e-05,
      "loss": 0.9500601768493653,
      "mean_token_accuracy": 0.7772382199764252,
      "num_tokens": 16990014.0,
      "step": 8560
    },
    {
      "entropy": 0.9873059459030629,
      "epoch": 0.15394979117079086,
      "grad_norm": 6.78125,
      "learning_rate": 4.767211049441301e-05,
      "loss": 1.0860373497009277,
      "mean_token_accuracy": 0.7473328277468682,
      "num_tokens": 17009763.0,
      "step": 8570
    },
    {
      "entropy": 0.9409085288643837,
      "epoch": 0.15412942920016168,
      "grad_norm": 5.46875,
      "learning_rate": 4.76660529233747e-05,
      "loss": 1.0001693725585938,
      "mean_token_accuracy": 0.7653929725289345,
      "num_tokens": 17029272.0,
      "step": 8580
    },
    {
      "entropy": 0.9125297427177429,
      "epoch": 0.1543090672295325,
      "grad_norm": 4.84375,
      "learning_rate": 4.765998786703554e-05,
      "loss": 0.9233849525451661,
      "mean_token_accuracy": 0.7878601789474488,
      "num_tokens": 17049691.0,
      "step": 8590
    },
    {
      "entropy": 0.8805228359997272,
      "epoch": 0.15448870525890332,
      "grad_norm": 6.71875,
      "learning_rate": 4.765391532739846e-05,
      "loss": 0.8344390869140625,
      "mean_token_accuracy": 0.7932497188448906,
      "num_tokens": 17069604.0,
      "step": 8600
    },
    {
      "entropy": 0.9685429468750953,
      "epoch": 0.15466834328827414,
      "grad_norm": 6.09375,
      "learning_rate": 4.764783530646888e-05,
      "loss": 0.988888168334961,
      "mean_token_accuracy": 0.7677136167883873,
      "num_tokens": 17090470.0,
      "step": 8610
    },
    {
      "entropy": 0.8634371489286423,
      "epoch": 0.15484798131764493,
      "grad_norm": 4.90625,
      "learning_rate": 4.764174780625469e-05,
      "loss": 0.9613308906555176,
      "mean_token_accuracy": 0.7739670544862747,
      "num_tokens": 17110593.0,
      "step": 8620
    },
    {
      "entropy": 0.9327140390872956,
      "epoch": 0.15502761934701575,
      "grad_norm": 7.125,
      "learning_rate": 4.763565282876623e-05,
      "loss": 0.9377048492431641,
      "mean_token_accuracy": 0.7768642783164978,
      "num_tokens": 17128944.0,
      "step": 8630
    },
    {
      "entropy": 0.9358147360384464,
      "epoch": 0.15520725737638658,
      "grad_norm": 4.59375,
      "learning_rate": 4.762955037601633e-05,
      "loss": 0.8977405548095703,
      "mean_token_accuracy": 0.7839958354830742,
      "num_tokens": 17149082.0,
      "step": 8640
    },
    {
      "entropy": 0.9750100076198578,
      "epoch": 0.1553868954057574,
      "grad_norm": 6.84375,
      "learning_rate": 4.762344045002028e-05,
      "loss": 1.030782413482666,
      "mean_token_accuracy": 0.7678866490721703,
      "num_tokens": 17169281.0,
      "step": 8650
    },
    {
      "entropy": 0.8895739287137985,
      "epoch": 0.15556653343512822,
      "grad_norm": 8.1875,
      "learning_rate": 4.761732305279585e-05,
      "loss": 0.9771529197692871,
      "mean_token_accuracy": 0.7717635035514832,
      "num_tokens": 17189593.0,
      "step": 8660
    },
    {
      "entropy": 0.9889240600168705,
      "epoch": 0.15574617146449904,
      "grad_norm": 5.84375,
      "learning_rate": 4.7611198186363244e-05,
      "loss": 1.029722309112549,
      "mean_token_accuracy": 0.7595301106572151,
      "num_tokens": 17210641.0,
      "step": 8670
    },
    {
      "entropy": 0.9961977578699589,
      "epoch": 0.15592580949386986,
      "grad_norm": 6.375,
      "learning_rate": 4.760506585274517e-05,
      "loss": 1.0377142906188965,
      "mean_token_accuracy": 0.7633031956851483,
      "num_tokens": 17230687.0,
      "step": 8680
    },
    {
      "entropy": 1.0272681072354317,
      "epoch": 0.15610544752324068,
      "grad_norm": 7.65625,
      "learning_rate": 4.759892605396679e-05,
      "loss": 1.0011177062988281,
      "mean_token_accuracy": 0.7630869328975678,
      "num_tokens": 17249267.0,
      "step": 8690
    },
    {
      "entropy": 0.9655700929462909,
      "epoch": 0.15628508555261147,
      "grad_norm": 4.84375,
      "learning_rate": 4.7592778792055714e-05,
      "loss": 0.9809362411499023,
      "mean_token_accuracy": 0.7759057402610778,
      "num_tokens": 17269204.0,
      "step": 8700
    },
    {
      "entropy": 0.9096764788031578,
      "epoch": 0.1564647235819823,
      "grad_norm": 6.1875,
      "learning_rate": 4.7586624069042044e-05,
      "loss": 0.9462522506713867,
      "mean_token_accuracy": 0.7753981962800026,
      "num_tokens": 17288534.0,
      "step": 8710
    },
    {
      "entropy": 0.9044617041945457,
      "epoch": 0.15664436161135312,
      "grad_norm": 6.90625,
      "learning_rate": 4.758046188695833e-05,
      "loss": 0.9258021354675293,
      "mean_token_accuracy": 0.7789418533444404,
      "num_tokens": 17308185.0,
      "step": 8720
    },
    {
      "entropy": 0.9491473779082298,
      "epoch": 0.15682399964072394,
      "grad_norm": 5.71875,
      "learning_rate": 4.757429224783958e-05,
      "loss": 0.9592753410339355,
      "mean_token_accuracy": 0.7788291737437248,
      "num_tokens": 17328573.0,
      "step": 8730
    },
    {
      "entropy": 0.9719256080687046,
      "epoch": 0.15700363767009476,
      "grad_norm": 6.09375,
      "learning_rate": 4.7568115153723284e-05,
      "loss": 1.0173357009887696,
      "mean_token_accuracy": 0.7668135076761246,
      "num_tokens": 17348027.0,
      "step": 8740
    },
    {
      "entropy": 0.9268810249865055,
      "epoch": 0.15718327569946558,
      "grad_norm": 4.90625,
      "learning_rate": 4.756193060664938e-05,
      "loss": 0.9744657516479492,
      "mean_token_accuracy": 0.7741603493690491,
      "num_tokens": 17367334.0,
      "step": 8750
    },
    {
      "entropy": 0.971621935069561,
      "epoch": 0.1573629137288364,
      "grad_norm": 7.0625,
      "learning_rate": 4.7555738608660264e-05,
      "loss": 0.9625894546508789,
      "mean_token_accuracy": 0.7770540103316307,
      "num_tokens": 17387198.0,
      "step": 8760
    },
    {
      "entropy": 1.0689078114926815,
      "epoch": 0.15754255175820722,
      "grad_norm": 8.125,
      "learning_rate": 4.754953916180081e-05,
      "loss": 1.066531276702881,
      "mean_token_accuracy": 0.7581532210111618,
      "num_tokens": 17408593.0,
      "step": 8770
    },
    {
      "entropy": 0.8720496878027916,
      "epoch": 0.15772218978757804,
      "grad_norm": 7.15625,
      "learning_rate": 4.7543332268118335e-05,
      "loss": 0.8786140441894531,
      "mean_token_accuracy": 0.7867275476455688,
      "num_tokens": 17428171.0,
      "step": 8780
    },
    {
      "entropy": 0.9861971028149128,
      "epoch": 0.15790182781694884,
      "grad_norm": 6.125,
      "learning_rate": 4.753711792966261e-05,
      "loss": 1.0357498168945312,
      "mean_token_accuracy": 0.7570379182696343,
      "num_tokens": 17447144.0,
      "step": 8790
    },
    {
      "entropy": 1.045317170023918,
      "epoch": 0.15808146584631966,
      "grad_norm": 6.75,
      "learning_rate": 4.753089614848589e-05,
      "loss": 1.0704665184020996,
      "mean_token_accuracy": 0.7500792741775513,
      "num_tokens": 17466393.0,
      "step": 8800
    },
    {
      "entropy": 0.8440519250929356,
      "epoch": 0.15826110387569048,
      "grad_norm": 9.4375,
      "learning_rate": 4.752466692664287e-05,
      "loss": 0.8856520652770996,
      "mean_token_accuracy": 0.7968454629182815,
      "num_tokens": 17486247.0,
      "step": 8810
    },
    {
      "entropy": 1.0042638331651688,
      "epoch": 0.1584407419050613,
      "grad_norm": 7.75,
      "learning_rate": 4.751843026619071e-05,
      "loss": 0.9723373413085937,
      "mean_token_accuracy": 0.7791522830724716,
      "num_tokens": 17504949.0,
      "step": 8820
    },
    {
      "entropy": 0.8787732116878033,
      "epoch": 0.15862037993443212,
      "grad_norm": 7.1875,
      "learning_rate": 4.751218616918901e-05,
      "loss": 0.9162117004394531,
      "mean_token_accuracy": 0.7846383064985275,
      "num_tokens": 17524718.0,
      "step": 8830
    },
    {
      "entropy": 1.0122891396284104,
      "epoch": 0.15880001796380294,
      "grad_norm": 6.9375,
      "learning_rate": 4.7505934637699856e-05,
      "loss": 1.0410573959350586,
      "mean_token_accuracy": 0.764104038476944,
      "num_tokens": 17546036.0,
      "step": 8840
    },
    {
      "entropy": 1.0013054005801678,
      "epoch": 0.15897965599317376,
      "grad_norm": 6.875,
      "learning_rate": 4.749967567378775e-05,
      "loss": 1.064315128326416,
      "mean_token_accuracy": 0.7516775041818619,
      "num_tokens": 17565785.0,
      "step": 8850
    },
    {
      "entropy": 0.96607446372509,
      "epoch": 0.15915929402254458,
      "grad_norm": 6.6875,
      "learning_rate": 4.7493409279519686e-05,
      "loss": 0.9831289291381836,
      "mean_token_accuracy": 0.7710923448204994,
      "num_tokens": 17587505.0,
      "step": 8860
    },
    {
      "entropy": 0.9975804023444652,
      "epoch": 0.15933893205191538,
      "grad_norm": 4.625,
      "learning_rate": 4.7487135456965084e-05,
      "loss": 1.0060144424438477,
      "mean_token_accuracy": 0.7647943995893002,
      "num_tokens": 17607190.0,
      "step": 8870
    },
    {
      "entropy": 0.9912250831723213,
      "epoch": 0.1595185700812862,
      "grad_norm": 4.4375,
      "learning_rate": 4.7480854208195835e-05,
      "loss": 1.0001302719116212,
      "mean_token_accuracy": 0.7621441900730133,
      "num_tokens": 17626816.0,
      "step": 8880
    },
    {
      "entropy": 0.8946340456604958,
      "epoch": 0.15969820811065702,
      "grad_norm": 5.09375,
      "learning_rate": 4.747456553528628e-05,
      "loss": 0.9108309745788574,
      "mean_token_accuracy": 0.7817248150706291,
      "num_tokens": 17647779.0,
      "step": 8890
    },
    {
      "entropy": 0.9393540628254413,
      "epoch": 0.15987784614002784,
      "grad_norm": 6.0,
      "learning_rate": 4.7468269440313194e-05,
      "loss": 0.974214744567871,
      "mean_token_accuracy": 0.7725089535117149,
      "num_tokens": 17667830.0,
      "step": 8900
    },
    {
      "entropy": 1.0443107917904855,
      "epoch": 0.16005748416939866,
      "grad_norm": 6.09375,
      "learning_rate": 4.746196592535584e-05,
      "loss": 1.0719220161437988,
      "mean_token_accuracy": 0.7534979403018951,
      "num_tokens": 17688361.0,
      "step": 8910
    },
    {
      "entropy": 1.042989766597748,
      "epoch": 0.16023712219876948,
      "grad_norm": 5.65625,
      "learning_rate": 4.745565499249589e-05,
      "loss": 1.0510982513427733,
      "mean_token_accuracy": 0.7477129012346267,
      "num_tokens": 17707540.0,
      "step": 8920
    },
    {
      "entropy": 1.0210182338953018,
      "epoch": 0.1604167602281403,
      "grad_norm": 7.21875,
      "learning_rate": 4.744933664381749e-05,
      "loss": 1.1169282913208007,
      "mean_token_accuracy": 0.7457792028784752,
      "num_tokens": 17726894.0,
      "step": 8930
    },
    {
      "entropy": 1.0933182634413243,
      "epoch": 0.16059639825751112,
      "grad_norm": 4.5625,
      "learning_rate": 4.7443010881407225e-05,
      "loss": 1.1368219375610351,
      "mean_token_accuracy": 0.7440774589776993,
      "num_tokens": 17749237.0,
      "step": 8940
    },
    {
      "entropy": 1.0087726727128028,
      "epoch": 0.16077603628688195,
      "grad_norm": 5.375,
      "learning_rate": 4.743667770735414e-05,
      "loss": 1.0164118766784669,
      "mean_token_accuracy": 0.760972361266613,
      "num_tokens": 17769183.0,
      "step": 8950
    },
    {
      "entropy": 0.9550929941236973,
      "epoch": 0.16095567431625274,
      "grad_norm": 5.9375,
      "learning_rate": 4.743033712374971e-05,
      "loss": 0.9461418151855469,
      "mean_token_accuracy": 0.7787161409854889,
      "num_tokens": 17789489.0,
      "step": 8960
    },
    {
      "entropy": 0.9016365617513656,
      "epoch": 0.16113531234562356,
      "grad_norm": 5.21875,
      "learning_rate": 4.742398913268788e-05,
      "loss": 0.8944819450378418,
      "mean_token_accuracy": 0.784300409257412,
      "num_tokens": 17808932.0,
      "step": 8970
    },
    {
      "entropy": 0.9215508490800858,
      "epoch": 0.16131495037499438,
      "grad_norm": 4.78125,
      "learning_rate": 4.741763373626502e-05,
      "loss": 0.9370976448059082,
      "mean_token_accuracy": 0.7738397717475891,
      "num_tokens": 17829564.0,
      "step": 8980
    },
    {
      "entropy": 0.8448221601545811,
      "epoch": 0.1614945884043652,
      "grad_norm": 7.78125,
      "learning_rate": 4.741127093657995e-05,
      "loss": 0.8962443351745606,
      "mean_token_accuracy": 0.7898091658949852,
      "num_tokens": 17849384.0,
      "step": 8990
    },
    {
      "epoch": 0.16167422643373602,
      "eval_entropy": 0.9303436948060989,
      "eval_loss": 0.9488416314125061,
      "eval_mean_token_accuracy": 0.7757836725711823,
      "eval_num_tokens": 17870229.0,
      "eval_runtime": 40.5228,
      "eval_samples_per_second": 49.355,
      "eval_steps_per_second": 6.169,
      "step": 9000
    },
    {
      "entropy": 0.9638120740652084,
      "epoch": 0.16185386446310684,
      "grad_norm": 5.3125,
      "learning_rate": 4.739852313583074e-05,
      "loss": 1.018975067138672,
      "mean_token_accuracy": 0.7673677865415811,
      "num_tokens": 17890018.0,
      "step": 9010
    },
    {
      "entropy": 0.9681854233145714,
      "epoch": 0.16203350249247767,
      "grad_norm": 6.0,
      "learning_rate": 4.739213813897645e-05,
      "loss": 0.9576973915100098,
      "mean_token_accuracy": 0.7843311294913292,
      "num_tokens": 17909464.0,
      "step": 9020
    },
    {
      "entropy": 0.9766761496663093,
      "epoch": 0.1622131405218485,
      "grad_norm": 7.8125,
      "learning_rate": 4.738574574727969e-05,
      "loss": 0.9760836601257324,
      "mean_token_accuracy": 0.7754037395119667,
      "num_tokens": 17929076.0,
      "step": 9030
    },
    {
      "entropy": 0.8228480659425259,
      "epoch": 0.16239277855121928,
      "grad_norm": 6.90625,
      "learning_rate": 4.7379345962851515e-05,
      "loss": 0.8559449195861817,
      "mean_token_accuracy": 0.8000741928815842,
      "num_tokens": 17948295.0,
      "step": 9040
    },
    {
      "entropy": 0.9463241461664438,
      "epoch": 0.1625724165805901,
      "grad_norm": 6.53125,
      "learning_rate": 4.737293878780539e-05,
      "loss": 1.0076125144958497,
      "mean_token_accuracy": 0.7696226395666599,
      "num_tokens": 17968971.0,
      "step": 9050
    },
    {
      "entropy": 0.8300053380429745,
      "epoch": 0.16275205460996092,
      "grad_norm": 6.71875,
      "learning_rate": 4.7366524224257245e-05,
      "loss": 0.8233118057250977,
      "mean_token_accuracy": 0.7992605149745942,
      "num_tokens": 17988264.0,
      "step": 9060
    },
    {
      "entropy": 0.9255936995148659,
      "epoch": 0.16293169263933174,
      "grad_norm": 6.65625,
      "learning_rate": 4.736010227432544e-05,
      "loss": 0.9705953598022461,
      "mean_token_accuracy": 0.7700245976448059,
      "num_tokens": 18007269.0,
      "step": 9070
    },
    {
      "entropy": 0.9908387616276741,
      "epoch": 0.16311133066870256,
      "grad_norm": 11.1875,
      "learning_rate": 4.7353672940130796e-05,
      "loss": 1.0327488899230957,
      "mean_token_accuracy": 0.7612299218773841,
      "num_tokens": 18026595.0,
      "step": 9080
    },
    {
      "entropy": 0.9577730923891068,
      "epoch": 0.16329096869807339,
      "grad_norm": 5.4375,
      "learning_rate": 4.734723622379653e-05,
      "loss": 0.9640662193298339,
      "mean_token_accuracy": 0.7827681139111519,
      "num_tokens": 18045736.0,
      "step": 9090
    },
    {
      "entropy": 0.9837521903216839,
      "epoch": 0.1634706067274442,
      "grad_norm": 8.6875,
      "learning_rate": 4.734079212744833e-05,
      "loss": 1.0170021057128906,
      "mean_token_accuracy": 0.7706265673041344,
      "num_tokens": 18067611.0,
      "step": 9100
    },
    {
      "entropy": 0.8206705145537854,
      "epoch": 0.16365024475681503,
      "grad_norm": 6.71875,
      "learning_rate": 4.733434065321432e-05,
      "loss": 0.8678516387939453,
      "mean_token_accuracy": 0.7948760494589806,
      "num_tokens": 18086958.0,
      "step": 9110
    },
    {
      "entropy": 0.9261166758835315,
      "epoch": 0.16382988278618585,
      "grad_norm": 6.40625,
      "learning_rate": 4.7327881803225046e-05,
      "loss": 0.9737388610839843,
      "mean_token_accuracy": 0.7739491254091263,
      "num_tokens": 18106731.0,
      "step": 9120
    },
    {
      "entropy": 0.8987494334578514,
      "epoch": 0.16400952081555664,
      "grad_norm": 9.25,
      "learning_rate": 4.732141557961349e-05,
      "loss": 0.9284511566162109,
      "mean_token_accuracy": 0.7756375715136528,
      "num_tokens": 18126809.0,
      "step": 9130
    },
    {
      "entropy": 1.0321485213935375,
      "epoch": 0.16418915884492746,
      "grad_norm": 6.40625,
      "learning_rate": 4.731494198451509e-05,
      "loss": 1.0619437217712402,
      "mean_token_accuracy": 0.7506299927830696,
      "num_tokens": 18147498.0,
      "step": 9140
    },
    {
      "entropy": 0.8776893839240074,
      "epoch": 0.16436879687429828,
      "grad_norm": 7.03125,
      "learning_rate": 4.7308461020067704e-05,
      "loss": 0.8966526985168457,
      "mean_token_accuracy": 0.7856161281466484,
      "num_tokens": 18166452.0,
      "step": 9150
    },
    {
      "entropy": 0.9587407387793064,
      "epoch": 0.1645484349036691,
      "grad_norm": 8.9375,
      "learning_rate": 4.7301972688411614e-05,
      "loss": 0.939057445526123,
      "mean_token_accuracy": 0.7788362771272659,
      "num_tokens": 18184290.0,
      "step": 9160
    },
    {
      "entropy": 0.8897178299725056,
      "epoch": 0.16472807293303993,
      "grad_norm": 7.6875,
      "learning_rate": 4.729547699168954e-05,
      "loss": 0.9725374221801758,
      "mean_token_accuracy": 0.7768823891878128,
      "num_tokens": 18204300.0,
      "step": 9170
    },
    {
      "entropy": 0.8844131723046302,
      "epoch": 0.16490771096241075,
      "grad_norm": 7.25,
      "learning_rate": 4.7288973932046656e-05,
      "loss": 0.9096429824829102,
      "mean_token_accuracy": 0.7802961796522141,
      "num_tokens": 18223806.0,
      "step": 9180
    },
    {
      "entropy": 1.0714459702372552,
      "epoch": 0.16508734899178157,
      "grad_norm": 6.375,
      "learning_rate": 4.728246351163054e-05,
      "loss": 1.106997299194336,
      "mean_token_accuracy": 0.7480820059776306,
      "num_tokens": 18243708.0,
      "step": 9190
    },
    {
      "entropy": 0.8125671908259392,
      "epoch": 0.1652669870211524,
      "grad_norm": 6.34375,
      "learning_rate": 4.727594573259121e-05,
      "loss": 0.7652768611907959,
      "mean_token_accuracy": 0.8130101934075356,
      "num_tokens": 18263476.0,
      "step": 9200
    },
    {
      "entropy": 0.8467555202543735,
      "epoch": 0.16544662505052318,
      "grad_norm": 7.3125,
      "learning_rate": 4.7269420597081114e-05,
      "loss": 0.920893669128418,
      "mean_token_accuracy": 0.7788765475153923,
      "num_tokens": 18284127.0,
      "step": 9210
    },
    {
      "entropy": 0.9326670832931996,
      "epoch": 0.165626263079894,
      "grad_norm": 6.59375,
      "learning_rate": 4.726288810725515e-05,
      "loss": 0.9445914268493653,
      "mean_token_accuracy": 0.7757194757461547,
      "num_tokens": 18303875.0,
      "step": 9220
    },
    {
      "entropy": 0.8629965014755726,
      "epoch": 0.16580590110926482,
      "grad_norm": 6.65625,
      "learning_rate": 4.72563482652706e-05,
      "loss": 0.9051396369934082,
      "mean_token_accuracy": 0.7843986138701439,
      "num_tokens": 18323468.0,
      "step": 9230
    },
    {
      "entropy": 0.8389853581786155,
      "epoch": 0.16598553913863565,
      "grad_norm": 6.34375,
      "learning_rate": 4.724980107328722e-05,
      "loss": 0.8194594383239746,
      "mean_token_accuracy": 0.798887787759304,
      "num_tokens": 18343394.0,
      "step": 9240
    },
    {
      "entropy": 1.0091021195054055,
      "epoch": 0.16616517716800647,
      "grad_norm": 6.5,
      "learning_rate": 4.724324653346716e-05,
      "loss": 1.043700885772705,
      "mean_token_accuracy": 0.7542940318584442,
      "num_tokens": 18363903.0,
      "step": 9250
    },
    {
      "entropy": 0.9561960935592652,
      "epoch": 0.1663448151973773,
      "grad_norm": 6.0,
      "learning_rate": 4.723668464797502e-05,
      "loss": 1.0119144439697265,
      "mean_token_accuracy": 0.7664878085255623,
      "num_tokens": 18382940.0,
      "step": 9260
    },
    {
      "entropy": 0.8921266764402389,
      "epoch": 0.1665244532267481,
      "grad_norm": 5.59375,
      "learning_rate": 4.723011541897781e-05,
      "loss": 0.8833674430847168,
      "mean_token_accuracy": 0.7899740055203438,
      "num_tokens": 18403071.0,
      "step": 9270
    },
    {
      "entropy": 0.8875182561576367,
      "epoch": 0.16670409125611893,
      "grad_norm": 5.6875,
      "learning_rate": 4.722353884864496e-05,
      "loss": 0.9245976448059082,
      "mean_token_accuracy": 0.7728003561496735,
      "num_tokens": 18423145.0,
      "step": 9280
    },
    {
      "entropy": 0.942279863357544,
      "epoch": 0.16688372928548975,
      "grad_norm": 6.0,
      "learning_rate": 4.721695493914835e-05,
      "loss": 1.0021965980529786,
      "mean_token_accuracy": 0.7701666370034218,
      "num_tokens": 18442701.0,
      "step": 9290
    },
    {
      "entropy": 1.0353110067546367,
      "epoch": 0.16706336731486054,
      "grad_norm": 11.0,
      "learning_rate": 4.7210363692662267e-05,
      "loss": 1.0737655639648438,
      "mean_token_accuracy": 0.7501712031662464,
      "num_tokens": 18463468.0,
      "step": 9300
    },
    {
      "entropy": 0.9296294204890728,
      "epoch": 0.16724300534423137,
      "grad_norm": 5.5625,
      "learning_rate": 4.720376511136342e-05,
      "loss": 0.8965011596679687,
      "mean_token_accuracy": 0.7824577704071999,
      "num_tokens": 18484080.0,
      "step": 9310
    },
    {
      "entropy": 0.8524242773652076,
      "epoch": 0.1674226433736022,
      "grad_norm": 4.84375,
      "learning_rate": 4.719715919743094e-05,
      "loss": 0.9343573570251464,
      "mean_token_accuracy": 0.793031319975853,
      "num_tokens": 18505286.0,
      "step": 9320
    },
    {
      "entropy": 0.911799605935812,
      "epoch": 0.167602281402973,
      "grad_norm": 6.03125,
      "learning_rate": 4.719054595304638e-05,
      "loss": 0.9171762466430664,
      "mean_token_accuracy": 0.7827555373311043,
      "num_tokens": 18525912.0,
      "step": 9330
    },
    {
      "entropy": 0.8643674038350582,
      "epoch": 0.16778191943234383,
      "grad_norm": 5.40625,
      "learning_rate": 4.718392538039374e-05,
      "loss": 0.8773012161254883,
      "mean_token_accuracy": 0.795585049688816,
      "num_tokens": 18545706.0,
      "step": 9340
    },
    {
      "entropy": 1.027641935646534,
      "epoch": 0.16796155746171465,
      "grad_norm": 6.59375,
      "learning_rate": 4.717729748165938e-05,
      "loss": 1.0094654083251953,
      "mean_token_accuracy": 0.7578551426529885,
      "num_tokens": 18565375.0,
      "step": 9350
    },
    {
      "entropy": 0.9796357691287995,
      "epoch": 0.16814119549108547,
      "grad_norm": 4.1875,
      "learning_rate": 4.717066225903214e-05,
      "loss": 0.9889277458190918,
      "mean_token_accuracy": 0.7676798954606057,
      "num_tokens": 18586569.0,
      "step": 9360
    },
    {
      "entropy": 0.8381679400801658,
      "epoch": 0.1683208335204563,
      "grad_norm": 6.75,
      "learning_rate": 4.7164019714703256e-05,
      "loss": 0.862794303894043,
      "mean_token_accuracy": 0.8010192960500717,
      "num_tokens": 18605744.0,
      "step": 9370
    },
    {
      "entropy": 0.903783343732357,
      "epoch": 0.16850047154982709,
      "grad_norm": 7.6875,
      "learning_rate": 4.715736985086636e-05,
      "loss": 0.9910968780517578,
      "mean_token_accuracy": 0.7659548781812191,
      "num_tokens": 18625789.0,
      "step": 9380
    },
    {
      "entropy": 0.8808243818581104,
      "epoch": 0.1686801095791979,
      "grad_norm": 6.4375,
      "learning_rate": 4.7150712669717545e-05,
      "loss": 0.8979572296142578,
      "mean_token_accuracy": 0.787866523861885,
      "num_tokens": 18645912.0,
      "step": 9390
    },
    {
      "entropy": 0.8958413884043693,
      "epoch": 0.16885974760856873,
      "grad_norm": 7.5625,
      "learning_rate": 4.714404817345528e-05,
      "loss": 0.9158196449279785,
      "mean_token_accuracy": 0.7902197211980819,
      "num_tokens": 18665868.0,
      "step": 9400
    },
    {
      "entropy": 0.862149441242218,
      "epoch": 0.16903938563793955,
      "grad_norm": 5.75,
      "learning_rate": 4.713737636428048e-05,
      "loss": 0.8268040657043457,
      "mean_token_accuracy": 0.7911340206861496,
      "num_tokens": 18686617.0,
      "step": 9410
    },
    {
      "entropy": 0.907992385327816,
      "epoch": 0.16921902366731037,
      "grad_norm": 4.90625,
      "learning_rate": 4.713069724439645e-05,
      "loss": 0.949774169921875,
      "mean_token_accuracy": 0.7834359690546989,
      "num_tokens": 18706887.0,
      "step": 9420
    },
    {
      "entropy": 0.9191290006041527,
      "epoch": 0.1693986616966812,
      "grad_norm": 7.40625,
      "learning_rate": 4.712401081600893e-05,
      "loss": 0.9316378593444824,
      "mean_token_accuracy": 0.7811064325273037,
      "num_tokens": 18727175.0,
      "step": 9430
    },
    {
      "entropy": 0.8583660207688808,
      "epoch": 0.169578299726052,
      "grad_norm": 5.875,
      "learning_rate": 4.711731708132606e-05,
      "loss": 0.8893906593322753,
      "mean_token_accuracy": 0.7887267261743546,
      "num_tokens": 18746133.0,
      "step": 9440
    },
    {
      "entropy": 0.9689084470272065,
      "epoch": 0.16975793775542283,
      "grad_norm": 8.25,
      "learning_rate": 4.7110616042558404e-05,
      "loss": 1.0509136199951172,
      "mean_token_accuracy": 0.7614096269011498,
      "num_tokens": 18765494.0,
      "step": 9450
    },
    {
      "entropy": 0.8895253948867321,
      "epoch": 0.16993757578479365,
      "grad_norm": 6.53125,
      "learning_rate": 4.7103907701918924e-05,
      "loss": 0.9014310836791992,
      "mean_token_accuracy": 0.7900927186012268,
      "num_tokens": 18785273.0,
      "step": 9460
    },
    {
      "entropy": 0.916766132414341,
      "epoch": 0.17011721381416445,
      "grad_norm": 4.96875,
      "learning_rate": 4.7097192061623e-05,
      "loss": 0.9033540725708008,
      "mean_token_accuracy": 0.7863207668066025,
      "num_tokens": 18804621.0,
      "step": 9470
    },
    {
      "entropy": 0.8936407186090947,
      "epoch": 0.17029685184353527,
      "grad_norm": 7.34375,
      "learning_rate": 4.7090469123888436e-05,
      "loss": 0.9383153915405273,
      "mean_token_accuracy": 0.7787775993347168,
      "num_tokens": 18823792.0,
      "step": 9480
    },
    {
      "entropy": 0.9060123920440674,
      "epoch": 0.1704764898729061,
      "grad_norm": 6.0,
      "learning_rate": 4.7083738890935426e-05,
      "loss": 0.8974112510681153,
      "mean_token_accuracy": 0.7896077901124954,
      "num_tokens": 18844352.0,
      "step": 9490
    },
    {
      "entropy": 0.9063649274408817,
      "epoch": 0.1706561279022769,
      "grad_norm": 7.0625,
      "learning_rate": 4.707700136498658e-05,
      "loss": 0.9537568092346191,
      "mean_token_accuracy": 0.7833113089203835,
      "num_tokens": 18864462.0,
      "step": 9500
    },
    {
      "entropy": 0.9782896302640438,
      "epoch": 0.17083576593164773,
      "grad_norm": 4.875,
      "learning_rate": 4.707025654826692e-05,
      "loss": 1.0257890701293946,
      "mean_token_accuracy": 0.7674552157521248,
      "num_tokens": 18885483.0,
      "step": 9510
    },
    {
      "entropy": 0.8746934436261654,
      "epoch": 0.17101540396101855,
      "grad_norm": 6.84375,
      "learning_rate": 4.706350444300387e-05,
      "loss": 0.9134779930114746,
      "mean_token_accuracy": 0.7908712387084961,
      "num_tokens": 18905364.0,
      "step": 9520
    },
    {
      "entropy": 0.8717833809554577,
      "epoch": 0.17119504199038937,
      "grad_norm": 4.5,
      "learning_rate": 4.705674505142726e-05,
      "loss": 0.8525161743164062,
      "mean_token_accuracy": 0.7905518040060997,
      "num_tokens": 18924291.0,
      "step": 9530
    },
    {
      "entropy": 1.027393075078726,
      "epoch": 0.1713746800197602,
      "grad_norm": 7.375,
      "learning_rate": 4.704997837576935e-05,
      "loss": 1.0328557968139649,
      "mean_token_accuracy": 0.7582243226468564,
      "num_tokens": 18944899.0,
      "step": 9540
    },
    {
      "entropy": 0.8280367687344551,
      "epoch": 0.171554318049131,
      "grad_norm": 8.5625,
      "learning_rate": 4.7043204418264756e-05,
      "loss": 0.888212013244629,
      "mean_token_accuracy": 0.7915822759270668,
      "num_tokens": 18965407.0,
      "step": 9550
    },
    {
      "entropy": 0.8664294868707657,
      "epoch": 0.1717339560785018,
      "grad_norm": 6.0625,
      "learning_rate": 4.703642318115056e-05,
      "loss": 0.8780263900756836,
      "mean_token_accuracy": 0.7851709812879563,
      "num_tokens": 18985761.0,
      "step": 9560
    },
    {
      "entropy": 0.8437274187803269,
      "epoch": 0.17191359410787263,
      "grad_norm": 10.0,
      "learning_rate": 4.702963466666619e-05,
      "loss": 0.8491219520568848,
      "mean_token_accuracy": 0.7917777583003044,
      "num_tokens": 19005278.0,
      "step": 9570
    },
    {
      "entropy": 0.7588127285242081,
      "epoch": 0.17209323213724345,
      "grad_norm": 6.4375,
      "learning_rate": 4.702283887705352e-05,
      "loss": 0.7774446964263916,
      "mean_token_accuracy": 0.8105572029948235,
      "num_tokens": 19025594.0,
      "step": 9580
    },
    {
      "entropy": 0.9226869255304336,
      "epoch": 0.17227287016661427,
      "grad_norm": 6.375,
      "learning_rate": 4.701603581455681e-05,
      "loss": 0.9948654174804688,
      "mean_token_accuracy": 0.7676158845424652,
      "num_tokens": 19045507.0,
      "step": 9590
    },
    {
      "entropy": 0.9335997827351094,
      "epoch": 0.1724525081959851,
      "grad_norm": 6.75,
      "learning_rate": 4.700922548142273e-05,
      "loss": 0.9896574020385742,
      "mean_token_accuracy": 0.7613599583506584,
      "num_tokens": 19065763.0,
      "step": 9600
    },
    {
      "entropy": 0.970546105504036,
      "epoch": 0.17263214622535591,
      "grad_norm": 5.1875,
      "learning_rate": 4.7002407879900314e-05,
      "loss": 0.9497195243835449,
      "mean_token_accuracy": 0.77253128439188,
      "num_tokens": 19086119.0,
      "step": 9610
    },
    {
      "entropy": 0.9360686421394349,
      "epoch": 0.17281178425472674,
      "grad_norm": 5.5,
      "learning_rate": 4.699558301224105e-05,
      "loss": 0.9997067451477051,
      "mean_token_accuracy": 0.7645293459296226,
      "num_tokens": 19105940.0,
      "step": 9620
    },
    {
      "entropy": 0.9390234097838401,
      "epoch": 0.17299142228409756,
      "grad_norm": 5.1875,
      "learning_rate": 4.6988750880698797e-05,
      "loss": 0.96348876953125,
      "mean_token_accuracy": 0.7760978445410729,
      "num_tokens": 19125875.0,
      "step": 9630
    },
    {
      "entropy": 1.0701192140579223,
      "epoch": 0.17317106031346835,
      "grad_norm": 5.8125,
      "learning_rate": 4.698191148752982e-05,
      "loss": 1.0500544548034667,
      "mean_token_accuracy": 0.754813103377819,
      "num_tokens": 19145327.0,
      "step": 9640
    },
    {
      "entropy": 0.9879483342170715,
      "epoch": 0.17335069834283917,
      "grad_norm": 6.0625,
      "learning_rate": 4.6975064834992765e-05,
      "loss": 0.9527148246765137,
      "mean_token_accuracy": 0.7701940312981606,
      "num_tokens": 19165501.0,
      "step": 9650
    },
    {
      "entropy": 0.79984863512218,
      "epoch": 0.17353033637221,
      "grad_norm": 7.1875,
      "learning_rate": 4.69682109253487e-05,
      "loss": 0.9002044677734375,
      "mean_token_accuracy": 0.7872171685099602,
      "num_tokens": 19184148.0,
      "step": 9660
    },
    {
      "entropy": 0.9505138382315635,
      "epoch": 0.1737099744015808,
      "grad_norm": 5.75,
      "learning_rate": 4.6961349760861095e-05,
      "loss": 0.9538670539855957,
      "mean_token_accuracy": 0.7678418636322022,
      "num_tokens": 19204565.0,
      "step": 9670
    },
    {
      "entropy": 0.9760745167732239,
      "epoch": 0.17388961243095163,
      "grad_norm": 6.03125,
      "learning_rate": 4.695448134379577e-05,
      "loss": 0.9559142112731933,
      "mean_token_accuracy": 0.7691164925694466,
      "num_tokens": 19225042.0,
      "step": 9680
    },
    {
      "entropy": 0.8791351959109306,
      "epoch": 0.17406925046032246,
      "grad_norm": 9.9375,
      "learning_rate": 4.6947605676420994e-05,
      "loss": 0.9054128646850585,
      "mean_token_accuracy": 0.786659762263298,
      "num_tokens": 19245213.0,
      "step": 9690
    },
    {
      "entropy": 0.9709884136915207,
      "epoch": 0.17424888848969328,
      "grad_norm": 4.03125,
      "learning_rate": 4.6940722761007396e-05,
      "loss": 0.9749642372131347,
      "mean_token_accuracy": 0.7780238419771195,
      "num_tokens": 19265047.0,
      "step": 9700
    },
    {
      "entropy": 0.9294934887439013,
      "epoch": 0.1744285265190641,
      "grad_norm": 6.28125,
      "learning_rate": 4.6933832599828e-05,
      "loss": 0.9403132438659668,
      "mean_token_accuracy": 0.789320471137762,
      "num_tokens": 19284890.0,
      "step": 9710
    },
    {
      "entropy": 0.8958697043359279,
      "epoch": 0.1746081645484349,
      "grad_norm": 9.1875,
      "learning_rate": 4.6926935195158257e-05,
      "loss": 0.9269455909729004,
      "mean_token_accuracy": 0.7735864460468292,
      "num_tokens": 19304494.0,
      "step": 9720
    },
    {
      "entropy": 0.9792730428278447,
      "epoch": 0.1747878025778057,
      "grad_norm": 7.59375,
      "learning_rate": 4.692003054927597e-05,
      "loss": 1.0214195251464844,
      "mean_token_accuracy": 0.7646785318851471,
      "num_tokens": 19324438.0,
      "step": 9730
    },
    {
      "entropy": 1.0013775184750557,
      "epoch": 0.17496744060717653,
      "grad_norm": 6.5625,
      "learning_rate": 4.6913118664461344e-05,
      "loss": 1.0185001373291016,
      "mean_token_accuracy": 0.7612957775592804,
      "num_tokens": 19344417.0,
      "step": 9740
    },
    {
      "entropy": 0.8812263227999211,
      "epoch": 0.17514707863654735,
      "grad_norm": 6.21875,
      "learning_rate": 4.690619954299699e-05,
      "loss": 0.8862649917602539,
      "mean_token_accuracy": 0.7906287685036659,
      "num_tokens": 19365669.0,
      "step": 9750
    },
    {
      "entropy": 0.9786417976021766,
      "epoch": 0.17532671666591818,
      "grad_norm": 5.25,
      "learning_rate": 4.6899273187167894e-05,
      "loss": 0.9972304344177246,
      "mean_token_accuracy": 0.7638727478682995,
      "num_tokens": 19386111.0,
      "step": 9760
    },
    {
      "entropy": 0.843319721519947,
      "epoch": 0.175506354695289,
      "grad_norm": 7.15625,
      "learning_rate": 4.689233959926142e-05,
      "loss": 0.8748242378234863,
      "mean_token_accuracy": 0.7886282086372376,
      "num_tokens": 19406243.0,
      "step": 9770
    },
    {
      "entropy": 0.8470712170004845,
      "epoch": 0.17568599272465982,
      "grad_norm": 5.40625,
      "learning_rate": 4.688539878156737e-05,
      "loss": 0.8864126205444336,
      "mean_token_accuracy": 0.7888110801577568,
      "num_tokens": 19426876.0,
      "step": 9780
    },
    {
      "entropy": 0.880895970761776,
      "epoch": 0.17586563075403064,
      "grad_norm": 5.09375,
      "learning_rate": 4.687845073637786e-05,
      "loss": 0.9161407470703125,
      "mean_token_accuracy": 0.7760112106800079,
      "num_tokens": 19448188.0,
      "step": 9790
    },
    {
      "entropy": 0.8215919353067875,
      "epoch": 0.17604526878340146,
      "grad_norm": 5.5625,
      "learning_rate": 4.687149546598747e-05,
      "loss": 0.8065370559692383,
      "mean_token_accuracy": 0.8095763623714447,
      "num_tokens": 19469161.0,
      "step": 9800
    },
    {
      "entropy": 0.940293101966381,
      "epoch": 0.17622490681277225,
      "grad_norm": 5.71875,
      "learning_rate": 4.6864532972693096e-05,
      "loss": 0.9980769157409668,
      "mean_token_accuracy": 0.7708943307399749,
      "num_tokens": 19490164.0,
      "step": 9810
    },
    {
      "entropy": 0.9293912209570407,
      "epoch": 0.17640454484214307,
      "grad_norm": 8.0625,
      "learning_rate": 4.685756325879407e-05,
      "loss": 0.9511631965637207,
      "mean_token_accuracy": 0.7740518391132355,
      "num_tokens": 19510823.0,
      "step": 9820
    },
    {
      "entropy": 0.819664042443037,
      "epoch": 0.1765841828715139,
      "grad_norm": 6.71875,
      "learning_rate": 4.685058632659208e-05,
      "loss": 0.8230952262878418,
      "mean_token_accuracy": 0.808774447441101,
      "num_tokens": 19531766.0,
      "step": 9830
    },
    {
      "entropy": 0.8843199007213116,
      "epoch": 0.17676382090088472,
      "grad_norm": 5.375,
      "learning_rate": 4.6843602178391216e-05,
      "loss": 0.9109560966491699,
      "mean_token_accuracy": 0.7788462847471237,
      "num_tokens": 19552359.0,
      "step": 9840
    },
    {
      "entropy": 0.7806415006518364,
      "epoch": 0.17694345893025554,
      "grad_norm": 5.28125,
      "learning_rate": 4.683661081649794e-05,
      "loss": 0.8133620262145996,
      "mean_token_accuracy": 0.8053049743175507,
      "num_tokens": 19572777.0,
      "step": 9850
    },
    {
      "entropy": 0.9454508267343045,
      "epoch": 0.17712309695962636,
      "grad_norm": 8.125,
      "learning_rate": 4.682961224322109e-05,
      "loss": 1.0305035591125489,
      "mean_token_accuracy": 0.7619006484746933,
      "num_tokens": 19591845.0,
      "step": 9860
    },
    {
      "entropy": 0.8355608589947223,
      "epoch": 0.17730273498899718,
      "grad_norm": 4.84375,
      "learning_rate": 4.6822606460871905e-05,
      "loss": 0.8281384468078613,
      "mean_token_accuracy": 0.7892985329031944,
      "num_tokens": 19612190.0,
      "step": 9870
    },
    {
      "entropy": 0.9059530779719353,
      "epoch": 0.177482373018368,
      "grad_norm": 7.5625,
      "learning_rate": 4.681559347176399e-05,
      "loss": 0.9679039001464844,
      "mean_token_accuracy": 0.7695346951484681,
      "num_tokens": 19631449.0,
      "step": 9880
    },
    {
      "entropy": 0.8924325972795486,
      "epoch": 0.1776620110477388,
      "grad_norm": 6.9375,
      "learning_rate": 4.6808573278213335e-05,
      "loss": 0.8547731399536133,
      "mean_token_accuracy": 0.7943399846553802,
      "num_tokens": 19651738.0,
      "step": 9890
    },
    {
      "entropy": 0.8809739574790001,
      "epoch": 0.17784164907710961,
      "grad_norm": 7.8125,
      "learning_rate": 4.680154588253831e-05,
      "loss": 0.8992229461669922,
      "mean_token_accuracy": 0.7830890074372292,
      "num_tokens": 19671214.0,
      "step": 9900
    },
    {
      "entropy": 0.9373166345059871,
      "epoch": 0.17802128710648044,
      "grad_norm": 5.8125,
      "learning_rate": 4.679451128705966e-05,
      "loss": 1.0422589302062988,
      "mean_token_accuracy": 0.7628077566623688,
      "num_tokens": 19690764.0,
      "step": 9910
    },
    {
      "entropy": 0.884357450157404,
      "epoch": 0.17820092513585126,
      "grad_norm": 6.8125,
      "learning_rate": 4.678746949410051e-05,
      "loss": 0.875083065032959,
      "mean_token_accuracy": 0.793388170003891,
      "num_tokens": 19710421.0,
      "step": 9920
    },
    {
      "entropy": 1.003797211498022,
      "epoch": 0.17838056316522208,
      "grad_norm": 6.40625,
      "learning_rate": 4.6780420505986364e-05,
      "loss": 1.0005805015563964,
      "mean_token_accuracy": 0.7616712465882302,
      "num_tokens": 19729586.0,
      "step": 9930
    },
    {
      "entropy": 0.9402335725724698,
      "epoch": 0.1785602011945929,
      "grad_norm": 5.375,
      "learning_rate": 4.67733643250451e-05,
      "loss": 1.0139225959777831,
      "mean_token_accuracy": 0.763370369374752,
      "num_tokens": 19750089.0,
      "step": 9940
    },
    {
      "entropy": 0.9041382670402527,
      "epoch": 0.17873983922396372,
      "grad_norm": 7.4375,
      "learning_rate": 4.6766300953606965e-05,
      "loss": 0.9334269523620605,
      "mean_token_accuracy": 0.7833434388041496,
      "num_tokens": 19769988.0,
      "step": 9950
    },
    {
      "entropy": 0.9178200788795948,
      "epoch": 0.17891947725333454,
      "grad_norm": 5.0625,
      "learning_rate": 4.6759230394004585e-05,
      "loss": 0.8636631011962891,
      "mean_token_accuracy": 0.7860185578465462,
      "num_tokens": 19789913.0,
      "step": 9960
    },
    {
      "entropy": 0.885036726295948,
      "epoch": 0.17909911528270536,
      "grad_norm": 6.96875,
      "learning_rate": 4.675215264857297e-05,
      "loss": 0.9375524520874023,
      "mean_token_accuracy": 0.7828322112560272,
      "num_tokens": 19809575.0,
      "step": 9970
    },
    {
      "entropy": 0.9421730369329453,
      "epoch": 0.17927875331207616,
      "grad_norm": 5.28125,
      "learning_rate": 4.674506771964948e-05,
      "loss": 0.963228988647461,
      "mean_token_accuracy": 0.7686885207891464,
      "num_tokens": 19829459.0,
      "step": 9980
    },
    {
      "entropy": 0.9244213327765465,
      "epoch": 0.17945839134144698,
      "grad_norm": 4.3125,
      "learning_rate": 4.6737975609573885e-05,
      "loss": 0.963899803161621,
      "mean_token_accuracy": 0.7784286767244339,
      "num_tokens": 19848553.0,
      "step": 9990
    },
    {
      "epoch": 0.1796380293708178,
      "eval_entropy": 0.8654658439159393,
      "eval_loss": 0.9178380370140076,
      "eval_mean_token_accuracy": 0.7827489011287689,
      "eval_num_tokens": 19868813.0,
      "eval_runtime": 40.5594,
      "eval_samples_per_second": 49.31,
      "eval_steps_per_second": 6.164,
      "step": 10000
    },
    {
      "entropy": 0.8970327652990818,
      "epoch": 0.17981766740018862,
      "grad_norm": 6.4375,
      "learning_rate": 4.672376985533718e-05,
      "loss": 0.9154790878295899,
      "mean_token_accuracy": 0.7790773957967758,
      "num_tokens": 19889070.0,
      "step": 10010
    },
    {
      "entropy": 0.8860844872891903,
      "epoch": 0.17999730542955944,
      "grad_norm": 5.5,
      "learning_rate": 4.6716656215867416e-05,
      "loss": 0.9297883033752441,
      "mean_token_accuracy": 0.7806318372488021,
      "num_tokens": 19909669.0,
      "step": 10020
    },
    {
      "entropy": 0.8904704309999942,
      "epoch": 0.18017694345893026,
      "grad_norm": 5.0625,
      "learning_rate": 4.6709535404628235e-05,
      "loss": 0.8861412048339844,
      "mean_token_accuracy": 0.7842861548066139,
      "num_tokens": 19929310.0,
      "step": 10030
    },
    {
      "entropy": 0.9421122550964356,
      "epoch": 0.18035658148830108,
      "grad_norm": 5.875,
      "learning_rate": 4.670240742397123e-05,
      "loss": 0.9656044006347656,
      "mean_token_accuracy": 0.7745575472712517,
      "num_tokens": 19949316.0,
      "step": 10040
    },
    {
      "entropy": 0.9575737029314041,
      "epoch": 0.1805362195176719,
      "grad_norm": 12.6875,
      "learning_rate": 4.669527227625036e-05,
      "loss": 0.9661746978759765,
      "mean_token_accuracy": 0.7781783565878868,
      "num_tokens": 19968635.0,
      "step": 10050
    },
    {
      "entropy": 0.8467212662100791,
      "epoch": 0.1807158575470427,
      "grad_norm": 6.75,
      "learning_rate": 4.6688129963821955e-05,
      "loss": 0.8847176551818847,
      "mean_token_accuracy": 0.7942148804664612,
      "num_tokens": 19987448.0,
      "step": 10060
    },
    {
      "entropy": 0.8960808664560318,
      "epoch": 0.18089549557641352,
      "grad_norm": 5.78125,
      "learning_rate": 4.668098048904472e-05,
      "loss": 0.9462654113769531,
      "mean_token_accuracy": 0.776871158182621,
      "num_tokens": 20007490.0,
      "step": 10070
    },
    {
      "entropy": 0.8117484986782074,
      "epoch": 0.18107513360578434,
      "grad_norm": 4.78125,
      "learning_rate": 4.667382385427971e-05,
      "loss": 0.810932731628418,
      "mean_token_accuracy": 0.8044671684503555,
      "num_tokens": 20027156.0,
      "step": 10080
    },
    {
      "entropy": 0.9018398560583591,
      "epoch": 0.18125477163515516,
      "grad_norm": 9.25,
      "learning_rate": 4.6666660061890355e-05,
      "loss": 0.9327230453491211,
      "mean_token_accuracy": 0.7876079708337784,
      "num_tokens": 20046362.0,
      "step": 10090
    },
    {
      "entropy": 0.869919040799141,
      "epoch": 0.18143440966452598,
      "grad_norm": 7.25,
      "learning_rate": 4.665948911424246e-05,
      "loss": 0.8856520652770996,
      "mean_token_accuracy": 0.7849762588739395,
      "num_tokens": 20066034.0,
      "step": 10100
    },
    {
      "entropy": 0.8612315021455288,
      "epoch": 0.1816140476938968,
      "grad_norm": 4.875,
      "learning_rate": 4.665231101370415e-05,
      "loss": 0.8797199249267578,
      "mean_token_accuracy": 0.7865251332521439,
      "num_tokens": 20086132.0,
      "step": 10110
    },
    {
      "entropy": 0.8215412184596061,
      "epoch": 0.18179368572326762,
      "grad_norm": 7.4375,
      "learning_rate": 4.664512576264596e-05,
      "loss": 0.8817085266113281,
      "mean_token_accuracy": 0.7978115871548652,
      "num_tokens": 20105458.0,
      "step": 10120
    },
    {
      "entropy": 0.8708537980914116,
      "epoch": 0.18197332375263844,
      "grad_norm": 5.1875,
      "learning_rate": 4.663793336344077e-05,
      "loss": 0.8897501945495605,
      "mean_token_accuracy": 0.7868052572011948,
      "num_tokens": 20124421.0,
      "step": 10130
    },
    {
      "entropy": 0.8606214374303818,
      "epoch": 0.18215296178200927,
      "grad_norm": 6.53125,
      "learning_rate": 4.663073381846381e-05,
      "loss": 0.8434826850891113,
      "mean_token_accuracy": 0.7889123961329461,
      "num_tokens": 20143074.0,
      "step": 10140
    },
    {
      "entropy": 0.8437647886574269,
      "epoch": 0.18233259981138006,
      "grad_norm": 5.9375,
      "learning_rate": 4.662352713009268e-05,
      "loss": 0.8886876106262207,
      "mean_token_accuracy": 0.7893482580780983,
      "num_tokens": 20162432.0,
      "step": 10150
    },
    {
      "entropy": 0.959248498827219,
      "epoch": 0.18251223784075088,
      "grad_norm": 7.96875,
      "learning_rate": 4.661631330070734e-05,
      "loss": 0.9278444290161133,
      "mean_token_accuracy": 0.782409805059433,
      "num_tokens": 20182124.0,
      "step": 10160
    },
    {
      "entropy": 0.8961891002953053,
      "epoch": 0.1826918758701217,
      "grad_norm": 6.15625,
      "learning_rate": 4.66090923326901e-05,
      "loss": 0.9482830047607422,
      "mean_token_accuracy": 0.7701358571648598,
      "num_tokens": 20201455.0,
      "step": 10170
    },
    {
      "entropy": 0.8502549394965172,
      "epoch": 0.18287151389949252,
      "grad_norm": 4.5,
      "learning_rate": 4.6601864228425636e-05,
      "loss": 0.9447774887084961,
      "mean_token_accuracy": 0.7777151077985763,
      "num_tokens": 20221383.0,
      "step": 10180
    },
    {
      "entropy": 0.9295808129012585,
      "epoch": 0.18305115192886334,
      "grad_norm": 6.15625,
      "learning_rate": 4.6594628990300973e-05,
      "loss": 0.9805933952331543,
      "mean_token_accuracy": 0.7741496138274669,
      "num_tokens": 20241710.0,
      "step": 10190
    },
    {
      "entropy": 0.9320115998387337,
      "epoch": 0.18323078995823416,
      "grad_norm": 6.125,
      "learning_rate": 4.65873866207055e-05,
      "loss": 0.9468727111816406,
      "mean_token_accuracy": 0.7794321238994598,
      "num_tokens": 20261922.0,
      "step": 10200
    },
    {
      "entropy": 0.9146525450050831,
      "epoch": 0.18341042798760498,
      "grad_norm": 4.96875,
      "learning_rate": 4.6580137122030965e-05,
      "loss": 0.9139816284179687,
      "mean_token_accuracy": 0.7824425846338272,
      "num_tokens": 20281555.0,
      "step": 10210
    },
    {
      "entropy": 0.9685615167021752,
      "epoch": 0.1835900660169758,
      "grad_norm": 8.5625,
      "learning_rate": 4.6572880496671444e-05,
      "loss": 1.0247797966003418,
      "mean_token_accuracy": 0.7732262901961804,
      "num_tokens": 20302381.0,
      "step": 10220
    },
    {
      "entropy": 0.9388386622071266,
      "epoch": 0.1837697040463466,
      "grad_norm": 5.1875,
      "learning_rate": 4.656561674702341e-05,
      "loss": 0.9076201438903808,
      "mean_token_accuracy": 0.7856024459004403,
      "num_tokens": 20322244.0,
      "step": 10230
    },
    {
      "entropy": 0.8644967660307884,
      "epoch": 0.18394934207571742,
      "grad_norm": 5.65625,
      "learning_rate": 4.6558345875485645e-05,
      "loss": 0.8878963470458985,
      "mean_token_accuracy": 0.7892375700175762,
      "num_tokens": 20341154.0,
      "step": 10240
    },
    {
      "entropy": 1.055906529724598,
      "epoch": 0.18412898010508824,
      "grad_norm": 7.375,
      "learning_rate": 4.65510678844593e-05,
      "loss": 1.0465203285217286,
      "mean_token_accuracy": 0.7560982845723629,
      "num_tokens": 20361875.0,
      "step": 10250
    },
    {
      "entropy": 0.8690978042781353,
      "epoch": 0.18430861813445906,
      "grad_norm": 6.625,
      "learning_rate": 4.6543782776347896e-05,
      "loss": 0.8669333457946777,
      "mean_token_accuracy": 0.805006018280983,
      "num_tokens": 20382291.0,
      "step": 10260
    },
    {
      "entropy": 0.846253851801157,
      "epoch": 0.18448825616382988,
      "grad_norm": 6.75,
      "learning_rate": 4.653649055355727e-05,
      "loss": 0.899083423614502,
      "mean_token_accuracy": 0.7814048767089844,
      "num_tokens": 20401817.0,
      "step": 10270
    },
    {
      "entropy": 0.8792815178632736,
      "epoch": 0.1846678941932007,
      "grad_norm": 5.46875,
      "learning_rate": 4.6529191218495636e-05,
      "loss": 0.9170869827270508,
      "mean_token_accuracy": 0.7820360198616981,
      "num_tokens": 20422421.0,
      "step": 10280
    },
    {
      "entropy": 0.8867128424346447,
      "epoch": 0.18484753222257153,
      "grad_norm": 4.90625,
      "learning_rate": 4.6521884773573556e-05,
      "loss": 0.9152216911315918,
      "mean_token_accuracy": 0.7900599136948585,
      "num_tokens": 20442493.0,
      "step": 10290
    },
    {
      "entropy": 0.9218067035079003,
      "epoch": 0.18502717025194235,
      "grad_norm": 6.28125,
      "learning_rate": 4.6514571221203905e-05,
      "loss": 0.8914688110351563,
      "mean_token_accuracy": 0.7879645988345146,
      "num_tokens": 20461711.0,
      "step": 10300
    },
    {
      "entropy": 0.8364969179034233,
      "epoch": 0.18520680828131317,
      "grad_norm": 6.375,
      "learning_rate": 4.650725056380195e-05,
      "loss": 0.8176088333129883,
      "mean_token_accuracy": 0.8037059500813484,
      "num_tokens": 20480729.0,
      "step": 10310
    },
    {
      "entropy": 0.8749627761542798,
      "epoch": 0.18538644631068396,
      "grad_norm": 6.84375,
      "learning_rate": 4.649992280378529e-05,
      "loss": 0.9180304527282714,
      "mean_token_accuracy": 0.7805110365152359,
      "num_tokens": 20500229.0,
      "step": 10320
    },
    {
      "entropy": 0.8077126622200013,
      "epoch": 0.18556608434005478,
      "grad_norm": 6.0625,
      "learning_rate": 4.649258794357385e-05,
      "loss": 0.854182243347168,
      "mean_token_accuracy": 0.7967185273766517,
      "num_tokens": 20520376.0,
      "step": 10330
    },
    {
      "entropy": 0.8610661804676056,
      "epoch": 0.1857457223694256,
      "grad_norm": 5.4375,
      "learning_rate": 4.648524598558992e-05,
      "loss": 0.8956315994262696,
      "mean_token_accuracy": 0.7910210475325584,
      "num_tokens": 20540065.0,
      "step": 10340
    },
    {
      "entropy": 0.9021451853215694,
      "epoch": 0.18592536039879642,
      "grad_norm": 7.78125,
      "learning_rate": 4.6477896932258136e-05,
      "loss": 0.9509499549865723,
      "mean_token_accuracy": 0.7669754132628441,
      "num_tokens": 20559944.0,
      "step": 10350
    },
    {
      "entropy": 0.9365390732884407,
      "epoch": 0.18610499842816725,
      "grad_norm": 4.9375,
      "learning_rate": 4.6470540786005454e-05,
      "loss": 0.9479766845703125,
      "mean_token_accuracy": 0.7729396760463715,
      "num_tokens": 20579411.0,
      "step": 10360
    },
    {
      "entropy": 0.9452847026288509,
      "epoch": 0.18628463645753807,
      "grad_norm": 4.5625,
      "learning_rate": 4.646317754926121e-05,
      "loss": 0.9594344139099121,
      "mean_token_accuracy": 0.7686725646257401,
      "num_tokens": 20598534.0,
      "step": 10370
    },
    {
      "entropy": 0.8386963322758675,
      "epoch": 0.1864642744869089,
      "grad_norm": 5.25,
      "learning_rate": 4.645580722445704e-05,
      "loss": 0.8169011116027832,
      "mean_token_accuracy": 0.8096451237797737,
      "num_tokens": 20619214.0,
      "step": 10380
    },
    {
      "entropy": 0.8733599707484245,
      "epoch": 0.1866439125162797,
      "grad_norm": 8.0625,
      "learning_rate": 4.644842981402695e-05,
      "loss": 0.8992310523986816,
      "mean_token_accuracy": 0.781498996913433,
      "num_tokens": 20638825.0,
      "step": 10390
    },
    {
      "entropy": 0.9407279476523399,
      "epoch": 0.1868235505456505,
      "grad_norm": 5.84375,
      "learning_rate": 4.6441045320407276e-05,
      "loss": 0.9471444129943848,
      "mean_token_accuracy": 0.7787516072392464,
      "num_tokens": 20657920.0,
      "step": 10400
    },
    {
      "entropy": 0.9154140867292881,
      "epoch": 0.18700318857502132,
      "grad_norm": 7.65625,
      "learning_rate": 4.6433653746036696e-05,
      "loss": 0.9598176956176758,
      "mean_token_accuracy": 0.7760003358125687,
      "num_tokens": 20677170.0,
      "step": 10410
    },
    {
      "entropy": 0.9073328129947186,
      "epoch": 0.18718282660439214,
      "grad_norm": 7.40625,
      "learning_rate": 4.642625509335622e-05,
      "loss": 0.9268363952636719,
      "mean_token_accuracy": 0.7803513795137406,
      "num_tokens": 20696038.0,
      "step": 10420
    },
    {
      "entropy": 0.8684262469410896,
      "epoch": 0.18736246463376297,
      "grad_norm": 5.46875,
      "learning_rate": 4.64188493648092e-05,
      "loss": 0.8868843078613281,
      "mean_token_accuracy": 0.7954951122403144,
      "num_tokens": 20716309.0,
      "step": 10430
    },
    {
      "entropy": 0.918014845252037,
      "epoch": 0.1875421026631338,
      "grad_norm": 6.34375,
      "learning_rate": 4.641143656284133e-05,
      "loss": 0.9020030975341797,
      "mean_token_accuracy": 0.7870040729641914,
      "num_tokens": 20735924.0,
      "step": 10440
    },
    {
      "entropy": 0.9092310547828675,
      "epoch": 0.1877217406925046,
      "grad_norm": 9.1875,
      "learning_rate": 4.640401668990063e-05,
      "loss": 0.9500041961669922,
      "mean_token_accuracy": 0.7762953087687492,
      "num_tokens": 20754844.0,
      "step": 10450
    },
    {
      "entropy": 0.9070415891706943,
      "epoch": 0.18790137872187543,
      "grad_norm": 5.4375,
      "learning_rate": 4.639658974843746e-05,
      "loss": 0.9467242240905762,
      "mean_token_accuracy": 0.7788534298539161,
      "num_tokens": 20773990.0,
      "step": 10460
    },
    {
      "entropy": 0.8963094502687454,
      "epoch": 0.18808101675124625,
      "grad_norm": 9.4375,
      "learning_rate": 4.638915574090451e-05,
      "loss": 0.9309506416320801,
      "mean_token_accuracy": 0.7770019412040711,
      "num_tokens": 20793639.0,
      "step": 10470
    },
    {
      "entropy": 1.0263412542641164,
      "epoch": 0.18826065478061707,
      "grad_norm": 6.96875,
      "learning_rate": 4.638171466975682e-05,
      "loss": 1.053122615814209,
      "mean_token_accuracy": 0.7624891124665737,
      "num_tokens": 20813716.0,
      "step": 10480
    },
    {
      "entropy": 0.9137916810810566,
      "epoch": 0.18844029280998786,
      "grad_norm": 7.09375,
      "learning_rate": 4.637426653745174e-05,
      "loss": 0.9340500831604004,
      "mean_token_accuracy": 0.7687035456299782,
      "num_tokens": 20834462.0,
      "step": 10490
    },
    {
      "entropy": 0.8767765007913113,
      "epoch": 0.18861993083935868,
      "grad_norm": 6.75,
      "learning_rate": 4.636681134644896e-05,
      "loss": 0.897179126739502,
      "mean_token_accuracy": 0.783487931638956,
      "num_tokens": 20854620.0,
      "step": 10500
    },
    {
      "entropy": 0.8499530285596848,
      "epoch": 0.1887995688687295,
      "grad_norm": 5.90625,
      "learning_rate": 4.6359349099210505e-05,
      "loss": 0.8415807723999024,
      "mean_token_accuracy": 0.7923376768827438,
      "num_tokens": 20874108.0,
      "step": 10510
    },
    {
      "entropy": 0.8776728577911854,
      "epoch": 0.18897920689810033,
      "grad_norm": 6.1875,
      "learning_rate": 4.6351879798200734e-05,
      "loss": 0.902304744720459,
      "mean_token_accuracy": 0.7939081758260726,
      "num_tokens": 20894464.0,
      "step": 10520
    },
    {
      "entropy": 0.8587953537702561,
      "epoch": 0.18915884492747115,
      "grad_norm": 5.71875,
      "learning_rate": 4.6344403445886334e-05,
      "loss": 0.9612351417541504,
      "mean_token_accuracy": 0.7729480594396592,
      "num_tokens": 20913768.0,
      "step": 10530
    },
    {
      "entropy": 0.9976669743657112,
      "epoch": 0.18933848295684197,
      "grad_norm": 6.5,
      "learning_rate": 4.63369200447363e-05,
      "loss": 1.0563933372497558,
      "mean_token_accuracy": 0.7569487959146499,
      "num_tokens": 20933343.0,
      "step": 10540
    },
    {
      "entropy": 0.923011091351509,
      "epoch": 0.1895181209862128,
      "grad_norm": 5.3125,
      "learning_rate": 4.632942959722198e-05,
      "loss": 0.9680864334106445,
      "mean_token_accuracy": 0.7764596194028854,
      "num_tokens": 20951847.0,
      "step": 10550
    },
    {
      "entropy": 0.9453989893198014,
      "epoch": 0.1896977590155836,
      "grad_norm": 7.15625,
      "learning_rate": 4.632193210581704e-05,
      "loss": 0.9448729515075683,
      "mean_token_accuracy": 0.7815139219164848,
      "num_tokens": 20971638.0,
      "step": 10560
    },
    {
      "entropy": 1.0035074219107627,
      "epoch": 0.1898773970449544,
      "grad_norm": 7.3125,
      "learning_rate": 4.631442757299748e-05,
      "loss": 1.0361597061157226,
      "mean_token_accuracy": 0.7540891155600548,
      "num_tokens": 20990794.0,
      "step": 10570
    },
    {
      "entropy": 0.9517493501305581,
      "epoch": 0.19005703507432523,
      "grad_norm": 8.0625,
      "learning_rate": 4.630691600124161e-05,
      "loss": 0.9442025184631347,
      "mean_token_accuracy": 0.7785703375935554,
      "num_tokens": 21011265.0,
      "step": 10580
    },
    {
      "entropy": 0.8879652462899685,
      "epoch": 0.19023667310369605,
      "grad_norm": 7.96875,
      "learning_rate": 4.6299397393030075e-05,
      "loss": 0.9373575210571289,
      "mean_token_accuracy": 0.7772521086037159,
      "num_tokens": 21031432.0,
      "step": 10590
    },
    {
      "entropy": 0.9365950301289558,
      "epoch": 0.19041631113306687,
      "grad_norm": 8.5625,
      "learning_rate": 4.629187175084584e-05,
      "loss": 0.9817299842834473,
      "mean_token_accuracy": 0.7737820461392403,
      "num_tokens": 21050058.0,
      "step": 10600
    },
    {
      "entropy": 0.9550454951822758,
      "epoch": 0.1905959491624377,
      "grad_norm": 5.21875,
      "learning_rate": 4.628433907717419e-05,
      "loss": 0.9584936141967774,
      "mean_token_accuracy": 0.7790491640567779,
      "num_tokens": 21070387.0,
      "step": 10610
    },
    {
      "entropy": 0.8827878668904304,
      "epoch": 0.1907755871918085,
      "grad_norm": 6.28125,
      "learning_rate": 4.6276799374502744e-05,
      "loss": 0.8645289421081543,
      "mean_token_accuracy": 0.7939226984977722,
      "num_tokens": 21090799.0,
      "step": 10620
    },
    {
      "entropy": 0.9570800736546516,
      "epoch": 0.19095522522117933,
      "grad_norm": 7.8125,
      "learning_rate": 4.6269252645321425e-05,
      "loss": 0.9892303466796875,
      "mean_token_accuracy": 0.7694451868534088,
      "num_tokens": 21111151.0,
      "step": 10630
    },
    {
      "entropy": 0.8552551865577698,
      "epoch": 0.19113486325055015,
      "grad_norm": 6.6875,
      "learning_rate": 4.6261698892122496e-05,
      "loss": 0.8567474365234375,
      "mean_token_accuracy": 0.79371307939291,
      "num_tokens": 21131310.0,
      "step": 10640
    },
    {
      "entropy": 0.8782812021672726,
      "epoch": 0.19131450127992095,
      "grad_norm": 5.3125,
      "learning_rate": 4.625413811740053e-05,
      "loss": 0.9345088005065918,
      "mean_token_accuracy": 0.7743212327361106,
      "num_tokens": 21152138.0,
      "step": 10650
    },
    {
      "entropy": 0.9206596046686173,
      "epoch": 0.19149413930929177,
      "grad_norm": 5.25,
      "learning_rate": 4.6246570323652404e-05,
      "loss": 0.958531379699707,
      "mean_token_accuracy": 0.7713661819696427,
      "num_tokens": 21171264.0,
      "step": 10660
    },
    {
      "entropy": 0.9246576435863971,
      "epoch": 0.1916737773386626,
      "grad_norm": 4.625,
      "learning_rate": 4.623899551337735e-05,
      "loss": 0.921489143371582,
      "mean_token_accuracy": 0.7835305579006672,
      "num_tokens": 21192167.0,
      "step": 10670
    },
    {
      "entropy": 0.959288053959608,
      "epoch": 0.1918534153680334,
      "grad_norm": 6.1875,
      "learning_rate": 4.6231413689076876e-05,
      "loss": 0.9701106071472168,
      "mean_token_accuracy": 0.770016510039568,
      "num_tokens": 21211220.0,
      "step": 10680
    },
    {
      "entropy": 0.8775778666138649,
      "epoch": 0.19203305339740423,
      "grad_norm": 9.6875,
      "learning_rate": 4.622382485325483e-05,
      "loss": 0.958392333984375,
      "mean_token_accuracy": 0.7752549842000007,
      "num_tokens": 21230874.0,
      "step": 10690
    },
    {
      "entropy": 0.9155103363096714,
      "epoch": 0.19221269142677505,
      "grad_norm": 5.375,
      "learning_rate": 4.621622900841738e-05,
      "loss": 0.9071863174438477,
      "mean_token_accuracy": 0.7879876345396042,
      "num_tokens": 21250996.0,
      "step": 10700
    },
    {
      "entropy": 0.9597749948501587,
      "epoch": 0.19239232945614587,
      "grad_norm": 7.9375,
      "learning_rate": 4.620862615707298e-05,
      "loss": 0.9698691368103027,
      "mean_token_accuracy": 0.7723872274160385,
      "num_tokens": 21270578.0,
      "step": 10710
    },
    {
      "entropy": 0.8198323681950569,
      "epoch": 0.1925719674855167,
      "grad_norm": 4.96875,
      "learning_rate": 4.620101630173243e-05,
      "loss": 0.7743246555328369,
      "mean_token_accuracy": 0.8123441755771637,
      "num_tokens": 21291284.0,
      "step": 10720
    },
    {
      "entropy": 0.8347475752234459,
      "epoch": 0.19275160551488751,
      "grad_norm": 4.78125,
      "learning_rate": 4.619339944490883e-05,
      "loss": 0.8893527984619141,
      "mean_token_accuracy": 0.7908472135663033,
      "num_tokens": 21310945.0,
      "step": 10730
    },
    {
      "entropy": 0.8385367393493652,
      "epoch": 0.1929312435442583,
      "grad_norm": 6.46875,
      "learning_rate": 4.6185775589117594e-05,
      "loss": 0.8540879249572754,
      "mean_token_accuracy": 0.7914605081081391,
      "num_tokens": 21331112.0,
      "step": 10740
    },
    {
      "entropy": 0.9175145834684372,
      "epoch": 0.19311088157362913,
      "grad_norm": 6.0,
      "learning_rate": 4.617814473687643e-05,
      "loss": 0.9755760192871094,
      "mean_token_accuracy": 0.7762549720704556,
      "num_tokens": 21351718.0,
      "step": 10750
    },
    {
      "entropy": 0.9445755057036876,
      "epoch": 0.19329051960299995,
      "grad_norm": 5.75,
      "learning_rate": 4.617050689070539e-05,
      "loss": 0.9090581893920898,
      "mean_token_accuracy": 0.7907473109662533,
      "num_tokens": 21370590.0,
      "step": 10760
    },
    {
      "entropy": 0.8613024458289147,
      "epoch": 0.19347015763237077,
      "grad_norm": 4.90625,
      "learning_rate": 4.616286205312681e-05,
      "loss": 0.854987621307373,
      "mean_token_accuracy": 0.791745288670063,
      "num_tokens": 21390147.0,
      "step": 10770
    },
    {
      "entropy": 0.7588732674717903,
      "epoch": 0.1936497956617416,
      "grad_norm": 7.125,
      "learning_rate": 4.615521022666535e-05,
      "loss": 0.8054081916809082,
      "mean_token_accuracy": 0.8077579349279403,
      "num_tokens": 21410136.0,
      "step": 10780
    },
    {
      "entropy": 0.9653812102973461,
      "epoch": 0.1938294336911124,
      "grad_norm": 7.375,
      "learning_rate": 4.6147551413847956e-05,
      "loss": 1.036971378326416,
      "mean_token_accuracy": 0.7621015064418316,
      "num_tokens": 21431212.0,
      "step": 10790
    },
    {
      "entropy": 0.98788276091218,
      "epoch": 0.19400907172048323,
      "grad_norm": 10.8125,
      "learning_rate": 4.613988561720391e-05,
      "loss": 0.9979410171508789,
      "mean_token_accuracy": 0.7652235180139542,
      "num_tokens": 21450484.0,
      "step": 10800
    },
    {
      "entropy": 0.8375085666775703,
      "epoch": 0.19418870974985405,
      "grad_norm": 6.125,
      "learning_rate": 4.6132212839264765e-05,
      "loss": 0.8567901611328125,
      "mean_token_accuracy": 0.7948850840330124,
      "num_tokens": 21471132.0,
      "step": 10810
    },
    {
      "entropy": 0.8053878769278526,
      "epoch": 0.19436834777922485,
      "grad_norm": 4.0,
      "learning_rate": 4.6124533082564434e-05,
      "loss": 0.8006709098815918,
      "mean_token_accuracy": 0.8106097310781479,
      "num_tokens": 21491697.0,
      "step": 10820
    },
    {
      "entropy": 1.0332276977598667,
      "epoch": 0.19454798580859567,
      "grad_norm": 7.53125,
      "learning_rate": 4.611684634963908e-05,
      "loss": 1.1347893714904784,
      "mean_token_accuracy": 0.7474535003304481,
      "num_tokens": 21512177.0,
      "step": 10830
    },
    {
      "entropy": 0.8964815348386764,
      "epoch": 0.1947276238379665,
      "grad_norm": 6.6875,
      "learning_rate": 4.610915264302719e-05,
      "loss": 0.8832280158996582,
      "mean_token_accuracy": 0.7851007223129273,
      "num_tokens": 21532670.0,
      "step": 10840
    },
    {
      "entropy": 1.0005129784345628,
      "epoch": 0.1949072618673373,
      "grad_norm": 8.125,
      "learning_rate": 4.610145196526956e-05,
      "loss": 1.020705795288086,
      "mean_token_accuracy": 0.7706126257777214,
      "num_tokens": 21552174.0,
      "step": 10850
    },
    {
      "entropy": 0.820348034799099,
      "epoch": 0.19508689989670813,
      "grad_norm": 6.96875,
      "learning_rate": 4.609374431890928e-05,
      "loss": 0.8059733390808106,
      "mean_token_accuracy": 0.803258153796196,
      "num_tokens": 21571350.0,
      "step": 10860
    },
    {
      "entropy": 0.8472366109490395,
      "epoch": 0.19526653792607895,
      "grad_norm": 5.40625,
      "learning_rate": 4.608602970649175e-05,
      "loss": 0.8730942726135253,
      "mean_token_accuracy": 0.7821216896176338,
      "num_tokens": 21590937.0,
      "step": 10870
    },
    {
      "entropy": 0.826553588360548,
      "epoch": 0.19544617595544977,
      "grad_norm": 8.75,
      "learning_rate": 4.6078308130564665e-05,
      "loss": 0.8561902046203613,
      "mean_token_accuracy": 0.7990699172019958,
      "num_tokens": 21612347.0,
      "step": 10880
    },
    {
      "entropy": 0.8690866716206074,
      "epoch": 0.1956258139848206,
      "grad_norm": 5.25,
      "learning_rate": 4.607057959367802e-05,
      "loss": 0.8949705123901367,
      "mean_token_accuracy": 0.7854206740856171,
      "num_tokens": 21631843.0,
      "step": 10890
    },
    {
      "entropy": 0.9460312694311142,
      "epoch": 0.19580545201419142,
      "grad_norm": 6.03125,
      "learning_rate": 4.60628440983841e-05,
      "loss": 1.0245882987976074,
      "mean_token_accuracy": 0.7560961097478867,
      "num_tokens": 21650983.0,
      "step": 10900
    },
    {
      "entropy": 0.9562560014426709,
      "epoch": 0.1959850900435622,
      "grad_norm": 4.34375,
      "learning_rate": 4.605510164723751e-05,
      "loss": 0.9722962379455566,
      "mean_token_accuracy": 0.7684106647968292,
      "num_tokens": 21670621.0,
      "step": 10910
    },
    {
      "entropy": 0.9785911172628403,
      "epoch": 0.19616472807293303,
      "grad_norm": 7.34375,
      "learning_rate": 4.604735224279513e-05,
      "loss": 0.9742871284484863,
      "mean_token_accuracy": 0.7737636178731918,
      "num_tokens": 21690360.0,
      "step": 10920
    },
    {
      "entropy": 0.8847896374762059,
      "epoch": 0.19634436610230385,
      "grad_norm": 6.9375,
      "learning_rate": 4.603959588761615e-05,
      "loss": 0.8559576988220214,
      "mean_token_accuracy": 0.7961952418088913,
      "num_tokens": 21709743.0,
      "step": 10930
    },
    {
      "entropy": 0.8677625961601734,
      "epoch": 0.19652400413167467,
      "grad_norm": 7.21875,
      "learning_rate": 4.603183258426204e-05,
      "loss": 0.9095634460449219,
      "mean_token_accuracy": 0.78458052277565,
      "num_tokens": 21730160.0,
      "step": 10940
    },
    {
      "entropy": 0.901097447425127,
      "epoch": 0.1967036421610455,
      "grad_norm": 6.9375,
      "learning_rate": 4.602406233529659e-05,
      "loss": 0.9543604850769043,
      "mean_token_accuracy": 0.776050640642643,
      "num_tokens": 21749938.0,
      "step": 10950
    },
    {
      "entropy": 0.9408169202506542,
      "epoch": 0.19688328019041632,
      "grad_norm": 7.0625,
      "learning_rate": 4.6016285143285856e-05,
      "loss": 0.9706070899963379,
      "mean_token_accuracy": 0.7788687959313393,
      "num_tokens": 21770002.0,
      "step": 10960
    },
    {
      "entropy": 0.8664716377854347,
      "epoch": 0.19706291821978714,
      "grad_norm": 10.0625,
      "learning_rate": 4.600850101079821e-05,
      "loss": 0.8894480705261231,
      "mean_token_accuracy": 0.7916534319519997,
      "num_tokens": 21789813.0,
      "step": 10970
    },
    {
      "entropy": 0.8398127883672715,
      "epoch": 0.19724255624915796,
      "grad_norm": 6.21875,
      "learning_rate": 4.6000709940404296e-05,
      "loss": 0.8119674682617187,
      "mean_token_accuracy": 0.801940806210041,
      "num_tokens": 21808766.0,
      "step": 10980
    },
    {
      "entropy": 0.8680760376155376,
      "epoch": 0.19742219427852875,
      "grad_norm": 6.625,
      "learning_rate": 4.599291193467706e-05,
      "loss": 0.9252339363098144,
      "mean_token_accuracy": 0.776945061981678,
      "num_tokens": 21829032.0,
      "step": 10990
    },
    {
      "epoch": 0.19760183230789957,
      "eval_entropy": 0.8788281654119492,
      "eval_loss": 0.8879156708717346,
      "eval_mean_token_accuracy": 0.7868806464672089,
      "eval_num_tokens": 21848854.0,
      "eval_runtime": 40.5208,
      "eval_samples_per_second": 49.357,
      "eval_steps_per_second": 6.17,
      "step": 11000
    },
    {
      "entropy": 0.9209602098912001,
      "epoch": 0.1977814703372704,
      "grad_norm": 5.59375,
      "learning_rate": 4.5977295127525876e-05,
      "loss": 0.9637662887573242,
      "mean_token_accuracy": 0.7798609986901284,
      "num_tokens": 21867872.0,
      "step": 11010
    },
    {
      "entropy": 0.9289896786212921,
      "epoch": 0.19796110836664121,
      "grad_norm": 5.75,
      "learning_rate": 4.5969476331259266e-05,
      "loss": 0.8969556808471679,
      "mean_token_accuracy": 0.7833624616265297,
      "num_tokens": 21887621.0,
      "step": 11020
    },
    {
      "entropy": 0.9317864462733269,
      "epoch": 0.19814074639601204,
      "grad_norm": 5.875,
      "learning_rate": 4.5961650609974006e-05,
      "loss": 0.9004011154174805,
      "mean_token_accuracy": 0.7834533080458641,
      "num_tokens": 21907573.0,
      "step": 11030
    },
    {
      "entropy": 0.9031454622745514,
      "epoch": 0.19832038442538286,
      "grad_norm": 5.40625,
      "learning_rate": 4.595381796625449e-05,
      "loss": 0.9555165290832519,
      "mean_token_accuracy": 0.7760496199131012,
      "num_tokens": 21926828.0,
      "step": 11040
    },
    {
      "entropy": 0.9250333406031132,
      "epoch": 0.19850002245475368,
      "grad_norm": 9.75,
      "learning_rate": 4.5945978402687395e-05,
      "loss": 0.9731258392333985,
      "mean_token_accuracy": 0.7761249348521233,
      "num_tokens": 21945502.0,
      "step": 11050
    },
    {
      "entropy": 0.8963647350668907,
      "epoch": 0.1986796604841245,
      "grad_norm": 11.125,
      "learning_rate": 4.593813192186168e-05,
      "loss": 0.9551509857177735,
      "mean_token_accuracy": 0.7708211421966553,
      "num_tokens": 21964156.0,
      "step": 11060
    },
    {
      "entropy": 0.8488388136029243,
      "epoch": 0.19885929851349532,
      "grad_norm": 6.4375,
      "learning_rate": 4.5930278526368584e-05,
      "loss": 0.8046700477600097,
      "mean_token_accuracy": 0.8030500337481499,
      "num_tokens": 21983577.0,
      "step": 11070
    },
    {
      "entropy": 0.9691312938928605,
      "epoch": 0.1990389365428661,
      "grad_norm": 6.71875,
      "learning_rate": 4.592241821880163e-05,
      "loss": 0.9761872291564941,
      "mean_token_accuracy": 0.7622840806841851,
      "num_tokens": 22004651.0,
      "step": 11080
    },
    {
      "entropy": 0.8757909446954727,
      "epoch": 0.19921857457223693,
      "grad_norm": 7.21875,
      "learning_rate": 4.5914551001756646e-05,
      "loss": 0.9096345901489258,
      "mean_token_accuracy": 0.7778163269162178,
      "num_tokens": 22024582.0,
      "step": 11090
    },
    {
      "entropy": 0.9655227281153203,
      "epoch": 0.19939821260160775,
      "grad_norm": 4.625,
      "learning_rate": 4.590667687783171e-05,
      "loss": 0.9641009330749511,
      "mean_token_accuracy": 0.7781957976520062,
      "num_tokens": 22044410.0,
      "step": 11100
    },
    {
      "entropy": 0.8858931280672551,
      "epoch": 0.19957785063097858,
      "grad_norm": 6.84375,
      "learning_rate": 4.5898795849627194e-05,
      "loss": 0.8693421363830567,
      "mean_token_accuracy": 0.7945321410894394,
      "num_tokens": 22064626.0,
      "step": 11110
    },
    {
      "entropy": 0.7739857137203217,
      "epoch": 0.1997574886603494,
      "grad_norm": 8.6875,
      "learning_rate": 4.589090791974576e-05,
      "loss": 0.7859329223632813,
      "mean_token_accuracy": 0.8048031345009804,
      "num_tokens": 22084977.0,
      "step": 11120
    },
    {
      "entropy": 0.825855728238821,
      "epoch": 0.19993712668972022,
      "grad_norm": 7.1875,
      "learning_rate": 4.588301309079234e-05,
      "loss": 0.9059236526489258,
      "mean_token_accuracy": 0.7829899162054061,
      "num_tokens": 22105204.0,
      "step": 11130
    },
    {
      "entropy": 0.7788939066231251,
      "epoch": 0.20011676471909104,
      "grad_norm": 5.8125,
      "learning_rate": 4.5875111365374134e-05,
      "loss": 0.7993866920471191,
      "mean_token_accuracy": 0.7998276114463806,
      "num_tokens": 22125193.0,
      "step": 11140
    },
    {
      "entropy": 0.8257860995829105,
      "epoch": 0.20029640274846186,
      "grad_norm": 4.78125,
      "learning_rate": 4.5867202746100636e-05,
      "loss": 0.8266702651977539,
      "mean_token_accuracy": 0.797323040664196,
      "num_tokens": 22144556.0,
      "step": 11150
    },
    {
      "entropy": 0.9007250122725964,
      "epoch": 0.20047604077783265,
      "grad_norm": 7.40625,
      "learning_rate": 4.58592872355836e-05,
      "loss": 0.9284053802490234,
      "mean_token_accuracy": 0.7715675190091134,
      "num_tokens": 22164655.0,
      "step": 11160
    },
    {
      "entropy": 0.8784724034368991,
      "epoch": 0.20065567880720347,
      "grad_norm": 5.34375,
      "learning_rate": 4.585136483643709e-05,
      "loss": 0.882784366607666,
      "mean_token_accuracy": 0.7878041610121727,
      "num_tokens": 22183691.0,
      "step": 11170
    },
    {
      "entropy": 0.8141251787543297,
      "epoch": 0.2008353168365743,
      "grad_norm": 6.6875,
      "learning_rate": 4.5843435551277405e-05,
      "loss": 0.8255696296691895,
      "mean_token_accuracy": 0.8025176003575325,
      "num_tokens": 22203634.0,
      "step": 11180
    },
    {
      "entropy": 0.8806927412748337,
      "epoch": 0.20101495486594512,
      "grad_norm": 4.875,
      "learning_rate": 4.583549938272313e-05,
      "loss": 0.8752194404602051,
      "mean_token_accuracy": 0.7877620980143547,
      "num_tokens": 22222963.0,
      "step": 11190
    },
    {
      "entropy": 0.8817830346524715,
      "epoch": 0.20119459289531594,
      "grad_norm": 5.5625,
      "learning_rate": 4.582755633339514e-05,
      "loss": 0.9269624710083008,
      "mean_token_accuracy": 0.7841905288398265,
      "num_tokens": 22243710.0,
      "step": 11200
    },
    {
      "entropy": 0.9529386624693871,
      "epoch": 0.20137423092468676,
      "grad_norm": 8.25,
      "learning_rate": 4.581960640591656e-05,
      "loss": 0.9944866180419922,
      "mean_token_accuracy": 0.7648677222430706,
      "num_tokens": 22263922.0,
      "step": 11210
    },
    {
      "entropy": 0.9114404521882534,
      "epoch": 0.20155386895405758,
      "grad_norm": 5.46875,
      "learning_rate": 4.58116496029128e-05,
      "loss": 0.9124053955078125,
      "mean_token_accuracy": 0.7908343091607094,
      "num_tokens": 22284002.0,
      "step": 11220
    },
    {
      "entropy": 0.8170178420841694,
      "epoch": 0.2017335069834284,
      "grad_norm": 5.53125,
      "learning_rate": 4.580368592701155e-05,
      "loss": 0.8187215805053711,
      "mean_token_accuracy": 0.8034049019217491,
      "num_tokens": 22304298.0,
      "step": 11230
    },
    {
      "entropy": 0.7807209081947803,
      "epoch": 0.20191314501279922,
      "grad_norm": 4.9375,
      "learning_rate": 4.579571538084273e-05,
      "loss": 0.7871910095214844,
      "mean_token_accuracy": 0.8143251970410347,
      "num_tokens": 22323858.0,
      "step": 11240
    },
    {
      "entropy": 0.8423394210636616,
      "epoch": 0.20209278304217002,
      "grad_norm": 5.5625,
      "learning_rate": 4.578773796703858e-05,
      "loss": 0.8645644187927246,
      "mean_token_accuracy": 0.7933366313576699,
      "num_tokens": 22344688.0,
      "step": 11250
    },
    {
      "entropy": 0.9253683909773827,
      "epoch": 0.20227242107154084,
      "grad_norm": 9.3125,
      "learning_rate": 4.5779753688233565e-05,
      "loss": 1.0047744750976562,
      "mean_token_accuracy": 0.7710051968693733,
      "num_tokens": 22366475.0,
      "step": 11260
    },
    {
      "entropy": 0.9232454262673855,
      "epoch": 0.20245205910091166,
      "grad_norm": 5.375,
      "learning_rate": 4.577176254706445e-05,
      "loss": 0.9102578163146973,
      "mean_token_accuracy": 0.7746404260396957,
      "num_tokens": 22387377.0,
      "step": 11270
    },
    {
      "entropy": 0.842565505951643,
      "epoch": 0.20263169713028248,
      "grad_norm": 6.03125,
      "learning_rate": 4.576376454617024e-05,
      "loss": 0.8164196014404297,
      "mean_token_accuracy": 0.8009149223566056,
      "num_tokens": 22406862.0,
      "step": 11280
    },
    {
      "entropy": 0.9664469242095948,
      "epoch": 0.2028113351596533,
      "grad_norm": 6.03125,
      "learning_rate": 4.575575968819223e-05,
      "loss": 1.031270694732666,
      "mean_token_accuracy": 0.7514821663498878,
      "num_tokens": 22427367.0,
      "step": 11290
    },
    {
      "entropy": 0.9795466922223568,
      "epoch": 0.20299097318902412,
      "grad_norm": 8.4375,
      "learning_rate": 4.574774797577395e-05,
      "loss": 1.0498817443847657,
      "mean_token_accuracy": 0.7623425647616386,
      "num_tokens": 22446980.0,
      "step": 11300
    },
    {
      "entropy": 0.8971272014081478,
      "epoch": 0.20317061121839494,
      "grad_norm": 7.03125,
      "learning_rate": 4.5739729411561225e-05,
      "loss": 0.859313678741455,
      "mean_token_accuracy": 0.7944761261343956,
      "num_tokens": 22466988.0,
      "step": 11310
    },
    {
      "entropy": 0.9614310465753079,
      "epoch": 0.20335024924776576,
      "grad_norm": 5.25,
      "learning_rate": 4.573170399820213e-05,
      "loss": 0.9974462509155273,
      "mean_token_accuracy": 0.7657989799976349,
      "num_tokens": 22487321.0,
      "step": 11320
    },
    {
      "entropy": 0.8890015535056591,
      "epoch": 0.20352988727713656,
      "grad_norm": 6.09375,
      "learning_rate": 4.572367173834698e-05,
      "loss": 0.8796669960021972,
      "mean_token_accuracy": 0.7878701902925969,
      "num_tokens": 22506986.0,
      "step": 11330
    },
    {
      "entropy": 0.8429915435612202,
      "epoch": 0.20370952530650738,
      "grad_norm": 5.1875,
      "learning_rate": 4.5715632634648395e-05,
      "loss": 0.8574431419372559,
      "mean_token_accuracy": 0.7864844635128975,
      "num_tokens": 22526077.0,
      "step": 11340
    },
    {
      "entropy": 0.8833885349333286,
      "epoch": 0.2038891633358782,
      "grad_norm": 5.8125,
      "learning_rate": 4.5707586689761214e-05,
      "loss": 0.9213557243347168,
      "mean_token_accuracy": 0.7824743866920472,
      "num_tokens": 22545477.0,
      "step": 11350
    },
    {
      "entropy": 0.8574072010815144,
      "epoch": 0.20406880136524902,
      "grad_norm": 5.34375,
      "learning_rate": 4.569953390634257e-05,
      "loss": 0.8702330589294434,
      "mean_token_accuracy": 0.7950220510363579,
      "num_tokens": 22565993.0,
      "step": 11360
    },
    {
      "entropy": 0.883940476924181,
      "epoch": 0.20424843939461984,
      "grad_norm": 4.59375,
      "learning_rate": 4.569147428705182e-05,
      "loss": 0.9097959518432617,
      "mean_token_accuracy": 0.7870344892144203,
      "num_tokens": 22586012.0,
      "step": 11370
    },
    {
      "entropy": 0.966283518821001,
      "epoch": 0.20442807742399066,
      "grad_norm": 8.0,
      "learning_rate": 4.568340783455059e-05,
      "loss": 1.0267705917358398,
      "mean_token_accuracy": 0.7695122942328453,
      "num_tokens": 22607119.0,
      "step": 11380
    },
    {
      "entropy": 0.9584083452820777,
      "epoch": 0.20460771545336148,
      "grad_norm": 6.5,
      "learning_rate": 4.567533455150279e-05,
      "loss": 0.9550557136535645,
      "mean_token_accuracy": 0.771448640525341,
      "num_tokens": 22626702.0,
      "step": 11390
    },
    {
      "entropy": 0.9120221249759197,
      "epoch": 0.2047873534827323,
      "grad_norm": 6.75,
      "learning_rate": 4.566725444057455e-05,
      "loss": 0.9075155258178711,
      "mean_token_accuracy": 0.7838271245360374,
      "num_tokens": 22646645.0,
      "step": 11400
    },
    {
      "entropy": 0.8010118879377842,
      "epoch": 0.20496699151210312,
      "grad_norm": 6.0625,
      "learning_rate": 4.5659167504434264e-05,
      "loss": 0.8274785995483398,
      "mean_token_accuracy": 0.7890014976263047,
      "num_tokens": 22666564.0,
      "step": 11410
    },
    {
      "entropy": 0.8687062926590443,
      "epoch": 0.20514662954147392,
      "grad_norm": 7.09375,
      "learning_rate": 4.56510737457526e-05,
      "loss": 0.9017355918884278,
      "mean_token_accuracy": 0.7844808578491211,
      "num_tokens": 22687941.0,
      "step": 11420
    },
    {
      "entropy": 0.885517680644989,
      "epoch": 0.20532626757084474,
      "grad_norm": 7.0625,
      "learning_rate": 4.564297316720245e-05,
      "loss": 0.9004813194274902,
      "mean_token_accuracy": 0.7821603134274483,
      "num_tokens": 22707713.0,
      "step": 11430
    },
    {
      "entropy": 0.8573585227131844,
      "epoch": 0.20550590560021556,
      "grad_norm": 5.8125,
      "learning_rate": 4.563486577145898e-05,
      "loss": 0.9219475746154785,
      "mean_token_accuracy": 0.7881103307008743,
      "num_tokens": 22728096.0,
      "step": 11440
    },
    {
      "entropy": 0.9365617819130421,
      "epoch": 0.20568554362958638,
      "grad_norm": 6.9375,
      "learning_rate": 4.562675156119959e-05,
      "loss": 0.9951620101928711,
      "mean_token_accuracy": 0.7604023084044457,
      "num_tokens": 22746751.0,
      "step": 11450
    },
    {
      "entropy": 0.8300738863646984,
      "epoch": 0.2058651816589572,
      "grad_norm": 6.0,
      "learning_rate": 4.561863053910395e-05,
      "loss": 0.8131514549255371,
      "mean_token_accuracy": 0.8002211332321167,
      "num_tokens": 22766279.0,
      "step": 11460
    },
    {
      "entropy": 0.9336564019322395,
      "epoch": 0.20604481968832802,
      "grad_norm": 5.9375,
      "learning_rate": 4.561050270785396e-05,
      "loss": 0.9160529136657715,
      "mean_token_accuracy": 0.7857787400484085,
      "num_tokens": 22785884.0,
      "step": 11470
    },
    {
      "entropy": 0.8281168967485428,
      "epoch": 0.20622445771769884,
      "grad_norm": 7.5625,
      "learning_rate": 4.5602368070133794e-05,
      "loss": 0.7916419982910157,
      "mean_token_accuracy": 0.8017245307564735,
      "num_tokens": 22806077.0,
      "step": 11480
    },
    {
      "entropy": 0.9007447838783265,
      "epoch": 0.20640409574706967,
      "grad_norm": 7.84375,
      "learning_rate": 4.5594226628629834e-05,
      "loss": 0.8879066467285156,
      "mean_token_accuracy": 0.786144457757473,
      "num_tokens": 22825538.0,
      "step": 11490
    },
    {
      "entropy": 0.7842489652335644,
      "epoch": 0.20658373377644046,
      "grad_norm": 5.59375,
      "learning_rate": 4.558607838603075e-05,
      "loss": 0.8031447410583497,
      "mean_token_accuracy": 0.8001901626586914,
      "num_tokens": 22846178.0,
      "step": 11500
    },
    {
      "entropy": 0.8617790050804615,
      "epoch": 0.20676337180581128,
      "grad_norm": 6.9375,
      "learning_rate": 4.557792334502743e-05,
      "loss": 0.9158389091491699,
      "mean_token_accuracy": 0.7877695441246033,
      "num_tokens": 22865653.0,
      "step": 11510
    },
    {
      "entropy": 0.8902119196951389,
      "epoch": 0.2069430098351821,
      "grad_norm": 7.5,
      "learning_rate": 4.556976150831302e-05,
      "loss": 0.9528593063354492,
      "mean_token_accuracy": 0.7729435041546822,
      "num_tokens": 22884719.0,
      "step": 11520
    },
    {
      "entropy": 0.8985680162906646,
      "epoch": 0.20712264786455292,
      "grad_norm": 6.0,
      "learning_rate": 4.556159287858292e-05,
      "loss": 0.9034319877624511,
      "mean_token_accuracy": 0.7807430028915405,
      "num_tokens": 22905792.0,
      "step": 11530
    },
    {
      "entropy": 0.859601828455925,
      "epoch": 0.20730228589392374,
      "grad_norm": 5.53125,
      "learning_rate": 4.555341745853474e-05,
      "loss": 0.8416665077209473,
      "mean_token_accuracy": 0.7908956751227378,
      "num_tokens": 22926267.0,
      "step": 11540
    },
    {
      "entropy": 0.9283218637108803,
      "epoch": 0.20748192392329456,
      "grad_norm": 6.21875,
      "learning_rate": 4.5545235250868365e-05,
      "loss": 0.9314607620239258,
      "mean_token_accuracy": 0.7908827289938927,
      "num_tokens": 22947863.0,
      "step": 11550
    },
    {
      "entropy": 0.8506497517228127,
      "epoch": 0.20766156195266539,
      "grad_norm": 7.3125,
      "learning_rate": 4.553704625828592e-05,
      "loss": 0.8783835411071778,
      "mean_token_accuracy": 0.7937017478048801,
      "num_tokens": 22969396.0,
      "step": 11560
    },
    {
      "entropy": 0.8327492535114288,
      "epoch": 0.2078411999820362,
      "grad_norm": 6.15625,
      "learning_rate": 4.5528850483491746e-05,
      "loss": 0.8268207550048828,
      "mean_token_accuracy": 0.7945083320140839,
      "num_tokens": 22988583.0,
      "step": 11570
    },
    {
      "entropy": 0.9341848261654377,
      "epoch": 0.20802083801140703,
      "grad_norm": 5.6875,
      "learning_rate": 4.552064792919245e-05,
      "loss": 0.9600367546081543,
      "mean_token_accuracy": 0.7718945637345314,
      "num_tokens": 23009288.0,
      "step": 11580
    },
    {
      "entropy": 0.7509714752435684,
      "epoch": 0.20820047604077782,
      "grad_norm": 7.1875,
      "learning_rate": 4.5512438598096854e-05,
      "loss": 0.7658008098602295,
      "mean_token_accuracy": 0.8125150874257088,
      "num_tokens": 23030198.0,
      "step": 11590
    },
    {
      "entropy": 0.8244590133428573,
      "epoch": 0.20838011407014864,
      "grad_norm": 5.65625,
      "learning_rate": 4.550422249291603e-05,
      "loss": 0.8507304191589355,
      "mean_token_accuracy": 0.7967900648713112,
      "num_tokens": 23049564.0,
      "step": 11600
    },
    {
      "entropy": 0.8053080555051565,
      "epoch": 0.20855975209951946,
      "grad_norm": 5.78125,
      "learning_rate": 4.54959996163633e-05,
      "loss": 0.8656981468200684,
      "mean_token_accuracy": 0.787071841955185,
      "num_tokens": 23069041.0,
      "step": 11610
    },
    {
      "entropy": 0.803099624067545,
      "epoch": 0.20873939012889028,
      "grad_norm": 6.46875,
      "learning_rate": 4.5487769971154205e-05,
      "loss": 0.7804460048675537,
      "mean_token_accuracy": 0.8118880361318588,
      "num_tokens": 23089370.0,
      "step": 11620
    },
    {
      "entropy": 0.9415870226919651,
      "epoch": 0.2089190281582611,
      "grad_norm": 7.53125,
      "learning_rate": 4.547953356000652e-05,
      "loss": 0.9741015434265137,
      "mean_token_accuracy": 0.7798790514469147,
      "num_tokens": 23109675.0,
      "step": 11630
    },
    {
      "entropy": 0.8991717837750912,
      "epoch": 0.20909866618763193,
      "grad_norm": 6.84375,
      "learning_rate": 4.5471290385640276e-05,
      "loss": 0.93502197265625,
      "mean_token_accuracy": 0.782210287451744,
      "num_tokens": 23130478.0,
      "step": 11640
    },
    {
      "entropy": 0.798111055791378,
      "epoch": 0.20927830421700275,
      "grad_norm": 5.8125,
      "learning_rate": 4.54630404507777e-05,
      "loss": 0.8036151885986328,
      "mean_token_accuracy": 0.7982562229037284,
      "num_tokens": 23151515.0,
      "step": 11650
    },
    {
      "entropy": 0.9028390303254128,
      "epoch": 0.20945794224637357,
      "grad_norm": 5.375,
      "learning_rate": 4.5454783758143295e-05,
      "loss": 0.9040911674499512,
      "mean_token_accuracy": 0.7812075838446617,
      "num_tokens": 23172837.0,
      "step": 11660
    },
    {
      "entropy": 0.8002609081566334,
      "epoch": 0.20963758027574436,
      "grad_norm": 5.40625,
      "learning_rate": 4.544652031046376e-05,
      "loss": 0.7831881523132325,
      "mean_token_accuracy": 0.8047621518373489,
      "num_tokens": 23193080.0,
      "step": 11670
    },
    {
      "entropy": 0.9469489753246307,
      "epoch": 0.20981721830511518,
      "grad_norm": 6.125,
      "learning_rate": 4.5438250110468036e-05,
      "loss": 0.9466521263122558,
      "mean_token_accuracy": 0.7654821455478669,
      "num_tokens": 23213176.0,
      "step": 11680
    },
    {
      "entropy": 0.796565118432045,
      "epoch": 0.209996856334486,
      "grad_norm": 5.46875,
      "learning_rate": 4.542997316088732e-05,
      "loss": 0.8797205924987793,
      "mean_token_accuracy": 0.7918878197669983,
      "num_tokens": 23233369.0,
      "step": 11690
    },
    {
      "entropy": 0.9146947503089905,
      "epoch": 0.21017649436385683,
      "grad_norm": 7.71875,
      "learning_rate": 4.542168946445499e-05,
      "loss": 0.9108059883117676,
      "mean_token_accuracy": 0.7837534680962562,
      "num_tokens": 23252571.0,
      "step": 11700
    },
    {
      "entropy": 0.8756265386939048,
      "epoch": 0.21035613239322765,
      "grad_norm": 4.65625,
      "learning_rate": 4.541339902390669e-05,
      "loss": 0.8936793327331543,
      "mean_token_accuracy": 0.7886108085513115,
      "num_tokens": 23272706.0,
      "step": 11710
    },
    {
      "entropy": 0.8471706986427308,
      "epoch": 0.21053577042259847,
      "grad_norm": 5.1875,
      "learning_rate": 4.5405101841980266e-05,
      "loss": 0.8751951217651367,
      "mean_token_accuracy": 0.7870098739862442,
      "num_tokens": 23293553.0,
      "step": 11720
    },
    {
      "entropy": 0.8902305915951729,
      "epoch": 0.2107154084519693,
      "grad_norm": 4.46875,
      "learning_rate": 4.539679792141582e-05,
      "loss": 0.869639778137207,
      "mean_token_accuracy": 0.7864952832460403,
      "num_tokens": 23312878.0,
      "step": 11730
    },
    {
      "entropy": 0.8602326273918152,
      "epoch": 0.2108950464813401,
      "grad_norm": 6.78125,
      "learning_rate": 4.538848726495566e-05,
      "loss": 0.888673210144043,
      "mean_token_accuracy": 0.782754035294056,
      "num_tokens": 23333078.0,
      "step": 11740
    },
    {
      "entropy": 0.8880679443478584,
      "epoch": 0.21107468451071093,
      "grad_norm": 6.875,
      "learning_rate": 4.53801698753443e-05,
      "loss": 0.8646439552307129,
      "mean_token_accuracy": 0.790368628501892,
      "num_tokens": 23353434.0,
      "step": 11750
    },
    {
      "entropy": 0.7796212345361709,
      "epoch": 0.21125432254008172,
      "grad_norm": 5.375,
      "learning_rate": 4.537184575532853e-05,
      "loss": 0.7964162826538086,
      "mean_token_accuracy": 0.8108895421028137,
      "num_tokens": 23374564.0,
      "step": 11760
    },
    {
      "entropy": 0.9269380301237107,
      "epoch": 0.21143396056945254,
      "grad_norm": 8.5625,
      "learning_rate": 4.53635149076573e-05,
      "loss": 0.9741859436035156,
      "mean_token_accuracy": 0.7672837674617767,
      "num_tokens": 23395041.0,
      "step": 11770
    },
    {
      "entropy": 0.8324027925729751,
      "epoch": 0.21161359859882337,
      "grad_norm": 4.9375,
      "learning_rate": 4.535517733508183e-05,
      "loss": 0.8666191101074219,
      "mean_token_accuracy": 0.7994116529822349,
      "num_tokens": 23414320.0,
      "step": 11780
    },
    {
      "entropy": 0.7509162478148937,
      "epoch": 0.2117932366281942,
      "grad_norm": 5.84375,
      "learning_rate": 4.534683304035554e-05,
      "loss": 0.8079517364501954,
      "mean_token_accuracy": 0.8080166667699814,
      "num_tokens": 23433143.0,
      "step": 11790
    },
    {
      "entropy": 0.9041401676833629,
      "epoch": 0.211972874657565,
      "grad_norm": 6.25,
      "learning_rate": 4.5338482026234064e-05,
      "loss": 0.9419396400451661,
      "mean_token_accuracy": 0.7839539527893067,
      "num_tokens": 23454367.0,
      "step": 11800
    },
    {
      "entropy": 0.8956077970564366,
      "epoch": 0.21215251268693583,
      "grad_norm": 6.28125,
      "learning_rate": 4.533012429547528e-05,
      "loss": 0.9027382850646972,
      "mean_token_accuracy": 0.7832041651010513,
      "num_tokens": 23473264.0,
      "step": 11810
    },
    {
      "entropy": 0.9157611280679703,
      "epoch": 0.21233215071630665,
      "grad_norm": 5.0625,
      "learning_rate": 4.532175985083926e-05,
      "loss": 0.8620236396789551,
      "mean_token_accuracy": 0.7872451275587082,
      "num_tokens": 23493342.0,
      "step": 11820
    },
    {
      "entropy": 0.8748748458921909,
      "epoch": 0.21251178874567747,
      "grad_norm": 5.375,
      "learning_rate": 4.53133886950883e-05,
      "loss": 0.9250937461853027,
      "mean_token_accuracy": 0.7745199605822564,
      "num_tokens": 23512932.0,
      "step": 11830
    },
    {
      "entropy": 0.8910929664969445,
      "epoch": 0.21269142677504826,
      "grad_norm": 8.1875,
      "learning_rate": 4.530501083098692e-05,
      "loss": 0.9051648139953613,
      "mean_token_accuracy": 0.7787199780344963,
      "num_tokens": 23531845.0,
      "step": 11840
    },
    {
      "entropy": 0.8718089506030082,
      "epoch": 0.21287106480441909,
      "grad_norm": 5.0625,
      "learning_rate": 4.5296626261301845e-05,
      "loss": 0.8654747009277344,
      "mean_token_accuracy": 0.7920468553900719,
      "num_tokens": 23551421.0,
      "step": 11850
    },
    {
      "entropy": 0.8071680948138237,
      "epoch": 0.2130507028337899,
      "grad_norm": 5.65625,
      "learning_rate": 4.5288234988802015e-05,
      "loss": 0.8205275535583496,
      "mean_token_accuracy": 0.7986840844154358,
      "num_tokens": 23570403.0,
      "step": 11860
    },
    {
      "entropy": 0.8409905366599559,
      "epoch": 0.21323034086316073,
      "grad_norm": 5.625,
      "learning_rate": 4.5279837016258595e-05,
      "loss": 0.8703806877136231,
      "mean_token_accuracy": 0.7924432069063186,
      "num_tokens": 23589410.0,
      "step": 11870
    },
    {
      "entropy": 0.8105420738458633,
      "epoch": 0.21340997889253155,
      "grad_norm": 7.21875,
      "learning_rate": 4.5271432346444955e-05,
      "loss": 0.8503334999084473,
      "mean_token_accuracy": 0.7964959919452668,
      "num_tokens": 23610230.0,
      "step": 11880
    },
    {
      "entropy": 0.8245964542031288,
      "epoch": 0.21358961692190237,
      "grad_norm": 5.5,
      "learning_rate": 4.526302098213667e-05,
      "loss": 0.8177347183227539,
      "mean_token_accuracy": 0.7970363959670067,
      "num_tokens": 23630470.0,
      "step": 11890
    },
    {
      "entropy": 0.7978442572057247,
      "epoch": 0.2137692549512732,
      "grad_norm": 5.125,
      "learning_rate": 4.525460292611154e-05,
      "loss": 0.853847885131836,
      "mean_token_accuracy": 0.7944505468010903,
      "num_tokens": 23650379.0,
      "step": 11900
    },
    {
      "entropy": 0.9227809779345989,
      "epoch": 0.213948892980644,
      "grad_norm": 6.0,
      "learning_rate": 4.5246178181149556e-05,
      "loss": 0.9348913192749023,
      "mean_token_accuracy": 0.7690330147743225,
      "num_tokens": 23671313.0,
      "step": 11910
    },
    {
      "entropy": 0.885598923265934,
      "epoch": 0.21412853101001483,
      "grad_norm": 5.5,
      "learning_rate": 4.5237746750032945e-05,
      "loss": 0.8872857093811035,
      "mean_token_accuracy": 0.7929845601320267,
      "num_tokens": 23690640.0,
      "step": 11920
    },
    {
      "entropy": 0.8461043626070023,
      "epoch": 0.21430816903938563,
      "grad_norm": 6.0625,
      "learning_rate": 4.522930863554612e-05,
      "loss": 0.84425048828125,
      "mean_token_accuracy": 0.7933999031782151,
      "num_tokens": 23710825.0,
      "step": 11930
    },
    {
      "entropy": 0.8892766140401364,
      "epoch": 0.21448780706875645,
      "grad_norm": 5.3125,
      "learning_rate": 4.52208638404757e-05,
      "loss": 0.9082716941833496,
      "mean_token_accuracy": 0.7829471364617347,
      "num_tokens": 23730619.0,
      "step": 11940
    },
    {
      "entropy": 0.8075975857675075,
      "epoch": 0.21466744509812727,
      "grad_norm": 5.3125,
      "learning_rate": 4.5212412367610524e-05,
      "loss": 0.8088004112243652,
      "mean_token_accuracy": 0.8022006094455719,
      "num_tokens": 23750363.0,
      "step": 11950
    },
    {
      "entropy": 0.7650886200368404,
      "epoch": 0.2148470831274981,
      "grad_norm": 6.75,
      "learning_rate": 4.520395421974163e-05,
      "loss": 0.7765495777130127,
      "mean_token_accuracy": 0.8091513708233833,
      "num_tokens": 23769728.0,
      "step": 11960
    },
    {
      "entropy": 0.7986391887068749,
      "epoch": 0.2150267211568689,
      "grad_norm": 4.8125,
      "learning_rate": 4.519548939966226e-05,
      "loss": 0.8482050895690918,
      "mean_token_accuracy": 0.7979520753026008,
      "num_tokens": 23788994.0,
      "step": 11970
    },
    {
      "entropy": 0.8329289734363556,
      "epoch": 0.21520635918623973,
      "grad_norm": 6.90625,
      "learning_rate": 4.518701791016786e-05,
      "loss": 0.8150861740112305,
      "mean_token_accuracy": 0.8043266505002975,
      "num_tokens": 23807849.0,
      "step": 11980
    },
    {
      "entropy": 0.7797232665121555,
      "epoch": 0.21538599721561055,
      "grad_norm": 5.0625,
      "learning_rate": 4.517853975405608e-05,
      "loss": 0.8143875122070312,
      "mean_token_accuracy": 0.8019890189170837,
      "num_tokens": 23828150.0,
      "step": 11990
    },
    {
      "epoch": 0.21556563524498137,
      "eval_entropy": 0.8464146839380264,
      "eval_loss": 0.8452970385551453,
      "eval_mean_token_accuracy": 0.7959438359737396,
      "eval_num_tokens": 23848467.0,
      "eval_runtime": 40.578,
      "eval_samples_per_second": 49.288,
      "eval_steps_per_second": 6.161,
      "step": 12000
    },
    {
      "entropy": 0.8479792516678571,
      "epoch": 0.21574527327435217,
      "grad_norm": 9.125,
      "learning_rate": 4.516156345318198e-05,
      "loss": 0.875247573852539,
      "mean_token_accuracy": 0.7886293981224298,
      "num_tokens": 23866943.0,
      "step": 12010
    },
    {
      "entropy": 0.8574517481029034,
      "epoch": 0.215924911303723,
      "grad_norm": 5.4375,
      "learning_rate": 4.515306531402596e-05,
      "loss": 0.844296932220459,
      "mean_token_accuracy": 0.7944005832076073,
      "num_tokens": 23888828.0,
      "step": 12020
    },
    {
      "entropy": 0.8414042226970195,
      "epoch": 0.2161045493330938,
      "grad_norm": 5.625,
      "learning_rate": 4.514456051946516e-05,
      "loss": 0.8315954208374023,
      "mean_token_accuracy": 0.8022475019097328,
      "num_tokens": 23908019.0,
      "step": 12030
    },
    {
      "entropy": 0.8386501006782054,
      "epoch": 0.21628418736246463,
      "grad_norm": 5.71875,
      "learning_rate": 4.513604907230823e-05,
      "loss": 0.8816123008728027,
      "mean_token_accuracy": 0.7946676641702652,
      "num_tokens": 23927515.0,
      "step": 12040
    },
    {
      "entropy": 0.8604972586035728,
      "epoch": 0.21646382539183545,
      "grad_norm": 4.15625,
      "learning_rate": 4.5127530975366e-05,
      "loss": 0.866706657409668,
      "mean_token_accuracy": 0.7919363409280777,
      "num_tokens": 23949695.0,
      "step": 12050
    },
    {
      "entropy": 0.8096953980624676,
      "epoch": 0.21664346342120627,
      "grad_norm": 5.03125,
      "learning_rate": 4.511900623145152e-05,
      "loss": 0.7925299644470215,
      "mean_token_accuracy": 0.8021791979670525,
      "num_tokens": 23969585.0,
      "step": 12060
    },
    {
      "entropy": 0.8854925207793712,
      "epoch": 0.2168231014505771,
      "grad_norm": 7.09375,
      "learning_rate": 4.5110474843380036e-05,
      "loss": 0.9172743797302246,
      "mean_token_accuracy": 0.7840204879641532,
      "num_tokens": 23989799.0,
      "step": 12070
    },
    {
      "entropy": 0.796972592920065,
      "epoch": 0.21700273947994791,
      "grad_norm": 7.09375,
      "learning_rate": 4.510193681396896e-05,
      "loss": 0.8046932220458984,
      "mean_token_accuracy": 0.8079536944627762,
      "num_tokens": 24009301.0,
      "step": 12080
    },
    {
      "entropy": 0.8570881865918636,
      "epoch": 0.21718237750931874,
      "grad_norm": 5.59375,
      "learning_rate": 4.5093392146037935e-05,
      "loss": 0.8699812889099121,
      "mean_token_accuracy": 0.7916162610054016,
      "num_tokens": 24029046.0,
      "step": 12090
    },
    {
      "entropy": 0.8185301721096039,
      "epoch": 0.21736201553868953,
      "grad_norm": 6.96875,
      "learning_rate": 4.5084840842408746e-05,
      "loss": 0.7953297615051269,
      "mean_token_accuracy": 0.8010001286864281,
      "num_tokens": 24048559.0,
      "step": 12100
    },
    {
      "entropy": 0.7530040487647056,
      "epoch": 0.21754165356806035,
      "grad_norm": 6.25,
      "learning_rate": 4.507628290590543e-05,
      "loss": 0.7911427974700928,
      "mean_token_accuracy": 0.8138585031032562,
      "num_tokens": 24068416.0,
      "step": 12110
    },
    {
      "entropy": 0.8224557250738144,
      "epoch": 0.21772129159743117,
      "grad_norm": 4.3125,
      "learning_rate": 4.5067718339354173e-05,
      "loss": 0.8501961708068848,
      "mean_token_accuracy": 0.8008611887693405,
      "num_tokens": 24088278.0,
      "step": 12120
    },
    {
      "entropy": 0.7375212855637073,
      "epoch": 0.217900929626802,
      "grad_norm": 5.1875,
      "learning_rate": 4.5059147145583346e-05,
      "loss": 0.743773078918457,
      "mean_token_accuracy": 0.8217915192246437,
      "num_tokens": 24108905.0,
      "step": 12130
    },
    {
      "entropy": 0.8700596071779728,
      "epoch": 0.2180805676561728,
      "grad_norm": 4.90625,
      "learning_rate": 4.5050569327423554e-05,
      "loss": 0.8951065063476562,
      "mean_token_accuracy": 0.7870243668556214,
      "num_tokens": 24128774.0,
      "step": 12140
    },
    {
      "entropy": 0.8434512503445148,
      "epoch": 0.21826020568554363,
      "grad_norm": 6.03125,
      "learning_rate": 4.5041984887707527e-05,
      "loss": 0.8785881042480469,
      "mean_token_accuracy": 0.7928961955010891,
      "num_tokens": 24149531.0,
      "step": 12150
    },
    {
      "entropy": 0.8176505111157895,
      "epoch": 0.21843984371491446,
      "grad_norm": 5.625,
      "learning_rate": 4.5033393829270235e-05,
      "loss": 0.8494326591491699,
      "mean_token_accuracy": 0.7927245572209358,
      "num_tokens": 24171301.0,
      "step": 12160
    },
    {
      "entropy": 0.8252288304269314,
      "epoch": 0.21861948174428528,
      "grad_norm": 6.53125,
      "learning_rate": 4.502479615494881e-05,
      "loss": 0.8436321258544922,
      "mean_token_accuracy": 0.7965697005391121,
      "num_tokens": 24190613.0,
      "step": 12170
    },
    {
      "entropy": 0.8238121353089809,
      "epoch": 0.21879911977365607,
      "grad_norm": 6.09375,
      "learning_rate": 4.501619186758258e-05,
      "loss": 0.8353986740112305,
      "mean_token_accuracy": 0.80070820748806,
      "num_tokens": 24211586.0,
      "step": 12180
    },
    {
      "entropy": 0.8209795475006103,
      "epoch": 0.2189787578030269,
      "grad_norm": 5.375,
      "learning_rate": 4.500758097001303e-05,
      "loss": 0.804349422454834,
      "mean_token_accuracy": 0.8051554724574089,
      "num_tokens": 24230729.0,
      "step": 12190
    },
    {
      "entropy": 0.7983332190662622,
      "epoch": 0.2191583958323977,
      "grad_norm": 6.28125,
      "learning_rate": 4.4998963465083856e-05,
      "loss": 0.7756563186645508,
      "mean_token_accuracy": 0.8130069330334664,
      "num_tokens": 24250505.0,
      "step": 12200
    },
    {
      "entropy": 0.8395957559347152,
      "epoch": 0.21933803386176853,
      "grad_norm": 6.9375,
      "learning_rate": 4.4990339355640934e-05,
      "loss": 0.8581814765930176,
      "mean_token_accuracy": 0.7993357323110104,
      "num_tokens": 24269312.0,
      "step": 12210
    },
    {
      "entropy": 0.8188572727143765,
      "epoch": 0.21951767189113935,
      "grad_norm": 6.0,
      "learning_rate": 4.4981708644532305e-05,
      "loss": 0.8869921684265136,
      "mean_token_accuracy": 0.791328576952219,
      "num_tokens": 24289015.0,
      "step": 12220
    },
    {
      "entropy": 0.8390635289251804,
      "epoch": 0.21969730992051018,
      "grad_norm": 6.375,
      "learning_rate": 4.49730713346082e-05,
      "loss": 0.8982131958007813,
      "mean_token_accuracy": 0.7822622925043106,
      "num_tokens": 24309304.0,
      "step": 12230
    },
    {
      "entropy": 0.8575118318200111,
      "epoch": 0.219876947949881,
      "grad_norm": 5.96875,
      "learning_rate": 4.496442742872104e-05,
      "loss": 0.8457438468933105,
      "mean_token_accuracy": 0.7942942529916763,
      "num_tokens": 24328393.0,
      "step": 12240
    },
    {
      "entropy": 0.7731442593038083,
      "epoch": 0.22005658597925182,
      "grad_norm": 4.0,
      "learning_rate": 4.495577692972539e-05,
      "loss": 0.7842674732208252,
      "mean_token_accuracy": 0.8094912648200989,
      "num_tokens": 24348741.0,
      "step": 12250
    },
    {
      "entropy": 0.7967588119208813,
      "epoch": 0.22023622400862264,
      "grad_norm": 4.625,
      "learning_rate": 4.494711984047804e-05,
      "loss": 0.8280070304870606,
      "mean_token_accuracy": 0.8015093073248863,
      "num_tokens": 24367604.0,
      "step": 12260
    },
    {
      "entropy": 0.8123065695166588,
      "epoch": 0.22041586203799343,
      "grad_norm": 7.0625,
      "learning_rate": 4.493845616383791e-05,
      "loss": 0.8543951988220215,
      "mean_token_accuracy": 0.7966227486729622,
      "num_tokens": 24387296.0,
      "step": 12270
    },
    {
      "entropy": 0.8806727811694145,
      "epoch": 0.22059550006736425,
      "grad_norm": 5.96875,
      "learning_rate": 4.492978590266613e-05,
      "loss": 0.8688224792480469,
      "mean_token_accuracy": 0.7905280292034149,
      "num_tokens": 24407301.0,
      "step": 12280
    },
    {
      "entropy": 0.8444134458899498,
      "epoch": 0.22077513809673507,
      "grad_norm": 6.25,
      "learning_rate": 4.492110905982598e-05,
      "loss": 0.8337883949279785,
      "mean_token_accuracy": 0.8072254166007042,
      "num_tokens": 24426197.0,
      "step": 12290
    },
    {
      "entropy": 0.8585028409957886,
      "epoch": 0.2209547761261059,
      "grad_norm": 8.25,
      "learning_rate": 4.491242563818294e-05,
      "loss": 0.9325284957885742,
      "mean_token_accuracy": 0.7722136795520782,
      "num_tokens": 24444720.0,
      "step": 12300
    },
    {
      "entropy": 0.9239864327013493,
      "epoch": 0.22113441415547672,
      "grad_norm": 6.375,
      "learning_rate": 4.490373564060465e-05,
      "loss": 0.9178210258483886,
      "mean_token_accuracy": 0.7798847004771232,
      "num_tokens": 24465029.0,
      "step": 12310
    },
    {
      "entropy": 0.8193739831447602,
      "epoch": 0.22131405218484754,
      "grad_norm": 5.84375,
      "learning_rate": 4.48950390699609e-05,
      "loss": 0.8450372695922852,
      "mean_token_accuracy": 0.7999195009469986,
      "num_tokens": 24484109.0,
      "step": 12320
    },
    {
      "entropy": 0.8811383880674839,
      "epoch": 0.22149369021421836,
      "grad_norm": 8.125,
      "learning_rate": 4.488633592912368e-05,
      "loss": 0.8541461944580078,
      "mean_token_accuracy": 0.7896482035517692,
      "num_tokens": 24504292.0,
      "step": 12330
    },
    {
      "entropy": 0.9137931719422341,
      "epoch": 0.22167332824358918,
      "grad_norm": 7.25,
      "learning_rate": 4.487762622096714e-05,
      "loss": 0.9298038482666016,
      "mean_token_accuracy": 0.7718917161226273,
      "num_tokens": 24524072.0,
      "step": 12340
    },
    {
      "entropy": 0.8275380201637745,
      "epoch": 0.22185296627295997,
      "grad_norm": 5.34375,
      "learning_rate": 4.4868909948367596e-05,
      "loss": 0.8407528877258301,
      "mean_token_accuracy": 0.8019433960318565,
      "num_tokens": 24544271.0,
      "step": 12350
    },
    {
      "entropy": 0.7840126506984234,
      "epoch": 0.2220326043023308,
      "grad_norm": 4.28125,
      "learning_rate": 4.4860187114203534e-05,
      "loss": 0.8254824638366699,
      "mean_token_accuracy": 0.8082240000367165,
      "num_tokens": 24564515.0,
      "step": 12360
    },
    {
      "entropy": 0.8121642455458641,
      "epoch": 0.22221224233170161,
      "grad_norm": 5.0625,
      "learning_rate": 4.485145772135561e-05,
      "loss": 0.7950341701507568,
      "mean_token_accuracy": 0.805321016907692,
      "num_tokens": 24585731.0,
      "step": 12370
    },
    {
      "entropy": 0.8478195108473301,
      "epoch": 0.22239188036107244,
      "grad_norm": 7.75,
      "learning_rate": 4.484272177270664e-05,
      "loss": 0.8351319313049317,
      "mean_token_accuracy": 0.7944676846265792,
      "num_tokens": 24605350.0,
      "step": 12380
    },
    {
      "entropy": 0.8992158904671669,
      "epoch": 0.22257151839044326,
      "grad_norm": 5.15625,
      "learning_rate": 4.483397927114162e-05,
      "loss": 0.8890450477600098,
      "mean_token_accuracy": 0.7909966140985489,
      "num_tokens": 24624754.0,
      "step": 12390
    },
    {
      "entropy": 0.8411101944744587,
      "epoch": 0.22275115641981408,
      "grad_norm": 4.78125,
      "learning_rate": 4.482523021954768e-05,
      "loss": 0.9016767501831054,
      "mean_token_accuracy": 0.7933079645037651,
      "num_tokens": 24644294.0,
      "step": 12400
    },
    {
      "entropy": 0.8896338380873203,
      "epoch": 0.2229307944491849,
      "grad_norm": 7.46875,
      "learning_rate": 4.481647462081414e-05,
      "loss": 0.8993163108825684,
      "mean_token_accuracy": 0.789176182448864,
      "num_tokens": 24663927.0,
      "step": 12410
    },
    {
      "entropy": 0.7691617041826249,
      "epoch": 0.22311043247855572,
      "grad_norm": 7.34375,
      "learning_rate": 4.4807712477832476e-05,
      "loss": 0.7881993293762207,
      "mean_token_accuracy": 0.8030701950192451,
      "num_tokens": 24683665.0,
      "step": 12420
    },
    {
      "entropy": 0.7974100209772587,
      "epoch": 0.22329007050792654,
      "grad_norm": 5.40625,
      "learning_rate": 4.4798943793496325e-05,
      "loss": 0.8093111038208007,
      "mean_token_accuracy": 0.8029956519603729,
      "num_tokens": 24704145.0,
      "step": 12430
    },
    {
      "entropy": 0.9408095441758633,
      "epoch": 0.22346970853729733,
      "grad_norm": 9.0,
      "learning_rate": 4.479016857070147e-05,
      "loss": 0.9576236724853515,
      "mean_token_accuracy": 0.7741116255521774,
      "num_tokens": 24723945.0,
      "step": 12440
    },
    {
      "entropy": 0.8627887338399887,
      "epoch": 0.22364934656666816,
      "grad_norm": 8.8125,
      "learning_rate": 4.478138681234587e-05,
      "loss": 0.9228914260864258,
      "mean_token_accuracy": 0.7747398644685746,
      "num_tokens": 24742979.0,
      "step": 12450
    },
    {
      "entropy": 0.9155248194932938,
      "epoch": 0.22382898459603898,
      "grad_norm": 6.25,
      "learning_rate": 4.477259852132964e-05,
      "loss": 0.8994830131530762,
      "mean_token_accuracy": 0.7837437972426414,
      "num_tokens": 24763656.0,
      "step": 12460
    },
    {
      "entropy": 0.8223226211965085,
      "epoch": 0.2240086226254098,
      "grad_norm": 8.25,
      "learning_rate": 4.476380370055506e-05,
      "loss": 0.851894760131836,
      "mean_token_accuracy": 0.7952192440629006,
      "num_tokens": 24783643.0,
      "step": 12470
    },
    {
      "entropy": 0.7979088582098484,
      "epoch": 0.22418826065478062,
      "grad_norm": 5.1875,
      "learning_rate": 4.4755002352926536e-05,
      "loss": 0.8254804611206055,
      "mean_token_accuracy": 0.7957828283309937,
      "num_tokens": 24803589.0,
      "step": 12480
    },
    {
      "entropy": 0.8761582091450691,
      "epoch": 0.22436789868415144,
      "grad_norm": 4.96875,
      "learning_rate": 4.4746194481350656e-05,
      "loss": 0.8574308395385742,
      "mean_token_accuracy": 0.7925078794360161,
      "num_tokens": 24824087.0,
      "step": 12490
    },
    {
      "entropy": 0.8385453555732966,
      "epoch": 0.22454753671352226,
      "grad_norm": 6.5625,
      "learning_rate": 4.473738008873617e-05,
      "loss": 0.8114017486572266,
      "mean_token_accuracy": 0.8024143561720848,
      "num_tokens": 24844133.0,
      "step": 12500
    },
    {
      "entropy": 0.8314867570996285,
      "epoch": 0.22472717474289308,
      "grad_norm": 5.90625,
      "learning_rate": 4.472855917799395e-05,
      "loss": 0.843207836151123,
      "mean_token_accuracy": 0.7980601876974106,
      "num_tokens": 24865332.0,
      "step": 12510
    },
    {
      "entropy": 0.7927067019045353,
      "epoch": 0.22490681277226388,
      "grad_norm": 5.40625,
      "learning_rate": 4.4719731752037046e-05,
      "loss": 0.8535285949707031,
      "mean_token_accuracy": 0.7969173178076744,
      "num_tokens": 24885192.0,
      "step": 12520
    },
    {
      "entropy": 0.8247241027653217,
      "epoch": 0.2250864508016347,
      "grad_norm": 4.3125,
      "learning_rate": 4.4710897813780654e-05,
      "loss": 0.857239055633545,
      "mean_token_accuracy": 0.7954294830560684,
      "num_tokens": 24904529.0,
      "step": 12530
    },
    {
      "entropy": 0.782114976644516,
      "epoch": 0.22526608883100552,
      "grad_norm": 6.59375,
      "learning_rate": 4.470205736614212e-05,
      "loss": 0.7867801666259766,
      "mean_token_accuracy": 0.8075640633702278,
      "num_tokens": 24923966.0,
      "step": 12540
    },
    {
      "entropy": 0.8624099388718605,
      "epoch": 0.22544572686037634,
      "grad_norm": 6.03125,
      "learning_rate": 4.469321041204093e-05,
      "loss": 0.8868124961853028,
      "mean_token_accuracy": 0.7884203091263771,
      "num_tokens": 24943864.0,
      "step": 12550
    },
    {
      "entropy": 0.7803060226142406,
      "epoch": 0.22562536488974716,
      "grad_norm": 6.03125,
      "learning_rate": 4.4684356954398725e-05,
      "loss": 0.8262887001037598,
      "mean_token_accuracy": 0.8020025327801704,
      "num_tokens": 24963770.0,
      "step": 12560
    },
    {
      "entropy": 0.8649819917976856,
      "epoch": 0.22580500291911798,
      "grad_norm": 6.5625,
      "learning_rate": 4.467549699613931e-05,
      "loss": 0.9401711463928223,
      "mean_token_accuracy": 0.7824192769825459,
      "num_tokens": 24983392.0,
      "step": 12570
    },
    {
      "entropy": 0.9410753034055233,
      "epoch": 0.2259846409484888,
      "grad_norm": 5.9375,
      "learning_rate": 4.466663054018861e-05,
      "loss": 1.0132086753845215,
      "mean_token_accuracy": 0.7755438074469566,
      "num_tokens": 25003105.0,
      "step": 12580
    },
    {
      "entropy": 0.877509132027626,
      "epoch": 0.22616427897785962,
      "grad_norm": 7.125,
      "learning_rate": 4.465775758947471e-05,
      "loss": 0.8720463752746582,
      "mean_token_accuracy": 0.7866893485188484,
      "num_tokens": 25021863.0,
      "step": 12590
    },
    {
      "entropy": 0.9339281156659126,
      "epoch": 0.22634391700723044,
      "grad_norm": 6.40625,
      "learning_rate": 4.4648878146927845e-05,
      "loss": 0.921604061126709,
      "mean_token_accuracy": 0.7866102978587151,
      "num_tokens": 25041952.0,
      "step": 12600
    },
    {
      "entropy": 0.904553259909153,
      "epoch": 0.22652355503660124,
      "grad_norm": 5.5625,
      "learning_rate": 4.463999221548039e-05,
      "loss": 0.8969251632690429,
      "mean_token_accuracy": 0.7922039024531842,
      "num_tokens": 25062308.0,
      "step": 12610
    },
    {
      "entropy": 0.7811459705233574,
      "epoch": 0.22670319306597206,
      "grad_norm": 5.28125,
      "learning_rate": 4.463109979806684e-05,
      "loss": 0.8446714401245117,
      "mean_token_accuracy": 0.8029682204127312,
      "num_tokens": 25080841.0,
      "step": 12620
    },
    {
      "entropy": 0.9139662891626358,
      "epoch": 0.22688283109534288,
      "grad_norm": 5.90625,
      "learning_rate": 4.462220089762387e-05,
      "loss": 0.9217382431030273,
      "mean_token_accuracy": 0.7815721467137337,
      "num_tokens": 25100418.0,
      "step": 12630
    },
    {
      "entropy": 0.8794430300593377,
      "epoch": 0.2270624691247137,
      "grad_norm": 6.53125,
      "learning_rate": 4.461329551709027e-05,
      "loss": 0.8852977752685547,
      "mean_token_accuracy": 0.7826037406921387,
      "num_tokens": 25120256.0,
      "step": 12640
    },
    {
      "entropy": 0.877899319678545,
      "epoch": 0.22724210715408452,
      "grad_norm": 7.4375,
      "learning_rate": 4.460438365940698e-05,
      "loss": 0.8553321838378907,
      "mean_token_accuracy": 0.7954798057675362,
      "num_tokens": 25139531.0,
      "step": 12650
    },
    {
      "entropy": 0.7066783405840397,
      "epoch": 0.22742174518345534,
      "grad_norm": 6.46875,
      "learning_rate": 4.459546532751708e-05,
      "loss": 0.7618896961212158,
      "mean_token_accuracy": 0.8173317685723305,
      "num_tokens": 25158992.0,
      "step": 12660
    },
    {
      "entropy": 0.7926855161786079,
      "epoch": 0.22760138321282616,
      "grad_norm": 7.875,
      "learning_rate": 4.4586540524365775e-05,
      "loss": 0.8324318885803222,
      "mean_token_accuracy": 0.8025322109460831,
      "num_tokens": 25178117.0,
      "step": 12670
    },
    {
      "entropy": 0.8477311000227928,
      "epoch": 0.22778102124219698,
      "grad_norm": 5.90625,
      "learning_rate": 4.457760925290043e-05,
      "loss": 0.8981081962585449,
      "mean_token_accuracy": 0.7870203867554665,
      "num_tokens": 25199253.0,
      "step": 12680
    },
    {
      "entropy": 0.79274223446846,
      "epoch": 0.22796065927156778,
      "grad_norm": 4.0,
      "learning_rate": 4.4568671516070523e-05,
      "loss": 0.8021099090576171,
      "mean_token_accuracy": 0.8121156215667724,
      "num_tokens": 25219354.0,
      "step": 12690
    },
    {
      "entropy": 0.9788394019007682,
      "epoch": 0.2281402973009386,
      "grad_norm": 5.375,
      "learning_rate": 4.455972731682769e-05,
      "loss": 0.9878940582275391,
      "mean_token_accuracy": 0.7678033530712127,
      "num_tokens": 25240953.0,
      "step": 12700
    },
    {
      "entropy": 0.9230789631605149,
      "epoch": 0.22831993533030942,
      "grad_norm": 6.125,
      "learning_rate": 4.455077665812567e-05,
      "loss": 0.9540598869323731,
      "mean_token_accuracy": 0.7733573749661445,
      "num_tokens": 25260118.0,
      "step": 12710
    },
    {
      "entropy": 0.8246907591819763,
      "epoch": 0.22849957335968024,
      "grad_norm": 5.6875,
      "learning_rate": 4.454181954292036e-05,
      "loss": 0.8094210624694824,
      "mean_token_accuracy": 0.8006536200642586,
      "num_tokens": 25279782.0,
      "step": 12720
    },
    {
      "entropy": 0.8479666955769062,
      "epoch": 0.22867921138905106,
      "grad_norm": 5.5,
      "learning_rate": 4.45328559741698e-05,
      "loss": 0.8293931961059571,
      "mean_token_accuracy": 0.7997950077056885,
      "num_tokens": 25299892.0,
      "step": 12730
    },
    {
      "entropy": 0.8010788090527058,
      "epoch": 0.22885884941842188,
      "grad_norm": 6.1875,
      "learning_rate": 4.452388595483412e-05,
      "loss": 0.7823635578155518,
      "mean_token_accuracy": 0.8061694011092186,
      "num_tokens": 25318977.0,
      "step": 12740
    },
    {
      "entropy": 0.8933595702052116,
      "epoch": 0.2290384874477927,
      "grad_norm": 5.0625,
      "learning_rate": 4.4514909487875624e-05,
      "loss": 0.9494735717773437,
      "mean_token_accuracy": 0.7759922675788402,
      "num_tokens": 25338326.0,
      "step": 12750
    },
    {
      "entropy": 0.8753757774829865,
      "epoch": 0.22921812547716353,
      "grad_norm": 10.75,
      "learning_rate": 4.450592657625871e-05,
      "loss": 0.9691245079040527,
      "mean_token_accuracy": 0.7673906400799752,
      "num_tokens": 25357082.0,
      "step": 12760
    },
    {
      "entropy": 0.8472864843904973,
      "epoch": 0.22939776350653435,
      "grad_norm": 6.0,
      "learning_rate": 4.449693722294992e-05,
      "loss": 0.8274935722351074,
      "mean_token_accuracy": 0.794508482515812,
      "num_tokens": 25376120.0,
      "step": 12770
    },
    {
      "entropy": 0.8558566257357597,
      "epoch": 0.22957740153590514,
      "grad_norm": 6.6875,
      "learning_rate": 4.448794143091793e-05,
      "loss": 0.8903715133666992,
      "mean_token_accuracy": 0.7848844036459923,
      "num_tokens": 25396376.0,
      "step": 12780
    },
    {
      "entropy": 0.8676431268453598,
      "epoch": 0.22975703956527596,
      "grad_norm": 5.34375,
      "learning_rate": 4.447893920313355e-05,
      "loss": 0.8301250457763671,
      "mean_token_accuracy": 0.8004786998033524,
      "num_tokens": 25416156.0,
      "step": 12790
    },
    {
      "entropy": 0.8249128334224224,
      "epoch": 0.22993667759464678,
      "grad_norm": 4.375,
      "learning_rate": 4.446993054256967e-05,
      "loss": 0.8252821922302246,
      "mean_token_accuracy": 0.8049804523587227,
      "num_tokens": 25436968.0,
      "step": 12800
    },
    {
      "entropy": 0.8511012464761734,
      "epoch": 0.2301163156240176,
      "grad_norm": 6.75,
      "learning_rate": 4.446091545220136e-05,
      "loss": 0.873869514465332,
      "mean_token_accuracy": 0.7804472506046295,
      "num_tokens": 25457658.0,
      "step": 12810
    },
    {
      "entropy": 0.8360457956790924,
      "epoch": 0.23029595365338842,
      "grad_norm": 6.25,
      "learning_rate": 4.445189393500577e-05,
      "loss": 0.8606691360473633,
      "mean_token_accuracy": 0.7959957838058471,
      "num_tokens": 25476351.0,
      "step": 12820
    },
    {
      "entropy": 0.831702147424221,
      "epoch": 0.23047559168275925,
      "grad_norm": 5.125,
      "learning_rate": 4.4442865993962215e-05,
      "loss": 0.8223786354064941,
      "mean_token_accuracy": 0.8010449886322022,
      "num_tokens": 25495937.0,
      "step": 12830
    },
    {
      "entropy": 0.8353251658380032,
      "epoch": 0.23065522971213007,
      "grad_norm": 9.75,
      "learning_rate": 4.443383163205208e-05,
      "loss": 0.8224445343017578,
      "mean_token_accuracy": 0.7952178120613098,
      "num_tokens": 25515360.0,
      "step": 12840
    },
    {
      "entropy": 0.7540961034595967,
      "epoch": 0.2308348677415009,
      "grad_norm": 4.71875,
      "learning_rate": 4.4424790852258925e-05,
      "loss": 0.7488266944885253,
      "mean_token_accuracy": 0.8184132441878319,
      "num_tokens": 25535112.0,
      "step": 12850
    },
    {
      "entropy": 0.854656008630991,
      "epoch": 0.23101450577087168,
      "grad_norm": 4.9375,
      "learning_rate": 4.4415743657568374e-05,
      "loss": 0.9016470909118652,
      "mean_token_accuracy": 0.7882017225027085,
      "num_tokens": 25553352.0,
      "step": 12860
    },
    {
      "entropy": 0.8081205911934376,
      "epoch": 0.2311941438002425,
      "grad_norm": 6.03125,
      "learning_rate": 4.4406690050968224e-05,
      "loss": 0.8460009574890137,
      "mean_token_accuracy": 0.7893910720944405,
      "num_tokens": 25573937.0,
      "step": 12870
    },
    {
      "entropy": 0.8350037544965744,
      "epoch": 0.23137378182961332,
      "grad_norm": 7.5625,
      "learning_rate": 4.439763003544835e-05,
      "loss": 0.8450263023376465,
      "mean_token_accuracy": 0.8000707551836967,
      "num_tokens": 25593783.0,
      "step": 12880
    },
    {
      "entropy": 0.8313210375607014,
      "epoch": 0.23155341985898414,
      "grad_norm": 6.9375,
      "learning_rate": 4.4388563614000756e-05,
      "loss": 0.830899715423584,
      "mean_token_accuracy": 0.8053499922156334,
      "num_tokens": 25612475.0,
      "step": 12890
    },
    {
      "entropy": 0.8646220825612545,
      "epoch": 0.23173305788835497,
      "grad_norm": 6.125,
      "learning_rate": 4.4379490789619574e-05,
      "loss": 0.8717094421386719,
      "mean_token_accuracy": 0.7907189726829529,
      "num_tokens": 25631839.0,
      "step": 12900
    },
    {
      "entropy": 0.8579128488898278,
      "epoch": 0.2319126959177258,
      "grad_norm": 5.84375,
      "learning_rate": 4.4370411565301015e-05,
      "loss": 0.8247678756713868,
      "mean_token_accuracy": 0.8076560109853744,
      "num_tokens": 25651009.0,
      "step": 12910
    },
    {
      "entropy": 0.8028069369494915,
      "epoch": 0.2320923339470966,
      "grad_norm": 5.84375,
      "learning_rate": 4.4361325944043455e-05,
      "loss": 0.8240917205810547,
      "mean_token_accuracy": 0.7966988205909729,
      "num_tokens": 25669922.0,
      "step": 12920
    },
    {
      "entropy": 0.7637861631810665,
      "epoch": 0.23227197197646743,
      "grad_norm": 5.25,
      "learning_rate": 4.435223392884734e-05,
      "loss": 0.8295699119567871,
      "mean_token_accuracy": 0.8057667642831803,
      "num_tokens": 25690500.0,
      "step": 12930
    },
    {
      "entropy": 0.802569330483675,
      "epoch": 0.23245161000583825,
      "grad_norm": 4.59375,
      "learning_rate": 4.434313552271525e-05,
      "loss": 0.8176230430603028,
      "mean_token_accuracy": 0.8051825419068337,
      "num_tokens": 25711981.0,
      "step": 12940
    },
    {
      "entropy": 0.8870356164872646,
      "epoch": 0.23263124803520904,
      "grad_norm": 5.0625,
      "learning_rate": 4.4334030728651854e-05,
      "loss": 0.8993595123291016,
      "mean_token_accuracy": 0.787372249364853,
      "num_tokens": 25731738.0,
      "step": 12950
    },
    {
      "entropy": 0.8762515470385551,
      "epoch": 0.23281088606457986,
      "grad_norm": 6.46875,
      "learning_rate": 4.4324919549663964e-05,
      "loss": 0.8683287620544433,
      "mean_token_accuracy": 0.7880354702472687,
      "num_tokens": 25751894.0,
      "step": 12960
    },
    {
      "entropy": 0.8842963829636574,
      "epoch": 0.23299052409395068,
      "grad_norm": 6.96875,
      "learning_rate": 4.431580198876046e-05,
      "loss": 0.9580388069152832,
      "mean_token_accuracy": 0.7765140473842621,
      "num_tokens": 25772582.0,
      "step": 12970
    },
    {
      "entropy": 0.7793270118534565,
      "epoch": 0.2331701621233215,
      "grad_norm": 5.28125,
      "learning_rate": 4.430667804895238e-05,
      "loss": 0.7752094268798828,
      "mean_token_accuracy": 0.8184760630130767,
      "num_tokens": 25791605.0,
      "step": 12980
    },
    {
      "entropy": 0.8661415778100491,
      "epoch": 0.23334980015269233,
      "grad_norm": 4.90625,
      "learning_rate": 4.429754773325281e-05,
      "loss": 0.8656119346618653,
      "mean_token_accuracy": 0.8004096314311028,
      "num_tokens": 25812293.0,
      "step": 12990
    },
    {
      "epoch": 0.23352943818206315,
      "eval_entropy": 0.8022053318023682,
      "eval_loss": 0.8216367959976196,
      "eval_mean_token_accuracy": 0.8034624741077423,
      "eval_num_tokens": 25831861.0,
      "eval_runtime": 40.5239,
      "eval_samples_per_second": 49.354,
      "eval_steps_per_second": 6.169,
      "step": 13000
    },
    {
      "entropy": 0.8294111751019955,
      "epoch": 0.23370907621143397,
      "grad_norm": 6.9375,
      "learning_rate": 4.4279267986242234e-05,
      "loss": 0.8443568229675293,
      "mean_token_accuracy": 0.7993490062654018,
      "num_tokens": 25850942.0,
      "step": 13010
    },
    {
      "entropy": 0.7961078353226185,
      "epoch": 0.2338887142408048,
      "grad_norm": 5.59375,
      "learning_rate": 4.427011856096798e-05,
      "loss": 0.839866828918457,
      "mean_token_accuracy": 0.808353939652443,
      "num_tokens": 25871033.0,
      "step": 13020
    },
    {
      "entropy": 0.8532736130058766,
      "epoch": 0.23406835227017558,
      "grad_norm": 5.65625,
      "learning_rate": 4.4260962771875746e-05,
      "loss": 0.8963590621948242,
      "mean_token_accuracy": 0.7935997501015664,
      "num_tokens": 25891614.0,
      "step": 13030
    },
    {
      "entropy": 0.778396725654602,
      "epoch": 0.2342479902995464,
      "grad_norm": 7.0,
      "learning_rate": 4.4251800621989184e-05,
      "loss": 0.7322530269622802,
      "mean_token_accuracy": 0.8226825997233391,
      "num_tokens": 25911040.0,
      "step": 13040
    },
    {
      "entropy": 0.826403621584177,
      "epoch": 0.23442762832891723,
      "grad_norm": 7.96875,
      "learning_rate": 4.424263211433402e-05,
      "loss": 0.8655001640319824,
      "mean_token_accuracy": 0.7976618558168411,
      "num_tokens": 25930698.0,
      "step": 13050
    },
    {
      "entropy": 0.8097852259874344,
      "epoch": 0.23460726635828805,
      "grad_norm": 6.03125,
      "learning_rate": 4.423345725193808e-05,
      "loss": 0.8429778099060059,
      "mean_token_accuracy": 0.7988743856549263,
      "num_tokens": 25951499.0,
      "step": 13060
    },
    {
      "entropy": 0.8602092564105988,
      "epoch": 0.23478690438765887,
      "grad_norm": 6.34375,
      "learning_rate": 4.4224276037831314e-05,
      "loss": 0.8577801704406738,
      "mean_token_accuracy": 0.7953047916293144,
      "num_tokens": 25971205.0,
      "step": 13070
    },
    {
      "entropy": 0.8327577278017998,
      "epoch": 0.2349665424170297,
      "grad_norm": 5.6875,
      "learning_rate": 4.421508847504574e-05,
      "loss": 0.8539166450500488,
      "mean_token_accuracy": 0.7880702629685402,
      "num_tokens": 25991494.0,
      "step": 13080
    },
    {
      "entropy": 0.8312172628939152,
      "epoch": 0.2351461804464005,
      "grad_norm": 7.5625,
      "learning_rate": 4.420589456661548e-05,
      "loss": 0.8287091255187988,
      "mean_token_accuracy": 0.8004522904753685,
      "num_tokens": 26010275.0,
      "step": 13090
    },
    {
      "entropy": 0.7685111060738563,
      "epoch": 0.23532581847577133,
      "grad_norm": 7.40625,
      "learning_rate": 4.419669431557676e-05,
      "loss": 0.7482511520385742,
      "mean_token_accuracy": 0.8192949175834656,
      "num_tokens": 26030801.0,
      "step": 13100
    },
    {
      "entropy": 0.8228398337960243,
      "epoch": 0.23550545650514215,
      "grad_norm": 7.09375,
      "learning_rate": 4.4187487724967905e-05,
      "loss": 0.9251422882080078,
      "mean_token_accuracy": 0.7824779823422432,
      "num_tokens": 26051236.0,
      "step": 13110
    },
    {
      "entropy": 0.7810831367969513,
      "epoch": 0.23568509453451295,
      "grad_norm": 6.15625,
      "learning_rate": 4.4178274797829324e-05,
      "loss": 0.8000071525573731,
      "mean_token_accuracy": 0.800956991314888,
      "num_tokens": 26070429.0,
      "step": 13120
    },
    {
      "entropy": 0.8344467595219612,
      "epoch": 0.23586473256388377,
      "grad_norm": 5.0,
      "learning_rate": 4.41690555372035e-05,
      "loss": 0.8428084373474121,
      "mean_token_accuracy": 0.8010011479258538,
      "num_tokens": 26090936.0,
      "step": 13130
    },
    {
      "entropy": 0.8300610117614269,
      "epoch": 0.2360443705932546,
      "grad_norm": 5.5,
      "learning_rate": 4.415982994613505e-05,
      "loss": 0.8459264755249023,
      "mean_token_accuracy": 0.7962104439735412,
      "num_tokens": 26110516.0,
      "step": 13140
    },
    {
      "entropy": 0.8765744000673295,
      "epoch": 0.2362240086226254,
      "grad_norm": 5.6875,
      "learning_rate": 4.4150598027670654e-05,
      "loss": 0.8630064964294434,
      "mean_token_accuracy": 0.7884830616414547,
      "num_tokens": 26132150.0,
      "step": 13150
    },
    {
      "entropy": 0.9264851808547974,
      "epoch": 0.23640364665199623,
      "grad_norm": 6.625,
      "learning_rate": 4.414135978485908e-05,
      "loss": 0.9343637466430664,
      "mean_token_accuracy": 0.776537823677063,
      "num_tokens": 26152423.0,
      "step": 13160
    },
    {
      "entropy": 0.8725615873932838,
      "epoch": 0.23658328468136705,
      "grad_norm": 6.46875,
      "learning_rate": 4.41321152207512e-05,
      "loss": 0.9374333381652832,
      "mean_token_accuracy": 0.7747077733278275,
      "num_tokens": 26171500.0,
      "step": 13170
    },
    {
      "entropy": 0.837825221568346,
      "epoch": 0.23676292271073787,
      "grad_norm": 4.9375,
      "learning_rate": 4.412286433839995e-05,
      "loss": 0.8546714782714844,
      "mean_token_accuracy": 0.795743516087532,
      "num_tokens": 26190506.0,
      "step": 13180
    },
    {
      "entropy": 0.8153563156723976,
      "epoch": 0.2369425607401087,
      "grad_norm": 6.03125,
      "learning_rate": 4.411360714086038e-05,
      "loss": 0.8052280426025391,
      "mean_token_accuracy": 0.801672175526619,
      "num_tokens": 26210199.0,
      "step": 13190
    },
    {
      "entropy": 0.7945010028779507,
      "epoch": 0.2371221987694795,
      "grad_norm": 4.3125,
      "learning_rate": 4.4104343631189614e-05,
      "loss": 0.8012173652648926,
      "mean_token_accuracy": 0.8045805335044861,
      "num_tokens": 26230514.0,
      "step": 13200
    },
    {
      "entropy": 0.8502820879220963,
      "epoch": 0.2373018367988503,
      "grad_norm": 5.625,
      "learning_rate": 4.409507381244686e-05,
      "loss": 0.8486570358276367,
      "mean_token_accuracy": 0.8026246711611748,
      "num_tokens": 26251549.0,
      "step": 13210
    },
    {
      "entropy": 0.8523926630616188,
      "epoch": 0.23748147482822113,
      "grad_norm": 6.4375,
      "learning_rate": 4.4085797687693394e-05,
      "loss": 0.9124655723571777,
      "mean_token_accuracy": 0.7897864326834678,
      "num_tokens": 26271989.0,
      "step": 13220
    },
    {
      "entropy": 0.8423567868769168,
      "epoch": 0.23766111285759195,
      "grad_norm": 5.09375,
      "learning_rate": 4.4076515259992614e-05,
      "loss": 0.8454095840454101,
      "mean_token_accuracy": 0.8019152909517289,
      "num_tokens": 26291672.0,
      "step": 13230
    },
    {
      "entropy": 0.8563370063900948,
      "epoch": 0.23784075088696277,
      "grad_norm": 5.84375,
      "learning_rate": 4.406722653240995e-05,
      "loss": 0.8178421974182128,
      "mean_token_accuracy": 0.804261839389801,
      "num_tokens": 26311822.0,
      "step": 13240
    },
    {
      "entropy": 0.8608871519565582,
      "epoch": 0.2380203889163336,
      "grad_norm": 6.90625,
      "learning_rate": 4.405793150801295e-05,
      "loss": 0.9165966033935546,
      "mean_token_accuracy": 0.7838574834167957,
      "num_tokens": 26331384.0,
      "step": 13250
    },
    {
      "entropy": 0.8039006620645524,
      "epoch": 0.2382000269457044,
      "grad_norm": 4.8125,
      "learning_rate": 4.404863018987122e-05,
      "loss": 0.8268921852111817,
      "mean_token_accuracy": 0.7975385501980782,
      "num_tokens": 26350874.0,
      "step": 13260
    },
    {
      "entropy": 0.751151566952467,
      "epoch": 0.23837966497507523,
      "grad_norm": 7.25,
      "learning_rate": 4.4039322581056474e-05,
      "loss": 0.7578027725219727,
      "mean_token_accuracy": 0.8109473019838334,
      "num_tokens": 26370804.0,
      "step": 13270
    },
    {
      "entropy": 0.8888045981526375,
      "epoch": 0.23855930300444605,
      "grad_norm": 7.25,
      "learning_rate": 4.4030008684642466e-05,
      "loss": 0.859557056427002,
      "mean_token_accuracy": 0.793322491645813,
      "num_tokens": 26389833.0,
      "step": 13280
    },
    {
      "entropy": 0.8526957459747792,
      "epoch": 0.23873894103381685,
      "grad_norm": 5.6875,
      "learning_rate": 4.4020688503705046e-05,
      "loss": 0.8818240165710449,
      "mean_token_accuracy": 0.790793451666832,
      "num_tokens": 26409188.0,
      "step": 13290
    },
    {
      "entropy": 0.904371589422226,
      "epoch": 0.23891857906318767,
      "grad_norm": 5.9375,
      "learning_rate": 4.401136204132213e-05,
      "loss": 0.9643376350402832,
      "mean_token_accuracy": 0.7745593547821045,
      "num_tokens": 26429527.0,
      "step": 13300
    },
    {
      "entropy": 0.7771104529500008,
      "epoch": 0.2390982170925585,
      "grad_norm": 5.3125,
      "learning_rate": 4.400202930057373e-05,
      "loss": 0.7797120094299317,
      "mean_token_accuracy": 0.8115101099014282,
      "num_tokens": 26449676.0,
      "step": 13310
    },
    {
      "entropy": 0.8480752810835839,
      "epoch": 0.2392778551219293,
      "grad_norm": 6.34375,
      "learning_rate": 4.3992690284541903e-05,
      "loss": 0.8843204498291015,
      "mean_token_accuracy": 0.790017269551754,
      "num_tokens": 26469282.0,
      "step": 13320
    },
    {
      "entropy": 0.8000416107475757,
      "epoch": 0.23945749315130013,
      "grad_norm": 6.03125,
      "learning_rate": 4.3983344996310805e-05,
      "loss": 0.8028103828430175,
      "mean_token_accuracy": 0.7986316993832588,
      "num_tokens": 26490018.0,
      "step": 13330
    },
    {
      "entropy": 0.9123258791863919,
      "epoch": 0.23963713118067095,
      "grad_norm": 5.65625,
      "learning_rate": 4.397399343896663e-05,
      "loss": 0.9376555442810058,
      "mean_token_accuracy": 0.7819564029574394,
      "num_tokens": 26510800.0,
      "step": 13340
    },
    {
      "entropy": 0.8541868910193443,
      "epoch": 0.23981676921004177,
      "grad_norm": 6.8125,
      "learning_rate": 4.3964635615597686e-05,
      "loss": 0.8546566963195801,
      "mean_token_accuracy": 0.7991875395178795,
      "num_tokens": 26531089.0,
      "step": 13350
    },
    {
      "entropy": 0.8048194207251071,
      "epoch": 0.2399964072394126,
      "grad_norm": 7.34375,
      "learning_rate": 4.395527152929432e-05,
      "loss": 0.8428640365600586,
      "mean_token_accuracy": 0.7993195682764054,
      "num_tokens": 26549926.0,
      "step": 13360
    },
    {
      "entropy": 0.7812407940626145,
      "epoch": 0.2401760452687834,
      "grad_norm": 5.0,
      "learning_rate": 4.394590118314894e-05,
      "loss": 0.7664508819580078,
      "mean_token_accuracy": 0.8158423021435738,
      "num_tokens": 26569704.0,
      "step": 13370
    },
    {
      "entropy": 0.8481205046176911,
      "epoch": 0.2403556832981542,
      "grad_norm": 5.90625,
      "learning_rate": 4.393652458025605e-05,
      "loss": 0.8584465026855469,
      "mean_token_accuracy": 0.7954918637871742,
      "num_tokens": 26589701.0,
      "step": 13380
    },
    {
      "entropy": 0.8093801572918892,
      "epoch": 0.24053532132752503,
      "grad_norm": 6.0625,
      "learning_rate": 4.392714172371221e-05,
      "loss": 0.8151741981506347,
      "mean_token_accuracy": 0.8042861700057984,
      "num_tokens": 26609809.0,
      "step": 13390
    },
    {
      "entropy": 0.8859588146209717,
      "epoch": 0.24071495935689585,
      "grad_norm": 6.96875,
      "learning_rate": 4.3917752616616015e-05,
      "loss": 0.9433646202087402,
      "mean_token_accuracy": 0.7855261400341987,
      "num_tokens": 26631465.0,
      "step": 13400
    },
    {
      "entropy": 0.7782510004937648,
      "epoch": 0.24089459738626667,
      "grad_norm": 6.6875,
      "learning_rate": 4.3908357262068176e-05,
      "loss": 0.7895389080047608,
      "mean_token_accuracy": 0.8005859911441803,
      "num_tokens": 26651394.0,
      "step": 13410
    },
    {
      "entropy": 0.8104953475296497,
      "epoch": 0.2410742354156375,
      "grad_norm": 4.9375,
      "learning_rate": 4.389895566317144e-05,
      "loss": 0.8134613990783691,
      "mean_token_accuracy": 0.8051000170409679,
      "num_tokens": 26671934.0,
      "step": 13420
    },
    {
      "entropy": 0.7244637757539749,
      "epoch": 0.24125387344500832,
      "grad_norm": 4.875,
      "learning_rate": 4.38895478230306e-05,
      "loss": 0.7035492897033692,
      "mean_token_accuracy": 0.8240381866693497,
      "num_tokens": 26691979.0,
      "step": 13430
    },
    {
      "entropy": 0.7952271088957786,
      "epoch": 0.24143351147437914,
      "grad_norm": 8.4375,
      "learning_rate": 4.388013374475254e-05,
      "loss": 0.793808126449585,
      "mean_token_accuracy": 0.8059951648116112,
      "num_tokens": 26712283.0,
      "step": 13440
    },
    {
      "entropy": 0.838923753798008,
      "epoch": 0.24161314950374996,
      "grad_norm": 6.625,
      "learning_rate": 4.387071343144619e-05,
      "loss": 0.87814302444458,
      "mean_token_accuracy": 0.7969618663191795,
      "num_tokens": 26732103.0,
      "step": 13450
    },
    {
      "entropy": 0.7875683717429638,
      "epoch": 0.24179278753312075,
      "grad_norm": 5.65625,
      "learning_rate": 4.386128688622253e-05,
      "loss": 0.8302737236022949,
      "mean_token_accuracy": 0.8018063828349113,
      "num_tokens": 26752388.0,
      "step": 13460
    },
    {
      "entropy": 0.8359189722687006,
      "epoch": 0.24197242556249157,
      "grad_norm": 6.28125,
      "learning_rate": 4.3851854112194623e-05,
      "loss": 0.8467391967773438,
      "mean_token_accuracy": 0.7976729929447174,
      "num_tokens": 26773124.0,
      "step": 13470
    },
    {
      "entropy": 0.8606241069734096,
      "epoch": 0.2421520635918624,
      "grad_norm": 5.65625,
      "learning_rate": 4.384241511247757e-05,
      "loss": 0.8593057632446289,
      "mean_token_accuracy": 0.797717472910881,
      "num_tokens": 26792855.0,
      "step": 13480
    },
    {
      "entropy": 0.8028083570301533,
      "epoch": 0.24233170162123321,
      "grad_norm": 3.40625,
      "learning_rate": 4.383296989018854e-05,
      "loss": 0.7776421546936035,
      "mean_token_accuracy": 0.8102691829204559,
      "num_tokens": 26812403.0,
      "step": 13490
    },
    {
      "entropy": 0.8435114089399576,
      "epoch": 0.24251133965060404,
      "grad_norm": 7.3125,
      "learning_rate": 4.382351844844674e-05,
      "loss": 0.9107912063598633,
      "mean_token_accuracy": 0.7920294776558876,
      "num_tokens": 26833051.0,
      "step": 13500
    },
    {
      "entropy": 0.8145574770867825,
      "epoch": 0.24269097767997486,
      "grad_norm": 9.25,
      "learning_rate": 4.381406079037344e-05,
      "loss": 0.8400983810424805,
      "mean_token_accuracy": 0.7998263776302338,
      "num_tokens": 26852623.0,
      "step": 13510
    },
    {
      "entropy": 0.7644591614603996,
      "epoch": 0.24287061570934568,
      "grad_norm": 6.09375,
      "learning_rate": 4.3804596919091964e-05,
      "loss": 0.7844414234161377,
      "mean_token_accuracy": 0.8130142956972122,
      "num_tokens": 26871633.0,
      "step": 13520
    },
    {
      "entropy": 0.8215132541954517,
      "epoch": 0.2430502537387165,
      "grad_norm": 5.78125,
      "learning_rate": 4.3795126837727695e-05,
      "loss": 0.8524012565612793,
      "mean_token_accuracy": 0.7911192134022713,
      "num_tokens": 26889910.0,
      "step": 13530
    },
    {
      "entropy": 0.8603561118245124,
      "epoch": 0.2432298917680873,
      "grad_norm": 6.65625,
      "learning_rate": 4.378565054940805e-05,
      "loss": 0.879394817352295,
      "mean_token_accuracy": 0.7950222387909889,
      "num_tokens": 26910226.0,
      "step": 13540
    },
    {
      "entropy": 0.8143448971211911,
      "epoch": 0.2434095297974581,
      "grad_norm": 7.40625,
      "learning_rate": 4.377616805726252e-05,
      "loss": 0.8625999450683594,
      "mean_token_accuracy": 0.7928274750709534,
      "num_tokens": 26930520.0,
      "step": 13550
    },
    {
      "entropy": 0.7778517529368401,
      "epoch": 0.24358916782682893,
      "grad_norm": 6.875,
      "learning_rate": 4.376667936442261e-05,
      "loss": 0.7624529361724853,
      "mean_token_accuracy": 0.8165528297424316,
      "num_tokens": 26949394.0,
      "step": 13560
    },
    {
      "entropy": 0.8629262465983629,
      "epoch": 0.24376880585619976,
      "grad_norm": 6.46875,
      "learning_rate": 4.37571844740219e-05,
      "loss": 0.8273058891296386,
      "mean_token_accuracy": 0.8002196326851845,
      "num_tokens": 26970106.0,
      "step": 13570
    },
    {
      "entropy": 0.8166915208101273,
      "epoch": 0.24394844388557058,
      "grad_norm": 6.75,
      "learning_rate": 4.3747683389196015e-05,
      "loss": 0.8216407775878907,
      "mean_token_accuracy": 0.8027824029326439,
      "num_tokens": 26990358.0,
      "step": 13580
    },
    {
      "entropy": 0.8272555425763131,
      "epoch": 0.2441280819149414,
      "grad_norm": 5.875,
      "learning_rate": 4.3738176113082614e-05,
      "loss": 0.8302549362182617,
      "mean_token_accuracy": 0.8105716541409492,
      "num_tokens": 27010640.0,
      "step": 13590
    },
    {
      "entropy": 0.8589202456176281,
      "epoch": 0.24430771994431222,
      "grad_norm": 6.46875,
      "learning_rate": 4.3728662648821406e-05,
      "loss": 0.8641993522644043,
      "mean_token_accuracy": 0.7981086507439613,
      "num_tokens": 27030213.0,
      "step": 13600
    },
    {
      "entropy": 0.7838783346116542,
      "epoch": 0.24448735797368304,
      "grad_norm": 6.71875,
      "learning_rate": 4.3719142999554146e-05,
      "loss": 0.8545126914978027,
      "mean_token_accuracy": 0.7930277362465858,
      "num_tokens": 27049607.0,
      "step": 13610
    },
    {
      "entropy": 0.751653042435646,
      "epoch": 0.24466699600305383,
      "grad_norm": 5.0,
      "learning_rate": 4.370961716842463e-05,
      "loss": 0.7441647052764893,
      "mean_token_accuracy": 0.8134953200817108,
      "num_tokens": 27069676.0,
      "step": 13620
    },
    {
      "entropy": 0.7849678728729487,
      "epoch": 0.24484663403242465,
      "grad_norm": 6.375,
      "learning_rate": 4.3700085158578694e-05,
      "loss": 0.7777293205261231,
      "mean_token_accuracy": 0.8170342177152634,
      "num_tokens": 27088185.0,
      "step": 13630
    },
    {
      "entropy": 0.7968810461461544,
      "epoch": 0.24502627206179547,
      "grad_norm": 6.96875,
      "learning_rate": 4.369054697316423e-05,
      "loss": 0.8433983802795411,
      "mean_token_accuracy": 0.7983433678746223,
      "num_tokens": 27108207.0,
      "step": 13640
    },
    {
      "entropy": 0.8658739365637302,
      "epoch": 0.2452059100911663,
      "grad_norm": 4.78125,
      "learning_rate": 4.3681002615331126e-05,
      "loss": 0.8854556083679199,
      "mean_token_accuracy": 0.7839902430772782,
      "num_tokens": 27129344.0,
      "step": 13650
    },
    {
      "entropy": 0.8249386437237263,
      "epoch": 0.24538554812053712,
      "grad_norm": 6.21875,
      "learning_rate": 4.3671452088231355e-05,
      "loss": 0.8174446105957032,
      "mean_token_accuracy": 0.8045562699437141,
      "num_tokens": 27149165.0,
      "step": 13660
    },
    {
      "entropy": 0.7624789878726006,
      "epoch": 0.24556518614990794,
      "grad_norm": 4.5625,
      "learning_rate": 4.366189539501891e-05,
      "loss": 0.7846019268035889,
      "mean_token_accuracy": 0.8127708435058594,
      "num_tokens": 27168936.0,
      "step": 13670
    },
    {
      "entropy": 0.7870650894939899,
      "epoch": 0.24574482417927876,
      "grad_norm": 6.46875,
      "learning_rate": 4.3652332538849824e-05,
      "loss": 0.8217618942260743,
      "mean_token_accuracy": 0.7953908115625381,
      "num_tokens": 27188793.0,
      "step": 13680
    },
    {
      "entropy": 0.763074304908514,
      "epoch": 0.24592446220864958,
      "grad_norm": 7.0,
      "learning_rate": 4.364276352288215e-05,
      "loss": 0.7522635459899902,
      "mean_token_accuracy": 0.8163917616009713,
      "num_tokens": 27208804.0,
      "step": 13690
    },
    {
      "entropy": 0.7606382876634598,
      "epoch": 0.2461041002380204,
      "grad_norm": 6.59375,
      "learning_rate": 4.3633188350275996e-05,
      "loss": 0.8136446952819825,
      "mean_token_accuracy": 0.8047461166977883,
      "num_tokens": 27228913.0,
      "step": 13700
    },
    {
      "entropy": 0.8110116370022297,
      "epoch": 0.2462837382673912,
      "grad_norm": 6.15625,
      "learning_rate": 4.362360702419348e-05,
      "loss": 0.800137996673584,
      "mean_token_accuracy": 0.8082036674022675,
      "num_tokens": 27248103.0,
      "step": 13710
    },
    {
      "entropy": 0.7978027477860451,
      "epoch": 0.24646337629676202,
      "grad_norm": 5.96875,
      "learning_rate": 4.361401954779879e-05,
      "loss": 0.8092119216918945,
      "mean_token_accuracy": 0.8066146284341812,
      "num_tokens": 27267757.0,
      "step": 13720
    },
    {
      "entropy": 0.7688148498535157,
      "epoch": 0.24664301432613284,
      "grad_norm": 6.28125,
      "learning_rate": 4.36044259242581e-05,
      "loss": 0.8275465965270996,
      "mean_token_accuracy": 0.8025032967329025,
      "num_tokens": 27287344.0,
      "step": 13730
    },
    {
      "entropy": 0.7337068825960159,
      "epoch": 0.24682265235550366,
      "grad_norm": 5.53125,
      "learning_rate": 4.359482615673964e-05,
      "loss": 0.7259703636169433,
      "mean_token_accuracy": 0.8188115805387497,
      "num_tokens": 27306101.0,
      "step": 13740
    },
    {
      "entropy": 0.7875922851264476,
      "epoch": 0.24700229038487448,
      "grad_norm": 5.6875,
      "learning_rate": 4.3585220248413674e-05,
      "loss": 0.7924707412719727,
      "mean_token_accuracy": 0.8108751147985458,
      "num_tokens": 27325117.0,
      "step": 13750
    },
    {
      "entropy": 0.7412275068461895,
      "epoch": 0.2471819284142453,
      "grad_norm": 4.84375,
      "learning_rate": 4.3575608202452474e-05,
      "loss": 0.7450665473937989,
      "mean_token_accuracy": 0.8127679198980331,
      "num_tokens": 27345952.0,
      "step": 13760
    },
    {
      "entropy": 0.8160237550735474,
      "epoch": 0.24736156644361612,
      "grad_norm": 5.5625,
      "learning_rate": 4.3565990022030346e-05,
      "loss": 0.8164488792419433,
      "mean_token_accuracy": 0.7953980743885041,
      "num_tokens": 27365071.0,
      "step": 13770
    },
    {
      "entropy": 0.8261899374425411,
      "epoch": 0.24754120447298694,
      "grad_norm": 4.375,
      "learning_rate": 4.355636571032364e-05,
      "loss": 0.8589550971984863,
      "mean_token_accuracy": 0.7909029290080071,
      "num_tokens": 27384532.0,
      "step": 13780
    },
    {
      "entropy": 0.7894202262163162,
      "epoch": 0.24772084250235774,
      "grad_norm": 4.9375,
      "learning_rate": 4.3546735270510695e-05,
      "loss": 0.8152304649353027,
      "mean_token_accuracy": 0.7971238806843758,
      "num_tokens": 27404764.0,
      "step": 13790
    },
    {
      "entropy": 0.8257817037403583,
      "epoch": 0.24790048053172856,
      "grad_norm": 7.375,
      "learning_rate": 4.3537098705771916e-05,
      "loss": 0.812558937072754,
      "mean_token_accuracy": 0.7981402933597564,
      "num_tokens": 27424630.0,
      "step": 13800
    },
    {
      "entropy": 0.7889108881354332,
      "epoch": 0.24808011856109938,
      "grad_norm": 8.25,
      "learning_rate": 4.352745601928969e-05,
      "loss": 0.8120283126831055,
      "mean_token_accuracy": 0.8006609439849853,
      "num_tokens": 27443682.0,
      "step": 13810
    },
    {
      "entropy": 0.8692940428853035,
      "epoch": 0.2482597565904702,
      "grad_norm": 6.0625,
      "learning_rate": 4.3517807214248464e-05,
      "loss": 0.862795639038086,
      "mean_token_accuracy": 0.7934139177203179,
      "num_tokens": 27463291.0,
      "step": 13820
    },
    {
      "entropy": 0.8374870076775551,
      "epoch": 0.24843939461984102,
      "grad_norm": 6.40625,
      "learning_rate": 4.3508152293834665e-05,
      "loss": 0.8341181755065918,
      "mean_token_accuracy": 0.7993769809603691,
      "num_tokens": 27483281.0,
      "step": 13830
    },
    {
      "entropy": 0.9041734732687473,
      "epoch": 0.24861903264921184,
      "grad_norm": 6.15625,
      "learning_rate": 4.349849126123678e-05,
      "loss": 0.9557903289794922,
      "mean_token_accuracy": 0.7831974193453789,
      "num_tokens": 27503024.0,
      "step": 13840
    },
    {
      "entropy": 0.819349580258131,
      "epoch": 0.24879867067858266,
      "grad_norm": 7.84375,
      "learning_rate": 4.3488824119645276e-05,
      "loss": 0.7811998844146728,
      "mean_token_accuracy": 0.8059081554412841,
      "num_tokens": 27523383.0,
      "step": 13850
    },
    {
      "entropy": 0.7715065360069275,
      "epoch": 0.24897830870795348,
      "grad_norm": 7.5625,
      "learning_rate": 4.347915087225268e-05,
      "loss": 0.7440221309661865,
      "mean_token_accuracy": 0.8082112684845925,
      "num_tokens": 27542745.0,
      "step": 13860
    },
    {
      "entropy": 0.7329110682010651,
      "epoch": 0.2491579467373243,
      "grad_norm": 5.59375,
      "learning_rate": 4.34694715222535e-05,
      "loss": 0.7399541854858398,
      "mean_token_accuracy": 0.8082781165838242,
      "num_tokens": 27562094.0,
      "step": 13870
    },
    {
      "entropy": 0.7984652914106846,
      "epoch": 0.2493375847666951,
      "grad_norm": 5.0,
      "learning_rate": 4.3459786072844266e-05,
      "loss": 0.8460657119750976,
      "mean_token_accuracy": 0.7990671768784523,
      "num_tokens": 27582008.0,
      "step": 13880
    },
    {
      "entropy": 0.746969722956419,
      "epoch": 0.24951722279606592,
      "grad_norm": 5.625,
      "learning_rate": 4.345009452722354e-05,
      "loss": 0.7735437870025634,
      "mean_token_accuracy": 0.818760558962822,
      "num_tokens": 27601466.0,
      "step": 13890
    },
    {
      "entropy": 0.7327978014945984,
      "epoch": 0.24969686082543674,
      "grad_norm": 6.3125,
      "learning_rate": 4.3440396888591874e-05,
      "loss": 0.7709096908569336,
      "mean_token_accuracy": 0.8040370464324951,
      "num_tokens": 27620665.0,
      "step": 13900
    },
    {
      "entropy": 0.8508526973426342,
      "epoch": 0.24987649885480756,
      "grad_norm": 4.8125,
      "learning_rate": 4.343069316015185e-05,
      "loss": 0.8447537422180176,
      "mean_token_accuracy": 0.7939540073275566,
      "num_tokens": 27640545.0,
      "step": 13910
    },
    {
      "entropy": 0.848307815939188,
      "epoch": 0.2500561368841784,
      "grad_norm": 5.96875,
      "learning_rate": 4.342098334510806e-05,
      "loss": 0.8779210090637207,
      "mean_token_accuracy": 0.7960585847496986,
      "num_tokens": 27660498.0,
      "step": 13920
    },
    {
      "entropy": 0.8527353264391422,
      "epoch": 0.2502357749135492,
      "grad_norm": 8.0625,
      "learning_rate": 4.341126744666709e-05,
      "loss": 0.9123054504394531,
      "mean_token_accuracy": 0.7886921599507332,
      "num_tokens": 27680538.0,
      "step": 13930
    },
    {
      "entropy": 0.839226158708334,
      "epoch": 0.25041541294292,
      "grad_norm": 7.75,
      "learning_rate": 4.3401545468037546e-05,
      "loss": 0.8788884162902832,
      "mean_token_accuracy": 0.7935563623905182,
      "num_tokens": 27700263.0,
      "step": 13940
    },
    {
      "entropy": 0.7657243803143501,
      "epoch": 0.25059505097229084,
      "grad_norm": 5.84375,
      "learning_rate": 4.3391817412430046e-05,
      "loss": 0.7886397838592529,
      "mean_token_accuracy": 0.8056871309876442,
      "num_tokens": 27719390.0,
      "step": 13950
    },
    {
      "entropy": 0.8648333057761193,
      "epoch": 0.25077468900166167,
      "grad_norm": 5.8125,
      "learning_rate": 4.338208328305722e-05,
      "loss": 0.8675991058349609,
      "mean_token_accuracy": 0.7935129538178444,
      "num_tokens": 27738841.0,
      "step": 13960
    },
    {
      "entropy": 0.7995679803192616,
      "epoch": 0.2509543270310325,
      "grad_norm": 7.28125,
      "learning_rate": 4.337234308313367e-05,
      "loss": 0.8029519081115722,
      "mean_token_accuracy": 0.8084412455558777,
      "num_tokens": 27757994.0,
      "step": 13970
    },
    {
      "entropy": 0.8524405397474766,
      "epoch": 0.2511339650604033,
      "grad_norm": 4.53125,
      "learning_rate": 4.3362596815876046e-05,
      "loss": 0.899988842010498,
      "mean_token_accuracy": 0.7921252772212029,
      "num_tokens": 27777845.0,
      "step": 13980
    },
    {
      "entropy": 0.845899411290884,
      "epoch": 0.25131360308977413,
      "grad_norm": 6.5625,
      "learning_rate": 4.335284448450298e-05,
      "loss": 0.8093595504760742,
      "mean_token_accuracy": 0.8086071893572807,
      "num_tokens": 27797731.0,
      "step": 13990
    },
    {
      "epoch": 0.25149324111914495,
      "eval_entropy": 0.7823118499517441,
      "eval_loss": 0.7997555136680603,
      "eval_mean_token_accuracy": 0.8076520009040833,
      "eval_num_tokens": 27817716.0,
      "eval_runtime": 40.5724,
      "eval_samples_per_second": 49.295,
      "eval_steps_per_second": 6.162,
      "step": 14000
    },
    {
      "entropy": 0.8305971156805754,
      "epoch": 0.2516728791485157,
      "grad_norm": 6.28125,
      "learning_rate": 4.3333321642295046e-05,
      "loss": 0.8534729957580567,
      "mean_token_accuracy": 0.7936193771660328,
      "num_tokens": 27837607.0,
      "step": 14010
    },
    {
      "entropy": 0.8847357168793678,
      "epoch": 0.25185251717788654,
      "grad_norm": 6.1875,
      "learning_rate": 4.332355113790747e-05,
      "loss": 0.9044944763183593,
      "mean_token_accuracy": 0.7891054704785347,
      "num_tokens": 27857516.0,
      "step": 14020
    },
    {
      "entropy": 0.7862937532365322,
      "epoch": 0.25203215520725736,
      "grad_norm": 6.0625,
      "learning_rate": 4.331377458229901e-05,
      "loss": 0.7968478202819824,
      "mean_token_accuracy": 0.8118302032351494,
      "num_tokens": 27877637.0,
      "step": 14030
    },
    {
      "entropy": 0.8244802989065647,
      "epoch": 0.2522117932366282,
      "grad_norm": 4.5625,
      "learning_rate": 4.3303991978698276e-05,
      "loss": 0.8252405166625977,
      "mean_token_accuracy": 0.8023489505052567,
      "num_tokens": 27898377.0,
      "step": 14040
    },
    {
      "entropy": 0.7420347094535827,
      "epoch": 0.252391431265999,
      "grad_norm": 5.03125,
      "learning_rate": 4.3294203330335935e-05,
      "loss": 0.7452284336090088,
      "mean_token_accuracy": 0.815617349743843,
      "num_tokens": 27917734.0,
      "step": 14050
    },
    {
      "entropy": 0.760425566136837,
      "epoch": 0.2525710692953698,
      "grad_norm": 8.125,
      "learning_rate": 4.328440864044461e-05,
      "loss": 0.7840786457061768,
      "mean_token_accuracy": 0.8156183704733848,
      "num_tokens": 27937442.0,
      "step": 14060
    },
    {
      "entropy": 0.7686777248978615,
      "epoch": 0.25275070732474064,
      "grad_norm": 7.3125,
      "learning_rate": 4.3274607912258916e-05,
      "loss": 0.7291486263275146,
      "mean_token_accuracy": 0.8176906436681748,
      "num_tokens": 27955575.0,
      "step": 14070
    },
    {
      "entropy": 0.8890803173184395,
      "epoch": 0.25293034535411146,
      "grad_norm": 7.4375,
      "learning_rate": 4.326480114901548e-05,
      "loss": 0.9546053886413575,
      "mean_token_accuracy": 0.7770086444914341,
      "num_tokens": 27976284.0,
      "step": 14080
    },
    {
      "entropy": 0.8317489318549633,
      "epoch": 0.2531099833834823,
      "grad_norm": 6.71875,
      "learning_rate": 4.3254988353952916e-05,
      "loss": 0.854953384399414,
      "mean_token_accuracy": 0.7921025469899178,
      "num_tokens": 27996070.0,
      "step": 14090
    },
    {
      "entropy": 0.833439689874649,
      "epoch": 0.2532896214128531,
      "grad_norm": 7.5625,
      "learning_rate": 4.324516953031182e-05,
      "loss": 0.8885133743286133,
      "mean_token_accuracy": 0.7922736361622811,
      "num_tokens": 28015615.0,
      "step": 14100
    },
    {
      "entropy": 0.8334063023328782,
      "epoch": 0.2534692594422239,
      "grad_norm": 5.8125,
      "learning_rate": 4.323534468133481e-05,
      "loss": 0.8776500701904297,
      "mean_token_accuracy": 0.7903510466217994,
      "num_tokens": 28034876.0,
      "step": 14110
    },
    {
      "entropy": 0.8881644174456597,
      "epoch": 0.25364889747159475,
      "grad_norm": 6.5,
      "learning_rate": 4.322551381026645e-05,
      "loss": 0.9214949607849121,
      "mean_token_accuracy": 0.7832420527935028,
      "num_tokens": 28053874.0,
      "step": 14120
    },
    {
      "entropy": 0.8123660005629063,
      "epoch": 0.25382853550096557,
      "grad_norm": 4.1875,
      "learning_rate": 4.321567692035331e-05,
      "loss": 0.7783702850341797,
      "mean_token_accuracy": 0.8071834981441498,
      "num_tokens": 28074653.0,
      "step": 14130
    },
    {
      "entropy": 0.8241522639989853,
      "epoch": 0.2540081735303364,
      "grad_norm": 6.0625,
      "learning_rate": 4.320583401484398e-05,
      "loss": 0.8078440666198731,
      "mean_token_accuracy": 0.8116596773266792,
      "num_tokens": 28094287.0,
      "step": 14140
    },
    {
      "entropy": 0.8052144214510918,
      "epoch": 0.2541878115597072,
      "grad_norm": 6.46875,
      "learning_rate": 4.3195985096988976e-05,
      "loss": 0.8097591400146484,
      "mean_token_accuracy": 0.7987451165914535,
      "num_tokens": 28114100.0,
      "step": 14150
    },
    {
      "entropy": 0.793202667683363,
      "epoch": 0.25436744958907803,
      "grad_norm": 7.0,
      "learning_rate": 4.318613017004086e-05,
      "loss": 0.8525226593017579,
      "mean_token_accuracy": 0.7962001897394657,
      "num_tokens": 28134178.0,
      "step": 14160
    },
    {
      "entropy": 0.8541349522769451,
      "epoch": 0.25454708761844885,
      "grad_norm": 6.46875,
      "learning_rate": 4.317626923725414e-05,
      "loss": 0.8765149116516113,
      "mean_token_accuracy": 0.7851223066449166,
      "num_tokens": 28154273.0,
      "step": 14170
    },
    {
      "entropy": 0.8576840594410896,
      "epoch": 0.2547267256478196,
      "grad_norm": 5.65625,
      "learning_rate": 4.316640230188531e-05,
      "loss": 0.8743958473205566,
      "mean_token_accuracy": 0.7917151406407357,
      "num_tokens": 28173566.0,
      "step": 14180
    },
    {
      "entropy": 0.8606802739202977,
      "epoch": 0.25490636367719044,
      "grad_norm": 6.65625,
      "learning_rate": 4.315652936719286e-05,
      "loss": 0.8259428024291993,
      "mean_token_accuracy": 0.7993944764137269,
      "num_tokens": 28193412.0,
      "step": 14190
    },
    {
      "entropy": 0.8613556973636151,
      "epoch": 0.25508600170656126,
      "grad_norm": 6.21875,
      "learning_rate": 4.3146650436437265e-05,
      "loss": 0.8245043754577637,
      "mean_token_accuracy": 0.7978127881884575,
      "num_tokens": 28213004.0,
      "step": 14200
    },
    {
      "entropy": 0.8105805769562722,
      "epoch": 0.2552656397359321,
      "grad_norm": 6.3125,
      "learning_rate": 4.313676551288095e-05,
      "loss": 0.8677256584167481,
      "mean_token_accuracy": 0.7978790134191514,
      "num_tokens": 28232502.0,
      "step": 14210
    },
    {
      "entropy": 0.7708905167877674,
      "epoch": 0.2554452777653029,
      "grad_norm": 6.21875,
      "learning_rate": 4.312687459978836e-05,
      "loss": 0.7871992111206054,
      "mean_token_accuracy": 0.8169694811105728,
      "num_tokens": 28252834.0,
      "step": 14220
    },
    {
      "entropy": 0.7973978713154792,
      "epoch": 0.2556249157946737,
      "grad_norm": 5.625,
      "learning_rate": 4.3116977700425894e-05,
      "loss": 0.766731071472168,
      "mean_token_accuracy": 0.8153269544243813,
      "num_tokens": 28272589.0,
      "step": 14230
    },
    {
      "entropy": 0.7794838853180408,
      "epoch": 0.25580455382404454,
      "grad_norm": 4.03125,
      "learning_rate": 4.310707481806192e-05,
      "loss": 0.7510892868041992,
      "mean_token_accuracy": 0.8205872982740402,
      "num_tokens": 28292506.0,
      "step": 14240
    },
    {
      "entropy": 0.7519724570214749,
      "epoch": 0.25598419185341537,
      "grad_norm": 5.15625,
      "learning_rate": 4.3097165955966814e-05,
      "loss": 0.7734851360321044,
      "mean_token_accuracy": 0.8113970935344696,
      "num_tokens": 28312154.0,
      "step": 14250
    },
    {
      "entropy": 0.7895194351673126,
      "epoch": 0.2561638298827862,
      "grad_norm": 6.625,
      "learning_rate": 4.308725111741288e-05,
      "loss": 0.805809497833252,
      "mean_token_accuracy": 0.8009387388825416,
      "num_tokens": 28331430.0,
      "step": 14260
    },
    {
      "entropy": 0.7499701671302319,
      "epoch": 0.256343467912157,
      "grad_norm": 6.84375,
      "learning_rate": 4.307733030567445e-05,
      "loss": 0.7558623313903808,
      "mean_token_accuracy": 0.8095523729920387,
      "num_tokens": 28350833.0,
      "step": 14270
    },
    {
      "entropy": 0.7204631559550763,
      "epoch": 0.25652310594152783,
      "grad_norm": 5.75,
      "learning_rate": 4.306740352402777e-05,
      "loss": 0.7321593761444092,
      "mean_token_accuracy": 0.8199307128787041,
      "num_tokens": 28371791.0,
      "step": 14280
    },
    {
      "entropy": 0.8380329694598914,
      "epoch": 0.25670274397089865,
      "grad_norm": 4.84375,
      "learning_rate": 4.305747077575112e-05,
      "loss": 0.9004829406738282,
      "mean_token_accuracy": 0.7955252528190613,
      "num_tokens": 28392587.0,
      "step": 14290
    },
    {
      "entropy": 0.7112549595534802,
      "epoch": 0.25688238200026947,
      "grad_norm": 7.0,
      "learning_rate": 4.3047532064124695e-05,
      "loss": 0.7641959667205811,
      "mean_token_accuracy": 0.8152240738272667,
      "num_tokens": 28414527.0,
      "step": 14300
    },
    {
      "entropy": 0.825827780365944,
      "epoch": 0.2570620200296403,
      "grad_norm": 6.03125,
      "learning_rate": 4.30375873924307e-05,
      "loss": 0.845159912109375,
      "mean_token_accuracy": 0.7914518535137176,
      "num_tokens": 28434858.0,
      "step": 14310
    },
    {
      "entropy": 0.8412707515060902,
      "epoch": 0.2572416580590111,
      "grad_norm": 6.34375,
      "learning_rate": 4.302763676395327e-05,
      "loss": 0.86558837890625,
      "mean_token_accuracy": 0.7979420840740203,
      "num_tokens": 28456334.0,
      "step": 14320
    },
    {
      "entropy": 0.7655456654727459,
      "epoch": 0.25742129608838193,
      "grad_norm": 4.65625,
      "learning_rate": 4.3017680181978556e-05,
      "loss": 0.7576260566711426,
      "mean_token_accuracy": 0.8161571398377419,
      "num_tokens": 28475542.0,
      "step": 14330
    },
    {
      "entropy": 0.8217014811933041,
      "epoch": 0.25760093411775276,
      "grad_norm": 5.84375,
      "learning_rate": 4.300771764979462e-05,
      "loss": 0.8637436866760254,
      "mean_token_accuracy": 0.7939350441098213,
      "num_tokens": 28495075.0,
      "step": 14340
    },
    {
      "entropy": 0.8074494399130344,
      "epoch": 0.2577805721471235,
      "grad_norm": 5.75,
      "learning_rate": 4.2997749170691536e-05,
      "loss": 0.8022321701049805,
      "mean_token_accuracy": 0.8168031826615334,
      "num_tokens": 28515172.0,
      "step": 14350
    },
    {
      "entropy": 0.7716741152107716,
      "epoch": 0.25796021017649434,
      "grad_norm": 5.5625,
      "learning_rate": 4.2987774747961315e-05,
      "loss": 0.7860188007354736,
      "mean_token_accuracy": 0.8156327664852142,
      "num_tokens": 28535338.0,
      "step": 14360
    },
    {
      "entropy": 0.783092538267374,
      "epoch": 0.25813984820586516,
      "grad_norm": 7.125,
      "learning_rate": 4.297779438489795e-05,
      "loss": 0.7879272937774658,
      "mean_token_accuracy": 0.8077037408947945,
      "num_tokens": 28553639.0,
      "step": 14370
    },
    {
      "entropy": 0.6922728888690471,
      "epoch": 0.258319486235236,
      "grad_norm": 6.15625,
      "learning_rate": 4.296780808479736e-05,
      "loss": 0.697773551940918,
      "mean_token_accuracy": 0.8285977512598037,
      "num_tokens": 28573390.0,
      "step": 14380
    },
    {
      "entropy": 0.7938649445772171,
      "epoch": 0.2584991242646068,
      "grad_norm": 4.84375,
      "learning_rate": 4.2957815850957466e-05,
      "loss": 0.8564668655395508,
      "mean_token_accuracy": 0.7936929121613503,
      "num_tokens": 28592908.0,
      "step": 14390
    },
    {
      "entropy": 0.822998633235693,
      "epoch": 0.2586787622939776,
      "grad_norm": 5.65625,
      "learning_rate": 4.294781768667813e-05,
      "loss": 0.8278387069702149,
      "mean_token_accuracy": 0.7978073790669441,
      "num_tokens": 28613648.0,
      "step": 14400
    },
    {
      "entropy": 0.7473748818039894,
      "epoch": 0.25885840032334845,
      "grad_norm": 6.6875,
      "learning_rate": 4.293781359526117e-05,
      "loss": 0.7390222549438477,
      "mean_token_accuracy": 0.8186459958553314,
      "num_tokens": 28633074.0,
      "step": 14410
    },
    {
      "entropy": 0.8443714216351509,
      "epoch": 0.25903803835271927,
      "grad_norm": 8.0,
      "learning_rate": 4.2927803580010365e-05,
      "loss": 0.849765396118164,
      "mean_token_accuracy": 0.7938869014382363,
      "num_tokens": 28651429.0,
      "step": 14420
    },
    {
      "entropy": 0.845733255892992,
      "epoch": 0.2592176763820901,
      "grad_norm": 5.53125,
      "learning_rate": 4.291778764423145e-05,
      "loss": 0.8720784187316895,
      "mean_token_accuracy": 0.7989844918251038,
      "num_tokens": 28670611.0,
      "step": 14430
    },
    {
      "entropy": 0.818676570057869,
      "epoch": 0.2593973144114609,
      "grad_norm": 5.78125,
      "learning_rate": 4.290776579123212e-05,
      "loss": 0.8249676704406739,
      "mean_token_accuracy": 0.8039656817913056,
      "num_tokens": 28691481.0,
      "step": 14440
    },
    {
      "entropy": 0.8130947772413493,
      "epoch": 0.25957695244083173,
      "grad_norm": 4.6875,
      "learning_rate": 4.2897738024322014e-05,
      "loss": 0.798153018951416,
      "mean_token_accuracy": 0.8037065446376801,
      "num_tokens": 28713083.0,
      "step": 14450
    },
    {
      "entropy": 0.7692267332226038,
      "epoch": 0.25975659047020255,
      "grad_norm": 5.8125,
      "learning_rate": 4.288770434681274e-05,
      "loss": 0.7926108837127686,
      "mean_token_accuracy": 0.8088784359395504,
      "num_tokens": 28734145.0,
      "step": 14460
    },
    {
      "entropy": 0.8078782632946968,
      "epoch": 0.2599362284995734,
      "grad_norm": 9.875,
      "learning_rate": 4.2877664762017835e-05,
      "loss": 0.7899657726287842,
      "mean_token_accuracy": 0.8087860003113747,
      "num_tokens": 28753182.0,
      "step": 14470
    },
    {
      "entropy": 0.7435506097972393,
      "epoch": 0.2601158665289442,
      "grad_norm": 6.09375,
      "learning_rate": 4.286761927325281e-05,
      "loss": 0.7472553253173828,
      "mean_token_accuracy": 0.8158667489886284,
      "num_tokens": 28772417.0,
      "step": 14480
    },
    {
      "entropy": 0.7945246532559395,
      "epoch": 0.260295504558315,
      "grad_norm": 6.25,
      "learning_rate": 4.285756788383512e-05,
      "loss": 0.867308521270752,
      "mean_token_accuracy": 0.7831429615616798,
      "num_tokens": 28792444.0,
      "step": 14490
    },
    {
      "entropy": 0.8455458812415599,
      "epoch": 0.26047514258768584,
      "grad_norm": 6.25,
      "learning_rate": 4.2847510597084136e-05,
      "loss": 0.8616435050964355,
      "mean_token_accuracy": 0.7934768483042717,
      "num_tokens": 28813559.0,
      "step": 14500
    },
    {
      "entropy": 0.8628276482224464,
      "epoch": 0.26065478061705666,
      "grad_norm": 6.09375,
      "learning_rate": 4.283744741632123e-05,
      "loss": 0.862397289276123,
      "mean_token_accuracy": 0.7975175499916076,
      "num_tokens": 28832767.0,
      "step": 14510
    },
    {
      "entropy": 0.8063305981457234,
      "epoch": 0.2608344186464274,
      "grad_norm": 8.0,
      "learning_rate": 4.28273783448697e-05,
      "loss": 0.8468155860900879,
      "mean_token_accuracy": 0.7887645468115807,
      "num_tokens": 28853012.0,
      "step": 14520
    },
    {
      "entropy": 0.7459668152034282,
      "epoch": 0.26101405667579824,
      "grad_norm": 5.5625,
      "learning_rate": 4.2817303386054763e-05,
      "loss": 0.7352814197540283,
      "mean_token_accuracy": 0.8190690964460373,
      "num_tokens": 28871306.0,
      "step": 14530
    },
    {
      "entropy": 0.7463353462517262,
      "epoch": 0.26119369470516907,
      "grad_norm": 6.65625,
      "learning_rate": 4.2807222543203604e-05,
      "loss": 0.7881595134735108,
      "mean_token_accuracy": 0.8142834976315498,
      "num_tokens": 28891443.0,
      "step": 14540
    },
    {
      "entropy": 0.8250531561672687,
      "epoch": 0.2613733327345399,
      "grad_norm": 8.3125,
      "learning_rate": 4.279713581964537e-05,
      "loss": 0.8437843322753906,
      "mean_token_accuracy": 0.7979848653078079,
      "num_tokens": 28911515.0,
      "step": 14550
    },
    {
      "entropy": 0.7826508209109306,
      "epoch": 0.2615529707639107,
      "grad_norm": 6.625,
      "learning_rate": 4.278704321871112e-05,
      "loss": 0.8044563293457031,
      "mean_token_accuracy": 0.8065446883440017,
      "num_tokens": 28930633.0,
      "step": 14560
    },
    {
      "entropy": 0.7261652827262879,
      "epoch": 0.26173260879328153,
      "grad_norm": 3.5625,
      "learning_rate": 4.277694474373385e-05,
      "loss": 0.7151919841766358,
      "mean_token_accuracy": 0.814508906006813,
      "num_tokens": 28952538.0,
      "step": 14570
    },
    {
      "entropy": 0.7804071493446827,
      "epoch": 0.26191224682265235,
      "grad_norm": 5.09375,
      "learning_rate": 4.276684039804851e-05,
      "loss": 0.8198032379150391,
      "mean_token_accuracy": 0.8068506315350532,
      "num_tokens": 28973957.0,
      "step": 14580
    },
    {
      "entropy": 0.8185569062829018,
      "epoch": 0.26209188485202317,
      "grad_norm": 4.0,
      "learning_rate": 4.2756730184992e-05,
      "loss": 0.8221435546875,
      "mean_token_accuracy": 0.800892961025238,
      "num_tokens": 28994054.0,
      "step": 14590
    },
    {
      "entropy": 0.826179300993681,
      "epoch": 0.262271522881394,
      "grad_norm": 5.65625,
      "learning_rate": 4.274661410790314e-05,
      "loss": 0.8319836616516113,
      "mean_token_accuracy": 0.8005240052938462,
      "num_tokens": 29014652.0,
      "step": 14600
    },
    {
      "entropy": 0.781681714951992,
      "epoch": 0.2624511609107648,
      "grad_norm": 4.75,
      "learning_rate": 4.2736492170122696e-05,
      "loss": 0.7349733829498291,
      "mean_token_accuracy": 0.8168729975819587,
      "num_tokens": 29033426.0,
      "step": 14610
    },
    {
      "entropy": 0.8356439761817456,
      "epoch": 0.26263079894013563,
      "grad_norm": 7.25,
      "learning_rate": 4.272636437499336e-05,
      "loss": 0.8564369201660156,
      "mean_token_accuracy": 0.800352169573307,
      "num_tokens": 29052896.0,
      "step": 14620
    },
    {
      "entropy": 0.8980650261044503,
      "epoch": 0.26281043696950646,
      "grad_norm": 8.8125,
      "learning_rate": 4.271623072585975e-05,
      "loss": 0.9067797660827637,
      "mean_token_accuracy": 0.789995738863945,
      "num_tokens": 29072045.0,
      "step": 14630
    },
    {
      "entropy": 0.7769496843218804,
      "epoch": 0.2629900749988773,
      "grad_norm": 6.90625,
      "learning_rate": 4.2706091226068455e-05,
      "loss": 0.7838994026184082,
      "mean_token_accuracy": 0.8106893107295037,
      "num_tokens": 29092078.0,
      "step": 14640
    },
    {
      "entropy": 0.7628645084798336,
      "epoch": 0.2631697130282481,
      "grad_norm": 7.0,
      "learning_rate": 4.269594587896796e-05,
      "loss": 0.8081195831298829,
      "mean_token_accuracy": 0.8002945467829704,
      "num_tokens": 29112156.0,
      "step": 14650
    },
    {
      "entropy": 0.8126539684832096,
      "epoch": 0.2633493510576189,
      "grad_norm": 8.9375,
      "learning_rate": 4.2685794687908695e-05,
      "loss": 0.8209115028381347,
      "mean_token_accuracy": 0.8036455735564232,
      "num_tokens": 29131066.0,
      "step": 14660
    },
    {
      "entropy": 0.7613566733896733,
      "epoch": 0.26352898908698974,
      "grad_norm": 6.65625,
      "learning_rate": 4.2675637656243014e-05,
      "loss": 0.7804110527038575,
      "mean_token_accuracy": 0.8108525827527047,
      "num_tokens": 29150631.0,
      "step": 14670
    },
    {
      "entropy": 0.7164460457861423,
      "epoch": 0.26370862711636056,
      "grad_norm": 5.90625,
      "learning_rate": 4.266547478732521e-05,
      "loss": 0.6783817291259766,
      "mean_token_accuracy": 0.8299393966794014,
      "num_tokens": 29169795.0,
      "step": 14680
    },
    {
      "entropy": 0.7525430284440517,
      "epoch": 0.2638882651457313,
      "grad_norm": 5.875,
      "learning_rate": 4.2655306084511494e-05,
      "loss": 0.7945614814758301,
      "mean_token_accuracy": 0.8117571890354156,
      "num_tokens": 29189503.0,
      "step": 14690
    },
    {
      "entropy": 0.7658544696867466,
      "epoch": 0.26406790317510215,
      "grad_norm": 5.84375,
      "learning_rate": 4.264513155116001e-05,
      "loss": 0.7698403835296631,
      "mean_token_accuracy": 0.8125215902924537,
      "num_tokens": 29209627.0,
      "step": 14700
    },
    {
      "entropy": 0.7455651178956032,
      "epoch": 0.26424754120447297,
      "grad_norm": 9.875,
      "learning_rate": 4.263495119063081e-05,
      "loss": 0.7787853717803955,
      "mean_token_accuracy": 0.8147596046328545,
      "num_tokens": 29228437.0,
      "step": 14710
    },
    {
      "entropy": 0.7897412188351154,
      "epoch": 0.2644271792338438,
      "grad_norm": 7.9375,
      "learning_rate": 4.26247650062859e-05,
      "loss": 0.8080427169799804,
      "mean_token_accuracy": 0.807109959423542,
      "num_tokens": 29247411.0,
      "step": 14720
    },
    {
      "entropy": 0.7684538669884204,
      "epoch": 0.2646068172632146,
      "grad_norm": 8.625,
      "learning_rate": 4.26145730014892e-05,
      "loss": 0.789597463607788,
      "mean_token_accuracy": 0.8073916092514992,
      "num_tokens": 29267354.0,
      "step": 14730
    },
    {
      "entropy": 0.8205644749104977,
      "epoch": 0.26478645529258543,
      "grad_norm": 7.59375,
      "learning_rate": 4.2604375179606526e-05,
      "loss": 0.8572630882263184,
      "mean_token_accuracy": 0.7883263632655144,
      "num_tokens": 29287836.0,
      "step": 14740
    },
    {
      "entropy": 0.8595778204500675,
      "epoch": 0.26496609332195625,
      "grad_norm": 5.28125,
      "learning_rate": 4.259417154400565e-05,
      "loss": 0.8462514877319336,
      "mean_token_accuracy": 0.7908521980047226,
      "num_tokens": 29308045.0,
      "step": 14750
    },
    {
      "entropy": 0.8849016591906548,
      "epoch": 0.2651457313513271,
      "grad_norm": 4.875,
      "learning_rate": 4.258396209805624e-05,
      "loss": 0.8981064796447754,
      "mean_token_accuracy": 0.7855641141533851,
      "num_tokens": 29328955.0,
      "step": 14760
    },
    {
      "entropy": 0.8084438286721707,
      "epoch": 0.2653253693806979,
      "grad_norm": 4.5625,
      "learning_rate": 4.25737468451299e-05,
      "loss": 0.8432988166809082,
      "mean_token_accuracy": 0.7984826803207398,
      "num_tokens": 29349679.0,
      "step": 14770
    },
    {
      "entropy": 0.7323750138282776,
      "epoch": 0.2655050074100687,
      "grad_norm": 6.03125,
      "learning_rate": 4.2563525788600144e-05,
      "loss": 0.724146842956543,
      "mean_token_accuracy": 0.8244312986731529,
      "num_tokens": 29369380.0,
      "step": 14780
    },
    {
      "entropy": 0.8129285268485547,
      "epoch": 0.26568464543943954,
      "grad_norm": 4.96875,
      "learning_rate": 4.255329893184239e-05,
      "loss": 0.8562200546264649,
      "mean_token_accuracy": 0.7958911299705506,
      "num_tokens": 29388300.0,
      "step": 14790
    },
    {
      "entropy": 0.8517115540802479,
      "epoch": 0.26586428346881036,
      "grad_norm": 8.375,
      "learning_rate": 4.2543066278234e-05,
      "loss": 0.9158024787902832,
      "mean_token_accuracy": 0.7789546191692353,
      "num_tokens": 29408773.0,
      "step": 14800
    },
    {
      "entropy": 0.8241821102797985,
      "epoch": 0.2660439214981812,
      "grad_norm": 6.8125,
      "learning_rate": 4.2532827831154225e-05,
      "loss": 0.8350062370300293,
      "mean_token_accuracy": 0.7896240040659904,
      "num_tokens": 29428186.0,
      "step": 14810
    },
    {
      "entropy": 0.8539618864655495,
      "epoch": 0.266223559527552,
      "grad_norm": 5.25,
      "learning_rate": 4.2522583593984245e-05,
      "loss": 0.8458507537841797,
      "mean_token_accuracy": 0.7927403837442398,
      "num_tokens": 29447999.0,
      "step": 14820
    },
    {
      "entropy": 0.886480239033699,
      "epoch": 0.2664031975569228,
      "grad_norm": 7.15625,
      "learning_rate": 4.251233357010713e-05,
      "loss": 0.9013381958007812,
      "mean_token_accuracy": 0.7853526636958122,
      "num_tokens": 29468914.0,
      "step": 14830
    },
    {
      "entropy": 0.8678180009126664,
      "epoch": 0.26658283558629364,
      "grad_norm": 5.09375,
      "learning_rate": 4.2502077762907896e-05,
      "loss": 0.8747787475585938,
      "mean_token_accuracy": 0.7956721991300583,
      "num_tokens": 29489661.0,
      "step": 14840
    },
    {
      "entropy": 0.8419894684106112,
      "epoch": 0.26676247361566446,
      "grad_norm": 7.5625,
      "learning_rate": 4.249181617577344e-05,
      "loss": 0.8622321128845215,
      "mean_token_accuracy": 0.7896248832345009,
      "num_tokens": 29509997.0,
      "step": 14850
    },
    {
      "entropy": 0.8152633763849735,
      "epoch": 0.26694211164503523,
      "grad_norm": 4.0,
      "learning_rate": 4.2481548812092575e-05,
      "loss": 0.8355317115783691,
      "mean_token_accuracy": 0.8034189760684967,
      "num_tokens": 29529308.0,
      "step": 14860
    },
    {
      "entropy": 0.8315967574715615,
      "epoch": 0.26712174967440605,
      "grad_norm": 4.96875,
      "learning_rate": 4.2471275675256034e-05,
      "loss": 0.8158427238464355,
      "mean_token_accuracy": 0.8048417374491692,
      "num_tokens": 29549575.0,
      "step": 14870
    },
    {
      "entropy": 0.7599101223051548,
      "epoch": 0.26730138770377687,
      "grad_norm": 4.96875,
      "learning_rate": 4.2460996768656434e-05,
      "loss": 0.7426119327545166,
      "mean_token_accuracy": 0.817649282515049,
      "num_tokens": 29570081.0,
      "step": 14880
    },
    {
      "entropy": 0.7584678359329701,
      "epoch": 0.2674810257331477,
      "grad_norm": 7.1875,
      "learning_rate": 4.245071209568832e-05,
      "loss": 0.8116582870483399,
      "mean_token_accuracy": 0.8024958059191704,
      "num_tokens": 29590816.0,
      "step": 14890
    },
    {
      "entropy": 0.7050860553979874,
      "epoch": 0.2676606637625185,
      "grad_norm": 5.625,
      "learning_rate": 4.244042165974813e-05,
      "loss": 0.724920129776001,
      "mean_token_accuracy": 0.8218887910246849,
      "num_tokens": 29610304.0,
      "step": 14900
    },
    {
      "entropy": 0.7423150785267353,
      "epoch": 0.26784030179188933,
      "grad_norm": 5.59375,
      "learning_rate": 4.2430125464234195e-05,
      "loss": 0.7800523281097412,
      "mean_token_accuracy": 0.815909343957901,
      "num_tokens": 29631452.0,
      "step": 14910
    },
    {
      "entropy": 0.8177873887121677,
      "epoch": 0.26801993982126016,
      "grad_norm": 6.3125,
      "learning_rate": 4.241982351254677e-05,
      "loss": 0.8582361221313477,
      "mean_token_accuracy": 0.79053515791893,
      "num_tokens": 29651311.0,
      "step": 14920
    },
    {
      "entropy": 0.8094145499169827,
      "epoch": 0.268199577850631,
      "grad_norm": 6.71875,
      "learning_rate": 4.2409515808088e-05,
      "loss": 0.8056672096252442,
      "mean_token_accuracy": 0.7977997109293937,
      "num_tokens": 29670275.0,
      "step": 14930
    },
    {
      "entropy": 0.7838593252003193,
      "epoch": 0.2683792158800018,
      "grad_norm": 6.5625,
      "learning_rate": 4.239920235426193e-05,
      "loss": 0.7865023612976074,
      "mean_token_accuracy": 0.8021491393446922,
      "num_tokens": 29690177.0,
      "step": 14940
    },
    {
      "entropy": 0.7969058059155941,
      "epoch": 0.2685588539093726,
      "grad_norm": 5.75,
      "learning_rate": 4.23888831544745e-05,
      "loss": 0.8203608512878418,
      "mean_token_accuracy": 0.8052724778652192,
      "num_tokens": 29709957.0,
      "step": 14950
    },
    {
      "entropy": 0.8394558370113373,
      "epoch": 0.26873849193874344,
      "grad_norm": 5.1875,
      "learning_rate": 4.237855821213356e-05,
      "loss": 0.8768033027648926,
      "mean_token_accuracy": 0.7964011982083321,
      "num_tokens": 29730036.0,
      "step": 14960
    },
    {
      "entropy": 0.7590589866042137,
      "epoch": 0.26891812996811426,
      "grad_norm": 5.8125,
      "learning_rate": 4.236822753064883e-05,
      "loss": 0.7614968776702881,
      "mean_token_accuracy": 0.8096046566963195,
      "num_tokens": 29748930.0,
      "step": 14970
    },
    {
      "entropy": 0.8132305979728699,
      "epoch": 0.2690977679974851,
      "grad_norm": 4.28125,
      "learning_rate": 4.235789111343196e-05,
      "loss": 0.819987678527832,
      "mean_token_accuracy": 0.7965493872761726,
      "num_tokens": 29768401.0,
      "step": 14980
    },
    {
      "entropy": 0.7061920069158077,
      "epoch": 0.2692774060268559,
      "grad_norm": 5.25,
      "learning_rate": 4.2347548963896476e-05,
      "loss": 0.7335346698760986,
      "mean_token_accuracy": 0.8200434759259224,
      "num_tokens": 29788539.0,
      "step": 14990
    },
    {
      "epoch": 0.2694570440562267,
      "eval_entropy": 0.782933717250824,
      "eval_loss": 0.7773110866546631,
      "eval_mean_token_accuracy": 0.8121837055683137,
      "eval_num_tokens": 29807778.0,
      "eval_runtime": 40.5759,
      "eval_samples_per_second": 49.29,
      "eval_steps_per_second": 6.161,
      "step": 15000
    },
    {
      "entropy": 0.858425565250218,
      "epoch": 0.26963668208559755,
      "grad_norm": 6.4375,
      "learning_rate": 4.232684748153323e-05,
      "loss": 0.8543200492858887,
      "mean_token_accuracy": 0.7928964242339134,
      "num_tokens": 29827572.0,
      "step": 15010
    },
    {
      "entropy": 0.8562903605401516,
      "epoch": 0.26981632011496837,
      "grad_norm": 6.1875,
      "learning_rate": 4.2316488155541976e-05,
      "loss": 0.8729925155639648,
      "mean_token_accuracy": 0.7820759490132332,
      "num_tokens": 29846918.0,
      "step": 15020
    },
    {
      "entropy": 0.6811912328004837,
      "epoch": 0.26999595814433913,
      "grad_norm": 4.84375,
      "learning_rate": 4.2306123110905135e-05,
      "loss": 0.723421049118042,
      "mean_token_accuracy": 0.821088133752346,
      "num_tokens": 29867640.0,
      "step": 15030
    },
    {
      "entropy": 0.8043637439608574,
      "epoch": 0.27017559617370995,
      "grad_norm": 7.78125,
      "learning_rate": 4.2295752351045695e-05,
      "loss": 0.8367928504943848,
      "mean_token_accuracy": 0.8034457430243492,
      "num_tokens": 29887707.0,
      "step": 15040
    },
    {
      "entropy": 0.7981558598577976,
      "epoch": 0.2703552342030808,
      "grad_norm": 7.03125,
      "learning_rate": 4.228537587938851e-05,
      "loss": 0.8467235565185547,
      "mean_token_accuracy": 0.7965778172016144,
      "num_tokens": 29906956.0,
      "step": 15050
    },
    {
      "entropy": 0.8807403571903706,
      "epoch": 0.2705348722324516,
      "grad_norm": 5.9375,
      "learning_rate": 4.227499369936034e-05,
      "loss": 0.8878775596618652,
      "mean_token_accuracy": 0.7872479170560837,
      "num_tokens": 29928219.0,
      "step": 15060
    },
    {
      "entropy": 0.7907146669924259,
      "epoch": 0.2707145102618224,
      "grad_norm": 6.125,
      "learning_rate": 4.226460581438984e-05,
      "loss": 0.7813410758972168,
      "mean_token_accuracy": 0.8135585740208626,
      "num_tokens": 29947772.0,
      "step": 15070
    },
    {
      "entropy": 0.8488945506513119,
      "epoch": 0.27089414829119324,
      "grad_norm": 5.375,
      "learning_rate": 4.225421222790751e-05,
      "loss": 0.844735050201416,
      "mean_token_accuracy": 0.7945253551006317,
      "num_tokens": 29968239.0,
      "step": 15080
    },
    {
      "entropy": 0.764475117623806,
      "epoch": 0.27107378632056406,
      "grad_norm": 5.125,
      "learning_rate": 4.224381294334576e-05,
      "loss": 0.7774603366851807,
      "mean_token_accuracy": 0.8092085734009743,
      "num_tokens": 29987550.0,
      "step": 15090
    },
    {
      "entropy": 0.7700480103492737,
      "epoch": 0.2712534243499349,
      "grad_norm": 6.25,
      "learning_rate": 4.2233407964138905e-05,
      "loss": 0.7915660381317139,
      "mean_token_accuracy": 0.8012581571936608,
      "num_tokens": 30007130.0,
      "step": 15100
    },
    {
      "entropy": 0.7512736760079861,
      "epoch": 0.2714330623793057,
      "grad_norm": 5.9375,
      "learning_rate": 4.2222997293723085e-05,
      "loss": 0.7560422897338868,
      "mean_token_accuracy": 0.812439477443695,
      "num_tokens": 30025409.0,
      "step": 15110
    },
    {
      "entropy": 0.76257349178195,
      "epoch": 0.2716127004086765,
      "grad_norm": 5.46875,
      "learning_rate": 4.2212580935536365e-05,
      "loss": 0.8023286819458008,
      "mean_token_accuracy": 0.8082122102379798,
      "num_tokens": 30044429.0,
      "step": 15120
    },
    {
      "entropy": 0.7564937815070152,
      "epoch": 0.27179233843804734,
      "grad_norm": 5.53125,
      "learning_rate": 4.220215889301867e-05,
      "loss": 0.8027561187744141,
      "mean_token_accuracy": 0.8062302619218826,
      "num_tokens": 30063577.0,
      "step": 15130
    },
    {
      "entropy": 0.8030178837478161,
      "epoch": 0.27197197646741816,
      "grad_norm": 5.0625,
      "learning_rate": 4.219173116961179e-05,
      "loss": 0.7878527641296387,
      "mean_token_accuracy": 0.8132232993841171,
      "num_tokens": 30082990.0,
      "step": 15140
    },
    {
      "entropy": 0.8668191500008107,
      "epoch": 0.272151614496789,
      "grad_norm": 7.59375,
      "learning_rate": 4.218129776875944e-05,
      "loss": 0.8677660942077636,
      "mean_token_accuracy": 0.7946762427687645,
      "num_tokens": 30102881.0,
      "step": 15150
    },
    {
      "entropy": 0.8510095737874508,
      "epoch": 0.2723312525261598,
      "grad_norm": 5.78125,
      "learning_rate": 4.2170858693907136e-05,
      "loss": 0.873196792602539,
      "mean_token_accuracy": 0.80057852268219,
      "num_tokens": 30123853.0,
      "step": 15160
    },
    {
      "entropy": 0.823201435059309,
      "epoch": 0.2725108905555306,
      "grad_norm": 5.5625,
      "learning_rate": 4.216041394850233e-05,
      "loss": 0.8839154243469238,
      "mean_token_accuracy": 0.7994728803634643,
      "num_tokens": 30144619.0,
      "step": 15170
    },
    {
      "entropy": 0.8226317308843136,
      "epoch": 0.27269052858490145,
      "grad_norm": 5.3125,
      "learning_rate": 4.214996353599432e-05,
      "loss": 0.8312336921691894,
      "mean_token_accuracy": 0.8007644861936569,
      "num_tokens": 30163965.0,
      "step": 15180
    },
    {
      "entropy": 0.7755617506802082,
      "epoch": 0.2728701666142722,
      "grad_norm": 4.5625,
      "learning_rate": 4.2139507459834276e-05,
      "loss": 0.7457169532775879,
      "mean_token_accuracy": 0.8187390372157097,
      "num_tokens": 30184265.0,
      "step": 15190
    },
    {
      "entropy": 0.7315781570971012,
      "epoch": 0.27304980464364303,
      "grad_norm": 5.65625,
      "learning_rate": 4.212904572347525e-05,
      "loss": 0.698231840133667,
      "mean_token_accuracy": 0.8214591443538666,
      "num_tokens": 30203910.0,
      "step": 15200
    },
    {
      "entropy": 0.8158537074923515,
      "epoch": 0.27322944267301386,
      "grad_norm": 4.90625,
      "learning_rate": 4.2118578330372136e-05,
      "loss": 0.7763810157775879,
      "mean_token_accuracy": 0.8141224846243859,
      "num_tokens": 30222866.0,
      "step": 15210
    },
    {
      "entropy": 0.705679365247488,
      "epoch": 0.2734090807023847,
      "grad_norm": 5.28125,
      "learning_rate": 4.210810528398174e-05,
      "loss": 0.7324028968811035,
      "mean_token_accuracy": 0.8257035061717033,
      "num_tokens": 30242837.0,
      "step": 15220
    },
    {
      "entropy": 0.7460914671421051,
      "epoch": 0.2735887187317555,
      "grad_norm": 6.625,
      "learning_rate": 4.2097626587762694e-05,
      "loss": 0.792588472366333,
      "mean_token_accuracy": 0.8025313034653664,
      "num_tokens": 30261863.0,
      "step": 15230
    },
    {
      "entropy": 0.8398551344871521,
      "epoch": 0.2737683567611263,
      "grad_norm": 7.6875,
      "learning_rate": 4.2087142245175523e-05,
      "loss": 0.8939518928527832,
      "mean_token_accuracy": 0.7907396376132965,
      "num_tokens": 30281434.0,
      "step": 15240
    },
    {
      "entropy": 0.819460928440094,
      "epoch": 0.27394799479049714,
      "grad_norm": 6.3125,
      "learning_rate": 4.20766522596826e-05,
      "loss": 0.8539933204650879,
      "mean_token_accuracy": 0.7936834156513214,
      "num_tokens": 30300472.0,
      "step": 15250
    },
    {
      "entropy": 0.7469527527689934,
      "epoch": 0.27412763281986796,
      "grad_norm": 5.125,
      "learning_rate": 4.206615663474817e-05,
      "loss": 0.8117154121398926,
      "mean_token_accuracy": 0.810653404891491,
      "num_tokens": 30320941.0,
      "step": 15260
    },
    {
      "entropy": 0.7652257144451141,
      "epoch": 0.2743072708492388,
      "grad_norm": 7.5,
      "learning_rate": 4.205565537383832e-05,
      "loss": 0.7320143222808838,
      "mean_token_accuracy": 0.8190082177519799,
      "num_tokens": 30340624.0,
      "step": 15270
    },
    {
      "entropy": 0.7558058016002178,
      "epoch": 0.2744869088786096,
      "grad_norm": 5.8125,
      "learning_rate": 4.2045148480421034e-05,
      "loss": 0.7971116542816162,
      "mean_token_accuracy": 0.8121700271964073,
      "num_tokens": 30360364.0,
      "step": 15280
    },
    {
      "entropy": 0.8277565866708756,
      "epoch": 0.2746665469079804,
      "grad_norm": 7.9375,
      "learning_rate": 4.203463595796614e-05,
      "loss": 0.8302898406982422,
      "mean_token_accuracy": 0.799428278207779,
      "num_tokens": 30379580.0,
      "step": 15290
    },
    {
      "entropy": 0.7397950187325477,
      "epoch": 0.27484618493735125,
      "grad_norm": 5.0,
      "learning_rate": 4.202411780994532e-05,
      "loss": 0.752332353591919,
      "mean_token_accuracy": 0.8206473425030708,
      "num_tokens": 30399545.0,
      "step": 15300
    },
    {
      "entropy": 0.7773168049752712,
      "epoch": 0.27502582296672207,
      "grad_norm": 7.4375,
      "learning_rate": 4.201359403983211e-05,
      "loss": 0.7869867801666259,
      "mean_token_accuracy": 0.8087172418832779,
      "num_tokens": 30418498.0,
      "step": 15310
    },
    {
      "entropy": 0.8848711736500263,
      "epoch": 0.2752054609960929,
      "grad_norm": 5.6875,
      "learning_rate": 4.2003064651101907e-05,
      "loss": 0.8747241020202636,
      "mean_token_accuracy": 0.7983499258756638,
      "num_tokens": 30437891.0,
      "step": 15320
    },
    {
      "entropy": 0.844631502777338,
      "epoch": 0.2753850990254637,
      "grad_norm": 6.125,
      "learning_rate": 4.1992529647231984e-05,
      "loss": 0.8646357536315918,
      "mean_token_accuracy": 0.7951238691806793,
      "num_tokens": 30458142.0,
      "step": 15330
    },
    {
      "entropy": 0.8027203001081944,
      "epoch": 0.27556473705483453,
      "grad_norm": 8.875,
      "learning_rate": 4.198198903170143e-05,
      "loss": 0.8388633728027344,
      "mean_token_accuracy": 0.794037452340126,
      "num_tokens": 30477217.0,
      "step": 15340
    },
    {
      "entropy": 0.7782227091491223,
      "epoch": 0.27574437508420535,
      "grad_norm": 7.96875,
      "learning_rate": 4.197144280799122e-05,
      "loss": 0.7912730693817138,
      "mean_token_accuracy": 0.8029851600527763,
      "num_tokens": 30497337.0,
      "step": 15350
    },
    {
      "entropy": 0.7280202258378268,
      "epoch": 0.2759240131135761,
      "grad_norm": 4.96875,
      "learning_rate": 4.1960890979584163e-05,
      "loss": 0.7232720851898193,
      "mean_token_accuracy": 0.8153430238366127,
      "num_tokens": 30516157.0,
      "step": 15360
    },
    {
      "entropy": 0.7394778750836849,
      "epoch": 0.27610365114294694,
      "grad_norm": 7.03125,
      "learning_rate": 4.195033354996492e-05,
      "loss": 0.7084841251373291,
      "mean_token_accuracy": 0.825533977150917,
      "num_tokens": 30536781.0,
      "step": 15370
    },
    {
      "entropy": 0.851001937687397,
      "epoch": 0.27628328917231776,
      "grad_norm": 8.375,
      "learning_rate": 4.193977052262001e-05,
      "loss": 0.9042911529541016,
      "mean_token_accuracy": 0.7801014602184295,
      "num_tokens": 30556029.0,
      "step": 15380
    },
    {
      "entropy": 0.7767828799784183,
      "epoch": 0.2764629272016886,
      "grad_norm": 6.84375,
      "learning_rate": 4.192920190103781e-05,
      "loss": 0.802911376953125,
      "mean_token_accuracy": 0.8073208600282669,
      "num_tokens": 30576326.0,
      "step": 15390
    },
    {
      "entropy": 0.7369447678327561,
      "epoch": 0.2766425652310594,
      "grad_norm": 5.4375,
      "learning_rate": 4.191862768870851e-05,
      "loss": 0.7825936794281005,
      "mean_token_accuracy": 0.8234935507178307,
      "num_tokens": 30596282.0,
      "step": 15400
    },
    {
      "entropy": 0.7223578862845897,
      "epoch": 0.2768222032604302,
      "grad_norm": 7.0,
      "learning_rate": 4.1908047889124184e-05,
      "loss": 0.7637592315673828,
      "mean_token_accuracy": 0.8166277825832366,
      "num_tokens": 30615588.0,
      "step": 15410
    },
    {
      "entropy": 0.8957057736814023,
      "epoch": 0.27700184128980104,
      "grad_norm": 4.84375,
      "learning_rate": 4.189746250577873e-05,
      "loss": 0.8960773468017578,
      "mean_token_accuracy": 0.7940184935927391,
      "num_tokens": 30635162.0,
      "step": 15420
    },
    {
      "entropy": 0.7907737970352173,
      "epoch": 0.27718147931917186,
      "grad_norm": 5.90625,
      "learning_rate": 4.188687154216788e-05,
      "loss": 0.7601071834564209,
      "mean_token_accuracy": 0.8129597023129463,
      "num_tokens": 30655417.0,
      "step": 15430
    },
    {
      "entropy": 0.8412348203361034,
      "epoch": 0.2773611173485427,
      "grad_norm": 5.5625,
      "learning_rate": 4.187627500178925e-05,
      "loss": 0.8645668983459472,
      "mean_token_accuracy": 0.7875519007444381,
      "num_tokens": 30674930.0,
      "step": 15440
    },
    {
      "entropy": 0.7478829987347126,
      "epoch": 0.2775407553779135,
      "grad_norm": 5.5,
      "learning_rate": 4.1865672888142246e-05,
      "loss": 0.7541036605834961,
      "mean_token_accuracy": 0.815546253323555,
      "num_tokens": 30694021.0,
      "step": 15450
    },
    {
      "entropy": 0.7349458284676075,
      "epoch": 0.2777203934072843,
      "grad_norm": 5.78125,
      "learning_rate": 4.1855065204728164e-05,
      "loss": 0.7484430313110352,
      "mean_token_accuracy": 0.8199443146586418,
      "num_tokens": 30713047.0,
      "step": 15460
    },
    {
      "entropy": 0.7155452094972133,
      "epoch": 0.27790003143665515,
      "grad_norm": 4.875,
      "learning_rate": 4.184445195505009e-05,
      "loss": 0.7334370136260986,
      "mean_token_accuracy": 0.8189377710223198,
      "num_tokens": 30733300.0,
      "step": 15470
    },
    {
      "entropy": 0.7888330042362213,
      "epoch": 0.27807966946602597,
      "grad_norm": 6.34375,
      "learning_rate": 4.183383314261299e-05,
      "loss": 0.8616847038269043,
      "mean_token_accuracy": 0.801086388528347,
      "num_tokens": 30753889.0,
      "step": 15480
    },
    {
      "entropy": 0.8318060167133808,
      "epoch": 0.2782593074953968,
      "grad_norm": 5.625,
      "learning_rate": 4.182320877092364e-05,
      "loss": 0.8092374801635742,
      "mean_token_accuracy": 0.8028967261314393,
      "num_tokens": 30774808.0,
      "step": 15490
    },
    {
      "entropy": 0.8503727957606315,
      "epoch": 0.2784389455247676,
      "grad_norm": 7.8125,
      "learning_rate": 4.1812578843490676e-05,
      "loss": 0.9135036468505859,
      "mean_token_accuracy": 0.7880162164568901,
      "num_tokens": 30795797.0,
      "step": 15500
    },
    {
      "entropy": 0.8089977629482746,
      "epoch": 0.27861858355413843,
      "grad_norm": 6.5,
      "learning_rate": 4.1801943363824545e-05,
      "loss": 0.8380992889404297,
      "mean_token_accuracy": 0.7971097275614738,
      "num_tokens": 30816803.0,
      "step": 15510
    },
    {
      "entropy": 0.7498241052031517,
      "epoch": 0.27879822158350925,
      "grad_norm": 6.90625,
      "learning_rate": 4.1791302335437535e-05,
      "loss": 0.767780876159668,
      "mean_token_accuracy": 0.8159568309783936,
      "num_tokens": 30836134.0,
      "step": 15520
    },
    {
      "entropy": 0.7281147390604019,
      "epoch": 0.27897785961288,
      "grad_norm": 5.34375,
      "learning_rate": 4.1780655761843776e-05,
      "loss": 0.7200061798095703,
      "mean_token_accuracy": 0.8257206797599792,
      "num_tokens": 30856097.0,
      "step": 15530
    },
    {
      "entropy": 0.8131218485534191,
      "epoch": 0.27915749764225084,
      "grad_norm": 5.625,
      "learning_rate": 4.177000364655922e-05,
      "loss": 0.8036087989807129,
      "mean_token_accuracy": 0.7982544377446175,
      "num_tokens": 30875604.0,
      "step": 15540
    },
    {
      "entropy": 0.7974972054362297,
      "epoch": 0.27933713567162166,
      "grad_norm": 4.75,
      "learning_rate": 4.175934599310165e-05,
      "loss": 0.7071460723876953,
      "mean_token_accuracy": 0.8259832620620727,
      "num_tokens": 30895446.0,
      "step": 15550
    },
    {
      "entropy": 0.6969527877867222,
      "epoch": 0.2795167737009925,
      "grad_norm": 4.21875,
      "learning_rate": 4.174868280499069e-05,
      "loss": 0.6668246746063232,
      "mean_token_accuracy": 0.8357454121112824,
      "num_tokens": 30915951.0,
      "step": 15560
    },
    {
      "entropy": 0.7876476876437664,
      "epoch": 0.2796964117303633,
      "grad_norm": 5.84375,
      "learning_rate": 4.1738014085747765e-05,
      "loss": 0.8730512619018554,
      "mean_token_accuracy": 0.7937133431434631,
      "num_tokens": 30935765.0,
      "step": 15570
    },
    {
      "entropy": 0.7450729414820672,
      "epoch": 0.2798760497597341,
      "grad_norm": 5.5,
      "learning_rate": 4.172733983889615e-05,
      "loss": 0.7500867366790771,
      "mean_token_accuracy": 0.8151269376277923,
      "num_tokens": 30955605.0,
      "step": 15580
    },
    {
      "entropy": 0.8328361734747887,
      "epoch": 0.28005568778910495,
      "grad_norm": 6.3125,
      "learning_rate": 4.171666006796094e-05,
      "loss": 0.88980712890625,
      "mean_token_accuracy": 0.8004471838474274,
      "num_tokens": 30973597.0,
      "step": 15590
    },
    {
      "entropy": 0.7893170610070228,
      "epoch": 0.28023532581847577,
      "grad_norm": 5.0,
      "learning_rate": 4.170597477646905e-05,
      "loss": 0.7996819019317627,
      "mean_token_accuracy": 0.8028216451406479,
      "num_tokens": 30993835.0,
      "step": 15600
    },
    {
      "entropy": 0.7736422643065453,
      "epoch": 0.2804149638478466,
      "grad_norm": 4.71875,
      "learning_rate": 4.169528396794922e-05,
      "loss": 0.7739620685577393,
      "mean_token_accuracy": 0.808749896287918,
      "num_tokens": 31013286.0,
      "step": 15610
    },
    {
      "entropy": 0.83072724416852,
      "epoch": 0.2805946018772174,
      "grad_norm": 5.5,
      "learning_rate": 4.1684587645932006e-05,
      "loss": 0.8459759712219238,
      "mean_token_accuracy": 0.7988006912171841,
      "num_tokens": 31034633.0,
      "step": 15620
    },
    {
      "entropy": 0.7503642030060291,
      "epoch": 0.28077423990658823,
      "grad_norm": 6.3125,
      "learning_rate": 4.167388581394981e-05,
      "loss": 0.7674846649169922,
      "mean_token_accuracy": 0.815148974955082,
      "num_tokens": 31055391.0,
      "step": 15630
    },
    {
      "entropy": 0.8070327155292034,
      "epoch": 0.28095387793595905,
      "grad_norm": 5.40625,
      "learning_rate": 4.166317847553681e-05,
      "loss": 0.8116891860961915,
      "mean_token_accuracy": 0.801799064874649,
      "num_tokens": 31074940.0,
      "step": 15640
    },
    {
      "entropy": 0.8238192707300186,
      "epoch": 0.28113351596532987,
      "grad_norm": 5.34375,
      "learning_rate": 4.165246563422906e-05,
      "loss": 0.8363788604736329,
      "mean_token_accuracy": 0.8007370710372925,
      "num_tokens": 31095470.0,
      "step": 15650
    },
    {
      "entropy": 0.7823977239429951,
      "epoch": 0.2813131539947007,
      "grad_norm": 5.96875,
      "learning_rate": 4.164174729356437e-05,
      "loss": 0.7903109073638916,
      "mean_token_accuracy": 0.810466168820858,
      "num_tokens": 31115044.0,
      "step": 15660
    },
    {
      "entropy": 0.7794207729399204,
      "epoch": 0.2814927920240715,
      "grad_norm": 6.25,
      "learning_rate": 4.163102345708241e-05,
      "loss": 0.7499467372894287,
      "mean_token_accuracy": 0.8065095126628876,
      "num_tokens": 31134441.0,
      "step": 15670
    },
    {
      "entropy": 0.7686565883457661,
      "epoch": 0.28167243005344234,
      "grad_norm": 5.5,
      "learning_rate": 4.162029412832464e-05,
      "loss": 0.7682111740112305,
      "mean_token_accuracy": 0.8164161577820778,
      "num_tokens": 31154435.0,
      "step": 15680
    },
    {
      "entropy": 0.7627410054206848,
      "epoch": 0.28185206808281316,
      "grad_norm": 5.375,
      "learning_rate": 4.160955931083434e-05,
      "loss": 0.8017070770263672,
      "mean_token_accuracy": 0.8144962951540947,
      "num_tokens": 31175256.0,
      "step": 15690
    },
    {
      "entropy": 0.724706993997097,
      "epoch": 0.2820317061121839,
      "grad_norm": 4.8125,
      "learning_rate": 4.1598819008156633e-05,
      "loss": 0.7209950923919678,
      "mean_token_accuracy": 0.8210929945111275,
      "num_tokens": 31194174.0,
      "step": 15700
    },
    {
      "entropy": 0.7097630426287651,
      "epoch": 0.28221134414155474,
      "grad_norm": 4.0625,
      "learning_rate": 4.15880732238384e-05,
      "loss": 0.7172952175140381,
      "mean_token_accuracy": 0.8239585056900978,
      "num_tokens": 31213803.0,
      "step": 15710
    },
    {
      "entropy": 0.7432207785546779,
      "epoch": 0.28239098217092556,
      "grad_norm": 6.59375,
      "learning_rate": 4.157732196142836e-05,
      "loss": 0.7617032527923584,
      "mean_token_accuracy": 0.8206196159124375,
      "num_tokens": 31233701.0,
      "step": 15720
    },
    {
      "entropy": 0.7092471659183502,
      "epoch": 0.2825706202002964,
      "grad_norm": 5.65625,
      "learning_rate": 4.156656522447706e-05,
      "loss": 0.7115270614624023,
      "mean_token_accuracy": 0.8199438467621804,
      "num_tokens": 31253320.0,
      "step": 15730
    },
    {
      "entropy": 0.7400734752416611,
      "epoch": 0.2827502582296672,
      "grad_norm": 5.65625,
      "learning_rate": 4.1555803016536815e-05,
      "loss": 0.7746194362640381,
      "mean_token_accuracy": 0.8121130377054214,
      "num_tokens": 31274566.0,
      "step": 15740
    },
    {
      "entropy": 0.7510441161692143,
      "epoch": 0.282929896259038,
      "grad_norm": 6.25,
      "learning_rate": 4.1545035341161776e-05,
      "loss": 0.8051490783691406,
      "mean_token_accuracy": 0.8095645785331727,
      "num_tokens": 31294004.0,
      "step": 15750
    },
    {
      "entropy": 0.8055024221539497,
      "epoch": 0.28310953428840885,
      "grad_norm": 5.5625,
      "learning_rate": 4.153426220190788e-05,
      "loss": 0.8285681724548339,
      "mean_token_accuracy": 0.8129273235797883,
      "num_tokens": 31313690.0,
      "step": 15760
    },
    {
      "entropy": 0.7768132604658604,
      "epoch": 0.28328917231777967,
      "grad_norm": 7.4375,
      "learning_rate": 4.15234836023329e-05,
      "loss": 0.7528494834899903,
      "mean_token_accuracy": 0.8132277220487595,
      "num_tokens": 31334113.0,
      "step": 15770
    },
    {
      "entropy": 0.8804208666086197,
      "epoch": 0.2834688103471505,
      "grad_norm": 5.5,
      "learning_rate": 4.151269954599637e-05,
      "loss": 0.9061985015869141,
      "mean_token_accuracy": 0.7852698013186454,
      "num_tokens": 31354142.0,
      "step": 15780
    },
    {
      "entropy": 0.7419680453836918,
      "epoch": 0.2836484483765213,
      "grad_norm": 9.1875,
      "learning_rate": 4.1501910036459656e-05,
      "loss": 0.7257208347320556,
      "mean_token_accuracy": 0.8212027832865715,
      "num_tokens": 31373620.0,
      "step": 15790
    },
    {
      "entropy": 0.6994263164699077,
      "epoch": 0.28382808640589213,
      "grad_norm": 5.75,
      "learning_rate": 4.1491115077285916e-05,
      "loss": 0.7097643852233887,
      "mean_token_accuracy": 0.8258636325597764,
      "num_tokens": 31393428.0,
      "step": 15800
    },
    {
      "entropy": 0.7774539552628994,
      "epoch": 0.28400772443526295,
      "grad_norm": 5.8125,
      "learning_rate": 4.148031467204011e-05,
      "loss": 0.7774507522583007,
      "mean_token_accuracy": 0.8088270738720894,
      "num_tokens": 31413716.0,
      "step": 15810
    },
    {
      "entropy": 0.8529713410884142,
      "epoch": 0.2841873624646338,
      "grad_norm": 6.5,
      "learning_rate": 4.146950882428898e-05,
      "loss": 0.8658102035522461,
      "mean_token_accuracy": 0.7921802893280983,
      "num_tokens": 31433879.0,
      "step": 15820
    },
    {
      "entropy": 0.7473576918244362,
      "epoch": 0.2843670004940046,
      "grad_norm": 6.15625,
      "learning_rate": 4.14586975376011e-05,
      "loss": 0.736475133895874,
      "mean_token_accuracy": 0.8213563725352288,
      "num_tokens": 31453155.0,
      "step": 15830
    },
    {
      "entropy": 0.7054384544491767,
      "epoch": 0.2845466385233754,
      "grad_norm": 5.1875,
      "learning_rate": 4.14478808155468e-05,
      "loss": 0.7497594356536865,
      "mean_token_accuracy": 0.8174060076475144,
      "num_tokens": 31472618.0,
      "step": 15840
    },
    {
      "entropy": 0.9327756479382515,
      "epoch": 0.28472627655274624,
      "grad_norm": 6.0625,
      "learning_rate": 4.1437058661698244e-05,
      "loss": 0.9854413986206054,
      "mean_token_accuracy": 0.7784182697534561,
      "num_tokens": 31492201.0,
      "step": 15850
    },
    {
      "entropy": 0.8623161919414997,
      "epoch": 0.28490591458211706,
      "grad_norm": 6.28125,
      "learning_rate": 4.1426231079629355e-05,
      "loss": 0.8703429222106933,
      "mean_token_accuracy": 0.7865716114640235,
      "num_tokens": 31513233.0,
      "step": 15860
    },
    {
      "entropy": 0.7912965916097164,
      "epoch": 0.2850855526114878,
      "grad_norm": 6.65625,
      "learning_rate": 4.141539807291587e-05,
      "loss": 0.78533935546875,
      "mean_token_accuracy": 0.8112850442528725,
      "num_tokens": 31533118.0,
      "step": 15870
    },
    {
      "entropy": 0.7576616831123829,
      "epoch": 0.28526519064085865,
      "grad_norm": 6.53125,
      "learning_rate": 4.1404559645135314e-05,
      "loss": 0.7483622550964355,
      "mean_token_accuracy": 0.815638393163681,
      "num_tokens": 31552064.0,
      "step": 15880
    },
    {
      "entropy": 0.7274096436798573,
      "epoch": 0.28544482867022947,
      "grad_norm": 5.78125,
      "learning_rate": 4.139371579986699e-05,
      "loss": 0.7241614818572998,
      "mean_token_accuracy": 0.8139476224780082,
      "num_tokens": 31571986.0,
      "step": 15890
    },
    {
      "entropy": 0.8361580953001976,
      "epoch": 0.2856244666996003,
      "grad_norm": 6.0625,
      "learning_rate": 4.1382866540692017e-05,
      "loss": 0.8124227523803711,
      "mean_token_accuracy": 0.8022423937916756,
      "num_tokens": 31591019.0,
      "step": 15900
    },
    {
      "entropy": 0.7711074855178595,
      "epoch": 0.2858041047289711,
      "grad_norm": 5.4375,
      "learning_rate": 4.1372011871193274e-05,
      "loss": 0.7568377017974853,
      "mean_token_accuracy": 0.819816367328167,
      "num_tokens": 31610266.0,
      "step": 15910
    },
    {
      "entropy": 0.8036399684846401,
      "epoch": 0.28598374275834193,
      "grad_norm": 6.3125,
      "learning_rate": 4.1361151794955436e-05,
      "loss": 0.8389735221862793,
      "mean_token_accuracy": 0.7925105318427086,
      "num_tokens": 31631074.0,
      "step": 15920
    },
    {
      "entropy": 0.7928057365119457,
      "epoch": 0.28616338078771275,
      "grad_norm": 6.0,
      "learning_rate": 4.1350286315564955e-05,
      "loss": 0.8213986396789551,
      "mean_token_accuracy": 0.7985964506864548,
      "num_tokens": 31651200.0,
      "step": 15930
    },
    {
      "entropy": 0.7588919274508953,
      "epoch": 0.28634301881708357,
      "grad_norm": 5.34375,
      "learning_rate": 4.13394154366101e-05,
      "loss": 0.7963966846466064,
      "mean_token_accuracy": 0.8093880817294121,
      "num_tokens": 31670520.0,
      "step": 15940
    },
    {
      "entropy": 0.7961931593716145,
      "epoch": 0.2865226568464544,
      "grad_norm": 6.84375,
      "learning_rate": 4.132853916168089e-05,
      "loss": 0.8064395904541015,
      "mean_token_accuracy": 0.8072598338127136,
      "num_tokens": 31690983.0,
      "step": 15950
    },
    {
      "entropy": 0.7187569998204708,
      "epoch": 0.2867022948758252,
      "grad_norm": 4.96875,
      "learning_rate": 4.1317657494369135e-05,
      "loss": 0.7582693099975586,
      "mean_token_accuracy": 0.8090129747986794,
      "num_tokens": 31711180.0,
      "step": 15960
    },
    {
      "entropy": 0.8032232850790024,
      "epoch": 0.28688193290519604,
      "grad_norm": 7.6875,
      "learning_rate": 4.130677043826843e-05,
      "loss": 0.8159518241882324,
      "mean_token_accuracy": 0.8023663058876991,
      "num_tokens": 31730481.0,
      "step": 15970
    },
    {
      "entropy": 0.7934968687593937,
      "epoch": 0.28706157093456686,
      "grad_norm": 6.71875,
      "learning_rate": 4.129587799697414e-05,
      "loss": 0.7949066638946534,
      "mean_token_accuracy": 0.8001003921031952,
      "num_tokens": 31750886.0,
      "step": 15980
    },
    {
      "entropy": 0.8237157940864563,
      "epoch": 0.2872412089639377,
      "grad_norm": 6.46875,
      "learning_rate": 4.1284980174083425e-05,
      "loss": 0.8106029510498047,
      "mean_token_accuracy": 0.8044885590672493,
      "num_tokens": 31771212.0,
      "step": 15990
    },
    {
      "epoch": 0.2874208469933085,
      "eval_entropy": 0.769939710021019,
      "eval_loss": 0.7637472748756409,
      "eval_mean_token_accuracy": 0.8148191969394684,
      "eval_num_tokens": 31791201.0,
      "eval_runtime": 40.556,
      "eval_samples_per_second": 49.315,
      "eval_steps_per_second": 6.164,
      "step": 16000
    },
    {
      "entropy": 0.8075591046363115,
      "epoch": 0.2876004850226793,
      "grad_norm": 6.0625,
      "learning_rate": 4.12631683979102e-05,
      "loss": 0.8072579383850098,
      "mean_token_accuracy": 0.8063986800611019,
      "num_tokens": 31811837.0,
      "step": 16010
    },
    {
      "entropy": 0.7812395192682743,
      "epoch": 0.28778012305205014,
      "grad_norm": 4.4375,
      "learning_rate": 4.1252254451830866e-05,
      "loss": 0.7616629600524902,
      "mean_token_accuracy": 0.8214369758963584,
      "num_tokens": 31830167.0,
      "step": 16020
    },
    {
      "entropy": 0.734915504604578,
      "epoch": 0.28795976108142096,
      "grad_norm": 7.3125,
      "learning_rate": 4.124133513856147e-05,
      "loss": 0.7374481201171875,
      "mean_token_accuracy": 0.8198157727718354,
      "num_tokens": 31849761.0,
      "step": 16030
    },
    {
      "entropy": 0.7607235573232174,
      "epoch": 0.2881393991107917,
      "grad_norm": 6.1875,
      "learning_rate": 4.123041046170802e-05,
      "loss": 0.7706562042236328,
      "mean_token_accuracy": 0.8223676219582557,
      "num_tokens": 31869937.0,
      "step": 16040
    },
    {
      "entropy": 0.7586753569543362,
      "epoch": 0.28831903714016255,
      "grad_norm": 6.1875,
      "learning_rate": 4.1219480424878326e-05,
      "loss": 0.7715690612792969,
      "mean_token_accuracy": 0.8111238434910775,
      "num_tokens": 31889669.0,
      "step": 16050
    },
    {
      "entropy": 0.6968930706381797,
      "epoch": 0.28849867516953337,
      "grad_norm": 6.125,
      "learning_rate": 4.120854503168195e-05,
      "loss": 0.756509256362915,
      "mean_token_accuracy": 0.8189873367547988,
      "num_tokens": 31909265.0,
      "step": 16060
    },
    {
      "entropy": 0.7779805593192577,
      "epoch": 0.2886783131989042,
      "grad_norm": 9.25,
      "learning_rate": 4.119760428573024e-05,
      "loss": 0.7882138252258301,
      "mean_token_accuracy": 0.8103580631315708,
      "num_tokens": 31929716.0,
      "step": 16070
    },
    {
      "entropy": 0.8281352058053016,
      "epoch": 0.288857951228275,
      "grad_norm": 6.25,
      "learning_rate": 4.1186658190636286e-05,
      "loss": 0.8437937736511231,
      "mean_token_accuracy": 0.797195915877819,
      "num_tokens": 31948424.0,
      "step": 16080
    },
    {
      "entropy": 0.6807680770754814,
      "epoch": 0.28903758925764583,
      "grad_norm": 6.53125,
      "learning_rate": 4.117570675001495e-05,
      "loss": 0.6521598339080811,
      "mean_token_accuracy": 0.8366071119904518,
      "num_tokens": 31968102.0,
      "step": 16090
    },
    {
      "entropy": 0.741829277575016,
      "epoch": 0.28921722728701665,
      "grad_norm": 7.0625,
      "learning_rate": 4.116474996748287e-05,
      "loss": 0.7737984657287598,
      "mean_token_accuracy": 0.8136377841234207,
      "num_tokens": 31988242.0,
      "step": 16100
    },
    {
      "entropy": 0.6802665669471025,
      "epoch": 0.2893968653163875,
      "grad_norm": 5.15625,
      "learning_rate": 4.115378784665847e-05,
      "loss": 0.6773580551147461,
      "mean_token_accuracy": 0.8326177462935448,
      "num_tokens": 32007244.0,
      "step": 16110
    },
    {
      "entropy": 0.7470757000148296,
      "epoch": 0.2895765033457583,
      "grad_norm": 5.9375,
      "learning_rate": 4.114282039116187e-05,
      "loss": 0.8040597915649415,
      "mean_token_accuracy": 0.7988247752189637,
      "num_tokens": 32026112.0,
      "step": 16120
    },
    {
      "entropy": 0.734778118878603,
      "epoch": 0.2897561413751291,
      "grad_norm": 7.34375,
      "learning_rate": 4.113184760461502e-05,
      "loss": 0.7510690689086914,
      "mean_token_accuracy": 0.8199361532926559,
      "num_tokens": 32046018.0,
      "step": 16130
    },
    {
      "entropy": 0.8198709115386009,
      "epoch": 0.28993577940449994,
      "grad_norm": 5.25,
      "learning_rate": 4.1120869490641595e-05,
      "loss": 0.8106560707092285,
      "mean_token_accuracy": 0.7973447903990746,
      "num_tokens": 32065297.0,
      "step": 16140
    },
    {
      "entropy": 0.7495136730372906,
      "epoch": 0.29011541743387076,
      "grad_norm": 7.0625,
      "learning_rate": 4.110988605286704e-05,
      "loss": 0.7754366397857666,
      "mean_token_accuracy": 0.8070261612534523,
      "num_tokens": 32085864.0,
      "step": 16150
    },
    {
      "entropy": 0.7799046531319618,
      "epoch": 0.2902950554632416,
      "grad_norm": 6.34375,
      "learning_rate": 4.109889729491856e-05,
      "loss": 0.7895327091217041,
      "mean_token_accuracy": 0.8026217311620713,
      "num_tokens": 32104997.0,
      "step": 16160
    },
    {
      "entropy": 0.703456562757492,
      "epoch": 0.2904746934926124,
      "grad_norm": 5.5,
      "learning_rate": 4.108790322042512e-05,
      "loss": 0.700212049484253,
      "mean_token_accuracy": 0.8223715513944626,
      "num_tokens": 32124715.0,
      "step": 16170
    },
    {
      "entropy": 0.790971789509058,
      "epoch": 0.2906543315219832,
      "grad_norm": 11.8125,
      "learning_rate": 4.107690383301741e-05,
      "loss": 0.8490647315979004,
      "mean_token_accuracy": 0.7925594851374627,
      "num_tokens": 32144759.0,
      "step": 16180
    },
    {
      "entropy": 0.774212346971035,
      "epoch": 0.29083396955135404,
      "grad_norm": 5.625,
      "learning_rate": 4.106589913632791e-05,
      "loss": 0.7576523303985596,
      "mean_token_accuracy": 0.8170502424240113,
      "num_tokens": 32165617.0,
      "step": 16190
    },
    {
      "entropy": 0.7632520660758019,
      "epoch": 0.29101360758072486,
      "grad_norm": 6.78125,
      "learning_rate": 4.105488913399086e-05,
      "loss": 0.7882227897644043,
      "mean_token_accuracy": 0.8019206464290619,
      "num_tokens": 32184924.0,
      "step": 16200
    },
    {
      "entropy": 0.7957564160227776,
      "epoch": 0.29119324561009563,
      "grad_norm": 6.625,
      "learning_rate": 4.104387382964221e-05,
      "loss": 0.8309175491333007,
      "mean_token_accuracy": 0.8002100974321366,
      "num_tokens": 32204341.0,
      "step": 16210
    },
    {
      "entropy": 0.815641226619482,
      "epoch": 0.29137288363946645,
      "grad_norm": 5.4375,
      "learning_rate": 4.1032853226919696e-05,
      "loss": 0.8366466522216797,
      "mean_token_accuracy": 0.8038297548890114,
      "num_tokens": 32224698.0,
      "step": 16220
    },
    {
      "entropy": 0.7388044387102127,
      "epoch": 0.2915525216688373,
      "grad_norm": 4.6875,
      "learning_rate": 4.102182732946279e-05,
      "loss": 0.7544495582580566,
      "mean_token_accuracy": 0.8144376754760743,
      "num_tokens": 32245140.0,
      "step": 16230
    },
    {
      "entropy": 0.7551039382815361,
      "epoch": 0.2917321596982081,
      "grad_norm": 6.5625,
      "learning_rate": 4.101079614091271e-05,
      "loss": 0.7650655746459961,
      "mean_token_accuracy": 0.8112197160720825,
      "num_tokens": 32264956.0,
      "step": 16240
    },
    {
      "entropy": 0.7910215213894844,
      "epoch": 0.2919117977275789,
      "grad_norm": 5.25,
      "learning_rate": 4.099975966491244e-05,
      "loss": 0.8058812141418457,
      "mean_token_accuracy": 0.8040731906890869,
      "num_tokens": 32284451.0,
      "step": 16250
    },
    {
      "entropy": 0.8513581395149231,
      "epoch": 0.29209143575694974,
      "grad_norm": 6.46875,
      "learning_rate": 4.098871790510669e-05,
      "loss": 0.8738054275512696,
      "mean_token_accuracy": 0.7896746918559074,
      "num_tokens": 32303906.0,
      "step": 16260
    },
    {
      "entropy": 0.7252016447484493,
      "epoch": 0.29227107378632056,
      "grad_norm": 6.1875,
      "learning_rate": 4.097767086514191e-05,
      "loss": 0.7252104759216309,
      "mean_token_accuracy": 0.8254316568374633,
      "num_tokens": 32323234.0,
      "step": 16270
    },
    {
      "entropy": 0.6794798195362091,
      "epoch": 0.2924507118156914,
      "grad_norm": 4.71875,
      "learning_rate": 4.096661854866633e-05,
      "loss": 0.6654901027679443,
      "mean_token_accuracy": 0.8359261587262153,
      "num_tokens": 32344109.0,
      "step": 16280
    },
    {
      "entropy": 0.7859842635691165,
      "epoch": 0.2926303498450622,
      "grad_norm": 4.9375,
      "learning_rate": 4.0955560959329874e-05,
      "loss": 0.7949409484863281,
      "mean_token_accuracy": 0.8133767902851105,
      "num_tokens": 32363043.0,
      "step": 16290
    },
    {
      "entropy": 0.756204704195261,
      "epoch": 0.292809987874433,
      "grad_norm": 5.0,
      "learning_rate": 4.0944498100784244e-05,
      "loss": 0.7917475223541259,
      "mean_token_accuracy": 0.8035780057311058,
      "num_tokens": 32382831.0,
      "step": 16300
    },
    {
      "entropy": 0.6658496014773846,
      "epoch": 0.29298962590380384,
      "grad_norm": 5.65625,
      "learning_rate": 4.093342997668287e-05,
      "loss": 0.7091598510742188,
      "mean_token_accuracy": 0.8197940289974213,
      "num_tokens": 32403481.0,
      "step": 16310
    },
    {
      "entropy": 0.7673735603690147,
      "epoch": 0.29316926393317466,
      "grad_norm": 5.03125,
      "learning_rate": 4.092235659068091e-05,
      "loss": 0.7712867259979248,
      "mean_token_accuracy": 0.8052166000008583,
      "num_tokens": 32423608.0,
      "step": 16320
    },
    {
      "entropy": 0.7582424409687519,
      "epoch": 0.2933489019625455,
      "grad_norm": 4.75,
      "learning_rate": 4.0911277946435285e-05,
      "loss": 0.759792709350586,
      "mean_token_accuracy": 0.8127274006605149,
      "num_tokens": 32443711.0,
      "step": 16330
    },
    {
      "entropy": 0.7317033216357232,
      "epoch": 0.2935285399919163,
      "grad_norm": 4.6875,
      "learning_rate": 4.090019404760462e-05,
      "loss": 0.6944607257843017,
      "mean_token_accuracy": 0.8303657665848732,
      "num_tokens": 32463328.0,
      "step": 16340
    },
    {
      "entropy": 0.7394083186984062,
      "epoch": 0.2937081780212871,
      "grad_norm": 5.96875,
      "learning_rate": 4.088910489784931e-05,
      "loss": 0.7875749588012695,
      "mean_token_accuracy": 0.8133788600564003,
      "num_tokens": 32484034.0,
      "step": 16350
    },
    {
      "entropy": 0.7232165172696113,
      "epoch": 0.29388781605065795,
      "grad_norm": 4.5625,
      "learning_rate": 4.087801050083146e-05,
      "loss": 0.7252594470977783,
      "mean_token_accuracy": 0.8206925451755523,
      "num_tokens": 32502926.0,
      "step": 16360
    },
    {
      "entropy": 0.7964564621448517,
      "epoch": 0.29406745408002877,
      "grad_norm": 6.46875,
      "learning_rate": 4.0866910860214915e-05,
      "loss": 0.8127152442932128,
      "mean_token_accuracy": 0.8036402523517608,
      "num_tokens": 32522225.0,
      "step": 16370
    },
    {
      "entropy": 0.7179344594478607,
      "epoch": 0.29424709210939953,
      "grad_norm": 7.34375,
      "learning_rate": 4.0855805979665245e-05,
      "loss": 0.7233580112457275,
      "mean_token_accuracy": 0.8175023674964905,
      "num_tokens": 32541469.0,
      "step": 16380
    },
    {
      "entropy": 0.7371633421629668,
      "epoch": 0.29442673013877035,
      "grad_norm": 5.59375,
      "learning_rate": 4.084469586284977e-05,
      "loss": 0.7301100730895996,
      "mean_token_accuracy": 0.8220718190073967,
      "num_tokens": 32561720.0,
      "step": 16390
    },
    {
      "entropy": 0.8346506595611572,
      "epoch": 0.2946063681681412,
      "grad_norm": 8.125,
      "learning_rate": 4.0833580513437505e-05,
      "loss": 0.8832295417785645,
      "mean_token_accuracy": 0.7864325046539307,
      "num_tokens": 32581819.0,
      "step": 16400
    },
    {
      "entropy": 0.7887408591806888,
      "epoch": 0.294786006197512,
      "grad_norm": 4.71875,
      "learning_rate": 4.0822459935099246e-05,
      "loss": 0.7814285278320312,
      "mean_token_accuracy": 0.8054888308048248,
      "num_tokens": 32602657.0,
      "step": 16410
    },
    {
      "entropy": 0.8148703306913376,
      "epoch": 0.2949656442268828,
      "grad_norm": 5.21875,
      "learning_rate": 4.081133413150746e-05,
      "loss": 0.8244067192077636,
      "mean_token_accuracy": 0.8036095216870308,
      "num_tokens": 32620929.0,
      "step": 16420
    },
    {
      "entropy": 0.6441010862588883,
      "epoch": 0.29514528225625364,
      "grad_norm": 4.59375,
      "learning_rate": 4.0800203106336365e-05,
      "loss": 0.6752524375915527,
      "mean_token_accuracy": 0.8291705891489982,
      "num_tokens": 32641281.0,
      "step": 16430
    },
    {
      "entropy": 0.7516821302473545,
      "epoch": 0.29532492028562446,
      "grad_norm": 8.125,
      "learning_rate": 4.07890668632619e-05,
      "loss": 0.7595185279846192,
      "mean_token_accuracy": 0.8179703846573829,
      "num_tokens": 32660949.0,
      "step": 16440
    },
    {
      "entropy": 0.6725540444254875,
      "epoch": 0.2955045583149953,
      "grad_norm": 11.9375,
      "learning_rate": 4.0777925405961746e-05,
      "loss": 0.6743653297424317,
      "mean_token_accuracy": 0.8307419419288635,
      "num_tokens": 32681234.0,
      "step": 16450
    },
    {
      "entropy": 0.759384922683239,
      "epoch": 0.2956841963443661,
      "grad_norm": 4.53125,
      "learning_rate": 4.0766778738115275e-05,
      "loss": 0.7526276111602783,
      "mean_token_accuracy": 0.8164209336042404,
      "num_tokens": 32702043.0,
      "step": 16460
    },
    {
      "entropy": 0.7467499770224094,
      "epoch": 0.2958638343737369,
      "grad_norm": 4.9375,
      "learning_rate": 4.0755626863403605e-05,
      "loss": 0.7757118701934814,
      "mean_token_accuracy": 0.8179592102766037,
      "num_tokens": 32722579.0,
      "step": 16470
    },
    {
      "entropy": 0.79761051684618,
      "epoch": 0.29604347240310774,
      "grad_norm": 5.46875,
      "learning_rate": 4.074446978550954e-05,
      "loss": 0.8320376396179199,
      "mean_token_accuracy": 0.8091583713889122,
      "num_tokens": 32742948.0,
      "step": 16480
    },
    {
      "entropy": 0.7638556130230427,
      "epoch": 0.29622311043247856,
      "grad_norm": 5.03125,
      "learning_rate": 4.0733307508117654e-05,
      "loss": 0.706824779510498,
      "mean_token_accuracy": 0.8254311710596085,
      "num_tokens": 32762214.0,
      "step": 16490
    },
    {
      "entropy": 0.7699882030487061,
      "epoch": 0.2964027484618494,
      "grad_norm": 9.0,
      "learning_rate": 4.072214003491418e-05,
      "loss": 0.790395975112915,
      "mean_token_accuracy": 0.8132738694548607,
      "num_tokens": 32781151.0,
      "step": 16500
    },
    {
      "entropy": 0.7158139243721962,
      "epoch": 0.2965823864912202,
      "grad_norm": 6.0625,
      "learning_rate": 4.0710967369587126e-05,
      "loss": 0.7447795391082763,
      "mean_token_accuracy": 0.8179330617189408,
      "num_tokens": 32801289.0,
      "step": 16510
    },
    {
      "entropy": 0.7331295058131218,
      "epoch": 0.29676202452059103,
      "grad_norm": 7.40625,
      "learning_rate": 4.069978951582616e-05,
      "loss": 0.7282636642456055,
      "mean_token_accuracy": 0.8215281695127488,
      "num_tokens": 32821438.0,
      "step": 16520
    },
    {
      "entropy": 0.6937525883316994,
      "epoch": 0.29694166254996185,
      "grad_norm": 4.78125,
      "learning_rate": 4.0688606477322695e-05,
      "loss": 0.6796750545501709,
      "mean_token_accuracy": 0.8300207406282425,
      "num_tokens": 32840712.0,
      "step": 16530
    },
    {
      "entropy": 0.7728459522128105,
      "epoch": 0.29712130057933267,
      "grad_norm": 8.6875,
      "learning_rate": 4.067741825776985e-05,
      "loss": 0.8052539825439453,
      "mean_token_accuracy": 0.8005506187677384,
      "num_tokens": 32861526.0,
      "step": 16540
    },
    {
      "entropy": 0.6426697306334972,
      "epoch": 0.29730093860870344,
      "grad_norm": 5.03125,
      "learning_rate": 4.0666224860862454e-05,
      "loss": 0.6739311695098877,
      "mean_token_accuracy": 0.8328271776437759,
      "num_tokens": 32882231.0,
      "step": 16550
    },
    {
      "entropy": 0.791443296521902,
      "epoch": 0.29748057663807426,
      "grad_norm": 5.40625,
      "learning_rate": 4.065502629029704e-05,
      "loss": 0.7856956481933594,
      "mean_token_accuracy": 0.8074075743556023,
      "num_tokens": 32901687.0,
      "step": 16560
    },
    {
      "entropy": 0.7727358289062977,
      "epoch": 0.2976602146674451,
      "grad_norm": 4.5625,
      "learning_rate": 4.064382254977187e-05,
      "loss": 0.7975558280944824,
      "mean_token_accuracy": 0.8042471423745156,
      "num_tokens": 32922222.0,
      "step": 16570
    },
    {
      "entropy": 0.7073809765279293,
      "epoch": 0.2978398526968159,
      "grad_norm": 6.5,
      "learning_rate": 4.063261364298688e-05,
      "loss": 0.714824914932251,
      "mean_token_accuracy": 0.8312507584691048,
      "num_tokens": 32941124.0,
      "step": 16580
    },
    {
      "entropy": 0.8149250656366348,
      "epoch": 0.2980194907261867,
      "grad_norm": 5.90625,
      "learning_rate": 4.0621399573643754e-05,
      "loss": 0.8715660095214843,
      "mean_token_accuracy": 0.8037385821342469,
      "num_tokens": 32961277.0,
      "step": 16590
    },
    {
      "entropy": 0.8408456407487392,
      "epoch": 0.29819912875555754,
      "grad_norm": 6.5,
      "learning_rate": 4.061018034544584e-05,
      "loss": 0.8403009414672852,
      "mean_token_accuracy": 0.7976683929562569,
      "num_tokens": 32980032.0,
      "step": 16600
    },
    {
      "entropy": 0.731586092710495,
      "epoch": 0.29837876678492836,
      "grad_norm": 5.28125,
      "learning_rate": 4.0598955962098205e-05,
      "loss": 0.7359925270080566,
      "mean_token_accuracy": 0.823723541200161,
      "num_tokens": 32999408.0,
      "step": 16610
    },
    {
      "entropy": 0.6712837897241115,
      "epoch": 0.2985584048142992,
      "grad_norm": 5.71875,
      "learning_rate": 4.058772642730764e-05,
      "loss": 0.6713774681091309,
      "mean_token_accuracy": 0.8351075023412704,
      "num_tokens": 33019089.0,
      "step": 16620
    },
    {
      "entropy": 0.8408644244074821,
      "epoch": 0.29873804284367,
      "grad_norm": 5.0,
      "learning_rate": 4.057649174478259e-05,
      "loss": 0.8882511138916016,
      "mean_token_accuracy": 0.79170081615448,
      "num_tokens": 33039611.0,
      "step": 16630
    },
    {
      "entropy": 0.7752871446311473,
      "epoch": 0.2989176808730408,
      "grad_norm": 5.53125,
      "learning_rate": 4.056525191823325e-05,
      "loss": 0.8054216384887696,
      "mean_token_accuracy": 0.8082266628742218,
      "num_tokens": 33058969.0,
      "step": 16640
    },
    {
      "entropy": 0.7794643566012383,
      "epoch": 0.29909731890241165,
      "grad_norm": 5.4375,
      "learning_rate": 4.055400695137149e-05,
      "loss": 0.7833192348480225,
      "mean_token_accuracy": 0.8092082172632218,
      "num_tokens": 33078854.0,
      "step": 16650
    },
    {
      "entropy": 0.8296513441950083,
      "epoch": 0.29927695693178247,
      "grad_norm": 5.53125,
      "learning_rate": 4.054275684791087e-05,
      "loss": 0.8304566383361817,
      "mean_token_accuracy": 0.8034424290060997,
      "num_tokens": 33099081.0,
      "step": 16660
    },
    {
      "entropy": 0.8378834009170533,
      "epoch": 0.2994565949611533,
      "grad_norm": 6.96875,
      "learning_rate": 4.053150161156666e-05,
      "loss": 0.832552719116211,
      "mean_token_accuracy": 0.8049099296331406,
      "num_tokens": 33119046.0,
      "step": 16670
    },
    {
      "entropy": 0.7404852122068405,
      "epoch": 0.2996362329905241,
      "grad_norm": 5.375,
      "learning_rate": 4.052024124605582e-05,
      "loss": 0.7401700973510742,
      "mean_token_accuracy": 0.8271591573953628,
      "num_tokens": 33137117.0,
      "step": 16680
    },
    {
      "entropy": 0.7142466381192207,
      "epoch": 0.29981587101989493,
      "grad_norm": 5.28125,
      "learning_rate": 4.0508975755097e-05,
      "loss": 0.715613317489624,
      "mean_token_accuracy": 0.8214541643857955,
      "num_tokens": 33157837.0,
      "step": 16690
    },
    {
      "entropy": 0.6744240231812,
      "epoch": 0.29999550904926575,
      "grad_norm": 5.5,
      "learning_rate": 4.049770514241057e-05,
      "loss": 0.6703217029571533,
      "mean_token_accuracy": 0.8276666879653931,
      "num_tokens": 33176952.0,
      "step": 16700
    },
    {
      "entropy": 0.7664406470954418,
      "epoch": 0.3001751470786366,
      "grad_norm": 7.15625,
      "learning_rate": 4.048642941171854e-05,
      "loss": 0.7839531898498535,
      "mean_token_accuracy": 0.8130544558167457,
      "num_tokens": 33196826.0,
      "step": 16710
    },
    {
      "entropy": 0.8311275757849217,
      "epoch": 0.30035478510800734,
      "grad_norm": 8.25,
      "learning_rate": 4.047514856674465e-05,
      "loss": 0.9031271934509277,
      "mean_token_accuracy": 0.7948875650763512,
      "num_tokens": 33217126.0,
      "step": 16720
    },
    {
      "entropy": 0.7347884178161621,
      "epoch": 0.30053442313737816,
      "grad_norm": 8.0625,
      "learning_rate": 4.046386261121432e-05,
      "loss": 0.7311437606811524,
      "mean_token_accuracy": 0.816559998691082,
      "num_tokens": 33236767.0,
      "step": 16730
    },
    {
      "entropy": 0.7218950673937797,
      "epoch": 0.300714061166749,
      "grad_norm": 5.375,
      "learning_rate": 4.0452571548854656e-05,
      "loss": 0.7324501037597656,
      "mean_token_accuracy": 0.8159527018666267,
      "num_tokens": 33255966.0,
      "step": 16740
    },
    {
      "entropy": 0.890400905907154,
      "epoch": 0.3008936991961198,
      "grad_norm": 7.59375,
      "learning_rate": 4.044127538339445e-05,
      "loss": 0.8462236404418946,
      "mean_token_accuracy": 0.785409215092659,
      "num_tokens": 33275751.0,
      "step": 16750
    },
    {
      "entropy": 0.7645746022462845,
      "epoch": 0.3010733372254906,
      "grad_norm": 5.15625,
      "learning_rate": 4.0429974118564175e-05,
      "loss": 0.7906983375549317,
      "mean_token_accuracy": 0.8026041612029076,
      "num_tokens": 33295250.0,
      "step": 16760
    },
    {
      "entropy": 0.7720949843525886,
      "epoch": 0.30125297525486144,
      "grad_norm": 6.0625,
      "learning_rate": 4.0418667758095995e-05,
      "loss": 0.7648301601409913,
      "mean_token_accuracy": 0.8153431460261344,
      "num_tokens": 33316131.0,
      "step": 16770
    },
    {
      "entropy": 0.6727171760052443,
      "epoch": 0.30143261328423226,
      "grad_norm": 4.71875,
      "learning_rate": 4.040735630572375e-05,
      "loss": 0.6772083282470703,
      "mean_token_accuracy": 0.8367044433951378,
      "num_tokens": 33335935.0,
      "step": 16780
    },
    {
      "entropy": 0.8031598135828972,
      "epoch": 0.3016122513136031,
      "grad_norm": 5.78125,
      "learning_rate": 4.039603976518298e-05,
      "loss": 0.8405230522155762,
      "mean_token_accuracy": 0.8036567464470863,
      "num_tokens": 33355415.0,
      "step": 16790
    },
    {
      "entropy": 0.8119475387036801,
      "epoch": 0.3017918893429739,
      "grad_norm": 5.40625,
      "learning_rate": 4.038471814021088e-05,
      "loss": 0.8467553138732911,
      "mean_token_accuracy": 0.8047595426440239,
      "num_tokens": 33375713.0,
      "step": 16800
    },
    {
      "entropy": 0.8614579483866691,
      "epoch": 0.30197152737234473,
      "grad_norm": 6.34375,
      "learning_rate": 4.0373391434546324e-05,
      "loss": 0.8881271362304688,
      "mean_token_accuracy": 0.7956246256828308,
      "num_tokens": 33397461.0,
      "step": 16810
    },
    {
      "entropy": 0.8732067100703717,
      "epoch": 0.30215116540171555,
      "grad_norm": 6.34375,
      "learning_rate": 4.03620596519299e-05,
      "loss": 0.8863471984863281,
      "mean_token_accuracy": 0.7939257130026818,
      "num_tokens": 33417282.0,
      "step": 16820
    },
    {
      "entropy": 0.7578549019992351,
      "epoch": 0.30233080343108637,
      "grad_norm": 7.21875,
      "learning_rate": 4.035072279610384e-05,
      "loss": 0.7369836330413818,
      "mean_token_accuracy": 0.8270233556628227,
      "num_tokens": 33437036.0,
      "step": 16830
    },
    {
      "entropy": 0.8801611095666886,
      "epoch": 0.3025104414604572,
      "grad_norm": 6.3125,
      "learning_rate": 4.0339380870812036e-05,
      "loss": 0.913973331451416,
      "mean_token_accuracy": 0.7854417368769646,
      "num_tokens": 33458374.0,
      "step": 16840
    },
    {
      "entropy": 0.6873669430613518,
      "epoch": 0.302690079489828,
      "grad_norm": 8.25,
      "learning_rate": 4.0328033879800106e-05,
      "loss": 0.6995107650756835,
      "mean_token_accuracy": 0.8247468620538712,
      "num_tokens": 33478410.0,
      "step": 16850
    },
    {
      "entropy": 0.7423675902187824,
      "epoch": 0.30286971751919883,
      "grad_norm": 5.09375,
      "learning_rate": 4.0316681826815295e-05,
      "loss": 0.7063494682312011,
      "mean_token_accuracy": 0.8239742875099182,
      "num_tokens": 33499461.0,
      "step": 16860
    },
    {
      "entropy": 0.7965792261064053,
      "epoch": 0.30304935554856965,
      "grad_norm": 6.5,
      "learning_rate": 4.0305324715606543e-05,
      "loss": 0.8206728935241699,
      "mean_token_accuracy": 0.8042060241103173,
      "num_tokens": 33518451.0,
      "step": 16870
    },
    {
      "entropy": 0.7631928488612175,
      "epoch": 0.3032289935779405,
      "grad_norm": 6.96875,
      "learning_rate": 4.0293962549924466e-05,
      "loss": 0.7700592517852783,
      "mean_token_accuracy": 0.8086611896753311,
      "num_tokens": 33538199.0,
      "step": 16880
    },
    {
      "entropy": 0.7288111381232738,
      "epoch": 0.30340863160731124,
      "grad_norm": 4.90625,
      "learning_rate": 4.028259533352131e-05,
      "loss": 0.7613233089447021,
      "mean_token_accuracy": 0.8105173423886299,
      "num_tokens": 33557175.0,
      "step": 16890
    },
    {
      "entropy": 0.678798970580101,
      "epoch": 0.30358826963668206,
      "grad_norm": 6.15625,
      "learning_rate": 4.027122307015104e-05,
      "loss": 0.7268945693969726,
      "mean_token_accuracy": 0.8203946098685264,
      "num_tokens": 33576453.0,
      "step": 16900
    },
    {
      "entropy": 0.7926782719790936,
      "epoch": 0.3037679076660529,
      "grad_norm": 6.03125,
      "learning_rate": 4.025984576356926e-05,
      "loss": 0.8218985557556152,
      "mean_token_accuracy": 0.8051017701625824,
      "num_tokens": 33595742.0,
      "step": 16910
    },
    {
      "entropy": 0.8191024236381054,
      "epoch": 0.3039475456954237,
      "grad_norm": 7.1875,
      "learning_rate": 4.0248463417533234e-05,
      "loss": 0.8528058052062988,
      "mean_token_accuracy": 0.7965284943580627,
      "num_tokens": 33614596.0,
      "step": 16920
    },
    {
      "entropy": 0.7082028731703758,
      "epoch": 0.3041271837247945,
      "grad_norm": 3.859375,
      "learning_rate": 4.023707603580191e-05,
      "loss": 0.7071847915649414,
      "mean_token_accuracy": 0.8188983902335167,
      "num_tokens": 33636056.0,
      "step": 16930
    },
    {
      "entropy": 0.7995364531874657,
      "epoch": 0.30430682175416535,
      "grad_norm": 4.53125,
      "learning_rate": 4.022568362213588e-05,
      "loss": 0.811736011505127,
      "mean_token_accuracy": 0.8118455722928047,
      "num_tokens": 33655485.0,
      "step": 16940
    },
    {
      "entropy": 0.7739044696092605,
      "epoch": 0.30448645978353617,
      "grad_norm": 6.65625,
      "learning_rate": 4.0214286180297414e-05,
      "loss": 0.7538593292236329,
      "mean_token_accuracy": 0.8164787590503693,
      "num_tokens": 33674896.0,
      "step": 16950
    },
    {
      "entropy": 0.8529946267604828,
      "epoch": 0.304666097812907,
      "grad_norm": 5.53125,
      "learning_rate": 4.020288371405043e-05,
      "loss": 0.8298836708068847,
      "mean_token_accuracy": 0.7960230588912964,
      "num_tokens": 33694522.0,
      "step": 16960
    },
    {
      "entropy": 0.7227740488946438,
      "epoch": 0.3048457358422778,
      "grad_norm": 4.71875,
      "learning_rate": 4.019147622716052e-05,
      "loss": 0.7262112140655518,
      "mean_token_accuracy": 0.8117696195840836,
      "num_tokens": 33713848.0,
      "step": 16970
    },
    {
      "entropy": 0.8260045371949672,
      "epoch": 0.30502537387164863,
      "grad_norm": 7.9375,
      "learning_rate": 4.018006372339491e-05,
      "loss": 0.821468448638916,
      "mean_token_accuracy": 0.7979263037443161,
      "num_tokens": 33734382.0,
      "step": 16980
    },
    {
      "entropy": 0.7398630760610103,
      "epoch": 0.30520501190101945,
      "grad_norm": 6.78125,
      "learning_rate": 4.01686462065225e-05,
      "loss": 0.7616140365600585,
      "mean_token_accuracy": 0.810047110915184,
      "num_tokens": 33754326.0,
      "step": 16990
    },
    {
      "epoch": 0.3053846499303903,
      "eval_entropy": 0.746131109714508,
      "eval_loss": 0.7473167777061462,
      "eval_mean_token_accuracy": 0.8186712293624878,
      "eval_num_tokens": 33773663.0,
      "eval_runtime": 40.5441,
      "eval_samples_per_second": 49.329,
      "eval_steps_per_second": 6.166,
      "step": 17000
    },
    {
      "entropy": 0.752722242474556,
      "epoch": 0.3055642879597611,
      "grad_norm": 6.3125,
      "learning_rate": 4.014579614854117e-05,
      "loss": 0.7757734298706055,
      "mean_token_accuracy": 0.8113306805491447,
      "num_tokens": 33793483.0,
      "step": 17010
    },
    {
      "entropy": 0.7913009047508239,
      "epoch": 0.3057439259891319,
      "grad_norm": 5.875,
      "learning_rate": 4.013436361497832e-05,
      "loss": 0.7943925857543945,
      "mean_token_accuracy": 0.8029153510928154,
      "num_tokens": 33812581.0,
      "step": 17020
    },
    {
      "entropy": 0.7796534996479749,
      "epoch": 0.30592356401850274,
      "grad_norm": 5.0,
      "learning_rate": 4.0122926083400795e-05,
      "loss": 0.7848310947418213,
      "mean_token_accuracy": 0.8077147454023361,
      "num_tokens": 33833401.0,
      "step": 17030
    },
    {
      "entropy": 0.7849590167403221,
      "epoch": 0.30610320204787356,
      "grad_norm": 7.5625,
      "learning_rate": 4.0111483557585774e-05,
      "loss": 0.7763415336608886,
      "mean_token_accuracy": 0.8105032786726951,
      "num_tokens": 33853978.0,
      "step": 17040
    },
    {
      "entropy": 0.7379834957420826,
      "epoch": 0.3062828400772444,
      "grad_norm": 6.65625,
      "learning_rate": 4.0100036041312064e-05,
      "loss": 0.7281891345977783,
      "mean_token_accuracy": 0.8167565539479256,
      "num_tokens": 33875489.0,
      "step": 17050
    },
    {
      "entropy": 0.7870235066860914,
      "epoch": 0.30646247810661514,
      "grad_norm": 7.28125,
      "learning_rate": 4.008858353836012e-05,
      "loss": 0.8118382453918457,
      "mean_token_accuracy": 0.8102093383669853,
      "num_tokens": 33895021.0,
      "step": 17060
    },
    {
      "entropy": 0.7490365158766508,
      "epoch": 0.30664211613598596,
      "grad_norm": 5.96875,
      "learning_rate": 4.007712605251206e-05,
      "loss": 0.7716125011444092,
      "mean_token_accuracy": 0.8152915611863136,
      "num_tokens": 33914380.0,
      "step": 17070
    },
    {
      "entropy": 0.7463030867278576,
      "epoch": 0.3068217541653568,
      "grad_norm": 4.5625,
      "learning_rate": 4.0065663587551624e-05,
      "loss": 0.7154333591461182,
      "mean_token_accuracy": 0.8198764607310295,
      "num_tokens": 33933039.0,
      "step": 17080
    },
    {
      "entropy": 0.862423637509346,
      "epoch": 0.3070013921947276,
      "grad_norm": 4.78125,
      "learning_rate": 4.005419614726421e-05,
      "loss": 0.8879791259765625,
      "mean_token_accuracy": 0.7910993218421936,
      "num_tokens": 33954758.0,
      "step": 17090
    },
    {
      "entropy": 0.7775037899613381,
      "epoch": 0.30718103022409843,
      "grad_norm": 6.4375,
      "learning_rate": 4.004272373543687e-05,
      "loss": 0.7843568801879883,
      "mean_token_accuracy": 0.8033167466521263,
      "num_tokens": 33974274.0,
      "step": 17100
    },
    {
      "entropy": 0.7304410509765148,
      "epoch": 0.30736066825346925,
      "grad_norm": 6.71875,
      "learning_rate": 4.003124635585828e-05,
      "loss": 0.7187887668609619,
      "mean_token_accuracy": 0.8203001841902733,
      "num_tokens": 33994673.0,
      "step": 17110
    },
    {
      "entropy": 0.7252669785171747,
      "epoch": 0.30754030628284007,
      "grad_norm": 4.9375,
      "learning_rate": 4.001976401231875e-05,
      "loss": 0.7237717151641846,
      "mean_token_accuracy": 0.822287268191576,
      "num_tokens": 34014016.0,
      "step": 17120
    },
    {
      "entropy": 0.6598985336720944,
      "epoch": 0.3077199443122109,
      "grad_norm": 3.9375,
      "learning_rate": 4.000827670861025e-05,
      "loss": 0.6747980117797852,
      "mean_token_accuracy": 0.8322231858968735,
      "num_tokens": 34033941.0,
      "step": 17130
    },
    {
      "entropy": 0.6381131887435914,
      "epoch": 0.3078995823415817,
      "grad_norm": 5.15625,
      "learning_rate": 3.999678444852638e-05,
      "loss": 0.6548916339874268,
      "mean_token_accuracy": 0.8374187156558037,
      "num_tokens": 34054314.0,
      "step": 17140
    },
    {
      "entropy": 0.7456407353281975,
      "epoch": 0.30807922037095253,
      "grad_norm": 7.84375,
      "learning_rate": 3.9985287235862385e-05,
      "loss": 0.7782154083251953,
      "mean_token_accuracy": 0.8080860584974289,
      "num_tokens": 34073926.0,
      "step": 17150
    },
    {
      "entropy": 0.7674588494002819,
      "epoch": 0.30825885840032335,
      "grad_norm": 7.8125,
      "learning_rate": 3.997378507441512e-05,
      "loss": 0.806364917755127,
      "mean_token_accuracy": 0.7930010914802551,
      "num_tokens": 34094428.0,
      "step": 17160
    },
    {
      "entropy": 0.7215720698237419,
      "epoch": 0.3084384964296942,
      "grad_norm": 5.25,
      "learning_rate": 3.99622779679831e-05,
      "loss": 0.7024405002593994,
      "mean_token_accuracy": 0.818800701200962,
      "num_tokens": 34114270.0,
      "step": 17170
    },
    {
      "entropy": 0.7715274028480053,
      "epoch": 0.308618134459065,
      "grad_norm": 6.875,
      "learning_rate": 3.995076592036646e-05,
      "loss": 0.8163088798522949,
      "mean_token_accuracy": 0.8038177058100701,
      "num_tokens": 34133286.0,
      "step": 17180
    },
    {
      "entropy": 0.8199379093945026,
      "epoch": 0.3087977724884358,
      "grad_norm": 6.4375,
      "learning_rate": 3.993924893536698e-05,
      "loss": 0.8404458045959473,
      "mean_token_accuracy": 0.8024330317974091,
      "num_tokens": 34152962.0,
      "step": 17190
    },
    {
      "entropy": 0.8414850845932961,
      "epoch": 0.30897741051780664,
      "grad_norm": 5.625,
      "learning_rate": 3.992772701678805e-05,
      "loss": 0.8890132904052734,
      "mean_token_accuracy": 0.7895616620779038,
      "num_tokens": 34174493.0,
      "step": 17200
    },
    {
      "entropy": 0.7950747579336166,
      "epoch": 0.30915704854717746,
      "grad_norm": 9.125,
      "learning_rate": 3.991620016843471e-05,
      "loss": 0.7995020866394043,
      "mean_token_accuracy": 0.8052304580807685,
      "num_tokens": 34195334.0,
      "step": 17210
    },
    {
      "entropy": 0.7398949205875397,
      "epoch": 0.3093366865765483,
      "grad_norm": 9.125,
      "learning_rate": 3.990466839411361e-05,
      "loss": 0.7757128238677978,
      "mean_token_accuracy": 0.8195069462060929,
      "num_tokens": 34214521.0,
      "step": 17220
    },
    {
      "entropy": 0.7580545455217361,
      "epoch": 0.30951632460591905,
      "grad_norm": 5.75,
      "learning_rate": 3.989313169763304e-05,
      "loss": 0.7773032665252686,
      "mean_token_accuracy": 0.8153793126344681,
      "num_tokens": 34235900.0,
      "step": 17230
    },
    {
      "entropy": 0.7163580529391765,
      "epoch": 0.30969596263528987,
      "grad_norm": 5.125,
      "learning_rate": 3.988159008280292e-05,
      "loss": 0.7127536296844482,
      "mean_token_accuracy": 0.8304613947868347,
      "num_tokens": 34254993.0,
      "step": 17240
    },
    {
      "entropy": 0.7382263816893101,
      "epoch": 0.3098756006646607,
      "grad_norm": 5.21875,
      "learning_rate": 3.987004355343477e-05,
      "loss": 0.7537627696990967,
      "mean_token_accuracy": 0.8176436588168144,
      "num_tokens": 34275022.0,
      "step": 17250
    },
    {
      "entropy": 0.7582098595798016,
      "epoch": 0.3100552386940315,
      "grad_norm": 6.90625,
      "learning_rate": 3.985849211334175e-05,
      "loss": 0.7572234630584717,
      "mean_token_accuracy": 0.8098289221525192,
      "num_tokens": 34295122.0,
      "step": 17260
    },
    {
      "entropy": 0.8045599684119225,
      "epoch": 0.31023487672340233,
      "grad_norm": 9.0,
      "learning_rate": 3.984693576633865e-05,
      "loss": 0.8160170555114746,
      "mean_token_accuracy": 0.7962881594896316,
      "num_tokens": 34314026.0,
      "step": 17270
    },
    {
      "entropy": 0.8006197161972523,
      "epoch": 0.31041451475277315,
      "grad_norm": 6.15625,
      "learning_rate": 3.9835374516241865e-05,
      "loss": 0.7819230079650878,
      "mean_token_accuracy": 0.8114110291004181,
      "num_tokens": 34333511.0,
      "step": 17280
    },
    {
      "entropy": 0.7648152396082878,
      "epoch": 0.310594152782144,
      "grad_norm": 5.0625,
      "learning_rate": 3.982380836686941e-05,
      "loss": 0.7909525871276856,
      "mean_token_accuracy": 0.8103834196925164,
      "num_tokens": 34353447.0,
      "step": 17290
    },
    {
      "entropy": 0.7797297820448875,
      "epoch": 0.3107737908115148,
      "grad_norm": 5.5625,
      "learning_rate": 3.9812237322040937e-05,
      "loss": 0.7945465564727783,
      "mean_token_accuracy": 0.8063509717583657,
      "num_tokens": 34373616.0,
      "step": 17300
    },
    {
      "entropy": 0.7823939166963101,
      "epoch": 0.3109534288408856,
      "grad_norm": 7.9375,
      "learning_rate": 3.980066138557768e-05,
      "loss": 0.7792452812194824,
      "mean_token_accuracy": 0.8049327626824379,
      "num_tokens": 34393492.0,
      "step": 17310
    },
    {
      "entropy": 0.7256773076951504,
      "epoch": 0.31113306687025644,
      "grad_norm": 4.21875,
      "learning_rate": 3.978908056130252e-05,
      "loss": 0.7022278785705567,
      "mean_token_accuracy": 0.8217857733368874,
      "num_tokens": 34413337.0,
      "step": 17320
    },
    {
      "entropy": 0.7851140700280667,
      "epoch": 0.31131270489962726,
      "grad_norm": 3.90625,
      "learning_rate": 3.977749485303994e-05,
      "loss": 0.7973503112792969,
      "mean_token_accuracy": 0.8053277119994163,
      "num_tokens": 34434119.0,
      "step": 17330
    },
    {
      "entropy": 0.6800014093518257,
      "epoch": 0.3114923429289981,
      "grad_norm": 5.53125,
      "learning_rate": 3.9765904264616036e-05,
      "loss": 0.7107971668243408,
      "mean_token_accuracy": 0.8358748331665993,
      "num_tokens": 34454343.0,
      "step": 17340
    },
    {
      "entropy": 0.7413935773074627,
      "epoch": 0.3116719809583689,
      "grad_norm": 6.0625,
      "learning_rate": 3.975430879985852e-05,
      "loss": 0.7587538719177246,
      "mean_token_accuracy": 0.8167598187923432,
      "num_tokens": 34475470.0,
      "step": 17350
    },
    {
      "entropy": 0.8088782027363777,
      "epoch": 0.3118516189877397,
      "grad_norm": 8.5,
      "learning_rate": 3.9742708462596714e-05,
      "loss": 0.8073749542236328,
      "mean_token_accuracy": 0.802432382106781,
      "num_tokens": 34495786.0,
      "step": 17360
    },
    {
      "entropy": 0.7750293768942356,
      "epoch": 0.31203125701711054,
      "grad_norm": 5.84375,
      "learning_rate": 3.9731103256661525e-05,
      "loss": 0.8018113136291504,
      "mean_token_accuracy": 0.8028248488903046,
      "num_tokens": 34515441.0,
      "step": 17370
    },
    {
      "entropy": 0.7360148802399635,
      "epoch": 0.31221089504648136,
      "grad_norm": 4.9375,
      "learning_rate": 3.971949318588551e-05,
      "loss": 0.7195657253265381,
      "mean_token_accuracy": 0.8264596506953239,
      "num_tokens": 34535470.0,
      "step": 17380
    },
    {
      "entropy": 0.7150483272969723,
      "epoch": 0.3123905330758522,
      "grad_norm": 6.71875,
      "learning_rate": 3.9707878254102816e-05,
      "loss": 0.6925189018249511,
      "mean_token_accuracy": 0.8243127301335335,
      "num_tokens": 34555607.0,
      "step": 17390
    },
    {
      "entropy": 0.7102046512067318,
      "epoch": 0.31257017110522295,
      "grad_norm": 5.90625,
      "learning_rate": 3.969625846514916e-05,
      "loss": 0.7089354038238526,
      "mean_token_accuracy": 0.8217187836766243,
      "num_tokens": 34575189.0,
      "step": 17400
    },
    {
      "entropy": 0.7956078216433525,
      "epoch": 0.31274980913459377,
      "grad_norm": 8.0,
      "learning_rate": 3.968463382286193e-05,
      "loss": 0.79115309715271,
      "mean_token_accuracy": 0.8102262482047081,
      "num_tokens": 34594153.0,
      "step": 17410
    },
    {
      "entropy": 0.7971115879714489,
      "epoch": 0.3129294471639646,
      "grad_norm": 6.25,
      "learning_rate": 3.9673004331080055e-05,
      "loss": 0.8102178573608398,
      "mean_token_accuracy": 0.8009338974952698,
      "num_tokens": 34613927.0,
      "step": 17420
    },
    {
      "entropy": 0.7245143163949251,
      "epoch": 0.3131090851933354,
      "grad_norm": 6.09375,
      "learning_rate": 3.9661369993644106e-05,
      "loss": 0.7183018684387207,
      "mean_token_accuracy": 0.8191718652844429,
      "num_tokens": 34634157.0,
      "step": 17430
    },
    {
      "entropy": 0.7351012133061886,
      "epoch": 0.31328872322270623,
      "grad_norm": 7.625,
      "learning_rate": 3.964973081439622e-05,
      "loss": 0.7541585922241211,
      "mean_token_accuracy": 0.8216377049684525,
      "num_tokens": 34653281.0,
      "step": 17440
    },
    {
      "entropy": 0.6604509472846984,
      "epoch": 0.31346836125207705,
      "grad_norm": 6.1875,
      "learning_rate": 3.963808679718018e-05,
      "loss": 0.6379446506500244,
      "mean_token_accuracy": 0.840779747068882,
      "num_tokens": 34673070.0,
      "step": 17450
    },
    {
      "entropy": 0.7899170063436032,
      "epoch": 0.3136479992814479,
      "grad_norm": 5.0625,
      "learning_rate": 3.962643794584131e-05,
      "loss": 0.783599853515625,
      "mean_token_accuracy": 0.8054482564330101,
      "num_tokens": 34693375.0,
      "step": 17460
    },
    {
      "entropy": 0.6961737267673016,
      "epoch": 0.3138276373108187,
      "grad_norm": 6.3125,
      "learning_rate": 3.9614784264226576e-05,
      "loss": 0.7157501220703125,
      "mean_token_accuracy": 0.8270401880145073,
      "num_tokens": 34712734.0,
      "step": 17470
    },
    {
      "entropy": 0.6720437578856945,
      "epoch": 0.3140072753401895,
      "grad_norm": 7.4375,
      "learning_rate": 3.9603125756184525e-05,
      "loss": 0.7000727176666259,
      "mean_token_accuracy": 0.8251109153032303,
      "num_tokens": 34732350.0,
      "step": 17480
    },
    {
      "entropy": 0.6779511567205191,
      "epoch": 0.31418691336956034,
      "grad_norm": 4.78125,
      "learning_rate": 3.959146242556529e-05,
      "loss": 0.6967905521392822,
      "mean_token_accuracy": 0.8350613325834274,
      "num_tokens": 34752358.0,
      "step": 17490
    },
    {
      "entropy": 0.7495093815028667,
      "epoch": 0.31436655139893116,
      "grad_norm": 5.53125,
      "learning_rate": 3.95797942762206e-05,
      "loss": 0.8074655532836914,
      "mean_token_accuracy": 0.8126159369945526,
      "num_tokens": 34772085.0,
      "step": 17500
    },
    {
      "entropy": 0.7551707163453102,
      "epoch": 0.314546189428302,
      "grad_norm": 5.875,
      "learning_rate": 3.956812131200378e-05,
      "loss": 0.7482800960540772,
      "mean_token_accuracy": 0.8156968057155609,
      "num_tokens": 34791957.0,
      "step": 17510
    },
    {
      "entropy": 0.7427793987095356,
      "epoch": 0.3147258274576728,
      "grad_norm": 7.28125,
      "learning_rate": 3.955644353676974e-05,
      "loss": 0.7400613307952881,
      "mean_token_accuracy": 0.8234319016337395,
      "num_tokens": 34810755.0,
      "step": 17520
    },
    {
      "entropy": 0.7959214210510254,
      "epoch": 0.3149054654870436,
      "grad_norm": 4.96875,
      "learning_rate": 3.9544760954374985e-05,
      "loss": 0.791872215270996,
      "mean_token_accuracy": 0.8023567460477352,
      "num_tokens": 34829525.0,
      "step": 17530
    },
    {
      "entropy": 0.7121578104794025,
      "epoch": 0.31508510351641444,
      "grad_norm": 5.9375,
      "learning_rate": 3.95330735686776e-05,
      "loss": 0.6979360103607177,
      "mean_token_accuracy": 0.8324784830212593,
      "num_tokens": 34848684.0,
      "step": 17540
    },
    {
      "entropy": 0.7069754101336002,
      "epoch": 0.31526474154578527,
      "grad_norm": 6.40625,
      "learning_rate": 3.952138138353726e-05,
      "loss": 0.7719102382659913,
      "mean_token_accuracy": 0.8122841611504554,
      "num_tokens": 34867438.0,
      "step": 17550
    },
    {
      "entropy": 0.7023864611983299,
      "epoch": 0.3154443795751561,
      "grad_norm": 8.3125,
      "learning_rate": 3.950968440281524e-05,
      "loss": 0.6983365058898926,
      "mean_token_accuracy": 0.8294951304793358,
      "num_tokens": 34886713.0,
      "step": 17560
    },
    {
      "entropy": 0.7242735005915165,
      "epoch": 0.31562401760452685,
      "grad_norm": 4.875,
      "learning_rate": 3.949798263037435e-05,
      "loss": 0.7602056503295899,
      "mean_token_accuracy": 0.8163086891174316,
      "num_tokens": 34907512.0,
      "step": 17570
    },
    {
      "entropy": 0.8385130237787962,
      "epoch": 0.3158036556338977,
      "grad_norm": 6.0625,
      "learning_rate": 3.948627607007906e-05,
      "loss": 0.9197768211364746,
      "mean_token_accuracy": 0.7968756482005119,
      "num_tokens": 34927968.0,
      "step": 17580
    },
    {
      "entropy": 0.761645146459341,
      "epoch": 0.3159832936632685,
      "grad_norm": 5.125,
      "learning_rate": 3.9474564725795324e-05,
      "loss": 0.7537056922912597,
      "mean_token_accuracy": 0.8109301716089249,
      "num_tokens": 34948685.0,
      "step": 17590
    },
    {
      "entropy": 0.7647321999073029,
      "epoch": 0.3161629316926393,
      "grad_norm": 6.71875,
      "learning_rate": 3.946284860139078e-05,
      "loss": 0.7590048789978028,
      "mean_token_accuracy": 0.8145085543394088,
      "num_tokens": 34969244.0,
      "step": 17600
    },
    {
      "entropy": 0.7169677264988422,
      "epoch": 0.31634256972201014,
      "grad_norm": 5.59375,
      "learning_rate": 3.945112770073457e-05,
      "loss": 0.7077847480773926,
      "mean_token_accuracy": 0.825789125263691,
      "num_tokens": 34988936.0,
      "step": 17610
    },
    {
      "entropy": 0.7381929822266102,
      "epoch": 0.31652220775138096,
      "grad_norm": 7.5625,
      "learning_rate": 3.9439402027697435e-05,
      "loss": 0.7341346263885498,
      "mean_token_accuracy": 0.8267239779233932,
      "num_tokens": 35008085.0,
      "step": 17620
    },
    {
      "entropy": 0.7316080778837204,
      "epoch": 0.3167018457807518,
      "grad_norm": 5.59375,
      "learning_rate": 3.94276715861517e-05,
      "loss": 0.7617347717285157,
      "mean_token_accuracy": 0.8139830753207207,
      "num_tokens": 35028054.0,
      "step": 17630
    },
    {
      "entropy": 0.7260509677231312,
      "epoch": 0.3168814838101226,
      "grad_norm": 8.6875,
      "learning_rate": 3.9415936379971264e-05,
      "loss": 0.7644793033599854,
      "mean_token_accuracy": 0.8167120516300201,
      "num_tokens": 35046774.0,
      "step": 17640
    },
    {
      "entropy": 0.790661808475852,
      "epoch": 0.3170611218394934,
      "grad_norm": 6.15625,
      "learning_rate": 3.940419641303159e-05,
      "loss": 0.8470956802368164,
      "mean_token_accuracy": 0.804165604710579,
      "num_tokens": 35067003.0,
      "step": 17650
    },
    {
      "entropy": 0.7736934401094914,
      "epoch": 0.31724075986886424,
      "grad_norm": 5.8125,
      "learning_rate": 3.939245168920971e-05,
      "loss": 0.8395459175109863,
      "mean_token_accuracy": 0.8051696851849556,
      "num_tokens": 35087491.0,
      "step": 17660
    },
    {
      "entropy": 0.7307654902338981,
      "epoch": 0.31742039789823506,
      "grad_norm": 5.0625,
      "learning_rate": 3.9380702212384234e-05,
      "loss": 0.7179835319519043,
      "mean_token_accuracy": 0.8296659529209137,
      "num_tokens": 35108480.0,
      "step": 17670
    },
    {
      "entropy": 0.7836914859712124,
      "epoch": 0.3176000359276059,
      "grad_norm": 6.09375,
      "learning_rate": 3.936894798643536e-05,
      "loss": 0.7921192646026611,
      "mean_token_accuracy": 0.8002183720469475,
      "num_tokens": 35129802.0,
      "step": 17680
    },
    {
      "entropy": 0.821904543042183,
      "epoch": 0.3177796739569767,
      "grad_norm": 6.3125,
      "learning_rate": 3.935718901524481e-05,
      "loss": 0.8243512153625489,
      "mean_token_accuracy": 0.8043264150619507,
      "num_tokens": 35149769.0,
      "step": 17690
    },
    {
      "entropy": 0.6292390547692776,
      "epoch": 0.3179593119863475,
      "grad_norm": 6.75,
      "learning_rate": 3.934542530269593e-05,
      "loss": 0.6188334941864013,
      "mean_token_accuracy": 0.8497886136174202,
      "num_tokens": 35168529.0,
      "step": 17700
    },
    {
      "entropy": 0.7929140143096447,
      "epoch": 0.31813895001571835,
      "grad_norm": 7.0625,
      "learning_rate": 3.933365685267358e-05,
      "loss": 0.7906225681304931,
      "mean_token_accuracy": 0.8013290375471115,
      "num_tokens": 35187658.0,
      "step": 17710
    },
    {
      "entropy": 0.8120412543416023,
      "epoch": 0.31831858804508917,
      "grad_norm": 6.53125,
      "learning_rate": 3.932188366906422e-05,
      "loss": 0.7878899574279785,
      "mean_token_accuracy": 0.8007497161626815,
      "num_tokens": 35206249.0,
      "step": 17720
    },
    {
      "entropy": 0.690728259831667,
      "epoch": 0.31849822607446,
      "grad_norm": 8.25,
      "learning_rate": 3.931010575575584e-05,
      "loss": 0.7184494495391845,
      "mean_token_accuracy": 0.8269483983516693,
      "num_tokens": 35224970.0,
      "step": 17730
    },
    {
      "entropy": 0.740854436904192,
      "epoch": 0.31867786410383075,
      "grad_norm": 6.625,
      "learning_rate": 3.929832311663803e-05,
      "loss": 0.7246941089630127,
      "mean_token_accuracy": 0.8198397517204284,
      "num_tokens": 35243346.0,
      "step": 17740
    },
    {
      "entropy": 0.6676546044647693,
      "epoch": 0.3188575021332016,
      "grad_norm": 5.84375,
      "learning_rate": 3.928653575560191e-05,
      "loss": 0.7035351276397706,
      "mean_token_accuracy": 0.8237126424908638,
      "num_tokens": 35263450.0,
      "step": 17750
    },
    {
      "entropy": 0.7323327511548996,
      "epoch": 0.3190371401625724,
      "grad_norm": 5.15625,
      "learning_rate": 3.927474367654017e-05,
      "loss": 0.7486181735992432,
      "mean_token_accuracy": 0.8162674978375435,
      "num_tokens": 35283680.0,
      "step": 17760
    },
    {
      "entropy": 0.7591222655028105,
      "epoch": 0.3192167781919432,
      "grad_norm": 7.1875,
      "learning_rate": 3.926294688334707e-05,
      "loss": 0.7804352283477783,
      "mean_token_accuracy": 0.8173931494355202,
      "num_tokens": 35303303.0,
      "step": 17770
    },
    {
      "entropy": 0.8280812725424767,
      "epoch": 0.31939641622131404,
      "grad_norm": 6.03125,
      "learning_rate": 3.925114537991841e-05,
      "loss": 0.8541572570800782,
      "mean_token_accuracy": 0.8000821828842163,
      "num_tokens": 35324250.0,
      "step": 17780
    },
    {
      "entropy": 0.716727501899004,
      "epoch": 0.31957605425068486,
      "grad_norm": 7.8125,
      "learning_rate": 3.9239339170151554e-05,
      "loss": 0.6860591411590576,
      "mean_token_accuracy": 0.8338948950171471,
      "num_tokens": 35344382.0,
      "step": 17790
    },
    {
      "entropy": 0.7611725151538848,
      "epoch": 0.3197556922800557,
      "grad_norm": 7.125,
      "learning_rate": 3.922752825794542e-05,
      "loss": 0.8029959678649903,
      "mean_token_accuracy": 0.8064067751169205,
      "num_tokens": 35362893.0,
      "step": 17800
    },
    {
      "entropy": 0.6786594569683075,
      "epoch": 0.3199353303094265,
      "grad_norm": 6.875,
      "learning_rate": 3.9215712647200464e-05,
      "loss": 0.6978197574615479,
      "mean_token_accuracy": 0.8251541584730149,
      "num_tokens": 35382799.0,
      "step": 17810
    },
    {
      "entropy": 0.6543627932667733,
      "epoch": 0.3201149683387973,
      "grad_norm": 5.46875,
      "learning_rate": 3.920389234181871e-05,
      "loss": 0.6348058700561523,
      "mean_token_accuracy": 0.8412888690829277,
      "num_tokens": 35402553.0,
      "step": 17820
    },
    {
      "entropy": 0.6935957431793213,
      "epoch": 0.32029460636816814,
      "grad_norm": 5.125,
      "learning_rate": 3.9192067345703755e-05,
      "loss": 0.6866061210632324,
      "mean_token_accuracy": 0.829162959754467,
      "num_tokens": 35422225.0,
      "step": 17830
    },
    {
      "entropy": 0.693574846163392,
      "epoch": 0.32047424439753897,
      "grad_norm": 6.5625,
      "learning_rate": 3.918023766276068e-05,
      "loss": 0.7246339797973633,
      "mean_token_accuracy": 0.8193977564573288,
      "num_tokens": 35441901.0,
      "step": 17840
    },
    {
      "entropy": 0.820479690283537,
      "epoch": 0.3206538824269098,
      "grad_norm": 6.09375,
      "learning_rate": 3.916840329689618e-05,
      "loss": 0.8252622604370117,
      "mean_token_accuracy": 0.7990131780505181,
      "num_tokens": 35461154.0,
      "step": 17850
    },
    {
      "entropy": 0.6803769111633301,
      "epoch": 0.3208335204562806,
      "grad_norm": 5.53125,
      "learning_rate": 3.915656425201845e-05,
      "loss": 0.7319106101989746,
      "mean_token_accuracy": 0.8188167631626129,
      "num_tokens": 35481281.0,
      "step": 17860
    },
    {
      "entropy": 0.8349010899662972,
      "epoch": 0.32101315848565143,
      "grad_norm": 5.71875,
      "learning_rate": 3.914472053203726e-05,
      "loss": 0.9022153854370117,
      "mean_token_accuracy": 0.789471959322691,
      "num_tokens": 35501401.0,
      "step": 17870
    },
    {
      "entropy": 0.7311196655035019,
      "epoch": 0.32119279651502225,
      "grad_norm": 5.75,
      "learning_rate": 3.913287214086392e-05,
      "loss": 0.7151325225830079,
      "mean_token_accuracy": 0.8244502767920494,
      "num_tokens": 35522449.0,
      "step": 17880
    },
    {
      "entropy": 0.87136470079422,
      "epoch": 0.32137243454439307,
      "grad_norm": 5.9375,
      "learning_rate": 3.912101908241126e-05,
      "loss": 0.867275333404541,
      "mean_token_accuracy": 0.7914594322443008,
      "num_tokens": 35542643.0,
      "step": 17890
    },
    {
      "entropy": 0.730673173815012,
      "epoch": 0.3215520725737639,
      "grad_norm": 4.71875,
      "learning_rate": 3.9109161360593684e-05,
      "loss": 0.7197350025177002,
      "mean_token_accuracy": 0.8230228215456009,
      "num_tokens": 35562064.0,
      "step": 17900
    },
    {
      "entropy": 0.7367616221308708,
      "epoch": 0.32173171060313466,
      "grad_norm": 6.59375,
      "learning_rate": 3.9097298979327104e-05,
      "loss": 0.765618896484375,
      "mean_token_accuracy": 0.8102297723293305,
      "num_tokens": 35581401.0,
      "step": 17910
    },
    {
      "entropy": 0.7334169682115317,
      "epoch": 0.3219113486325055,
      "grad_norm": 5.0,
      "learning_rate": 3.9085431942528995e-05,
      "loss": 0.6692475795745849,
      "mean_token_accuracy": 0.8380904793739319,
      "num_tokens": 35601578.0,
      "step": 17920
    },
    {
      "entropy": 0.711165975034237,
      "epoch": 0.3220909866618763,
      "grad_norm": 5.03125,
      "learning_rate": 3.9073560254118357e-05,
      "loss": 0.7349391460418702,
      "mean_token_accuracy": 0.8185041308403015,
      "num_tokens": 35621624.0,
      "step": 17930
    },
    {
      "entropy": 0.6822538442909718,
      "epoch": 0.3222706246912471,
      "grad_norm": 6.125,
      "learning_rate": 3.906168391801574e-05,
      "loss": 0.7448233604431153,
      "mean_token_accuracy": 0.8198145613074302,
      "num_tokens": 35640193.0,
      "step": 17940
    },
    {
      "entropy": 0.7884551115334034,
      "epoch": 0.32245026272061794,
      "grad_norm": 5.71875,
      "learning_rate": 3.904980293814319e-05,
      "loss": 0.8254501342773437,
      "mean_token_accuracy": 0.8034142166376114,
      "num_tokens": 35659808.0,
      "step": 17950
    },
    {
      "entropy": 0.8055154643952847,
      "epoch": 0.32262990074998876,
      "grad_norm": 4.71875,
      "learning_rate": 3.9037917318424345e-05,
      "loss": 0.8211849212646485,
      "mean_token_accuracy": 0.8013218566775322,
      "num_tokens": 35680537.0,
      "step": 17960
    },
    {
      "entropy": 0.769489198178053,
      "epoch": 0.3228095387793596,
      "grad_norm": 6.15625,
      "learning_rate": 3.902602706278434e-05,
      "loss": 0.737896203994751,
      "mean_token_accuracy": 0.8228942841291428,
      "num_tokens": 35700398.0,
      "step": 17970
    },
    {
      "entropy": 0.6962581276893616,
      "epoch": 0.3229891768087304,
      "grad_norm": 5.25,
      "learning_rate": 3.901413217514983e-05,
      "loss": 0.6619221687316894,
      "mean_token_accuracy": 0.8340642988681793,
      "num_tokens": 35720628.0,
      "step": 17980
    },
    {
      "entropy": 0.6967524424195289,
      "epoch": 0.3231688148381012,
      "grad_norm": 6.65625,
      "learning_rate": 3.9002232659449036e-05,
      "loss": 0.6605311393737793,
      "mean_token_accuracy": 0.8388112917542457,
      "num_tokens": 35740644.0,
      "step": 17990
    },
    {
      "epoch": 0.32334845286747205,
      "eval_entropy": 0.6859483782052994,
      "eval_loss": 0.7286707758903503,
      "eval_mean_token_accuracy": 0.8220490045547485,
      "eval_num_tokens": 35760000.0,
      "eval_runtime": 40.5099,
      "eval_samples_per_second": 49.371,
      "eval_steps_per_second": 6.171,
      "step": 18000
    },
    {
      "entropy": 0.7104107137769461,
      "epoch": 0.32352809089684287,
      "grad_norm": 7.21875,
      "learning_rate": 3.897841975956901e-05,
      "loss": 0.7437191963195801,
      "mean_token_accuracy": 0.8219703547656536,
      "num_tokens": 35779311.0,
      "step": 18010
    },
    {
      "entropy": 0.6965253315865993,
      "epoch": 0.3237077289262137,
      "grad_norm": 5.15625,
      "learning_rate": 3.896650638325382e-05,
      "loss": 0.738174295425415,
      "mean_token_accuracy": 0.8219190165400505,
      "num_tokens": 35798459.0,
      "step": 18020
    },
    {
      "entropy": 0.7180504962801934,
      "epoch": 0.3238873669555845,
      "grad_norm": 8.1875,
      "learning_rate": 3.895458839460041e-05,
      "loss": 0.7187300682067871,
      "mean_token_accuracy": 0.823338258266449,
      "num_tokens": 35818198.0,
      "step": 18030
    },
    {
      "entropy": 0.7858592428267002,
      "epoch": 0.32406700498495533,
      "grad_norm": 6.0625,
      "learning_rate": 3.8942665797544614e-05,
      "loss": 0.7939746379852295,
      "mean_token_accuracy": 0.8107775658369064,
      "num_tokens": 35838842.0,
      "step": 18040
    },
    {
      "entropy": 0.7111394345760346,
      "epoch": 0.32424664301432615,
      "grad_norm": 7.125,
      "learning_rate": 3.893073859602378e-05,
      "loss": 0.7317216873168946,
      "mean_token_accuracy": 0.8147927239537239,
      "num_tokens": 35859226.0,
      "step": 18050
    },
    {
      "entropy": 0.7522515088319779,
      "epoch": 0.324426281043697,
      "grad_norm": 5.375,
      "learning_rate": 3.891880679397679e-05,
      "loss": 0.7336560249328613,
      "mean_token_accuracy": 0.815828162431717,
      "num_tokens": 35878908.0,
      "step": 18060
    },
    {
      "entropy": 0.6881562776863575,
      "epoch": 0.3246059190730678,
      "grad_norm": 5.375,
      "learning_rate": 3.890687039534402e-05,
      "loss": 0.7237205505371094,
      "mean_token_accuracy": 0.8199845224618911,
      "num_tokens": 35898344.0,
      "step": 18070
    },
    {
      "entropy": 0.7639813840389251,
      "epoch": 0.32478555710243856,
      "grad_norm": 4.71875,
      "learning_rate": 3.889492940406739e-05,
      "loss": 0.7334863185882569,
      "mean_token_accuracy": 0.8067487761378288,
      "num_tokens": 35917475.0,
      "step": 18080
    },
    {
      "entropy": 0.7383829429745674,
      "epoch": 0.3249651951318094,
      "grad_norm": 6.8125,
      "learning_rate": 3.888298382409032e-05,
      "loss": 0.7514029026031495,
      "mean_token_accuracy": 0.8088784500956535,
      "num_tokens": 35937271.0,
      "step": 18090
    },
    {
      "entropy": 0.6948116440325975,
      "epoch": 0.3251448331611802,
      "grad_norm": 6.53125,
      "learning_rate": 3.887103365935776e-05,
      "loss": 0.7240853786468506,
      "mean_token_accuracy": 0.8247540324926377,
      "num_tokens": 35956289.0,
      "step": 18100
    },
    {
      "entropy": 0.7187845341861248,
      "epoch": 0.325324471190551,
      "grad_norm": 9.125,
      "learning_rate": 3.8859078913816164e-05,
      "loss": 0.7233653545379639,
      "mean_token_accuracy": 0.8141455739736557,
      "num_tokens": 35975111.0,
      "step": 18110
    },
    {
      "entropy": 0.7458083823323249,
      "epoch": 0.32550410921992184,
      "grad_norm": 7.8125,
      "learning_rate": 3.8847119591413496e-05,
      "loss": 0.7706594467163086,
      "mean_token_accuracy": 0.8207385942339898,
      "num_tokens": 35994414.0,
      "step": 18120
    },
    {
      "entropy": 0.8206540107727051,
      "epoch": 0.32568374724929267,
      "grad_norm": 8.1875,
      "learning_rate": 3.8835155696099246e-05,
      "loss": 0.8289780616760254,
      "mean_token_accuracy": 0.798843277990818,
      "num_tokens": 36015479.0,
      "step": 18130
    },
    {
      "entropy": 0.7936149846762419,
      "epoch": 0.3258633852786635,
      "grad_norm": 6.34375,
      "learning_rate": 3.882318723182439e-05,
      "loss": 0.7851647853851318,
      "mean_token_accuracy": 0.8097789123654365,
      "num_tokens": 36035585.0,
      "step": 18140
    },
    {
      "entropy": 0.7167164027690888,
      "epoch": 0.3260430233080343,
      "grad_norm": 6.53125,
      "learning_rate": 3.881121420254144e-05,
      "loss": 0.7050783634185791,
      "mean_token_accuracy": 0.8215075165033341,
      "num_tokens": 36055767.0,
      "step": 18150
    },
    {
      "entropy": 0.6607315503060818,
      "epoch": 0.32622266133740513,
      "grad_norm": 4.625,
      "learning_rate": 3.879923661220439e-05,
      "loss": 0.6287544250488282,
      "mean_token_accuracy": 0.8367185801267624,
      "num_tokens": 36074978.0,
      "step": 18160
    },
    {
      "entropy": 0.6782990470528603,
      "epoch": 0.32640229936677595,
      "grad_norm": 5.8125,
      "learning_rate": 3.878725446476876e-05,
      "loss": 0.6678516864776611,
      "mean_token_accuracy": 0.8337928794324398,
      "num_tokens": 36094330.0,
      "step": 18170
    },
    {
      "entropy": 0.6288402430713177,
      "epoch": 0.32658193739614677,
      "grad_norm": 5.0,
      "learning_rate": 3.877526776419158e-05,
      "loss": 0.6523642539978027,
      "mean_token_accuracy": 0.8342447802424431,
      "num_tokens": 36113661.0,
      "step": 18180
    },
    {
      "entropy": 0.7460531532764435,
      "epoch": 0.3267615754255176,
      "grad_norm": 7.21875,
      "learning_rate": 3.876327651443135e-05,
      "loss": 0.7760750770568847,
      "mean_token_accuracy": 0.8118333652615547,
      "num_tokens": 36133901.0,
      "step": 18190
    },
    {
      "entropy": 0.7059107288718224,
      "epoch": 0.3269412134548884,
      "grad_norm": 8.1875,
      "learning_rate": 3.87512807194481e-05,
      "loss": 0.7308043003082275,
      "mean_token_accuracy": 0.8239887073636055,
      "num_tokens": 36154322.0,
      "step": 18200
    },
    {
      "entropy": 0.6194564238190651,
      "epoch": 0.32712085148425923,
      "grad_norm": 5.75,
      "learning_rate": 3.8739280383203366e-05,
      "loss": 0.7034087657928467,
      "mean_token_accuracy": 0.8321375474333763,
      "num_tokens": 36174064.0,
      "step": 18210
    },
    {
      "entropy": 0.7423336267471313,
      "epoch": 0.32730048951363006,
      "grad_norm": 7.0625,
      "learning_rate": 3.872727550966017e-05,
      "loss": 0.7456801891326904,
      "mean_token_accuracy": 0.8181778118014336,
      "num_tokens": 36193095.0,
      "step": 18220
    },
    {
      "entropy": 0.7759594857692719,
      "epoch": 0.3274801275430009,
      "grad_norm": 7.125,
      "learning_rate": 3.871526610278302e-05,
      "loss": 0.774388599395752,
      "mean_token_accuracy": 0.8100086480379105,
      "num_tokens": 36213463.0,
      "step": 18230
    },
    {
      "entropy": 0.7618721894919872,
      "epoch": 0.3276597655723717,
      "grad_norm": 7.3125,
      "learning_rate": 3.8703252166537954e-05,
      "loss": 0.7587786197662354,
      "mean_token_accuracy": 0.8235827222466469,
      "num_tokens": 36233781.0,
      "step": 18240
    },
    {
      "entropy": 0.7541228655725718,
      "epoch": 0.32783940360174246,
      "grad_norm": 5.96875,
      "learning_rate": 3.869123370489248e-05,
      "loss": 0.8078394889831543,
      "mean_token_accuracy": 0.8117158576846123,
      "num_tokens": 36254098.0,
      "step": 18250
    },
    {
      "entropy": 0.7939810432493687,
      "epoch": 0.3280190416311133,
      "grad_norm": 6.03125,
      "learning_rate": 3.867921072181561e-05,
      "loss": 0.7747293949127197,
      "mean_token_accuracy": 0.8127890095114708,
      "num_tokens": 36273992.0,
      "step": 18260
    },
    {
      "entropy": 0.7231539063155651,
      "epoch": 0.3281986796604841,
      "grad_norm": 5.625,
      "learning_rate": 3.866718322127784e-05,
      "loss": 0.7358824729919433,
      "mean_token_accuracy": 0.8163224875926971,
      "num_tokens": 36293821.0,
      "step": 18270
    },
    {
      "entropy": 0.6936972655355931,
      "epoch": 0.3283783176898549,
      "grad_norm": 5.46875,
      "learning_rate": 3.865515120725117e-05,
      "loss": 0.6835323810577393,
      "mean_token_accuracy": 0.833130057156086,
      "num_tokens": 36313991.0,
      "step": 18280
    },
    {
      "entropy": 0.7611787430942059,
      "epoch": 0.32855795571922575,
      "grad_norm": 6.09375,
      "learning_rate": 3.8643114683709106e-05,
      "loss": 0.781134557723999,
      "mean_token_accuracy": 0.8132867559790611,
      "num_tokens": 36333701.0,
      "step": 18290
    },
    {
      "entropy": 0.7220565572381019,
      "epoch": 0.32873759374859657,
      "grad_norm": 4.96875,
      "learning_rate": 3.863107365462659e-05,
      "loss": 0.7168581008911132,
      "mean_token_accuracy": 0.8252742826938629,
      "num_tokens": 36353354.0,
      "step": 18300
    },
    {
      "entropy": 0.7133948326110839,
      "epoch": 0.3289172317779674,
      "grad_norm": 6.4375,
      "learning_rate": 3.86190281239801e-05,
      "loss": 0.7207342147827148,
      "mean_token_accuracy": 0.8175678163766861,
      "num_tokens": 36373590.0,
      "step": 18310
    },
    {
      "entropy": 0.7170808725059032,
      "epoch": 0.3290968698073382,
      "grad_norm": 7.125,
      "learning_rate": 3.860697809574759e-05,
      "loss": 0.7516814231872558,
      "mean_token_accuracy": 0.8184340953826904,
      "num_tokens": 36392673.0,
      "step": 18320
    },
    {
      "entropy": 0.6643745338544249,
      "epoch": 0.32927650783670903,
      "grad_norm": 5.96875,
      "learning_rate": 3.8594923573908496e-05,
      "loss": 0.6682701110839844,
      "mean_token_accuracy": 0.8350428402423858,
      "num_tokens": 36411923.0,
      "step": 18330
    },
    {
      "entropy": 0.7081220515072346,
      "epoch": 0.32945614586607985,
      "grad_norm": 4.21875,
      "learning_rate": 3.8582864562443735e-05,
      "loss": 0.726106071472168,
      "mean_token_accuracy": 0.8204600110650062,
      "num_tokens": 36432530.0,
      "step": 18340
    },
    {
      "entropy": 0.7445301353931427,
      "epoch": 0.3296357838954507,
      "grad_norm": 9.625,
      "learning_rate": 3.857080106533569e-05,
      "loss": 0.823037052154541,
      "mean_token_accuracy": 0.8115371927618981,
      "num_tokens": 36452150.0,
      "step": 18350
    },
    {
      "entropy": 0.8014833815395832,
      "epoch": 0.3298154219248215,
      "grad_norm": 7.1875,
      "learning_rate": 3.8558733086568276e-05,
      "loss": 0.8193591117858887,
      "mean_token_accuracy": 0.8090257301926613,
      "num_tokens": 36471767.0,
      "step": 18360
    },
    {
      "entropy": 0.8053083054721355,
      "epoch": 0.3299950599541923,
      "grad_norm": 6.46875,
      "learning_rate": 3.854666063012684e-05,
      "loss": 0.8458587646484375,
      "mean_token_accuracy": 0.8012135088443756,
      "num_tokens": 36492233.0,
      "step": 18370
    },
    {
      "entropy": 0.824552272260189,
      "epoch": 0.33017469798356314,
      "grad_norm": 7.03125,
      "learning_rate": 3.853458369999822e-05,
      "loss": 0.7995263576507569,
      "mean_token_accuracy": 0.8024511843919754,
      "num_tokens": 36511468.0,
      "step": 18380
    },
    {
      "entropy": 0.7469260133802891,
      "epoch": 0.33035433601293396,
      "grad_norm": 5.90625,
      "learning_rate": 3.8522502300170744e-05,
      "loss": 0.722861909866333,
      "mean_token_accuracy": 0.8246075689792634,
      "num_tokens": 36530853.0,
      "step": 18390
    },
    {
      "entropy": 0.7714269630610943,
      "epoch": 0.3305339740423048,
      "grad_norm": 5.15625,
      "learning_rate": 3.85104164346342e-05,
      "loss": 0.7363530158996582,
      "mean_token_accuracy": 0.8184788078069687,
      "num_tokens": 36551393.0,
      "step": 18400
    },
    {
      "entropy": 0.7279658913612366,
      "epoch": 0.3307136120716756,
      "grad_norm": 8.0625,
      "learning_rate": 3.849832610737986e-05,
      "loss": 0.7606022357940674,
      "mean_token_accuracy": 0.8112672522664071,
      "num_tokens": 36570769.0,
      "step": 18410
    },
    {
      "entropy": 0.7019796539098024,
      "epoch": 0.33089325010104637,
      "grad_norm": 7.34375,
      "learning_rate": 3.848623132240047e-05,
      "loss": 0.7645049571990967,
      "mean_token_accuracy": 0.8167028889060021,
      "num_tokens": 36590872.0,
      "step": 18420
    },
    {
      "entropy": 0.7192273221909999,
      "epoch": 0.3310728881304172,
      "grad_norm": 4.625,
      "learning_rate": 3.847413208369025e-05,
      "loss": 0.7425713539123535,
      "mean_token_accuracy": 0.8125694468617439,
      "num_tokens": 36610178.0,
      "step": 18430
    },
    {
      "entropy": 0.7617163226008415,
      "epoch": 0.331252526159788,
      "grad_norm": 7.40625,
      "learning_rate": 3.8462028395244874e-05,
      "loss": 0.7606909275054932,
      "mean_token_accuracy": 0.8147780910134316,
      "num_tokens": 36630216.0,
      "step": 18440
    },
    {
      "entropy": 0.7540189482271671,
      "epoch": 0.33143216418915883,
      "grad_norm": 4.4375,
      "learning_rate": 3.844992026106151e-05,
      "loss": 0.7246033191680908,
      "mean_token_accuracy": 0.828866283595562,
      "num_tokens": 36650214.0,
      "step": 18450
    },
    {
      "entropy": 0.7657978005707264,
      "epoch": 0.33161180221852965,
      "grad_norm": 7.84375,
      "learning_rate": 3.843780768513877e-05,
      "loss": 0.7647784233093262,
      "mean_token_accuracy": 0.817056518048048,
      "num_tokens": 36670995.0,
      "step": 18460
    },
    {
      "entropy": 0.726831492781639,
      "epoch": 0.33179144024790047,
      "grad_norm": 5.34375,
      "learning_rate": 3.8425690671476756e-05,
      "loss": 0.7299631595611572,
      "mean_token_accuracy": 0.8184500336647034,
      "num_tokens": 36690742.0,
      "step": 18470
    },
    {
      "entropy": 0.7030957140028476,
      "epoch": 0.3319710782772713,
      "grad_norm": 7.0,
      "learning_rate": 3.8413569224077026e-05,
      "loss": 0.7092371940612793,
      "mean_token_accuracy": 0.8225024163722991,
      "num_tokens": 36709945.0,
      "step": 18480
    },
    {
      "entropy": 0.698981449007988,
      "epoch": 0.3321507163066421,
      "grad_norm": 6.1875,
      "learning_rate": 3.840144334694259e-05,
      "loss": 0.7193443775177002,
      "mean_token_accuracy": 0.8260339245200157,
      "num_tokens": 36730552.0,
      "step": 18490
    },
    {
      "entropy": 0.7802863925695419,
      "epoch": 0.33233035433601293,
      "grad_norm": 7.125,
      "learning_rate": 3.838931304407793e-05,
      "loss": 0.7969265937805176,
      "mean_token_accuracy": 0.8059286683797836,
      "num_tokens": 36751378.0,
      "step": 18500
    },
    {
      "entropy": 0.681403387337923,
      "epoch": 0.33250999236538376,
      "grad_norm": 5.90625,
      "learning_rate": 3.837717831948901e-05,
      "loss": 0.6884686470031738,
      "mean_token_accuracy": 0.8318066015839577,
      "num_tokens": 36770624.0,
      "step": 18510
    },
    {
      "entropy": 0.6782195389270782,
      "epoch": 0.3326896303947546,
      "grad_norm": 7.46875,
      "learning_rate": 3.836503917718323e-05,
      "loss": 0.7010732650756836,
      "mean_token_accuracy": 0.8273067072033882,
      "num_tokens": 36790879.0,
      "step": 18520
    },
    {
      "entropy": 0.7966554440557957,
      "epoch": 0.3328692684241254,
      "grad_norm": 5.9375,
      "learning_rate": 3.835289562116944e-05,
      "loss": 0.7911303997039795,
      "mean_token_accuracy": 0.8021986454725265,
      "num_tokens": 36811172.0,
      "step": 18530
    },
    {
      "entropy": 0.7638285987079143,
      "epoch": 0.3330489064534962,
      "grad_norm": 4.9375,
      "learning_rate": 3.8340747655457966e-05,
      "loss": 0.7817434310913086,
      "mean_token_accuracy": 0.8101646333932877,
      "num_tokens": 36831743.0,
      "step": 18540
    },
    {
      "entropy": 0.7218019172549248,
      "epoch": 0.33322854448286704,
      "grad_norm": 5.03125,
      "learning_rate": 3.8328595284060594e-05,
      "loss": 0.7030789852142334,
      "mean_token_accuracy": 0.8340276136994362,
      "num_tokens": 36850942.0,
      "step": 18550
    },
    {
      "entropy": 0.8480454966425895,
      "epoch": 0.33340818251223786,
      "grad_norm": 6.53125,
      "learning_rate": 3.831643851099056e-05,
      "loss": 0.9133129119873047,
      "mean_token_accuracy": 0.7898767620325089,
      "num_tokens": 36871385.0,
      "step": 18560
    },
    {
      "entropy": 0.7373027011752129,
      "epoch": 0.3335878205416087,
      "grad_norm": 6.90625,
      "learning_rate": 3.8304277340262536e-05,
      "loss": 0.7270811080932618,
      "mean_token_accuracy": 0.8201663658022881,
      "num_tokens": 36891429.0,
      "step": 18570
    },
    {
      "entropy": 0.6780092939734459,
      "epoch": 0.3337674585709795,
      "grad_norm": 5.03125,
      "learning_rate": 3.829211177589267e-05,
      "loss": 0.6929347515106201,
      "mean_token_accuracy": 0.8263748154044152,
      "num_tokens": 36911038.0,
      "step": 18580
    },
    {
      "entropy": 0.7557262569665909,
      "epoch": 0.33394709660035027,
      "grad_norm": 5.875,
      "learning_rate": 3.827994182189855e-05,
      "loss": 0.772157859802246,
      "mean_token_accuracy": 0.8099320515990257,
      "num_tokens": 36930958.0,
      "step": 18590
    },
    {
      "entropy": 0.6927871316671371,
      "epoch": 0.3341267346297211,
      "grad_norm": 5.5,
      "learning_rate": 3.826776748229922e-05,
      "loss": 0.6662344455718994,
      "mean_token_accuracy": 0.835584020614624,
      "num_tokens": 36951474.0,
      "step": 18600
    },
    {
      "entropy": 0.6624749101698398,
      "epoch": 0.3343063726590919,
      "grad_norm": 8.1875,
      "learning_rate": 3.825558876111518e-05,
      "loss": 0.6888839721679687,
      "mean_token_accuracy": 0.8293868422508239,
      "num_tokens": 36972228.0,
      "step": 18610
    },
    {
      "entropy": 0.7181465312838554,
      "epoch": 0.33448601068846273,
      "grad_norm": 4.9375,
      "learning_rate": 3.824340566236833e-05,
      "loss": 0.7347521305084228,
      "mean_token_accuracy": 0.8214095681905746,
      "num_tokens": 36992018.0,
      "step": 18620
    },
    {
      "entropy": 0.7762895479798317,
      "epoch": 0.33466564871783355,
      "grad_norm": 7.09375,
      "learning_rate": 3.823121819008208e-05,
      "loss": 0.7590028285980225,
      "mean_token_accuracy": 0.8116882398724556,
      "num_tokens": 37012143.0,
      "step": 18630
    },
    {
      "entropy": 0.7358211122453213,
      "epoch": 0.3348452867472044,
      "grad_norm": 6.5,
      "learning_rate": 3.821902634828124e-05,
      "loss": 0.7745433330535889,
      "mean_token_accuracy": 0.8135696589946747,
      "num_tokens": 37031625.0,
      "step": 18640
    },
    {
      "entropy": 0.7558896265923977,
      "epoch": 0.3350249247765752,
      "grad_norm": 5.03125,
      "learning_rate": 3.8206830140992094e-05,
      "loss": 0.7886178016662597,
      "mean_token_accuracy": 0.810863482952118,
      "num_tokens": 37054140.0,
      "step": 18650
    },
    {
      "entropy": 0.6869462393224239,
      "epoch": 0.335204562805946,
      "grad_norm": 5.375,
      "learning_rate": 3.8194629572242335e-05,
      "loss": 0.7197296142578125,
      "mean_token_accuracy": 0.829002232849598,
      "num_tokens": 37073511.0,
      "step": 18660
    },
    {
      "entropy": 0.7334383629262448,
      "epoch": 0.33538420083531684,
      "grad_norm": 6.59375,
      "learning_rate": 3.818242464606112e-05,
      "loss": 0.759697437286377,
      "mean_token_accuracy": 0.8188296809792519,
      "num_tokens": 37093285.0,
      "step": 18670
    },
    {
      "entropy": 0.7271647840738297,
      "epoch": 0.33556383886468766,
      "grad_norm": 5.9375,
      "learning_rate": 3.817021536647905e-05,
      "loss": 0.7197713375091552,
      "mean_token_accuracy": 0.8217777356505394,
      "num_tokens": 37112477.0,
      "step": 18680
    },
    {
      "entropy": 0.7619598247110844,
      "epoch": 0.3357434768940585,
      "grad_norm": 6.4375,
      "learning_rate": 3.815800173752813e-05,
      "loss": 0.7630502700805664,
      "mean_token_accuracy": 0.8127833500504493,
      "num_tokens": 37132284.0,
      "step": 18690
    },
    {
      "entropy": 0.7578432653099298,
      "epoch": 0.3359231149234293,
      "grad_norm": 7.34375,
      "learning_rate": 3.814578376324183e-05,
      "loss": 0.731998872756958,
      "mean_token_accuracy": 0.8185982555150986,
      "num_tokens": 37152362.0,
      "step": 18700
    },
    {
      "entropy": 0.6884110160171986,
      "epoch": 0.3361027529528001,
      "grad_norm": 6.90625,
      "learning_rate": 3.8133561447655046e-05,
      "loss": 0.6964850425720215,
      "mean_token_accuracy": 0.8265047192573547,
      "num_tokens": 37172611.0,
      "step": 18710
    },
    {
      "entropy": 0.8253501400351524,
      "epoch": 0.33628239098217094,
      "grad_norm": 8.125,
      "learning_rate": 3.812133479480413e-05,
      "loss": 0.8615789413452148,
      "mean_token_accuracy": 0.7984562411904335,
      "num_tokens": 37193416.0,
      "step": 18720
    },
    {
      "entropy": 0.7341925159096718,
      "epoch": 0.33646202901154176,
      "grad_norm": 6.25,
      "learning_rate": 3.8109103808726816e-05,
      "loss": 0.7278830051422119,
      "mean_token_accuracy": 0.8197360396385193,
      "num_tokens": 37213556.0,
      "step": 18730
    },
    {
      "entropy": 0.7634441427886486,
      "epoch": 0.3366416670409126,
      "grad_norm": 6.90625,
      "learning_rate": 3.8096868493462324e-05,
      "loss": 0.7745267391204834,
      "mean_token_accuracy": 0.8052794098854065,
      "num_tokens": 37233035.0,
      "step": 18740
    },
    {
      "entropy": 0.7776296749711037,
      "epoch": 0.3368213050702834,
      "grad_norm": 4.28125,
      "learning_rate": 3.808462885305126e-05,
      "loss": 0.7788728713989258,
      "mean_token_accuracy": 0.8059475466609001,
      "num_tokens": 37252712.0,
      "step": 18750
    },
    {
      "entropy": 0.7568254709243775,
      "epoch": 0.33700094309965417,
      "grad_norm": 6.0,
      "learning_rate": 3.8072384891535694e-05,
      "loss": 0.7935106754302979,
      "mean_token_accuracy": 0.8023427307605744,
      "num_tokens": 37272399.0,
      "step": 18760
    },
    {
      "entropy": 0.6368025138974189,
      "epoch": 0.337180581129025,
      "grad_norm": 5.03125,
      "learning_rate": 3.8060136612959096e-05,
      "loss": 0.6402888298034668,
      "mean_token_accuracy": 0.8436927735805512,
      "num_tokens": 37292500.0,
      "step": 18770
    },
    {
      "entropy": 0.7777565490454436,
      "epoch": 0.3373602191583958,
      "grad_norm": 3.875,
      "learning_rate": 3.8047884021366374e-05,
      "loss": 0.7925205230712891,
      "mean_token_accuracy": 0.808977472782135,
      "num_tokens": 37312100.0,
      "step": 18780
    },
    {
      "entropy": 0.7706734105944634,
      "epoch": 0.33753985718776663,
      "grad_norm": 6.875,
      "learning_rate": 3.803562712080386e-05,
      "loss": 0.8321983337402343,
      "mean_token_accuracy": 0.8087673798203469,
      "num_tokens": 37331489.0,
      "step": 18790
    },
    {
      "entropy": 0.7277456097304821,
      "epoch": 0.33771949521713746,
      "grad_norm": 4.9375,
      "learning_rate": 3.80233659153193e-05,
      "loss": 0.7134997367858886,
      "mean_token_accuracy": 0.8213506639003754,
      "num_tokens": 37350969.0,
      "step": 18800
    },
    {
      "entropy": 0.6118907302618026,
      "epoch": 0.3378991332465083,
      "grad_norm": 5.15625,
      "learning_rate": 3.801110040896187e-05,
      "loss": 0.6330528736114502,
      "mean_token_accuracy": 0.8412870526313782,
      "num_tokens": 37371179.0,
      "step": 18810
    },
    {
      "entropy": 0.7131917484104633,
      "epoch": 0.3380787712758791,
      "grad_norm": 4.0625,
      "learning_rate": 3.799883060578218e-05,
      "loss": 0.6921073436737061,
      "mean_token_accuracy": 0.8304278865456581,
      "num_tokens": 37390184.0,
      "step": 18820
    },
    {
      "entropy": 0.7184038236737251,
      "epoch": 0.3382584093052499,
      "grad_norm": 5.40625,
      "learning_rate": 3.798655650983223e-05,
      "loss": 0.6857271194458008,
      "mean_token_accuracy": 0.8249371990561485,
      "num_tokens": 37409401.0,
      "step": 18830
    },
    {
      "entropy": 0.7210898119956255,
      "epoch": 0.33843804733462074,
      "grad_norm": 8.9375,
      "learning_rate": 3.7974278125165455e-05,
      "loss": 0.7216170787811279,
      "mean_token_accuracy": 0.8264225035905838,
      "num_tokens": 37428922.0,
      "step": 18840
    },
    {
      "entropy": 0.7202691189944744,
      "epoch": 0.33861768536399156,
      "grad_norm": 5.90625,
      "learning_rate": 3.79619954558367e-05,
      "loss": 0.7232740879058838,
      "mean_token_accuracy": 0.8209880575537681,
      "num_tokens": 37448123.0,
      "step": 18850
    },
    {
      "entropy": 0.7499488189816474,
      "epoch": 0.3387973233933624,
      "grad_norm": 4.25,
      "learning_rate": 3.7949708505902234e-05,
      "loss": 0.7715535163879395,
      "mean_token_accuracy": 0.8093914017081261,
      "num_tokens": 37468812.0,
      "step": 18860
    },
    {
      "entropy": 0.6953777208924293,
      "epoch": 0.3389769614227332,
      "grad_norm": 6.03125,
      "learning_rate": 3.793741727941973e-05,
      "loss": 0.7301736354827881,
      "mean_token_accuracy": 0.8261926978826523,
      "num_tokens": 37487471.0,
      "step": 18870
    },
    {
      "entropy": 0.7437459662556648,
      "epoch": 0.339156599452104,
      "grad_norm": 5.71875,
      "learning_rate": 3.792512178044828e-05,
      "loss": 0.7270897388458252,
      "mean_token_accuracy": 0.8254898443818093,
      "num_tokens": 37507350.0,
      "step": 18880
    },
    {
      "entropy": 0.7612410224974155,
      "epoch": 0.33933623748147485,
      "grad_norm": 7.65625,
      "learning_rate": 3.791282201304838e-05,
      "loss": 0.7653296947479248,
      "mean_token_accuracy": 0.8047315880656243,
      "num_tokens": 37528072.0,
      "step": 18890
    },
    {
      "entropy": 0.6096017561852932,
      "epoch": 0.33951587551084567,
      "grad_norm": 5.75,
      "learning_rate": 3.7900517981281944e-05,
      "loss": 0.6406901359558106,
      "mean_token_accuracy": 0.8357881680130959,
      "num_tokens": 37547134.0,
      "step": 18900
    },
    {
      "entropy": 0.7701665848493576,
      "epoch": 0.3396955135402165,
      "grad_norm": 9.3125,
      "learning_rate": 3.78882096892123e-05,
      "loss": 0.8009776115417481,
      "mean_token_accuracy": 0.8100209400057793,
      "num_tokens": 37567547.0,
      "step": 18910
    },
    {
      "entropy": 0.7547654867172241,
      "epoch": 0.3398751515695873,
      "grad_norm": 7.875,
      "learning_rate": 3.787589714090414e-05,
      "loss": 0.7779419898986817,
      "mean_token_accuracy": 0.8086513087153435,
      "num_tokens": 37588144.0,
      "step": 18920
    },
    {
      "entropy": 0.7244872614741326,
      "epoch": 0.3400547895989581,
      "grad_norm": 5.46875,
      "learning_rate": 3.786358034042363e-05,
      "loss": 0.698306655883789,
      "mean_token_accuracy": 0.8251514658331871,
      "num_tokens": 37609079.0,
      "step": 18930
    },
    {
      "entropy": 0.761848258972168,
      "epoch": 0.3402344276283289,
      "grad_norm": 5.75,
      "learning_rate": 3.785125929183829e-05,
      "loss": 0.750943946838379,
      "mean_token_accuracy": 0.8199173092842102,
      "num_tokens": 37629767.0,
      "step": 18940
    },
    {
      "entropy": 0.6687883049249649,
      "epoch": 0.3404140656576997,
      "grad_norm": 7.96875,
      "learning_rate": 3.783893399921705e-05,
      "loss": 0.6762603759765625,
      "mean_token_accuracy": 0.8328309640288353,
      "num_tokens": 37649492.0,
      "step": 18950
    },
    {
      "entropy": 0.7742831759154797,
      "epoch": 0.34059370368707054,
      "grad_norm": 4.53125,
      "learning_rate": 3.782660446663026e-05,
      "loss": 0.8442475318908691,
      "mean_token_accuracy": 0.7992448180913925,
      "num_tokens": 37669884.0,
      "step": 18960
    },
    {
      "entropy": 0.6961066674441099,
      "epoch": 0.34077334171644136,
      "grad_norm": 6.375,
      "learning_rate": 3.781427069814966e-05,
      "loss": 0.7144382953643799,
      "mean_token_accuracy": 0.8219897776842118,
      "num_tokens": 37689319.0,
      "step": 18970
    },
    {
      "entropy": 0.7304068498313427,
      "epoch": 0.3409529797458122,
      "grad_norm": 6.15625,
      "learning_rate": 3.7801932697848384e-05,
      "loss": 0.7871774673461914,
      "mean_token_accuracy": 0.8083600834012031,
      "num_tokens": 37708412.0,
      "step": 18980
    },
    {
      "entropy": 0.6915356785058975,
      "epoch": 0.341132617775183,
      "grad_norm": 8.0,
      "learning_rate": 3.7789590469800964e-05,
      "loss": 0.6875615596771241,
      "mean_token_accuracy": 0.8225994825363159,
      "num_tokens": 37727832.0,
      "step": 18990
    },
    {
      "epoch": 0.3413122558045538,
      "eval_entropy": 0.7105240641832352,
      "eval_loss": 0.7102656364440918,
      "eval_mean_token_accuracy": 0.8266999173164368,
      "eval_num_tokens": 37748444.0,
      "eval_runtime": 40.5106,
      "eval_samples_per_second": 49.37,
      "eval_steps_per_second": 6.171,
      "step": 19000
    },
    {
      "entropy": 0.7470948297530413,
      "epoch": 0.34149189383392464,
      "grad_norm": 5.375,
      "learning_rate": 3.776489334677283e-05,
      "loss": 0.7552666187286377,
      "mean_token_accuracy": 0.8163947820663452,
      "num_tokens": 37767868.0,
      "step": 19010
    },
    {
      "entropy": 0.658749309182167,
      "epoch": 0.34167153186329546,
      "grad_norm": 4.28125,
      "learning_rate": 3.7752538459948166e-05,
      "loss": 0.6398323059082032,
      "mean_token_accuracy": 0.8333108991384506,
      "num_tokens": 37788923.0,
      "step": 19020
    },
    {
      "entropy": 0.7044443979859352,
      "epoch": 0.3418511698926663,
      "grad_norm": 5.96875,
      "learning_rate": 3.7740179361689456e-05,
      "loss": 0.6969527721405029,
      "mean_token_accuracy": 0.8253401532769203,
      "num_tokens": 37808391.0,
      "step": 19030
    },
    {
      "entropy": 0.7176072269678115,
      "epoch": 0.3420308079220371,
      "grad_norm": 9.1875,
      "learning_rate": 3.7727816056078194e-05,
      "loss": 0.738358449935913,
      "mean_token_accuracy": 0.8177365466952324,
      "num_tokens": 37828939.0,
      "step": 19040
    },
    {
      "entropy": 0.7564432226121426,
      "epoch": 0.3422104459514079,
      "grad_norm": 7.15625,
      "learning_rate": 3.771544854719728e-05,
      "loss": 0.7708708763122558,
      "mean_token_accuracy": 0.8114937514066696,
      "num_tokens": 37849027.0,
      "step": 19050
    },
    {
      "entropy": 0.7048233859241009,
      "epoch": 0.34239008398077875,
      "grad_norm": 5.625,
      "learning_rate": 3.7703076839131e-05,
      "loss": 0.7374229431152344,
      "mean_token_accuracy": 0.8205429792404175,
      "num_tokens": 37869459.0,
      "step": 19060
    },
    {
      "entropy": 0.7019929453730583,
      "epoch": 0.34256972201014957,
      "grad_norm": 4.3125,
      "learning_rate": 3.769070093596502e-05,
      "loss": 0.7498816013336181,
      "mean_token_accuracy": 0.8127054795622826,
      "num_tokens": 37889501.0,
      "step": 19070
    },
    {
      "entropy": 0.7260166116058826,
      "epoch": 0.3427493600395204,
      "grad_norm": 4.59375,
      "learning_rate": 3.767832084178639e-05,
      "loss": 0.7173238277435303,
      "mean_token_accuracy": 0.8211311534047127,
      "num_tokens": 37909116.0,
      "step": 19080
    },
    {
      "entropy": 0.7571039728820324,
      "epoch": 0.3429289980688912,
      "grad_norm": 5.1875,
      "learning_rate": 3.766593656068353e-05,
      "loss": 0.7422631740570068,
      "mean_token_accuracy": 0.820719176530838,
      "num_tokens": 37930223.0,
      "step": 19090
    },
    {
      "entropy": 0.6945526666939259,
      "epoch": 0.343108636098262,
      "grad_norm": 6.0625,
      "learning_rate": 3.7653548096746286e-05,
      "loss": 0.6704851150512695,
      "mean_token_accuracy": 0.8353661194443702,
      "num_tokens": 37949311.0,
      "step": 19100
    },
    {
      "entropy": 0.6877373330295086,
      "epoch": 0.3432882741276328,
      "grad_norm": 6.125,
      "learning_rate": 3.764115545406586e-05,
      "loss": 0.6937294006347656,
      "mean_token_accuracy": 0.8236712366342545,
      "num_tokens": 37969216.0,
      "step": 19110
    },
    {
      "entropy": 0.7838771343231201,
      "epoch": 0.3434679121570036,
      "grad_norm": 6.8125,
      "learning_rate": 3.7628758636734805e-05,
      "loss": 0.8437059402465821,
      "mean_token_accuracy": 0.8050981879234314,
      "num_tokens": 37988391.0,
      "step": 19120
    },
    {
      "entropy": 0.7158090360462666,
      "epoch": 0.34364755018637444,
      "grad_norm": 5.28125,
      "learning_rate": 3.76163576488471e-05,
      "loss": 0.7070069789886475,
      "mean_token_accuracy": 0.8265893533825874,
      "num_tokens": 38007629.0,
      "step": 19130
    },
    {
      "entropy": 0.7143278494477272,
      "epoch": 0.34382718821574526,
      "grad_norm": 6.0,
      "learning_rate": 3.760395249449809e-05,
      "loss": 0.7046169757843017,
      "mean_token_accuracy": 0.8205106377601623,
      "num_tokens": 38027354.0,
      "step": 19140
    },
    {
      "entropy": 0.7360103152692318,
      "epoch": 0.3440068262451161,
      "grad_norm": 7.3125,
      "learning_rate": 3.759154317778447e-05,
      "loss": 0.7412358283996582,
      "mean_token_accuracy": 0.821279838681221,
      "num_tokens": 38047049.0,
      "step": 19150
    },
    {
      "entropy": 0.6841146789491177,
      "epoch": 0.3441864642744869,
      "grad_norm": 6.0,
      "learning_rate": 3.7579129702804344e-05,
      "loss": 0.7308036327362061,
      "mean_token_accuracy": 0.8284485906362533,
      "num_tokens": 38066016.0,
      "step": 19160
    },
    {
      "entropy": 0.7355218008160591,
      "epoch": 0.3443661023038577,
      "grad_norm": 8.625,
      "learning_rate": 3.756671207365716e-05,
      "loss": 0.7159568309783936,
      "mean_token_accuracy": 0.824302400648594,
      "num_tokens": 38085862.0,
      "step": 19170
    },
    {
      "entropy": 0.6563275642693043,
      "epoch": 0.34454574033322855,
      "grad_norm": 6.96875,
      "learning_rate": 3.755429029444375e-05,
      "loss": 0.6665488243103027,
      "mean_token_accuracy": 0.829937107861042,
      "num_tokens": 38105712.0,
      "step": 19180
    },
    {
      "entropy": 0.7235178224742412,
      "epoch": 0.34472537836259937,
      "grad_norm": 6.28125,
      "learning_rate": 3.754186436926632e-05,
      "loss": 0.7619231224060059,
      "mean_token_accuracy": 0.816053093969822,
      "num_tokens": 38125983.0,
      "step": 19190
    },
    {
      "entropy": 0.7046797551214695,
      "epoch": 0.3449050163919702,
      "grad_norm": 7.875,
      "learning_rate": 3.7529434302228464e-05,
      "loss": 0.7277954578399658,
      "mean_token_accuracy": 0.8219369724392891,
      "num_tokens": 38146031.0,
      "step": 19200
    },
    {
      "entropy": 0.7113085374236107,
      "epoch": 0.345084654421341,
      "grad_norm": 4.8125,
      "learning_rate": 3.751700009743508e-05,
      "loss": 0.7442852973937988,
      "mean_token_accuracy": 0.817693717777729,
      "num_tokens": 38165427.0,
      "step": 19210
    },
    {
      "entropy": 0.7748551547527314,
      "epoch": 0.34526429245071183,
      "grad_norm": 5.59375,
      "learning_rate": 3.7504561758992505e-05,
      "loss": 0.7610948085784912,
      "mean_token_accuracy": 0.829310166835785,
      "num_tokens": 38184749.0,
      "step": 19220
    },
    {
      "entropy": 0.7265764161944389,
      "epoch": 0.34544393048008265,
      "grad_norm": 7.71875,
      "learning_rate": 3.74921192910084e-05,
      "loss": 0.7449968338012696,
      "mean_token_accuracy": 0.8206524759531021,
      "num_tokens": 38204974.0,
      "step": 19230
    },
    {
      "entropy": 0.7975287966430187,
      "epoch": 0.34562356850945347,
      "grad_norm": 8.25,
      "learning_rate": 3.747967269759181e-05,
      "loss": 0.7726430892944336,
      "mean_token_accuracy": 0.8185803309082985,
      "num_tokens": 38225103.0,
      "step": 19240
    },
    {
      "entropy": 0.5731569331139326,
      "epoch": 0.3458032065388243,
      "grad_norm": 4.84375,
      "learning_rate": 3.7467221982853125e-05,
      "loss": 0.5851688385009766,
      "mean_token_accuracy": 0.8555969774723053,
      "num_tokens": 38244715.0,
      "step": 19250
    },
    {
      "entropy": 0.7497602261602878,
      "epoch": 0.3459828445681951,
      "grad_norm": 5.0625,
      "learning_rate": 3.74547671509041e-05,
      "loss": 0.7365609169006347,
      "mean_token_accuracy": 0.8170225068926811,
      "num_tokens": 38265205.0,
      "step": 19260
    },
    {
      "entropy": 0.7458068639039993,
      "epoch": 0.3461624825975659,
      "grad_norm": 5.5,
      "learning_rate": 3.744230820585786e-05,
      "loss": 0.7546461582183838,
      "mean_token_accuracy": 0.8166630506515503,
      "num_tokens": 38285777.0,
      "step": 19270
    },
    {
      "entropy": 0.7627834901213646,
      "epoch": 0.3463421206269367,
      "grad_norm": 4.9375,
      "learning_rate": 3.742984515182888e-05,
      "loss": 0.7814962863922119,
      "mean_token_accuracy": 0.8142510116100311,
      "num_tokens": 38305896.0,
      "step": 19280
    },
    {
      "entropy": 0.7436657659709454,
      "epoch": 0.3465217586563075,
      "grad_norm": 6.0625,
      "learning_rate": 3.7417377992933006e-05,
      "loss": 0.7709619045257569,
      "mean_token_accuracy": 0.8094367116689682,
      "num_tokens": 38325186.0,
      "step": 19290
    },
    {
      "entropy": 0.6691153690218925,
      "epoch": 0.34670139668567834,
      "grad_norm": 6.28125,
      "learning_rate": 3.74049067332874e-05,
      "loss": 0.6791462898254395,
      "mean_token_accuracy": 0.8314015701413154,
      "num_tokens": 38345465.0,
      "step": 19300
    },
    {
      "entropy": 0.7306291438639164,
      "epoch": 0.34688103471504916,
      "grad_norm": 9.0625,
      "learning_rate": 3.7392431377010634e-05,
      "loss": 0.714200210571289,
      "mean_token_accuracy": 0.8198609828948975,
      "num_tokens": 38366031.0,
      "step": 19310
    },
    {
      "entropy": 0.6979415692389012,
      "epoch": 0.34706067274442,
      "grad_norm": 5.84375,
      "learning_rate": 3.7379951928222584e-05,
      "loss": 0.7082321643829346,
      "mean_token_accuracy": 0.8262175038456917,
      "num_tokens": 38386115.0,
      "step": 19320
    },
    {
      "entropy": 0.7768630117177964,
      "epoch": 0.3472403107737908,
      "grad_norm": 6.15625,
      "learning_rate": 3.7367468391044514e-05,
      "loss": 0.7835107326507569,
      "mean_token_accuracy": 0.8064766526222229,
      "num_tokens": 38404826.0,
      "step": 19330
    },
    {
      "entropy": 0.6705552920699119,
      "epoch": 0.3474199488031616,
      "grad_norm": 4.4375,
      "learning_rate": 3.7354980769599014e-05,
      "loss": 0.6855151176452636,
      "mean_token_accuracy": 0.8339838474988938,
      "num_tokens": 38424704.0,
      "step": 19340
    },
    {
      "entropy": 0.6821355912834406,
      "epoch": 0.34759958683253245,
      "grad_norm": 5.59375,
      "learning_rate": 3.734248906801003e-05,
      "loss": 0.6492209434509277,
      "mean_token_accuracy": 0.8339143961668014,
      "num_tokens": 38444330.0,
      "step": 19350
    },
    {
      "entropy": 0.7818419747054577,
      "epoch": 0.34777922486190327,
      "grad_norm": 4.65625,
      "learning_rate": 3.732999329040286e-05,
      "loss": 0.8500520706176757,
      "mean_token_accuracy": 0.7986829802393913,
      "num_tokens": 38463562.0,
      "step": 19360
    },
    {
      "entropy": 0.8236159533262253,
      "epoch": 0.3479588628912741,
      "grad_norm": 6.59375,
      "learning_rate": 3.731749344090414e-05,
      "loss": 0.8518532752990723,
      "mean_token_accuracy": 0.7995924435555934,
      "num_tokens": 38483998.0,
      "step": 19370
    },
    {
      "entropy": 0.7961986035108566,
      "epoch": 0.3481385009206449,
      "grad_norm": 6.625,
      "learning_rate": 3.7304989523641866e-05,
      "loss": 0.7961482524871826,
      "mean_token_accuracy": 0.8040114253759384,
      "num_tokens": 38503596.0,
      "step": 19380
    },
    {
      "entropy": 0.7170285671949387,
      "epoch": 0.34831813895001573,
      "grad_norm": 5.9375,
      "learning_rate": 3.729248154274535e-05,
      "loss": 0.6814708709716797,
      "mean_token_accuracy": 0.8306064814329147,
      "num_tokens": 38523550.0,
      "step": 19390
    },
    {
      "entropy": 0.6218185223639011,
      "epoch": 0.34849777697938655,
      "grad_norm": 6.4375,
      "learning_rate": 3.727996950234528e-05,
      "loss": 0.5942700862884521,
      "mean_token_accuracy": 0.8479880779981613,
      "num_tokens": 38543445.0,
      "step": 19400
    },
    {
      "entropy": 0.6584331452846527,
      "epoch": 0.3486774150087574,
      "grad_norm": 6.6875,
      "learning_rate": 3.7267453406573646e-05,
      "loss": 0.7164645671844483,
      "mean_token_accuracy": 0.828114801645279,
      "num_tokens": 38563211.0,
      "step": 19410
    },
    {
      "entropy": 0.7123130120337009,
      "epoch": 0.3488570530381282,
      "grad_norm": 6.25,
      "learning_rate": 3.7254933259563826e-05,
      "loss": 0.7497681140899658,
      "mean_token_accuracy": 0.8183339387178421,
      "num_tokens": 38582003.0,
      "step": 19420
    },
    {
      "entropy": 0.7531318098306656,
      "epoch": 0.349036691067499,
      "grad_norm": 8.5,
      "learning_rate": 3.724240906545049e-05,
      "loss": 0.745562219619751,
      "mean_token_accuracy": 0.8220388367772102,
      "num_tokens": 38601447.0,
      "step": 19430
    },
    {
      "entropy": 0.7783000510185957,
      "epoch": 0.3492163290968698,
      "grad_norm": 5.75,
      "learning_rate": 3.722988082836967e-05,
      "loss": 0.7921068668365479,
      "mean_token_accuracy": 0.8143302634358406,
      "num_tokens": 38622222.0,
      "step": 19440
    },
    {
      "entropy": 0.7307706788182259,
      "epoch": 0.3493959671262406,
      "grad_norm": 6.21875,
      "learning_rate": 3.721734855245872e-05,
      "loss": 0.7375482082366943,
      "mean_token_accuracy": 0.823814032971859,
      "num_tokens": 38643011.0,
      "step": 19450
    },
    {
      "entropy": 0.7250261522829533,
      "epoch": 0.3495756051556114,
      "grad_norm": 6.0625,
      "learning_rate": 3.720481224185633e-05,
      "loss": 0.7338265419006348,
      "mean_token_accuracy": 0.8206143364310264,
      "num_tokens": 38663207.0,
      "step": 19460
    },
    {
      "entropy": 0.6972012229263782,
      "epoch": 0.34975524318498225,
      "grad_norm": 6.90625,
      "learning_rate": 3.7192271900702554e-05,
      "loss": 0.7050270080566406,
      "mean_token_accuracy": 0.8321330428123475,
      "num_tokens": 38682920.0,
      "step": 19470
    },
    {
      "entropy": 0.6967833112925291,
      "epoch": 0.34993488121435307,
      "grad_norm": 4.5,
      "learning_rate": 3.717972753313871e-05,
      "loss": 0.6925751209259033,
      "mean_token_accuracy": 0.8247395232319832,
      "num_tokens": 38702535.0,
      "step": 19480
    },
    {
      "entropy": 0.7052656009793281,
      "epoch": 0.3501145192437239,
      "grad_norm": 5.75,
      "learning_rate": 3.7167179143307514e-05,
      "loss": 0.6996980667114258,
      "mean_token_accuracy": 0.8199231758713722,
      "num_tokens": 38721777.0,
      "step": 19490
    },
    {
      "entropy": 0.7232063189148903,
      "epoch": 0.3502941572730947,
      "grad_norm": 6.40625,
      "learning_rate": 3.7154626735352964e-05,
      "loss": 0.7731166839599609,
      "mean_token_accuracy": 0.8144383385777474,
      "num_tokens": 38740962.0,
      "step": 19500
    },
    {
      "entropy": 0.6854423105716705,
      "epoch": 0.35047379530246553,
      "grad_norm": 5.84375,
      "learning_rate": 3.714207031342042e-05,
      "loss": 0.709846019744873,
      "mean_token_accuracy": 0.8282036945223809,
      "num_tokens": 38762405.0,
      "step": 19510
    },
    {
      "entropy": 0.677189652621746,
      "epoch": 0.35065343333183635,
      "grad_norm": 4.46875,
      "learning_rate": 3.712950988165653e-05,
      "loss": 0.6914024829864502,
      "mean_token_accuracy": 0.8244973734021187,
      "num_tokens": 38782050.0,
      "step": 19520
    },
    {
      "entropy": 0.7527848079800605,
      "epoch": 0.35083307136120717,
      "grad_norm": 6.46875,
      "learning_rate": 3.7116945444209306e-05,
      "loss": 0.7600464344024658,
      "mean_token_accuracy": 0.8162539556622506,
      "num_tokens": 38801977.0,
      "step": 19530
    },
    {
      "entropy": 0.7454664569348097,
      "epoch": 0.351012709390578,
      "grad_norm": 6.40625,
      "learning_rate": 3.710437700522805e-05,
      "loss": 0.7491289138793945,
      "mean_token_accuracy": 0.8286859825253486,
      "num_tokens": 38820532.0,
      "step": 19540
    },
    {
      "entropy": 0.7071416556835175,
      "epoch": 0.3511923474199488,
      "grad_norm": 6.125,
      "learning_rate": 3.70918045688634e-05,
      "loss": 0.6822431087493896,
      "mean_token_accuracy": 0.8370253548026085,
      "num_tokens": 38842067.0,
      "step": 19550
    },
    {
      "entropy": 0.7024726055562496,
      "epoch": 0.35137198544931963,
      "grad_norm": 5.25,
      "learning_rate": 3.707922813926731e-05,
      "loss": 0.7256744384765625,
      "mean_token_accuracy": 0.8155663028359413,
      "num_tokens": 38862431.0,
      "step": 19560
    },
    {
      "entropy": 0.7440385214984417,
      "epoch": 0.35155162347869046,
      "grad_norm": 5.0,
      "learning_rate": 3.706664772059306e-05,
      "loss": 0.7175159454345703,
      "mean_token_accuracy": 0.8276472702622414,
      "num_tokens": 38883744.0,
      "step": 19570
    },
    {
      "entropy": 0.7138817705214023,
      "epoch": 0.3517312615080613,
      "grad_norm": 6.0,
      "learning_rate": 3.705406331699525e-05,
      "loss": 0.7633171081542969,
      "mean_token_accuracy": 0.8222905993461609,
      "num_tokens": 38904011.0,
      "step": 19580
    },
    {
      "entropy": 0.7213800169527531,
      "epoch": 0.3519108995374321,
      "grad_norm": 5.0625,
      "learning_rate": 3.704147493262978e-05,
      "loss": 0.7514737606048584,
      "mean_token_accuracy": 0.8108336687088012,
      "num_tokens": 38923084.0,
      "step": 19590
    },
    {
      "entropy": 0.7661073513329029,
      "epoch": 0.3520905375668029,
      "grad_norm": 6.5625,
      "learning_rate": 3.7028882571653874e-05,
      "loss": 0.7500514030456543,
      "mean_token_accuracy": 0.8162610396742821,
      "num_tokens": 38942573.0,
      "step": 19600
    },
    {
      "entropy": 0.7282956473529338,
      "epoch": 0.3522701755961737,
      "grad_norm": 6.25,
      "learning_rate": 3.7016286238226067e-05,
      "loss": 0.7041528701782227,
      "mean_token_accuracy": 0.8238245025277138,
      "num_tokens": 38961629.0,
      "step": 19610
    },
    {
      "entropy": 0.7249262992292642,
      "epoch": 0.3524498136255445,
      "grad_norm": 5.4375,
      "learning_rate": 3.7003685936506215e-05,
      "loss": 0.7533349990844727,
      "mean_token_accuracy": 0.8187106624245644,
      "num_tokens": 38981060.0,
      "step": 19620
    },
    {
      "entropy": 0.7481336526572704,
      "epoch": 0.3526294516549153,
      "grad_norm": 5.28125,
      "learning_rate": 3.699108167065546e-05,
      "loss": 0.7613324642181396,
      "mean_token_accuracy": 0.813330452144146,
      "num_tokens": 39001040.0,
      "step": 19630
    },
    {
      "entropy": 0.7565899334847928,
      "epoch": 0.35280908968428615,
      "grad_norm": 6.09375,
      "learning_rate": 3.697847344483629e-05,
      "loss": 0.7764715671539306,
      "mean_token_accuracy": 0.8162884220480919,
      "num_tokens": 39021544.0,
      "step": 19640
    },
    {
      "entropy": 0.7485118616372347,
      "epoch": 0.35298872771365697,
      "grad_norm": 5.125,
      "learning_rate": 3.696586126321248e-05,
      "loss": 0.6814226150512696,
      "mean_token_accuracy": 0.8276503384113312,
      "num_tokens": 39041079.0,
      "step": 19650
    },
    {
      "entropy": 0.7217489499598742,
      "epoch": 0.3531683657430278,
      "grad_norm": 5.6875,
      "learning_rate": 3.6953245129949085e-05,
      "loss": 0.749444580078125,
      "mean_token_accuracy": 0.8240980982780457,
      "num_tokens": 39060446.0,
      "step": 19660
    },
    {
      "entropy": 0.6897683598101139,
      "epoch": 0.3533480037723986,
      "grad_norm": 4.625,
      "learning_rate": 3.694062504921253e-05,
      "loss": 0.6841042518615723,
      "mean_token_accuracy": 0.8274684309959411,
      "num_tokens": 39081505.0,
      "step": 19670
    },
    {
      "entropy": 0.6393590599298478,
      "epoch": 0.35352764180176943,
      "grad_norm": 6.09375,
      "learning_rate": 3.692800102517047e-05,
      "loss": 0.6473001003265381,
      "mean_token_accuracy": 0.8371409267187119,
      "num_tokens": 39100466.0,
      "step": 19680
    },
    {
      "entropy": 0.7079867459833622,
      "epoch": 0.35370727983114025,
      "grad_norm": 6.53125,
      "learning_rate": 3.691537306199193e-05,
      "loss": 0.7407841682434082,
      "mean_token_accuracy": 0.8245193272829056,
      "num_tokens": 39122009.0,
      "step": 19690
    },
    {
      "entropy": 0.6577744990587234,
      "epoch": 0.3538869178605111,
      "grad_norm": 5.8125,
      "learning_rate": 3.690274116384719e-05,
      "loss": 0.649692153930664,
      "mean_token_accuracy": 0.8371589824557304,
      "num_tokens": 39141733.0,
      "step": 19700
    },
    {
      "entropy": 0.7200697928667068,
      "epoch": 0.3540665558898819,
      "grad_norm": 7.40625,
      "learning_rate": 3.6890105334907843e-05,
      "loss": 0.7775652885437012,
      "mean_token_accuracy": 0.8101379722356796,
      "num_tokens": 39161180.0,
      "step": 19710
    },
    {
      "entropy": 0.6804958343505859,
      "epoch": 0.3542461939192527,
      "grad_norm": 5.0,
      "learning_rate": 3.687746557934677e-05,
      "loss": 0.6561213970184326,
      "mean_token_accuracy": 0.8360828205943107,
      "num_tokens": 39181958.0,
      "step": 19720
    },
    {
      "entropy": 0.7525444965809583,
      "epoch": 0.35442583194862354,
      "grad_norm": 6.59375,
      "learning_rate": 3.686482190133818e-05,
      "loss": 0.7912158489227294,
      "mean_token_accuracy": 0.8218772232532501,
      "num_tokens": 39202302.0,
      "step": 19730
    },
    {
      "entropy": 0.6691570825874805,
      "epoch": 0.35460546997799436,
      "grad_norm": 5.875,
      "learning_rate": 3.685217430505755e-05,
      "loss": 0.6586970329284668,
      "mean_token_accuracy": 0.832605043053627,
      "num_tokens": 39221488.0,
      "step": 19740
    },
    {
      "entropy": 0.7555838368833065,
      "epoch": 0.3547851080073652,
      "grad_norm": 5.59375,
      "learning_rate": 3.683952279468164e-05,
      "loss": 0.7917445182800293,
      "mean_token_accuracy": 0.81813974827528,
      "num_tokens": 39241628.0,
      "step": 19750
    },
    {
      "entropy": 0.6395846024155617,
      "epoch": 0.354964746036736,
      "grad_norm": 6.4375,
      "learning_rate": 3.6826867374388543e-05,
      "loss": 0.6577080249786377,
      "mean_token_accuracy": 0.8358124822378159,
      "num_tokens": 39260693.0,
      "step": 19760
    },
    {
      "entropy": 0.6996312208473683,
      "epoch": 0.3551443840661068,
      "grad_norm": 6.03125,
      "learning_rate": 3.681420804835761e-05,
      "loss": 0.6488876819610596,
      "mean_token_accuracy": 0.8398793131113053,
      "num_tokens": 39281467.0,
      "step": 19770
    },
    {
      "entropy": 0.6990426421165467,
      "epoch": 0.3553240220954776,
      "grad_norm": 5.71875,
      "learning_rate": 3.68015448207695e-05,
      "loss": 0.7335831642150878,
      "mean_token_accuracy": 0.821978397667408,
      "num_tokens": 39301169.0,
      "step": 19780
    },
    {
      "entropy": 0.7376824270933866,
      "epoch": 0.3555036601248484,
      "grad_norm": 5.65625,
      "learning_rate": 3.6788877695806144e-05,
      "loss": 0.7520842552185059,
      "mean_token_accuracy": 0.8214549958705902,
      "num_tokens": 39321292.0,
      "step": 19790
    },
    {
      "entropy": 0.6470536831766367,
      "epoch": 0.35568329815421923,
      "grad_norm": 5.3125,
      "learning_rate": 3.677620667765078e-05,
      "loss": 0.6641606330871582,
      "mean_token_accuracy": 0.8322773054242134,
      "num_tokens": 39341290.0,
      "step": 19800
    },
    {
      "entropy": 0.7078388050198555,
      "epoch": 0.35586293618359005,
      "grad_norm": 7.09375,
      "learning_rate": 3.6763531770487894e-05,
      "loss": 0.72921781539917,
      "mean_token_accuracy": 0.8244057804346084,
      "num_tokens": 39361365.0,
      "step": 19810
    },
    {
      "entropy": 0.8154718972742557,
      "epoch": 0.35604257421296087,
      "grad_norm": 6.46875,
      "learning_rate": 3.675085297850332e-05,
      "loss": 0.8529019355773926,
      "mean_token_accuracy": 0.8069949239492417,
      "num_tokens": 39381863.0,
      "step": 19820
    },
    {
      "entropy": 0.68941012583673,
      "epoch": 0.3562222122423317,
      "grad_norm": 5.46875,
      "learning_rate": 3.673817030588411e-05,
      "loss": 0.6655060768127441,
      "mean_token_accuracy": 0.8340241625905037,
      "num_tokens": 39402349.0,
      "step": 19830
    },
    {
      "entropy": 0.7359036855399609,
      "epoch": 0.3564018502717025,
      "grad_norm": 7.0,
      "learning_rate": 3.672548375681864e-05,
      "loss": 0.6846307754516602,
      "mean_token_accuracy": 0.8293835997581482,
      "num_tokens": 39422871.0,
      "step": 19840
    },
    {
      "entropy": 0.7430458717048168,
      "epoch": 0.35658148830107333,
      "grad_norm": 8.75,
      "learning_rate": 3.6712793335496536e-05,
      "loss": 0.8190174102783203,
      "mean_token_accuracy": 0.8075822867453099,
      "num_tokens": 39442900.0,
      "step": 19850
    },
    {
      "entropy": 0.6816018588840962,
      "epoch": 0.35676112633044416,
      "grad_norm": 3.875,
      "learning_rate": 3.6700099046108735e-05,
      "loss": 0.6835995197296143,
      "mean_token_accuracy": 0.8273237034678459,
      "num_tokens": 39462698.0,
      "step": 19860
    },
    {
      "entropy": 0.7785143665969372,
      "epoch": 0.356940764359815,
      "grad_norm": 5.84375,
      "learning_rate": 3.668740089284743e-05,
      "loss": 0.791728162765503,
      "mean_token_accuracy": 0.8137085393071175,
      "num_tokens": 39482219.0,
      "step": 19870
    },
    {
      "entropy": 0.6846207953989506,
      "epoch": 0.3571204023891858,
      "grad_norm": 4.71875,
      "learning_rate": 3.66746988799061e-05,
      "loss": 0.6944001197814942,
      "mean_token_accuracy": 0.8188466727733612,
      "num_tokens": 39502012.0,
      "step": 19880
    },
    {
      "entropy": 0.8880552910268307,
      "epoch": 0.3573000404185566,
      "grad_norm": 6.71875,
      "learning_rate": 3.6661993011479475e-05,
      "loss": 0.9361356735229492,
      "mean_token_accuracy": 0.791209614276886,
      "num_tokens": 39521430.0,
      "step": 19890
    },
    {
      "entropy": 0.7826593890786171,
      "epoch": 0.35747967844792744,
      "grad_norm": 4.9375,
      "learning_rate": 3.6649283291763584e-05,
      "loss": 0.7490047931671142,
      "mean_token_accuracy": 0.8159813627600669,
      "num_tokens": 39542135.0,
      "step": 19900
    },
    {
      "entropy": 0.7004616983234883,
      "epoch": 0.35765931647729826,
      "grad_norm": 6.84375,
      "learning_rate": 3.6636569724955736e-05,
      "loss": 0.7144822597503662,
      "mean_token_accuracy": 0.8229681357741356,
      "num_tokens": 39561941.0,
      "step": 19910
    },
    {
      "entropy": 0.7251121565699578,
      "epoch": 0.3578389545066691,
      "grad_norm": 5.90625,
      "learning_rate": 3.662385231525447e-05,
      "loss": 0.7363756179809571,
      "mean_token_accuracy": 0.8226905658841133,
      "num_tokens": 39581015.0,
      "step": 19920
    },
    {
      "entropy": 0.6704909421503544,
      "epoch": 0.3580185925360399,
      "grad_norm": 4.96875,
      "learning_rate": 3.661113106685965e-05,
      "loss": 0.6460696220397949,
      "mean_token_accuracy": 0.8354404449462891,
      "num_tokens": 39600564.0,
      "step": 19930
    },
    {
      "entropy": 0.660530687123537,
      "epoch": 0.3581982305654107,
      "grad_norm": 5.5625,
      "learning_rate": 3.659840598397234e-05,
      "loss": 0.6877734184265136,
      "mean_token_accuracy": 0.8326182022690773,
      "num_tokens": 39620032.0,
      "step": 19940
    },
    {
      "entropy": 0.7667974092066288,
      "epoch": 0.3583778685947815,
      "grad_norm": 6.15625,
      "learning_rate": 3.658567707079494e-05,
      "loss": 0.7793787956237793,
      "mean_token_accuracy": 0.804989618062973,
      "num_tokens": 39640175.0,
      "step": 19950
    },
    {
      "entropy": 0.7590945478528738,
      "epoch": 0.3585575066241523,
      "grad_norm": 5.5625,
      "learning_rate": 3.6572944331531064e-05,
      "loss": 0.7614902973175048,
      "mean_token_accuracy": 0.8122261539101601,
      "num_tokens": 39660011.0,
      "step": 19960
    },
    {
      "entropy": 0.7114308450371027,
      "epoch": 0.35873714465352313,
      "grad_norm": 5.8125,
      "learning_rate": 3.6560207770385615e-05,
      "loss": 0.7441701889038086,
      "mean_token_accuracy": 0.8226582407951355,
      "num_tokens": 39679296.0,
      "step": 19970
    },
    {
      "entropy": 0.6666687287390232,
      "epoch": 0.35891678268289395,
      "grad_norm": 6.46875,
      "learning_rate": 3.654746739156474e-05,
      "loss": 0.681654977798462,
      "mean_token_accuracy": 0.8284374788403511,
      "num_tokens": 39699090.0,
      "step": 19980
    },
    {
      "entropy": 0.7235564418137074,
      "epoch": 0.3590964207122648,
      "grad_norm": 7.09375,
      "learning_rate": 3.6534723199275875e-05,
      "loss": 0.7334863662719726,
      "mean_token_accuracy": 0.8140013873577118,
      "num_tokens": 39719927.0,
      "step": 19990
    },
    {
      "epoch": 0.3592760587416356,
      "eval_entropy": 0.6944533010721207,
      "eval_loss": 0.6942692995071411,
      "eval_mean_token_accuracy": 0.8306707270145416,
      "eval_num_tokens": 39740177.0,
      "eval_runtime": 40.5397,
      "eval_samples_per_second": 49.334,
      "eval_steps_per_second": 6.167,
      "step": 20000
    },
    {
      "entropy": 0.7013344779610634,
      "epoch": 0.3594556967710064,
      "grad_norm": 5.71875,
      "learning_rate": 3.650922339113009e-05,
      "loss": 0.6944020748138428,
      "mean_token_accuracy": 0.8249864757061005,
      "num_tokens": 39760159.0,
      "step": 20010
    },
    {
      "entropy": 0.6237478792667389,
      "epoch": 0.35963533480037724,
      "grad_norm": 6.53125,
      "learning_rate": 3.6496467783694326e-05,
      "loss": 0.5890854358673095,
      "mean_token_accuracy": 0.8490129664540291,
      "num_tokens": 39779550.0,
      "step": 20020
    },
    {
      "entropy": 0.6795018419623375,
      "epoch": 0.35981497282974806,
      "grad_norm": 5.15625,
      "learning_rate": 3.64837083796328e-05,
      "loss": 0.6895266056060791,
      "mean_token_accuracy": 0.821311953663826,
      "num_tokens": 39800849.0,
      "step": 20030
    },
    {
      "entropy": 0.6331407427787781,
      "epoch": 0.3599946108591189,
      "grad_norm": 6.8125,
      "learning_rate": 3.647094518315923e-05,
      "loss": 0.640320873260498,
      "mean_token_accuracy": 0.8492522686719894,
      "num_tokens": 39821414.0,
      "step": 20040
    },
    {
      "entropy": 0.684369832277298,
      "epoch": 0.3601742488884897,
      "grad_norm": 4.59375,
      "learning_rate": 3.6458178198488574e-05,
      "loss": 0.7153933048248291,
      "mean_token_accuracy": 0.8293255597352982,
      "num_tokens": 39842582.0,
      "step": 20050
    },
    {
      "entropy": 0.8237659387290478,
      "epoch": 0.3603538869178605,
      "grad_norm": 5.6875,
      "learning_rate": 3.644540742983703e-05,
      "loss": 0.861668872833252,
      "mean_token_accuracy": 0.8008444100618363,
      "num_tokens": 39861715.0,
      "step": 20060
    },
    {
      "entropy": 0.7737053610384464,
      "epoch": 0.36053352494723134,
      "grad_norm": 10.125,
      "learning_rate": 3.643263288142205e-05,
      "loss": 0.7810943126678467,
      "mean_token_accuracy": 0.8098478958010673,
      "num_tokens": 39881751.0,
      "step": 20070
    },
    {
      "entropy": 0.7501410365104675,
      "epoch": 0.36071316297660216,
      "grad_norm": 6.09375,
      "learning_rate": 3.641985455746235e-05,
      "loss": 0.7198531150817871,
      "mean_token_accuracy": 0.8267974510788918,
      "num_tokens": 39902124.0,
      "step": 20080
    },
    {
      "entropy": 0.7151833653450013,
      "epoch": 0.360892801005973,
      "grad_norm": 5.78125,
      "learning_rate": 3.640707246217785e-05,
      "loss": 0.712099552154541,
      "mean_token_accuracy": 0.8209892630577087,
      "num_tokens": 39921721.0,
      "step": 20090
    },
    {
      "entropy": 0.6613160043954849,
      "epoch": 0.3610724390353438,
      "grad_norm": 5.9375,
      "learning_rate": 3.639428659978978e-05,
      "loss": 0.6676843643188477,
      "mean_token_accuracy": 0.8368137508630753,
      "num_tokens": 39941695.0,
      "step": 20100
    },
    {
      "entropy": 0.6396045289933682,
      "epoch": 0.3612520770647146,
      "grad_norm": 4.3125,
      "learning_rate": 3.638149697452056e-05,
      "loss": 0.7175408363342285,
      "mean_token_accuracy": 0.8297272369265556,
      "num_tokens": 39960680.0,
      "step": 20110
    },
    {
      "entropy": 0.7199223957955837,
      "epoch": 0.3614317150940854,
      "grad_norm": 5.90625,
      "learning_rate": 3.6368703590593876e-05,
      "loss": 0.7145095825195312,
      "mean_token_accuracy": 0.8290554374456406,
      "num_tokens": 39981024.0,
      "step": 20120
    },
    {
      "entropy": 0.765744150429964,
      "epoch": 0.3616113531234562,
      "grad_norm": 8.1875,
      "learning_rate": 3.635590645223465e-05,
      "loss": 0.7662083625793457,
      "mean_token_accuracy": 0.8059632152318954,
      "num_tokens": 40000335.0,
      "step": 20130
    },
    {
      "entropy": 0.6582271385937929,
      "epoch": 0.36179099115282704,
      "grad_norm": 5.75,
      "learning_rate": 3.634310556366905e-05,
      "loss": 0.6708737373352051,
      "mean_token_accuracy": 0.8384749323129654,
      "num_tokens": 40020498.0,
      "step": 20140
    },
    {
      "entropy": 0.7423563003540039,
      "epoch": 0.36197062918219786,
      "grad_norm": 5.0625,
      "learning_rate": 3.633030092912447e-05,
      "loss": 0.7253873825073243,
      "mean_token_accuracy": 0.8240020588040352,
      "num_tokens": 40040171.0,
      "step": 20150
    },
    {
      "entropy": 0.625378180295229,
      "epoch": 0.3621502672115687,
      "grad_norm": 4.90625,
      "learning_rate": 3.631749255282954e-05,
      "loss": 0.6063027858734131,
      "mean_token_accuracy": 0.8435919210314751,
      "num_tokens": 40059885.0,
      "step": 20160
    },
    {
      "entropy": 0.6876794174313545,
      "epoch": 0.3623299052409395,
      "grad_norm": 5.40625,
      "learning_rate": 3.630468043901416e-05,
      "loss": 0.7047016620635986,
      "mean_token_accuracy": 0.8221633091568947,
      "num_tokens": 40080094.0,
      "step": 20170
    },
    {
      "entropy": 0.6922564268112182,
      "epoch": 0.3625095432703103,
      "grad_norm": 5.59375,
      "learning_rate": 3.6291864591909414e-05,
      "loss": 0.6729543685913086,
      "mean_token_accuracy": 0.8316970765590668,
      "num_tokens": 40099321.0,
      "step": 20180
    },
    {
      "entropy": 0.7137661136686801,
      "epoch": 0.36268918129968114,
      "grad_norm": 5.625,
      "learning_rate": 3.6279045015747656e-05,
      "loss": 0.7086482048034668,
      "mean_token_accuracy": 0.8198181629180908,
      "num_tokens": 40118280.0,
      "step": 20190
    },
    {
      "entropy": 0.7512328773736954,
      "epoch": 0.36286881932905196,
      "grad_norm": 5.46875,
      "learning_rate": 3.626622171476246e-05,
      "loss": 0.7443258762359619,
      "mean_token_accuracy": 0.818686555325985,
      "num_tokens": 40137749.0,
      "step": 20200
    },
    {
      "entropy": 0.7105993315577507,
      "epoch": 0.3630484573584228,
      "grad_norm": 5.71875,
      "learning_rate": 3.6253394693188614e-05,
      "loss": 0.746551513671875,
      "mean_token_accuracy": 0.8173611864447594,
      "num_tokens": 40158263.0,
      "step": 20210
    },
    {
      "entropy": 0.7092254210263491,
      "epoch": 0.3632280953877936,
      "grad_norm": 9.1875,
      "learning_rate": 3.624056395526216e-05,
      "loss": 0.7148334980010986,
      "mean_token_accuracy": 0.8275555238127709,
      "num_tokens": 40177840.0,
      "step": 20220
    },
    {
      "entropy": 0.7027041845023632,
      "epoch": 0.3634077334171644,
      "grad_norm": 4.34375,
      "learning_rate": 3.622772950522036e-05,
      "loss": 0.7142897129058838,
      "mean_token_accuracy": 0.8278562650084496,
      "num_tokens": 40198084.0,
      "step": 20230
    },
    {
      "entropy": 0.705822728574276,
      "epoch": 0.36358737144653525,
      "grad_norm": 5.15625,
      "learning_rate": 3.621489134730171e-05,
      "loss": 0.7521200180053711,
      "mean_token_accuracy": 0.8146567225456238,
      "num_tokens": 40217199.0,
      "step": 20240
    },
    {
      "entropy": 0.7429266415536404,
      "epoch": 0.36376700947590607,
      "grad_norm": 5.46875,
      "learning_rate": 3.6202049485745894e-05,
      "loss": 0.7194550037384033,
      "mean_token_accuracy": 0.8256473749876022,
      "num_tokens": 40237375.0,
      "step": 20250
    },
    {
      "entropy": 0.7113359659910202,
      "epoch": 0.3639466475052769,
      "grad_norm": 5.40625,
      "learning_rate": 3.618920392479385e-05,
      "loss": 0.722438907623291,
      "mean_token_accuracy": 0.8275696858763695,
      "num_tokens": 40257209.0,
      "step": 20260
    },
    {
      "entropy": 0.6595685303211212,
      "epoch": 0.3641262855346477,
      "grad_norm": 5.625,
      "learning_rate": 3.617635466868774e-05,
      "loss": 0.6976892948150635,
      "mean_token_accuracy": 0.8322662249207496,
      "num_tokens": 40276219.0,
      "step": 20270
    },
    {
      "entropy": 0.669034144282341,
      "epoch": 0.36430592356401853,
      "grad_norm": 5.875,
      "learning_rate": 3.616350172167094e-05,
      "loss": 0.6463847160339355,
      "mean_token_accuracy": 0.8287868216633797,
      "num_tokens": 40296230.0,
      "step": 20280
    },
    {
      "entropy": 0.7146677948534489,
      "epoch": 0.3644855615933893,
      "grad_norm": 4.03125,
      "learning_rate": 3.6150645087988037e-05,
      "loss": 0.68258056640625,
      "mean_token_accuracy": 0.8271263122558594,
      "num_tokens": 40315585.0,
      "step": 20290
    },
    {
      "entropy": 0.6644920468330383,
      "epoch": 0.3646651996227601,
      "grad_norm": 3.953125,
      "learning_rate": 3.6137784771884836e-05,
      "loss": 0.6472136974334717,
      "mean_token_accuracy": 0.8422638654708863,
      "num_tokens": 40335838.0,
      "step": 20300
    },
    {
      "entropy": 0.6753864359110594,
      "epoch": 0.36484483765213094,
      "grad_norm": 5.34375,
      "learning_rate": 3.612492077760838e-05,
      "loss": 0.6792931079864502,
      "mean_token_accuracy": 0.8317681014537811,
      "num_tokens": 40356436.0,
      "step": 20310
    },
    {
      "entropy": 0.6847208485007286,
      "epoch": 0.36502447568150176,
      "grad_norm": 6.0625,
      "learning_rate": 3.6112053109406896e-05,
      "loss": 0.7046985626220703,
      "mean_token_accuracy": 0.8303431078791619,
      "num_tokens": 40375830.0,
      "step": 20320
    },
    {
      "entropy": 0.6820194914937019,
      "epoch": 0.3652041137108726,
      "grad_norm": 7.0625,
      "learning_rate": 3.609918177152985e-05,
      "loss": 0.6546438694000244,
      "mean_token_accuracy": 0.8450311154127121,
      "num_tokens": 40395857.0,
      "step": 20330
    },
    {
      "entropy": 0.6877770833671093,
      "epoch": 0.3653837517402434,
      "grad_norm": 7.09375,
      "learning_rate": 3.6086306768227906e-05,
      "loss": 0.6881216049194336,
      "mean_token_accuracy": 0.8294637113809585,
      "num_tokens": 40415066.0,
      "step": 20340
    },
    {
      "entropy": 0.7158541504293681,
      "epoch": 0.3655633897696142,
      "grad_norm": 7.59375,
      "learning_rate": 3.607342810375293e-05,
      "loss": 0.7619421005249023,
      "mean_token_accuracy": 0.8182507574558258,
      "num_tokens": 40435507.0,
      "step": 20350
    },
    {
      "entropy": 0.6468711584806442,
      "epoch": 0.36574302779898504,
      "grad_norm": 8.1875,
      "learning_rate": 3.6060545782358026e-05,
      "loss": 0.682933235168457,
      "mean_token_accuracy": 0.8350359842181205,
      "num_tokens": 40455309.0,
      "step": 20360
    },
    {
      "entropy": 0.6519595451653004,
      "epoch": 0.36592266582835586,
      "grad_norm": 6.0625,
      "learning_rate": 3.604765980829747e-05,
      "loss": 0.6527403831481934,
      "mean_token_accuracy": 0.8345248341560364,
      "num_tokens": 40475262.0,
      "step": 20370
    },
    {
      "entropy": 0.7483419723808765,
      "epoch": 0.3661023038577267,
      "grad_norm": 5.6875,
      "learning_rate": 3.603477018582678e-05,
      "loss": 0.7906387329101563,
      "mean_token_accuracy": 0.8112152606248856,
      "num_tokens": 40496653.0,
      "step": 20380
    },
    {
      "entropy": 0.7134924802929163,
      "epoch": 0.3662819418870975,
      "grad_norm": 7.09375,
      "learning_rate": 3.6021876919202644e-05,
      "loss": 0.6915002822875976,
      "mean_token_accuracy": 0.8235787972807884,
      "num_tokens": 40517642.0,
      "step": 20390
    },
    {
      "entropy": 0.6885743692517281,
      "epoch": 0.3664615799164683,
      "grad_norm": 5.96875,
      "learning_rate": 3.600898001268298e-05,
      "loss": 0.7159062385559082,
      "mean_token_accuracy": 0.8227503642439842,
      "num_tokens": 40536493.0,
      "step": 20400
    },
    {
      "entropy": 0.7345754653215408,
      "epoch": 0.36664121794583915,
      "grad_norm": 7.03125,
      "learning_rate": 3.59960794705269e-05,
      "loss": 0.7159864425659179,
      "mean_token_accuracy": 0.8271230682730675,
      "num_tokens": 40557248.0,
      "step": 20410
    },
    {
      "entropy": 0.6364628583192825,
      "epoch": 0.36682085597520997,
      "grad_norm": 3.953125,
      "learning_rate": 3.598317529699471e-05,
      "loss": 0.6005306720733643,
      "mean_token_accuracy": 0.8479072451591492,
      "num_tokens": 40577382.0,
      "step": 20420
    },
    {
      "entropy": 0.7025217309594154,
      "epoch": 0.3670004940045808,
      "grad_norm": 4.96875,
      "learning_rate": 3.597026749634792e-05,
      "loss": 0.7096610069274902,
      "mean_token_accuracy": 0.8188622146844864,
      "num_tokens": 40597578.0,
      "step": 20430
    },
    {
      "entropy": 0.6743962246924639,
      "epoch": 0.3671801320339516,
      "grad_norm": 6.4375,
      "learning_rate": 3.5957356072849254e-05,
      "loss": 0.7003351211547851,
      "mean_token_accuracy": 0.8283336594700813,
      "num_tokens": 40616509.0,
      "step": 20440
    },
    {
      "entropy": 0.7375138387084007,
      "epoch": 0.36735977006332243,
      "grad_norm": 5.84375,
      "learning_rate": 3.594444103076259e-05,
      "loss": 0.7750445365905761,
      "mean_token_accuracy": 0.8222465500235557,
      "num_tokens": 40635970.0,
      "step": 20450
    },
    {
      "entropy": 0.7050664879381656,
      "epoch": 0.3675394080926932,
      "grad_norm": 7.90625,
      "learning_rate": 3.593152237435304e-05,
      "loss": 0.7686359882354736,
      "mean_token_accuracy": 0.8198610424995423,
      "num_tokens": 40654679.0,
      "step": 20460
    },
    {
      "entropy": 0.792995760589838,
      "epoch": 0.367719046122064,
      "grad_norm": 6.59375,
      "learning_rate": 3.591860010788691e-05,
      "loss": 0.8202888488769531,
      "mean_token_accuracy": 0.8088745936751366,
      "num_tokens": 40674425.0,
      "step": 20470
    },
    {
      "entropy": 0.7411613710224628,
      "epoch": 0.36789868415143484,
      "grad_norm": 6.40625,
      "learning_rate": 3.590567423563167e-05,
      "loss": 0.7281792640686036,
      "mean_token_accuracy": 0.8291094839572907,
      "num_tokens": 40693753.0,
      "step": 20480
    },
    {
      "entropy": 0.6900273330509663,
      "epoch": 0.36807832218080566,
      "grad_norm": 6.03125,
      "learning_rate": 3.5892744761856e-05,
      "loss": 0.6661017894744873,
      "mean_token_accuracy": 0.8430843695998191,
      "num_tokens": 40713165.0,
      "step": 20490
    },
    {
      "entropy": 0.8103882588446141,
      "epoch": 0.3682579602101765,
      "grad_norm": 7.6875,
      "learning_rate": 3.587981169082977e-05,
      "loss": 0.8332761764526367,
      "mean_token_accuracy": 0.8005744189023971,
      "num_tokens": 40732306.0,
      "step": 20500
    },
    {
      "entropy": 0.7376714583486319,
      "epoch": 0.3684375982395473,
      "grad_norm": 9.5,
      "learning_rate": 3.586687502682402e-05,
      "loss": 0.7614424705505372,
      "mean_token_accuracy": 0.8152658015489578,
      "num_tokens": 40752499.0,
      "step": 20510
    },
    {
      "entropy": 0.661192062497139,
      "epoch": 0.3686172362689181,
      "grad_norm": 5.71875,
      "learning_rate": 3.5853934774111005e-05,
      "loss": 0.6855201244354248,
      "mean_token_accuracy": 0.8343196496367454,
      "num_tokens": 40772307.0,
      "step": 20520
    },
    {
      "entropy": 0.7207657627761364,
      "epoch": 0.36879687429828895,
      "grad_norm": 6.6875,
      "learning_rate": 3.584099093696415e-05,
      "loss": 0.7796384811401367,
      "mean_token_accuracy": 0.813531118631363,
      "num_tokens": 40792695.0,
      "step": 20530
    },
    {
      "entropy": 0.698410940170288,
      "epoch": 0.36897651232765977,
      "grad_norm": 3.84375,
      "learning_rate": 3.582804351965805e-05,
      "loss": 0.6850183010101318,
      "mean_token_accuracy": 0.832542960345745,
      "num_tokens": 40811777.0,
      "step": 20540
    },
    {
      "entropy": 0.7131990224123002,
      "epoch": 0.3691561503570306,
      "grad_norm": 5.0,
      "learning_rate": 3.581509252646851e-05,
      "loss": 0.6880014419555665,
      "mean_token_accuracy": 0.8246005326509476,
      "num_tokens": 40832110.0,
      "step": 20550
    },
    {
      "entropy": 0.6933339335024357,
      "epoch": 0.3693357883864014,
      "grad_norm": 6.6875,
      "learning_rate": 3.58021379616725e-05,
      "loss": 0.7115252017974854,
      "mean_token_accuracy": 0.8217625468969345,
      "num_tokens": 40853350.0,
      "step": 20560
    },
    {
      "entropy": 0.7036506719887257,
      "epoch": 0.36951542641577223,
      "grad_norm": 5.6875,
      "learning_rate": 3.578917982954818e-05,
      "loss": 0.7371609687805176,
      "mean_token_accuracy": 0.8202033147215844,
      "num_tokens": 40872550.0,
      "step": 20570
    },
    {
      "entropy": 0.6599099144339562,
      "epoch": 0.36969506444514305,
      "grad_norm": 7.40625,
      "learning_rate": 3.577621813437485e-05,
      "loss": 0.6824038982391357,
      "mean_token_accuracy": 0.8300283327698708,
      "num_tokens": 40891089.0,
      "step": 20580
    },
    {
      "entropy": 0.6858442336320877,
      "epoch": 0.3698747024745139,
      "grad_norm": 5.78125,
      "learning_rate": 3.576325288043305e-05,
      "loss": 0.7259114742279053,
      "mean_token_accuracy": 0.8260356232523918,
      "num_tokens": 40910666.0,
      "step": 20590
    },
    {
      "entropy": 0.7528995238244534,
      "epoch": 0.3700543405038847,
      "grad_norm": 6.59375,
      "learning_rate": 3.575028407200443e-05,
      "loss": 0.736287784576416,
      "mean_token_accuracy": 0.820683453977108,
      "num_tokens": 40930315.0,
      "step": 20600
    },
    {
      "entropy": 0.7175080262124538,
      "epoch": 0.3702339785332555,
      "grad_norm": 7.90625,
      "learning_rate": 3.5737311713371884e-05,
      "loss": 0.7299849033355713,
      "mean_token_accuracy": 0.8204699426889419,
      "num_tokens": 40949011.0,
      "step": 20610
    },
    {
      "entropy": 0.7660007044672966,
      "epoch": 0.37041361656262634,
      "grad_norm": 5.09375,
      "learning_rate": 3.572433580881942e-05,
      "loss": 0.7598221778869629,
      "mean_token_accuracy": 0.8238486498594284,
      "num_tokens": 40969887.0,
      "step": 20620
    },
    {
      "entropy": 0.7726422816514968,
      "epoch": 0.3705932545919971,
      "grad_norm": 5.65625,
      "learning_rate": 3.571135636263222e-05,
      "loss": 0.7469907283782959,
      "mean_token_accuracy": 0.8175657704472542,
      "num_tokens": 40989996.0,
      "step": 20630
    },
    {
      "entropy": 0.7317269049584866,
      "epoch": 0.3707728926213679,
      "grad_norm": 5.5,
      "learning_rate": 3.569837337909669e-05,
      "loss": 0.6978418827056885,
      "mean_token_accuracy": 0.8275679588317871,
      "num_tokens": 41010153.0,
      "step": 20640
    },
    {
      "entropy": 0.7007546126842499,
      "epoch": 0.37095253065073874,
      "grad_norm": 6.65625,
      "learning_rate": 3.568538686250033e-05,
      "loss": 0.6966262817382812,
      "mean_token_accuracy": 0.8213423326611519,
      "num_tokens": 41029092.0,
      "step": 20650
    },
    {
      "entropy": 0.6580459583550692,
      "epoch": 0.37113216868010956,
      "grad_norm": 7.46875,
      "learning_rate": 3.567239681713187e-05,
      "loss": 0.6834336280822754,
      "mean_token_accuracy": 0.8368600651621818,
      "num_tokens": 41048745.0,
      "step": 20660
    },
    {
      "entropy": 0.6332890424877405,
      "epoch": 0.3713118067094804,
      "grad_norm": 8.125,
      "learning_rate": 3.565940324728116e-05,
      "loss": 0.7018272876739502,
      "mean_token_accuracy": 0.8295303642749786,
      "num_tokens": 41067650.0,
      "step": 20670
    },
    {
      "entropy": 0.6855477657169103,
      "epoch": 0.3714914447388512,
      "grad_norm": 8.3125,
      "learning_rate": 3.564640615723924e-05,
      "loss": 0.666628360748291,
      "mean_token_accuracy": 0.8349691465497017,
      "num_tokens": 41088330.0,
      "step": 20680
    },
    {
      "entropy": 0.7081369586288929,
      "epoch": 0.371671082768222,
      "grad_norm": 6.0625,
      "learning_rate": 3.5633405551298316e-05,
      "loss": 0.7413527488708496,
      "mean_token_accuracy": 0.820302052795887,
      "num_tokens": 41109029.0,
      "step": 20690
    },
    {
      "entropy": 0.6725036777555943,
      "epoch": 0.37185072079759285,
      "grad_norm": 4.71875,
      "learning_rate": 3.562040143375174e-05,
      "loss": 0.672006893157959,
      "mean_token_accuracy": 0.8392054289579391,
      "num_tokens": 41128571.0,
      "step": 20700
    },
    {
      "entropy": 0.7486584067344666,
      "epoch": 0.37203035882696367,
      "grad_norm": 7.40625,
      "learning_rate": 3.5607393808894016e-05,
      "loss": 0.7569601535797119,
      "mean_token_accuracy": 0.8119803681969643,
      "num_tokens": 41148207.0,
      "step": 20710
    },
    {
      "entropy": 0.7085374869406224,
      "epoch": 0.3722099968563345,
      "grad_norm": 6.84375,
      "learning_rate": 3.5594382681020836e-05,
      "loss": 0.7080005168914795,
      "mean_token_accuracy": 0.8279252201318741,
      "num_tokens": 41168933.0,
      "step": 20720
    },
    {
      "entropy": 0.7667396642267704,
      "epoch": 0.3723896348857053,
      "grad_norm": 6.8125,
      "learning_rate": 3.558136805442901e-05,
      "loss": 0.7845785617828369,
      "mean_token_accuracy": 0.806319385766983,
      "num_tokens": 41188641.0,
      "step": 20730
    },
    {
      "entropy": 0.6925453685224057,
      "epoch": 0.37256927291507613,
      "grad_norm": 4.375,
      "learning_rate": 3.556834993341654e-05,
      "loss": 0.6717335700988769,
      "mean_token_accuracy": 0.8295074760913849,
      "num_tokens": 41207707.0,
      "step": 20740
    },
    {
      "entropy": 0.6259505048394203,
      "epoch": 0.37274891094444695,
      "grad_norm": 4.34375,
      "learning_rate": 3.555532832228257e-05,
      "loss": 0.6191555023193359,
      "mean_token_accuracy": 0.8462332755327224,
      "num_tokens": 41227780.0,
      "step": 20750
    },
    {
      "entropy": 0.775537907332182,
      "epoch": 0.3729285489738178,
      "grad_norm": 6.34375,
      "learning_rate": 3.554230322532738e-05,
      "loss": 0.7648780822753907,
      "mean_token_accuracy": 0.817702554166317,
      "num_tokens": 41247434.0,
      "step": 20760
    },
    {
      "entropy": 0.6931198447942734,
      "epoch": 0.3731081870031886,
      "grad_norm": 5.03125,
      "learning_rate": 3.5529274646852425e-05,
      "loss": 0.6568755626678466,
      "mean_token_accuracy": 0.8285251230001449,
      "num_tokens": 41267381.0,
      "step": 20770
    },
    {
      "entropy": 0.5807228919118643,
      "epoch": 0.3732878250325594,
      "grad_norm": 6.96875,
      "learning_rate": 3.551624259116029e-05,
      "loss": 0.6186048984527588,
      "mean_token_accuracy": 0.8535513564944267,
      "num_tokens": 41289015.0,
      "step": 20780
    },
    {
      "entropy": 0.6589427012950182,
      "epoch": 0.37346746306193024,
      "grad_norm": 6.78125,
      "learning_rate": 3.550320706255474e-05,
      "loss": 0.7088614463806152,
      "mean_token_accuracy": 0.828295425325632,
      "num_tokens": 41307776.0,
      "step": 20790
    },
    {
      "entropy": 0.7005051389336586,
      "epoch": 0.373647101091301,
      "grad_norm": 5.53125,
      "learning_rate": 3.549016806534064e-05,
      "loss": 0.7337851524353027,
      "mean_token_accuracy": 0.8246641531586647,
      "num_tokens": 41327389.0,
      "step": 20800
    },
    {
      "entropy": 0.6386764496564865,
      "epoch": 0.3738267391206718,
      "grad_norm": 4.5,
      "learning_rate": 3.5477125603824026e-05,
      "loss": 0.6172388553619385,
      "mean_token_accuracy": 0.848934157192707,
      "num_tokens": 41347016.0,
      "step": 20810
    },
    {
      "entropy": 0.6635836806148291,
      "epoch": 0.37400637715004265,
      "grad_norm": 4.8125,
      "learning_rate": 3.546407968231208e-05,
      "loss": 0.6792052268981934,
      "mean_token_accuracy": 0.8319901213049888,
      "num_tokens": 41366991.0,
      "step": 20820
    },
    {
      "entropy": 0.7323547806590796,
      "epoch": 0.37418601517941347,
      "grad_norm": 5.78125,
      "learning_rate": 3.545103030511314e-05,
      "loss": 0.6899195671081543,
      "mean_token_accuracy": 0.8316585719585419,
      "num_tokens": 41388285.0,
      "step": 20830
    },
    {
      "entropy": 0.6686758436262608,
      "epoch": 0.3743656532087843,
      "grad_norm": 6.59375,
      "learning_rate": 3.5437977476536656e-05,
      "loss": 0.6565890789031983,
      "mean_token_accuracy": 0.8307007864117623,
      "num_tokens": 41408508.0,
      "step": 20840
    },
    {
      "entropy": 0.7113072119653225,
      "epoch": 0.3745452912381551,
      "grad_norm": 9.375,
      "learning_rate": 3.542492120089322e-05,
      "loss": 0.7894554138183594,
      "mean_token_accuracy": 0.8116436541080475,
      "num_tokens": 41428503.0,
      "step": 20850
    },
    {
      "entropy": 0.6463107161223889,
      "epoch": 0.37472492926752593,
      "grad_norm": 4.6875,
      "learning_rate": 3.541186148249459e-05,
      "loss": 0.6480932235717773,
      "mean_token_accuracy": 0.8367932558059692,
      "num_tokens": 41449152.0,
      "step": 20860
    },
    {
      "entropy": 0.6425647713243962,
      "epoch": 0.37490456729689675,
      "grad_norm": 7.21875,
      "learning_rate": 3.539879832565364e-05,
      "loss": 0.6824722766876221,
      "mean_token_accuracy": 0.8280842825770378,
      "num_tokens": 41468498.0,
      "step": 20870
    },
    {
      "entropy": 0.6355410009622574,
      "epoch": 0.3750842053262676,
      "grad_norm": 5.5,
      "learning_rate": 3.538573173468439e-05,
      "loss": 0.636577558517456,
      "mean_token_accuracy": 0.8400583997368812,
      "num_tokens": 41488070.0,
      "step": 20880
    },
    {
      "entropy": 0.7394064828753472,
      "epoch": 0.3752638433556384,
      "grad_norm": 7.0625,
      "learning_rate": 3.537266171390197e-05,
      "loss": 0.7315902709960938,
      "mean_token_accuracy": 0.8213183268904686,
      "num_tokens": 41507345.0,
      "step": 20890
    },
    {
      "entropy": 0.6322452001273632,
      "epoch": 0.3754434813850092,
      "grad_norm": 9.4375,
      "learning_rate": 3.535958826762267e-05,
      "loss": 0.657258939743042,
      "mean_token_accuracy": 0.8430497974157334,
      "num_tokens": 41527834.0,
      "step": 20900
    },
    {
      "entropy": 0.8347547523677349,
      "epoch": 0.37562311941438004,
      "grad_norm": 8.25,
      "learning_rate": 3.53465114001639e-05,
      "loss": 0.8560379981994629,
      "mean_token_accuracy": 0.7973290704190731,
      "num_tokens": 41548289.0,
      "step": 20910
    },
    {
      "entropy": 0.6853037364780903,
      "epoch": 0.37580275744375086,
      "grad_norm": 6.21875,
      "learning_rate": 3.5333431115844205e-05,
      "loss": 0.6840935707092285,
      "mean_token_accuracy": 0.8291462883353233,
      "num_tokens": 41568451.0,
      "step": 20920
    },
    {
      "entropy": 0.7625200234353542,
      "epoch": 0.3759823954731217,
      "grad_norm": 6.65625,
      "learning_rate": 3.532034741898325e-05,
      "loss": 0.7132830142974853,
      "mean_token_accuracy": 0.8310846492648125,
      "num_tokens": 41588340.0,
      "step": 20930
    },
    {
      "entropy": 0.7042336642742157,
      "epoch": 0.3761620335024925,
      "grad_norm": 6.46875,
      "learning_rate": 3.530726031390184e-05,
      "loss": 0.7298715591430665,
      "mean_token_accuracy": 0.8205323219299316,
      "num_tokens": 41609047.0,
      "step": 20940
    },
    {
      "entropy": 0.7789061367511749,
      "epoch": 0.3763416715318633,
      "grad_norm": 5.34375,
      "learning_rate": 3.5294169804921885e-05,
      "loss": 0.8108373641967773,
      "mean_token_accuracy": 0.8094974607229233,
      "num_tokens": 41628103.0,
      "step": 20950
    },
    {
      "entropy": 0.6398904502391816,
      "epoch": 0.37652130956123414,
      "grad_norm": 4.84375,
      "learning_rate": 3.5281075896366445e-05,
      "loss": 0.63114013671875,
      "mean_token_accuracy": 0.8428157761693,
      "num_tokens": 41648200.0,
      "step": 20960
    },
    {
      "entropy": 0.703277637436986,
      "epoch": 0.3767009475906049,
      "grad_norm": 6.59375,
      "learning_rate": 3.526797859255968e-05,
      "loss": 0.7367339611053467,
      "mean_token_accuracy": 0.8178547725081444,
      "num_tokens": 41668428.0,
      "step": 20970
    },
    {
      "entropy": 0.7693057917058468,
      "epoch": 0.3768805856199757,
      "grad_norm": 4.90625,
      "learning_rate": 3.5254877897826874e-05,
      "loss": 0.7364476203918457,
      "mean_token_accuracy": 0.8227125436067582,
      "num_tokens": 41688528.0,
      "step": 20980
    },
    {
      "entropy": 0.7052096277475357,
      "epoch": 0.37706022364934655,
      "grad_norm": 5.4375,
      "learning_rate": 3.524177381649445e-05,
      "loss": 0.668573522567749,
      "mean_token_accuracy": 0.8438011229038238,
      "num_tokens": 41708018.0,
      "step": 20990
    },
    {
      "epoch": 0.37723986167871737,
      "eval_entropy": 0.6488430182933808,
      "eval_loss": 0.6855563521385193,
      "eval_mean_token_accuracy": 0.8325611810684204,
      "eval_num_tokens": 41728160.0,
      "eval_runtime": 40.6107,
      "eval_samples_per_second": 49.248,
      "eval_steps_per_second": 6.156,
      "step": 21000
    },
    {
      "entropy": 0.685401535127312,
      "epoch": 0.3774194997080882,
      "grad_norm": 7.125,
      "learning_rate": 3.521555551134197e-05,
      "loss": 0.6980848789215088,
      "mean_token_accuracy": 0.8320543691515923,
      "num_tokens": 41747646.0,
      "step": 21010
    },
    {
      "entropy": 0.6122336108237505,
      "epoch": 0.377599137737459,
      "grad_norm": 6.875,
      "learning_rate": 3.520244129618032e-05,
      "loss": 0.6310426235198975,
      "mean_token_accuracy": 0.8563805148005486,
      "num_tokens": 41766901.0,
      "step": 21020
    },
    {
      "entropy": 0.7632059790194035,
      "epoch": 0.37777877576682983,
      "grad_norm": 8.5,
      "learning_rate": 3.5189323711735854e-05,
      "loss": 0.8137436866760254,
      "mean_token_accuracy": 0.8099944233894348,
      "num_tokens": 41786149.0,
      "step": 21030
    },
    {
      "entropy": 0.7002090759575367,
      "epoch": 0.37795841379620065,
      "grad_norm": 4.90625,
      "learning_rate": 3.517620276234056e-05,
      "loss": 0.6996436595916748,
      "mean_token_accuracy": 0.8254104003310203,
      "num_tokens": 41807072.0,
      "step": 21040
    },
    {
      "entropy": 0.6723452359437943,
      "epoch": 0.3781380518255715,
      "grad_norm": 5.1875,
      "learning_rate": 3.516307845232755e-05,
      "loss": 0.6553958892822266,
      "mean_token_accuracy": 0.8418293669819832,
      "num_tokens": 41827055.0,
      "step": 21050
    },
    {
      "entropy": 0.7408423908054829,
      "epoch": 0.3783176898549423,
      "grad_norm": 5.1875,
      "learning_rate": 3.514995078603102e-05,
      "loss": 0.7837141036987305,
      "mean_token_accuracy": 0.7975616917014122,
      "num_tokens": 41847147.0,
      "step": 21060
    },
    {
      "entropy": 0.7634066000580788,
      "epoch": 0.3784973278843131,
      "grad_norm": 6.84375,
      "learning_rate": 3.513681976778629e-05,
      "loss": 0.768970251083374,
      "mean_token_accuracy": 0.8190920308232308,
      "num_tokens": 41866899.0,
      "step": 21070
    },
    {
      "entropy": 0.645480614900589,
      "epoch": 0.37867696591368394,
      "grad_norm": 5.5,
      "learning_rate": 3.512368540192979e-05,
      "loss": 0.6082379341125488,
      "mean_token_accuracy": 0.8418129622936249,
      "num_tokens": 41886739.0,
      "step": 21080
    },
    {
      "entropy": 0.6537938177585602,
      "epoch": 0.37885660394305476,
      "grad_norm": 8.25,
      "learning_rate": 3.5110547692799026e-05,
      "loss": 0.6309549808502197,
      "mean_token_accuracy": 0.8478370845317841,
      "num_tokens": 41907319.0,
      "step": 21090
    },
    {
      "entropy": 0.6578036092221737,
      "epoch": 0.3790362419724256,
      "grad_norm": 6.1875,
      "learning_rate": 3.5097406644732664e-05,
      "loss": 0.6945337772369384,
      "mean_token_accuracy": 0.8212285339832306,
      "num_tokens": 41926696.0,
      "step": 21100
    },
    {
      "entropy": 0.7340669080615043,
      "epoch": 0.3792158800017964,
      "grad_norm": 4.9375,
      "learning_rate": 3.508426226207043e-05,
      "loss": 0.7506400585174561,
      "mean_token_accuracy": 0.817342446744442,
      "num_tokens": 41945922.0,
      "step": 21110
    },
    {
      "entropy": 0.6796356346458197,
      "epoch": 0.3793955180311672,
      "grad_norm": 6.53125,
      "learning_rate": 3.5071114549153166e-05,
      "loss": 0.699470043182373,
      "mean_token_accuracy": 0.8282402493059635,
      "num_tokens": 41965065.0,
      "step": 21120
    },
    {
      "entropy": 0.7802211627364158,
      "epoch": 0.379575156060538,
      "grad_norm": 9.0,
      "learning_rate": 3.505796351032279e-05,
      "loss": 0.8359064102172852,
      "mean_token_accuracy": 0.8052878767251969,
      "num_tokens": 41984673.0,
      "step": 21130
    },
    {
      "entropy": 0.6757221980020404,
      "epoch": 0.3797547940899088,
      "grad_norm": 3.25,
      "learning_rate": 3.504480914992235e-05,
      "loss": 0.6891399383544922,
      "mean_token_accuracy": 0.8288613453507423,
      "num_tokens": 42005196.0,
      "step": 21140
    },
    {
      "entropy": 0.6448501594364643,
      "epoch": 0.37993443211927963,
      "grad_norm": 6.5625,
      "learning_rate": 3.503165147229599e-05,
      "loss": 0.6328178882598877,
      "mean_token_accuracy": 0.838468787074089,
      "num_tokens": 42025158.0,
      "step": 21150
    },
    {
      "entropy": 0.6921123370528222,
      "epoch": 0.38011407014865045,
      "grad_norm": 4.1875,
      "learning_rate": 3.501849048178892e-05,
      "loss": 0.6879634380340576,
      "mean_token_accuracy": 0.8307283446192741,
      "num_tokens": 42044844.0,
      "step": 21160
    },
    {
      "entropy": 0.7336290996521712,
      "epoch": 0.3802937081780213,
      "grad_norm": 5.34375,
      "learning_rate": 3.500532618274748e-05,
      "loss": 0.7146403312683105,
      "mean_token_accuracy": 0.8240015953779221,
      "num_tokens": 42064728.0,
      "step": 21170
    },
    {
      "entropy": 0.6398572064936161,
      "epoch": 0.3804733462073921,
      "grad_norm": 7.625,
      "learning_rate": 3.499215857951906e-05,
      "loss": 0.6443054676055908,
      "mean_token_accuracy": 0.8377064347267151,
      "num_tokens": 42083813.0,
      "step": 21180
    },
    {
      "entropy": 0.6949018970131874,
      "epoch": 0.3806529842367629,
      "grad_norm": 8.3125,
      "learning_rate": 3.497898767645219e-05,
      "loss": 0.7198577404022217,
      "mean_token_accuracy": 0.815118819475174,
      "num_tokens": 42103212.0,
      "step": 21190
    },
    {
      "entropy": 0.7099394038319587,
      "epoch": 0.38083262226613374,
      "grad_norm": 3.421875,
      "learning_rate": 3.496581347789646e-05,
      "loss": 0.711375617980957,
      "mean_token_accuracy": 0.8298701912164688,
      "num_tokens": 42123187.0,
      "step": 21200
    },
    {
      "entropy": 0.6876069072633981,
      "epoch": 0.38101226029550456,
      "grad_norm": 7.59375,
      "learning_rate": 3.4952635988202545e-05,
      "loss": 0.7006659030914306,
      "mean_token_accuracy": 0.8265680730342865,
      "num_tokens": 42142599.0,
      "step": 21210
    },
    {
      "entropy": 0.6791185714304447,
      "epoch": 0.3811918983248754,
      "grad_norm": 4.875,
      "learning_rate": 3.4939455211722224e-05,
      "loss": 0.7223628520965576,
      "mean_token_accuracy": 0.8260677263140679,
      "num_tokens": 42161820.0,
      "step": 21220
    },
    {
      "entropy": 0.6870696507394314,
      "epoch": 0.3813715363542462,
      "grad_norm": 4.8125,
      "learning_rate": 3.492627115280836e-05,
      "loss": 0.734947919845581,
      "mean_token_accuracy": 0.8313745379447937,
      "num_tokens": 42181353.0,
      "step": 21230
    },
    {
      "entropy": 0.7005032584071159,
      "epoch": 0.381551174383617,
      "grad_norm": 7.59375,
      "learning_rate": 3.491308381581489e-05,
      "loss": 0.7267261505126953,
      "mean_token_accuracy": 0.8278243407607079,
      "num_tokens": 42201403.0,
      "step": 21240
    },
    {
      "entropy": 0.7523579679429531,
      "epoch": 0.38173081241298784,
      "grad_norm": 5.90625,
      "learning_rate": 3.489989320509682e-05,
      "loss": 0.7507474422454834,
      "mean_token_accuracy": 0.8187444761395455,
      "num_tokens": 42220674.0,
      "step": 21250
    },
    {
      "entropy": 0.7516706213355064,
      "epoch": 0.38191045044235866,
      "grad_norm": 7.34375,
      "learning_rate": 3.4886699325010265e-05,
      "loss": 0.768617820739746,
      "mean_token_accuracy": 0.817050164937973,
      "num_tokens": 42240755.0,
      "step": 21260
    },
    {
      "entropy": 0.7213110692799092,
      "epoch": 0.3820900884717295,
      "grad_norm": 7.78125,
      "learning_rate": 3.487350217991242e-05,
      "loss": 0.7076373100280762,
      "mean_token_accuracy": 0.83057651668787,
      "num_tokens": 42260267.0,
      "step": 21270
    },
    {
      "entropy": 0.7579997472465039,
      "epoch": 0.3822697265011003,
      "grad_norm": 6.5625,
      "learning_rate": 3.486030177416153e-05,
      "loss": 0.7041111946105957,
      "mean_token_accuracy": 0.8322078928351402,
      "num_tokens": 42281265.0,
      "step": 21280
    },
    {
      "entropy": 0.7179070495069027,
      "epoch": 0.3824493645304711,
      "grad_norm": 5.15625,
      "learning_rate": 3.484709811211695e-05,
      "loss": 0.6804269313812256,
      "mean_token_accuracy": 0.8361951395869255,
      "num_tokens": 42300232.0,
      "step": 21290
    },
    {
      "entropy": 0.722075591981411,
      "epoch": 0.3826290025598419,
      "grad_norm": 6.09375,
      "learning_rate": 3.483389119813909e-05,
      "loss": 0.8238638877868653,
      "mean_token_accuracy": 0.808764573931694,
      "num_tokens": 42319913.0,
      "step": 21300
    },
    {
      "entropy": 0.6796317979693413,
      "epoch": 0.3828086405892127,
      "grad_norm": 4.53125,
      "learning_rate": 3.482068103658942e-05,
      "loss": 0.6493636131286621,
      "mean_token_accuracy": 0.8387590900063515,
      "num_tokens": 42339530.0,
      "step": 21310
    },
    {
      "entropy": 0.5769486486911773,
      "epoch": 0.38298827861858353,
      "grad_norm": 4.75,
      "learning_rate": 3.480746763183053e-05,
      "loss": 0.557643461227417,
      "mean_token_accuracy": 0.8530028179287911,
      "num_tokens": 42359621.0,
      "step": 21320
    },
    {
      "entropy": 0.7712953083217144,
      "epoch": 0.38316791664795435,
      "grad_norm": 6.84375,
      "learning_rate": 3.479425098822604e-05,
      "loss": 0.8029804229736328,
      "mean_token_accuracy": 0.8159344509243965,
      "num_tokens": 42379493.0,
      "step": 21330
    },
    {
      "entropy": 0.749280133843422,
      "epoch": 0.3833475546773252,
      "grad_norm": 5.5625,
      "learning_rate": 3.4781031110140646e-05,
      "loss": 0.7428476333618164,
      "mean_token_accuracy": 0.815738758444786,
      "num_tokens": 42400146.0,
      "step": 21340
    },
    {
      "entropy": 0.6972391869872808,
      "epoch": 0.383527192706696,
      "grad_norm": 3.9375,
      "learning_rate": 3.476780800194011e-05,
      "loss": 0.6858940124511719,
      "mean_token_accuracy": 0.8398605063557625,
      "num_tokens": 42421680.0,
      "step": 21350
    },
    {
      "entropy": 0.7181228488683701,
      "epoch": 0.3837068307360668,
      "grad_norm": 4.21875,
      "learning_rate": 3.475458166799129e-05,
      "loss": 0.7356170177459717,
      "mean_token_accuracy": 0.8195689767599106,
      "num_tokens": 42442691.0,
      "step": 21360
    },
    {
      "entropy": 0.7038106553256511,
      "epoch": 0.38388646876543764,
      "grad_norm": 6.53125,
      "learning_rate": 3.474135211266207e-05,
      "loss": 0.7087418079376221,
      "mean_token_accuracy": 0.8259187340736389,
      "num_tokens": 42462515.0,
      "step": 21370
    },
    {
      "entropy": 0.6616566590964794,
      "epoch": 0.38406610679480846,
      "grad_norm": 4.34375,
      "learning_rate": 3.472811934032143e-05,
      "loss": 0.6580923080444336,
      "mean_token_accuracy": 0.8408215939998627,
      "num_tokens": 42481697.0,
      "step": 21380
    },
    {
      "entropy": 0.6362639755010605,
      "epoch": 0.3842457448241793,
      "grad_norm": 6.0,
      "learning_rate": 3.471488335533939e-05,
      "loss": 0.663828182220459,
      "mean_token_accuracy": 0.834761281311512,
      "num_tokens": 42501885.0,
      "step": 21390
    },
    {
      "entropy": 0.6487002413719892,
      "epoch": 0.3844253828535501,
      "grad_norm": 4.53125,
      "learning_rate": 3.470164416208703e-05,
      "loss": 0.6778441905975342,
      "mean_token_accuracy": 0.83646130412817,
      "num_tokens": 42521868.0,
      "step": 21400
    },
    {
      "entropy": 0.7919644050300121,
      "epoch": 0.3846050208829209,
      "grad_norm": 6.875,
      "learning_rate": 3.468840176493651e-05,
      "loss": 0.8010330200195312,
      "mean_token_accuracy": 0.8065056562423706,
      "num_tokens": 42542181.0,
      "step": 21410
    },
    {
      "entropy": 0.7306326061487198,
      "epoch": 0.38478465891229174,
      "grad_norm": 7.0,
      "learning_rate": 3.467515616826104e-05,
      "loss": 0.7506579875946044,
      "mean_token_accuracy": 0.8228535532951355,
      "num_tokens": 42562267.0,
      "step": 21420
    },
    {
      "entropy": 0.7638312131166458,
      "epoch": 0.38496429694166256,
      "grad_norm": 7.28125,
      "learning_rate": 3.466190737643487e-05,
      "loss": 0.7362929344177246,
      "mean_token_accuracy": 0.8181168526411057,
      "num_tokens": 42582569.0,
      "step": 21430
    },
    {
      "entropy": 0.7458190105855464,
      "epoch": 0.3851439349710334,
      "grad_norm": 7.3125,
      "learning_rate": 3.464865539383333e-05,
      "loss": 0.8107745170593261,
      "mean_token_accuracy": 0.8001433417201043,
      "num_tokens": 42603822.0,
      "step": 21440
    },
    {
      "entropy": 0.740352027118206,
      "epoch": 0.3853235730004042,
      "grad_norm": 7.09375,
      "learning_rate": 3.4635400224832784e-05,
      "loss": 0.7381250381469726,
      "mean_token_accuracy": 0.8173072606325149,
      "num_tokens": 42623256.0,
      "step": 21450
    },
    {
      "entropy": 0.7283120695501566,
      "epoch": 0.38550321102977503,
      "grad_norm": 5.8125,
      "learning_rate": 3.462214187381065e-05,
      "loss": 0.7106196403503418,
      "mean_token_accuracy": 0.8298001632094383,
      "num_tokens": 42643838.0,
      "step": 21460
    },
    {
      "entropy": 0.6796266309916973,
      "epoch": 0.3856828490591458,
      "grad_norm": 5.9375,
      "learning_rate": 3.460888034514542e-05,
      "loss": 0.6561205387115479,
      "mean_token_accuracy": 0.8299140691757202,
      "num_tokens": 42663244.0,
      "step": 21470
    },
    {
      "entropy": 0.6110217899084092,
      "epoch": 0.3858624870885166,
      "grad_norm": 6.375,
      "learning_rate": 3.459561564321661e-05,
      "loss": 0.6209378242492676,
      "mean_token_accuracy": 0.8427578493952751,
      "num_tokens": 42682135.0,
      "step": 21480
    },
    {
      "entropy": 0.7800540171563626,
      "epoch": 0.38604212511788744,
      "grad_norm": 5.59375,
      "learning_rate": 3.458234777240479e-05,
      "loss": 0.8154523849487305,
      "mean_token_accuracy": 0.8081024348735809,
      "num_tokens": 42702257.0,
      "step": 21490
    },
    {
      "entropy": 0.6480796031653882,
      "epoch": 0.38622176314725826,
      "grad_norm": 6.78125,
      "learning_rate": 3.456907673709158e-05,
      "loss": 0.6605803012847901,
      "mean_token_accuracy": 0.8390386402606964,
      "num_tokens": 42723596.0,
      "step": 21500
    },
    {
      "entropy": 0.6994588017463684,
      "epoch": 0.3864014011766291,
      "grad_norm": 5.15625,
      "learning_rate": 3.4555802541659644e-05,
      "loss": 0.6819421291351319,
      "mean_token_accuracy": 0.8299408167600631,
      "num_tokens": 42743880.0,
      "step": 21510
    },
    {
      "entropy": 0.6927497975528241,
      "epoch": 0.3865810392059999,
      "grad_norm": 7.28125,
      "learning_rate": 3.45425251904927e-05,
      "loss": 0.6864865779876709,
      "mean_token_accuracy": 0.8331621691584588,
      "num_tokens": 42763940.0,
      "step": 21520
    },
    {
      "entropy": 0.6850131340324879,
      "epoch": 0.3867606772353707,
      "grad_norm": 7.375,
      "learning_rate": 3.452924468797548e-05,
      "loss": 0.6442245960235595,
      "mean_token_accuracy": 0.8404936403036117,
      "num_tokens": 42784356.0,
      "step": 21530
    },
    {
      "entropy": 0.7827148750424385,
      "epoch": 0.38694031526474154,
      "grad_norm": 8.1875,
      "learning_rate": 3.451596103849379e-05,
      "loss": 0.7670826435089111,
      "mean_token_accuracy": 0.8089598819613457,
      "num_tokens": 42804281.0,
      "step": 21540
    },
    {
      "entropy": 0.7187535740435124,
      "epoch": 0.38711995329411236,
      "grad_norm": 6.6875,
      "learning_rate": 3.4502674246434444e-05,
      "loss": 0.7305371761322021,
      "mean_token_accuracy": 0.8233100593090057,
      "num_tokens": 42824583.0,
      "step": 21550
    },
    {
      "entropy": 0.718814817070961,
      "epoch": 0.3872995913234832,
      "grad_norm": 6.5,
      "learning_rate": 3.448938431618532e-05,
      "loss": 0.771711778640747,
      "mean_token_accuracy": 0.8186303213238716,
      "num_tokens": 42844602.0,
      "step": 21560
    },
    {
      "entropy": 0.71777877882123,
      "epoch": 0.387479229352854,
      "grad_norm": 4.25,
      "learning_rate": 3.4476091252135324e-05,
      "loss": 0.7546377182006836,
      "mean_token_accuracy": 0.8187700524926186,
      "num_tokens": 42865399.0,
      "step": 21570
    },
    {
      "entropy": 0.756792512908578,
      "epoch": 0.3876588673822248,
      "grad_norm": 8.875,
      "learning_rate": 3.446279505867438e-05,
      "loss": 0.7395129203796387,
      "mean_token_accuracy": 0.8236748024821281,
      "num_tokens": 42885154.0,
      "step": 21580
    },
    {
      "entropy": 0.6950151212513447,
      "epoch": 0.38783850541159565,
      "grad_norm": 7.84375,
      "learning_rate": 3.4449495740193486e-05,
      "loss": 0.6975533962249756,
      "mean_token_accuracy": 0.8226354211568833,
      "num_tokens": 42904441.0,
      "step": 21590
    },
    {
      "entropy": 0.7311376191675663,
      "epoch": 0.38801814344096647,
      "grad_norm": 7.34375,
      "learning_rate": 3.443619330108462e-05,
      "loss": 0.7514359951019287,
      "mean_token_accuracy": 0.8153172314167023,
      "num_tokens": 42923706.0,
      "step": 21600
    },
    {
      "entropy": 0.6920048259198666,
      "epoch": 0.3881977814703373,
      "grad_norm": 5.40625,
      "learning_rate": 3.442288774574083e-05,
      "loss": 0.6697358131408692,
      "mean_token_accuracy": 0.8313635379076004,
      "num_tokens": 42943242.0,
      "step": 21610
    },
    {
      "entropy": 0.7440321501344442,
      "epoch": 0.3883774194997081,
      "grad_norm": 9.6875,
      "learning_rate": 3.440957907855617e-05,
      "loss": 0.7557305335998535,
      "mean_token_accuracy": 0.8166138425469398,
      "num_tokens": 42962456.0,
      "step": 21620
    },
    {
      "entropy": 0.6744061037898064,
      "epoch": 0.38855705752907893,
      "grad_norm": 9.5,
      "learning_rate": 3.439626730392575e-05,
      "loss": 0.6839993953704834,
      "mean_token_accuracy": 0.8325913473963737,
      "num_tokens": 42981771.0,
      "step": 21630
    },
    {
      "entropy": 0.6937509968876838,
      "epoch": 0.3887366955584497,
      "grad_norm": 5.125,
      "learning_rate": 3.438295242624567e-05,
      "loss": 0.6781824588775635,
      "mean_token_accuracy": 0.8350390091538429,
      "num_tokens": 43001487.0,
      "step": 21640
    },
    {
      "entropy": 0.7251948766410351,
      "epoch": 0.3889163335878205,
      "grad_norm": 4.46875,
      "learning_rate": 3.436963444991309e-05,
      "loss": 0.6926638603210449,
      "mean_token_accuracy": 0.8267672210931778,
      "num_tokens": 43022761.0,
      "step": 21650
    },
    {
      "entropy": 0.6618501082062721,
      "epoch": 0.38909597161719134,
      "grad_norm": 6.3125,
      "learning_rate": 3.435631337932615e-05,
      "loss": 0.6908493995666504,
      "mean_token_accuracy": 0.8245184361934662,
      "num_tokens": 43042096.0,
      "step": 21660
    },
    {
      "entropy": 0.6265833333134652,
      "epoch": 0.38927560964656216,
      "grad_norm": 5.125,
      "learning_rate": 3.434298921888406e-05,
      "loss": 0.662623929977417,
      "mean_token_accuracy": 0.8327146396040916,
      "num_tokens": 43061869.0,
      "step": 21670
    },
    {
      "entropy": 0.7083549693226814,
      "epoch": 0.389455247675933,
      "grad_norm": 5.65625,
      "learning_rate": 3.432966197298702e-05,
      "loss": 0.7210670471191406,
      "mean_token_accuracy": 0.8178113684058189,
      "num_tokens": 43081676.0,
      "step": 21680
    },
    {
      "entropy": 0.7019871361553669,
      "epoch": 0.3896348857053038,
      "grad_norm": 7.21875,
      "learning_rate": 3.431633164603626e-05,
      "loss": 0.6966697216033936,
      "mean_token_accuracy": 0.8228415340185166,
      "num_tokens": 43101399.0,
      "step": 21690
    },
    {
      "entropy": 0.6654799848794937,
      "epoch": 0.3898145237346746,
      "grad_norm": 6.8125,
      "learning_rate": 3.430299824243402e-05,
      "loss": 0.6256399154663086,
      "mean_token_accuracy": 0.8519033700227737,
      "num_tokens": 43121284.0,
      "step": 21700
    },
    {
      "entropy": 0.692875400185585,
      "epoch": 0.38999416176404544,
      "grad_norm": 6.0,
      "learning_rate": 3.428966176658356e-05,
      "loss": 0.7363029479980469,
      "mean_token_accuracy": 0.8216084897518158,
      "num_tokens": 43141299.0,
      "step": 21710
    },
    {
      "entropy": 0.6670462653040886,
      "epoch": 0.39017379979341626,
      "grad_norm": 8.6875,
      "learning_rate": 3.427632222288917e-05,
      "loss": 0.7002129554748535,
      "mean_token_accuracy": 0.8275974094867706,
      "num_tokens": 43162254.0,
      "step": 21720
    },
    {
      "entropy": 0.687794029712677,
      "epoch": 0.3903534378227871,
      "grad_norm": 5.28125,
      "learning_rate": 3.4262979615756106e-05,
      "loss": 0.743415117263794,
      "mean_token_accuracy": 0.8172914445400238,
      "num_tokens": 43183365.0,
      "step": 21730
    },
    {
      "entropy": 0.6906866550445556,
      "epoch": 0.3905330758521579,
      "grad_norm": 6.375,
      "learning_rate": 3.42496339495907e-05,
      "loss": 0.7134065628051758,
      "mean_token_accuracy": 0.8251887768507004,
      "num_tokens": 43202230.0,
      "step": 21740
    },
    {
      "entropy": 0.6972950875759125,
      "epoch": 0.39071271388152873,
      "grad_norm": 6.6875,
      "learning_rate": 3.423628522880026e-05,
      "loss": 0.673663854598999,
      "mean_token_accuracy": 0.8404061913490295,
      "num_tokens": 43221269.0,
      "step": 21750
    },
    {
      "entropy": 0.6280014872550964,
      "epoch": 0.39089235191089955,
      "grad_norm": 7.375,
      "learning_rate": 3.422293345779309e-05,
      "loss": 0.6238652229309082,
      "mean_token_accuracy": 0.851250422000885,
      "num_tokens": 43240875.0,
      "step": 21760
    },
    {
      "entropy": 0.6681252829730511,
      "epoch": 0.39107198994027037,
      "grad_norm": 5.53125,
      "learning_rate": 3.4209578640978513e-05,
      "loss": 0.6600646495819091,
      "mean_token_accuracy": 0.8339387580752373,
      "num_tokens": 43261344.0,
      "step": 21770
    },
    {
      "entropy": 0.7529622189700603,
      "epoch": 0.3912516279696412,
      "grad_norm": 4.5,
      "learning_rate": 3.4196220782766875e-05,
      "loss": 0.7635057449340821,
      "mean_token_accuracy": 0.820761026442051,
      "num_tokens": 43282661.0,
      "step": 21780
    },
    {
      "entropy": 0.6740679532289505,
      "epoch": 0.391431265999012,
      "grad_norm": 8.1875,
      "learning_rate": 3.4182859887569514e-05,
      "loss": 0.7343880653381347,
      "mean_token_accuracy": 0.8247687846422196,
      "num_tokens": 43301423.0,
      "step": 21790
    },
    {
      "entropy": 0.6474896594882011,
      "epoch": 0.39161090402838283,
      "grad_norm": 5.15625,
      "learning_rate": 3.416949595979876e-05,
      "loss": 0.6172186851501464,
      "mean_token_accuracy": 0.8379372462630272,
      "num_tokens": 43321050.0,
      "step": 21800
    },
    {
      "entropy": 0.7663525573909282,
      "epoch": 0.3917905420577536,
      "grad_norm": 5.125,
      "learning_rate": 3.415612900386795e-05,
      "loss": 0.7770637512207031,
      "mean_token_accuracy": 0.8147229582071305,
      "num_tokens": 43342756.0,
      "step": 21810
    },
    {
      "entropy": 0.6903288178145885,
      "epoch": 0.3919701800871244,
      "grad_norm": 5.8125,
      "learning_rate": 3.414275902419144e-05,
      "loss": 0.6720218658447266,
      "mean_token_accuracy": 0.8323619931936264,
      "num_tokens": 43362577.0,
      "step": 21820
    },
    {
      "entropy": 0.671393821388483,
      "epoch": 0.39214981811649524,
      "grad_norm": 5.03125,
      "learning_rate": 3.4129386025184553e-05,
      "loss": 0.6806766986846924,
      "mean_token_accuracy": 0.8317648664116859,
      "num_tokens": 43382636.0,
      "step": 21830
    },
    {
      "entropy": 0.6655507545918227,
      "epoch": 0.39232945614586606,
      "grad_norm": 4.21875,
      "learning_rate": 3.4116010011263645e-05,
      "loss": 0.6139698505401612,
      "mean_token_accuracy": 0.8451815262436867,
      "num_tokens": 43402011.0,
      "step": 21840
    },
    {
      "entropy": 0.6672481700778008,
      "epoch": 0.3925090941752369,
      "grad_norm": 4.53125,
      "learning_rate": 3.410263098684603e-05,
      "loss": 0.6430671691894532,
      "mean_token_accuracy": 0.8357481941580772,
      "num_tokens": 43421257.0,
      "step": 21850
    },
    {
      "entropy": 0.5748035836964845,
      "epoch": 0.3926887322046077,
      "grad_norm": 4.875,
      "learning_rate": 3.408924895635004e-05,
      "loss": 0.6040756225585937,
      "mean_token_accuracy": 0.8473412156105041,
      "num_tokens": 43441235.0,
      "step": 21860
    },
    {
      "entropy": 0.6708726942539215,
      "epoch": 0.3928683702339785,
      "grad_norm": 6.28125,
      "learning_rate": 3.407586392419501e-05,
      "loss": 0.6487235069274903,
      "mean_token_accuracy": 0.8345163777470589,
      "num_tokens": 43460822.0,
      "step": 21870
    },
    {
      "entropy": 0.6514489538967609,
      "epoch": 0.39304800826334935,
      "grad_norm": 5.1875,
      "learning_rate": 3.406247589480122e-05,
      "loss": 0.6552809238433838,
      "mean_token_accuracy": 0.8357023641467094,
      "num_tokens": 43479995.0,
      "step": 21880
    },
    {
      "entropy": 0.6578537121415138,
      "epoch": 0.39322764629272017,
      "grad_norm": 5.4375,
      "learning_rate": 3.404908487259e-05,
      "loss": 0.6744859695434571,
      "mean_token_accuracy": 0.8255391091108322,
      "num_tokens": 43500845.0,
      "step": 21890
    },
    {
      "entropy": 0.6408430073410273,
      "epoch": 0.393407284322091,
      "grad_norm": 7.125,
      "learning_rate": 3.403569086198362e-05,
      "loss": 0.6577876567840576,
      "mean_token_accuracy": 0.8369144916534423,
      "num_tokens": 43520227.0,
      "step": 21900
    },
    {
      "entropy": 0.6405914753675461,
      "epoch": 0.3935869223514618,
      "grad_norm": 4.46875,
      "learning_rate": 3.4022293867405366e-05,
      "loss": 0.6383644580841065,
      "mean_token_accuracy": 0.8361352548003197,
      "num_tokens": 43541165.0,
      "step": 21910
    },
    {
      "entropy": 0.7254486445337534,
      "epoch": 0.39376656038083263,
      "grad_norm": 6.53125,
      "learning_rate": 3.400889389327948e-05,
      "loss": 0.7279854774475097,
      "mean_token_accuracy": 0.828614880144596,
      "num_tokens": 43560236.0,
      "step": 21920
    },
    {
      "entropy": 0.6784035690128803,
      "epoch": 0.39394619841020345,
      "grad_norm": 6.875,
      "learning_rate": 3.399549094403124e-05,
      "loss": 0.7065106391906738,
      "mean_token_accuracy": 0.8331379547715188,
      "num_tokens": 43579568.0,
      "step": 21930
    },
    {
      "entropy": 0.6221677821129561,
      "epoch": 0.3941258364395743,
      "grad_norm": 6.90625,
      "learning_rate": 3.3982085024086846e-05,
      "loss": 0.6614912509918213,
      "mean_token_accuracy": 0.8389544501900673,
      "num_tokens": 43598805.0,
      "step": 21940
    },
    {
      "entropy": 0.7338356442749501,
      "epoch": 0.3943054744689451,
      "grad_norm": 6.65625,
      "learning_rate": 3.396867613787352e-05,
      "loss": 0.7375948429107666,
      "mean_token_accuracy": 0.8159846365451813,
      "num_tokens": 43619071.0,
      "step": 21950
    },
    {
      "entropy": 0.7576738998293877,
      "epoch": 0.3944851124983159,
      "grad_norm": 6.40625,
      "learning_rate": 3.395526428981944e-05,
      "loss": 0.7416477680206299,
      "mean_token_accuracy": 0.8244003638625145,
      "num_tokens": 43640629.0,
      "step": 21960
    },
    {
      "entropy": 0.6696250911802053,
      "epoch": 0.39466475052768674,
      "grad_norm": 5.96875,
      "learning_rate": 3.394184948435379e-05,
      "loss": 0.6080710411071777,
      "mean_token_accuracy": 0.8495638355612755,
      "num_tokens": 43659640.0,
      "step": 21970
    },
    {
      "entropy": 0.5898547060787678,
      "epoch": 0.3948443885570575,
      "grad_norm": 9.5,
      "learning_rate": 3.392843172590669e-05,
      "loss": 0.5898168563842774,
      "mean_token_accuracy": 0.8504234462976455,
      "num_tokens": 43679856.0,
      "step": 21980
    },
    {
      "entropy": 0.7593280732631683,
      "epoch": 0.3950240265864283,
      "grad_norm": 6.15625,
      "learning_rate": 3.391501101890927e-05,
      "loss": 0.8135687828063964,
      "mean_token_accuracy": 0.8096916928887368,
      "num_tokens": 43698471.0,
      "step": 21990
    },
    {
      "epoch": 0.39520366461579914,
      "eval_entropy": 0.6558015918135643,
      "eval_loss": 0.673142671585083,
      "eval_mean_token_accuracy": 0.8351366941928864,
      "eval_num_tokens": 43719262.0,
      "eval_runtime": 40.4837,
      "eval_samples_per_second": 49.403,
      "eval_steps_per_second": 6.175,
      "step": 22000
    },
    {
      "entropy": 0.7241968221962451,
      "epoch": 0.39538330264516997,
      "grad_norm": 5.84375,
      "learning_rate": 3.388816077699281e-05,
      "loss": 0.7623065948486328,
      "mean_token_accuracy": 0.819486765563488,
      "num_tokens": 43739976.0,
      "step": 22010
    },
    {
      "entropy": 0.6493478201329708,
      "epoch": 0.3955629406745408,
      "grad_norm": 5.46875,
      "learning_rate": 3.387473125094087e-05,
      "loss": 0.6746752738952637,
      "mean_token_accuracy": 0.8314756393432617,
      "num_tokens": 43760308.0,
      "step": 22020
    },
    {
      "entropy": 0.6624151483178139,
      "epoch": 0.3957425787039116,
      "grad_norm": 4.8125,
      "learning_rate": 3.386129879407281e-05,
      "loss": 0.671884822845459,
      "mean_token_accuracy": 0.8360084101557732,
      "num_tokens": 43779747.0,
      "step": 22030
    },
    {
      "entropy": 0.6607727073132992,
      "epoch": 0.39592221673328243,
      "grad_norm": 5.5,
      "learning_rate": 3.384786341082459e-05,
      "loss": 0.645862627029419,
      "mean_token_accuracy": 0.8417955219745636,
      "num_tokens": 43799610.0,
      "step": 22040
    },
    {
      "entropy": 0.7024108413606882,
      "epoch": 0.39610185476265325,
      "grad_norm": 5.28125,
      "learning_rate": 3.383442510563316e-05,
      "loss": 0.6677959442138672,
      "mean_token_accuracy": 0.8274481117725372,
      "num_tokens": 43820592.0,
      "step": 22050
    },
    {
      "entropy": 0.6819233074784279,
      "epoch": 0.39628149279202407,
      "grad_norm": 4.65625,
      "learning_rate": 3.382098388293642e-05,
      "loss": 0.7044548034667969,
      "mean_token_accuracy": 0.82572370916605,
      "num_tokens": 43840837.0,
      "step": 22060
    },
    {
      "entropy": 0.6408405773341656,
      "epoch": 0.3964611308213949,
      "grad_norm": 5.28125,
      "learning_rate": 3.3807539747173236e-05,
      "loss": 0.623042631149292,
      "mean_token_accuracy": 0.8414137437939644,
      "num_tokens": 43860241.0,
      "step": 22070
    },
    {
      "entropy": 0.6417080465704202,
      "epoch": 0.3966407688507657,
      "grad_norm": 5.625,
      "learning_rate": 3.3794092702783426e-05,
      "loss": 0.6433993339538574,
      "mean_token_accuracy": 0.8341434553265572,
      "num_tokens": 43879191.0,
      "step": 22080
    },
    {
      "entropy": 0.6420730002224445,
      "epoch": 0.39682040688013653,
      "grad_norm": 6.25,
      "learning_rate": 3.37806427542078e-05,
      "loss": 0.6256831645965576,
      "mean_token_accuracy": 0.8470543757081032,
      "num_tokens": 43898693.0,
      "step": 22090
    },
    {
      "entropy": 0.6462625682353973,
      "epoch": 0.39700004490950735,
      "grad_norm": 7.40625,
      "learning_rate": 3.37671899058881e-05,
      "loss": 0.6173986911773681,
      "mean_token_accuracy": 0.8404306560754776,
      "num_tokens": 43919459.0,
      "step": 22100
    },
    {
      "entropy": 0.7910645864903927,
      "epoch": 0.3971796829388782,
      "grad_norm": 8.375,
      "learning_rate": 3.3753734162267025e-05,
      "loss": 0.7814685821533203,
      "mean_token_accuracy": 0.8104552209377289,
      "num_tokens": 43938663.0,
      "step": 22110
    },
    {
      "entropy": 0.6860839754343033,
      "epoch": 0.397359320968249,
      "grad_norm": 7.125,
      "learning_rate": 3.3740275527788247e-05,
      "loss": 0.7195610523223877,
      "mean_token_accuracy": 0.8275859281420708,
      "num_tokens": 43958629.0,
      "step": 22120
    },
    {
      "entropy": 0.6875795148313045,
      "epoch": 0.3975389589976198,
      "grad_norm": 5.84375,
      "learning_rate": 3.3726814006896375e-05,
      "loss": 0.7132916450500488,
      "mean_token_accuracy": 0.8207760348916053,
      "num_tokens": 43978071.0,
      "step": 22130
    },
    {
      "entropy": 0.6830442637205124,
      "epoch": 0.39771859702699064,
      "grad_norm": 8.0,
      "learning_rate": 3.3713349604036984e-05,
      "loss": 0.7069227695465088,
      "mean_token_accuracy": 0.8318000227212906,
      "num_tokens": 43997192.0,
      "step": 22140
    },
    {
      "entropy": 0.6524615310132503,
      "epoch": 0.3978982350563614,
      "grad_norm": 6.25,
      "learning_rate": 3.36998823236566e-05,
      "loss": 0.6561416149139404,
      "mean_token_accuracy": 0.8385264322161674,
      "num_tokens": 44016078.0,
      "step": 22150
    },
    {
      "entropy": 0.612352479249239,
      "epoch": 0.3980778730857322,
      "grad_norm": 4.78125,
      "learning_rate": 3.368641217020268e-05,
      "loss": 0.6088404178619384,
      "mean_token_accuracy": 0.8454895094037056,
      "num_tokens": 44035660.0,
      "step": 22160
    },
    {
      "entropy": 0.6279764913022519,
      "epoch": 0.39825751111510305,
      "grad_norm": 5.75,
      "learning_rate": 3.3672939148123654e-05,
      "loss": 0.6485654354095459,
      "mean_token_accuracy": 0.8316551372408867,
      "num_tokens": 44056453.0,
      "step": 22170
    },
    {
      "entropy": 0.6700321987271309,
      "epoch": 0.39843714914447387,
      "grad_norm": 8.3125,
      "learning_rate": 3.3659463261868907e-05,
      "loss": 0.6924183368682861,
      "mean_token_accuracy": 0.823557086288929,
      "num_tokens": 44076029.0,
      "step": 22180
    },
    {
      "entropy": 0.6503889881074428,
      "epoch": 0.3986167871738447,
      "grad_norm": 6.90625,
      "learning_rate": 3.364598451588873e-05,
      "loss": 0.664320707321167,
      "mean_token_accuracy": 0.8362871915102005,
      "num_tokens": 44094977.0,
      "step": 22190
    },
    {
      "entropy": 0.6858331218361855,
      "epoch": 0.3987964252032155,
      "grad_norm": 6.03125,
      "learning_rate": 3.3632502914634395e-05,
      "loss": 0.683863353729248,
      "mean_token_accuracy": 0.8401730537414551,
      "num_tokens": 44115534.0,
      "step": 22200
    },
    {
      "entropy": 0.6879073277115821,
      "epoch": 0.39897606323258633,
      "grad_norm": 7.875,
      "learning_rate": 3.361901846255809e-05,
      "loss": 0.7091107845306397,
      "mean_token_accuracy": 0.8244135081768036,
      "num_tokens": 44135521.0,
      "step": 22210
    },
    {
      "entropy": 0.7438757218420505,
      "epoch": 0.39915570126195715,
      "grad_norm": 6.1875,
      "learning_rate": 3.360553116411298e-05,
      "loss": 0.7478530406951904,
      "mean_token_accuracy": 0.8174891158938408,
      "num_tokens": 44155326.0,
      "step": 22220
    },
    {
      "entropy": 0.7605732206255198,
      "epoch": 0.399335339291328,
      "grad_norm": 5.21875,
      "learning_rate": 3.3592041023753126e-05,
      "loss": 0.8048529624938965,
      "mean_token_accuracy": 0.8003992080688477,
      "num_tokens": 44174855.0,
      "step": 22230
    },
    {
      "entropy": 0.6924356527626514,
      "epoch": 0.3995149773206988,
      "grad_norm": 4.8125,
      "learning_rate": 3.3578548045933575e-05,
      "loss": 0.6906275749206543,
      "mean_token_accuracy": 0.8295157432556153,
      "num_tokens": 44195500.0,
      "step": 22240
    },
    {
      "entropy": 0.6545244716107845,
      "epoch": 0.3996946153500696,
      "grad_norm": 4.53125,
      "learning_rate": 3.3565052235110254e-05,
      "loss": 0.6450164318084717,
      "mean_token_accuracy": 0.8418580979108811,
      "num_tokens": 44216116.0,
      "step": 22250
    },
    {
      "entropy": 0.6838923864066601,
      "epoch": 0.39987425337944044,
      "grad_norm": 6.25,
      "learning_rate": 3.355155359574008e-05,
      "loss": 0.6862600803375244,
      "mean_token_accuracy": 0.8321245431900024,
      "num_tokens": 44235391.0,
      "step": 22260
    },
    {
      "entropy": 0.687882962077856,
      "epoch": 0.40005389140881126,
      "grad_norm": 5.59375,
      "learning_rate": 3.3538052132280877e-05,
      "loss": 0.6652538299560546,
      "mean_token_accuracy": 0.8352219730615615,
      "num_tokens": 44256196.0,
      "step": 22270
    },
    {
      "entropy": 0.6697224445641041,
      "epoch": 0.4002335294381821,
      "grad_norm": 6.15625,
      "learning_rate": 3.35245478491914e-05,
      "loss": 0.7327306270599365,
      "mean_token_accuracy": 0.8231875389814377,
      "num_tokens": 44275319.0,
      "step": 22280
    },
    {
      "entropy": 0.6583179734647274,
      "epoch": 0.4004131674675529,
      "grad_norm": 4.4375,
      "learning_rate": 3.3511040750931355e-05,
      "loss": 0.6128525733947754,
      "mean_token_accuracy": 0.8432210311293602,
      "num_tokens": 44295317.0,
      "step": 22290
    },
    {
      "entropy": 0.6696611866354942,
      "epoch": 0.4005928054969237,
      "grad_norm": 6.0,
      "learning_rate": 3.349753084196135e-05,
      "loss": 0.6502727508544922,
      "mean_token_accuracy": 0.8459016531705856,
      "num_tokens": 44315549.0,
      "step": 22300
    },
    {
      "entropy": 0.6934429407119751,
      "epoch": 0.40077244352629454,
      "grad_norm": 5.1875,
      "learning_rate": 3.348401812674294e-05,
      "loss": 0.6869980812072753,
      "mean_token_accuracy": 0.8297330781817436,
      "num_tokens": 44334589.0,
      "step": 22310
    },
    {
      "entropy": 0.5777344413101673,
      "epoch": 0.4009520815556653,
      "grad_norm": 5.0,
      "learning_rate": 3.3470502609738595e-05,
      "loss": 0.5970641613006592,
      "mean_token_accuracy": 0.8529870539903641,
      "num_tokens": 44355100.0,
      "step": 22320
    },
    {
      "entropy": 0.5983030380681157,
      "epoch": 0.40113171958503613,
      "grad_norm": 5.0,
      "learning_rate": 3.345698429541174e-05,
      "loss": 0.6491535186767579,
      "mean_token_accuracy": 0.8428195133805275,
      "num_tokens": 44376101.0,
      "step": 22330
    },
    {
      "entropy": 0.6968772262334824,
      "epoch": 0.40131135761440695,
      "grad_norm": 5.6875,
      "learning_rate": 3.344346318822668e-05,
      "loss": 0.7004656314849853,
      "mean_token_accuracy": 0.8284920811653137,
      "num_tokens": 44395549.0,
      "step": 22340
    },
    {
      "entropy": 0.7113131783902645,
      "epoch": 0.40149099564377777,
      "grad_norm": 5.5,
      "learning_rate": 3.342993929264866e-05,
      "loss": 0.7103156566619873,
      "mean_token_accuracy": 0.8222610637545585,
      "num_tokens": 44414808.0,
      "step": 22350
    },
    {
      "entropy": 0.6783322386443615,
      "epoch": 0.4016706336731486,
      "grad_norm": 6.5625,
      "learning_rate": 3.341641261314386e-05,
      "loss": 0.7103092193603515,
      "mean_token_accuracy": 0.8326554521918297,
      "num_tokens": 44433371.0,
      "step": 22360
    },
    {
      "entropy": 0.698417105525732,
      "epoch": 0.4018502717025194,
      "grad_norm": 6.875,
      "learning_rate": 3.340288315417937e-05,
      "loss": 0.6647182941436768,
      "mean_token_accuracy": 0.8376531884074211,
      "num_tokens": 44453164.0,
      "step": 22370
    },
    {
      "entropy": 0.6952787030488252,
      "epoch": 0.40202990973189023,
      "grad_norm": 4.71875,
      "learning_rate": 3.338935092022319e-05,
      "loss": 0.7135073661804199,
      "mean_token_accuracy": 0.8245546475052834,
      "num_tokens": 44474193.0,
      "step": 22380
    },
    {
      "entropy": 0.805926077067852,
      "epoch": 0.40220954776126105,
      "grad_norm": 5.53125,
      "learning_rate": 3.3375815915744254e-05,
      "loss": 0.8032584190368652,
      "mean_token_accuracy": 0.8070807561278344,
      "num_tokens": 44494379.0,
      "step": 22390
    },
    {
      "entropy": 0.6957412242889405,
      "epoch": 0.4023891857906319,
      "grad_norm": 5.84375,
      "learning_rate": 3.336227814521239e-05,
      "loss": 0.7394069194793701,
      "mean_token_accuracy": 0.816690368950367,
      "num_tokens": 44513991.0,
      "step": 22400
    },
    {
      "entropy": 0.7332220442593098,
      "epoch": 0.4025688238200027,
      "grad_norm": 7.1875,
      "learning_rate": 3.334873761309834e-05,
      "loss": 0.7164353370666504,
      "mean_token_accuracy": 0.8268209457397461,
      "num_tokens": 44534466.0,
      "step": 22410
    },
    {
      "entropy": 0.7292749412357807,
      "epoch": 0.4027484618493735,
      "grad_norm": 5.65625,
      "learning_rate": 3.3335194323873784e-05,
      "loss": 0.7544412612915039,
      "mean_token_accuracy": 0.8137175813317299,
      "num_tokens": 44554688.0,
      "step": 22420
    },
    {
      "entropy": 0.8354973644018173,
      "epoch": 0.40292809987874434,
      "grad_norm": 7.59375,
      "learning_rate": 3.332164828201129e-05,
      "loss": 0.8762054443359375,
      "mean_token_accuracy": 0.8022147446870804,
      "num_tokens": 44574184.0,
      "step": 22430
    },
    {
      "entropy": 0.7801049770787358,
      "epoch": 0.40310773790811516,
      "grad_norm": 5.5,
      "learning_rate": 3.330809949198435e-05,
      "loss": 0.7766165256500244,
      "mean_token_accuracy": 0.81794403642416,
      "num_tokens": 44593140.0,
      "step": 22440
    },
    {
      "entropy": 0.6667418234050274,
      "epoch": 0.403287375937486,
      "grad_norm": 4.03125,
      "learning_rate": 3.329454795826733e-05,
      "loss": 0.7061066627502441,
      "mean_token_accuracy": 0.8328410059213638,
      "num_tokens": 44612913.0,
      "step": 22450
    },
    {
      "entropy": 0.7023803547024727,
      "epoch": 0.4034670139668568,
      "grad_norm": 6.25,
      "learning_rate": 3.3280993685335546e-05,
      "loss": 0.6761770725250245,
      "mean_token_accuracy": 0.8256873741745949,
      "num_tokens": 44633349.0,
      "step": 22460
    },
    {
      "entropy": 0.6428582966327667,
      "epoch": 0.4036466519962276,
      "grad_norm": 5.59375,
      "learning_rate": 3.3267436677665204e-05,
      "loss": 0.6783720970153808,
      "mean_token_accuracy": 0.8358080565929413,
      "num_tokens": 44652771.0,
      "step": 22470
    },
    {
      "entropy": 0.6104120582342147,
      "epoch": 0.40382629002559844,
      "grad_norm": 6.90625,
      "learning_rate": 3.325387693973338e-05,
      "loss": 0.568636703491211,
      "mean_token_accuracy": 0.851903049647808,
      "num_tokens": 44670926.0,
      "step": 22480
    },
    {
      "entropy": 0.6643792770802974,
      "epoch": 0.4040059280549692,
      "grad_norm": 8.0625,
      "learning_rate": 3.324031447601812e-05,
      "loss": 0.6976274013519287,
      "mean_token_accuracy": 0.8361451581120491,
      "num_tokens": 44690911.0,
      "step": 22490
    },
    {
      "entropy": 0.6068355873227119,
      "epoch": 0.40418556608434003,
      "grad_norm": 5.5,
      "learning_rate": 3.32267492909983e-05,
      "loss": 0.6128767967224121,
      "mean_token_accuracy": 0.8496202155947685,
      "num_tokens": 44710635.0,
      "step": 22500
    },
    {
      "entropy": 0.7108094185590744,
      "epoch": 0.40436520411371085,
      "grad_norm": 6.375,
      "learning_rate": 3.321318138915373e-05,
      "loss": 0.7066905498504639,
      "mean_token_accuracy": 0.8265850663185119,
      "num_tokens": 44730342.0,
      "step": 22510
    },
    {
      "entropy": 0.6774210140109063,
      "epoch": 0.4045448421430817,
      "grad_norm": 4.8125,
      "learning_rate": 3.319961077496512e-05,
      "loss": 0.7050983428955078,
      "mean_token_accuracy": 0.8301229119300843,
      "num_tokens": 44750413.0,
      "step": 22520
    },
    {
      "entropy": 0.7404642555862665,
      "epoch": 0.4047244801724525,
      "grad_norm": 4.65625,
      "learning_rate": 3.3186037452914065e-05,
      "loss": 0.7653717517852783,
      "mean_token_accuracy": 0.8150559157133103,
      "num_tokens": 44769818.0,
      "step": 22530
    },
    {
      "entropy": 0.6418867006897926,
      "epoch": 0.4049041182018233,
      "grad_norm": 10.3125,
      "learning_rate": 3.317246142748305e-05,
      "loss": 0.6636004447937012,
      "mean_token_accuracy": 0.8357202738523484,
      "num_tokens": 44789614.0,
      "step": 22540
    },
    {
      "entropy": 0.6306140393018722,
      "epoch": 0.40508375623119414,
      "grad_norm": 5.125,
      "learning_rate": 3.315888270315547e-05,
      "loss": 0.6421436309814453,
      "mean_token_accuracy": 0.8493367597460747,
      "num_tokens": 44809124.0,
      "step": 22550
    },
    {
      "entropy": 0.6423268489539623,
      "epoch": 0.40526339426056496,
      "grad_norm": 4.9375,
      "learning_rate": 3.314530128441559e-05,
      "loss": 0.6546293735504151,
      "mean_token_accuracy": 0.8408963426947593,
      "num_tokens": 44828970.0,
      "step": 22560
    },
    {
      "entropy": 0.7043904002755881,
      "epoch": 0.4054430322899358,
      "grad_norm": 5.875,
      "learning_rate": 3.313171717574858e-05,
      "loss": 0.7079294681549072,
      "mean_token_accuracy": 0.832610534131527,
      "num_tokens": 44848759.0,
      "step": 22570
    },
    {
      "entropy": 0.7322651073336601,
      "epoch": 0.4056226703193066,
      "grad_norm": 5.46875,
      "learning_rate": 3.3118130381640506e-05,
      "loss": 0.7366083145141602,
      "mean_token_accuracy": 0.8156790941953659,
      "num_tokens": 44867911.0,
      "step": 22580
    },
    {
      "entropy": 0.6543777473270893,
      "epoch": 0.4058023083486774,
      "grad_norm": 7.15625,
      "learning_rate": 3.310454090657828e-05,
      "loss": 0.6609643459320068,
      "mean_token_accuracy": 0.8365061476826667,
      "num_tokens": 44888307.0,
      "step": 22590
    },
    {
      "entropy": 0.697276008501649,
      "epoch": 0.40598194637804824,
      "grad_norm": 6.5,
      "learning_rate": 3.3090948755049774e-05,
      "loss": 0.7440574645996094,
      "mean_token_accuracy": 0.8174171894788742,
      "num_tokens": 44908101.0,
      "step": 22600
    },
    {
      "entropy": 0.5557753995060921,
      "epoch": 0.40616158440741906,
      "grad_norm": 4.1875,
      "learning_rate": 3.3077353931543655e-05,
      "loss": 0.5601234436035156,
      "mean_token_accuracy": 0.8668026849627495,
      "num_tokens": 44928837.0,
      "step": 22610
    },
    {
      "entropy": 0.709277143701911,
      "epoch": 0.4063412224367899,
      "grad_norm": 7.1875,
      "learning_rate": 3.306375644054954e-05,
      "loss": 0.7436286926269531,
      "mean_token_accuracy": 0.8214550584554672,
      "num_tokens": 44948537.0,
      "step": 22620
    },
    {
      "entropy": 0.6570791564881802,
      "epoch": 0.4065208604661607,
      "grad_norm": 4.4375,
      "learning_rate": 3.3050156286557884e-05,
      "loss": 0.6490310668945313,
      "mean_token_accuracy": 0.8325426861643791,
      "num_tokens": 44968760.0,
      "step": 22630
    },
    {
      "entropy": 0.6308296252042055,
      "epoch": 0.4067004984955315,
      "grad_norm": 5.0625,
      "learning_rate": 3.303655347406005e-05,
      "loss": 0.6484262466430664,
      "mean_token_accuracy": 0.8365836650133133,
      "num_tokens": 44989323.0,
      "step": 22640
    },
    {
      "entropy": 0.6776929520070553,
      "epoch": 0.40688013652490235,
      "grad_norm": 6.0,
      "learning_rate": 3.302294800754828e-05,
      "loss": 0.677354383468628,
      "mean_token_accuracy": 0.8280426368117333,
      "num_tokens": 45009314.0,
      "step": 22650
    },
    {
      "entropy": 0.7251378327608109,
      "epoch": 0.4070597745542731,
      "grad_norm": 4.8125,
      "learning_rate": 3.300933989151566e-05,
      "loss": 0.72687668800354,
      "mean_token_accuracy": 0.8290361434221267,
      "num_tokens": 45029529.0,
      "step": 22660
    },
    {
      "entropy": 0.7583900839090347,
      "epoch": 0.40723941258364393,
      "grad_norm": 5.59375,
      "learning_rate": 3.2995729130456185e-05,
      "loss": 0.8117211341857911,
      "mean_token_accuracy": 0.808470395207405,
      "num_tokens": 45050530.0,
      "step": 22670
    },
    {
      "entropy": 0.7688450172543526,
      "epoch": 0.40741905061301475,
      "grad_norm": 7.40625,
      "learning_rate": 3.2982115728864714e-05,
      "loss": 0.7603821277618408,
      "mean_token_accuracy": 0.8212204411625862,
      "num_tokens": 45070168.0,
      "step": 22680
    },
    {
      "entropy": 0.7614038497209549,
      "epoch": 0.4075986886423856,
      "grad_norm": 8.3125,
      "learning_rate": 3.296849969123696e-05,
      "loss": 0.7804455280303955,
      "mean_token_accuracy": 0.8184756860136986,
      "num_tokens": 45089616.0,
      "step": 22690
    },
    {
      "entropy": 0.6373643200844527,
      "epoch": 0.4077783266717564,
      "grad_norm": 7.40625,
      "learning_rate": 3.295488102206955e-05,
      "loss": 0.6532629489898681,
      "mean_token_accuracy": 0.8422436520457268,
      "num_tokens": 45109399.0,
      "step": 22700
    },
    {
      "entropy": 0.6100007340312004,
      "epoch": 0.4079579647011272,
      "grad_norm": 6.0,
      "learning_rate": 3.294125972585992e-05,
      "loss": 0.5999004364013671,
      "mean_token_accuracy": 0.842499227821827,
      "num_tokens": 45128804.0,
      "step": 22710
    },
    {
      "entropy": 0.6053519256412982,
      "epoch": 0.40813760273049804,
      "grad_norm": 5.53125,
      "learning_rate": 3.292763580710642e-05,
      "loss": 0.6307381629943848,
      "mean_token_accuracy": 0.8362287029623985,
      "num_tokens": 45150081.0,
      "step": 22720
    },
    {
      "entropy": 0.6801414672285319,
      "epoch": 0.40831724075986886,
      "grad_norm": 6.09375,
      "learning_rate": 3.291400927030825e-05,
      "loss": 0.6554404735565186,
      "mean_token_accuracy": 0.832933883368969,
      "num_tokens": 45169850.0,
      "step": 22730
    },
    {
      "entropy": 0.7439732313156128,
      "epoch": 0.4084968787892397,
      "grad_norm": 6.53125,
      "learning_rate": 3.2900380119965466e-05,
      "loss": 0.8080075263977051,
      "mean_token_accuracy": 0.8092029258608818,
      "num_tokens": 45189503.0,
      "step": 22740
    },
    {
      "entropy": 0.6725748501718044,
      "epoch": 0.4086765168186105,
      "grad_norm": 5.59375,
      "learning_rate": 3.288674836057902e-05,
      "loss": 0.6501432418823242,
      "mean_token_accuracy": 0.8382540047168732,
      "num_tokens": 45209829.0,
      "step": 22750
    },
    {
      "entropy": 0.753080402687192,
      "epoch": 0.4088561548479813,
      "grad_norm": 7.3125,
      "learning_rate": 3.2873113996650676e-05,
      "loss": 0.7526428699493408,
      "mean_token_accuracy": 0.8169500544667244,
      "num_tokens": 45228555.0,
      "step": 22760
    },
    {
      "entropy": 0.711393267661333,
      "epoch": 0.40903579287735214,
      "grad_norm": 6.78125,
      "learning_rate": 3.285947703268311e-05,
      "loss": 0.6784804344177247,
      "mean_token_accuracy": 0.838080070912838,
      "num_tokens": 45247888.0,
      "step": 22770
    },
    {
      "entropy": 0.6723963245749474,
      "epoch": 0.40921543090672297,
      "grad_norm": 5.75,
      "learning_rate": 3.2845837473179795e-05,
      "loss": 0.7027187347412109,
      "mean_token_accuracy": 0.833698658645153,
      "num_tokens": 45268022.0,
      "step": 22780
    },
    {
      "entropy": 0.7531605176627636,
      "epoch": 0.4093950689360938,
      "grad_norm": 5.21875,
      "learning_rate": 3.283219532264514e-05,
      "loss": 0.7299594402313232,
      "mean_token_accuracy": 0.8188507795333863,
      "num_tokens": 45289034.0,
      "step": 22790
    },
    {
      "entropy": 0.6526213444769382,
      "epoch": 0.4095747069654646,
      "grad_norm": 5.53125,
      "learning_rate": 3.281855058558434e-05,
      "loss": 0.6303492069244385,
      "mean_token_accuracy": 0.843523895740509,
      "num_tokens": 45308137.0,
      "step": 22800
    },
    {
      "entropy": 0.6062753543257713,
      "epoch": 0.40975434499483543,
      "grad_norm": 5.90625,
      "learning_rate": 3.280490326650348e-05,
      "loss": 0.6138824462890625,
      "mean_token_accuracy": 0.8411237925291062,
      "num_tokens": 45328314.0,
      "step": 22810
    },
    {
      "entropy": 0.6476704373955726,
      "epoch": 0.40993398302420625,
      "grad_norm": 7.34375,
      "learning_rate": 3.279125336990948e-05,
      "loss": 0.6581060409545898,
      "mean_token_accuracy": 0.8376688912510872,
      "num_tokens": 45348715.0,
      "step": 22820
    },
    {
      "entropy": 0.6718829549849034,
      "epoch": 0.410113621053577,
      "grad_norm": 6.09375,
      "learning_rate": 3.2777600900310137e-05,
      "loss": 0.7055031299591065,
      "mean_token_accuracy": 0.833826319873333,
      "num_tokens": 45367995.0,
      "step": 22830
    },
    {
      "entropy": 0.629845866933465,
      "epoch": 0.41029325908294784,
      "grad_norm": 6.375,
      "learning_rate": 3.276394586221407e-05,
      "loss": 0.6474076747894287,
      "mean_token_accuracy": 0.8385332003235817,
      "num_tokens": 45388897.0,
      "step": 22840
    },
    {
      "entropy": 0.7140257030725479,
      "epoch": 0.41047289711231866,
      "grad_norm": 6.75,
      "learning_rate": 3.2750288260130755e-05,
      "loss": 0.7341724872589112,
      "mean_token_accuracy": 0.8218767404556274,
      "num_tokens": 45408007.0,
      "step": 22850
    },
    {
      "entropy": 0.716847138106823,
      "epoch": 0.4106525351416895,
      "grad_norm": 6.90625,
      "learning_rate": 3.273662809857053e-05,
      "loss": 0.7314211368560791,
      "mean_token_accuracy": 0.8170238375663758,
      "num_tokens": 45429022.0,
      "step": 22860
    },
    {
      "entropy": 0.7257310207933187,
      "epoch": 0.4108321731710603,
      "grad_norm": 6.53125,
      "learning_rate": 3.272296538204454e-05,
      "loss": 0.7288665771484375,
      "mean_token_accuracy": 0.8246805012226105,
      "num_tokens": 45449753.0,
      "step": 22870
    },
    {
      "entropy": 0.7496883995831013,
      "epoch": 0.4110118112004311,
      "grad_norm": 6.15625,
      "learning_rate": 3.270930011506482e-05,
      "loss": 0.7539916038513184,
      "mean_token_accuracy": 0.814261469244957,
      "num_tokens": 45470533.0,
      "step": 22880
    },
    {
      "entropy": 0.6454012591391802,
      "epoch": 0.41119144922980194,
      "grad_norm": 4.5625,
      "learning_rate": 3.269563230214422e-05,
      "loss": 0.6186908721923828,
      "mean_token_accuracy": 0.838475202023983,
      "num_tokens": 45490313.0,
      "step": 22890
    },
    {
      "entropy": 0.6866658806800843,
      "epoch": 0.41137108725917276,
      "grad_norm": 4.15625,
      "learning_rate": 3.268196194779644e-05,
      "loss": 0.6965051651000976,
      "mean_token_accuracy": 0.8315447926521301,
      "num_tokens": 45509896.0,
      "step": 22900
    },
    {
      "entropy": 0.6137972973287106,
      "epoch": 0.4115507252885436,
      "grad_norm": 5.46875,
      "learning_rate": 3.2668289056536004e-05,
      "loss": 0.6033967971801758,
      "mean_token_accuracy": 0.8553759783506394,
      "num_tokens": 45530136.0,
      "step": 22910
    },
    {
      "entropy": 0.650493610650301,
      "epoch": 0.4117303633179144,
      "grad_norm": 5.8125,
      "learning_rate": 3.26546136328783e-05,
      "loss": 0.666453742980957,
      "mean_token_accuracy": 0.836171281337738,
      "num_tokens": 45549535.0,
      "step": 22920
    },
    {
      "entropy": 0.733617166429758,
      "epoch": 0.4119100013472852,
      "grad_norm": 8.9375,
      "learning_rate": 3.264093568133954e-05,
      "loss": 0.7418248653411865,
      "mean_token_accuracy": 0.8189740777015686,
      "num_tokens": 45569261.0,
      "step": 22930
    },
    {
      "entropy": 0.6186542198061943,
      "epoch": 0.41208963937665605,
      "grad_norm": 7.46875,
      "learning_rate": 3.2627255206436744e-05,
      "loss": 0.6323249340057373,
      "mean_token_accuracy": 0.8455927580595016,
      "num_tokens": 45588745.0,
      "step": 22940
    },
    {
      "entropy": 0.6665212947875261,
      "epoch": 0.41226927740602687,
      "grad_norm": 6.09375,
      "learning_rate": 3.2613572212687824e-05,
      "loss": 0.6764096260070801,
      "mean_token_accuracy": 0.8316299125552178,
      "num_tokens": 45608895.0,
      "step": 22950
    },
    {
      "entropy": 0.6451776951551438,
      "epoch": 0.4124489154353977,
      "grad_norm": 4.21875,
      "learning_rate": 3.259988670461147e-05,
      "loss": 0.6428620338439941,
      "mean_token_accuracy": 0.8476417437195778,
      "num_tokens": 45628294.0,
      "step": 22960
    },
    {
      "entropy": 0.7484756927937269,
      "epoch": 0.4126285534647685,
      "grad_norm": 6.6875,
      "learning_rate": 3.2586198686727237e-05,
      "loss": 0.7448271751403809,
      "mean_token_accuracy": 0.8054756566882133,
      "num_tokens": 45647797.0,
      "step": 22970
    },
    {
      "entropy": 0.660018665716052,
      "epoch": 0.41280819149413933,
      "grad_norm": 6.0,
      "learning_rate": 3.257250816355547e-05,
      "loss": 0.7036425590515136,
      "mean_token_accuracy": 0.8346416994929313,
      "num_tokens": 45667586.0,
      "step": 22980
    },
    {
      "entropy": 0.6594592332839966,
      "epoch": 0.41298782952351015,
      "grad_norm": 6.25,
      "learning_rate": 3.255881513961739e-05,
      "loss": 0.6726462841033936,
      "mean_token_accuracy": 0.8292291700839997,
      "num_tokens": 45687366.0,
      "step": 22990
    },
    {
      "epoch": 0.4131674675528809,
      "eval_entropy": 0.6609513915777206,
      "eval_loss": 0.6571119427680969,
      "eval_mean_token_accuracy": 0.839023366689682,
      "eval_num_tokens": 45708492.0,
      "eval_runtime": 40.548,
      "eval_samples_per_second": 49.324,
      "eval_steps_per_second": 6.166,
      "step": 23000
    },
    {
      "entropy": 0.6640272188931704,
      "epoch": 0.41334710558225174,
      "grad_norm": 5.4375,
      "learning_rate": 3.2531421607531165e-05,
      "loss": 0.6746596336364746,
      "mean_token_accuracy": 0.8314566947519779,
      "num_tokens": 45728240.0,
      "step": 23010
    },
    {
      "entropy": 0.6499909503385425,
      "epoch": 0.41352674361162256,
      "grad_norm": 7.65625,
      "learning_rate": 3.251772110842955e-05,
      "loss": 0.6314692974090577,
      "mean_token_accuracy": 0.8525784566998482,
      "num_tokens": 45747909.0,
      "step": 23020
    },
    {
      "entropy": 0.7167673274874687,
      "epoch": 0.4137063816409934,
      "grad_norm": 6.9375,
      "learning_rate": 3.2504018126654636e-05,
      "loss": 0.7251205444335938,
      "mean_token_accuracy": 0.8285655096173287,
      "num_tokens": 45766710.0,
      "step": 23030
    },
    {
      "entropy": 0.6369082137942315,
      "epoch": 0.4138860196703642,
      "grad_norm": 5.5625,
      "learning_rate": 3.249031266673174e-05,
      "loss": 0.6498454093933106,
      "mean_token_accuracy": 0.8428271114826202,
      "num_tokens": 45785833.0,
      "step": 23040
    },
    {
      "entropy": 0.6905918989330531,
      "epoch": 0.414065657699735,
      "grad_norm": 5.96875,
      "learning_rate": 3.2476604733186994e-05,
      "loss": 0.7275197982788086,
      "mean_token_accuracy": 0.8224400177597999,
      "num_tokens": 45807177.0,
      "step": 23050
    },
    {
      "entropy": 0.7335435926914216,
      "epoch": 0.41424529572910584,
      "grad_norm": 5.5,
      "learning_rate": 3.246289433054736e-05,
      "loss": 0.7611394405364991,
      "mean_token_accuracy": 0.8114820539951324,
      "num_tokens": 45827820.0,
      "step": 23060
    },
    {
      "entropy": 0.6114605758339167,
      "epoch": 0.41442493375847667,
      "grad_norm": 5.84375,
      "learning_rate": 3.244918146334056e-05,
      "loss": 0.6043206214904785,
      "mean_token_accuracy": 0.848127818107605,
      "num_tokens": 45846895.0,
      "step": 23070
    },
    {
      "entropy": 0.7020625710487366,
      "epoch": 0.4146045717878475,
      "grad_norm": 6.5625,
      "learning_rate": 3.24354661360952e-05,
      "loss": 0.7163649082183838,
      "mean_token_accuracy": 0.8233471050858497,
      "num_tokens": 45866193.0,
      "step": 23080
    },
    {
      "entropy": 0.7127120420336723,
      "epoch": 0.4147842098172183,
      "grad_norm": 5.84375,
      "learning_rate": 3.242174835334066e-05,
      "loss": 0.7266265869140625,
      "mean_token_accuracy": 0.8179249033331871,
      "num_tokens": 45884912.0,
      "step": 23090
    },
    {
      "entropy": 0.7246527075767517,
      "epoch": 0.41496384784658913,
      "grad_norm": 7.90625,
      "learning_rate": 3.240802811960714e-05,
      "loss": 0.7663551330566406,
      "mean_token_accuracy": 0.8252000197768211,
      "num_tokens": 45904113.0,
      "step": 23100
    },
    {
      "entropy": 0.7118993975222111,
      "epoch": 0.41514348587595995,
      "grad_norm": 6.15625,
      "learning_rate": 3.239430543942565e-05,
      "loss": 0.7112181186676025,
      "mean_token_accuracy": 0.8305133000016213,
      "num_tokens": 45925037.0,
      "step": 23110
    },
    {
      "entropy": 0.6859762005507946,
      "epoch": 0.41532312390533077,
      "grad_norm": 4.53125,
      "learning_rate": 3.2380580317328003e-05,
      "loss": 0.6995909690856934,
      "mean_token_accuracy": 0.8377453327178955,
      "num_tokens": 45944901.0,
      "step": 23120
    },
    {
      "entropy": 0.702526556700468,
      "epoch": 0.4155027619347016,
      "grad_norm": 6.34375,
      "learning_rate": 3.236685275784682e-05,
      "loss": 0.6959921360015869,
      "mean_token_accuracy": 0.8270902886986733,
      "num_tokens": 45965654.0,
      "step": 23130
    },
    {
      "entropy": 0.6819639127701521,
      "epoch": 0.4156823999640724,
      "grad_norm": 4.96875,
      "learning_rate": 3.2353122765515534e-05,
      "loss": 0.6668210983276367,
      "mean_token_accuracy": 0.83558998554945,
      "num_tokens": 45985910.0,
      "step": 23140
    },
    {
      "entropy": 0.6115058541297913,
      "epoch": 0.41586203799344323,
      "grad_norm": 7.46875,
      "learning_rate": 3.2339390344868384e-05,
      "loss": 0.6231845855712891,
      "mean_token_accuracy": 0.8413219466805458,
      "num_tokens": 46005543.0,
      "step": 23150
    },
    {
      "entropy": 0.6468209695070982,
      "epoch": 0.41604167602281406,
      "grad_norm": 7.4375,
      "learning_rate": 3.2325655500440385e-05,
      "loss": 0.643020486831665,
      "mean_token_accuracy": 0.8409248143434525,
      "num_tokens": 46026257.0,
      "step": 23160
    },
    {
      "entropy": 0.5919083263725042,
      "epoch": 0.4162213140521848,
      "grad_norm": 4.875,
      "learning_rate": 3.231191823676738e-05,
      "loss": 0.541643762588501,
      "mean_token_accuracy": 0.8708536520600318,
      "num_tokens": 46047386.0,
      "step": 23170
    },
    {
      "entropy": 0.6307024881243706,
      "epoch": 0.41640095208155564,
      "grad_norm": 5.9375,
      "learning_rate": 3.2298178558386e-05,
      "loss": 0.6454501628875733,
      "mean_token_accuracy": 0.8373385310173035,
      "num_tokens": 46066855.0,
      "step": 23180
    },
    {
      "entropy": 0.6024470072239637,
      "epoch": 0.41658059011092646,
      "grad_norm": 6.78125,
      "learning_rate": 3.2284436469833676e-05,
      "loss": 0.612937068939209,
      "mean_token_accuracy": 0.8493049532175064,
      "num_tokens": 46086517.0,
      "step": 23190
    },
    {
      "entropy": 0.6759311430156231,
      "epoch": 0.4167602281402973,
      "grad_norm": 8.0,
      "learning_rate": 3.227069197564863e-05,
      "loss": 0.7639508247375488,
      "mean_token_accuracy": 0.8115621119737625,
      "num_tokens": 46106135.0,
      "step": 23200
    },
    {
      "entropy": 0.6196597956120968,
      "epoch": 0.4169398661696681,
      "grad_norm": 8.5,
      "learning_rate": 3.225694508036989e-05,
      "loss": 0.6441765308380127,
      "mean_token_accuracy": 0.8415669411420822,
      "num_tokens": 46125533.0,
      "step": 23210
    },
    {
      "entropy": 0.6639901414513588,
      "epoch": 0.4171195041990389,
      "grad_norm": 4.375,
      "learning_rate": 3.224319578853726e-05,
      "loss": 0.6453790187835693,
      "mean_token_accuracy": 0.8375705406069756,
      "num_tokens": 46146831.0,
      "step": 23220
    },
    {
      "entropy": 0.6392422068864108,
      "epoch": 0.41729914222840975,
      "grad_norm": 5.0625,
      "learning_rate": 3.222944410469135e-05,
      "loss": 0.6557307243347168,
      "mean_token_accuracy": 0.8410643219947815,
      "num_tokens": 46165976.0,
      "step": 23230
    },
    {
      "entropy": 0.7951401233673095,
      "epoch": 0.41747878025778057,
      "grad_norm": 6.65625,
      "learning_rate": 3.2215690033373555e-05,
      "loss": 0.8281031608581543,
      "mean_token_accuracy": 0.8034468427300453,
      "num_tokens": 46185598.0,
      "step": 23240
    },
    {
      "entropy": 0.6511195970699191,
      "epoch": 0.4176584182871514,
      "grad_norm": 5.78125,
      "learning_rate": 3.220193357912606e-05,
      "loss": 0.6580522537231446,
      "mean_token_accuracy": 0.8436175793409347,
      "num_tokens": 46206444.0,
      "step": 23250
    },
    {
      "entropy": 0.6845143094658852,
      "epoch": 0.4178380563165222,
      "grad_norm": 5.4375,
      "learning_rate": 3.218817474649183e-05,
      "loss": 0.6704878330230712,
      "mean_token_accuracy": 0.8357115671038627,
      "num_tokens": 46226102.0,
      "step": 23260
    },
    {
      "entropy": 0.7219809021800756,
      "epoch": 0.41801769434589303,
      "grad_norm": 6.375,
      "learning_rate": 3.217441354001462e-05,
      "loss": 0.7451582908630371,
      "mean_token_accuracy": 0.8247045442461968,
      "num_tokens": 46246800.0,
      "step": 23270
    },
    {
      "entropy": 0.6332199849188328,
      "epoch": 0.41819733237526385,
      "grad_norm": 5.03125,
      "learning_rate": 3.2160649964238976e-05,
      "loss": 0.6463354587554931,
      "mean_token_accuracy": 0.8427874118089675,
      "num_tokens": 46266296.0,
      "step": 23280
    },
    {
      "entropy": 0.6888075068593025,
      "epoch": 0.4183769704046347,
      "grad_norm": 7.21875,
      "learning_rate": 3.214688402371021e-05,
      "loss": 0.6718083381652832,
      "mean_token_accuracy": 0.8235295474529266,
      "num_tokens": 46285333.0,
      "step": 23290
    },
    {
      "entropy": 0.6196028664708138,
      "epoch": 0.4185566084340055,
      "grad_norm": 6.1875,
      "learning_rate": 3.2133115722974435e-05,
      "loss": 0.6293376445770263,
      "mean_token_accuracy": 0.8440947338938714,
      "num_tokens": 46305708.0,
      "step": 23300
    },
    {
      "entropy": 0.7481485478579998,
      "epoch": 0.4187362464633763,
      "grad_norm": 4.5625,
      "learning_rate": 3.2119345066578523e-05,
      "loss": 0.7710437774658203,
      "mean_token_accuracy": 0.8181166261434555,
      "num_tokens": 46326098.0,
      "step": 23310
    },
    {
      "entropy": 0.6420524008572102,
      "epoch": 0.41891588449274714,
      "grad_norm": 6.9375,
      "learning_rate": 3.210557205907015e-05,
      "loss": 0.6342854499816895,
      "mean_token_accuracy": 0.8384896740317345,
      "num_tokens": 46344915.0,
      "step": 23320
    },
    {
      "entropy": 0.7156854037195444,
      "epoch": 0.41909552252211796,
      "grad_norm": 5.375,
      "learning_rate": 3.209179670499773e-05,
      "loss": 0.6809807777404785,
      "mean_token_accuracy": 0.8372182667255401,
      "num_tokens": 46364522.0,
      "step": 23330
    },
    {
      "entropy": 0.7047219440340996,
      "epoch": 0.4192751605514887,
      "grad_norm": 7.40625,
      "learning_rate": 3.20780190089105e-05,
      "loss": 0.7081789970397949,
      "mean_token_accuracy": 0.8230768047273159,
      "num_tokens": 46385845.0,
      "step": 23340
    },
    {
      "entropy": 0.7965566612780094,
      "epoch": 0.41945479858085954,
      "grad_norm": 5.40625,
      "learning_rate": 3.2064238975358426e-05,
      "loss": 0.8140080451965332,
      "mean_token_accuracy": 0.8055187836289406,
      "num_tokens": 46406830.0,
      "step": 23350
    },
    {
      "entropy": 0.5983365148305893,
      "epoch": 0.41963443661023037,
      "grad_norm": 6.96875,
      "learning_rate": 3.2050456608892264e-05,
      "loss": 0.5742587089538574,
      "mean_token_accuracy": 0.8593401685357094,
      "num_tokens": 46426832.0,
      "step": 23360
    },
    {
      "entropy": 0.7125523172318935,
      "epoch": 0.4198140746396012,
      "grad_norm": 6.84375,
      "learning_rate": 3.203667191406356e-05,
      "loss": 0.7139479637145996,
      "mean_token_accuracy": 0.821300408244133,
      "num_tokens": 46446471.0,
      "step": 23370
    },
    {
      "entropy": 0.6192107323557139,
      "epoch": 0.419993712668972,
      "grad_norm": 6.625,
      "learning_rate": 3.2022884895424594e-05,
      "loss": 0.6201735019683838,
      "mean_token_accuracy": 0.8412507727742196,
      "num_tokens": 46465851.0,
      "step": 23380
    },
    {
      "entropy": 0.6526543114334344,
      "epoch": 0.42017335069834283,
      "grad_norm": 6.40625,
      "learning_rate": 3.2009095557528434e-05,
      "loss": 0.674851942062378,
      "mean_token_accuracy": 0.8391532301902771,
      "num_tokens": 46485708.0,
      "step": 23390
    },
    {
      "entropy": 0.6175659246742725,
      "epoch": 0.42035298872771365,
      "grad_norm": 4.875,
      "learning_rate": 3.1995303904928906e-05,
      "loss": 0.6282788276672363,
      "mean_token_accuracy": 0.8451520740985871,
      "num_tokens": 46505589.0,
      "step": 23400
    },
    {
      "entropy": 0.7528997153043747,
      "epoch": 0.42053262675708447,
      "grad_norm": 7.46875,
      "learning_rate": 3.1981509942180616e-05,
      "loss": 0.7615347385406495,
      "mean_token_accuracy": 0.8222849503159523,
      "num_tokens": 46525755.0,
      "step": 23410
    },
    {
      "entropy": 0.5969504952430725,
      "epoch": 0.4207122647864553,
      "grad_norm": 5.28125,
      "learning_rate": 3.19677136738389e-05,
      "loss": 0.5787000179290771,
      "mean_token_accuracy": 0.8569890528917312,
      "num_tokens": 46545136.0,
      "step": 23420
    },
    {
      "entropy": 0.5864482589066029,
      "epoch": 0.4208919028158261,
      "grad_norm": 6.3125,
      "learning_rate": 3.19539151044599e-05,
      "loss": 0.6217243671417236,
      "mean_token_accuracy": 0.842398451268673,
      "num_tokens": 46564088.0,
      "step": 23430
    },
    {
      "entropy": 0.6566278148442507,
      "epoch": 0.42107154084519693,
      "grad_norm": 7.0625,
      "learning_rate": 3.194011423860048e-05,
      "loss": 0.628873348236084,
      "mean_token_accuracy": 0.8416550204157829,
      "num_tokens": 46585061.0,
      "step": 23440
    },
    {
      "entropy": 0.7693922020494938,
      "epoch": 0.42125117887456776,
      "grad_norm": 5.21875,
      "learning_rate": 3.1926311080818276e-05,
      "loss": 0.7796463966369629,
      "mean_token_accuracy": 0.8079635210335254,
      "num_tokens": 46604801.0,
      "step": 23450
    },
    {
      "entropy": 0.6586019679903984,
      "epoch": 0.4214308169039386,
      "grad_norm": 6.6875,
      "learning_rate": 3.191250563567169e-05,
      "loss": 0.6638798713684082,
      "mean_token_accuracy": 0.8277608349919319,
      "num_tokens": 46624277.0,
      "step": 23460
    },
    {
      "entropy": 0.6861205764114857,
      "epoch": 0.4216104549333094,
      "grad_norm": 5.25,
      "learning_rate": 3.189869790771987e-05,
      "loss": 0.7087327003479004,
      "mean_token_accuracy": 0.8244765803217888,
      "num_tokens": 46643745.0,
      "step": 23470
    },
    {
      "entropy": 0.7769518010318279,
      "epoch": 0.4217900929626802,
      "grad_norm": 7.6875,
      "learning_rate": 3.188488790152271e-05,
      "loss": 0.7951416969299316,
      "mean_token_accuracy": 0.8073859304189682,
      "num_tokens": 46662940.0,
      "step": 23480
    },
    {
      "entropy": 0.5699319195002317,
      "epoch": 0.42196973099205104,
      "grad_norm": 7.3125,
      "learning_rate": 3.187107562164087e-05,
      "loss": 0.577358341217041,
      "mean_token_accuracy": 0.852861861884594,
      "num_tokens": 46683294.0,
      "step": 23490
    },
    {
      "entropy": 0.6043197866529226,
      "epoch": 0.42214936902142186,
      "grad_norm": 4.5,
      "learning_rate": 3.1857261072635756e-05,
      "loss": 0.6108151912689209,
      "mean_token_accuracy": 0.8447893992066383,
      "num_tokens": 46702759.0,
      "step": 23500
    },
    {
      "entropy": 0.7254449337720871,
      "epoch": 0.4223290070507926,
      "grad_norm": 5.875,
      "learning_rate": 3.184344425906952e-05,
      "loss": 0.7229170799255371,
      "mean_token_accuracy": 0.8244763568043709,
      "num_tokens": 46722507.0,
      "step": 23510
    },
    {
      "entropy": 0.6677377238869667,
      "epoch": 0.42250864508016345,
      "grad_norm": 5.875,
      "learning_rate": 3.182962518550507e-05,
      "loss": 0.6493424892425537,
      "mean_token_accuracy": 0.834862295538187,
      "num_tokens": 46742663.0,
      "step": 23520
    },
    {
      "entropy": 0.7454547628760337,
      "epoch": 0.42268828310953427,
      "grad_norm": 6.46875,
      "learning_rate": 3.1815803856506065e-05,
      "loss": 0.7162877082824707,
      "mean_token_accuracy": 0.8231628760695457,
      "num_tokens": 46763001.0,
      "step": 23530
    },
    {
      "entropy": 0.5990430530160665,
      "epoch": 0.4228679211389051,
      "grad_norm": 5.6875,
      "learning_rate": 3.180198027663688e-05,
      "loss": 0.6127132415771485,
      "mean_token_accuracy": 0.8470400109887123,
      "num_tokens": 46783269.0,
      "step": 23540
    },
    {
      "entropy": 0.6272448725998402,
      "epoch": 0.4230475591682759,
      "grad_norm": 5.0625,
      "learning_rate": 3.178815445046267e-05,
      "loss": 0.6132073402404785,
      "mean_token_accuracy": 0.8410403013229371,
      "num_tokens": 46803280.0,
      "step": 23550
    },
    {
      "entropy": 0.6414046663790941,
      "epoch": 0.42322719719764673,
      "grad_norm": 6.3125,
      "learning_rate": 3.1774326382549294e-05,
      "loss": 0.6087607383728028,
      "mean_token_accuracy": 0.838011734187603,
      "num_tokens": 46822857.0,
      "step": 23560
    },
    {
      "entropy": 0.660580787062645,
      "epoch": 0.42340683522701755,
      "grad_norm": 5.84375,
      "learning_rate": 3.17604960774634e-05,
      "loss": 0.678853702545166,
      "mean_token_accuracy": 0.8306648895144463,
      "num_tokens": 46842970.0,
      "step": 23570
    },
    {
      "entropy": 0.7083369355648756,
      "epoch": 0.4235864732563884,
      "grad_norm": 5.21875,
      "learning_rate": 3.174666353977231e-05,
      "loss": 0.7372464179992676,
      "mean_token_accuracy": 0.8184053733944893,
      "num_tokens": 46863624.0,
      "step": 23580
    },
    {
      "entropy": 0.7065844684839249,
      "epoch": 0.4237661112857592,
      "grad_norm": 7.875,
      "learning_rate": 3.173282877404416e-05,
      "loss": 0.7306058406829834,
      "mean_token_accuracy": 0.8294106855988502,
      "num_tokens": 46884669.0,
      "step": 23590
    },
    {
      "entropy": 0.6485109269618988,
      "epoch": 0.42394574931513,
      "grad_norm": 6.65625,
      "learning_rate": 3.171899178484775e-05,
      "loss": 0.6996689796447754,
      "mean_token_accuracy": 0.8309333324432373,
      "num_tokens": 46904201.0,
      "step": 23600
    },
    {
      "entropy": 0.586810477823019,
      "epoch": 0.42412538734450084,
      "grad_norm": 5.59375,
      "learning_rate": 3.170515257675267e-05,
      "loss": 0.6223575592041015,
      "mean_token_accuracy": 0.8495154276490211,
      "num_tokens": 46924700.0,
      "step": 23610
    },
    {
      "entropy": 0.67276396676898,
      "epoch": 0.42430502537387166,
      "grad_norm": 6.59375,
      "learning_rate": 3.1691311154329196e-05,
      "loss": 0.7173723220825196,
      "mean_token_accuracy": 0.8314430415630341,
      "num_tokens": 46944195.0,
      "step": 23620
    },
    {
      "entropy": 0.7477695576846599,
      "epoch": 0.4244846634032425,
      "grad_norm": 5.3125,
      "learning_rate": 3.167746752214838e-05,
      "loss": 0.7162967681884765,
      "mean_token_accuracy": 0.8267581835389137,
      "num_tokens": 46965342.0,
      "step": 23630
    },
    {
      "entropy": 0.7500372648239135,
      "epoch": 0.4246643014326133,
      "grad_norm": 7.03125,
      "learning_rate": 3.166362168478196e-05,
      "loss": 0.7602426052093506,
      "mean_token_accuracy": 0.8255150020122528,
      "num_tokens": 46984530.0,
      "step": 23640
    },
    {
      "entropy": 0.6396973673254251,
      "epoch": 0.4248439394619841,
      "grad_norm": 5.53125,
      "learning_rate": 3.1649773646802445e-05,
      "loss": 0.6249341011047364,
      "mean_token_accuracy": 0.8477486088871956,
      "num_tokens": 47004567.0,
      "step": 23650
    },
    {
      "entropy": 0.7148826811462641,
      "epoch": 0.42502357749135494,
      "grad_norm": 4.375,
      "learning_rate": 3.1635923412783044e-05,
      "loss": 0.7322266101837158,
      "mean_token_accuracy": 0.8240624606609345,
      "num_tokens": 47023577.0,
      "step": 23660
    },
    {
      "entropy": 0.7269294366240502,
      "epoch": 0.42520321552072576,
      "grad_norm": 5.59375,
      "learning_rate": 3.162207098729768e-05,
      "loss": 0.7517065525054931,
      "mean_token_accuracy": 0.8258009478449821,
      "num_tokens": 47043125.0,
      "step": 23670
    },
    {
      "entropy": 0.6598549596965313,
      "epoch": 0.42538285355009653,
      "grad_norm": 5.59375,
      "learning_rate": 3.160821637492104e-05,
      "loss": 0.6774972438812256,
      "mean_token_accuracy": 0.8381332844495774,
      "num_tokens": 47062853.0,
      "step": 23680
    },
    {
      "entropy": 0.7702537305653095,
      "epoch": 0.42556249157946735,
      "grad_norm": 5.75,
      "learning_rate": 3.15943595802285e-05,
      "loss": 0.755970048904419,
      "mean_token_accuracy": 0.8167732790112495,
      "num_tokens": 47083174.0,
      "step": 23690
    },
    {
      "entropy": 0.55663131326437,
      "epoch": 0.42574212960883817,
      "grad_norm": 4.75,
      "learning_rate": 3.158050060779617e-05,
      "loss": 0.5648555278778076,
      "mean_token_accuracy": 0.8610092669725418,
      "num_tokens": 47104474.0,
      "step": 23700
    },
    {
      "entropy": 0.6427796367555857,
      "epoch": 0.425921767638209,
      "grad_norm": 6.0625,
      "learning_rate": 3.1566639462200865e-05,
      "loss": 0.6340445518493653,
      "mean_token_accuracy": 0.8362182572484016,
      "num_tokens": 47123616.0,
      "step": 23710
    },
    {
      "entropy": 0.6005171479657292,
      "epoch": 0.4261014056675798,
      "grad_norm": 4.09375,
      "learning_rate": 3.155277614802015e-05,
      "loss": 0.6307083606719971,
      "mean_token_accuracy": 0.8414289563894272,
      "num_tokens": 47145000.0,
      "step": 23720
    },
    {
      "entropy": 0.6968921110033989,
      "epoch": 0.42628104369695063,
      "grad_norm": 5.03125,
      "learning_rate": 3.153891066983226e-05,
      "loss": 0.6963856220245361,
      "mean_token_accuracy": 0.8283406108617782,
      "num_tokens": 47163130.0,
      "step": 23730
    },
    {
      "entropy": 0.6803260661661625,
      "epoch": 0.42646068172632146,
      "grad_norm": 6.40625,
      "learning_rate": 3.1525043032216185e-05,
      "loss": 0.6678486824035644,
      "mean_token_accuracy": 0.8259052127599716,
      "num_tokens": 47183823.0,
      "step": 23740
    },
    {
      "entropy": 0.660888458788395,
      "epoch": 0.4266403197556923,
      "grad_norm": 6.0,
      "learning_rate": 3.1511173239751615e-05,
      "loss": 0.6798188209533691,
      "mean_token_accuracy": 0.8293200120329857,
      "num_tokens": 47203718.0,
      "step": 23750
    },
    {
      "entropy": 0.7096694614738226,
      "epoch": 0.4268199577850631,
      "grad_norm": 6.9375,
      "learning_rate": 3.149730129701893e-05,
      "loss": 0.6745379447937012,
      "mean_token_accuracy": 0.8326949834823608,
      "num_tokens": 47223142.0,
      "step": 23760
    },
    {
      "entropy": 0.7273824464529752,
      "epoch": 0.4269995958144339,
      "grad_norm": 6.65625,
      "learning_rate": 3.1483427208599256e-05,
      "loss": 0.7428422927856445,
      "mean_token_accuracy": 0.8209651917219162,
      "num_tokens": 47243548.0,
      "step": 23770
    },
    {
      "entropy": 0.6760349325835705,
      "epoch": 0.42717923384380474,
      "grad_norm": 6.1875,
      "learning_rate": 3.146955097907441e-05,
      "loss": 0.6873424530029297,
      "mean_token_accuracy": 0.8322339043021202,
      "num_tokens": 47262744.0,
      "step": 23780
    },
    {
      "entropy": 0.6197375915944576,
      "epoch": 0.42735887187317556,
      "grad_norm": 5.21875,
      "learning_rate": 3.145567261302692e-05,
      "loss": 0.6541970729827881,
      "mean_token_accuracy": 0.8423280358314514,
      "num_tokens": 47282814.0,
      "step": 23790
    },
    {
      "entropy": 0.6615794908255339,
      "epoch": 0.4275385099025464,
      "grad_norm": 6.125,
      "learning_rate": 3.1441792115039996e-05,
      "loss": 0.6347053050994873,
      "mean_token_accuracy": 0.8358678922057152,
      "num_tokens": 47302536.0,
      "step": 23800
    },
    {
      "entropy": 0.6702492091804743,
      "epoch": 0.4277181479319172,
      "grad_norm": 5.59375,
      "learning_rate": 3.142790948969759e-05,
      "loss": 0.6672655582427979,
      "mean_token_accuracy": 0.8402595117688179,
      "num_tokens": 47321623.0,
      "step": 23810
    },
    {
      "entropy": 0.6698753476142884,
      "epoch": 0.427897785961288,
      "grad_norm": 7.6875,
      "learning_rate": 3.141402474158434e-05,
      "loss": 0.6738217353820801,
      "mean_token_accuracy": 0.8343053162097931,
      "num_tokens": 47342292.0,
      "step": 23820
    },
    {
      "entropy": 0.6731942117214202,
      "epoch": 0.42807742399065885,
      "grad_norm": 6.125,
      "learning_rate": 3.140013787528559e-05,
      "loss": 0.6803479194641113,
      "mean_token_accuracy": 0.8351604998111725,
      "num_tokens": 47363443.0,
      "step": 23830
    },
    {
      "entropy": 0.6426824301481247,
      "epoch": 0.42825706202002967,
      "grad_norm": 6.5,
      "learning_rate": 3.1386248895387354e-05,
      "loss": 0.61573805809021,
      "mean_token_accuracy": 0.8445127591490745,
      "num_tokens": 47384622.0,
      "step": 23840
    },
    {
      "entropy": 0.6753062032163143,
      "epoch": 0.42843670004940043,
      "grad_norm": 5.84375,
      "learning_rate": 3.137235780647639e-05,
      "loss": 0.6749578952789307,
      "mean_token_accuracy": 0.8357361450791359,
      "num_tokens": 47403592.0,
      "step": 23850
    },
    {
      "entropy": 0.6939686320722103,
      "epoch": 0.42861633807877125,
      "grad_norm": 6.0625,
      "learning_rate": 3.135846461314011e-05,
      "loss": 0.7325024604797363,
      "mean_token_accuracy": 0.8239819511771203,
      "num_tokens": 47423913.0,
      "step": 23860
    },
    {
      "entropy": 0.696803268790245,
      "epoch": 0.4287959761081421,
      "grad_norm": 7.3125,
      "learning_rate": 3.134456931996666e-05,
      "loss": 0.7140926837921142,
      "mean_token_accuracy": 0.8170573860406876,
      "num_tokens": 47444523.0,
      "step": 23870
    },
    {
      "entropy": 0.6037124387919903,
      "epoch": 0.4289756141375129,
      "grad_norm": 9.0,
      "learning_rate": 3.1330671931544855e-05,
      "loss": 0.606038761138916,
      "mean_token_accuracy": 0.8527293592691422,
      "num_tokens": 47463990.0,
      "step": 23880
    },
    {
      "entropy": 0.6957373976707458,
      "epoch": 0.4291552521668837,
      "grad_norm": 6.4375,
      "learning_rate": 3.13167724524642e-05,
      "loss": 0.6865151882171631,
      "mean_token_accuracy": 0.82889783680439,
      "num_tokens": 47483852.0,
      "step": 23890
    },
    {
      "entropy": 0.6792431350797414,
      "epoch": 0.42933489019625454,
      "grad_norm": 6.53125,
      "learning_rate": 3.1302870887314906e-05,
      "loss": 0.7170518398284912,
      "mean_token_accuracy": 0.8275602132081985,
      "num_tokens": 47503603.0,
      "step": 23900
    },
    {
      "entropy": 0.6037205047905445,
      "epoch": 0.42951452822562536,
      "grad_norm": 7.75,
      "learning_rate": 3.128896724068786e-05,
      "loss": 0.6428238868713378,
      "mean_token_accuracy": 0.8414407536387444,
      "num_tokens": 47523826.0,
      "step": 23910
    },
    {
      "entropy": 0.6739425085484981,
      "epoch": 0.4296941662549962,
      "grad_norm": 5.5625,
      "learning_rate": 3.1275061517174656e-05,
      "loss": 0.6554346561431885,
      "mean_token_accuracy": 0.8329007640480995,
      "num_tokens": 47543826.0,
      "step": 23920
    },
    {
      "entropy": 0.6566864341497421,
      "epoch": 0.429873804284367,
      "grad_norm": 6.0625,
      "learning_rate": 3.1261153721367535e-05,
      "loss": 0.6901873588562012,
      "mean_token_accuracy": 0.8337967872619629,
      "num_tokens": 47562575.0,
      "step": 23930
    },
    {
      "entropy": 0.7516804128885269,
      "epoch": 0.4300534423137378,
      "grad_norm": 6.96875,
      "learning_rate": 3.1247243857859464e-05,
      "loss": 0.7447842597961426,
      "mean_token_accuracy": 0.8130014762282372,
      "num_tokens": 47581688.0,
      "step": 23940
    },
    {
      "entropy": 0.6308729652315378,
      "epoch": 0.43023308034310864,
      "grad_norm": 6.34375,
      "learning_rate": 3.123333193124406e-05,
      "loss": 0.6159263610839844,
      "mean_token_accuracy": 0.8465975895524025,
      "num_tokens": 47602053.0,
      "step": 23950
    },
    {
      "entropy": 0.6302910849452019,
      "epoch": 0.43041271837247946,
      "grad_norm": 7.1875,
      "learning_rate": 3.121941794611566e-05,
      "loss": 0.6234625339508056,
      "mean_token_accuracy": 0.8500958800315856,
      "num_tokens": 47622221.0,
      "step": 23960
    },
    {
      "entropy": 0.6727326799184083,
      "epoch": 0.4305923564018503,
      "grad_norm": 6.625,
      "learning_rate": 3.1205501907069234e-05,
      "loss": 0.6773569107055664,
      "mean_token_accuracy": 0.8358851373195648,
      "num_tokens": 47642303.0,
      "step": 23970
    },
    {
      "entropy": 0.5640832461416722,
      "epoch": 0.4307719944312211,
      "grad_norm": 5.25,
      "learning_rate": 3.119158381870046e-05,
      "loss": 0.5562549114227295,
      "mean_token_accuracy": 0.8653771936893463,
      "num_tokens": 47662365.0,
      "step": 23980
    },
    {
      "entropy": 0.646189521998167,
      "epoch": 0.4309516324605919,
      "grad_norm": 5.125,
      "learning_rate": 3.11776636856057e-05,
      "loss": 0.6979965209960938,
      "mean_token_accuracy": 0.8403136804699898,
      "num_tokens": 47681648.0,
      "step": 23990
    },
    {
      "epoch": 0.43113127048996275,
      "eval_entropy": 0.643974608361721,
      "eval_loss": 0.6478521823883057,
      "eval_mean_token_accuracy": 0.8415909221172333,
      "eval_num_tokens": 47700497.0,
      "eval_runtime": 40.5431,
      "eval_samples_per_second": 49.33,
      "eval_steps_per_second": 6.166,
      "step": 24000
    },
    {
      "entropy": 0.6116340771317482,
      "epoch": 0.43131090851933357,
      "grad_norm": 7.40625,
      "learning_rate": 3.114981730362695e-05,
      "loss": 0.6225731849670411,
      "mean_token_accuracy": 0.8486511081457138,
      "num_tokens": 47720176.0,
      "step": 24010
    },
    {
      "entropy": 0.6417730018496514,
      "epoch": 0.43149054654870433,
      "grad_norm": 8.1875,
      "learning_rate": 3.113589106393904e-05,
      "loss": 0.6425131797790528,
      "mean_token_accuracy": 0.8422294422984123,
      "num_tokens": 47740766.0,
      "step": 24020
    },
    {
      "entropy": 0.7078488174825907,
      "epoch": 0.43167018457807516,
      "grad_norm": 7.34375,
      "learning_rate": 3.112196279791726e-05,
      "loss": 0.7469717502593994,
      "mean_token_accuracy": 0.8250307664275169,
      "num_tokens": 47761590.0,
      "step": 24030
    },
    {
      "entropy": 0.6866001222282648,
      "epoch": 0.431849822607446,
      "grad_norm": 5.3125,
      "learning_rate": 3.1108032510161326e-05,
      "loss": 0.6446122169494629,
      "mean_token_accuracy": 0.833529357612133,
      "num_tokens": 47779748.0,
      "step": 24040
    },
    {
      "entropy": 0.6868690174072981,
      "epoch": 0.4320294606368168,
      "grad_norm": 7.3125,
      "learning_rate": 3.109410020527161e-05,
      "loss": 0.7091449737548828,
      "mean_token_accuracy": 0.8321635752916337,
      "num_tokens": 47798969.0,
      "step": 24050
    },
    {
      "entropy": 0.6160532560199499,
      "epoch": 0.4322090986661876,
      "grad_norm": 6.21875,
      "learning_rate": 3.108016588784916e-05,
      "loss": 0.6060786724090577,
      "mean_token_accuracy": 0.8443858772516251,
      "num_tokens": 47818210.0,
      "step": 24060
    },
    {
      "entropy": 0.6329162698239088,
      "epoch": 0.43238873669555844,
      "grad_norm": 7.125,
      "learning_rate": 3.1066229562495685e-05,
      "loss": 0.6312871932983398,
      "mean_token_accuracy": 0.8438546836376191,
      "num_tokens": 47838313.0,
      "step": 24070
    },
    {
      "entropy": 0.672634968161583,
      "epoch": 0.43256837472492926,
      "grad_norm": 6.40625,
      "learning_rate": 3.1052291233813555e-05,
      "loss": 0.6814215183258057,
      "mean_token_accuracy": 0.8314599946141243,
      "num_tokens": 47857746.0,
      "step": 24080
    },
    {
      "entropy": 0.7047202907502651,
      "epoch": 0.4327480127543001,
      "grad_norm": 8.625,
      "learning_rate": 3.1038350906405804e-05,
      "loss": 0.7244247913360595,
      "mean_token_accuracy": 0.8213862523436546,
      "num_tokens": 47878735.0,
      "step": 24090
    },
    {
      "entropy": 0.6799656547605991,
      "epoch": 0.4329276507836709,
      "grad_norm": 5.21875,
      "learning_rate": 3.1024408584876124e-05,
      "loss": 0.6469698905944824,
      "mean_token_accuracy": 0.836238905787468,
      "num_tokens": 47897996.0,
      "step": 24100
    },
    {
      "entropy": 0.7024842206388712,
      "epoch": 0.4331072888130417,
      "grad_norm": 7.6875,
      "learning_rate": 3.1010464273828864e-05,
      "loss": 0.7120854377746582,
      "mean_token_accuracy": 0.8228103414177894,
      "num_tokens": 47917720.0,
      "step": 24110
    },
    {
      "entropy": 0.6569418989121913,
      "epoch": 0.43328692684241255,
      "grad_norm": 8.5,
      "learning_rate": 3.0996517977869045e-05,
      "loss": 0.6930888652801513,
      "mean_token_accuracy": 0.8321496114134789,
      "num_tokens": 47937335.0,
      "step": 24120
    },
    {
      "entropy": 0.7343391198664904,
      "epoch": 0.43346656487178337,
      "grad_norm": 7.3125,
      "learning_rate": 3.09825697016023e-05,
      "loss": 0.7747005939483642,
      "mean_token_accuracy": 0.8177273198962212,
      "num_tokens": 47957523.0,
      "step": 24130
    },
    {
      "entropy": 0.6938234955072403,
      "epoch": 0.4336462029011542,
      "grad_norm": 6.03125,
      "learning_rate": 3.0968619449634984e-05,
      "loss": 0.7045555591583252,
      "mean_token_accuracy": 0.8318816423416138,
      "num_tokens": 47977998.0,
      "step": 24140
    },
    {
      "entropy": 0.6539994336664676,
      "epoch": 0.433825840930525,
      "grad_norm": 7.5,
      "learning_rate": 3.0954667226574036e-05,
      "loss": 0.6392402648925781,
      "mean_token_accuracy": 0.8388224467635155,
      "num_tokens": 47997819.0,
      "step": 24150
    },
    {
      "entropy": 0.6357131343334913,
      "epoch": 0.43400547895989583,
      "grad_norm": 4.8125,
      "learning_rate": 3.0940713037027106e-05,
      "loss": 0.6495887756347656,
      "mean_token_accuracy": 0.8364812210202217,
      "num_tokens": 48017141.0,
      "step": 24160
    },
    {
      "entropy": 0.5694891676306725,
      "epoch": 0.43418511698926665,
      "grad_norm": 6.375,
      "learning_rate": 3.092675688560244e-05,
      "loss": 0.5821140289306641,
      "mean_token_accuracy": 0.8563517794013024,
      "num_tokens": 48038509.0,
      "step": 24170
    },
    {
      "entropy": 0.6492311626672744,
      "epoch": 0.43436475501863747,
      "grad_norm": 7.0,
      "learning_rate": 3.0912798776908974e-05,
      "loss": 0.6166582584381104,
      "mean_token_accuracy": 0.8455361276865005,
      "num_tokens": 48059405.0,
      "step": 24180
    },
    {
      "entropy": 0.6579222697764635,
      "epoch": 0.43454439304800824,
      "grad_norm": 6.0,
      "learning_rate": 3.0898838715556256e-05,
      "loss": 0.6397764205932617,
      "mean_token_accuracy": 0.8406086131930351,
      "num_tokens": 48078804.0,
      "step": 24190
    },
    {
      "entropy": 0.6347151096910238,
      "epoch": 0.43472403107737906,
      "grad_norm": 5.0625,
      "learning_rate": 3.0884876706154516e-05,
      "loss": 0.6415602207183838,
      "mean_token_accuracy": 0.8425936251878738,
      "num_tokens": 48098535.0,
      "step": 24200
    },
    {
      "entropy": 0.6497670665383339,
      "epoch": 0.4349036691067499,
      "grad_norm": 5.875,
      "learning_rate": 3.087091275331459e-05,
      "loss": 0.6485769748687744,
      "mean_token_accuracy": 0.8367728665471077,
      "num_tokens": 48117842.0,
      "step": 24210
    },
    {
      "entropy": 0.6299391526728868,
      "epoch": 0.4350833071361207,
      "grad_norm": 4.59375,
      "learning_rate": 3.085694686164798e-05,
      "loss": 0.6326821327209473,
      "mean_token_accuracy": 0.8517057940363884,
      "num_tokens": 48138103.0,
      "step": 24220
    },
    {
      "entropy": 0.5950762785971164,
      "epoch": 0.4352629451654915,
      "grad_norm": 8.0625,
      "learning_rate": 3.0842979035766825e-05,
      "loss": 0.5773335933685303,
      "mean_token_accuracy": 0.859959015250206,
      "num_tokens": 48157555.0,
      "step": 24230
    },
    {
      "entropy": 0.6722312860190869,
      "epoch": 0.43544258319486234,
      "grad_norm": 5.875,
      "learning_rate": 3.0829009280283884e-05,
      "loss": 0.7104061603546142,
      "mean_token_accuracy": 0.8225480541586876,
      "num_tokens": 48176636.0,
      "step": 24240
    },
    {
      "entropy": 0.6431870773434639,
      "epoch": 0.43562222122423316,
      "grad_norm": 6.75,
      "learning_rate": 3.081503759981259e-05,
      "loss": 0.6344115734100342,
      "mean_token_accuracy": 0.839464607834816,
      "num_tokens": 48196475.0,
      "step": 24250
    },
    {
      "entropy": 0.6827650159597397,
      "epoch": 0.435801859253604,
      "grad_norm": 8.375,
      "learning_rate": 3.080106399896697e-05,
      "loss": 0.7154043197631836,
      "mean_token_accuracy": 0.8314108595252037,
      "num_tokens": 48215659.0,
      "step": 24260
    },
    {
      "entropy": 0.6553299807012081,
      "epoch": 0.4359814972829748,
      "grad_norm": 6.09375,
      "learning_rate": 3.0787088482361725e-05,
      "loss": 0.6391956329345703,
      "mean_token_accuracy": 0.8403597295284271,
      "num_tokens": 48234982.0,
      "step": 24270
    },
    {
      "entropy": 0.6779052548110485,
      "epoch": 0.4361611353123456,
      "grad_norm": 8.25,
      "learning_rate": 3.0773111054612146e-05,
      "loss": 0.6637829780578614,
      "mean_token_accuracy": 0.838368509709835,
      "num_tokens": 48254121.0,
      "step": 24280
    },
    {
      "entropy": 0.5827049601823091,
      "epoch": 0.43634077334171645,
      "grad_norm": 7.96875,
      "learning_rate": 3.07591317203342e-05,
      "loss": 0.5852439880371094,
      "mean_token_accuracy": 0.8524028226733208,
      "num_tokens": 48273354.0,
      "step": 24290
    },
    {
      "entropy": 0.616769940406084,
      "epoch": 0.43652041137108727,
      "grad_norm": 6.65625,
      "learning_rate": 3.0745150484144444e-05,
      "loss": 0.6197027683258056,
      "mean_token_accuracy": 0.8483280181884766,
      "num_tokens": 48291969.0,
      "step": 24300
    },
    {
      "entropy": 0.6108016103506089,
      "epoch": 0.4367000494004581,
      "grad_norm": 5.875,
      "learning_rate": 3.073116735066009e-05,
      "loss": 0.5824091911315918,
      "mean_token_accuracy": 0.8571629509329796,
      "num_tokens": 48311133.0,
      "step": 24310
    },
    {
      "entropy": 0.5841566108167171,
      "epoch": 0.4368796874298289,
      "grad_norm": 4.875,
      "learning_rate": 3.071718232449896e-05,
      "loss": 0.5923855304718018,
      "mean_token_accuracy": 0.8529880329966545,
      "num_tokens": 48331149.0,
      "step": 24320
    },
    {
      "entropy": 0.7286870822310447,
      "epoch": 0.43705932545919973,
      "grad_norm": 8.625,
      "learning_rate": 3.070319541027952e-05,
      "loss": 0.7580495834350586,
      "mean_token_accuracy": 0.8108010180294514,
      "num_tokens": 48351572.0,
      "step": 24330
    },
    {
      "entropy": 0.6825996853411198,
      "epoch": 0.43723896348857055,
      "grad_norm": 8.75,
      "learning_rate": 3.068920661262085e-05,
      "loss": 0.6995247840881348,
      "mean_token_accuracy": 0.8319066911935806,
      "num_tokens": 48370946.0,
      "step": 24340
    },
    {
      "entropy": 0.6596868000924587,
      "epoch": 0.4374186015179414,
      "grad_norm": 7.4375,
      "learning_rate": 3.067521593614262e-05,
      "loss": 0.6630935192108154,
      "mean_token_accuracy": 0.8351659640669823,
      "num_tokens": 48390397.0,
      "step": 24350
    },
    {
      "entropy": 0.6398804992437362,
      "epoch": 0.43759823954731214,
      "grad_norm": 6.28125,
      "learning_rate": 3.066122338546518e-05,
      "loss": 0.6383000373840332,
      "mean_token_accuracy": 0.8420041784644127,
      "num_tokens": 48409555.0,
      "step": 24360
    },
    {
      "entropy": 0.7143633153289557,
      "epoch": 0.43777787757668296,
      "grad_norm": 5.46875,
      "learning_rate": 3.064722896520946e-05,
      "loss": 0.723311471939087,
      "mean_token_accuracy": 0.8258228704333306,
      "num_tokens": 48429848.0,
      "step": 24370
    },
    {
      "entropy": 0.6875798180699348,
      "epoch": 0.4379575156060538,
      "grad_norm": 8.6875,
      "learning_rate": 3.063323267999702e-05,
      "loss": 0.7160614013671875,
      "mean_token_accuracy": 0.8254805788397789,
      "num_tokens": 48448674.0,
      "step": 24380
    },
    {
      "entropy": 0.651991369575262,
      "epoch": 0.4381371536354246,
      "grad_norm": 7.5,
      "learning_rate": 3.0619234534450034e-05,
      "loss": 0.6259730815887451,
      "mean_token_accuracy": 0.8404921442270279,
      "num_tokens": 48467701.0,
      "step": 24390
    },
    {
      "entropy": 0.6553060866892337,
      "epoch": 0.4383167916647954,
      "grad_norm": 5.5,
      "learning_rate": 3.060523453319127e-05,
      "loss": 0.6829468250274658,
      "mean_token_accuracy": 0.8249972000718117,
      "num_tokens": 48487228.0,
      "step": 24400
    },
    {
      "entropy": 0.6708332661539316,
      "epoch": 0.43849642969416625,
      "grad_norm": 6.03125,
      "learning_rate": 3.059123268084415e-05,
      "loss": 0.6162095069885254,
      "mean_token_accuracy": 0.8462602972984314,
      "num_tokens": 48507170.0,
      "step": 24410
    },
    {
      "entropy": 0.6859062768518924,
      "epoch": 0.43867606772353707,
      "grad_norm": 6.71875,
      "learning_rate": 3.057722898203267e-05,
      "loss": 0.7276244640350342,
      "mean_token_accuracy": 0.8227898523211479,
      "num_tokens": 48526543.0,
      "step": 24420
    },
    {
      "entropy": 0.6548083253204823,
      "epoch": 0.4388557057529079,
      "grad_norm": 5.65625,
      "learning_rate": 3.0563223441381464e-05,
      "loss": 0.6397836685180665,
      "mean_token_accuracy": 0.8447485193610191,
      "num_tokens": 48546188.0,
      "step": 24430
    },
    {
      "entropy": 0.6609334945678711,
      "epoch": 0.4390353437822787,
      "grad_norm": 5.65625,
      "learning_rate": 3.0549216063515754e-05,
      "loss": 0.6716468334197998,
      "mean_token_accuracy": 0.8284130379557609,
      "num_tokens": 48566666.0,
      "step": 24440
    },
    {
      "entropy": 0.6348645225167274,
      "epoch": 0.43921498181164953,
      "grad_norm": 5.84375,
      "learning_rate": 3.053520685306138e-05,
      "loss": 0.6639882564544678,
      "mean_token_accuracy": 0.8381089508533478,
      "num_tokens": 48586426.0,
      "step": 24450
    },
    {
      "entropy": 0.7350969210267067,
      "epoch": 0.43939461984102035,
      "grad_norm": 6.1875,
      "learning_rate": 3.052119581464477e-05,
      "loss": 0.7468497276306152,
      "mean_token_accuracy": 0.8118630766868591,
      "num_tokens": 48606440.0,
      "step": 24460
    },
    {
      "entropy": 0.6784627448767424,
      "epoch": 0.43957425787039117,
      "grad_norm": 6.4375,
      "learning_rate": 3.0507182952892983e-05,
      "loss": 0.6952576637268066,
      "mean_token_accuracy": 0.8315318375825882,
      "num_tokens": 48627786.0,
      "step": 24470
    },
    {
      "entropy": 0.5751110419631005,
      "epoch": 0.439753895899762,
      "grad_norm": 6.34375,
      "learning_rate": 3.049316827243366e-05,
      "loss": 0.575243616104126,
      "mean_token_accuracy": 0.8466513440012932,
      "num_tokens": 48647046.0,
      "step": 24480
    },
    {
      "entropy": 0.5471175953745842,
      "epoch": 0.4399335339291328,
      "grad_norm": 5.71875,
      "learning_rate": 3.0479151777895042e-05,
      "loss": 0.5089127063751221,
      "mean_token_accuracy": 0.8720040857791901,
      "num_tokens": 48666933.0,
      "step": 24490
    },
    {
      "entropy": 0.6096704304218292,
      "epoch": 0.44011317195850364,
      "grad_norm": 4.6875,
      "learning_rate": 3.0465133473905987e-05,
      "loss": 0.5958727359771728,
      "mean_token_accuracy": 0.8473158493638039,
      "num_tokens": 48687232.0,
      "step": 24500
    },
    {
      "entropy": 0.5894104741513729,
      "epoch": 0.44029280998787446,
      "grad_norm": 5.40625,
      "learning_rate": 3.0451113365095933e-05,
      "loss": 0.5703695297241211,
      "mean_token_accuracy": 0.8509207278490066,
      "num_tokens": 48706691.0,
      "step": 24510
    },
    {
      "entropy": 0.5600882176309824,
      "epoch": 0.4404724480172453,
      "grad_norm": 5.28125,
      "learning_rate": 3.043709145609493e-05,
      "loss": 0.5430359840393066,
      "mean_token_accuracy": 0.8666277050971984,
      "num_tokens": 48726035.0,
      "step": 24520
    },
    {
      "entropy": 0.6792589701712132,
      "epoch": 0.44065208604661604,
      "grad_norm": 6.21875,
      "learning_rate": 3.042306775153359e-05,
      "loss": 0.6821471214294433,
      "mean_token_accuracy": 0.8340050727128983,
      "num_tokens": 48746045.0,
      "step": 24530
    },
    {
      "entropy": 0.6152332082390786,
      "epoch": 0.44083172407598686,
      "grad_norm": 8.125,
      "learning_rate": 3.0409042256043164e-05,
      "loss": 0.6431283473968505,
      "mean_token_accuracy": 0.8401598826050758,
      "num_tokens": 48766269.0,
      "step": 24540
    },
    {
      "entropy": 0.7134300217032432,
      "epoch": 0.4410113621053577,
      "grad_norm": 6.6875,
      "learning_rate": 3.0395014974255454e-05,
      "loss": 0.7360266208648681,
      "mean_token_accuracy": 0.818088473379612,
      "num_tokens": 48785585.0,
      "step": 24550
    },
    {
      "entropy": 0.6341461170464754,
      "epoch": 0.4411910001347285,
      "grad_norm": 6.84375,
      "learning_rate": 3.0380985910802885e-05,
      "loss": 0.6664112091064454,
      "mean_token_accuracy": 0.8312144741415978,
      "num_tokens": 48806205.0,
      "step": 24560
    },
    {
      "entropy": 0.6343520622700453,
      "epoch": 0.4413706381640993,
      "grad_norm": 5.78125,
      "learning_rate": 3.036695507031844e-05,
      "loss": 0.6160861492156983,
      "mean_token_accuracy": 0.8393663197755814,
      "num_tokens": 48825212.0,
      "step": 24570
    },
    {
      "entropy": 0.6558980982750654,
      "epoch": 0.44155027619347015,
      "grad_norm": 8.0625,
      "learning_rate": 3.0352922457435707e-05,
      "loss": 0.6846744537353515,
      "mean_token_accuracy": 0.8371525272727013,
      "num_tokens": 48844033.0,
      "step": 24580
    },
    {
      "entropy": 0.7383323162794113,
      "epoch": 0.44172991422284097,
      "grad_norm": 9.0,
      "learning_rate": 3.033888807678886e-05,
      "loss": 0.7554762840270997,
      "mean_token_accuracy": 0.810349977016449,
      "num_tokens": 48863759.0,
      "step": 24590
    },
    {
      "entropy": 0.6707283407449722,
      "epoch": 0.4419095522522118,
      "grad_norm": 6.21875,
      "learning_rate": 3.0324851933012644e-05,
      "loss": 0.6680423736572265,
      "mean_token_accuracy": 0.8358715415000916,
      "num_tokens": 48882898.0,
      "step": 24600
    },
    {
      "entropy": 0.6419709332287311,
      "epoch": 0.4420891902815826,
      "grad_norm": 6.5625,
      "learning_rate": 3.0310814030742406e-05,
      "loss": 0.6242598056793213,
      "mean_token_accuracy": 0.8465228125452995,
      "num_tokens": 48903964.0,
      "step": 24610
    },
    {
      "entropy": 0.6546909786760807,
      "epoch": 0.44226882831095343,
      "grad_norm": 7.65625,
      "learning_rate": 3.0296774374614046e-05,
      "loss": 0.6794456005096435,
      "mean_token_accuracy": 0.8325377434492112,
      "num_tokens": 48924771.0,
      "step": 24620
    },
    {
      "entropy": 0.6202659197151661,
      "epoch": 0.44244846634032425,
      "grad_norm": 7.6875,
      "learning_rate": 3.0282732969264082e-05,
      "loss": 0.6098038196563721,
      "mean_token_accuracy": 0.8548227608203888,
      "num_tokens": 48945621.0,
      "step": 24630
    },
    {
      "entropy": 0.6240485176444054,
      "epoch": 0.4426281043696951,
      "grad_norm": 6.8125,
      "learning_rate": 3.026868981932956e-05,
      "loss": 0.6937830924987793,
      "mean_token_accuracy": 0.8319880947470665,
      "num_tokens": 48965034.0,
      "step": 24640
    },
    {
      "entropy": 0.7398942716419696,
      "epoch": 0.4428077423990659,
      "grad_norm": 7.125,
      "learning_rate": 3.0254644929448156e-05,
      "loss": 0.7947087287902832,
      "mean_token_accuracy": 0.8174772322177887,
      "num_tokens": 48986022.0,
      "step": 24650
    },
    {
      "entropy": 0.6909448966383934,
      "epoch": 0.4429873804284367,
      "grad_norm": 8.5,
      "learning_rate": 3.0240598304258078e-05,
      "loss": 0.6848007678985596,
      "mean_token_accuracy": 0.8385121032595635,
      "num_tokens": 49005820.0,
      "step": 24660
    },
    {
      "entropy": 0.6841110631823539,
      "epoch": 0.44316701845780754,
      "grad_norm": 6.09375,
      "learning_rate": 3.0226549948398126e-05,
      "loss": 0.6714051723480224,
      "mean_token_accuracy": 0.8297701433300972,
      "num_tokens": 49026623.0,
      "step": 24670
    },
    {
      "entropy": 0.6862603195011616,
      "epoch": 0.44334665648717836,
      "grad_norm": 4.96875,
      "learning_rate": 3.021249986650766e-05,
      "loss": 0.7057745456695557,
      "mean_token_accuracy": 0.8246428772807122,
      "num_tokens": 49046777.0,
      "step": 24680
    },
    {
      "entropy": 0.6537864200770855,
      "epoch": 0.4435262945165492,
      "grad_norm": 6.375,
      "learning_rate": 3.0198448063226635e-05,
      "loss": 0.6506913661956787,
      "mean_token_accuracy": 0.8443417236208915,
      "num_tokens": 49066261.0,
      "step": 24690
    },
    {
      "entropy": 0.7047199092805385,
      "epoch": 0.44370593254591995,
      "grad_norm": 6.8125,
      "learning_rate": 3.018439454319555e-05,
      "loss": 0.7087321281433105,
      "mean_token_accuracy": 0.8283670008182525,
      "num_tokens": 49086343.0,
      "step": 24700
    },
    {
      "entropy": 0.6323735993355513,
      "epoch": 0.44388557057529077,
      "grad_norm": 5.75,
      "learning_rate": 3.017033931105548e-05,
      "loss": 0.6258208274841308,
      "mean_token_accuracy": 0.8504282802343368,
      "num_tokens": 49105564.0,
      "step": 24710
    },
    {
      "entropy": 0.6451583333313465,
      "epoch": 0.4440652086046616,
      "grad_norm": 5.6875,
      "learning_rate": 3.0156282371448053e-05,
      "loss": 0.6097512722015381,
      "mean_token_accuracy": 0.8471232742071152,
      "num_tokens": 49124530.0,
      "step": 24720
    },
    {
      "entropy": 0.6597867157310248,
      "epoch": 0.4442448466340324,
      "grad_norm": 6.75,
      "learning_rate": 3.0142223729015488e-05,
      "loss": 0.6635383129119873,
      "mean_token_accuracy": 0.8407457947731019,
      "num_tokens": 49144100.0,
      "step": 24730
    },
    {
      "entropy": 0.6090733125805855,
      "epoch": 0.44442448466340323,
      "grad_norm": 8.25,
      "learning_rate": 3.0128163388400544e-05,
      "loss": 0.6453722953796387,
      "mean_token_accuracy": 0.834476524591446,
      "num_tokens": 49164086.0,
      "step": 24740
    },
    {
      "entropy": 0.7093131553381682,
      "epoch": 0.44460412269277405,
      "grad_norm": 5.9375,
      "learning_rate": 3.011410135424655e-05,
      "loss": 0.7638623714447021,
      "mean_token_accuracy": 0.8188984617590904,
      "num_tokens": 49184118.0,
      "step": 24750
    },
    {
      "entropy": 0.7432346932590008,
      "epoch": 0.44478376072214487,
      "grad_norm": 6.8125,
      "learning_rate": 3.0100037631197393e-05,
      "loss": 0.7585359573364258,
      "mean_token_accuracy": 0.8147079572081566,
      "num_tokens": 49204244.0,
      "step": 24760
    },
    {
      "entropy": 0.6366347067058087,
      "epoch": 0.4449633987515157,
      "grad_norm": 4.375,
      "learning_rate": 3.0085972223897503e-05,
      "loss": 0.6233406066894531,
      "mean_token_accuracy": 0.8406553149223328,
      "num_tokens": 49225128.0,
      "step": 24770
    },
    {
      "entropy": 0.7200527563691139,
      "epoch": 0.4451430367808865,
      "grad_norm": 6.125,
      "learning_rate": 3.0071905136991895e-05,
      "loss": 0.6957122802734375,
      "mean_token_accuracy": 0.8265476405620575,
      "num_tokens": 49245120.0,
      "step": 24780
    },
    {
      "entropy": 0.5662011288106441,
      "epoch": 0.44532267481025734,
      "grad_norm": 6.34375,
      "learning_rate": 3.0057836375126116e-05,
      "loss": 0.5451754093170166,
      "mean_token_accuracy": 0.860263068974018,
      "num_tokens": 49264458.0,
      "step": 24790
    },
    {
      "entropy": 0.7338626340031624,
      "epoch": 0.44550231283962816,
      "grad_norm": 5.71875,
      "learning_rate": 3.004376594294628e-05,
      "loss": 0.7422623634338379,
      "mean_token_accuracy": 0.8200034037232399,
      "num_tokens": 49283471.0,
      "step": 24800
    },
    {
      "entropy": 0.7104780457913875,
      "epoch": 0.445681950868999,
      "grad_norm": 6.0625,
      "learning_rate": 3.0029693845099028e-05,
      "loss": 0.7348772525787354,
      "mean_token_accuracy": 0.8238610781729221,
      "num_tokens": 49302765.0,
      "step": 24810
    },
    {
      "entropy": 0.6248294528573751,
      "epoch": 0.4458615888983698,
      "grad_norm": 7.84375,
      "learning_rate": 3.0015620086231595e-05,
      "loss": 0.5852408885955811,
      "mean_token_accuracy": 0.8594656825065613,
      "num_tokens": 49322787.0,
      "step": 24820
    },
    {
      "entropy": 0.5683596175163984,
      "epoch": 0.4460412269277406,
      "grad_norm": 5.5625,
      "learning_rate": 3.0001544670991716e-05,
      "loss": 0.5722835540771485,
      "mean_token_accuracy": 0.8501962542533874,
      "num_tokens": 49342315.0,
      "step": 24830
    },
    {
      "entropy": 0.6376261048018932,
      "epoch": 0.44622086495711144,
      "grad_norm": 5.84375,
      "learning_rate": 2.998746760402771e-05,
      "loss": 0.6271524429321289,
      "mean_token_accuracy": 0.8414926365017891,
      "num_tokens": 49361358.0,
      "step": 24840
    },
    {
      "entropy": 0.6727514907717704,
      "epoch": 0.44640050298648226,
      "grad_norm": 6.53125,
      "learning_rate": 2.9973388889988418e-05,
      "loss": 0.7007735729217529,
      "mean_token_accuracy": 0.8299114316701889,
      "num_tokens": 49380593.0,
      "step": 24850
    },
    {
      "entropy": 0.6736725352704525,
      "epoch": 0.4465801410158531,
      "grad_norm": 6.1875,
      "learning_rate": 2.995930853352324e-05,
      "loss": 0.7083977699279785,
      "mean_token_accuracy": 0.8343409970402718,
      "num_tokens": 49400816.0,
      "step": 24860
    },
    {
      "entropy": 0.6473349504172802,
      "epoch": 0.44675977904522385,
      "grad_norm": 6.90625,
      "learning_rate": 2.9945226539282113e-05,
      "loss": 0.6853185176849366,
      "mean_token_accuracy": 0.8376095771789551,
      "num_tokens": 49420615.0,
      "step": 24870
    },
    {
      "entropy": 0.600874425470829,
      "epoch": 0.44693941707459467,
      "grad_norm": 6.78125,
      "learning_rate": 2.9931142911915512e-05,
      "loss": 0.6223116874694824,
      "mean_token_accuracy": 0.8470140606164932,
      "num_tokens": 49439231.0,
      "step": 24880
    },
    {
      "entropy": 0.642533902078867,
      "epoch": 0.4471190551039655,
      "grad_norm": 5.65625,
      "learning_rate": 2.991705765607445e-05,
      "loss": 0.6592984676361084,
      "mean_token_accuracy": 0.8327147260308265,
      "num_tokens": 49458789.0,
      "step": 24890
    },
    {
      "entropy": 0.7062618792057037,
      "epoch": 0.4472986931333363,
      "grad_norm": 5.6875,
      "learning_rate": 2.9902970776410484e-05,
      "loss": 0.7163824081420899,
      "mean_token_accuracy": 0.832091948390007,
      "num_tokens": 49478639.0,
      "step": 24900
    },
    {
      "entropy": 0.6589072704315185,
      "epoch": 0.44747833116270713,
      "grad_norm": 6.375,
      "learning_rate": 2.9888882277575707e-05,
      "loss": 0.6528713703155518,
      "mean_token_accuracy": 0.8422851070761681,
      "num_tokens": 49498727.0,
      "step": 24910
    },
    {
      "entropy": 0.6145786799490451,
      "epoch": 0.44765796919207795,
      "grad_norm": 4.96875,
      "learning_rate": 2.9874792164222736e-05,
      "loss": 0.587584114074707,
      "mean_token_accuracy": 0.8557287961244583,
      "num_tokens": 49517793.0,
      "step": 24920
    },
    {
      "entropy": 0.636996253579855,
      "epoch": 0.4478376072214488,
      "grad_norm": 7.15625,
      "learning_rate": 2.9860700441004746e-05,
      "loss": 0.6754997253417969,
      "mean_token_accuracy": 0.8291559904813767,
      "num_tokens": 49537370.0,
      "step": 24930
    },
    {
      "entropy": 0.612726590782404,
      "epoch": 0.4480172452508196,
      "grad_norm": 4.15625,
      "learning_rate": 2.9846607112575414e-05,
      "loss": 0.6053715229034424,
      "mean_token_accuracy": 0.8496888294816017,
      "num_tokens": 49556941.0,
      "step": 24940
    },
    {
      "entropy": 0.6087141796946526,
      "epoch": 0.4481968832801904,
      "grad_norm": 6.46875,
      "learning_rate": 2.9832512183588962e-05,
      "loss": 0.6325457572937012,
      "mean_token_accuracy": 0.8408694788813591,
      "num_tokens": 49576599.0,
      "step": 24950
    },
    {
      "entropy": 0.64008983746171,
      "epoch": 0.44837652130956124,
      "grad_norm": 6.46875,
      "learning_rate": 2.9818415658700145e-05,
      "loss": 0.6202704906463623,
      "mean_token_accuracy": 0.8437325850129127,
      "num_tokens": 49596679.0,
      "step": 24960
    },
    {
      "entropy": 0.604790635406971,
      "epoch": 0.44855615933893206,
      "grad_norm": 9.625,
      "learning_rate": 2.9804317542564232e-05,
      "loss": 0.6460066795349121,
      "mean_token_accuracy": 0.843159033358097,
      "num_tokens": 49615915.0,
      "step": 24970
    },
    {
      "entropy": 0.5462105844169856,
      "epoch": 0.4487357973683029,
      "grad_norm": 6.09375,
      "learning_rate": 2.9790217839837024e-05,
      "loss": 0.583113431930542,
      "mean_token_accuracy": 0.8569620534777641,
      "num_tokens": 49635193.0,
      "step": 24980
    },
    {
      "entropy": 0.6619690850377082,
      "epoch": 0.4489154353976737,
      "grad_norm": 5.0625,
      "learning_rate": 2.9776116555174848e-05,
      "loss": 0.6237630367279052,
      "mean_token_accuracy": 0.8396276220679283,
      "num_tokens": 49655352.0,
      "step": 24990
    },
    {
      "epoch": 0.4490950734270445,
      "eval_entropy": 0.6292575550675392,
      "eval_loss": 0.6323664784431458,
      "eval_mean_token_accuracy": 0.8447910149097443,
      "eval_num_tokens": 49674772.0,
      "eval_runtime": 40.5094,
      "eval_samples_per_second": 49.371,
      "eval_steps_per_second": 6.171,
      "step": 25000
    },
    {
      "entropy": 0.638684949092567,
      "epoch": 0.44927471145641534,
      "grad_norm": 4.78125,
      "learning_rate": 2.9747909258673506e-05,
      "loss": 0.6595239639282227,
      "mean_token_accuracy": 0.836548687517643,
      "num_tokens": 49694382.0,
      "step": 25010
    },
    {
      "entropy": 0.7055387504398822,
      "epoch": 0.44945434948578616,
      "grad_norm": 5.78125,
      "learning_rate": 2.9733803256149612e-05,
      "loss": 0.7578374862670898,
      "mean_token_accuracy": 0.8209755778312683,
      "num_tokens": 49714637.0,
      "step": 25020
    },
    {
      "entropy": 0.6387829348444939,
      "epoch": 0.449633987515157,
      "grad_norm": 6.0,
      "learning_rate": 2.9719695690321243e-05,
      "loss": 0.6398033142089844,
      "mean_token_accuracy": 0.8389146134257317,
      "num_tokens": 49733850.0,
      "step": 25030
    },
    {
      "entropy": 0.643056233972311,
      "epoch": 0.44981362554452775,
      "grad_norm": 7.90625,
      "learning_rate": 2.9705586565847347e-05,
      "loss": 0.6638535499572754,
      "mean_token_accuracy": 0.8431431859731674,
      "num_tokens": 49753411.0,
      "step": 25040
    },
    {
      "entropy": 0.5584363348782062,
      "epoch": 0.44999326357389857,
      "grad_norm": 8.125,
      "learning_rate": 2.969147588738735e-05,
      "loss": 0.5573175430297852,
      "mean_token_accuracy": 0.8553016409277916,
      "num_tokens": 49772143.0,
      "step": 25050
    },
    {
      "entropy": 0.6667997360229492,
      "epoch": 0.4501729016032694,
      "grad_norm": 9.1875,
      "learning_rate": 2.967736365960121e-05,
      "loss": 0.6630809307098389,
      "mean_token_accuracy": 0.8395236819982529,
      "num_tokens": 49791977.0,
      "step": 25060
    },
    {
      "entropy": 0.674366020783782,
      "epoch": 0.4503525396326402,
      "grad_norm": 6.59375,
      "learning_rate": 2.966324988714938e-05,
      "loss": 0.6595981597900391,
      "mean_token_accuracy": 0.8447778016328812,
      "num_tokens": 49810932.0,
      "step": 25070
    },
    {
      "entropy": 0.6593878965824842,
      "epoch": 0.45053217766201104,
      "grad_norm": 6.53125,
      "learning_rate": 2.9649134574692834e-05,
      "loss": 0.6435557842254639,
      "mean_token_accuracy": 0.8381580978631973,
      "num_tokens": 49831365.0,
      "step": 25080
    },
    {
      "entropy": 0.6742773965001106,
      "epoch": 0.45071181569138186,
      "grad_norm": 5.59375,
      "learning_rate": 2.9635017726893057e-05,
      "loss": 0.6652705669403076,
      "mean_token_accuracy": 0.8452425882220268,
      "num_tokens": 49850210.0,
      "step": 25090
    },
    {
      "entropy": 0.6782958213239908,
      "epoch": 0.4508914537207527,
      "grad_norm": 7.125,
      "learning_rate": 2.9620899348412028e-05,
      "loss": 0.6942856311798096,
      "mean_token_accuracy": 0.8309721395373344,
      "num_tokens": 49870124.0,
      "step": 25100
    },
    {
      "entropy": 0.666462529078126,
      "epoch": 0.4510710917501235,
      "grad_norm": 6.4375,
      "learning_rate": 2.960677944391226e-05,
      "loss": 0.6675957679748535,
      "mean_token_accuracy": 0.8296088501811028,
      "num_tokens": 49890310.0,
      "step": 25110
    },
    {
      "entropy": 0.629999179393053,
      "epoch": 0.4512507297794943,
      "grad_norm": 5.15625,
      "learning_rate": 2.9592658018056717e-05,
      "loss": 0.6155696868896484,
      "mean_token_accuracy": 0.849763223528862,
      "num_tokens": 49910540.0,
      "step": 25120
    },
    {
      "entropy": 0.6640816450119018,
      "epoch": 0.45143036780886514,
      "grad_norm": 6.71875,
      "learning_rate": 2.957853507550893e-05,
      "loss": 0.6632445335388184,
      "mean_token_accuracy": 0.8374566316604615,
      "num_tokens": 49929541.0,
      "step": 25130
    },
    {
      "entropy": 0.6631264202296734,
      "epoch": 0.45161000583823596,
      "grad_norm": 5.34375,
      "learning_rate": 2.9564410620932876e-05,
      "loss": 0.6752084732055664,
      "mean_token_accuracy": 0.8368457749485969,
      "num_tokens": 49948952.0,
      "step": 25140
    },
    {
      "entropy": 0.656722454726696,
      "epoch": 0.4517896438676068,
      "grad_norm": 9.1875,
      "learning_rate": 2.9550284658993066e-05,
      "loss": 0.6669083595275879,
      "mean_token_accuracy": 0.834371255338192,
      "num_tokens": 49968963.0,
      "step": 25150
    },
    {
      "entropy": 0.631779707968235,
      "epoch": 0.4519692818969776,
      "grad_norm": 4.8125,
      "learning_rate": 2.95361571943545e-05,
      "loss": 0.6154253005981445,
      "mean_token_accuracy": 0.8475814059376716,
      "num_tokens": 49990747.0,
      "step": 25160
    },
    {
      "entropy": 0.7300349719822407,
      "epoch": 0.4521489199263484,
      "grad_norm": 7.78125,
      "learning_rate": 2.9522028231682664e-05,
      "loss": 0.727361011505127,
      "mean_token_accuracy": 0.8220221385359764,
      "num_tokens": 50010828.0,
      "step": 25170
    },
    {
      "entropy": 0.5861820664256812,
      "epoch": 0.45232855795571925,
      "grad_norm": 10.25,
      "learning_rate": 2.950789777564355e-05,
      "loss": 0.6004938125610352,
      "mean_token_accuracy": 0.8485207810997963,
      "num_tokens": 50031216.0,
      "step": 25180
    },
    {
      "entropy": 0.5503634229302407,
      "epoch": 0.45250819598509007,
      "grad_norm": 7.59375,
      "learning_rate": 2.949376583090363e-05,
      "loss": 0.526863956451416,
      "mean_token_accuracy": 0.8697680130600929,
      "num_tokens": 50051311.0,
      "step": 25190
    },
    {
      "entropy": 0.6283725298941135,
      "epoch": 0.4526878340144609,
      "grad_norm": 7.875,
      "learning_rate": 2.9479632402129904e-05,
      "loss": 0.63141770362854,
      "mean_token_accuracy": 0.8399059519171714,
      "num_tokens": 50071451.0,
      "step": 25200
    },
    {
      "entropy": 0.627512489259243,
      "epoch": 0.45286747204383165,
      "grad_norm": 6.53125,
      "learning_rate": 2.94654974939898e-05,
      "loss": 0.6458745956420898,
      "mean_token_accuracy": 0.8321477234363556,
      "num_tokens": 50091831.0,
      "step": 25210
    },
    {
      "entropy": 0.6924109339714051,
      "epoch": 0.4530471100732025,
      "grad_norm": 6.5,
      "learning_rate": 2.9451361111151293e-05,
      "loss": 0.6783711433410644,
      "mean_token_accuracy": 0.8283683940768242,
      "num_tokens": 50111958.0,
      "step": 25220
    },
    {
      "entropy": 0.6518695548176765,
      "epoch": 0.4532267481025733,
      "grad_norm": 6.53125,
      "learning_rate": 2.943722325828281e-05,
      "loss": 0.5924414157867431,
      "mean_token_accuracy": 0.8441137000918388,
      "num_tokens": 50132302.0,
      "step": 25230
    },
    {
      "entropy": 0.6240761388093233,
      "epoch": 0.4534063861319441,
      "grad_norm": 7.25,
      "learning_rate": 2.9423083940053285e-05,
      "loss": 0.6506386756896972,
      "mean_token_accuracy": 0.8441406473517418,
      "num_tokens": 50152830.0,
      "step": 25240
    },
    {
      "entropy": 0.6854124598205089,
      "epoch": 0.45358602416131494,
      "grad_norm": 10.6875,
      "learning_rate": 2.9408943161132124e-05,
      "loss": 0.7455312728881835,
      "mean_token_accuracy": 0.8255221836268902,
      "num_tokens": 50172015.0,
      "step": 25250
    },
    {
      "entropy": 0.5960974767804146,
      "epoch": 0.45376566219068576,
      "grad_norm": 6.25,
      "learning_rate": 2.9394800926189205e-05,
      "loss": 0.6277171611785889,
      "mean_token_accuracy": 0.8464951530098915,
      "num_tokens": 50191436.0,
      "step": 25260
    },
    {
      "entropy": 0.5768345851451159,
      "epoch": 0.4539453002200566,
      "grad_norm": 6.09375,
      "learning_rate": 2.9380657239894917e-05,
      "loss": 0.5919283390045166,
      "mean_token_accuracy": 0.8517137989401817,
      "num_tokens": 50211268.0,
      "step": 25270
    },
    {
      "entropy": 0.5933682758361101,
      "epoch": 0.4541249382494274,
      "grad_norm": 5.90625,
      "learning_rate": 2.9366512106920095e-05,
      "loss": 0.5840302467346191,
      "mean_token_accuracy": 0.8495647132396698,
      "num_tokens": 50231493.0,
      "step": 25280
    },
    {
      "entropy": 0.6076404370367527,
      "epoch": 0.4543045762787982,
      "grad_norm": 5.875,
      "learning_rate": 2.935236553193609e-05,
      "loss": 0.6132025241851806,
      "mean_token_accuracy": 0.8435410961508751,
      "num_tokens": 50251284.0,
      "step": 25290
    },
    {
      "entropy": 0.7315272923558951,
      "epoch": 0.45448421430816904,
      "grad_norm": 5.21875,
      "learning_rate": 2.9338217519614675e-05,
      "loss": 0.7790263175964356,
      "mean_token_accuracy": 0.820195198059082,
      "num_tokens": 50270751.0,
      "step": 25300
    },
    {
      "entropy": 0.6670715101063251,
      "epoch": 0.45466385233753986,
      "grad_norm": 6.1875,
      "learning_rate": 2.932406807462815e-05,
      "loss": 0.6770613193511963,
      "mean_token_accuracy": 0.8367348626255989,
      "num_tokens": 50291656.0,
      "step": 25310
    },
    {
      "entropy": 0.6457675602287054,
      "epoch": 0.4548434903669107,
      "grad_norm": 5.40625,
      "learning_rate": 2.9309917201649266e-05,
      "loss": 0.612489652633667,
      "mean_token_accuracy": 0.844483008980751,
      "num_tokens": 50310552.0,
      "step": 25320
    },
    {
      "entropy": 0.6395229611545801,
      "epoch": 0.4550231283962815,
      "grad_norm": 6.46875,
      "learning_rate": 2.9295764905351247e-05,
      "loss": 0.6768513202667237,
      "mean_token_accuracy": 0.848475331068039,
      "num_tokens": 50330173.0,
      "step": 25330
    },
    {
      "entropy": 0.5990995306521654,
      "epoch": 0.4552027664256523,
      "grad_norm": 9.3125,
      "learning_rate": 2.9281611190407783e-05,
      "loss": 0.601787805557251,
      "mean_token_accuracy": 0.8497443243861198,
      "num_tokens": 50349617.0,
      "step": 25340
    },
    {
      "entropy": 0.693512424826622,
      "epoch": 0.45538240445502315,
      "grad_norm": 5.84375,
      "learning_rate": 2.9267456061493037e-05,
      "loss": 0.7133592128753662,
      "mean_token_accuracy": 0.828156964480877,
      "num_tokens": 50370045.0,
      "step": 25350
    },
    {
      "entropy": 0.6325917668640614,
      "epoch": 0.45556204248439397,
      "grad_norm": 7.96875,
      "learning_rate": 2.925329952328163e-05,
      "loss": 0.6935085773468017,
      "mean_token_accuracy": 0.829437917470932,
      "num_tokens": 50390556.0,
      "step": 25360
    },
    {
      "entropy": 0.6913409329950809,
      "epoch": 0.4557416805137648,
      "grad_norm": 6.375,
      "learning_rate": 2.9239141580448675e-05,
      "loss": 0.7083559513092041,
      "mean_token_accuracy": 0.8345638900995255,
      "num_tokens": 50410548.0,
      "step": 25370
    },
    {
      "entropy": 0.637165478989482,
      "epoch": 0.45592131854313556,
      "grad_norm": 7.46875,
      "learning_rate": 2.9224982237669723e-05,
      "loss": 0.6610452651977539,
      "mean_token_accuracy": 0.8335252806544304,
      "num_tokens": 50429588.0,
      "step": 25380
    },
    {
      "entropy": 0.6131776116788388,
      "epoch": 0.4561009565725064,
      "grad_norm": 4.71875,
      "learning_rate": 2.9210821499620787e-05,
      "loss": 0.5746246337890625,
      "mean_token_accuracy": 0.8555165618658066,
      "num_tokens": 50449342.0,
      "step": 25390
    },
    {
      "entropy": 0.6258339460939168,
      "epoch": 0.4562805946018772,
      "grad_norm": 7.625,
      "learning_rate": 2.9196659370978363e-05,
      "loss": 0.6317354679107666,
      "mean_token_accuracy": 0.8398290351033211,
      "num_tokens": 50468952.0,
      "step": 25400
    },
    {
      "entropy": 0.5593230076134205,
      "epoch": 0.456460232631248,
      "grad_norm": 4.875,
      "learning_rate": 2.9182495856419377e-05,
      "loss": 0.5120494842529297,
      "mean_token_accuracy": 0.8629510864615441,
      "num_tokens": 50490052.0,
      "step": 25410
    },
    {
      "entropy": 0.6611490480601787,
      "epoch": 0.45663987066061884,
      "grad_norm": 5.625,
      "learning_rate": 2.9168330960621243e-05,
      "loss": 0.7093032836914063,
      "mean_token_accuracy": 0.8329256534576416,
      "num_tokens": 50510165.0,
      "step": 25420
    },
    {
      "entropy": 0.6370770268142223,
      "epoch": 0.45681950868998966,
      "grad_norm": 5.3125,
      "learning_rate": 2.9154164688261803e-05,
      "loss": 0.6011889934539795,
      "mean_token_accuracy": 0.8522447913885116,
      "num_tokens": 50529615.0,
      "step": 25430
    },
    {
      "entropy": 0.645000410079956,
      "epoch": 0.4569991467193605,
      "grad_norm": 6.125,
      "learning_rate": 2.9139997044019375e-05,
      "loss": 0.6536627769470215,
      "mean_token_accuracy": 0.8377035349607468,
      "num_tokens": 50550132.0,
      "step": 25440
    },
    {
      "entropy": 0.7422636657953262,
      "epoch": 0.4571787847487313,
      "grad_norm": 4.8125,
      "learning_rate": 2.9125828032572718e-05,
      "loss": 0.8207412719726562,
      "mean_token_accuracy": 0.8069488435983658,
      "num_tokens": 50571378.0,
      "step": 25450
    },
    {
      "entropy": 0.6084121130406857,
      "epoch": 0.4573584227781021,
      "grad_norm": 5.5625,
      "learning_rate": 2.9111657658601043e-05,
      "loss": 0.621738052368164,
      "mean_token_accuracy": 0.846948866546154,
      "num_tokens": 50591551.0,
      "step": 25460
    },
    {
      "entropy": 0.5886890947818756,
      "epoch": 0.45753806080747295,
      "grad_norm": 6.15625,
      "learning_rate": 2.909748592678402e-05,
      "loss": 0.5699174404144287,
      "mean_token_accuracy": 0.8514479652047158,
      "num_tokens": 50611327.0,
      "step": 25470
    },
    {
      "entropy": 0.6397239524871111,
      "epoch": 0.45771769883684377,
      "grad_norm": 3.78125,
      "learning_rate": 2.908331284180176e-05,
      "loss": 0.6495782852172851,
      "mean_token_accuracy": 0.8374035209417343,
      "num_tokens": 50631131.0,
      "step": 25480
    },
    {
      "entropy": 0.6873040594160557,
      "epoch": 0.4578973368662146,
      "grad_norm": 5.28125,
      "learning_rate": 2.9069138408334817e-05,
      "loss": 0.6738361358642578,
      "mean_token_accuracy": 0.8331977978348732,
      "num_tokens": 50651039.0,
      "step": 25490
    },
    {
      "entropy": 0.6267948567867279,
      "epoch": 0.4580769748955854,
      "grad_norm": 5.28125,
      "learning_rate": 2.9054962631064196e-05,
      "loss": 0.6212169170379639,
      "mean_token_accuracy": 0.840568694472313,
      "num_tokens": 50672373.0,
      "step": 25500
    },
    {
      "entropy": 0.6677661683410406,
      "epoch": 0.45825661292495623,
      "grad_norm": 5.78125,
      "learning_rate": 2.9040785514671358e-05,
      "loss": 0.7249666213989258,
      "mean_token_accuracy": 0.825019097328186,
      "num_tokens": 50693258.0,
      "step": 25510
    },
    {
      "entropy": 0.5941379114985466,
      "epoch": 0.45843625095432705,
      "grad_norm": 5.0625,
      "learning_rate": 2.9026607063838186e-05,
      "loss": 0.572938060760498,
      "mean_token_accuracy": 0.8503668770194054,
      "num_tokens": 50712640.0,
      "step": 25520
    },
    {
      "entropy": 0.6403566155582666,
      "epoch": 0.4586158889836979,
      "grad_norm": 6.5625,
      "learning_rate": 2.9012427283247005e-05,
      "loss": 0.632210636138916,
      "mean_token_accuracy": 0.8391883745789528,
      "num_tokens": 50731832.0,
      "step": 25530
    },
    {
      "entropy": 0.600138621032238,
      "epoch": 0.4587955270130687,
      "grad_norm": 7.0625,
      "learning_rate": 2.899824617758059e-05,
      "loss": 0.5980629444122314,
      "mean_token_accuracy": 0.8472205772995949,
      "num_tokens": 50752141.0,
      "step": 25540
    },
    {
      "entropy": 0.7062356412410736,
      "epoch": 0.45897516504243946,
      "grad_norm": 6.625,
      "learning_rate": 2.898406375152215e-05,
      "loss": 0.7078725814819335,
      "mean_token_accuracy": 0.8208790302276612,
      "num_tokens": 50771039.0,
      "step": 25550
    },
    {
      "entropy": 0.6620148822665215,
      "epoch": 0.4591548030718103,
      "grad_norm": 8.0,
      "learning_rate": 2.896988000975533e-05,
      "loss": 0.6740068435668946,
      "mean_token_accuracy": 0.8431305274367332,
      "num_tokens": 50790402.0,
      "step": 25560
    },
    {
      "entropy": 0.5858115360140801,
      "epoch": 0.4593344411011811,
      "grad_norm": 6.15625,
      "learning_rate": 2.8955694956964212e-05,
      "loss": 0.6215967178344727,
      "mean_token_accuracy": 0.8493713468313218,
      "num_tokens": 50808872.0,
      "step": 25570
    },
    {
      "entropy": 0.7574009954929352,
      "epoch": 0.4595140791305519,
      "grad_norm": 7.09375,
      "learning_rate": 2.8941508597833295e-05,
      "loss": 0.7576849460601807,
      "mean_token_accuracy": 0.8182817742228508,
      "num_tokens": 50828996.0,
      "step": 25580
    },
    {
      "entropy": 0.6275342617183923,
      "epoch": 0.45969371715992274,
      "grad_norm": 7.71875,
      "learning_rate": 2.8927320937047536e-05,
      "loss": 0.6255920886993408,
      "mean_token_accuracy": 0.8419717267155647,
      "num_tokens": 50848034.0,
      "step": 25590
    },
    {
      "entropy": 0.6065071992576122,
      "epoch": 0.45987335518929356,
      "grad_norm": 4.5,
      "learning_rate": 2.8913131979292298e-05,
      "loss": 0.5861433029174805,
      "mean_token_accuracy": 0.8565980672836304,
      "num_tokens": 50868084.0,
      "step": 25600
    },
    {
      "entropy": 0.6757667608559131,
      "epoch": 0.4600529932186644,
      "grad_norm": 7.21875,
      "learning_rate": 2.8898941729253386e-05,
      "loss": 0.6780172348022461,
      "mean_token_accuracy": 0.8281867980957032,
      "num_tokens": 50886944.0,
      "step": 25610
    },
    {
      "entropy": 0.6902949158102274,
      "epoch": 0.4602326312480352,
      "grad_norm": 7.28125,
      "learning_rate": 2.888475019161704e-05,
      "loss": 0.6881210803985596,
      "mean_token_accuracy": 0.8257127717137337,
      "num_tokens": 50906240.0,
      "step": 25620
    },
    {
      "entropy": 0.6011094443500042,
      "epoch": 0.460412269277406,
      "grad_norm": 7.875,
      "learning_rate": 2.8870557371069884e-05,
      "loss": 0.6133516788482666,
      "mean_token_accuracy": 0.8461831539869309,
      "num_tokens": 50926740.0,
      "step": 25630
    },
    {
      "entropy": 0.6040114372968673,
      "epoch": 0.46059190730677685,
      "grad_norm": 5.25,
      "learning_rate": 2.8856363272299018e-05,
      "loss": 0.5971450328826904,
      "mean_token_accuracy": 0.8495015546679496,
      "num_tokens": 50945992.0,
      "step": 25640
    },
    {
      "entropy": 0.6934925317764282,
      "epoch": 0.46077154533614767,
      "grad_norm": 8.0625,
      "learning_rate": 2.8842167899991933e-05,
      "loss": 0.7048426151275635,
      "mean_token_accuracy": 0.8351691097021103,
      "num_tokens": 50965222.0,
      "step": 25650
    },
    {
      "entropy": 0.6329482924193144,
      "epoch": 0.4609511833655185,
      "grad_norm": 4.6875,
      "learning_rate": 2.882797125883657e-05,
      "loss": 0.6606556892395019,
      "mean_token_accuracy": 0.8396263301372529,
      "num_tokens": 50986202.0,
      "step": 25660
    },
    {
      "entropy": 0.5530278362333775,
      "epoch": 0.4611308213948893,
      "grad_norm": 6.96875,
      "learning_rate": 2.8813773353521223e-05,
      "loss": 0.5789451599121094,
      "mean_token_accuracy": 0.8534066557884217,
      "num_tokens": 51005457.0,
      "step": 25670
    },
    {
      "entropy": 0.6640733372420072,
      "epoch": 0.46131045942426013,
      "grad_norm": 5.15625,
      "learning_rate": 2.8799574188734685e-05,
      "loss": 0.6331408023834229,
      "mean_token_accuracy": 0.8493509814143181,
      "num_tokens": 51024106.0,
      "step": 25680
    },
    {
      "entropy": 0.7068748187273741,
      "epoch": 0.46149009745363095,
      "grad_norm": 6.03125,
      "learning_rate": 2.8785373769166113e-05,
      "loss": 0.7153264999389648,
      "mean_token_accuracy": 0.8256227418780326,
      "num_tokens": 51045094.0,
      "step": 25690
    },
    {
      "entropy": 0.5893899440765381,
      "epoch": 0.4616697354830018,
      "grad_norm": 5.28125,
      "learning_rate": 2.87711720995051e-05,
      "loss": 0.5735009193420411,
      "mean_token_accuracy": 0.8588044553995132,
      "num_tokens": 51064576.0,
      "step": 25700
    },
    {
      "entropy": 0.6313374683260917,
      "epoch": 0.4618493735123726,
      "grad_norm": 6.71875,
      "learning_rate": 2.875696918444164e-05,
      "loss": 0.6082898139953613,
      "mean_token_accuracy": 0.846271063387394,
      "num_tokens": 51085428.0,
      "step": 25710
    },
    {
      "entropy": 0.5920072223991155,
      "epoch": 0.46202901154174336,
      "grad_norm": 6.5,
      "learning_rate": 2.8742765028666142e-05,
      "loss": 0.6229467868804932,
      "mean_token_accuracy": 0.8434670001268387,
      "num_tokens": 51104251.0,
      "step": 25720
    },
    {
      "entropy": 0.6229615122079849,
      "epoch": 0.4622086495711142,
      "grad_norm": 6.625,
      "learning_rate": 2.8728559636869435e-05,
      "loss": 0.6378118991851807,
      "mean_token_accuracy": 0.8415612027049064,
      "num_tokens": 51124070.0,
      "step": 25730
    },
    {
      "entropy": 0.6763359542936087,
      "epoch": 0.462388287600485,
      "grad_norm": 6.15625,
      "learning_rate": 2.8714353013742736e-05,
      "loss": 0.6679882526397705,
      "mean_token_accuracy": 0.8349309280514717,
      "num_tokens": 51143765.0,
      "step": 25740
    },
    {
      "entropy": 0.669699078425765,
      "epoch": 0.4625679256298558,
      "grad_norm": 6.21875,
      "learning_rate": 2.8700145163977697e-05,
      "loss": 0.6854621887207031,
      "mean_token_accuracy": 0.8312308505177498,
      "num_tokens": 51163628.0,
      "step": 25750
    },
    {
      "entropy": 0.6926385834813118,
      "epoch": 0.46274756365922665,
      "grad_norm": 6.125,
      "learning_rate": 2.8685936092266334e-05,
      "loss": 0.6914441585540771,
      "mean_token_accuracy": 0.8288914635777473,
      "num_tokens": 51183202.0,
      "step": 25760
    },
    {
      "entropy": 0.6186100497841835,
      "epoch": 0.46292720168859747,
      "grad_norm": 6.71875,
      "learning_rate": 2.8671725803301114e-05,
      "loss": 0.6753859043121337,
      "mean_token_accuracy": 0.8333504065871239,
      "num_tokens": 51202883.0,
      "step": 25770
    },
    {
      "entropy": 0.6608428992331028,
      "epoch": 0.4631068397179683,
      "grad_norm": 5.34375,
      "learning_rate": 2.8657514301774864e-05,
      "loss": 0.6533806800842286,
      "mean_token_accuracy": 0.8425754457712173,
      "num_tokens": 51223523.0,
      "step": 25780
    },
    {
      "entropy": 0.6102218974381686,
      "epoch": 0.4632864777473391,
      "grad_norm": 5.46875,
      "learning_rate": 2.8643301592380843e-05,
      "loss": 0.6256093978881836,
      "mean_token_accuracy": 0.8441969484090805,
      "num_tokens": 51243262.0,
      "step": 25790
    },
    {
      "entropy": 0.8202163949608803,
      "epoch": 0.46346611577670993,
      "grad_norm": 6.1875,
      "learning_rate": 2.8629087679812693e-05,
      "loss": 0.7995344638824463,
      "mean_token_accuracy": 0.811702573299408,
      "num_tokens": 51263221.0,
      "step": 25800
    },
    {
      "entropy": 0.64853769633919,
      "epoch": 0.46364575380608075,
      "grad_norm": 9.0,
      "learning_rate": 2.861487256876445e-05,
      "loss": 0.6470497608184814,
      "mean_token_accuracy": 0.8406313449144364,
      "num_tokens": 51282773.0,
      "step": 25810
    },
    {
      "entropy": 0.7274887137115001,
      "epoch": 0.4638253918354516,
      "grad_norm": 5.84375,
      "learning_rate": 2.860065626393056e-05,
      "loss": 0.766373586654663,
      "mean_token_accuracy": 0.8170804768800736,
      "num_tokens": 51301855.0,
      "step": 25820
    },
    {
      "entropy": 0.657333143800497,
      "epoch": 0.4640050298648224,
      "grad_norm": 7.6875,
      "learning_rate": 2.8586438770005846e-05,
      "loss": 0.6868473529815674,
      "mean_token_accuracy": 0.8316267430782318,
      "num_tokens": 51321462.0,
      "step": 25830
    },
    {
      "entropy": 0.6249050987884402,
      "epoch": 0.4641846678941932,
      "grad_norm": 5.09375,
      "learning_rate": 2.8572220091685552e-05,
      "loss": 0.6120130538940429,
      "mean_token_accuracy": 0.8491566121578217,
      "num_tokens": 51341980.0,
      "step": 25840
    },
    {
      "entropy": 0.6045302528887987,
      "epoch": 0.46436430592356404,
      "grad_norm": 5.28125,
      "learning_rate": 2.8558000233665274e-05,
      "loss": 0.5650796890258789,
      "mean_token_accuracy": 0.8537971988320351,
      "num_tokens": 51362139.0,
      "step": 25850
    },
    {
      "entropy": 0.6725374810397625,
      "epoch": 0.46454394395293486,
      "grad_norm": 7.125,
      "learning_rate": 2.854377920064103e-05,
      "loss": 0.6571398258209229,
      "mean_token_accuracy": 0.8342468813061714,
      "num_tokens": 51382179.0,
      "step": 25860
    },
    {
      "entropy": 0.5756803572177887,
      "epoch": 0.4647235819823057,
      "grad_norm": 5.09375,
      "learning_rate": 2.8529556997309198e-05,
      "loss": 0.5507289886474609,
      "mean_token_accuracy": 0.8598543912172317,
      "num_tokens": 51401893.0,
      "step": 25870
    },
    {
      "entropy": 0.6082211524248123,
      "epoch": 0.4649032200116765,
      "grad_norm": 8.9375,
      "learning_rate": 2.8515333628366586e-05,
      "loss": 0.6282366275787353,
      "mean_token_accuracy": 0.8454065442085266,
      "num_tokens": 51421743.0,
      "step": 25880
    },
    {
      "entropy": 0.6294764690101147,
      "epoch": 0.46508285804104726,
      "grad_norm": 4.4375,
      "learning_rate": 2.850110909851033e-05,
      "loss": 0.6501500606536865,
      "mean_token_accuracy": 0.835572762787342,
      "num_tokens": 51440908.0,
      "step": 25890
    },
    {
      "entropy": 0.6060632511973381,
      "epoch": 0.4652624960704181,
      "grad_norm": 7.875,
      "learning_rate": 2.8486883412437998e-05,
      "loss": 0.6195859909057617,
      "mean_token_accuracy": 0.8399563387036324,
      "num_tokens": 51460793.0,
      "step": 25900
    },
    {
      "entropy": 0.6406848393380642,
      "epoch": 0.4654421340997889,
      "grad_norm": 7.21875,
      "learning_rate": 2.8472656574847508e-05,
      "loss": 0.6472595691680908,
      "mean_token_accuracy": 0.8409404948353767,
      "num_tokens": 51480438.0,
      "step": 25910
    },
    {
      "entropy": 0.6251276426017285,
      "epoch": 0.46562177212915973,
      "grad_norm": 6.15625,
      "learning_rate": 2.845842859043718e-05,
      "loss": 0.6895182609558106,
      "mean_token_accuracy": 0.832318739593029,
      "num_tokens": 51502089.0,
      "step": 25920
    },
    {
      "entropy": 0.6077856305986643,
      "epoch": 0.46580141015853055,
      "grad_norm": 6.34375,
      "learning_rate": 2.8444199463905703e-05,
      "loss": 0.5685009002685547,
      "mean_token_accuracy": 0.854977959394455,
      "num_tokens": 51522751.0,
      "step": 25930
    },
    {
      "entropy": 0.5694128222763538,
      "epoch": 0.46598104818790137,
      "grad_norm": 4.84375,
      "learning_rate": 2.8429969199952128e-05,
      "loss": 0.5380094051361084,
      "mean_token_accuracy": 0.8683346420526504,
      "num_tokens": 51542375.0,
      "step": 25940
    },
    {
      "entropy": 0.6811686966568231,
      "epoch": 0.4661606862172722,
      "grad_norm": 5.9375,
      "learning_rate": 2.841573780327592e-05,
      "loss": 0.6538914680480957,
      "mean_token_accuracy": 0.8386504486203193,
      "num_tokens": 51562880.0,
      "step": 25950
    },
    {
      "entropy": 0.6149995278567075,
      "epoch": 0.466340324246643,
      "grad_norm": 5.6875,
      "learning_rate": 2.8401505278576877e-05,
      "loss": 0.5841248989105224,
      "mean_token_accuracy": 0.8548590317368507,
      "num_tokens": 51584160.0,
      "step": 25960
    },
    {
      "entropy": 0.6974666863679886,
      "epoch": 0.46651996227601383,
      "grad_norm": 7.0,
      "learning_rate": 2.838727163055521e-05,
      "loss": 0.7242869377136231,
      "mean_token_accuracy": 0.817896381765604,
      "num_tokens": 51604486.0,
      "step": 25970
    },
    {
      "entropy": 0.6544119141995907,
      "epoch": 0.46669960030538465,
      "grad_norm": 6.5625,
      "learning_rate": 2.8373036863911457e-05,
      "loss": 0.6475766658782959,
      "mean_token_accuracy": 0.8401797011494636,
      "num_tokens": 51623418.0,
      "step": 25980
    },
    {
      "entropy": 0.6466917738318443,
      "epoch": 0.4668792383347555,
      "grad_norm": 7.1875,
      "learning_rate": 2.8358800983346556e-05,
      "loss": 0.6784713745117188,
      "mean_token_accuracy": 0.8389300957322121,
      "num_tokens": 51643318.0,
      "step": 25990
    },
    {
      "epoch": 0.4670588763641263,
      "eval_entropy": 0.6190758219957352,
      "eval_loss": 0.624586284160614,
      "eval_mean_token_accuracy": 0.8457643120288849,
      "eval_num_tokens": 51662963.0,
      "eval_runtime": 40.5724,
      "eval_samples_per_second": 49.295,
      "eval_steps_per_second": 6.162,
      "step": 26000
    }
  ],
  "logging_steps": 10,
  "max_steps": 55668,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 4.9179453258262e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}