{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5625,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.4598965495824814,
      "epoch": 0.0005333333333333334,
      "grad_norm": 0.8618218898773193,
      "learning_rate": 0.0,
      "loss": 0.4677274227142334,
      "mean_token_accuracy": 0.8419764935970306,
      "num_tokens": 8192.0,
      "step": 1
    },
    {
      "entropy": 0.4768468721045388,
      "epoch": 0.005333333333333333,
      "grad_norm": 1.1119650602340698,
      "learning_rate": 3.6e-05,
      "loss": 0.49025758107503253,
      "mean_token_accuracy": 0.8352630999353197,
      "num_tokens": 81920.0,
      "step": 10
    },
    {
      "entropy": 0.47109493613243103,
      "epoch": 0.010666666666666666,
      "grad_norm": 1.1768184900283813,
      "learning_rate": 7.6e-05,
      "loss": 0.4978835105895996,
      "mean_token_accuracy": 0.8347235828638077,
      "num_tokens": 163840.0,
      "step": 20
    },
    {
      "entropy": 0.4710109308362007,
      "epoch": 0.016,
      "grad_norm": 0.7324985265731812,
      "learning_rate": 0.000116,
      "loss": 0.4964810848236084,
      "mean_token_accuracy": 0.8343199610710144,
      "num_tokens": 245760.0,
      "step": 30
    },
    {
      "entropy": 0.4771491393446922,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.7841163873672485,
      "learning_rate": 0.00015600000000000002,
      "loss": 0.5007925987243652,
      "mean_token_accuracy": 0.835090509057045,
      "num_tokens": 327680.0,
      "step": 40
    },
    {
      "entropy": 0.47855959087610245,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.5469140410423279,
      "learning_rate": 0.000196,
      "loss": 0.5098973274230957,
      "mean_token_accuracy": 0.8313233852386475,
      "num_tokens": 409600.0,
      "step": 50
    },
    {
      "entropy": 0.4701948702335358,
      "epoch": 0.032,
      "grad_norm": 0.5480951070785522,
      "learning_rate": 0.00019999871393152016,
      "loss": 0.5013056755065918,
      "mean_token_accuracy": 0.8358610570430756,
      "num_tokens": 491520.0,
      "step": 60
    },
    {
      "entropy": 0.4816762119531631,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.483894407749176,
      "learning_rate": 0.0001999942683051699,
      "loss": 0.5129937648773193,
      "mean_token_accuracy": 0.8314701557159424,
      "num_tokens": 573440.0,
      "step": 70
    },
    {
      "entropy": 0.4728733152151108,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.4553426504135132,
      "learning_rate": 0.0001999866473846968,
      "loss": 0.49852399826049804,
      "mean_token_accuracy": 0.8351516604423523,
      "num_tokens": 655360.0,
      "step": 80
    },
    {
      "entropy": 0.46917213648557665,
      "epoch": 0.048,
      "grad_norm": 0.4663492441177368,
      "learning_rate": 0.00019997585141210138,
      "loss": 0.49323554039001466,
      "mean_token_accuracy": 0.8362279832363129,
      "num_tokens": 737280.0,
      "step": 90
    },
    {
      "entropy": 0.47362413704395295,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.3850269615650177,
      "learning_rate": 0.00019996188073020713,
      "loss": 0.5015204906463623,
      "mean_token_accuracy": 0.8351516574621201,
      "num_tokens": 819200.0,
      "step": 100
    },
    {
      "entropy": 0.4821264028549194,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.45479482412338257,
      "learning_rate": 0.00019994473578264973,
      "loss": 0.5033886909484864,
      "mean_token_accuracy": 0.8326198607683182,
      "num_tokens": 901120.0,
      "step": 110
    },
    {
      "entropy": 0.4650668889284134,
      "epoch": 0.064,
      "grad_norm": 0.47631603479385376,
      "learning_rate": 0.00019992441711386293,
      "loss": 0.4962433338165283,
      "mean_token_accuracy": 0.8367906093597413,
      "num_tokens": 983040.0,
      "step": 120
    },
    {
      "entropy": 0.4852884247899055,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.4958484470844269,
      "learning_rate": 0.00019990092536906122,
      "loss": 0.507628345489502,
      "mean_token_accuracy": 0.8318615466356277,
      "num_tokens": 1064960.0,
      "step": 130
    },
    {
      "entropy": 0.478881211578846,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.41259217262268066,
      "learning_rate": 0.00019987426129421926,
      "loss": 0.5079506874084473,
      "mean_token_accuracy": 0.8304305255413056,
      "num_tokens": 1146880.0,
      "step": 140
    },
    {
      "entropy": 0.47110913693904877,
      "epoch": 0.08,
      "grad_norm": 0.49864017963409424,
      "learning_rate": 0.00019984442573604846,
      "loss": 0.4969972610473633,
      "mean_token_accuracy": 0.8364114463329315,
      "num_tokens": 1228800.0,
      "step": 150
    },
    {
      "entropy": 0.47135204523801805,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.3627927899360657,
      "learning_rate": 0.00019981141964196984,
      "loss": 0.4969362258911133,
      "mean_token_accuracy": 0.8347847402095795,
      "num_tokens": 1310720.0,
      "step": 160
    },
    {
      "entropy": 0.4681368634104729,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.3932296633720398,
      "learning_rate": 0.00019977524406008406,
      "loss": 0.49262151718139646,
      "mean_token_accuracy": 0.8375856161117554,
      "num_tokens": 1392640.0,
      "step": 170
    },
    {
      "entropy": 0.47689022272825243,
      "epoch": 0.096,
      "grad_norm": 0.4345407485961914,
      "learning_rate": 0.00019973590013913802,
      "loss": 0.5007582187652588,
      "mean_token_accuracy": 0.8347725033760071,
      "num_tokens": 1474560.0,
      "step": 180
    },
    {
      "entropy": 0.4713203594088554,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.5165418982505798,
      "learning_rate": 0.00019969338912848857,
      "loss": 0.49573826789855957,
      "mean_token_accuracy": 0.8354452043771744,
      "num_tokens": 1556480.0,
      "step": 190
    },
    {
      "entropy": 0.47541218250989914,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.36708351969718933,
      "learning_rate": 0.00019964771237806272,
      "loss": 0.49782729148864746,
      "mean_token_accuracy": 0.8340264171361923,
      "num_tokens": 1638400.0,
      "step": 200
    },
    {
      "entropy": 0.45323160141706464,
      "epoch": 0.112,
      "grad_norm": 0.3474246859550476,
      "learning_rate": 0.00019959887133831478,
      "loss": 0.47730298042297364,
      "mean_token_accuracy": 0.8411081194877624,
      "num_tokens": 1720320.0,
      "step": 210
    },
    {
      "entropy": 0.4694709673523903,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.3597964346408844,
      "learning_rate": 0.0001995468675601803,
      "loss": 0.49341282844543455,
      "mean_token_accuracy": 0.8383194714784622,
      "num_tokens": 1802240.0,
      "step": 220
    },
    {
      "entropy": 0.4739027410745621,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.4527409076690674,
      "learning_rate": 0.0001994917026950268,
      "loss": 0.49920153617858887,
      "mean_token_accuracy": 0.8331335633993149,
      "num_tokens": 1884160.0,
      "step": 230
    },
    {
      "entropy": 0.47076870054006575,
      "epoch": 0.128,
      "grad_norm": 0.35778895020484924,
      "learning_rate": 0.0001994333784946015,
      "loss": 0.4972074508666992,
      "mean_token_accuracy": 0.8346257328987121,
      "num_tokens": 1966080.0,
      "step": 240
    },
    {
      "entropy": 0.4743542864918709,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.31897658109664917,
      "learning_rate": 0.00019937189681097536,
      "loss": 0.4985039710998535,
      "mean_token_accuracy": 0.8341854184865951,
      "num_tokens": 2048000.0,
      "step": 250
    },
    {
      "entropy": 0.475962033867836,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.32596156001091003,
      "learning_rate": 0.00019930725959648466,
      "loss": 0.5011369705200195,
      "mean_token_accuracy": 0.8342465788125992,
      "num_tokens": 2129920.0,
      "step": 260
    },
    {
      "entropy": 0.47928410172462466,
      "epoch": 0.144,
      "grad_norm": 0.3736521005630493,
      "learning_rate": 0.00019923946890366864,
      "loss": 0.5007104396820068,
      "mean_token_accuracy": 0.8340631127357483,
      "num_tokens": 2211840.0,
      "step": 270
    },
    {
      "entropy": 0.4498627737164497,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.37184008955955505,
      "learning_rate": 0.00019916852688520463,
      "loss": 0.47563848495483396,
      "mean_token_accuracy": 0.8403131157159806,
      "num_tokens": 2293760.0,
      "step": 280
    },
    {
      "entropy": 0.45984665900468824,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.33217671513557434,
      "learning_rate": 0.0001990944357938395,
      "loss": 0.4853631019592285,
      "mean_token_accuracy": 0.8384173214435577,
      "num_tokens": 2375680.0,
      "step": 290
    },
    {
      "entropy": 0.473418790102005,
      "epoch": 0.16,
      "grad_norm": 0.37292230129241943,
      "learning_rate": 0.0001990171979823181,
      "loss": 0.4935110092163086,
      "mean_token_accuracy": 0.8368028342723847,
      "num_tokens": 2457600.0,
      "step": 300
    },
    {
      "entropy": 0.46503881365060806,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.28929391503334045,
      "learning_rate": 0.00019893681590330874,
      "loss": 0.4931675910949707,
      "mean_token_accuracy": 0.8358855098485947,
      "num_tokens": 2539520.0,
      "step": 310
    },
    {
      "entropy": 0.4774789586663246,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.31040647625923157,
      "learning_rate": 0.00019885329210932517,
      "loss": 0.502250862121582,
      "mean_token_accuracy": 0.8329501003026962,
      "num_tokens": 2621440.0,
      "step": 320
    },
    {
      "entropy": 0.47055001705884936,
      "epoch": 0.176,
      "grad_norm": 0.36547258496284485,
      "learning_rate": 0.0001987666292526455,
      "loss": 0.4946347713470459,
      "mean_token_accuracy": 0.8376223087310791,
      "num_tokens": 2703360.0,
      "step": 330
    },
    {
      "entropy": 0.4748024299740791,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.335625022649765,
      "learning_rate": 0.0001986768300852281,
      "loss": 0.4987218379974365,
      "mean_token_accuracy": 0.8339163362979889,
      "num_tokens": 2785280.0,
      "step": 340
    },
    {
      "entropy": 0.4574843019247055,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.8638505339622498,
      "learning_rate": 0.00019858389745862405,
      "loss": 0.48145270347595215,
      "mean_token_accuracy": 0.839469176530838,
      "num_tokens": 2867200.0,
      "step": 350
    },
    {
      "entropy": 0.47080869525671004,
      "epoch": 0.192,
      "grad_norm": 0.33534178137779236,
      "learning_rate": 0.0001984878343238868,
      "loss": 0.4966005802154541,
      "mean_token_accuracy": 0.8357876688241959,
      "num_tokens": 2949120.0,
      "step": 360
    },
    {
      "entropy": 0.47144559025764465,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.35321077704429626,
      "learning_rate": 0.00019838864373147814,
      "loss": 0.4953292369842529,
      "mean_token_accuracy": 0.8352128237485885,
      "num_tokens": 3031040.0,
      "step": 370
    },
    {
      "entropy": 0.48720123171806334,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.35722675919532776,
      "learning_rate": 0.00019828632883117176,
      "loss": 0.5126425743103027,
      "mean_token_accuracy": 0.8297700613737107,
      "num_tokens": 3112960.0,
      "step": 380
    },
    {
      "entropy": 0.47570254653692245,
      "epoch": 0.208,
      "grad_norm": 0.516044557094574,
      "learning_rate": 0.00019818089287195286,
      "loss": 0.49929466247558596,
      "mean_token_accuracy": 0.8334760189056396,
      "num_tokens": 3194880.0,
      "step": 390
    },
    {
      "entropy": 0.45633259862661363,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.2613847255706787,
      "learning_rate": 0.00019807233920191516,
      "loss": 0.4783585548400879,
      "mean_token_accuracy": 0.8396404057741165,
      "num_tokens": 3276800.0,
      "step": 400
    },
    {
      "entropy": 0.4616701290011406,
      "epoch": 0.21866666666666668,
      "grad_norm": 0.3630369305610657,
      "learning_rate": 0.0001979606712681546,
      "loss": 0.4853507518768311,
      "mean_token_accuracy": 0.8378180056810379,
      "num_tokens": 3358720.0,
      "step": 410
    },
    {
      "entropy": 0.4638349115848541,
      "epoch": 0.224,
      "grad_norm": 0.32545581459999084,
      "learning_rate": 0.00019784589261665966,
      "loss": 0.4868640422821045,
      "mean_token_accuracy": 0.8398238748311997,
      "num_tokens": 3440640.0,
      "step": 420
    },
    {
      "entropy": 0.4723006457090378,
      "epoch": 0.22933333333333333,
      "grad_norm": 0.2839032709598541,
      "learning_rate": 0.0001977280068921991,
      "loss": 0.49332942962646487,
      "mean_token_accuracy": 0.8366927623748779,
      "num_tokens": 3522560.0,
      "step": 430
    },
    {
      "entropy": 0.4808267056941986,
      "epoch": 0.23466666666666666,
      "grad_norm": 0.32518133521080017,
      "learning_rate": 0.0001976070178382059,
      "loss": 0.5029919624328614,
      "mean_token_accuracy": 0.8347602784633636,
      "num_tokens": 3604480.0,
      "step": 440
    },
    {
      "entropy": 0.47628943622112274,
      "epoch": 0.24,
      "grad_norm": 0.34521424770355225,
      "learning_rate": 0.00019748292929665869,
      "loss": 0.5040892124176025,
      "mean_token_accuracy": 0.8313600778579712,
      "num_tokens": 3686400.0,
      "step": 450
    },
    {
      "entropy": 0.46220199316740035,
      "epoch": 0.24533333333333332,
      "grad_norm": 0.3326896131038666,
      "learning_rate": 0.00019735574520795938,
      "loss": 0.4836634635925293,
      "mean_token_accuracy": 0.837903618812561,
      "num_tokens": 3768320.0,
      "step": 460
    },
    {
      "entropy": 0.4717408582568169,
      "epoch": 0.25066666666666665,
      "grad_norm": 0.26027098298072815,
      "learning_rate": 0.00019722546961080842,
      "loss": 0.49863033294677733,
      "mean_token_accuracy": 0.8350905060768128,
      "num_tokens": 3850240.0,
      "step": 470
    },
    {
      "entropy": 0.46862735599279404,
      "epoch": 0.256,
      "grad_norm": 0.2998311221599579,
      "learning_rate": 0.00019709210664207625,
      "loss": 0.4927967548370361,
      "mean_token_accuracy": 0.8348092019557953,
      "num_tokens": 3932160.0,
      "step": 480
    },
    {
      "entropy": 0.46283133924007414,
      "epoch": 0.2613333333333333,
      "grad_norm": 0.274065762758255,
      "learning_rate": 0.00019695566053667217,
      "loss": 0.4877967357635498,
      "mean_token_accuracy": 0.836545991897583,
      "num_tokens": 4014080.0,
      "step": 490
    },
    {
      "entropy": 0.4567370221018791,
      "epoch": 0.26666666666666666,
      "grad_norm": 0.30564069747924805,
      "learning_rate": 0.00019681613562740963,
      "loss": 0.4802112579345703,
      "mean_token_accuracy": 0.84118150472641,
      "num_tokens": 4096000.0,
      "step": 500
    },
    {
      "entropy": 0.46357189267873766,
      "epoch": 0.272,
      "grad_norm": 0.319296270608902,
      "learning_rate": 0.0001966735363448689,
      "loss": 0.487426233291626,
      "mean_token_accuracy": 0.8382827788591385,
      "num_tokens": 4177920.0,
      "step": 510
    },
    {
      "entropy": 0.4628389373421669,
      "epoch": 0.2773333333333333,
      "grad_norm": 0.3266727328300476,
      "learning_rate": 0.00019652786721725604,
      "loss": 0.4883276462554932,
      "mean_token_accuracy": 0.8371086061000824,
      "num_tokens": 4259840.0,
      "step": 520
    },
    {
      "entropy": 0.4788906291127205,
      "epoch": 0.2826666666666667,
      "grad_norm": 0.4103166162967682,
      "learning_rate": 0.00019637913287025955,
      "loss": 0.5023647308349609,
      "mean_token_accuracy": 0.8328767120838165,
      "num_tokens": 4341760.0,
      "step": 530
    },
    {
      "entropy": 0.4703716978430748,
      "epoch": 0.288,
      "grad_norm": 0.38001173734664917,
      "learning_rate": 0.00019622733802690306,
      "loss": 0.49620709419250486,
      "mean_token_accuracy": 0.8345890462398529,
      "num_tokens": 4423680.0,
      "step": 540
    },
    {
      "entropy": 0.4588580518960953,
      "epoch": 0.29333333333333333,
      "grad_norm": 0.32616063952445984,
      "learning_rate": 0.0001960724875073956,
      "loss": 0.48500909805297854,
      "mean_token_accuracy": 0.8382216215133667,
      "num_tokens": 4505600.0,
      "step": 550
    },
    {
      "entropy": 0.4704396888613701,
      "epoch": 0.2986666666666667,
      "grad_norm": 0.32692262530326843,
      "learning_rate": 0.00019591458622897841,
      "loss": 0.4929184913635254,
      "mean_token_accuracy": 0.8346991211175918,
      "num_tokens": 4587520.0,
      "step": 560
    },
    {
      "entropy": 0.4703381538391113,
      "epoch": 0.304,
      "grad_norm": 0.30592527985572815,
      "learning_rate": 0.00019575363920576897,
      "loss": 0.49500484466552735,
      "mean_token_accuracy": 0.8349070399999619,
      "num_tokens": 4669440.0,
      "step": 570
    },
    {
      "entropy": 0.47023057043552396,
      "epoch": 0.30933333333333335,
      "grad_norm": 0.300794392824173,
      "learning_rate": 0.00019558965154860158,
      "loss": 0.49615039825439455,
      "mean_token_accuracy": 0.8345523416996002,
      "num_tokens": 4751360.0,
      "step": 580
    },
    {
      "entropy": 0.47680234611034394,
      "epoch": 0.31466666666666665,
      "grad_norm": 0.3349059820175171,
      "learning_rate": 0.0001954226284648651,
      "loss": 0.5006989479064942,
      "mean_token_accuracy": 0.834882590174675,
      "num_tokens": 4833280.0,
      "step": 590
    },
    {
      "entropy": 0.4700273126363754,
      "epoch": 0.32,
      "grad_norm": 0.44947975873947144,
      "learning_rate": 0.00019525257525833772,
      "loss": 0.49590649604797366,
      "mean_token_accuracy": 0.8349804282188416,
      "num_tokens": 4915200.0,
      "step": 600
    },
    {
      "entropy": 0.46182160526514054,
      "epoch": 0.3253333333333333,
      "grad_norm": 0.31675371527671814,
      "learning_rate": 0.00019507949732901843,
      "loss": 0.4854272365570068,
      "mean_token_accuracy": 0.8387597769498825,
      "num_tokens": 4997120.0,
      "step": 610
    },
    {
      "entropy": 0.4660103768110275,
      "epoch": 0.33066666666666666,
      "grad_norm": 0.3232131600379944,
      "learning_rate": 0.00019490340017295555,
      "loss": 0.4887096881866455,
      "mean_token_accuracy": 0.8356286644935608,
      "num_tokens": 5079040.0,
      "step": 620
    },
    {
      "entropy": 0.46997269839048383,
      "epoch": 0.336,
      "grad_norm": 0.29394158720970154,
      "learning_rate": 0.00019472428938207222,
      "loss": 0.49637699127197266,
      "mean_token_accuracy": 0.835335123538971,
      "num_tokens": 5160960.0,
      "step": 630
    },
    {
      "entropy": 0.4736437723040581,
      "epoch": 0.3413333333333333,
      "grad_norm": 0.3401532471179962,
      "learning_rate": 0.0001945421706439889,
      "loss": 0.4949938297271729,
      "mean_token_accuracy": 0.8349559634923935,
      "num_tokens": 5242880.0,
      "step": 640
    },
    {
      "entropy": 0.4688768327236176,
      "epoch": 0.3466666666666667,
      "grad_norm": 0.30354562401771545,
      "learning_rate": 0.00019435704974184263,
      "loss": 0.4957862854003906,
      "mean_token_accuracy": 0.8341487288475037,
      "num_tokens": 5324800.0,
      "step": 650
    },
    {
      "entropy": 0.4714811623096466,
      "epoch": 0.352,
      "grad_norm": 0.33926305174827576,
      "learning_rate": 0.00019416893255410348,
      "loss": 0.4908115386962891,
      "mean_token_accuracy": 0.8354329735040664,
      "num_tokens": 5406720.0,
      "step": 660
    },
    {
      "entropy": 0.46530274748802186,
      "epoch": 0.35733333333333334,
      "grad_norm": 0.29678741097450256,
      "learning_rate": 0.0001939778250543878,
      "loss": 0.49260859489440917,
      "mean_token_accuracy": 0.8360445171594619,
      "num_tokens": 5488640.0,
      "step": 670
    },
    {
      "entropy": 0.46746390461921694,
      "epoch": 0.3626666666666667,
      "grad_norm": 0.33242908120155334,
      "learning_rate": 0.0001937837333112687,
      "loss": 0.4908029079437256,
      "mean_token_accuracy": 0.8367294520139694,
      "num_tokens": 5570560.0,
      "step": 680
    },
    {
      "entropy": 0.47267923355102537,
      "epoch": 0.368,
      "grad_norm": 0.31784218549728394,
      "learning_rate": 0.00019358666348808306,
      "loss": 0.4970839977264404,
      "mean_token_accuracy": 0.8348581224679947,
      "num_tokens": 5652480.0,
      "step": 690
    },
    {
      "entropy": 0.4630097344517708,
      "epoch": 0.37333333333333335,
      "grad_norm": 0.32503777742385864,
      "learning_rate": 0.00019338662184273617,
      "loss": 0.4922970771789551,
      "mean_token_accuracy": 0.8364481449127197,
      "num_tokens": 5734400.0,
      "step": 700
    },
    {
      "entropy": 0.4575432151556015,
      "epoch": 0.37866666666666665,
      "grad_norm": 0.3280612826347351,
      "learning_rate": 0.00019318361472750272,
      "loss": 0.47847614288330076,
      "mean_token_accuracy": 0.8396037191152572,
      "num_tokens": 5816320.0,
      "step": 710
    },
    {
      "entropy": 0.46167032420635223,
      "epoch": 0.384,
      "grad_norm": 0.2691577970981598,
      "learning_rate": 0.00019297764858882514,
      "loss": 0.48371038436889646,
      "mean_token_accuracy": 0.8382583141326905,
      "num_tokens": 5898240.0,
      "step": 720
    },
    {
      "entropy": 0.45909654200077055,
      "epoch": 0.3893333333333333,
      "grad_norm": 0.2929236590862274,
      "learning_rate": 0.00019276872996710907,
      "loss": 0.48665547370910645,
      "mean_token_accuracy": 0.8385273963212967,
      "num_tokens": 5980160.0,
      "step": 730
    },
    {
      "entropy": 0.4599571406841278,
      "epoch": 0.39466666666666667,
      "grad_norm": 0.3496358394622803,
      "learning_rate": 0.00019255686549651544,
      "loss": 0.4863029956817627,
      "mean_token_accuracy": 0.838050389289856,
      "num_tokens": 6062080.0,
      "step": 740
    },
    {
      "entropy": 0.4732843995094299,
      "epoch": 0.4,
      "grad_norm": 0.3045860230922699,
      "learning_rate": 0.00019234206190474993,
      "loss": 0.49899816513061523,
      "mean_token_accuracy": 0.8332925587892532,
      "num_tokens": 6144000.0,
      "step": 750
    },
    {
      "entropy": 0.45619571655988694,
      "epoch": 0.4053333333333333,
      "grad_norm": 0.29448261857032776,
      "learning_rate": 0.00019212432601284934,
      "loss": 0.48101630210876467,
      "mean_token_accuracy": 0.8395792603492737,
      "num_tokens": 6225920.0,
      "step": 760
    },
    {
      "entropy": 0.4570640236139297,
      "epoch": 0.4106666666666667,
      "grad_norm": 0.31912145018577576,
      "learning_rate": 0.00019190366473496498,
      "loss": 0.48552522659301756,
      "mean_token_accuracy": 0.8383316993713379,
      "num_tokens": 6307840.0,
      "step": 770
    },
    {
      "entropy": 0.4614878877997398,
      "epoch": 0.416,
      "grad_norm": 0.29299449920654297,
      "learning_rate": 0.00019168008507814302,
      "loss": 0.48582162857055666,
      "mean_token_accuracy": 0.8369862973690033,
      "num_tokens": 6389760.0,
      "step": 780
    },
    {
      "entropy": 0.4764199286699295,
      "epoch": 0.42133333333333334,
      "grad_norm": 0.3088685870170593,
      "learning_rate": 0.00019145359414210212,
      "loss": 0.500695276260376,
      "mean_token_accuracy": 0.8338184893131256,
      "num_tokens": 6471680.0,
      "step": 790
    },
    {
      "entropy": 0.4658344119787216,
      "epoch": 0.4266666666666667,
      "grad_norm": 0.40950676798820496,
      "learning_rate": 0.0001912241991190079,
      "loss": 0.490970516204834,
      "mean_token_accuracy": 0.8368028283119202,
      "num_tokens": 6553600.0,
      "step": 800
    },
    {
      "entropy": 0.4629588112235069,
      "epoch": 0.432,
      "grad_norm": 0.31325265765190125,
      "learning_rate": 0.0001909919072932445,
      "loss": 0.49127984046936035,
      "mean_token_accuracy": 0.8364481419324875,
      "num_tokens": 6635520.0,
      "step": 810
    },
    {
      "entropy": 0.4665198639035225,
      "epoch": 0.43733333333333335,
      "grad_norm": 0.2922220230102539,
      "learning_rate": 0.0001907567260411835,
      "loss": 0.4937121391296387,
      "mean_token_accuracy": 0.8348703503608703,
      "num_tokens": 6717440.0,
      "step": 820
    },
    {
      "entropy": 0.47019302994012835,
      "epoch": 0.44266666666666665,
      "grad_norm": 0.2724287509918213,
      "learning_rate": 0.00019051866283094934,
      "loss": 0.4938626766204834,
      "mean_token_accuracy": 0.835971137881279,
      "num_tokens": 6799360.0,
      "step": 830
    },
    {
      "entropy": 0.4524460420012474,
      "epoch": 0.448,
      "grad_norm": 0.2934575378894806,
      "learning_rate": 0.00019027772522218247,
      "loss": 0.4761967182159424,
      "mean_token_accuracy": 0.8413405060768128,
      "num_tokens": 6881280.0,
      "step": 840
    },
    {
      "entropy": 0.46517165899276736,
      "epoch": 0.4533333333333333,
      "grad_norm": 0.25503990054130554,
      "learning_rate": 0.0001900339208657992,
      "loss": 0.48938493728637694,
      "mean_token_accuracy": 0.8362157583236695,
      "num_tokens": 6963200.0,
      "step": 850
    },
    {
      "entropy": 0.46393668055534365,
      "epoch": 0.45866666666666667,
      "grad_norm": 0.3126053810119629,
      "learning_rate": 0.00018978725750374862,
      "loss": 0.4887409210205078,
      "mean_token_accuracy": 0.8374510794878006,
      "num_tokens": 7045120.0,
      "step": 860
    },
    {
      "entropy": 0.4659988895058632,
      "epoch": 0.464,
      "grad_norm": 0.3315013349056244,
      "learning_rate": 0.0001895377429687669,
      "loss": 0.48931269645690917,
      "mean_token_accuracy": 0.8364481389522552,
      "num_tokens": 7127040.0,
      "step": 870
    },
    {
      "entropy": 0.47046023458242414,
      "epoch": 0.4693333333333333,
      "grad_norm": 0.2909882664680481,
      "learning_rate": 0.00018928538518412857,
      "loss": 0.4963551044464111,
      "mean_token_accuracy": 0.8360567569732666,
      "num_tokens": 7208960.0,
      "step": 880
    },
    {
      "entropy": 0.4718432009220123,
      "epoch": 0.4746666666666667,
      "grad_norm": 0.31447282433509827,
      "learning_rate": 0.00018903019216339487,
      "loss": 0.49306254386901854,
      "mean_token_accuracy": 0.8334637969732285,
      "num_tokens": 7290880.0,
      "step": 890
    },
    {
      "entropy": 0.46723778545856476,
      "epoch": 0.48,
      "grad_norm": 0.27959972620010376,
      "learning_rate": 0.0001887721720101592,
      "loss": 0.49190969467163087,
      "mean_token_accuracy": 0.8366560727357865,
      "num_tokens": 7372800.0,
      "step": 900
    },
    {
      "entropy": 0.4639134585857391,
      "epoch": 0.48533333333333334,
      "grad_norm": 0.2751777470111847,
      "learning_rate": 0.00018851133291779003,
      "loss": 0.488355827331543,
      "mean_token_accuracy": 0.8373165339231491,
      "num_tokens": 7454720.0,
      "step": 910
    },
    {
      "entropy": 0.47239599376916885,
      "epoch": 0.49066666666666664,
      "grad_norm": 0.30166029930114746,
      "learning_rate": 0.00018824768316917038,
      "loss": 0.4971829891204834,
      "mean_token_accuracy": 0.8345034152269364,
      "num_tokens": 7536640.0,
      "step": 920
    },
    {
      "entropy": 0.4682949334383011,
      "epoch": 0.496,
      "grad_norm": 0.2700556516647339,
      "learning_rate": 0.00018798123113643515,
      "loss": 0.4912562847137451,
      "mean_token_accuracy": 0.8360812187194824,
      "num_tokens": 7618560.0,
      "step": 930
    },
    {
      "entropy": 0.47007360607385634,
      "epoch": 0.5013333333333333,
      "grad_norm": 0.2699863314628601,
      "learning_rate": 0.00018771198528070502,
      "loss": 0.49548869132995604,
      "mean_token_accuracy": 0.8343199610710144,
      "num_tokens": 7700480.0,
      "step": 940
    },
    {
      "entropy": 0.46558715850114823,
      "epoch": 0.5066666666666667,
      "grad_norm": 0.2879142463207245,
      "learning_rate": 0.0001874399541518179,
      "loss": 0.4899310111999512,
      "mean_token_accuracy": 0.8356408983469009,
      "num_tokens": 7782400.0,
      "step": 950
    },
    {
      "entropy": 0.4690485090017319,
      "epoch": 0.512,
      "grad_norm": 0.31532764434814453,
      "learning_rate": 0.00018716514638805734,
      "loss": 0.49160404205322267,
      "mean_token_accuracy": 0.8357876718044281,
      "num_tokens": 7864320.0,
      "step": 960
    },
    {
      "entropy": 0.4711877599358559,
      "epoch": 0.5173333333333333,
      "grad_norm": 0.2866258919239044,
      "learning_rate": 0.00018688757071587829,
      "loss": 0.4993025302886963,
      "mean_token_accuracy": 0.8326198667287826,
      "num_tokens": 7946240.0,
      "step": 970
    },
    {
      "entropy": 0.4613691926002502,
      "epoch": 0.5226666666666666,
      "grad_norm": 0.2963973581790924,
      "learning_rate": 0.0001866072359496299,
      "loss": 0.48441057205200194,
      "mean_token_accuracy": 0.8382827818393708,
      "num_tokens": 8028160.0,
      "step": 980
    },
    {
      "entropy": 0.45830214470624925,
      "epoch": 0.528,
      "grad_norm": 0.30914536118507385,
      "learning_rate": 0.00018632415099127588,
      "loss": 0.48667316436767577,
      "mean_token_accuracy": 0.8380626201629638,
      "num_tokens": 8110080.0,
      "step": 990
    },
    {
      "entropy": 0.4772834450006485,
      "epoch": 0.5333333333333333,
      "grad_norm": 0.3332306444644928,
      "learning_rate": 0.00018603832483011145,
      "loss": 0.498505687713623,
      "mean_token_accuracy": 0.8316291540861129,
      "num_tokens": 8192000.0,
      "step": 1000
    },
    {
      "entropy": 0.46559437811374665,
      "epoch": 0.5386666666666666,
      "grad_norm": 0.30015191435813904,
      "learning_rate": 0.00018574976654247812,
      "loss": 0.48902440071105957,
      "mean_token_accuracy": 0.8363258332014084,
      "num_tokens": 8273920.0,
      "step": 1010
    },
    {
      "entropy": 0.451631960272789,
      "epoch": 0.544,
      "grad_norm": 0.30106401443481445,
      "learning_rate": 0.0001854584852914755,
      "loss": 0.47860217094421387,
      "mean_token_accuracy": 0.8401908040046692,
      "num_tokens": 8355840.0,
      "step": 1020
    },
    {
      "entropy": 0.45579866170883176,
      "epoch": 0.5493333333333333,
      "grad_norm": 0.3621468245983124,
      "learning_rate": 0.00018516449032667009,
      "loss": 0.4805305004119873,
      "mean_token_accuracy": 0.8400195688009262,
      "num_tokens": 8437760.0,
      "step": 1030
    },
    {
      "entropy": 0.4544841006398201,
      "epoch": 0.5546666666666666,
      "grad_norm": 0.2794948220252991,
      "learning_rate": 0.00018486779098380194,
      "loss": 0.4778414249420166,
      "mean_token_accuracy": 0.8401296436786652,
      "num_tokens": 8519680.0,
      "step": 1040
    },
    {
      "entropy": 0.4718469813466072,
      "epoch": 0.56,
      "grad_norm": 0.30742916464805603,
      "learning_rate": 0.00018456839668448777,
      "loss": 0.4994341850280762,
      "mean_token_accuracy": 0.8344544976949692,
      "num_tokens": 8601600.0,
      "step": 1050
    },
    {
      "entropy": 0.45879508554935455,
      "epoch": 0.5653333333333334,
      "grad_norm": 0.31825685501098633,
      "learning_rate": 0.00018426631693592218,
      "loss": 0.4842658042907715,
      "mean_token_accuracy": 0.8375611513853073,
      "num_tokens": 8683520.0,
      "step": 1060
    },
    {
      "entropy": 0.462214232981205,
      "epoch": 0.5706666666666667,
      "grad_norm": 0.3559083342552185,
      "learning_rate": 0.0001839615613305754,
      "loss": 0.4861720085144043,
      "mean_token_accuracy": 0.8365215271711349,
      "num_tokens": 8765440.0,
      "step": 1070
    },
    {
      "entropy": 0.4669525742530823,
      "epoch": 0.576,
      "grad_norm": 0.32343730330467224,
      "learning_rate": 0.000183654139545889,
      "loss": 0.4933466911315918,
      "mean_token_accuracy": 0.8351271957159042,
      "num_tokens": 8847360.0,
      "step": 1080
    },
    {
      "entropy": 0.46452919840812684,
      "epoch": 0.5813333333333334,
      "grad_norm": 0.278026282787323,
      "learning_rate": 0.00018334406134396834,
      "loss": 0.49009218215942385,
      "mean_token_accuracy": 0.83616683781147,
      "num_tokens": 8929280.0,
      "step": 1090
    },
    {
      "entropy": 0.4648389250040054,
      "epoch": 0.5866666666666667,
      "grad_norm": 0.3153679370880127,
      "learning_rate": 0.00018303133657127273,
      "loss": 0.49022712707519533,
      "mean_token_accuracy": 0.835322892665863,
      "num_tokens": 9011200.0,
      "step": 1100
    },
    {
      "entropy": 0.4719624638557434,
      "epoch": 0.592,
      "grad_norm": 0.30147644877433777,
      "learning_rate": 0.00018271597515830265,
      "loss": 0.4942080020904541,
      "mean_token_accuracy": 0.8336228013038636,
      "num_tokens": 9093120.0,
      "step": 1110
    },
    {
      "entropy": 0.4562485605478287,
      "epoch": 0.5973333333333334,
      "grad_norm": 0.28926217555999756,
      "learning_rate": 0.00018239798711928453,
      "loss": 0.48056983947753906,
      "mean_token_accuracy": 0.8390777915716171,
      "num_tokens": 9175040.0,
      "step": 1120
    },
    {
      "entropy": 0.4592498913407326,
      "epoch": 0.6026666666666667,
      "grad_norm": 0.2862878739833832,
      "learning_rate": 0.00018207738255185266,
      "loss": 0.4865690231323242,
      "mean_token_accuracy": 0.836350291967392,
      "num_tokens": 9256960.0,
      "step": 1130
    },
    {
      "entropy": 0.46526293754577636,
      "epoch": 0.608,
      "grad_norm": 0.4841032326221466,
      "learning_rate": 0.00018175417163672852,
      "loss": 0.4909062385559082,
      "mean_token_accuracy": 0.8361912935972213,
      "num_tokens": 9338880.0,
      "step": 1140
    },
    {
      "entropy": 0.4709723249077797,
      "epoch": 0.6133333333333333,
      "grad_norm": 0.34231436252593994,
      "learning_rate": 0.00018142836463739759,
      "loss": 0.49631567001342775,
      "mean_token_accuracy": 0.8354941248893738,
      "num_tokens": 9420800.0,
      "step": 1150
    },
    {
      "entropy": 0.47033472508192065,
      "epoch": 0.6186666666666667,
      "grad_norm": 0.3496396839618683,
      "learning_rate": 0.00018109997189978339,
      "loss": 0.49637880325317385,
      "mean_token_accuracy": 0.8337695717811584,
      "num_tokens": 9502720.0,
      "step": 1160
    },
    {
      "entropy": 0.4585022464394569,
      "epoch": 0.624,
      "grad_norm": 0.3378464877605438,
      "learning_rate": 0.0001807690038519189,
      "loss": 0.4835620403289795,
      "mean_token_accuracy": 0.8369985342025756,
      "num_tokens": 9584640.0,
      "step": 1170
    },
    {
      "entropy": 0.46430673748254775,
      "epoch": 0.6293333333333333,
      "grad_norm": 0.2734282314777374,
      "learning_rate": 0.00018043547100361545,
      "loss": 0.48804349899291993,
      "mean_token_accuracy": 0.8371575355529786,
      "num_tokens": 9666560.0,
      "step": 1180
    },
    {
      "entropy": 0.4538321003317833,
      "epoch": 0.6346666666666667,
      "grad_norm": 0.31381797790527344,
      "learning_rate": 0.0001800993839461291,
      "loss": 0.4797845363616943,
      "mean_token_accuracy": 0.839359101653099,
      "num_tokens": 9748480.0,
      "step": 1190
    },
    {
      "entropy": 0.46689577400684357,
      "epoch": 0.64,
      "grad_norm": 0.3240891695022583,
      "learning_rate": 0.00017976075335182407,
      "loss": 0.49417672157287595,
      "mean_token_accuracy": 0.8348948121070862,
      "num_tokens": 9830400.0,
      "step": 1200
    },
    {
      "entropy": 0.45945043563842775,
      "epoch": 0.6453333333333333,
      "grad_norm": 0.30571091175079346,
      "learning_rate": 0.0001794195899738341,
      "loss": 0.48449225425720216,
      "mean_token_accuracy": 0.8381604701280594,
      "num_tokens": 9912320.0,
      "step": 1210
    },
    {
      "entropy": 0.45976146310567856,
      "epoch": 0.6506666666666666,
      "grad_norm": 0.31973016262054443,
      "learning_rate": 0.00017907590464572083,
      "loss": 0.48612003326416015,
      "mean_token_accuracy": 0.8371330708265304,
      "num_tokens": 9994240.0,
      "step": 1220
    },
    {
      "entropy": 0.4679174989461899,
      "epoch": 0.656,
      "grad_norm": 0.25371766090393066,
      "learning_rate": 0.0001787297082811298,
      "loss": 0.4935025691986084,
      "mean_token_accuracy": 0.8343444257974625,
      "num_tokens": 10076160.0,
      "step": 1230
    },
    {
      "entropy": 0.4576866328716278,
      "epoch": 0.6613333333333333,
      "grad_norm": 0.2980435788631439,
      "learning_rate": 0.0001783810118734439,
      "loss": 0.48330230712890626,
      "mean_token_accuracy": 0.8372920751571655,
      "num_tokens": 10158080.0,
      "step": 1240
    },
    {
      "entropy": 0.47016931921243665,
      "epoch": 0.6666666666666666,
      "grad_norm": 0.34448325634002686,
      "learning_rate": 0.00017802982649543444,
      "loss": 0.49863452911376954,
      "mean_token_accuracy": 0.8339530318975449,
      "num_tokens": 10240000.0,
      "step": 1250
    },
    {
      "entropy": 0.47134701162576675,
      "epoch": 0.672,
      "grad_norm": 0.29068523645401,
      "learning_rate": 0.00017767616329890918,
      "loss": 0.4956683158874512,
      "mean_token_accuracy": 0.8342588067054748,
      "num_tokens": 10321920.0,
      "step": 1260
    },
    {
      "entropy": 0.47508572340011596,
      "epoch": 0.6773333333333333,
      "grad_norm": 0.3611260652542114,
      "learning_rate": 0.00017732003351435854,
      "loss": 0.49777750968933104,
      "mean_token_accuracy": 0.8324853211641312,
      "num_tokens": 10403840.0,
      "step": 1270
    },
    {
      "entropy": 0.46301001161336897,
      "epoch": 0.6826666666666666,
      "grad_norm": 0.268515408039093,
      "learning_rate": 0.00017696144845059881,
      "loss": 0.48827152252197265,
      "mean_token_accuracy": 0.8357754439115525,
      "num_tokens": 10485760.0,
      "step": 1280
    },
    {
      "entropy": 0.454975888133049,
      "epoch": 0.688,
      "grad_norm": 0.32114407420158386,
      "learning_rate": 0.0001766004194944131,
      "loss": 0.4804567337036133,
      "mean_token_accuracy": 0.8386986255645752,
      "num_tokens": 10567680.0,
      "step": 1290
    },
    {
      "entropy": 0.45681490153074267,
      "epoch": 0.6933333333333334,
      "grad_norm": 0.30865970253944397,
      "learning_rate": 0.0001762369581101898,
      "loss": 0.4790478706359863,
      "mean_token_accuracy": 0.8393713295459747,
      "num_tokens": 10649600.0,
      "step": 1300
    },
    {
      "entropy": 0.46000348776578903,
      "epoch": 0.6986666666666667,
      "grad_norm": 0.2983383536338806,
      "learning_rate": 0.00017587107583955832,
      "loss": 0.4848841667175293,
      "mean_token_accuracy": 0.8373899221420288,
      "num_tokens": 10731520.0,
      "step": 1310
    },
    {
      "entropy": 0.46640499234199523,
      "epoch": 0.704,
      "grad_norm": 0.2804456055164337,
      "learning_rate": 0.0001755027843010229,
      "loss": 0.490816068649292,
      "mean_token_accuracy": 0.8355797469615937,
      "num_tokens": 10813440.0,
      "step": 1320
    },
    {
      "entropy": 0.46081168949604034,
      "epoch": 0.7093333333333334,
      "grad_norm": 0.29534995555877686,
      "learning_rate": 0.00017513209518959345,
      "loss": 0.48174343109130857,
      "mean_token_accuracy": 0.8382216215133667,
      "num_tokens": 10895360.0,
      "step": 1330
    },
    {
      "entropy": 0.4528317332267761,
      "epoch": 0.7146666666666667,
      "grad_norm": 0.30809423327445984,
      "learning_rate": 0.00017475902027641415,
      "loss": 0.4788063526153564,
      "mean_token_accuracy": 0.8405210316181183,
      "num_tokens": 10977280.0,
      "step": 1340
    },
    {
      "entropy": 0.4585173100233078,
      "epoch": 0.72,
      "grad_norm": 0.2678697109222412,
      "learning_rate": 0.0001743835714083899,
      "loss": 0.48354172706604004,
      "mean_token_accuracy": 0.8382093995809555,
      "num_tokens": 11059200.0,
      "step": 1350
    },
    {
      "entropy": 0.46523621380329133,
      "epoch": 0.7253333333333334,
      "grad_norm": 0.2942293882369995,
      "learning_rate": 0.0001740057605078098,
      "loss": 0.49128217697143556,
      "mean_token_accuracy": 0.8345523416996002,
      "num_tokens": 11141120.0,
      "step": 1360
    },
    {
      "entropy": 0.46702390909194946,
      "epoch": 0.7306666666666667,
      "grad_norm": 0.27172449231147766,
      "learning_rate": 0.00017362559957196893,
      "loss": 0.4921745300292969,
      "mean_token_accuracy": 0.8357509821653366,
      "num_tokens": 11223040.0,
      "step": 1370
    },
    {
      "entropy": 0.455118428170681,
      "epoch": 0.736,
      "grad_norm": 0.29810631275177,
      "learning_rate": 0.00017324310067278704,
      "loss": 0.47878117561340333,
      "mean_token_accuracy": 0.8403742641210556,
      "num_tokens": 11304960.0,
      "step": 1380
    },
    {
      "entropy": 0.47074327766895296,
      "epoch": 0.7413333333333333,
      "grad_norm": 0.353583425283432,
      "learning_rate": 0.0001728582759564253,
      "loss": 0.49146604537963867,
      "mean_token_accuracy": 0.8346501857042312,
      "num_tokens": 11386880.0,
      "step": 1390
    },
    {
      "entropy": 0.46050443798303603,
      "epoch": 0.7466666666666667,
      "grad_norm": 0.33743584156036377,
      "learning_rate": 0.00017247113764290082,
      "loss": 0.48703465461730955,
      "mean_token_accuracy": 0.8353718131780624,
      "num_tokens": 11468800.0,
      "step": 1400
    },
    {
      "entropy": 0.4616937756538391,
      "epoch": 0.752,
      "grad_norm": 0.3148867189884186,
      "learning_rate": 0.00017208169802569826,
      "loss": 0.48604726791381836,
      "mean_token_accuracy": 0.8367049872875214,
      "num_tokens": 11550720.0,
      "step": 1410
    },
    {
      "entropy": 0.4618809401988983,
      "epoch": 0.7573333333333333,
      "grad_norm": 0.31237900257110596,
      "learning_rate": 0.00017168996947137976,
      "loss": 0.48597126007080077,
      "mean_token_accuracy": 0.8373165309429169,
      "num_tokens": 11632640.0,
      "step": 1420
    },
    {
      "entropy": 0.45168851464986803,
      "epoch": 0.7626666666666667,
      "grad_norm": 0.26867756247520447,
      "learning_rate": 0.00017129596441919204,
      "loss": 0.4792360782623291,
      "mean_token_accuracy": 0.8391267091035843,
      "num_tokens": 11714560.0,
      "step": 1430
    },
    {
      "entropy": 0.4700272098183632,
      "epoch": 0.768,
      "grad_norm": 0.32062190771102905,
      "learning_rate": 0.00017089969538067155,
      "loss": 0.4931813716888428,
      "mean_token_accuracy": 0.8345768064260483,
      "num_tokens": 11796480.0,
      "step": 1440
    },
    {
      "entropy": 0.45127319544553757,
      "epoch": 0.7733333333333333,
      "grad_norm": 0.25563088059425354,
      "learning_rate": 0.00017050117493924694,
      "loss": 0.47654094696044924,
      "mean_token_accuracy": 0.8394936442375183,
      "num_tokens": 11878400.0,
      "step": 1450
    },
    {
      "entropy": 0.4657404705882072,
      "epoch": 0.7786666666666666,
      "grad_norm": 0.28902241587638855,
      "learning_rate": 0.00017010041574983973,
      "loss": 0.4899322509765625,
      "mean_token_accuracy": 0.8358121335506439,
      "num_tokens": 11960320.0,
      "step": 1460
    },
    {
      "entropy": 0.46490112245082854,
      "epoch": 0.784,
      "grad_norm": 0.32127004861831665,
      "learning_rate": 0.00016969743053846237,
      "loss": 0.4905268669128418,
      "mean_token_accuracy": 0.8352984338998795,
      "num_tokens": 12042240.0,
      "step": 1470
    },
    {
      "entropy": 0.4582438513636589,
      "epoch": 0.7893333333333333,
      "grad_norm": 0.298467218875885,
      "learning_rate": 0.000169292232101814,
      "loss": 0.4845992088317871,
      "mean_token_accuracy": 0.8368395328521728,
      "num_tokens": 12124160.0,
      "step": 1480
    },
    {
      "entropy": 0.4644436791539192,
      "epoch": 0.7946666666666666,
      "grad_norm": 0.3031492233276367,
      "learning_rate": 0.0001688848333068743,
      "loss": 0.488465690612793,
      "mean_token_accuracy": 0.8364359110593795,
      "num_tokens": 12206080.0,
      "step": 1490
    },
    {
      "entropy": 0.45517687648534777,
      "epoch": 0.8,
      "grad_norm": 0.31369954347610474,
      "learning_rate": 0.00016847524709049475,
      "loss": 0.48354415893554686,
      "mean_token_accuracy": 0.8375244617462159,
      "num_tokens": 12288000.0,
      "step": 1500
    },
    {
      "entropy": 0.4555055618286133,
      "epoch": 0.8053333333333333,
      "grad_norm": 0.27002349495887756,
      "learning_rate": 0.0001680634864589879,
      "loss": 0.47807936668395995,
      "mean_token_accuracy": 0.8374877661466599,
      "num_tokens": 12369920.0,
      "step": 1510
    },
    {
      "entropy": 0.4547290876507759,
      "epoch": 0.8106666666666666,
      "grad_norm": 0.38989922404289246,
      "learning_rate": 0.0001676495644877143,
      "loss": 0.4785407066345215,
      "mean_token_accuracy": 0.838759782910347,
      "num_tokens": 12451840.0,
      "step": 1520
    },
    {
      "entropy": 0.4581173598766327,
      "epoch": 0.816,
      "grad_norm": 0.2933461368083954,
      "learning_rate": 0.00016723349432066738,
      "loss": 0.4821171283721924,
      "mean_token_accuracy": 0.8379403173923492,
      "num_tokens": 12533760.0,
      "step": 1530
    },
    {
      "entropy": 0.4574388861656189,
      "epoch": 0.8213333333333334,
      "grad_norm": 0.3411906063556671,
      "learning_rate": 0.00016681528917005596,
      "loss": 0.4809244155883789,
      "mean_token_accuracy": 0.8376956939697265,
      "num_tokens": 12615680.0,
      "step": 1540
    },
    {
      "entropy": 0.4678387776017189,
      "epoch": 0.8266666666666667,
      "grad_norm": 0.2791672646999359,
      "learning_rate": 0.0001663949623158848,
      "loss": 0.49310989379882814,
      "mean_token_accuracy": 0.8355430424213409,
      "num_tokens": 12697600.0,
      "step": 1550
    },
    {
      "entropy": 0.4656101256608963,
      "epoch": 0.832,
      "grad_norm": 0.3118132948875427,
      "learning_rate": 0.0001659725271055328,
      "loss": 0.4908637523651123,
      "mean_token_accuracy": 0.8354451984167099,
      "num_tokens": 12779520.0,
      "step": 1560
    },
    {
      "entropy": 0.4631996288895607,
      "epoch": 0.8373333333333334,
      "grad_norm": 0.2814883589744568,
      "learning_rate": 0.00016554799695332924,
      "loss": 0.48832063674926757,
      "mean_token_accuracy": 0.8357020586729049,
      "num_tokens": 12861440.0,
      "step": 1570
    },
    {
      "entropy": 0.47060815542936324,
      "epoch": 0.8426666666666667,
      "grad_norm": 0.3350438177585602,
      "learning_rate": 0.00016512138534012776,
      "loss": 0.49778051376342775,
      "mean_token_accuracy": 0.8336105734109879,
      "num_tokens": 12943360.0,
      "step": 1580
    },
    {
      "entropy": 0.46530798375606536,
      "epoch": 0.848,
      "grad_norm": 0.29841533303260803,
      "learning_rate": 0.00016469270581287836,
      "loss": 0.4874864101409912,
      "mean_token_accuracy": 0.8364970713853837,
      "num_tokens": 13025280.0,
      "step": 1590
    },
    {
      "entropy": 0.4572747603058815,
      "epoch": 0.8533333333333334,
      "grad_norm": 0.30778223276138306,
      "learning_rate": 0.0001642619719841971,
      "loss": 0.48714232444763184,
      "mean_token_accuracy": 0.8364970654249191,
      "num_tokens": 13107200.0,
      "step": 1600
    },
    {
      "entropy": 0.45623271018266676,
      "epoch": 0.8586666666666667,
      "grad_norm": 0.3078499734401703,
      "learning_rate": 0.0001638291975319339,
      "loss": 0.4790012359619141,
      "mean_token_accuracy": 0.8382216185331345,
      "num_tokens": 13189120.0,
      "step": 1610
    },
    {
      "entropy": 0.44585090428590773,
      "epoch": 0.864,
      "grad_norm": 0.3154717981815338,
      "learning_rate": 0.00016339439619873813,
      "loss": 0.46976747512817385,
      "mean_token_accuracy": 0.8412671208381652,
      "num_tokens": 13271040.0,
      "step": 1620
    },
    {
      "entropy": 0.45995187759399414,
      "epoch": 0.8693333333333333,
      "grad_norm": 0.3001498281955719,
      "learning_rate": 0.00016295758179162242,
      "loss": 0.48497829437255857,
      "mean_token_accuracy": 0.8371575325727463,
      "num_tokens": 13352960.0,
      "step": 1630
    },
    {
      "entropy": 0.4521583795547485,
      "epoch": 0.8746666666666667,
      "grad_norm": 0.29989272356033325,
      "learning_rate": 0.00016251876818152393,
      "loss": 0.4744596481323242,
      "mean_token_accuracy": 0.8407289624214173,
      "num_tokens": 13434880.0,
      "step": 1640
    },
    {
      "entropy": 0.45803203284740446,
      "epoch": 0.88,
      "grad_norm": 0.289101779460907,
      "learning_rate": 0.00016207796930286424,
      "loss": 0.4807485580444336,
      "mean_token_accuracy": 0.8380503863096237,
      "num_tokens": 13516800.0,
      "step": 1650
    },
    {
      "entropy": 0.4540840357542038,
      "epoch": 0.8853333333333333,
      "grad_norm": 0.2910986542701721,
      "learning_rate": 0.0001616351991531064,
      "loss": 0.4767774105072021,
      "mean_token_accuracy": 0.8387964785099029,
      "num_tokens": 13598720.0,
      "step": 1660
    },
    {
      "entropy": 0.4497424304485321,
      "epoch": 0.8906666666666667,
      "grad_norm": 0.3054055869579315,
      "learning_rate": 0.0001611904717923109,
      "loss": 0.4732930660247803,
      "mean_token_accuracy": 0.8405577272176743,
      "num_tokens": 13680640.0,
      "step": 1670
    },
    {
      "entropy": 0.4625942513346672,
      "epoch": 0.896,
      "grad_norm": 0.2968590259552002,
      "learning_rate": 0.00016074380134268888,
      "loss": 0.4878493309020996,
      "mean_token_accuracy": 0.835188364982605,
      "num_tokens": 13762560.0,
      "step": 1680
    },
    {
      "entropy": 0.4662549033761024,
      "epoch": 0.9013333333333333,
      "grad_norm": 0.2677576541900635,
      "learning_rate": 0.0001602952019881539,
      "loss": 0.49081854820251464,
      "mean_token_accuracy": 0.8343444228172302,
      "num_tokens": 13844480.0,
      "step": 1690
    },
    {
      "entropy": 0.4566287204623222,
      "epoch": 0.9066666666666666,
      "grad_norm": 0.24112841486930847,
      "learning_rate": 0.0001598446879738714,
      "loss": 0.47814083099365234,
      "mean_token_accuracy": 0.8376467674970627,
      "num_tokens": 13926400.0,
      "step": 1700
    },
    {
      "entropy": 0.4560120806097984,
      "epoch": 0.912,
      "grad_norm": 0.29659050703048706,
      "learning_rate": 0.00015939227360580634,
      "loss": 0.4790761470794678,
      "mean_token_accuracy": 0.8382705450057983,
      "num_tokens": 14008320.0,
      "step": 1710
    },
    {
      "entropy": 0.4614653393626213,
      "epoch": 0.9173333333333333,
      "grad_norm": 0.26346153020858765,
      "learning_rate": 0.000158937973250269,
      "loss": 0.48452486991882326,
      "mean_token_accuracy": 0.836607140302658,
      "num_tokens": 14090240.0,
      "step": 1720
    },
    {
      "entropy": 0.46237656772136687,
      "epoch": 0.9226666666666666,
      "grad_norm": 0.2936680018901825,
      "learning_rate": 0.00015848180133345877,
      "loss": 0.4856091022491455,
      "mean_token_accuracy": 0.8362279891967773,
      "num_tokens": 14172160.0,
      "step": 1730
    },
    {
      "entropy": 0.4579738602042198,
      "epoch": 0.928,
      "grad_norm": 0.27517297863960266,
      "learning_rate": 0.0001580237723410059,
      "loss": 0.48234920501708983,
      "mean_token_accuracy": 0.8365582197904586,
      "num_tokens": 14254080.0,
      "step": 1740
    },
    {
      "entropy": 0.45451856702566146,
      "epoch": 0.9333333333333333,
      "grad_norm": 0.3157144784927368,
      "learning_rate": 0.00015756390081751178,
      "loss": 0.48070473670959474,
      "mean_token_accuracy": 0.8386007815599441,
      "num_tokens": 14336000.0,
      "step": 1750
    },
    {
      "entropy": 0.4789514258503914,
      "epoch": 0.9386666666666666,
      "grad_norm": 0.3633069396018982,
      "learning_rate": 0.0001571022013660869,
      "loss": 0.5019748210906982,
      "mean_token_accuracy": 0.8305528372526169,
      "num_tokens": 14417920.0,
      "step": 1760
    },
    {
      "entropy": 0.45176429450511935,
      "epoch": 0.944,
      "grad_norm": 0.28711122274398804,
      "learning_rate": 0.0001566386886478871,
      "loss": 0.47701144218444824,
      "mean_token_accuracy": 0.8390533298254013,
      "num_tokens": 14499840.0,
      "step": 1770
    },
    {
      "entropy": 0.4583976581692696,
      "epoch": 0.9493333333333334,
      "grad_norm": 0.315071702003479,
      "learning_rate": 0.0001561733773816481,
      "loss": 0.48304171562194825,
      "mean_token_accuracy": 0.8380137056112289,
      "num_tokens": 14581760.0,
      "step": 1780
    },
    {
      "entropy": 0.45052715092897416,
      "epoch": 0.9546666666666667,
      "grad_norm": 0.23746146261692047,
      "learning_rate": 0.00015570628234321814,
      "loss": 0.4737369537353516,
      "mean_token_accuracy": 0.8416462779045105,
      "num_tokens": 14663680.0,
      "step": 1790
    },
    {
      "entropy": 0.45896155536174776,
      "epoch": 0.96,
      "grad_norm": 0.2549385726451874,
      "learning_rate": 0.00015523741836508866,
      "loss": 0.4805328369140625,
      "mean_token_accuracy": 0.8365949124097825,
      "num_tokens": 14745600.0,
      "step": 1800
    },
    {
      "entropy": 0.4620462045073509,
      "epoch": 0.9653333333333334,
      "grad_norm": 0.2818538248538971,
      "learning_rate": 0.00015476680033592332,
      "loss": 0.486307954788208,
      "mean_token_accuracy": 0.8363747537136078,
      "num_tokens": 14827520.0,
      "step": 1810
    },
    {
      "entropy": 0.45291213840246203,
      "epoch": 0.9706666666666667,
      "grad_norm": 0.2865256369113922,
      "learning_rate": 0.00015429444320008538,
      "loss": 0.47402515411376955,
      "mean_token_accuracy": 0.8393468648195267,
      "num_tokens": 14909440.0,
      "step": 1820
    },
    {
      "entropy": 0.4530337184667587,
      "epoch": 0.976,
      "grad_norm": 0.33044490218162537,
      "learning_rate": 0.00015382036195716293,
      "loss": 0.47954673767089845,
      "mean_token_accuracy": 0.8391144841909408,
      "num_tokens": 14991360.0,
      "step": 1830
    },
    {
      "entropy": 0.45839793235063553,
      "epoch": 0.9813333333333333,
      "grad_norm": 0.27488183975219727,
      "learning_rate": 0.00015334457166149267,
      "loss": 0.47922191619873045,
      "mean_token_accuracy": 0.83831946849823,
      "num_tokens": 15073280.0,
      "step": 1840
    },
    {
      "entropy": 0.4551577404141426,
      "epoch": 0.9866666666666667,
      "grad_norm": 0.3314872980117798,
      "learning_rate": 0.00015286708742168184,
      "loss": 0.4800881862640381,
      "mean_token_accuracy": 0.8381115466356277,
      "num_tokens": 15155200.0,
      "step": 1850
    },
    {
      "entropy": 0.46016068756580353,
      "epoch": 0.992,
      "grad_norm": 0.2758563458919525,
      "learning_rate": 0.00015238792440012858,
      "loss": 0.4828464508056641,
      "mean_token_accuracy": 0.8374143809080123,
      "num_tokens": 15237120.0,
      "step": 1860
    },
    {
      "entropy": 0.45787719786167147,
      "epoch": 0.9973333333333333,
      "grad_norm": 0.25891438126564026,
      "learning_rate": 0.00015190709781254025,
      "loss": 0.4816002368927002,
      "mean_token_accuracy": 0.8372431457042694,
      "num_tokens": 15319040.0,
      "step": 1870
    },
    {
      "entropy": 0.47158827930688857,
      "epoch": 1.0026666666666666,
      "grad_norm": 0.24438625574111938,
      "learning_rate": 0.0001514246229274503,
      "loss": 0.4930420398712158,
      "mean_token_accuracy": 0.8336717247962951,
      "num_tokens": 15400960.0,
      "step": 1880
    },
    {
      "entropy": 0.45009716153144835,
      "epoch": 1.008,
      "grad_norm": 0.26381391286849976,
      "learning_rate": 0.00015094051506573362,
      "loss": 0.4731320381164551,
      "mean_token_accuracy": 0.8403864920139312,
      "num_tokens": 15482880.0,
      "step": 1890
    },
    {
      "entropy": 0.4532205924391747,
      "epoch": 1.0133333333333334,
      "grad_norm": 0.35160163044929504,
      "learning_rate": 0.00015045478960011983,
      "loss": 0.47685995101928713,
      "mean_token_accuracy": 0.8391389429569245,
      "num_tokens": 15564800.0,
      "step": 1900
    },
    {
      "entropy": 0.4642463266849518,
      "epoch": 1.0186666666666666,
      "grad_norm": 0.3478567600250244,
      "learning_rate": 0.00014996746195470508,
      "loss": 0.4873044013977051,
      "mean_token_accuracy": 0.8356409013271332,
      "num_tokens": 15646720.0,
      "step": 1910
    },
    {
      "entropy": 0.45275849252939226,
      "epoch": 1.024,
      "grad_norm": 0.2523830235004425,
      "learning_rate": 0.00014947854760446246,
      "loss": 0.47649006843566893,
      "mean_token_accuracy": 0.8383806347846985,
      "num_tokens": 15728640.0,
      "step": 1920
    },
    {
      "entropy": 0.455821231007576,
      "epoch": 1.0293333333333334,
      "grad_norm": 0.2520630657672882,
      "learning_rate": 0.00014898806207475047,
      "loss": 0.48169946670532227,
      "mean_token_accuracy": 0.8364726066589355,
      "num_tokens": 15810560.0,
      "step": 1930
    },
    {
      "entropy": 0.46014908105134966,
      "epoch": 1.0346666666666666,
      "grad_norm": 0.27884477376937866,
      "learning_rate": 0.00014849602094082006,
      "loss": 0.4830568790435791,
      "mean_token_accuracy": 0.8373654633760452,
      "num_tokens": 15892480.0,
      "step": 1940
    },
    {
      "entropy": 0.4596777975559235,
      "epoch": 1.04,
      "grad_norm": 0.29964160919189453,
      "learning_rate": 0.00014800243982731996,
      "loss": 0.4837499618530273,
      "mean_token_accuracy": 0.8360567510128021,
      "num_tokens": 15974400.0,
      "step": 1950
    },
    {
      "entropy": 0.4516203835606575,
      "epoch": 1.0453333333333332,
      "grad_norm": 0.28172388672828674,
      "learning_rate": 0.0001475073344078006,
      "loss": 0.4748356819152832,
      "mean_token_accuracy": 0.8389799416065216,
      "num_tokens": 16056320.0,
      "step": 1960
    },
    {
      "entropy": 0.4465511843562126,
      "epoch": 1.0506666666666666,
      "grad_norm": 0.31371772289276123,
      "learning_rate": 0.0001470107204042165,
      "loss": 0.46893014907836916,
      "mean_token_accuracy": 0.8421110540628434,
      "num_tokens": 16138240.0,
      "step": 1970
    },
    {
      "entropy": 0.4643126592040062,
      "epoch": 1.056,
      "grad_norm": 0.3073205351829529,
      "learning_rate": 0.0001465126135864267,
      "loss": 0.4890984058380127,
      "mean_token_accuracy": 0.834650194644928,
      "num_tokens": 16220160.0,
      "step": 1980
    },
    {
      "entropy": 0.46284378319978714,
      "epoch": 1.0613333333333332,
      "grad_norm": 0.25584980845451355,
      "learning_rate": 0.00014601302977169435,
      "loss": 0.48810677528381347,
      "mean_token_accuracy": 0.8347969740629196,
      "num_tokens": 16302080.0,
      "step": 1990
    },
    {
      "entropy": 0.4598368525505066,
      "epoch": 1.0666666666666667,
      "grad_norm": 0.23851245641708374,
      "learning_rate": 0.00014551198482418425,
      "loss": 0.48256711959838866,
      "mean_token_accuracy": 0.8383316993713379,
      "num_tokens": 16384000.0,
      "step": 2000
    },
    {
      "entropy": 0.463072144985199,
      "epoch": 1.072,
      "grad_norm": 0.24018454551696777,
      "learning_rate": 0.0001450094946544591,
      "loss": 0.48588037490844727,
      "mean_token_accuracy": 0.8369985342025756,
      "num_tokens": 16465920.0,
      "step": 2010
    },
    {
      "entropy": 0.4554838389158249,
      "epoch": 1.0773333333333333,
      "grad_norm": 0.26262980699539185,
      "learning_rate": 0.00014450557521897435,
      "loss": 0.47637052536010743,
      "mean_token_accuracy": 0.8392734855413437,
      "num_tokens": 16547840.0,
      "step": 2020
    },
    {
      "entropy": 0.46434304863214493,
      "epoch": 1.0826666666666667,
      "grad_norm": 0.2791703939437866,
      "learning_rate": 0.0001440002425195713,
      "loss": 0.4899892807006836,
      "mean_token_accuracy": 0.8342221170663834,
      "num_tokens": 16629760.0,
      "step": 2030
    },
    {
      "entropy": 0.46599327176809313,
      "epoch": 1.088,
      "grad_norm": 0.2516133487224579,
      "learning_rate": 0.00014349351260296928,
      "loss": 0.4881533145904541,
      "mean_token_accuracy": 0.8354085087776184,
      "num_tokens": 16711680.0,
      "step": 2040
    },
    {
      "entropy": 0.4491189017891884,
      "epoch": 1.0933333333333333,
      "grad_norm": 0.3033424913883209,
      "learning_rate": 0.00014298540156025585,
      "loss": 0.47525844573974607,
      "mean_token_accuracy": 0.840325340628624,
      "num_tokens": 16793600.0,
      "step": 2050
    },
    {
      "entropy": 0.44756190329790113,
      "epoch": 1.0986666666666667,
      "grad_norm": 0.2564951181411743,
      "learning_rate": 0.00014247592552637584,
      "loss": 0.46916990280151366,
      "mean_token_accuracy": 0.8411081224679947,
      "num_tokens": 16875520.0,
      "step": 2060
    },
    {
      "entropy": 0.4501015737652779,
      "epoch": 1.104,
      "grad_norm": 0.26202088594436646,
      "learning_rate": 0.00014196510067961916,
      "loss": 0.4716159820556641,
      "mean_token_accuracy": 0.8391389429569245,
      "num_tokens": 16957440.0,
      "step": 2070
    },
    {
      "entropy": 0.4566748946905136,
      "epoch": 1.1093333333333333,
      "grad_norm": 0.27088451385498047,
      "learning_rate": 0.0001414529432411068,
      "loss": 0.47965083122253416,
      "mean_token_accuracy": 0.8379158496856689,
      "num_tokens": 17039360.0,
      "step": 2080
    },
    {
      "entropy": 0.454061321914196,
      "epoch": 1.1146666666666667,
      "grad_norm": 0.24049293994903564,
      "learning_rate": 0.00014093946947427608,
      "loss": 0.4773685455322266,
      "mean_token_accuracy": 0.8391756355762482,
      "num_tokens": 17121280.0,
      "step": 2090
    },
    {
      "entropy": 0.4559741377830505,
      "epoch": 1.12,
      "grad_norm": 0.26275789737701416,
      "learning_rate": 0.00014042469568436382,
      "loss": 0.4781862735748291,
      "mean_token_accuracy": 0.8372553765773774,
      "num_tokens": 17203200.0,
      "step": 2100
    },
    {
      "entropy": 0.45210961401462557,
      "epoch": 1.1253333333333333,
      "grad_norm": 0.2621893286705017,
      "learning_rate": 0.00013990863821788897,
      "loss": 0.4754960060119629,
      "mean_token_accuracy": 0.8409368902444839,
      "num_tokens": 17285120.0,
      "step": 2110
    },
    {
      "entropy": 0.46071210950613023,
      "epoch": 1.1306666666666667,
      "grad_norm": 0.2665644586086273,
      "learning_rate": 0.00013939131346213306,
      "loss": 0.48028388023376467,
      "mean_token_accuracy": 0.8362035155296326,
      "num_tokens": 17367040.0,
      "step": 2120
    },
    {
      "entropy": 0.46311234384775163,
      "epoch": 1.1360000000000001,
      "grad_norm": 0.2657349407672882,
      "learning_rate": 0.00013887273784462032,
      "loss": 0.48636879920959475,
      "mean_token_accuracy": 0.8359466671943665,
      "num_tokens": 17448960.0,
      "step": 2130
    },
    {
      "entropy": 0.45575994104146955,
      "epoch": 1.1413333333333333,
      "grad_norm": 0.31427016854286194,
      "learning_rate": 0.00013835292783259567,
      "loss": 0.4794747829437256,
      "mean_token_accuracy": 0.8384173214435577,
      "num_tokens": 17530880.0,
      "step": 2140
    },
    {
      "entropy": 0.4551381945610046,
      "epoch": 1.1466666666666667,
      "grad_norm": 0.28378787636756897,
      "learning_rate": 0.000137831899932502,
      "loss": 0.47565302848815916,
      "mean_token_accuracy": 0.8382827818393708,
      "num_tokens": 17612800.0,
      "step": 2150
    },
    {
      "entropy": 0.4576284572482109,
      "epoch": 1.152,
      "grad_norm": 0.2831465005874634,
      "learning_rate": 0.00013730967068945586,
      "loss": 0.4819480419158936,
      "mean_token_accuracy": 0.8363502949476243,
      "num_tokens": 17694720.0,
      "step": 2160
    },
    {
      "entropy": 0.4545293599367142,
      "epoch": 1.1573333333333333,
      "grad_norm": 0.2564343810081482,
      "learning_rate": 0.0001367862566867223,
      "loss": 0.4760554313659668,
      "mean_token_accuracy": 0.8397627204656601,
      "num_tokens": 17776640.0,
      "step": 2170
    },
    {
      "entropy": 0.4436169371008873,
      "epoch": 1.1626666666666667,
      "grad_norm": 0.34586217999458313,
      "learning_rate": 0.00013626167454518798,
      "loss": 0.4654336452484131,
      "mean_token_accuracy": 0.8422333627939225,
      "num_tokens": 17858560.0,
      "step": 2180
    },
    {
      "entropy": 0.45734113454818726,
      "epoch": 1.168,
      "grad_norm": 0.2863917946815491,
      "learning_rate": 0.00013573594092283363,
      "loss": 0.47915191650390626,
      "mean_token_accuracy": 0.8372431486845017,
      "num_tokens": 17940480.0,
      "step": 2190
    },
    {
      "entropy": 0.45436161011457443,
      "epoch": 1.1733333333333333,
      "grad_norm": 0.2896290123462677,
      "learning_rate": 0.00013520907251420486,
      "loss": 0.4764246940612793,
      "mean_token_accuracy": 0.8400929540395736,
      "num_tokens": 18022400.0,
      "step": 2200
    },
    {
      "entropy": 0.4608956053853035,
      "epoch": 1.1786666666666668,
      "grad_norm": 0.3019149899482727,
      "learning_rate": 0.00013468108604988238,
      "loss": 0.4831700325012207,
      "mean_token_accuracy": 0.836545991897583,
      "num_tokens": 18104320.0,
      "step": 2210
    },
    {
      "entropy": 0.4481834441423416,
      "epoch": 1.184,
      "grad_norm": 0.2716231644153595,
      "learning_rate": 0.00013415199829595024,
      "loss": 0.47463126182556153,
      "mean_token_accuracy": 0.8393468677997589,
      "num_tokens": 18186240.0,
      "step": 2220
    },
    {
      "entropy": 0.44566879272460935,
      "epoch": 1.1893333333333334,
      "grad_norm": 0.3151785135269165,
      "learning_rate": 0.00013362182605346375,
      "loss": 0.46677536964416505,
      "mean_token_accuracy": 0.8417441308498382,
      "num_tokens": 18268160.0,
      "step": 2230
    },
    {
      "entropy": 0.4517290756106377,
      "epoch": 1.1946666666666665,
      "grad_norm": 0.2718968987464905,
      "learning_rate": 0.00013309058615791606,
      "loss": 0.4716118335723877,
      "mean_token_accuracy": 0.8401296406984329,
      "num_tokens": 18350080.0,
      "step": 2240
    },
    {
      "entropy": 0.4497871816158295,
      "epoch": 1.2,
      "grad_norm": 0.25722774863243103,
      "learning_rate": 0.0001325582954787031,
      "loss": 0.4738447666168213,
      "mean_token_accuracy": 0.840288656949997,
      "num_tokens": 18432000.0,
      "step": 2250
    },
    {
      "entropy": 0.4530187055468559,
      "epoch": 1.2053333333333334,
      "grad_norm": 0.302322655916214,
      "learning_rate": 0.00013202497091858843,
      "loss": 0.4760636329650879,
      "mean_token_accuracy": 0.8383561640977859,
      "num_tokens": 18513920.0,
      "step": 2260
    },
    {
      "entropy": 0.4625497832894325,
      "epoch": 1.2106666666666666,
      "grad_norm": 0.25169312953948975,
      "learning_rate": 0.00013149062941316612,
      "loss": 0.4825915813446045,
      "mean_token_accuracy": 0.8359466671943665,
      "num_tokens": 18595840.0,
      "step": 2270
    },
    {
      "entropy": 0.45275418907403947,
      "epoch": 1.216,
      "grad_norm": 0.287241131067276,
      "learning_rate": 0.00013095528793032317,
      "loss": 0.4763150691986084,
      "mean_token_accuracy": 0.8392123311758042,
      "num_tokens": 18677760.0,
      "step": 2280
    },
    {
      "entropy": 0.46166481375694274,
      "epoch": 1.2213333333333334,
      "grad_norm": 0.2802252471446991,
      "learning_rate": 0.00013041896346970057,
      "loss": 0.4830777645111084,
      "mean_token_accuracy": 0.8361179023981095,
      "num_tokens": 18759680.0,
      "step": 2290
    },
    {
      "entropy": 0.45855253785848615,
      "epoch": 1.2266666666666666,
      "grad_norm": 0.25347864627838135,
      "learning_rate": 0.00012988167306215355,
      "loss": 0.4811714649200439,
      "mean_token_accuracy": 0.8375122308731079,
      "num_tokens": 18841600.0,
      "step": 2300
    },
    {
      "entropy": 0.4519998535513878,
      "epoch": 1.232,
      "grad_norm": 0.25941354036331177,
      "learning_rate": 0.00012934343376921076,
      "loss": 0.47400102615356443,
      "mean_token_accuracy": 0.839065557718277,
      "num_tokens": 18923520.0,
      "step": 2310
    },
    {
      "entropy": 0.45826067179441454,
      "epoch": 1.2373333333333334,
      "grad_norm": 0.25858303904533386,
      "learning_rate": 0.0001288042626825325,
      "loss": 0.4782583236694336,
      "mean_token_accuracy": 0.8367049902677536,
      "num_tokens": 19005440.0,
      "step": 2320
    },
    {
      "entropy": 0.46559557765722276,
      "epoch": 1.2426666666666666,
      "grad_norm": 0.2688336670398712,
      "learning_rate": 0.0001282641769233679,
      "loss": 0.4879752159118652,
      "mean_token_accuracy": 0.8360200554132462,
      "num_tokens": 19087360.0,
      "step": 2330
    },
    {
      "entropy": 0.4653527677059174,
      "epoch": 1.248,
      "grad_norm": 0.25175878405570984,
      "learning_rate": 0.00012772319364201136,
      "loss": 0.486749267578125,
      "mean_token_accuracy": 0.834307724237442,
      "num_tokens": 19169280.0,
      "step": 2340
    },
    {
      "entropy": 0.4516812652349472,
      "epoch": 1.2533333333333334,
      "grad_norm": 0.2519511580467224,
      "learning_rate": 0.00012718133001725783,
      "loss": 0.4739360332489014,
      "mean_token_accuracy": 0.8411937415599823,
      "num_tokens": 19251200.0,
      "step": 2350
    },
    {
      "entropy": 0.46116370260715484,
      "epoch": 1.2586666666666666,
      "grad_norm": 0.28249990940093994,
      "learning_rate": 0.0001266386032558574,
      "loss": 0.4816390037536621,
      "mean_token_accuracy": 0.8372431486845017,
      "num_tokens": 19333120.0,
      "step": 2360
    },
    {
      "entropy": 0.45341705083847045,
      "epoch": 1.264,
      "grad_norm": 0.26403769850730896,
      "learning_rate": 0.00012609503059196882,
      "loss": 0.47722954750061036,
      "mean_token_accuracy": 0.8380748510360718,
      "num_tokens": 19415040.0,
      "step": 2370
    },
    {
      "entropy": 0.4603477954864502,
      "epoch": 1.2693333333333334,
      "grad_norm": 0.3148926794528961,
      "learning_rate": 0.0001255506292866123,
      "loss": 0.48342156410217285,
      "mean_token_accuracy": 0.8367539137601853,
      "num_tokens": 19496960.0,
      "step": 2380
    },
    {
      "entropy": 0.4594887912273407,
      "epoch": 1.2746666666666666,
      "grad_norm": 0.3305438756942749,
      "learning_rate": 0.00012500541662712134,
      "loss": 0.4802074909210205,
      "mean_token_accuracy": 0.8359955966472625,
      "num_tokens": 19578880.0,
      "step": 2390
    },
    {
      "entropy": 0.4564969763159752,
      "epoch": 1.28,
      "grad_norm": 0.3165145814418793,
      "learning_rate": 0.00012445940992659383,
      "loss": 0.47836689949035643,
      "mean_token_accuracy": 0.8378913849592209,
      "num_tokens": 19660800.0,
      "step": 2400
    },
    {
      "entropy": 0.46479198932647703,
      "epoch": 1.2853333333333334,
      "grad_norm": 0.3122265934944153,
      "learning_rate": 0.0001239126265233422,
      "loss": 0.48781919479370117,
      "mean_token_accuracy": 0.8335616409778595,
      "num_tokens": 19742720.0,
      "step": 2410
    },
    {
      "entropy": 0.45316422283649443,
      "epoch": 1.2906666666666666,
      "grad_norm": 0.256700724363327,
      "learning_rate": 0.00012336508378034292,
      "loss": 0.4735256195068359,
      "mean_token_accuracy": 0.8383439272642136,
      "num_tokens": 19824640.0,
      "step": 2420
    },
    {
      "entropy": 0.45666906982660294,
      "epoch": 1.296,
      "grad_norm": 0.2838805913925171,
      "learning_rate": 0.0001228167990846851,
      "loss": 0.48002142906188966,
      "mean_token_accuracy": 0.8373654574155808,
      "num_tokens": 19906560.0,
      "step": 2430
    },
    {
      "entropy": 0.45154699981212615,
      "epoch": 1.3013333333333335,
      "grad_norm": 0.25855761766433716,
      "learning_rate": 0.0001222677898470184,
      "loss": 0.4715434074401855,
      "mean_token_accuracy": 0.8396771043539047,
      "num_tokens": 19988480.0,
      "step": 2440
    },
    {
      "entropy": 0.45676193833351136,
      "epoch": 1.3066666666666666,
      "grad_norm": 0.2858618497848511,
      "learning_rate": 0.00012171807350100009,
      "loss": 0.47927446365356446,
      "mean_token_accuracy": 0.8368639945983887,
      "num_tokens": 20070400.0,
      "step": 2450
    },
    {
      "entropy": 0.46133691966533663,
      "epoch": 1.312,
      "grad_norm": 0.271575927734375,
      "learning_rate": 0.00012116766750274153,
      "loss": 0.4815515518188477,
      "mean_token_accuracy": 0.8367539137601853,
      "num_tokens": 20152320.0,
      "step": 2460
    },
    {
      "entropy": 0.44794601798057554,
      "epoch": 1.3173333333333335,
      "grad_norm": 0.306148499250412,
      "learning_rate": 0.00012061658933025388,
      "loss": 0.4697227954864502,
      "mean_token_accuracy": 0.8407778859138488,
      "num_tokens": 20234240.0,
      "step": 2470
    },
    {
      "entropy": 0.45496426373720167,
      "epoch": 1.3226666666666667,
      "grad_norm": 0.26636722683906555,
      "learning_rate": 0.00012006485648289292,
      "loss": 0.4774960517883301,
      "mean_token_accuracy": 0.8379892438650132,
      "num_tokens": 20316160.0,
      "step": 2480
    },
    {
      "entropy": 0.44859183877706527,
      "epoch": 1.328,
      "grad_norm": 0.2544063925743103,
      "learning_rate": 0.0001195124864808035,
      "loss": 0.468443489074707,
      "mean_token_accuracy": 0.840484344959259,
      "num_tokens": 20398080.0,
      "step": 2490
    },
    {
      "entropy": 0.4587726712226868,
      "epoch": 1.3333333333333333,
      "grad_norm": 0.27513957023620605,
      "learning_rate": 0.00011895949686436324,
      "loss": 0.48148107528686523,
      "mean_token_accuracy": 0.8370841473340989,
      "num_tokens": 20480000.0,
      "step": 2500
    },
    {
      "entropy": 0.45128536522388457,
      "epoch": 1.3386666666666667,
      "grad_norm": 0.22385765612125397,
      "learning_rate": 0.00011840590519362542,
      "loss": 0.4761323928833008,
      "mean_token_accuracy": 0.8380748510360718,
      "num_tokens": 20561920.0,
      "step": 2510
    },
    {
      "entropy": 0.45485699772834776,
      "epoch": 1.3439999999999999,
      "grad_norm": 0.28511685132980347,
      "learning_rate": 0.0001178517290477614,
      "loss": 0.4757716178894043,
      "mean_token_accuracy": 0.8388576328754425,
      "num_tokens": 20643840.0,
      "step": 2520
    },
    {
      "entropy": 0.4568641513586044,
      "epoch": 1.3493333333333333,
      "grad_norm": 0.2950267791748047,
      "learning_rate": 0.00011729698602450241,
      "loss": 0.4761605739593506,
      "mean_token_accuracy": 0.839200097322464,
      "num_tokens": 20725760.0,
      "step": 2530
    },
    {
      "entropy": 0.4652768924832344,
      "epoch": 1.3546666666666667,
      "grad_norm": 0.2946852445602417,
      "learning_rate": 0.00011674169373958075,
      "loss": 0.49057469367980955,
      "mean_token_accuracy": 0.8329134047031402,
      "num_tokens": 20807680.0,
      "step": 2540
    },
    {
      "entropy": 0.45410631895065307,
      "epoch": 1.3599999999999999,
      "grad_norm": 0.2633431553840637,
      "learning_rate": 0.00011618586982617041,
      "loss": 0.4750985145568848,
      "mean_token_accuracy": 0.8373532265424728,
      "num_tokens": 20889600.0,
      "step": 2550
    },
    {
      "entropy": 0.4602236285805702,
      "epoch": 1.3653333333333333,
      "grad_norm": 0.27011752128601074,
      "learning_rate": 0.00011562953193432706,
      "loss": 0.4808211326599121,
      "mean_token_accuracy": 0.8370107531547546,
      "num_tokens": 20971520.0,
      "step": 2560
    },
    {
      "entropy": 0.4602127745747566,
      "epoch": 1.3706666666666667,
      "grad_norm": 0.27184605598449707,
      "learning_rate": 0.00011507269773042769,
      "loss": 0.47937779426574706,
      "mean_token_accuracy": 0.835934442281723,
      "num_tokens": 21053440.0,
      "step": 2570
    },
    {
      "entropy": 0.45975601971149443,
      "epoch": 1.376,
      "grad_norm": 0.2988097667694092,
      "learning_rate": 0.0001145153848966095,
      "loss": 0.4820730686187744,
      "mean_token_accuracy": 0.8374633014202117,
      "num_tokens": 21135360.0,
      "step": 2580
    },
    {
      "entropy": 0.45733146369457245,
      "epoch": 1.3813333333333333,
      "grad_norm": 0.3084525763988495,
      "learning_rate": 0.00011395761113020855,
      "loss": 0.48058257102966306,
      "mean_token_accuracy": 0.8370719224214553,
      "num_tokens": 21217280.0,
      "step": 2590
    },
    {
      "entropy": 0.455784310400486,
      "epoch": 1.3866666666666667,
      "grad_norm": 0.28003671765327454,
      "learning_rate": 0.00011339939414319763,
      "loss": 0.4752959251403809,
      "mean_token_accuracy": 0.838588547706604,
      "num_tokens": 21299200.0,
      "step": 2600
    },
    {
      "entropy": 0.4473429873585701,
      "epoch": 1.392,
      "grad_norm": 0.25539201498031616,
      "learning_rate": 0.00011284075166162393,
      "loss": 0.46677088737487793,
      "mean_token_accuracy": 0.8413160413503646,
      "num_tokens": 21381120.0,
      "step": 2610
    },
    {
      "entropy": 0.45877171605825423,
      "epoch": 1.3973333333333333,
      "grad_norm": 0.2671823501586914,
      "learning_rate": 0.00011228170142504613,
      "loss": 0.4821139335632324,
      "mean_token_accuracy": 0.8361668348312378,
      "num_tokens": 21463040.0,
      "step": 2620
    },
    {
      "entropy": 0.46888451725244523,
      "epoch": 1.4026666666666667,
      "grad_norm": 0.2629116475582123,
      "learning_rate": 0.00011172226118597104,
      "loss": 0.49042186737060545,
      "mean_token_accuracy": 0.8342832714319229,
      "num_tokens": 21544960.0,
      "step": 2630
    },
    {
      "entropy": 0.4596924975514412,
      "epoch": 1.408,
      "grad_norm": 0.2614009380340576,
      "learning_rate": 0.00011116244870928992,
      "loss": 0.48156285285949707,
      "mean_token_accuracy": 0.8361545979976654,
      "num_tokens": 21626880.0,
      "step": 2640
    },
    {
      "entropy": 0.45862935185432435,
      "epoch": 1.4133333333333333,
      "grad_norm": 0.2913917899131775,
      "learning_rate": 0.00011060228177171434,
      "loss": 0.4771733283996582,
      "mean_token_accuracy": 0.8374266177415848,
      "num_tokens": 21708800.0,
      "step": 2650
    },
    {
      "entropy": 0.4604292169213295,
      "epoch": 1.4186666666666667,
      "grad_norm": 0.2865404188632965,
      "learning_rate": 0.00011004177816121171,
      "loss": 0.48269853591918943,
      "mean_token_accuracy": 0.8351761251688004,
      "num_tokens": 21790720.0,
      "step": 2660
    },
    {
      "entropy": 0.4481583312153816,
      "epoch": 1.424,
      "grad_norm": 0.27249985933303833,
      "learning_rate": 0.00010948095567644041,
      "loss": 0.4679734230041504,
      "mean_token_accuracy": 0.8396526426076889,
      "num_tokens": 21872640.0,
      "step": 2670
    },
    {
      "entropy": 0.45001966059207915,
      "epoch": 1.4293333333333333,
      "grad_norm": 0.30754292011260986,
      "learning_rate": 0.00010891983212618455,
      "loss": 0.4728683948516846,
      "mean_token_accuracy": 0.8407044976949691,
      "num_tokens": 21954560.0,
      "step": 2680
    },
    {
      "entropy": 0.45329019874334336,
      "epoch": 1.4346666666666668,
      "grad_norm": 0.2550135552883148,
      "learning_rate": 0.0001083584253287886,
      "loss": 0.47282838821411133,
      "mean_token_accuracy": 0.8387842506170273,
      "num_tokens": 22036480.0,
      "step": 2690
    },
    {
      "entropy": 0.44966589063405993,
      "epoch": 1.44,
      "grad_norm": 0.25526052713394165,
      "learning_rate": 0.00010779675311159134,
      "loss": 0.4705975532531738,
      "mean_token_accuracy": 0.8407411932945251,
      "num_tokens": 22118400.0,
      "step": 2700
    },
    {
      "entropy": 0.4490411922335625,
      "epoch": 1.4453333333333334,
      "grad_norm": 0.2486724704504013,
      "learning_rate": 0.00010723483331035999,
      "loss": 0.4715598106384277,
      "mean_token_accuracy": 0.8399339526891708,
      "num_tokens": 22200320.0,
      "step": 2710
    },
    {
      "entropy": 0.4678569436073303,
      "epoch": 1.4506666666666668,
      "grad_norm": 0.2844434976577759,
      "learning_rate": 0.00010667268376872374,
      "loss": 0.4862506866455078,
      "mean_token_accuracy": 0.8346746563911438,
      "num_tokens": 22282240.0,
      "step": 2720
    },
    {
      "entropy": 0.468651220202446,
      "epoch": 1.456,
      "grad_norm": 0.24747876822948456,
      "learning_rate": 0.00010611032233760707,
      "loss": 0.4911623477935791,
      "mean_token_accuracy": 0.8339285761117935,
      "num_tokens": 22364160.0,
      "step": 2730
    },
    {
      "entropy": 0.43975051641464236,
      "epoch": 1.4613333333333334,
      "grad_norm": 0.274668425321579,
      "learning_rate": 0.00010554776687466304,
      "loss": 0.4601419448852539,
      "mean_token_accuracy": 0.8426247596740722,
      "num_tokens": 22446080.0,
      "step": 2740
    },
    {
      "entropy": 0.4610159292817116,
      "epoch": 1.4666666666666668,
      "grad_norm": 0.27988290786743164,
      "learning_rate": 0.00010498503524370602,
      "loss": 0.47959046363830565,
      "mean_token_accuracy": 0.8381237655878067,
      "num_tokens": 22528000.0,
      "step": 2750
    },
    {
      "entropy": 0.4528918221592903,
      "epoch": 1.472,
      "grad_norm": 0.2467563897371292,
      "learning_rate": 0.00010442214531414473,
      "loss": 0.4767890453338623,
      "mean_token_accuracy": 0.8378791570663452,
      "num_tokens": 22609920.0,
      "step": 2760
    },
    {
      "entropy": 0.4701098412275314,
      "epoch": 1.4773333333333334,
      "grad_norm": 0.2544519007205963,
      "learning_rate": 0.00010385911496041449,
      "loss": 0.4897502899169922,
      "mean_token_accuracy": 0.8342588007450104,
      "num_tokens": 22691840.0,
      "step": 2770
    },
    {
      "entropy": 0.4421505406498909,
      "epoch": 1.4826666666666668,
      "grad_norm": 0.23783396184444427,
      "learning_rate": 0.00010329596206140982,
      "loss": 0.4636669635772705,
      "mean_token_accuracy": 0.8420498996973038,
      "num_tokens": 22773760.0,
      "step": 2780
    },
    {
      "entropy": 0.44983386248350143,
      "epoch": 1.488,
      "grad_norm": 0.2746131420135498,
      "learning_rate": 0.00010273270449991657,
      "loss": 0.46945581436157224,
      "mean_token_accuracy": 0.841169273853302,
      "num_tokens": 22855680.0,
      "step": 2790
    },
    {
      "entropy": 0.45045541822910307,
      "epoch": 1.4933333333333334,
      "grad_norm": 0.24367398023605347,
      "learning_rate": 0.00010216936016204415,
      "loss": 0.47146148681640626,
      "mean_token_accuracy": 0.8398238748311997,
      "num_tokens": 22937600.0,
      "step": 2800
    },
    {
      "entropy": 0.44976121336221697,
      "epoch": 1.4986666666666666,
      "grad_norm": 0.2660079896450043,
      "learning_rate": 0.00010160594693665759,
      "loss": 0.4694420337677002,
      "mean_token_accuracy": 0.8408023446798325,
      "num_tokens": 23019520.0,
      "step": 2810
    },
    {
      "entropy": 0.44268113672733306,
      "epoch": 1.504,
      "grad_norm": 0.24104079604148865,
      "learning_rate": 0.00010104248271480942,
      "loss": 0.46308207511901855,
      "mean_token_accuracy": 0.8417074382305145,
      "num_tokens": 23101440.0,
      "step": 2820
    },
    {
      "entropy": 0.4463737368583679,
      "epoch": 1.5093333333333332,
      "grad_norm": 0.3583536446094513,
      "learning_rate": 0.00010047898538917142,
      "loss": 0.4680209159851074,
      "mean_token_accuracy": 0.841915363073349,
      "num_tokens": 23183360.0,
      "step": 2830
    },
    {
      "entropy": 0.4463341489434242,
      "epoch": 1.5146666666666668,
      "grad_norm": 0.23940922319889069,
      "learning_rate": 9.991547285346673e-05,
      "loss": 0.4663118362426758,
      "mean_token_accuracy": 0.8412793576717377,
      "num_tokens": 23265280.0,
      "step": 2840
    },
    {
      "entropy": 0.45618336647748947,
      "epoch": 1.52,
      "grad_norm": 0.2739761173725128,
      "learning_rate": 9.93519630019014e-05,
      "loss": 0.4754621505737305,
      "mean_token_accuracy": 0.8386374741792679,
      "num_tokens": 23347200.0,
      "step": 2850
    },
    {
      "entropy": 0.44509555846452714,
      "epoch": 1.5253333333333332,
      "grad_norm": 0.29173797369003296,
      "learning_rate": 9.878847372859627e-05,
      "loss": 0.4643296718597412,
      "mean_token_accuracy": 0.8414750456809997,
      "num_tokens": 23429120.0,
      "step": 2860
    },
    {
      "entropy": 0.44736330956220627,
      "epoch": 1.5306666666666666,
      "grad_norm": 0.28082218766212463,
      "learning_rate": 9.822502292701871e-05,
      "loss": 0.47009940147399903,
      "mean_token_accuracy": 0.839848330616951,
      "num_tokens": 23511040.0,
      "step": 2870
    },
    {
      "entropy": 0.44525205343961716,
      "epoch": 1.536,
      "grad_norm": 0.2465215027332306,
      "learning_rate": 9.766162848941445e-05,
      "loss": 0.4627802848815918,
      "mean_token_accuracy": 0.8423067510128022,
      "num_tokens": 23592960.0,
      "step": 2880
    },
    {
      "entropy": 0.45830114632844926,
      "epoch": 1.5413333333333332,
      "grad_norm": 0.29059696197509766,
      "learning_rate": 9.709830830623937e-05,
      "loss": 0.4771836757659912,
      "mean_token_accuracy": 0.8390410989522934,
      "num_tokens": 23674880.0,
      "step": 2890
    },
    {
      "entropy": 0.46312975883483887,
      "epoch": 1.5466666666666666,
      "grad_norm": 0.30086666345596313,
      "learning_rate": 9.653508026559144e-05,
      "loss": 0.48303637504577634,
      "mean_token_accuracy": 0.8355675101280212,
      "num_tokens": 23756800.0,
      "step": 2900
    },
    {
      "entropy": 0.4574548527598381,
      "epoch": 1.552,
      "grad_norm": 0.24203862249851227,
      "learning_rate": 9.597196225264264e-05,
      "loss": 0.4773427963256836,
      "mean_token_accuracy": 0.8376712322235107,
      "num_tokens": 23838720.0,
      "step": 2910
    },
    {
      "entropy": 0.46411854922771456,
      "epoch": 1.5573333333333332,
      "grad_norm": 0.25727152824401855,
      "learning_rate": 9.540897214907105e-05,
      "loss": 0.48354244232177734,
      "mean_token_accuracy": 0.8340753436088562,
      "num_tokens": 23920640.0,
      "step": 2920
    },
    {
      "entropy": 0.45165770351886747,
      "epoch": 1.5626666666666666,
      "grad_norm": 0.2625143229961395,
      "learning_rate": 9.484612783249309e-05,
      "loss": 0.47493581771850585,
      "mean_token_accuracy": 0.8390044033527374,
      "num_tokens": 24002560.0,
      "step": 2930
    },
    {
      "entropy": 0.45827767848968504,
      "epoch": 1.568,
      "grad_norm": 0.3130404055118561,
      "learning_rate": 9.428344717589568e-05,
      "loss": 0.4779340744018555,
      "mean_token_accuracy": 0.8363992184400558,
      "num_tokens": 24084480.0,
      "step": 2940
    },
    {
      "entropy": 0.4503240793943405,
      "epoch": 1.5733333333333333,
      "grad_norm": 0.27369004487991333,
      "learning_rate": 9.372094804706867e-05,
      "loss": 0.47219367027282716,
      "mean_token_accuracy": 0.8394080221652984,
      "num_tokens": 24166400.0,
      "step": 2950
    },
    {
      "entropy": 0.4558084771037102,
      "epoch": 1.5786666666666667,
      "grad_norm": 0.26175567507743835,
      "learning_rate": 9.315864830803774e-05,
      "loss": 0.47450528144836424,
      "mean_token_accuracy": 0.8377201616764068,
      "num_tokens": 24248320.0,
      "step": 2960
    },
    {
      "entropy": 0.4532435715198517,
      "epoch": 1.584,
      "grad_norm": 0.2650909423828125,
      "learning_rate": 9.259656581449685e-05,
      "loss": 0.4723830699920654,
      "mean_token_accuracy": 0.839701560139656,
      "num_tokens": 24330240.0,
      "step": 2970
    },
    {
      "entropy": 0.46193648129701614,
      "epoch": 1.5893333333333333,
      "grad_norm": 0.2406294196844101,
      "learning_rate": 9.203471841524145e-05,
      "loss": 0.4809102535247803,
      "mean_token_accuracy": 0.8352739691734314,
      "num_tokens": 24412160.0,
      "step": 2980
    },
    {
      "entropy": 0.45854454636573794,
      "epoch": 1.5946666666666667,
      "grad_norm": 0.23232723772525787,
      "learning_rate": 9.147312395160157e-05,
      "loss": 0.47540898323059083,
      "mean_token_accuracy": 0.8373899191617966,
      "num_tokens": 24494080.0,
      "step": 2990
    },
    {
      "entropy": 0.4575690686702728,
      "epoch": 1.6,
      "grad_norm": 0.26282423734664917,
      "learning_rate": 9.091180025687534e-05,
      "loss": 0.47870402336120604,
      "mean_token_accuracy": 0.8371330678462983,
      "num_tokens": 24576000.0,
      "step": 3000
    },
    {
      "entropy": 0.4535189256072044,
      "epoch": 1.6053333333333333,
      "grad_norm": 0.3063070476055145,
      "learning_rate": 9.035076515576277e-05,
      "loss": 0.4781663417816162,
      "mean_token_accuracy": 0.8371575355529786,
      "num_tokens": 24657920.0,
      "step": 3010
    },
    {
      "entropy": 0.46885926723480226,
      "epoch": 1.6106666666666667,
      "grad_norm": 0.2803836166858673,
      "learning_rate": 8.979003646379953e-05,
      "loss": 0.49016809463500977,
      "mean_token_accuracy": 0.8332681000232697,
      "num_tokens": 24739840.0,
      "step": 3020
    },
    {
      "entropy": 0.46059383898973466,
      "epoch": 1.616,
      "grad_norm": 0.25507381558418274,
      "learning_rate": 8.922963198679137e-05,
      "loss": 0.4826096534729004,
      "mean_token_accuracy": 0.8356164455413818,
      "num_tokens": 24821760.0,
      "step": 3030
    },
    {
      "entropy": 0.45469827204942703,
      "epoch": 1.6213333333333333,
      "grad_norm": 0.23571361601352692,
      "learning_rate": 8.866956952024879e-05,
      "loss": 0.47269234657287595,
      "mean_token_accuracy": 0.8383928567171097,
      "num_tokens": 24903680.0,
      "step": 3040
    },
    {
      "entropy": 0.44722938984632493,
      "epoch": 1.6266666666666667,
      "grad_norm": 0.23081284761428833,
      "learning_rate": 8.81098668488217e-05,
      "loss": 0.4672435760498047,
      "mean_token_accuracy": 0.8406433433294296,
      "num_tokens": 24985600.0,
      "step": 3050
    },
    {
      "entropy": 0.45836658775806427,
      "epoch": 1.6320000000000001,
      "grad_norm": 0.3021607995033264,
      "learning_rate": 8.755054174573477e-05,
      "loss": 0.47626209259033203,
      "mean_token_accuracy": 0.8377568483352661,
      "num_tokens": 25067520.0,
      "step": 3060
    },
    {
      "entropy": 0.4558746457099915,
      "epoch": 1.6373333333333333,
      "grad_norm": 0.2360406368970871,
      "learning_rate": 8.699161197222323e-05,
      "loss": 0.47612819671630857,
      "mean_token_accuracy": 0.8378180056810379,
      "num_tokens": 25149440.0,
      "step": 3070
    },
    {
      "entropy": 0.4472507417201996,
      "epoch": 1.6426666666666667,
      "grad_norm": 0.25049057602882385,
      "learning_rate": 8.64330952769686e-05,
      "loss": 0.46475839614868164,
      "mean_token_accuracy": 0.8398483335971832,
      "num_tokens": 25231360.0,
      "step": 3080
    },
    {
      "entropy": 0.44390463680028913,
      "epoch": 1.6480000000000001,
      "grad_norm": 0.2740257978439331,
      "learning_rate": 8.587500939553521e-05,
      "loss": 0.46585235595703123,
      "mean_token_accuracy": 0.8406678080558777,
      "num_tokens": 25313280.0,
      "step": 3090
    },
    {
      "entropy": 0.4431780934333801,
      "epoch": 1.6533333333333333,
      "grad_norm": 0.27888405323028564,
      "learning_rate": 8.531737204980706e-05,
      "loss": 0.4624992847442627,
      "mean_token_accuracy": 0.8414261221885682,
      "num_tokens": 25395200.0,
      "step": 3100
    },
    {
      "entropy": 0.44821561872959137,
      "epoch": 1.6586666666666665,
      "grad_norm": 0.2523767650127411,
      "learning_rate": 8.476020094742491e-05,
      "loss": 0.47049641609191895,
      "mean_token_accuracy": 0.8386130124330521,
      "num_tokens": 25477120.0,
      "step": 3110
    },
    {
      "entropy": 0.45478425174951553,
      "epoch": 1.6640000000000001,
      "grad_norm": 0.2492104470729828,
      "learning_rate": 8.420351378122417e-05,
      "loss": 0.4721405506134033,
      "mean_token_accuracy": 0.8376834660768508,
      "num_tokens": 25559040.0,
      "step": 3120
    },
    {
      "entropy": 0.4579773709177971,
      "epoch": 1.6693333333333333,
      "grad_norm": 0.26135653257369995,
      "learning_rate": 8.364732822867291e-05,
      "loss": 0.48019938468933104,
      "mean_token_accuracy": 0.8367661476135254,
      "num_tokens": 25640960.0,
      "step": 3130
    },
    {
      "entropy": 0.45195942372083664,
      "epoch": 1.6746666666666665,
      "grad_norm": 0.23364049196243286,
      "learning_rate": 8.309166195131057e-05,
      "loss": 0.4704467296600342,
      "mean_token_accuracy": 0.8384417831897736,
      "num_tokens": 25722880.0,
      "step": 3140
    },
    {
      "entropy": 0.45350766032934187,
      "epoch": 1.6800000000000002,
      "grad_norm": 0.2811194956302643,
      "learning_rate": 8.253653259418724e-05,
      "loss": 0.4747741222381592,
      "mean_token_accuracy": 0.8393957853317261,
      "num_tokens": 25804800.0,
      "step": 3150
    },
    {
      "entropy": 0.4388273522257805,
      "epoch": 1.6853333333333333,
      "grad_norm": 0.2690880596637726,
      "learning_rate": 8.19819577853031e-05,
      "loss": 0.4582062244415283,
      "mean_token_accuracy": 0.8430772960186005,
      "num_tokens": 25886720.0,
      "step": 3160
    },
    {
      "entropy": 0.45715883523225787,
      "epoch": 1.6906666666666665,
      "grad_norm": 0.26028507947921753,
      "learning_rate": 8.142795513504882e-05,
      "loss": 0.4765753269195557,
      "mean_token_accuracy": 0.8370107591152192,
      "num_tokens": 25968640.0,
      "step": 3170
    },
    {
      "entropy": 0.4536979079246521,
      "epoch": 1.696,
      "grad_norm": 0.2740664780139923,
      "learning_rate": 8.087454223564637e-05,
      "loss": 0.47188363075256345,
      "mean_token_accuracy": 0.839408016204834,
      "num_tokens": 26050560.0,
      "step": 3180
    },
    {
      "entropy": 0.4482908099889755,
      "epoch": 1.7013333333333334,
      "grad_norm": 0.2595570981502533,
      "learning_rate": 8.032173666059027e-05,
      "loss": 0.4694857597351074,
      "mean_token_accuracy": 0.8407045125961303,
      "num_tokens": 26132480.0,
      "step": 3190
    },
    {
      "entropy": 0.45770729929208753,
      "epoch": 1.7066666666666666,
      "grad_norm": 0.2796715199947357,
      "learning_rate": 7.976955596408955e-05,
      "loss": 0.4778446197509766,
      "mean_token_accuracy": 0.8380748510360718,
      "num_tokens": 26214400.0,
      "step": 3200
    },
    {
      "entropy": 0.4419079482555389,
      "epoch": 1.712,
      "grad_norm": 0.24227827787399292,
      "learning_rate": 7.921801768051051e-05,
      "loss": 0.46060910224914553,
      "mean_token_accuracy": 0.8418786704540253,
      "num_tokens": 26296320.0,
      "step": 3210
    },
    {
      "entropy": 0.4490420952439308,
      "epoch": 1.7173333333333334,
      "grad_norm": 0.26219332218170166,
      "learning_rate": 7.866713932381968e-05,
      "loss": 0.4692113399505615,
      "mean_token_accuracy": 0.8401541084051132,
      "num_tokens": 26378240.0,
      "step": 3220
    },
    {
      "entropy": 0.4525995865464211,
      "epoch": 1.7226666666666666,
      "grad_norm": 0.2324022650718689,
      "learning_rate": 7.811693838702773e-05,
      "loss": 0.47092609405517577,
      "mean_token_accuracy": 0.8388943135738373,
      "num_tokens": 26460160.0,
      "step": 3230
    },
    {
      "entropy": 0.44824724942445754,
      "epoch": 1.728,
      "grad_norm": 0.24255606532096863,
      "learning_rate": 7.756743234163417e-05,
      "loss": 0.46755127906799315,
      "mean_token_accuracy": 0.8412181943655014,
      "num_tokens": 26542080.0,
      "step": 3240
    },
    {
      "entropy": 0.43668646961450575,
      "epoch": 1.7333333333333334,
      "grad_norm": 0.24743597209453583,
      "learning_rate": 7.701863863707228e-05,
      "loss": 0.45528078079223633,
      "mean_token_accuracy": 0.8442514657974243,
      "num_tokens": 26624000.0,
      "step": 3250
    },
    {
      "entropy": 0.4531865745782852,
      "epoch": 1.7386666666666666,
      "grad_norm": 0.2378353327512741,
      "learning_rate": 7.647057470015508e-05,
      "loss": 0.47452712059020996,
      "mean_token_accuracy": 0.8388209342956543,
      "num_tokens": 26705920.0,
      "step": 3260
    },
    {
      "entropy": 0.4488813027739525,
      "epoch": 1.744,
      "grad_norm": 0.2644421458244324,
      "learning_rate": 7.592325793452222e-05,
      "loss": 0.4670372009277344,
      "mean_token_accuracy": 0.8407778859138488,
      "num_tokens": 26787840.0,
      "step": 3270
    },
    {
      "entropy": 0.45321867763996126,
      "epoch": 1.7493333333333334,
      "grad_norm": 0.2989591956138611,
      "learning_rate": 7.537670572008684e-05,
      "loss": 0.4726859569549561,
      "mean_token_accuracy": 0.8389799416065216,
      "num_tokens": 26869760.0,
      "step": 3280
    },
    {
      "entropy": 0.45405653268098833,
      "epoch": 1.7546666666666666,
      "grad_norm": 0.2519940137863159,
      "learning_rate": 7.483093541248411e-05,
      "loss": 0.4742014408111572,
      "mean_token_accuracy": 0.8378180027008056,
      "num_tokens": 26951680.0,
      "step": 3290
    },
    {
      "entropy": 0.45349525213241576,
      "epoch": 1.76,
      "grad_norm": 0.23866058886051178,
      "learning_rate": 7.428596434251988e-05,
      "loss": 0.47319612503051756,
      "mean_token_accuracy": 0.8374021530151368,
      "num_tokens": 27033600.0,
      "step": 3300
    },
    {
      "entropy": 0.458782896399498,
      "epoch": 1.7653333333333334,
      "grad_norm": 0.2633265554904938,
      "learning_rate": 7.374180981562036e-05,
      "loss": 0.4797355175018311,
      "mean_token_accuracy": 0.8360567480325699,
      "num_tokens": 27115520.0,
      "step": 3310
    },
    {
      "entropy": 0.46364671885967257,
      "epoch": 1.7706666666666666,
      "grad_norm": 0.24553151428699493,
      "learning_rate": 7.319848911128272e-05,
      "loss": 0.48137893676757815,
      "mean_token_accuracy": 0.835225048661232,
      "num_tokens": 27197440.0,
      "step": 3320
    },
    {
      "entropy": 0.45163214951753616,
      "epoch": 1.776,
      "grad_norm": 0.2736968696117401,
      "learning_rate": 7.265601948252622e-05,
      "loss": 0.4711438179016113,
      "mean_token_accuracy": 0.8408512681722641,
      "num_tokens": 27279360.0,
      "step": 3330
    },
    {
      "entropy": 0.45708479434251786,
      "epoch": 1.7813333333333334,
      "grad_norm": 0.27659282088279724,
      "learning_rate": 7.211441815534438e-05,
      "loss": 0.4765739440917969,
      "mean_token_accuracy": 0.838380628824234,
      "num_tokens": 27361280.0,
      "step": 3340
    },
    {
      "entropy": 0.4524995177984238,
      "epoch": 1.7866666666666666,
      "grad_norm": 0.23820243775844574,
      "learning_rate": 7.157370232815822e-05,
      "loss": 0.47139592170715333,
      "mean_token_accuracy": 0.8389432430267334,
      "num_tokens": 27443200.0,
      "step": 3350
    },
    {
      "entropy": 0.4549313947558403,
      "epoch": 1.792,
      "grad_norm": 0.23141273856163025,
      "learning_rate": 7.103388917126965e-05,
      "loss": 0.4748246669769287,
      "mean_token_accuracy": 0.8387475579977035,
      "num_tokens": 27525120.0,
      "step": 3360
    },
    {
      "entropy": 0.4555491656064987,
      "epoch": 1.7973333333333334,
      "grad_norm": 0.2520790100097656,
      "learning_rate": 7.049499582631662e-05,
      "loss": 0.4766997337341309,
      "mean_token_accuracy": 0.8379892438650132,
      "num_tokens": 27607040.0,
      "step": 3370
    },
    {
      "entropy": 0.4520802959799767,
      "epoch": 1.8026666666666666,
      "grad_norm": 0.21699775755405426,
      "learning_rate": 6.995703940572876e-05,
      "loss": 0.47013354301452637,
      "mean_token_accuracy": 0.8393835633993149,
      "num_tokens": 27688960.0,
      "step": 3380
    },
    {
      "entropy": 0.4411054328083992,
      "epoch": 1.808,
      "grad_norm": 0.25057393312454224,
      "learning_rate": 6.942003699218374e-05,
      "loss": 0.4595449447631836,
      "mean_token_accuracy": 0.8437744587659836,
      "num_tokens": 27770880.0,
      "step": 3390
    },
    {
      "entropy": 0.46269181966781614,
      "epoch": 1.8133333333333335,
      "grad_norm": 0.25000807642936707,
      "learning_rate": 6.8884005638065e-05,
      "loss": 0.48174448013305665,
      "mean_token_accuracy": 0.8355919748544693,
      "num_tokens": 27852800.0,
      "step": 3400
    },
    {
      "entropy": 0.45039884448051454,
      "epoch": 1.8186666666666667,
      "grad_norm": 0.25356537103652954,
      "learning_rate": 6.834896236492033e-05,
      "loss": 0.46697049140930175,
      "mean_token_accuracy": 0.8410347402095795,
      "num_tokens": 27934720.0,
      "step": 3410
    },
    {
      "entropy": 0.45715008974075316,
      "epoch": 1.8239999999999998,
      "grad_norm": 0.21722032129764557,
      "learning_rate": 6.781492416292113e-05,
      "loss": 0.4742414951324463,
      "mean_token_accuracy": 0.8370963782072067,
      "num_tokens": 28016640.0,
      "step": 3420
    },
    {
      "entropy": 0.444775764644146,
      "epoch": 1.8293333333333335,
      "grad_norm": 0.28689178824424744,
      "learning_rate": 6.728190799032297e-05,
      "loss": 0.4651144027709961,
      "mean_token_accuracy": 0.8404476463794708,
      "num_tokens": 28098560.0,
      "step": 3430
    },
    {
      "entropy": 0.4539962843060493,
      "epoch": 1.8346666666666667,
      "grad_norm": 0.26667651534080505,
      "learning_rate": 6.674993077292729e-05,
      "loss": 0.4711462497711182,
      "mean_token_accuracy": 0.8391511708498001,
      "num_tokens": 28180480.0,
      "step": 3440
    },
    {
      "entropy": 0.4594313636422157,
      "epoch": 1.8399999999999999,
      "grad_norm": 0.2558381259441376,
      "learning_rate": 6.621900940354355e-05,
      "loss": 0.47672715187072756,
      "mean_token_accuracy": 0.8360200583934784,
      "num_tokens": 28262400.0,
      "step": 3450
    },
    {
      "entropy": 0.4491269409656525,
      "epoch": 1.8453333333333335,
      "grad_norm": 0.25986447930336,
      "learning_rate": 6.568916074145324e-05,
      "loss": 0.4719265937805176,
      "mean_token_accuracy": 0.8384295463562011,
      "num_tokens": 28344320.0,
      "step": 3460
    },
    {
      "entropy": 0.4569529816508293,
      "epoch": 1.8506666666666667,
      "grad_norm": 0.22530515491962433,
      "learning_rate": 6.516040161187407e-05,
      "loss": 0.47388343811035155,
      "mean_token_accuracy": 0.8380381613969803,
      "num_tokens": 28426240.0,
      "step": 3470
    },
    {
      "entropy": 0.4521230891346931,
      "epoch": 1.8559999999999999,
      "grad_norm": 0.2698580026626587,
      "learning_rate": 6.463274880542595e-05,
      "loss": 0.46836180686950685,
      "mean_token_accuracy": 0.8396281808614731,
      "num_tokens": 28508160.0,
      "step": 3480
    },
    {
      "entropy": 0.4573205426335335,
      "epoch": 1.8613333333333333,
      "grad_norm": 0.2569490373134613,
      "learning_rate": 6.410621907759791e-05,
      "loss": 0.47612714767456055,
      "mean_token_accuracy": 0.8375856161117554,
      "num_tokens": 28590080.0,
      "step": 3490
    },
    {
      "entropy": 0.4520913749933243,
      "epoch": 1.8666666666666667,
      "grad_norm": 0.2499508410692215,
      "learning_rate": 6.358082914821574e-05,
      "loss": 0.47176947593688967,
      "mean_token_accuracy": 0.8387964755296707,
      "num_tokens": 28672000.0,
      "step": 3500
    },
    {
      "entropy": 0.45928225517272947,
      "epoch": 1.8719999999999999,
      "grad_norm": 0.2763667702674866,
      "learning_rate": 6.305659570091122e-05,
      "loss": 0.4785407066345215,
      "mean_token_accuracy": 0.8375733911991119,
      "num_tokens": 28753920.0,
      "step": 3510
    },
    {
      "entropy": 0.4611460089683533,
      "epoch": 1.8773333333333333,
      "grad_norm": 0.23890498280525208,
      "learning_rate": 6.253353538259236e-05,
      "loss": 0.47818994522094727,
      "mean_token_accuracy": 0.8357754439115525,
      "num_tokens": 28835840.0,
      "step": 3520
    },
    {
      "entropy": 0.44563424587249756,
      "epoch": 1.8826666666666667,
      "grad_norm": 0.2487446367740631,
      "learning_rate": 6.201166480291476e-05,
      "loss": 0.46471896171569826,
      "mean_token_accuracy": 0.8413405090570449,
      "num_tokens": 28917760.0,
      "step": 3530
    },
    {
      "entropy": 0.4476030975580215,
      "epoch": 1.888,
      "grad_norm": 0.26070529222488403,
      "learning_rate": 6.149100053375404e-05,
      "loss": 0.4690867900848389,
      "mean_token_accuracy": 0.8408145785331727,
      "num_tokens": 28999680.0,
      "step": 3540
    },
    {
      "entropy": 0.4452753260731697,
      "epoch": 1.8933333333333333,
      "grad_norm": 0.2441214919090271,
      "learning_rate": 6.0971559108679855e-05,
      "loss": 0.461823558807373,
      "mean_token_accuracy": 0.8412059664726257,
      "num_tokens": 29081600.0,
      "step": 3550
    },
    {
      "entropy": 0.45318825989961625,
      "epoch": 1.8986666666666667,
      "grad_norm": 0.26395824551582336,
      "learning_rate": 6.045335702243065e-05,
      "loss": 0.4709752082824707,
      "mean_token_accuracy": 0.8400929480791092,
      "num_tokens": 29163520.0,
      "step": 3560
    },
    {
      "entropy": 0.45433855801820755,
      "epoch": 1.904,
      "grad_norm": 0.24897560477256775,
      "learning_rate": 5.993641073038995e-05,
      "loss": 0.4730278968811035,
      "mean_token_accuracy": 0.838857638835907,
      "num_tokens": 29245440.0,
      "step": 3570
    },
    {
      "entropy": 0.45538086295127866,
      "epoch": 1.9093333333333333,
      "grad_norm": 0.2463449090719223,
      "learning_rate": 5.9420736648063956e-05,
      "loss": 0.471648645401001,
      "mean_token_accuracy": 0.8385640919208527,
      "num_tokens": 29327360.0,
      "step": 3580
    },
    {
      "entropy": 0.45286301225423814,
      "epoch": 1.9146666666666667,
      "grad_norm": 0.27083930373191833,
      "learning_rate": 5.890635115056001e-05,
      "loss": 0.4696188926696777,
      "mean_token_accuracy": 0.8394447207450867,
      "num_tokens": 29409280.0,
      "step": 3590
    },
    {
      "entropy": 0.44802253842353823,
      "epoch": 1.92,
      "grad_norm": 0.24196985363960266,
      "learning_rate": 5.839327057206678e-05,
      "loss": 0.46616101264953613,
      "mean_token_accuracy": 0.8407411962747574,
      "num_tokens": 29491200.0,
      "step": 3600
    },
    {
      "entropy": 0.4546609878540039,
      "epoch": 1.9253333333333333,
      "grad_norm": 0.2525026500225067,
      "learning_rate": 5.788151120533567e-05,
      "loss": 0.4738184452056885,
      "mean_token_accuracy": 0.8373654633760452,
      "num_tokens": 29573120.0,
      "step": 3610
    },
    {
      "entropy": 0.4548776403069496,
      "epoch": 1.9306666666666668,
      "grad_norm": 0.2817035913467407,
      "learning_rate": 5.7371089301163086e-05,
      "loss": 0.47237591743469237,
      "mean_token_accuracy": 0.8389187842607498,
      "num_tokens": 29655040.0,
      "step": 3620
    },
    {
      "entropy": 0.4530694574117661,
      "epoch": 1.936,
      "grad_norm": 0.2742132842540741,
      "learning_rate": 5.6862021067874915e-05,
      "loss": 0.47298660278320315,
      "mean_token_accuracy": 0.8375244617462159,
      "num_tokens": 29736960.0,
      "step": 3630
    },
    {
      "entropy": 0.4617029786109924,
      "epoch": 1.9413333333333334,
      "grad_norm": 0.2624448835849762,
      "learning_rate": 5.635432267081133e-05,
      "loss": 0.47890653610229494,
      "mean_token_accuracy": 0.8358855187892914,
      "num_tokens": 29818880.0,
      "step": 3640
    },
    {
      "entropy": 0.44912499338388445,
      "epoch": 1.9466666666666668,
      "grad_norm": 0.2520037293434143,
      "learning_rate": 5.584801023181365e-05,
      "loss": 0.4675039291381836,
      "mean_token_accuracy": 0.8395792573690415,
      "num_tokens": 29900800.0,
      "step": 3650
    },
    {
      "entropy": 0.44727917909622195,
      "epoch": 1.952,
      "grad_norm": 0.25161853432655334,
      "learning_rate": 5.534309982871271e-05,
      "loss": 0.4663228988647461,
      "mean_token_accuracy": 0.8409001976251602,
      "num_tokens": 29982720.0,
      "step": 3660
    },
    {
      "entropy": 0.45230296552181243,
      "epoch": 1.9573333333333334,
      "grad_norm": 0.3486286997795105,
      "learning_rate": 5.483960749481771e-05,
      "loss": 0.4689676761627197,
      "mean_token_accuracy": 0.8387353181838989,
      "num_tokens": 30064640.0,
      "step": 3670
    },
    {
      "entropy": 0.4443836480379105,
      "epoch": 1.9626666666666668,
      "grad_norm": 0.22573833167552948,
      "learning_rate": 5.433754921840759e-05,
      "loss": 0.46163601875305177,
      "mean_token_accuracy": 0.8423679083585739,
      "num_tokens": 30146560.0,
      "step": 3680
    },
    {
      "entropy": 0.4542616128921509,
      "epoch": 1.968,
      "grad_norm": 0.24735628068447113,
      "learning_rate": 5.38369409422231e-05,
      "loss": 0.47411108016967773,
      "mean_token_accuracy": 0.8380870819091797,
      "num_tokens": 30228480.0,
      "step": 3690
    },
    {
      "entropy": 0.4505576819181442,
      "epoch": 1.9733333333333334,
      "grad_norm": 0.28003963828086853,
      "learning_rate": 5.33377985629605e-05,
      "loss": 0.4678519248962402,
      "mean_token_accuracy": 0.8398850351572037,
      "num_tokens": 30310400.0,
      "step": 3700
    },
    {
      "entropy": 0.4529769614338875,
      "epoch": 1.9786666666666668,
      "grad_norm": 0.2528955638408661,
      "learning_rate": 5.284013793076694e-05,
      "loss": 0.4710858345031738,
      "mean_token_accuracy": 0.8388943254947663,
      "num_tokens": 30392320.0,
      "step": 3710
    },
    {
      "entropy": 0.46223973035812377,
      "epoch": 1.984,
      "grad_norm": 0.24886520206928253,
      "learning_rate": 5.2343974848737064e-05,
      "loss": 0.480480432510376,
      "mean_token_accuracy": 0.8363258361816406,
      "num_tokens": 30474240.0,
      "step": 3720
    },
    {
      "entropy": 0.4638181492686272,
      "epoch": 1.9893333333333332,
      "grad_norm": 0.24633975327014923,
      "learning_rate": 5.1849325072411024e-05,
      "loss": 0.48080787658691404,
      "mean_token_accuracy": 0.8352495074272156,
      "num_tokens": 30556160.0,
      "step": 3730
    },
    {
      "entropy": 0.452928726375103,
      "epoch": 1.9946666666666668,
      "grad_norm": 0.23061209917068481,
      "learning_rate": 5.1356204309274435e-05,
      "loss": 0.47095208168029784,
      "mean_token_accuracy": 0.8385396242141724,
      "num_tokens": 30638080.0,
      "step": 3740
    },
    {
      "entropy": 0.46319368183612825,
      "epoch": 2.0,
      "grad_norm": 0.2585753798484802,
      "learning_rate": 5.086462821825948e-05,
      "loss": 0.4810307502746582,
      "mean_token_accuracy": 0.8346991151571274,
      "num_tokens": 30720000.0,
      "step": 3750
    },
    {
      "entropy": 0.44506675750017166,
      "epoch": 2.005333333333333,
      "grad_norm": 0.22186453640460968,
      "learning_rate": 5.03746124092475e-05,
      "loss": 0.4607857227325439,
      "mean_token_accuracy": 0.8417074292898178,
      "num_tokens": 30801920.0,
      "step": 3760
    },
    {
      "entropy": 0.44696380794048307,
      "epoch": 2.010666666666667,
      "grad_norm": 0.23102766275405884,
      "learning_rate": 4.988617244257362e-05,
      "loss": 0.4633793830871582,
      "mean_token_accuracy": 0.8409368902444839,
      "num_tokens": 30883840.0,
      "step": 3770
    },
    {
      "entropy": 0.4573482796549797,
      "epoch": 2.016,
      "grad_norm": 0.2631347179412842,
      "learning_rate": 4.9399323828532375e-05,
      "loss": 0.47553558349609376,
      "mean_token_accuracy": 0.8380259335041046,
      "num_tokens": 30965760.0,
      "step": 3780
    },
    {
      "entropy": 0.45846093595027926,
      "epoch": 2.021333333333333,
      "grad_norm": 0.24107612669467926,
      "learning_rate": 4.891408202688531e-05,
      "loss": 0.47466244697570803,
      "mean_token_accuracy": 0.8376956939697265,
      "num_tokens": 31047680.0,
      "step": 3790
    },
    {
      "entropy": 0.45362787544727323,
      "epoch": 2.026666666666667,
      "grad_norm": 0.2581658959388733,
      "learning_rate": 4.843046244637005e-05,
      "loss": 0.4721574306488037,
      "mean_token_accuracy": 0.8384784817695617,
      "num_tokens": 31129600.0,
      "step": 3800
    },
    {
      "entropy": 0.45568518191576,
      "epoch": 2.032,
      "grad_norm": 0.24644684791564941,
      "learning_rate": 4.7948480444210866e-05,
      "loss": 0.47399563789367677,
      "mean_token_accuracy": 0.8376834660768508,
      "num_tokens": 31211520.0,
      "step": 3810
    },
    {
      "entropy": 0.44302789866924286,
      "epoch": 2.037333333333333,
      "grad_norm": 0.2578761875629425,
      "learning_rate": 4.7468151325631225e-05,
      "loss": 0.4603163242340088,
      "mean_token_accuracy": 0.8425513684749604,
      "num_tokens": 31293440.0,
      "step": 3820
    },
    {
      "entropy": 0.44779601246118544,
      "epoch": 2.042666666666667,
      "grad_norm": 0.23922692239284515,
      "learning_rate": 4.69894903433677e-05,
      "loss": 0.46392064094543456,
      "mean_token_accuracy": 0.841475048661232,
      "num_tokens": 31375360.0,
      "step": 3830
    },
    {
      "entropy": 0.43781903088092805,
      "epoch": 2.048,
      "grad_norm": 0.2260008156299591,
      "learning_rate": 4.651251269718544e-05,
      "loss": 0.45696425437927246,
      "mean_token_accuracy": 0.8434442281723022,
      "num_tokens": 31457280.0,
      "step": 3840
    },
    {
      "entropy": 0.45533082336187364,
      "epoch": 2.0533333333333332,
      "grad_norm": 0.2638528645038605,
      "learning_rate": 4.603723353339579e-05,
      "loss": 0.4712835788726807,
      "mean_token_accuracy": 0.837976998090744,
      "num_tokens": 31539200.0,
      "step": 3850
    },
    {
      "entropy": 0.4518005207180977,
      "epoch": 2.058666666666667,
      "grad_norm": 0.23908574879169464,
      "learning_rate": 4.5563667944375155e-05,
      "loss": 0.47028241157531736,
      "mean_token_accuracy": 0.838246089220047,
      "num_tokens": 31621120.0,
      "step": 3860
    },
    {
      "entropy": 0.4511943057179451,
      "epoch": 2.064,
      "grad_norm": 0.2564878463745117,
      "learning_rate": 4.5091830968085824e-05,
      "loss": 0.46828432083129884,
      "mean_token_accuracy": 0.8395303398370743,
      "num_tokens": 31703040.0,
      "step": 3870
    },
    {
      "entropy": 0.4469565153121948,
      "epoch": 2.0693333333333332,
      "grad_norm": 0.23527298867702484,
      "learning_rate": 4.4621737587598276e-05,
      "loss": 0.4634071350097656,
      "mean_token_accuracy": 0.8416829705238342,
      "num_tokens": 31784960.0,
      "step": 3880
    },
    {
      "entropy": 0.45741556882858275,
      "epoch": 2.074666666666667,
      "grad_norm": 0.25746968388557434,
      "learning_rate": 4.415340273061564e-05,
      "loss": 0.47378177642822267,
      "mean_token_accuracy": 0.8365092903375626,
      "num_tokens": 31866880.0,
      "step": 3890
    },
    {
      "entropy": 0.45066776126623154,
      "epoch": 2.08,
      "grad_norm": 0.24657104909420013,
      "learning_rate": 4.368684126899957e-05,
      "loss": 0.46764178276062013,
      "mean_token_accuracy": 0.8400807201862335,
      "num_tokens": 31948800.0,
      "step": 3900
    },
    {
      "entropy": 0.4564896494150162,
      "epoch": 2.0853333333333333,
      "grad_norm": 0.2448369711637497,
      "learning_rate": 4.322206801829782e-05,
      "loss": 0.4762216091156006,
      "mean_token_accuracy": 0.8366193681955337,
      "num_tokens": 32030720.0,
      "step": 3910
    },
    {
      "entropy": 0.44439963549375533,
      "epoch": 2.0906666666666665,
      "grad_norm": 0.2505202889442444,
      "learning_rate": 4.275909773727409e-05,
      "loss": 0.4604054927825928,
      "mean_token_accuracy": 0.8418786644935607,
      "num_tokens": 32112640.0,
      "step": 3920
    },
    {
      "entropy": 0.4499572977423668,
      "epoch": 2.096,
      "grad_norm": 0.2527288496494293,
      "learning_rate": 4.229794512743917e-05,
      "loss": 0.46473493576049807,
      "mean_token_accuracy": 0.8408390402793884,
      "num_tokens": 32194560.0,
      "step": 3930
    },
    {
      "entropy": 0.45977194756269457,
      "epoch": 2.1013333333333333,
      "grad_norm": 0.27111220359802246,
      "learning_rate": 4.183862483258405e-05,
      "loss": 0.47615480422973633,
      "mean_token_accuracy": 0.8361912965774536,
      "num_tokens": 32276480.0,
      "step": 3940
    },
    {
      "entropy": 0.4550243988633156,
      "epoch": 2.1066666666666665,
      "grad_norm": 0.24795487523078918,
      "learning_rate": 4.138115143831509e-05,
      "loss": 0.46946077346801757,
      "mean_token_accuracy": 0.8385273963212967,
      "num_tokens": 32358400.0,
      "step": 3950
    },
    {
      "entropy": 0.443332539498806,
      "epoch": 2.112,
      "grad_norm": 0.2566700875759125,
      "learning_rate": 4.092553947159074e-05,
      "loss": 0.46065611839294435,
      "mean_token_accuracy": 0.8428938269615174,
      "num_tokens": 32440320.0,
      "step": 3960
    },
    {
      "entropy": 0.44889726489782333,
      "epoch": 2.1173333333333333,
      "grad_norm": 0.24460414052009583,
      "learning_rate": 4.047180340026026e-05,
      "loss": 0.4646409511566162,
      "mean_token_accuracy": 0.8403865069150924,
      "num_tokens": 32522240.0,
      "step": 3970
    },
    {
      "entropy": 0.46501262187957765,
      "epoch": 2.1226666666666665,
      "grad_norm": 0.25299856066703796,
      "learning_rate": 4.001995763260431e-05,
      "loss": 0.47845892906188964,
      "mean_token_accuracy": 0.8348458856344223,
      "num_tokens": 32604160.0,
      "step": 3980
    },
    {
      "entropy": 0.4521528258919716,
      "epoch": 2.128,
      "grad_norm": 0.27291417121887207,
      "learning_rate": 3.957001651687734e-05,
      "loss": 0.470092248916626,
      "mean_token_accuracy": 0.8382583111524582,
      "num_tokens": 32686080.0,
      "step": 3990
    },
    {
      "entropy": 0.4502758398652077,
      "epoch": 2.1333333333333333,
      "grad_norm": 0.2502305805683136,
      "learning_rate": 3.912199434085212e-05,
      "loss": 0.4687769412994385,
      "mean_token_accuracy": 0.8403131097555161,
      "num_tokens": 32768000.0,
      "step": 4000
    },
    {
      "entropy": 0.44980519264936447,
      "epoch": 2.1386666666666665,
      "grad_norm": 0.2605154514312744,
      "learning_rate": 3.867590533136594e-05,
      "loss": 0.4647549629211426,
      "mean_token_accuracy": 0.841413888335228,
      "num_tokens": 32849920.0,
      "step": 4010
    },
    {
      "entropy": 0.4395098477602005,
      "epoch": 2.144,
      "grad_norm": 0.27497583627700806,
      "learning_rate": 3.823176365386876e-05,
      "loss": 0.4572429656982422,
      "mean_token_accuracy": 0.8426859080791473,
      "num_tokens": 32931840.0,
      "step": 4020
    },
    {
      "entropy": 0.45462971478700637,
      "epoch": 2.1493333333333333,
      "grad_norm": 0.24453990161418915,
      "learning_rate": 3.7789583411973564e-05,
      "loss": 0.47185516357421875,
      "mean_token_accuracy": 0.8383561611175537,
      "num_tokens": 33013760.0,
      "step": 4030
    },
    {
      "entropy": 0.46014717817306516,
      "epoch": 2.1546666666666665,
      "grad_norm": 0.2591027021408081,
      "learning_rate": 3.7349378647008426e-05,
      "loss": 0.47922072410583494,
      "mean_token_accuracy": 0.8350293606519699,
      "num_tokens": 33095680.0,
      "step": 4040
    },
    {
      "entropy": 0.4442116320133209,
      "epoch": 2.16,
      "grad_norm": 0.2650836706161499,
      "learning_rate": 3.691116333757052e-05,
      "loss": 0.4594129085540771,
      "mean_token_accuracy": 0.8436888515949249,
      "num_tokens": 33177600.0,
      "step": 4050
    },
    {
      "entropy": 0.44998047798871993,
      "epoch": 2.1653333333333333,
      "grad_norm": 0.23278743028640747,
      "learning_rate": 3.6474951399082444e-05,
      "loss": 0.46572365760803225,
      "mean_token_accuracy": 0.8395303308963775,
      "num_tokens": 33259520.0,
      "step": 4060
    },
    {
      "entropy": 0.44763067811727525,
      "epoch": 2.1706666666666665,
      "grad_norm": 0.2440081089735031,
      "learning_rate": 3.6040756683350215e-05,
      "loss": 0.46425209045410154,
      "mean_token_accuracy": 0.8403620302677155,
      "num_tokens": 33341440.0,
      "step": 4070
    },
    {
      "entropy": 0.4510946825146675,
      "epoch": 2.176,
      "grad_norm": 0.2567829191684723,
      "learning_rate": 3.56085929781233e-05,
      "loss": 0.4698475360870361,
      "mean_token_accuracy": 0.8388454049825669,
      "num_tokens": 33423360.0,
      "step": 4080
    },
    {
      "entropy": 0.45450502038002016,
      "epoch": 2.1813333333333333,
      "grad_norm": 0.23825280368328094,
      "learning_rate": 3.5178474006657134e-05,
      "loss": 0.46937408447265627,
      "mean_token_accuracy": 0.8378913938999176,
      "num_tokens": 33505280.0,
      "step": 4090
    },
    {
      "entropy": 0.45174714624881745,
      "epoch": 2.1866666666666665,
      "grad_norm": 0.2661704123020172,
      "learning_rate": 3.475041342727693e-05,
      "loss": 0.46913671493530273,
      "mean_token_accuracy": 0.8385640919208527,
      "num_tokens": 33587200.0,
      "step": 4100
    },
    {
      "entropy": 0.4488513618707657,
      "epoch": 2.192,
      "grad_norm": 0.2576528489589691,
      "learning_rate": 3.4324424832944145e-05,
      "loss": 0.4654860496520996,
      "mean_token_accuracy": 0.8401663362979889,
      "num_tokens": 33669120.0,
      "step": 4110
    },
    {
      "entropy": 0.450703439116478,
      "epoch": 2.1973333333333334,
      "grad_norm": 0.2588220536708832,
      "learning_rate": 3.3900521750825044e-05,
      "loss": 0.46711506843566897,
      "mean_token_accuracy": 0.8394202500581741,
      "num_tokens": 33751040.0,
      "step": 4120
    },
    {
      "entropy": 0.45834354907274244,
      "epoch": 2.2026666666666666,
      "grad_norm": 0.242472305893898,
      "learning_rate": 3.3478717641860725e-05,
      "loss": 0.47342720031738283,
      "mean_token_accuracy": 0.8371453016996384,
      "num_tokens": 33832960.0,
      "step": 4130
    },
    {
      "entropy": 0.44646052867174146,
      "epoch": 2.208,
      "grad_norm": 0.24574773013591766,
      "learning_rate": 3.305902590034002e-05,
      "loss": 0.4635519504547119,
      "mean_token_accuracy": 0.8407167315483093,
      "num_tokens": 33914880.0,
      "step": 4140
    },
    {
      "entropy": 0.45711341947317125,
      "epoch": 2.2133333333333334,
      "grad_norm": 0.2685025632381439,
      "learning_rate": 3.2641459853474054e-05,
      "loss": 0.4719253063201904,
      "mean_token_accuracy": 0.8380381613969803,
      "num_tokens": 33996800.0,
      "step": 4150
    },
    {
      "entropy": 0.44123495519161227,
      "epoch": 2.2186666666666666,
      "grad_norm": 0.23826037347316742,
      "learning_rate": 3.222603276097289e-05,
      "loss": 0.4564843654632568,
      "mean_token_accuracy": 0.8426492214202881,
      "num_tokens": 34078720.0,
      "step": 4160
    },
    {
      "entropy": 0.4482108950614929,
      "epoch": 2.224,
      "grad_norm": 0.2652825117111206,
      "learning_rate": 3.181275781462475e-05,
      "loss": 0.4643686294555664,
      "mean_token_accuracy": 0.84058218896389,
      "num_tokens": 34160640.0,
      "step": 4170
    },
    {
      "entropy": 0.4622217774391174,
      "epoch": 2.2293333333333334,
      "grad_norm": 0.28008148074150085,
      "learning_rate": 3.1401648137876925e-05,
      "loss": 0.4789405345916748,
      "mean_token_accuracy": 0.8359833627939224,
      "num_tokens": 34242560.0,
      "step": 4180
    },
    {
      "entropy": 0.45667475163936616,
      "epoch": 2.2346666666666666,
      "grad_norm": 0.274131178855896,
      "learning_rate": 3.099271678541899e-05,
      "loss": 0.47229709625244143,
      "mean_token_accuracy": 0.8380381673574447,
      "num_tokens": 34324480.0,
      "step": 4190
    },
    {
      "entropy": 0.4538781255483627,
      "epoch": 2.24,
      "grad_norm": 0.26839420199394226,
      "learning_rate": 3.0585976742768554e-05,
      "loss": 0.46877193450927734,
      "mean_token_accuracy": 0.8392490148544312,
      "num_tokens": 34406400.0,
      "step": 4200
    },
    {
      "entropy": 0.45335969179868696,
      "epoch": 2.2453333333333334,
      "grad_norm": 0.25544652342796326,
      "learning_rate": 3.018144092585853e-05,
      "loss": 0.4683972358703613,
      "mean_token_accuracy": 0.837866923213005,
      "num_tokens": 34488320.0,
      "step": 4210
    },
    {
      "entropy": 0.4529318079352379,
      "epoch": 2.2506666666666666,
      "grad_norm": 0.24646338820457458,
      "learning_rate": 2.977912218062716e-05,
      "loss": 0.4701500415802002,
      "mean_token_accuracy": 0.8389799445867538,
      "num_tokens": 34570240.0,
      "step": 4220
    },
    {
      "entropy": 0.45443865209817885,
      "epoch": 2.2560000000000002,
      "grad_norm": 0.25814223289489746,
      "learning_rate": 2.9379033282610315e-05,
      "loss": 0.4709921836853027,
      "mean_token_accuracy": 0.8378302365541458,
      "num_tokens": 34652160.0,
      "step": 4230
    },
    {
      "entropy": 0.4512682601809502,
      "epoch": 2.2613333333333334,
      "grad_norm": 0.26942750811576843,
      "learning_rate": 2.8981186936535332e-05,
      "loss": 0.46540265083312987,
      "mean_token_accuracy": 0.8396159470081329,
      "num_tokens": 34734080.0,
      "step": 4240
    },
    {
      "entropy": 0.4528087452054024,
      "epoch": 2.2666666666666666,
      "grad_norm": 0.27728861570358276,
      "learning_rate": 2.858559577591803e-05,
      "loss": 0.46912856101989747,
      "mean_token_accuracy": 0.8397871822118759,
      "num_tokens": 34816000.0,
      "step": 4250
    },
    {
      "entropy": 0.45357600450515745,
      "epoch": 2.2720000000000002,
      "grad_norm": 0.27805355191230774,
      "learning_rate": 2.8192272362661308e-05,
      "loss": 0.46867995262145995,
      "mean_token_accuracy": 0.8380626201629638,
      "num_tokens": 34897920.0,
      "step": 4260
    },
    {
      "entropy": 0.4450182169675827,
      "epoch": 2.2773333333333334,
      "grad_norm": 0.3053533732891083,
      "learning_rate": 2.780122918665622e-05,
      "loss": 0.4611020565032959,
      "mean_token_accuracy": 0.8422578305006028,
      "num_tokens": 34979840.0,
      "step": 4270
    },
    {
      "entropy": 0.4475100502371788,
      "epoch": 2.2826666666666666,
      "grad_norm": 0.24690933525562286,
      "learning_rate": 2.74124786653855e-05,
      "loss": 0.463855504989624,
      "mean_token_accuracy": 0.8394447177648544,
      "num_tokens": 35061760.0,
      "step": 4280
    },
    {
      "entropy": 0.4467692539095879,
      "epoch": 2.288,
      "grad_norm": 0.24188663065433502,
      "learning_rate": 2.7026033143529206e-05,
      "loss": 0.46448593139648436,
      "mean_token_accuracy": 0.8401785671710968,
      "num_tokens": 35143680.0,
      "step": 4290
    },
    {
      "entropy": 0.4438614472746849,
      "epoch": 2.2933333333333334,
      "grad_norm": 0.28143206238746643,
      "learning_rate": 2.6641904892572556e-05,
      "loss": 0.4591333866119385,
      "mean_token_accuracy": 0.842502447962761,
      "num_tokens": 35225600.0,
      "step": 4300
    },
    {
      "entropy": 0.45180450528860094,
      "epoch": 2.2986666666666666,
      "grad_norm": 0.2468089759349823,
      "learning_rate": 2.626010611041658e-05,
      "loss": 0.4694178104400635,
      "mean_token_accuracy": 0.8382460862398148,
      "num_tokens": 35307520.0,
      "step": 4310
    },
    {
      "entropy": 0.4481292858719826,
      "epoch": 2.304,
      "grad_norm": 0.25983795523643494,
      "learning_rate": 2.58806489209904e-05,
      "loss": 0.46465325355529785,
      "mean_token_accuracy": 0.8413038194179535,
      "num_tokens": 35389440.0,
      "step": 4320
    },
    {
      "entropy": 0.455538783967495,
      "epoch": 2.3093333333333335,
      "grad_norm": 0.2795232832431793,
      "learning_rate": 2.5503545373866523e-05,
      "loss": 0.47109217643737794,
      "mean_token_accuracy": 0.8379525452852249,
      "num_tokens": 35471360.0,
      "step": 4330
    },
    {
      "entropy": 0.4505577817559242,
      "epoch": 2.3146666666666667,
      "grad_norm": 0.26285114884376526,
      "learning_rate": 2.51288074438781e-05,
      "loss": 0.4664009094238281,
      "mean_token_accuracy": 0.8397260248661041,
      "num_tokens": 35553280.0,
      "step": 4340
    },
    {
      "entropy": 0.44685496389865875,
      "epoch": 2.32,
      "grad_norm": 0.24313150346279144,
      "learning_rate": 2.4756447030738593e-05,
      "loss": 0.4611032009124756,
      "mean_token_accuracy": 0.8404965728521347,
      "num_tokens": 35635200.0,
      "step": 4350
    },
    {
      "entropy": 0.454390786588192,
      "epoch": 2.3253333333333335,
      "grad_norm": 0.27514687180519104,
      "learning_rate": 2.438647595866409e-05,
      "loss": 0.4706882953643799,
      "mean_token_accuracy": 0.8387964814901352,
      "num_tokens": 35717120.0,
      "step": 4360
    },
    {
      "entropy": 0.4499600291252136,
      "epoch": 2.3306666666666667,
      "grad_norm": 0.26181483268737793,
      "learning_rate": 2.4018905975997686e-05,
      "loss": 0.46518454551696775,
      "mean_token_accuracy": 0.8404721111059189,
      "num_tokens": 35799040.0,
      "step": 4370
    },
    {
      "entropy": 0.45308135002851485,
      "epoch": 2.336,
      "grad_norm": 0.29293519258499146,
      "learning_rate": 2.36537487548364e-05,
      "loss": 0.46730494499206543,
      "mean_token_accuracy": 0.8390655547380448,
      "num_tokens": 35880960.0,
      "step": 4380
    },
    {
      "entropy": 0.4502210706472397,
      "epoch": 2.3413333333333335,
      "grad_norm": 0.26994070410728455,
      "learning_rate": 2.3291015890660672e-05,
      "loss": 0.46766109466552735,
      "mean_token_accuracy": 0.8376467764377594,
      "num_tokens": 35962880.0,
      "step": 4390
    },
    {
      "entropy": 0.44541915953159333,
      "epoch": 2.3466666666666667,
      "grad_norm": 0.26039642095565796,
      "learning_rate": 2.2930718901966042e-05,
      "loss": 0.4604947090148926,
      "mean_token_accuracy": 0.8416585206985474,
      "num_tokens": 36044800.0,
      "step": 4400
    },
    {
      "entropy": 0.4566008448600769,
      "epoch": 2.352,
      "grad_norm": 0.25916898250579834,
      "learning_rate": 2.257286922989732e-05,
      "loss": 0.47378830909729003,
      "mean_token_accuracy": 0.8367906153202057,
      "num_tokens": 36126720.0,
      "step": 4410
    },
    {
      "entropy": 0.44468457251787186,
      "epoch": 2.3573333333333335,
      "grad_norm": 0.25393351912498474,
      "learning_rate": 2.2217478237885468e-05,
      "loss": 0.46176724433898925,
      "mean_token_accuracy": 0.8413894355297089,
      "num_tokens": 36208640.0,
      "step": 4420
    },
    {
      "entropy": 0.4476290255784988,
      "epoch": 2.3626666666666667,
      "grad_norm": 0.2615390121936798,
      "learning_rate": 2.186455721128663e-05,
      "loss": 0.4604832649230957,
      "mean_token_accuracy": 0.8401296466588974,
      "num_tokens": 36290560.0,
      "step": 4430
    },
    {
      "entropy": 0.45934602469205854,
      "epoch": 2.368,
      "grad_norm": 0.2526385486125946,
      "learning_rate": 2.1514117357023755e-05,
      "loss": 0.47376585006713867,
      "mean_token_accuracy": 0.8380136936903,
      "num_tokens": 36372480.0,
      "step": 4440
    },
    {
      "entropy": 0.44678495824337006,
      "epoch": 2.3733333333333335,
      "grad_norm": 0.27249231934547424,
      "learning_rate": 2.1166169803230752e-05,
      "loss": 0.46298727989196775,
      "mean_token_accuracy": 0.8416952073574067,
      "num_tokens": 36454400.0,
      "step": 4450
    },
    {
      "entropy": 0.4526365488767624,
      "epoch": 2.3786666666666667,
      "grad_norm": 0.2729928195476532,
      "learning_rate": 2.082072559889915e-05,
      "loss": 0.4689302444458008,
      "mean_token_accuracy": 0.8395058691501618,
      "num_tokens": 36536320.0,
      "step": 4460
    },
    {
      "entropy": 0.45859123170375826,
      "epoch": 2.384,
      "grad_norm": 0.28589633107185364,
      "learning_rate": 2.0477795713527214e-05,
      "loss": 0.4733113288879395,
      "mean_token_accuracy": 0.8366438329219819,
      "num_tokens": 36618240.0,
      "step": 4470
    },
    {
      "entropy": 0.4517447352409363,
      "epoch": 2.389333333333333,
      "grad_norm": 0.25276878476142883,
      "learning_rate": 2.0137391036771625e-05,
      "loss": 0.46679229736328126,
      "mean_token_accuracy": 0.8389799445867538,
      "num_tokens": 36700160.0,
      "step": 4480
    },
    {
      "entropy": 0.4585997238755226,
      "epoch": 2.3946666666666667,
      "grad_norm": 0.26167282462120056,
      "learning_rate": 1.979952237810159e-05,
      "loss": 0.4741063594818115,
      "mean_token_accuracy": 0.8366071492433548,
      "num_tokens": 36782080.0,
      "step": 4490
    },
    {
      "entropy": 0.44781620651483534,
      "epoch": 2.4,
      "grad_norm": 0.24818018078804016,
      "learning_rate": 1.946420046645574e-05,
      "loss": 0.46209282875061036,
      "mean_token_accuracy": 0.8398727953433991,
      "num_tokens": 36864000.0,
      "step": 4500
    },
    {
      "entropy": 0.4667114421725273,
      "epoch": 2.405333333333333,
      "grad_norm": 0.257123202085495,
      "learning_rate": 1.913143594990138e-05,
      "loss": 0.4828639030456543,
      "mean_token_accuracy": 0.8328400194644928,
      "num_tokens": 36945920.0,
      "step": 4510
    },
    {
      "entropy": 0.4615085259079933,
      "epoch": 2.4106666666666667,
      "grad_norm": 0.2584381103515625,
      "learning_rate": 1.8801239395296245e-05,
      "loss": 0.47613911628723143,
      "mean_token_accuracy": 0.8375,
      "num_tokens": 37027840.0,
      "step": 4520
    },
    {
      "entropy": 0.44482574313879014,
      "epoch": 2.416,
      "grad_norm": 0.2563045024871826,
      "learning_rate": 1.8473621287953146e-05,
      "loss": 0.45844345092773436,
      "mean_token_accuracy": 0.8423189789056778,
      "num_tokens": 37109760.0,
      "step": 4530
    },
    {
      "entropy": 0.4473595589399338,
      "epoch": 2.421333333333333,
      "grad_norm": 0.25328364968299866,
      "learning_rate": 1.814859203130689e-05,
      "loss": 0.4624685287475586,
      "mean_token_accuracy": 0.8398483335971832,
      "num_tokens": 37191680.0,
      "step": 4540
    },
    {
      "entropy": 0.45830340385437013,
      "epoch": 2.4266666666666667,
      "grad_norm": 0.26038241386413574,
      "learning_rate": 1.7826161946583963e-05,
      "loss": 0.4733546733856201,
      "mean_token_accuracy": 0.8360445171594619,
      "num_tokens": 37273600.0,
      "step": 4550
    },
    {
      "entropy": 0.44374454617500303,
      "epoch": 2.432,
      "grad_norm": 0.250342458486557,
      "learning_rate": 1.750634127247468e-05,
      "loss": 0.46002516746520994,
      "mean_token_accuracy": 0.8427837640047073,
      "num_tokens": 37355520.0,
      "step": 4560
    },
    {
      "entropy": 0.45987376421689985,
      "epoch": 2.437333333333333,
      "grad_norm": 0.26905587315559387,
      "learning_rate": 1.7189140164808258e-05,
      "loss": 0.4722400665283203,
      "mean_token_accuracy": 0.836766141653061,
      "num_tokens": 37437440.0,
      "step": 4570
    },
    {
      "entropy": 0.45264346301555636,
      "epoch": 2.4426666666666668,
      "grad_norm": 0.2980540692806244,
      "learning_rate": 1.6874568696230175e-05,
      "loss": 0.4682675838470459,
      "mean_token_accuracy": 0.83780577480793,
      "num_tokens": 37519360.0,
      "step": 4580
    },
    {
      "entropy": 0.44761130660772325,
      "epoch": 2.448,
      "grad_norm": 0.2728835344314575,
      "learning_rate": 1.6562636855882285e-05,
      "loss": 0.4630136966705322,
      "mean_token_accuracy": 0.8400684922933579,
      "num_tokens": 37601280.0,
      "step": 4590
    },
    {
      "entropy": 0.4487395152449608,
      "epoch": 2.453333333333333,
      "grad_norm": 0.24033136665821075,
      "learning_rate": 1.625335454908574e-05,
      "loss": 0.4643992900848389,
      "mean_token_accuracy": 0.8405454963445663,
      "num_tokens": 37683200.0,
      "step": 4600
    },
    {
      "entropy": 0.44841051995754244,
      "epoch": 2.458666666666667,
      "grad_norm": 0.27655768394470215,
      "learning_rate": 1.5946731597026408e-05,
      "loss": 0.4650374412536621,
      "mean_token_accuracy": 0.8392367899417877,
      "num_tokens": 37765120.0,
      "step": 4610
    },
    {
      "entropy": 0.4566493570804596,
      "epoch": 2.464,
      "grad_norm": 0.2565130293369293,
      "learning_rate": 1.564277773644289e-05,
      "loss": 0.46881656646728515,
      "mean_token_accuracy": 0.8386130183935165,
      "num_tokens": 37847040.0,
      "step": 4620
    },
    {
      "entropy": 0.4436029031872749,
      "epoch": 2.469333333333333,
      "grad_norm": 0.262909471988678,
      "learning_rate": 1.5341502619317515e-05,
      "loss": 0.4580474853515625,
      "mean_token_accuracy": 0.8416829764842987,
      "num_tokens": 37928960.0,
      "step": 4630
    },
    {
      "entropy": 0.44710765928030016,
      "epoch": 2.474666666666667,
      "grad_norm": 0.26413020491600037,
      "learning_rate": 1.5042915812569703e-05,
      "loss": 0.45969138145446775,
      "mean_token_accuracy": 0.8421599745750428,
      "num_tokens": 38010880.0,
      "step": 4640
    },
    {
      "entropy": 0.4467643514275551,
      "epoch": 2.48,
      "grad_norm": 0.27428650856018066,
      "learning_rate": 1.4747026797752229e-05,
      "loss": 0.4625056743621826,
      "mean_token_accuracy": 0.8408268183469773,
      "num_tokens": 38092800.0,
      "step": 4650
    },
    {
      "entropy": 0.46534535139799116,
      "epoch": 2.485333333333333,
      "grad_norm": 0.28595706820487976,
      "learning_rate": 1.4453844970750174e-05,
      "loss": 0.47984962463378905,
      "mean_token_accuracy": 0.8357020586729049,
      "num_tokens": 38174720.0,
      "step": 4660
    },
    {
      "entropy": 0.45867831707000734,
      "epoch": 2.490666666666667,
      "grad_norm": 0.27034512162208557,
      "learning_rate": 1.4163379641482422e-05,
      "loss": 0.4715856075286865,
      "mean_token_accuracy": 0.8379280745983124,
      "num_tokens": 38256640.0,
      "step": 4670
    },
    {
      "entropy": 0.4538570076227188,
      "epoch": 2.496,
      "grad_norm": 0.26876991987228394,
      "learning_rate": 1.3875640033606196e-05,
      "loss": 0.4691471576690674,
      "mean_token_accuracy": 0.8371330678462983,
      "num_tokens": 38338560.0,
      "step": 4680
    },
    {
      "entropy": 0.4556427806615829,
      "epoch": 2.501333333333333,
      "grad_norm": 0.2881099581718445,
      "learning_rate": 1.3590635284224097e-05,
      "loss": 0.4712162971496582,
      "mean_token_accuracy": 0.8369618386030198,
      "num_tokens": 38420480.0,
      "step": 4690
    },
    {
      "entropy": 0.4579096272587776,
      "epoch": 2.506666666666667,
      "grad_norm": 0.29092371463775635,
      "learning_rate": 1.3308374443593851e-05,
      "loss": 0.47023835182189944,
      "mean_token_accuracy": 0.8373165339231491,
      "num_tokens": 38502400.0,
      "step": 4700
    },
    {
      "entropy": 0.44774961173534394,
      "epoch": 2.512,
      "grad_norm": 0.2891729772090912,
      "learning_rate": 1.3028866474841117e-05,
      "loss": 0.4632315158843994,
      "mean_token_accuracy": 0.8399828732013702,
      "num_tokens": 38584320.0,
      "step": 4710
    },
    {
      "entropy": 0.4536259397864342,
      "epoch": 2.517333333333333,
      "grad_norm": 0.2757751941680908,
      "learning_rate": 1.2752120253674749e-05,
      "loss": 0.47016267776489257,
      "mean_token_accuracy": 0.8375122308731079,
      "num_tokens": 38666240.0,
      "step": 4720
    },
    {
      "entropy": 0.4506521672010422,
      "epoch": 2.522666666666667,
      "grad_norm": 0.28497275710105896,
      "learning_rate": 1.2478144568104876e-05,
      "loss": 0.4660027503967285,
      "mean_token_accuracy": 0.8386741697788238,
      "num_tokens": 38748160.0,
      "step": 4730
    },
    {
      "entropy": 0.4502082496881485,
      "epoch": 2.528,
      "grad_norm": 0.24598681926727295,
      "learning_rate": 1.2206948118164096e-05,
      "loss": 0.46458234786987307,
      "mean_token_accuracy": 0.8400440245866776,
      "num_tokens": 38830080.0,
      "step": 4740
    },
    {
      "entropy": 0.43981405049562455,
      "epoch": 2.533333333333333,
      "grad_norm": 0.26732921600341797,
      "learning_rate": 1.1938539515630886e-05,
      "loss": 0.45707154273986816,
      "mean_token_accuracy": 0.8425391346216202,
      "num_tokens": 38912000.0,
      "step": 4750
    },
    {
      "entropy": 0.4527917131781578,
      "epoch": 2.538666666666667,
      "grad_norm": 0.28040680289268494,
      "learning_rate": 1.1672927283756319e-05,
      "loss": 0.46486268043518064,
      "mean_token_accuracy": 0.839371332526207,
      "num_tokens": 38993920.0,
      "step": 4760
    },
    {
      "entropy": 0.4444227755069733,
      "epoch": 2.544,
      "grad_norm": 0.2760142683982849,
      "learning_rate": 1.1410119856993517e-05,
      "loss": 0.46082253456115724,
      "mean_token_accuracy": 0.8411448150873184,
      "num_tokens": 39075840.0,
      "step": 4770
    },
    {
      "entropy": 0.44712996631860735,
      "epoch": 2.5493333333333332,
      "grad_norm": 0.30131396651268005,
      "learning_rate": 1.1150125580729509e-05,
      "loss": 0.46242609024047854,
      "mean_token_accuracy": 0.8411815077066421,
      "num_tokens": 39157760.0,
      "step": 4780
    },
    {
      "entropy": 0.4572935774922371,
      "epoch": 2.554666666666667,
      "grad_norm": 0.2738524377346039,
      "learning_rate": 1.0892952711020509e-05,
      "loss": 0.4695780277252197,
      "mean_token_accuracy": 0.8377935409545898,
      "num_tokens": 39239680.0,
      "step": 4790
    },
    {
      "entropy": 0.453690730035305,
      "epoch": 2.56,
      "grad_norm": 0.2725144326686859,
      "learning_rate": 1.0638609414329625e-05,
      "loss": 0.46935582160949707,
      "mean_token_accuracy": 0.8377201527357101,
      "num_tokens": 39321600.0,
      "step": 4800
    },
    {
      "entropy": 0.45265552699565886,
      "epoch": 2.5653333333333332,
      "grad_norm": 0.2913009226322174,
      "learning_rate": 1.0387103767267503e-05,
      "loss": 0.46540656089782717,
      "mean_token_accuracy": 0.8396771043539047,
      "num_tokens": 39403520.0,
      "step": 4810
    },
    {
      "entropy": 0.44755236953496935,
      "epoch": 2.570666666666667,
      "grad_norm": 0.26273438334465027,
      "learning_rate": 1.0138443756335958e-05,
      "loss": 0.4629544734954834,
      "mean_token_accuracy": 0.8411937326192855,
      "num_tokens": 39485440.0,
      "step": 4820
    },
    {
      "entropy": 0.4530874103307724,
      "epoch": 2.576,
      "grad_norm": 0.29529908299446106,
      "learning_rate": 9.892637277674289e-06,
      "loss": 0.4639744281768799,
      "mean_token_accuracy": 0.8414016604423523,
      "num_tokens": 39567360.0,
      "step": 4830
    },
    {
      "entropy": 0.43311204165220263,
      "epoch": 2.5813333333333333,
      "grad_norm": 0.25197288393974304,
      "learning_rate": 9.64969213680852e-06,
      "loss": 0.4501542091369629,
      "mean_token_accuracy": 0.8444716155529022,
      "num_tokens": 39649280.0,
      "step": 4840
    },
    {
      "entropy": 0.4600949317216873,
      "epoch": 2.586666666666667,
      "grad_norm": 0.2921645939350128,
      "learning_rate": 9.409616048403668e-06,
      "loss": 0.47548599243164064,
      "mean_token_accuracy": 0.8365582168102265,
      "num_tokens": 39731200.0,
      "step": 4850
    },
    {
      "entropy": 0.44240161329507827,
      "epoch": 2.592,
      "grad_norm": 0.267330139875412,
      "learning_rate": 9.172416636018588e-06,
      "loss": 0.45755791664123535,
      "mean_token_accuracy": 0.8414505869150162,
      "num_tokens": 39813120.0,
      "step": 4860
    },
    {
      "entropy": 0.447795481979847,
      "epoch": 2.5973333333333333,
      "grad_norm": 0.2813604772090912,
      "learning_rate": 8.938101431863966e-06,
      "loss": 0.4616422176361084,
      "mean_token_accuracy": 0.8420499056577683,
      "num_tokens": 39895040.0,
      "step": 4870
    },
    {
      "entropy": 0.4525182291865349,
      "epoch": 2.602666666666667,
      "grad_norm": 0.2886878550052643,
      "learning_rate": 8.706677876563263e-06,
      "loss": 0.4693244457244873,
      "mean_token_accuracy": 0.8387720167636872,
      "num_tokens": 39976960.0,
      "step": 4880
    },
    {
      "entropy": 0.4551100000739098,
      "epoch": 2.608,
      "grad_norm": 0.26703986525535583,
      "learning_rate": 8.478153318916171e-06,
      "loss": 0.4676485538482666,
      "mean_token_accuracy": 0.838515168428421,
      "num_tokens": 40058880.0,
      "step": 4890
    },
    {
      "entropy": 0.44708543419837954,
      "epoch": 2.6133333333333333,
      "grad_norm": 0.2868178188800812,
      "learning_rate": 8.252535015665497e-06,
      "loss": 0.4631068229675293,
      "mean_token_accuracy": 0.840154105424881,
      "num_tokens": 40140800.0,
      "step": 4900
    },
    {
      "entropy": 0.45270776450634004,
      "epoch": 2.618666666666667,
      "grad_norm": 0.2741808593273163,
      "learning_rate": 8.029830131266637e-06,
      "loss": 0.4664623737335205,
      "mean_token_accuracy": 0.8386986225843429,
      "num_tokens": 40222720.0,
      "step": 4910
    },
    {
      "entropy": 0.4639578312635422,
      "epoch": 2.624,
      "grad_norm": 0.3059973418712616,
      "learning_rate": 7.810045737659989e-06,
      "loss": 0.4747736930847168,
      "mean_token_accuracy": 0.8357387483119965,
      "num_tokens": 40304640.0,
      "step": 4920
    },
    {
      "entropy": 0.4539079934358597,
      "epoch": 2.6293333333333333,
      "grad_norm": 0.27267950773239136,
      "learning_rate": 7.593188814046537e-06,
      "loss": 0.4701559066772461,
      "mean_token_accuracy": 0.8380014628171921,
      "num_tokens": 40386560.0,
      "step": 4930
    },
    {
      "entropy": 0.4451843246817589,
      "epoch": 2.634666666666667,
      "grad_norm": 0.26455599069595337,
      "learning_rate": 7.37926624666615e-06,
      "loss": 0.4587543964385986,
      "mean_token_accuracy": 0.8420865952968597,
      "num_tokens": 40468480.0,
      "step": 4940
    },
    {
      "entropy": 0.45154883712530136,
      "epoch": 2.64,
      "grad_norm": 0.2962474822998047,
      "learning_rate": 7.168284828578842e-06,
      "loss": 0.4668453216552734,
      "mean_token_accuracy": 0.8391756296157837,
      "num_tokens": 40550400.0,
      "step": 4950
    },
    {
      "entropy": 0.4520912140607834,
      "epoch": 2.6453333333333333,
      "grad_norm": 0.31778860092163086,
      "learning_rate": 6.960251259449235e-06,
      "loss": 0.4684610366821289,
      "mean_token_accuracy": 0.8384417772293091,
      "num_tokens": 40632320.0,
      "step": 4960
    },
    {
      "entropy": 0.4528982549905777,
      "epoch": 2.6506666666666665,
      "grad_norm": 0.28653034567832947,
      "learning_rate": 6.755172145333666e-06,
      "loss": 0.4672276496887207,
      "mean_token_accuracy": 0.8388454049825669,
      "num_tokens": 40714240.0,
      "step": 4970
    },
    {
      "entropy": 0.4561873242259026,
      "epoch": 2.656,
      "grad_norm": 0.31481459736824036,
      "learning_rate": 6.553053998470382e-06,
      "loss": 0.46719932556152344,
      "mean_token_accuracy": 0.8387964725494385,
      "num_tokens": 40796160.0,
      "step": 4980
    },
    {
      "entropy": 0.4549671173095703,
      "epoch": 2.6613333333333333,
      "grad_norm": 0.295692503452301,
      "learning_rate": 6.353903237072989e-06,
      "loss": 0.46936726570129395,
      "mean_token_accuracy": 0.8386619329452515,
      "num_tokens": 40878080.0,
      "step": 4990
    },
    {
      "entropy": 0.4517622709274292,
      "epoch": 2.6666666666666665,
      "grad_norm": 0.3115569055080414,
      "learning_rate": 6.157726185126322e-06,
      "loss": 0.46626715660095214,
      "mean_token_accuracy": 0.8389554798603058,
      "num_tokens": 40960000.0,
      "step": 5000
    },
    {
      "entropy": 0.45392653048038484,
      "epoch": 2.672,
      "grad_norm": 0.26847726106643677,
      "learning_rate": 5.964529072185843e-06,
      "loss": 0.4665029525756836,
      "mean_token_accuracy": 0.8403498023748398,
      "num_tokens": 41041920.0,
      "step": 5010
    },
    {
      "entropy": 0.4514739602804184,
      "epoch": 2.6773333333333333,
      "grad_norm": 0.3133648931980133,
      "learning_rate": 5.774318033179804e-06,
      "loss": 0.4645054817199707,
      "mean_token_accuracy": 0.8398116379976273,
      "num_tokens": 41123840.0,
      "step": 5020
    },
    {
      "entropy": 0.4498090326786041,
      "epoch": 2.6826666666666665,
      "grad_norm": 0.2844656705856323,
      "learning_rate": 5.587099108214288e-06,
      "loss": 0.46469731330871583,
      "mean_token_accuracy": 0.8397260248661041,
      "num_tokens": 41205760.0,
      "step": 5030
    },
    {
      "entropy": 0.4490039795637131,
      "epoch": 2.6879999999999997,
      "grad_norm": 0.2590588331222534,
      "learning_rate": 5.402878242381615e-06,
      "loss": 0.4642624378204346,
      "mean_token_accuracy": 0.8398361027240753,
      "num_tokens": 41287680.0,
      "step": 5040
    },
    {
      "entropy": 0.45719897001981735,
      "epoch": 2.6933333333333334,
      "grad_norm": 0.2700011730194092,
      "learning_rate": 5.2216612855713885e-06,
      "loss": 0.4711441516876221,
      "mean_token_accuracy": 0.8374877721071243,
      "num_tokens": 41369600.0,
      "step": 5050
    },
    {
      "entropy": 0.46288054138422013,
      "epoch": 2.6986666666666665,
      "grad_norm": 0.28052258491516113,
      "learning_rate": 5.04345399228483e-06,
      "loss": 0.4738938331604004,
      "mean_token_accuracy": 0.8369618386030198,
      "num_tokens": 41451520.0,
      "step": 5060
    },
    {
      "entropy": 0.4502833381295204,
      "epoch": 2.7039999999999997,
      "grad_norm": 0.26064032316207886,
      "learning_rate": 4.868262021451986e-06,
      "loss": 0.4665383338928223,
      "mean_token_accuracy": 0.8388087064027786,
      "num_tokens": 41533440.0,
      "step": 5070
    },
    {
      "entropy": 0.44439722001552584,
      "epoch": 2.7093333333333334,
      "grad_norm": 0.2844536006450653,
      "learning_rate": 4.696090936252096e-06,
      "loss": 0.4589056491851807,
      "mean_token_accuracy": 0.841621819138527,
      "num_tokens": 41615360.0,
      "step": 5080
    },
    {
      "entropy": 0.4538668513298035,
      "epoch": 2.7146666666666666,
      "grad_norm": 0.2985025644302368,
      "learning_rate": 4.526946203936844e-06,
      "loss": 0.4662965774536133,
      "mean_token_accuracy": 0.8394447147846222,
      "num_tokens": 41697280.0,
      "step": 5090
    },
    {
      "entropy": 0.46122039705514906,
      "epoch": 2.7199999999999998,
      "grad_norm": 0.2884168326854706,
      "learning_rate": 4.360833195656866e-06,
      "loss": 0.4751310348510742,
      "mean_token_accuracy": 0.8365582257509232,
      "num_tokens": 41779200.0,
      "step": 5100
    },
    {
      "entropy": 0.45998714566230775,
      "epoch": 2.7253333333333334,
      "grad_norm": 0.30457043647766113,
      "learning_rate": 4.197757186291063e-06,
      "loss": 0.4748219013214111,
      "mean_token_accuracy": 0.8359955966472625,
      "num_tokens": 41861120.0,
      "step": 5110
    },
    {
      "entropy": 0.44479849487543105,
      "epoch": 2.7306666666666666,
      "grad_norm": 0.29238763451576233,
      "learning_rate": 4.03772335427921e-06,
      "loss": 0.4585710048675537,
      "mean_token_accuracy": 0.8424046009778976,
      "num_tokens": 41943040.0,
      "step": 5120
    },
    {
      "entropy": 0.4618475392460823,
      "epoch": 2.7359999999999998,
      "grad_norm": 0.3157424330711365,
      "learning_rate": 3.880736781457451e-06,
      "loss": 0.4785804271697998,
      "mean_token_accuracy": 0.8357999056577683,
      "num_tokens": 42024960.0,
      "step": 5130
    },
    {
      "entropy": 0.4608913615345955,
      "epoch": 2.7413333333333334,
      "grad_norm": 0.28729718923568726,
      "learning_rate": 3.726802452896916e-06,
      "loss": 0.4764893054962158,
      "mean_token_accuracy": 0.8351516604423523,
      "num_tokens": 42106880.0,
      "step": 5140
    },
    {
      "entropy": 0.44625203758478166,
      "epoch": 2.7466666666666666,
      "grad_norm": 0.32014942169189453,
      "learning_rate": 3.5759252567454602e-06,
      "loss": 0.46076245307922364,
      "mean_token_accuracy": 0.8423189759254456,
      "num_tokens": 42188800.0,
      "step": 5150
    },
    {
      "entropy": 0.4436600223183632,
      "epoch": 2.752,
      "grad_norm": 0.2793176770210266,
      "learning_rate": 3.4281099840724697e-06,
      "loss": 0.45792279243469236,
      "mean_token_accuracy": 0.8423434436321259,
      "num_tokens": 42270720.0,
      "step": 5160
    },
    {
      "entropy": 0.4526971012353897,
      "epoch": 2.7573333333333334,
      "grad_norm": 0.2883247435092926,
      "learning_rate": 3.2833613287166052e-06,
      "loss": 0.46686615943908694,
      "mean_token_accuracy": 0.8379892349243164,
      "num_tokens": 42352640.0,
      "step": 5170
    },
    {
      "entropy": 0.4579934701323509,
      "epoch": 2.7626666666666666,
      "grad_norm": 0.2697383761405945,
      "learning_rate": 3.1416838871368924e-06,
      "loss": 0.4703409194946289,
      "mean_token_accuracy": 0.83637475669384,
      "num_tokens": 42434560.0,
      "step": 5180
    },
    {
      "entropy": 0.45192168056964876,
      "epoch": 2.768,
      "grad_norm": 0.32307878136634827,
      "learning_rate": 3.003082158266657e-06,
      "loss": 0.4644754409790039,
      "mean_token_accuracy": 0.839163401722908,
      "num_tokens": 42516480.0,
      "step": 5190
    },
    {
      "entropy": 0.455105622112751,
      "epoch": 2.7733333333333334,
      "grad_norm": 0.329609751701355,
      "learning_rate": 2.867560543370751e-06,
      "loss": 0.46916804313659666,
      "mean_token_accuracy": 0.8372798383235931,
      "num_tokens": 42598400.0,
      "step": 5200
    },
    {
      "entropy": 0.4457827642560005,
      "epoch": 2.7786666666666666,
      "grad_norm": 0.29751548171043396,
      "learning_rate": 2.735123345905677e-06,
      "loss": 0.4584932804107666,
      "mean_token_accuracy": 0.8421722054481506,
      "num_tokens": 42680320.0,
      "step": 5210
    },
    {
      "entropy": 0.4453358739614487,
      "epoch": 2.784,
      "grad_norm": 0.2674790918827057,
      "learning_rate": 2.6057747713830293e-06,
      "loss": 0.455474328994751,
      "mean_token_accuracy": 0.8424412995576859,
      "num_tokens": 42762240.0,
      "step": 5220
    },
    {
      "entropy": 0.4565201625227928,
      "epoch": 2.7893333333333334,
      "grad_norm": 0.30781424045562744,
      "learning_rate": 2.4795189272359354e-06,
      "loss": 0.4705306053161621,
      "mean_token_accuracy": 0.8383317023515702,
      "num_tokens": 42844160.0,
      "step": 5230
    },
    {
      "entropy": 0.46073125302791595,
      "epoch": 2.7946666666666666,
      "grad_norm": 0.2797315716743469,
      "learning_rate": 2.3563598226885808e-06,
      "loss": 0.47438869476318357,
      "mean_token_accuracy": 0.8359711349010468,
      "num_tokens": 42926080.0,
      "step": 5240
    },
    {
      "entropy": 0.45304558873176576,
      "epoch": 2.8,
      "grad_norm": 0.30737870931625366,
      "learning_rate": 2.2363013686289236e-06,
      "loss": 0.46637873649597167,
      "mean_token_accuracy": 0.8387108594179153,
      "num_tokens": 43008000.0,
      "step": 5250
    },
    {
      "entropy": 0.4527182370424271,
      "epoch": 2.8053333333333335,
      "grad_norm": 0.28889599442481995,
      "learning_rate": 2.1193473774845486e-06,
      "loss": 0.4643861770629883,
      "mean_token_accuracy": 0.8389554768800735,
      "num_tokens": 43089920.0,
      "step": 5260
    },
    {
      "entropy": 0.4460126429796219,
      "epoch": 2.8106666666666666,
      "grad_norm": 0.2682284116744995,
      "learning_rate": 2.0055015631014864e-06,
      "loss": 0.45789422988891604,
      "mean_token_accuracy": 0.8437010794878006,
      "num_tokens": 43171840.0,
      "step": 5270
    },
    {
      "entropy": 0.4479100704193115,
      "epoch": 2.816,
      "grad_norm": 0.3133639097213745,
      "learning_rate": 1.8947675406264298e-06,
      "loss": 0.46192097663879395,
      "mean_token_accuracy": 0.8403497993946075,
      "num_tokens": 43253760.0,
      "step": 5280
    },
    {
      "entropy": 0.4563271626830101,
      "epoch": 2.8213333333333335,
      "grad_norm": 0.3027438819408417,
      "learning_rate": 1.7871488263918267e-06,
      "loss": 0.47133522033691405,
      "mean_token_accuracy": 0.8384173214435577,
      "num_tokens": 43335680.0,
      "step": 5290
    },
    {
      "entropy": 0.45812766402959826,
      "epoch": 2.8266666666666667,
      "grad_norm": 0.26634904742240906,
      "learning_rate": 1.6826488378042572e-06,
      "loss": 0.47136754989624025,
      "mean_token_accuracy": 0.837157529592514,
      "num_tokens": 43417600.0,
      "step": 5300
    },
    {
      "entropy": 0.4506367400288582,
      "epoch": 2.832,
      "grad_norm": 0.2991483509540558,
      "learning_rate": 1.5812708932359198e-06,
      "loss": 0.46463494300842284,
      "mean_token_accuracy": 0.8391022473573685,
      "num_tokens": 43499520.0,
      "step": 5310
    },
    {
      "entropy": 0.450600166618824,
      "epoch": 2.8373333333333335,
      "grad_norm": 0.3214839696884155,
      "learning_rate": 1.4830182119192403e-06,
      "loss": 0.46474614143371584,
      "mean_token_accuracy": 0.839567020535469,
      "num_tokens": 43581440.0,
      "step": 5320
    },
    {
      "entropy": 0.4504974618554115,
      "epoch": 2.8426666666666667,
      "grad_norm": 0.3197139799594879,
      "learning_rate": 1.3878939138446622e-06,
      "loss": 0.46619091033935545,
      "mean_token_accuracy": 0.8390777856111526,
      "num_tokens": 43663360.0,
      "step": 5330
    },
    {
      "entropy": 0.44753725826740265,
      "epoch": 2.848,
      "grad_norm": 0.29300546646118164,
      "learning_rate": 1.2959010196615829e-06,
      "loss": 0.46111645698547366,
      "mean_token_accuracy": 0.8416707426309585,
      "num_tokens": 43745280.0,
      "step": 5340
    },
    {
      "entropy": 0.46371145695447924,
      "epoch": 2.8533333333333335,
      "grad_norm": 0.31372490525245667,
      "learning_rate": 1.207042450582374e-06,
      "loss": 0.4735804557800293,
      "mean_token_accuracy": 0.8351638913154602,
      "num_tokens": 43827200.0,
      "step": 5350
    },
    {
      "entropy": 0.46155391335487367,
      "epoch": 2.8586666666666667,
      "grad_norm": 0.2849220037460327,
      "learning_rate": 1.1213210282897014e-06,
      "loss": 0.47335071563720704,
      "mean_token_accuracy": 0.8359833717346191,
      "num_tokens": 43909120.0,
      "step": 5360
    },
    {
      "entropy": 0.4528530493378639,
      "epoch": 2.864,
      "grad_norm": 0.3159361183643341,
      "learning_rate": 1.0387394748468726e-06,
      "loss": 0.46541948318481446,
      "mean_token_accuracy": 0.8400684952735901,
      "num_tokens": 43991040.0,
      "step": 5370
    },
    {
      "entropy": 0.4492829918861389,
      "epoch": 2.8693333333333335,
      "grad_norm": 0.2818220853805542,
      "learning_rate": 9.59300412611397e-07,
      "loss": 0.4634866714477539,
      "mean_token_accuracy": 0.8403864949941635,
      "num_tokens": 44072960.0,
      "step": 5380
    },
    {
      "entropy": 0.4488120019435883,
      "epoch": 2.8746666666666667,
      "grad_norm": 0.2998656928539276,
      "learning_rate": 8.830063641517395e-07,
      "loss": 0.46211652755737304,
      "mean_token_accuracy": 0.8402886390686035,
      "num_tokens": 44154880.0,
      "step": 5390
    },
    {
      "entropy": 0.4545857846736908,
      "epoch": 2.88,
      "grad_norm": 0.30847206711769104,
      "learning_rate": 8.09859752167208e-07,
      "loss": 0.46671667098999026,
      "mean_token_accuracy": 0.8386619389057159,
      "num_tokens": 44236800.0,
      "step": 5400
    },
    {
      "entropy": 0.44052227288484574,
      "epoch": 2.8853333333333335,
      "grad_norm": 0.2986578047275543,
      "learning_rate": 7.398628994109813e-07,
      "loss": 0.45644354820251465,
      "mean_token_accuracy": 0.8433830738067627,
      "num_tokens": 44318720.0,
      "step": 5410
    },
    {
      "entropy": 0.4432691514492035,
      "epoch": 2.8906666666666667,
      "grad_norm": 0.27898529171943665,
      "learning_rate": 6.730180286164345e-07,
      "loss": 0.45602927207946775,
      "mean_token_accuracy": 0.8424046039581299,
      "num_tokens": 44400640.0,
      "step": 5420
    },
    {
      "entropy": 0.4528514310717583,
      "epoch": 2.896,
      "grad_norm": 0.3124619722366333,
      "learning_rate": 6.093272624264845e-07,
      "loss": 0.4649935722351074,
      "mean_token_accuracy": 0.8396648734807968,
      "num_tokens": 44482560.0,
      "step": 5430
    },
    {
      "entropy": 0.4640558212995529,
      "epoch": 2.9013333333333335,
      "grad_norm": 0.3163350522518158,
      "learning_rate": 5.487926233261998e-07,
      "loss": 0.4764530181884766,
      "mean_token_accuracy": 0.8357142865657806,
      "num_tokens": 44564480.0,
      "step": 5440
    },
    {
      "entropy": 0.4557580292224884,
      "epoch": 2.9066666666666667,
      "grad_norm": 0.31627100706100464,
      "learning_rate": 4.914160335786066e-07,
      "loss": 0.46790485382080077,
      "mean_token_accuracy": 0.8372431457042694,
      "num_tokens": 44646400.0,
      "step": 5450
    },
    {
      "entropy": 0.4475193291902542,
      "epoch": 2.912,
      "grad_norm": 0.286876380443573,
      "learning_rate": 4.371993151635945e-07,
      "loss": 0.4607985019683838,
      "mean_token_accuracy": 0.8413282811641694,
      "num_tokens": 44728320.0,
      "step": 5460
    },
    {
      "entropy": 0.44934196919202807,
      "epoch": 2.9173333333333336,
      "grad_norm": 0.27818113565444946,
      "learning_rate": 3.86144189720139e-07,
      "loss": 0.4626044750213623,
      "mean_token_accuracy": 0.8408268064260482,
      "num_tokens": 44810240.0,
      "step": 5470
    },
    {
      "entropy": 0.44739541560411455,
      "epoch": 2.9226666666666667,
      "grad_norm": 0.31981348991394043,
      "learning_rate": 3.382522784915687e-07,
      "loss": 0.4610436916351318,
      "mean_token_accuracy": 0.840631115436554,
      "num_tokens": 44892160.0,
      "step": 5480
    },
    {
      "entropy": 0.45534133464097976,
      "epoch": 2.928,
      "grad_norm": 0.29233241081237793,
      "learning_rate": 2.935251022740615e-07,
      "loss": 0.4685820579528809,
      "mean_token_accuracy": 0.8385029345750808,
      "num_tokens": 44974080.0,
      "step": 5490
    },
    {
      "entropy": 0.45534287989139555,
      "epoch": 2.9333333333333336,
      "grad_norm": 0.2768395245075226,
      "learning_rate": 2.519640813684498e-07,
      "loss": 0.46599640846252444,
      "mean_token_accuracy": 0.8385763227939605,
      "num_tokens": 45056000.0,
      "step": 5500
    },
    {
      "entropy": 0.4526157841086388,
      "epoch": 2.9386666666666668,
      "grad_norm": 0.2910348176956177,
      "learning_rate": 2.135705355350237e-07,
      "loss": 0.466368293762207,
      "mean_token_accuracy": 0.8379158526659012,
      "num_tokens": 45137920.0,
      "step": 5510
    },
    {
      "entropy": 0.4534804970026016,
      "epoch": 2.944,
      "grad_norm": 0.30979251861572266,
      "learning_rate": 1.7834568395166395e-07,
      "loss": 0.4673768043518066,
      "mean_token_accuracy": 0.8381360054016114,
      "num_tokens": 45219840.0,
      "step": 5520
    },
    {
      "entropy": 0.4612312138080597,
      "epoch": 2.9493333333333336,
      "grad_norm": 0.33005690574645996,
      "learning_rate": 1.4629064517514003e-07,
      "loss": 0.4743774890899658,
      "mean_token_accuracy": 0.8374021559953689,
      "num_tokens": 45301760.0,
      "step": 5530
    },
    {
      "entropy": 0.4531767278909683,
      "epoch": 2.9546666666666668,
      "grad_norm": 0.2805468440055847,
      "learning_rate": 1.1740643710553833e-07,
      "loss": 0.4644935131072998,
      "mean_token_accuracy": 0.8385151654481888,
      "num_tokens": 45383680.0,
      "step": 5540
    },
    {
      "entropy": 0.4589652240276337,
      "epoch": 2.96,
      "grad_norm": 0.2965894341468811,
      "learning_rate": 9.169397695399928e-08,
      "loss": 0.4720900058746338,
      "mean_token_accuracy": 0.8371086090803146,
      "num_tokens": 45465600.0,
      "step": 5550
    },
    {
      "entropy": 0.4469110697507858,
      "epoch": 2.9653333333333336,
      "grad_norm": 0.2930002808570862,
      "learning_rate": 6.91540812135516e-08,
      "loss": 0.45932540893554685,
      "mean_token_accuracy": 0.8421232908964157,
      "num_tokens": 45547520.0,
      "step": 5560
    },
    {
      "entropy": 0.44865818321704865,
      "epoch": 2.970666666666667,
      "grad_norm": 0.3220630884170532,
      "learning_rate": 4.9787465633222006e-08,
      "loss": 0.45930633544921873,
      "mean_token_accuracy": 0.8401174157857895,
      "num_tokens": 45629440.0,
      "step": 5570
    },
    {
      "entropy": 0.4525710791349411,
      "epoch": 2.976,
      "grad_norm": 0.3257077932357788,
      "learning_rate": 3.3594745195242304e-08,
      "loss": 0.4674736499786377,
      "mean_token_accuracy": 0.8400195658206939,
      "num_tokens": 45711360.0,
      "step": 5580
    },
    {
      "entropy": 0.44830994606018065,
      "epoch": 2.981333333333333,
      "grad_norm": 0.2981422543525696,
      "learning_rate": 2.0576434095609388e-08,
      "loss": 0.4635817527770996,
      "mean_token_accuracy": 0.8397749543190003,
      "num_tokens": 45793280.0,
      "step": 5590
    },
    {
      "entropy": 0.4499359354376793,
      "epoch": 2.986666666666667,
      "grad_norm": 0.28879281878471375,
      "learning_rate": 1.0732945727665034e-08,
      "loss": 0.4629524230957031,
      "mean_token_accuracy": 0.8390288650989532,
      "num_tokens": 45875200.0,
      "step": 5600
    },
    {
      "entropy": 0.44630577266216276,
      "epoch": 2.992,
      "grad_norm": 0.2833120822906494,
      "learning_rate": 4.0645926690396905e-09,
      "loss": 0.4590027332305908,
      "mean_token_accuracy": 0.8410714268684387,
      "num_tokens": 45957120.0,
      "step": 5610
    },
    {
      "entropy": 0.46539789736270903,
      "epoch": 2.997333333333333,
      "grad_norm": 0.29831957817077637,
      "learning_rate": 5.71586671693769e-10,
      "loss": 0.4785189151763916,
      "mean_token_accuracy": 0.8346991151571274,
      "num_tokens": 46039040.0,
      "step": 5620
    },
    {
      "entropy": 0.4503598093986511,
      "epoch": 3.0,
      "mean_token_accuracy": 0.8388209342956543,
      "num_tokens": 46080000.0,
      "step": 5625,
      "total_flos": 3.193730629632e+16,
      "train_loss": 0.47722113210890027,
      "train_runtime": 5619.0378,
      "train_samples_per_second": 16.017,
      "train_steps_per_second": 1.001
    }
  ],
  "logging_steps": 10,
  "max_steps": 5625,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.193730629632e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}