{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1954,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0005117707267144319,
      "grad_norm": 11.5,
      "learning_rate": 0.0,
      "loss": 0.6615302562713623,
      "mean_token_accuracy": 0.8735949993133545,
      "num_tokens": 2320079.0,
      "step": 1
    },
    {
      "epoch": 0.0010235414534288639,
      "grad_norm": 10.3125,
      "learning_rate": 6.779661016949153e-07,
      "loss": 0.6237350106239319,
      "mean_token_accuracy": 0.8759393692016602,
      "num_tokens": 4696276.0,
      "step": 2
    },
    {
      "epoch": 0.0015353121801432957,
      "grad_norm": 10.5,
      "learning_rate": 1.3559322033898307e-06,
      "loss": 0.6227678060531616,
      "mean_token_accuracy": 0.8779875636100769,
      "num_tokens": 7025849.0,
      "step": 3
    },
    {
      "epoch": 0.0020470829068577278,
      "grad_norm": 10.0,
      "learning_rate": 2.033898305084746e-06,
      "loss": 0.599985659122467,
      "mean_token_accuracy": 0.8816940784454346,
      "num_tokens": 9403873.0,
      "step": 4
    },
    {
      "epoch": 0.00255885363357216,
      "grad_norm": 9.9375,
      "learning_rate": 2.7118644067796613e-06,
      "loss": 0.599875807762146,
      "mean_token_accuracy": 0.8815739154815674,
      "num_tokens": 11727019.0,
      "step": 5
    },
    {
      "epoch": 0.0030706243602865915,
      "grad_norm": 9.8125,
      "learning_rate": 3.3898305084745763e-06,
      "loss": 0.5986174941062927,
      "mean_token_accuracy": 0.8789271116256714,
      "num_tokens": 14066866.0,
      "step": 6
    },
    {
      "epoch": 0.0035823950870010235,
      "grad_norm": 9.4375,
      "learning_rate": 4.067796610169492e-06,
      "loss": 0.5987590551376343,
      "mean_token_accuracy": 0.8753452301025391,
      "num_tokens": 16466175.0,
      "step": 7
    },
    {
      "epoch": 0.0040941658137154556,
      "grad_norm": 9.4375,
      "learning_rate": 4.745762711864408e-06,
      "loss": 0.5874762535095215,
      "mean_token_accuracy": 0.8820503354072571,
      "num_tokens": 18765502.0,
      "step": 8
    },
    {
      "epoch": 0.004605936540429887,
      "grad_norm": 8.0625,
      "learning_rate": 5.423728813559323e-06,
      "loss": 0.5618375539779663,
      "mean_token_accuracy": 0.8773337602615356,
      "num_tokens": 21099871.0,
      "step": 9
    },
    {
      "epoch": 0.00511770726714432,
      "grad_norm": 7.5625,
      "learning_rate": 6.1016949152542385e-06,
      "loss": 0.5771276950836182,
      "mean_token_accuracy": 0.8724632263183594,
      "num_tokens": 23407837.0,
      "step": 10
    },
    {
      "epoch": 0.005629477993858751,
      "grad_norm": 6.5,
      "learning_rate": 6.779661016949153e-06,
      "loss": 0.5555146336555481,
      "mean_token_accuracy": 0.8743457794189453,
      "num_tokens": 25661091.0,
      "step": 11
    },
    {
      "epoch": 0.006141248720573183,
      "grad_norm": 5.96875,
      "learning_rate": 7.4576271186440685e-06,
      "loss": 0.5288861989974976,
      "mean_token_accuracy": 0.8811137080192566,
      "num_tokens": 27892031.0,
      "step": 12
    },
    {
      "epoch": 0.006653019447287615,
      "grad_norm": 5.0,
      "learning_rate": 8.135593220338983e-06,
      "loss": 0.5005252361297607,
      "mean_token_accuracy": 0.882656455039978,
      "num_tokens": 30237454.0,
      "step": 13
    },
    {
      "epoch": 0.007164790174002047,
      "grad_norm": 4.625,
      "learning_rate": 8.8135593220339e-06,
      "loss": 0.5040340423583984,
      "mean_token_accuracy": 0.8786897659301758,
      "num_tokens": 32513258.0,
      "step": 14
    },
    {
      "epoch": 0.007676560900716479,
      "grad_norm": 4.03125,
      "learning_rate": 9.491525423728815e-06,
      "loss": 0.4535851776599884,
      "mean_token_accuracy": 0.8873242735862732,
      "num_tokens": 34880583.0,
      "step": 15
    },
    {
      "epoch": 0.008188331627430911,
      "grad_norm": 4.03125,
      "learning_rate": 1.016949152542373e-05,
      "loss": 0.44062256813049316,
      "mean_token_accuracy": 0.884108304977417,
      "num_tokens": 37239381.0,
      "step": 16
    },
    {
      "epoch": 0.008700102354145343,
      "grad_norm": 4.09375,
      "learning_rate": 1.0847457627118645e-05,
      "loss": 0.41929304599761963,
      "mean_token_accuracy": 0.8841854333877563,
      "num_tokens": 39497263.0,
      "step": 17
    },
    {
      "epoch": 0.009211873080859774,
      "grad_norm": 3.59375,
      "learning_rate": 1.1525423728813561e-05,
      "loss": 0.3988308310508728,
      "mean_token_accuracy": 0.8874022960662842,
      "num_tokens": 41782456.0,
      "step": 18
    },
    {
      "epoch": 0.009723643807574206,
      "grad_norm": 2.25,
      "learning_rate": 1.2203389830508477e-05,
      "loss": 0.37021446228027344,
      "mean_token_accuracy": 0.8906716108322144,
      "num_tokens": 44152883.0,
      "step": 19
    },
    {
      "epoch": 0.01023541453428864,
      "grad_norm": 1.203125,
      "learning_rate": 1.288135593220339e-05,
      "loss": 0.33961254358291626,
      "mean_token_accuracy": 0.8958112597465515,
      "num_tokens": 46521545.0,
      "step": 20
    },
    {
      "epoch": 0.010747185261003071,
      "grad_norm": 0.98046875,
      "learning_rate": 1.3559322033898305e-05,
      "loss": 0.34668752551078796,
      "mean_token_accuracy": 0.8941212892532349,
      "num_tokens": 48857983.0,
      "step": 21
    },
    {
      "epoch": 0.011258955987717503,
      "grad_norm": 0.95703125,
      "learning_rate": 1.4237288135593221e-05,
      "loss": 0.3559458255767822,
      "mean_token_accuracy": 0.8909661769866943,
      "num_tokens": 51139956.0,
      "step": 22
    },
    {
      "epoch": 0.011770726714431934,
      "grad_norm": 0.91015625,
      "learning_rate": 1.4915254237288137e-05,
      "loss": 0.32786354422569275,
      "mean_token_accuracy": 0.8965020775794983,
      "num_tokens": 53516594.0,
      "step": 23
    },
    {
      "epoch": 0.012282497441146366,
      "grad_norm": 0.83984375,
      "learning_rate": 1.5593220338983053e-05,
      "loss": 0.34585118293762207,
      "mean_token_accuracy": 0.8866230249404907,
      "num_tokens": 55860072.0,
      "step": 24
    },
    {
      "epoch": 0.0127942681678608,
      "grad_norm": 0.77734375,
      "learning_rate": 1.6271186440677967e-05,
      "loss": 0.3193508982658386,
      "mean_token_accuracy": 0.8939554691314697,
      "num_tokens": 58213864.0,
      "step": 25
    },
    {
      "epoch": 0.01330603889457523,
      "grad_norm": 0.7578125,
      "learning_rate": 1.694915254237288e-05,
      "loss": 0.3014819622039795,
      "mean_token_accuracy": 0.8997341394424438,
      "num_tokens": 60528111.0,
      "step": 26
    },
    {
      "epoch": 0.013817809621289662,
      "grad_norm": 0.73046875,
      "learning_rate": 1.76271186440678e-05,
      "loss": 0.31091025471687317,
      "mean_token_accuracy": 0.8942556381225586,
      "num_tokens": 62823035.0,
      "step": 27
    },
    {
      "epoch": 0.014329580348004094,
      "grad_norm": 0.73046875,
      "learning_rate": 1.8305084745762713e-05,
      "loss": 0.29065370559692383,
      "mean_token_accuracy": 0.9010685682296753,
      "num_tokens": 65123249.0,
      "step": 28
    },
    {
      "epoch": 0.014841351074718526,
      "grad_norm": 0.74609375,
      "learning_rate": 1.898305084745763e-05,
      "loss": 0.2881646454334259,
      "mean_token_accuracy": 0.9003605842590332,
      "num_tokens": 67424276.0,
      "step": 29
    },
    {
      "epoch": 0.015353121801432957,
      "grad_norm": 0.8203125,
      "learning_rate": 1.9661016949152545e-05,
      "loss": 0.30136775970458984,
      "mean_token_accuracy": 0.8926438689231873,
      "num_tokens": 69734742.0,
      "step": 30
    },
    {
      "epoch": 0.01586489252814739,
      "grad_norm": 0.84765625,
      "learning_rate": 2.033898305084746e-05,
      "loss": 0.2823793292045593,
      "mean_token_accuracy": 0.8976598978042603,
      "num_tokens": 72079235.0,
      "step": 31
    },
    {
      "epoch": 0.016376663254861822,
      "grad_norm": 0.9296875,
      "learning_rate": 2.1016949152542376e-05,
      "loss": 0.2715657949447632,
      "mean_token_accuracy": 0.8980191946029663,
      "num_tokens": 74397265.0,
      "step": 32
    },
    {
      "epoch": 0.016888433981576252,
      "grad_norm": 0.95703125,
      "learning_rate": 2.169491525423729e-05,
      "loss": 0.25339755415916443,
      "mean_token_accuracy": 0.9006052017211914,
      "num_tokens": 76770466.0,
      "step": 33
    },
    {
      "epoch": 0.017400204708290685,
      "grad_norm": 0.90234375,
      "learning_rate": 2.2372881355932205e-05,
      "loss": 0.26326876878738403,
      "mean_token_accuracy": 0.8995774984359741,
      "num_tokens": 79142833.0,
      "step": 34
    },
    {
      "epoch": 0.01791197543500512,
      "grad_norm": 0.66015625,
      "learning_rate": 2.3050847457627122e-05,
      "loss": 0.2743386924266815,
      "mean_token_accuracy": 0.8940516710281372,
      "num_tokens": 81566064.0,
      "step": 35
    },
    {
      "epoch": 0.01842374616171955,
      "grad_norm": 0.357421875,
      "learning_rate": 2.3728813559322036e-05,
      "loss": 0.25603926181793213,
      "mean_token_accuracy": 0.900063157081604,
      "num_tokens": 83917037.0,
      "step": 36
    },
    {
      "epoch": 0.018935516888433982,
      "grad_norm": 0.298828125,
      "learning_rate": 2.4406779661016954e-05,
      "loss": 0.25153395533561707,
      "mean_token_accuracy": 0.9015868306159973,
      "num_tokens": 86182220.0,
      "step": 37
    },
    {
      "epoch": 0.019447287615148412,
      "grad_norm": 0.26953125,
      "learning_rate": 2.5084745762711865e-05,
      "loss": 0.2470320463180542,
      "mean_token_accuracy": 0.9035332798957825,
      "num_tokens": 88490982.0,
      "step": 38
    },
    {
      "epoch": 0.019959058341862845,
      "grad_norm": 0.283203125,
      "learning_rate": 2.576271186440678e-05,
      "loss": 0.24254797399044037,
      "mean_token_accuracy": 0.9051077961921692,
      "num_tokens": 90810721.0,
      "step": 39
    },
    {
      "epoch": 0.02047082906857728,
      "grad_norm": 0.28125,
      "learning_rate": 2.6440677966101696e-05,
      "loss": 0.2485657036304474,
      "mean_token_accuracy": 0.9023720026016235,
      "num_tokens": 93104606.0,
      "step": 40
    },
    {
      "epoch": 0.02098259979529171,
      "grad_norm": 0.27734375,
      "learning_rate": 2.711864406779661e-05,
      "loss": 0.24406373500823975,
      "mean_token_accuracy": 0.903762936592102,
      "num_tokens": 95387513.0,
      "step": 41
    },
    {
      "epoch": 0.021494370522006142,
      "grad_norm": 0.26171875,
      "learning_rate": 2.7796610169491528e-05,
      "loss": 0.25467628240585327,
      "mean_token_accuracy": 0.9009064435958862,
      "num_tokens": 97663850.0,
      "step": 42
    },
    {
      "epoch": 0.022006141248720572,
      "grad_norm": 0.2578125,
      "learning_rate": 2.8474576271186442e-05,
      "loss": 0.2538940906524658,
      "mean_token_accuracy": 0.9008869528770447,
      "num_tokens": 100025394.0,
      "step": 43
    },
    {
      "epoch": 0.022517911975435005,
      "grad_norm": 0.265625,
      "learning_rate": 2.9152542372881356e-05,
      "loss": 0.2588855028152466,
      "mean_token_accuracy": 0.8986419439315796,
      "num_tokens": 102346673.0,
      "step": 44
    },
    {
      "epoch": 0.02302968270214944,
      "grad_norm": 0.2470703125,
      "learning_rate": 2.9830508474576274e-05,
      "loss": 0.24579349160194397,
      "mean_token_accuracy": 0.9038864374160767,
      "num_tokens": 104679401.0,
      "step": 45
    },
    {
      "epoch": 0.02354145342886387,
      "grad_norm": 0.23828125,
      "learning_rate": 3.0508474576271188e-05,
      "loss": 0.2431679666042328,
      "mean_token_accuracy": 0.9045011401176453,
      "num_tokens": 106962025.0,
      "step": 46
    },
    {
      "epoch": 0.024053224155578302,
      "grad_norm": 0.236328125,
      "learning_rate": 3.1186440677966106e-05,
      "loss": 0.23930755257606506,
      "mean_token_accuracy": 0.9057190418243408,
      "num_tokens": 109249155.0,
      "step": 47
    },
    {
      "epoch": 0.02456499488229273,
      "grad_norm": 0.234375,
      "learning_rate": 3.186440677966102e-05,
      "loss": 0.2506660223007202,
      "mean_token_accuracy": 0.9016631841659546,
      "num_tokens": 111564988.0,
      "step": 48
    },
    {
      "epoch": 0.025076765609007165,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.2542372881355934e-05,
      "loss": 0.24912282824516296,
      "mean_token_accuracy": 0.9023576378822327,
      "num_tokens": 113877895.0,
      "step": 49
    },
    {
      "epoch": 0.0255885363357216,
      "grad_norm": 0.244140625,
      "learning_rate": 3.322033898305085e-05,
      "loss": 0.2605428099632263,
      "mean_token_accuracy": 0.8973718881607056,
      "num_tokens": 116194059.0,
      "step": 50
    },
    {
      "epoch": 0.026100307062436028,
      "grad_norm": 0.228515625,
      "learning_rate": 3.389830508474576e-05,
      "loss": 0.23745524883270264,
      "mean_token_accuracy": 0.9060723781585693,
      "num_tokens": 118560566.0,
      "step": 51
    },
    {
      "epoch": 0.02661207778915046,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.457627118644068e-05,
      "loss": 0.24948260188102722,
      "mean_token_accuracy": 0.9019341468811035,
      "num_tokens": 120910639.0,
      "step": 52
    },
    {
      "epoch": 0.02712384851586489,
      "grad_norm": 0.240234375,
      "learning_rate": 3.52542372881356e-05,
      "loss": 0.2526223659515381,
      "mean_token_accuracy": 0.9017540216445923,
      "num_tokens": 123241486.0,
      "step": 53
    },
    {
      "epoch": 0.027635619242579325,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.593220338983051e-05,
      "loss": 0.2567969560623169,
      "mean_token_accuracy": 0.8988749980926514,
      "num_tokens": 125578004.0,
      "step": 54
    },
    {
      "epoch": 0.028147389969293755,
      "grad_norm": 0.224609375,
      "learning_rate": 3.6610169491525426e-05,
      "loss": 0.24165990948677063,
      "mean_token_accuracy": 0.9049586057662964,
      "num_tokens": 127933479.0,
      "step": 55
    },
    {
      "epoch": 0.028659160696008188,
      "grad_norm": 0.2265625,
      "learning_rate": 3.728813559322034e-05,
      "loss": 0.24376311898231506,
      "mean_token_accuracy": 0.904253363609314,
      "num_tokens": 130261031.0,
      "step": 56
    },
    {
      "epoch": 0.02917093142272262,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.796610169491526e-05,
      "loss": 0.2403348982334137,
      "mean_token_accuracy": 0.9053685665130615,
      "num_tokens": 132583771.0,
      "step": 57
    },
    {
      "epoch": 0.02968270214943705,
      "grad_norm": 0.240234375,
      "learning_rate": 3.8644067796610175e-05,
      "loss": 0.25105395913124084,
      "mean_token_accuracy": 0.9011954069137573,
      "num_tokens": 134825045.0,
      "step": 58
    },
    {
      "epoch": 0.030194472876151485,
      "grad_norm": 0.2265625,
      "learning_rate": 3.932203389830509e-05,
      "loss": 0.23845598101615906,
      "mean_token_accuracy": 0.9067290425300598,
      "num_tokens": 137157670.0,
      "step": 59
    },
    {
      "epoch": 0.030706243602865915,
      "grad_norm": 0.2099609375,
      "learning_rate": 4e-05,
      "loss": 0.24293529987335205,
      "mean_token_accuracy": 0.9043503999710083,
      "num_tokens": 139536201.0,
      "step": 60
    },
    {
      "epoch": 0.031218014329580348,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.999997251591916e-05,
      "loss": 0.23977050185203552,
      "mean_token_accuracy": 0.9054509997367859,
      "num_tokens": 141908204.0,
      "step": 61
    },
    {
      "epoch": 0.03172978505629478,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.9999890063752174e-05,
      "loss": 0.23753559589385986,
      "mean_token_accuracy": 0.9050326347351074,
      "num_tokens": 144215775.0,
      "step": 62
    },
    {
      "epoch": 0.03224155578300921,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.999975264372565e-05,
      "loss": 0.2436637580394745,
      "mean_token_accuracy": 0.9037324786186218,
      "num_tokens": 146530159.0,
      "step": 63
    },
    {
      "epoch": 0.032753326509723645,
      "grad_norm": 0.22265625,
      "learning_rate": 3.9999560256217275e-05,
      "loss": 0.24489296972751617,
      "mean_token_accuracy": 0.903905987739563,
      "num_tokens": 148842642.0,
      "step": 64
    },
    {
      "epoch": 0.03326509723643808,
      "grad_norm": 0.23046875,
      "learning_rate": 3.999931290175582e-05,
      "loss": 0.23913097381591797,
      "mean_token_accuracy": 0.9058101177215576,
      "num_tokens": 151184479.0,
      "step": 65
    },
    {
      "epoch": 0.033776867963152504,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.999901058102109e-05,
      "loss": 0.2515897750854492,
      "mean_token_accuracy": 0.9008671045303345,
      "num_tokens": 153545286.0,
      "step": 66
    },
    {
      "epoch": 0.03428863868986694,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.999865329484402e-05,
      "loss": 0.24594755470752716,
      "mean_token_accuracy": 0.9027431011199951,
      "num_tokens": 155820945.0,
      "step": 67
    },
    {
      "epoch": 0.03480040941658137,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.999824104420656e-05,
      "loss": 0.24103160202503204,
      "mean_token_accuracy": 0.9051599502563477,
      "num_tokens": 158110609.0,
      "step": 68
    },
    {
      "epoch": 0.035312180143295804,
      "grad_norm": 0.216796875,
      "learning_rate": 3.999777383024174e-05,
      "loss": 0.2393469214439392,
      "mean_token_accuracy": 0.9063411951065063,
      "num_tokens": 160440227.0,
      "step": 69
    },
    {
      "epoch": 0.03582395087001024,
      "grad_norm": 0.22265625,
      "learning_rate": 3.9997251654233656e-05,
      "loss": 0.24396447837352753,
      "mean_token_accuracy": 0.9038181900978088,
      "num_tokens": 162748188.0,
      "step": 70
    },
    {
      "epoch": 0.036335721596724664,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.999667451761746e-05,
      "loss": 0.24223962426185608,
      "mean_token_accuracy": 0.9045360088348389,
      "num_tokens": 165058400.0,
      "step": 71
    },
    {
      "epoch": 0.0368474923234391,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.9996042421979376e-05,
      "loss": 0.24551153182983398,
      "mean_token_accuracy": 0.9035066366195679,
      "num_tokens": 167364428.0,
      "step": 72
    },
    {
      "epoch": 0.03735926305015353,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.999535536905663e-05,
      "loss": 0.2292848825454712,
      "mean_token_accuracy": 0.9096421003341675,
      "num_tokens": 169707082.0,
      "step": 73
    },
    {
      "epoch": 0.037871033776867964,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.999461336073755e-05,
      "loss": 0.24630841612815857,
      "mean_token_accuracy": 0.9031414985656738,
      "num_tokens": 172003774.0,
      "step": 74
    },
    {
      "epoch": 0.0383828045035824,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.9993816399061474e-05,
      "loss": 0.2562379240989685,
      "mean_token_accuracy": 0.8988590240478516,
      "num_tokens": 174388310.0,
      "step": 75
    },
    {
      "epoch": 0.038894575230296824,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.999296448621877e-05,
      "loss": 0.23270204663276672,
      "mean_token_accuracy": 0.9078875184059143,
      "num_tokens": 176648737.0,
      "step": 76
    },
    {
      "epoch": 0.03940634595701126,
      "grad_norm": 0.54296875,
      "learning_rate": 3.9992057624550835e-05,
      "loss": 0.24843892455101013,
      "mean_token_accuracy": 0.9016368389129639,
      "num_tokens": 178980169.0,
      "step": 77
    },
    {
      "epoch": 0.03991811668372569,
      "grad_norm": 0.220703125,
      "learning_rate": 3.999109581655011e-05,
      "loss": 0.25261902809143066,
      "mean_token_accuracy": 0.900001049041748,
      "num_tokens": 181321193.0,
      "step": 78
    },
    {
      "epoch": 0.040429887410440124,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.9990079064860035e-05,
      "loss": 0.25579142570495605,
      "mean_token_accuracy": 0.8987343311309814,
      "num_tokens": 183588553.0,
      "step": 79
    },
    {
      "epoch": 0.04094165813715456,
      "grad_norm": 0.2109375,
      "learning_rate": 3.998900737227505e-05,
      "loss": 0.23632150888442993,
      "mean_token_accuracy": 0.9070842266082764,
      "num_tokens": 185910143.0,
      "step": 80
    },
    {
      "epoch": 0.041453428863868984,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.9987880741740615e-05,
      "loss": 0.23823469877243042,
      "mean_token_accuracy": 0.9051164388656616,
      "num_tokens": 188218412.0,
      "step": 81
    },
    {
      "epoch": 0.04196519959058342,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.998669917635315e-05,
      "loss": 0.24175620079040527,
      "mean_token_accuracy": 0.9043114185333252,
      "num_tokens": 190552553.0,
      "step": 82
    },
    {
      "epoch": 0.04247697031729785,
      "grad_norm": 0.21484375,
      "learning_rate": 3.99854626793601e-05,
      "loss": 0.23264633119106293,
      "mean_token_accuracy": 0.9081540107727051,
      "num_tokens": 192804452.0,
      "step": 83
    },
    {
      "epoch": 0.042988741044012284,
      "grad_norm": 0.255859375,
      "learning_rate": 3.9984171254159856e-05,
      "loss": 0.24963194131851196,
      "mean_token_accuracy": 0.9016275405883789,
      "num_tokens": 195022368.0,
      "step": 84
    },
    {
      "epoch": 0.04350051177072672,
      "grad_norm": 0.220703125,
      "learning_rate": 3.998282490430178e-05,
      "loss": 0.24411894381046295,
      "mean_token_accuracy": 0.9035032987594604,
      "num_tokens": 197385945.0,
      "step": 85
    },
    {
      "epoch": 0.044012282497441144,
      "grad_norm": 0.2265625,
      "learning_rate": 3.998142363348619e-05,
      "loss": 0.25204041600227356,
      "mean_token_accuracy": 0.9000582695007324,
      "num_tokens": 199729117.0,
      "step": 86
    },
    {
      "epoch": 0.04452405322415558,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.997996744556435e-05,
      "loss": 0.2458759844303131,
      "mean_token_accuracy": 0.9026727676391602,
      "num_tokens": 201978227.0,
      "step": 87
    },
    {
      "epoch": 0.04503582395087001,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.997845634453846e-05,
      "loss": 0.2460886538028717,
      "mean_token_accuracy": 0.9026058316230774,
      "num_tokens": 204270340.0,
      "step": 88
    },
    {
      "epoch": 0.045547594677584444,
      "grad_norm": 0.236328125,
      "learning_rate": 3.997689033456165e-05,
      "loss": 0.2375563383102417,
      "mean_token_accuracy": 0.9055942296981812,
      "num_tokens": 206538425.0,
      "step": 89
    },
    {
      "epoch": 0.04605936540429888,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.997526941993794e-05,
      "loss": 0.24444015324115753,
      "mean_token_accuracy": 0.9034237861633301,
      "num_tokens": 208848710.0,
      "step": 90
    },
    {
      "epoch": 0.0465711361310133,
      "grad_norm": 0.228515625,
      "learning_rate": 3.997359360512227e-05,
      "loss": 0.2320779711008072,
      "mean_token_accuracy": 0.9084529876708984,
      "num_tokens": 211154342.0,
      "step": 91
    },
    {
      "epoch": 0.04708290685772774,
      "grad_norm": 0.21875,
      "learning_rate": 3.997186289472047e-05,
      "loss": 0.2342875897884369,
      "mean_token_accuracy": 0.9067320823669434,
      "num_tokens": 213491261.0,
      "step": 92
    },
    {
      "epoch": 0.04759467758444217,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.997007729348923e-05,
      "loss": 0.24909600615501404,
      "mean_token_accuracy": 0.9017957448959351,
      "num_tokens": 215781690.0,
      "step": 93
    },
    {
      "epoch": 0.048106448311156604,
      "grad_norm": 0.23046875,
      "learning_rate": 3.996823680633612e-05,
      "loss": 0.23682336509227753,
      "mean_token_accuracy": 0.9051960706710815,
      "num_tokens": 218120952.0,
      "step": 94
    },
    {
      "epoch": 0.04861821903787104,
      "grad_norm": 0.22265625,
      "learning_rate": 3.996634143831954e-05,
      "loss": 0.2347905933856964,
      "mean_token_accuracy": 0.9073508977890015,
      "num_tokens": 220528355.0,
      "step": 95
    },
    {
      "epoch": 0.04912998976458546,
      "grad_norm": 0.228515625,
      "learning_rate": 3.9964391194648736e-05,
      "loss": 0.24354062974452972,
      "mean_token_accuracy": 0.9036295413970947,
      "num_tokens": 222884211.0,
      "step": 96
    },
    {
      "epoch": 0.0496417604912999,
      "grad_norm": 0.21875,
      "learning_rate": 3.996238608068378e-05,
      "loss": 0.24138852953910828,
      "mean_token_accuracy": 0.9039537906646729,
      "num_tokens": 225280810.0,
      "step": 97
    },
    {
      "epoch": 0.05015353121801433,
      "grad_norm": 0.23046875,
      "learning_rate": 3.996032610193554e-05,
      "loss": 0.24951758980751038,
      "mean_token_accuracy": 0.9016480445861816,
      "num_tokens": 227646916.0,
      "step": 98
    },
    {
      "epoch": 0.05066530194472876,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.9958211264065675e-05,
      "loss": 0.23976950347423553,
      "mean_token_accuracy": 0.9051105380058289,
      "num_tokens": 229923315.0,
      "step": 99
    },
    {
      "epoch": 0.0511770726714432,
      "grad_norm": 0.22265625,
      "learning_rate": 3.995604157288663e-05,
      "loss": 0.23906734585762024,
      "mean_token_accuracy": 0.9047994613647461,
      "num_tokens": 232279923.0,
      "step": 100
    },
    {
      "epoch": 0.05168884339815762,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.995381703436159e-05,
      "loss": 0.23211261630058289,
      "mean_token_accuracy": 0.9082344174385071,
      "num_tokens": 234629507.0,
      "step": 101
    },
    {
      "epoch": 0.052200614124872056,
      "grad_norm": 0.23828125,
      "learning_rate": 3.9951537654604504e-05,
      "loss": 0.245839923620224,
      "mean_token_accuracy": 0.9024314880371094,
      "num_tokens": 236918155.0,
      "step": 102
    },
    {
      "epoch": 0.05271238485158649,
      "grad_norm": 0.23046875,
      "learning_rate": 3.994920343988004e-05,
      "loss": 0.24258410930633545,
      "mean_token_accuracy": 0.9043809175491333,
      "num_tokens": 239327994.0,
      "step": 103
    },
    {
      "epoch": 0.05322415557830092,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.994681439660357e-05,
      "loss": 0.25015968084335327,
      "mean_token_accuracy": 0.9001978635787964,
      "num_tokens": 241642381.0,
      "step": 104
    },
    {
      "epoch": 0.05373592630501536,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.994437053134115e-05,
      "loss": 0.24693500995635986,
      "mean_token_accuracy": 0.9024827480316162,
      "num_tokens": 243958762.0,
      "step": 105
    },
    {
      "epoch": 0.05424769703172978,
      "grad_norm": 0.240234375,
      "learning_rate": 3.994187185080953e-05,
      "loss": 0.2501673102378845,
      "mean_token_accuracy": 0.9010171890258789,
      "num_tokens": 246272153.0,
      "step": 106
    },
    {
      "epoch": 0.054759467758444216,
      "grad_norm": 0.25,
      "learning_rate": 3.99393183618761e-05,
      "loss": 0.24285846948623657,
      "mean_token_accuracy": 0.9041951894760132,
      "num_tokens": 248681747.0,
      "step": 107
    },
    {
      "epoch": 0.05527123848515865,
      "grad_norm": 0.228515625,
      "learning_rate": 3.9936710071558896e-05,
      "loss": 0.23713770508766174,
      "mean_token_accuracy": 0.9055157899856567,
      "num_tokens": 250983115.0,
      "step": 108
    },
    {
      "epoch": 0.05578300921187308,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.993404698702655e-05,
      "loss": 0.2447194755077362,
      "mean_token_accuracy": 0.9031585454940796,
      "num_tokens": 253279515.0,
      "step": 109
    },
    {
      "epoch": 0.05629477993858751,
      "grad_norm": 0.228515625,
      "learning_rate": 3.993132911559833e-05,
      "loss": 0.24290750920772552,
      "mean_token_accuracy": 0.9037627577781677,
      "num_tokens": 255669018.0,
      "step": 110
    },
    {
      "epoch": 0.05680655066530194,
      "grad_norm": 0.220703125,
      "learning_rate": 3.992855646474402e-05,
      "loss": 0.22462835907936096,
      "mean_token_accuracy": 0.9111753106117249,
      "num_tokens": 258013902.0,
      "step": 111
    },
    {
      "epoch": 0.057318321392016376,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.992572904208403e-05,
      "loss": 0.24201668798923492,
      "mean_token_accuracy": 0.904655396938324,
      "num_tokens": 260416625.0,
      "step": 112
    },
    {
      "epoch": 0.05783009211873081,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.992284685538924e-05,
      "loss": 0.25579050183296204,
      "mean_token_accuracy": 0.8983302116394043,
      "num_tokens": 262734660.0,
      "step": 113
    },
    {
      "epoch": 0.05834186284544524,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.991990991258111e-05,
      "loss": 0.24900555610656738,
      "mean_token_accuracy": 0.9004928469657898,
      "num_tokens": 265042510.0,
      "step": 114
    },
    {
      "epoch": 0.05885363357215967,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.9916918221731535e-05,
      "loss": 0.2456320971250534,
      "mean_token_accuracy": 0.9025330543518066,
      "num_tokens": 267321620.0,
      "step": 115
    },
    {
      "epoch": 0.0593654042988741,
      "grad_norm": 0.2265625,
      "learning_rate": 3.99138717910629e-05,
      "loss": 0.2373812198638916,
      "mean_token_accuracy": 0.9067170023918152,
      "num_tokens": 269669727.0,
      "step": 116
    },
    {
      "epoch": 0.059877175025588536,
      "grad_norm": 0.220703125,
      "learning_rate": 3.991077062894805e-05,
      "loss": 0.245025634765625,
      "mean_token_accuracy": 0.9023817181587219,
      "num_tokens": 271897855.0,
      "step": 117
    },
    {
      "epoch": 0.06038894575230297,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.990761474391024e-05,
      "loss": 0.24305002391338348,
      "mean_token_accuracy": 0.9033375978469849,
      "num_tokens": 274186684.0,
      "step": 118
    },
    {
      "epoch": 0.0609007164790174,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.990440414462313e-05,
      "loss": 0.23775458335876465,
      "mean_token_accuracy": 0.9058787822723389,
      "num_tokens": 276472922.0,
      "step": 119
    },
    {
      "epoch": 0.06141248720573183,
      "grad_norm": 0.236328125,
      "learning_rate": 3.990113883991076e-05,
      "loss": 0.2594411373138428,
      "mean_token_accuracy": 0.8971904516220093,
      "num_tokens": 278711264.0,
      "step": 120
    },
    {
      "epoch": 0.06192425793244626,
      "grad_norm": 0.220703125,
      "learning_rate": 3.989781883874752e-05,
      "loss": 0.2472670078277588,
      "mean_token_accuracy": 0.9024604558944702,
      "num_tokens": 281050658.0,
      "step": 121
    },
    {
      "epoch": 0.062436028659160696,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.989444415025812e-05,
      "loss": 0.2575949728488922,
      "mean_token_accuracy": 0.8981107473373413,
      "num_tokens": 283353984.0,
      "step": 122
    },
    {
      "epoch": 0.06294779938587512,
      "grad_norm": 0.234375,
      "learning_rate": 3.9891014783717587e-05,
      "loss": 0.24273017048835754,
      "mean_token_accuracy": 0.9032453298568726,
      "num_tokens": 285686229.0,
      "step": 123
    },
    {
      "epoch": 0.06345957011258956,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.988753074855122e-05,
      "loss": 0.26370328664779663,
      "mean_token_accuracy": 0.8952413201332092,
      "num_tokens": 288089790.0,
      "step": 124
    },
    {
      "epoch": 0.06397134083930399,
      "grad_norm": 0.21484375,
      "learning_rate": 3.988399205433457e-05,
      "loss": 0.23635610938072205,
      "mean_token_accuracy": 0.9053285717964172,
      "num_tokens": 290449478.0,
      "step": 125
    },
    {
      "epoch": 0.06448311156601842,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.988039871079341e-05,
      "loss": 0.22739923000335693,
      "mean_token_accuracy": 0.9096212387084961,
      "num_tokens": 292775572.0,
      "step": 126
    },
    {
      "epoch": 0.06499488229273286,
      "grad_norm": 0.2109375,
      "learning_rate": 3.9876750727803716e-05,
      "loss": 0.23607537150382996,
      "mean_token_accuracy": 0.9054372310638428,
      "num_tokens": 295075567.0,
      "step": 127
    },
    {
      "epoch": 0.06550665301944729,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.987304811539163e-05,
      "loss": 0.2510735094547272,
      "mean_token_accuracy": 0.9006474018096924,
      "num_tokens": 297427349.0,
      "step": 128
    },
    {
      "epoch": 0.06601842374616172,
      "grad_norm": 0.224609375,
      "learning_rate": 3.9869290883733456e-05,
      "loss": 0.23523449897766113,
      "mean_token_accuracy": 0.9075653553009033,
      "num_tokens": 299685847.0,
      "step": 129
    },
    {
      "epoch": 0.06653019447287616,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.986547904315558e-05,
      "loss": 0.24752739071846008,
      "mean_token_accuracy": 0.9018803238868713,
      "num_tokens": 302002986.0,
      "step": 130
    },
    {
      "epoch": 0.06704196519959059,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.986161260413451e-05,
      "loss": 0.25192034244537354,
      "mean_token_accuracy": 0.8999905586242676,
      "num_tokens": 304298748.0,
      "step": 131
    },
    {
      "epoch": 0.06755373592630501,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.985769157729679e-05,
      "loss": 0.24674037098884583,
      "mean_token_accuracy": 0.9028269052505493,
      "num_tokens": 306574575.0,
      "step": 132
    },
    {
      "epoch": 0.06806550665301944,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.9853715973419006e-05,
      "loss": 0.24268773198127747,
      "mean_token_accuracy": 0.9040877819061279,
      "num_tokens": 308923493.0,
      "step": 133
    },
    {
      "epoch": 0.06857727737973388,
      "grad_norm": 0.228515625,
      "learning_rate": 3.984968580342774e-05,
      "loss": 0.24677950143814087,
      "mean_token_accuracy": 0.9014135599136353,
      "num_tokens": 311199791.0,
      "step": 134
    },
    {
      "epoch": 0.06908904810644831,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.984560107839954e-05,
      "loss": 0.24123777449131012,
      "mean_token_accuracy": 0.9045393466949463,
      "num_tokens": 313515755.0,
      "step": 135
    },
    {
      "epoch": 0.06960081883316274,
      "grad_norm": 0.255859375,
      "learning_rate": 3.98414618095609e-05,
      "loss": 0.24962928891181946,
      "mean_token_accuracy": 0.8996661305427551,
      "num_tokens": 315848042.0,
      "step": 136
    },
    {
      "epoch": 0.07011258955987718,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.983726800828823e-05,
      "loss": 0.23512694239616394,
      "mean_token_accuracy": 0.9060874581336975,
      "num_tokens": 318239641.0,
      "step": 137
    },
    {
      "epoch": 0.07062436028659161,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.983301968610779e-05,
      "loss": 0.2494986206293106,
      "mean_token_accuracy": 0.9015611410140991,
      "num_tokens": 320617945.0,
      "step": 138
    },
    {
      "epoch": 0.07113613101330604,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.9828716854695706e-05,
      "loss": 0.23691149055957794,
      "mean_token_accuracy": 0.9062352180480957,
      "num_tokens": 322975099.0,
      "step": 139
    },
    {
      "epoch": 0.07164790174002048,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.9824359525877924e-05,
      "loss": 0.24398863315582275,
      "mean_token_accuracy": 0.9035800695419312,
      "num_tokens": 325291081.0,
      "step": 140
    },
    {
      "epoch": 0.07215967246673491,
      "grad_norm": 0.2451171875,
      "learning_rate": 3.981994771163016e-05,
      "loss": 0.2516748309135437,
      "mean_token_accuracy": 0.8992185592651367,
      "num_tokens": 327597560.0,
      "step": 141
    },
    {
      "epoch": 0.07267144319344933,
      "grad_norm": 0.23046875,
      "learning_rate": 3.981548142407787e-05,
      "loss": 0.24472585320472717,
      "mean_token_accuracy": 0.9026576280593872,
      "num_tokens": 329843577.0,
      "step": 142
    },
    {
      "epoch": 0.07318321392016376,
      "grad_norm": 0.212890625,
      "learning_rate": 3.981096067549624e-05,
      "loss": 0.2427264153957367,
      "mean_token_accuracy": 0.9044903516769409,
      "num_tokens": 332212626.0,
      "step": 143
    },
    {
      "epoch": 0.0736949846468782,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.980638547831014e-05,
      "loss": 0.24787260591983795,
      "mean_token_accuracy": 0.9018590450286865,
      "num_tokens": 334577137.0,
      "step": 144
    },
    {
      "epoch": 0.07420675537359263,
      "grad_norm": 0.21875,
      "learning_rate": 3.980175584509407e-05,
      "loss": 0.24098140001296997,
      "mean_token_accuracy": 0.9035793542861938,
      "num_tokens": 336989127.0,
      "step": 145
    },
    {
      "epoch": 0.07471852610030706,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.979707178857215e-05,
      "loss": 0.26297232508659363,
      "mean_token_accuracy": 0.8967742919921875,
      "num_tokens": 339310698.0,
      "step": 146
    },
    {
      "epoch": 0.0752302968270215,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.979233332161809e-05,
      "loss": 0.23618894815444946,
      "mean_token_accuracy": 0.906849205493927,
      "num_tokens": 341633693.0,
      "step": 147
    },
    {
      "epoch": 0.07574206755373593,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.978754045725512e-05,
      "loss": 0.23620736598968506,
      "mean_token_accuracy": 0.9067492485046387,
      "num_tokens": 344048260.0,
      "step": 148
    },
    {
      "epoch": 0.07625383828045036,
      "grad_norm": 0.234375,
      "learning_rate": 3.978269320865599e-05,
      "loss": 0.24367862939834595,
      "mean_token_accuracy": 0.9038439393043518,
      "num_tokens": 346291448.0,
      "step": 149
    },
    {
      "epoch": 0.0767656090071648,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.977779158914292e-05,
      "loss": 0.24385759234428406,
      "mean_token_accuracy": 0.9024319648742676,
      "num_tokens": 348689796.0,
      "step": 150
    },
    {
      "epoch": 0.07727737973387923,
      "grad_norm": 0.2265625,
      "learning_rate": 3.9772835612187556e-05,
      "loss": 0.2407391369342804,
      "mean_token_accuracy": 0.9042386412620544,
      "num_tokens": 350978595.0,
      "step": 151
    },
    {
      "epoch": 0.07778915046059365,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.976782529141094e-05,
      "loss": 0.23678825795650482,
      "mean_token_accuracy": 0.9058905243873596,
      "num_tokens": 353255328.0,
      "step": 152
    },
    {
      "epoch": 0.07830092118730808,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.9762760640583495e-05,
      "loss": 0.23502452671527863,
      "mean_token_accuracy": 0.9063423871994019,
      "num_tokens": 355605843.0,
      "step": 153
    },
    {
      "epoch": 0.07881269191402251,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.975764167362494e-05,
      "loss": 0.2486424446105957,
      "mean_token_accuracy": 0.9014183878898621,
      "num_tokens": 357944973.0,
      "step": 154
    },
    {
      "epoch": 0.07932446264073695,
      "grad_norm": 0.208984375,
      "learning_rate": 3.975246840460428e-05,
      "loss": 0.22794106602668762,
      "mean_token_accuracy": 0.9088218212127686,
      "num_tokens": 360277043.0,
      "step": 155
    },
    {
      "epoch": 0.07983623336745138,
      "grad_norm": 0.228515625,
      "learning_rate": 3.974724084773977e-05,
      "loss": 0.2334713488817215,
      "mean_token_accuracy": 0.9075926542282104,
      "num_tokens": 362609833.0,
      "step": 156
    },
    {
      "epoch": 0.08034800409416581,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.9741959017398885e-05,
      "loss": 0.24186921119689941,
      "mean_token_accuracy": 0.9038417339324951,
      "num_tokens": 364983170.0,
      "step": 157
    },
    {
      "epoch": 0.08085977482088025,
      "grad_norm": 0.22265625,
      "learning_rate": 3.973662292809823e-05,
      "loss": 0.2411707043647766,
      "mean_token_accuracy": 0.9033752679824829,
      "num_tokens": 367278433.0,
      "step": 158
    },
    {
      "epoch": 0.08137154554759468,
      "grad_norm": 0.232421875,
      "learning_rate": 3.973123259450357e-05,
      "loss": 0.2413295954465866,
      "mean_token_accuracy": 0.9041519165039062,
      "num_tokens": 369591001.0,
      "step": 159
    },
    {
      "epoch": 0.08188331627430911,
      "grad_norm": 0.236328125,
      "learning_rate": 3.972578803142973e-05,
      "loss": 0.24333271384239197,
      "mean_token_accuracy": 0.9042398929595947,
      "num_tokens": 371839586.0,
      "step": 160
    },
    {
      "epoch": 0.08239508700102355,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.9720289253840605e-05,
      "loss": 0.23233577609062195,
      "mean_token_accuracy": 0.9080957174301147,
      "num_tokens": 374123581.0,
      "step": 161
    },
    {
      "epoch": 0.08290685772773797,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.971473627684906e-05,
      "loss": 0.2532513439655304,
      "mean_token_accuracy": 0.8993710875511169,
      "num_tokens": 376422053.0,
      "step": 162
    },
    {
      "epoch": 0.0834186284544524,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.970912911571696e-05,
      "loss": 0.24000269174575806,
      "mean_token_accuracy": 0.9042823314666748,
      "num_tokens": 378628808.0,
      "step": 163
    },
    {
      "epoch": 0.08393039918116683,
      "grad_norm": 0.23828125,
      "learning_rate": 3.9703467785855064e-05,
      "loss": 0.25049230456352234,
      "mean_token_accuracy": 0.9007202386856079,
      "num_tokens": 380902727.0,
      "step": 164
    },
    {
      "epoch": 0.08444216990788127,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.969775230282302e-05,
      "loss": 0.2438296526670456,
      "mean_token_accuracy": 0.9033468961715698,
      "num_tokens": 383141482.0,
      "step": 165
    },
    {
      "epoch": 0.0849539406345957,
      "grad_norm": 0.2109375,
      "learning_rate": 3.9691982682329304e-05,
      "loss": 0.2399730682373047,
      "mean_token_accuracy": 0.9052138328552246,
      "num_tokens": 385530301.0,
      "step": 166
    },
    {
      "epoch": 0.08546571136131013,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.9686158940231194e-05,
      "loss": 0.2350412756204605,
      "mean_token_accuracy": 0.9067797064781189,
      "num_tokens": 387833601.0,
      "step": 167
    },
    {
      "epoch": 0.08597748208802457,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.96802810925347e-05,
      "loss": 0.2500113844871521,
      "mean_token_accuracy": 0.9005874395370483,
      "num_tokens": 390090536.0,
      "step": 168
    },
    {
      "epoch": 0.086489252814739,
      "grad_norm": 0.220703125,
      "learning_rate": 3.967434915539455e-05,
      "loss": 0.24514105916023254,
      "mean_token_accuracy": 0.9026144742965698,
      "num_tokens": 392416689.0,
      "step": 169
    },
    {
      "epoch": 0.08700102354145343,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.966836314511413e-05,
      "loss": 0.23948493599891663,
      "mean_token_accuracy": 0.9041406512260437,
      "num_tokens": 394753087.0,
      "step": 170
    },
    {
      "epoch": 0.08751279426816785,
      "grad_norm": 0.228515625,
      "learning_rate": 3.966232307814545e-05,
      "loss": 0.24536272883415222,
      "mean_token_accuracy": 0.9028125405311584,
      "num_tokens": 397110347.0,
      "step": 171
    },
    {
      "epoch": 0.08802456499488229,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.9656228971089066e-05,
      "loss": 0.2509254217147827,
      "mean_token_accuracy": 0.900165319442749,
      "num_tokens": 399392741.0,
      "step": 172
    },
    {
      "epoch": 0.08853633572159672,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.965008084069406e-05,
      "loss": 0.24496576189994812,
      "mean_token_accuracy": 0.9026360511779785,
      "num_tokens": 401713028.0,
      "step": 173
    },
    {
      "epoch": 0.08904810644831115,
      "grad_norm": 0.23828125,
      "learning_rate": 3.964387870385802e-05,
      "loss": 0.2560853362083435,
      "mean_token_accuracy": 0.8977952003479004,
      "num_tokens": 403976883.0,
      "step": 174
    },
    {
      "epoch": 0.08955987717502559,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.963762257762695e-05,
      "loss": 0.25204896926879883,
      "mean_token_accuracy": 0.9003223180770874,
      "num_tokens": 406266242.0,
      "step": 175
    },
    {
      "epoch": 0.09007164790174002,
      "grad_norm": 0.2265625,
      "learning_rate": 3.963131247919523e-05,
      "loss": 0.2376379370689392,
      "mean_token_accuracy": 0.9049273133277893,
      "num_tokens": 408649860.0,
      "step": 176
    },
    {
      "epoch": 0.09058341862845445,
      "grad_norm": 0.220703125,
      "learning_rate": 3.9624948425905586e-05,
      "loss": 0.24343842267990112,
      "mean_token_accuracy": 0.9033228158950806,
      "num_tokens": 410957173.0,
      "step": 177
    },
    {
      "epoch": 0.09109518935516889,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.961853043524905e-05,
      "loss": 0.24352236092090607,
      "mean_token_accuracy": 0.9031516313552856,
      "num_tokens": 413361883.0,
      "step": 178
    },
    {
      "epoch": 0.09160696008188332,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.961205852486485e-05,
      "loss": 0.23091845214366913,
      "mean_token_accuracy": 0.9076954126358032,
      "num_tokens": 415622840.0,
      "step": 179
    },
    {
      "epoch": 0.09211873080859775,
      "grad_norm": 0.234375,
      "learning_rate": 3.960553271254045e-05,
      "loss": 0.24649059772491455,
      "mean_token_accuracy": 0.9018585681915283,
      "num_tokens": 417926120.0,
      "step": 180
    },
    {
      "epoch": 0.09263050153531217,
      "grad_norm": 0.208984375,
      "learning_rate": 3.9598953016211455e-05,
      "loss": 0.23450930416584015,
      "mean_token_accuracy": 0.9068349003791809,
      "num_tokens": 420174777.0,
      "step": 181
    },
    {
      "epoch": 0.0931422722620266,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.959231945396156e-05,
      "loss": 0.24614956974983215,
      "mean_token_accuracy": 0.9020057320594788,
      "num_tokens": 422554503.0,
      "step": 182
    },
    {
      "epoch": 0.09365404298874104,
      "grad_norm": 0.232421875,
      "learning_rate": 3.958563204402247e-05,
      "loss": 0.24184201657772064,
      "mean_token_accuracy": 0.9030746817588806,
      "num_tokens": 424907943.0,
      "step": 183
    },
    {
      "epoch": 0.09416581371545547,
      "grad_norm": 0.2470703125,
      "learning_rate": 3.957889080477394e-05,
      "loss": 0.2645432651042938,
      "mean_token_accuracy": 0.8945055603981018,
      "num_tokens": 427230853.0,
      "step": 184
    },
    {
      "epoch": 0.09467758444216991,
      "grad_norm": 0.216796875,
      "learning_rate": 3.9572095754743646e-05,
      "loss": 0.2329862117767334,
      "mean_token_accuracy": 0.9077844023704529,
      "num_tokens": 429590198.0,
      "step": 185
    },
    {
      "epoch": 0.09518935516888434,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.956524691260715e-05,
      "loss": 0.23751172423362732,
      "mean_token_accuracy": 0.9065521359443665,
      "num_tokens": 431967770.0,
      "step": 186
    },
    {
      "epoch": 0.09570112589559877,
      "grad_norm": 0.224609375,
      "learning_rate": 3.9558344297187877e-05,
      "loss": 0.24001538753509521,
      "mean_token_accuracy": 0.9043587446212769,
      "num_tokens": 434347305.0,
      "step": 187
    },
    {
      "epoch": 0.09621289662231321,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.955138792745702e-05,
      "loss": 0.23923087120056152,
      "mean_token_accuracy": 0.9045559167861938,
      "num_tokens": 436708529.0,
      "step": 188
    },
    {
      "epoch": 0.09672466734902764,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.9544377822533526e-05,
      "loss": 0.24397693574428558,
      "mean_token_accuracy": 0.9028058052062988,
      "num_tokens": 438996292.0,
      "step": 189
    },
    {
      "epoch": 0.09723643807574207,
      "grad_norm": 0.224609375,
      "learning_rate": 3.953731400168402e-05,
      "loss": 0.24463525414466858,
      "mean_token_accuracy": 0.9030151963233948,
      "num_tokens": 441362134.0,
      "step": 190
    },
    {
      "epoch": 0.0977482088024565,
      "grad_norm": 0.234375,
      "learning_rate": 3.953019648432277e-05,
      "loss": 0.2523895502090454,
      "mean_token_accuracy": 0.9001028537750244,
      "num_tokens": 443651229.0,
      "step": 191
    },
    {
      "epoch": 0.09825997952917093,
      "grad_norm": 0.2490234375,
      "learning_rate": 3.9523025290011615e-05,
      "loss": 0.2461979240179062,
      "mean_token_accuracy": 0.9017066955566406,
      "num_tokens": 445892243.0,
      "step": 192
    },
    {
      "epoch": 0.09877175025588536,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.951580043845993e-05,
      "loss": 0.24612081050872803,
      "mean_token_accuracy": 0.9023141860961914,
      "num_tokens": 448210886.0,
      "step": 193
    },
    {
      "epoch": 0.0992835209825998,
      "grad_norm": 0.2265625,
      "learning_rate": 3.950852194952455e-05,
      "loss": 0.24193325638771057,
      "mean_token_accuracy": 0.9037947654724121,
      "num_tokens": 450581857.0,
      "step": 194
    },
    {
      "epoch": 0.09979529170931423,
      "grad_norm": 0.3046875,
      "learning_rate": 3.950118984320972e-05,
      "loss": 0.23432588577270508,
      "mean_token_accuracy": 0.9069265723228455,
      "num_tokens": 452886151.0,
      "step": 195
    },
    {
      "epoch": 0.10030706243602866,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.949380413966709e-05,
      "loss": 0.23968566954135895,
      "mean_token_accuracy": 0.9041038751602173,
      "num_tokens": 455248081.0,
      "step": 196
    },
    {
      "epoch": 0.1008188331627431,
      "grad_norm": 0.2421875,
      "learning_rate": 3.948636485919556e-05,
      "loss": 0.25988954305648804,
      "mean_token_accuracy": 0.8967081308364868,
      "num_tokens": 457528984.0,
      "step": 197
    },
    {
      "epoch": 0.10133060388945753,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.9478872022241326e-05,
      "loss": 0.24214377999305725,
      "mean_token_accuracy": 0.904679536819458,
      "num_tokens": 459875862.0,
      "step": 198
    },
    {
      "epoch": 0.10184237461617196,
      "grad_norm": 0.220703125,
      "learning_rate": 3.947132564939775e-05,
      "loss": 0.23286911845207214,
      "mean_token_accuracy": 0.9072979688644409,
      "num_tokens": 462239646.0,
      "step": 199
    },
    {
      "epoch": 0.1023541453428864,
      "grad_norm": 0.21875,
      "learning_rate": 3.946372576140535e-05,
      "loss": 0.23759622871875763,
      "mean_token_accuracy": 0.9054461717605591,
      "num_tokens": 464560367.0,
      "step": 200
    },
    {
      "epoch": 0.10286591606960081,
      "grad_norm": 0.2265625,
      "learning_rate": 3.945607237915172e-05,
      "loss": 0.24943985044956207,
      "mean_token_accuracy": 0.9012597799301147,
      "num_tokens": 466872301.0,
      "step": 201
    },
    {
      "epoch": 0.10337768679631525,
      "grad_norm": 0.24609375,
      "learning_rate": 3.944836552367147e-05,
      "loss": 0.24607722461223602,
      "mean_token_accuracy": 0.9016529321670532,
      "num_tokens": 469176731.0,
      "step": 202
    },
    {
      "epoch": 0.10388945752302968,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.9440605216146196e-05,
      "loss": 0.24603869020938873,
      "mean_token_accuracy": 0.9023638963699341,
      "num_tokens": 471494197.0,
      "step": 203
    },
    {
      "epoch": 0.10440122824974411,
      "grad_norm": 0.2265625,
      "learning_rate": 3.9432791477904386e-05,
      "loss": 0.2362613081932068,
      "mean_token_accuracy": 0.9055600166320801,
      "num_tokens": 473911852.0,
      "step": 204
    },
    {
      "epoch": 0.10491299897645855,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.9424924330421374e-05,
      "loss": 0.24371454119682312,
      "mean_token_accuracy": 0.9037792682647705,
      "num_tokens": 476201068.0,
      "step": 205
    },
    {
      "epoch": 0.10542476970317298,
      "grad_norm": 0.2265625,
      "learning_rate": 3.9417003795319296e-05,
      "loss": 0.24258068203926086,
      "mean_token_accuracy": 0.9032790660858154,
      "num_tokens": 478541287.0,
      "step": 206
    },
    {
      "epoch": 0.10593654042988741,
      "grad_norm": 0.224609375,
      "learning_rate": 3.940902989436702e-05,
      "loss": 0.23777678608894348,
      "mean_token_accuracy": 0.906508207321167,
      "num_tokens": 480958439.0,
      "step": 207
    },
    {
      "epoch": 0.10644831115660185,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.940100264948007e-05,
      "loss": 0.24303403496742249,
      "mean_token_accuracy": 0.9037009477615356,
      "num_tokens": 483300322.0,
      "step": 208
    },
    {
      "epoch": 0.10696008188331628,
      "grad_norm": 0.2421875,
      "learning_rate": 3.93929220827206e-05,
      "loss": 0.24931487441062927,
      "mean_token_accuracy": 0.9011132121086121,
      "num_tokens": 485632610.0,
      "step": 209
    },
    {
      "epoch": 0.10747185261003071,
      "grad_norm": 0.23046875,
      "learning_rate": 3.93847882162973e-05,
      "loss": 0.23398777842521667,
      "mean_token_accuracy": 0.9063008427619934,
      "num_tokens": 487880094.0,
      "step": 210
    },
    {
      "epoch": 0.10798362333674513,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.9376601072565354e-05,
      "loss": 0.25359076261520386,
      "mean_token_accuracy": 0.8982406854629517,
      "num_tokens": 490200798.0,
      "step": 211
    },
    {
      "epoch": 0.10849539406345957,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.9368360674026384e-05,
      "loss": 0.2520386576652527,
      "mean_token_accuracy": 0.899967610836029,
      "num_tokens": 492452806.0,
      "step": 212
    },
    {
      "epoch": 0.109007164790174,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.936006704332836e-05,
      "loss": 0.23607973754405975,
      "mean_token_accuracy": 0.905978798866272,
      "num_tokens": 494737472.0,
      "step": 213
    },
    {
      "epoch": 0.10951893551688843,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.935172020326555e-05,
      "loss": 0.26264360547065735,
      "mean_token_accuracy": 0.8960838913917542,
      "num_tokens": 497069270.0,
      "step": 214
    },
    {
      "epoch": 0.11003070624360287,
      "grad_norm": 0.21875,
      "learning_rate": 3.93433201767785e-05,
      "loss": 0.23800143599510193,
      "mean_token_accuracy": 0.9048517942428589,
      "num_tokens": 499443976.0,
      "step": 215
    },
    {
      "epoch": 0.1105424769703173,
      "grad_norm": 0.228515625,
      "learning_rate": 3.9334866986953895e-05,
      "loss": 0.23571544885635376,
      "mean_token_accuracy": 0.906517744064331,
      "num_tokens": 501959497.0,
      "step": 216
    },
    {
      "epoch": 0.11105424769703173,
      "grad_norm": 0.2265625,
      "learning_rate": 3.932636065702456e-05,
      "loss": 0.25068461894989014,
      "mean_token_accuracy": 0.899833619594574,
      "num_tokens": 504277966.0,
      "step": 217
    },
    {
      "epoch": 0.11156601842374617,
      "grad_norm": 0.2265625,
      "learning_rate": 3.9317801210369354e-05,
      "loss": 0.23850904405117035,
      "mean_token_accuracy": 0.9052789807319641,
      "num_tokens": 506586393.0,
      "step": 218
    },
    {
      "epoch": 0.1120777891504606,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.930918867051313e-05,
      "loss": 0.25066208839416504,
      "mean_token_accuracy": 0.9004342555999756,
      "num_tokens": 508845783.0,
      "step": 219
    },
    {
      "epoch": 0.11258955987717502,
      "grad_norm": 0.232421875,
      "learning_rate": 3.930052306112667e-05,
      "loss": 0.2505097985267639,
      "mean_token_accuracy": 0.9000257849693298,
      "num_tokens": 511133130.0,
      "step": 220
    },
    {
      "epoch": 0.11310133060388945,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.92918044060266e-05,
      "loss": 0.25210145115852356,
      "mean_token_accuracy": 0.9000192880630493,
      "num_tokens": 513386395.0,
      "step": 221
    },
    {
      "epoch": 0.11361310133060389,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.928303272917534e-05,
      "loss": 0.24281346797943115,
      "mean_token_accuracy": 0.9035021066665649,
      "num_tokens": 515692517.0,
      "step": 222
    },
    {
      "epoch": 0.11412487205731832,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.927420805468104e-05,
      "loss": 0.23728840053081512,
      "mean_token_accuracy": 0.9046974182128906,
      "num_tokens": 518048570.0,
      "step": 223
    },
    {
      "epoch": 0.11463664278403275,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.92653304067975e-05,
      "loss": 0.2408822774887085,
      "mean_token_accuracy": 0.9040480852127075,
      "num_tokens": 520377611.0,
      "step": 224
    },
    {
      "epoch": 0.11514841351074719,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.9256399809924125e-05,
      "loss": 0.24657976627349854,
      "mean_token_accuracy": 0.9027004837989807,
      "num_tokens": 522656469.0,
      "step": 225
    },
    {
      "epoch": 0.11566018423746162,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.9247416288605847e-05,
      "loss": 0.24549952149391174,
      "mean_token_accuracy": 0.9020711183547974,
      "num_tokens": 525035297.0,
      "step": 226
    },
    {
      "epoch": 0.11617195496417605,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.923837986753304e-05,
      "loss": 0.23946134746074677,
      "mean_token_accuracy": 0.9044755697250366,
      "num_tokens": 527276266.0,
      "step": 227
    },
    {
      "epoch": 0.11668372569089049,
      "grad_norm": 0.228515625,
      "learning_rate": 3.922929057154147e-05,
      "loss": 0.2389141172170639,
      "mean_token_accuracy": 0.9048819541931152,
      "num_tokens": 529624606.0,
      "step": 228
    },
    {
      "epoch": 0.11719549641760492,
      "grad_norm": 0.236328125,
      "learning_rate": 3.922014842561225e-05,
      "loss": 0.25452756881713867,
      "mean_token_accuracy": 0.8986743688583374,
      "num_tokens": 531868760.0,
      "step": 229
    },
    {
      "epoch": 0.11770726714431934,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.921095345487171e-05,
      "loss": 0.2530708611011505,
      "mean_token_accuracy": 0.8991786241531372,
      "num_tokens": 534113949.0,
      "step": 230
    },
    {
      "epoch": 0.11821903787103377,
      "grad_norm": 0.2265625,
      "learning_rate": 3.920170568459138e-05,
      "loss": 0.2461683750152588,
      "mean_token_accuracy": 0.9018513560295105,
      "num_tokens": 536459753.0,
      "step": 231
    },
    {
      "epoch": 0.1187308085977482,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.919240514018793e-05,
      "loss": 0.23934483528137207,
      "mean_token_accuracy": 0.9040737152099609,
      "num_tokens": 538789620.0,
      "step": 232
    },
    {
      "epoch": 0.11924257932446264,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.9183051847223035e-05,
      "loss": 0.24763987958431244,
      "mean_token_accuracy": 0.9008349180221558,
      "num_tokens": 541141815.0,
      "step": 233
    },
    {
      "epoch": 0.11975435005117707,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.9173645831403363e-05,
      "loss": 0.2560185492038727,
      "mean_token_accuracy": 0.8987752199172974,
      "num_tokens": 543480623.0,
      "step": 234
    },
    {
      "epoch": 0.1202661207778915,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.9164187118580484e-05,
      "loss": 0.2373552918434143,
      "mean_token_accuracy": 0.9049652814865112,
      "num_tokens": 545806733.0,
      "step": 235
    },
    {
      "epoch": 0.12077789150460594,
      "grad_norm": 0.2265625,
      "learning_rate": 3.91546757347508e-05,
      "loss": 0.24879631400108337,
      "mean_token_accuracy": 0.901096522808075,
      "num_tokens": 548196969.0,
      "step": 236
    },
    {
      "epoch": 0.12128966223132037,
      "grad_norm": 0.248046875,
      "learning_rate": 3.9145111706055476e-05,
      "loss": 0.24674512445926666,
      "mean_token_accuracy": 0.9020344018936157,
      "num_tokens": 550414042.0,
      "step": 237
    },
    {
      "epoch": 0.1218014329580348,
      "grad_norm": 0.234375,
      "learning_rate": 3.9135495058780364e-05,
      "loss": 0.24662131071090698,
      "mean_token_accuracy": 0.9012952446937561,
      "num_tokens": 552718125.0,
      "step": 238
    },
    {
      "epoch": 0.12231320368474924,
      "grad_norm": 0.2265625,
      "learning_rate": 3.912582581935594e-05,
      "loss": 0.24000178277492523,
      "mean_token_accuracy": 0.9042918682098389,
      "num_tokens": 555114581.0,
      "step": 239
    },
    {
      "epoch": 0.12282497441146366,
      "grad_norm": 0.251953125,
      "learning_rate": 3.9116104014357216e-05,
      "loss": 0.26257240772247314,
      "mean_token_accuracy": 0.8949935436248779,
      "num_tokens": 557403163.0,
      "step": 240
    },
    {
      "epoch": 0.12333674513817809,
      "grad_norm": 0.23828125,
      "learning_rate": 3.910632967050368e-05,
      "loss": 0.24393190443515778,
      "mean_token_accuracy": 0.902656078338623,
      "num_tokens": 559736879.0,
      "step": 241
    },
    {
      "epoch": 0.12384851586489252,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.909650281465922e-05,
      "loss": 0.23649246990680695,
      "mean_token_accuracy": 0.9056961536407471,
      "num_tokens": 562135953.0,
      "step": 242
    },
    {
      "epoch": 0.12436028659160696,
      "grad_norm": 0.244140625,
      "learning_rate": 3.908662347383204e-05,
      "loss": 0.25284868478775024,
      "mean_token_accuracy": 0.8986271619796753,
      "num_tokens": 564460530.0,
      "step": 243
    },
    {
      "epoch": 0.12487205731832139,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.907669167517461e-05,
      "loss": 0.2411130964756012,
      "mean_token_accuracy": 0.90416020154953,
      "num_tokens": 566727330.0,
      "step": 244
    },
    {
      "epoch": 0.1253838280450358,
      "grad_norm": 0.23828125,
      "learning_rate": 3.906670744598355e-05,
      "loss": 0.25052309036254883,
      "mean_token_accuracy": 0.899774432182312,
      "num_tokens": 569027373.0,
      "step": 245
    },
    {
      "epoch": 0.12589559877175024,
      "grad_norm": 0.23828125,
      "learning_rate": 3.905667081369962e-05,
      "loss": 0.24171078205108643,
      "mean_token_accuracy": 0.9036694169044495,
      "num_tokens": 571314142.0,
      "step": 246
    },
    {
      "epoch": 0.12640736949846468,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.9046581805907554e-05,
      "loss": 0.2403813898563385,
      "mean_token_accuracy": 0.9040638208389282,
      "num_tokens": 573610387.0,
      "step": 247
    },
    {
      "epoch": 0.1269191402251791,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.903644045033608e-05,
      "loss": 0.24874547123908997,
      "mean_token_accuracy": 0.9000816941261292,
      "num_tokens": 575982046.0,
      "step": 248
    },
    {
      "epoch": 0.12743091095189354,
      "grad_norm": 0.22265625,
      "learning_rate": 3.9026246774857786e-05,
      "loss": 0.23651516437530518,
      "mean_token_accuracy": 0.905229926109314,
      "num_tokens": 578278459.0,
      "step": 249
    },
    {
      "epoch": 0.12794268167860798,
      "grad_norm": 0.212890625,
      "learning_rate": 3.901600080748903e-05,
      "loss": 0.23908546566963196,
      "mean_token_accuracy": 0.9044920206069946,
      "num_tokens": 580588880.0,
      "step": 250
    },
    {
      "epoch": 0.1284544524053224,
      "grad_norm": 0.224609375,
      "learning_rate": 3.900570257638993e-05,
      "loss": 0.24699902534484863,
      "mean_token_accuracy": 0.9030405282974243,
      "num_tokens": 582903028.0,
      "step": 251
    },
    {
      "epoch": 0.12896622313203684,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.899535210986423e-05,
      "loss": 0.25645768642425537,
      "mean_token_accuracy": 0.8976227045059204,
      "num_tokens": 585138415.0,
      "step": 252
    },
    {
      "epoch": 0.12947799385875128,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.898494943635922e-05,
      "loss": 0.2261674404144287,
      "mean_token_accuracy": 0.9105488061904907,
      "num_tokens": 587416429.0,
      "step": 253
    },
    {
      "epoch": 0.1299897645854657,
      "grad_norm": 0.228515625,
      "learning_rate": 3.8974494584465705e-05,
      "loss": 0.24532634019851685,
      "mean_token_accuracy": 0.9024957418441772,
      "num_tokens": 589814708.0,
      "step": 254
    },
    {
      "epoch": 0.13050153531218014,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.896398758291788e-05,
      "loss": 0.24492083489894867,
      "mean_token_accuracy": 0.9024583697319031,
      "num_tokens": 592181543.0,
      "step": 255
    },
    {
      "epoch": 0.13101330603889458,
      "grad_norm": 0.228515625,
      "learning_rate": 3.8953428460593276e-05,
      "loss": 0.24431923031806946,
      "mean_token_accuracy": 0.9021969437599182,
      "num_tokens": 594455859.0,
      "step": 256
    },
    {
      "epoch": 0.131525076765609,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.8942817246512665e-05,
      "loss": 0.24174588918685913,
      "mean_token_accuracy": 0.904079258441925,
      "num_tokens": 596761000.0,
      "step": 257
    },
    {
      "epoch": 0.13203684749232344,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.8932153969839997e-05,
      "loss": 0.24932000041007996,
      "mean_token_accuracy": 0.9009469151496887,
      "num_tokens": 599107864.0,
      "step": 258
    },
    {
      "epoch": 0.13254861821903788,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.8921438659882305e-05,
      "loss": 0.23109044134616852,
      "mean_token_accuracy": 0.9065721035003662,
      "num_tokens": 601406032.0,
      "step": 259
    },
    {
      "epoch": 0.1330603889457523,
      "grad_norm": 0.234375,
      "learning_rate": 3.891067134608963e-05,
      "loss": 0.23812396824359894,
      "mean_token_accuracy": 0.9057261943817139,
      "num_tokens": 603798571.0,
      "step": 260
    },
    {
      "epoch": 0.13357215967246674,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.889985205805496e-05,
      "loss": 0.24448290467262268,
      "mean_token_accuracy": 0.9017740488052368,
      "num_tokens": 606073369.0,
      "step": 261
    },
    {
      "epoch": 0.13408393039918118,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.88889808255141e-05,
      "loss": 0.24974820017814636,
      "mean_token_accuracy": 0.9002279043197632,
      "num_tokens": 608390662.0,
      "step": 262
    },
    {
      "epoch": 0.1345957011258956,
      "grad_norm": 0.232421875,
      "learning_rate": 3.887805767834563e-05,
      "loss": 0.24345102906227112,
      "mean_token_accuracy": 0.9028403759002686,
      "num_tokens": 610630177.0,
      "step": 263
    },
    {
      "epoch": 0.13510747185261002,
      "grad_norm": 0.234375,
      "learning_rate": 3.8867082646570827e-05,
      "loss": 0.2577478885650635,
      "mean_token_accuracy": 0.8987982273101807,
      "num_tokens": 612923092.0,
      "step": 264
    },
    {
      "epoch": 0.13561924257932445,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.885605576035355e-05,
      "loss": 0.2369365394115448,
      "mean_token_accuracy": 0.9051200151443481,
      "num_tokens": 615191366.0,
      "step": 265
    },
    {
      "epoch": 0.13613101330603888,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.8844977050000194e-05,
      "loss": 0.25867852568626404,
      "mean_token_accuracy": 0.8966612815856934,
      "num_tokens": 617468276.0,
      "step": 266
    },
    {
      "epoch": 0.13664278403275332,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.883384654595956e-05,
      "loss": 0.24485568702220917,
      "mean_token_accuracy": 0.9039015769958496,
      "num_tokens": 619819553.0,
      "step": 267
    },
    {
      "epoch": 0.13715455475946775,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.882266427882282e-05,
      "loss": 0.24303387105464935,
      "mean_token_accuracy": 0.9037355184555054,
      "num_tokens": 622164827.0,
      "step": 268
    },
    {
      "epoch": 0.13766632548618218,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.881143027932341e-05,
      "loss": 0.2430010735988617,
      "mean_token_accuracy": 0.9025978446006775,
      "num_tokens": 624495867.0,
      "step": 269
    },
    {
      "epoch": 0.13817809621289662,
      "grad_norm": 0.232421875,
      "learning_rate": 3.880014457833695e-05,
      "loss": 0.2390325367450714,
      "mean_token_accuracy": 0.9053988456726074,
      "num_tokens": 626768460.0,
      "step": 270
    },
    {
      "epoch": 0.13868986693961105,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.8788807206881146e-05,
      "loss": 0.23508915305137634,
      "mean_token_accuracy": 0.9065207242965698,
      "num_tokens": 629150763.0,
      "step": 271
    },
    {
      "epoch": 0.13920163766632548,
      "grad_norm": 0.23046875,
      "learning_rate": 3.8777418196115715e-05,
      "loss": 0.24325430393218994,
      "mean_token_accuracy": 0.9028247594833374,
      "num_tokens": 631470517.0,
      "step": 272
    },
    {
      "epoch": 0.13971340839303992,
      "grad_norm": 0.2265625,
      "learning_rate": 3.876597757734232e-05,
      "loss": 0.23866796493530273,
      "mean_token_accuracy": 0.9044126868247986,
      "num_tokens": 633745523.0,
      "step": 273
    },
    {
      "epoch": 0.14022517911975435,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.875448538200444e-05,
      "loss": 0.24252457916736603,
      "mean_token_accuracy": 0.9033201336860657,
      "num_tokens": 636084796.0,
      "step": 274
    },
    {
      "epoch": 0.14073694984646878,
      "grad_norm": 0.232421875,
      "learning_rate": 3.874294164168732e-05,
      "loss": 0.2509998083114624,
      "mean_token_accuracy": 0.8997800350189209,
      "num_tokens": 638393865.0,
      "step": 275
    },
    {
      "epoch": 0.14124872057318322,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.8731346388117875e-05,
      "loss": 0.23807352781295776,
      "mean_token_accuracy": 0.9054867029190063,
      "num_tokens": 640756148.0,
      "step": 276
    },
    {
      "epoch": 0.14176049129989765,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.871969965316458e-05,
      "loss": 0.2576558291912079,
      "mean_token_accuracy": 0.898151159286499,
      "num_tokens": 643060877.0,
      "step": 277
    },
    {
      "epoch": 0.14227226202661208,
      "grad_norm": 0.23046875,
      "learning_rate": 3.870800146883743e-05,
      "loss": 0.23458275198936462,
      "mean_token_accuracy": 0.9061093330383301,
      "num_tokens": 645398779.0,
      "step": 278
    },
    {
      "epoch": 0.14278403275332652,
      "grad_norm": 0.21875,
      "learning_rate": 3.8696251867287806e-05,
      "loss": 0.24690546095371246,
      "mean_token_accuracy": 0.9020704030990601,
      "num_tokens": 647759051.0,
      "step": 279
    },
    {
      "epoch": 0.14329580348004095,
      "grad_norm": 0.212890625,
      "learning_rate": 3.86844508808084e-05,
      "loss": 0.23661227524280548,
      "mean_token_accuracy": 0.9052391648292542,
      "num_tokens": 650168557.0,
      "step": 280
    },
    {
      "epoch": 0.14380757420675538,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.867259854183315e-05,
      "loss": 0.24386972188949585,
      "mean_token_accuracy": 0.9031212329864502,
      "num_tokens": 652446232.0,
      "step": 281
    },
    {
      "epoch": 0.14431934493346982,
      "grad_norm": 0.23046875,
      "learning_rate": 3.866069488293711e-05,
      "loss": 0.2385876178741455,
      "mean_token_accuracy": 0.9040321111679077,
      "num_tokens": 654711331.0,
      "step": 282
    },
    {
      "epoch": 0.14483111566018425,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.86487399368364e-05,
      "loss": 0.24863524734973907,
      "mean_token_accuracy": 0.9006595611572266,
      "num_tokens": 657007165.0,
      "step": 283
    },
    {
      "epoch": 0.14534288638689866,
      "grad_norm": 0.2265625,
      "learning_rate": 3.863673373638808e-05,
      "loss": 0.24834966659545898,
      "mean_token_accuracy": 0.900520920753479,
      "num_tokens": 659426803.0,
      "step": 284
    },
    {
      "epoch": 0.1458546571136131,
      "grad_norm": 0.228515625,
      "learning_rate": 3.86246763145901e-05,
      "loss": 0.2375149130821228,
      "mean_token_accuracy": 0.9051332473754883,
      "num_tokens": 661787400.0,
      "step": 285
    },
    {
      "epoch": 0.14636642784032752,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.861256770458117e-05,
      "loss": 0.2356548309326172,
      "mean_token_accuracy": 0.9059121608734131,
      "num_tokens": 664111478.0,
      "step": 286
    },
    {
      "epoch": 0.14687819856704196,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.8600407939640696e-05,
      "loss": 0.24046719074249268,
      "mean_token_accuracy": 0.9043402671813965,
      "num_tokens": 666403478.0,
      "step": 287
    },
    {
      "epoch": 0.1473899692937564,
      "grad_norm": 0.216796875,
      "learning_rate": 3.858819705318867e-05,
      "loss": 0.23155397176742554,
      "mean_token_accuracy": 0.9079502820968628,
      "num_tokens": 668729813.0,
      "step": 288
    },
    {
      "epoch": 0.14790174002047082,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.857593507878559e-05,
      "loss": 0.2560300827026367,
      "mean_token_accuracy": 0.8984227180480957,
      "num_tokens": 670991130.0,
      "step": 289
    },
    {
      "epoch": 0.14841351074718526,
      "grad_norm": 0.228515625,
      "learning_rate": 3.856362205013237e-05,
      "loss": 0.23610028624534607,
      "mean_token_accuracy": 0.905639111995697,
      "num_tokens": 673304890.0,
      "step": 290
    },
    {
      "epoch": 0.1489252814738997,
      "grad_norm": 0.216796875,
      "learning_rate": 3.855125800107024e-05,
      "loss": 0.23371656239032745,
      "mean_token_accuracy": 0.9072609543800354,
      "num_tokens": 675652246.0,
      "step": 291
    },
    {
      "epoch": 0.14943705220061412,
      "grad_norm": 0.234375,
      "learning_rate": 3.8538842965580635e-05,
      "loss": 0.24607157707214355,
      "mean_token_accuracy": 0.9015742540359497,
      "num_tokens": 678000733.0,
      "step": 292
    },
    {
      "epoch": 0.14994882292732856,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.852637697778516e-05,
      "loss": 0.2585951089859009,
      "mean_token_accuracy": 0.8974155187606812,
      "num_tokens": 680402241.0,
      "step": 293
    },
    {
      "epoch": 0.150460593654043,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.851386007194543e-05,
      "loss": 0.2247713953256607,
      "mean_token_accuracy": 0.9102572202682495,
      "num_tokens": 682793676.0,
      "step": 294
    },
    {
      "epoch": 0.15097236438075742,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.8501292282463006e-05,
      "loss": 0.25294244289398193,
      "mean_token_accuracy": 0.8989990949630737,
      "num_tokens": 685056227.0,
      "step": 295
    },
    {
      "epoch": 0.15148413510747186,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.84886736438793e-05,
      "loss": 0.2406432032585144,
      "mean_token_accuracy": 0.9038178324699402,
      "num_tokens": 687340974.0,
      "step": 296
    },
    {
      "epoch": 0.1519959058341863,
      "grad_norm": 0.228515625,
      "learning_rate": 3.8476004190875486e-05,
      "loss": 0.2459370493888855,
      "mean_token_accuracy": 0.9015297889709473,
      "num_tokens": 689564923.0,
      "step": 297
    },
    {
      "epoch": 0.15250767656090072,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.846328395827239e-05,
      "loss": 0.24155974388122559,
      "mean_token_accuracy": 0.9032509326934814,
      "num_tokens": 691928141.0,
      "step": 298
    },
    {
      "epoch": 0.15301944728761516,
      "grad_norm": 0.2265625,
      "learning_rate": 3.84505129810304e-05,
      "loss": 0.23955386877059937,
      "mean_token_accuracy": 0.9046834707260132,
      "num_tokens": 694305131.0,
      "step": 299
    },
    {
      "epoch": 0.1535312180143296,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.843769129424938e-05,
      "loss": 0.2479058802127838,
      "mean_token_accuracy": 0.9010123014450073,
      "num_tokens": 696641638.0,
      "step": 300
    },
    {
      "epoch": 0.15404298874104402,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.8424818933168555e-05,
      "loss": 0.2468447983264923,
      "mean_token_accuracy": 0.9001972675323486,
      "num_tokens": 698852282.0,
      "step": 301
    },
    {
      "epoch": 0.15455475946775846,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.841189593316642e-05,
      "loss": 0.2378157079219818,
      "mean_token_accuracy": 0.9053705334663391,
      "num_tokens": 701189642.0,
      "step": 302
    },
    {
      "epoch": 0.15506653019447286,
      "grad_norm": 0.228515625,
      "learning_rate": 3.8398922329760656e-05,
      "loss": 0.2497168630361557,
      "mean_token_accuracy": 0.9004157781600952,
      "num_tokens": 703472152.0,
      "step": 303
    },
    {
      "epoch": 0.1555783009211873,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.838589815860802e-05,
      "loss": 0.24089187383651733,
      "mean_token_accuracy": 0.9038037061691284,
      "num_tokens": 705775228.0,
      "step": 304
    },
    {
      "epoch": 0.15609007164790173,
      "grad_norm": 0.21484375,
      "learning_rate": 3.837282345550425e-05,
      "loss": 0.24044910073280334,
      "mean_token_accuracy": 0.9039671421051025,
      "num_tokens": 708091038.0,
      "step": 305
    },
    {
      "epoch": 0.15660184237461616,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.8359698256383964e-05,
      "loss": 0.23290538787841797,
      "mean_token_accuracy": 0.9076389670372009,
      "num_tokens": 710391260.0,
      "step": 306
    },
    {
      "epoch": 0.1571136131013306,
      "grad_norm": 0.236328125,
      "learning_rate": 3.834652259732057e-05,
      "loss": 0.2503116726875305,
      "mean_token_accuracy": 0.9002904891967773,
      "num_tokens": 712750878.0,
      "step": 307
    },
    {
      "epoch": 0.15762538382804503,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.833329651452615e-05,
      "loss": 0.24079462885856628,
      "mean_token_accuracy": 0.9043090343475342,
      "num_tokens": 715016310.0,
      "step": 308
    },
    {
      "epoch": 0.15813715455475946,
      "grad_norm": 0.232421875,
      "learning_rate": 3.832002004435138e-05,
      "loss": 0.24995118379592896,
      "mean_token_accuracy": 0.9014227390289307,
      "num_tokens": 717366939.0,
      "step": 309
    },
    {
      "epoch": 0.1586489252814739,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.830669322328541e-05,
      "loss": 0.23532240092754364,
      "mean_token_accuracy": 0.906798243522644,
      "num_tokens": 719706608.0,
      "step": 310
    },
    {
      "epoch": 0.15916069600818833,
      "grad_norm": 0.224609375,
      "learning_rate": 3.82933160879558e-05,
      "loss": 0.24664545059204102,
      "mean_token_accuracy": 0.901218831539154,
      "num_tokens": 721996460.0,
      "step": 311
    },
    {
      "epoch": 0.15967246673490276,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.827988867512836e-05,
      "loss": 0.24682888388633728,
      "mean_token_accuracy": 0.9005022048950195,
      "num_tokens": 724366953.0,
      "step": 312
    },
    {
      "epoch": 0.1601842374616172,
      "grad_norm": 0.244140625,
      "learning_rate": 3.826641102170711e-05,
      "loss": 0.24089287221431732,
      "mean_token_accuracy": 0.9040721654891968,
      "num_tokens": 726704602.0,
      "step": 313
    },
    {
      "epoch": 0.16069600818833163,
      "grad_norm": 0.25,
      "learning_rate": 3.825288316473414e-05,
      "loss": 0.24661597609519958,
      "mean_token_accuracy": 0.9019604325294495,
      "num_tokens": 729042393.0,
      "step": 314
    },
    {
      "epoch": 0.16120777891504606,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.823930514138952e-05,
      "loss": 0.23302510380744934,
      "mean_token_accuracy": 0.9062788486480713,
      "num_tokens": 731294925.0,
      "step": 315
    },
    {
      "epoch": 0.1617195496417605,
      "grad_norm": 0.248046875,
      "learning_rate": 3.8225676988991193e-05,
      "loss": 0.24521690607070923,
      "mean_token_accuracy": 0.9017184972763062,
      "num_tokens": 733586224.0,
      "step": 316
    },
    {
      "epoch": 0.16223132036847493,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.8211998744994893e-05,
      "loss": 0.25062423944473267,
      "mean_token_accuracy": 0.9007511138916016,
      "num_tokens": 735942683.0,
      "step": 317
    },
    {
      "epoch": 0.16274309109518936,
      "grad_norm": 0.205078125,
      "learning_rate": 3.819827044699401e-05,
      "loss": 0.24442115426063538,
      "mean_token_accuracy": 0.9024111032485962,
      "num_tokens": 738146879.0,
      "step": 318
    },
    {
      "epoch": 0.1632548618219038,
      "grad_norm": 0.2451171875,
      "learning_rate": 3.818449213271952e-05,
      "loss": 0.2519283890724182,
      "mean_token_accuracy": 0.8997936248779297,
      "num_tokens": 740344198.0,
      "step": 319
    },
    {
      "epoch": 0.16376663254861823,
      "grad_norm": 0.232421875,
      "learning_rate": 3.817066384003984e-05,
      "loss": 0.24167296290397644,
      "mean_token_accuracy": 0.9032808542251587,
      "num_tokens": 742688136.0,
      "step": 320
    },
    {
      "epoch": 0.16427840327533266,
      "grad_norm": 0.23046875,
      "learning_rate": 3.815678560696076e-05,
      "loss": 0.24623367190361023,
      "mean_token_accuracy": 0.9021510481834412,
      "num_tokens": 745029421.0,
      "step": 321
    },
    {
      "epoch": 0.1647901740020471,
      "grad_norm": 0.212890625,
      "learning_rate": 3.814285747162534e-05,
      "loss": 0.22775866091251373,
      "mean_token_accuracy": 0.9090938568115234,
      "num_tokens": 747424886.0,
      "step": 322
    },
    {
      "epoch": 0.1653019447287615,
      "grad_norm": 0.21875,
      "learning_rate": 3.812887947231378e-05,
      "loss": 0.23189744353294373,
      "mean_token_accuracy": 0.9078401923179626,
      "num_tokens": 749711920.0,
      "step": 323
    },
    {
      "epoch": 0.16581371545547594,
      "grad_norm": 0.220703125,
      "learning_rate": 3.811485164744332e-05,
      "loss": 0.24300794303417206,
      "mean_token_accuracy": 0.9032754898071289,
      "num_tokens": 752053976.0,
      "step": 324
    },
    {
      "epoch": 0.16632548618219037,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.8100774035568135e-05,
      "loss": 0.2298419326543808,
      "mean_token_accuracy": 0.9081000089645386,
      "num_tokens": 754396404.0,
      "step": 325
    },
    {
      "epoch": 0.1668372569089048,
      "grad_norm": 0.2265625,
      "learning_rate": 3.808664667537927e-05,
      "loss": 0.24145059287548065,
      "mean_token_accuracy": 0.9037232398986816,
      "num_tokens": 756736990.0,
      "step": 326
    },
    {
      "epoch": 0.16734902763561924,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.8072469605704465e-05,
      "loss": 0.24170780181884766,
      "mean_token_accuracy": 0.902938961982727,
      "num_tokens": 759039503.0,
      "step": 327
    },
    {
      "epoch": 0.16786079836233367,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.8058242865508094e-05,
      "loss": 0.22731032967567444,
      "mean_token_accuracy": 0.9088147282600403,
      "num_tokens": 761411867.0,
      "step": 328
    },
    {
      "epoch": 0.1683725690890481,
      "grad_norm": 0.234375,
      "learning_rate": 3.8043966493891045e-05,
      "loss": 0.254142165184021,
      "mean_token_accuracy": 0.8991057872772217,
      "num_tokens": 763780971.0,
      "step": 329
    },
    {
      "epoch": 0.16888433981576254,
      "grad_norm": 0.236328125,
      "learning_rate": 3.802964053009061e-05,
      "loss": 0.24407893419265747,
      "mean_token_accuracy": 0.9023332595825195,
      "num_tokens": 766065092.0,
      "step": 330
    },
    {
      "epoch": 0.16939611054247697,
      "grad_norm": 0.2265625,
      "learning_rate": 3.8015265013480386e-05,
      "loss": 0.24811218678951263,
      "mean_token_accuracy": 0.9014322757720947,
      "num_tokens": 768356723.0,
      "step": 331
    },
    {
      "epoch": 0.1699078812691914,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.800083998357016e-05,
      "loss": 0.24287313222885132,
      "mean_token_accuracy": 0.9030367732048035,
      "num_tokens": 770704635.0,
      "step": 332
    },
    {
      "epoch": 0.17041965199590584,
      "grad_norm": 0.23046875,
      "learning_rate": 3.79863654800058e-05,
      "loss": 0.24590522050857544,
      "mean_token_accuracy": 0.9021560549736023,
      "num_tokens": 772993629.0,
      "step": 333
    },
    {
      "epoch": 0.17093142272262027,
      "grad_norm": 0.228515625,
      "learning_rate": 3.797184154256916e-05,
      "loss": 0.2427978217601776,
      "mean_token_accuracy": 0.9036859273910522,
      "num_tokens": 775322024.0,
      "step": 334
    },
    {
      "epoch": 0.1714431934493347,
      "grad_norm": 0.23828125,
      "learning_rate": 3.795726821117793e-05,
      "loss": 0.25739073753356934,
      "mean_token_accuracy": 0.8976163864135742,
      "num_tokens": 777627599.0,
      "step": 335
    },
    {
      "epoch": 0.17195496417604914,
      "grad_norm": 0.240234375,
      "learning_rate": 3.794264552588557e-05,
      "loss": 0.24701356887817383,
      "mean_token_accuracy": 0.9015910625457764,
      "num_tokens": 779994454.0,
      "step": 336
    },
    {
      "epoch": 0.17246673490276357,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.79279735268812e-05,
      "loss": 0.22736479341983795,
      "mean_token_accuracy": 0.9083849191665649,
      "num_tokens": 782300307.0,
      "step": 337
    },
    {
      "epoch": 0.172978505629478,
      "grad_norm": 0.228515625,
      "learning_rate": 3.7913252254489454e-05,
      "loss": 0.24969004094600677,
      "mean_token_accuracy": 0.9005051255226135,
      "num_tokens": 784552569.0,
      "step": 338
    },
    {
      "epoch": 0.17349027635619244,
      "grad_norm": 0.232421875,
      "learning_rate": 3.7898481749170404e-05,
      "loss": 0.2461336851119995,
      "mean_token_accuracy": 0.9007627964019775,
      "num_tokens": 786935195.0,
      "step": 339
    },
    {
      "epoch": 0.17400204708290687,
      "grad_norm": 0.248046875,
      "learning_rate": 3.788366205151941e-05,
      "loss": 0.26372602581977844,
      "mean_token_accuracy": 0.8945207595825195,
      "num_tokens": 789151979.0,
      "step": 340
    },
    {
      "epoch": 0.1745138178096213,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.786879320226707e-05,
      "loss": 0.2538994550704956,
      "mean_token_accuracy": 0.8983903527259827,
      "num_tokens": 791470063.0,
      "step": 341
    },
    {
      "epoch": 0.1750255885363357,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.785387524227902e-05,
      "loss": 0.24439497292041779,
      "mean_token_accuracy": 0.9028741717338562,
      "num_tokens": 793804217.0,
      "step": 342
    },
    {
      "epoch": 0.17553735926305014,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.783890821255593e-05,
      "loss": 0.23874230682849884,
      "mean_token_accuracy": 0.9042048454284668,
      "num_tokens": 796116641.0,
      "step": 343
    },
    {
      "epoch": 0.17604912998976457,
      "grad_norm": 0.216796875,
      "learning_rate": 3.78238921542333e-05,
      "loss": 0.23060342669487,
      "mean_token_accuracy": 0.9076826572418213,
      "num_tokens": 798395179.0,
      "step": 344
    },
    {
      "epoch": 0.176560900716479,
      "grad_norm": 0.2470703125,
      "learning_rate": 3.7808827108581374e-05,
      "loss": 0.2600717842578888,
      "mean_token_accuracy": 0.8965385556221008,
      "num_tokens": 800831759.0,
      "step": 345
    },
    {
      "epoch": 0.17707267144319344,
      "grad_norm": 0.22265625,
      "learning_rate": 3.779371311700506e-05,
      "loss": 0.2385694682598114,
      "mean_token_accuracy": 0.9049931168556213,
      "num_tokens": 803122577.0,
      "step": 346
    },
    {
      "epoch": 0.17758444216990787,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.777855022104376e-05,
      "loss": 0.23724478483200073,
      "mean_token_accuracy": 0.9054543375968933,
      "num_tokens": 805512806.0,
      "step": 347
    },
    {
      "epoch": 0.1780962128966223,
      "grad_norm": 0.234375,
      "learning_rate": 3.776333846237131e-05,
      "loss": 0.2315598428249359,
      "mean_token_accuracy": 0.9078267812728882,
      "num_tokens": 807791359.0,
      "step": 348
    },
    {
      "epoch": 0.17860798362333674,
      "grad_norm": 0.22265625,
      "learning_rate": 3.7748077882795826e-05,
      "loss": 0.23856602609157562,
      "mean_token_accuracy": 0.9047994613647461,
      "num_tokens": 810116334.0,
      "step": 349
    },
    {
      "epoch": 0.17911975435005117,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.77327685242596e-05,
      "loss": 0.2510479688644409,
      "mean_token_accuracy": 0.8999359011650085,
      "num_tokens": 812387655.0,
      "step": 350
    },
    {
      "epoch": 0.1796315250767656,
      "grad_norm": 0.232421875,
      "learning_rate": 3.771741042883902e-05,
      "loss": 0.2360416054725647,
      "mean_token_accuracy": 0.905266284942627,
      "num_tokens": 814700290.0,
      "step": 351
    },
    {
      "epoch": 0.18014329580348004,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.770200363874438e-05,
      "loss": 0.24439957737922668,
      "mean_token_accuracy": 0.9026628136634827,
      "num_tokens": 816997440.0,
      "step": 352
    },
    {
      "epoch": 0.18065506653019447,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.768654819631984e-05,
      "loss": 0.2365715503692627,
      "mean_token_accuracy": 0.9059075117111206,
      "num_tokens": 819356349.0,
      "step": 353
    },
    {
      "epoch": 0.1811668372569089,
      "grad_norm": 0.240234375,
      "learning_rate": 3.7671044144043244e-05,
      "loss": 0.2522013187408447,
      "mean_token_accuracy": 0.8994003534317017,
      "num_tokens": 821613338.0,
      "step": 354
    },
    {
      "epoch": 0.18167860798362334,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.7655491524526075e-05,
      "loss": 0.24765610694885254,
      "mean_token_accuracy": 0.8999534845352173,
      "num_tokens": 823957849.0,
      "step": 355
    },
    {
      "epoch": 0.18219037871033777,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.7639890380513266e-05,
      "loss": 0.2579483389854431,
      "mean_token_accuracy": 0.8970052003860474,
      "num_tokens": 826294505.0,
      "step": 356
    },
    {
      "epoch": 0.1827021494370522,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.762424075488313e-05,
      "loss": 0.23767563700675964,
      "mean_token_accuracy": 0.904742956161499,
      "num_tokens": 828595378.0,
      "step": 357
    },
    {
      "epoch": 0.18321392016376664,
      "grad_norm": 0.255859375,
      "learning_rate": 3.7608542690647225e-05,
      "loss": 0.23741160333156586,
      "mean_token_accuracy": 0.9043813347816467,
      "num_tokens": 830932276.0,
      "step": 358
    },
    {
      "epoch": 0.18372569089048107,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.759279623095024e-05,
      "loss": 0.2445419877767563,
      "mean_token_accuracy": 0.9028075933456421,
      "num_tokens": 833173010.0,
      "step": 359
    },
    {
      "epoch": 0.1842374616171955,
      "grad_norm": 0.2421875,
      "learning_rate": 3.757700141906987e-05,
      "loss": 0.24986325204372406,
      "mean_token_accuracy": 0.9007747173309326,
      "num_tokens": 835437644.0,
      "step": 360
    },
    {
      "epoch": 0.18474923234390994,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.75611582984167e-05,
      "loss": 0.23528356850147247,
      "mean_token_accuracy": 0.9062751531600952,
      "num_tokens": 837787018.0,
      "step": 361
    },
    {
      "epoch": 0.18526100307062435,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.7545266912534095e-05,
      "loss": 0.24495801329612732,
      "mean_token_accuracy": 0.9027127623558044,
      "num_tokens": 840048540.0,
      "step": 362
    },
    {
      "epoch": 0.18577277379733878,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.752932730509807e-05,
      "loss": 0.2522444725036621,
      "mean_token_accuracy": 0.8996013402938843,
      "num_tokens": 842400340.0,
      "step": 363
    },
    {
      "epoch": 0.1862845445240532,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.751333951991717e-05,
      "loss": 0.24304912984371185,
      "mean_token_accuracy": 0.9031103253364563,
      "num_tokens": 844686507.0,
      "step": 364
    },
    {
      "epoch": 0.18679631525076765,
      "grad_norm": 0.240234375,
      "learning_rate": 3.749730360093235e-05,
      "loss": 0.2344396859407425,
      "mean_token_accuracy": 0.9056589603424072,
      "num_tokens": 847069498.0,
      "step": 365
    },
    {
      "epoch": 0.18730808597748208,
      "grad_norm": 0.236328125,
      "learning_rate": 3.7481219592216856e-05,
      "loss": 0.24086922407150269,
      "mean_token_accuracy": 0.9032906293869019,
      "num_tokens": 849402956.0,
      "step": 366
    },
    {
      "epoch": 0.18781985670419651,
      "grad_norm": 0.220703125,
      "learning_rate": 3.746508753797612e-05,
      "loss": 0.2390659749507904,
      "mean_token_accuracy": 0.9038146734237671,
      "num_tokens": 851654899.0,
      "step": 367
    },
    {
      "epoch": 0.18833162743091095,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.744890748254761e-05,
      "loss": 0.24440377950668335,
      "mean_token_accuracy": 0.9012736082077026,
      "num_tokens": 853928719.0,
      "step": 368
    },
    {
      "epoch": 0.18884339815762538,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.7432679470400704e-05,
      "loss": 0.25158968567848206,
      "mean_token_accuracy": 0.9001213908195496,
      "num_tokens": 856246435.0,
      "step": 369
    },
    {
      "epoch": 0.18935516888433981,
      "grad_norm": 0.232421875,
      "learning_rate": 3.741640354613661e-05,
      "loss": 0.23392273485660553,
      "mean_token_accuracy": 0.906129777431488,
      "num_tokens": 858653031.0,
      "step": 370
    },
    {
      "epoch": 0.18986693961105425,
      "grad_norm": 0.23046875,
      "learning_rate": 3.740007975448821e-05,
      "loss": 0.23196980357170105,
      "mean_token_accuracy": 0.9081213474273682,
      "num_tokens": 860989194.0,
      "step": 371
    },
    {
      "epoch": 0.19037871033776868,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.7383708140319953e-05,
      "loss": 0.23316840827465057,
      "mean_token_accuracy": 0.9083614349365234,
      "num_tokens": 863333274.0,
      "step": 372
    },
    {
      "epoch": 0.19089048106448311,
      "grad_norm": 0.220703125,
      "learning_rate": 3.736728874862771e-05,
      "loss": 0.23878034949302673,
      "mean_token_accuracy": 0.9049158096313477,
      "num_tokens": 865756751.0,
      "step": 373
    },
    {
      "epoch": 0.19140225179119755,
      "grad_norm": 0.23828125,
      "learning_rate": 3.735082162453867e-05,
      "loss": 0.2537354826927185,
      "mean_token_accuracy": 0.8995856046676636,
      "num_tokens": 868072418.0,
      "step": 374
    },
    {
      "epoch": 0.19191402251791198,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.733430681331121e-05,
      "loss": 0.2437945455312729,
      "mean_token_accuracy": 0.9028054475784302,
      "num_tokens": 870443352.0,
      "step": 375
    },
    {
      "epoch": 0.19242579324462641,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.7317744360334765e-05,
      "loss": 0.2475128322839737,
      "mean_token_accuracy": 0.9017243385314941,
      "num_tokens": 872737076.0,
      "step": 376
    },
    {
      "epoch": 0.19293756397134085,
      "grad_norm": 0.220703125,
      "learning_rate": 3.7301134311129724e-05,
      "loss": 0.23508289456367493,
      "mean_token_accuracy": 0.9055121541023254,
      "num_tokens": 875080359.0,
      "step": 377
    },
    {
      "epoch": 0.19344933469805528,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.7284476711347274e-05,
      "loss": 0.24122078716754913,
      "mean_token_accuracy": 0.9042385816574097,
      "num_tokens": 877363508.0,
      "step": 378
    },
    {
      "epoch": 0.19396110542476971,
      "grad_norm": 0.228515625,
      "learning_rate": 3.7267771606769304e-05,
      "loss": 0.24215635657310486,
      "mean_token_accuracy": 0.9023236036300659,
      "num_tokens": 879656182.0,
      "step": 379
    },
    {
      "epoch": 0.19447287615148415,
      "grad_norm": 0.216796875,
      "learning_rate": 3.7251019043308244e-05,
      "loss": 0.24022677540779114,
      "mean_token_accuracy": 0.904537558555603,
      "num_tokens": 881881668.0,
      "step": 380
    },
    {
      "epoch": 0.19498464687819858,
      "grad_norm": 0.220703125,
      "learning_rate": 3.723421906700699e-05,
      "loss": 0.24244694411754608,
      "mean_token_accuracy": 0.9034748077392578,
      "num_tokens": 884242562.0,
      "step": 381
    },
    {
      "epoch": 0.195496417604913,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.721737172403873e-05,
      "loss": 0.23603053390979767,
      "mean_token_accuracy": 0.905598521232605,
      "num_tokens": 886579146.0,
      "step": 382
    },
    {
      "epoch": 0.19600818833162742,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.7200477060706825e-05,
      "loss": 0.23616501688957214,
      "mean_token_accuracy": 0.9052140116691589,
      "num_tokens": 888849652.0,
      "step": 383
    },
    {
      "epoch": 0.19651995905834185,
      "grad_norm": 0.23046875,
      "learning_rate": 3.718353512344473e-05,
      "loss": 0.23420850932598114,
      "mean_token_accuracy": 0.9064357280731201,
      "num_tokens": 891204221.0,
      "step": 384
    },
    {
      "epoch": 0.1970317297850563,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.716654595881577e-05,
      "loss": 0.2488848716020584,
      "mean_token_accuracy": 0.9011543989181519,
      "num_tokens": 893571403.0,
      "step": 385
    },
    {
      "epoch": 0.19754350051177072,
      "grad_norm": 0.2451171875,
      "learning_rate": 3.7149509613513125e-05,
      "loss": 0.2397938370704651,
      "mean_token_accuracy": 0.9034260511398315,
      "num_tokens": 895830321.0,
      "step": 386
    },
    {
      "epoch": 0.19805527123848515,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.713242613435962e-05,
      "loss": 0.2476431131362915,
      "mean_token_accuracy": 0.900909960269928,
      "num_tokens": 898161111.0,
      "step": 387
    },
    {
      "epoch": 0.1985670419651996,
      "grad_norm": 0.2265625,
      "learning_rate": 3.711529556830763e-05,
      "loss": 0.2354734241962433,
      "mean_token_accuracy": 0.9064542055130005,
      "num_tokens": 900473937.0,
      "step": 388
    },
    {
      "epoch": 0.19907881269191402,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.7098117962438926e-05,
      "loss": 0.24063530564308167,
      "mean_token_accuracy": 0.9041630029678345,
      "num_tokens": 902770705.0,
      "step": 389
    },
    {
      "epoch": 0.19959058341862845,
      "grad_norm": 0.2109375,
      "learning_rate": 3.708089336396459e-05,
      "loss": 0.23458059132099152,
      "mean_token_accuracy": 0.9057543277740479,
      "num_tokens": 905077554.0,
      "step": 390
    },
    {
      "epoch": 0.2001023541453429,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.706362182022485e-05,
      "loss": 0.24183547496795654,
      "mean_token_accuracy": 0.9034228324890137,
      "num_tokens": 907462620.0,
      "step": 391
    },
    {
      "epoch": 0.20061412487205732,
      "grad_norm": 0.2265625,
      "learning_rate": 3.7046303378688955e-05,
      "loss": 0.24525779485702515,
      "mean_token_accuracy": 0.9025380611419678,
      "num_tokens": 909779706.0,
      "step": 392
    },
    {
      "epoch": 0.20112589559877175,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.7028938086955045e-05,
      "loss": 0.2386317253112793,
      "mean_token_accuracy": 0.9044380784034729,
      "num_tokens": 912089688.0,
      "step": 393
    },
    {
      "epoch": 0.2016376663254862,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.7011525992750036e-05,
      "loss": 0.2529582977294922,
      "mean_token_accuracy": 0.8988164663314819,
      "num_tokens": 914426956.0,
      "step": 394
    },
    {
      "epoch": 0.20214943705220062,
      "grad_norm": 0.240234375,
      "learning_rate": 3.699406714392946e-05,
      "loss": 0.24768118560314178,
      "mean_token_accuracy": 0.9013988971710205,
      "num_tokens": 916702998.0,
      "step": 395
    },
    {
      "epoch": 0.20266120777891505,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.697656158847737e-05,
      "loss": 0.2506449222564697,
      "mean_token_accuracy": 0.8997914791107178,
      "num_tokens": 919018440.0,
      "step": 396
    },
    {
      "epoch": 0.2031729785056295,
      "grad_norm": 0.212890625,
      "learning_rate": 3.695900937450615e-05,
      "loss": 0.2311909794807434,
      "mean_token_accuracy": 0.9084502458572388,
      "num_tokens": 921339794.0,
      "step": 397
    },
    {
      "epoch": 0.20368474923234392,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.6941410550256474e-05,
      "loss": 0.24381963908672333,
      "mean_token_accuracy": 0.90239018201828,
      "num_tokens": 923695494.0,
      "step": 398
    },
    {
      "epoch": 0.20419651995905835,
      "grad_norm": 0.240234375,
      "learning_rate": 3.692376516409709e-05,
      "loss": 0.23586870729923248,
      "mean_token_accuracy": 0.9052355289459229,
      "num_tokens": 926013362.0,
      "step": 399
    },
    {
      "epoch": 0.2047082906857728,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.6906073264524705e-05,
      "loss": 0.23754334449768066,
      "mean_token_accuracy": 0.9052695035934448,
      "num_tokens": 928363190.0,
      "step": 400
    },
    {
      "epoch": 0.2052200614124872,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.688833490016389e-05,
      "loss": 0.2354251742362976,
      "mean_token_accuracy": 0.9057842493057251,
      "num_tokens": 930623445.0,
      "step": 401
    },
    {
      "epoch": 0.20573183213920163,
      "grad_norm": 0.236328125,
      "learning_rate": 3.687055011976691e-05,
      "loss": 0.24204140901565552,
      "mean_token_accuracy": 0.9032788276672363,
      "num_tokens": 932978445.0,
      "step": 402
    },
    {
      "epoch": 0.20624360286591606,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.685271897221359e-05,
      "loss": 0.24975529313087463,
      "mean_token_accuracy": 0.9002896547317505,
      "num_tokens": 935278271.0,
      "step": 403
    },
    {
      "epoch": 0.2067553735926305,
      "grad_norm": 0.228515625,
      "learning_rate": 3.683484150651121e-05,
      "loss": 0.236271932721138,
      "mean_token_accuracy": 0.9062798023223877,
      "num_tokens": 937605736.0,
      "step": 404
    },
    {
      "epoch": 0.20726714431934493,
      "grad_norm": 0.212890625,
      "learning_rate": 3.681691777179434e-05,
      "loss": 0.2312353551387787,
      "mean_token_accuracy": 0.9079550504684448,
      "num_tokens": 939945551.0,
      "step": 405
    },
    {
      "epoch": 0.20777891504605936,
      "grad_norm": 0.2470703125,
      "learning_rate": 3.679894781732471e-05,
      "loss": 0.2420002967119217,
      "mean_token_accuracy": 0.9036163091659546,
      "num_tokens": 942210245.0,
      "step": 406
    },
    {
      "epoch": 0.2082906857727738,
      "grad_norm": 0.2431640625,
      "learning_rate": 3.678093169249109e-05,
      "loss": 0.23406867682933807,
      "mean_token_accuracy": 0.9065150022506714,
      "num_tokens": 944587874.0,
      "step": 407
    },
    {
      "epoch": 0.20880245649948823,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.676286944680915e-05,
      "loss": 0.22046592831611633,
      "mean_token_accuracy": 0.91209876537323,
      "num_tokens": 946987563.0,
      "step": 408
    },
    {
      "epoch": 0.20931422722620266,
      "grad_norm": 0.234375,
      "learning_rate": 3.674476112992131e-05,
      "loss": 0.23706169426441193,
      "mean_token_accuracy": 0.9049497842788696,
      "num_tokens": 949308679.0,
      "step": 409
    },
    {
      "epoch": 0.2098259979529171,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.672660679159661e-05,
      "loss": 0.2504193186759949,
      "mean_token_accuracy": 0.9004606008529663,
      "num_tokens": 951632858.0,
      "step": 410
    },
    {
      "epoch": 0.21033776867963153,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.670840648173058e-05,
      "loss": 0.22704721987247467,
      "mean_token_accuracy": 0.9088304042816162,
      "num_tokens": 954005311.0,
      "step": 411
    },
    {
      "epoch": 0.21084953940634596,
      "grad_norm": 0.234375,
      "learning_rate": 3.66901602503451e-05,
      "loss": 0.2583014667034149,
      "mean_token_accuracy": 0.8969918489456177,
      "num_tokens": 956308945.0,
      "step": 412
    },
    {
      "epoch": 0.2113613101330604,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.6671868147588264e-05,
      "loss": 0.2257426679134369,
      "mean_token_accuracy": 0.90947425365448,
      "num_tokens": 958628752.0,
      "step": 413
    },
    {
      "epoch": 0.21187308085977483,
      "grad_norm": 0.2490234375,
      "learning_rate": 3.665353022373423e-05,
      "loss": 0.2552053928375244,
      "mean_token_accuracy": 0.898165762424469,
      "num_tokens": 960989968.0,
      "step": 414
    },
    {
      "epoch": 0.21238485158648926,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.6635146529183105e-05,
      "loss": 0.24765229225158691,
      "mean_token_accuracy": 0.9016796350479126,
      "num_tokens": 963302936.0,
      "step": 415
    },
    {
      "epoch": 0.2128966223132037,
      "grad_norm": 0.21484375,
      "learning_rate": 3.661671711446077e-05,
      "loss": 0.23721420764923096,
      "mean_token_accuracy": 0.9053990244865417,
      "num_tokens": 965587682.0,
      "step": 416
    },
    {
      "epoch": 0.21340839303991813,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.6598242030218794e-05,
      "loss": 0.240750253200531,
      "mean_token_accuracy": 0.9047630429267883,
      "num_tokens": 967956093.0,
      "step": 417
    },
    {
      "epoch": 0.21392016376663256,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.657972132723423e-05,
      "loss": 0.24396032094955444,
      "mean_token_accuracy": 0.9025865793228149,
      "num_tokens": 970305274.0,
      "step": 418
    },
    {
      "epoch": 0.214431934493347,
      "grad_norm": 0.21484375,
      "learning_rate": 3.6561155056409537e-05,
      "loss": 0.22670894861221313,
      "mean_token_accuracy": 0.9092156887054443,
      "num_tokens": 972630225.0,
      "step": 419
    },
    {
      "epoch": 0.21494370522006143,
      "grad_norm": 0.255859375,
      "learning_rate": 3.65425432687724e-05,
      "loss": 0.25253862142562866,
      "mean_token_accuracy": 0.8988114595413208,
      "num_tokens": 974867480.0,
      "step": 420
    },
    {
      "epoch": 0.21545547594677583,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.652388601547562e-05,
      "loss": 0.2465483546257019,
      "mean_token_accuracy": 0.9019961357116699,
      "num_tokens": 977280776.0,
      "step": 421
    },
    {
      "epoch": 0.21596724667349027,
      "grad_norm": 0.22265625,
      "learning_rate": 3.650518334779692e-05,
      "loss": 0.24689501523971558,
      "mean_token_accuracy": 0.9014147520065308,
      "num_tokens": 979620483.0,
      "step": 422
    },
    {
      "epoch": 0.2164790174002047,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.648643531713888e-05,
      "loss": 0.2553960680961609,
      "mean_token_accuracy": 0.8979020714759827,
      "num_tokens": 981938611.0,
      "step": 423
    },
    {
      "epoch": 0.21699078812691913,
      "grad_norm": 0.240234375,
      "learning_rate": 3.646764197502874e-05,
      "loss": 0.24548465013504028,
      "mean_token_accuracy": 0.901494026184082,
      "num_tokens": 984228267.0,
      "step": 424
    },
    {
      "epoch": 0.21750255885363357,
      "grad_norm": 0.2265625,
      "learning_rate": 3.644880337311826e-05,
      "loss": 0.24032264947891235,
      "mean_token_accuracy": 0.9044836163520813,
      "num_tokens": 986600824.0,
      "step": 425
    },
    {
      "epoch": 0.218014329580348,
      "grad_norm": 0.23046875,
      "learning_rate": 3.642991956318362e-05,
      "loss": 0.2415490746498108,
      "mean_token_accuracy": 0.9032592177391052,
      "num_tokens": 988882158.0,
      "step": 426
    },
    {
      "epoch": 0.21852610030706243,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.641099059712523e-05,
      "loss": 0.2364172637462616,
      "mean_token_accuracy": 0.9059511423110962,
      "num_tokens": 991293475.0,
      "step": 427
    },
    {
      "epoch": 0.21903787103377687,
      "grad_norm": 0.212890625,
      "learning_rate": 3.6392016526967606e-05,
      "loss": 0.23218265175819397,
      "mean_token_accuracy": 0.9073692560195923,
      "num_tokens": 993658940.0,
      "step": 428
    },
    {
      "epoch": 0.2195496417604913,
      "grad_norm": 0.220703125,
      "learning_rate": 3.637299740485925e-05,
      "loss": 0.24905461072921753,
      "mean_token_accuracy": 0.90073561668396,
      "num_tokens": 995960143.0,
      "step": 429
    },
    {
      "epoch": 0.22006141248720573,
      "grad_norm": 0.232421875,
      "learning_rate": 3.6353933283072456e-05,
      "loss": 0.23349443078041077,
      "mean_token_accuracy": 0.9054442644119263,
      "num_tokens": 998327876.0,
      "step": 430
    },
    {
      "epoch": 0.22057318321392017,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.633482421400322e-05,
      "loss": 0.24457991123199463,
      "mean_token_accuracy": 0.9029472470283508,
      "num_tokens": 1000670700.0,
      "step": 431
    },
    {
      "epoch": 0.2210849539406346,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.6315670250171066e-05,
      "loss": 0.24769577383995056,
      "mean_token_accuracy": 0.9014680981636047,
      "num_tokens": 1002916895.0,
      "step": 432
    },
    {
      "epoch": 0.22159672466734903,
      "grad_norm": 0.38671875,
      "learning_rate": 3.6296471444218895e-05,
      "loss": 0.2509588599205017,
      "mean_token_accuracy": 0.8990687131881714,
      "num_tokens": 1005259526.0,
      "step": 433
    },
    {
      "epoch": 0.22210849539406347,
      "grad_norm": 0.2265625,
      "learning_rate": 3.627722784891286e-05,
      "loss": 0.24002444744110107,
      "mean_token_accuracy": 0.9040156006813049,
      "num_tokens": 1007554084.0,
      "step": 434
    },
    {
      "epoch": 0.2226202661207779,
      "grad_norm": 0.234375,
      "learning_rate": 3.625793951714222e-05,
      "loss": 0.2349294126033783,
      "mean_token_accuracy": 0.9065636396408081,
      "num_tokens": 1009878846.0,
      "step": 435
    },
    {
      "epoch": 0.22313203684749233,
      "grad_norm": 0.220703125,
      "learning_rate": 3.623860650191917e-05,
      "loss": 0.24390318989753723,
      "mean_token_accuracy": 0.9037121534347534,
      "num_tokens": 1012239593.0,
      "step": 436
    },
    {
      "epoch": 0.22364380757420677,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.621922885637874e-05,
      "loss": 0.23274153470993042,
      "mean_token_accuracy": 0.9068530797958374,
      "num_tokens": 1014624602.0,
      "step": 437
    },
    {
      "epoch": 0.2241555783009212,
      "grad_norm": 0.2421875,
      "learning_rate": 3.619980663377861e-05,
      "loss": 0.24109569191932678,
      "mean_token_accuracy": 0.9034782648086548,
      "num_tokens": 1016945426.0,
      "step": 438
    },
    {
      "epoch": 0.22466734902763563,
      "grad_norm": 0.232421875,
      "learning_rate": 3.6180339887498953e-05,
      "loss": 0.2557511329650879,
      "mean_token_accuracy": 0.8983403444290161,
      "num_tokens": 1019195051.0,
      "step": 439
    },
    {
      "epoch": 0.22517911975435004,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.616082867104235e-05,
      "loss": 0.24063870310783386,
      "mean_token_accuracy": 0.9045221209526062,
      "num_tokens": 1021457892.0,
      "step": 440
    },
    {
      "epoch": 0.22569089048106447,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.6141273038033574e-05,
      "loss": 0.2539862394332886,
      "mean_token_accuracy": 0.8991597294807434,
      "num_tokens": 1023768112.0,
      "step": 441
    },
    {
      "epoch": 0.2262026612077789,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.612167304221949e-05,
      "loss": 0.24151253700256348,
      "mean_token_accuracy": 0.9027745723724365,
      "num_tokens": 1026133126.0,
      "step": 442
    },
    {
      "epoch": 0.22671443193449334,
      "grad_norm": 0.228515625,
      "learning_rate": 3.610202873746889e-05,
      "loss": 0.24297219514846802,
      "mean_token_accuracy": 0.9028961062431335,
      "num_tokens": 1028380449.0,
      "step": 443
    },
    {
      "epoch": 0.22722620266120777,
      "grad_norm": 0.224609375,
      "learning_rate": 3.6082340177772335e-05,
      "loss": 0.2419261932373047,
      "mean_token_accuracy": 0.9034960865974426,
      "num_tokens": 1030699149.0,
      "step": 444
    },
    {
      "epoch": 0.2277379733879222,
      "grad_norm": 0.2060546875,
      "learning_rate": 3.606260741724202e-05,
      "loss": 0.22728073596954346,
      "mean_token_accuracy": 0.9080263376235962,
      "num_tokens": 1033069875.0,
      "step": 445
    },
    {
      "epoch": 0.22824974411463664,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.6042830510111635e-05,
      "loss": 0.2607341408729553,
      "mean_token_accuracy": 0.8959197998046875,
      "num_tokens": 1035352997.0,
      "step": 446
    },
    {
      "epoch": 0.22876151484135107,
      "grad_norm": 0.248046875,
      "learning_rate": 3.602300951073617e-05,
      "loss": 0.2586715817451477,
      "mean_token_accuracy": 0.8966549634933472,
      "num_tokens": 1037665166.0,
      "step": 447
    },
    {
      "epoch": 0.2292732855680655,
      "grad_norm": 0.21484375,
      "learning_rate": 3.600314447359184e-05,
      "loss": 0.22965756058692932,
      "mean_token_accuracy": 0.9084504246711731,
      "num_tokens": 1040074501.0,
      "step": 448
    },
    {
      "epoch": 0.22978505629477994,
      "grad_norm": 0.232421875,
      "learning_rate": 3.5983235453275865e-05,
      "loss": 0.24916194379329681,
      "mean_token_accuracy": 0.9002809524536133,
      "num_tokens": 1042429724.0,
      "step": 449
    },
    {
      "epoch": 0.23029682702149437,
      "grad_norm": 0.2109375,
      "learning_rate": 3.596328250450636e-05,
      "loss": 0.2383897453546524,
      "mean_token_accuracy": 0.9046846628189087,
      "num_tokens": 1044861990.0,
      "step": 450
    },
    {
      "epoch": 0.2308085977482088,
      "grad_norm": 0.22265625,
      "learning_rate": 3.594328568212216e-05,
      "loss": 0.24011099338531494,
      "mean_token_accuracy": 0.903791069984436,
      "num_tokens": 1047119901.0,
      "step": 451
    },
    {
      "epoch": 0.23132036847492324,
      "grad_norm": 0.921875,
      "learning_rate": 3.592324504108271e-05,
      "loss": 0.2562912702560425,
      "mean_token_accuracy": 0.8995200395584106,
      "num_tokens": 1049407890.0,
      "step": 452
    },
    {
      "epoch": 0.23183213920163767,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.5903160636467864e-05,
      "loss": 0.2566264271736145,
      "mean_token_accuracy": 0.8979303240776062,
      "num_tokens": 1051717968.0,
      "step": 453
    },
    {
      "epoch": 0.2323439099283521,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.5883032523477755e-05,
      "loss": 0.246356800198555,
      "mean_token_accuracy": 0.9017155766487122,
      "num_tokens": 1053981750.0,
      "step": 454
    },
    {
      "epoch": 0.23285568065506654,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.586286075743266e-05,
      "loss": 0.2518104910850525,
      "mean_token_accuracy": 0.8998588919639587,
      "num_tokens": 1056339525.0,
      "step": 455
    },
    {
      "epoch": 0.23336745138178097,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.584264539377282e-05,
      "loss": 0.23439809679985046,
      "mean_token_accuracy": 0.9058735370635986,
      "num_tokens": 1058655419.0,
      "step": 456
    },
    {
      "epoch": 0.2338792221084954,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.58223864880583e-05,
      "loss": 0.24640057981014252,
      "mean_token_accuracy": 0.90175461769104,
      "num_tokens": 1060988375.0,
      "step": 457
    },
    {
      "epoch": 0.23439099283520984,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.580208409596885e-05,
      "loss": 0.24519798159599304,
      "mean_token_accuracy": 0.9008749723434448,
      "num_tokens": 1063328398.0,
      "step": 458
    },
    {
      "epoch": 0.23490276356192427,
      "grad_norm": 0.38671875,
      "learning_rate": 3.578173827330372e-05,
      "loss": 0.2299843728542328,
      "mean_token_accuracy": 0.9083817005157471,
      "num_tokens": 1065662527.0,
      "step": 459
    },
    {
      "epoch": 0.23541453428863868,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.576134907598154e-05,
      "loss": 0.24683886766433716,
      "mean_token_accuracy": 0.9021903276443481,
      "num_tokens": 1067952451.0,
      "step": 460
    },
    {
      "epoch": 0.2359263050153531,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.5740916560040145e-05,
      "loss": 0.2639543116092682,
      "mean_token_accuracy": 0.8951137065887451,
      "num_tokens": 1070282164.0,
      "step": 461
    },
    {
      "epoch": 0.23643807574206754,
      "grad_norm": 0.244140625,
      "learning_rate": 3.5720440781636414e-05,
      "loss": 0.24139979481697083,
      "mean_token_accuracy": 0.9032844305038452,
      "num_tokens": 1072607845.0,
      "step": 462
    },
    {
      "epoch": 0.23694984646878198,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.5699921797046156e-05,
      "loss": 0.2473992109298706,
      "mean_token_accuracy": 0.9018906354904175,
      "num_tokens": 1074925338.0,
      "step": 463
    },
    {
      "epoch": 0.2374616171954964,
      "grad_norm": 0.84375,
      "learning_rate": 3.5679359662663904e-05,
      "loss": 0.24547743797302246,
      "mean_token_accuracy": 0.9030252695083618,
      "num_tokens": 1077130195.0,
      "step": 464
    },
    {
      "epoch": 0.23797338792221084,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.56587544350028e-05,
      "loss": 0.2349417805671692,
      "mean_token_accuracy": 0.9058421850204468,
      "num_tokens": 1079393219.0,
      "step": 465
    },
    {
      "epoch": 0.23848515864892528,
      "grad_norm": 0.228515625,
      "learning_rate": 3.563810617069443e-05,
      "loss": 0.24476391077041626,
      "mean_token_accuracy": 0.9038570523262024,
      "num_tokens": 1081749847.0,
      "step": 466
    },
    {
      "epoch": 0.2389969293756397,
      "grad_norm": 0.22265625,
      "learning_rate": 3.561741492648863e-05,
      "loss": 0.24334202706813812,
      "mean_token_accuracy": 0.9022039175033569,
      "num_tokens": 1083980046.0,
      "step": 467
    },
    {
      "epoch": 0.23950870010235414,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.5596680759253394e-05,
      "loss": 0.2295098602771759,
      "mean_token_accuracy": 0.907668948173523,
      "num_tokens": 1086236824.0,
      "step": 468
    },
    {
      "epoch": 0.24002047082906858,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.5575903725974674e-05,
      "loss": 0.2492680698633194,
      "mean_token_accuracy": 0.901369571685791,
      "num_tokens": 1088565123.0,
      "step": 469
    },
    {
      "epoch": 0.240532241555783,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.555508388375623e-05,
      "loss": 0.24024854600429535,
      "mean_token_accuracy": 0.9044923782348633,
      "num_tokens": 1090941733.0,
      "step": 470
    },
    {
      "epoch": 0.24104401228249744,
      "grad_norm": 0.228515625,
      "learning_rate": 3.553422128981949e-05,
      "loss": 0.23286883533000946,
      "mean_token_accuracy": 0.9070909023284912,
      "num_tokens": 1093214709.0,
      "step": 471
    },
    {
      "epoch": 0.24155578300921188,
      "grad_norm": 0.220703125,
      "learning_rate": 3.5513316001503376e-05,
      "loss": 0.24813956022262573,
      "mean_token_accuracy": 0.9018540382385254,
      "num_tokens": 1095491943.0,
      "step": 472
    },
    {
      "epoch": 0.2420675537359263,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.5492368076264155e-05,
      "loss": 0.2413458228111267,
      "mean_token_accuracy": 0.9036571979522705,
      "num_tokens": 1097809176.0,
      "step": 473
    },
    {
      "epoch": 0.24257932446264074,
      "grad_norm": 0.2578125,
      "learning_rate": 3.5471377571675266e-05,
      "loss": 0.249013289809227,
      "mean_token_accuracy": 0.9014924168586731,
      "num_tokens": 1100138390.0,
      "step": 474
    },
    {
      "epoch": 0.24309109518935518,
      "grad_norm": 0.23046875,
      "learning_rate": 3.545034454542718e-05,
      "loss": 0.24226942658424377,
      "mean_token_accuracy": 0.9032390117645264,
      "num_tokens": 1102483745.0,
      "step": 475
    },
    {
      "epoch": 0.2436028659160696,
      "grad_norm": 0.22265625,
      "learning_rate": 3.542926905532725e-05,
      "loss": 0.2264949083328247,
      "mean_token_accuracy": 0.9103409647941589,
      "num_tokens": 1104754745.0,
      "step": 476
    },
    {
      "epoch": 0.24411463664278404,
      "grad_norm": 0.28515625,
      "learning_rate": 3.5408151159299505e-05,
      "loss": 0.244118332862854,
      "mean_token_accuracy": 0.9027391672134399,
      "num_tokens": 1107021378.0,
      "step": 477
    },
    {
      "epoch": 0.24462640736949848,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.538699091538456e-05,
      "loss": 0.24700713157653809,
      "mean_token_accuracy": 0.9011213779449463,
      "num_tokens": 1109338734.0,
      "step": 478
    },
    {
      "epoch": 0.24513817809621288,
      "grad_norm": 0.24609375,
      "learning_rate": 3.536578838173939e-05,
      "loss": 0.2473280131816864,
      "mean_token_accuracy": 0.9009669423103333,
      "num_tokens": 1111622688.0,
      "step": 479
    },
    {
      "epoch": 0.24564994882292732,
      "grad_norm": 0.24609375,
      "learning_rate": 3.5344543616637205e-05,
      "loss": 0.24605457484722137,
      "mean_token_accuracy": 0.9022561311721802,
      "num_tokens": 1113926195.0,
      "step": 480
    },
    {
      "epoch": 0.24616171954964175,
      "grad_norm": 0.251953125,
      "learning_rate": 3.5323256678467306e-05,
      "loss": 0.2537941634654999,
      "mean_token_accuracy": 0.8988264799118042,
      "num_tokens": 1116276788.0,
      "step": 481
    },
    {
      "epoch": 0.24667349027635618,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.5301927625734865e-05,
      "loss": 0.25913769006729126,
      "mean_token_accuracy": 0.897028923034668,
      "num_tokens": 1118578720.0,
      "step": 482
    },
    {
      "epoch": 0.24718526100307062,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.5280556517060836e-05,
      "loss": 0.24431279301643372,
      "mean_token_accuracy": 0.9028832912445068,
      "num_tokens": 1120898863.0,
      "step": 483
    },
    {
      "epoch": 0.24769703172978505,
      "grad_norm": 0.2490234375,
      "learning_rate": 3.525914341118175e-05,
      "loss": 0.2524384558200836,
      "mean_token_accuracy": 0.9002180099487305,
      "num_tokens": 1123224688.0,
      "step": 484
    },
    {
      "epoch": 0.24820880245649948,
      "grad_norm": 0.24609375,
      "learning_rate": 3.523768836694955e-05,
      "loss": 0.2564295828342438,
      "mean_token_accuracy": 0.8975832462310791,
      "num_tokens": 1125478783.0,
      "step": 485
    },
    {
      "epoch": 0.24872057318321392,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.521619144333145e-05,
      "loss": 0.22366863489151,
      "mean_token_accuracy": 0.91010582447052,
      "num_tokens": 1127785566.0,
      "step": 486
    },
    {
      "epoch": 0.24923234390992835,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.519465269940979e-05,
      "loss": 0.2406628429889679,
      "mean_token_accuracy": 0.9041343927383423,
      "num_tokens": 1130155980.0,
      "step": 487
    },
    {
      "epoch": 0.24974411463664278,
      "grad_norm": 0.216796875,
      "learning_rate": 3.5173072194381804e-05,
      "loss": 0.22920985519886017,
      "mean_token_accuracy": 0.9089385867118835,
      "num_tokens": 1132475086.0,
      "step": 488
    },
    {
      "epoch": 0.2502558853633572,
      "grad_norm": 0.224609375,
      "learning_rate": 3.5151449987559546e-05,
      "loss": 0.23717355728149414,
      "mean_token_accuracy": 0.9053545594215393,
      "num_tokens": 1134732629.0,
      "step": 489
    },
    {
      "epoch": 0.2507676560900716,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.512978613836965e-05,
      "loss": 0.24618548154830933,
      "mean_token_accuracy": 0.9020051956176758,
      "num_tokens": 1137042806.0,
      "step": 490
    },
    {
      "epoch": 0.25127942681678606,
      "grad_norm": 0.2412109375,
      "learning_rate": 3.510808070635322e-05,
      "loss": 0.2464095950126648,
      "mean_token_accuracy": 0.9017961025238037,
      "num_tokens": 1139380266.0,
      "step": 491
    },
    {
      "epoch": 0.2517911975435005,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.5086333751165636e-05,
      "loss": 0.23715856671333313,
      "mean_token_accuracy": 0.9050180912017822,
      "num_tokens": 1141764884.0,
      "step": 492
    },
    {
      "epoch": 0.2523029682702149,
      "grad_norm": 0.2109375,
      "learning_rate": 3.5064545332576414e-05,
      "loss": 0.23294416069984436,
      "mean_token_accuracy": 0.9067354202270508,
      "num_tokens": 1144127274.0,
      "step": 493
    },
    {
      "epoch": 0.25281473899692936,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.504271551046901e-05,
      "loss": 0.23124143481254578,
      "mean_token_accuracy": 0.9086431264877319,
      "num_tokens": 1146454468.0,
      "step": 494
    },
    {
      "epoch": 0.2533265097236438,
      "grad_norm": 0.234375,
      "learning_rate": 3.502084434484069e-05,
      "loss": 0.2530425190925598,
      "mean_token_accuracy": 0.8996856808662415,
      "num_tokens": 1148784975.0,
      "step": 495
    },
    {
      "epoch": 0.2538382804503582,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.499893189580234e-05,
      "loss": 0.23520028591156006,
      "mean_token_accuracy": 0.9057403802871704,
      "num_tokens": 1151123966.0,
      "step": 496
    },
    {
      "epoch": 0.25435005117707266,
      "grad_norm": 0.22265625,
      "learning_rate": 3.497697822357832e-05,
      "loss": 0.22223903238773346,
      "mean_token_accuracy": 0.910506010055542,
      "num_tokens": 1153375488.0,
      "step": 497
    },
    {
      "epoch": 0.2548618219037871,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.4954983388506264e-05,
      "loss": 0.2403850108385086,
      "mean_token_accuracy": 0.9043492078781128,
      "num_tokens": 1155706987.0,
      "step": 498
    },
    {
      "epoch": 0.2553735926305015,
      "grad_norm": 0.224609375,
      "learning_rate": 3.493294745103697e-05,
      "loss": 0.2553556561470032,
      "mean_token_accuracy": 0.8980218172073364,
      "num_tokens": 1158011746.0,
      "step": 499
    },
    {
      "epoch": 0.25588536335721596,
      "grad_norm": 0.208984375,
      "learning_rate": 3.491087047173418e-05,
      "loss": 0.23822295665740967,
      "mean_token_accuracy": 0.9052966833114624,
      "num_tokens": 1160325657.0,
      "step": 500
    },
    {
      "epoch": 0.2563971340839304,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.488875251127444e-05,
      "loss": 0.2465994954109192,
      "mean_token_accuracy": 0.9017128944396973,
      "num_tokens": 1162618357.0,
      "step": 501
    },
    {
      "epoch": 0.2569089048106448,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.486659363044693e-05,
      "loss": 0.23991400003433228,
      "mean_token_accuracy": 0.9041088819503784,
      "num_tokens": 1164900858.0,
      "step": 502
    },
    {
      "epoch": 0.25742067553735926,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.48443938901533e-05,
      "loss": 0.2441604882478714,
      "mean_token_accuracy": 0.9028209447860718,
      "num_tokens": 1167180046.0,
      "step": 503
    },
    {
      "epoch": 0.2579324462640737,
      "grad_norm": 0.265625,
      "learning_rate": 3.4822153351407513e-05,
      "loss": 0.24480074644088745,
      "mean_token_accuracy": 0.9013353586196899,
      "num_tokens": 1169502231.0,
      "step": 504
    },
    {
      "epoch": 0.2584442169907881,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.4799872075335614e-05,
      "loss": 0.238721564412117,
      "mean_token_accuracy": 0.9039852619171143,
      "num_tokens": 1171786017.0,
      "step": 505
    },
    {
      "epoch": 0.25895598771750256,
      "grad_norm": 0.23828125,
      "learning_rate": 3.477755012317567e-05,
      "loss": 0.2518821358680725,
      "mean_token_accuracy": 0.9000849723815918,
      "num_tokens": 1174132836.0,
      "step": 506
    },
    {
      "epoch": 0.259467758444217,
      "grad_norm": 0.23828125,
      "learning_rate": 3.475518755627749e-05,
      "loss": 0.2544906735420227,
      "mean_token_accuracy": 0.8981234431266785,
      "num_tokens": 1176523460.0,
      "step": 507
    },
    {
      "epoch": 0.2599795291709314,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.473278443610256e-05,
      "loss": 0.25335317850112915,
      "mean_token_accuracy": 0.8988422155380249,
      "num_tokens": 1178832165.0,
      "step": 508
    },
    {
      "epoch": 0.26049129989764586,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.4710340824223774e-05,
      "loss": 0.25439125299453735,
      "mean_token_accuracy": 0.8989884257316589,
      "num_tokens": 1181119852.0,
      "step": 509
    },
    {
      "epoch": 0.2610030706243603,
      "grad_norm": 0.234375,
      "learning_rate": 3.4687856782325346e-05,
      "loss": 0.2469266653060913,
      "mean_token_accuracy": 0.9005539417266846,
      "num_tokens": 1183347179.0,
      "step": 510
    },
    {
      "epoch": 0.2615148413510747,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.46653323722026e-05,
      "loss": 0.2617835998535156,
      "mean_token_accuracy": 0.8959383964538574,
      "num_tokens": 1185637405.0,
      "step": 511
    },
    {
      "epoch": 0.26202661207778916,
      "grad_norm": 0.224609375,
      "learning_rate": 3.46427676557618e-05,
      "loss": 0.24276477098464966,
      "mean_token_accuracy": 0.9031580686569214,
      "num_tokens": 1188008251.0,
      "step": 512
    },
    {
      "epoch": 0.2625383828045036,
      "grad_norm": 0.23828125,
      "learning_rate": 3.462016269502e-05,
      "loss": 0.2432727813720703,
      "mean_token_accuracy": 0.9025658369064331,
      "num_tokens": 1190295675.0,
      "step": 513
    },
    {
      "epoch": 0.263050153531218,
      "grad_norm": 0.23828125,
      "learning_rate": 3.459751755210486e-05,
      "loss": 0.25503653287887573,
      "mean_token_accuracy": 0.8984668850898743,
      "num_tokens": 1192610458.0,
      "step": 514
    },
    {
      "epoch": 0.26356192425793246,
      "grad_norm": 0.22265625,
      "learning_rate": 3.4574832289254474e-05,
      "loss": 0.24925944209098816,
      "mean_token_accuracy": 0.9004257917404175,
      "num_tokens": 1194973228.0,
      "step": 515
    },
    {
      "epoch": 0.2640736949846469,
      "grad_norm": 0.216796875,
      "learning_rate": 3.455210696881719e-05,
      "loss": 0.24223284423351288,
      "mean_token_accuracy": 0.9032938480377197,
      "num_tokens": 1197315802.0,
      "step": 516
    },
    {
      "epoch": 0.2645854657113613,
      "grad_norm": 0.2265625,
      "learning_rate": 3.4529341653251474e-05,
      "loss": 0.25214481353759766,
      "mean_token_accuracy": 0.899933934211731,
      "num_tokens": 1199669089.0,
      "step": 517
    },
    {
      "epoch": 0.26509723643807576,
      "grad_norm": 0.232421875,
      "learning_rate": 3.45065364051257e-05,
      "loss": 0.24507631361484528,
      "mean_token_accuracy": 0.9011919498443604,
      "num_tokens": 1201916236.0,
      "step": 518
    },
    {
      "epoch": 0.2656090071647902,
      "grad_norm": 0.212890625,
      "learning_rate": 3.448369128711799e-05,
      "loss": 0.23652759194374084,
      "mean_token_accuracy": 0.9058696031570435,
      "num_tokens": 1204172976.0,
      "step": 519
    },
    {
      "epoch": 0.2661207778915046,
      "grad_norm": 0.2265625,
      "learning_rate": 3.4460806362016064e-05,
      "loss": 0.2389504611492157,
      "mean_token_accuracy": 0.9031352996826172,
      "num_tokens": 1206576642.0,
      "step": 520
    },
    {
      "epoch": 0.26663254861821906,
      "grad_norm": 0.220703125,
      "learning_rate": 3.4437881692717035e-05,
      "loss": 0.23572736978530884,
      "mean_token_accuracy": 0.9057648181915283,
      "num_tokens": 1208862205.0,
      "step": 521
    },
    {
      "epoch": 0.2671443193449335,
      "grad_norm": 0.23046875,
      "learning_rate": 3.441491734222723e-05,
      "loss": 0.2421688437461853,
      "mean_token_accuracy": 0.9025044441223145,
      "num_tokens": 1211213188.0,
      "step": 522
    },
    {
      "epoch": 0.2676560900716479,
      "grad_norm": 0.220703125,
      "learning_rate": 3.439191337366208e-05,
      "loss": 0.24411529302597046,
      "mean_token_accuracy": 0.9024232625961304,
      "num_tokens": 1213613706.0,
      "step": 523
    },
    {
      "epoch": 0.26816786079836236,
      "grad_norm": 0.33984375,
      "learning_rate": 3.436886985024586e-05,
      "loss": 0.2461850792169571,
      "mean_token_accuracy": 0.9020059108734131,
      "num_tokens": 1215901562.0,
      "step": 524
    },
    {
      "epoch": 0.2686796315250768,
      "grad_norm": 0.20703125,
      "learning_rate": 3.4345786835311584e-05,
      "loss": 0.22401735186576843,
      "mean_token_accuracy": 0.9105480909347534,
      "num_tokens": 1218218246.0,
      "step": 525
    },
    {
      "epoch": 0.2691914022517912,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.432266439230079e-05,
      "loss": 0.23760315775871277,
      "mean_token_accuracy": 0.9052151441574097,
      "num_tokens": 1220526144.0,
      "step": 526
    },
    {
      "epoch": 0.26970317297850566,
      "grad_norm": 0.220703125,
      "learning_rate": 3.42995025847634e-05,
      "loss": 0.258918434381485,
      "mean_token_accuracy": 0.8972591161727905,
      "num_tokens": 1222864351.0,
      "step": 527
    },
    {
      "epoch": 0.27021494370522003,
      "grad_norm": 0.203125,
      "learning_rate": 3.42763014763575e-05,
      "loss": 0.2253480851650238,
      "mean_token_accuracy": 0.9104202389717102,
      "num_tokens": 1225219501.0,
      "step": 528
    },
    {
      "epoch": 0.27072671443193447,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.4253061130849215e-05,
      "loss": 0.26072022318840027,
      "mean_token_accuracy": 0.8957808017730713,
      "num_tokens": 1227534580.0,
      "step": 529
    },
    {
      "epoch": 0.2712384851586489,
      "grad_norm": 0.2890625,
      "learning_rate": 3.422978161211249e-05,
      "loss": 0.23931275308132172,
      "mean_token_accuracy": 0.9046899676322937,
      "num_tokens": 1229803686.0,
      "step": 530
    },
    {
      "epoch": 0.27175025588536333,
      "grad_norm": 0.220703125,
      "learning_rate": 3.4206462984128944e-05,
      "loss": 0.24310852587223053,
      "mean_token_accuracy": 0.9028220176696777,
      "num_tokens": 1232164947.0,
      "step": 531
    },
    {
      "epoch": 0.27226202661207777,
      "grad_norm": 0.216796875,
      "learning_rate": 3.418310531098769e-05,
      "loss": 0.24602502584457397,
      "mean_token_accuracy": 0.9020251035690308,
      "num_tokens": 1234546278.0,
      "step": 532
    },
    {
      "epoch": 0.2727737973387922,
      "grad_norm": 0.228515625,
      "learning_rate": 3.4159708656885136e-05,
      "loss": 0.24877920746803284,
      "mean_token_accuracy": 0.9005723595619202,
      "num_tokens": 1236843261.0,
      "step": 533
    },
    {
      "epoch": 0.27328556806550663,
      "grad_norm": 0.22265625,
      "learning_rate": 3.413627308612484e-05,
      "loss": 0.23154538869857788,
      "mean_token_accuracy": 0.9082626104354858,
      "num_tokens": 1239114489.0,
      "step": 534
    },
    {
      "epoch": 0.27379733879222107,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.411279866311732e-05,
      "loss": 0.2536682188510895,
      "mean_token_accuracy": 0.8990767002105713,
      "num_tokens": 1241548148.0,
      "step": 535
    },
    {
      "epoch": 0.2743091095189355,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.408928545237985e-05,
      "loss": 0.24886426329612732,
      "mean_token_accuracy": 0.9012750387191772,
      "num_tokens": 1243942070.0,
      "step": 536
    },
    {
      "epoch": 0.27482088024564993,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.406573351853635e-05,
      "loss": 0.2454976886510849,
      "mean_token_accuracy": 0.9024184346199036,
      "num_tokens": 1246207719.0,
      "step": 537
    },
    {
      "epoch": 0.27533265097236437,
      "grad_norm": 0.52734375,
      "learning_rate": 3.4042142926317134e-05,
      "loss": 0.234938383102417,
      "mean_token_accuracy": 0.9060052633285522,
      "num_tokens": 1248502688.0,
      "step": 538
    },
    {
      "epoch": 0.2758444216990788,
      "grad_norm": 0.21875,
      "learning_rate": 3.401851374055879e-05,
      "loss": 0.23078405857086182,
      "mean_token_accuracy": 0.9075149893760681,
      "num_tokens": 1250835289.0,
      "step": 539
    },
    {
      "epoch": 0.27635619242579323,
      "grad_norm": 0.216796875,
      "learning_rate": 3.3994846026203934e-05,
      "loss": 0.22821998596191406,
      "mean_token_accuracy": 0.9081479907035828,
      "num_tokens": 1253246841.0,
      "step": 540
    },
    {
      "epoch": 0.27686796315250767,
      "grad_norm": 0.216796875,
      "learning_rate": 3.3971139848301136e-05,
      "loss": 0.2527444362640381,
      "mean_token_accuracy": 0.8992606401443481,
      "num_tokens": 1255577196.0,
      "step": 541
    },
    {
      "epoch": 0.2773797338792221,
      "grad_norm": 0.212890625,
      "learning_rate": 3.3947395272004633e-05,
      "loss": 0.231692373752594,
      "mean_token_accuracy": 0.9075101613998413,
      "num_tokens": 1257865048.0,
      "step": 542
    },
    {
      "epoch": 0.27789150460593653,
      "grad_norm": 0.22265625,
      "learning_rate": 3.39236123625742e-05,
      "loss": 0.2429943084716797,
      "mean_token_accuracy": 0.903406023979187,
      "num_tokens": 1260132654.0,
      "step": 543
    },
    {
      "epoch": 0.27840327533265097,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.3899791185375e-05,
      "loss": 0.25038912892341614,
      "mean_token_accuracy": 0.8995441198348999,
      "num_tokens": 1262481342.0,
      "step": 544
    },
    {
      "epoch": 0.2789150460593654,
      "grad_norm": 0.22265625,
      "learning_rate": 3.3875931805877334e-05,
      "loss": 0.24212628602981567,
      "mean_token_accuracy": 0.9032609462738037,
      "num_tokens": 1264830197.0,
      "step": 545
    },
    {
      "epoch": 0.27942681678607983,
      "grad_norm": 0.224609375,
      "learning_rate": 3.3852034289656516e-05,
      "loss": 0.2441590279340744,
      "mean_token_accuracy": 0.9015325903892517,
      "num_tokens": 1267202221.0,
      "step": 546
    },
    {
      "epoch": 0.27993858751279427,
      "grad_norm": 0.22265625,
      "learning_rate": 3.3828098702392676e-05,
      "loss": 0.2414209097623825,
      "mean_token_accuracy": 0.9036802053451538,
      "num_tokens": 1269542457.0,
      "step": 547
    },
    {
      "epoch": 0.2804503582395087,
      "grad_norm": 0.21484375,
      "learning_rate": 3.380412510987057e-05,
      "loss": 0.23633268475532532,
      "mean_token_accuracy": 0.906451404094696,
      "num_tokens": 1271864308.0,
      "step": 548
    },
    {
      "epoch": 0.28096212896622313,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.378011357797942e-05,
      "loss": 0.23326073586940765,
      "mean_token_accuracy": 0.907046377658844,
      "num_tokens": 1274111755.0,
      "step": 549
    },
    {
      "epoch": 0.28147389969293757,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.37560641727127e-05,
      "loss": 0.24384915828704834,
      "mean_token_accuracy": 0.9025110006332397,
      "num_tokens": 1276450354.0,
      "step": 550
    },
    {
      "epoch": 0.281985670419652,
      "grad_norm": 0.21875,
      "learning_rate": 3.373197696016801e-05,
      "loss": 0.2380177229642868,
      "mean_token_accuracy": 0.904766321182251,
      "num_tokens": 1278845610.0,
      "step": 551
    },
    {
      "epoch": 0.28249744114636643,
      "grad_norm": 0.22265625,
      "learning_rate": 3.370785200654683e-05,
      "loss": 0.23893848061561584,
      "mean_token_accuracy": 0.9042294025421143,
      "num_tokens": 1281133434.0,
      "step": 552
    },
    {
      "epoch": 0.28300921187308087,
      "grad_norm": 0.212890625,
      "learning_rate": 3.368368937815438e-05,
      "loss": 0.24242790043354034,
      "mean_token_accuracy": 0.9030978679656982,
      "num_tokens": 1283503791.0,
      "step": 553
    },
    {
      "epoch": 0.2835209825997953,
      "grad_norm": 0.21875,
      "learning_rate": 3.365948914139942e-05,
      "loss": 0.23803919553756714,
      "mean_token_accuracy": 0.905780553817749,
      "num_tokens": 1285871057.0,
      "step": 554
    },
    {
      "epoch": 0.28403275332650973,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.363525136279407e-05,
      "loss": 0.2506956160068512,
      "mean_token_accuracy": 0.899477481842041,
      "num_tokens": 1288257128.0,
      "step": 555
    },
    {
      "epoch": 0.28454452405322417,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.361097610895365e-05,
      "loss": 0.24204733967781067,
      "mean_token_accuracy": 0.9034696817398071,
      "num_tokens": 1290543377.0,
      "step": 556
    },
    {
      "epoch": 0.2850562947799386,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.358666344659646e-05,
      "loss": 0.24132026731967926,
      "mean_token_accuracy": 0.9031459093093872,
      "num_tokens": 1292883810.0,
      "step": 557
    },
    {
      "epoch": 0.28556806550665303,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.3562313442543616e-05,
      "loss": 0.24152176082134247,
      "mean_token_accuracy": 0.9031692743301392,
      "num_tokens": 1295201596.0,
      "step": 558
    },
    {
      "epoch": 0.28607983623336747,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.353792616371886e-05,
      "loss": 0.25351986289024353,
      "mean_token_accuracy": 0.8986018896102905,
      "num_tokens": 1297565510.0,
      "step": 559
    },
    {
      "epoch": 0.2865916069600819,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.3513501677148394e-05,
      "loss": 0.22850850224494934,
      "mean_token_accuracy": 0.9077621102333069,
      "num_tokens": 1299952326.0,
      "step": 560
    },
    {
      "epoch": 0.28710337768679633,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.348904004996068e-05,
      "loss": 0.2533406913280487,
      "mean_token_accuracy": 0.898931622505188,
      "num_tokens": 1302327454.0,
      "step": 561
    },
    {
      "epoch": 0.28761514841351077,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.3464541349386234e-05,
      "loss": 0.24140101671218872,
      "mean_token_accuracy": 0.903323233127594,
      "num_tokens": 1304568085.0,
      "step": 562
    },
    {
      "epoch": 0.2881269191402252,
      "grad_norm": 0.216796875,
      "learning_rate": 3.34400056427575e-05,
      "loss": 0.24925443530082703,
      "mean_token_accuracy": 0.9002532958984375,
      "num_tokens": 1306857364.0,
      "step": 563
    },
    {
      "epoch": 0.28863868986693964,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.341543299750861e-05,
      "loss": 0.23040398955345154,
      "mean_token_accuracy": 0.9080673456192017,
      "num_tokens": 1309216292.0,
      "step": 564
    },
    {
      "epoch": 0.28915046059365407,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.33908234811752e-05,
      "loss": 0.23440861701965332,
      "mean_token_accuracy": 0.9064846038818359,
      "num_tokens": 1311598206.0,
      "step": 565
    },
    {
      "epoch": 0.2896622313203685,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.336617716139429e-05,
      "loss": 0.2492506206035614,
      "mean_token_accuracy": 0.8995347023010254,
      "num_tokens": 1313900063.0,
      "step": 566
    },
    {
      "epoch": 0.2901740020470829,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.334149410590402e-05,
      "loss": 0.23845365643501282,
      "mean_token_accuracy": 0.9048078060150146,
      "num_tokens": 1316193791.0,
      "step": 567
    },
    {
      "epoch": 0.2906857727737973,
      "grad_norm": 0.2265625,
      "learning_rate": 3.331677438254349e-05,
      "loss": 0.24486872553825378,
      "mean_token_accuracy": 0.9024873375892639,
      "num_tokens": 1318472044.0,
      "step": 568
    },
    {
      "epoch": 0.29119754350051175,
      "grad_norm": 0.236328125,
      "learning_rate": 3.329201805925259e-05,
      "loss": 0.24889004230499268,
      "mean_token_accuracy": 0.900090217590332,
      "num_tokens": 1320753703.0,
      "step": 569
    },
    {
      "epoch": 0.2917093142272262,
      "grad_norm": 0.21484375,
      "learning_rate": 3.32672252040718e-05,
      "loss": 0.23213379085063934,
      "mean_token_accuracy": 0.9081496596336365,
      "num_tokens": 1323114712.0,
      "step": 570
    },
    {
      "epoch": 0.2922210849539406,
      "grad_norm": 0.2265625,
      "learning_rate": 3.3242395885142e-05,
      "loss": 0.24411334097385406,
      "mean_token_accuracy": 0.9030811786651611,
      "num_tokens": 1325452733.0,
      "step": 571
    },
    {
      "epoch": 0.29273285568065505,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.3217530170704305e-05,
      "loss": 0.2500269412994385,
      "mean_token_accuracy": 0.8997089862823486,
      "num_tokens": 1327768647.0,
      "step": 572
    },
    {
      "epoch": 0.2932446264073695,
      "grad_norm": 0.22265625,
      "learning_rate": 3.319262812909983e-05,
      "loss": 0.24306321144104004,
      "mean_token_accuracy": 0.9036771059036255,
      "num_tokens": 1330057787.0,
      "step": 573
    },
    {
      "epoch": 0.2937563971340839,
      "grad_norm": 0.234375,
      "learning_rate": 3.316768982876955e-05,
      "loss": 0.24905052781105042,
      "mean_token_accuracy": 0.9009923934936523,
      "num_tokens": 1332416042.0,
      "step": 574
    },
    {
      "epoch": 0.29426816786079835,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.3142715338254094e-05,
      "loss": 0.24036094546318054,
      "mean_token_accuracy": 0.9040223956108093,
      "num_tokens": 1334767376.0,
      "step": 575
    },
    {
      "epoch": 0.2947799385875128,
      "grad_norm": 0.2470703125,
      "learning_rate": 3.311770472619357e-05,
      "loss": 0.2547222375869751,
      "mean_token_accuracy": 0.8984191417694092,
      "num_tokens": 1337153478.0,
      "step": 576
    },
    {
      "epoch": 0.2952917093142272,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.309265806132731e-05,
      "loss": 0.25136467814445496,
      "mean_token_accuracy": 0.9003368616104126,
      "num_tokens": 1339496902.0,
      "step": 577
    },
    {
      "epoch": 0.29580348004094165,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.306757541249379e-05,
      "loss": 0.23967060446739197,
      "mean_token_accuracy": 0.9038301706314087,
      "num_tokens": 1341835100.0,
      "step": 578
    },
    {
      "epoch": 0.2963152507676561,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.304245684863038e-05,
      "loss": 0.23299473524093628,
      "mean_token_accuracy": 0.9068459272384644,
      "num_tokens": 1344206467.0,
      "step": 579
    },
    {
      "epoch": 0.2968270214943705,
      "grad_norm": 0.240234375,
      "learning_rate": 3.301730243877312e-05,
      "loss": 0.23509997129440308,
      "mean_token_accuracy": 0.9053818583488464,
      "num_tokens": 1346474415.0,
      "step": 580
    },
    {
      "epoch": 0.29733879222108495,
      "grad_norm": 0.234375,
      "learning_rate": 3.29921122520566e-05,
      "loss": 0.24475213885307312,
      "mean_token_accuracy": 0.9019044637680054,
      "num_tokens": 1348892237.0,
      "step": 581
    },
    {
      "epoch": 0.2978505629477994,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.296688635771374e-05,
      "loss": 0.2307039052248001,
      "mean_token_accuracy": 0.9076792001724243,
      "num_tokens": 1351199426.0,
      "step": 582
    },
    {
      "epoch": 0.2983623336745138,
      "grad_norm": 0.2265625,
      "learning_rate": 3.294162482507559e-05,
      "loss": 0.24835218489170074,
      "mean_token_accuracy": 0.9007115364074707,
      "num_tokens": 1353518768.0,
      "step": 583
    },
    {
      "epoch": 0.29887410440122825,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.291632772357114e-05,
      "loss": 0.24255964159965515,
      "mean_token_accuracy": 0.9030742645263672,
      "num_tokens": 1355864457.0,
      "step": 584
    },
    {
      "epoch": 0.2993858751279427,
      "grad_norm": 0.21484375,
      "learning_rate": 3.289099512272716e-05,
      "loss": 0.23161974549293518,
      "mean_token_accuracy": 0.9069007039070129,
      "num_tokens": 1358247031.0,
      "step": 585
    },
    {
      "epoch": 0.2998976458546571,
      "grad_norm": 0.23046875,
      "learning_rate": 3.286562709216798e-05,
      "loss": 0.24821877479553223,
      "mean_token_accuracy": 0.9006421566009521,
      "num_tokens": 1360528538.0,
      "step": 586
    },
    {
      "epoch": 0.30040941658137155,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.284022370161529e-05,
      "loss": 0.2462676763534546,
      "mean_token_accuracy": 0.9016884565353394,
      "num_tokens": 1362881248.0,
      "step": 587
    },
    {
      "epoch": 0.300921187308086,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.281478502088796e-05,
      "loss": 0.24905790388584137,
      "mean_token_accuracy": 0.9009596109390259,
      "num_tokens": 1365261144.0,
      "step": 588
    },
    {
      "epoch": 0.3014329580348004,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.278931111990189e-05,
      "loss": 0.23999112844467163,
      "mean_token_accuracy": 0.9040271639823914,
      "num_tokens": 1367621455.0,
      "step": 589
    },
    {
      "epoch": 0.30194472876151485,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.276380206866975e-05,
      "loss": 0.2396191656589508,
      "mean_token_accuracy": 0.9039372205734253,
      "num_tokens": 1369906544.0,
      "step": 590
    },
    {
      "epoch": 0.3024564994882293,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.2738257937300815e-05,
      "loss": 0.24827340245246887,
      "mean_token_accuracy": 0.9002362489700317,
      "num_tokens": 1372227634.0,
      "step": 591
    },
    {
      "epoch": 0.3029682702149437,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.2712678796000794e-05,
      "loss": 0.24329856038093567,
      "mean_token_accuracy": 0.9019630551338196,
      "num_tokens": 1374597864.0,
      "step": 592
    },
    {
      "epoch": 0.30348004094165815,
      "grad_norm": 0.232421875,
      "learning_rate": 3.2687064715071586e-05,
      "loss": 0.24841907620429993,
      "mean_token_accuracy": 0.9010759592056274,
      "num_tokens": 1376902959.0,
      "step": 593
    },
    {
      "epoch": 0.3039918116683726,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.266141576491116e-05,
      "loss": 0.2295304238796234,
      "mean_token_accuracy": 0.9080773591995239,
      "num_tokens": 1379252784.0,
      "step": 594
    },
    {
      "epoch": 0.304503582395087,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.263573201601328e-05,
      "loss": 0.23561805486679077,
      "mean_token_accuracy": 0.9063023328781128,
      "num_tokens": 1381583552.0,
      "step": 595
    },
    {
      "epoch": 0.30501535312180145,
      "grad_norm": 0.21484375,
      "learning_rate": 3.261001353896738e-05,
      "loss": 0.23630577325820923,
      "mean_token_accuracy": 0.9048407077789307,
      "num_tokens": 1383850242.0,
      "step": 596
    },
    {
      "epoch": 0.3055271238485159,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.258426040445833e-05,
      "loss": 0.22259709239006042,
      "mean_token_accuracy": 0.9110952019691467,
      "num_tokens": 1386217562.0,
      "step": 597
    },
    {
      "epoch": 0.3060388945752303,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.255847268326625e-05,
      "loss": 0.2513413727283478,
      "mean_token_accuracy": 0.8980776071548462,
      "num_tokens": 1388480354.0,
      "step": 598
    },
    {
      "epoch": 0.30655066530194475,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.2532650446266326e-05,
      "loss": 0.23419791460037231,
      "mean_token_accuracy": 0.9068004488945007,
      "num_tokens": 1390775311.0,
      "step": 599
    },
    {
      "epoch": 0.3070624360286592,
      "grad_norm": 0.25390625,
      "learning_rate": 3.2506793764428586e-05,
      "loss": 0.24252808094024658,
      "mean_token_accuracy": 0.9025326371192932,
      "num_tokens": 1393023930.0,
      "step": 600
    },
    {
      "epoch": 0.3075742067553736,
      "grad_norm": 0.232421875,
      "learning_rate": 3.2480902708817765e-05,
      "loss": 0.2411651611328125,
      "mean_token_accuracy": 0.9042178392410278,
      "num_tokens": 1395304364.0,
      "step": 601
    },
    {
      "epoch": 0.30808597748208805,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.245497735059304e-05,
      "loss": 0.24166527390480042,
      "mean_token_accuracy": 0.9035400748252869,
      "num_tokens": 1397723252.0,
      "step": 602
    },
    {
      "epoch": 0.3085977482088025,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.242901776100787e-05,
      "loss": 0.23842638731002808,
      "mean_token_accuracy": 0.9044324159622192,
      "num_tokens": 1400076624.0,
      "step": 603
    },
    {
      "epoch": 0.3091095189355169,
      "grad_norm": 0.228515625,
      "learning_rate": 3.240302401140982e-05,
      "loss": 0.2509409189224243,
      "mean_token_accuracy": 0.8990161418914795,
      "num_tokens": 1402341322.0,
      "step": 604
    },
    {
      "epoch": 0.30962128966223135,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.2376996173240295e-05,
      "loss": 0.2321169376373291,
      "mean_token_accuracy": 0.9071183204650879,
      "num_tokens": 1404729743.0,
      "step": 605
    },
    {
      "epoch": 0.3101330603889457,
      "grad_norm": 0.234375,
      "learning_rate": 3.235093431803444e-05,
      "loss": 0.25112560391426086,
      "mean_token_accuracy": 0.8997725248336792,
      "num_tokens": 1407123527.0,
      "step": 606
    },
    {
      "epoch": 0.31064483111566016,
      "grad_norm": 0.220703125,
      "learning_rate": 3.232483851742085e-05,
      "loss": 0.24101826548576355,
      "mean_token_accuracy": 0.9038769602775574,
      "num_tokens": 1409482017.0,
      "step": 607
    },
    {
      "epoch": 0.3111566018423746,
      "grad_norm": 0.228515625,
      "learning_rate": 3.229870884312145e-05,
      "loss": 0.23882240056991577,
      "mean_token_accuracy": 0.9038627743721008,
      "num_tokens": 1411871992.0,
      "step": 608
    },
    {
      "epoch": 0.311668372569089,
      "grad_norm": 0.216796875,
      "learning_rate": 3.227254536695123e-05,
      "loss": 0.24141624569892883,
      "mean_token_accuracy": 0.9039629697799683,
      "num_tokens": 1414167622.0,
      "step": 609
    },
    {
      "epoch": 0.31218014329580346,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.224634816081811e-05,
      "loss": 0.24027304351329803,
      "mean_token_accuracy": 0.9036896824836731,
      "num_tokens": 1416503989.0,
      "step": 610
    },
    {
      "epoch": 0.3126919140225179,
      "grad_norm": 0.224609375,
      "learning_rate": 3.2220117296722716e-05,
      "loss": 0.2451549470424652,
      "mean_token_accuracy": 0.9017866849899292,
      "num_tokens": 1418778396.0,
      "step": 611
    },
    {
      "epoch": 0.3132036847492323,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.219385284675814e-05,
      "loss": 0.2508337199687958,
      "mean_token_accuracy": 0.8995389938354492,
      "num_tokens": 1421124772.0,
      "step": 612
    },
    {
      "epoch": 0.31371545547594676,
      "grad_norm": 0.2392578125,
      "learning_rate": 3.216755488310983e-05,
      "loss": 0.24410291016101837,
      "mean_token_accuracy": 0.9030972719192505,
      "num_tokens": 1423455939.0,
      "step": 613
    },
    {
      "epoch": 0.3142272262026612,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.214122347805531e-05,
      "loss": 0.24226468801498413,
      "mean_token_accuracy": 0.9035427570343018,
      "num_tokens": 1425671770.0,
      "step": 614
    },
    {
      "epoch": 0.3147389969293756,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.211485870396403e-05,
      "loss": 0.24737964570522308,
      "mean_token_accuracy": 0.9009989500045776,
      "num_tokens": 1427973311.0,
      "step": 615
    },
    {
      "epoch": 0.31525076765609006,
      "grad_norm": 0.22265625,
      "learning_rate": 3.208846063329715e-05,
      "loss": 0.24609681963920593,
      "mean_token_accuracy": 0.9018851518630981,
      "num_tokens": 1430346850.0,
      "step": 616
    },
    {
      "epoch": 0.3157625383828045,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.206202933860734e-05,
      "loss": 0.24578604102134705,
      "mean_token_accuracy": 0.9013868570327759,
      "num_tokens": 1432670894.0,
      "step": 617
    },
    {
      "epoch": 0.3162743091095189,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.203556489253859e-05,
      "loss": 0.25513842701911926,
      "mean_token_accuracy": 0.8981077671051025,
      "num_tokens": 1434965346.0,
      "step": 618
    },
    {
      "epoch": 0.31678607983623336,
      "grad_norm": 0.234375,
      "learning_rate": 3.200906736782598e-05,
      "loss": 0.22746647894382477,
      "mean_token_accuracy": 0.908621072769165,
      "num_tokens": 1437274036.0,
      "step": 619
    },
    {
      "epoch": 0.3172978505629478,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.1982536837295544e-05,
      "loss": 0.233872652053833,
      "mean_token_accuracy": 0.907141387462616,
      "num_tokens": 1439530566.0,
      "step": 620
    },
    {
      "epoch": 0.3178096212896622,
      "grad_norm": 0.212890625,
      "learning_rate": 3.195597337386399e-05,
      "loss": 0.23293691873550415,
      "mean_token_accuracy": 0.9070836305618286,
      "num_tokens": 1441787416.0,
      "step": 621
    },
    {
      "epoch": 0.31832139201637666,
      "grad_norm": 0.220703125,
      "learning_rate": 3.192937705053856e-05,
      "loss": 0.2250894457101822,
      "mean_token_accuracy": 0.9087742567062378,
      "num_tokens": 1444103039.0,
      "step": 622
    },
    {
      "epoch": 0.3188331627430911,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.19027479404168e-05,
      "loss": 0.23495540022850037,
      "mean_token_accuracy": 0.9059388637542725,
      "num_tokens": 1446465978.0,
      "step": 623
    },
    {
      "epoch": 0.3193449334698055,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.1876086116686384e-05,
      "loss": 0.25172722339630127,
      "mean_token_accuracy": 0.9000471234321594,
      "num_tokens": 1448875447.0,
      "step": 624
    },
    {
      "epoch": 0.31985670419651996,
      "grad_norm": 0.236328125,
      "learning_rate": 3.184939165262487e-05,
      "loss": 0.2494128942489624,
      "mean_token_accuracy": 0.9006774425506592,
      "num_tokens": 1451144879.0,
      "step": 625
    },
    {
      "epoch": 0.3203684749232344,
      "grad_norm": 0.220703125,
      "learning_rate": 3.1822664621599544e-05,
      "loss": 0.23646180331707,
      "mean_token_accuracy": 0.9062004089355469,
      "num_tokens": 1453477454.0,
      "step": 626
    },
    {
      "epoch": 0.3208802456499488,
      "grad_norm": 0.224609375,
      "learning_rate": 3.17959050970672e-05,
      "loss": 0.24730537831783295,
      "mean_token_accuracy": 0.9014170169830322,
      "num_tokens": 1455850098.0,
      "step": 627
    },
    {
      "epoch": 0.32139201637666326,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.1769113152573927e-05,
      "loss": 0.24850961565971375,
      "mean_token_accuracy": 0.9006792306900024,
      "num_tokens": 1458165268.0,
      "step": 628
    },
    {
      "epoch": 0.3219037871033777,
      "grad_norm": 0.208984375,
      "learning_rate": 3.1742288861754925e-05,
      "loss": 0.22685515880584717,
      "mean_token_accuracy": 0.9101665019989014,
      "num_tokens": 1460485608.0,
      "step": 629
    },
    {
      "epoch": 0.3224155578300921,
      "grad_norm": 0.228515625,
      "learning_rate": 3.171543229833428e-05,
      "loss": 0.23168088495731354,
      "mean_token_accuracy": 0.9057867527008057,
      "num_tokens": 1462896671.0,
      "step": 630
    },
    {
      "epoch": 0.32292732855680656,
      "grad_norm": 0.21484375,
      "learning_rate": 3.16885435361248e-05,
      "loss": 0.23565831780433655,
      "mean_token_accuracy": 0.905713677406311,
      "num_tokens": 1465262820.0,
      "step": 631
    },
    {
      "epoch": 0.323439099283521,
      "grad_norm": 0.236328125,
      "learning_rate": 3.166162264902777e-05,
      "loss": 0.2437942922115326,
      "mean_token_accuracy": 0.9026268124580383,
      "num_tokens": 1467555362.0,
      "step": 632
    },
    {
      "epoch": 0.3239508700102354,
      "grad_norm": 0.224609375,
      "learning_rate": 3.1634669711032774e-05,
      "loss": 0.2439117729663849,
      "mean_token_accuracy": 0.9020536541938782,
      "num_tokens": 1469914722.0,
      "step": 633
    },
    {
      "epoch": 0.32446264073694986,
      "grad_norm": 0.2373046875,
      "learning_rate": 3.160768479621749e-05,
      "loss": 0.2573942244052887,
      "mean_token_accuracy": 0.8974725008010864,
      "num_tokens": 1472203510.0,
      "step": 634
    },
    {
      "epoch": 0.3249744114636643,
      "grad_norm": 0.2353515625,
      "learning_rate": 3.1580667978747476e-05,
      "loss": 0.25029733777046204,
      "mean_token_accuracy": 0.9004098773002625,
      "num_tokens": 1474494591.0,
      "step": 635
    },
    {
      "epoch": 0.3254861821903787,
      "grad_norm": 0.2451171875,
      "learning_rate": 3.155361933287597e-05,
      "loss": 0.24800394475460052,
      "mean_token_accuracy": 0.9003521203994751,
      "num_tokens": 1476846921.0,
      "step": 636
    },
    {
      "epoch": 0.32599795291709316,
      "grad_norm": 0.29296875,
      "learning_rate": 3.1526538932943685e-05,
      "loss": 0.23540963232517242,
      "mean_token_accuracy": 0.904874861240387,
      "num_tokens": 1479168755.0,
      "step": 637
    },
    {
      "epoch": 0.3265097236438076,
      "grad_norm": 0.224609375,
      "learning_rate": 3.149942685337861e-05,
      "loss": 0.2482079714536667,
      "mean_token_accuracy": 0.9011715650558472,
      "num_tokens": 1481513749.0,
      "step": 638
    },
    {
      "epoch": 0.327021494370522,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.1472283168695814e-05,
      "loss": 0.23626822233200073,
      "mean_token_accuracy": 0.9060275554656982,
      "num_tokens": 1483929087.0,
      "step": 639
    },
    {
      "epoch": 0.32753326509723646,
      "grad_norm": 0.23046875,
      "learning_rate": 3.144510795349721e-05,
      "loss": 0.2379792034626007,
      "mean_token_accuracy": 0.9045366048812866,
      "num_tokens": 1486177434.0,
      "step": 640
    },
    {
      "epoch": 0.3280450358239509,
      "grad_norm": 0.212890625,
      "learning_rate": 3.141790128247138e-05,
      "loss": 0.23451635241508484,
      "mean_token_accuracy": 0.9061591029167175,
      "num_tokens": 1488516010.0,
      "step": 641
    },
    {
      "epoch": 0.3285568065506653,
      "grad_norm": 0.255859375,
      "learning_rate": 3.139066323039336e-05,
      "loss": 0.24409064650535583,
      "mean_token_accuracy": 0.90277099609375,
      "num_tokens": 1490848645.0,
      "step": 642
    },
    {
      "epoch": 0.32906857727737976,
      "grad_norm": 0.2314453125,
      "learning_rate": 3.1363393872124434e-05,
      "loss": 0.24439433217048645,
      "mean_token_accuracy": 0.9028987288475037,
      "num_tokens": 1493194515.0,
      "step": 643
    },
    {
      "epoch": 0.3295803480040942,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.133609328261193e-05,
      "loss": 0.22682851552963257,
      "mean_token_accuracy": 0.9091569185256958,
      "num_tokens": 1495483160.0,
      "step": 644
    },
    {
      "epoch": 0.33009211873080857,
      "grad_norm": 0.21484375,
      "learning_rate": 3.1308761536888995e-05,
      "loss": 0.22784388065338135,
      "mean_token_accuracy": 0.9095223546028137,
      "num_tokens": 1497895185.0,
      "step": 645
    },
    {
      "epoch": 0.330603889457523,
      "grad_norm": 0.228515625,
      "learning_rate": 3.128139871007444e-05,
      "loss": 0.268596351146698,
      "mean_token_accuracy": 0.894147515296936,
      "num_tokens": 1500186376.0,
      "step": 646
    },
    {
      "epoch": 0.33111566018423744,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.125400487737246e-05,
      "loss": 0.23798827826976776,
      "mean_token_accuracy": 0.9050635099411011,
      "num_tokens": 1502437449.0,
      "step": 647
    },
    {
      "epoch": 0.33162743091095187,
      "grad_norm": 0.2294921875,
      "learning_rate": 3.12265801140725e-05,
      "loss": 0.24123576283454895,
      "mean_token_accuracy": 0.9037752747535706,
      "num_tokens": 1504822949.0,
      "step": 648
    },
    {
      "epoch": 0.3321392016376663,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.1199124495549e-05,
      "loss": 0.24198491871356964,
      "mean_token_accuracy": 0.9043118357658386,
      "num_tokens": 1507140021.0,
      "step": 649
    },
    {
      "epoch": 0.33265097236438074,
      "grad_norm": 0.2421875,
      "learning_rate": 3.11716380972612e-05,
      "loss": 0.2556699514389038,
      "mean_token_accuracy": 0.8980397582054138,
      "num_tokens": 1509458152.0,
      "step": 650
    },
    {
      "epoch": 0.33316274309109517,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.114412099475293e-05,
      "loss": 0.2433251142501831,
      "mean_token_accuracy": 0.9018635749816895,
      "num_tokens": 1511797267.0,
      "step": 651
    },
    {
      "epoch": 0.3336745138178096,
      "grad_norm": 0.216796875,
      "learning_rate": 3.1116573263652434e-05,
      "loss": 0.22825486958026886,
      "mean_token_accuracy": 0.9085249900817871,
      "num_tokens": 1514100163.0,
      "step": 652
    },
    {
      "epoch": 0.33418628454452404,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.108899497967211e-05,
      "loss": 0.2238548994064331,
      "mean_token_accuracy": 0.9099162817001343,
      "num_tokens": 1516392693.0,
      "step": 653
    },
    {
      "epoch": 0.33469805527123847,
      "grad_norm": 0.232421875,
      "learning_rate": 3.106138621860834e-05,
      "loss": 0.2525849938392639,
      "mean_token_accuracy": 0.899003267288208,
      "num_tokens": 1518744842.0,
      "step": 654
    },
    {
      "epoch": 0.3352098259979529,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.103374705634126e-05,
      "loss": 0.25230008363723755,
      "mean_token_accuracy": 0.8991031646728516,
      "num_tokens": 1521094524.0,
      "step": 655
    },
    {
      "epoch": 0.33572159672466734,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.1006077568834576e-05,
      "loss": 0.24269869923591614,
      "mean_token_accuracy": 0.902784526348114,
      "num_tokens": 1523316851.0,
      "step": 656
    },
    {
      "epoch": 0.33623336745138177,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.0978377832135323e-05,
      "loss": 0.2413584291934967,
      "mean_token_accuracy": 0.9032823443412781,
      "num_tokens": 1525628477.0,
      "step": 657
    },
    {
      "epoch": 0.3367451381780962,
      "grad_norm": 0.224609375,
      "learning_rate": 3.0950647922373685e-05,
      "loss": 0.24098674952983856,
      "mean_token_accuracy": 0.9035335779190063,
      "num_tokens": 1528027340.0,
      "step": 658
    },
    {
      "epoch": 0.33725690890481064,
      "grad_norm": 0.21875,
      "learning_rate": 3.092288791576277e-05,
      "loss": 0.24431684613227844,
      "mean_token_accuracy": 0.9024189710617065,
      "num_tokens": 1530285068.0,
      "step": 659
    },
    {
      "epoch": 0.33776867963152507,
      "grad_norm": 0.2001953125,
      "learning_rate": 3.0895097888598406e-05,
      "loss": 0.2387009859085083,
      "mean_token_accuracy": 0.9040940403938293,
      "num_tokens": 1532586433.0,
      "step": 660
    },
    {
      "epoch": 0.3382804503582395,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.086727791725892e-05,
      "loss": 0.2552155554294586,
      "mean_token_accuracy": 0.897842288017273,
      "num_tokens": 1534913523.0,
      "step": 661
    },
    {
      "epoch": 0.33879222108495394,
      "grad_norm": 0.220703125,
      "learning_rate": 3.083942807820496e-05,
      "loss": 0.23023727536201477,
      "mean_token_accuracy": 0.9074360728263855,
      "num_tokens": 1537299684.0,
      "step": 662
    },
    {
      "epoch": 0.33930399181166837,
      "grad_norm": 0.220703125,
      "learning_rate": 3.081154844797924e-05,
      "loss": 0.24087801575660706,
      "mean_token_accuracy": 0.9033225774765015,
      "num_tokens": 1539526667.0,
      "step": 663
    },
    {
      "epoch": 0.3398157625383828,
      "grad_norm": 0.21875,
      "learning_rate": 3.078363910320636e-05,
      "loss": 0.23045027256011963,
      "mean_token_accuracy": 0.9075684547424316,
      "num_tokens": 1541970845.0,
      "step": 664
    },
    {
      "epoch": 0.34032753326509724,
      "grad_norm": 0.216796875,
      "learning_rate": 3.07557001205926e-05,
      "loss": 0.2398047149181366,
      "mean_token_accuracy": 0.9043534994125366,
      "num_tokens": 1544271924.0,
      "step": 665
    },
    {
      "epoch": 0.34083930399181167,
      "grad_norm": 0.2333984375,
      "learning_rate": 3.0727731576925665e-05,
      "loss": 0.2458709329366684,
      "mean_token_accuracy": 0.900982141494751,
      "num_tokens": 1546623251.0,
      "step": 666
    },
    {
      "epoch": 0.3413510747185261,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.069973354907455e-05,
      "loss": 0.234120711684227,
      "mean_token_accuracy": 0.9062849283218384,
      "num_tokens": 1548865322.0,
      "step": 667
    },
    {
      "epoch": 0.34186284544524054,
      "grad_norm": 0.212890625,
      "learning_rate": 3.0671706113989245e-05,
      "loss": 0.24309557676315308,
      "mean_token_accuracy": 0.9028342962265015,
      "num_tokens": 1551197550.0,
      "step": 668
    },
    {
      "epoch": 0.34237461617195497,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.064364934870058e-05,
      "loss": 0.24217969179153442,
      "mean_token_accuracy": 0.9031239748001099,
      "num_tokens": 1553505878.0,
      "step": 669
    },
    {
      "epoch": 0.3428863868986694,
      "grad_norm": 0.2060546875,
      "learning_rate": 3.061556333032001e-05,
      "loss": 0.2333216369152069,
      "mean_token_accuracy": 0.9069995284080505,
      "num_tokens": 1555804070.0,
      "step": 670
    },
    {
      "epoch": 0.34339815762538384,
      "grad_norm": 0.228515625,
      "learning_rate": 3.058744813603935e-05,
      "loss": 0.2486981600522995,
      "mean_token_accuracy": 0.9002174139022827,
      "num_tokens": 1558124283.0,
      "step": 671
    },
    {
      "epoch": 0.34390992835209827,
      "grad_norm": 0.2216796875,
      "learning_rate": 3.055930384313065e-05,
      "loss": 0.2407969832420349,
      "mean_token_accuracy": 0.9043532609939575,
      "num_tokens": 1560423428.0,
      "step": 672
    },
    {
      "epoch": 0.3444216990788127,
      "grad_norm": 0.2265625,
      "learning_rate": 3.05311305289459e-05,
      "loss": 0.2416815608739853,
      "mean_token_accuracy": 0.9028774499893188,
      "num_tokens": 1562766711.0,
      "step": 673
    },
    {
      "epoch": 0.34493346980552714,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.0502928270916877e-05,
      "loss": 0.23584765195846558,
      "mean_token_accuracy": 0.9061226844787598,
      "num_tokens": 1565072107.0,
      "step": 674
    },
    {
      "epoch": 0.34544524053224157,
      "grad_norm": 0.2265625,
      "learning_rate": 3.047469714655488e-05,
      "loss": 0.24882128834724426,
      "mean_token_accuracy": 0.900031566619873,
      "num_tokens": 1567307260.0,
      "step": 675
    },
    {
      "epoch": 0.345957011258956,
      "grad_norm": 0.220703125,
      "learning_rate": 3.0446437233450563e-05,
      "loss": 0.2441668063402176,
      "mean_token_accuracy": 0.9016845226287842,
      "num_tokens": 1569612442.0,
      "step": 676
    },
    {
      "epoch": 0.34646878198567044,
      "grad_norm": 0.21484375,
      "learning_rate": 3.041814860927371e-05,
      "loss": 0.25491148233413696,
      "mean_token_accuracy": 0.8980976343154907,
      "num_tokens": 1572002369.0,
      "step": 677
    },
    {
      "epoch": 0.34698055271238487,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.038983135177299e-05,
      "loss": 0.23887769877910614,
      "mean_token_accuracy": 0.9036267995834351,
      "num_tokens": 1574331722.0,
      "step": 678
    },
    {
      "epoch": 0.3474923234390993,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.036148553877579e-05,
      "loss": 0.2420348823070526,
      "mean_token_accuracy": 0.9042450189590454,
      "num_tokens": 1576645670.0,
      "step": 679
    },
    {
      "epoch": 0.34800409416581374,
      "grad_norm": 0.228515625,
      "learning_rate": 3.033311124818797e-05,
      "loss": 0.2434910088777542,
      "mean_token_accuracy": 0.902186393737793,
      "num_tokens": 1578897780.0,
      "step": 680
    },
    {
      "epoch": 0.34851586489252817,
      "grad_norm": 0.23046875,
      "learning_rate": 3.030470855799367e-05,
      "loss": 0.24218632280826569,
      "mean_token_accuracy": 0.9016110897064209,
      "num_tokens": 1581239408.0,
      "step": 681
    },
    {
      "epoch": 0.3490276356192426,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.027627754625505e-05,
      "loss": 0.23476898670196533,
      "mean_token_accuracy": 0.9064728021621704,
      "num_tokens": 1583596463.0,
      "step": 682
    },
    {
      "epoch": 0.34953940634595704,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.0247818291112154e-05,
      "loss": 0.2369702160358429,
      "mean_token_accuracy": 0.9047074913978577,
      "num_tokens": 1585952904.0,
      "step": 683
    },
    {
      "epoch": 0.3500511770726714,
      "grad_norm": 0.2109375,
      "learning_rate": 3.0219330870782625e-05,
      "loss": 0.24848270416259766,
      "mean_token_accuracy": 0.900421142578125,
      "num_tokens": 1588296937.0,
      "step": 684
    },
    {
      "epoch": 0.35056294779938585,
      "grad_norm": 0.2255859375,
      "learning_rate": 3.0190815363561506e-05,
      "loss": 0.24300053715705872,
      "mean_token_accuracy": 0.9028089046478271,
      "num_tokens": 1590640133.0,
      "step": 685
    },
    {
      "epoch": 0.3510747185261003,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.0162271847821066e-05,
      "loss": 0.24400517344474792,
      "mean_token_accuracy": 0.9014276266098022,
      "num_tokens": 1592966997.0,
      "step": 686
    },
    {
      "epoch": 0.3515864892528147,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.0133700402010516e-05,
      "loss": 0.24908116459846497,
      "mean_token_accuracy": 0.9009978771209717,
      "num_tokens": 1595304655.0,
      "step": 687
    },
    {
      "epoch": 0.35209825997952915,
      "grad_norm": 0.2275390625,
      "learning_rate": 3.0105101104655864e-05,
      "loss": 0.23560413718223572,
      "mean_token_accuracy": 0.9061902165412903,
      "num_tokens": 1597650910.0,
      "step": 688
    },
    {
      "epoch": 0.3526100307062436,
      "grad_norm": 0.224609375,
      "learning_rate": 3.0076474034359646e-05,
      "loss": 0.2369040995836258,
      "mean_token_accuracy": 0.9052839875221252,
      "num_tokens": 1599977639.0,
      "step": 689
    },
    {
      "epoch": 0.353121801432958,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.0047819269800725e-05,
      "loss": 0.23666158318519592,
      "mean_token_accuracy": 0.9051854014396667,
      "num_tokens": 1602320418.0,
      "step": 690
    },
    {
      "epoch": 0.35363357215967245,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.00191368897341e-05,
      "loss": 0.2322743535041809,
      "mean_token_accuracy": 0.9062018394470215,
      "num_tokens": 1604650309.0,
      "step": 691
    },
    {
      "epoch": 0.3541453428863869,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.9990426972990647e-05,
      "loss": 0.23806977272033691,
      "mean_token_accuracy": 0.904265820980072,
      "num_tokens": 1606885392.0,
      "step": 692
    },
    {
      "epoch": 0.3546571136131013,
      "grad_norm": 0.21875,
      "learning_rate": 2.996168959847694e-05,
      "loss": 0.2475142925977707,
      "mean_token_accuracy": 0.9013664722442627,
      "num_tokens": 1609193196.0,
      "step": 693
    },
    {
      "epoch": 0.35516888433981575,
      "grad_norm": 0.2265625,
      "learning_rate": 2.9932924845175006e-05,
      "loss": 0.24700018763542175,
      "mean_token_accuracy": 0.9011731147766113,
      "num_tokens": 1611455156.0,
      "step": 694
    },
    {
      "epoch": 0.3556806550665302,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.9904132792142127e-05,
      "loss": 0.24108916521072388,
      "mean_token_accuracy": 0.902315616607666,
      "num_tokens": 1613814257.0,
      "step": 695
    },
    {
      "epoch": 0.3561924257932446,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.9875313518510616e-05,
      "loss": 0.2289903312921524,
      "mean_token_accuracy": 0.9084707498550415,
      "num_tokens": 1616131839.0,
      "step": 696
    },
    {
      "epoch": 0.35670419651995905,
      "grad_norm": 0.439453125,
      "learning_rate": 2.9846467103487598e-05,
      "loss": 0.24391049146652222,
      "mean_token_accuracy": 0.9026439189910889,
      "num_tokens": 1618525012.0,
      "step": 697
    },
    {
      "epoch": 0.3572159672466735,
      "grad_norm": 0.234375,
      "learning_rate": 2.981759362635479e-05,
      "loss": 0.2595597505569458,
      "mean_token_accuracy": 0.8964878916740417,
      "num_tokens": 1620905611.0,
      "step": 698
    },
    {
      "epoch": 0.3577277379733879,
      "grad_norm": 0.21484375,
      "learning_rate": 2.978869316646829e-05,
      "loss": 0.2316192388534546,
      "mean_token_accuracy": 0.9073943495750427,
      "num_tokens": 1623325766.0,
      "step": 699
    },
    {
      "epoch": 0.35823950870010235,
      "grad_norm": 0.212890625,
      "learning_rate": 2.9759765803258363e-05,
      "loss": 0.23769330978393555,
      "mean_token_accuracy": 0.904839038848877,
      "num_tokens": 1625658152.0,
      "step": 700
    },
    {
      "epoch": 0.3587512794268168,
      "grad_norm": 0.2109375,
      "learning_rate": 2.97308116162292e-05,
      "loss": 0.23297451436519623,
      "mean_token_accuracy": 0.9072391390800476,
      "num_tokens": 1628031583.0,
      "step": 701
    },
    {
      "epoch": 0.3592630501535312,
      "grad_norm": 0.22265625,
      "learning_rate": 2.970183068495872e-05,
      "loss": 0.24940824508666992,
      "mean_token_accuracy": 0.9004422426223755,
      "num_tokens": 1630302226.0,
      "step": 702
    },
    {
      "epoch": 0.35977482088024565,
      "grad_norm": 0.228515625,
      "learning_rate": 2.9672823089098357e-05,
      "loss": 0.25964057445526123,
      "mean_token_accuracy": 0.896355926990509,
      "num_tokens": 1632573013.0,
      "step": 703
    },
    {
      "epoch": 0.3602865916069601,
      "grad_norm": 0.224609375,
      "learning_rate": 2.964378890837282e-05,
      "loss": 0.239279642701149,
      "mean_token_accuracy": 0.9039829969406128,
      "num_tokens": 1634974035.0,
      "step": 704
    },
    {
      "epoch": 0.3607983623336745,
      "grad_norm": 0.22265625,
      "learning_rate": 2.9614728222579882e-05,
      "loss": 0.2461889684200287,
      "mean_token_accuracy": 0.9012339115142822,
      "num_tokens": 1637216635.0,
      "step": 705
    },
    {
      "epoch": 0.36131013306038895,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.9585641111590173e-05,
      "loss": 0.2400927096605301,
      "mean_token_accuracy": 0.902631938457489,
      "num_tokens": 1639466716.0,
      "step": 706
    },
    {
      "epoch": 0.3618219037871034,
      "grad_norm": 0.2333984375,
      "learning_rate": 2.9556527655346943e-05,
      "loss": 0.2398812174797058,
      "mean_token_accuracy": 0.90440833568573,
      "num_tokens": 1641760143.0,
      "step": 707
    },
    {
      "epoch": 0.3623336745138178,
      "grad_norm": 0.236328125,
      "learning_rate": 2.9527387933865838e-05,
      "loss": 0.23721109330654144,
      "mean_token_accuracy": 0.9042433500289917,
      "num_tokens": 1644120061.0,
      "step": 708
    },
    {
      "epoch": 0.36284544524053225,
      "grad_norm": 0.224609375,
      "learning_rate": 2.9498222027234723e-05,
      "loss": 0.2331499457359314,
      "mean_token_accuracy": 0.9067851901054382,
      "num_tokens": 1646413982.0,
      "step": 709
    },
    {
      "epoch": 0.3633572159672467,
      "grad_norm": 0.236328125,
      "learning_rate": 2.9469030015613396e-05,
      "loss": 0.24486148357391357,
      "mean_token_accuracy": 0.9014126062393188,
      "num_tokens": 1648584239.0,
      "step": 710
    },
    {
      "epoch": 0.3638689866939611,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.9439811979233434e-05,
      "loss": 0.24951231479644775,
      "mean_token_accuracy": 0.8999561071395874,
      "num_tokens": 1650826212.0,
      "step": 711
    },
    {
      "epoch": 0.36438075742067555,
      "grad_norm": 0.2294921875,
      "learning_rate": 2.9410567998397907e-05,
      "loss": 0.24284642934799194,
      "mean_token_accuracy": 0.9031785726547241,
      "num_tokens": 1653174942.0,
      "step": 712
    },
    {
      "epoch": 0.36489252814739,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.9381298153481215e-05,
      "loss": 0.239113450050354,
      "mean_token_accuracy": 0.9050037264823914,
      "num_tokens": 1655409777.0,
      "step": 713
    },
    {
      "epoch": 0.3654042988741044,
      "grad_norm": 0.2412109375,
      "learning_rate": 2.9352002524928847e-05,
      "loss": 0.23492088913917542,
      "mean_token_accuracy": 0.9060993194580078,
      "num_tokens": 1657661757.0,
      "step": 714
    },
    {
      "epoch": 0.36591606960081885,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.9322681193257127e-05,
      "loss": 0.23672103881835938,
      "mean_token_accuracy": 0.905315101146698,
      "num_tokens": 1659926809.0,
      "step": 715
    },
    {
      "epoch": 0.3664278403275333,
      "grad_norm": 0.21484375,
      "learning_rate": 2.9293334239053056e-05,
      "loss": 0.2286120355129242,
      "mean_token_accuracy": 0.9081072807312012,
      "num_tokens": 1662277272.0,
      "step": 716
    },
    {
      "epoch": 0.3669396110542477,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.926396174297402e-05,
      "loss": 0.2349155843257904,
      "mean_token_accuracy": 0.9055910706520081,
      "num_tokens": 1664526534.0,
      "step": 717
    },
    {
      "epoch": 0.36745138178096215,
      "grad_norm": 0.2373046875,
      "learning_rate": 2.923456378574765e-05,
      "loss": 0.25440317392349243,
      "mean_token_accuracy": 0.8994500041007996,
      "num_tokens": 1666829824.0,
      "step": 718
    },
    {
      "epoch": 0.3679631525076766,
      "grad_norm": 0.232421875,
      "learning_rate": 2.9205140448171514e-05,
      "loss": 0.23598453402519226,
      "mean_token_accuracy": 0.9055377840995789,
      "num_tokens": 1669118601.0,
      "step": 719
    },
    {
      "epoch": 0.368474923234391,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.9175691811112954e-05,
      "loss": 0.24427597224712372,
      "mean_token_accuracy": 0.9024015665054321,
      "num_tokens": 1671399800.0,
      "step": 720
    },
    {
      "epoch": 0.36898669396110545,
      "grad_norm": 0.23046875,
      "learning_rate": 2.9146217955508844e-05,
      "loss": 0.24498499929904938,
      "mean_token_accuracy": 0.9023171663284302,
      "num_tokens": 1673697732.0,
      "step": 721
    },
    {
      "epoch": 0.3694984646878199,
      "grad_norm": 0.220703125,
      "learning_rate": 2.911671896236536e-05,
      "loss": 0.24232915043830872,
      "mean_token_accuracy": 0.9036654233932495,
      "num_tokens": 1675986208.0,
      "step": 722
    },
    {
      "epoch": 0.37001023541453426,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.9087194912757783e-05,
      "loss": 0.2322159707546234,
      "mean_token_accuracy": 0.9073365330696106,
      "num_tokens": 1678394048.0,
      "step": 723
    },
    {
      "epoch": 0.3705220061412487,
      "grad_norm": 0.220703125,
      "learning_rate": 2.905764588783024e-05,
      "loss": 0.2522595226764679,
      "mean_token_accuracy": 0.8992304801940918,
      "num_tokens": 1680657821.0,
      "step": 724
    },
    {
      "epoch": 0.3710337768679631,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.9028071968795517e-05,
      "loss": 0.2400766909122467,
      "mean_token_accuracy": 0.9036146402359009,
      "num_tokens": 1682997206.0,
      "step": 725
    },
    {
      "epoch": 0.37154554759467756,
      "grad_norm": 0.20703125,
      "learning_rate": 2.899847323693481e-05,
      "loss": 0.22264322638511658,
      "mean_token_accuracy": 0.9107201099395752,
      "num_tokens": 1685384238.0,
      "step": 726
    },
    {
      "epoch": 0.372057318321392,
      "grad_norm": 0.228515625,
      "learning_rate": 2.8968849773597515e-05,
      "loss": 0.24084073305130005,
      "mean_token_accuracy": 0.9033567905426025,
      "num_tokens": 1687690952.0,
      "step": 727
    },
    {
      "epoch": 0.3725690890481064,
      "grad_norm": 0.21875,
      "learning_rate": 2.8939201660200998e-05,
      "loss": 0.2480347454547882,
      "mean_token_accuracy": 0.900590181350708,
      "num_tokens": 1690033425.0,
      "step": 728
    },
    {
      "epoch": 0.37308085977482086,
      "grad_norm": 0.212890625,
      "learning_rate": 2.8909528978230367e-05,
      "loss": 0.23108139634132385,
      "mean_token_accuracy": 0.9078389406204224,
      "num_tokens": 1692357831.0,
      "step": 729
    },
    {
      "epoch": 0.3735926305015353,
      "grad_norm": 0.208984375,
      "learning_rate": 2.8879831809238272e-05,
      "loss": 0.23502245545387268,
      "mean_token_accuracy": 0.9063891172409058,
      "num_tokens": 1694704801.0,
      "step": 730
    },
    {
      "epoch": 0.3741044012282497,
      "grad_norm": 0.2265625,
      "learning_rate": 2.885011023484464e-05,
      "loss": 0.24789008498191833,
      "mean_token_accuracy": 0.9010317325592041,
      "num_tokens": 1696981295.0,
      "step": 731
    },
    {
      "epoch": 0.37461617195496416,
      "grad_norm": 0.224609375,
      "learning_rate": 2.882036433673649e-05,
      "loss": 0.2426455318927765,
      "mean_token_accuracy": 0.9027731418609619,
      "num_tokens": 1699281282.0,
      "step": 732
    },
    {
      "epoch": 0.3751279426816786,
      "grad_norm": 0.234375,
      "learning_rate": 2.8790594196667693e-05,
      "loss": 0.2583252191543579,
      "mean_token_accuracy": 0.8972660899162292,
      "num_tokens": 1701592236.0,
      "step": 733
    },
    {
      "epoch": 0.37563971340839303,
      "grad_norm": 0.2109375,
      "learning_rate": 2.8760799896458748e-05,
      "loss": 0.23822692036628723,
      "mean_token_accuracy": 0.9043693542480469,
      "num_tokens": 1703878512.0,
      "step": 734
    },
    {
      "epoch": 0.37615148413510746,
      "grad_norm": 0.21484375,
      "learning_rate": 2.8730981517996536e-05,
      "loss": 0.2353169023990631,
      "mean_token_accuracy": 0.9052903652191162,
      "num_tokens": 1706237273.0,
      "step": 735
    },
    {
      "epoch": 0.3766632548618219,
      "grad_norm": 0.20703125,
      "learning_rate": 2.8701139143234137e-05,
      "loss": 0.231239914894104,
      "mean_token_accuracy": 0.9077514410018921,
      "num_tokens": 1708544506.0,
      "step": 736
    },
    {
      "epoch": 0.37717502558853633,
      "grad_norm": 0.2275390625,
      "learning_rate": 2.8671272854190584e-05,
      "loss": 0.24149535596370697,
      "mean_token_accuracy": 0.9030716419219971,
      "num_tokens": 1710853760.0,
      "step": 737
    },
    {
      "epoch": 0.37768679631525076,
      "grad_norm": 0.2265625,
      "learning_rate": 2.8641382732950607e-05,
      "loss": 0.23814260959625244,
      "mean_token_accuracy": 0.9047002792358398,
      "num_tokens": 1713132343.0,
      "step": 738
    },
    {
      "epoch": 0.3781985670419652,
      "grad_norm": 0.21484375,
      "learning_rate": 2.861146886166447e-05,
      "loss": 0.251039057970047,
      "mean_token_accuracy": 0.8997892737388611,
      "num_tokens": 1715384149.0,
      "step": 739
    },
    {
      "epoch": 0.37871033776867963,
      "grad_norm": 0.220703125,
      "learning_rate": 2.85815313225477e-05,
      "loss": 0.24105879664421082,
      "mean_token_accuracy": 0.904567301273346,
      "num_tokens": 1717663276.0,
      "step": 740
    },
    {
      "epoch": 0.37922210849539406,
      "grad_norm": 0.224609375,
      "learning_rate": 2.855157019788087e-05,
      "loss": 0.24465064704418182,
      "mean_token_accuracy": 0.9023328423500061,
      "num_tokens": 1719940056.0,
      "step": 741
    },
    {
      "epoch": 0.3797338792221085,
      "grad_norm": 0.21875,
      "learning_rate": 2.852158557000937e-05,
      "loss": 0.23708850145339966,
      "mean_token_accuracy": 0.9055869579315186,
      "num_tokens": 1722237548.0,
      "step": 742
    },
    {
      "epoch": 0.38024564994882293,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.84915775213432e-05,
      "loss": 0.2417065054178238,
      "mean_token_accuracy": 0.9032176733016968,
      "num_tokens": 1724574093.0,
      "step": 743
    },
    {
      "epoch": 0.38075742067553736,
      "grad_norm": 0.2265625,
      "learning_rate": 2.8461546134356726e-05,
      "loss": 0.2519155442714691,
      "mean_token_accuracy": 0.8999332785606384,
      "num_tokens": 1726840835.0,
      "step": 744
    },
    {
      "epoch": 0.3812691914022518,
      "grad_norm": 0.2333984375,
      "learning_rate": 2.843149149158845e-05,
      "loss": 0.2557094097137451,
      "mean_token_accuracy": 0.8970787525177002,
      "num_tokens": 1729132914.0,
      "step": 745
    },
    {
      "epoch": 0.38178096212896623,
      "grad_norm": 0.2353515625,
      "learning_rate": 2.84014136756408e-05,
      "loss": 0.23917049169540405,
      "mean_token_accuracy": 0.9041876196861267,
      "num_tokens": 1731452243.0,
      "step": 746
    },
    {
      "epoch": 0.38229273285568066,
      "grad_norm": 0.224609375,
      "learning_rate": 2.8371312769179872e-05,
      "loss": 0.2554776668548584,
      "mean_token_accuracy": 0.8984086513519287,
      "num_tokens": 1733768943.0,
      "step": 747
    },
    {
      "epoch": 0.3828045035823951,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.834118885493527e-05,
      "loss": 0.24543851613998413,
      "mean_token_accuracy": 0.902124285697937,
      "num_tokens": 1736029030.0,
      "step": 748
    },
    {
      "epoch": 0.38331627430910953,
      "grad_norm": 0.21875,
      "learning_rate": 2.8311042015699778e-05,
      "loss": 0.2391567826271057,
      "mean_token_accuracy": 0.9044284820556641,
      "num_tokens": 1738325256.0,
      "step": 749
    },
    {
      "epoch": 0.38382804503582396,
      "grad_norm": 0.2265625,
      "learning_rate": 2.8280872334329223e-05,
      "loss": 0.23765408992767334,
      "mean_token_accuracy": 0.9051293134689331,
      "num_tokens": 1740630987.0,
      "step": 750
    },
    {
      "epoch": 0.3843398157625384,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.82506798937422e-05,
      "loss": 0.2398764044046402,
      "mean_token_accuracy": 0.9049874544143677,
      "num_tokens": 1742978007.0,
      "step": 751
    },
    {
      "epoch": 0.38485158648925283,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.8220464776919856e-05,
      "loss": 0.2378014624118805,
      "mean_token_accuracy": 0.9055244326591492,
      "num_tokens": 1745229238.0,
      "step": 752
    },
    {
      "epoch": 0.38536335721596726,
      "grad_norm": 0.21875,
      "learning_rate": 2.8190227066905673e-05,
      "loss": 0.23691287636756897,
      "mean_token_accuracy": 0.9052901268005371,
      "num_tokens": 1747530674.0,
      "step": 753
    },
    {
      "epoch": 0.3858751279426817,
      "grad_norm": 0.216796875,
      "learning_rate": 2.815996684680519e-05,
      "loss": 0.2294151335954666,
      "mean_token_accuracy": 0.9088376760482788,
      "num_tokens": 1749869102.0,
      "step": 754
    },
    {
      "epoch": 0.38638689866939613,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.812968419978587e-05,
      "loss": 0.2410891354084015,
      "mean_token_accuracy": 0.903997540473938,
      "num_tokens": 1752183117.0,
      "step": 755
    },
    {
      "epoch": 0.38689866939611056,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.809937920907677e-05,
      "loss": 0.24046054482460022,
      "mean_token_accuracy": 0.9039179086685181,
      "num_tokens": 1754564596.0,
      "step": 756
    },
    {
      "epoch": 0.387410440122825,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.806905195796838e-05,
      "loss": 0.2396499216556549,
      "mean_token_accuracy": 0.9034593105316162,
      "num_tokens": 1756915756.0,
      "step": 757
    },
    {
      "epoch": 0.38792221084953943,
      "grad_norm": 0.23828125,
      "learning_rate": 2.803870252981235e-05,
      "loss": 0.24510972201824188,
      "mean_token_accuracy": 0.9023751616477966,
      "num_tokens": 1759282667.0,
      "step": 758
    },
    {
      "epoch": 0.38843398157625386,
      "grad_norm": 0.30078125,
      "learning_rate": 2.80083310080213e-05,
      "loss": 0.25275835394859314,
      "mean_token_accuracy": 0.8988927602767944,
      "num_tokens": 1761623536.0,
      "step": 759
    },
    {
      "epoch": 0.3889457523029683,
      "grad_norm": 0.208984375,
      "learning_rate": 2.797793747606857e-05,
      "loss": 0.23179687559604645,
      "mean_token_accuracy": 0.9074491262435913,
      "num_tokens": 1763977983.0,
      "step": 760
    },
    {
      "epoch": 0.38945752302968273,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.7947522017487988e-05,
      "loss": 0.25633811950683594,
      "mean_token_accuracy": 0.8976446390151978,
      "num_tokens": 1766310990.0,
      "step": 761
    },
    {
      "epoch": 0.38996929375639716,
      "grad_norm": 0.2109375,
      "learning_rate": 2.791708471587364e-05,
      "loss": 0.24148055911064148,
      "mean_token_accuracy": 0.9037688970565796,
      "num_tokens": 1768706223.0,
      "step": 762
    },
    {
      "epoch": 0.39048106448311154,
      "grad_norm": 0.20703125,
      "learning_rate": 2.788662565487965e-05,
      "loss": 0.22813335061073303,
      "mean_token_accuracy": 0.9076387286186218,
      "num_tokens": 1771038987.0,
      "step": 763
    },
    {
      "epoch": 0.390992835209826,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.7856144918219964e-05,
      "loss": 0.2500133812427521,
      "mean_token_accuracy": 0.8996018171310425,
      "num_tokens": 1773339808.0,
      "step": 764
    },
    {
      "epoch": 0.3915046059365404,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.7825642589668066e-05,
      "loss": 0.23866717517375946,
      "mean_token_accuracy": 0.9039770364761353,
      "num_tokens": 1775652927.0,
      "step": 765
    },
    {
      "epoch": 0.39201637666325484,
      "grad_norm": 0.21875,
      "learning_rate": 2.7795118753056814e-05,
      "loss": 0.250081330537796,
      "mean_token_accuracy": 0.9004420042037964,
      "num_tokens": 1777971190.0,
      "step": 766
    },
    {
      "epoch": 0.3925281473899693,
      "grad_norm": 0.208984375,
      "learning_rate": 2.7764573492278162e-05,
      "loss": 0.23372815549373627,
      "mean_token_accuracy": 0.9068757891654968,
      "num_tokens": 1780178331.0,
      "step": 767
    },
    {
      "epoch": 0.3930399181166837,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.7734006891282957e-05,
      "loss": 0.23472312092781067,
      "mean_token_accuracy": 0.9060715436935425,
      "num_tokens": 1782578508.0,
      "step": 768
    },
    {
      "epoch": 0.39355168884339814,
      "grad_norm": 0.22265625,
      "learning_rate": 2.7703419034080684e-05,
      "loss": 0.2354522943496704,
      "mean_token_accuracy": 0.9059188365936279,
      "num_tokens": 1784860512.0,
      "step": 769
    },
    {
      "epoch": 0.3940634595701126,
      "grad_norm": 0.212890625,
      "learning_rate": 2.7672810004739262e-05,
      "loss": 0.22936893999576569,
      "mean_token_accuracy": 0.9078748226165771,
      "num_tokens": 1787222008.0,
      "step": 770
    },
    {
      "epoch": 0.394575230296827,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.7642179887384798e-05,
      "loss": 0.2501637935638428,
      "mean_token_accuracy": 0.9004693031311035,
      "num_tokens": 1789524226.0,
      "step": 771
    },
    {
      "epoch": 0.39508700102354144,
      "grad_norm": 0.21484375,
      "learning_rate": 2.761152876620135e-05,
      "loss": 0.2349257618188858,
      "mean_token_accuracy": 0.9055708050727844,
      "num_tokens": 1791805965.0,
      "step": 772
    },
    {
      "epoch": 0.3955987717502559,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.7580856725430704e-05,
      "loss": 0.22980362176895142,
      "mean_token_accuracy": 0.9076554775238037,
      "num_tokens": 1794181227.0,
      "step": 773
    },
    {
      "epoch": 0.3961105424769703,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.7550163849372152e-05,
      "loss": 0.23345798254013062,
      "mean_token_accuracy": 0.9067044854164124,
      "num_tokens": 1796471594.0,
      "step": 774
    },
    {
      "epoch": 0.39662231320368474,
      "grad_norm": 0.2109375,
      "learning_rate": 2.7519450222382235e-05,
      "loss": 0.23691195249557495,
      "mean_token_accuracy": 0.9055696725845337,
      "num_tokens": 1798844573.0,
      "step": 775
    },
    {
      "epoch": 0.3971340839303992,
      "grad_norm": 0.2275390625,
      "learning_rate": 2.7488715928874542e-05,
      "loss": 0.2383270561695099,
      "mean_token_accuracy": 0.9040813446044922,
      "num_tokens": 1801135801.0,
      "step": 776
    },
    {
      "epoch": 0.3976458546571136,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.7457961053319447e-05,
      "loss": 0.2401023507118225,
      "mean_token_accuracy": 0.9046177268028259,
      "num_tokens": 1803425959.0,
      "step": 777
    },
    {
      "epoch": 0.39815762538382804,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.7427185680243903e-05,
      "loss": 0.23365935683250427,
      "mean_token_accuracy": 0.9070382714271545,
      "num_tokens": 1805740337.0,
      "step": 778
    },
    {
      "epoch": 0.3986693961105425,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.7396389894231196e-05,
      "loss": 0.23288217186927795,
      "mean_token_accuracy": 0.9066125750541687,
      "num_tokens": 1807995033.0,
      "step": 779
    },
    {
      "epoch": 0.3991811668372569,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.7365573779920708e-05,
      "loss": 0.24669694900512695,
      "mean_token_accuracy": 0.9015425443649292,
      "num_tokens": 1810328548.0,
      "step": 780
    },
    {
      "epoch": 0.39969293756397134,
      "grad_norm": 0.2265625,
      "learning_rate": 2.7334737422007697e-05,
      "loss": 0.2477458417415619,
      "mean_token_accuracy": 0.9013952612876892,
      "num_tokens": 1812676411.0,
      "step": 781
    },
    {
      "epoch": 0.4002047082906858,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.7303880905243054e-05,
      "loss": 0.24141286313533783,
      "mean_token_accuracy": 0.903466522693634,
      "num_tokens": 1815050167.0,
      "step": 782
    },
    {
      "epoch": 0.4007164790174002,
      "grad_norm": 0.21484375,
      "learning_rate": 2.7273004314433093e-05,
      "loss": 0.24114467203617096,
      "mean_token_accuracy": 0.9026352167129517,
      "num_tokens": 1817326186.0,
      "step": 783
    },
    {
      "epoch": 0.40122824974411464,
      "grad_norm": 0.22265625,
      "learning_rate": 2.724210773443927e-05,
      "loss": 0.24285440146923065,
      "mean_token_accuracy": 0.9029049277305603,
      "num_tokens": 1819587725.0,
      "step": 784
    },
    {
      "epoch": 0.4017400204708291,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.721119125017801e-05,
      "loss": 0.23684142529964447,
      "mean_token_accuracy": 0.9048891663551331,
      "num_tokens": 1821855006.0,
      "step": 785
    },
    {
      "epoch": 0.4022517911975435,
      "grad_norm": 0.228515625,
      "learning_rate": 2.7180254946620416e-05,
      "loss": 0.25105780363082886,
      "mean_token_accuracy": 0.8997485637664795,
      "num_tokens": 1824214597.0,
      "step": 786
    },
    {
      "epoch": 0.40276356192425794,
      "grad_norm": 0.212890625,
      "learning_rate": 2.7149298908792087e-05,
      "loss": 0.2339065670967102,
      "mean_token_accuracy": 0.9058884978294373,
      "num_tokens": 1826606869.0,
      "step": 787
    },
    {
      "epoch": 0.4032753326509724,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.7118323221772837e-05,
      "loss": 0.2420683056116104,
      "mean_token_accuracy": 0.9026799201965332,
      "num_tokens": 1828848988.0,
      "step": 788
    },
    {
      "epoch": 0.4037871033776868,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.7087327970696496e-05,
      "loss": 0.2332175374031067,
      "mean_token_accuracy": 0.9072391390800476,
      "num_tokens": 1831158463.0,
      "step": 789
    },
    {
      "epoch": 0.40429887410440124,
      "grad_norm": 0.2265625,
      "learning_rate": 2.705631324075067e-05,
      "loss": 0.23152995109558105,
      "mean_token_accuracy": 0.9070245623588562,
      "num_tokens": 1833498656.0,
      "step": 790
    },
    {
      "epoch": 0.4048106448311157,
      "grad_norm": 0.2109375,
      "learning_rate": 2.702527911717648e-05,
      "loss": 0.23069503903388977,
      "mean_token_accuracy": 0.9083107709884644,
      "num_tokens": 1835840637.0,
      "step": 791
    },
    {
      "epoch": 0.4053224155578301,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.6994225685268375e-05,
      "loss": 0.23311814665794373,
      "mean_token_accuracy": 0.9070239067077637,
      "num_tokens": 1838159746.0,
      "step": 792
    },
    {
      "epoch": 0.40583418628454454,
      "grad_norm": 0.21484375,
      "learning_rate": 2.6963153030373854e-05,
      "loss": 0.2461758255958557,
      "mean_token_accuracy": 0.9009677767753601,
      "num_tokens": 1840422253.0,
      "step": 793
    },
    {
      "epoch": 0.406345957011259,
      "grad_norm": 0.21484375,
      "learning_rate": 2.6932061237893258e-05,
      "loss": 0.22817286849021912,
      "mean_token_accuracy": 0.9087069034576416,
      "num_tokens": 1842789964.0,
      "step": 794
    },
    {
      "epoch": 0.4068577277379734,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.6900950393279507e-05,
      "loss": 0.23170748353004456,
      "mean_token_accuracy": 0.9066638350486755,
      "num_tokens": 1845109648.0,
      "step": 795
    },
    {
      "epoch": 0.40736949846468784,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.686982058203791e-05,
      "loss": 0.2415587604045868,
      "mean_token_accuracy": 0.9029146432876587,
      "num_tokens": 1847373265.0,
      "step": 796
    },
    {
      "epoch": 0.4078812691914023,
      "grad_norm": 0.2109375,
      "learning_rate": 2.683867188972589e-05,
      "loss": 0.23718330264091492,
      "mean_token_accuracy": 0.9051872491836548,
      "num_tokens": 1849691653.0,
      "step": 797
    },
    {
      "epoch": 0.4083930399181167,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.6807504401952758e-05,
      "loss": 0.24405694007873535,
      "mean_token_accuracy": 0.902409017086029,
      "num_tokens": 1852053896.0,
      "step": 798
    },
    {
      "epoch": 0.40890481064483114,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.6776318204379498e-05,
      "loss": 0.23485472798347473,
      "mean_token_accuracy": 0.9066416621208191,
      "num_tokens": 1854371002.0,
      "step": 799
    },
    {
      "epoch": 0.4094165813715456,
      "grad_norm": 0.22265625,
      "learning_rate": 2.6745113382718503e-05,
      "loss": 0.2353123128414154,
      "mean_token_accuracy": 0.9060661792755127,
      "num_tokens": 1856725560.0,
      "step": 800
    },
    {
      "epoch": 0.40992835209826,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.671389002273336e-05,
      "loss": 0.2425578534603119,
      "mean_token_accuracy": 0.9034452438354492,
      "num_tokens": 1859161851.0,
      "step": 801
    },
    {
      "epoch": 0.4104401228249744,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.66826482102386e-05,
      "loss": 0.2462906837463379,
      "mean_token_accuracy": 0.9013774394989014,
      "num_tokens": 1861517005.0,
      "step": 802
    },
    {
      "epoch": 0.4109518935516888,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.6651388031099473e-05,
      "loss": 0.24759438633918762,
      "mean_token_accuracy": 0.9006645679473877,
      "num_tokens": 1863810537.0,
      "step": 803
    },
    {
      "epoch": 0.41146366427840325,
      "grad_norm": 0.2294921875,
      "learning_rate": 2.6620109571231716e-05,
      "loss": 0.24231931567192078,
      "mean_token_accuracy": 0.9030505418777466,
      "num_tokens": 1866156236.0,
      "step": 804
    },
    {
      "epoch": 0.4119754350051177,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.6588812916601293e-05,
      "loss": 0.2332892119884491,
      "mean_token_accuracy": 0.9067400693893433,
      "num_tokens": 1868499531.0,
      "step": 805
    },
    {
      "epoch": 0.4124872057318321,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.6557498153224183e-05,
      "loss": 0.23552373051643372,
      "mean_token_accuracy": 0.9048362970352173,
      "num_tokens": 1870741738.0,
      "step": 806
    },
    {
      "epoch": 0.41299897645854655,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.652616536716613e-05,
      "loss": 0.24909988045692444,
      "mean_token_accuracy": 0.9012943506240845,
      "num_tokens": 1873067281.0,
      "step": 807
    },
    {
      "epoch": 0.413510747185261,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.6494814644542434e-05,
      "loss": 0.23888584971427917,
      "mean_token_accuracy": 0.9046968817710876,
      "num_tokens": 1875523833.0,
      "step": 808
    },
    {
      "epoch": 0.4140225179119754,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.6463446071517655e-05,
      "loss": 0.2519075274467468,
      "mean_token_accuracy": 0.8993628025054932,
      "num_tokens": 1877897262.0,
      "step": 809
    },
    {
      "epoch": 0.41453428863868985,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.6432059734305447e-05,
      "loss": 0.24122852087020874,
      "mean_token_accuracy": 0.9027490019798279,
      "num_tokens": 1880210302.0,
      "step": 810
    },
    {
      "epoch": 0.4150460593654043,
      "grad_norm": 0.201171875,
      "learning_rate": 2.640065571916827e-05,
      "loss": 0.21772930026054382,
      "mean_token_accuracy": 0.9127544164657593,
      "num_tokens": 1882562566.0,
      "step": 811
    },
    {
      "epoch": 0.4155578300921187,
      "grad_norm": 0.21875,
      "learning_rate": 2.636923411241716e-05,
      "loss": 0.23687930405139923,
      "mean_token_accuracy": 0.9054446220397949,
      "num_tokens": 1884958610.0,
      "step": 812
    },
    {
      "epoch": 0.41606960081883315,
      "grad_norm": 0.220703125,
      "learning_rate": 2.6337795000411533e-05,
      "loss": 0.24087852239608765,
      "mean_token_accuracy": 0.9032421708106995,
      "num_tokens": 1887349903.0,
      "step": 813
    },
    {
      "epoch": 0.4165813715455476,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.6306338469558884e-05,
      "loss": 0.23323535919189453,
      "mean_token_accuracy": 0.9064413905143738,
      "num_tokens": 1889667124.0,
      "step": 814
    },
    {
      "epoch": 0.417093142272262,
      "grad_norm": 0.2265625,
      "learning_rate": 2.6274864606314613e-05,
      "loss": 0.24287615716457367,
      "mean_token_accuracy": 0.9032970070838928,
      "num_tokens": 1891880268.0,
      "step": 815
    },
    {
      "epoch": 0.41760491299897645,
      "grad_norm": 0.23046875,
      "learning_rate": 2.6243373497181723e-05,
      "loss": 0.24675028026103973,
      "mean_token_accuracy": 0.9009292125701904,
      "num_tokens": 1894241759.0,
      "step": 816
    },
    {
      "epoch": 0.4181166837256909,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.6211865228710646e-05,
      "loss": 0.24235588312149048,
      "mean_token_accuracy": 0.9025608897209167,
      "num_tokens": 1896695234.0,
      "step": 817
    },
    {
      "epoch": 0.4186284544524053,
      "grad_norm": 0.22265625,
      "learning_rate": 2.618033988749895e-05,
      "loss": 0.2332221269607544,
      "mean_token_accuracy": 0.9068978428840637,
      "num_tokens": 1898987393.0,
      "step": 818
    },
    {
      "epoch": 0.41914022517911975,
      "grad_norm": 0.2294921875,
      "learning_rate": 2.614879756019115e-05,
      "loss": 0.2513318359851837,
      "mean_token_accuracy": 0.899950385093689,
      "num_tokens": 1901348793.0,
      "step": 819
    },
    {
      "epoch": 0.4196519959058342,
      "grad_norm": 0.232421875,
      "learning_rate": 2.6117238333478425e-05,
      "loss": 0.2509998679161072,
      "mean_token_accuracy": 0.9000905156135559,
      "num_tokens": 1903634790.0,
      "step": 820
    },
    {
      "epoch": 0.4201637666325486,
      "grad_norm": 0.224609375,
      "learning_rate": 2.6085662294098414e-05,
      "loss": 0.23784857988357544,
      "mean_token_accuracy": 0.9039379358291626,
      "num_tokens": 1905912919.0,
      "step": 821
    },
    {
      "epoch": 0.42067553735926305,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.6054069528834948e-05,
      "loss": 0.25552624464035034,
      "mean_token_accuracy": 0.8971079587936401,
      "num_tokens": 1908317822.0,
      "step": 822
    },
    {
      "epoch": 0.4211873080859775,
      "grad_norm": 0.220703125,
      "learning_rate": 2.602246012451785e-05,
      "loss": 0.24463434517383575,
      "mean_token_accuracy": 0.901896595954895,
      "num_tokens": 1910666538.0,
      "step": 823
    },
    {
      "epoch": 0.4216990788126919,
      "grad_norm": 0.22265625,
      "learning_rate": 2.5990834168022665e-05,
      "loss": 0.2430872768163681,
      "mean_token_accuracy": 0.9034345149993896,
      "num_tokens": 1912927211.0,
      "step": 824
    },
    {
      "epoch": 0.42221084953940635,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.5959191746270415e-05,
      "loss": 0.2513027787208557,
      "mean_token_accuracy": 0.8988285064697266,
      "num_tokens": 1915205586.0,
      "step": 825
    },
    {
      "epoch": 0.4227226202661208,
      "grad_norm": 0.224609375,
      "learning_rate": 2.5927532946227393e-05,
      "loss": 0.23942482471466064,
      "mean_token_accuracy": 0.9031716585159302,
      "num_tokens": 1917388060.0,
      "step": 826
    },
    {
      "epoch": 0.4232343909928352,
      "grad_norm": 0.2353515625,
      "learning_rate": 2.5895857854904912e-05,
      "loss": 0.2533252239227295,
      "mean_token_accuracy": 0.899280309677124,
      "num_tokens": 1919722082.0,
      "step": 827
    },
    {
      "epoch": 0.42374616171954965,
      "grad_norm": 0.216796875,
      "learning_rate": 2.586416655935903e-05,
      "loss": 0.25325891375541687,
      "mean_token_accuracy": 0.8986974954605103,
      "num_tokens": 1922037532.0,
      "step": 828
    },
    {
      "epoch": 0.4242579324462641,
      "grad_norm": 0.21875,
      "learning_rate": 2.583245914669037e-05,
      "loss": 0.23335722088813782,
      "mean_token_accuracy": 0.9066333174705505,
      "num_tokens": 1924410677.0,
      "step": 829
    },
    {
      "epoch": 0.4247697031729785,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.580073570404384e-05,
      "loss": 0.24733424186706543,
      "mean_token_accuracy": 0.9016789197921753,
      "num_tokens": 1926731332.0,
      "step": 830
    },
    {
      "epoch": 0.42528147389969295,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.5768996318608413e-05,
      "loss": 0.2406269609928131,
      "mean_token_accuracy": 0.9039403200149536,
      "num_tokens": 1928988830.0,
      "step": 831
    },
    {
      "epoch": 0.4257932446264074,
      "grad_norm": 0.20703125,
      "learning_rate": 2.5737241077616858e-05,
      "loss": 0.2352488487958908,
      "mean_token_accuracy": 0.9060033559799194,
      "num_tokens": 1931278561.0,
      "step": 832
    },
    {
      "epoch": 0.4263050153531218,
      "grad_norm": 0.22265625,
      "learning_rate": 2.5705470068345548e-05,
      "loss": 0.2475416511297226,
      "mean_token_accuracy": 0.9012844562530518,
      "num_tokens": 1933628147.0,
      "step": 833
    },
    {
      "epoch": 0.42681678607983625,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.567368337811418e-05,
      "loss": 0.23840981721878052,
      "mean_token_accuracy": 0.9048683643341064,
      "num_tokens": 1935923747.0,
      "step": 834
    },
    {
      "epoch": 0.4273285568065507,
      "grad_norm": 0.20703125,
      "learning_rate": 2.5641881094285545e-05,
      "loss": 0.23180106282234192,
      "mean_token_accuracy": 0.9076418876647949,
      "num_tokens": 1938190250.0,
      "step": 835
    },
    {
      "epoch": 0.4278403275332651,
      "grad_norm": 0.22265625,
      "learning_rate": 2.5610063304265306e-05,
      "loss": 0.24199697375297546,
      "mean_token_accuracy": 0.9032477140426636,
      "num_tokens": 1940366766.0,
      "step": 836
    },
    {
      "epoch": 0.42835209825997955,
      "grad_norm": 0.220703125,
      "learning_rate": 2.5578230095501715e-05,
      "loss": 0.24248391389846802,
      "mean_token_accuracy": 0.9030354619026184,
      "num_tokens": 1942735679.0,
      "step": 837
    },
    {
      "epoch": 0.428863868986694,
      "grad_norm": 0.197265625,
      "learning_rate": 2.5546381555485448e-05,
      "loss": 0.23023414611816406,
      "mean_token_accuracy": 0.9088187217712402,
      "num_tokens": 1945137681.0,
      "step": 838
    },
    {
      "epoch": 0.4293756397134084,
      "grad_norm": 0.21875,
      "learning_rate": 2.551451777174927e-05,
      "loss": 0.2416648119688034,
      "mean_token_accuracy": 0.9029020071029663,
      "num_tokens": 1947433612.0,
      "step": 839
    },
    {
      "epoch": 0.42988741044012285,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.548263883186787e-05,
      "loss": 0.24912726879119873,
      "mean_token_accuracy": 0.9006466865539551,
      "num_tokens": 1949714774.0,
      "step": 840
    },
    {
      "epoch": 0.43039918116683723,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.5450744823457578e-05,
      "loss": 0.24519827961921692,
      "mean_token_accuracy": 0.9024604558944702,
      "num_tokens": 1952065884.0,
      "step": 841
    },
    {
      "epoch": 0.43091095189355166,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.5418835834176153e-05,
      "loss": 0.23246580362319946,
      "mean_token_accuracy": 0.9074460864067078,
      "num_tokens": 1954407900.0,
      "step": 842
    },
    {
      "epoch": 0.4314227226202661,
      "grad_norm": 0.216796875,
      "learning_rate": 2.538691195172251e-05,
      "loss": 0.23542174696922302,
      "mean_token_accuracy": 0.9058051705360413,
      "num_tokens": 1956701393.0,
      "step": 843
    },
    {
      "epoch": 0.43193449334698053,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.535497326383651e-05,
      "loss": 0.23297828435897827,
      "mean_token_accuracy": 0.9065630435943604,
      "num_tokens": 1958981490.0,
      "step": 844
    },
    {
      "epoch": 0.43244626407369496,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.5323019858298702e-05,
      "loss": 0.23436512053012848,
      "mean_token_accuracy": 0.906700611114502,
      "num_tokens": 1961362502.0,
      "step": 845
    },
    {
      "epoch": 0.4329580348004094,
      "grad_norm": 0.2353515625,
      "learning_rate": 2.5291051822930084e-05,
      "loss": 0.243670254945755,
      "mean_token_accuracy": 0.902607798576355,
      "num_tokens": 1963698678.0,
      "step": 846
    },
    {
      "epoch": 0.43346980552712383,
      "grad_norm": 0.205078125,
      "learning_rate": 2.5259069245591862e-05,
      "loss": 0.24633143842220306,
      "mean_token_accuracy": 0.9017248153686523,
      "num_tokens": 1965945100.0,
      "step": 847
    },
    {
      "epoch": 0.43398157625383826,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.522707221418521e-05,
      "loss": 0.2385929524898529,
      "mean_token_accuracy": 0.9045177698135376,
      "num_tokens": 1968256163.0,
      "step": 848
    },
    {
      "epoch": 0.4344933469805527,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.5195060816651027e-05,
      "loss": 0.2357458919286728,
      "mean_token_accuracy": 0.9051198959350586,
      "num_tokens": 1970564842.0,
      "step": 849
    },
    {
      "epoch": 0.43500511770726713,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.51630351409697e-05,
      "loss": 0.24326607584953308,
      "mean_token_accuracy": 0.9030519723892212,
      "num_tokens": 1972864872.0,
      "step": 850
    },
    {
      "epoch": 0.43551688843398156,
      "grad_norm": 0.228515625,
      "learning_rate": 2.513099527516085e-05,
      "loss": 0.24299931526184082,
      "mean_token_accuracy": 0.9025289416313171,
      "num_tokens": 1975074471.0,
      "step": 851
    },
    {
      "epoch": 0.436028659160696,
      "grad_norm": 0.203125,
      "learning_rate": 2.5098941307283103e-05,
      "loss": 0.22911803424358368,
      "mean_token_accuracy": 0.9086661338806152,
      "num_tokens": 1977454590.0,
      "step": 852
    },
    {
      "epoch": 0.43654042988741043,
      "grad_norm": 0.2265625,
      "learning_rate": 2.5066873325433845e-05,
      "loss": 0.2555042505264282,
      "mean_token_accuracy": 0.8977828621864319,
      "num_tokens": 1979740400.0,
      "step": 853
    },
    {
      "epoch": 0.43705220061412486,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.5034791417748987e-05,
      "loss": 0.23387137055397034,
      "mean_token_accuracy": 0.906950831413269,
      "num_tokens": 1982054090.0,
      "step": 854
    },
    {
      "epoch": 0.4375639713408393,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.5002695672402685e-05,
      "loss": 0.24106261134147644,
      "mean_token_accuracy": 0.9036184549331665,
      "num_tokens": 1984409901.0,
      "step": 855
    },
    {
      "epoch": 0.43807574206755373,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.4970586177607155e-05,
      "loss": 0.23481762409210205,
      "mean_token_accuracy": 0.9048020839691162,
      "num_tokens": 1986793012.0,
      "step": 856
    },
    {
      "epoch": 0.43858751279426816,
      "grad_norm": 0.22265625,
      "learning_rate": 2.4938463021612402e-05,
      "loss": 0.24429544806480408,
      "mean_token_accuracy": 0.9017348289489746,
      "num_tokens": 1989157303.0,
      "step": 857
    },
    {
      "epoch": 0.4390992835209826,
      "grad_norm": 0.22265625,
      "learning_rate": 2.4906326292705947e-05,
      "loss": 0.2507425546646118,
      "mean_token_accuracy": 0.8996096849441528,
      "num_tokens": 1991566258.0,
      "step": 858
    },
    {
      "epoch": 0.43961105424769703,
      "grad_norm": 0.2265625,
      "learning_rate": 2.4874176079212645e-05,
      "loss": 0.2476508617401123,
      "mean_token_accuracy": 0.901013970375061,
      "num_tokens": 1993896127.0,
      "step": 859
    },
    {
      "epoch": 0.44012282497441146,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.4842012469494404e-05,
      "loss": 0.24817486107349396,
      "mean_token_accuracy": 0.8995844125747681,
      "num_tokens": 1996286187.0,
      "step": 860
    },
    {
      "epoch": 0.4406345957011259,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.480983555194995e-05,
      "loss": 0.24054773151874542,
      "mean_token_accuracy": 0.9034702777862549,
      "num_tokens": 1998625771.0,
      "step": 861
    },
    {
      "epoch": 0.44114636642784033,
      "grad_norm": 0.2333984375,
      "learning_rate": 2.477764541501458e-05,
      "loss": 0.24857361614704132,
      "mean_token_accuracy": 0.9008734226226807,
      "num_tokens": 2000987763.0,
      "step": 862
    },
    {
      "epoch": 0.44165813715455476,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.4745442147159927e-05,
      "loss": 0.24742761254310608,
      "mean_token_accuracy": 0.9017316102981567,
      "num_tokens": 2003263161.0,
      "step": 863
    },
    {
      "epoch": 0.4421699078812692,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.471322583689372e-05,
      "loss": 0.2432578206062317,
      "mean_token_accuracy": 0.9024760127067566,
      "num_tokens": 2005600352.0,
      "step": 864
    },
    {
      "epoch": 0.44268167860798363,
      "grad_norm": 0.216796875,
      "learning_rate": 2.4680996572759514e-05,
      "loss": 0.23977509140968323,
      "mean_token_accuracy": 0.9042637348175049,
      "num_tokens": 2007904799.0,
      "step": 865
    },
    {
      "epoch": 0.44319344933469806,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.464875444333649e-05,
      "loss": 0.2245306521654129,
      "mean_token_accuracy": 0.9086848497390747,
      "num_tokens": 2010260994.0,
      "step": 866
    },
    {
      "epoch": 0.4437052200614125,
      "grad_norm": 0.2314453125,
      "learning_rate": 2.461649953723916e-05,
      "loss": 0.2577112317085266,
      "mean_token_accuracy": 0.8973844051361084,
      "num_tokens": 2012610556.0,
      "step": 867
    },
    {
      "epoch": 0.44421699078812693,
      "grad_norm": 0.21484375,
      "learning_rate": 2.4584231943117196e-05,
      "loss": 0.2444671392440796,
      "mean_token_accuracy": 0.9024501442909241,
      "num_tokens": 2014916487.0,
      "step": 868
    },
    {
      "epoch": 0.44472876151484136,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.4551951749655086e-05,
      "loss": 0.24965384602546692,
      "mean_token_accuracy": 0.9003263115882874,
      "num_tokens": 2017197061.0,
      "step": 869
    },
    {
      "epoch": 0.4452405322415558,
      "grad_norm": 0.203125,
      "learning_rate": 2.4519659045572005e-05,
      "loss": 0.22202329337596893,
      "mean_token_accuracy": 0.9113013744354248,
      "num_tokens": 2019522474.0,
      "step": 870
    },
    {
      "epoch": 0.44575230296827023,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.4487353919621452e-05,
      "loss": 0.24411368370056152,
      "mean_token_accuracy": 0.9018816351890564,
      "num_tokens": 2021817654.0,
      "step": 871
    },
    {
      "epoch": 0.44626407369498466,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.4455036460591113e-05,
      "loss": 0.228296160697937,
      "mean_token_accuracy": 0.9084150195121765,
      "num_tokens": 2024153934.0,
      "step": 872
    },
    {
      "epoch": 0.4467758444216991,
      "grad_norm": 0.232421875,
      "learning_rate": 2.442270675730255e-05,
      "loss": 0.2451878935098648,
      "mean_token_accuracy": 0.9011752605438232,
      "num_tokens": 2026421555.0,
      "step": 873
    },
    {
      "epoch": 0.44728761514841353,
      "grad_norm": 0.21484375,
      "learning_rate": 2.4390364898610983e-05,
      "loss": 0.25157493352890015,
      "mean_token_accuracy": 0.8989896774291992,
      "num_tokens": 2028744533.0,
      "step": 874
    },
    {
      "epoch": 0.44779938587512796,
      "grad_norm": 0.232421875,
      "learning_rate": 2.4358010973405035e-05,
      "loss": 0.26127728819847107,
      "mean_token_accuracy": 0.8948214054107666,
      "num_tokens": 2031101180.0,
      "step": 875
    },
    {
      "epoch": 0.4483111566018424,
      "grad_norm": 0.2275390625,
      "learning_rate": 2.4325645070606494e-05,
      "loss": 0.25488847494125366,
      "mean_token_accuracy": 0.8980970978736877,
      "num_tokens": 2033392848.0,
      "step": 876
    },
    {
      "epoch": 0.44882292732855683,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.429326727917008e-05,
      "loss": 0.23262593150138855,
      "mean_token_accuracy": 0.9074921607971191,
      "num_tokens": 2035777264.0,
      "step": 877
    },
    {
      "epoch": 0.44933469805527126,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.4260877688083156e-05,
      "loss": 0.24084976315498352,
      "mean_token_accuracy": 0.9027462005615234,
      "num_tokens": 2038059506.0,
      "step": 878
    },
    {
      "epoch": 0.4498464687819857,
      "grad_norm": 0.228515625,
      "learning_rate": 2.4228476386365552e-05,
      "loss": 0.24548330903053284,
      "mean_token_accuracy": 0.9003421664237976,
      "num_tokens": 2040352270.0,
      "step": 879
    },
    {
      "epoch": 0.4503582395087001,
      "grad_norm": 0.2109375,
      "learning_rate": 2.4196063463069272e-05,
      "loss": 0.23449887335300446,
      "mean_token_accuracy": 0.906043529510498,
      "num_tokens": 2042680655.0,
      "step": 880
    },
    {
      "epoch": 0.4508700102354145,
      "grad_norm": 0.216796875,
      "learning_rate": 2.4163639007278243e-05,
      "loss": 0.24064227938652039,
      "mean_token_accuracy": 0.9024906158447266,
      "num_tokens": 2045001779.0,
      "step": 881
    },
    {
      "epoch": 0.45138178096212894,
      "grad_norm": 0.2265625,
      "learning_rate": 2.4131203108108104e-05,
      "loss": 0.2534548342227936,
      "mean_token_accuracy": 0.8986525535583496,
      "num_tokens": 2047338441.0,
      "step": 882
    },
    {
      "epoch": 0.4518935516888434,
      "grad_norm": 0.23046875,
      "learning_rate": 2.4098755854705945e-05,
      "loss": 0.2520129084587097,
      "mean_token_accuracy": 0.8996820449829102,
      "num_tokens": 2049605396.0,
      "step": 883
    },
    {
      "epoch": 0.4524053224155578,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.4066297336250063e-05,
      "loss": 0.229930117726326,
      "mean_token_accuracy": 0.908171534538269,
      "num_tokens": 2051897913.0,
      "step": 884
    },
    {
      "epoch": 0.45291709314227224,
      "grad_norm": 0.21484375,
      "learning_rate": 2.403382764194971e-05,
      "loss": 0.23793238401412964,
      "mean_token_accuracy": 0.9054166674613953,
      "num_tokens": 2054131972.0,
      "step": 885
    },
    {
      "epoch": 0.4534288638689867,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.400134686104485e-05,
      "loss": 0.24680352210998535,
      "mean_token_accuracy": 0.9008627533912659,
      "num_tokens": 2056409146.0,
      "step": 886
    },
    {
      "epoch": 0.4539406345957011,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.3968855082805938e-05,
      "loss": 0.2466234564781189,
      "mean_token_accuracy": 0.9016786217689514,
      "num_tokens": 2058682645.0,
      "step": 887
    },
    {
      "epoch": 0.45445240532241554,
      "grad_norm": 0.220703125,
      "learning_rate": 2.3936352396533625e-05,
      "loss": 0.2427237331867218,
      "mean_token_accuracy": 0.9028965830802917,
      "num_tokens": 2060986001.0,
      "step": 888
    },
    {
      "epoch": 0.45496417604913,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.3903838891558566e-05,
      "loss": 0.24414557218551636,
      "mean_token_accuracy": 0.9022408127784729,
      "num_tokens": 2063337135.0,
      "step": 889
    },
    {
      "epoch": 0.4554759467758444,
      "grad_norm": 0.205078125,
      "learning_rate": 2.3871314657241132e-05,
      "loss": 0.2324252873659134,
      "mean_token_accuracy": 0.907265305519104,
      "num_tokens": 2065744835.0,
      "step": 890
    },
    {
      "epoch": 0.45598771750255884,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.3838779782971204e-05,
      "loss": 0.22330494225025177,
      "mean_token_accuracy": 0.909885823726654,
      "num_tokens": 2068067892.0,
      "step": 891
    },
    {
      "epoch": 0.4564994882292733,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.3806234358167884e-05,
      "loss": 0.24085098505020142,
      "mean_token_accuracy": 0.9028595685958862,
      "num_tokens": 2070371402.0,
      "step": 892
    },
    {
      "epoch": 0.4570112589559877,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.3773678472279285e-05,
      "loss": 0.2492145150899887,
      "mean_token_accuracy": 0.9004114270210266,
      "num_tokens": 2072729861.0,
      "step": 893
    },
    {
      "epoch": 0.45752302968270214,
      "grad_norm": 0.2314453125,
      "learning_rate": 2.3741112214782268e-05,
      "loss": 0.2452218234539032,
      "mean_token_accuracy": 0.9021847248077393,
      "num_tokens": 2075054496.0,
      "step": 894
    },
    {
      "epoch": 0.4580348004094166,
      "grad_norm": 0.23046875,
      "learning_rate": 2.370853567518219e-05,
      "loss": 0.2586902976036072,
      "mean_token_accuracy": 0.8958449959754944,
      "num_tokens": 2077437199.0,
      "step": 895
    },
    {
      "epoch": 0.458546571136131,
      "grad_norm": 0.21875,
      "learning_rate": 2.3675948943012686e-05,
      "loss": 0.2462834119796753,
      "mean_token_accuracy": 0.9016544818878174,
      "num_tokens": 2079752019.0,
      "step": 896
    },
    {
      "epoch": 0.45905834186284544,
      "grad_norm": 0.208984375,
      "learning_rate": 2.3643352107835385e-05,
      "loss": 0.2387780249118805,
      "mean_token_accuracy": 0.9043389558792114,
      "num_tokens": 2082102413.0,
      "step": 897
    },
    {
      "epoch": 0.4595701125895599,
      "grad_norm": 0.212890625,
      "learning_rate": 2.3610745259239706e-05,
      "loss": 0.24922625720500946,
      "mean_token_accuracy": 0.9002403616905212,
      "num_tokens": 2084415503.0,
      "step": 898
    },
    {
      "epoch": 0.4600818833162743,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.357812848684256e-05,
      "loss": 0.23388271033763885,
      "mean_token_accuracy": 0.9063811302185059,
      "num_tokens": 2086692531.0,
      "step": 899
    },
    {
      "epoch": 0.46059365404298874,
      "grad_norm": 0.220703125,
      "learning_rate": 2.3545501880288158e-05,
      "loss": 0.23414306342601776,
      "mean_token_accuracy": 0.906059980392456,
      "num_tokens": 2089007514.0,
      "step": 900
    },
    {
      "epoch": 0.4611054247697032,
      "grad_norm": 0.212890625,
      "learning_rate": 2.3512865529247725e-05,
      "loss": 0.242130309343338,
      "mean_token_accuracy": 0.9024388790130615,
      "num_tokens": 2091314650.0,
      "step": 901
    },
    {
      "epoch": 0.4616171954964176,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.3480219523419264e-05,
      "loss": 0.2394711673259735,
      "mean_token_accuracy": 0.9045045375823975,
      "num_tokens": 2093673143.0,
      "step": 902
    },
    {
      "epoch": 0.46212896622313204,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.344756395252734e-05,
      "loss": 0.23821371793746948,
      "mean_token_accuracy": 0.9048156142234802,
      "num_tokens": 2096048935.0,
      "step": 903
    },
    {
      "epoch": 0.4626407369498465,
      "grad_norm": 0.2109375,
      "learning_rate": 2.3414898906322776e-05,
      "loss": 0.23315371572971344,
      "mean_token_accuracy": 0.9065717458724976,
      "num_tokens": 2098435713.0,
      "step": 904
    },
    {
      "epoch": 0.4631525076765609,
      "grad_norm": 0.22265625,
      "learning_rate": 2.338222447458245e-05,
      "loss": 0.2510300874710083,
      "mean_token_accuracy": 0.899447500705719,
      "num_tokens": 2100688614.0,
      "step": 905
    },
    {
      "epoch": 0.46366427840327534,
      "grad_norm": 0.21484375,
      "learning_rate": 2.3349540747109033e-05,
      "loss": 0.2418459802865982,
      "mean_token_accuracy": 0.9039223194122314,
      "num_tokens": 2102994562.0,
      "step": 906
    },
    {
      "epoch": 0.4641760491299898,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.3316847813730755e-05,
      "loss": 0.23242053389549255,
      "mean_token_accuracy": 0.9073715209960938,
      "num_tokens": 2105409211.0,
      "step": 907
    },
    {
      "epoch": 0.4646878198567042,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.3284145764301128e-05,
      "loss": 0.23046347498893738,
      "mean_token_accuracy": 0.9081612825393677,
      "num_tokens": 2107758754.0,
      "step": 908
    },
    {
      "epoch": 0.46519959058341864,
      "grad_norm": 0.203125,
      "learning_rate": 2.3251434688698727e-05,
      "loss": 0.22762297093868256,
      "mean_token_accuracy": 0.9084324836730957,
      "num_tokens": 2110202999.0,
      "step": 909
    },
    {
      "epoch": 0.4657113613101331,
      "grad_norm": 0.21484375,
      "learning_rate": 2.3218714676826952e-05,
      "loss": 0.22502781450748444,
      "mean_token_accuracy": 0.9091217517852783,
      "num_tokens": 2112481762.0,
      "step": 910
    },
    {
      "epoch": 0.4662231320368475,
      "grad_norm": 0.224609375,
      "learning_rate": 2.3185985818613732e-05,
      "loss": 0.24317046999931335,
      "mean_token_accuracy": 0.902564287185669,
      "num_tokens": 2114832087.0,
      "step": 911
    },
    {
      "epoch": 0.46673490276356194,
      "grad_norm": 0.2294921875,
      "learning_rate": 2.3153248204011333e-05,
      "loss": 0.255307137966156,
      "mean_token_accuracy": 0.8976502418518066,
      "num_tokens": 2117175690.0,
      "step": 912
    },
    {
      "epoch": 0.4672466734902764,
      "grad_norm": 0.21484375,
      "learning_rate": 2.3120501922996078e-05,
      "loss": 0.23302549123764038,
      "mean_token_accuracy": 0.9064841270446777,
      "num_tokens": 2119546889.0,
      "step": 913
    },
    {
      "epoch": 0.4677584442169908,
      "grad_norm": 0.2265625,
      "learning_rate": 2.3087747065568117e-05,
      "loss": 0.23747578263282776,
      "mean_token_accuracy": 0.9047103524208069,
      "num_tokens": 2121853581.0,
      "step": 914
    },
    {
      "epoch": 0.46827021494370524,
      "grad_norm": 0.22265625,
      "learning_rate": 2.305498372175115e-05,
      "loss": 0.2523254156112671,
      "mean_token_accuracy": 0.8991957306861877,
      "num_tokens": 2124165424.0,
      "step": 915
    },
    {
      "epoch": 0.4687819856704197,
      "grad_norm": 0.2109375,
      "learning_rate": 2.302221198159223e-05,
      "loss": 0.23020127415657043,
      "mean_token_accuracy": 0.9068570137023926,
      "num_tokens": 2126450505.0,
      "step": 916
    },
    {
      "epoch": 0.4692937563971341,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.298943193516147e-05,
      "loss": 0.24708376824855804,
      "mean_token_accuracy": 0.9006466269493103,
      "num_tokens": 2128793727.0,
      "step": 917
    },
    {
      "epoch": 0.46980552712384854,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.2956643672551804e-05,
      "loss": 0.24288558959960938,
      "mean_token_accuracy": 0.9031857848167419,
      "num_tokens": 2131204074.0,
      "step": 918
    },
    {
      "epoch": 0.4703172978505629,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.2923847283878777e-05,
      "loss": 0.24008920788764954,
      "mean_token_accuracy": 0.9038950204849243,
      "num_tokens": 2133577002.0,
      "step": 919
    },
    {
      "epoch": 0.47082906857727735,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.2891042859280227e-05,
      "loss": 0.24225588142871857,
      "mean_token_accuracy": 0.9024549722671509,
      "num_tokens": 2135867289.0,
      "step": 920
    },
    {
      "epoch": 0.4713408393039918,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.285823048891612e-05,
      "loss": 0.2356199026107788,
      "mean_token_accuracy": 0.905686616897583,
      "num_tokens": 2138211311.0,
      "step": 921
    },
    {
      "epoch": 0.4718526100307062,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.2825410262968223e-05,
      "loss": 0.23472660779953003,
      "mean_token_accuracy": 0.9058970212936401,
      "num_tokens": 2140554938.0,
      "step": 922
    },
    {
      "epoch": 0.47236438075742065,
      "grad_norm": 0.2109375,
      "learning_rate": 2.2792582271639922e-05,
      "loss": 0.24154981970787048,
      "mean_token_accuracy": 0.9032675623893738,
      "num_tokens": 2142890883.0,
      "step": 923
    },
    {
      "epoch": 0.4728761514841351,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.2759746605155934e-05,
      "loss": 0.25006264448165894,
      "mean_token_accuracy": 0.9006325602531433,
      "num_tokens": 2145156712.0,
      "step": 924
    },
    {
      "epoch": 0.4733879222108495,
      "grad_norm": 0.220703125,
      "learning_rate": 2.272690335376206e-05,
      "loss": 0.2486068159341812,
      "mean_token_accuracy": 0.9000714421272278,
      "num_tokens": 2147480178.0,
      "step": 925
    },
    {
      "epoch": 0.47389969293756395,
      "grad_norm": 0.216796875,
      "learning_rate": 2.2694052607724966e-05,
      "loss": 0.2481604963541031,
      "mean_token_accuracy": 0.9006718397140503,
      "num_tokens": 2149790917.0,
      "step": 926
    },
    {
      "epoch": 0.4744114636642784,
      "grad_norm": 0.224609375,
      "learning_rate": 2.26611944573319e-05,
      "loss": 0.23617756366729736,
      "mean_token_accuracy": 0.905422568321228,
      "num_tokens": 2152103226.0,
      "step": 927
    },
    {
      "epoch": 0.4749232343909928,
      "grad_norm": 0.212890625,
      "learning_rate": 2.2628328992890484e-05,
      "loss": 0.2353382557630539,
      "mean_token_accuracy": 0.9056463241577148,
      "num_tokens": 2154396875.0,
      "step": 928
    },
    {
      "epoch": 0.47543500511770725,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.2595456304728404e-05,
      "loss": 0.24457982182502747,
      "mean_token_accuracy": 0.9023004770278931,
      "num_tokens": 2156738381.0,
      "step": 929
    },
    {
      "epoch": 0.4759467758444217,
      "grad_norm": 5.40625,
      "learning_rate": 2.256257648319325e-05,
      "loss": 0.3827543258666992,
      "mean_token_accuracy": 0.8960498571395874,
      "num_tokens": 2159077064.0,
      "step": 930
    },
    {
      "epoch": 0.4764585465711361,
      "grad_norm": 0.21484375,
      "learning_rate": 2.252968961865216e-05,
      "loss": 0.23312145471572876,
      "mean_token_accuracy": 0.9067120552062988,
      "num_tokens": 2161380419.0,
      "step": 931
    },
    {
      "epoch": 0.47697031729785055,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.249679580149167e-05,
      "loss": 0.24088236689567566,
      "mean_token_accuracy": 0.9037264585494995,
      "num_tokens": 2163714396.0,
      "step": 932
    },
    {
      "epoch": 0.477482088024565,
      "grad_norm": 0.2265625,
      "learning_rate": 2.2463895122117428e-05,
      "loss": 0.2476264238357544,
      "mean_token_accuracy": 0.9013989567756653,
      "num_tokens": 2165974414.0,
      "step": 933
    },
    {
      "epoch": 0.4779938587512794,
      "grad_norm": 0.21875,
      "learning_rate": 2.2430987670953904e-05,
      "loss": 0.22714769840240479,
      "mean_token_accuracy": 0.9092563986778259,
      "num_tokens": 2168324297.0,
      "step": 934
    },
    {
      "epoch": 0.47850562947799385,
      "grad_norm": 0.2333984375,
      "learning_rate": 2.239807353844421e-05,
      "loss": 0.2613288164138794,
      "mean_token_accuracy": 0.8962797522544861,
      "num_tokens": 2170649285.0,
      "step": 935
    },
    {
      "epoch": 0.4790174002047083,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.236515281504982e-05,
      "loss": 0.2610067129135132,
      "mean_token_accuracy": 0.8964434266090393,
      "num_tokens": 2172971294.0,
      "step": 936
    },
    {
      "epoch": 0.4795291709314227,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.2332225591250313e-05,
      "loss": 0.2245626151561737,
      "mean_token_accuracy": 0.9104319214820862,
      "num_tokens": 2175276326.0,
      "step": 937
    },
    {
      "epoch": 0.48004094165813715,
      "grad_norm": 0.2265625,
      "learning_rate": 2.2299291957543137e-05,
      "loss": 0.25086504220962524,
      "mean_token_accuracy": 0.8997451663017273,
      "num_tokens": 2177612146.0,
      "step": 938
    },
    {
      "epoch": 0.4805527123848516,
      "grad_norm": 0.2275390625,
      "learning_rate": 2.2266352004443358e-05,
      "loss": 0.24927952885627747,
      "mean_token_accuracy": 0.9004132747650146,
      "num_tokens": 2179947266.0,
      "step": 939
    },
    {
      "epoch": 0.481064483111566,
      "grad_norm": 0.220703125,
      "learning_rate": 2.2233405822483408e-05,
      "loss": 0.23952211439609528,
      "mean_token_accuracy": 0.9038504362106323,
      "num_tokens": 2182261671.0,
      "step": 940
    },
    {
      "epoch": 0.48157625383828045,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.220045350221284e-05,
      "loss": 0.2520080804824829,
      "mean_token_accuracy": 0.899161159992218,
      "num_tokens": 2184536563.0,
      "step": 941
    },
    {
      "epoch": 0.4820880245649949,
      "grad_norm": 0.21484375,
      "learning_rate": 2.2167495134198073e-05,
      "loss": 0.23782022297382355,
      "mean_token_accuracy": 0.9055663347244263,
      "num_tokens": 2186842596.0,
      "step": 942
    },
    {
      "epoch": 0.4825997952917093,
      "grad_norm": 0.2294921875,
      "learning_rate": 2.2134530809022162e-05,
      "loss": 0.2520570158958435,
      "mean_token_accuracy": 0.8996591567993164,
      "num_tokens": 2189192787.0,
      "step": 943
    },
    {
      "epoch": 0.48311156601842375,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.210156061728452e-05,
      "loss": 0.2506716549396515,
      "mean_token_accuracy": 0.8996166586875916,
      "num_tokens": 2191535742.0,
      "step": 944
    },
    {
      "epoch": 0.4836233367451382,
      "grad_norm": 0.20703125,
      "learning_rate": 2.2068584649600684e-05,
      "loss": 0.22747543454170227,
      "mean_token_accuracy": 0.9086758494377136,
      "num_tokens": 2193856550.0,
      "step": 945
    },
    {
      "epoch": 0.4841351074718526,
      "grad_norm": 0.22265625,
      "learning_rate": 2.2035602996602074e-05,
      "loss": 0.24261368811130524,
      "mean_token_accuracy": 0.9028343558311462,
      "num_tokens": 2196203261.0,
      "step": 946
    },
    {
      "epoch": 0.48464687819856705,
      "grad_norm": 0.220703125,
      "learning_rate": 2.200261574893574e-05,
      "loss": 0.2387663722038269,
      "mean_token_accuracy": 0.9040470719337463,
      "num_tokens": 2198517817.0,
      "step": 947
    },
    {
      "epoch": 0.4851586489252815,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.1969622997264084e-05,
      "loss": 0.24347040057182312,
      "mean_token_accuracy": 0.9021696448326111,
      "num_tokens": 2200757848.0,
      "step": 948
    },
    {
      "epoch": 0.4856704196519959,
      "grad_norm": 0.21484375,
      "learning_rate": 2.1936624832264665e-05,
      "loss": 0.2403295487165451,
      "mean_token_accuracy": 0.9044336080551147,
      "num_tokens": 2203068154.0,
      "step": 949
    },
    {
      "epoch": 0.48618219037871035,
      "grad_norm": 0.224609375,
      "learning_rate": 2.1903621344629892e-05,
      "loss": 0.24066472053527832,
      "mean_token_accuracy": 0.9037690758705139,
      "num_tokens": 2205432590.0,
      "step": 950
    },
    {
      "epoch": 0.4866939611054248,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.187061262506683e-05,
      "loss": 0.22863197326660156,
      "mean_token_accuracy": 0.9088470935821533,
      "num_tokens": 2207816531.0,
      "step": 951
    },
    {
      "epoch": 0.4872057318321392,
      "grad_norm": 0.224609375,
      "learning_rate": 2.183759876429691e-05,
      "loss": 0.24785694479942322,
      "mean_token_accuracy": 0.9006286263465881,
      "num_tokens": 2210143433.0,
      "step": 952
    },
    {
      "epoch": 0.48771750255885365,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.180457985305569e-05,
      "loss": 0.23105363547801971,
      "mean_token_accuracy": 0.9079128503799438,
      "num_tokens": 2212486856.0,
      "step": 953
    },
    {
      "epoch": 0.4882292732855681,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.1771555982092615e-05,
      "loss": 0.24187663197517395,
      "mean_token_accuracy": 0.9035848379135132,
      "num_tokens": 2214809632.0,
      "step": 954
    },
    {
      "epoch": 0.4887410440122825,
      "grad_norm": 0.203125,
      "learning_rate": 2.173852724217075e-05,
      "loss": 0.2299591302871704,
      "mean_token_accuracy": 0.9078748226165771,
      "num_tokens": 2217168566.0,
      "step": 955
    },
    {
      "epoch": 0.48925281473899696,
      "grad_norm": 0.208984375,
      "learning_rate": 2.1705493724066567e-05,
      "loss": 0.23360417783260345,
      "mean_token_accuracy": 0.9060325622558594,
      "num_tokens": 2219483962.0,
      "step": 956
    },
    {
      "epoch": 0.4897645854657114,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.1672455518569637e-05,
      "loss": 0.2469150722026825,
      "mean_token_accuracy": 0.9009710550308228,
      "num_tokens": 2221791290.0,
      "step": 957
    },
    {
      "epoch": 0.49027635619242577,
      "grad_norm": 0.203125,
      "learning_rate": 2.1639412716482444e-05,
      "loss": 0.22991932928562164,
      "mean_token_accuracy": 0.9073364734649658,
      "num_tokens": 2224143845.0,
      "step": 958
    },
    {
      "epoch": 0.4907881269191402,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.1606365408620086e-05,
      "loss": 0.24906453490257263,
      "mean_token_accuracy": 0.9007817506790161,
      "num_tokens": 2226502757.0,
      "step": 959
    },
    {
      "epoch": 0.49129989764585463,
      "grad_norm": 0.220703125,
      "learning_rate": 2.1573313685810056e-05,
      "loss": 0.2396901249885559,
      "mean_token_accuracy": 0.9042616486549377,
      "num_tokens": 2228898445.0,
      "step": 960
    },
    {
      "epoch": 0.49181166837256907,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.1540257638891968e-05,
      "loss": 0.23958274722099304,
      "mean_token_accuracy": 0.9041556715965271,
      "num_tokens": 2231228772.0,
      "step": 961
    },
    {
      "epoch": 0.4923234390992835,
      "grad_norm": 0.19921875,
      "learning_rate": 2.1507197358717337e-05,
      "loss": 0.24283327162265778,
      "mean_token_accuracy": 0.9031631350517273,
      "num_tokens": 2233607950.0,
      "step": 962
    },
    {
      "epoch": 0.49283520982599793,
      "grad_norm": 0.232421875,
      "learning_rate": 2.14741329361493e-05,
      "loss": 0.2505471706390381,
      "mean_token_accuracy": 0.8992401361465454,
      "num_tokens": 2235923811.0,
      "step": 963
    },
    {
      "epoch": 0.49334698055271237,
      "grad_norm": 0.216796875,
      "learning_rate": 2.1441064462062387e-05,
      "loss": 0.24110093712806702,
      "mean_token_accuracy": 0.9034450054168701,
      "num_tokens": 2238211571.0,
      "step": 964
    },
    {
      "epoch": 0.4938587512794268,
      "grad_norm": 0.224609375,
      "learning_rate": 2.140799202734225e-05,
      "loss": 0.2554323673248291,
      "mean_token_accuracy": 0.8972551822662354,
      "num_tokens": 2240514869.0,
      "step": 965
    },
    {
      "epoch": 0.49437052200614123,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.1374915722885444e-05,
      "loss": 0.2304881513118744,
      "mean_token_accuracy": 0.9078360199928284,
      "num_tokens": 2242878217.0,
      "step": 966
    },
    {
      "epoch": 0.49488229273285567,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.1341835639599154e-05,
      "loss": 0.2389930635690689,
      "mean_token_accuracy": 0.9047361612319946,
      "num_tokens": 2245217195.0,
      "step": 967
    },
    {
      "epoch": 0.4953940634595701,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.1308751868400947e-05,
      "loss": 0.255341112613678,
      "mean_token_accuracy": 0.8986247777938843,
      "num_tokens": 2247498937.0,
      "step": 968
    },
    {
      "epoch": 0.49590583418628453,
      "grad_norm": 0.20703125,
      "learning_rate": 2.1275664500218514e-05,
      "loss": 0.22911149263381958,
      "mean_token_accuracy": 0.9076403379440308,
      "num_tokens": 2249859524.0,
      "step": 969
    },
    {
      "epoch": 0.49641760491299897,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.1242573625989466e-05,
      "loss": 0.2472180426120758,
      "mean_token_accuracy": 0.9006569385528564,
      "num_tokens": 2252219372.0,
      "step": 970
    },
    {
      "epoch": 0.4969293756397134,
      "grad_norm": 0.21484375,
      "learning_rate": 2.120947933666101e-05,
      "loss": 0.23480567336082458,
      "mean_token_accuracy": 0.9057515859603882,
      "num_tokens": 2254485639.0,
      "step": 971
    },
    {
      "epoch": 0.49744114636642783,
      "grad_norm": 0.208984375,
      "learning_rate": 2.1176381723189775e-05,
      "loss": 0.2414826899766922,
      "mean_token_accuracy": 0.9039485454559326,
      "num_tokens": 2256836951.0,
      "step": 972
    },
    {
      "epoch": 0.49795291709314227,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.1143280876541498e-05,
      "loss": 0.2386275827884674,
      "mean_token_accuracy": 0.9039566516876221,
      "num_tokens": 2259137635.0,
      "step": 973
    },
    {
      "epoch": 0.4984646878198567,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.1110176887690824e-05,
      "loss": 0.22159408032894135,
      "mean_token_accuracy": 0.9109582901000977,
      "num_tokens": 2261438763.0,
      "step": 974
    },
    {
      "epoch": 0.49897645854657113,
      "grad_norm": 0.2109375,
      "learning_rate": 2.1077069847621012e-05,
      "loss": 0.23979297280311584,
      "mean_token_accuracy": 0.9037314057350159,
      "num_tokens": 2263804278.0,
      "step": 975
    },
    {
      "epoch": 0.49948822927328557,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.1043959847323715e-05,
      "loss": 0.2485138177871704,
      "mean_token_accuracy": 0.9002581238746643,
      "num_tokens": 2266160016.0,
      "step": 976
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.212890625,
      "learning_rate": 2.1010846977798745e-05,
      "loss": 0.23735478520393372,
      "mean_token_accuracy": 0.9052770137786865,
      "num_tokens": 2268491249.0,
      "step": 977
    },
    {
      "epoch": 0.5005117707267144,
      "grad_norm": 0.216796875,
      "learning_rate": 2.0977731330053766e-05,
      "loss": 0.23228245973587036,
      "mean_token_accuracy": 0.9065968990325928,
      "num_tokens": 2270834410.0,
      "step": 978
    },
    {
      "epoch": 0.5010235414534289,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.09446129951041e-05,
      "loss": 0.23551523685455322,
      "mean_token_accuracy": 0.9056209325790405,
      "num_tokens": 2273232076.0,
      "step": 979
    },
    {
      "epoch": 0.5015353121801432,
      "grad_norm": 0.201171875,
      "learning_rate": 2.091149206397243e-05,
      "loss": 0.22903989255428314,
      "mean_token_accuracy": 0.9083459973335266,
      "num_tokens": 2275638007.0,
      "step": 980
    },
    {
      "epoch": 0.5020470829068577,
      "grad_norm": 0.37109375,
      "learning_rate": 2.0878368627688618e-05,
      "loss": 0.24206386506557465,
      "mean_token_accuracy": 0.9030288457870483,
      "num_tokens": 2277849124.0,
      "step": 981
    },
    {
      "epoch": 0.5025588536335721,
      "grad_norm": 0.205078125,
      "learning_rate": 2.0845242777289362e-05,
      "loss": 0.22638970613479614,
      "mean_token_accuracy": 0.9100583791732788,
      "num_tokens": 2280182447.0,
      "step": 982
    },
    {
      "epoch": 0.5030706243602866,
      "grad_norm": 0.216796875,
      "learning_rate": 2.0812114603818023e-05,
      "loss": 0.24335363507270813,
      "mean_token_accuracy": 0.9028633832931519,
      "num_tokens": 2282546082.0,
      "step": 983
    },
    {
      "epoch": 0.503582395087001,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.0778984198324337e-05,
      "loss": 0.23735658824443817,
      "mean_token_accuracy": 0.9038137197494507,
      "num_tokens": 2284868360.0,
      "step": 984
    },
    {
      "epoch": 0.5040941658137155,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.074585165186419e-05,
      "loss": 0.2405419647693634,
      "mean_token_accuracy": 0.9029441475868225,
      "num_tokens": 2287193042.0,
      "step": 985
    },
    {
      "epoch": 0.5046059365404298,
      "grad_norm": 0.212890625,
      "learning_rate": 2.0712717055499332e-05,
      "loss": 0.2446592152118683,
      "mean_token_accuracy": 0.9018350839614868,
      "num_tokens": 2289596474.0,
      "step": 986
    },
    {
      "epoch": 0.5051177072671443,
      "grad_norm": 0.474609375,
      "learning_rate": 2.0679580500297147e-05,
      "loss": 0.24656769633293152,
      "mean_token_accuracy": 0.9020823240280151,
      "num_tokens": 2291945691.0,
      "step": 987
    },
    {
      "epoch": 0.5056294779938587,
      "grad_norm": 0.22265625,
      "learning_rate": 2.0646442077330433e-05,
      "loss": 0.2531547546386719,
      "mean_token_accuracy": 0.8989205360412598,
      "num_tokens": 2294273539.0,
      "step": 988
    },
    {
      "epoch": 0.5061412487205732,
      "grad_norm": 0.22265625,
      "learning_rate": 2.0613301877677086e-05,
      "loss": 0.24289318919181824,
      "mean_token_accuracy": 0.9026070833206177,
      "num_tokens": 2296644512.0,
      "step": 989
    },
    {
      "epoch": 0.5066530194472876,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.05801599924199e-05,
      "loss": 0.23883798718452454,
      "mean_token_accuracy": 0.9039129018783569,
      "num_tokens": 2298945242.0,
      "step": 990
    },
    {
      "epoch": 0.5071647901740021,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.05470165126463e-05,
      "loss": 0.24075280129909515,
      "mean_token_accuracy": 0.9042339324951172,
      "num_tokens": 2301348084.0,
      "step": 991
    },
    {
      "epoch": 0.5076765609007164,
      "grad_norm": 0.22265625,
      "learning_rate": 2.0513871529448092e-05,
      "loss": 0.24703040719032288,
      "mean_token_accuracy": 0.9010108113288879,
      "num_tokens": 2303672410.0,
      "step": 992
    },
    {
      "epoch": 0.5081883316274309,
      "grad_norm": 0.22265625,
      "learning_rate": 2.0480725133921222e-05,
      "loss": 0.2351382076740265,
      "mean_token_accuracy": 0.9065154194831848,
      "num_tokens": 2306022607.0,
      "step": 993
    },
    {
      "epoch": 0.5087001023541453,
      "grad_norm": 0.2333984375,
      "learning_rate": 2.0447577417165503e-05,
      "loss": 0.2403746247291565,
      "mean_token_accuracy": 0.9041292667388916,
      "num_tokens": 2308342793.0,
      "step": 994
    },
    {
      "epoch": 0.5092118730808598,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.0414428470284394e-05,
      "loss": 0.24167266488075256,
      "mean_token_accuracy": 0.9033075571060181,
      "num_tokens": 2310691213.0,
      "step": 995
    },
    {
      "epoch": 0.5097236438075742,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.0381278384384728e-05,
      "loss": 0.24317389726638794,
      "mean_token_accuracy": 0.9023122787475586,
      "num_tokens": 2313076243.0,
      "step": 996
    },
    {
      "epoch": 0.5102354145342887,
      "grad_norm": 0.236328125,
      "learning_rate": 2.034812725057647e-05,
      "loss": 0.25204697251319885,
      "mean_token_accuracy": 0.8990522027015686,
      "num_tokens": 2315441843.0,
      "step": 997
    },
    {
      "epoch": 0.510747185261003,
      "grad_norm": 0.2255859375,
      "learning_rate": 2.0314975159972465e-05,
      "loss": 0.24428069591522217,
      "mean_token_accuracy": 0.9024900197982788,
      "num_tokens": 2317672823.0,
      "step": 998
    },
    {
      "epoch": 0.5112589559877175,
      "grad_norm": 0.21484375,
      "learning_rate": 2.0281822203688174e-05,
      "loss": 0.24255579710006714,
      "mean_token_accuracy": 0.9024190306663513,
      "num_tokens": 2319921889.0,
      "step": 999
    },
    {
      "epoch": 0.5117707267144319,
      "grad_norm": 0.2216796875,
      "learning_rate": 2.024866847284147e-05,
      "loss": 0.23944807052612305,
      "mean_token_accuracy": 0.9040535092353821,
      "num_tokens": 2322270467.0,
      "step": 1000
    },
    {
      "epoch": 0.5122824974411464,
      "grad_norm": 0.208984375,
      "learning_rate": 2.0215514058552316e-05,
      "loss": 0.24436265230178833,
      "mean_token_accuracy": 0.9019676446914673,
      "num_tokens": 2324682028.0,
      "step": 1001
    },
    {
      "epoch": 0.5127942681678608,
      "grad_norm": 0.21875,
      "learning_rate": 2.0182359051942587e-05,
      "loss": 0.23873907327651978,
      "mean_token_accuracy": 0.9051768183708191,
      "num_tokens": 2327021273.0,
      "step": 1002
    },
    {
      "epoch": 0.5133060388945753,
      "grad_norm": 0.220703125,
      "learning_rate": 2.014920354413576e-05,
      "loss": 0.2424958497285843,
      "mean_token_accuracy": 0.9026447534561157,
      "num_tokens": 2329389741.0,
      "step": 1003
    },
    {
      "epoch": 0.5138178096212896,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.0116047626256714e-05,
      "loss": 0.2460601031780243,
      "mean_token_accuracy": 0.9012932777404785,
      "num_tokens": 2331713453.0,
      "step": 1004
    },
    {
      "epoch": 0.5143295803480041,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.0082891389431433e-05,
      "loss": 0.2358800768852234,
      "mean_token_accuracy": 0.9051352143287659,
      "num_tokens": 2334084947.0,
      "step": 1005
    },
    {
      "epoch": 0.5148413510747185,
      "grad_norm": 0.205078125,
      "learning_rate": 2.0049734924786776e-05,
      "loss": 0.24391266703605652,
      "mean_token_accuracy": 0.9020912051200867,
      "num_tokens": 2336444751.0,
      "step": 1006
    },
    {
      "epoch": 0.515353121801433,
      "grad_norm": 0.216796875,
      "learning_rate": 2.0016578323450262e-05,
      "loss": 0.24506732821464539,
      "mean_token_accuracy": 0.9019192457199097,
      "num_tokens": 2338833500.0,
      "step": 1007
    },
    {
      "epoch": 0.5158648925281474,
      "grad_norm": 0.22265625,
      "learning_rate": 1.9983421676549745e-05,
      "loss": 0.2566121816635132,
      "mean_token_accuracy": 0.8977725505828857,
      "num_tokens": 2341135979.0,
      "step": 1008
    },
    {
      "epoch": 0.5163766632548619,
      "grad_norm": 0.2109375,
      "learning_rate": 1.9950265075213228e-05,
      "loss": 0.23776935040950775,
      "mean_token_accuracy": 0.9049922227859497,
      "num_tokens": 2343457634.0,
      "step": 1009
    },
    {
      "epoch": 0.5168884339815762,
      "grad_norm": 0.2373046875,
      "learning_rate": 1.991710861056858e-05,
      "loss": 0.24995961785316467,
      "mean_token_accuracy": 0.900090754032135,
      "num_tokens": 2345792921.0,
      "step": 1010
    },
    {
      "epoch": 0.5174002047082907,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.98839523737433e-05,
      "loss": 0.23581629991531372,
      "mean_token_accuracy": 0.9046155214309692,
      "num_tokens": 2348089478.0,
      "step": 1011
    },
    {
      "epoch": 0.5179119754350051,
      "grad_norm": 0.20703125,
      "learning_rate": 1.985079645586424e-05,
      "loss": 0.23893576860427856,
      "mean_token_accuracy": 0.9039298295974731,
      "num_tokens": 2350441363.0,
      "step": 1012
    },
    {
      "epoch": 0.5184237461617196,
      "grad_norm": 0.20703125,
      "learning_rate": 1.9817640948057416e-05,
      "loss": 0.22692367434501648,
      "mean_token_accuracy": 0.9095773696899414,
      "num_tokens": 2352741761.0,
      "step": 1013
    },
    {
      "epoch": 0.518935516888434,
      "grad_norm": 0.212890625,
      "learning_rate": 1.9784485941447687e-05,
      "loss": 0.24796724319458008,
      "mean_token_accuracy": 0.9014312028884888,
      "num_tokens": 2355024811.0,
      "step": 1014
    },
    {
      "epoch": 0.5194472876151485,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.975133152715854e-05,
      "loss": 0.24468523263931274,
      "mean_token_accuracy": 0.9020880460739136,
      "num_tokens": 2357342606.0,
      "step": 1015
    },
    {
      "epoch": 0.5199590583418628,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.9718177796311833e-05,
      "loss": 0.23898480832576752,
      "mean_token_accuracy": 0.9038752913475037,
      "num_tokens": 2359701520.0,
      "step": 1016
    },
    {
      "epoch": 0.5204708290685772,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.9685024840027545e-05,
      "loss": 0.2516195774078369,
      "mean_token_accuracy": 0.8992189168930054,
      "num_tokens": 2362040680.0,
      "step": 1017
    },
    {
      "epoch": 0.5209825997952917,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.965187274942354e-05,
      "loss": 0.23949061334133148,
      "mean_token_accuracy": 0.9040783643722534,
      "num_tokens": 2364393905.0,
      "step": 1018
    },
    {
      "epoch": 0.5214943705220061,
      "grad_norm": 0.212890625,
      "learning_rate": 1.9618721615615272e-05,
      "loss": 0.232496976852417,
      "mean_token_accuracy": 0.9069885611534119,
      "num_tokens": 2366663785.0,
      "step": 1019
    },
    {
      "epoch": 0.5220061412487206,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.9585571529715606e-05,
      "loss": 0.24938343465328217,
      "mean_token_accuracy": 0.899982213973999,
      "num_tokens": 2368958938.0,
      "step": 1020
    },
    {
      "epoch": 0.522517911975435,
      "grad_norm": 0.2265625,
      "learning_rate": 1.95524225828345e-05,
      "loss": 0.2524535655975342,
      "mean_token_accuracy": 0.8990520238876343,
      "num_tokens": 2371268214.0,
      "step": 1021
    },
    {
      "epoch": 0.5230296827021494,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.9519274866078785e-05,
      "loss": 0.25808417797088623,
      "mean_token_accuracy": 0.8968746662139893,
      "num_tokens": 2373560085.0,
      "step": 1022
    },
    {
      "epoch": 0.5235414534288638,
      "grad_norm": 0.212890625,
      "learning_rate": 1.9486128470551915e-05,
      "loss": 0.23457393050193787,
      "mean_token_accuracy": 0.906116247177124,
      "num_tokens": 2375854874.0,
      "step": 1023
    },
    {
      "epoch": 0.5240532241555783,
      "grad_norm": 0.21484375,
      "learning_rate": 1.9452983487353706e-05,
      "loss": 0.2514454424381256,
      "mean_token_accuracy": 0.8991702795028687,
      "num_tokens": 2378229428.0,
      "step": 1024
    },
    {
      "epoch": 0.5245649948822927,
      "grad_norm": 0.22265625,
      "learning_rate": 1.9419840007580107e-05,
      "loss": 0.25161534547805786,
      "mean_token_accuracy": 0.8989554047584534,
      "num_tokens": 2380541898.0,
      "step": 1025
    },
    {
      "epoch": 0.5250767656090072,
      "grad_norm": 0.208984375,
      "learning_rate": 1.938669812232292e-05,
      "loss": 0.2263299822807312,
      "mean_token_accuracy": 0.9082002639770508,
      "num_tokens": 2382875867.0,
      "step": 1026
    },
    {
      "epoch": 0.5255885363357216,
      "grad_norm": 0.203125,
      "learning_rate": 1.9353557922669567e-05,
      "loss": 0.24835680425167084,
      "mean_token_accuracy": 0.900672435760498,
      "num_tokens": 2385203299.0,
      "step": 1027
    },
    {
      "epoch": 0.526100307062436,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.9320419499702853e-05,
      "loss": 0.24051877856254578,
      "mean_token_accuracy": 0.9037377834320068,
      "num_tokens": 2387616330.0,
      "step": 1028
    },
    {
      "epoch": 0.5266120777891504,
      "grad_norm": 0.21484375,
      "learning_rate": 1.9287282944500674e-05,
      "loss": 0.23745164275169373,
      "mean_token_accuracy": 0.9046916961669922,
      "num_tokens": 2389944328.0,
      "step": 1029
    },
    {
      "epoch": 0.5271238485158649,
      "grad_norm": 0.205078125,
      "learning_rate": 1.9254148348135817e-05,
      "loss": 0.23312783241271973,
      "mean_token_accuracy": 0.9066191911697388,
      "num_tokens": 2392327705.0,
      "step": 1030
    },
    {
      "epoch": 0.5276356192425793,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.9221015801675667e-05,
      "loss": 0.22487768530845642,
      "mean_token_accuracy": 0.9093253016471863,
      "num_tokens": 2394608651.0,
      "step": 1031
    },
    {
      "epoch": 0.5281473899692938,
      "grad_norm": 0.2314453125,
      "learning_rate": 1.9187885396181987e-05,
      "loss": 0.2543412744998932,
      "mean_token_accuracy": 0.8986214995384216,
      "num_tokens": 2396974292.0,
      "step": 1032
    },
    {
      "epoch": 0.5286591606960082,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.9154757222710648e-05,
      "loss": 0.23992332816123962,
      "mean_token_accuracy": 0.9032690525054932,
      "num_tokens": 2399317725.0,
      "step": 1033
    },
    {
      "epoch": 0.5291709314227226,
      "grad_norm": 0.2333984375,
      "learning_rate": 1.9121631372311385e-05,
      "loss": 0.25686925649642944,
      "mean_token_accuracy": 0.8970204591751099,
      "num_tokens": 2401744200.0,
      "step": 1034
    },
    {
      "epoch": 0.529682702149437,
      "grad_norm": 0.216796875,
      "learning_rate": 1.908850793602757e-05,
      "loss": 0.2396707534790039,
      "mean_token_accuracy": 0.9048184156417847,
      "num_tokens": 2404018907.0,
      "step": 1035
    },
    {
      "epoch": 0.5301944728761515,
      "grad_norm": 0.2109375,
      "learning_rate": 1.9055387004895903e-05,
      "loss": 0.23393571376800537,
      "mean_token_accuracy": 0.9057960510253906,
      "num_tokens": 2406417702.0,
      "step": 1036
    },
    {
      "epoch": 0.5307062436028659,
      "grad_norm": 0.20703125,
      "learning_rate": 1.9022268669946237e-05,
      "loss": 0.23629304766654968,
      "mean_token_accuracy": 0.9052932262420654,
      "num_tokens": 2408709333.0,
      "step": 1037
    },
    {
      "epoch": 0.5312180143295804,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.8989153022201258e-05,
      "loss": 0.23688778281211853,
      "mean_token_accuracy": 0.9045931100845337,
      "num_tokens": 2411048060.0,
      "step": 1038
    },
    {
      "epoch": 0.5317297850562948,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.8956040152676288e-05,
      "loss": 0.2425532042980194,
      "mean_token_accuracy": 0.9024364948272705,
      "num_tokens": 2413291255.0,
      "step": 1039
    },
    {
      "epoch": 0.5322415557830092,
      "grad_norm": 0.205078125,
      "learning_rate": 1.8922930152378998e-05,
      "loss": 0.24960795044898987,
      "mean_token_accuracy": 0.9007670283317566,
      "num_tokens": 2415641282.0,
      "step": 1040
    },
    {
      "epoch": 0.5327533265097236,
      "grad_norm": 0.201171875,
      "learning_rate": 1.888982311230919e-05,
      "loss": 0.23080599308013916,
      "mean_token_accuracy": 0.9070900082588196,
      "num_tokens": 2417927804.0,
      "step": 1041
    },
    {
      "epoch": 0.5332650972364381,
      "grad_norm": 0.208984375,
      "learning_rate": 1.8856719123458505e-05,
      "loss": 0.22647303342819214,
      "mean_token_accuracy": 0.9086161851882935,
      "num_tokens": 2420243743.0,
      "step": 1042
    },
    {
      "epoch": 0.5337768679631525,
      "grad_norm": 0.212890625,
      "learning_rate": 1.8823618276810228e-05,
      "loss": 0.24495983123779297,
      "mean_token_accuracy": 0.9023796319961548,
      "num_tokens": 2422583488.0,
      "step": 1043
    },
    {
      "epoch": 0.534288638689867,
      "grad_norm": 0.2001953125,
      "learning_rate": 1.8790520663338993e-05,
      "loss": 0.2228817343711853,
      "mean_token_accuracy": 0.9101508259773254,
      "num_tokens": 2424898021.0,
      "step": 1044
    },
    {
      "epoch": 0.5348004094165814,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.875742637401054e-05,
      "loss": 0.2410866916179657,
      "mean_token_accuracy": 0.9034696817398071,
      "num_tokens": 2427276334.0,
      "step": 1045
    },
    {
      "epoch": 0.5353121801432958,
      "grad_norm": 0.22265625,
      "learning_rate": 1.8724335499781493e-05,
      "loss": 0.24748753011226654,
      "mean_token_accuracy": 0.9016665816307068,
      "num_tokens": 2429644289.0,
      "step": 1046
    },
    {
      "epoch": 0.5358239508700102,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.8691248131599063e-05,
      "loss": 0.23413631319999695,
      "mean_token_accuracy": 0.9061568975448608,
      "num_tokens": 2432028315.0,
      "step": 1047
    },
    {
      "epoch": 0.5363357215967247,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.8658164360400856e-05,
      "loss": 0.2449459731578827,
      "mean_token_accuracy": 0.9024395942687988,
      "num_tokens": 2434363289.0,
      "step": 1048
    },
    {
      "epoch": 0.5368474923234391,
      "grad_norm": 0.216796875,
      "learning_rate": 1.862508427711456e-05,
      "loss": 0.2450811117887497,
      "mean_token_accuracy": 0.901879072189331,
      "num_tokens": 2436719908.0,
      "step": 1049
    },
    {
      "epoch": 0.5373592630501536,
      "grad_norm": 0.208984375,
      "learning_rate": 1.8592007972657752e-05,
      "loss": 0.21894767880439758,
      "mean_token_accuracy": 0.9114576578140259,
      "num_tokens": 2439009488.0,
      "step": 1050
    },
    {
      "epoch": 0.537871033776868,
      "grad_norm": 0.22265625,
      "learning_rate": 1.855893553793762e-05,
      "loss": 0.24506866931915283,
      "mean_token_accuracy": 0.9018581509590149,
      "num_tokens": 2441245880.0,
      "step": 1051
    },
    {
      "epoch": 0.5383828045035824,
      "grad_norm": 0.21875,
      "learning_rate": 1.8525867063850704e-05,
      "loss": 0.24321720004081726,
      "mean_token_accuracy": 0.903563380241394,
      "num_tokens": 2443521361.0,
      "step": 1052
    },
    {
      "epoch": 0.5388945752302968,
      "grad_norm": 0.208984375,
      "learning_rate": 1.849280264128267e-05,
      "loss": 0.23793555796146393,
      "mean_token_accuracy": 0.9048739075660706,
      "num_tokens": 2445863519.0,
      "step": 1053
    },
    {
      "epoch": 0.5394063459570113,
      "grad_norm": 0.205078125,
      "learning_rate": 1.8459742361108035e-05,
      "loss": 0.24046635627746582,
      "mean_token_accuracy": 0.9039778113365173,
      "num_tokens": 2448163967.0,
      "step": 1054
    },
    {
      "epoch": 0.5399181166837257,
      "grad_norm": 0.21875,
      "learning_rate": 1.842668631418995e-05,
      "loss": 0.24589504301548004,
      "mean_token_accuracy": 0.9010632038116455,
      "num_tokens": 2450408520.0,
      "step": 1055
    },
    {
      "epoch": 0.5404298874104401,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.8393634591379917e-05,
      "loss": 0.2510853707790375,
      "mean_token_accuracy": 0.8995676040649414,
      "num_tokens": 2452719176.0,
      "step": 1056
    },
    {
      "epoch": 0.5409416581371546,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.8360587283517556e-05,
      "loss": 0.23586755990982056,
      "mean_token_accuracy": 0.9049806594848633,
      "num_tokens": 2455043519.0,
      "step": 1057
    },
    {
      "epoch": 0.5414534288638689,
      "grad_norm": 0.20703125,
      "learning_rate": 1.8327544481430367e-05,
      "loss": 0.2386358380317688,
      "mean_token_accuracy": 0.9042952060699463,
      "num_tokens": 2457401169.0,
      "step": 1058
    },
    {
      "epoch": 0.5419651995905834,
      "grad_norm": 0.2109375,
      "learning_rate": 1.8294506275933436e-05,
      "loss": 0.24809345602989197,
      "mean_token_accuracy": 0.9002325534820557,
      "num_tokens": 2459783139.0,
      "step": 1059
    },
    {
      "epoch": 0.5424769703172978,
      "grad_norm": 0.220703125,
      "learning_rate": 1.8261472757829253e-05,
      "loss": 0.24084602296352386,
      "mean_token_accuracy": 0.9038909673690796,
      "num_tokens": 2461976098.0,
      "step": 1060
    },
    {
      "epoch": 0.5429887410440123,
      "grad_norm": 0.212890625,
      "learning_rate": 1.822844401790739e-05,
      "loss": 0.24166733026504517,
      "mean_token_accuracy": 0.9030689597129822,
      "num_tokens": 2464353202.0,
      "step": 1061
    },
    {
      "epoch": 0.5435005117707267,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.819542014694432e-05,
      "loss": 0.24798062443733215,
      "mean_token_accuracy": 0.9010701775550842,
      "num_tokens": 2466667978.0,
      "step": 1062
    },
    {
      "epoch": 0.5440122824974412,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.8162401235703096e-05,
      "loss": 0.2220243215560913,
      "mean_token_accuracy": 0.9107509255409241,
      "num_tokens": 2469022197.0,
      "step": 1063
    },
    {
      "epoch": 0.5445240532241555,
      "grad_norm": 0.2109375,
      "learning_rate": 1.812938737493317e-05,
      "loss": 0.23943200707435608,
      "mean_token_accuracy": 0.9043347239494324,
      "num_tokens": 2471450177.0,
      "step": 1064
    },
    {
      "epoch": 0.54503582395087,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.809637865537011e-05,
      "loss": 0.23722779750823975,
      "mean_token_accuracy": 0.9051599502563477,
      "num_tokens": 2473749616.0,
      "step": 1065
    },
    {
      "epoch": 0.5455475946775844,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.806337516773534e-05,
      "loss": 0.23565496504306793,
      "mean_token_accuracy": 0.9051924347877502,
      "num_tokens": 2476141289.0,
      "step": 1066
    },
    {
      "epoch": 0.5460593654042989,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.8030377002735923e-05,
      "loss": 0.22768619656562805,
      "mean_token_accuracy": 0.9085276126861572,
      "num_tokens": 2478311207.0,
      "step": 1067
    },
    {
      "epoch": 0.5465711361310133,
      "grad_norm": 0.21484375,
      "learning_rate": 1.7997384251064266e-05,
      "loss": 0.24653352797031403,
      "mean_token_accuracy": 0.901559591293335,
      "num_tokens": 2480609801.0,
      "step": 1068
    },
    {
      "epoch": 0.5470829068577278,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.7964397003397932e-05,
      "loss": 0.23930546641349792,
      "mean_token_accuracy": 0.9039927124977112,
      "num_tokens": 2482941206.0,
      "step": 1069
    },
    {
      "epoch": 0.5475946775844421,
      "grad_norm": 0.224609375,
      "learning_rate": 1.7931415350399323e-05,
      "loss": 0.24840064346790314,
      "mean_token_accuracy": 0.9011787176132202,
      "num_tokens": 2485205764.0,
      "step": 1070
    },
    {
      "epoch": 0.5481064483111566,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.789843938271549e-05,
      "loss": 0.234029620885849,
      "mean_token_accuracy": 0.9050673842430115,
      "num_tokens": 2487546668.0,
      "step": 1071
    },
    {
      "epoch": 0.548618219037871,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.786546919097784e-05,
      "loss": 0.24353347718715668,
      "mean_token_accuracy": 0.9023292064666748,
      "num_tokens": 2489808974.0,
      "step": 1072
    },
    {
      "epoch": 0.5491299897645855,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.7832504865801927e-05,
      "loss": 0.2490781545639038,
      "mean_token_accuracy": 0.9008018374443054,
      "num_tokens": 2492134993.0,
      "step": 1073
    },
    {
      "epoch": 0.5496417604912999,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.779954649778717e-05,
      "loss": 0.24056583642959595,
      "mean_token_accuracy": 0.9036996364593506,
      "num_tokens": 2494392003.0,
      "step": 1074
    },
    {
      "epoch": 0.5501535312180144,
      "grad_norm": 0.220703125,
      "learning_rate": 1.7766594177516595e-05,
      "loss": 0.22794106602668762,
      "mean_token_accuracy": 0.9093185663223267,
      "num_tokens": 2496700375.0,
      "step": 1075
    },
    {
      "epoch": 0.5506653019447287,
      "grad_norm": 0.203125,
      "learning_rate": 1.773364799555665e-05,
      "loss": 0.2433292418718338,
      "mean_token_accuracy": 0.9025551676750183,
      "num_tokens": 2499021507.0,
      "step": 1076
    },
    {
      "epoch": 0.5511770726714432,
      "grad_norm": 0.208984375,
      "learning_rate": 1.7700708042456866e-05,
      "loss": 0.2254454642534256,
      "mean_token_accuracy": 0.9093138575553894,
      "num_tokens": 2501444053.0,
      "step": 1077
    },
    {
      "epoch": 0.5516888433981576,
      "grad_norm": 0.2265625,
      "learning_rate": 1.7667774408749694e-05,
      "loss": 0.2487165927886963,
      "mean_token_accuracy": 0.9008100032806396,
      "num_tokens": 2503744695.0,
      "step": 1078
    },
    {
      "epoch": 0.5522006141248721,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.763484718495018e-05,
      "loss": 0.23697692155838013,
      "mean_token_accuracy": 0.9053808450698853,
      "num_tokens": 2506046040.0,
      "step": 1079
    },
    {
      "epoch": 0.5527123848515865,
      "grad_norm": 0.208984375,
      "learning_rate": 1.760192646155579e-05,
      "loss": 0.23883779346942902,
      "mean_token_accuracy": 0.9044649600982666,
      "num_tokens": 2508464038.0,
      "step": 1080
    },
    {
      "epoch": 0.553224155578301,
      "grad_norm": 0.21484375,
      "learning_rate": 1.7569012329046103e-05,
      "loss": 0.23809722065925598,
      "mean_token_accuracy": 0.9042915105819702,
      "num_tokens": 2510809904.0,
      "step": 1081
    },
    {
      "epoch": 0.5537359263050153,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.753610487788258e-05,
      "loss": 0.23244613409042358,
      "mean_token_accuracy": 0.9067838191986084,
      "num_tokens": 2513249186.0,
      "step": 1082
    },
    {
      "epoch": 0.5542476970317298,
      "grad_norm": 0.205078125,
      "learning_rate": 1.7503204198508334e-05,
      "loss": 0.2436913549900055,
      "mean_token_accuracy": 0.9019400477409363,
      "num_tokens": 2515605454.0,
      "step": 1083
    },
    {
      "epoch": 0.5547594677584442,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.7470310381347845e-05,
      "loss": 0.25070685148239136,
      "mean_token_accuracy": 0.8998407125473022,
      "num_tokens": 2517823338.0,
      "step": 1084
    },
    {
      "epoch": 0.5552712384851587,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.7437423516806764e-05,
      "loss": 0.2266395390033722,
      "mean_token_accuracy": 0.908598780632019,
      "num_tokens": 2520184893.0,
      "step": 1085
    },
    {
      "epoch": 0.5557830092118731,
      "grad_norm": 0.212890625,
      "learning_rate": 1.74045436952716e-05,
      "loss": 0.231015145778656,
      "mean_token_accuracy": 0.9076524972915649,
      "num_tokens": 2522549427.0,
      "step": 1086
    },
    {
      "epoch": 0.5562947799385876,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.737167100710952e-05,
      "loss": 0.2377699911594391,
      "mean_token_accuracy": 0.9040229916572571,
      "num_tokens": 2524885893.0,
      "step": 1087
    },
    {
      "epoch": 0.5568065506653019,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.73388055426681e-05,
      "loss": 0.24127113819122314,
      "mean_token_accuracy": 0.9029889106750488,
      "num_tokens": 2527101496.0,
      "step": 1088
    },
    {
      "epoch": 0.5573183213920164,
      "grad_norm": 0.21484375,
      "learning_rate": 1.730594739227504e-05,
      "loss": 0.23872314393520355,
      "mean_token_accuracy": 0.9045621752738953,
      "num_tokens": 2529433675.0,
      "step": 1089
    },
    {
      "epoch": 0.5578300921187308,
      "grad_norm": 0.21484375,
      "learning_rate": 1.7273096646237947e-05,
      "loss": 0.24281030893325806,
      "mean_token_accuracy": 0.9028335809707642,
      "num_tokens": 2531765817.0,
      "step": 1090
    },
    {
      "epoch": 0.5583418628454453,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.7240253394844073e-05,
      "loss": 0.23587429523468018,
      "mean_token_accuracy": 0.9057948589324951,
      "num_tokens": 2534066475.0,
      "step": 1091
    },
    {
      "epoch": 0.5588536335721597,
      "grad_norm": 0.220703125,
      "learning_rate": 1.7207417728360084e-05,
      "loss": 0.24670276045799255,
      "mean_token_accuracy": 0.9017560482025146,
      "num_tokens": 2536355718.0,
      "step": 1092
    },
    {
      "epoch": 0.5593654042988742,
      "grad_norm": 0.208984375,
      "learning_rate": 1.717458973703178e-05,
      "loss": 0.24236103892326355,
      "mean_token_accuracy": 0.9027793407440186,
      "num_tokens": 2538688450.0,
      "step": 1093
    },
    {
      "epoch": 0.5598771750255885,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.7141769511083882e-05,
      "loss": 0.2379949986934662,
      "mean_token_accuracy": 0.9042934775352478,
      "num_tokens": 2540998001.0,
      "step": 1094
    },
    {
      "epoch": 0.5603889457523029,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.7108957140719776e-05,
      "loss": 0.2490566372871399,
      "mean_token_accuracy": 0.9000282287597656,
      "num_tokens": 2543419719.0,
      "step": 1095
    },
    {
      "epoch": 0.5609007164790174,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.7076152716121227e-05,
      "loss": 0.2431260049343109,
      "mean_token_accuracy": 0.9025271534919739,
      "num_tokens": 2545693738.0,
      "step": 1096
    },
    {
      "epoch": 0.5614124872057318,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.70433563274482e-05,
      "loss": 0.2368476837873459,
      "mean_token_accuracy": 0.9049715995788574,
      "num_tokens": 2548092004.0,
      "step": 1097
    },
    {
      "epoch": 0.5619242579324463,
      "grad_norm": 0.21875,
      "learning_rate": 1.7010568064838538e-05,
      "loss": 0.23619848489761353,
      "mean_token_accuracy": 0.9053382873535156,
      "num_tokens": 2550424376.0,
      "step": 1098
    },
    {
      "epoch": 0.5624360286591606,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.6977788018407777e-05,
      "loss": 0.24532219767570496,
      "mean_token_accuracy": 0.9023861289024353,
      "num_tokens": 2552798961.0,
      "step": 1099
    },
    {
      "epoch": 0.5629477993858751,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.6945016278248854e-05,
      "loss": 0.23708729445934296,
      "mean_token_accuracy": 0.9040791392326355,
      "num_tokens": 2555128108.0,
      "step": 1100
    },
    {
      "epoch": 0.5634595701125895,
      "grad_norm": 0.21875,
      "learning_rate": 1.6912252934431897e-05,
      "loss": 0.24672022461891174,
      "mean_token_accuracy": 0.9012787938117981,
      "num_tokens": 2557362773.0,
      "step": 1101
    },
    {
      "epoch": 0.563971340839304,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.6879498077003926e-05,
      "loss": 0.23643174767494202,
      "mean_token_accuracy": 0.9047388434410095,
      "num_tokens": 2559716780.0,
      "step": 1102
    },
    {
      "epoch": 0.5644831115660184,
      "grad_norm": 0.21484375,
      "learning_rate": 1.684675179598867e-05,
      "loss": 0.24871838092803955,
      "mean_token_accuracy": 0.9005886912345886,
      "num_tokens": 2562045769.0,
      "step": 1103
    },
    {
      "epoch": 0.5649948822927329,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.6814014181386275e-05,
      "loss": 0.24231809377670288,
      "mean_token_accuracy": 0.9028476476669312,
      "num_tokens": 2564327215.0,
      "step": 1104
    },
    {
      "epoch": 0.5655066530194472,
      "grad_norm": 0.21875,
      "learning_rate": 1.6781285323173055e-05,
      "loss": 0.24393004179000854,
      "mean_token_accuracy": 0.9020379781723022,
      "num_tokens": 2566567076.0,
      "step": 1105
    },
    {
      "epoch": 0.5660184237461617,
      "grad_norm": 0.2001953125,
      "learning_rate": 1.6748565311301276e-05,
      "loss": 0.2332708239555359,
      "mean_token_accuracy": 0.9072904586791992,
      "num_tokens": 2568890902.0,
      "step": 1106
    },
    {
      "epoch": 0.5665301944728761,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.6715854235698882e-05,
      "loss": 0.25027868151664734,
      "mean_token_accuracy": 0.8992896676063538,
      "num_tokens": 2571237706.0,
      "step": 1107
    },
    {
      "epoch": 0.5670419651995906,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.6683152186269255e-05,
      "loss": 0.2385050356388092,
      "mean_token_accuracy": 0.9038647413253784,
      "num_tokens": 2573567369.0,
      "step": 1108
    },
    {
      "epoch": 0.567553735926305,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.665045925289097e-05,
      "loss": 0.2507104277610779,
      "mean_token_accuracy": 0.8996255397796631,
      "num_tokens": 2575876788.0,
      "step": 1109
    },
    {
      "epoch": 0.5680655066530195,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.6617775525417553e-05,
      "loss": 0.24322745203971863,
      "mean_token_accuracy": 0.902401328086853,
      "num_tokens": 2578189108.0,
      "step": 1110
    },
    {
      "epoch": 0.5685772773797338,
      "grad_norm": 0.216796875,
      "learning_rate": 1.658510109367723e-05,
      "loss": 0.24861006438732147,
      "mean_token_accuracy": 0.9011671543121338,
      "num_tokens": 2580483067.0,
      "step": 1111
    },
    {
      "epoch": 0.5690890481064483,
      "grad_norm": 0.208984375,
      "learning_rate": 1.6552436047472662e-05,
      "loss": 0.2465353012084961,
      "mean_token_accuracy": 0.9017524123191833,
      "num_tokens": 2582804999.0,
      "step": 1112
    },
    {
      "epoch": 0.5696008188331627,
      "grad_norm": 0.2109375,
      "learning_rate": 1.651978047658074e-05,
      "loss": 0.25906696915626526,
      "mean_token_accuracy": 0.8965818285942078,
      "num_tokens": 2585056954.0,
      "step": 1113
    },
    {
      "epoch": 0.5701125895598772,
      "grad_norm": 0.205078125,
      "learning_rate": 1.6487134470752286e-05,
      "loss": 0.2334526777267456,
      "mean_token_accuracy": 0.9062738418579102,
      "num_tokens": 2587373881.0,
      "step": 1114
    },
    {
      "epoch": 0.5706243602865916,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.6454498119711852e-05,
      "loss": 0.24545079469680786,
      "mean_token_accuracy": 0.9017007350921631,
      "num_tokens": 2589716308.0,
      "step": 1115
    },
    {
      "epoch": 0.5711361310133061,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.6421871513157445e-05,
      "loss": 0.24768389761447906,
      "mean_token_accuracy": 0.9004713296890259,
      "num_tokens": 2592110028.0,
      "step": 1116
    },
    {
      "epoch": 0.5716479017400204,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.6389254740760297e-05,
      "loss": 0.23157872259616852,
      "mean_token_accuracy": 0.9071974754333496,
      "num_tokens": 2594428431.0,
      "step": 1117
    },
    {
      "epoch": 0.5721596724667349,
      "grad_norm": 0.21875,
      "learning_rate": 1.6356647892164615e-05,
      "loss": 0.24348820745944977,
      "mean_token_accuracy": 0.9024673104286194,
      "num_tokens": 2596693932.0,
      "step": 1118
    },
    {
      "epoch": 0.5726714431934493,
      "grad_norm": 0.1953125,
      "learning_rate": 1.6324051056987317e-05,
      "loss": 0.2144147753715515,
      "mean_token_accuracy": 0.913762092590332,
      "num_tokens": 2598963310.0,
      "step": 1119
    },
    {
      "epoch": 0.5731832139201638,
      "grad_norm": 0.2265625,
      "learning_rate": 1.6291464324817812e-05,
      "loss": 0.24252286553382874,
      "mean_token_accuracy": 0.9030930399894714,
      "num_tokens": 2601223622.0,
      "step": 1120
    },
    {
      "epoch": 0.5736949846468782,
      "grad_norm": 0.212890625,
      "learning_rate": 1.625888778521774e-05,
      "loss": 0.23277021944522858,
      "mean_token_accuracy": 0.9064062237739563,
      "num_tokens": 2603679910.0,
      "step": 1121
    },
    {
      "epoch": 0.5742067553735927,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.622632152772072e-05,
      "loss": 0.23773489892482758,
      "mean_token_accuracy": 0.9047682285308838,
      "num_tokens": 2606125527.0,
      "step": 1122
    },
    {
      "epoch": 0.574718526100307,
      "grad_norm": 0.208984375,
      "learning_rate": 1.619376564183212e-05,
      "loss": 0.23127667605876923,
      "mean_token_accuracy": 0.9070393443107605,
      "num_tokens": 2608480743.0,
      "step": 1123
    },
    {
      "epoch": 0.5752302968270215,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.6161220217028803e-05,
      "loss": 0.24100711941719055,
      "mean_token_accuracy": 0.9038695096969604,
      "num_tokens": 2610785385.0,
      "step": 1124
    },
    {
      "epoch": 0.5757420675537359,
      "grad_norm": 0.205078125,
      "learning_rate": 1.612868534275887e-05,
      "loss": 0.22695697844028473,
      "mean_token_accuracy": 0.9095764756202698,
      "num_tokens": 2613110914.0,
      "step": 1125
    },
    {
      "epoch": 0.5762538382804504,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.6096161108441437e-05,
      "loss": 0.2536354660987854,
      "mean_token_accuracy": 0.8978942632675171,
      "num_tokens": 2615374030.0,
      "step": 1126
    },
    {
      "epoch": 0.5767656090071648,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.606364760346638e-05,
      "loss": 0.24987366795539856,
      "mean_token_accuracy": 0.8997305035591125,
      "num_tokens": 2617597198.0,
      "step": 1127
    },
    {
      "epoch": 0.5772773797338793,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.603114491719407e-05,
      "loss": 0.23766952753067017,
      "mean_token_accuracy": 0.9047482013702393,
      "num_tokens": 2619931261.0,
      "step": 1128
    },
    {
      "epoch": 0.5777891504605936,
      "grad_norm": 0.208984375,
      "learning_rate": 1.5998653138955153e-05,
      "loss": 0.24788798391819,
      "mean_token_accuracy": 0.9013940095901489,
      "num_tokens": 2622327324.0,
      "step": 1129
    },
    {
      "epoch": 0.5783009211873081,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.5966172358050296e-05,
      "loss": 0.24045248329639435,
      "mean_token_accuracy": 0.9041882753372192,
      "num_tokens": 2624676738.0,
      "step": 1130
    },
    {
      "epoch": 0.5788126919140225,
      "grad_norm": 0.208984375,
      "learning_rate": 1.5933702663749944e-05,
      "loss": 0.24429473280906677,
      "mean_token_accuracy": 0.9020708203315735,
      "num_tokens": 2627041501.0,
      "step": 1131
    },
    {
      "epoch": 0.579324462640737,
      "grad_norm": 0.212890625,
      "learning_rate": 1.5901244145294055e-05,
      "loss": 0.2449023723602295,
      "mean_token_accuracy": 0.901771068572998,
      "num_tokens": 2629362170.0,
      "step": 1132
    },
    {
      "epoch": 0.5798362333674514,
      "grad_norm": 0.205078125,
      "learning_rate": 1.58687968918919e-05,
      "loss": 0.24376901984214783,
      "mean_token_accuracy": 0.9027917385101318,
      "num_tokens": 2631683752.0,
      "step": 1133
    },
    {
      "epoch": 0.5803480040941658,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.5836360992721764e-05,
      "loss": 0.22909775376319885,
      "mean_token_accuracy": 0.9078489542007446,
      "num_tokens": 2634021086.0,
      "step": 1134
    },
    {
      "epoch": 0.5808597748208802,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.5803936536930734e-05,
      "loss": 0.2422669231891632,
      "mean_token_accuracy": 0.9035046100616455,
      "num_tokens": 2636261195.0,
      "step": 1135
    },
    {
      "epoch": 0.5813715455475946,
      "grad_norm": 0.220703125,
      "learning_rate": 1.577152361363445e-05,
      "loss": 0.25669169425964355,
      "mean_token_accuracy": 0.8974829912185669,
      "num_tokens": 2638573103.0,
      "step": 1136
    },
    {
      "epoch": 0.5818833162743091,
      "grad_norm": 0.21484375,
      "learning_rate": 1.5739122311916847e-05,
      "loss": 0.2474186271429062,
      "mean_token_accuracy": 0.9002084136009216,
      "num_tokens": 2640994881.0,
      "step": 1137
    },
    {
      "epoch": 0.5823950870010235,
      "grad_norm": 0.20703125,
      "learning_rate": 1.5706732720829935e-05,
      "loss": 0.23448091745376587,
      "mean_token_accuracy": 0.9059063196182251,
      "num_tokens": 2643227418.0,
      "step": 1138
    },
    {
      "epoch": 0.582906857727738,
      "grad_norm": 0.208984375,
      "learning_rate": 1.5674354929393513e-05,
      "loss": 0.2408888041973114,
      "mean_token_accuracy": 0.9031467437744141,
      "num_tokens": 2645588755.0,
      "step": 1139
    },
    {
      "epoch": 0.5834186284544524,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.5641989026594968e-05,
      "loss": 0.23802098631858826,
      "mean_token_accuracy": 0.9046595692634583,
      "num_tokens": 2647947623.0,
      "step": 1140
    },
    {
      "epoch": 0.5839303991811668,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.5609635101389023e-05,
      "loss": 0.24602627754211426,
      "mean_token_accuracy": 0.9012717604637146,
      "num_tokens": 2650385017.0,
      "step": 1141
    },
    {
      "epoch": 0.5844421699078812,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.5577293242697452e-05,
      "loss": 0.23383881151676178,
      "mean_token_accuracy": 0.9062581062316895,
      "num_tokens": 2652647299.0,
      "step": 1142
    },
    {
      "epoch": 0.5849539406345957,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.5544963539408894e-05,
      "loss": 0.23052066564559937,
      "mean_token_accuracy": 0.9069335460662842,
      "num_tokens": 2654881076.0,
      "step": 1143
    },
    {
      "epoch": 0.5854657113613101,
      "grad_norm": 0.212890625,
      "learning_rate": 1.5512646080378555e-05,
      "loss": 0.24414518475532532,
      "mean_token_accuracy": 0.9023891091346741,
      "num_tokens": 2657189392.0,
      "step": 1144
    },
    {
      "epoch": 0.5859774820880246,
      "grad_norm": 0.21484375,
      "learning_rate": 1.5480340954428008e-05,
      "loss": 0.24646379053592682,
      "mean_token_accuracy": 0.9010924100875854,
      "num_tokens": 2659559044.0,
      "step": 1145
    },
    {
      "epoch": 0.586489252814739,
      "grad_norm": 0.212890625,
      "learning_rate": 1.5448048250344917e-05,
      "loss": 0.23363427817821503,
      "mean_token_accuracy": 0.9060872793197632,
      "num_tokens": 2661860101.0,
      "step": 1146
    },
    {
      "epoch": 0.5870010235414534,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.5415768056882807e-05,
      "loss": 0.2405618280172348,
      "mean_token_accuracy": 0.9040665626525879,
      "num_tokens": 2664266425.0,
      "step": 1147
    },
    {
      "epoch": 0.5875127942681678,
      "grad_norm": 0.216796875,
      "learning_rate": 1.538350046276084e-05,
      "loss": 0.24424511194229126,
      "mean_token_accuracy": 0.9012519717216492,
      "num_tokens": 2666584453.0,
      "step": 1148
    },
    {
      "epoch": 0.5880245649948823,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.535124555666352e-05,
      "loss": 0.24418145418167114,
      "mean_token_accuracy": 0.9019414186477661,
      "num_tokens": 2668867139.0,
      "step": 1149
    },
    {
      "epoch": 0.5885363357215967,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.5319003427240493e-05,
      "loss": 0.24064208567142487,
      "mean_token_accuracy": 0.9032343029975891,
      "num_tokens": 2671137368.0,
      "step": 1150
    },
    {
      "epoch": 0.5890481064483112,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.5286774163106287e-05,
      "loss": 0.22513285279273987,
      "mean_token_accuracy": 0.9095910787582397,
      "num_tokens": 2673499631.0,
      "step": 1151
    },
    {
      "epoch": 0.5895598771750256,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.5254557852840077e-05,
      "loss": 0.22479213774204254,
      "mean_token_accuracy": 0.9088853597640991,
      "num_tokens": 2675797460.0,
      "step": 1152
    },
    {
      "epoch": 0.59007164790174,
      "grad_norm": 0.212890625,
      "learning_rate": 1.5222354584985425e-05,
      "loss": 0.23685865104198456,
      "mean_token_accuracy": 0.9039855599403381,
      "num_tokens": 2678114667.0,
      "step": 1153
    },
    {
      "epoch": 0.5905834186284544,
      "grad_norm": 0.19921875,
      "learning_rate": 1.5190164448050058e-05,
      "loss": 0.23238936066627502,
      "mean_token_accuracy": 0.9064977169036865,
      "num_tokens": 2680511896.0,
      "step": 1154
    },
    {
      "epoch": 0.5910951893551689,
      "grad_norm": 0.21875,
      "learning_rate": 1.5157987530505597e-05,
      "loss": 0.24987168610095978,
      "mean_token_accuracy": 0.8998004198074341,
      "num_tokens": 2682791764.0,
      "step": 1155
    },
    {
      "epoch": 0.5916069600818833,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.5125823920787357e-05,
      "loss": 0.23521879315376282,
      "mean_token_accuracy": 0.9055279493331909,
      "num_tokens": 2685067268.0,
      "step": 1156
    },
    {
      "epoch": 0.5921187308085978,
      "grad_norm": 0.208984375,
      "learning_rate": 1.5093673707294058e-05,
      "loss": 0.24032309651374817,
      "mean_token_accuracy": 0.9040378332138062,
      "num_tokens": 2687347698.0,
      "step": 1157
    },
    {
      "epoch": 0.5926305015353122,
      "grad_norm": 0.2109375,
      "learning_rate": 1.5061536978387605e-05,
      "loss": 0.23971155285835266,
      "mean_token_accuracy": 0.9034240245819092,
      "num_tokens": 2689699343.0,
      "step": 1158
    },
    {
      "epoch": 0.5931422722620266,
      "grad_norm": 0.212890625,
      "learning_rate": 1.5029413822392846e-05,
      "loss": 0.2384820133447647,
      "mean_token_accuracy": 0.9045255184173584,
      "num_tokens": 2692020662.0,
      "step": 1159
    },
    {
      "epoch": 0.593654042988741,
      "grad_norm": 0.21484375,
      "learning_rate": 1.4997304327597318e-05,
      "loss": 0.24896807968616486,
      "mean_token_accuracy": 0.9003876447677612,
      "num_tokens": 2694302047.0,
      "step": 1160
    },
    {
      "epoch": 0.5941658137154555,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.4965208582251023e-05,
      "loss": 0.23809699714183807,
      "mean_token_accuracy": 0.904249906539917,
      "num_tokens": 2696622885.0,
      "step": 1161
    },
    {
      "epoch": 0.5946775844421699,
      "grad_norm": 0.208984375,
      "learning_rate": 1.4933126674566154e-05,
      "loss": 0.23821359872817993,
      "mean_token_accuracy": 0.9050553441047668,
      "num_tokens": 2699031608.0,
      "step": 1162
    },
    {
      "epoch": 0.5951893551688844,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.4901058692716899e-05,
      "loss": 0.2375257909297943,
      "mean_token_accuracy": 0.9045150279998779,
      "num_tokens": 2701410750.0,
      "step": 1163
    },
    {
      "epoch": 0.5957011258955988,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.4869004724839158e-05,
      "loss": 0.2458113729953766,
      "mean_token_accuracy": 0.9017479419708252,
      "num_tokens": 2703761952.0,
      "step": 1164
    },
    {
      "epoch": 0.5962128966223132,
      "grad_norm": 0.228515625,
      "learning_rate": 1.4836964859030305e-05,
      "loss": 0.24472102522850037,
      "mean_token_accuracy": 0.9023573398590088,
      "num_tokens": 2706040111.0,
      "step": 1165
    },
    {
      "epoch": 0.5967246673490276,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.4804939183348982e-05,
      "loss": 0.24038773775100708,
      "mean_token_accuracy": 0.9036481380462646,
      "num_tokens": 2708332058.0,
      "step": 1166
    },
    {
      "epoch": 0.5972364380757421,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.4772927785814797e-05,
      "loss": 0.2387123852968216,
      "mean_token_accuracy": 0.9049074649810791,
      "num_tokens": 2710726988.0,
      "step": 1167
    },
    {
      "epoch": 0.5977482088024565,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.4740930754408145e-05,
      "loss": 0.23840895295143127,
      "mean_token_accuracy": 0.9042227268218994,
      "num_tokens": 2713146190.0,
      "step": 1168
    },
    {
      "epoch": 0.598259979529171,
      "grad_norm": 0.21875,
      "learning_rate": 1.4708948177069922e-05,
      "loss": 0.25360772013664246,
      "mean_token_accuracy": 0.8974609375,
      "num_tokens": 2715463336.0,
      "step": 1169
    },
    {
      "epoch": 0.5987717502558854,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.4676980141701298e-05,
      "loss": 0.24813157320022583,
      "mean_token_accuracy": 0.900977611541748,
      "num_tokens": 2717727939.0,
      "step": 1170
    },
    {
      "epoch": 0.5992835209825998,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.4645026736163494e-05,
      "loss": 0.23974081873893738,
      "mean_token_accuracy": 0.9033565521240234,
      "num_tokens": 2720011679.0,
      "step": 1171
    },
    {
      "epoch": 0.5997952917093142,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.4613088048277493e-05,
      "loss": 0.23849806189537048,
      "mean_token_accuracy": 0.9043401479721069,
      "num_tokens": 2722389745.0,
      "step": 1172
    },
    {
      "epoch": 0.6003070624360286,
      "grad_norm": 0.2255859375,
      "learning_rate": 1.4581164165823855e-05,
      "loss": 0.24164807796478271,
      "mean_token_accuracy": 0.9036725759506226,
      "num_tokens": 2724649183.0,
      "step": 1173
    },
    {
      "epoch": 0.6008188331627431,
      "grad_norm": 0.208984375,
      "learning_rate": 1.4549255176542425e-05,
      "loss": 0.2364886999130249,
      "mean_token_accuracy": 0.9048072099685669,
      "num_tokens": 2726995492.0,
      "step": 1174
    },
    {
      "epoch": 0.6013306038894575,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.4517361168132138e-05,
      "loss": 0.2359144687652588,
      "mean_token_accuracy": 0.905745267868042,
      "num_tokens": 2729341356.0,
      "step": 1175
    },
    {
      "epoch": 0.601842374616172,
      "grad_norm": 0.201171875,
      "learning_rate": 1.4485482228250735e-05,
      "loss": 0.22118720412254333,
      "mean_token_accuracy": 0.9113866686820984,
      "num_tokens": 2731743893.0,
      "step": 1176
    },
    {
      "epoch": 0.6023541453428863,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.4453618444514552e-05,
      "loss": 0.2339686155319214,
      "mean_token_accuracy": 0.9058175683021545,
      "num_tokens": 2734137320.0,
      "step": 1177
    },
    {
      "epoch": 0.6028659160696008,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.4421769904498285e-05,
      "loss": 0.25213971734046936,
      "mean_token_accuracy": 0.8986042737960815,
      "num_tokens": 2736441211.0,
      "step": 1178
    },
    {
      "epoch": 0.6033776867963152,
      "grad_norm": 0.224609375,
      "learning_rate": 1.43899366957347e-05,
      "loss": 0.24562975764274597,
      "mean_token_accuracy": 0.9010947942733765,
      "num_tokens": 2738766741.0,
      "step": 1179
    },
    {
      "epoch": 0.6038894575230297,
      "grad_norm": 0.232421875,
      "learning_rate": 1.435811890571446e-05,
      "loss": 0.24440884590148926,
      "mean_token_accuracy": 0.9021767973899841,
      "num_tokens": 2741084566.0,
      "step": 1180
    },
    {
      "epoch": 0.6044012282497441,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.4326316621885826e-05,
      "loss": 0.25855952501296997,
      "mean_token_accuracy": 0.8970465660095215,
      "num_tokens": 2743449917.0,
      "step": 1181
    },
    {
      "epoch": 0.6049129989764586,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.4294529931654459e-05,
      "loss": 0.241653174161911,
      "mean_token_accuracy": 0.903138279914856,
      "num_tokens": 2745787174.0,
      "step": 1182
    },
    {
      "epoch": 0.6054247697031729,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.4262758922383147e-05,
      "loss": 0.22813111543655396,
      "mean_token_accuracy": 0.9080767631530762,
      "num_tokens": 2748120513.0,
      "step": 1183
    },
    {
      "epoch": 0.6059365404298874,
      "grad_norm": 0.2109375,
      "learning_rate": 1.42310036813916e-05,
      "loss": 0.24120007455348969,
      "mean_token_accuracy": 0.9038897752761841,
      "num_tokens": 2750437392.0,
      "step": 1184
    },
    {
      "epoch": 0.6064483111566018,
      "grad_norm": 0.21484375,
      "learning_rate": 1.419926429595616e-05,
      "loss": 0.24330668151378632,
      "mean_token_accuracy": 0.9011092782020569,
      "num_tokens": 2752713952.0,
      "step": 1185
    },
    {
      "epoch": 0.6069600818833163,
      "grad_norm": 0.216796875,
      "learning_rate": 1.4167540853309633e-05,
      "loss": 0.2514837086200714,
      "mean_token_accuracy": 0.8997951745986938,
      "num_tokens": 2754969099.0,
      "step": 1186
    },
    {
      "epoch": 0.6074718526100307,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.4135833440640977e-05,
      "loss": 0.24876756966114044,
      "mean_token_accuracy": 0.9006786346435547,
      "num_tokens": 2757312972.0,
      "step": 1187
    },
    {
      "epoch": 0.6079836233367452,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.4104142145095096e-05,
      "loss": 0.24057689309120178,
      "mean_token_accuracy": 0.903491199016571,
      "num_tokens": 2759577990.0,
      "step": 1188
    },
    {
      "epoch": 0.6084953940634595,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.4072467053772612e-05,
      "loss": 0.24999216198921204,
      "mean_token_accuracy": 0.8998757004737854,
      "num_tokens": 2761909843.0,
      "step": 1189
    },
    {
      "epoch": 0.609007164790174,
      "grad_norm": 0.2109375,
      "learning_rate": 1.404080825372959e-05,
      "loss": 0.25338688492774963,
      "mean_token_accuracy": 0.898749053478241,
      "num_tokens": 2764224520.0,
      "step": 1190
    },
    {
      "epoch": 0.6095189355168884,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.4009165831977347e-05,
      "loss": 0.25480204820632935,
      "mean_token_accuracy": 0.8986008167266846,
      "num_tokens": 2766508599.0,
      "step": 1191
    },
    {
      "epoch": 0.6100307062436029,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.3977539875482151e-05,
      "loss": 0.22839517891407013,
      "mean_token_accuracy": 0.9077802896499634,
      "num_tokens": 2768840222.0,
      "step": 1192
    },
    {
      "epoch": 0.6105424769703173,
      "grad_norm": 0.2255859375,
      "learning_rate": 1.3945930471165055e-05,
      "loss": 0.25379258394241333,
      "mean_token_accuracy": 0.898489236831665,
      "num_tokens": 2771128823.0,
      "step": 1193
    },
    {
      "epoch": 0.6110542476970318,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.3914337705901594e-05,
      "loss": 0.2331770956516266,
      "mean_token_accuracy": 0.906539797782898,
      "num_tokens": 2773413781.0,
      "step": 1194
    },
    {
      "epoch": 0.6115660184237461,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.388276166652158e-05,
      "loss": 0.24477235972881317,
      "mean_token_accuracy": 0.9019397497177124,
      "num_tokens": 2775764047.0,
      "step": 1195
    },
    {
      "epoch": 0.6120777891504606,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.3851202439808857e-05,
      "loss": 0.2324167937040329,
      "mean_token_accuracy": 0.9066228270530701,
      "num_tokens": 2778050991.0,
      "step": 1196
    },
    {
      "epoch": 0.612589559877175,
      "grad_norm": 0.2109375,
      "learning_rate": 1.3819660112501054e-05,
      "loss": 0.24216875433921814,
      "mean_token_accuracy": 0.9032064080238342,
      "num_tokens": 2780364179.0,
      "step": 1197
    },
    {
      "epoch": 0.6131013306038895,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.3788134771289366e-05,
      "loss": 0.2470916509628296,
      "mean_token_accuracy": 0.9012836217880249,
      "num_tokens": 2782715916.0,
      "step": 1198
    },
    {
      "epoch": 0.6136131013306039,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.3756626502818282e-05,
      "loss": 0.2422672063112259,
      "mean_token_accuracy": 0.903227686882019,
      "num_tokens": 2785018994.0,
      "step": 1199
    },
    {
      "epoch": 0.6141248720573184,
      "grad_norm": 0.203125,
      "learning_rate": 1.3725135393685389e-05,
      "loss": 0.2343059480190277,
      "mean_token_accuracy": 0.9052448272705078,
      "num_tokens": 2787403519.0,
      "step": 1200
    },
    {
      "epoch": 0.6146366427840327,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.3693661530441116e-05,
      "loss": 0.23233872652053833,
      "mean_token_accuracy": 0.9070663452148438,
      "num_tokens": 2789802930.0,
      "step": 1201
    },
    {
      "epoch": 0.6151484135107472,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.366220499958847e-05,
      "loss": 0.2442513108253479,
      "mean_token_accuracy": 0.9028111696243286,
      "num_tokens": 2792108489.0,
      "step": 1202
    },
    {
      "epoch": 0.6156601842374616,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.3630765887582845e-05,
      "loss": 0.2537516951560974,
      "mean_token_accuracy": 0.8983054161071777,
      "num_tokens": 2794440313.0,
      "step": 1203
    },
    {
      "epoch": 0.6161719549641761,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.3599344280831738e-05,
      "loss": 0.25090837478637695,
      "mean_token_accuracy": 0.8993459343910217,
      "num_tokens": 2796805102.0,
      "step": 1204
    },
    {
      "epoch": 0.6166837256908905,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.356794026569456e-05,
      "loss": 0.2422456294298172,
      "mean_token_accuracy": 0.9023822546005249,
      "num_tokens": 2799168363.0,
      "step": 1205
    },
    {
      "epoch": 0.617195496417605,
      "grad_norm": 0.201171875,
      "learning_rate": 1.3536553928482348e-05,
      "loss": 0.22978681325912476,
      "mean_token_accuracy": 0.9070588946342468,
      "num_tokens": 2801452520.0,
      "step": 1206
    },
    {
      "epoch": 0.6177072671443193,
      "grad_norm": 0.2294921875,
      "learning_rate": 1.3505185355457569e-05,
      "loss": 0.2617835998535156,
      "mean_token_accuracy": 0.8955304622650146,
      "num_tokens": 2803784133.0,
      "step": 1207
    },
    {
      "epoch": 0.6182190378710338,
      "grad_norm": 0.208984375,
      "learning_rate": 1.347383463283387e-05,
      "loss": 0.23017628490924835,
      "mean_token_accuracy": 0.9072146415710449,
      "num_tokens": 2806093148.0,
      "step": 1208
    },
    {
      "epoch": 0.6187308085977482,
      "grad_norm": 0.21484375,
      "learning_rate": 1.3442501846775823e-05,
      "loss": 0.23716974258422852,
      "mean_token_accuracy": 0.9038214087486267,
      "num_tokens": 2808373671.0,
      "step": 1209
    },
    {
      "epoch": 0.6192425793244627,
      "grad_norm": 0.197265625,
      "learning_rate": 1.3411187083398716e-05,
      "loss": 0.228618323802948,
      "mean_token_accuracy": 0.9083126783370972,
      "num_tokens": 2810690932.0,
      "step": 1210
    },
    {
      "epoch": 0.6197543500511771,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.337989042876829e-05,
      "loss": 0.22553783655166626,
      "mean_token_accuracy": 0.9096112251281738,
      "num_tokens": 2812994841.0,
      "step": 1211
    },
    {
      "epoch": 0.6202661207778914,
      "grad_norm": 0.203125,
      "learning_rate": 1.3348611968900534e-05,
      "loss": 0.2385912835597992,
      "mean_token_accuracy": 0.9038527607917786,
      "num_tokens": 2815399055.0,
      "step": 1212
    },
    {
      "epoch": 0.6207778915046059,
      "grad_norm": 0.201171875,
      "learning_rate": 1.3317351789761406e-05,
      "loss": 0.23558856546878815,
      "mean_token_accuracy": 0.9052136540412903,
      "num_tokens": 2817723080.0,
      "step": 1213
    },
    {
      "epoch": 0.6212896622313203,
      "grad_norm": 0.21484375,
      "learning_rate": 1.3286109977266647e-05,
      "loss": 0.2367328405380249,
      "mean_token_accuracy": 0.9043396711349487,
      "num_tokens": 2820075972.0,
      "step": 1214
    },
    {
      "epoch": 0.6218014329580348,
      "grad_norm": 0.21484375,
      "learning_rate": 1.3254886617281499e-05,
      "loss": 0.23439635336399078,
      "mean_token_accuracy": 0.9062700867652893,
      "num_tokens": 2822363335.0,
      "step": 1215
    },
    {
      "epoch": 0.6223132036847492,
      "grad_norm": 0.216796875,
      "learning_rate": 1.3223681795620502e-05,
      "loss": 0.23629826307296753,
      "mean_token_accuracy": 0.9048300981521606,
      "num_tokens": 2824687030.0,
      "step": 1216
    },
    {
      "epoch": 0.6228249744114637,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.3192495598047246e-05,
      "loss": 0.23537464439868927,
      "mean_token_accuracy": 0.9058517217636108,
      "num_tokens": 2827035812.0,
      "step": 1217
    },
    {
      "epoch": 0.623336745138178,
      "grad_norm": 0.201171875,
      "learning_rate": 1.3161328110274117e-05,
      "loss": 0.23202106356620789,
      "mean_token_accuracy": 0.9066104888916016,
      "num_tokens": 2829344922.0,
      "step": 1218
    },
    {
      "epoch": 0.6238485158648925,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.3130179417962095e-05,
      "loss": 0.23711642622947693,
      "mean_token_accuracy": 0.90513014793396,
      "num_tokens": 2831701284.0,
      "step": 1219
    },
    {
      "epoch": 0.6243602865916069,
      "grad_norm": 0.203125,
      "learning_rate": 1.3099049606720498e-05,
      "loss": 0.24805745482444763,
      "mean_token_accuracy": 0.9008549451828003,
      "num_tokens": 2834076413.0,
      "step": 1220
    },
    {
      "epoch": 0.6248720573183214,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.3067938762106754e-05,
      "loss": 0.2427055835723877,
      "mean_token_accuracy": 0.9024381041526794,
      "num_tokens": 2836373425.0,
      "step": 1221
    },
    {
      "epoch": 0.6253838280450358,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.3036846969626146e-05,
      "loss": 0.23589447140693665,
      "mean_token_accuracy": 0.9046304225921631,
      "num_tokens": 2838672465.0,
      "step": 1222
    },
    {
      "epoch": 0.6258955987717503,
      "grad_norm": 0.208984375,
      "learning_rate": 1.3005774314731625e-05,
      "loss": 0.2439790964126587,
      "mean_token_accuracy": 0.9023952484130859,
      "num_tokens": 2840987298.0,
      "step": 1223
    },
    {
      "epoch": 0.6264073694984647,
      "grad_norm": 0.20703125,
      "learning_rate": 1.297472088282352e-05,
      "loss": 0.24890512228012085,
      "mean_token_accuracy": 0.9000982642173767,
      "num_tokens": 2843342472.0,
      "step": 1224
    },
    {
      "epoch": 0.6269191402251791,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.2943686759249337e-05,
      "loss": 0.2541946768760681,
      "mean_token_accuracy": 0.8980565667152405,
      "num_tokens": 2845690215.0,
      "step": 1225
    },
    {
      "epoch": 0.6274309109518935,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.291267202930351e-05,
      "loss": 0.24058011174201965,
      "mean_token_accuracy": 0.9041627645492554,
      "num_tokens": 2848006493.0,
      "step": 1226
    },
    {
      "epoch": 0.627942681678608,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.288167677822717e-05,
      "loss": 0.23980367183685303,
      "mean_token_accuracy": 0.9041623473167419,
      "num_tokens": 2850328257.0,
      "step": 1227
    },
    {
      "epoch": 0.6284544524053224,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.2850701091207921e-05,
      "loss": 0.22569718956947327,
      "mean_token_accuracy": 0.9084949493408203,
      "num_tokens": 2852668176.0,
      "step": 1228
    },
    {
      "epoch": 0.6289662231320369,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.2819745053379586e-05,
      "loss": 0.2525040805339813,
      "mean_token_accuracy": 0.8994076251983643,
      "num_tokens": 2855126767.0,
      "step": 1229
    },
    {
      "epoch": 0.6294779938587513,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.2788808749821992e-05,
      "loss": 0.2360077053308487,
      "mean_token_accuracy": 0.9056181907653809,
      "num_tokens": 2857514250.0,
      "step": 1230
    },
    {
      "epoch": 0.6299897645854657,
      "grad_norm": 0.2109375,
      "learning_rate": 1.2757892265560731e-05,
      "loss": 0.24257981777191162,
      "mean_token_accuracy": 0.9034539461135864,
      "num_tokens": 2859812030.0,
      "step": 1231
    },
    {
      "epoch": 0.6305015353121801,
      "grad_norm": 0.2109375,
      "learning_rate": 1.2726995685566912e-05,
      "loss": 0.23912790417671204,
      "mean_token_accuracy": 0.9039080739021301,
      "num_tokens": 2862200306.0,
      "step": 1232
    },
    {
      "epoch": 0.6310133060388946,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.2696119094756951e-05,
      "loss": 0.25526416301727295,
      "mean_token_accuracy": 0.8985037803649902,
      "num_tokens": 2864541729.0,
      "step": 1233
    },
    {
      "epoch": 0.631525076765609,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.2665262577992312e-05,
      "loss": 0.2279675006866455,
      "mean_token_accuracy": 0.9083495140075684,
      "num_tokens": 2866881656.0,
      "step": 1234
    },
    {
      "epoch": 0.6320368474923235,
      "grad_norm": 0.201171875,
      "learning_rate": 1.2634426220079304e-05,
      "loss": 0.2259557545185089,
      "mean_token_accuracy": 0.9096567630767822,
      "num_tokens": 2869250740.0,
      "step": 1235
    },
    {
      "epoch": 0.6325486182190379,
      "grad_norm": 0.2109375,
      "learning_rate": 1.2603610105768812e-05,
      "loss": 0.2503730058670044,
      "mean_token_accuracy": 0.9001143574714661,
      "num_tokens": 2871609643.0,
      "step": 1236
    },
    {
      "epoch": 0.6330603889457523,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.2572814319756096e-05,
      "loss": 0.25225234031677246,
      "mean_token_accuracy": 0.8982993364334106,
      "num_tokens": 2873966799.0,
      "step": 1237
    },
    {
      "epoch": 0.6335721596724667,
      "grad_norm": 0.197265625,
      "learning_rate": 1.2542038946680555e-05,
      "loss": 0.23388472199440002,
      "mean_token_accuracy": 0.9063358306884766,
      "num_tokens": 2876239109.0,
      "step": 1238
    },
    {
      "epoch": 0.6340839303991812,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.251128407112546e-05,
      "loss": 0.23375898599624634,
      "mean_token_accuracy": 0.9061154127120972,
      "num_tokens": 2878575326.0,
      "step": 1239
    },
    {
      "epoch": 0.6345957011258956,
      "grad_norm": 0.205078125,
      "learning_rate": 1.2480549777617768e-05,
      "loss": 0.2360427975654602,
      "mean_token_accuracy": 0.9045852422714233,
      "num_tokens": 2880929478.0,
      "step": 1240
    },
    {
      "epoch": 0.6351074718526101,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.2449836150627851e-05,
      "loss": 0.22810447216033936,
      "mean_token_accuracy": 0.9085752964019775,
      "num_tokens": 2883274924.0,
      "step": 1241
    },
    {
      "epoch": 0.6356192425793245,
      "grad_norm": 0.216796875,
      "learning_rate": 1.2419143274569303e-05,
      "loss": 0.23543781042099,
      "mean_token_accuracy": 0.9048992991447449,
      "num_tokens": 2885550005.0,
      "step": 1242
    },
    {
      "epoch": 0.6361310133060389,
      "grad_norm": 0.205078125,
      "learning_rate": 1.2388471233798658e-05,
      "loss": 0.2378784865140915,
      "mean_token_accuracy": 0.9044396281242371,
      "num_tokens": 2887819900.0,
      "step": 1243
    },
    {
      "epoch": 0.6366427840327533,
      "grad_norm": 0.205078125,
      "learning_rate": 1.235782011261521e-05,
      "loss": 0.2336754947900772,
      "mean_token_accuracy": 0.9071619510650635,
      "num_tokens": 2890123207.0,
      "step": 1244
    },
    {
      "epoch": 0.6371545547594678,
      "grad_norm": 0.1953125,
      "learning_rate": 1.232718999526074e-05,
      "loss": 0.22792543470859528,
      "mean_token_accuracy": 0.9085842370986938,
      "num_tokens": 2892491980.0,
      "step": 1245
    },
    {
      "epoch": 0.6376663254861822,
      "grad_norm": 0.20703125,
      "learning_rate": 1.2296580965919317e-05,
      "loss": 0.2432270646095276,
      "mean_token_accuracy": 0.9026405811309814,
      "num_tokens": 2894740837.0,
      "step": 1246
    },
    {
      "epoch": 0.6381780962128967,
      "grad_norm": 0.203125,
      "learning_rate": 1.2265993108717051e-05,
      "loss": 0.23249228298664093,
      "mean_token_accuracy": 0.9072756171226501,
      "num_tokens": 2897165576.0,
      "step": 1247
    },
    {
      "epoch": 0.638689866939611,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.223542650772184e-05,
      "loss": 0.241900235414505,
      "mean_token_accuracy": 0.9029766321182251,
      "num_tokens": 2899535551.0,
      "step": 1248
    },
    {
      "epoch": 0.6392016376663255,
      "grad_norm": 0.21484375,
      "learning_rate": 1.2204881246943191e-05,
      "loss": 0.23738700151443481,
      "mean_token_accuracy": 0.9039027690887451,
      "num_tokens": 2901882620.0,
      "step": 1249
    },
    {
      "epoch": 0.6397134083930399,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.2174357410331939e-05,
      "loss": 0.22756317257881165,
      "mean_token_accuracy": 0.9084951877593994,
      "num_tokens": 2904201202.0,
      "step": 1250
    },
    {
      "epoch": 0.6402251791197543,
      "grad_norm": 0.2109375,
      "learning_rate": 1.2143855081780048e-05,
      "loss": 0.24107195436954498,
      "mean_token_accuracy": 0.9030890464782715,
      "num_tokens": 2906531196.0,
      "step": 1251
    },
    {
      "epoch": 0.6407369498464688,
      "grad_norm": 0.22265625,
      "learning_rate": 1.211337434512035e-05,
      "loss": 0.2502373158931732,
      "mean_token_accuracy": 0.8997039794921875,
      "num_tokens": 2908826188.0,
      "step": 1252
    },
    {
      "epoch": 0.6412487205731832,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.2082915284126364e-05,
      "loss": 0.23074743151664734,
      "mean_token_accuracy": 0.9073284864425659,
      "num_tokens": 2911222579.0,
      "step": 1253
    },
    {
      "epoch": 0.6417604912998977,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.205247798251202e-05,
      "loss": 0.24645006656646729,
      "mean_token_accuracy": 0.901300847530365,
      "num_tokens": 2913509361.0,
      "step": 1254
    },
    {
      "epoch": 0.642272262026612,
      "grad_norm": 0.208984375,
      "learning_rate": 1.2022062523931432e-05,
      "loss": 0.24301284551620483,
      "mean_token_accuracy": 0.9005380272865295,
      "num_tokens": 2915895429.0,
      "step": 1255
    },
    {
      "epoch": 0.6427840327533265,
      "grad_norm": 0.2109375,
      "learning_rate": 1.1991668991978704e-05,
      "loss": 0.2378939390182495,
      "mean_token_accuracy": 0.9037714004516602,
      "num_tokens": 2918187107.0,
      "step": 1256
    },
    {
      "epoch": 0.6432958034800409,
      "grad_norm": 0.220703125,
      "learning_rate": 1.1961297470187656e-05,
      "loss": 0.2532947063446045,
      "mean_token_accuracy": 0.897946834564209,
      "num_tokens": 2920540483.0,
      "step": 1257
    },
    {
      "epoch": 0.6438075742067554,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.193094804203163e-05,
      "loss": 0.24371767044067383,
      "mean_token_accuracy": 0.9023501873016357,
      "num_tokens": 2922860429.0,
      "step": 1258
    },
    {
      "epoch": 0.6443193449334698,
      "grad_norm": 0.20703125,
      "learning_rate": 1.1900620790923236e-05,
      "loss": 0.2213759422302246,
      "mean_token_accuracy": 0.910429060459137,
      "num_tokens": 2925161182.0,
      "step": 1259
    },
    {
      "epoch": 0.6448311156601843,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.1870315800214131e-05,
      "loss": 0.23911526799201965,
      "mean_token_accuracy": 0.904130756855011,
      "num_tokens": 2927493083.0,
      "step": 1260
    },
    {
      "epoch": 0.6453428863868986,
      "grad_norm": 0.2109375,
      "learning_rate": 1.1840033153194811e-05,
      "loss": 0.25219935178756714,
      "mean_token_accuracy": 0.8985886573791504,
      "num_tokens": 2929808947.0,
      "step": 1261
    },
    {
      "epoch": 0.6458546571136131,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.1809772933094337e-05,
      "loss": 0.2457803636789322,
      "mean_token_accuracy": 0.9019883871078491,
      "num_tokens": 2932125791.0,
      "step": 1262
    },
    {
      "epoch": 0.6463664278403275,
      "grad_norm": 0.1962890625,
      "learning_rate": 1.1779535223080147e-05,
      "loss": 0.23383237421512604,
      "mean_token_accuracy": 0.9064129590988159,
      "num_tokens": 2934565107.0,
      "step": 1263
    },
    {
      "epoch": 0.646878198567042,
      "grad_norm": 0.2109375,
      "learning_rate": 1.1749320106257804e-05,
      "loss": 0.24028030037879944,
      "mean_token_accuracy": 0.9039646983146667,
      "num_tokens": 2937015975.0,
      "step": 1264
    },
    {
      "epoch": 0.6473899692937564,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.1719127665670786e-05,
      "loss": 0.2301207184791565,
      "mean_token_accuracy": 0.9067972898483276,
      "num_tokens": 2939281929.0,
      "step": 1265
    },
    {
      "epoch": 0.6479017400204709,
      "grad_norm": 0.2265625,
      "learning_rate": 1.168895798430023e-05,
      "loss": 0.257017582654953,
      "mean_token_accuracy": 0.8976175785064697,
      "num_tokens": 2941597808.0,
      "step": 1266
    },
    {
      "epoch": 0.6484135107471852,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.1658811145064734e-05,
      "loss": 0.24629956483840942,
      "mean_token_accuracy": 0.9011189341545105,
      "num_tokens": 2944004561.0,
      "step": 1267
    },
    {
      "epoch": 0.6489252814738997,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.1628687230820124e-05,
      "loss": 0.23866420984268188,
      "mean_token_accuracy": 0.9042989015579224,
      "num_tokens": 2946290425.0,
      "step": 1268
    },
    {
      "epoch": 0.6494370522006141,
      "grad_norm": 0.2265625,
      "learning_rate": 1.1598586324359212e-05,
      "loss": 0.23540395498275757,
      "mean_token_accuracy": 0.9051147699356079,
      "num_tokens": 2948608398.0,
      "step": 1269
    },
    {
      "epoch": 0.6499488229273286,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.1568508508411558e-05,
      "loss": 0.23464567959308624,
      "mean_token_accuracy": 0.905079185962677,
      "num_tokens": 2950922664.0,
      "step": 1270
    },
    {
      "epoch": 0.650460593654043,
      "grad_norm": 0.2109375,
      "learning_rate": 1.1538453865643277e-05,
      "loss": 0.2353413701057434,
      "mean_token_accuracy": 0.9055534601211548,
      "num_tokens": 2953203507.0,
      "step": 1271
    },
    {
      "epoch": 0.6509723643807575,
      "grad_norm": 0.21875,
      "learning_rate": 1.15084224786568e-05,
      "loss": 0.23862890899181366,
      "mean_token_accuracy": 0.9043632745742798,
      "num_tokens": 2955543258.0,
      "step": 1272
    },
    {
      "epoch": 0.6514841351074718,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.1478414429990638e-05,
      "loss": 0.24988803267478943,
      "mean_token_accuracy": 0.9002877473831177,
      "num_tokens": 2957818409.0,
      "step": 1273
    },
    {
      "epoch": 0.6519959058341863,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.144842980211914e-05,
      "loss": 0.23850972950458527,
      "mean_token_accuracy": 0.9043269157409668,
      "num_tokens": 2960155047.0,
      "step": 1274
    },
    {
      "epoch": 0.6525076765609007,
      "grad_norm": 0.21484375,
      "learning_rate": 1.14184686774523e-05,
      "loss": 0.24808132648468018,
      "mean_token_accuracy": 0.8999494314193726,
      "num_tokens": 2962421857.0,
      "step": 1275
    },
    {
      "epoch": 0.6530194472876152,
      "grad_norm": 0.19921875,
      "learning_rate": 1.1388531138335535e-05,
      "loss": 0.230534628033638,
      "mean_token_accuracy": 0.9061236381530762,
      "num_tokens": 2964821373.0,
      "step": 1276
    },
    {
      "epoch": 0.6535312180143296,
      "grad_norm": 0.205078125,
      "learning_rate": 1.1358617267049396e-05,
      "loss": 0.23905575275421143,
      "mean_token_accuracy": 0.9038777351379395,
      "num_tokens": 2967153089.0,
      "step": 1277
    },
    {
      "epoch": 0.654042988741044,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.1328727145809423e-05,
      "loss": 0.2245817631483078,
      "mean_token_accuracy": 0.9096016883850098,
      "num_tokens": 2969503809.0,
      "step": 1278
    },
    {
      "epoch": 0.6545547594677584,
      "grad_norm": 0.2109375,
      "learning_rate": 1.1298860856765862e-05,
      "loss": 0.2348024845123291,
      "mean_token_accuracy": 0.905834972858429,
      "num_tokens": 2971865232.0,
      "step": 1279
    },
    {
      "epoch": 0.6550665301944729,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.1269018482003473e-05,
      "loss": 0.24083885550498962,
      "mean_token_accuracy": 0.9037882089614868,
      "num_tokens": 2974241846.0,
      "step": 1280
    },
    {
      "epoch": 0.6555783009211873,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.123920010354126e-05,
      "loss": 0.2570626437664032,
      "mean_token_accuracy": 0.897019624710083,
      "num_tokens": 2976509062.0,
      "step": 1281
    },
    {
      "epoch": 0.6560900716479018,
      "grad_norm": 0.2294921875,
      "learning_rate": 1.1209405803332303e-05,
      "loss": 0.24472883343696594,
      "mean_token_accuracy": 0.9013326168060303,
      "num_tokens": 2978803010.0,
      "step": 1282
    },
    {
      "epoch": 0.6566018423746162,
      "grad_norm": 0.2236328125,
      "learning_rate": 1.1179635663263514e-05,
      "loss": 0.23859640955924988,
      "mean_token_accuracy": 0.9042452573776245,
      "num_tokens": 2981099627.0,
      "step": 1283
    },
    {
      "epoch": 0.6571136131013307,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.1149889765155368e-05,
      "loss": 0.25900721549987793,
      "mean_token_accuracy": 0.8959206342697144,
      "num_tokens": 2983368542.0,
      "step": 1284
    },
    {
      "epoch": 0.657625383828045,
      "grad_norm": 0.20703125,
      "learning_rate": 1.1120168190761735e-05,
      "loss": 0.23756825923919678,
      "mean_token_accuracy": 0.9050242900848389,
      "num_tokens": 2985610485.0,
      "step": 1285
    },
    {
      "epoch": 0.6581371545547595,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.1090471021769635e-05,
      "loss": 0.24206197261810303,
      "mean_token_accuracy": 0.902626097202301,
      "num_tokens": 2987924988.0,
      "step": 1286
    },
    {
      "epoch": 0.6586489252814739,
      "grad_norm": 0.21484375,
      "learning_rate": 1.1060798339799012e-05,
      "loss": 0.24297462403774261,
      "mean_token_accuracy": 0.9024270176887512,
      "num_tokens": 2990285495.0,
      "step": 1287
    },
    {
      "epoch": 0.6591606960081884,
      "grad_norm": 0.21875,
      "learning_rate": 1.1031150226402491e-05,
      "loss": 0.2403492033481598,
      "mean_token_accuracy": 0.9038851261138916,
      "num_tokens": 2992606982.0,
      "step": 1288
    },
    {
      "epoch": 0.6596724667349028,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.1001526763065193e-05,
      "loss": 0.25285273790359497,
      "mean_token_accuracy": 0.8988425731658936,
      "num_tokens": 2994869057.0,
      "step": 1289
    },
    {
      "epoch": 0.6601842374616171,
      "grad_norm": 0.1962890625,
      "learning_rate": 1.0971928031204486e-05,
      "loss": 0.22897624969482422,
      "mean_token_accuracy": 0.9076515436172485,
      "num_tokens": 2997281469.0,
      "step": 1290
    },
    {
      "epoch": 0.6606960081883316,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.0942354112169764e-05,
      "loss": 0.22851766645908356,
      "mean_token_accuracy": 0.9072650671005249,
      "num_tokens": 2999646862.0,
      "step": 1291
    },
    {
      "epoch": 0.661207778915046,
      "grad_norm": 0.2001953125,
      "learning_rate": 1.0912805087242222e-05,
      "loss": 0.2259330302476883,
      "mean_token_accuracy": 0.9098363518714905,
      "num_tokens": 3002033933.0,
      "step": 1292
    },
    {
      "epoch": 0.6617195496417605,
      "grad_norm": 0.2109375,
      "learning_rate": 1.0883281037634644e-05,
      "loss": 0.2377098798751831,
      "mean_token_accuracy": 0.9046051502227783,
      "num_tokens": 3004340708.0,
      "step": 1293
    },
    {
      "epoch": 0.6622313203684749,
      "grad_norm": 0.2109375,
      "learning_rate": 1.0853782044491168e-05,
      "loss": 0.23413708806037903,
      "mean_token_accuracy": 0.906644880771637,
      "num_tokens": 3006702597.0,
      "step": 1294
    },
    {
      "epoch": 0.6627430910951894,
      "grad_norm": 0.212890625,
      "learning_rate": 1.0824308188887054e-05,
      "loss": 0.24485185742378235,
      "mean_token_accuracy": 0.9009273052215576,
      "num_tokens": 3008949334.0,
      "step": 1295
    },
    {
      "epoch": 0.6632548618219037,
      "grad_norm": 0.19140625,
      "learning_rate": 1.0794859551828494e-05,
      "loss": 0.22269585728645325,
      "mean_token_accuracy": 0.9101850986480713,
      "num_tokens": 3011296410.0,
      "step": 1296
    },
    {
      "epoch": 0.6637666325486182,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.0765436214252355e-05,
      "loss": 0.23681242763996124,
      "mean_token_accuracy": 0.9045847654342651,
      "num_tokens": 3013640716.0,
      "step": 1297
    },
    {
      "epoch": 0.6642784032753326,
      "grad_norm": 0.208984375,
      "learning_rate": 1.073603825702598e-05,
      "loss": 0.2482609897851944,
      "mean_token_accuracy": 0.9016401171684265,
      "num_tokens": 3015914922.0,
      "step": 1298
    },
    {
      "epoch": 0.6647901740020471,
      "grad_norm": 0.212890625,
      "learning_rate": 1.0706665760946952e-05,
      "loss": 0.23775359988212585,
      "mean_token_accuracy": 0.9047305583953857,
      "num_tokens": 3018204441.0,
      "step": 1299
    },
    {
      "epoch": 0.6653019447287615,
      "grad_norm": 0.212890625,
      "learning_rate": 1.0677318806742876e-05,
      "loss": 0.22727128863334656,
      "mean_token_accuracy": 0.9087015986442566,
      "num_tokens": 3020535918.0,
      "step": 1300
    },
    {
      "epoch": 0.665813715455476,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.0647997475071165e-05,
      "loss": 0.2409413605928421,
      "mean_token_accuracy": 0.9033545255661011,
      "num_tokens": 3022894839.0,
      "step": 1301
    },
    {
      "epoch": 0.6663254861821903,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.0618701846518791e-05,
      "loss": 0.23980164527893066,
      "mean_token_accuracy": 0.9048198461532593,
      "num_tokens": 3025148463.0,
      "step": 1302
    },
    {
      "epoch": 0.6668372569089048,
      "grad_norm": 0.212890625,
      "learning_rate": 1.05894320016021e-05,
      "loss": 0.24121439456939697,
      "mean_token_accuracy": 0.9035086035728455,
      "num_tokens": 3027455204.0,
      "step": 1303
    },
    {
      "epoch": 0.6673490276356192,
      "grad_norm": 0.212890625,
      "learning_rate": 1.0560188020766573e-05,
      "loss": 0.23755863308906555,
      "mean_token_accuracy": 0.9044212102890015,
      "num_tokens": 3029731672.0,
      "step": 1304
    },
    {
      "epoch": 0.6678607983623337,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.0530969984386604e-05,
      "loss": 0.23188909888267517,
      "mean_token_accuracy": 0.9066227674484253,
      "num_tokens": 3032114070.0,
      "step": 1305
    },
    {
      "epoch": 0.6683725690890481,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.0501777972765282e-05,
      "loss": 0.2366810441017151,
      "mean_token_accuracy": 0.9058021306991577,
      "num_tokens": 3034446671.0,
      "step": 1306
    },
    {
      "epoch": 0.6688843398157626,
      "grad_norm": 0.208984375,
      "learning_rate": 1.047261206613416e-05,
      "loss": 0.2438490390777588,
      "mean_token_accuracy": 0.9023560881614685,
      "num_tokens": 3036768947.0,
      "step": 1307
    },
    {
      "epoch": 0.6693961105424769,
      "grad_norm": 0.208984375,
      "learning_rate": 1.0443472344653069e-05,
      "loss": 0.23903599381446838,
      "mean_token_accuracy": 0.9037718176841736,
      "num_tokens": 3039062495.0,
      "step": 1308
    },
    {
      "epoch": 0.6699078812691914,
      "grad_norm": 0.2109375,
      "learning_rate": 1.0414358888409834e-05,
      "loss": 0.24435776472091675,
      "mean_token_accuracy": 0.9019410014152527,
      "num_tokens": 3041339830.0,
      "step": 1309
    },
    {
      "epoch": 0.6704196519959058,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.038527177742012e-05,
      "loss": 0.23791205883026123,
      "mean_token_accuracy": 0.9048508405685425,
      "num_tokens": 3043645809.0,
      "step": 1310
    },
    {
      "epoch": 0.6709314227226203,
      "grad_norm": 0.203125,
      "learning_rate": 1.0356211091627184e-05,
      "loss": 0.2416505068540573,
      "mean_token_accuracy": 0.902978777885437,
      "num_tokens": 3045983253.0,
      "step": 1311
    },
    {
      "epoch": 0.6714431934493347,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.0327176910901655e-05,
      "loss": 0.2453463077545166,
      "mean_token_accuracy": 0.901100754737854,
      "num_tokens": 3048354487.0,
      "step": 1312
    },
    {
      "epoch": 0.6719549641760492,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.0298169315041282e-05,
      "loss": 0.23418360948562622,
      "mean_token_accuracy": 0.9064454436302185,
      "num_tokens": 3050753879.0,
      "step": 1313
    },
    {
      "epoch": 0.6724667349027635,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.0269188383770805e-05,
      "loss": 0.23124009370803833,
      "mean_token_accuracy": 0.9071859121322632,
      "num_tokens": 3053058254.0,
      "step": 1314
    },
    {
      "epoch": 0.672978505629478,
      "grad_norm": 0.2109375,
      "learning_rate": 1.0240234196741647e-05,
      "loss": 0.24222639203071594,
      "mean_token_accuracy": 0.9034215211868286,
      "num_tokens": 3055393540.0,
      "step": 1315
    },
    {
      "epoch": 0.6734902763561924,
      "grad_norm": 0.19921875,
      "learning_rate": 1.0211306833531716e-05,
      "loss": 0.23504209518432617,
      "mean_token_accuracy": 0.9054377675056458,
      "num_tokens": 3057726898.0,
      "step": 1316
    },
    {
      "epoch": 0.6740020470829069,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.0182406373645217e-05,
      "loss": 0.2276429682970047,
      "mean_token_accuracy": 0.908586859703064,
      "num_tokens": 3059991575.0,
      "step": 1317
    },
    {
      "epoch": 0.6745138178096213,
      "grad_norm": 0.208984375,
      "learning_rate": 1.0153532896512405e-05,
      "loss": 0.24414746463298798,
      "mean_token_accuracy": 0.9019554257392883,
      "num_tokens": 3062357505.0,
      "step": 1318
    },
    {
      "epoch": 0.6750255885363358,
      "grad_norm": 0.20703125,
      "learning_rate": 1.0124686481489394e-05,
      "loss": 0.23452962934970856,
      "mean_token_accuracy": 0.9055185914039612,
      "num_tokens": 3064656472.0,
      "step": 1319
    },
    {
      "epoch": 0.6755373592630501,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.0095867207857873e-05,
      "loss": 0.24715928733348846,
      "mean_token_accuracy": 0.9009056091308594,
      "num_tokens": 3066960106.0,
      "step": 1320
    },
    {
      "epoch": 0.6760491299897646,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.0067075154824994e-05,
      "loss": 0.24180763959884644,
      "mean_token_accuracy": 0.902402937412262,
      "num_tokens": 3069309851.0,
      "step": 1321
    },
    {
      "epoch": 0.676560900716479,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.0038310401523067e-05,
      "loss": 0.24972456693649292,
      "mean_token_accuracy": 0.900030255317688,
      "num_tokens": 3071608548.0,
      "step": 1322
    },
    {
      "epoch": 0.6770726714431935,
      "grad_norm": 0.208984375,
      "learning_rate": 1.0009573027009356e-05,
      "loss": 0.24095606803894043,
      "mean_token_accuracy": 0.9036855697631836,
      "num_tokens": 3073885663.0,
      "step": 1323
    },
    {
      "epoch": 0.6775844421699079,
      "grad_norm": 0.2109375,
      "learning_rate": 9.980863110265904e-06,
      "loss": 0.2384411096572876,
      "mean_token_accuracy": 0.9047838449478149,
      "num_tokens": 3076083325.0,
      "step": 1324
    },
    {
      "epoch": 0.6780962128966224,
      "grad_norm": 0.21875,
      "learning_rate": 9.952180730199277e-06,
      "loss": 0.25128230452537537,
      "mean_token_accuracy": 0.8992025852203369,
      "num_tokens": 3078459924.0,
      "step": 1325
    },
    {
      "epoch": 0.6786079836233367,
      "grad_norm": 0.220703125,
      "learning_rate": 9.923525965640366e-06,
      "loss": 0.24104222655296326,
      "mean_token_accuracy": 0.9032478928565979,
      "num_tokens": 3080859626.0,
      "step": 1326
    },
    {
      "epoch": 0.6791197543500512,
      "grad_norm": 0.205078125,
      "learning_rate": 9.894898895344143e-06,
      "loss": 0.2217009961605072,
      "mean_token_accuracy": 0.911152184009552,
      "num_tokens": 3083147067.0,
      "step": 1327
    },
    {
      "epoch": 0.6796315250767656,
      "grad_norm": 0.212890625,
      "learning_rate": 9.866299597989482e-06,
      "loss": 0.25562235713005066,
      "mean_token_accuracy": 0.8976737260818481,
      "num_tokens": 3085458310.0,
      "step": 1328
    },
    {
      "epoch": 0.68014329580348,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.837728152178943e-06,
      "loss": 0.2423228919506073,
      "mean_token_accuracy": 0.903155505657196,
      "num_tokens": 3087855740.0,
      "step": 1329
    },
    {
      "epoch": 0.6806550665301945,
      "grad_norm": 0.20703125,
      "learning_rate": 9.809184636438499e-06,
      "loss": 0.24334196746349335,
      "mean_token_accuracy": 0.90272057056427,
      "num_tokens": 3090218660.0,
      "step": 1330
    },
    {
      "epoch": 0.6811668372569089,
      "grad_norm": 0.208984375,
      "learning_rate": 9.780669129217381e-06,
      "loss": 0.23861688375473022,
      "mean_token_accuracy": 0.9046429395675659,
      "num_tokens": 3092466661.0,
      "step": 1331
    },
    {
      "epoch": 0.6816786079836233,
      "grad_norm": 0.21484375,
      "learning_rate": 9.752181708887846e-06,
      "loss": 0.24636128544807434,
      "mean_token_accuracy": 0.9013132452964783,
      "num_tokens": 3094706441.0,
      "step": 1332
    },
    {
      "epoch": 0.6821903787103377,
      "grad_norm": 0.19921875,
      "learning_rate": 9.723722453744957e-06,
      "loss": 0.2403607815504074,
      "mean_token_accuracy": 0.9028465151786804,
      "num_tokens": 3097049674.0,
      "step": 1333
    },
    {
      "epoch": 0.6827021494370522,
      "grad_norm": 0.216796875,
      "learning_rate": 9.695291442006341e-06,
      "loss": 0.24951383471488953,
      "mean_token_accuracy": 0.9006697535514832,
      "num_tokens": 3099388907.0,
      "step": 1334
    },
    {
      "epoch": 0.6832139201637666,
      "grad_norm": 0.1962890625,
      "learning_rate": 9.666888751812025e-06,
      "loss": 0.2351902723312378,
      "mean_token_accuracy": 0.9055284857749939,
      "num_tokens": 3101738170.0,
      "step": 1335
    },
    {
      "epoch": 0.6837256908904811,
      "grad_norm": 0.2255859375,
      "learning_rate": 9.638514461224215e-06,
      "loss": 0.24801850318908691,
      "mean_token_accuracy": 0.9006017446517944,
      "num_tokens": 3104038207.0,
      "step": 1336
    },
    {
      "epoch": 0.6842374616171955,
      "grad_norm": 0.22265625,
      "learning_rate": 9.610168648227014e-06,
      "loss": 0.24569794535636902,
      "mean_token_accuracy": 0.9016390442848206,
      "num_tokens": 3106333732.0,
      "step": 1337
    },
    {
      "epoch": 0.6847492323439099,
      "grad_norm": 0.2099609375,
      "learning_rate": 9.581851390726295e-06,
      "loss": 0.23271030187606812,
      "mean_token_accuracy": 0.9064342975616455,
      "num_tokens": 3108671169.0,
      "step": 1338
    },
    {
      "epoch": 0.6852610030706243,
      "grad_norm": 0.20703125,
      "learning_rate": 9.553562766549437e-06,
      "loss": 0.23991617560386658,
      "mean_token_accuracy": 0.9040896892547607,
      "num_tokens": 3111021764.0,
      "step": 1339
    },
    {
      "epoch": 0.6857727737973388,
      "grad_norm": 0.2138671875,
      "learning_rate": 9.525302853445131e-06,
      "loss": 0.24720986187458038,
      "mean_token_accuracy": 0.9005013704299927,
      "num_tokens": 3113322150.0,
      "step": 1340
    },
    {
      "epoch": 0.6862845445240532,
      "grad_norm": 0.205078125,
      "learning_rate": 9.497071729083133e-06,
      "loss": 0.22926515340805054,
      "mean_token_accuracy": 0.9077858924865723,
      "num_tokens": 3115671620.0,
      "step": 1341
    },
    {
      "epoch": 0.6867963152507677,
      "grad_norm": 0.224609375,
      "learning_rate": 9.468869471054103e-06,
      "loss": 0.23745758831501007,
      "mean_token_accuracy": 0.9045853614807129,
      "num_tokens": 3117955248.0,
      "step": 1342
    },
    {
      "epoch": 0.687308085977482,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.440696156869355e-06,
      "loss": 0.23408912122249603,
      "mean_token_accuracy": 0.9058077335357666,
      "num_tokens": 3120286254.0,
      "step": 1343
    },
    {
      "epoch": 0.6878198567041965,
      "grad_norm": 0.205078125,
      "learning_rate": 9.412551863960653e-06,
      "loss": 0.233529731631279,
      "mean_token_accuracy": 0.9070298075675964,
      "num_tokens": 3122570428.0,
      "step": 1344
    },
    {
      "epoch": 0.6883316274309109,
      "grad_norm": 0.216796875,
      "learning_rate": 9.384436669679999e-06,
      "loss": 0.24274151027202606,
      "mean_token_accuracy": 0.9020358324050903,
      "num_tokens": 3124843541.0,
      "step": 1345
    },
    {
      "epoch": 0.6888433981576254,
      "grad_norm": 0.2041015625,
      "learning_rate": 9.356350651299422e-06,
      "loss": 0.24723242223262787,
      "mean_token_accuracy": 0.901509165763855,
      "num_tokens": 3127162819.0,
      "step": 1346
    },
    {
      "epoch": 0.6893551688843398,
      "grad_norm": 0.20703125,
      "learning_rate": 9.328293886010767e-06,
      "loss": 0.24078062176704407,
      "mean_token_accuracy": 0.9019098281860352,
      "num_tokens": 3129498555.0,
      "step": 1347
    },
    {
      "epoch": 0.6898669396110543,
      "grad_norm": 0.193359375,
      "learning_rate": 9.300266450925459e-06,
      "loss": 0.2315809726715088,
      "mean_token_accuracy": 0.9066898822784424,
      "num_tokens": 3131884765.0,
      "step": 1348
    },
    {
      "epoch": 0.6903787103377687,
      "grad_norm": 3.6875,
      "learning_rate": 9.272268423074338e-06,
      "loss": 0.27196720242500305,
      "mean_token_accuracy": 0.9006954431533813,
      "num_tokens": 3134216590.0,
      "step": 1349
    },
    {
      "epoch": 0.6908904810644831,
      "grad_norm": 0.220703125,
      "learning_rate": 9.244299879407408e-06,
      "loss": 0.2456618994474411,
      "mean_token_accuracy": 0.9019328355789185,
      "num_tokens": 3136512226.0,
      "step": 1350
    },
    {
      "epoch": 0.6914022517911975,
      "grad_norm": 0.19140625,
      "learning_rate": 9.216360896793642e-06,
      "loss": 0.2265632301568985,
      "mean_token_accuracy": 0.9089025259017944,
      "num_tokens": 3138905911.0,
      "step": 1351
    },
    {
      "epoch": 0.691914022517912,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.188451552020763e-06,
      "loss": 0.24352310597896576,
      "mean_token_accuracy": 0.9024208784103394,
      "num_tokens": 3141207987.0,
      "step": 1352
    },
    {
      "epoch": 0.6924257932446264,
      "grad_norm": 0.22265625,
      "learning_rate": 9.16057192179504e-06,
      "loss": 0.2541433274745941,
      "mean_token_accuracy": 0.8979525566101074,
      "num_tokens": 3143453213.0,
      "step": 1353
    },
    {
      "epoch": 0.6929375639713409,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.132722082741085e-06,
      "loss": 0.2419171929359436,
      "mean_token_accuracy": 0.9029834270477295,
      "num_tokens": 3145813333.0,
      "step": 1354
    },
    {
      "epoch": 0.6934493346980553,
      "grad_norm": 0.2060546875,
      "learning_rate": 9.104902111401604e-06,
      "loss": 0.2398536205291748,
      "mean_token_accuracy": 0.9039478302001953,
      "num_tokens": 3148188462.0,
      "step": 1355
    },
    {
      "epoch": 0.6939611054247697,
      "grad_norm": 0.20703125,
      "learning_rate": 9.077112084237236e-06,
      "loss": 0.23878054320812225,
      "mean_token_accuracy": 0.9039744138717651,
      "num_tokens": 3150468153.0,
      "step": 1356
    },
    {
      "epoch": 0.6944728761514841,
      "grad_norm": 0.203125,
      "learning_rate": 9.049352077626321e-06,
      "loss": 0.23360854387283325,
      "mean_token_accuracy": 0.906240701675415,
      "num_tokens": 3152854736.0,
      "step": 1357
    },
    {
      "epoch": 0.6949846468781986,
      "grad_norm": 0.2177734375,
      "learning_rate": 9.021622167864682e-06,
      "loss": 0.242182657122612,
      "mean_token_accuracy": 0.9029110670089722,
      "num_tokens": 3155139801.0,
      "step": 1358
    },
    {
      "epoch": 0.695496417604913,
      "grad_norm": 0.21484375,
      "learning_rate": 8.993922431165429e-06,
      "loss": 0.24490024149417877,
      "mean_token_accuracy": 0.9013450741767883,
      "num_tokens": 3157496377.0,
      "step": 1359
    },
    {
      "epoch": 0.6960081883316275,
      "grad_norm": 0.2041015625,
      "learning_rate": 8.966252943658742e-06,
      "loss": 0.2447492778301239,
      "mean_token_accuracy": 0.90185546875,
      "num_tokens": 3159816391.0,
      "step": 1360
    },
    {
      "epoch": 0.6965199590583419,
      "grad_norm": 0.208984375,
      "learning_rate": 8.93861378139167e-06,
      "loss": 0.24788379669189453,
      "mean_token_accuracy": 0.9005329012870789,
      "num_tokens": 3162179878.0,
      "step": 1361
    },
    {
      "epoch": 0.6970317297850563,
      "grad_norm": 0.2158203125,
      "learning_rate": 8.911005020327897e-06,
      "loss": 0.2357151210308075,
      "mean_token_accuracy": 0.905322253704071,
      "num_tokens": 3164509464.0,
      "step": 1362
    },
    {
      "epoch": 0.6975435005117707,
      "grad_norm": 0.2197265625,
      "learning_rate": 8.88342673634757e-06,
      "loss": 0.24004599452018738,
      "mean_token_accuracy": 0.9032361507415771,
      "num_tokens": 3166721578.0,
      "step": 1363
    },
    {
      "epoch": 0.6980552712384852,
      "grad_norm": 0.208984375,
      "learning_rate": 8.855879005247073e-06,
      "loss": 0.23673835396766663,
      "mean_token_accuracy": 0.9052594900131226,
      "num_tokens": 3169087915.0,
      "step": 1364
    },
    {
      "epoch": 0.6985670419651996,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.82836190273881e-06,
      "loss": 0.23122775554656982,
      "mean_token_accuracy": 0.9063940048217773,
      "num_tokens": 3171303257.0,
      "step": 1365
    },
    {
      "epoch": 0.6990788126919141,
      "grad_norm": 0.212890625,
      "learning_rate": 8.800875504451003e-06,
      "loss": 0.23669832944869995,
      "mean_token_accuracy": 0.9053252935409546,
      "num_tokens": 3173669446.0,
      "step": 1366
    },
    {
      "epoch": 0.6995905834186285,
      "grad_norm": 0.20703125,
      "learning_rate": 8.773419885927499e-06,
      "loss": 0.24365505576133728,
      "mean_token_accuracy": 0.9025254249572754,
      "num_tokens": 3176037124.0,
      "step": 1367
    },
    {
      "epoch": 0.7001023541453428,
      "grad_norm": 0.2275390625,
      "learning_rate": 8.745995122627544e-06,
      "loss": 0.2447820007801056,
      "mean_token_accuracy": 0.9018913507461548,
      "num_tokens": 3178310832.0,
      "step": 1368
    },
    {
      "epoch": 0.7006141248720573,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.718601289925567e-06,
      "loss": 0.23365047574043274,
      "mean_token_accuracy": 0.905461311340332,
      "num_tokens": 3180635424.0,
      "step": 1369
    },
    {
      "epoch": 0.7011258955987717,
      "grad_norm": 0.2138671875,
      "learning_rate": 8.691238463111006e-06,
      "loss": 0.23187586665153503,
      "mean_token_accuracy": 0.9068617820739746,
      "num_tokens": 3183047665.0,
      "step": 1370
    },
    {
      "epoch": 0.7016376663254862,
      "grad_norm": 0.208984375,
      "learning_rate": 8.663906717388078e-06,
      "loss": 0.24338220059871674,
      "mean_token_accuracy": 0.9020247459411621,
      "num_tokens": 3185353074.0,
      "step": 1371
    },
    {
      "epoch": 0.7021494370522006,
      "grad_norm": 0.2138671875,
      "learning_rate": 8.636606127875576e-06,
      "loss": 0.23596347868442535,
      "mean_token_accuracy": 0.9057803750038147,
      "num_tokens": 3187660349.0,
      "step": 1372
    },
    {
      "epoch": 0.702661207778915,
      "grad_norm": 0.208984375,
      "learning_rate": 8.609336769606642e-06,
      "loss": 0.24136127531528473,
      "mean_token_accuracy": 0.9038076400756836,
      "num_tokens": 3189969567.0,
      "step": 1373
    },
    {
      "epoch": 0.7031729785056294,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.582098717528621e-06,
      "loss": 0.23653998970985413,
      "mean_token_accuracy": 0.9048939943313599,
      "num_tokens": 3192316904.0,
      "step": 1374
    },
    {
      "epoch": 0.7036847492323439,
      "grad_norm": 0.2373046875,
      "learning_rate": 8.554892046502801e-06,
      "loss": 0.24771855771541595,
      "mean_token_accuracy": 0.900834858417511,
      "num_tokens": 3194629628.0,
      "step": 1375
    },
    {
      "epoch": 0.7041965199590583,
      "grad_norm": 0.2177734375,
      "learning_rate": 8.527716831304194e-06,
      "loss": 0.2507264316082001,
      "mean_token_accuracy": 0.8996508121490479,
      "num_tokens": 3196956016.0,
      "step": 1376
    },
    {
      "epoch": 0.7047082906857728,
      "grad_norm": 0.2158203125,
      "learning_rate": 8.500573146621394e-06,
      "loss": 0.23188209533691406,
      "mean_token_accuracy": 0.9063490033149719,
      "num_tokens": 3199251681.0,
      "step": 1377
    },
    {
      "epoch": 0.7052200614124872,
      "grad_norm": 0.205078125,
      "learning_rate": 8.473461067056322e-06,
      "loss": 0.2383233904838562,
      "mean_token_accuracy": 0.9040708541870117,
      "num_tokens": 3201625296.0,
      "step": 1378
    },
    {
      "epoch": 0.7057318321392017,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.446380667124042e-06,
      "loss": 0.2557116746902466,
      "mean_token_accuracy": 0.8973208665847778,
      "num_tokens": 3203838809.0,
      "step": 1379
    },
    {
      "epoch": 0.706243602865916,
      "grad_norm": 0.2158203125,
      "learning_rate": 8.419332021252524e-06,
      "loss": 0.235560804605484,
      "mean_token_accuracy": 0.9053171277046204,
      "num_tokens": 3206205346.0,
      "step": 1380
    },
    {
      "epoch": 0.7067553735926305,
      "grad_norm": 0.1982421875,
      "learning_rate": 8.392315203782505e-06,
      "loss": 0.23199588060379028,
      "mean_token_accuracy": 0.9063323140144348,
      "num_tokens": 3208583492.0,
      "step": 1381
    },
    {
      "epoch": 0.7072671443193449,
      "grad_norm": 0.212890625,
      "learning_rate": 8.365330288967231e-06,
      "loss": 0.24587494134902954,
      "mean_token_accuracy": 0.9005568623542786,
      "num_tokens": 3210871911.0,
      "step": 1382
    },
    {
      "epoch": 0.7077789150460594,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.338377350972235e-06,
      "loss": 0.25031578540802,
      "mean_token_accuracy": 0.8995984792709351,
      "num_tokens": 3213203724.0,
      "step": 1383
    },
    {
      "epoch": 0.7082906857727738,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.311456463875208e-06,
      "loss": 0.2417197972536087,
      "mean_token_accuracy": 0.9028545022010803,
      "num_tokens": 3215522841.0,
      "step": 1384
    },
    {
      "epoch": 0.7088024564994883,
      "grad_norm": 0.2041015625,
      "learning_rate": 8.284567701665724e-06,
      "loss": 0.2266548126935959,
      "mean_token_accuracy": 0.9090459942817688,
      "num_tokens": 3217844478.0,
      "step": 1385
    },
    {
      "epoch": 0.7093142272262026,
      "grad_norm": 0.205078125,
      "learning_rate": 8.257711138245088e-06,
      "loss": 0.23606054484844208,
      "mean_token_accuracy": 0.9047561883926392,
      "num_tokens": 3220244976.0,
      "step": 1386
    },
    {
      "epoch": 0.7098259979529171,
      "grad_norm": 0.20703125,
      "learning_rate": 8.23088684742608e-06,
      "loss": 0.23847833275794983,
      "mean_token_accuracy": 0.9045600891113281,
      "num_tokens": 3222580740.0,
      "step": 1387
    },
    {
      "epoch": 0.7103377686796315,
      "grad_norm": 0.212890625,
      "learning_rate": 8.204094902932797e-06,
      "loss": 0.23473644256591797,
      "mean_token_accuracy": 0.9065552353858948,
      "num_tokens": 3224887999.0,
      "step": 1388
    },
    {
      "epoch": 0.710849539406346,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.177335378400458e-06,
      "loss": 0.24608084559440613,
      "mean_token_accuracy": 0.9015039205551147,
      "num_tokens": 3227222975.0,
      "step": 1389
    },
    {
      "epoch": 0.7113613101330604,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.150608347375135e-06,
      "loss": 0.24145427346229553,
      "mean_token_accuracy": 0.9032613039016724,
      "num_tokens": 3229580885.0,
      "step": 1390
    },
    {
      "epoch": 0.7118730808597749,
      "grad_norm": 0.2109375,
      "learning_rate": 8.123913883313624e-06,
      "loss": 0.23602724075317383,
      "mean_token_accuracy": 0.9051846265792847,
      "num_tokens": 3231822246.0,
      "step": 1391
    },
    {
      "epoch": 0.7123848515864892,
      "grad_norm": 0.2080078125,
      "learning_rate": 8.097252059583203e-06,
      "loss": 0.24641293287277222,
      "mean_token_accuracy": 0.9005869626998901,
      "num_tokens": 3234241199.0,
      "step": 1392
    },
    {
      "epoch": 0.7128966223132037,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.070622949461454e-06,
      "loss": 0.2453475296497345,
      "mean_token_accuracy": 0.9013621807098389,
      "num_tokens": 3236552238.0,
      "step": 1393
    },
    {
      "epoch": 0.7134083930399181,
      "grad_norm": 0.2119140625,
      "learning_rate": 8.044026626136022e-06,
      "loss": 0.23935642838478088,
      "mean_token_accuracy": 0.9047154188156128,
      "num_tokens": 3238816968.0,
      "step": 1394
    },
    {
      "epoch": 0.7139201637666326,
      "grad_norm": 0.205078125,
      "learning_rate": 8.017463162704458e-06,
      "loss": 0.24471712112426758,
      "mean_token_accuracy": 0.9018443822860718,
      "num_tokens": 3241102755.0,
      "step": 1395
    },
    {
      "epoch": 0.714431934493347,
      "grad_norm": 0.2138671875,
      "learning_rate": 7.990932632174022e-06,
      "loss": 0.24463540315628052,
      "mean_token_accuracy": 0.9011511206626892,
      "num_tokens": 3243466272.0,
      "step": 1396
    },
    {
      "epoch": 0.7149437052200615,
      "grad_norm": 0.201171875,
      "learning_rate": 7.964435107461417e-06,
      "loss": 0.22573699057102203,
      "mean_token_accuracy": 0.9086453914642334,
      "num_tokens": 3245783198.0,
      "step": 1397
    },
    {
      "epoch": 0.7154554759467758,
      "grad_norm": 0.20703125,
      "learning_rate": 7.93797066139266e-06,
      "loss": 0.23870599269866943,
      "mean_token_accuracy": 0.9041187763214111,
      "num_tokens": 3248118140.0,
      "step": 1398
    },
    {
      "epoch": 0.7159672466734903,
      "grad_norm": 0.201171875,
      "learning_rate": 7.911539366702849e-06,
      "loss": 0.23370197415351868,
      "mean_token_accuracy": 0.9053938388824463,
      "num_tokens": 3250376383.0,
      "step": 1399
    },
    {
      "epoch": 0.7164790174002047,
      "grad_norm": 0.2119140625,
      "learning_rate": 7.885141296035976e-06,
      "loss": 0.231037899851799,
      "mean_token_accuracy": 0.9072940349578857,
      "num_tokens": 3252744226.0,
      "step": 1400
    },
    {
      "epoch": 0.7169907881269192,
      "grad_norm": 0.2021484375,
      "learning_rate": 7.858776521944698e-06,
      "loss": 0.2432289719581604,
      "mean_token_accuracy": 0.9024935960769653,
      "num_tokens": 3255066159.0,
      "step": 1401
    },
    {
      "epoch": 0.7175025588536336,
      "grad_norm": 0.203125,
      "learning_rate": 7.832445116890177e-06,
      "loss": 0.23766840994358063,
      "mean_token_accuracy": 0.9049115180969238,
      "num_tokens": 3257442146.0,
      "step": 1402
    },
    {
      "epoch": 0.718014329580348,
      "grad_norm": 0.2060546875,
      "learning_rate": 7.806147153241863e-06,
      "loss": 0.23876744508743286,
      "mean_token_accuracy": 0.9046311378479004,
      "num_tokens": 3259748855.0,
      "step": 1403
    },
    {
      "epoch": 0.7185261003070624,
      "grad_norm": 0.1982421875,
      "learning_rate": 7.779882703277293e-06,
      "loss": 0.23331347107887268,
      "mean_token_accuracy": 0.9062086939811707,
      "num_tokens": 3262023356.0,
      "step": 1404
    },
    {
      "epoch": 0.7190378710337769,
      "grad_norm": 0.2041015625,
      "learning_rate": 7.753651839181888e-06,
      "loss": 0.23151686787605286,
      "mean_token_accuracy": 0.9073910713195801,
      "num_tokens": 3264384077.0,
      "step": 1405
    },
    {
      "epoch": 0.7195496417604913,
      "grad_norm": 0.203125,
      "learning_rate": 7.727454633048772e-06,
      "loss": 0.2308761179447174,
      "mean_token_accuracy": 0.9077283143997192,
      "num_tokens": 3266764747.0,
      "step": 1406
    },
    {
      "epoch": 0.7200614124872057,
      "grad_norm": 0.19921875,
      "learning_rate": 7.701291156878563e-06,
      "loss": 0.24761204421520233,
      "mean_token_accuracy": 0.9005869626998901,
      "num_tokens": 3269106329.0,
      "step": 1407
    },
    {
      "epoch": 0.7205731832139202,
      "grad_norm": 0.2119140625,
      "learning_rate": 7.675161482579157e-06,
      "loss": 0.242473304271698,
      "mean_token_accuracy": 0.9028928279876709,
      "num_tokens": 3271428995.0,
      "step": 1408
    },
    {
      "epoch": 0.7210849539406345,
      "grad_norm": 0.2041015625,
      "learning_rate": 7.649065681965569e-06,
      "loss": 0.22883516550064087,
      "mean_token_accuracy": 0.9084866046905518,
      "num_tokens": 3273773277.0,
      "step": 1409
    },
    {
      "epoch": 0.721596724667349,
      "grad_norm": 0.2119140625,
      "learning_rate": 7.623003826759707e-06,
      "loss": 0.24212776124477386,
      "mean_token_accuracy": 0.9038840532302856,
      "num_tokens": 3276062677.0,
      "step": 1410
    },
    {
      "epoch": 0.7221084953940634,
      "grad_norm": 0.216796875,
      "learning_rate": 7.596975988590187e-06,
      "loss": 0.2353907972574234,
      "mean_token_accuracy": 0.9051687717437744,
      "num_tokens": 3278350298.0,
      "step": 1411
    },
    {
      "epoch": 0.7226202661207779,
      "grad_norm": 0.20703125,
      "learning_rate": 7.570982238992128e-06,
      "loss": 0.23786267638206482,
      "mean_token_accuracy": 0.9048346281051636,
      "num_tokens": 3280639098.0,
      "step": 1412
    },
    {
      "epoch": 0.7231320368474923,
      "grad_norm": 0.212890625,
      "learning_rate": 7.545022649406963e-06,
      "loss": 0.23708005249500275,
      "mean_token_accuracy": 0.9043331146240234,
      "num_tokens": 3282973445.0,
      "step": 1413
    },
    {
      "epoch": 0.7236438075742068,
      "grad_norm": 0.220703125,
      "learning_rate": 7.519097291182244e-06,
      "loss": 0.24295803904533386,
      "mean_token_accuracy": 0.9031884670257568,
      "num_tokens": 3285259891.0,
      "step": 1414
    },
    {
      "epoch": 0.7241555783009211,
      "grad_norm": 0.212890625,
      "learning_rate": 7.49320623557142e-06,
      "loss": 0.23913902044296265,
      "mean_token_accuracy": 0.903611958026886,
      "num_tokens": 3287483396.0,
      "step": 1415
    },
    {
      "epoch": 0.7246673490276356,
      "grad_norm": 0.2197265625,
      "learning_rate": 7.467349553733687e-06,
      "loss": 0.253231942653656,
      "mean_token_accuracy": 0.8984534740447998,
      "num_tokens": 3289762751.0,
      "step": 1416
    },
    {
      "epoch": 0.72517911975435,
      "grad_norm": 0.220703125,
      "learning_rate": 7.441527316733757e-06,
      "loss": 0.24526646733283997,
      "mean_token_accuracy": 0.9019604921340942,
      "num_tokens": 3292078452.0,
      "step": 1417
    },
    {
      "epoch": 0.7256908904810645,
      "grad_norm": 0.2197265625,
      "learning_rate": 7.415739595541676e-06,
      "loss": 0.2486463487148285,
      "mean_token_accuracy": 0.9002640247344971,
      "num_tokens": 3294410598.0,
      "step": 1418
    },
    {
      "epoch": 0.7262026612077789,
      "grad_norm": 0.208984375,
      "learning_rate": 7.389986461032623e-06,
      "loss": 0.24028301239013672,
      "mean_token_accuracy": 0.903761625289917,
      "num_tokens": 3296711138.0,
      "step": 1419
    },
    {
      "epoch": 0.7267144319344934,
      "grad_norm": 0.21484375,
      "learning_rate": 7.364267983986722e-06,
      "loss": 0.25343775749206543,
      "mean_token_accuracy": 0.8980152606964111,
      "num_tokens": 3299068620.0,
      "step": 1420
    },
    {
      "epoch": 0.7272262026612077,
      "grad_norm": 0.201171875,
      "learning_rate": 7.338584235088853e-06,
      "loss": 0.23052768409252167,
      "mean_token_accuracy": 0.907322883605957,
      "num_tokens": 3301412606.0,
      "step": 1421
    },
    {
      "epoch": 0.7277379733879222,
      "grad_norm": 0.2109375,
      "learning_rate": 7.31293528492842e-06,
      "loss": 0.2299337089061737,
      "mean_token_accuracy": 0.9074872732162476,
      "num_tokens": 3303727973.0,
      "step": 1422
    },
    {
      "epoch": 0.7282497441146366,
      "grad_norm": 0.203125,
      "learning_rate": 7.287321203999216e-06,
      "loss": 0.2408350706100464,
      "mean_token_accuracy": 0.9029034376144409,
      "num_tokens": 3306037220.0,
      "step": 1423
    },
    {
      "epoch": 0.7287615148413511,
      "grad_norm": 0.2080078125,
      "learning_rate": 7.261742062699188e-06,
      "loss": 0.2321765124797821,
      "mean_token_accuracy": 0.9069453477859497,
      "num_tokens": 3308263647.0,
      "step": 1424
    },
    {
      "epoch": 0.7292732855680655,
      "grad_norm": 0.2060546875,
      "learning_rate": 7.2361979313302536e-06,
      "loss": 0.25000208616256714,
      "mean_token_accuracy": 0.89937424659729,
      "num_tokens": 3310561788.0,
      "step": 1425
    },
    {
      "epoch": 0.72978505629478,
      "grad_norm": 0.2177734375,
      "learning_rate": 7.210688880098109e-06,
      "loss": 0.2420661747455597,
      "mean_token_accuracy": 0.9036683440208435,
      "num_tokens": 3312913267.0,
      "step": 1426
    },
    {
      "epoch": 0.7302968270214943,
      "grad_norm": 0.201171875,
      "learning_rate": 7.185214979112038e-06,
      "loss": 0.222885400056839,
      "mean_token_accuracy": 0.910603404045105,
      "num_tokens": 3315204760.0,
      "step": 1427
    },
    {
      "epoch": 0.7308085977482088,
      "grad_norm": 0.208984375,
      "learning_rate": 7.159776298384724e-06,
      "loss": 0.23907789587974548,
      "mean_token_accuracy": 0.9044662117958069,
      "num_tokens": 3317467642.0,
      "step": 1428
    },
    {
      "epoch": 0.7313203684749232,
      "grad_norm": 0.2197265625,
      "learning_rate": 7.134372907832026e-06,
      "loss": 0.23700915277004242,
      "mean_token_accuracy": 0.9047907590866089,
      "num_tokens": 3319711729.0,
      "step": 1429
    },
    {
      "epoch": 0.7318321392016377,
      "grad_norm": 0.212890625,
      "learning_rate": 7.1090048772728385e-06,
      "loss": 0.24689772725105286,
      "mean_token_accuracy": 0.9012579917907715,
      "num_tokens": 3322059174.0,
      "step": 1430
    },
    {
      "epoch": 0.7323439099283521,
      "grad_norm": 0.2109375,
      "learning_rate": 7.0836722764288564e-06,
      "loss": 0.24583087861537933,
      "mean_token_accuracy": 0.9008960723876953,
      "num_tokens": 3324415948.0,
      "step": 1431
    },
    {
      "epoch": 0.7328556806550666,
      "grad_norm": 0.20703125,
      "learning_rate": 7.058375174924419e-06,
      "loss": 0.24422569572925568,
      "mean_token_accuracy": 0.9023270010948181,
      "num_tokens": 3326674833.0,
      "step": 1432
    },
    {
      "epoch": 0.7333674513817809,
      "grad_norm": 0.19140625,
      "learning_rate": 7.033113642286258e-06,
      "loss": 0.2304789125919342,
      "mean_token_accuracy": 0.9073271751403809,
      "num_tokens": 3329063210.0,
      "step": 1433
    },
    {
      "epoch": 0.7338792221084954,
      "grad_norm": 0.205078125,
      "learning_rate": 7.007887747943396e-06,
      "loss": 0.23883390426635742,
      "mean_token_accuracy": 0.903802216053009,
      "num_tokens": 3331473245.0,
      "step": 1434
    },
    {
      "epoch": 0.7343909928352098,
      "grad_norm": 0.22265625,
      "learning_rate": 6.982697561226888e-06,
      "loss": 0.25801628828048706,
      "mean_token_accuracy": 0.8961836099624634,
      "num_tokens": 3333795328.0,
      "step": 1435
    },
    {
      "epoch": 0.7349027635619243,
      "grad_norm": 0.1982421875,
      "learning_rate": 6.957543151369628e-06,
      "loss": 0.23488792777061462,
      "mean_token_accuracy": 0.9056859612464905,
      "num_tokens": 3336186111.0,
      "step": 1436
    },
    {
      "epoch": 0.7354145342886387,
      "grad_norm": 0.19921875,
      "learning_rate": 6.93242458750621e-06,
      "loss": 0.2350383698940277,
      "mean_token_accuracy": 0.9052417278289795,
      "num_tokens": 3338522146.0,
      "step": 1437
    },
    {
      "epoch": 0.7359263050153532,
      "grad_norm": 0.1982421875,
      "learning_rate": 6.907341938672696e-06,
      "loss": 0.23638111352920532,
      "mean_token_accuracy": 0.9039426445960999,
      "num_tokens": 3340849716.0,
      "step": 1438
    },
    {
      "epoch": 0.7364380757420675,
      "grad_norm": 0.2197265625,
      "learning_rate": 6.882295273806449e-06,
      "loss": 0.2503410577774048,
      "mean_token_accuracy": 0.899316668510437,
      "num_tokens": 3343196896.0,
      "step": 1439
    },
    {
      "epoch": 0.736949846468782,
      "grad_norm": 0.208984375,
      "learning_rate": 6.8572846617459045e-06,
      "loss": 0.23434507846832275,
      "mean_token_accuracy": 0.9052055478096008,
      "num_tokens": 3345542812.0,
      "step": 1440
    },
    {
      "epoch": 0.7374616171954964,
      "grad_norm": 0.216796875,
      "learning_rate": 6.832310171230449e-06,
      "loss": 0.2518743574619293,
      "mean_token_accuracy": 0.899559736251831,
      "num_tokens": 3347870501.0,
      "step": 1441
    },
    {
      "epoch": 0.7379733879222109,
      "grad_norm": 0.21484375,
      "learning_rate": 6.807371870900179e-06,
      "loss": 0.24081964790821075,
      "mean_token_accuracy": 0.9028434157371521,
      "num_tokens": 3350208608.0,
      "step": 1442
    },
    {
      "epoch": 0.7384851586489253,
      "grad_norm": 0.205078125,
      "learning_rate": 6.782469829295701e-06,
      "loss": 0.23452921211719513,
      "mean_token_accuracy": 0.9060171842575073,
      "num_tokens": 3352519374.0,
      "step": 1443
    },
    {
      "epoch": 0.7389969293756398,
      "grad_norm": 0.2080078125,
      "learning_rate": 6.757604114858001e-06,
      "loss": 0.23252585530281067,
      "mean_token_accuracy": 0.9065992832183838,
      "num_tokens": 3354890436.0,
      "step": 1444
    },
    {
      "epoch": 0.7395087001023541,
      "grad_norm": 0.201171875,
      "learning_rate": 6.732774795928203e-06,
      "loss": 0.2275145947933197,
      "mean_token_accuracy": 0.9084676504135132,
      "num_tokens": 3357197636.0,
      "step": 1445
    },
    {
      "epoch": 0.7400204708290685,
      "grad_norm": 0.21484375,
      "learning_rate": 6.707981940747419e-06,
      "loss": 0.2423272728919983,
      "mean_token_accuracy": 0.9027924537658691,
      "num_tokens": 3359497890.0,
      "step": 1446
    },
    {
      "epoch": 0.740532241555783,
      "grad_norm": 0.2158203125,
      "learning_rate": 6.683225617456519e-06,
      "loss": 0.25342661142349243,
      "mean_token_accuracy": 0.8990950584411621,
      "num_tokens": 3361756398.0,
      "step": 1447
    },
    {
      "epoch": 0.7410440122824974,
      "grad_norm": 0.2138671875,
      "learning_rate": 6.658505894095979e-06,
      "loss": 0.23116308450698853,
      "mean_token_accuracy": 0.9061205387115479,
      "num_tokens": 3364165101.0,
      "step": 1448
    },
    {
      "epoch": 0.7415557830092119,
      "grad_norm": 0.2177734375,
      "learning_rate": 6.63382283860571e-06,
      "loss": 0.2519935667514801,
      "mean_token_accuracy": 0.898392379283905,
      "num_tokens": 3366460286.0,
      "step": 1449
    },
    {
      "epoch": 0.7420675537359263,
      "grad_norm": 0.205078125,
      "learning_rate": 6.609176518824798e-06,
      "loss": 0.23662185668945312,
      "mean_token_accuracy": 0.9047504663467407,
      "num_tokens": 3368789956.0,
      "step": 1450
    },
    {
      "epoch": 0.7425793244626407,
      "grad_norm": 0.20703125,
      "learning_rate": 6.584567002491398e-06,
      "loss": 0.24247139692306519,
      "mean_token_accuracy": 0.9029272198677063,
      "num_tokens": 3371206811.0,
      "step": 1451
    },
    {
      "epoch": 0.7430910951893551,
      "grad_norm": 0.2216796875,
      "learning_rate": 6.559994357242501e-06,
      "loss": 0.24988959729671478,
      "mean_token_accuracy": 0.9000434875488281,
      "num_tokens": 3373523165.0,
      "step": 1452
    },
    {
      "epoch": 0.7436028659160696,
      "grad_norm": 0.201171875,
      "learning_rate": 6.53545865061377e-06,
      "loss": 0.238291934132576,
      "mean_token_accuracy": 0.9042469263076782,
      "num_tokens": 3375847988.0,
      "step": 1453
    },
    {
      "epoch": 0.744114636642784,
      "grad_norm": 0.2060546875,
      "learning_rate": 6.510959950039328e-06,
      "loss": 0.23719573020935059,
      "mean_token_accuracy": 0.9048639535903931,
      "num_tokens": 3378069018.0,
      "step": 1454
    },
    {
      "epoch": 0.7446264073694985,
      "grad_norm": 0.1962890625,
      "learning_rate": 6.486498322851602e-06,
      "loss": 0.22480593621730804,
      "mean_token_accuracy": 0.9094630479812622,
      "num_tokens": 3380398403.0,
      "step": 1455
    },
    {
      "epoch": 0.7451381780962129,
      "grad_norm": 0.2041015625,
      "learning_rate": 6.462073836281142e-06,
      "loss": 0.25007081031799316,
      "mean_token_accuracy": 0.8996092081069946,
      "num_tokens": 3382788894.0,
      "step": 1456
    },
    {
      "epoch": 0.7456499488229273,
      "grad_norm": 0.2060546875,
      "learning_rate": 6.43768655745639e-06,
      "loss": 0.24678850173950195,
      "mean_token_accuracy": 0.9017046689987183,
      "num_tokens": 3385125748.0,
      "step": 1457
    },
    {
      "epoch": 0.7461617195496417,
      "grad_norm": 0.22265625,
      "learning_rate": 6.413336553403542e-06,
      "loss": 0.23912662267684937,
      "mean_token_accuracy": 0.9036432504653931,
      "num_tokens": 3387358218.0,
      "step": 1458
    },
    {
      "epoch": 0.7466734902763562,
      "grad_norm": 0.208984375,
      "learning_rate": 6.389023891046351e-06,
      "loss": 0.2367551475763321,
      "mean_token_accuracy": 0.9060007333755493,
      "num_tokens": 3389705177.0,
      "step": 1459
    },
    {
      "epoch": 0.7471852610030706,
      "grad_norm": 0.205078125,
      "learning_rate": 6.364748637205935e-06,
      "loss": 0.23884378373622894,
      "mean_token_accuracy": 0.9042931795120239,
      "num_tokens": 3392047517.0,
      "step": 1460
    },
    {
      "epoch": 0.7476970317297851,
      "grad_norm": 0.21484375,
      "learning_rate": 6.34051085860059e-06,
      "loss": 0.2405683994293213,
      "mean_token_accuracy": 0.9021862745285034,
      "num_tokens": 3394290239.0,
      "step": 1461
    },
    {
      "epoch": 0.7482088024564995,
      "grad_norm": 0.2099609375,
      "learning_rate": 6.316310621845627e-06,
      "loss": 0.24024197459220886,
      "mean_token_accuracy": 0.9043231010437012,
      "num_tokens": 3396663833.0,
      "step": 1462
    },
    {
      "epoch": 0.7487205731832139,
      "grad_norm": 0.21484375,
      "learning_rate": 6.292147993453175e-06,
      "loss": 0.24579370021820068,
      "mean_token_accuracy": 0.9002299904823303,
      "num_tokens": 3398919290.0,
      "step": 1463
    },
    {
      "epoch": 0.7492323439099283,
      "grad_norm": 0.2109375,
      "learning_rate": 6.2680230398319915e-06,
      "loss": 0.24033096432685852,
      "mean_token_accuracy": 0.9030387997627258,
      "num_tokens": 3401230644.0,
      "step": 1464
    },
    {
      "epoch": 0.7497441146366428,
      "grad_norm": 0.2236328125,
      "learning_rate": 6.243935827287299e-06,
      "loss": 0.25999951362609863,
      "mean_token_accuracy": 0.8956894874572754,
      "num_tokens": 3403560369.0,
      "step": 1465
    },
    {
      "epoch": 0.7502558853633572,
      "grad_norm": 0.208984375,
      "learning_rate": 6.219886422020587e-06,
      "loss": 0.24716627597808838,
      "mean_token_accuracy": 0.9011995792388916,
      "num_tokens": 3405938273.0,
      "step": 1466
    },
    {
      "epoch": 0.7507676560900717,
      "grad_norm": 0.203125,
      "learning_rate": 6.195874890129439e-06,
      "loss": 0.22263939678668976,
      "mean_token_accuracy": 0.9096802473068237,
      "num_tokens": 3408298859.0,
      "step": 1467
    },
    {
      "epoch": 0.7512794268167861,
      "grad_norm": 0.208984375,
      "learning_rate": 6.171901297607332e-06,
      "loss": 0.2463497519493103,
      "mean_token_accuracy": 0.9023362398147583,
      "num_tokens": 3410661474.0,
      "step": 1468
    },
    {
      "epoch": 0.7517911975435005,
      "grad_norm": 0.2109375,
      "learning_rate": 6.1479657103434886e-06,
      "loss": 0.23987260460853577,
      "mean_token_accuracy": 0.9036012887954712,
      "num_tokens": 3413066595.0,
      "step": 1469
    },
    {
      "epoch": 0.7523029682702149,
      "grad_norm": 0.2109375,
      "learning_rate": 6.12406819412267e-06,
      "loss": 0.24927088618278503,
      "mean_token_accuracy": 0.9000421166419983,
      "num_tokens": 3415410323.0,
      "step": 1470
    },
    {
      "epoch": 0.7528147389969294,
      "grad_norm": 0.197265625,
      "learning_rate": 6.1002088146250015e-06,
      "loss": 0.24125051498413086,
      "mean_token_accuracy": 0.9041305184364319,
      "num_tokens": 3417711767.0,
      "step": 1471
    },
    {
      "epoch": 0.7533265097236438,
      "grad_norm": 0.2158203125,
      "learning_rate": 6.076387637425798e-06,
      "loss": 0.25007227063179016,
      "mean_token_accuracy": 0.900786817073822,
      "num_tokens": 3420072847.0,
      "step": 1472
    },
    {
      "epoch": 0.7538382804503583,
      "grad_norm": 0.2021484375,
      "learning_rate": 6.052604727995371e-06,
      "loss": 0.22310294210910797,
      "mean_token_accuracy": 0.9096604585647583,
      "num_tokens": 3422387771.0,
      "step": 1473
    },
    {
      "epoch": 0.7543500511770727,
      "grad_norm": 0.220703125,
      "learning_rate": 6.02886015169887e-06,
      "loss": 0.24500447511672974,
      "mean_token_accuracy": 0.9014773368835449,
      "num_tokens": 3424732544.0,
      "step": 1474
    },
    {
      "epoch": 0.7548618219037871,
      "grad_norm": 0.2080078125,
      "learning_rate": 6.005153973796067e-06,
      "loss": 0.2349054515361786,
      "mean_token_accuracy": 0.9055255651473999,
      "num_tokens": 3427056465.0,
      "step": 1475
    },
    {
      "epoch": 0.7553735926305015,
      "grad_norm": 0.201171875,
      "learning_rate": 5.9814862594412225e-06,
      "loss": 0.23629289865493774,
      "mean_token_accuracy": 0.905005693435669,
      "num_tokens": 3429413524.0,
      "step": 1476
    },
    {
      "epoch": 0.755885363357216,
      "grad_norm": 0.220703125,
      "learning_rate": 5.957857073682869e-06,
      "loss": 0.24342885613441467,
      "mean_token_accuracy": 0.9018474817276001,
      "num_tokens": 3431707376.0,
      "step": 1477
    },
    {
      "epoch": 0.7563971340839304,
      "grad_norm": 0.2109375,
      "learning_rate": 5.934266481463653e-06,
      "loss": 0.22868846356868744,
      "mean_token_accuracy": 0.907988429069519,
      "num_tokens": 3434060201.0,
      "step": 1478
    },
    {
      "epoch": 0.7569089048106449,
      "grad_norm": 0.2158203125,
      "learning_rate": 5.910714547620151e-06,
      "loss": 0.24137243628501892,
      "mean_token_accuracy": 0.9027203321456909,
      "num_tokens": 3436418761.0,
      "step": 1479
    },
    {
      "epoch": 0.7574206755373593,
      "grad_norm": 0.205078125,
      "learning_rate": 5.887201336882689e-06,
      "loss": 0.23513838648796082,
      "mean_token_accuracy": 0.9059600830078125,
      "num_tokens": 3438721847.0,
      "step": 1480
    },
    {
      "epoch": 0.7579324462640737,
      "grad_norm": 0.2109375,
      "learning_rate": 5.863726913875167e-06,
      "loss": 0.227595716714859,
      "mean_token_accuracy": 0.9083376526832581,
      "num_tokens": 3441077714.0,
      "step": 1481
    },
    {
      "epoch": 0.7584442169907881,
      "grad_norm": 0.21484375,
      "learning_rate": 5.84029134311487e-06,
      "loss": 0.23583558201789856,
      "mean_token_accuracy": 0.9048398733139038,
      "num_tokens": 3443384235.0,
      "step": 1482
    },
    {
      "epoch": 0.7589559877175026,
      "grad_norm": 0.2109375,
      "learning_rate": 5.816894689012316e-06,
      "loss": 0.2547760009765625,
      "mean_token_accuracy": 0.8980388641357422,
      "num_tokens": 3445696297.0,
      "step": 1483
    },
    {
      "epoch": 0.759467758444217,
      "grad_norm": 0.197265625,
      "learning_rate": 5.793537015871058e-06,
      "loss": 0.2364497184753418,
      "mean_token_accuracy": 0.9050335884094238,
      "num_tokens": 3448023992.0,
      "step": 1484
    },
    {
      "epoch": 0.7599795291709315,
      "grad_norm": 0.2021484375,
      "learning_rate": 5.77021838788752e-06,
      "loss": 0.22897160053253174,
      "mean_token_accuracy": 0.9074708223342896,
      "num_tokens": 3450381164.0,
      "step": 1485
    },
    {
      "epoch": 0.7604912998976459,
      "grad_norm": 0.2109375,
      "learning_rate": 5.746938869150789e-06,
      "loss": 0.22962792217731476,
      "mean_token_accuracy": 0.907524824142456,
      "num_tokens": 3452706709.0,
      "step": 1486
    },
    {
      "epoch": 0.7610030706243602,
      "grad_norm": 0.20703125,
      "learning_rate": 5.7236985236425005e-06,
      "loss": 0.24416764080524445,
      "mean_token_accuracy": 0.9023318290710449,
      "num_tokens": 3455073955.0,
      "step": 1487
    },
    {
      "epoch": 0.7615148413510747,
      "grad_norm": 0.201171875,
      "learning_rate": 5.700497415236608e-06,
      "loss": 0.23139488697052002,
      "mean_token_accuracy": 0.9070785045623779,
      "num_tokens": 3457385629.0,
      "step": 1488
    },
    {
      "epoch": 0.7620266120777891,
      "grad_norm": 0.2236328125,
      "learning_rate": 5.677335607699215e-06,
      "loss": 0.23588722944259644,
      "mean_token_accuracy": 0.9048452377319336,
      "num_tokens": 3459645473.0,
      "step": 1489
    },
    {
      "epoch": 0.7625383828045036,
      "grad_norm": 0.208984375,
      "learning_rate": 5.654213164688424e-06,
      "loss": 0.23417575657367706,
      "mean_token_accuracy": 0.9061224460601807,
      "num_tokens": 3461913170.0,
      "step": 1490
    },
    {
      "epoch": 0.763050153531218,
      "grad_norm": 0.205078125,
      "learning_rate": 5.631130149754145e-06,
      "loss": 0.2365921139717102,
      "mean_token_accuracy": 0.905544102191925,
      "num_tokens": 3464223664.0,
      "step": 1491
    },
    {
      "epoch": 0.7635619242579325,
      "grad_norm": 0.212890625,
      "learning_rate": 5.6080866263379315e-06,
      "loss": 0.24188300967216492,
      "mean_token_accuracy": 0.9035829901695251,
      "num_tokens": 3466436040.0,
      "step": 1492
    },
    {
      "epoch": 0.7640736949846468,
      "grad_norm": 0.20703125,
      "learning_rate": 5.585082657772769e-06,
      "loss": 0.23448029160499573,
      "mean_token_accuracy": 0.9052103757858276,
      "num_tokens": 3468729935.0,
      "step": 1493
    },
    {
      "epoch": 0.7645854657113613,
      "grad_norm": 0.2158203125,
      "learning_rate": 5.562118307282968e-06,
      "loss": 0.24291197955608368,
      "mean_token_accuracy": 0.9027431607246399,
      "num_tokens": 3471049737.0,
      "step": 1494
    },
    {
      "epoch": 0.7650972364380757,
      "grad_norm": 0.2177734375,
      "learning_rate": 5.539193637983937e-06,
      "loss": 0.2469116598367691,
      "mean_token_accuracy": 0.9007672071456909,
      "num_tokens": 3473291856.0,
      "step": 1495
    },
    {
      "epoch": 0.7656090071647902,
      "grad_norm": 0.2099609375,
      "learning_rate": 5.516308712882011e-06,
      "loss": 0.24126318097114563,
      "mean_token_accuracy": 0.9034344553947449,
      "num_tokens": 3475498353.0,
      "step": 1496
    },
    {
      "epoch": 0.7661207778915046,
      "grad_norm": 0.216796875,
      "learning_rate": 5.493463594874304e-06,
      "loss": 0.24325433373451233,
      "mean_token_accuracy": 0.9025585055351257,
      "num_tokens": 3477741934.0,
      "step": 1497
    },
    {
      "epoch": 0.7666325486182191,
      "grad_norm": 0.1953125,
      "learning_rate": 5.47065834674853e-06,
      "loss": 0.22745537757873535,
      "mean_token_accuracy": 0.908199667930603,
      "num_tokens": 3480100480.0,
      "step": 1498
    },
    {
      "epoch": 0.7671443193449334,
      "grad_norm": 0.216796875,
      "learning_rate": 5.447893031182818e-06,
      "loss": 0.2417261004447937,
      "mean_token_accuracy": 0.9018349647521973,
      "num_tokens": 3482338479.0,
      "step": 1499
    },
    {
      "epoch": 0.7676560900716479,
      "grad_norm": 0.2216796875,
      "learning_rate": 5.425167710745538e-06,
      "loss": 0.25646334886550903,
      "mean_token_accuracy": 0.8970831632614136,
      "num_tokens": 3484677192.0,
      "step": 1500
    },
    {
      "epoch": 0.7681678607983623,
      "grad_norm": 0.2138671875,
      "learning_rate": 5.402482447895138e-06,
      "loss": 0.2340346872806549,
      "mean_token_accuracy": 0.9054619073867798,
      "num_tokens": 3487020818.0,
      "step": 1501
    },
    {
      "epoch": 0.7686796315250768,
      "grad_norm": 0.208984375,
      "learning_rate": 5.379837304980005e-06,
      "loss": 0.24705086648464203,
      "mean_token_accuracy": 0.900261402130127,
      "num_tokens": 3489359895.0,
      "step": 1502
    },
    {
      "epoch": 0.7691914022517912,
      "grad_norm": 0.208984375,
      "learning_rate": 5.357232344238206e-06,
      "loss": 0.24051588773727417,
      "mean_token_accuracy": 0.9031257629394531,
      "num_tokens": 3491642755.0,
      "step": 1503
    },
    {
      "epoch": 0.7697031729785057,
      "grad_norm": 0.20703125,
      "learning_rate": 5.334667627797407e-06,
      "loss": 0.2377435863018036,
      "mean_token_accuracy": 0.9044411182403564,
      "num_tokens": 3493921296.0,
      "step": 1504
    },
    {
      "epoch": 0.77021494370522,
      "grad_norm": 0.20703125,
      "learning_rate": 5.312143217674659e-06,
      "loss": 0.22924938797950745,
      "mean_token_accuracy": 0.9083636999130249,
      "num_tokens": 3496209805.0,
      "step": 1505
    },
    {
      "epoch": 0.7707267144319345,
      "grad_norm": 0.203125,
      "learning_rate": 5.289659175776236e-06,
      "loss": 0.23915940523147583,
      "mean_token_accuracy": 0.9045486450195312,
      "num_tokens": 3498496994.0,
      "step": 1506
    },
    {
      "epoch": 0.7712384851586489,
      "grad_norm": 0.2080078125,
      "learning_rate": 5.26721556389745e-06,
      "loss": 0.23650911450386047,
      "mean_token_accuracy": 0.9052979946136475,
      "num_tokens": 3500870818.0,
      "step": 1507
    },
    {
      "epoch": 0.7717502558853634,
      "grad_norm": 0.205078125,
      "learning_rate": 5.244812443722507e-06,
      "loss": 0.23056916892528534,
      "mean_token_accuracy": 0.9065139889717102,
      "num_tokens": 3503191574.0,
      "step": 1508
    },
    {
      "epoch": 0.7722620266120778,
      "grad_norm": 0.2158203125,
      "learning_rate": 5.222449876824338e-06,
      "loss": 0.24914701282978058,
      "mean_token_accuracy": 0.9006263017654419,
      "num_tokens": 3505514066.0,
      "step": 1509
    },
    {
      "epoch": 0.7727737973387923,
      "grad_norm": 0.205078125,
      "learning_rate": 5.200127924664387e-06,
      "loss": 0.23518598079681396,
      "mean_token_accuracy": 0.9055591225624084,
      "num_tokens": 3507845251.0,
      "step": 1510
    },
    {
      "epoch": 0.7732855680655066,
      "grad_norm": 0.203125,
      "learning_rate": 5.177846648592492e-06,
      "loss": 0.24178743362426758,
      "mean_token_accuracy": 0.9025528430938721,
      "num_tokens": 3510044204.0,
      "step": 1511
    },
    {
      "epoch": 0.7737973387922211,
      "grad_norm": 0.2177734375,
      "learning_rate": 5.155606109846696e-06,
      "loss": 0.23651574552059174,
      "mean_token_accuracy": 0.9051347970962524,
      "num_tokens": 3512378946.0,
      "step": 1512
    },
    {
      "epoch": 0.7743091095189355,
      "grad_norm": 0.2041015625,
      "learning_rate": 5.133406369553078e-06,
      "loss": 0.23321270942687988,
      "mean_token_accuracy": 0.9056087732315063,
      "num_tokens": 3514729168.0,
      "step": 1513
    },
    {
      "epoch": 0.77482088024565,
      "grad_norm": 0.208984375,
      "learning_rate": 5.111247488725571e-06,
      "loss": 0.2378334105014801,
      "mean_token_accuracy": 0.9043527245521545,
      "num_tokens": 3517043832.0,
      "step": 1514
    },
    {
      "epoch": 0.7753326509723644,
      "grad_norm": 0.2197265625,
      "learning_rate": 5.089129528265831e-06,
      "loss": 0.24403595924377441,
      "mean_token_accuracy": 0.901374101638794,
      "num_tokens": 3519266245.0,
      "step": 1515
    },
    {
      "epoch": 0.7758444216990789,
      "grad_norm": 0.212890625,
      "learning_rate": 5.067052548963039e-06,
      "loss": 0.23769521713256836,
      "mean_token_accuracy": 0.9045549631118774,
      "num_tokens": 3521673336.0,
      "step": 1516
    },
    {
      "epoch": 0.7763561924257932,
      "grad_norm": 0.2138671875,
      "learning_rate": 5.045016611493741e-06,
      "loss": 0.24627208709716797,
      "mean_token_accuracy": 0.9019386172294617,
      "num_tokens": 3523926681.0,
      "step": 1517
    },
    {
      "epoch": 0.7768679631525077,
      "grad_norm": 0.20703125,
      "learning_rate": 5.023021776421688e-06,
      "loss": 0.23794905841350555,
      "mean_token_accuracy": 0.9044660329818726,
      "num_tokens": 3526191977.0,
      "step": 1518
    },
    {
      "epoch": 0.7773797338792221,
      "grad_norm": 0.2197265625,
      "learning_rate": 5.001068104197662e-06,
      "loss": 0.2643579840660095,
      "mean_token_accuracy": 0.8947698473930359,
      "num_tokens": 3528498812.0,
      "step": 1519
    },
    {
      "epoch": 0.7778915046059366,
      "grad_norm": 0.2109375,
      "learning_rate": 4.979155655159318e-06,
      "loss": 0.24964302778244019,
      "mean_token_accuracy": 0.9005386829376221,
      "num_tokens": 3530771349.0,
      "step": 1520
    },
    {
      "epoch": 0.778403275332651,
      "grad_norm": 0.20703125,
      "learning_rate": 4.957284489530998e-06,
      "loss": 0.2241249531507492,
      "mean_token_accuracy": 0.9096237421035767,
      "num_tokens": 3533129352.0,
      "step": 1521
    },
    {
      "epoch": 0.7789150460593655,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.935454667423596e-06,
      "loss": 0.2314164936542511,
      "mean_token_accuracy": 0.9073408842086792,
      "num_tokens": 3535545877.0,
      "step": 1522
    },
    {
      "epoch": 0.7794268167860798,
      "grad_norm": 0.2080078125,
      "learning_rate": 4.91366624883437e-06,
      "loss": 0.25201737880706787,
      "mean_token_accuracy": 0.8986425995826721,
      "num_tokens": 3537911809.0,
      "step": 1523
    },
    {
      "epoch": 0.7799385875127943,
      "grad_norm": 0.2138671875,
      "learning_rate": 4.891919293646788e-06,
      "loss": 0.2428901195526123,
      "mean_token_accuracy": 0.903080403804779,
      "num_tokens": 3540242326.0,
      "step": 1524
    },
    {
      "epoch": 0.7804503582395087,
      "grad_norm": 0.205078125,
      "learning_rate": 4.870213861630355e-06,
      "loss": 0.23086829483509064,
      "mean_token_accuracy": 0.907194197177887,
      "num_tokens": 3542650330.0,
      "step": 1525
    },
    {
      "epoch": 0.7809621289662231,
      "grad_norm": 0.2138671875,
      "learning_rate": 4.848550012440456e-06,
      "loss": 0.24243557453155518,
      "mean_token_accuracy": 0.9027186632156372,
      "num_tokens": 3545009477.0,
      "step": 1526
    },
    {
      "epoch": 0.7814738996929376,
      "grad_norm": 0.21875,
      "learning_rate": 4.826927805618198e-06,
      "loss": 0.2459157258272171,
      "mean_token_accuracy": 0.9016135931015015,
      "num_tokens": 3547341164.0,
      "step": 1527
    },
    {
      "epoch": 0.781985670419652,
      "grad_norm": 0.203125,
      "learning_rate": 4.805347300590217e-06,
      "loss": 0.22939318418502808,
      "mean_token_accuracy": 0.9083825349807739,
      "num_tokens": 3549672335.0,
      "step": 1528
    },
    {
      "epoch": 0.7824974411463664,
      "grad_norm": 0.2158203125,
      "learning_rate": 4.7838085566685495e-06,
      "loss": 0.23353661596775055,
      "mean_token_accuracy": 0.9062701463699341,
      "num_tokens": 3551924911.0,
      "step": 1529
    },
    {
      "epoch": 0.7830092118730808,
      "grad_norm": 0.208984375,
      "learning_rate": 4.762311633050456e-06,
      "loss": 0.25534456968307495,
      "mean_token_accuracy": 0.8973588347434998,
      "num_tokens": 3554195983.0,
      "step": 1530
    },
    {
      "epoch": 0.7835209825997953,
      "grad_norm": 0.2001953125,
      "learning_rate": 4.740856588818257e-06,
      "loss": 0.2343323975801468,
      "mean_token_accuracy": 0.9053955078125,
      "num_tokens": 3556533028.0,
      "step": 1531
    },
    {
      "epoch": 0.7840327533265097,
      "grad_norm": 0.203125,
      "learning_rate": 4.719443482939166e-06,
      "loss": 0.24788135290145874,
      "mean_token_accuracy": 0.9004840850830078,
      "num_tokens": 3558813851.0,
      "step": 1532
    },
    {
      "epoch": 0.7845445240532242,
      "grad_norm": 0.203125,
      "learning_rate": 4.6980723742651365e-06,
      "loss": 0.23409909009933472,
      "mean_token_accuracy": 0.906121551990509,
      "num_tokens": 3561145487.0,
      "step": 1533
    },
    {
      "epoch": 0.7850562947799385,
      "grad_norm": 0.203125,
      "learning_rate": 4.676743321532706e-06,
      "loss": 0.2351270318031311,
      "mean_token_accuracy": 0.9052785634994507,
      "num_tokens": 3563524522.0,
      "step": 1534
    },
    {
      "epoch": 0.785568065506653,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.6554563833628e-06,
      "loss": 0.23396769165992737,
      "mean_token_accuracy": 0.9062046408653259,
      "num_tokens": 3565778668.0,
      "step": 1535
    },
    {
      "epoch": 0.7860798362333674,
      "grad_norm": 0.2109375,
      "learning_rate": 4.634211618260618e-06,
      "loss": 0.23863154649734497,
      "mean_token_accuracy": 0.9040769934654236,
      "num_tokens": 3568148362.0,
      "step": 1536
    },
    {
      "epoch": 0.7865916069600819,
      "grad_norm": 0.2177734375,
      "learning_rate": 4.613009084615445e-06,
      "loss": 0.23575536906719208,
      "mean_token_accuracy": 0.9049249887466431,
      "num_tokens": 3570476069.0,
      "step": 1537
    },
    {
      "epoch": 0.7871033776867963,
      "grad_norm": 0.212890625,
      "learning_rate": 4.591848840700497e-06,
      "loss": 0.2392391562461853,
      "mean_token_accuracy": 0.9039806723594666,
      "num_tokens": 3572719844.0,
      "step": 1538
    },
    {
      "epoch": 0.7876151484135108,
      "grad_norm": 0.2099609375,
      "learning_rate": 4.5707309446727565e-06,
      "loss": 0.25053662061691284,
      "mean_token_accuracy": 0.8994553089141846,
      "num_tokens": 3574994449.0,
      "step": 1539
    },
    {
      "epoch": 0.7881269191402251,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.5496554545728215e-06,
      "loss": 0.23649203777313232,
      "mean_token_accuracy": 0.9048459529876709,
      "num_tokens": 3577252033.0,
      "step": 1540
    },
    {
      "epoch": 0.7886386898669396,
      "grad_norm": 0.208984375,
      "learning_rate": 4.5286224283247446e-06,
      "loss": 0.24663949012756348,
      "mean_token_accuracy": 0.9010806083679199,
      "num_tokens": 3579540874.0,
      "step": 1541
    },
    {
      "epoch": 0.789150460593654,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.507631923735854e-06,
      "loss": 0.23850780725479126,
      "mean_token_accuracy": 0.9035087823867798,
      "num_tokens": 3581805028.0,
      "step": 1542
    },
    {
      "epoch": 0.7896622313203685,
      "grad_norm": 0.2177734375,
      "learning_rate": 4.4866839984966285e-06,
      "loss": 0.24341434240341187,
      "mean_token_accuracy": 0.9021118879318237,
      "num_tokens": 3584048041.0,
      "step": 1543
    },
    {
      "epoch": 0.7901740020470829,
      "grad_norm": 0.212890625,
      "learning_rate": 4.465778710180513e-06,
      "loss": 0.2445431351661682,
      "mean_token_accuracy": 0.9016284942626953,
      "num_tokens": 3586366966.0,
      "step": 1544
    },
    {
      "epoch": 0.7906857727737974,
      "grad_norm": 0.208984375,
      "learning_rate": 4.444916116243778e-06,
      "loss": 0.2412737011909485,
      "mean_token_accuracy": 0.9037530422210693,
      "num_tokens": 3588735938.0,
      "step": 1545
    },
    {
      "epoch": 0.7911975435005117,
      "grad_norm": 0.2041015625,
      "learning_rate": 4.424096274025331e-06,
      "loss": 0.24709485471248627,
      "mean_token_accuracy": 0.9012623429298401,
      "num_tokens": 3591101938.0,
      "step": 1546
    },
    {
      "epoch": 0.7917093142272262,
      "grad_norm": 0.205078125,
      "learning_rate": 4.403319240746606e-06,
      "loss": 0.24166277050971985,
      "mean_token_accuracy": 0.9023483991622925,
      "num_tokens": 3593362758.0,
      "step": 1547
    },
    {
      "epoch": 0.7922210849539406,
      "grad_norm": 0.205078125,
      "learning_rate": 4.382585073511374e-06,
      "loss": 0.23581963777542114,
      "mean_token_accuracy": 0.9051092863082886,
      "num_tokens": 3595759048.0,
      "step": 1548
    },
    {
      "epoch": 0.7927328556806551,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.361893829305575e-06,
      "loss": 0.24852445721626282,
      "mean_token_accuracy": 0.9006964564323425,
      "num_tokens": 3598064803.0,
      "step": 1549
    },
    {
      "epoch": 0.7932446264073695,
      "grad_norm": 0.216796875,
      "learning_rate": 4.341245564997196e-06,
      "loss": 0.24950459599494934,
      "mean_token_accuracy": 0.900267481803894,
      "num_tokens": 3600462968.0,
      "step": 1550
    },
    {
      "epoch": 0.793756397134084,
      "grad_norm": 0.201171875,
      "learning_rate": 4.3206403373360975e-06,
      "loss": 0.2246776819229126,
      "mean_token_accuracy": 0.9102871417999268,
      "num_tokens": 3602769023.0,
      "step": 1551
    },
    {
      "epoch": 0.7942681678607983,
      "grad_norm": 0.203125,
      "learning_rate": 4.300078202953855e-06,
      "loss": 0.23132693767547607,
      "mean_token_accuracy": 0.906851589679718,
      "num_tokens": 3605112779.0,
      "step": 1552
    },
    {
      "epoch": 0.7947799385875128,
      "grad_norm": 0.2109375,
      "learning_rate": 4.2795592183635895e-06,
      "loss": 0.254802942276001,
      "mean_token_accuracy": 0.8983184099197388,
      "num_tokens": 3607424444.0,
      "step": 1553
    },
    {
      "epoch": 0.7952917093142272,
      "grad_norm": 0.2099609375,
      "learning_rate": 4.259083439959861e-06,
      "loss": 0.24731895327568054,
      "mean_token_accuracy": 0.9010987281799316,
      "num_tokens": 3609687532.0,
      "step": 1554
    },
    {
      "epoch": 0.7958034800409417,
      "grad_norm": 0.220703125,
      "learning_rate": 4.238650924018466e-06,
      "loss": 0.24537138640880585,
      "mean_token_accuracy": 0.9014303088188171,
      "num_tokens": 3611938775.0,
      "step": 1555
    },
    {
      "epoch": 0.7963152507676561,
      "grad_norm": 0.205078125,
      "learning_rate": 4.2182617266962825e-06,
      "loss": 0.23160792887210846,
      "mean_token_accuracy": 0.9070290923118591,
      "num_tokens": 3614303763.0,
      "step": 1556
    },
    {
      "epoch": 0.7968270214943706,
      "grad_norm": 0.1982421875,
      "learning_rate": 4.197915904031153e-06,
      "loss": 0.2322225570678711,
      "mean_token_accuracy": 0.9064878225326538,
      "num_tokens": 3616683344.0,
      "step": 1557
    },
    {
      "epoch": 0.797338792221085,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.177613511941701e-06,
      "loss": 0.2521055340766907,
      "mean_token_accuracy": 0.8994864225387573,
      "num_tokens": 3619018046.0,
      "step": 1558
    },
    {
      "epoch": 0.7978505629477994,
      "grad_norm": 0.2177734375,
      "learning_rate": 4.157354606227189e-06,
      "loss": 0.24327819049358368,
      "mean_token_accuracy": 0.9023374319076538,
      "num_tokens": 3621314200.0,
      "step": 1559
    },
    {
      "epoch": 0.7983623336745138,
      "grad_norm": 0.19921875,
      "learning_rate": 4.137139242567347e-06,
      "loss": 0.23265421390533447,
      "mean_token_accuracy": 0.9064282178878784,
      "num_tokens": 3623686091.0,
      "step": 1560
    },
    {
      "epoch": 0.7988741044012283,
      "grad_norm": 0.201171875,
      "learning_rate": 4.116967476522245e-06,
      "loss": 0.23006278276443481,
      "mean_token_accuracy": 0.9078693389892578,
      "num_tokens": 3626074352.0,
      "step": 1561
    },
    {
      "epoch": 0.7993858751279427,
      "grad_norm": 0.20703125,
      "learning_rate": 4.096839363532141e-06,
      "loss": 0.24099129438400269,
      "mean_token_accuracy": 0.9010360836982727,
      "num_tokens": 3628408258.0,
      "step": 1562
    },
    {
      "epoch": 0.7998976458546572,
      "grad_norm": 0.2109375,
      "learning_rate": 4.076754958917293e-06,
      "loss": 0.24903997778892517,
      "mean_token_accuracy": 0.8998479247093201,
      "num_tokens": 3630707357.0,
      "step": 1563
    },
    {
      "epoch": 0.8004094165813715,
      "grad_norm": 0.21484375,
      "learning_rate": 4.056714317877839e-06,
      "loss": 0.24168553948402405,
      "mean_token_accuracy": 0.903470516204834,
      "num_tokens": 3633005828.0,
      "step": 1564
    },
    {
      "epoch": 0.8009211873080859,
      "grad_norm": 0.21875,
      "learning_rate": 4.036717495493645e-06,
      "loss": 0.2370963841676712,
      "mean_token_accuracy": 0.904360294342041,
      "num_tokens": 3635364282.0,
      "step": 1565
    },
    {
      "epoch": 0.8014329580348004,
      "grad_norm": 0.20703125,
      "learning_rate": 4.016764546724141e-06,
      "loss": 0.22969384491443634,
      "mean_token_accuracy": 0.9073910713195801,
      "num_tokens": 3637716058.0,
      "step": 1566
    },
    {
      "epoch": 0.8019447287615148,
      "grad_norm": 0.203125,
      "learning_rate": 3.996855526408165e-06,
      "loss": 0.24585890769958496,
      "mean_token_accuracy": 0.9010179042816162,
      "num_tokens": 3639991713.0,
      "step": 1567
    },
    {
      "epoch": 0.8024564994882293,
      "grad_norm": 0.1962890625,
      "learning_rate": 3.976990489263828e-06,
      "loss": 0.2335507571697235,
      "mean_token_accuracy": 0.906141459941864,
      "num_tokens": 3642356168.0,
      "step": 1568
    },
    {
      "epoch": 0.8029682702149437,
      "grad_norm": 0.21875,
      "learning_rate": 3.957169489888373e-06,
      "loss": 0.25550952553749084,
      "mean_token_accuracy": 0.8971536159515381,
      "num_tokens": 3644615567.0,
      "step": 1569
    },
    {
      "epoch": 0.8034800409416581,
      "grad_norm": 0.21484375,
      "learning_rate": 3.9373925827579795e-06,
      "loss": 0.23803305625915527,
      "mean_token_accuracy": 0.9043164253234863,
      "num_tokens": 3646940051.0,
      "step": 1570
    },
    {
      "epoch": 0.8039918116683725,
      "grad_norm": 0.21484375,
      "learning_rate": 3.917659822227666e-06,
      "loss": 0.23781761527061462,
      "mean_token_accuracy": 0.9044642448425293,
      "num_tokens": 3649249585.0,
      "step": 1571
    },
    {
      "epoch": 0.804503582395087,
      "grad_norm": 0.2041015625,
      "learning_rate": 3.897971262531111e-06,
      "loss": 0.23150736093521118,
      "mean_token_accuracy": 0.9072446823120117,
      "num_tokens": 3651568866.0,
      "step": 1572
    },
    {
      "epoch": 0.8050153531218014,
      "grad_norm": 0.2060546875,
      "learning_rate": 3.878326957780514e-06,
      "loss": 0.23785394430160522,
      "mean_token_accuracy": 0.9040374755859375,
      "num_tokens": 3653972412.0,
      "step": 1573
    },
    {
      "epoch": 0.8055271238485159,
      "grad_norm": 0.216796875,
      "learning_rate": 3.858726961966432e-06,
      "loss": 0.2556914687156677,
      "mean_token_accuracy": 0.8971824049949646,
      "num_tokens": 3656347771.0,
      "step": 1574
    },
    {
      "epoch": 0.8060388945752303,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.83917132895766e-06,
      "loss": 0.24157454073429108,
      "mean_token_accuracy": 0.9034042358398438,
      "num_tokens": 3658636805.0,
      "step": 1575
    },
    {
      "epoch": 0.8065506653019447,
      "grad_norm": 0.224609375,
      "learning_rate": 3.819660112501053e-06,
      "loss": 0.2501552700996399,
      "mean_token_accuracy": 0.9001556634902954,
      "num_tokens": 3660942901.0,
      "step": 1576
    },
    {
      "epoch": 0.8070624360286591,
      "grad_norm": 0.201171875,
      "learning_rate": 3.8001933662213987e-06,
      "loss": 0.22934730350971222,
      "mean_token_accuracy": 0.9074376225471497,
      "num_tokens": 3663343736.0,
      "step": 1577
    },
    {
      "epoch": 0.8075742067553736,
      "grad_norm": 0.220703125,
      "learning_rate": 3.780771143621258e-06,
      "loss": 0.24903255701065063,
      "mean_token_accuracy": 0.9002914428710938,
      "num_tokens": 3665694923.0,
      "step": 1578
    },
    {
      "epoch": 0.808085977482088,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.7613934980808277e-06,
      "loss": 0.2322985827922821,
      "mean_token_accuracy": 0.9071157574653625,
      "num_tokens": 3667994200.0,
      "step": 1579
    },
    {
      "epoch": 0.8085977482088025,
      "grad_norm": 0.1943359375,
      "learning_rate": 3.742060482857788e-06,
      "loss": 0.22572113573551178,
      "mean_token_accuracy": 0.9094648361206055,
      "num_tokens": 3670286842.0,
      "step": 1580
    },
    {
      "epoch": 0.8091095189355169,
      "grad_norm": 0.201171875,
      "learning_rate": 3.722772151087144e-06,
      "loss": 0.23458947241306305,
      "mean_token_accuracy": 0.9061301350593567,
      "num_tokens": 3672577879.0,
      "step": 1581
    },
    {
      "epoch": 0.8096212896622313,
      "grad_norm": 0.203125,
      "learning_rate": 3.7035285557811106e-06,
      "loss": 0.24048560857772827,
      "mean_token_accuracy": 0.903782308101654,
      "num_tokens": 3674950693.0,
      "step": 1582
    },
    {
      "epoch": 0.8101330603889457,
      "grad_norm": 0.205078125,
      "learning_rate": 3.6843297498289364e-06,
      "loss": 0.24792324006557465,
      "mean_token_accuracy": 0.9006071090698242,
      "num_tokens": 3677269215.0,
      "step": 1583
    },
    {
      "epoch": 0.8106448311156602,
      "grad_norm": 0.20703125,
      "learning_rate": 3.6651757859967797e-06,
      "loss": 0.23056018352508545,
      "mean_token_accuracy": 0.9079487919807434,
      "num_tokens": 3679610038.0,
      "step": 1584
    },
    {
      "epoch": 0.8111566018423746,
      "grad_norm": 0.201171875,
      "learning_rate": 3.646066716927543e-06,
      "loss": 0.23635540902614594,
      "mean_token_accuracy": 0.9056214690208435,
      "num_tokens": 3681909179.0,
      "step": 1585
    },
    {
      "epoch": 0.8116683725690891,
      "grad_norm": 0.189453125,
      "learning_rate": 3.627002595140754e-06,
      "loss": 0.227906733751297,
      "mean_token_accuracy": 0.9078863859176636,
      "num_tokens": 3684268106.0,
      "step": 1586
    },
    {
      "epoch": 0.8121801432958035,
      "grad_norm": 0.216796875,
      "learning_rate": 3.607983473032399e-06,
      "loss": 0.2495579719543457,
      "mean_token_accuracy": 0.8993854522705078,
      "num_tokens": 3686592616.0,
      "step": 1587
    },
    {
      "epoch": 0.812691914022518,
      "grad_norm": 0.21484375,
      "learning_rate": 3.589009402874779e-06,
      "loss": 0.25757190585136414,
      "mean_token_accuracy": 0.8962908983230591,
      "num_tokens": 3688856334.0,
      "step": 1588
    },
    {
      "epoch": 0.8132036847492323,
      "grad_norm": 0.2041015625,
      "learning_rate": 3.570080436816383e-06,
      "loss": 0.2448480725288391,
      "mean_token_accuracy": 0.9018843770027161,
      "num_tokens": 3691198913.0,
      "step": 1589
    },
    {
      "epoch": 0.8137154554759468,
      "grad_norm": 0.2041015625,
      "learning_rate": 3.551196626881741e-06,
      "loss": 0.23913419246673584,
      "mean_token_accuracy": 0.9043669700622559,
      "num_tokens": 3693596601.0,
      "step": 1590
    },
    {
      "epoch": 0.8142272262026612,
      "grad_norm": 0.201171875,
      "learning_rate": 3.532358024971265e-06,
      "loss": 0.22717821598052979,
      "mean_token_accuracy": 0.9083351492881775,
      "num_tokens": 3695827201.0,
      "step": 1591
    },
    {
      "epoch": 0.8147389969293757,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.5135646828611192e-06,
      "loss": 0.22754395008087158,
      "mean_token_accuracy": 0.9085190892219543,
      "num_tokens": 3698268144.0,
      "step": 1592
    },
    {
      "epoch": 0.8152507676560901,
      "grad_norm": 0.2109375,
      "learning_rate": 3.494816652203079e-06,
      "loss": 0.24814772605895996,
      "mean_token_accuracy": 0.9005944132804871,
      "num_tokens": 3700662000.0,
      "step": 1593
    },
    {
      "epoch": 0.8157625383828045,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.4761139845243896e-06,
      "loss": 0.237248033285141,
      "mean_token_accuracy": 0.9048383235931396,
      "num_tokens": 3702974313.0,
      "step": 1594
    },
    {
      "epoch": 0.8162743091095189,
      "grad_norm": 0.212890625,
      "learning_rate": 3.4574567312276018e-06,
      "loss": 0.24407298862934113,
      "mean_token_accuracy": 0.9022369384765625,
      "num_tokens": 3705231833.0,
      "step": 1595
    },
    {
      "epoch": 0.8167860798362334,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.4388449435904693e-06,
      "loss": 0.22898811101913452,
      "mean_token_accuracy": 0.9078034162521362,
      "num_tokens": 3707591071.0,
      "step": 1596
    },
    {
      "epoch": 0.8172978505629478,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.4202786727657756e-06,
      "loss": 0.2309764325618744,
      "mean_token_accuracy": 0.9058740139007568,
      "num_tokens": 3709882310.0,
      "step": 1597
    },
    {
      "epoch": 0.8178096212896623,
      "grad_norm": 0.2109375,
      "learning_rate": 3.401757969781212e-06,
      "loss": 0.23748385906219482,
      "mean_token_accuracy": 0.9050218462944031,
      "num_tokens": 3712217832.0,
      "step": 1598
    },
    {
      "epoch": 0.8183213920163767,
      "grad_norm": 0.20703125,
      "learning_rate": 3.3832828855392296e-06,
      "loss": 0.2378753423690796,
      "mean_token_accuracy": 0.9056394100189209,
      "num_tokens": 3714572721.0,
      "step": 1599
    },
    {
      "epoch": 0.8188331627430911,
      "grad_norm": 0.2158203125,
      "learning_rate": 3.3648534708168957e-06,
      "loss": 0.23782376945018768,
      "mean_token_accuracy": 0.9052440524101257,
      "num_tokens": 3716851960.0,
      "step": 1600
    },
    {
      "epoch": 0.8193449334698055,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.3464697762657707e-06,
      "loss": 0.23778709769248962,
      "mean_token_accuracy": 0.904093861579895,
      "num_tokens": 3719208891.0,
      "step": 1601
    },
    {
      "epoch": 0.81985670419652,
      "grad_norm": 0.220703125,
      "learning_rate": 3.328131852411738e-06,
      "loss": 0.24655070900917053,
      "mean_token_accuracy": 0.9008020758628845,
      "num_tokens": 3721513994.0,
      "step": 1602
    },
    {
      "epoch": 0.8203684749232344,
      "grad_norm": 0.2060546875,
      "learning_rate": 3.309839749654902e-06,
      "loss": 0.2513682246208191,
      "mean_token_accuracy": 0.8991979360580444,
      "num_tokens": 3723832638.0,
      "step": 1603
    },
    {
      "epoch": 0.8208802456499488,
      "grad_norm": 0.21875,
      "learning_rate": 3.2915935182694246e-06,
      "loss": 0.2483924776315689,
      "mean_token_accuracy": 0.8998574018478394,
      "num_tokens": 3726131055.0,
      "step": 1604
    },
    {
      "epoch": 0.8213920163766633,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.2733932084033992e-06,
      "loss": 0.24295374751091003,
      "mean_token_accuracy": 0.9021866917610168,
      "num_tokens": 3728397337.0,
      "step": 1605
    },
    {
      "epoch": 0.8219037871033776,
      "grad_norm": 0.212890625,
      "learning_rate": 3.255238870078692e-06,
      "loss": 0.24717354774475098,
      "mean_token_accuracy": 0.9009761810302734,
      "num_tokens": 3730752474.0,
      "step": 1606
    },
    {
      "epoch": 0.8224155578300921,
      "grad_norm": 0.2041015625,
      "learning_rate": 3.237130553190848e-06,
      "loss": 0.24535490572452545,
      "mean_token_accuracy": 0.9016485810279846,
      "num_tokens": 3733191604.0,
      "step": 1607
    },
    {
      "epoch": 0.8229273285568065,
      "grad_norm": 0.201171875,
      "learning_rate": 3.2190683075089124e-06,
      "loss": 0.24841918051242828,
      "mean_token_accuracy": 0.8998646140098572,
      "num_tokens": 3735512809.0,
      "step": 1608
    },
    {
      "epoch": 0.823439099283521,
      "grad_norm": 0.21484375,
      "learning_rate": 3.201052182675295e-06,
      "loss": 0.2358197271823883,
      "mean_token_accuracy": 0.9052588939666748,
      "num_tokens": 3737693544.0,
      "step": 1609
    },
    {
      "epoch": 0.8239508700102354,
      "grad_norm": 0.212890625,
      "learning_rate": 3.183082228205665e-06,
      "loss": 0.23444458842277527,
      "mean_token_accuracy": 0.9063079953193665,
      "num_tokens": 3739984248.0,
      "step": 1610
    },
    {
      "epoch": 0.8244626407369499,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.165158493488789e-06,
      "loss": 0.2434852123260498,
      "mean_token_accuracy": 0.9024927020072937,
      "num_tokens": 3742250717.0,
      "step": 1611
    },
    {
      "epoch": 0.8249744114636642,
      "grad_norm": 0.22265625,
      "learning_rate": 3.147281027786413e-06,
      "loss": 0.2515159845352173,
      "mean_token_accuracy": 0.8992050290107727,
      "num_tokens": 3744633971.0,
      "step": 1612
    },
    {
      "epoch": 0.8254861821903787,
      "grad_norm": 0.20703125,
      "learning_rate": 3.129449880233091e-06,
      "loss": 0.2388344705104828,
      "mean_token_accuracy": 0.9039435386657715,
      "num_tokens": 3746949059.0,
      "step": 1613
    },
    {
      "epoch": 0.8259979529170931,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.1116650998361074e-06,
      "loss": 0.22920507192611694,
      "mean_token_accuracy": 0.9077956080436707,
      "num_tokens": 3749291515.0,
      "step": 1614
    },
    {
      "epoch": 0.8265097236438076,
      "grad_norm": 0.1982421875,
      "learning_rate": 3.093926735475299e-06,
      "loss": 0.23465125262737274,
      "mean_token_accuracy": 0.9055113792419434,
      "num_tokens": 3751652802.0,
      "step": 1615
    },
    {
      "epoch": 0.827021494370522,
      "grad_norm": 0.20703125,
      "learning_rate": 3.076234835902916e-06,
      "loss": 0.24503272771835327,
      "mean_token_accuracy": 0.9022438526153564,
      "num_tokens": 3754014829.0,
      "step": 1616
    },
    {
      "epoch": 0.8275332650972365,
      "grad_norm": 0.203125,
      "learning_rate": 3.0585894497435276e-06,
      "loss": 0.22923770546913147,
      "mean_token_accuracy": 0.9080894589424133,
      "num_tokens": 3756382284.0,
      "step": 1617
    },
    {
      "epoch": 0.8280450358239508,
      "grad_norm": 0.2021484375,
      "learning_rate": 3.0409906254938515e-06,
      "loss": 0.24037735164165497,
      "mean_token_accuracy": 0.9037420749664307,
      "num_tokens": 3758641812.0,
      "step": 1618
    },
    {
      "epoch": 0.8285568065506653,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.0234384115226454e-06,
      "loss": 0.24109932780265808,
      "mean_token_accuracy": 0.9029752016067505,
      "num_tokens": 3760938291.0,
      "step": 1619
    },
    {
      "epoch": 0.8290685772773797,
      "grad_norm": 0.2138671875,
      "learning_rate": 3.005932856070548e-06,
      "loss": 0.24192117154598236,
      "mean_token_accuracy": 0.9036969542503357,
      "num_tokens": 3763185448.0,
      "step": 1620
    },
    {
      "epoch": 0.8295803480040942,
      "grad_norm": 0.208984375,
      "learning_rate": 2.9884740072499663e-06,
      "loss": 0.24086302518844604,
      "mean_token_accuracy": 0.9035009741783142,
      "num_tokens": 3765502788.0,
      "step": 1621
    },
    {
      "epoch": 0.8300921187308086,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.9710619130449593e-06,
      "loss": 0.2418001890182495,
      "mean_token_accuracy": 0.9034029841423035,
      "num_tokens": 3767840673.0,
      "step": 1622
    },
    {
      "epoch": 0.8306038894575231,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.9536966213110484e-06,
      "loss": 0.23301640152931213,
      "mean_token_accuracy": 0.9067903757095337,
      "num_tokens": 3770179399.0,
      "step": 1623
    },
    {
      "epoch": 0.8311156601842374,
      "grad_norm": 0.1982421875,
      "learning_rate": 2.936378179775152e-06,
      "loss": 0.23583662509918213,
      "mean_token_accuracy": 0.9048396348953247,
      "num_tokens": 3772479919.0,
      "step": 1624
    },
    {
      "epoch": 0.8316274309109519,
      "grad_norm": 0.21484375,
      "learning_rate": 2.9191066360354116e-06,
      "loss": 0.2516755759716034,
      "mean_token_accuracy": 0.8995091915130615,
      "num_tokens": 3774710488.0,
      "step": 1625
    },
    {
      "epoch": 0.8321392016376663,
      "grad_norm": 0.205078125,
      "learning_rate": 2.9018820375610836e-06,
      "loss": 0.23344296216964722,
      "mean_token_accuracy": 0.90610671043396,
      "num_tokens": 3777011905.0,
      "step": 1626
    },
    {
      "epoch": 0.8326509723643808,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.8847044316923823e-06,
      "loss": 0.24298813939094543,
      "mean_token_accuracy": 0.9018492698669434,
      "num_tokens": 3779290788.0,
      "step": 1627
    },
    {
      "epoch": 0.8331627430910952,
      "grad_norm": 0.21484375,
      "learning_rate": 2.8675738656403805e-06,
      "loss": 0.24271036684513092,
      "mean_token_accuracy": 0.9027586579322815,
      "num_tokens": 3781530663.0,
      "step": 1628
    },
    {
      "epoch": 0.8336745138178097,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.8504903864868772e-06,
      "loss": 0.23833243548870087,
      "mean_token_accuracy": 0.9037951827049255,
      "num_tokens": 3783878233.0,
      "step": 1629
    },
    {
      "epoch": 0.834186284544524,
      "grad_norm": 0.205078125,
      "learning_rate": 2.8334540411842336e-06,
      "loss": 0.22805550694465637,
      "mean_token_accuracy": 0.9075877666473389,
      "num_tokens": 3786168258.0,
      "step": 1630
    },
    {
      "epoch": 0.8346980552712385,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.8164648765552806e-06,
      "loss": 0.24164406955242157,
      "mean_token_accuracy": 0.9037020206451416,
      "num_tokens": 3788501621.0,
      "step": 1631
    },
    {
      "epoch": 0.8352098259979529,
      "grad_norm": 0.203125,
      "learning_rate": 2.7995229392931733e-06,
      "loss": 0.22387836873531342,
      "mean_token_accuracy": 0.9103869199752808,
      "num_tokens": 3790830450.0,
      "step": 1632
    },
    {
      "epoch": 0.8357215967246674,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.7826282759612787e-06,
      "loss": 0.23120032250881195,
      "mean_token_accuracy": 0.9068602323532104,
      "num_tokens": 3793220977.0,
      "step": 1633
    },
    {
      "epoch": 0.8362333674513818,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.7657809329930162e-06,
      "loss": 0.2498580515384674,
      "mean_token_accuracy": 0.89909827709198,
      "num_tokens": 3795615205.0,
      "step": 1634
    },
    {
      "epoch": 0.8367451381780963,
      "grad_norm": 0.208984375,
      "learning_rate": 2.74898095669176e-06,
      "loss": 0.23680391907691956,
      "mean_token_accuracy": 0.9050635099411011,
      "num_tokens": 3797900792.0,
      "step": 1635
    },
    {
      "epoch": 0.8372569089048106,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.732228393230705e-06,
      "loss": 0.23092803359031677,
      "mean_token_accuracy": 0.9069736003875732,
      "num_tokens": 3800270500.0,
      "step": 1636
    },
    {
      "epoch": 0.8377686796315251,
      "grad_norm": 0.20703125,
      "learning_rate": 2.7155232886527307e-06,
      "loss": 0.22776909172534943,
      "mean_token_accuracy": 0.9088363647460938,
      "num_tokens": 3802618943.0,
      "step": 1637
    },
    {
      "epoch": 0.8382804503582395,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.6988656888702804e-06,
      "loss": 0.22597147524356842,
      "mean_token_accuracy": 0.9090520143508911,
      "num_tokens": 3804872249.0,
      "step": 1638
    },
    {
      "epoch": 0.838792221084954,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.6822556396652366e-06,
      "loss": 0.24293166399002075,
      "mean_token_accuracy": 0.9027612209320068,
      "num_tokens": 3807183726.0,
      "step": 1639
    },
    {
      "epoch": 0.8393039918116684,
      "grad_norm": 0.208984375,
      "learning_rate": 2.665693186688798e-06,
      "loss": 0.2280910164117813,
      "mean_token_accuracy": 0.9081389904022217,
      "num_tokens": 3809558704.0,
      "step": 1640
    },
    {
      "epoch": 0.8398157625383829,
      "grad_norm": 0.205078125,
      "learning_rate": 2.6491783754613365e-06,
      "loss": 0.2280825823545456,
      "mean_token_accuracy": 0.9079378843307495,
      "num_tokens": 3811900374.0,
      "step": 1641
    },
    {
      "epoch": 0.8403275332650972,
      "grad_norm": 0.208984375,
      "learning_rate": 2.632711251372293e-06,
      "loss": 0.2422739565372467,
      "mean_token_accuracy": 0.9026845693588257,
      "num_tokens": 3814128057.0,
      "step": 1642
    },
    {
      "epoch": 0.8408393039918116,
      "grad_norm": 0.212890625,
      "learning_rate": 2.616291859680049e-06,
      "loss": 0.2505464553833008,
      "mean_token_accuracy": 0.8996585607528687,
      "num_tokens": 3816355263.0,
      "step": 1643
    },
    {
      "epoch": 0.8413510747185261,
      "grad_norm": 0.203125,
      "learning_rate": 2.599920245511789e-06,
      "loss": 0.23582160472869873,
      "mean_token_accuracy": 0.9055758714675903,
      "num_tokens": 3818681373.0,
      "step": 1644
    },
    {
      "epoch": 0.8418628454452405,
      "grad_norm": 0.20703125,
      "learning_rate": 2.583596453863395e-06,
      "loss": 0.24285855889320374,
      "mean_token_accuracy": 0.902846097946167,
      "num_tokens": 3820962592.0,
      "step": 1645
    },
    {
      "epoch": 0.842374616171955,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.5673205295993044e-06,
      "loss": 0.24714215099811554,
      "mean_token_accuracy": 0.900741696357727,
      "num_tokens": 3823266275.0,
      "step": 1646
    },
    {
      "epoch": 0.8428863868986693,
      "grad_norm": 0.20703125,
      "learning_rate": 2.5510925174524027e-06,
      "loss": 0.2364179790019989,
      "mean_token_accuracy": 0.9044582843780518,
      "num_tokens": 3825631145.0,
      "step": 1647
    },
    {
      "epoch": 0.8433981576253838,
      "grad_norm": 0.20703125,
      "learning_rate": 2.534912462023884e-06,
      "loss": 0.23628106713294983,
      "mean_token_accuracy": 0.9055162668228149,
      "num_tokens": 3827973456.0,
      "step": 1648
    },
    {
      "epoch": 0.8439099283520982,
      "grad_norm": 0.1884765625,
      "learning_rate": 2.5187804077831456e-06,
      "loss": 0.21700745820999146,
      "mean_token_accuracy": 0.9131079316139221,
      "num_tokens": 3830340168.0,
      "step": 1649
    },
    {
      "epoch": 0.8444216990788127,
      "grad_norm": 0.216796875,
      "learning_rate": 2.5026963990676567e-06,
      "loss": 0.24574947357177734,
      "mean_token_accuracy": 0.900434136390686,
      "num_tokens": 3832611418.0,
      "step": 1650
    },
    {
      "epoch": 0.8449334698055271,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.4866604800828344e-06,
      "loss": 0.2428516447544098,
      "mean_token_accuracy": 0.9019399881362915,
      "num_tokens": 3834957573.0,
      "step": 1651
    },
    {
      "epoch": 0.8454452405322416,
      "grad_norm": 0.201171875,
      "learning_rate": 2.4706726949019323e-06,
      "loss": 0.24797488749027252,
      "mean_token_accuracy": 0.9005900621414185,
      "num_tokens": 3837275840.0,
      "step": 1652
    },
    {
      "epoch": 0.845957011258956,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.454733087465906e-06,
      "loss": 0.24074354767799377,
      "mean_token_accuracy": 0.9032886624336243,
      "num_tokens": 3839677148.0,
      "step": 1653
    },
    {
      "epoch": 0.8464687819856704,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.4388417015833057e-06,
      "loss": 0.23312735557556152,
      "mean_token_accuracy": 0.9060136079788208,
      "num_tokens": 3841982142.0,
      "step": 1654
    },
    {
      "epoch": 0.8469805527123848,
      "grad_norm": 0.20703125,
      "learning_rate": 2.422998580930138e-06,
      "loss": 0.24044573307037354,
      "mean_token_accuracy": 0.9014824628829956,
      "num_tokens": 3844304990.0,
      "step": 1655
    },
    {
      "epoch": 0.8474923234390993,
      "grad_norm": 0.20703125,
      "learning_rate": 2.407203769049766e-06,
      "loss": 0.23355062305927277,
      "mean_token_accuracy": 0.9065085649490356,
      "num_tokens": 3846599327.0,
      "step": 1656
    },
    {
      "epoch": 0.8480040941658137,
      "grad_norm": 0.21484375,
      "learning_rate": 2.391457309352778e-06,
      "loss": 0.22606658935546875,
      "mean_token_accuracy": 0.9085984230041504,
      "num_tokens": 3848844253.0,
      "step": 1657
    },
    {
      "epoch": 0.8485158648925282,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.3757592451168732e-06,
      "loss": 0.2389550656080246,
      "mean_token_accuracy": 0.9039148092269897,
      "num_tokens": 3851156841.0,
      "step": 1658
    },
    {
      "epoch": 0.8490276356192425,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.360109619486737e-06,
      "loss": 0.2314838469028473,
      "mean_token_accuracy": 0.9068527221679688,
      "num_tokens": 3853478972.0,
      "step": 1659
    },
    {
      "epoch": 0.849539406345957,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.3445084754739277e-06,
      "loss": 0.24113894999027252,
      "mean_token_accuracy": 0.9030579328536987,
      "num_tokens": 3855679211.0,
      "step": 1660
    },
    {
      "epoch": 0.8500511770726714,
      "grad_norm": 0.2177734375,
      "learning_rate": 2.328955855956758e-06,
      "loss": 0.23001277446746826,
      "mean_token_accuracy": 0.9071686267852783,
      "num_tokens": 3858071448.0,
      "step": 1661
    },
    {
      "epoch": 0.8505629477993859,
      "grad_norm": 0.2021484375,
      "learning_rate": 2.313451803680169e-06,
      "loss": 0.23918521404266357,
      "mean_token_accuracy": 0.9050883650779724,
      "num_tokens": 3860446540.0,
      "step": 1662
    },
    {
      "epoch": 0.8510747185261003,
      "grad_norm": 0.212890625,
      "learning_rate": 2.297996361255621e-06,
      "loss": 0.23880332708358765,
      "mean_token_accuracy": 0.9033644199371338,
      "num_tokens": 3862733600.0,
      "step": 1663
    },
    {
      "epoch": 0.8515864892528148,
      "grad_norm": 0.203125,
      "learning_rate": 2.2825895711609824e-06,
      "loss": 0.23089128732681274,
      "mean_token_accuracy": 0.907694399356842,
      "num_tokens": 3865070495.0,
      "step": 1664
    },
    {
      "epoch": 0.8520982599795291,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.2672314757404034e-06,
      "loss": 0.24228185415267944,
      "mean_token_accuracy": 0.902880072593689,
      "num_tokens": 3867446735.0,
      "step": 1665
    },
    {
      "epoch": 0.8526100307062436,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.2519221172041816e-06,
      "loss": 0.24188882112503052,
      "mean_token_accuracy": 0.9031442403793335,
      "num_tokens": 3869780149.0,
      "step": 1666
    },
    {
      "epoch": 0.853121801432958,
      "grad_norm": 0.201171875,
      "learning_rate": 2.236661537628697e-06,
      "loss": 0.23764187097549438,
      "mean_token_accuracy": 0.9039548635482788,
      "num_tokens": 3872129646.0,
      "step": 1667
    },
    {
      "epoch": 0.8536335721596725,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.2214497789562485e-06,
      "loss": 0.23780623078346252,
      "mean_token_accuracy": 0.9042505621910095,
      "num_tokens": 3874488326.0,
      "step": 1668
    },
    {
      "epoch": 0.8541453428863869,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.2062868829949478e-06,
      "loss": 0.22719678282737732,
      "mean_token_accuracy": 0.9086159467697144,
      "num_tokens": 3876822873.0,
      "step": 1669
    },
    {
      "epoch": 0.8546571136131014,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.1911728914186294e-06,
      "loss": 0.2517358362674713,
      "mean_token_accuracy": 0.8997058868408203,
      "num_tokens": 3879153410.0,
      "step": 1670
    },
    {
      "epoch": 0.8551688843398157,
      "grad_norm": 0.2109375,
      "learning_rate": 2.176107845766704e-06,
      "loss": 0.22537535429000854,
      "mean_token_accuracy": 0.909585177898407,
      "num_tokens": 3881396126.0,
      "step": 1671
    },
    {
      "epoch": 0.8556806550665302,
      "grad_norm": 0.20703125,
      "learning_rate": 2.1610917874440717e-06,
      "loss": 0.23936589062213898,
      "mean_token_accuracy": 0.9030795097351074,
      "num_tokens": 3883684742.0,
      "step": 1672
    },
    {
      "epoch": 0.8561924257932446,
      "grad_norm": 0.2119140625,
      "learning_rate": 2.146124757720982e-06,
      "loss": 0.2361324429512024,
      "mean_token_accuracy": 0.9054278135299683,
      "num_tokens": 3885983201.0,
      "step": 1673
    },
    {
      "epoch": 0.8567041965199591,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.1312067977329364e-06,
      "loss": 0.24453355371952057,
      "mean_token_accuracy": 0.9005939960479736,
      "num_tokens": 3888302169.0,
      "step": 1674
    },
    {
      "epoch": 0.8572159672466735,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.1163379484805933e-06,
      "loss": 0.2474808245897293,
      "mean_token_accuracy": 0.9013611078262329,
      "num_tokens": 3890744159.0,
      "step": 1675
    },
    {
      "epoch": 0.857727737973388,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.101518250829602e-06,
      "loss": 0.2363223433494568,
      "mean_token_accuracy": 0.9047167301177979,
      "num_tokens": 3893042279.0,
      "step": 1676
    },
    {
      "epoch": 0.8582395087001023,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.0867477455105467e-06,
      "loss": 0.2427113950252533,
      "mean_token_accuracy": 0.9020309448242188,
      "num_tokens": 3895395354.0,
      "step": 1677
    },
    {
      "epoch": 0.8587512794268168,
      "grad_norm": 0.2158203125,
      "learning_rate": 2.0720264731188023e-06,
      "loss": 0.24348124861717224,
      "mean_token_accuracy": 0.9021406173706055,
      "num_tokens": 3897719959.0,
      "step": 1678
    },
    {
      "epoch": 0.8592630501535312,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.057354474114435e-06,
      "loss": 0.24091975390911102,
      "mean_token_accuracy": 0.9031281471252441,
      "num_tokens": 3900101609.0,
      "step": 1679
    },
    {
      "epoch": 0.8597748208802457,
      "grad_norm": 0.205078125,
      "learning_rate": 2.042731788822081e-06,
      "loss": 0.23068687319755554,
      "mean_token_accuracy": 0.9069765210151672,
      "num_tokens": 3902416723.0,
      "step": 1680
    },
    {
      "epoch": 0.8602865916069601,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.0281584574308445e-06,
      "loss": 0.2462964802980423,
      "mean_token_accuracy": 0.9008943438529968,
      "num_tokens": 3904720378.0,
      "step": 1681
    },
    {
      "epoch": 0.8607983623336745,
      "grad_norm": 0.2001953125,
      "learning_rate": 2.0136345199941986e-06,
      "loss": 0.23467101156711578,
      "mean_token_accuracy": 0.905407726764679,
      "num_tokens": 3907049686.0,
      "step": 1682
    },
    {
      "epoch": 0.861310133060389,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.9991600164298418e-06,
      "loss": 0.2385522872209549,
      "mean_token_accuracy": 0.904025137424469,
      "num_tokens": 3909336177.0,
      "step": 1683
    },
    {
      "epoch": 0.8618219037871033,
      "grad_norm": 0.2109375,
      "learning_rate": 1.9847349865196165e-06,
      "loss": 0.24262794852256775,
      "mean_token_accuracy": 0.9032038450241089,
      "num_tokens": 3911627258.0,
      "step": 1684
    },
    {
      "epoch": 0.8623336745138178,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.9703594699093954e-06,
      "loss": 0.233658105134964,
      "mean_token_accuracy": 0.906119704246521,
      "num_tokens": 3913940857.0,
      "step": 1685
    },
    {
      "epoch": 0.8628454452405322,
      "grad_norm": 0.201171875,
      "learning_rate": 1.956033506108965e-06,
      "loss": 0.22655510902404785,
      "mean_token_accuracy": 0.9091532230377197,
      "num_tokens": 3916160263.0,
      "step": 1686
    },
    {
      "epoch": 0.8633572159672467,
      "grad_norm": 0.212890625,
      "learning_rate": 1.941757134491915e-06,
      "loss": 0.23523655533790588,
      "mean_token_accuracy": 0.9050002098083496,
      "num_tokens": 3918429724.0,
      "step": 1687
    },
    {
      "epoch": 0.8638689866939611,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.9275303942955426e-06,
      "loss": 0.24023278057575226,
      "mean_token_accuracy": 0.9028614163398743,
      "num_tokens": 3920724413.0,
      "step": 1688
    },
    {
      "epoch": 0.8643807574206755,
      "grad_norm": 0.216796875,
      "learning_rate": 1.913353324620735e-06,
      "loss": 0.24519719183444977,
      "mean_token_accuracy": 0.9014844298362732,
      "num_tokens": 3923088837.0,
      "step": 1689
    },
    {
      "epoch": 0.8648925281473899,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.8992259644318677e-06,
      "loss": 0.23995667695999146,
      "mean_token_accuracy": 0.9037377834320068,
      "num_tokens": 3925410032.0,
      "step": 1690
    },
    {
      "epoch": 0.8654042988741044,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.8851483525566893e-06,
      "loss": 0.22993138432502747,
      "mean_token_accuracy": 0.9076539874076843,
      "num_tokens": 3927704817.0,
      "step": 1691
    },
    {
      "epoch": 0.8659160696008188,
      "grad_norm": 0.1962890625,
      "learning_rate": 1.8711205276862254e-06,
      "loss": 0.22202640771865845,
      "mean_token_accuracy": 0.9096530675888062,
      "num_tokens": 3930022152.0,
      "step": 1692
    },
    {
      "epoch": 0.8664278403275333,
      "grad_norm": 0.1943359375,
      "learning_rate": 1.8571425283746624e-06,
      "loss": 0.23909637331962585,
      "mean_token_accuracy": 0.9041479825973511,
      "num_tokens": 3932349605.0,
      "step": 1693
    },
    {
      "epoch": 0.8669396110542477,
      "grad_norm": 0.201171875,
      "learning_rate": 1.8432143930392432e-06,
      "loss": 0.23716169595718384,
      "mean_token_accuracy": 0.904295802116394,
      "num_tokens": 3934691712.0,
      "step": 1694
    },
    {
      "epoch": 0.8674513817809621,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.8293361599601667e-06,
      "loss": 0.24437931180000305,
      "mean_token_accuracy": 0.9016886949539185,
      "num_tokens": 3937049841.0,
      "step": 1695
    },
    {
      "epoch": 0.8679631525076765,
      "grad_norm": 0.220703125,
      "learning_rate": 1.815507867280486e-06,
      "loss": 0.23723554611206055,
      "mean_token_accuracy": 0.9048513770103455,
      "num_tokens": 3939341430.0,
      "step": 1696
    },
    {
      "epoch": 0.868474923234391,
      "grad_norm": 0.212890625,
      "learning_rate": 1.8017295530059908e-06,
      "loss": 0.2323664128780365,
      "mean_token_accuracy": 0.9059238433837891,
      "num_tokens": 3941572150.0,
      "step": 1697
    },
    {
      "epoch": 0.8689866939611054,
      "grad_norm": 0.208984375,
      "learning_rate": 1.7880012550051096e-06,
      "loss": 0.2372923195362091,
      "mean_token_accuracy": 0.9039558172225952,
      "num_tokens": 3943866351.0,
      "step": 1698
    },
    {
      "epoch": 0.8694984646878199,
      "grad_norm": 0.203125,
      "learning_rate": 1.7743230110088094e-06,
      "loss": 0.23580823838710785,
      "mean_token_accuracy": 0.9054337739944458,
      "num_tokens": 3946207364.0,
      "step": 1699
    },
    {
      "epoch": 0.8700102354145343,
      "grad_norm": 0.20703125,
      "learning_rate": 1.7606948586104877e-06,
      "loss": 0.23092412948608398,
      "mean_token_accuracy": 0.9077162146568298,
      "num_tokens": 3948605019.0,
      "step": 1700
    },
    {
      "epoch": 0.8705220061412487,
      "grad_norm": 0.20703125,
      "learning_rate": 1.7471168352658652e-06,
      "loss": 0.2377311885356903,
      "mean_token_accuracy": 0.9047372341156006,
      "num_tokens": 3950919708.0,
      "step": 1701
    },
    {
      "epoch": 0.8710337768679631,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.7335889782928928e-06,
      "loss": 0.23976588249206543,
      "mean_token_accuracy": 0.9035115838050842,
      "num_tokens": 3953238980.0,
      "step": 1702
    },
    {
      "epoch": 0.8715455475946776,
      "grad_norm": 0.205078125,
      "learning_rate": 1.7201113248716427e-06,
      "loss": 0.2263752520084381,
      "mean_token_accuracy": 0.908869743347168,
      "num_tokens": 3955605497.0,
      "step": 1703
    },
    {
      "epoch": 0.872057318321392,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.7066839120442047e-06,
      "loss": 0.242161363363266,
      "mean_token_accuracy": 0.9020880460739136,
      "num_tokens": 3957834257.0,
      "step": 1704
    },
    {
      "epoch": 0.8725690890481065,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.6933067767145894e-06,
      "loss": 0.24617214500904083,
      "mean_token_accuracy": 0.9008675217628479,
      "num_tokens": 3960097316.0,
      "step": 1705
    },
    {
      "epoch": 0.8730808597748209,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.6799799556486253e-06,
      "loss": 0.2434813231229782,
      "mean_token_accuracy": 0.9022713303565979,
      "num_tokens": 3962345199.0,
      "step": 1706
    },
    {
      "epoch": 0.8735926305015353,
      "grad_norm": 0.22265625,
      "learning_rate": 1.6667034854738572e-06,
      "loss": 0.24644508957862854,
      "mean_token_accuracy": 0.9005682468414307,
      "num_tokens": 3964598759.0,
      "step": 1707
    },
    {
      "epoch": 0.8741044012282497,
      "grad_norm": 0.212890625,
      "learning_rate": 1.6534774026794353e-06,
      "loss": 0.248079314827919,
      "mean_token_accuracy": 0.9012099504470825,
      "num_tokens": 3966904559.0,
      "step": 1708
    },
    {
      "epoch": 0.8746161719549642,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.6403017436160395e-06,
      "loss": 0.2530434727668762,
      "mean_token_accuracy": 0.8994832038879395,
      "num_tokens": 3969245556.0,
      "step": 1709
    },
    {
      "epoch": 0.8751279426816786,
      "grad_norm": 0.197265625,
      "learning_rate": 1.6271765444957543e-06,
      "loss": 0.22800755500793457,
      "mean_token_accuracy": 0.9083436131477356,
      "num_tokens": 3971612959.0,
      "step": 1710
    },
    {
      "epoch": 0.8756397134083931,
      "grad_norm": 0.19921875,
      "learning_rate": 1.614101841391984e-06,
      "loss": 0.2203054428100586,
      "mean_token_accuracy": 0.9119775295257568,
      "num_tokens": 3973948547.0,
      "step": 1711
    },
    {
      "epoch": 0.8761514841351075,
      "grad_norm": 0.205078125,
      "learning_rate": 1.6010776702393483e-06,
      "loss": 0.2350594401359558,
      "mean_token_accuracy": 0.905418872833252,
      "num_tokens": 3976218420.0,
      "step": 1712
    },
    {
      "epoch": 0.876663254861822,
      "grad_norm": 0.21484375,
      "learning_rate": 1.5881040668335846e-06,
      "loss": 0.2382453978061676,
      "mean_token_accuracy": 0.9039679765701294,
      "num_tokens": 3978542995.0,
      "step": 1713
    },
    {
      "epoch": 0.8771750255885363,
      "grad_norm": 0.205078125,
      "learning_rate": 1.5751810668314504e-06,
      "loss": 0.2359733134508133,
      "mean_token_accuracy": 0.9051816463470459,
      "num_tokens": 3980834295.0,
      "step": 1714
    },
    {
      "epoch": 0.8776867963152508,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.5623087057506215e-06,
      "loss": 0.23910388350486755,
      "mean_token_accuracy": 0.9034892916679382,
      "num_tokens": 3983224711.0,
      "step": 1715
    },
    {
      "epoch": 0.8781985670419652,
      "grad_norm": 0.2177734375,
      "learning_rate": 1.5494870189695997e-06,
      "loss": 0.24414414167404175,
      "mean_token_accuracy": 0.9022042751312256,
      "num_tokens": 3985562405.0,
      "step": 1716
    },
    {
      "epoch": 0.8787103377686797,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.5367160417276128e-06,
      "loss": 0.24116411805152893,
      "mean_token_accuracy": 0.9035447835922241,
      "num_tokens": 3987862900.0,
      "step": 1717
    },
    {
      "epoch": 0.8792221084953941,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.5239958091245211e-06,
      "loss": 0.24182429909706116,
      "mean_token_accuracy": 0.9024230241775513,
      "num_tokens": 3990147435.0,
      "step": 1718
    },
    {
      "epoch": 0.8797338792221086,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.511326356120706e-06,
      "loss": 0.24073803424835205,
      "mean_token_accuracy": 0.9037317037582397,
      "num_tokens": 3992458183.0,
      "step": 1719
    },
    {
      "epoch": 0.8802456499488229,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.4987077175370025e-06,
      "loss": 0.23993617296218872,
      "mean_token_accuracy": 0.9040924906730652,
      "num_tokens": 3994784791.0,
      "step": 1720
    },
    {
      "epoch": 0.8807574206755373,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.4861399280545797e-06,
      "loss": 0.244277685880661,
      "mean_token_accuracy": 0.9020311236381531,
      "num_tokens": 3997074499.0,
      "step": 1721
    },
    {
      "epoch": 0.8812691914022518,
      "grad_norm": 0.203125,
      "learning_rate": 1.473623022214845e-06,
      "loss": 0.23243603110313416,
      "mean_token_accuracy": 0.9065407514572144,
      "num_tokens": 3999389053.0,
      "step": 1722
    },
    {
      "epoch": 0.8817809621289662,
      "grad_norm": 0.205078125,
      "learning_rate": 1.4611570344193672e-06,
      "loss": 0.22941100597381592,
      "mean_token_accuracy": 0.9079477190971375,
      "num_tokens": 4001765727.0,
      "step": 1723
    },
    {
      "epoch": 0.8822927328556807,
      "grad_norm": 0.21875,
      "learning_rate": 1.448741998929768e-06,
      "loss": 0.2406134009361267,
      "mean_token_accuracy": 0.9037305116653442,
      "num_tokens": 4004086564.0,
      "step": 1724
    },
    {
      "epoch": 0.882804503582395,
      "grad_norm": 0.2109375,
      "learning_rate": 1.4363779498676356e-06,
      "loss": 0.24752014875411987,
      "mean_token_accuracy": 0.9010401964187622,
      "num_tokens": 4006417615.0,
      "step": 1725
    },
    {
      "epoch": 0.8833162743091095,
      "grad_norm": 0.205078125,
      "learning_rate": 1.4240649212144098e-06,
      "loss": 0.24073445796966553,
      "mean_token_accuracy": 0.9031964540481567,
      "num_tokens": 4008675338.0,
      "step": 1726
    },
    {
      "epoch": 0.8838280450358239,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.4118029468113314e-06,
      "loss": 0.22726912796497345,
      "mean_token_accuracy": 0.9091330766677856,
      "num_tokens": 4010969602.0,
      "step": 1727
    },
    {
      "epoch": 0.8843398157625384,
      "grad_norm": 0.208984375,
      "learning_rate": 1.3995920603593073e-06,
      "loss": 0.23983415961265564,
      "mean_token_accuracy": 0.9037268161773682,
      "num_tokens": 4013272198.0,
      "step": 1728
    },
    {
      "epoch": 0.8848515864892528,
      "grad_norm": 0.2109375,
      "learning_rate": 1.3874322954188313e-06,
      "loss": 0.2507326602935791,
      "mean_token_accuracy": 0.8998594880104065,
      "num_tokens": 4015562224.0,
      "step": 1729
    },
    {
      "epoch": 0.8853633572159673,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.3753236854099017e-06,
      "loss": 0.2427779734134674,
      "mean_token_accuracy": 0.9024338722229004,
      "num_tokens": 4017797154.0,
      "step": 1730
    },
    {
      "epoch": 0.8858751279426816,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.363266263611922e-06,
      "loss": 0.2431245893239975,
      "mean_token_accuracy": 0.9031453132629395,
      "num_tokens": 4020126611.0,
      "step": 1731
    },
    {
      "epoch": 0.8863868986693961,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.3512600631636086e-06,
      "loss": 0.2402227222919464,
      "mean_token_accuracy": 0.9029800891876221,
      "num_tokens": 4022469165.0,
      "step": 1732
    },
    {
      "epoch": 0.8868986693961105,
      "grad_norm": 0.2265625,
      "learning_rate": 1.3393051170628967e-06,
      "loss": 0.2458072304725647,
      "mean_token_accuracy": 0.9013087749481201,
      "num_tokens": 4024721795.0,
      "step": 1733
    },
    {
      "epoch": 0.887410440122825,
      "grad_norm": 0.203125,
      "learning_rate": 1.3274014581668549e-06,
      "loss": 0.2266545444726944,
      "mean_token_accuracy": 0.9089148640632629,
      "num_tokens": 4027031640.0,
      "step": 1734
    },
    {
      "epoch": 0.8879222108495394,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.3155491191916036e-06,
      "loss": 0.24884209036827087,
      "mean_token_accuracy": 0.9006580710411072,
      "num_tokens": 4029344639.0,
      "step": 1735
    },
    {
      "epoch": 0.8884339815762539,
      "grad_norm": 0.216796875,
      "learning_rate": 1.3037481327121993e-06,
      "loss": 0.2381162941455841,
      "mean_token_accuracy": 0.9042055010795593,
      "num_tokens": 4031658688.0,
      "step": 1736
    },
    {
      "epoch": 0.8889457523029682,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.2919985311625727e-06,
      "loss": 0.24220749735832214,
      "mean_token_accuracy": 0.9022798538208008,
      "num_tokens": 4033878062.0,
      "step": 1737
    },
    {
      "epoch": 0.8894575230296827,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.2803003468354214e-06,
      "loss": 0.23370875418186188,
      "mean_token_accuracy": 0.9059466123580933,
      "num_tokens": 4036179289.0,
      "step": 1738
    },
    {
      "epoch": 0.8899692937563971,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.2686536118821336e-06,
      "loss": 0.245024636387825,
      "mean_token_accuracy": 0.9015016555786133,
      "num_tokens": 4038506469.0,
      "step": 1739
    },
    {
      "epoch": 0.8904810644831116,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.2570583583126862e-06,
      "loss": 0.24145953357219696,
      "mean_token_accuracy": 0.902708888053894,
      "num_tokens": 4040915775.0,
      "step": 1740
    },
    {
      "epoch": 0.890992835209826,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.2455146179955625e-06,
      "loss": 0.24414116144180298,
      "mean_token_accuracy": 0.9012308120727539,
      "num_tokens": 4043198302.0,
      "step": 1741
    },
    {
      "epoch": 0.8915046059365405,
      "grad_norm": 0.208984375,
      "learning_rate": 1.2340224226576857e-06,
      "loss": 0.23732039332389832,
      "mean_token_accuracy": 0.9042455554008484,
      "num_tokens": 4045480847.0,
      "step": 1742
    },
    {
      "epoch": 0.8920163766632548,
      "grad_norm": 0.228515625,
      "learning_rate": 1.2225818038842885e-06,
      "loss": 0.2524119019508362,
      "mean_token_accuracy": 0.8986272811889648,
      "num_tokens": 4047809345.0,
      "step": 1743
    },
    {
      "epoch": 0.8925281473899693,
      "grad_norm": 0.220703125,
      "learning_rate": 1.2111927931188605e-06,
      "loss": 0.2433866560459137,
      "mean_token_accuracy": 0.9019815921783447,
      "num_tokens": 4050111898.0,
      "step": 1744
    },
    {
      "epoch": 0.8930399181166837,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.1998554216630543e-06,
      "loss": 0.24553222954273224,
      "mean_token_accuracy": 0.9014458656311035,
      "num_tokens": 4052385044.0,
      "step": 1745
    },
    {
      "epoch": 0.8935516888433982,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.1885697206765934e-06,
      "loss": 0.24411624670028687,
      "mean_token_accuracy": 0.9024089574813843,
      "num_tokens": 4054801479.0,
      "step": 1746
    },
    {
      "epoch": 0.8940634595701126,
      "grad_norm": 0.22265625,
      "learning_rate": 1.1773357211771864e-06,
      "loss": 0.2572208642959595,
      "mean_token_accuracy": 0.8967999219894409,
      "num_tokens": 4057173003.0,
      "step": 1747
    },
    {
      "epoch": 0.8945752302968271,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.1661534540404484e-06,
      "loss": 0.24531105160713196,
      "mean_token_accuracy": 0.9011081457138062,
      "num_tokens": 4059510177.0,
      "step": 1748
    },
    {
      "epoch": 0.8950870010235414,
      "grad_norm": 0.2265625,
      "learning_rate": 1.1550229499998134e-06,
      "loss": 0.2557903230190277,
      "mean_token_accuracy": 0.8978636264801025,
      "num_tokens": 4061714218.0,
      "step": 1749
    },
    {
      "epoch": 0.8955987717502559,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.1439442396464507e-06,
      "loss": 0.24562552571296692,
      "mean_token_accuracy": 0.9011756181716919,
      "num_tokens": 4064204114.0,
      "step": 1750
    },
    {
      "epoch": 0.8961105424769703,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.1329173534291771e-06,
      "loss": 0.2353825867176056,
      "mean_token_accuracy": 0.904625654220581,
      "num_tokens": 4066509557.0,
      "step": 1751
    },
    {
      "epoch": 0.8966223132036848,
      "grad_norm": 0.1943359375,
      "learning_rate": 1.1219423216543722e-06,
      "loss": 0.23026278614997864,
      "mean_token_accuracy": 0.9077019691467285,
      "num_tokens": 4068881422.0,
      "step": 1752
    },
    {
      "epoch": 0.8971340839303992,
      "grad_norm": 0.205078125,
      "learning_rate": 1.1110191744859101e-06,
      "loss": 0.22593189775943756,
      "mean_token_accuracy": 0.909376859664917,
      "num_tokens": 4071249748.0,
      "step": 1753
    },
    {
      "epoch": 0.8976458546571137,
      "grad_norm": 0.2138671875,
      "learning_rate": 1.1001479419450467e-06,
      "loss": 0.23497574031352997,
      "mean_token_accuracy": 0.9056342840194702,
      "num_tokens": 4073567953.0,
      "step": 1754
    },
    {
      "epoch": 0.898157625383828,
      "grad_norm": 0.205078125,
      "learning_rate": 1.0893286539103709e-06,
      "loss": 0.24017499387264252,
      "mean_token_accuracy": 0.9021708965301514,
      "num_tokens": 4075874569.0,
      "step": 1755
    },
    {
      "epoch": 0.8986693961105425,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.0785613401177008e-06,
      "loss": 0.24595317244529724,
      "mean_token_accuracy": 0.9015190601348877,
      "num_tokens": 4078259696.0,
      "step": 1756
    },
    {
      "epoch": 0.8991811668372569,
      "grad_norm": 0.2216796875,
      "learning_rate": 1.067846030160009e-06,
      "loss": 0.24744781851768494,
      "mean_token_accuracy": 0.9006550312042236,
      "num_tokens": 4080601025.0,
      "step": 1757
    },
    {
      "epoch": 0.8996929375639714,
      "grad_norm": 0.19921875,
      "learning_rate": 1.057182753487338e-06,
      "loss": 0.2387484908103943,
      "mean_token_accuracy": 0.9037823677062988,
      "num_tokens": 4083017004.0,
      "step": 1758
    },
    {
      "epoch": 0.9002047082906858,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.0465715394067266e-06,
      "loss": 0.23855194449424744,
      "mean_token_accuracy": 0.9034793376922607,
      "num_tokens": 4085372062.0,
      "step": 1759
    },
    {
      "epoch": 0.9007164790174002,
      "grad_norm": 0.203125,
      "learning_rate": 1.0360124170821217e-06,
      "loss": 0.24589216709136963,
      "mean_token_accuracy": 0.9006357789039612,
      "num_tokens": 4087681700.0,
      "step": 1760
    },
    {
      "epoch": 0.9012282497441146,
      "grad_norm": 0.1982421875,
      "learning_rate": 1.025505415534296e-06,
      "loss": 0.2246972769498825,
      "mean_token_accuracy": 0.9096884727478027,
      "num_tokens": 4090058895.0,
      "step": 1761
    },
    {
      "epoch": 0.901740020470829,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.0150505636407804e-06,
      "loss": 0.23806020617485046,
      "mean_token_accuracy": 0.9042937159538269,
      "num_tokens": 4092362739.0,
      "step": 1762
    },
    {
      "epoch": 0.9022517911975435,
      "grad_norm": 0.2080078125,
      "learning_rate": 1.0046478901357726e-06,
      "loss": 0.23852694034576416,
      "mean_token_accuracy": 0.9047539234161377,
      "num_tokens": 4094700612.0,
      "step": 1763
    },
    {
      "epoch": 0.9027635619242579,
      "grad_norm": 0.2001953125,
      "learning_rate": 9.94297423610069e-07,
      "loss": 0.23672278225421906,
      "mean_token_accuracy": 0.904776930809021,
      "num_tokens": 4097008477.0,
      "step": 1764
    },
    {
      "epoch": 0.9032753326509724,
      "grad_norm": 0.2001953125,
      "learning_rate": 9.839991925109693e-07,
      "loss": 0.2272939682006836,
      "mean_token_accuracy": 0.9089174866676331,
      "num_tokens": 4099411282.0,
      "step": 1765
    },
    {
      "epoch": 0.9037871033776868,
      "grad_norm": 0.2099609375,
      "learning_rate": 9.7375322514222e-07,
      "loss": 0.24138498306274414,
      "mean_token_accuracy": 0.90294349193573,
      "num_tokens": 4101763951.0,
      "step": 1766
    },
    {
      "epoch": 0.9042988741044012,
      "grad_norm": 0.203125,
      "learning_rate": 9.635595496639217e-07,
      "loss": 0.2355237901210785,
      "mean_token_accuracy": 0.9055802226066589,
      "num_tokens": 4104106204.0,
      "step": 1767
    },
    {
      "epoch": 0.9048106448311156,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.534181940924502e-07,
      "loss": 0.24767813086509705,
      "mean_token_accuracy": 0.9015418291091919,
      "num_tokens": 4106408901.0,
      "step": 1768
    },
    {
      "epoch": 0.9053224155578301,
      "grad_norm": 0.2099609375,
      "learning_rate": 9.43329186300388e-07,
      "loss": 0.24542298913002014,
      "mean_token_accuracy": 0.9011844992637634,
      "num_tokens": 4108666031.0,
      "step": 1769
    },
    {
      "epoch": 0.9058341862845445,
      "grad_norm": 0.2041015625,
      "learning_rate": 9.332925540164495e-07,
      "loss": 0.23915928602218628,
      "mean_token_accuracy": 0.9043545722961426,
      "num_tokens": 4111016634.0,
      "step": 1770
    },
    {
      "epoch": 0.906345957011259,
      "grad_norm": 0.201171875,
      "learning_rate": 9.233083248253961e-07,
      "loss": 0.22725586593151093,
      "mean_token_accuracy": 0.9080440402030945,
      "num_tokens": 4113365496.0,
      "step": 1771
    },
    {
      "epoch": 0.9068577277379734,
      "grad_norm": 0.2080078125,
      "learning_rate": 9.133765261679616e-07,
      "loss": 0.24865686893463135,
      "mean_token_accuracy": 0.9004112482070923,
      "num_tokens": 4115675864.0,
      "step": 1772
    },
    {
      "epoch": 0.9073694984646878,
      "grad_norm": 0.2138671875,
      "learning_rate": 9.034971853407826e-07,
      "loss": 0.25074058771133423,
      "mean_token_accuracy": 0.8995054960250854,
      "num_tokens": 4118027311.0,
      "step": 1773
    },
    {
      "epoch": 0.9078812691914022,
      "grad_norm": 0.2109375,
      "learning_rate": 8.936703294963234e-07,
      "loss": 0.22573450207710266,
      "mean_token_accuracy": 0.9086707234382629,
      "num_tokens": 4120315646.0,
      "step": 1774
    },
    {
      "epoch": 0.9083930399181167,
      "grad_norm": 0.2021484375,
      "learning_rate": 8.838959856427865e-07,
      "loss": 0.2276623249053955,
      "mean_token_accuracy": 0.9086920022964478,
      "num_tokens": 4122618450.0,
      "step": 1775
    },
    {
      "epoch": 0.9089048106448311,
      "grad_norm": 0.19140625,
      "learning_rate": 8.741741806440629e-07,
      "loss": 0.22084596753120422,
      "mean_token_accuracy": 0.9111152291297913,
      "num_tokens": 4124963858.0,
      "step": 1776
    },
    {
      "epoch": 0.9094165813715456,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.645049412196371e-07,
      "loss": 0.2379627823829651,
      "mean_token_accuracy": 0.904704213142395,
      "num_tokens": 4127324524.0,
      "step": 1777
    },
    {
      "epoch": 0.90992835209826,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.548882939445291e-07,
      "loss": 0.2287905067205429,
      "mean_token_accuracy": 0.9081518650054932,
      "num_tokens": 4129626847.0,
      "step": 1778
    },
    {
      "epoch": 0.9104401228249744,
      "grad_norm": 0.2109375,
      "learning_rate": 8.453242652492011e-07,
      "loss": 0.24042832851409912,
      "mean_token_accuracy": 0.9037703275680542,
      "num_tokens": 4131900414.0,
      "step": 1779
    },
    {
      "epoch": 0.9109518935516888,
      "grad_norm": 0.2001953125,
      "learning_rate": 8.358128814195177e-07,
      "loss": 0.23994745314121246,
      "mean_token_accuracy": 0.9039919376373291,
      "num_tokens": 4134312962.0,
      "step": 1780
    },
    {
      "epoch": 0.9114636642784033,
      "grad_norm": 0.201171875,
      "learning_rate": 8.263541685966392e-07,
      "loss": 0.23516535758972168,
      "mean_token_accuracy": 0.9058958888053894,
      "num_tokens": 4136641906.0,
      "step": 1781
    },
    {
      "epoch": 0.9119754350051177,
      "grad_norm": 0.201171875,
      "learning_rate": 8.169481527769663e-07,
      "loss": 0.23374596238136292,
      "mean_token_accuracy": 0.9061974287033081,
      "num_tokens": 4138977134.0,
      "step": 1782
    },
    {
      "epoch": 0.9124872057318322,
      "grad_norm": 0.2099609375,
      "learning_rate": 8.075948598120687e-07,
      "loss": 0.24684570729732513,
      "mean_token_accuracy": 0.9014596343040466,
      "num_tokens": 4141255074.0,
      "step": 1783
    },
    {
      "epoch": 0.9129989764585466,
      "grad_norm": 0.2041015625,
      "learning_rate": 7.982943154086165e-07,
      "loss": 0.24626663327217102,
      "mean_token_accuracy": 0.9014739394187927,
      "num_tokens": 4143633183.0,
      "step": 1784
    },
    {
      "epoch": 0.913510747185261,
      "grad_norm": 0.20703125,
      "learning_rate": 7.89046545128298e-07,
      "loss": 0.23007969558238983,
      "mean_token_accuracy": 0.9076204299926758,
      "num_tokens": 4145889095.0,
      "step": 1785
    },
    {
      "epoch": 0.9140225179119754,
      "grad_norm": 0.224609375,
      "learning_rate": 7.798515743877577e-07,
      "loss": 0.2621467411518097,
      "mean_token_accuracy": 0.8951679468154907,
      "num_tokens": 4148223817.0,
      "step": 1786
    },
    {
      "epoch": 0.9145342886386899,
      "grad_norm": 0.205078125,
      "learning_rate": 7.707094284585314e-07,
      "loss": 0.23182465136051178,
      "mean_token_accuracy": 0.9062235355377197,
      "num_tokens": 4150543796.0,
      "step": 1787
    },
    {
      "epoch": 0.9150460593654043,
      "grad_norm": 0.201171875,
      "learning_rate": 7.616201324669692e-07,
      "loss": 0.2203071117401123,
      "mean_token_accuracy": 0.9100933074951172,
      "num_tokens": 4152919752.0,
      "step": 1788
    },
    {
      "epoch": 0.9155578300921188,
      "grad_norm": 0.2021484375,
      "learning_rate": 7.525837113941569e-07,
      "loss": 0.22591134905815125,
      "mean_token_accuracy": 0.9089967608451843,
      "num_tokens": 4155197289.0,
      "step": 1789
    },
    {
      "epoch": 0.9160696008188332,
      "grad_norm": 0.2177734375,
      "learning_rate": 7.436001900758749e-07,
      "loss": 0.2431735396385193,
      "mean_token_accuracy": 0.9021673202514648,
      "num_tokens": 4157458396.0,
      "step": 1790
    },
    {
      "epoch": 0.9165813715455476,
      "grad_norm": 0.205078125,
      "learning_rate": 7.346695932025038e-07,
      "loss": 0.22726017236709595,
      "mean_token_accuracy": 0.90886390209198,
      "num_tokens": 4159804139.0,
      "step": 1791
    },
    {
      "epoch": 0.917093142272262,
      "grad_norm": 0.2138671875,
      "learning_rate": 7.257919453189677e-07,
      "loss": 0.23983271420001984,
      "mean_token_accuracy": 0.9042177796363831,
      "num_tokens": 4162131544.0,
      "step": 1792
    },
    {
      "epoch": 0.9176049129989765,
      "grad_norm": 0.216796875,
      "learning_rate": 7.169672708246644e-07,
      "loss": 0.24487905204296112,
      "mean_token_accuracy": 0.9007759094238281,
      "num_tokens": 4164420860.0,
      "step": 1793
    },
    {
      "epoch": 0.9181166837256909,
      "grad_norm": 0.2138671875,
      "learning_rate": 7.08195593973402e-07,
      "loss": 0.24046722054481506,
      "mean_token_accuracy": 0.9035676717758179,
      "num_tokens": 4166881977.0,
      "step": 1794
    },
    {
      "epoch": 0.9186284544524054,
      "grad_norm": 0.2109375,
      "learning_rate": 6.994769388733313e-07,
      "loss": 0.2446351796388626,
      "mean_token_accuracy": 0.9016678333282471,
      "num_tokens": 4169170654.0,
      "step": 1795
    },
    {
      "epoch": 0.9191402251791198,
      "grad_norm": 0.2197265625,
      "learning_rate": 6.908113294868691e-07,
      "loss": 0.250029057264328,
      "mean_token_accuracy": 0.899800181388855,
      "num_tokens": 4171486718.0,
      "step": 1796
    },
    {
      "epoch": 0.9196519959058342,
      "grad_norm": 0.23828125,
      "learning_rate": 6.821987896306481e-07,
      "loss": 0.25022679567337036,
      "mean_token_accuracy": 0.8998500108718872,
      "num_tokens": 4173708718.0,
      "step": 1797
    },
    {
      "epoch": 0.9201637666325486,
      "grad_norm": 0.2080078125,
      "learning_rate": 6.736393429754429e-07,
      "loss": 0.2438860386610031,
      "mean_token_accuracy": 0.9020729660987854,
      "num_tokens": 4176061746.0,
      "step": 1798
    },
    {
      "epoch": 0.920675537359263,
      "grad_norm": 0.201171875,
      "learning_rate": 6.651330130461087e-07,
      "loss": 0.23754999041557312,
      "mean_token_accuracy": 0.903588056564331,
      "num_tokens": 4178418635.0,
      "step": 1799
    },
    {
      "epoch": 0.9211873080859775,
      "grad_norm": 0.208984375,
      "learning_rate": 6.566798232215066e-07,
      "loss": 0.2344556748867035,
      "mean_token_accuracy": 0.906150758266449,
      "num_tokens": 4180714607.0,
      "step": 1800
    },
    {
      "epoch": 0.9216990788126919,
      "grad_norm": 0.2060546875,
      "learning_rate": 6.482797967344501e-07,
      "loss": 0.23019728064537048,
      "mean_token_accuracy": 0.9076136350631714,
      "num_tokens": 4183090218.0,
      "step": 1801
    },
    {
      "epoch": 0.9222108495394064,
      "grad_norm": 0.2021484375,
      "learning_rate": 6.399329566716495e-07,
      "loss": 0.23415428400039673,
      "mean_token_accuracy": 0.9057151079177856,
      "num_tokens": 4185399782.0,
      "step": 1802
    },
    {
      "epoch": 0.9227226202661207,
      "grad_norm": 0.201171875,
      "learning_rate": 6.316393259736187e-07,
      "loss": 0.24380245804786682,
      "mean_token_accuracy": 0.9014829397201538,
      "num_tokens": 4187731845.0,
      "step": 1803
    },
    {
      "epoch": 0.9232343909928352,
      "grad_norm": 0.2177734375,
      "learning_rate": 6.233989274346441e-07,
      "loss": 0.24257655441761017,
      "mean_token_accuracy": 0.9027398824691772,
      "num_tokens": 4190013424.0,
      "step": 1804
    },
    {
      "epoch": 0.9237461617195496,
      "grad_norm": 0.20703125,
      "learning_rate": 6.152117837027006e-07,
      "loss": 0.24223364889621735,
      "mean_token_accuracy": 0.903228759765625,
      "num_tokens": 4192284873.0,
      "step": 1805
    },
    {
      "epoch": 0.9242579324462641,
      "grad_norm": 0.212890625,
      "learning_rate": 6.070779172794039e-07,
      "loss": 0.2462259829044342,
      "mean_token_accuracy": 0.9009907245635986,
      "num_tokens": 4194585576.0,
      "step": 1806
    },
    {
      "epoch": 0.9247697031729785,
      "grad_norm": 0.2119140625,
      "learning_rate": 5.989973505199343e-07,
      "loss": 0.2411918342113495,
      "mean_token_accuracy": 0.903308093547821,
      "num_tokens": 4196905882.0,
      "step": 1807
    },
    {
      "epoch": 0.925281473899693,
      "grad_norm": 0.2060546875,
      "learning_rate": 5.909701056329864e-07,
      "loss": 0.23583948612213135,
      "mean_token_accuracy": 0.905494213104248,
      "num_tokens": 4199236097.0,
      "step": 1808
    },
    {
      "epoch": 0.9257932446264073,
      "grad_norm": 0.2041015625,
      "learning_rate": 5.82996204680708e-07,
      "loss": 0.2447032332420349,
      "mean_token_accuracy": 0.9019079208374023,
      "num_tokens": 4201480182.0,
      "step": 1809
    },
    {
      "epoch": 0.9263050153531218,
      "grad_norm": 0.2080078125,
      "learning_rate": 5.75075669578633e-07,
      "loss": 0.24633444845676422,
      "mean_token_accuracy": 0.9010613560676575,
      "num_tokens": 4203803033.0,
      "step": 1810
    },
    {
      "epoch": 0.9268167860798362,
      "grad_norm": 0.2099609375,
      "learning_rate": 5.672085220956214e-07,
      "loss": 0.24297131597995758,
      "mean_token_accuracy": 0.9018105864524841,
      "num_tokens": 4206151075.0,
      "step": 1811
    },
    {
      "epoch": 0.9273285568065507,
      "grad_norm": 0.212890625,
      "learning_rate": 5.593947838538083e-07,
      "loss": 0.25089481472969055,
      "mean_token_accuracy": 0.8985912799835205,
      "num_tokens": 4208463982.0,
      "step": 1812
    },
    {
      "epoch": 0.9278403275332651,
      "grad_norm": 0.2138671875,
      "learning_rate": 5.51634476328533e-07,
      "loss": 0.25064563751220703,
      "mean_token_accuracy": 0.8990923166275024,
      "num_tokens": 4210769118.0,
      "step": 1813
    },
    {
      "epoch": 0.9283520982599796,
      "grad_norm": 0.2099609375,
      "learning_rate": 5.43927620848288e-07,
      "loss": 0.23145544528961182,
      "mean_token_accuracy": 0.9071863889694214,
      "num_tokens": 4213084880.0,
      "step": 1814
    },
    {
      "epoch": 0.9288638689866939,
      "grad_norm": 0.212890625,
      "learning_rate": 5.362742385946562e-07,
      "loss": 0.24774497747421265,
      "mean_token_accuracy": 0.9010523557662964,
      "num_tokens": 4215363351.0,
      "step": 1815
    },
    {
      "epoch": 0.9293756397134084,
      "grad_norm": 0.20703125,
      "learning_rate": 5.28674350602254e-07,
      "loss": 0.23351147770881653,
      "mean_token_accuracy": 0.906771183013916,
      "num_tokens": 4217737173.0,
      "step": 1816
    },
    {
      "epoch": 0.9298874104401228,
      "grad_norm": 0.2041015625,
      "learning_rate": 5.211279777586797e-07,
      "loss": 0.22095508873462677,
      "mean_token_accuracy": 0.9109306931495667,
      "num_tokens": 4219988970.0,
      "step": 1817
    },
    {
      "epoch": 0.9303991811668373,
      "grad_norm": 0.212890625,
      "learning_rate": 5.136351408044405e-07,
      "loss": 0.24047806859016418,
      "mean_token_accuracy": 0.9039925932884216,
      "num_tokens": 4222273419.0,
      "step": 1818
    },
    {
      "epoch": 0.9309109518935517,
      "grad_norm": 0.212890625,
      "learning_rate": 5.061958603329143e-07,
      "loss": 0.24951577186584473,
      "mean_token_accuracy": 0.9002627730369568,
      "num_tokens": 4224593714.0,
      "step": 1819
    },
    {
      "epoch": 0.9314227226202662,
      "grad_norm": 0.2080078125,
      "learning_rate": 4.988101567902792e-07,
      "loss": 0.2279396951198578,
      "mean_token_accuracy": 0.9085761904716492,
      "num_tokens": 4226909369.0,
      "step": 1820
    },
    {
      "epoch": 0.9319344933469805,
      "grad_norm": 0.208984375,
      "learning_rate": 4.91478050475458e-07,
      "loss": 0.23580925166606903,
      "mean_token_accuracy": 0.9053167104721069,
      "num_tokens": 4229281772.0,
      "step": 1821
    },
    {
      "epoch": 0.932446264073695,
      "grad_norm": 0.2138671875,
      "learning_rate": 4.841995615400729e-07,
      "loss": 0.24506716430187225,
      "mean_token_accuracy": 0.9013609886169434,
      "num_tokens": 4231555352.0,
      "step": 1822
    },
    {
      "epoch": 0.9329580348004094,
      "grad_norm": 0.205078125,
      "learning_rate": 4.769747099883849e-07,
      "loss": 0.22468651831150055,
      "mean_token_accuracy": 0.9102567434310913,
      "num_tokens": 4233896598.0,
      "step": 1823
    },
    {
      "epoch": 0.9334698055271239,
      "grad_norm": 0.2109375,
      "learning_rate": 4.6980351567723183e-07,
      "loss": 0.2460867166519165,
      "mean_token_accuracy": 0.9019665718078613,
      "num_tokens": 4236158262.0,
      "step": 1824
    },
    {
      "epoch": 0.9339815762538383,
      "grad_norm": 0.2197265625,
      "learning_rate": 4.626859983159837e-07,
      "loss": 0.2445887327194214,
      "mean_token_accuracy": 0.9016522169113159,
      "num_tokens": 4238553217.0,
      "step": 1825
    },
    {
      "epoch": 0.9344933469805528,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.556221774664793e-07,
      "loss": 0.23162530362606049,
      "mean_token_accuracy": 0.9065478444099426,
      "num_tokens": 4240839569.0,
      "step": 1826
    },
    {
      "epoch": 0.9350051177072671,
      "grad_norm": 0.208984375,
      "learning_rate": 4.4861207254298654e-07,
      "loss": 0.24220070242881775,
      "mean_token_accuracy": 0.9027045965194702,
      "num_tokens": 4243247837.0,
      "step": 1827
    },
    {
      "epoch": 0.9355168884339816,
      "grad_norm": 0.2138671875,
      "learning_rate": 4.4165570281212934e-07,
      "loss": 0.24767296016216278,
      "mean_token_accuracy": 0.9001665115356445,
      "num_tokens": 4245471192.0,
      "step": 1828
    },
    {
      "epoch": 0.936028659160696,
      "grad_norm": 0.19921875,
      "learning_rate": 4.3475308739285184e-07,
      "loss": 0.23595654964447021,
      "mean_token_accuracy": 0.9054543972015381,
      "num_tokens": 4247775285.0,
      "step": 1829
    },
    {
      "epoch": 0.9365404298874105,
      "grad_norm": 0.22265625,
      "learning_rate": 4.2790424525635866e-07,
      "loss": 0.22523005306720734,
      "mean_token_accuracy": 0.9096208810806274,
      "num_tokens": 4250154734.0,
      "step": 1830
    },
    {
      "epoch": 0.9370522006141249,
      "grad_norm": 0.224609375,
      "learning_rate": 4.2110919522606153e-07,
      "loss": 0.24768555164337158,
      "mean_token_accuracy": 0.9002323746681213,
      "num_tokens": 4252488091.0,
      "step": 1831
    },
    {
      "epoch": 0.9375639713408394,
      "grad_norm": 0.20703125,
      "learning_rate": 4.1436795597753485e-07,
      "loss": 0.23622006177902222,
      "mean_token_accuracy": 0.905249834060669,
      "num_tokens": 4254733154.0,
      "step": 1832
    },
    {
      "epoch": 0.9380757420675537,
      "grad_norm": 0.2001953125,
      "learning_rate": 4.076805460384514e-07,
      "loss": 0.22452300786972046,
      "mean_token_accuracy": 0.9087485671043396,
      "num_tokens": 4257013345.0,
      "step": 1833
    },
    {
      "epoch": 0.9385875127942682,
      "grad_norm": 0.2041015625,
      "learning_rate": 4.0104698378854444e-07,
      "loss": 0.2288627326488495,
      "mean_token_accuracy": 0.9073005318641663,
      "num_tokens": 4259359485.0,
      "step": 1834
    },
    {
      "epoch": 0.9390992835209826,
      "grad_norm": 0.2109375,
      "learning_rate": 3.944672874595479e-07,
      "loss": 0.24240367114543915,
      "mean_token_accuracy": 0.9014340043067932,
      "num_tokens": 4261714989.0,
      "step": 1835
    },
    {
      "epoch": 0.9396110542476971,
      "grad_norm": 0.2001953125,
      "learning_rate": 3.8794147513515403e-07,
      "loss": 0.23006820678710938,
      "mean_token_accuracy": 0.9076758027076721,
      "num_tokens": 4264092029.0,
      "step": 1836
    },
    {
      "epoch": 0.9401228249744115,
      "grad_norm": 0.208984375,
      "learning_rate": 3.8146956475095807e-07,
      "loss": 0.25301915407180786,
      "mean_token_accuracy": 0.8989365100860596,
      "num_tokens": 4266424069.0,
      "step": 1837
    },
    {
      "epoch": 0.9406345957011258,
      "grad_norm": 0.205078125,
      "learning_rate": 3.7505157409441384e-07,
      "loss": 0.23898211121559143,
      "mean_token_accuracy": 0.9042384624481201,
      "num_tokens": 4268711387.0,
      "step": 1838
    },
    {
      "epoch": 0.9411463664278403,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.686875208047713e-07,
      "loss": 0.2464844286441803,
      "mean_token_accuracy": 0.9007272720336914,
      "num_tokens": 4271019841.0,
      "step": 1839
    },
    {
      "epoch": 0.9416581371545547,
      "grad_norm": 0.203125,
      "learning_rate": 3.623774223730503e-07,
      "loss": 0.23183003067970276,
      "mean_token_accuracy": 0.9067622423171997,
      "num_tokens": 4273408095.0,
      "step": 1840
    },
    {
      "epoch": 0.9421699078812692,
      "grad_norm": 0.2041015625,
      "learning_rate": 3.561212961419802e-07,
      "loss": 0.2327738106250763,
      "mean_token_accuracy": 0.9060686230659485,
      "num_tokens": 4275698274.0,
      "step": 1841
    },
    {
      "epoch": 0.9426816786079836,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.4991915930594255e-07,
      "loss": 0.23882006108760834,
      "mean_token_accuracy": 0.9042930006980896,
      "num_tokens": 4278090204.0,
      "step": 1842
    },
    {
      "epoch": 0.9431934493346981,
      "grad_norm": 0.20703125,
      "learning_rate": 3.4377102891094194e-07,
      "loss": 0.24218299984931946,
      "mean_token_accuracy": 0.9028744101524353,
      "num_tokens": 4280389345.0,
      "step": 1843
    },
    {
      "epoch": 0.9437052200614124,
      "grad_norm": 0.2119140625,
      "learning_rate": 3.376769218545528e-07,
      "loss": 0.24414214491844177,
      "mean_token_accuracy": 0.9017541408538818,
      "num_tokens": 4282774127.0,
      "step": 1844
    },
    {
      "epoch": 0.9442169907881269,
      "grad_norm": 0.2236328125,
      "learning_rate": 3.3163685488586617e-07,
      "loss": 0.2592739760875702,
      "mean_token_accuracy": 0.8959935903549194,
      "num_tokens": 4285186379.0,
      "step": 1845
    },
    {
      "epoch": 0.9447287615148413,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.256508446054474e-07,
      "loss": 0.229769766330719,
      "mean_token_accuracy": 0.9070974588394165,
      "num_tokens": 4287570177.0,
      "step": 1846
    },
    {
      "epoch": 0.9452405322415558,
      "grad_norm": 0.2060546875,
      "learning_rate": 3.197189074653029e-07,
      "loss": 0.24879905581474304,
      "mean_token_accuracy": 0.900162935256958,
      "num_tokens": 4289923347.0,
      "step": 1847
    },
    {
      "epoch": 0.9457523029682702,
      "grad_norm": 0.2197265625,
      "learning_rate": 3.138410597688113e-07,
      "loss": 0.24884918332099915,
      "mean_token_accuracy": 0.9006398320198059,
      "num_tokens": 4292247432.0,
      "step": 1848
    },
    {
      "epoch": 0.9462640736949847,
      "grad_norm": 0.189453125,
      "learning_rate": 3.080173176706991e-07,
      "loss": 0.2152806967496872,
      "mean_token_accuracy": 0.9129062294960022,
      "num_tokens": 4294606024.0,
      "step": 1849
    },
    {
      "epoch": 0.946775844421699,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.022476971769828e-07,
      "loss": 0.24492430686950684,
      "mean_token_accuracy": 0.9020473957061768,
      "num_tokens": 4296900496.0,
      "step": 1850
    },
    {
      "epoch": 0.9472876151484135,
      "grad_norm": 0.205078125,
      "learning_rate": 2.965322141449378e-07,
      "loss": 0.2320486605167389,
      "mean_token_accuracy": 0.9071225523948669,
      "num_tokens": 4299205020.0,
      "step": 1851
    },
    {
      "epoch": 0.9477993858751279,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.908708842830454e-07,
      "loss": 0.2555650770664215,
      "mean_token_accuracy": 0.8966361284255981,
      "num_tokens": 4301482346.0,
      "step": 1852
    },
    {
      "epoch": 0.9483111566018424,
      "grad_norm": 0.2138671875,
      "learning_rate": 2.8526372315094366e-07,
      "loss": 0.2424306869506836,
      "mean_token_accuracy": 0.9028494358062744,
      "num_tokens": 4303801653.0,
      "step": 1853
    },
    {
      "epoch": 0.9488229273285568,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.797107461594029e-07,
      "loss": 0.23908714950084686,
      "mean_token_accuracy": 0.9046997427940369,
      "num_tokens": 4306234832.0,
      "step": 1854
    },
    {
      "epoch": 0.9493346980552713,
      "grad_norm": 0.2109375,
      "learning_rate": 2.7421196857027267e-07,
      "loss": 0.24698995053768158,
      "mean_token_accuracy": 0.9005534052848816,
      "num_tokens": 4308554551.0,
      "step": 1855
    },
    {
      "epoch": 0.9498464687819856,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.687674054964351e-07,
      "loss": 0.23998281359672546,
      "mean_token_accuracy": 0.9032370448112488,
      "num_tokens": 4310838180.0,
      "step": 1856
    },
    {
      "epoch": 0.9503582395087001,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.6337707190177143e-07,
      "loss": 0.24901746213436127,
      "mean_token_accuracy": 0.9001930356025696,
      "num_tokens": 4313169544.0,
      "step": 1857
    },
    {
      "epoch": 0.9508700102354145,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.5804098260111765e-07,
      "loss": 0.24582360684871674,
      "mean_token_accuracy": 0.9017804861068726,
      "num_tokens": 4315546895.0,
      "step": 1858
    },
    {
      "epoch": 0.951381780962129,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.5275915226022683e-07,
      "loss": 0.24259379506111145,
      "mean_token_accuracy": 0.9026422500610352,
      "num_tokens": 4317867629.0,
      "step": 1859
    },
    {
      "epoch": 0.9518935516888434,
      "grad_norm": 0.2041015625,
      "learning_rate": 2.4753159539572247e-07,
      "loss": 0.22942094504833221,
      "mean_token_accuracy": 0.9083530306816101,
      "num_tokens": 4320213241.0,
      "step": 1860
    },
    {
      "epoch": 0.9524053224155579,
      "grad_norm": 0.201171875,
      "learning_rate": 2.423583263750628e-07,
      "loss": 0.23639509081840515,
      "mean_token_accuracy": 0.9045007228851318,
      "num_tokens": 4322511210.0,
      "step": 1861
    },
    {
      "epoch": 0.9529170931422722,
      "grad_norm": 0.212890625,
      "learning_rate": 2.3723935941650566e-07,
      "loss": 0.24576647579669952,
      "mean_token_accuracy": 0.9009370803833008,
      "num_tokens": 4324815594.0,
      "step": 1862
    },
    {
      "epoch": 0.9534288638689867,
      "grad_norm": 0.2197265625,
      "learning_rate": 2.321747085890591e-07,
      "loss": 0.24352094531059265,
      "mean_token_accuracy": 0.9016268253326416,
      "num_tokens": 4327159791.0,
      "step": 1863
    },
    {
      "epoch": 0.9539406345957011,
      "grad_norm": 0.2109375,
      "learning_rate": 2.2716438781245077e-07,
      "loss": 0.24669376015663147,
      "mean_token_accuracy": 0.9011505842208862,
      "num_tokens": 4329431284.0,
      "step": 1864
    },
    {
      "epoch": 0.9544524053224156,
      "grad_norm": 0.21875,
      "learning_rate": 2.2220841085708766e-07,
      "loss": 0.242165207862854,
      "mean_token_accuracy": 0.9024611711502075,
      "num_tokens": 4331716270.0,
      "step": 1865
    },
    {
      "epoch": 0.95496417604913,
      "grad_norm": 0.2060546875,
      "learning_rate": 2.1730679134401633e-07,
      "loss": 0.23393496870994568,
      "mean_token_accuracy": 0.9058694243431091,
      "num_tokens": 4334046941.0,
      "step": 1866
    },
    {
      "epoch": 0.9554759467758445,
      "grad_norm": 0.21875,
      "learning_rate": 2.12459542744885e-07,
      "loss": 0.25998759269714355,
      "mean_token_accuracy": 0.8963778614997864,
      "num_tokens": 4336317852.0,
      "step": 1867
    },
    {
      "epoch": 0.9559877175025588,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.076666783819148e-07,
      "loss": 0.24574214220046997,
      "mean_token_accuracy": 0.9019118547439575,
      "num_tokens": 4338617237.0,
      "step": 1868
    },
    {
      "epoch": 0.9564994882292733,
      "grad_norm": 0.2080078125,
      "learning_rate": 2.029282114278508e-07,
      "loss": 0.244873046875,
      "mean_token_accuracy": 0.9016914367675781,
      "num_tokens": 4340988437.0,
      "step": 1869
    },
    {
      "epoch": 0.9570112589559877,
      "grad_norm": 0.2109375,
      "learning_rate": 1.982441549059333e-07,
      "loss": 0.24116575717926025,
      "mean_token_accuracy": 0.9026810526847839,
      "num_tokens": 4343372114.0,
      "step": 1870
    },
    {
      "epoch": 0.9575230296827022,
      "grad_norm": 0.189453125,
      "learning_rate": 1.9361452168986216e-07,
      "loss": 0.2202812135219574,
      "mean_token_accuracy": 0.9114246368408203,
      "num_tokens": 4345771190.0,
      "step": 1871
    },
    {
      "epoch": 0.9580348004094166,
      "grad_norm": 0.203125,
      "learning_rate": 1.8903932450375917e-07,
      "loss": 0.23655346035957336,
      "mean_token_accuracy": 0.9050624370574951,
      "num_tokens": 4348130513.0,
      "step": 1872
    },
    {
      "epoch": 0.9585465711361311,
      "grad_norm": 0.2109375,
      "learning_rate": 1.8451857592213463e-07,
      "loss": 0.24427402019500732,
      "mean_token_accuracy": 0.9025135636329651,
      "num_tokens": 4350426232.0,
      "step": 1873
    },
    {
      "epoch": 0.9590583418628454,
      "grad_norm": 0.20703125,
      "learning_rate": 1.8005228836984745e-07,
      "loss": 0.24683673679828644,
      "mean_token_accuracy": 0.9011588096618652,
      "num_tokens": 4352723915.0,
      "step": 1874
    },
    {
      "epoch": 0.9595701125895599,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.7564047412207852e-07,
      "loss": 0.23948240280151367,
      "mean_token_accuracy": 0.9040013551712036,
      "num_tokens": 4355126804.0,
      "step": 1875
    },
    {
      "epoch": 0.9600818833162743,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.7128314530429734e-07,
      "loss": 0.2483426183462143,
      "mean_token_accuracy": 0.900321364402771,
      "num_tokens": 4357411361.0,
      "step": 1876
    },
    {
      "epoch": 0.9605936540429887,
      "grad_norm": 0.203125,
      "learning_rate": 1.6698031389221549e-07,
      "loss": 0.24036595225334167,
      "mean_token_accuracy": 0.9030441045761108,
      "num_tokens": 4359723128.0,
      "step": 1877
    },
    {
      "epoch": 0.9611054247697032,
      "grad_norm": 0.2158203125,
      "learning_rate": 1.6273199171177534e-07,
      "loss": 0.2514555752277374,
      "mean_token_accuracy": 0.8989853858947754,
      "num_tokens": 4362047280.0,
      "step": 1878
    },
    {
      "epoch": 0.9616171954964176,
      "grad_norm": 0.1962890625,
      "learning_rate": 1.5853819043909702e-07,
      "loss": 0.21989592909812927,
      "mean_token_accuracy": 0.911870002746582,
      "num_tokens": 4364419317.0,
      "step": 1879
    },
    {
      "epoch": 0.962128966223132,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.5439892160046045e-07,
      "loss": 0.23366782069206238,
      "mean_token_accuracy": 0.9057061672210693,
      "num_tokens": 4366681009.0,
      "step": 1880
    },
    {
      "epoch": 0.9626407369498464,
      "grad_norm": 0.197265625,
      "learning_rate": 1.5031419657226322e-07,
      "loss": 0.23205488920211792,
      "mean_token_accuracy": 0.9065949320793152,
      "num_tokens": 4369048013.0,
      "step": 1881
    },
    {
      "epoch": 0.9631525076765609,
      "grad_norm": 0.1943359375,
      "learning_rate": 1.4628402658099615e-07,
      "loss": 0.2310076206922531,
      "mean_token_accuracy": 0.907619833946228,
      "num_tokens": 4371405632.0,
      "step": 1882
    },
    {
      "epoch": 0.9636642784032753,
      "grad_norm": 0.2197265625,
      "learning_rate": 1.4230842270321454e-07,
      "loss": 0.2340496927499771,
      "mean_token_accuracy": 0.9064990282058716,
      "num_tokens": 4373790468.0,
      "step": 1883
    },
    {
      "epoch": 0.9641760491299898,
      "grad_norm": 0.205078125,
      "learning_rate": 1.3838739586549354e-07,
      "loss": 0.23382988572120667,
      "mean_token_accuracy": 0.9055960178375244,
      "num_tokens": 4376076178.0,
      "step": 1884
    },
    {
      "epoch": 0.9646878198567042,
      "grad_norm": 0.205078125,
      "learning_rate": 1.3452095684442167e-07,
      "loss": 0.23047101497650146,
      "mean_token_accuracy": 0.907785952091217,
      "num_tokens": 4378336542.0,
      "step": 1885
    },
    {
      "epoch": 0.9651995905834186,
      "grad_norm": 0.2109375,
      "learning_rate": 1.307091162665497e-07,
      "loss": 0.2432878613471985,
      "mean_token_accuracy": 0.9028106331825256,
      "num_tokens": 4380700489.0,
      "step": 1886
    },
    {
      "epoch": 0.965711361310133,
      "grad_norm": 0.2041015625,
      "learning_rate": 1.269518846083706e-07,
      "loss": 0.2314404547214508,
      "mean_token_accuracy": 0.9070481061935425,
      "num_tokens": 4383087408.0,
      "step": 1887
    },
    {
      "epoch": 0.9662231320368475,
      "grad_norm": 0.203125,
      "learning_rate": 1.2324927219628636e-07,
      "loss": 0.22960352897644043,
      "mean_token_accuracy": 0.9079437255859375,
      "num_tokens": 4385363253.0,
      "step": 1888
    },
    {
      "epoch": 0.9667349027635619,
      "grad_norm": 0.20703125,
      "learning_rate": 1.1960128920659454e-07,
      "loss": 0.23420953750610352,
      "mean_token_accuracy": 0.9055393934249878,
      "num_tokens": 4387743167.0,
      "step": 1889
    },
    {
      "epoch": 0.9672466734902764,
      "grad_norm": 0.21484375,
      "learning_rate": 1.1600794566543505e-07,
      "loss": 0.2542899250984192,
      "mean_token_accuracy": 0.8979686498641968,
      "num_tokens": 4390040186.0,
      "step": 1890
    },
    {
      "epoch": 0.9677584442169908,
      "grad_norm": 0.20703125,
      "learning_rate": 1.1246925144878352e-07,
      "loss": 0.23831850290298462,
      "mean_token_accuracy": 0.9043866991996765,
      "num_tokens": 4392382324.0,
      "step": 1891
    },
    {
      "epoch": 0.9682702149437052,
      "grad_norm": 0.21484375,
      "learning_rate": 1.0898521628241566e-07,
      "loss": 0.24255505204200745,
      "mean_token_accuracy": 0.903046727180481,
      "num_tokens": 4394816536.0,
      "step": 1892
    },
    {
      "epoch": 0.9687819856704196,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.0555584974188293e-07,
      "loss": 0.2371337115764618,
      "mean_token_accuracy": 0.9045852422714233,
      "num_tokens": 4397160349.0,
      "step": 1893
    },
    {
      "epoch": 0.9692937563971341,
      "grad_norm": 0.2109375,
      "learning_rate": 1.0218116125248589e-07,
      "loss": 0.24374675750732422,
      "mean_token_accuracy": 0.9017643332481384,
      "num_tokens": 4399449264.0,
      "step": 1894
    },
    {
      "epoch": 0.9698055271238485,
      "grad_norm": 0.2041015625,
      "learning_rate": 9.886116008924085e-08,
      "loss": 0.24251684546470642,
      "mean_token_accuracy": 0.9032410383224487,
      "num_tokens": 4401760734.0,
      "step": 1895
    },
    {
      "epoch": 0.970317297850563,
      "grad_norm": 0.2021484375,
      "learning_rate": 9.5595855376871e-08,
      "loss": 0.24621185660362244,
      "mean_token_accuracy": 0.9011445045471191,
      "num_tokens": 4404155495.0,
      "step": 1896
    },
    {
      "epoch": 0.9708290685772774,
      "grad_norm": 0.201171875,
      "learning_rate": 9.2385256089762e-08,
      "loss": 0.2312861680984497,
      "mean_token_accuracy": 0.9072175025939941,
      "num_tokens": 4406466540.0,
      "step": 1897
    },
    {
      "epoch": 0.9713408393039918,
      "grad_norm": 0.2060546875,
      "learning_rate": 8.922937105195317e-08,
      "loss": 0.23264849185943604,
      "mean_token_accuracy": 0.9063292741775513,
      "num_tokens": 4408774692.0,
      "step": 1898
    },
    {
      "epoch": 0.9718526100307062,
      "grad_norm": 0.208984375,
      "learning_rate": 8.612820893710405e-08,
      "loss": 0.24374312162399292,
      "mean_token_accuracy": 0.9022489190101624,
      "num_tokens": 4411103623.0,
      "step": 1899
    },
    {
      "epoch": 0.9723643807574207,
      "grad_norm": 0.203125,
      "learning_rate": 8.30817782684723e-08,
      "loss": 0.23233643174171448,
      "mean_token_accuracy": 0.9059531092643738,
      "num_tokens": 4413415462.0,
      "step": 1900
    },
    {
      "epoch": 0.9728761514841351,
      "grad_norm": 0.2236328125,
      "learning_rate": 8.00900874188959e-08,
      "loss": 0.26903262734413147,
      "mean_token_accuracy": 0.8929834365844727,
      "num_tokens": 4415739459.0,
      "step": 1901
    },
    {
      "epoch": 0.9733879222108496,
      "grad_norm": 0.20703125,
      "learning_rate": 7.715314461075763e-08,
      "loss": 0.2502642273902893,
      "mean_token_accuracy": 0.9000949859619141,
      "num_tokens": 4418108371.0,
      "step": 1902
    },
    {
      "epoch": 0.973899692937564,
      "grad_norm": 0.2138671875,
      "learning_rate": 7.427095791597838e-08,
      "loss": 0.2364313006401062,
      "mean_token_accuracy": 0.9049532413482666,
      "num_tokens": 4420355561.0,
      "step": 1903
    },
    {
      "epoch": 0.9744114636642784,
      "grad_norm": 0.2099609375,
      "learning_rate": 7.144353525598391e-08,
      "loss": 0.24504435062408447,
      "mean_token_accuracy": 0.9020323753356934,
      "num_tokens": 4422735411.0,
      "step": 1904
    },
    {
      "epoch": 0.9749232343909928,
      "grad_norm": 0.197265625,
      "learning_rate": 6.867088440168034e-08,
      "loss": 0.23235927522182465,
      "mean_token_accuracy": 0.9074075222015381,
      "num_tokens": 4425249737.0,
      "step": 1905
    },
    {
      "epoch": 0.9754350051177073,
      "grad_norm": 0.2119140625,
      "learning_rate": 6.595301297344981e-08,
      "loss": 0.2413080930709839,
      "mean_token_accuracy": 0.9030349850654602,
      "num_tokens": 4427637276.0,
      "step": 1906
    },
    {
      "epoch": 0.9759467758444217,
      "grad_norm": 0.2099609375,
      "learning_rate": 6.328992844110815e-08,
      "loss": 0.24177658557891846,
      "mean_token_accuracy": 0.9024568796157837,
      "num_tokens": 4430020507.0,
      "step": 1907
    },
    {
      "epoch": 0.9764585465711362,
      "grad_norm": 0.2119140625,
      "learning_rate": 6.06816381239006e-08,
      "loss": 0.24208110570907593,
      "mean_token_accuracy": 0.902471661567688,
      "num_tokens": 4432336924.0,
      "step": 1908
    },
    {
      "epoch": 0.9769703172978506,
      "grad_norm": 0.2021484375,
      "learning_rate": 5.812814919047283e-08,
      "loss": 0.23694056272506714,
      "mean_token_accuracy": 0.9049849510192871,
      "num_tokens": 4434624774.0,
      "step": 1909
    },
    {
      "epoch": 0.977482088024565,
      "grad_norm": 0.2109375,
      "learning_rate": 5.5629468658853214e-08,
      "loss": 0.23598945140838623,
      "mean_token_accuracy": 0.9056123495101929,
      "num_tokens": 4436958377.0,
      "step": 1910
    },
    {
      "epoch": 0.9779938587512794,
      "grad_norm": 0.2109375,
      "learning_rate": 5.31856033964373e-08,
      "loss": 0.2350405752658844,
      "mean_token_accuracy": 0.905903697013855,
      "num_tokens": 4439314158.0,
      "step": 1911
    },
    {
      "epoch": 0.9785056294779939,
      "grad_norm": 0.2060546875,
      "learning_rate": 5.0796560119963364e-08,
      "loss": 0.2340988963842392,
      "mean_token_accuracy": 0.9066669940948486,
      "num_tokens": 4441627488.0,
      "step": 1912
    },
    {
      "epoch": 0.9790174002047083,
      "grad_norm": 0.197265625,
      "learning_rate": 4.846234539549688e-08,
      "loss": 0.22186754643917084,
      "mean_token_accuracy": 0.9109378457069397,
      "num_tokens": 4444004334.0,
      "step": 1913
    },
    {
      "epoch": 0.9795291709314228,
      "grad_norm": 0.2060546875,
      "learning_rate": 4.618296563841274e-08,
      "loss": 0.2351597398519516,
      "mean_token_accuracy": 0.9059134721755981,
      "num_tokens": 4446293262.0,
      "step": 1914
    },
    {
      "epoch": 0.9800409416581372,
      "grad_norm": 0.2216796875,
      "learning_rate": 4.39584271133775e-08,
      "loss": 0.24450939893722534,
      "mean_token_accuracy": 0.9014853835105896,
      "num_tokens": 4448505450.0,
      "step": 1915
    },
    {
      "epoch": 0.9805527123848515,
      "grad_norm": 0.21484375,
      "learning_rate": 4.1788735934329416e-08,
      "loss": 0.2407625913619995,
      "mean_token_accuracy": 0.9040027856826782,
      "num_tokens": 4450904115.0,
      "step": 1916
    },
    {
      "epoch": 0.981064483111566,
      "grad_norm": 0.208984375,
      "learning_rate": 3.9673898064465086e-08,
      "loss": 0.23535382747650146,
      "mean_token_accuracy": 0.9056427478790283,
      "num_tokens": 4453270664.0,
      "step": 1917
    },
    {
      "epoch": 0.9815762538382804,
      "grad_norm": 0.208984375,
      "learning_rate": 3.7613919316223934e-08,
      "loss": 0.2372298389673233,
      "mean_token_accuracy": 0.905026912689209,
      "num_tokens": 4455616709.0,
      "step": 1918
    },
    {
      "epoch": 0.9820880245649949,
      "grad_norm": 0.2099609375,
      "learning_rate": 3.5608805351265985e-08,
      "loss": 0.24346919357776642,
      "mean_token_accuracy": 0.9023795127868652,
      "num_tokens": 4457951382.0,
      "step": 1919
    },
    {
      "epoch": 0.9825997952917093,
      "grad_norm": 0.201171875,
      "learning_rate": 3.365856168046522e-08,
      "loss": 0.2374979853630066,
      "mean_token_accuracy": 0.9052866101264954,
      "num_tokens": 4460276443.0,
      "step": 1920
    },
    {
      "epoch": 0.9831115660184238,
      "grad_norm": 0.2177734375,
      "learning_rate": 3.176319366388514e-08,
      "loss": 0.2500947415828705,
      "mean_token_accuracy": 0.8995829820632935,
      "num_tokens": 4462596692.0,
      "step": 1921
    },
    {
      "epoch": 0.9836233367451381,
      "grad_norm": 0.205078125,
      "learning_rate": 2.992270651077211e-08,
      "loss": 0.24528098106384277,
      "mean_token_accuracy": 0.9018374681472778,
      "num_tokens": 4464919768.0,
      "step": 1922
    },
    {
      "epoch": 0.9841351074718526,
      "grad_norm": 0.208984375,
      "learning_rate": 2.8137105279533174e-08,
      "loss": 0.23430700600147247,
      "mean_token_accuracy": 0.9063045382499695,
      "num_tokens": 4467223356.0,
      "step": 1923
    },
    {
      "epoch": 0.984646878198567,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.6406394877731557e-08,
      "loss": 0.25115054845809937,
      "mean_token_accuracy": 0.9002994298934937,
      "num_tokens": 4469566310.0,
      "step": 1924
    },
    {
      "epoch": 0.9851586489252815,
      "grad_norm": 0.201171875,
      "learning_rate": 2.4730580062064523e-08,
      "loss": 0.22931015491485596,
      "mean_token_accuracy": 0.9077868461608887,
      "num_tokens": 4471900557.0,
      "step": 1925
    },
    {
      "epoch": 0.9856704196519959,
      "grad_norm": 0.2109375,
      "learning_rate": 2.3109665438356688e-08,
      "loss": 0.25012001395225525,
      "mean_token_accuracy": 0.8986424803733826,
      "num_tokens": 4474164608.0,
      "step": 1926
    },
    {
      "epoch": 0.9861821903787104,
      "grad_norm": 0.201171875,
      "learning_rate": 2.1543655461542244e-08,
      "loss": 0.229749396443367,
      "mean_token_accuracy": 0.9082704186439514,
      "num_tokens": 4476527204.0,
      "step": 1927
    },
    {
      "epoch": 0.9866939611054247,
      "grad_norm": 0.19921875,
      "learning_rate": 2.0032554435653885e-08,
      "loss": 0.22860166430473328,
      "mean_token_accuracy": 0.9089881181716919,
      "num_tokens": 4478808676.0,
      "step": 1928
    },
    {
      "epoch": 0.9872057318321392,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.8576366513816113e-08,
      "loss": 0.24762189388275146,
      "mean_token_accuracy": 0.9008810520172119,
      "num_tokens": 4481282023.0,
      "step": 1929
    },
    {
      "epoch": 0.9877175025588536,
      "grad_norm": 0.2021484375,
      "learning_rate": 1.7175095698225286e-08,
      "loss": 0.23651094734668732,
      "mean_token_accuracy": 0.9049453735351562,
      "num_tokens": 4483620937.0,
      "step": 1930
    },
    {
      "epoch": 0.9882292732855681,
      "grad_norm": 0.21484375,
      "learning_rate": 1.5828745840149594e-08,
      "loss": 0.24639737606048584,
      "mean_token_accuracy": 0.9014164209365845,
      "num_tokens": 4485993450.0,
      "step": 1931
    },
    {
      "epoch": 0.9887410440122825,
      "grad_norm": 0.205078125,
      "learning_rate": 1.453732063990243e-08,
      "loss": 0.23971056938171387,
      "mean_token_accuracy": 0.9037513732910156,
      "num_tokens": 4488339673.0,
      "step": 1932
    },
    {
      "epoch": 0.989252814738997,
      "grad_norm": 0.2060546875,
      "learning_rate": 1.3300823646851258e-08,
      "loss": 0.24533474445343018,
      "mean_token_accuracy": 0.9019055366516113,
      "num_tokens": 4490642612.0,
      "step": 1933
    },
    {
      "epoch": 0.9897645854657113,
      "grad_norm": 0.2099609375,
      "learning_rate": 1.2119258259393196e-08,
      "loss": 0.23759736120700836,
      "mean_token_accuracy": 0.9045272469520569,
      "num_tokens": 4492888992.0,
      "step": 1934
    },
    {
      "epoch": 0.9902763561924258,
      "grad_norm": 0.2119140625,
      "learning_rate": 1.0992627724952798e-08,
      "loss": 0.25161486864089966,
      "mean_token_accuracy": 0.8989132642745972,
      "num_tokens": 4495133133.0,
      "step": 1935
    },
    {
      "epoch": 0.9907881269191402,
      "grad_norm": 0.208984375,
      "learning_rate": 9.920935139968724e-09,
      "loss": 0.23531435430049896,
      "mean_token_accuracy": 0.9055323600769043,
      "num_tokens": 4497463013.0,
      "step": 1936
    },
    {
      "epoch": 0.9912998976458547,
      "grad_norm": 0.2216796875,
      "learning_rate": 8.904183449889303e-09,
      "loss": 0.24456465244293213,
      "mean_token_accuracy": 0.9024140238761902,
      "num_tokens": 4499864174.0,
      "step": 1937
    },
    {
      "epoch": 0.9918116683725691,
      "grad_norm": 0.2060546875,
      "learning_rate": 7.942375449165873e-09,
      "loss": 0.2417791783809662,
      "mean_token_accuracy": 0.9029876589775085,
      "num_tokens": 4502137354.0,
      "step": 1938
    },
    {
      "epoch": 0.9923234390992836,
      "grad_norm": 0.220703125,
      "learning_rate": 7.035513781237236e-09,
      "loss": 0.24936886131763458,
      "mean_token_accuracy": 0.8993229866027832,
      "num_tokens": 4504497095.0,
      "step": 1939
    },
    {
      "epoch": 0.9928352098259979,
      "grad_norm": 0.2109375,
      "learning_rate": 6.183600938529655e-09,
      "loss": 0.23786425590515137,
      "mean_token_accuracy": 0.9047465324401855,
      "num_tokens": 4506837099.0,
      "step": 1940
    },
    {
      "epoch": 0.9933469805527124,
      "grad_norm": 0.212890625,
      "learning_rate": 5.3866392624501995e-09,
      "loss": 0.25195011496543884,
      "mean_token_accuracy": 0.8981287479400635,
      "num_tokens": 4509123326.0,
      "step": 1941
    },
    {
      "epoch": 0.9938587512794268,
      "grad_norm": 0.2177734375,
      "learning_rate": 4.644630943368977e-09,
      "loss": 0.23868995904922485,
      "mean_token_accuracy": 0.9036577939987183,
      "num_tokens": 4511481024.0,
      "step": 1942
    },
    {
      "epoch": 0.9943705220061413,
      "grad_norm": 0.197265625,
      "learning_rate": 3.957578020632458e-09,
      "loss": 0.23438501358032227,
      "mean_token_accuracy": 0.9061693549156189,
      "num_tokens": 4513857160.0,
      "step": 1943
    },
    {
      "epoch": 0.9948822927328557,
      "grad_norm": 0.2080078125,
      "learning_rate": 3.325482382539047e-09,
      "loss": 0.24364961683750153,
      "mean_token_accuracy": 0.9023924469947815,
      "num_tokens": 4516158150.0,
      "step": 1944
    },
    {
      "epoch": 0.9953940634595702,
      "grad_norm": 0.203125,
      "learning_rate": 2.748345766350191e-09,
      "loss": 0.23208262026309967,
      "mean_token_accuracy": 0.9067127704620361,
      "num_tokens": 4518499538.0,
      "step": 1945
    },
    {
      "epoch": 0.9959058341862845,
      "grad_norm": 0.21875,
      "learning_rate": 2.2261697582681706e-09,
      "loss": 0.2413972020149231,
      "mean_token_accuracy": 0.903131902217865,
      "num_tokens": 4520864689.0,
      "step": 1946
    },
    {
      "epoch": 0.996417604912999,
      "grad_norm": 0.19921875,
      "learning_rate": 1.7589557934472034e-09,
      "loss": 0.23442962765693665,
      "mean_token_accuracy": 0.9058358073234558,
      "num_tokens": 4523216166.0,
      "step": 1947
    },
    {
      "epoch": 0.9969293756397134,
      "grad_norm": 0.21484375,
      "learning_rate": 1.346705155982342e-09,
      "loss": 0.24571941792964935,
      "mean_token_accuracy": 0.9015001654624939,
      "num_tokens": 4525631179.0,
      "step": 1948
    },
    {
      "epoch": 0.9974411463664279,
      "grad_norm": 0.21875,
      "learning_rate": 9.894189789050324e-10,
      "loss": 0.24400079250335693,
      "mean_token_accuracy": 0.9024295210838318,
      "num_tokens": 4528015179.0,
      "step": 1949
    },
    {
      "epoch": 0.9979529170931423,
      "grad_norm": 0.205078125,
      "learning_rate": 6.870982441875562e-10,
      "loss": 0.23915861546993256,
      "mean_token_accuracy": 0.9037104845046997,
      "num_tokens": 4530335507.0,
      "step": 1950
    },
    {
      "epoch": 0.9984646878198568,
      "grad_norm": 0.2080078125,
      "learning_rate": 4.397437827274864e-10,
      "loss": 0.2414853274822235,
      "mean_token_accuracy": 0.9032497406005859,
      "num_tokens": 4532611566.0,
      "step": 1951
    },
    {
      "epoch": 0.9989764585465711,
      "grad_norm": 0.2099609375,
      "learning_rate": 2.4735627435434894e-10,
      "loss": 0.24067628383636475,
      "mean_token_accuracy": 0.9041061997413635,
      "num_tokens": 4534923346.0,
      "step": 1952
    },
    {
      "epoch": 0.9994882292732856,
      "grad_norm": 0.197265625,
      "learning_rate": 1.0993624783184331e-10,
      "loss": 0.23372124135494232,
      "mean_token_accuracy": 0.9064845442771912,
      "num_tokens": 4537277003.0,
      "step": 1953
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.2236328125,
      "learning_rate": 2.7484080842299366e-11,
      "loss": 0.2508823275566101,
      "mean_token_accuracy": 0.8991467952728271,
      "num_tokens": 4539598413.0,
      "step": 1954
    }
  ],
  "logging_steps": 1,
  "max_steps": 1954,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.838893369331391e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}