{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6,
  "eval_steps": 100,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 0.5395457262173295,
      "epoch": 0.016,
      "grad_norm": 3200.0,
      "learning_rate": 4.5000000000000003e-07,
      "loss": 92.5745,
      "mean_token_accuracy": 0.4849457062780857,
      "num_tokens": 195524.0,
      "step": 10
    },
    {
      "entropy": 0.5515131525695324,
      "epoch": 0.032,
      "grad_norm": 2912.0,
      "learning_rate": 9.500000000000001e-07,
      "loss": 88.931,
      "mean_token_accuracy": 0.4874983370304108,
      "num_tokens": 390903.0,
      "step": 20
    },
    {
      "entropy": 0.6169477274641395,
      "epoch": 0.048,
      "grad_norm": 2912.0,
      "learning_rate": 1.45e-06,
      "loss": 78.3695,
      "mean_token_accuracy": 0.4966813003644347,
      "num_tokens": 589868.0,
      "step": 30
    },
    {
      "entropy": 0.8348326332867145,
      "epoch": 0.064,
      "grad_norm": 1160.0,
      "learning_rate": 1.9500000000000004e-06,
      "loss": 60.9248,
      "mean_token_accuracy": 0.505222667939961,
      "num_tokens": 791190.0,
      "step": 40
    },
    {
      "entropy": 1.3063700653612613,
      "epoch": 0.08,
      "grad_norm": 247.0,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 46.7906,
      "mean_token_accuracy": 0.5185849336907268,
      "num_tokens": 989860.0,
      "step": 50
    },
    {
      "entropy": 1.959995786100626,
      "epoch": 0.096,
      "grad_norm": 81.0,
      "learning_rate": 2.95e-06,
      "loss": 39.2759,
      "mean_token_accuracy": 0.5406044816598297,
      "num_tokens": 1181777.0,
      "step": 60
    },
    {
      "entropy": 2.191839651763439,
      "epoch": 0.112,
      "grad_norm": 57.0,
      "learning_rate": 3.45e-06,
      "loss": 34.17,
      "mean_token_accuracy": 0.5669999368488788,
      "num_tokens": 1385513.0,
      "step": 70
    },
    {
      "entropy": 1.863330227136612,
      "epoch": 0.128,
      "grad_norm": 48.25,
      "learning_rate": 3.95e-06,
      "loss": 31.3942,
      "mean_token_accuracy": 0.5891917455941439,
      "num_tokens": 1582368.0,
      "step": 80
    },
    {
      "entropy": 1.8903757609426974,
      "epoch": 0.144,
      "grad_norm": 39.25,
      "learning_rate": 4.450000000000001e-06,
      "loss": 29.9275,
      "mean_token_accuracy": 0.6045342192053795,
      "num_tokens": 1773764.0,
      "step": 90
    },
    {
      "entropy": 1.8238266378641128,
      "epoch": 0.16,
      "grad_norm": 36.25,
      "learning_rate": 4.95e-06,
      "loss": 29.3315,
      "mean_token_accuracy": 0.6066708620637655,
      "num_tokens": 1970077.0,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 1.7597194662094116,
      "eval_biology_loss": 1.708297848701477,
      "eval_biology_mean_token_accuracy": 0.6260299117565155,
      "eval_biology_num_tokens": 1970077.0,
      "eval_biology_runtime": 36.35,
      "eval_biology_samples_per_second": 13.755,
      "eval_biology_steps_per_second": 3.439,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 1.278447434425354,
      "eval_chemistry_loss": 1.24350905418396,
      "eval_chemistry_mean_token_accuracy": 0.717205593585968,
      "eval_chemistry_num_tokens": 1970077.0,
      "eval_chemistry_runtime": 44.9852,
      "eval_chemistry_samples_per_second": 11.115,
      "eval_chemistry_steps_per_second": 2.779,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_math_entropy": 0.9187946987152099,
      "eval_math_loss": 1.206222414970398,
      "eval_math_mean_token_accuracy": 0.7349825048446655,
      "eval_math_num_tokens": 1970077.0,
      "eval_math_runtime": 46.1658,
      "eval_math_samples_per_second": 10.831,
      "eval_math_steps_per_second": 2.708,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 1.2150983595848084,
      "eval_physics_loss": 1.2188719511032104,
      "eval_physics_mean_token_accuracy": 0.727249002456665,
      "eval_physics_num_tokens": 1970077.0,
      "eval_physics_runtime": 52.9985,
      "eval_physics_samples_per_second": 9.434,
      "eval_physics_steps_per_second": 2.359,
      "step": 100
    },
    {
      "entropy": 1.7616284050047397,
      "epoch": 0.176,
      "grad_norm": 33.25,
      "learning_rate": 5.450000000000001e-06,
      "loss": 27.9906,
      "mean_token_accuracy": 0.6186339979991317,
      "num_tokens": 2168354.0,
      "step": 110
    },
    {
      "entropy": 1.7145348988473414,
      "epoch": 0.192,
      "grad_norm": 36.25,
      "learning_rate": 5.950000000000001e-06,
      "loss": 27.271,
      "mean_token_accuracy": 0.6236628353595733,
      "num_tokens": 2365822.0,
      "step": 120
    },
    {
      "entropy": 1.7000788144767285,
      "epoch": 0.208,
      "grad_norm": 34.75,
      "learning_rate": 6.450000000000001e-06,
      "loss": 26.7483,
      "mean_token_accuracy": 0.6299795724451542,
      "num_tokens": 2558762.0,
      "step": 130
    },
    {
      "entropy": 1.6474159061908722,
      "epoch": 0.224,
      "grad_norm": 30.875,
      "learning_rate": 6.95e-06,
      "loss": 25.9891,
      "mean_token_accuracy": 0.6365062229335308,
      "num_tokens": 2755347.0,
      "step": 140
    },
    {
      "entropy": 1.6314032576978206,
      "epoch": 0.24,
      "grad_norm": 31.125,
      "learning_rate": 7.450000000000001e-06,
      "loss": 25.1799,
      "mean_token_accuracy": 0.64104891307652,
      "num_tokens": 2947346.0,
      "step": 150
    },
    {
      "entropy": 1.609092117100954,
      "epoch": 0.256,
      "grad_norm": 28.5,
      "learning_rate": 7.950000000000002e-06,
      "loss": 24.976,
      "mean_token_accuracy": 0.6415446903556585,
      "num_tokens": 3139957.0,
      "step": 160
    },
    {
      "entropy": 1.5253454238176345,
      "epoch": 0.272,
      "grad_norm": 28.875,
      "learning_rate": 8.45e-06,
      "loss": 24.4431,
      "mean_token_accuracy": 0.6438875071704387,
      "num_tokens": 3335951.0,
      "step": 170
    },
    {
      "entropy": 1.5003480702638625,
      "epoch": 0.288,
      "grad_norm": 31.5,
      "learning_rate": 8.95e-06,
      "loss": 23.9226,
      "mean_token_accuracy": 0.6480991121381521,
      "num_tokens": 3539731.0,
      "step": 180
    },
    {
      "entropy": 1.4835937578231095,
      "epoch": 0.304,
      "grad_norm": 29.375,
      "learning_rate": 9.450000000000001e-06,
      "loss": 23.7553,
      "mean_token_accuracy": 0.6533681858330965,
      "num_tokens": 3733488.0,
      "step": 190
    },
    {
      "entropy": 1.501522469520569,
      "epoch": 0.32,
      "grad_norm": 31.625,
      "learning_rate": 9.950000000000001e-06,
      "loss": 24.0263,
      "mean_token_accuracy": 0.6483922835439444,
      "num_tokens": 3920545.0,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 1.430639804840088,
      "eval_biology_loss": 1.4240669012069702,
      "eval_biology_mean_token_accuracy": 0.6643721928596497,
      "eval_biology_num_tokens": 3920545.0,
      "eval_biology_runtime": 36.631,
      "eval_biology_samples_per_second": 13.65,
      "eval_biology_steps_per_second": 3.412,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 1.068233790397644,
      "eval_chemistry_loss": 1.0446134805679321,
      "eval_chemistry_mean_token_accuracy": 0.7419922127723694,
      "eval_chemistry_num_tokens": 3920545.0,
      "eval_chemistry_runtime": 45.0089,
      "eval_chemistry_samples_per_second": 11.109,
      "eval_chemistry_steps_per_second": 2.777,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_math_entropy": 0.8016578803062439,
      "eval_math_loss": 1.064239740371704,
      "eval_math_mean_token_accuracy": 0.7532512550354004,
      "eval_math_num_tokens": 3920545.0,
      "eval_math_runtime": 46.1998,
      "eval_math_samples_per_second": 10.823,
      "eval_math_steps_per_second": 2.706,
      "step": 200
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 1.0390429463386535,
      "eval_physics_loss": 1.033566951751709,
      "eval_physics_mean_token_accuracy": 0.751745885848999,
      "eval_physics_num_tokens": 3920545.0,
      "eval_physics_runtime": 53.0382,
      "eval_physics_samples_per_second": 9.427,
      "eval_physics_steps_per_second": 2.357,
      "step": 200
    },
    {
      "entropy": 1.4409266009926796,
      "epoch": 0.336,
      "grad_norm": 30.5,
      "learning_rate": 1.045e-05,
      "loss": 23.1469,
      "mean_token_accuracy": 0.6563692390918732,
      "num_tokens": 4114077.0,
      "step": 210
    },
    {
      "entropy": 1.4634413696825503,
      "epoch": 0.352,
      "grad_norm": 29.875,
      "learning_rate": 1.095e-05,
      "loss": 23.4319,
      "mean_token_accuracy": 0.6559967070817947,
      "num_tokens": 4306949.0,
      "step": 220
    },
    {
      "entropy": 1.4390382423996926,
      "epoch": 0.368,
      "grad_norm": 29.5,
      "learning_rate": 1.145e-05,
      "loss": 23.1058,
      "mean_token_accuracy": 0.6596628129482269,
      "num_tokens": 4504001.0,
      "step": 230
    },
    {
      "entropy": 1.4329517252743245,
      "epoch": 0.384,
      "grad_norm": 30.625,
      "learning_rate": 1.195e-05,
      "loss": 23.0101,
      "mean_token_accuracy": 0.6596732661128044,
      "num_tokens": 4693812.0,
      "step": 240
    },
    {
      "entropy": 1.4356985121965409,
      "epoch": 0.4,
      "grad_norm": 26.0,
      "learning_rate": 1.2450000000000003e-05,
      "loss": 22.9403,
      "mean_token_accuracy": 0.6609152596443891,
      "num_tokens": 4887094.0,
      "step": 250
    },
    {
      "entropy": 1.445911981910467,
      "epoch": 0.416,
      "grad_norm": 28.0,
      "learning_rate": 1.295e-05,
      "loss": 23.2032,
      "mean_token_accuracy": 0.6575265824794769,
      "num_tokens": 5085369.0,
      "step": 260
    },
    {
      "entropy": 1.4276192136108876,
      "epoch": 0.432,
      "grad_norm": 30.25,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 22.9807,
      "mean_token_accuracy": 0.6605408634990454,
      "num_tokens": 5271275.0,
      "step": 270
    },
    {
      "entropy": 1.4254673644900322,
      "epoch": 0.448,
      "grad_norm": 32.0,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 22.9563,
      "mean_token_accuracy": 0.6602664466947317,
      "num_tokens": 5460559.0,
      "step": 280
    },
    {
      "entropy": 1.3786304764449597,
      "epoch": 0.464,
      "grad_norm": 28.375,
      "learning_rate": 1.4450000000000002e-05,
      "loss": 22.0498,
      "mean_token_accuracy": 0.6699710045009851,
      "num_tokens": 5653809.0,
      "step": 290
    },
    {
      "entropy": 1.3980159699916839,
      "epoch": 0.48,
      "grad_norm": 26.0,
      "learning_rate": 1.4950000000000003e-05,
      "loss": 22.5067,
      "mean_token_accuracy": 0.6643361520022154,
      "num_tokens": 5850176.0,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_biology_entropy": 1.378262495994568,
      "eval_biology_loss": 1.34170663356781,
      "eval_biology_mean_token_accuracy": 0.6778094511032104,
      "eval_biology_num_tokens": 5850176.0,
      "eval_biology_runtime": 36.5741,
      "eval_biology_samples_per_second": 13.671,
      "eval_biology_steps_per_second": 3.418,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_chemistry_entropy": 1.0265604996681212,
      "eval_chemistry_loss": 0.9883458018302917,
      "eval_chemistry_mean_token_accuracy": 0.7520135464668274,
      "eval_chemistry_num_tokens": 5850176.0,
      "eval_chemistry_runtime": 44.9771,
      "eval_chemistry_samples_per_second": 11.117,
      "eval_chemistry_steps_per_second": 2.779,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_math_entropy": 0.7964297127723694,
      "eval_math_loss": 1.0260037183761597,
      "eval_math_mean_token_accuracy": 0.759199161529541,
      "eval_math_num_tokens": 5850176.0,
      "eval_math_runtime": 46.2056,
      "eval_math_samples_per_second": 10.821,
      "eval_math_steps_per_second": 2.705,
      "step": 300
    },
    {
      "epoch": 0.48,
      "eval_physics_entropy": 1.0048614959716797,
      "eval_physics_loss": 0.9760251045227051,
      "eval_physics_mean_token_accuracy": 0.761619930267334,
      "eval_physics_num_tokens": 5850176.0,
      "eval_physics_runtime": 53.0249,
      "eval_physics_samples_per_second": 9.43,
      "eval_physics_steps_per_second": 2.357,
      "step": 300
    },
    {
      "entropy": 1.3288711309432983,
      "epoch": 0.496,
      "grad_norm": 26.375,
      "learning_rate": 1.545e-05,
      "loss": 21.2235,
      "mean_token_accuracy": 0.6774784430861474,
      "num_tokens": 6046503.0,
      "step": 310
    },
    {
      "entropy": 1.385315388813615,
      "epoch": 0.512,
      "grad_norm": 27.125,
      "learning_rate": 1.595e-05,
      "loss": 22.2247,
      "mean_token_accuracy": 0.6680108610540628,
      "num_tokens": 6240456.0,
      "step": 320
    },
    {
      "entropy": 1.3709059737622737,
      "epoch": 0.528,
      "grad_norm": 24.75,
      "learning_rate": 1.645e-05,
      "loss": 22.1079,
      "mean_token_accuracy": 0.6704635102301836,
      "num_tokens": 6430555.0,
      "step": 330
    },
    {
      "entropy": 1.3751415818929673,
      "epoch": 0.544,
      "grad_norm": 26.25,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 21.961,
      "mean_token_accuracy": 0.6710947025567293,
      "num_tokens": 6626006.0,
      "step": 340
    },
    {
      "entropy": 1.3602795064449311,
      "epoch": 0.56,
      "grad_norm": 30.625,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 21.9474,
      "mean_token_accuracy": 0.6700355879962444,
      "num_tokens": 6820754.0,
      "step": 350
    },
    {
      "entropy": 1.344116109609604,
      "epoch": 0.576,
      "grad_norm": 26.375,
      "learning_rate": 1.795e-05,
      "loss": 21.5572,
      "mean_token_accuracy": 0.6739088613539934,
      "num_tokens": 7021844.0,
      "step": 360
    },
    {
      "entropy": 1.3490931887179614,
      "epoch": 0.592,
      "grad_norm": 26.125,
      "learning_rate": 1.845e-05,
      "loss": 21.7554,
      "mean_token_accuracy": 0.6715496938675642,
      "num_tokens": 7213951.0,
      "step": 370
    },
    {
      "entropy": 1.3264936488121748,
      "epoch": 0.608,
      "grad_norm": 27.25,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 21.3421,
      "mean_token_accuracy": 0.6760978367179632,
      "num_tokens": 7416773.0,
      "step": 380
    },
    {
      "entropy": 1.348380234092474,
      "epoch": 0.624,
      "grad_norm": 28.125,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 21.6293,
      "mean_token_accuracy": 0.6746726848185063,
      "num_tokens": 7612843.0,
      "step": 390
    },
    {
      "entropy": 1.3465636018663645,
      "epoch": 0.64,
      "grad_norm": 26.25,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 21.6324,
      "mean_token_accuracy": 0.6738044265657663,
      "num_tokens": 7801633.0,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_biology_entropy": 1.3016406693458558,
      "eval_biology_loss": 1.2880380153656006,
      "eval_biology_mean_token_accuracy": 0.6870926094055175,
      "eval_biology_num_tokens": 7801633.0,
      "eval_biology_runtime": 36.6216,
      "eval_biology_samples_per_second": 13.653,
      "eval_biology_steps_per_second": 3.413,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_chemistry_entropy": 0.9742186889648438,
      "eval_chemistry_loss": 0.954098641872406,
      "eval_chemistry_mean_token_accuracy": 0.7588451662063599,
      "eval_chemistry_num_tokens": 7801633.0,
      "eval_chemistry_runtime": 44.9361,
      "eval_chemistry_samples_per_second": 11.127,
      "eval_chemistry_steps_per_second": 2.782,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_math_entropy": 0.7798485422134399,
      "eval_math_loss": 1.0040466785430908,
      "eval_math_mean_token_accuracy": 0.7620619239807129,
      "eval_math_num_tokens": 7801633.0,
      "eval_math_runtime": 46.2072,
      "eval_math_samples_per_second": 10.821,
      "eval_math_steps_per_second": 2.705,
      "step": 400
    },
    {
      "epoch": 0.64,
      "eval_physics_entropy": 0.9594037117958069,
      "eval_physics_loss": 0.9417626261711121,
      "eval_physics_mean_token_accuracy": 0.7674873929023742,
      "eval_physics_num_tokens": 7801633.0,
      "eval_physics_runtime": 53.0463,
      "eval_physics_samples_per_second": 9.426,
      "eval_physics_steps_per_second": 2.356,
      "step": 400
    },
    {
      "entropy": 1.3527625199407338,
      "epoch": 0.656,
      "grad_norm": 27.875,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 21.651,
      "mean_token_accuracy": 0.6724541377276182,
      "num_tokens": 7995843.0,
      "step": 410
    },
    {
      "entropy": 1.3500655848532914,
      "epoch": 0.672,
      "grad_norm": 27.125,
      "learning_rate": 1.9894444444444446e-05,
      "loss": 21.6896,
      "mean_token_accuracy": 0.673105026781559,
      "num_tokens": 8183103.0,
      "step": 420
    },
    {
      "entropy": 1.3231504444032907,
      "epoch": 0.688,
      "grad_norm": 23.5,
      "learning_rate": 1.983888888888889e-05,
      "loss": 21.3727,
      "mean_token_accuracy": 0.6781029276549816,
      "num_tokens": 8385976.0,
      "step": 430
    },
    {
      "entropy": 1.3333053421229124,
      "epoch": 0.704,
      "grad_norm": 23.5,
      "learning_rate": 1.9783333333333334e-05,
      "loss": 21.2722,
      "mean_token_accuracy": 0.6781596153974533,
      "num_tokens": 8578431.0,
      "step": 440
    },
    {
      "entropy": 1.2669135373085738,
      "epoch": 0.72,
      "grad_norm": 23.375,
      "learning_rate": 1.972777777777778e-05,
      "loss": 20.3907,
      "mean_token_accuracy": 0.6872118715196848,
      "num_tokens": 8781342.0,
      "step": 450
    },
    {
      "entropy": 1.2877357687801123,
      "epoch": 0.736,
      "grad_norm": 23.875,
      "learning_rate": 1.9672222222222222e-05,
      "loss": 20.6999,
      "mean_token_accuracy": 0.6825123306363821,
      "num_tokens": 8977918.0,
      "step": 460
    },
    {
      "entropy": 1.3120017491281033,
      "epoch": 0.752,
      "grad_norm": 24.5,
      "learning_rate": 1.9616666666666668e-05,
      "loss": 21.1222,
      "mean_token_accuracy": 0.6796863108873368,
      "num_tokens": 9169322.0,
      "step": 470
    },
    {
      "entropy": 1.263747826963663,
      "epoch": 0.768,
      "grad_norm": 26.375,
      "learning_rate": 1.9561111111111114e-05,
      "loss": 20.3726,
      "mean_token_accuracy": 0.6869186259806156,
      "num_tokens": 9368141.0,
      "step": 480
    },
    {
      "entropy": 1.3065229691565037,
      "epoch": 0.784,
      "grad_norm": 26.5,
      "learning_rate": 1.9505555555555556e-05,
      "loss": 20.8536,
      "mean_token_accuracy": 0.683413552865386,
      "num_tokens": 9565236.0,
      "step": 490
    },
    {
      "entropy": 1.3189398009330033,
      "epoch": 0.8,
      "grad_norm": 25.5,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 21.2576,
      "mean_token_accuracy": 0.6789898939430714,
      "num_tokens": 9761227.0,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_biology_entropy": 1.2551557726860045,
      "eval_biology_loss": 1.2525163888931274,
      "eval_biology_mean_token_accuracy": 0.6931379227638245,
      "eval_biology_num_tokens": 9761227.0,
      "eval_biology_runtime": 36.6065,
      "eval_biology_samples_per_second": 13.659,
      "eval_biology_steps_per_second": 3.415,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_chemistry_entropy": 0.9375595455169677,
      "eval_chemistry_loss": 0.9276740550994873,
      "eval_chemistry_mean_token_accuracy": 0.7639822545051574,
      "eval_chemistry_num_tokens": 9761227.0,
      "eval_chemistry_runtime": 44.5343,
      "eval_chemistry_samples_per_second": 11.227,
      "eval_chemistry_steps_per_second": 2.807,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_math_entropy": 0.7748257055282592,
      "eval_math_loss": 0.9847215414047241,
      "eval_math_mean_token_accuracy": 0.7653367323875427,
      "eval_math_num_tokens": 9761227.0,
      "eval_math_runtime": 45.9419,
      "eval_math_samples_per_second": 10.883,
      "eval_math_steps_per_second": 2.721,
      "step": 500
    },
    {
      "epoch": 0.8,
      "eval_physics_entropy": 0.9384980492591858,
      "eval_physics_loss": 0.9171136021614075,
      "eval_physics_mean_token_accuracy": 0.7720685515403748,
      "eval_physics_num_tokens": 9761227.0,
      "eval_physics_runtime": 52.888,
      "eval_physics_samples_per_second": 9.454,
      "eval_physics_steps_per_second": 2.363,
      "step": 500
    },
    {
      "entropy": 1.2767818640917539,
      "epoch": 0.816,
      "grad_norm": 25.5,
      "learning_rate": 1.9394444444444448e-05,
      "loss": 20.4527,
      "mean_token_accuracy": 0.6852765031158924,
      "num_tokens": 9958727.0,
      "step": 510
    },
    {
      "entropy": 1.2603609178215265,
      "epoch": 0.832,
      "grad_norm": 23.0,
      "learning_rate": 1.933888888888889e-05,
      "loss": 20.3411,
      "mean_token_accuracy": 0.6862579733133316,
      "num_tokens": 10155771.0,
      "step": 520
    },
    {
      "entropy": 1.2630852952599525,
      "epoch": 0.848,
      "grad_norm": 26.375,
      "learning_rate": 1.9283333333333332e-05,
      "loss": 20.316,
      "mean_token_accuracy": 0.6886717110872269,
      "num_tokens": 10357721.0,
      "step": 530
    },
    {
      "entropy": 1.2910415723919868,
      "epoch": 0.864,
      "grad_norm": 25.0,
      "learning_rate": 1.9227777777777778e-05,
      "loss": 20.7324,
      "mean_token_accuracy": 0.6839874558150768,
      "num_tokens": 10552495.0,
      "step": 540
    },
    {
      "entropy": 1.2793519891798497,
      "epoch": 0.88,
      "grad_norm": 26.625,
      "learning_rate": 1.9172222222222224e-05,
      "loss": 20.4504,
      "mean_token_accuracy": 0.6864676021039486,
      "num_tokens": 10748749.0,
      "step": 550
    },
    {
      "entropy": 1.2906367748975753,
      "epoch": 0.896,
      "grad_norm": 24.875,
      "learning_rate": 1.911666666666667e-05,
      "loss": 20.7703,
      "mean_token_accuracy": 0.6835371732711792,
      "num_tokens": 10943319.0,
      "step": 560
    },
    {
      "entropy": 1.2809050466865302,
      "epoch": 0.912,
      "grad_norm": 26.0,
      "learning_rate": 1.9061111111111112e-05,
      "loss": 20.5319,
      "mean_token_accuracy": 0.6867790374904871,
      "num_tokens": 11136935.0,
      "step": 570
    },
    {
      "entropy": 1.283418832719326,
      "epoch": 0.928,
      "grad_norm": 23.875,
      "learning_rate": 1.9005555555555557e-05,
      "loss": 20.6537,
      "mean_token_accuracy": 0.6832854464650154,
      "num_tokens": 11331098.0,
      "step": 580
    },
    {
      "entropy": 1.2303649988025427,
      "epoch": 0.944,
      "grad_norm": 21.5,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 19.7901,
      "mean_token_accuracy": 0.6927386663854123,
      "num_tokens": 11530550.0,
      "step": 590
    },
    {
      "entropy": 1.2655960384756326,
      "epoch": 0.96,
      "grad_norm": 26.125,
      "learning_rate": 1.8894444444444446e-05,
      "loss": 20.362,
      "mean_token_accuracy": 0.6886723652482033,
      "num_tokens": 11729645.0,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_biology_entropy": 1.2511886906623841,
      "eval_biology_loss": 1.2273433208465576,
      "eval_biology_mean_token_accuracy": 0.6978591928482055,
      "eval_biology_num_tokens": 11729645.0,
      "eval_biology_runtime": 36.8245,
      "eval_biology_samples_per_second": 13.578,
      "eval_biology_steps_per_second": 3.394,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_chemistry_entropy": 0.9398184275627136,
      "eval_chemistry_loss": 0.9139939546585083,
      "eval_chemistry_mean_token_accuracy": 0.7667500405311585,
      "eval_chemistry_num_tokens": 11729645.0,
      "eval_chemistry_runtime": 45.002,
      "eval_chemistry_samples_per_second": 11.111,
      "eval_chemistry_steps_per_second": 2.778,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_math_entropy": 0.7878352689743042,
      "eval_math_loss": 0.9694741368293762,
      "eval_math_mean_token_accuracy": 0.767361653804779,
      "eval_math_num_tokens": 11729645.0,
      "eval_math_runtime": 46.2016,
      "eval_math_samples_per_second": 10.822,
      "eval_math_steps_per_second": 2.706,
      "step": 600
    },
    {
      "epoch": 0.96,
      "eval_physics_entropy": 0.9437616720199585,
      "eval_physics_loss": 0.9044761061668396,
      "eval_physics_mean_token_accuracy": 0.7742767133712769,
      "eval_physics_num_tokens": 11729645.0,
      "eval_physics_runtime": 53.0008,
      "eval_physics_samples_per_second": 9.434,
      "eval_physics_steps_per_second": 2.358,
      "step": 600
    },
    {
      "entropy": 1.2491024732589722,
      "epoch": 0.976,
      "grad_norm": 25.875,
      "learning_rate": 1.883888888888889e-05,
      "loss": 19.9555,
      "mean_token_accuracy": 0.6900846060365439,
      "num_tokens": 11924644.0,
      "step": 610
    },
    {
      "entropy": 1.2545549362897872,
      "epoch": 0.992,
      "grad_norm": 23.0,
      "learning_rate": 1.8783333333333337e-05,
      "loss": 20.317,
      "mean_token_accuracy": 0.686628095433116,
      "num_tokens": 12123059.0,
      "step": 620
    },
    {
      "entropy": 1.2396365478634834,
      "epoch": 1.008,
      "grad_norm": 22.5,
      "learning_rate": 1.872777777777778e-05,
      "loss": 19.6736,
      "mean_token_accuracy": 0.6955156117677689,
      "num_tokens": 12319366.0,
      "step": 630
    },
    {
      "entropy": 1.143751036375761,
      "epoch": 1.024,
      "grad_norm": 24.5,
      "learning_rate": 1.8672222222222222e-05,
      "loss": 18.51,
      "mean_token_accuracy": 0.7079542689025402,
      "num_tokens": 12524183.0,
      "step": 640
    },
    {
      "entropy": 1.1911311563104392,
      "epoch": 1.04,
      "grad_norm": 25.25,
      "learning_rate": 1.8616666666666667e-05,
      "loss": 19.0515,
      "mean_token_accuracy": 0.6993662625551224,
      "num_tokens": 12718593.0,
      "step": 650
    },
    {
      "entropy": 1.1548987552523613,
      "epoch": 1.056,
      "grad_norm": 25.0,
      "learning_rate": 1.8561111111111113e-05,
      "loss": 18.5231,
      "mean_token_accuracy": 0.707820488885045,
      "num_tokens": 12917803.0,
      "step": 660
    },
    {
      "entropy": 1.1886027652770281,
      "epoch": 1.072,
      "grad_norm": 26.0,
      "learning_rate": 1.8505555555555556e-05,
      "loss": 18.9314,
      "mean_token_accuracy": 0.7034936733543873,
      "num_tokens": 13105826.0,
      "step": 670
    },
    {
      "entropy": 1.1603233687579633,
      "epoch": 1.088,
      "grad_norm": 24.5,
      "learning_rate": 1.845e-05,
      "loss": 18.681,
      "mean_token_accuracy": 0.7059972729533911,
      "num_tokens": 13298619.0,
      "step": 680
    },
    {
      "entropy": 1.1554178480058908,
      "epoch": 1.104,
      "grad_norm": 24.0,
      "learning_rate": 1.8394444444444447e-05,
      "loss": 18.6001,
      "mean_token_accuracy": 0.7077722396701575,
      "num_tokens": 13491486.0,
      "step": 690
    },
    {
      "entropy": 1.1779750995337963,
      "epoch": 1.12,
      "grad_norm": 26.125,
      "learning_rate": 1.833888888888889e-05,
      "loss": 18.886,
      "mean_token_accuracy": 0.7039439618587494,
      "num_tokens": 13674663.0,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_biology_entropy": 1.115936914920807,
      "eval_biology_loss": 1.2197552919387817,
      "eval_biology_mean_token_accuracy": 0.6998812022209168,
      "eval_biology_num_tokens": 13674663.0,
      "eval_biology_runtime": 36.6382,
      "eval_biology_samples_per_second": 13.647,
      "eval_biology_steps_per_second": 3.412,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_chemistry_entropy": 0.8371734113693238,
      "eval_chemistry_loss": 0.9145113229751587,
      "eval_chemistry_mean_token_accuracy": 0.7663986759185791,
      "eval_chemistry_num_tokens": 13674663.0,
      "eval_chemistry_runtime": 45.0064,
      "eval_chemistry_samples_per_second": 11.11,
      "eval_chemistry_steps_per_second": 2.777,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_math_entropy": 0.7305963957309722,
      "eval_math_loss": 0.9764987826347351,
      "eval_math_mean_token_accuracy": 0.766656831741333,
      "eval_math_num_tokens": 13674663.0,
      "eval_math_runtime": 46.2254,
      "eval_math_samples_per_second": 10.817,
      "eval_math_steps_per_second": 2.704,
      "step": 700
    },
    {
      "epoch": 1.12,
      "eval_physics_entropy": 0.852522144317627,
      "eval_physics_loss": 0.9007405638694763,
      "eval_physics_mean_token_accuracy": 0.7750241093635559,
      "eval_physics_num_tokens": 13674663.0,
      "eval_physics_runtime": 53.0213,
      "eval_physics_samples_per_second": 9.43,
      "eval_physics_steps_per_second": 2.358,
      "step": 700
    },
    {
      "entropy": 1.1570473805069923,
      "epoch": 1.1360000000000001,
      "grad_norm": 30.875,
      "learning_rate": 1.8283333333333335e-05,
      "loss": 18.7252,
      "mean_token_accuracy": 0.7053195346146822,
      "num_tokens": 13869134.0,
      "step": 710
    },
    {
      "entropy": 1.153097653761506,
      "epoch": 1.152,
      "grad_norm": 23.125,
      "learning_rate": 1.822777777777778e-05,
      "loss": 18.3474,
      "mean_token_accuracy": 0.7079035714268684,
      "num_tokens": 14078365.0,
      "step": 720
    },
    {
      "entropy": 1.1558022141456603,
      "epoch": 1.168,
      "grad_norm": 25.375,
      "learning_rate": 1.8172222222222223e-05,
      "loss": 18.6158,
      "mean_token_accuracy": 0.706475806236267,
      "num_tokens": 14266831.0,
      "step": 730
    },
    {
      "entropy": 1.1689991034567355,
      "epoch": 1.184,
      "grad_norm": 24.625,
      "learning_rate": 1.811666666666667e-05,
      "loss": 18.7482,
      "mean_token_accuracy": 0.7053946666419506,
      "num_tokens": 14465660.0,
      "step": 740
    },
    {
      "entropy": 1.1589057791978121,
      "epoch": 1.2,
      "grad_norm": 26.125,
      "learning_rate": 1.806111111111111e-05,
      "loss": 18.4761,
      "mean_token_accuracy": 0.7090217288583517,
      "num_tokens": 14653228.0,
      "step": 750
    },
    {
      "entropy": 1.1286497369408608,
      "epoch": 1.216,
      "grad_norm": 24.625,
      "learning_rate": 1.8005555555555557e-05,
      "loss": 18.0937,
      "mean_token_accuracy": 0.710760310664773,
      "num_tokens": 14857782.0,
      "step": 760
    },
    {
      "entropy": 1.1454383581876755,
      "epoch": 1.232,
      "grad_norm": 23.0,
      "learning_rate": 1.795e-05,
      "loss": 18.3666,
      "mean_token_accuracy": 0.7111793085932732,
      "num_tokens": 15047356.0,
      "step": 770
    },
    {
      "entropy": 1.1489330865442753,
      "epoch": 1.248,
      "grad_norm": 24.5,
      "learning_rate": 1.7894444444444445e-05,
      "loss": 18.4807,
      "mean_token_accuracy": 0.7070541866123676,
      "num_tokens": 15241098.0,
      "step": 780
    },
    {
      "entropy": 1.1616758443415165,
      "epoch": 1.264,
      "grad_norm": 23.0,
      "learning_rate": 1.783888888888889e-05,
      "loss": 18.6418,
      "mean_token_accuracy": 0.7069107063114644,
      "num_tokens": 15437657.0,
      "step": 790
    },
    {
      "entropy": 1.1660578336566687,
      "epoch": 1.28,
      "grad_norm": 26.875,
      "learning_rate": 1.7783333333333333e-05,
      "loss": 18.8491,
      "mean_token_accuracy": 0.7026976983994245,
      "num_tokens": 15630795.0,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_biology_entropy": 1.1199735798835755,
      "eval_biology_loss": 1.2088388204574585,
      "eval_biology_mean_token_accuracy": 0.7015871267318725,
      "eval_biology_num_tokens": 15630795.0,
      "eval_biology_runtime": 36.9629,
      "eval_biology_samples_per_second": 13.527,
      "eval_biology_steps_per_second": 3.382,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_chemistry_entropy": 0.8543037304878235,
      "eval_chemistry_loss": 0.9098371863365173,
      "eval_chemistry_mean_token_accuracy": 0.7672378678321838,
      "eval_chemistry_num_tokens": 15630795.0,
      "eval_chemistry_runtime": 44.9998,
      "eval_chemistry_samples_per_second": 11.111,
      "eval_chemistry_steps_per_second": 2.778,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_math_entropy": 0.7368179533481598,
      "eval_math_loss": 0.9706032872200012,
      "eval_math_mean_token_accuracy": 0.7682779974937439,
      "eval_math_num_tokens": 15630795.0,
      "eval_math_runtime": 46.2307,
      "eval_math_samples_per_second": 10.815,
      "eval_math_steps_per_second": 2.704,
      "step": 800
    },
    {
      "epoch": 1.28,
      "eval_physics_entropy": 0.8641948227882386,
      "eval_physics_loss": 0.8950499296188354,
      "eval_physics_mean_token_accuracy": 0.7758663225173951,
      "eval_physics_num_tokens": 15630795.0,
      "eval_physics_runtime": 53.0442,
      "eval_physics_samples_per_second": 9.426,
      "eval_physics_steps_per_second": 2.357,
      "step": 800
    },
    {
      "entropy": 1.1806952260434627,
      "epoch": 1.296,
      "grad_norm": 26.125,
      "learning_rate": 1.772777777777778e-05,
      "loss": 18.9295,
      "mean_token_accuracy": 0.70291638225317,
      "num_tokens": 15827105.0,
      "step": 810
    },
    {
      "entropy": 1.1602427676320075,
      "epoch": 1.312,
      "grad_norm": 27.625,
      "learning_rate": 1.7672222222222224e-05,
      "loss": 18.7911,
      "mean_token_accuracy": 0.7041201893240213,
      "num_tokens": 16019645.0,
      "step": 820
    },
    {
      "entropy": 1.1715985681861638,
      "epoch": 1.328,
      "grad_norm": 24.375,
      "learning_rate": 1.761666666666667e-05,
      "loss": 18.8262,
      "mean_token_accuracy": 0.7034816723316908,
      "num_tokens": 16221726.0,
      "step": 830
    },
    {
      "entropy": 1.1461844369769096,
      "epoch": 1.3439999999999999,
      "grad_norm": 25.5,
      "learning_rate": 1.7561111111111113e-05,
      "loss": 18.2642,
      "mean_token_accuracy": 0.7110818449407816,
      "num_tokens": 16427594.0,
      "step": 840
    },
    {
      "entropy": 1.1582913786172866,
      "epoch": 1.3599999999999999,
      "grad_norm": 25.25,
      "learning_rate": 1.7505555555555558e-05,
      "loss": 18.5772,
      "mean_token_accuracy": 0.7078793905675411,
      "num_tokens": 16621605.0,
      "step": 850
    },
    {
      "entropy": 1.1942507456988096,
      "epoch": 1.376,
      "grad_norm": 27.875,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 19.0764,
      "mean_token_accuracy": 0.701082956790924,
      "num_tokens": 16813444.0,
      "step": 860
    },
    {
      "entropy": 1.1594038642942905,
      "epoch": 1.392,
      "grad_norm": 24.125,
      "learning_rate": 1.7394444444444446e-05,
      "loss": 18.7411,
      "mean_token_accuracy": 0.7061580993235111,
      "num_tokens": 17006509.0,
      "step": 870
    },
    {
      "entropy": 1.1641225919127465,
      "epoch": 1.408,
      "grad_norm": 21.5,
      "learning_rate": 1.733888888888889e-05,
      "loss": 18.5028,
      "mean_token_accuracy": 0.7065019059926272,
      "num_tokens": 17197870.0,
      "step": 880
    },
    {
      "entropy": 1.153414449468255,
      "epoch": 1.424,
      "grad_norm": 24.125,
      "learning_rate": 1.7283333333333334e-05,
      "loss": 18.5571,
      "mean_token_accuracy": 0.7072413295507431,
      "num_tokens": 17394390.0,
      "step": 890
    },
    {
      "entropy": 1.1581655487418174,
      "epoch": 1.44,
      "grad_norm": 28.5,
      "learning_rate": 1.722777777777778e-05,
      "loss": 18.5762,
      "mean_token_accuracy": 0.7053147312253714,
      "num_tokens": 17587777.0,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_biology_entropy": 1.1328148250579835,
      "eval_biology_loss": 1.1959257125854492,
      "eval_biology_mean_token_accuracy": 0.7035633249282837,
      "eval_biology_num_tokens": 17587777.0,
      "eval_biology_runtime": 38.7737,
      "eval_biology_samples_per_second": 12.895,
      "eval_biology_steps_per_second": 3.224,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_chemistry_entropy": 0.8504280371665954,
      "eval_chemistry_loss": 0.9063042402267456,
      "eval_chemistry_mean_token_accuracy": 0.7678593263626099,
      "eval_chemistry_num_tokens": 17587777.0,
      "eval_chemistry_runtime": 44.9861,
      "eval_chemistry_samples_per_second": 11.115,
      "eval_chemistry_steps_per_second": 2.779,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_math_entropy": 0.7459685401916504,
      "eval_math_loss": 0.9690774083137512,
      "eval_math_mean_token_accuracy": 0.7676572284698486,
      "eval_math_num_tokens": 17587777.0,
      "eval_math_runtime": 82.6785,
      "eval_math_samples_per_second": 6.048,
      "eval_math_steps_per_second": 1.512,
      "step": 900
    },
    {
      "epoch": 1.44,
      "eval_physics_entropy": 0.8569722015857697,
      "eval_physics_loss": 0.8932082653045654,
      "eval_physics_mean_token_accuracy": 0.7766451034545898,
      "eval_physics_num_tokens": 17587777.0,
      "eval_physics_runtime": 52.7335,
      "eval_physics_samples_per_second": 9.482,
      "eval_physics_steps_per_second": 2.37,
      "step": 900
    },
    {
      "entropy": 1.1505886979401112,
      "epoch": 1.456,
      "grad_norm": 24.25,
      "learning_rate": 1.7172222222222223e-05,
      "loss": 18.5378,
      "mean_token_accuracy": 0.7084831334650517,
      "num_tokens": 17788456.0,
      "step": 910
    },
    {
      "entropy": 1.1548077858984471,
      "epoch": 1.472,
      "grad_norm": 24.125,
      "learning_rate": 1.7116666666666668e-05,
      "loss": 18.4342,
      "mean_token_accuracy": 0.7085961502045393,
      "num_tokens": 17984063.0,
      "step": 920
    },
    {
      "entropy": 1.1482487738132476,
      "epoch": 1.488,
      "grad_norm": 24.0,
      "learning_rate": 1.7061111111111114e-05,
      "loss": 18.4515,
      "mean_token_accuracy": 0.7099530510604382,
      "num_tokens": 18175640.0,
      "step": 930
    },
    {
      "entropy": 1.1528287250548601,
      "epoch": 1.504,
      "grad_norm": 23.375,
      "learning_rate": 1.7005555555555556e-05,
      "loss": 18.5572,
      "mean_token_accuracy": 0.7075925685465336,
      "num_tokens": 18367857.0,
      "step": 940
    },
    {
      "entropy": 1.1533676456660031,
      "epoch": 1.52,
      "grad_norm": 23.5,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 18.4716,
      "mean_token_accuracy": 0.7079793553799391,
      "num_tokens": 18569146.0,
      "step": 950
    },
    {
      "entropy": 1.1715267714112998,
      "epoch": 1.536,
      "grad_norm": 26.25,
      "learning_rate": 1.6894444444444448e-05,
      "loss": 18.754,
      "mean_token_accuracy": 0.7050810724496841,
      "num_tokens": 18755079.0,
      "step": 960
    },
    {
      "entropy": 1.1350098505616188,
      "epoch": 1.552,
      "grad_norm": 26.0,
      "learning_rate": 1.683888888888889e-05,
      "loss": 18.4393,
      "mean_token_accuracy": 0.7091891258955002,
      "num_tokens": 18956248.0,
      "step": 970
    },
    {
      "entropy": 1.1361894980072975,
      "epoch": 1.568,
      "grad_norm": 25.625,
      "learning_rate": 1.6783333333333332e-05,
      "loss": 18.2272,
      "mean_token_accuracy": 0.711289395019412,
      "num_tokens": 19150315.0,
      "step": 980
    },
    {
      "entropy": 1.1439239967614412,
      "epoch": 1.584,
      "grad_norm": 26.125,
      "learning_rate": 1.6727777777777778e-05,
      "loss": 18.3302,
      "mean_token_accuracy": 0.7097273994237184,
      "num_tokens": 19344260.0,
      "step": 990
    },
    {
      "entropy": 1.1727315995842218,
      "epoch": 1.6,
      "grad_norm": 23.625,
      "learning_rate": 1.6672222222222224e-05,
      "loss": 18.7555,
      "mean_token_accuracy": 0.7035728722810746,
      "num_tokens": 19532552.0,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_biology_entropy": 1.0895025281906128,
      "eval_biology_loss": 1.1883716583251953,
      "eval_biology_mean_token_accuracy": 0.705087408065796,
      "eval_biology_num_tokens": 19532552.0,
      "eval_biology_runtime": 36.662,
      "eval_biology_samples_per_second": 13.638,
      "eval_biology_steps_per_second": 3.41,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_chemistry_entropy": 0.8066284563541413,
      "eval_chemistry_loss": 0.9020781517028809,
      "eval_chemistry_mean_token_accuracy": 0.7695949592590332,
      "eval_chemistry_num_tokens": 19532552.0,
      "eval_chemistry_runtime": 44.602,
      "eval_chemistry_samples_per_second": 11.21,
      "eval_chemistry_steps_per_second": 2.803,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_math_entropy": 0.6928402831554413,
      "eval_math_loss": 0.9739019870758057,
      "eval_math_mean_token_accuracy": 0.7690589771270752,
      "eval_math_num_tokens": 19532552.0,
      "eval_math_runtime": 46.1135,
      "eval_math_samples_per_second": 10.843,
      "eval_math_steps_per_second": 2.711,
      "step": 1000
    },
    {
      "epoch": 1.6,
      "eval_physics_entropy": 0.8003968651294708,
      "eval_physics_loss": 0.889771044254303,
      "eval_physics_mean_token_accuracy": 0.7777439093589783,
      "eval_physics_num_tokens": 19532552.0,
      "eval_physics_runtime": 53.0769,
      "eval_physics_samples_per_second": 9.42,
      "eval_physics_steps_per_second": 2.355,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.2020838985302374e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}