{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.2,
  "eval_steps": 100,
  "global_step": 2000,
  "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
    },
    {
      "entropy": 1.1351401790976525,
      "epoch": 1.616,
      "grad_norm": 23.375,
      "learning_rate": 1.6616666666666666e-05,
      "loss": 18.25,
      "mean_token_accuracy": 0.7113314840942622,
      "num_tokens": 19732719.0,
      "step": 1010
    },
    {
      "entropy": 1.1716969449073076,
      "epoch": 1.6320000000000001,
      "grad_norm": 26.75,
      "learning_rate": 1.6561111111111112e-05,
      "loss": 18.7782,
      "mean_token_accuracy": 0.7055815242230892,
      "num_tokens": 19926830.0,
      "step": 1020
    },
    {
      "entropy": 1.1589280366897583,
      "epoch": 1.6480000000000001,
      "grad_norm": 22.5,
      "learning_rate": 1.6505555555555558e-05,
      "loss": 18.5404,
      "mean_token_accuracy": 0.7074035726487636,
      "num_tokens": 20118800.0,
      "step": 1030
    },
    {
      "entropy": 1.1163798790425061,
      "epoch": 1.6640000000000001,
      "grad_norm": 22.25,
      "learning_rate": 1.645e-05,
      "loss": 18.0098,
      "mean_token_accuracy": 0.7143486425280571,
      "num_tokens": 20320511.0,
      "step": 1040
    },
    {
      "entropy": 1.143330230936408,
      "epoch": 1.6800000000000002,
      "grad_norm": 24.625,
      "learning_rate": 1.6394444444444446e-05,
      "loss": 18.3667,
      "mean_token_accuracy": 0.7101939171552658,
      "num_tokens": 20513393.0,
      "step": 1050
    },
    {
      "entropy": 1.1441194768995047,
      "epoch": 1.696,
      "grad_norm": 24.625,
      "learning_rate": 1.633888888888889e-05,
      "loss": 18.3598,
      "mean_token_accuracy": 0.7109571672976017,
      "num_tokens": 20707237.0,
      "step": 1060
    },
    {
      "entropy": 1.1500865355134011,
      "epoch": 1.712,
      "grad_norm": 25.125,
      "learning_rate": 1.6283333333333334e-05,
      "loss": 18.3793,
      "mean_token_accuracy": 0.7098943468183279,
      "num_tokens": 20910419.0,
      "step": 1070
    },
    {
      "entropy": 1.151506431400776,
      "epoch": 1.728,
      "grad_norm": 23.875,
      "learning_rate": 1.622777777777778e-05,
      "loss": 18.6016,
      "mean_token_accuracy": 0.706824541464448,
      "num_tokens": 21107498.0,
      "step": 1080
    },
    {
      "entropy": 1.1211910132318734,
      "epoch": 1.744,
      "grad_norm": 23.125,
      "learning_rate": 1.6172222222222222e-05,
      "loss": 17.943,
      "mean_token_accuracy": 0.7139425069093704,
      "num_tokens": 21303955.0,
      "step": 1090
    },
    {
      "entropy": 1.135467940941453,
      "epoch": 1.76,
      "grad_norm": 26.625,
      "learning_rate": 1.6116666666666668e-05,
      "loss": 18.3203,
      "mean_token_accuracy": 0.711310513690114,
      "num_tokens": 21499572.0,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_biology_entropy": 1.128408634185791,
      "eval_biology_loss": 1.1783721446990967,
      "eval_biology_mean_token_accuracy": 0.7069245343208312,
      "eval_biology_num_tokens": 21499572.0,
      "eval_biology_runtime": 36.7197,
      "eval_biology_samples_per_second": 13.617,
      "eval_biology_steps_per_second": 3.404,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_chemistry_entropy": 0.8473764100074768,
      "eval_chemistry_loss": 0.8974772691726685,
      "eval_chemistry_mean_token_accuracy": 0.7699919505119324,
      "eval_chemistry_num_tokens": 21499572.0,
      "eval_chemistry_runtime": 45.1422,
      "eval_chemistry_samples_per_second": 11.076,
      "eval_chemistry_steps_per_second": 2.769,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_math_entropy": 0.7425255510807037,
      "eval_math_loss": 0.964664101600647,
      "eval_math_mean_token_accuracy": 0.7689636731147766,
      "eval_math_num_tokens": 21499572.0,
      "eval_math_runtime": 46.3198,
      "eval_math_samples_per_second": 10.795,
      "eval_math_steps_per_second": 2.699,
      "step": 1100
    },
    {
      "epoch": 1.76,
      "eval_physics_entropy": 0.8533617742061615,
      "eval_physics_loss": 0.8856339454650879,
      "eval_physics_mean_token_accuracy": 0.7777909712791443,
      "eval_physics_num_tokens": 21499572.0,
      "eval_physics_runtime": 53.4691,
      "eval_physics_samples_per_second": 9.351,
      "eval_physics_steps_per_second": 2.338,
      "step": 1100
    },
    {
      "entropy": 1.1444791179150342,
      "epoch": 1.776,
      "grad_norm": 25.125,
      "learning_rate": 1.6061111111111113e-05,
      "loss": 18.4154,
      "mean_token_accuracy": 0.7087540179491043,
      "num_tokens": 21692804.0,
      "step": 1110
    },
    {
      "entropy": 1.1155106097459793,
      "epoch": 1.792,
      "grad_norm": 26.75,
      "learning_rate": 1.6005555555555556e-05,
      "loss": 17.8771,
      "mean_token_accuracy": 0.7151219062507153,
      "num_tokens": 21894218.0,
      "step": 1120
    },
    {
      "entropy": 1.1439542584121227,
      "epoch": 1.808,
      "grad_norm": 26.0,
      "learning_rate": 1.595e-05,
      "loss": 18.2808,
      "mean_token_accuracy": 0.7120830215513706,
      "num_tokens": 22082522.0,
      "step": 1130
    },
    {
      "entropy": 1.164598362520337,
      "epoch": 1.8239999999999998,
      "grad_norm": 24.875,
      "learning_rate": 1.5894444444444447e-05,
      "loss": 18.6609,
      "mean_token_accuracy": 0.7053720869123936,
      "num_tokens": 22278933.0,
      "step": 1140
    },
    {
      "entropy": 1.1762568525969983,
      "epoch": 1.8399999999999999,
      "grad_norm": 25.875,
      "learning_rate": 1.583888888888889e-05,
      "loss": 18.8552,
      "mean_token_accuracy": 0.7038435664027929,
      "num_tokens": 22473801.0,
      "step": 1150
    },
    {
      "entropy": 1.124564103782177,
      "epoch": 1.8559999999999999,
      "grad_norm": 24.625,
      "learning_rate": 1.5783333333333335e-05,
      "loss": 18.1974,
      "mean_token_accuracy": 0.7101016439497471,
      "num_tokens": 22677853.0,
      "step": 1160
    },
    {
      "entropy": 1.1182593822479248,
      "epoch": 1.8719999999999999,
      "grad_norm": 24.75,
      "learning_rate": 1.572777777777778e-05,
      "loss": 17.8834,
      "mean_token_accuracy": 0.7157654833048582,
      "num_tokens": 22874965.0,
      "step": 1170
    },
    {
      "entropy": 1.1652609888464212,
      "epoch": 1.888,
      "grad_norm": 26.375,
      "learning_rate": 1.5672222222222223e-05,
      "loss": 18.6854,
      "mean_token_accuracy": 0.7054859042167664,
      "num_tokens": 23068892.0,
      "step": 1180
    },
    {
      "entropy": 1.1525317531079053,
      "epoch": 1.904,
      "grad_norm": 25.5,
      "learning_rate": 1.561666666666667e-05,
      "loss": 18.4917,
      "mean_token_accuracy": 0.7093732491135597,
      "num_tokens": 23263827.0,
      "step": 1190
    },
    {
      "entropy": 1.1493665751069784,
      "epoch": 1.92,
      "grad_norm": 23.25,
      "learning_rate": 1.556111111111111e-05,
      "loss": 18.5319,
      "mean_token_accuracy": 0.7083175975829363,
      "num_tokens": 23463627.0,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_biology_entropy": 1.1280886573791504,
      "eval_biology_loss": 1.168996810913086,
      "eval_biology_mean_token_accuracy": 0.7084329080581665,
      "eval_biology_num_tokens": 23463627.0,
      "eval_biology_runtime": 36.7535,
      "eval_biology_samples_per_second": 13.604,
      "eval_biology_steps_per_second": 3.401,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_chemistry_entropy": 0.8494512581825256,
      "eval_chemistry_loss": 0.8931704759597778,
      "eval_chemistry_mean_token_accuracy": 0.7704657254219055,
      "eval_chemistry_num_tokens": 23463627.0,
      "eval_chemistry_runtime": 45.1599,
      "eval_chemistry_samples_per_second": 11.072,
      "eval_chemistry_steps_per_second": 2.768,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_math_entropy": 0.7389046318531036,
      "eval_math_loss": 0.9631180167198181,
      "eval_math_mean_token_accuracy": 0.7691067876815796,
      "eval_math_num_tokens": 23463627.0,
      "eval_math_runtime": 46.3688,
      "eval_math_samples_per_second": 10.783,
      "eval_math_steps_per_second": 2.696,
      "step": 1200
    },
    {
      "epoch": 1.92,
      "eval_physics_entropy": 0.8575768961906433,
      "eval_physics_loss": 0.8806874752044678,
      "eval_physics_mean_token_accuracy": 0.7783133115768432,
      "eval_physics_num_tokens": 23463627.0,
      "eval_physics_runtime": 53.2169,
      "eval_physics_samples_per_second": 9.396,
      "eval_physics_steps_per_second": 2.349,
      "step": 1200
    },
    {
      "entropy": 1.119696567952633,
      "epoch": 1.936,
      "grad_norm": 25.875,
      "learning_rate": 1.5505555555555557e-05,
      "loss": 17.8962,
      "mean_token_accuracy": 0.7145485125482083,
      "num_tokens": 23660418.0,
      "step": 1210
    },
    {
      "entropy": 1.1398029860109091,
      "epoch": 1.952,
      "grad_norm": 24.25,
      "learning_rate": 1.545e-05,
      "loss": 18.2843,
      "mean_token_accuracy": 0.7102924332022666,
      "num_tokens": 23858145.0,
      "step": 1220
    },
    {
      "entropy": 1.1528537224978208,
      "epoch": 1.968,
      "grad_norm": 25.0,
      "learning_rate": 1.5394444444444445e-05,
      "loss": 18.4096,
      "mean_token_accuracy": 0.7092894580215215,
      "num_tokens": 24053364.0,
      "step": 1230
    },
    {
      "entropy": 1.130320394039154,
      "epoch": 1.984,
      "grad_norm": 24.0,
      "learning_rate": 1.533888888888889e-05,
      "loss": 18.2555,
      "mean_token_accuracy": 0.7108469780534505,
      "num_tokens": 24249465.0,
      "step": 1240
    },
    {
      "entropy": 1.1447110261768103,
      "epoch": 2.0,
      "grad_norm": 25.375,
      "learning_rate": 1.5283333333333333e-05,
      "loss": 18.3602,
      "mean_token_accuracy": 0.7106993485242128,
      "num_tokens": 24442582.0,
      "step": 1250
    },
    {
      "entropy": 1.018304006755352,
      "epoch": 2.016,
      "grad_norm": 32.25,
      "learning_rate": 1.5227777777777779e-05,
      "loss": 16.0943,
      "mean_token_accuracy": 0.736331307515502,
      "num_tokens": 24632353.0,
      "step": 1260
    },
    {
      "entropy": 1.0125610791146755,
      "epoch": 2.032,
      "grad_norm": 29.25,
      "learning_rate": 1.5172222222222223e-05,
      "loss": 16.1426,
      "mean_token_accuracy": 0.7360313270241022,
      "num_tokens": 24822715.0,
      "step": 1270
    },
    {
      "entropy": 0.9830857388675213,
      "epoch": 2.048,
      "grad_norm": 30.0,
      "learning_rate": 1.5116666666666669e-05,
      "loss": 15.7118,
      "mean_token_accuracy": 0.7416322931647301,
      "num_tokens": 25023118.0,
      "step": 1280
    },
    {
      "entropy": 0.9931403815746307,
      "epoch": 2.064,
      "grad_norm": 29.25,
      "learning_rate": 1.5061111111111113e-05,
      "loss": 15.8433,
      "mean_token_accuracy": 0.7399392265826463,
      "num_tokens": 25217414.0,
      "step": 1290
    },
    {
      "entropy": 0.9739402402192354,
      "epoch": 2.08,
      "grad_norm": 28.875,
      "learning_rate": 1.5005555555555557e-05,
      "loss": 15.654,
      "mean_token_accuracy": 0.7406358070671558,
      "num_tokens": 25408961.0,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_biology_entropy": 1.0040340127944947,
      "eval_biology_loss": 1.2068090438842773,
      "eval_biology_mean_token_accuracy": 0.7046228342056274,
      "eval_biology_num_tokens": 25408961.0,
      "eval_biology_runtime": 36.7953,
      "eval_biology_samples_per_second": 13.589,
      "eval_biology_steps_per_second": 3.397,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_chemistry_entropy": 0.7516332559585571,
      "eval_chemistry_loss": 0.9387209415435791,
      "eval_chemistry_mean_token_accuracy": 0.7650035581588746,
      "eval_chemistry_num_tokens": 25408961.0,
      "eval_chemistry_runtime": 45.2093,
      "eval_chemistry_samples_per_second": 11.06,
      "eval_chemistry_steps_per_second": 2.765,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_math_entropy": 0.6802132661342621,
      "eval_math_loss": 1.0038195848464966,
      "eval_math_mean_token_accuracy": 0.7657432436943055,
      "eval_math_num_tokens": 25408961.0,
      "eval_math_runtime": 46.4381,
      "eval_math_samples_per_second": 10.767,
      "eval_math_steps_per_second": 2.692,
      "step": 1300
    },
    {
      "epoch": 2.08,
      "eval_physics_entropy": 0.7668055052757263,
      "eval_physics_loss": 0.9190961718559265,
      "eval_physics_mean_token_accuracy": 0.774013831615448,
      "eval_physics_num_tokens": 25408961.0,
      "eval_physics_runtime": 53.1969,
      "eval_physics_samples_per_second": 9.399,
      "eval_physics_steps_per_second": 2.35,
      "step": 1300
    },
    {
      "entropy": 0.9855834160000085,
      "epoch": 2.096,
      "grad_norm": 30.125,
      "learning_rate": 1.4950000000000003e-05,
      "loss": 15.7302,
      "mean_token_accuracy": 0.7416285384446383,
      "num_tokens": 25600511.0,
      "step": 1310
    },
    {
      "entropy": 0.9560547668486834,
      "epoch": 2.112,
      "grad_norm": 30.0,
      "learning_rate": 1.4894444444444445e-05,
      "loss": 15.2923,
      "mean_token_accuracy": 0.7467750016599893,
      "num_tokens": 25796565.0,
      "step": 1320
    },
    {
      "entropy": 0.9734839156270028,
      "epoch": 2.128,
      "grad_norm": 28.375,
      "learning_rate": 1.4838888888888889e-05,
      "loss": 15.4659,
      "mean_token_accuracy": 0.7434854444116354,
      "num_tokens": 25991156.0,
      "step": 1330
    },
    {
      "entropy": 0.9701934043318033,
      "epoch": 2.144,
      "grad_norm": 32.0,
      "learning_rate": 1.4783333333333335e-05,
      "loss": 15.6079,
      "mean_token_accuracy": 0.742589833214879,
      "num_tokens": 26193237.0,
      "step": 1340
    },
    {
      "entropy": 0.9971341345459223,
      "epoch": 2.16,
      "grad_norm": 30.875,
      "learning_rate": 1.4727777777777779e-05,
      "loss": 15.9608,
      "mean_token_accuracy": 0.7389167688786984,
      "num_tokens": 26387993.0,
      "step": 1350
    },
    {
      "entropy": 1.0051845651119948,
      "epoch": 2.176,
      "grad_norm": 30.125,
      "learning_rate": 1.4672222222222223e-05,
      "loss": 16.0894,
      "mean_token_accuracy": 0.736288172006607,
      "num_tokens": 26585269.0,
      "step": 1360
    },
    {
      "entropy": 0.956790243089199,
      "epoch": 2.192,
      "grad_norm": 27.0,
      "learning_rate": 1.4616666666666668e-05,
      "loss": 15.321,
      "mean_token_accuracy": 0.7460867285728454,
      "num_tokens": 26776318.0,
      "step": 1370
    },
    {
      "entropy": 0.9826645214110613,
      "epoch": 2.208,
      "grad_norm": 29.0,
      "learning_rate": 1.4561111111111112e-05,
      "loss": 15.6504,
      "mean_token_accuracy": 0.7417031954973936,
      "num_tokens": 26974420.0,
      "step": 1380
    },
    {
      "entropy": 1.0038280066102743,
      "epoch": 2.224,
      "grad_norm": 28.375,
      "learning_rate": 1.4505555555555557e-05,
      "loss": 16.0751,
      "mean_token_accuracy": 0.7366515222936869,
      "num_tokens": 27167577.0,
      "step": 1390
    },
    {
      "entropy": 0.9645163491368294,
      "epoch": 2.24,
      "grad_norm": 29.625,
      "learning_rate": 1.4450000000000002e-05,
      "loss": 15.4475,
      "mean_token_accuracy": 0.7438073411583901,
      "num_tokens": 27364189.0,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_biology_entropy": 0.981437744140625,
      "eval_biology_loss": 1.207616925239563,
      "eval_biology_mean_token_accuracy": 0.704860345363617,
      "eval_biology_num_tokens": 27364189.0,
      "eval_biology_runtime": 37.0146,
      "eval_biology_samples_per_second": 13.508,
      "eval_biology_steps_per_second": 3.377,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_chemistry_entropy": 0.7361237547397613,
      "eval_chemistry_loss": 0.9404470324516296,
      "eval_chemistry_mean_token_accuracy": 0.7649986271858216,
      "eval_chemistry_num_tokens": 27364189.0,
      "eval_chemistry_runtime": 45.8584,
      "eval_chemistry_samples_per_second": 10.903,
      "eval_chemistry_steps_per_second": 2.726,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_math_entropy": 0.6734062578678132,
      "eval_math_loss": 1.008291482925415,
      "eval_math_mean_token_accuracy": 0.7654335222244263,
      "eval_math_num_tokens": 27364189.0,
      "eval_math_runtime": 46.5839,
      "eval_math_samples_per_second": 10.733,
      "eval_math_steps_per_second": 2.683,
      "step": 1400
    },
    {
      "epoch": 2.24,
      "eval_physics_entropy": 0.7516936962604522,
      "eval_physics_loss": 0.9206180572509766,
      "eval_physics_mean_token_accuracy": 0.7742253971099854,
      "eval_physics_num_tokens": 27364189.0,
      "eval_physics_runtime": 53.4062,
      "eval_physics_samples_per_second": 9.362,
      "eval_physics_steps_per_second": 2.341,
      "step": 1400
    },
    {
      "entropy": 0.95200383849442,
      "epoch": 2.2560000000000002,
      "grad_norm": 28.25,
      "learning_rate": 1.4394444444444446e-05,
      "loss": 15.29,
      "mean_token_accuracy": 0.7460251439362764,
      "num_tokens": 27557387.0,
      "step": 1410
    },
    {
      "entropy": 0.9579467125236988,
      "epoch": 2.2720000000000002,
      "grad_norm": 29.75,
      "learning_rate": 1.433888888888889e-05,
      "loss": 15.3949,
      "mean_token_accuracy": 0.7428190916776657,
      "num_tokens": 27755725.0,
      "step": 1420
    },
    {
      "entropy": 1.0049206741154193,
      "epoch": 2.288,
      "grad_norm": 29.375,
      "learning_rate": 1.4283333333333333e-05,
      "loss": 16.0165,
      "mean_token_accuracy": 0.7357172459363938,
      "num_tokens": 27950694.0,
      "step": 1430
    },
    {
      "entropy": 1.005023618414998,
      "epoch": 2.304,
      "grad_norm": 27.375,
      "learning_rate": 1.4227777777777778e-05,
      "loss": 16.031,
      "mean_token_accuracy": 0.7371877241879702,
      "num_tokens": 28150719.0,
      "step": 1440
    },
    {
      "entropy": 0.9545973517000675,
      "epoch": 2.32,
      "grad_norm": 27.375,
      "learning_rate": 1.4172222222222222e-05,
      "loss": 15.3343,
      "mean_token_accuracy": 0.7450725197792053,
      "num_tokens": 28348652.0,
      "step": 1450
    },
    {
      "entropy": 0.9779899045825005,
      "epoch": 2.336,
      "grad_norm": 29.125,
      "learning_rate": 1.4116666666666668e-05,
      "loss": 15.6932,
      "mean_token_accuracy": 0.7409903839230537,
      "num_tokens": 28542945.0,
      "step": 1460
    },
    {
      "entropy": 0.9990467399358749,
      "epoch": 2.352,
      "grad_norm": 30.0,
      "learning_rate": 1.4061111111111112e-05,
      "loss": 16.0289,
      "mean_token_accuracy": 0.737147618085146,
      "num_tokens": 28731927.0,
      "step": 1470
    },
    {
      "entropy": 0.9743069138377904,
      "epoch": 2.368,
      "grad_norm": 28.25,
      "learning_rate": 1.4005555555555556e-05,
      "loss": 15.5588,
      "mean_token_accuracy": 0.7433720890432596,
      "num_tokens": 28929298.0,
      "step": 1480
    },
    {
      "entropy": 0.9916814502328635,
      "epoch": 2.384,
      "grad_norm": 30.875,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 15.7992,
      "mean_token_accuracy": 0.738391650095582,
      "num_tokens": 29121142.0,
      "step": 1490
    },
    {
      "entropy": 0.9762042462825775,
      "epoch": 2.4,
      "grad_norm": 27.75,
      "learning_rate": 1.3894444444444446e-05,
      "loss": 15.6866,
      "mean_token_accuracy": 0.7411531064659357,
      "num_tokens": 29317919.0,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_biology_entropy": 0.9761668286323547,
      "eval_biology_loss": 1.204858422279358,
      "eval_biology_mean_token_accuracy": 0.7061952071189881,
      "eval_biology_num_tokens": 29317919.0,
      "eval_biology_runtime": 36.8411,
      "eval_biology_samples_per_second": 13.572,
      "eval_biology_steps_per_second": 3.393,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_chemistry_entropy": 0.7322783548831939,
      "eval_chemistry_loss": 0.9399828314781189,
      "eval_chemistry_mean_token_accuracy": 0.7658818020820618,
      "eval_chemistry_num_tokens": 29317919.0,
      "eval_chemistry_runtime": 44.9217,
      "eval_chemistry_samples_per_second": 11.13,
      "eval_chemistry_steps_per_second": 2.783,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_math_entropy": 0.6698096024990082,
      "eval_math_loss": 1.0045055150985718,
      "eval_math_mean_token_accuracy": 0.7665768723487854,
      "eval_math_num_tokens": 29317919.0,
      "eval_math_runtime": 46.3425,
      "eval_math_samples_per_second": 10.789,
      "eval_math_steps_per_second": 2.697,
      "step": 1500
    },
    {
      "epoch": 2.4,
      "eval_physics_entropy": 0.7526926681995392,
      "eval_physics_loss": 0.9185658097267151,
      "eval_physics_mean_token_accuracy": 0.7750369157791138,
      "eval_physics_num_tokens": 29317919.0,
      "eval_physics_runtime": 53.26,
      "eval_physics_samples_per_second": 9.388,
      "eval_physics_steps_per_second": 2.347,
      "step": 1500
    },
    {
      "entropy": 0.9726948961615562,
      "epoch": 2.416,
      "grad_norm": 30.75,
      "learning_rate": 1.383888888888889e-05,
      "loss": 15.6556,
      "mean_token_accuracy": 0.7424319490790368,
      "num_tokens": 29518541.0,
      "step": 1510
    },
    {
      "entropy": 0.976997298002243,
      "epoch": 2.432,
      "grad_norm": 32.0,
      "learning_rate": 1.3783333333333336e-05,
      "loss": 15.6639,
      "mean_token_accuracy": 0.7415778063237667,
      "num_tokens": 29720768.0,
      "step": 1520
    },
    {
      "entropy": 1.0026599559932947,
      "epoch": 2.448,
      "grad_norm": 31.5,
      "learning_rate": 1.372777777777778e-05,
      "loss": 16.037,
      "mean_token_accuracy": 0.7365970708429813,
      "num_tokens": 29916619.0,
      "step": 1530
    },
    {
      "entropy": 0.9530468691140414,
      "epoch": 2.464,
      "grad_norm": 31.75,
      "learning_rate": 1.3672222222222222e-05,
      "loss": 15.2741,
      "mean_token_accuracy": 0.7459127757698297,
      "num_tokens": 30121198.0,
      "step": 1540
    },
    {
      "entropy": 0.9706476211547852,
      "epoch": 2.48,
      "grad_norm": 28.875,
      "learning_rate": 1.3616666666666666e-05,
      "loss": 15.4971,
      "mean_token_accuracy": 0.7433239962905646,
      "num_tokens": 30325508.0,
      "step": 1550
    },
    {
      "entropy": 0.9861213160678745,
      "epoch": 2.496,
      "grad_norm": 28.625,
      "learning_rate": 1.3561111111111112e-05,
      "loss": 15.7964,
      "mean_token_accuracy": 0.7396648865193128,
      "num_tokens": 30517445.0,
      "step": 1560
    },
    {
      "entropy": 0.9939322113990784,
      "epoch": 2.512,
      "grad_norm": 33.5,
      "learning_rate": 1.3505555555555556e-05,
      "loss": 15.9554,
      "mean_token_accuracy": 0.7374067515134811,
      "num_tokens": 30714049.0,
      "step": 1570
    },
    {
      "entropy": 0.9829558525234461,
      "epoch": 2.528,
      "grad_norm": 29.375,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 15.7034,
      "mean_token_accuracy": 0.7400100238621234,
      "num_tokens": 30911317.0,
      "step": 1580
    },
    {
      "entropy": 0.97488385848701,
      "epoch": 2.544,
      "grad_norm": 28.125,
      "learning_rate": 1.3394444444444446e-05,
      "loss": 15.7159,
      "mean_token_accuracy": 0.7403594095259904,
      "num_tokens": 31108983.0,
      "step": 1590
    },
    {
      "entropy": 0.9890310749411583,
      "epoch": 2.56,
      "grad_norm": 30.125,
      "learning_rate": 1.333888888888889e-05,
      "loss": 15.8407,
      "mean_token_accuracy": 0.7394855204969645,
      "num_tokens": 31298994.0,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_biology_entropy": 0.9767315549850464,
      "eval_biology_loss": 1.1978344917297363,
      "eval_biology_mean_token_accuracy": 0.7075373558998108,
      "eval_biology_num_tokens": 31298994.0,
      "eval_biology_runtime": 36.8405,
      "eval_biology_samples_per_second": 13.572,
      "eval_biology_steps_per_second": 3.393,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_chemistry_entropy": 0.7359724206924438,
      "eval_chemistry_loss": 0.9340105652809143,
      "eval_chemistry_mean_token_accuracy": 0.7665223875045777,
      "eval_chemistry_num_tokens": 31298994.0,
      "eval_chemistry_runtime": 45.3527,
      "eval_chemistry_samples_per_second": 11.025,
      "eval_chemistry_steps_per_second": 2.756,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_math_entropy": 0.6756473398208618,
      "eval_math_loss": 0.998451292514801,
      "eval_math_mean_token_accuracy": 0.7671175785064698,
      "eval_math_num_tokens": 31298994.0,
      "eval_math_runtime": 46.427,
      "eval_math_samples_per_second": 10.77,
      "eval_math_steps_per_second": 2.692,
      "step": 1600
    },
    {
      "epoch": 2.56,
      "eval_physics_entropy": 0.7543728523254395,
      "eval_physics_loss": 0.9128937721252441,
      "eval_physics_mean_token_accuracy": 0.7758728971481323,
      "eval_physics_num_tokens": 31298994.0,
      "eval_physics_runtime": 53.705,
      "eval_physics_samples_per_second": 9.31,
      "eval_physics_steps_per_second": 2.328,
      "step": 1600
    },
    {
      "entropy": 0.9974417526274919,
      "epoch": 2.576,
      "grad_norm": 30.0,
      "learning_rate": 1.3283333333333335e-05,
      "loss": 15.8827,
      "mean_token_accuracy": 0.7387637589126825,
      "num_tokens": 31488512.0,
      "step": 1610
    },
    {
      "entropy": 1.0062089327722787,
      "epoch": 2.592,
      "grad_norm": 29.875,
      "learning_rate": 1.322777777777778e-05,
      "loss": 16.2346,
      "mean_token_accuracy": 0.7342662557959556,
      "num_tokens": 31684651.0,
      "step": 1620
    },
    {
      "entropy": 1.0037442568689585,
      "epoch": 2.608,
      "grad_norm": 29.875,
      "learning_rate": 1.3172222222222224e-05,
      "loss": 16.0345,
      "mean_token_accuracy": 0.7362603198736906,
      "num_tokens": 31876008.0,
      "step": 1630
    },
    {
      "entropy": 0.9704636950045824,
      "epoch": 2.624,
      "grad_norm": 32.0,
      "learning_rate": 1.311666666666667e-05,
      "loss": 15.619,
      "mean_token_accuracy": 0.7409613560885191,
      "num_tokens": 32069617.0,
      "step": 1640
    },
    {
      "entropy": 1.0037952404469253,
      "epoch": 2.64,
      "grad_norm": 29.0,
      "learning_rate": 1.3061111111111112e-05,
      "loss": 16.0021,
      "mean_token_accuracy": 0.736137805506587,
      "num_tokens": 32262932.0,
      "step": 1650
    },
    {
      "entropy": 0.9621167879551649,
      "epoch": 2.656,
      "grad_norm": 28.625,
      "learning_rate": 1.3005555555555556e-05,
      "loss": 15.384,
      "mean_token_accuracy": 0.7444965790957212,
      "num_tokens": 32465495.0,
      "step": 1660
    },
    {
      "entropy": 0.9599390238523483,
      "epoch": 2.672,
      "grad_norm": 30.25,
      "learning_rate": 1.295e-05,
      "loss": 15.5228,
      "mean_token_accuracy": 0.7436116680502891,
      "num_tokens": 32665458.0,
      "step": 1670
    },
    {
      "entropy": 0.9774678956717253,
      "epoch": 2.6879999999999997,
      "grad_norm": 27.875,
      "learning_rate": 1.2894444444444445e-05,
      "loss": 15.5753,
      "mean_token_accuracy": 0.7433817163109779,
      "num_tokens": 32867135.0,
      "step": 1680
    },
    {
      "entropy": 0.9774248816072941,
      "epoch": 2.7039999999999997,
      "grad_norm": 30.125,
      "learning_rate": 1.283888888888889e-05,
      "loss": 15.7087,
      "mean_token_accuracy": 0.7409440513700247,
      "num_tokens": 33070366.0,
      "step": 1690
    },
    {
      "entropy": 0.9762023743242025,
      "epoch": 2.7199999999999998,
      "grad_norm": 33.75,
      "learning_rate": 1.2783333333333333e-05,
      "loss": 15.6031,
      "mean_token_accuracy": 0.7418079309165477,
      "num_tokens": 33269122.0,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_biology_entropy": 0.9562785272598266,
      "eval_biology_loss": 1.2001115083694458,
      "eval_biology_mean_token_accuracy": 0.7079961094856262,
      "eval_biology_num_tokens": 33269122.0,
      "eval_biology_runtime": 36.8873,
      "eval_biology_samples_per_second": 13.555,
      "eval_biology_steps_per_second": 3.389,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_chemistry_entropy": 0.7180265989303589,
      "eval_chemistry_loss": 0.9368557333946228,
      "eval_chemistry_mean_token_accuracy": 0.7668515272140503,
      "eval_chemistry_num_tokens": 33269122.0,
      "eval_chemistry_runtime": 45.2091,
      "eval_chemistry_samples_per_second": 11.06,
      "eval_chemistry_steps_per_second": 2.765,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_math_entropy": 0.6551018316745758,
      "eval_math_loss": 1.0031293630599976,
      "eval_math_mean_token_accuracy": 0.7675854053497314,
      "eval_math_num_tokens": 33269122.0,
      "eval_math_runtime": 46.4714,
      "eval_math_samples_per_second": 10.759,
      "eval_math_steps_per_second": 2.69,
      "step": 1700
    },
    {
      "epoch": 2.7199999999999998,
      "eval_physics_entropy": 0.7281462309360505,
      "eval_physics_loss": 0.9170361161231995,
      "eval_physics_mean_token_accuracy": 0.7760837173461914,
      "eval_physics_num_tokens": 33269122.0,
      "eval_physics_runtime": 53.6243,
      "eval_physics_samples_per_second": 9.324,
      "eval_physics_steps_per_second": 2.331,
      "step": 1700
    },
    {
      "entropy": 1.0005533527582884,
      "epoch": 2.7359999999999998,
      "grad_norm": 27.5,
      "learning_rate": 1.272777777777778e-05,
      "loss": 15.9595,
      "mean_token_accuracy": 0.7374090943485498,
      "num_tokens": 33456592.0,
      "step": 1710
    },
    {
      "entropy": 0.9925159979611635,
      "epoch": 2.752,
      "grad_norm": 30.5,
      "learning_rate": 1.2672222222222223e-05,
      "loss": 15.9212,
      "mean_token_accuracy": 0.7368109248578548,
      "num_tokens": 33645862.0,
      "step": 1720
    },
    {
      "entropy": 1.0035406082868576,
      "epoch": 2.768,
      "grad_norm": 29.125,
      "learning_rate": 1.2616666666666669e-05,
      "loss": 16.0741,
      "mean_token_accuracy": 0.7346370995044709,
      "num_tokens": 33835740.0,
      "step": 1730
    },
    {
      "entropy": 0.9657413493841887,
      "epoch": 2.784,
      "grad_norm": 27.875,
      "learning_rate": 1.2561111111111113e-05,
      "loss": 15.5009,
      "mean_token_accuracy": 0.7434774361550808,
      "num_tokens": 34034702.0,
      "step": 1740
    },
    {
      "entropy": 0.9740934349596501,
      "epoch": 2.8,
      "grad_norm": 28.125,
      "learning_rate": 1.2505555555555557e-05,
      "loss": 15.5637,
      "mean_token_accuracy": 0.7438909105956555,
      "num_tokens": 34235603.0,
      "step": 1750
    },
    {
      "entropy": 0.9742557104676962,
      "epoch": 2.816,
      "grad_norm": 29.875,
      "learning_rate": 1.2450000000000003e-05,
      "loss": 15.6336,
      "mean_token_accuracy": 0.7424697510898113,
      "num_tokens": 34434833.0,
      "step": 1760
    },
    {
      "entropy": 1.0037282764911652,
      "epoch": 2.832,
      "grad_norm": 31.625,
      "learning_rate": 1.2394444444444445e-05,
      "loss": 16.0886,
      "mean_token_accuracy": 0.7369506523013115,
      "num_tokens": 34625581.0,
      "step": 1770
    },
    {
      "entropy": 0.9910386119037866,
      "epoch": 2.848,
      "grad_norm": 32.0,
      "learning_rate": 1.2338888888888889e-05,
      "loss": 15.8463,
      "mean_token_accuracy": 0.7397053975611925,
      "num_tokens": 34816521.0,
      "step": 1780
    },
    {
      "entropy": 0.9725664932280779,
      "epoch": 2.864,
      "grad_norm": 30.5,
      "learning_rate": 1.2283333333333333e-05,
      "loss": 15.7461,
      "mean_token_accuracy": 0.7406340092420578,
      "num_tokens": 35013631.0,
      "step": 1790
    },
    {
      "entropy": 1.00489255823195,
      "epoch": 2.88,
      "grad_norm": 32.0,
      "learning_rate": 1.2227777777777779e-05,
      "loss": 15.9662,
      "mean_token_accuracy": 0.7373752295970917,
      "num_tokens": 35207800.0,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_biology_entropy": 0.9769085459709167,
      "eval_biology_loss": 1.194201946258545,
      "eval_biology_mean_token_accuracy": 0.7088253183364868,
      "eval_biology_num_tokens": 35207800.0,
      "eval_biology_runtime": 36.6978,
      "eval_biology_samples_per_second": 13.625,
      "eval_biology_steps_per_second": 3.406,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_chemistry_entropy": 0.724792768239975,
      "eval_chemistry_loss": 0.9388518333435059,
      "eval_chemistry_mean_token_accuracy": 0.7662976326942443,
      "eval_chemistry_num_tokens": 35207800.0,
      "eval_chemistry_runtime": 45.0996,
      "eval_chemistry_samples_per_second": 11.087,
      "eval_chemistry_steps_per_second": 2.772,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_math_entropy": 0.6623785650730133,
      "eval_math_loss": 1.0053212642669678,
      "eval_math_mean_token_accuracy": 0.7665670537948608,
      "eval_math_num_tokens": 35207800.0,
      "eval_math_runtime": 46.2886,
      "eval_math_samples_per_second": 10.802,
      "eval_math_steps_per_second": 2.7,
      "step": 1800
    },
    {
      "epoch": 2.88,
      "eval_physics_entropy": 0.7309899318218231,
      "eval_physics_loss": 0.9189478754997253,
      "eval_physics_mean_token_accuracy": 0.7758140830993653,
      "eval_physics_num_tokens": 35207800.0,
      "eval_physics_runtime": 53.3951,
      "eval_physics_samples_per_second": 9.364,
      "eval_physics_steps_per_second": 2.341,
      "step": 1800
    },
    {
      "entropy": 0.9926395244896412,
      "epoch": 2.896,
      "grad_norm": 29.875,
      "learning_rate": 1.2172222222222223e-05,
      "loss": 15.8639,
      "mean_token_accuracy": 0.7381080884486437,
      "num_tokens": 35402995.0,
      "step": 1810
    },
    {
      "entropy": 1.0010335687547922,
      "epoch": 2.912,
      "grad_norm": 33.0,
      "learning_rate": 1.2116666666666667e-05,
      "loss": 16.0889,
      "mean_token_accuracy": 0.7379562932997942,
      "num_tokens": 35595456.0,
      "step": 1820
    },
    {
      "entropy": 0.9953980639576911,
      "epoch": 2.928,
      "grad_norm": 32.75,
      "learning_rate": 1.2061111111111113e-05,
      "loss": 15.9704,
      "mean_token_accuracy": 0.7373797506093979,
      "num_tokens": 35787309.0,
      "step": 1830
    },
    {
      "entropy": 0.9820944830775261,
      "epoch": 2.944,
      "grad_norm": 27.875,
      "learning_rate": 1.2005555555555557e-05,
      "loss": 15.7722,
      "mean_token_accuracy": 0.740547452121973,
      "num_tokens": 35982686.0,
      "step": 1840
    },
    {
      "entropy": 0.9821060415357351,
      "epoch": 2.96,
      "grad_norm": 31.625,
      "learning_rate": 1.195e-05,
      "loss": 15.6224,
      "mean_token_accuracy": 0.7415538590401412,
      "num_tokens": 36179520.0,
      "step": 1850
    },
    {
      "entropy": 0.9771638073027133,
      "epoch": 2.976,
      "grad_norm": 28.5,
      "learning_rate": 1.1894444444444447e-05,
      "loss": 15.6643,
      "mean_token_accuracy": 0.7411342523992062,
      "num_tokens": 36373689.0,
      "step": 1860
    },
    {
      "entropy": 0.9860306549817324,
      "epoch": 2.992,
      "grad_norm": 33.5,
      "learning_rate": 1.183888888888889e-05,
      "loss": 15.7875,
      "mean_token_accuracy": 0.7410633962601423,
      "num_tokens": 36569612.0,
      "step": 1870
    },
    {
      "entropy": 0.9603324174880982,
      "epoch": 3.008,
      "grad_norm": 37.25,
      "learning_rate": 1.1783333333333333e-05,
      "loss": 14.9135,
      "mean_token_accuracy": 0.7520980384200812,
      "num_tokens": 36759194.0,
      "step": 1880
    },
    {
      "entropy": 0.8478883054107428,
      "epoch": 3.024,
      "grad_norm": 34.75,
      "learning_rate": 1.1727777777777779e-05,
      "loss": 13.7212,
      "mean_token_accuracy": 0.7676899693906307,
      "num_tokens": 36947293.0,
      "step": 1890
    },
    {
      "entropy": 0.8174442987889051,
      "epoch": 3.04,
      "grad_norm": 37.75,
      "learning_rate": 1.1672222222222223e-05,
      "loss": 12.9695,
      "mean_token_accuracy": 0.7781106177717447,
      "num_tokens": 37136901.0,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_biology_entropy": 0.8307687420845031,
      "eval_biology_loss": 1.3037042617797852,
      "eval_biology_mean_token_accuracy": 0.7012130594253541,
      "eval_biology_num_tokens": 37136901.0,
      "eval_biology_runtime": 36.7227,
      "eval_biology_samples_per_second": 13.616,
      "eval_biology_steps_per_second": 3.404,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_chemistry_entropy": 0.6082699120044708,
      "eval_chemistry_loss": 1.0519765615463257,
      "eval_chemistry_mean_token_accuracy": 0.7578758845329284,
      "eval_chemistry_num_tokens": 37136901.0,
      "eval_chemistry_runtime": 45.1209,
      "eval_chemistry_samples_per_second": 11.081,
      "eval_chemistry_steps_per_second": 2.77,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_math_entropy": 0.5899905214309692,
      "eval_math_loss": 1.0844311714172363,
      "eval_math_mean_token_accuracy": 0.7623905053138733,
      "eval_math_num_tokens": 37136901.0,
      "eval_math_runtime": 46.314,
      "eval_math_samples_per_second": 10.796,
      "eval_math_steps_per_second": 2.699,
      "step": 1900
    },
    {
      "epoch": 3.04,
      "eval_physics_entropy": 0.6185110075473785,
      "eval_physics_loss": 1.0144134759902954,
      "eval_physics_mean_token_accuracy": 0.7686533064842224,
      "eval_physics_num_tokens": 37136901.0,
      "eval_physics_runtime": 53.1821,
      "eval_physics_samples_per_second": 9.402,
      "eval_physics_steps_per_second": 2.35,
      "step": 1900
    },
    {
      "entropy": 0.8187011457979679,
      "epoch": 3.056,
      "grad_norm": 37.25,
      "learning_rate": 1.1616666666666667e-05,
      "loss": 13.1003,
      "mean_token_accuracy": 0.7760679375380277,
      "num_tokens": 37331588.0,
      "step": 1910
    },
    {
      "entropy": 0.8076904013752937,
      "epoch": 3.072,
      "grad_norm": 37.5,
      "learning_rate": 1.1561111111111112e-05,
      "loss": 12.8615,
      "mean_token_accuracy": 0.781244232878089,
      "num_tokens": 37521211.0,
      "step": 1920
    },
    {
      "entropy": 0.82519664876163,
      "epoch": 3.088,
      "grad_norm": 37.0,
      "learning_rate": 1.1505555555555556e-05,
      "loss": 13.0928,
      "mean_token_accuracy": 0.7768605265766382,
      "num_tokens": 37713328.0,
      "step": 1930
    },
    {
      "entropy": 0.8123729152604937,
      "epoch": 3.104,
      "grad_norm": 33.75,
      "learning_rate": 1.145e-05,
      "loss": 12.9409,
      "mean_token_accuracy": 0.780734346807003,
      "num_tokens": 37913470.0,
      "step": 1940
    },
    {
      "entropy": 0.8159066649153829,
      "epoch": 3.12,
      "grad_norm": 37.0,
      "learning_rate": 1.1394444444444446e-05,
      "loss": 13.0653,
      "mean_token_accuracy": 0.7791258115321398,
      "num_tokens": 38106718.0,
      "step": 1950
    },
    {
      "entropy": 0.8273790344595909,
      "epoch": 3.136,
      "grad_norm": 40.75,
      "learning_rate": 1.133888888888889e-05,
      "loss": 13.1437,
      "mean_token_accuracy": 0.7766635850071907,
      "num_tokens": 38295701.0,
      "step": 1960
    },
    {
      "entropy": 0.81372896656394,
      "epoch": 3.152,
      "grad_norm": 37.5,
      "learning_rate": 1.1283333333333334e-05,
      "loss": 12.9402,
      "mean_token_accuracy": 0.7791113253682852,
      "num_tokens": 38493951.0,
      "step": 1970
    },
    {
      "entropy": 0.8379074672237039,
      "epoch": 3.168,
      "grad_norm": 39.0,
      "learning_rate": 1.122777777777778e-05,
      "loss": 13.4123,
      "mean_token_accuracy": 0.7723229706287384,
      "num_tokens": 38688390.0,
      "step": 1980
    },
    {
      "entropy": 0.8237785773351789,
      "epoch": 3.184,
      "grad_norm": 34.25,
      "learning_rate": 1.1172222222222222e-05,
      "loss": 13.1071,
      "mean_token_accuracy": 0.7771053221076727,
      "num_tokens": 38880157.0,
      "step": 1990
    },
    {
      "entropy": 0.8222423151135445,
      "epoch": 3.2,
      "grad_norm": 35.75,
      "learning_rate": 1.1116666666666666e-05,
      "loss": 13.2091,
      "mean_token_accuracy": 0.7758903849869967,
      "num_tokens": 39081782.0,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_biology_entropy": 0.8496361360549927,
      "eval_biology_loss": 1.2998476028442383,
      "eval_biology_mean_token_accuracy": 0.7012469072341919,
      "eval_biology_num_tokens": 39081782.0,
      "eval_biology_runtime": 36.6144,
      "eval_biology_samples_per_second": 13.656,
      "eval_biology_steps_per_second": 3.414,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_chemistry_entropy": 0.6230687329769135,
      "eval_chemistry_loss": 1.0478407144546509,
      "eval_chemistry_mean_token_accuracy": 0.7584491949081421,
      "eval_chemistry_num_tokens": 39081782.0,
      "eval_chemistry_runtime": 44.6898,
      "eval_chemistry_samples_per_second": 11.188,
      "eval_chemistry_steps_per_second": 2.797,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_math_entropy": 0.6067575147151947,
      "eval_math_loss": 1.082250714302063,
      "eval_math_mean_token_accuracy": 0.7613544554710389,
      "eval_math_num_tokens": 39081782.0,
      "eval_math_runtime": 46.1231,
      "eval_math_samples_per_second": 10.841,
      "eval_math_steps_per_second": 2.71,
      "step": 2000
    },
    {
      "epoch": 3.2,
      "eval_physics_entropy": 0.6437831990718842,
      "eval_physics_loss": 1.007546067237854,
      "eval_physics_mean_token_accuracy": 0.7685807209014892,
      "eval_physics_num_tokens": 39081782.0,
      "eval_physics_runtime": 53.088,
      "eval_physics_samples_per_second": 9.418,
      "eval_physics_steps_per_second": 2.355,
      "step": 2000
    }
  ],
  "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": 4.4082567723215933e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}