{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.4,
  "eval_steps": 100,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 1.0975568080320954,
      "epoch": 0.004,
      "grad_norm": 1928.0,
      "learning_rate": 4.5000000000000003e-07,
      "loss": 51.036,
      "mean_token_accuracy": 0.5464097319170833,
      "num_tokens": 51831.0,
      "step": 10
    },
    {
      "entropy": 1.2103175908327102,
      "epoch": 0.008,
      "grad_norm": 1012.0,
      "learning_rate": 9.500000000000001e-07,
      "loss": 47.7535,
      "mean_token_accuracy": 0.549801179766655,
      "num_tokens": 103172.0,
      "step": 20
    },
    {
      "entropy": 1.4519254021346568,
      "epoch": 0.012,
      "grad_norm": 1024.0,
      "learning_rate": 1.45e-06,
      "loss": 45.5399,
      "mean_token_accuracy": 0.5487965773791075,
      "num_tokens": 147915.0,
      "step": 30
    },
    {
      "entropy": 1.515171455964446,
      "epoch": 0.016,
      "grad_norm": 440.0,
      "learning_rate": 1.9500000000000004e-06,
      "loss": 40.0298,
      "mean_token_accuracy": 0.5707932172343135,
      "num_tokens": 195524.0,
      "step": 40
    },
    {
      "entropy": 1.529548142105341,
      "epoch": 0.02,
      "grad_norm": 328.0,
      "learning_rate": 2.4500000000000003e-06,
      "loss": 33.3046,
      "mean_token_accuracy": 0.5914752574637532,
      "num_tokens": 243747.0,
      "step": 50
    },
    {
      "entropy": 1.5671692434698343,
      "epoch": 0.024,
      "grad_norm": 241.0,
      "learning_rate": 2.95e-06,
      "loss": 30.3037,
      "mean_token_accuracy": 0.6058902809396386,
      "num_tokens": 296999.0,
      "step": 60
    },
    {
      "entropy": 1.549248830974102,
      "epoch": 0.028,
      "grad_norm": 235.0,
      "learning_rate": 3.45e-06,
      "loss": 29.0295,
      "mean_token_accuracy": 0.6220679895952343,
      "num_tokens": 342114.0,
      "step": 70
    },
    {
      "entropy": 1.42005056142807,
      "epoch": 0.032,
      "grad_norm": 167.0,
      "learning_rate": 3.95e-06,
      "loss": 26.6089,
      "mean_token_accuracy": 0.6380933992564678,
      "num_tokens": 390903.0,
      "step": 80
    },
    {
      "entropy": 1.5127948723733424,
      "epoch": 0.036,
      "grad_norm": 167.0,
      "learning_rate": 4.450000000000001e-06,
      "loss": 24.8172,
      "mean_token_accuracy": 0.6465719180181623,
      "num_tokens": 440359.0,
      "step": 90
    },
    {
      "entropy": 1.5200586583465339,
      "epoch": 0.04,
      "grad_norm": 142.0,
      "learning_rate": 4.95e-06,
      "loss": 24.207,
      "mean_token_accuracy": 0.6497391745448112,
      "num_tokens": 488481.0,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_biology_entropy": 1.4843950763940812,
      "eval_biology_loss": 1.4266694784164429,
      "eval_biology_mean_token_accuracy": 0.6619762739539147,
      "eval_biology_num_tokens": 488481.0,
      "eval_biology_runtime": 36.5287,
      "eval_biology_samples_per_second": 13.688,
      "eval_biology_steps_per_second": 13.688,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_chemistry_entropy": 1.091414896786213,
      "eval_chemistry_loss": 1.0560342073440552,
      "eval_chemistry_mean_token_accuracy": 0.7318743930459023,
      "eval_chemistry_num_tokens": 488481.0,
      "eval_chemistry_runtime": 39.96,
      "eval_chemistry_samples_per_second": 12.513,
      "eval_chemistry_steps_per_second": 12.513,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_math_entropy": 0.9103146089911461,
      "eval_math_loss": 1.0316376686096191,
      "eval_math_mean_token_accuracy": 0.7397149602174758,
      "eval_math_num_tokens": 488481.0,
      "eval_math_runtime": 41.0822,
      "eval_math_samples_per_second": 12.171,
      "eval_math_steps_per_second": 12.171,
      "step": 100
    },
    {
      "epoch": 0.04,
      "eval_physics_entropy": 0.9860039167404174,
      "eval_physics_loss": 0.9796567559242249,
      "eval_physics_mean_token_accuracy": 0.7483667035698891,
      "eval_physics_num_tokens": 488481.0,
      "eval_physics_runtime": 45.3249,
      "eval_physics_samples_per_second": 11.031,
      "eval_physics_steps_per_second": 11.031,
      "step": 100
    },
    {
      "entropy": 1.444656402617693,
      "epoch": 0.044,
      "grad_norm": 121.0,
      "learning_rate": 5.450000000000001e-06,
      "loss": 22.4665,
      "mean_token_accuracy": 0.6580426702275872,
      "num_tokens": 537471.0,
      "step": 110
    },
    {
      "entropy": 1.3339354760944844,
      "epoch": 0.048,
      "grad_norm": 129.0,
      "learning_rate": 5.950000000000001e-06,
      "loss": 21.4085,
      "mean_token_accuracy": 0.6670540764927864,
      "num_tokens": 589868.0,
      "step": 120
    },
    {
      "entropy": 1.3668562419712544,
      "epoch": 0.052,
      "grad_norm": 124.5,
      "learning_rate": 6.450000000000001e-06,
      "loss": 21.7438,
      "mean_token_accuracy": 0.6594993283972144,
      "num_tokens": 639512.0,
      "step": 130
    },
    {
      "entropy": 1.3065196067094802,
      "epoch": 0.056,
      "grad_norm": 118.5,
      "learning_rate": 6.95e-06,
      "loss": 20.9541,
      "mean_token_accuracy": 0.6683343779295683,
      "num_tokens": 690754.0,
      "step": 140
    },
    {
      "entropy": 1.3160824533551931,
      "epoch": 0.06,
      "grad_norm": 115.5,
      "learning_rate": 7.450000000000001e-06,
      "loss": 21.525,
      "mean_token_accuracy": 0.664486956782639,
      "num_tokens": 737075.0,
      "step": 150
    },
    {
      "entropy": 1.3052487019449472,
      "epoch": 0.064,
      "grad_norm": 113.0,
      "learning_rate": 7.950000000000002e-06,
      "loss": 20.5473,
      "mean_token_accuracy": 0.672793828509748,
      "num_tokens": 791190.0,
      "step": 160
    },
    {
      "entropy": 1.29992272593081,
      "epoch": 0.068,
      "grad_norm": 100.0,
      "learning_rate": 8.45e-06,
      "loss": 21.0557,
      "mean_token_accuracy": 0.6683677529916168,
      "num_tokens": 841344.0,
      "step": 170
    },
    {
      "entropy": 1.3232017546892165,
      "epoch": 0.072,
      "grad_norm": 114.5,
      "learning_rate": 8.95e-06,
      "loss": 21.0887,
      "mean_token_accuracy": 0.6686861604452133,
      "num_tokens": 890075.0,
      "step": 180
    },
    {
      "entropy": 1.2425255428999662,
      "epoch": 0.076,
      "grad_norm": 106.0,
      "learning_rate": 9.450000000000001e-06,
      "loss": 19.9137,
      "mean_token_accuracy": 0.6810434710234403,
      "num_tokens": 940599.0,
      "step": 190
    },
    {
      "entropy": 1.256486358679831,
      "epoch": 0.08,
      "grad_norm": 117.5,
      "learning_rate": 9.950000000000001e-06,
      "loss": 20.0961,
      "mean_token_accuracy": 0.6819845270365477,
      "num_tokens": 989860.0,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_biology_entropy": 1.241776801943779,
      "eval_biology_loss": 1.2506648302078247,
      "eval_biology_mean_token_accuracy": 0.6855908244848251,
      "eval_biology_num_tokens": 989860.0,
      "eval_biology_runtime": 36.0247,
      "eval_biology_samples_per_second": 13.879,
      "eval_biology_steps_per_second": 13.879,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_chemistry_entropy": 0.9190568276047707,
      "eval_chemistry_loss": 0.9248532652854919,
      "eval_chemistry_mean_token_accuracy": 0.7533808985352516,
      "eval_chemistry_num_tokens": 989860.0,
      "eval_chemistry_runtime": 40.487,
      "eval_chemistry_samples_per_second": 12.35,
      "eval_chemistry_steps_per_second": 12.35,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_math_entropy": 0.7673128904104233,
      "eval_math_loss": 0.9293037056922913,
      "eval_math_mean_token_accuracy": 0.759642232298851,
      "eval_math_num_tokens": 989860.0,
      "eval_math_runtime": 41.5439,
      "eval_math_samples_per_second": 12.035,
      "eval_math_steps_per_second": 12.035,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_physics_entropy": 0.8287813304066658,
      "eval_physics_loss": 0.8567265868186951,
      "eval_physics_mean_token_accuracy": 0.770364968419075,
      "eval_physics_num_tokens": 989860.0,
      "eval_physics_runtime": 46.6277,
      "eval_physics_samples_per_second": 10.723,
      "eval_physics_steps_per_second": 10.723,
      "step": 200
    },
    {
      "entropy": 1.2084858063608408,
      "epoch": 0.084,
      "grad_norm": 99.5,
      "learning_rate": 1.045e-05,
      "loss": 19.469,
      "mean_token_accuracy": 0.6908657155930996,
      "num_tokens": 1041199.0,
      "step": 210
    },
    {
      "entropy": 1.3143941588699817,
      "epoch": 0.088,
      "grad_norm": 139.0,
      "learning_rate": 1.095e-05,
      "loss": 21.209,
      "mean_token_accuracy": 0.670545787923038,
      "num_tokens": 1087099.0,
      "step": 220
    },
    {
      "entropy": 1.2587614998221397,
      "epoch": 0.092,
      "grad_norm": 108.0,
      "learning_rate": 1.145e-05,
      "loss": 20.1256,
      "mean_token_accuracy": 0.6842886801809073,
      "num_tokens": 1135090.0,
      "step": 230
    },
    {
      "entropy": 1.2565513662993908,
      "epoch": 0.096,
      "grad_norm": 106.0,
      "learning_rate": 1.195e-05,
      "loss": 20.4044,
      "mean_token_accuracy": 0.6776034630835056,
      "num_tokens": 1181777.0,
      "step": 240
    },
    {
      "entropy": 1.2938017453998327,
      "epoch": 0.1,
      "grad_norm": 112.0,
      "learning_rate": 1.2450000000000003e-05,
      "loss": 20.3714,
      "mean_token_accuracy": 0.6780621778219939,
      "num_tokens": 1229242.0,
      "step": 250
    },
    {
      "entropy": 1.2320711638778448,
      "epoch": 0.104,
      "grad_norm": 94.5,
      "learning_rate": 1.295e-05,
      "loss": 20.052,
      "mean_token_accuracy": 0.6799502000212669,
      "num_tokens": 1283881.0,
      "step": 260
    },
    {
      "entropy": 1.2415869060903788,
      "epoch": 0.108,
      "grad_norm": 107.0,
      "learning_rate": 1.3450000000000002e-05,
      "loss": 19.9151,
      "mean_token_accuracy": 0.6853171430528164,
      "num_tokens": 1332507.0,
      "step": 270
    },
    {
      "entropy": 1.242128773033619,
      "epoch": 0.112,
      "grad_norm": 92.5,
      "learning_rate": 1.3950000000000002e-05,
      "loss": 19.9203,
      "mean_token_accuracy": 0.685390866547823,
      "num_tokens": 1385513.0,
      "step": 280
    },
    {
      "entropy": 1.2662223283201457,
      "epoch": 0.116,
      "grad_norm": 100.5,
      "learning_rate": 1.4450000000000002e-05,
      "loss": 20.3929,
      "mean_token_accuracy": 0.6766710925847292,
      "num_tokens": 1432251.0,
      "step": 290
    },
    {
      "entropy": 1.198459953069687,
      "epoch": 0.12,
      "grad_norm": 101.0,
      "learning_rate": 1.4950000000000003e-05,
      "loss": 19.3584,
      "mean_token_accuracy": 0.68702756986022,
      "num_tokens": 1483667.0,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_biology_entropy": 1.2469175053834916,
      "eval_biology_loss": 1.2332121133804321,
      "eval_biology_mean_token_accuracy": 0.6868851281404496,
      "eval_biology_num_tokens": 1483667.0,
      "eval_biology_runtime": 36.8445,
      "eval_biology_samples_per_second": 13.571,
      "eval_biology_steps_per_second": 13.571,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_chemistry_entropy": 0.9295264416337014,
      "eval_chemistry_loss": 0.9073873162269592,
      "eval_chemistry_mean_token_accuracy": 0.7570922123789787,
      "eval_chemistry_num_tokens": 1483667.0,
      "eval_chemistry_runtime": 40.4179,
      "eval_chemistry_samples_per_second": 12.371,
      "eval_chemistry_steps_per_second": 12.371,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_math_entropy": 0.7820550892949104,
      "eval_math_loss": 0.9024779200553894,
      "eval_math_mean_token_accuracy": 0.7648742213249207,
      "eval_math_num_tokens": 1483667.0,
      "eval_math_runtime": 41.3487,
      "eval_math_samples_per_second": 12.092,
      "eval_math_steps_per_second": 12.092,
      "step": 300
    },
    {
      "epoch": 0.12,
      "eval_physics_entropy": 0.8381168316006661,
      "eval_physics_loss": 0.8357300758361816,
      "eval_physics_mean_token_accuracy": 0.7741321437358856,
      "eval_physics_num_tokens": 1483667.0,
      "eval_physics_runtime": 46.0597,
      "eval_physics_samples_per_second": 10.855,
      "eval_physics_steps_per_second": 10.855,
      "step": 300
    },
    {
      "entropy": 1.268108993396163,
      "epoch": 0.124,
      "grad_norm": 87.0,
      "learning_rate": 1.545e-05,
      "loss": 20.169,
      "mean_token_accuracy": 0.6794606490060687,
      "num_tokens": 1534286.0,
      "step": 310
    },
    {
      "entropy": 1.2664785664528608,
      "epoch": 0.128,
      "grad_norm": 95.0,
      "learning_rate": 1.595e-05,
      "loss": 20.1543,
      "mean_token_accuracy": 0.6814448466524482,
      "num_tokens": 1582368.0,
      "step": 320
    },
    {
      "entropy": 1.2443583220243455,
      "epoch": 0.132,
      "grad_norm": 100.5,
      "learning_rate": 1.645e-05,
      "loss": 20.1453,
      "mean_token_accuracy": 0.6805377397686243,
      "num_tokens": 1630878.0,
      "step": 330
    },
    {
      "entropy": 1.295349196344614,
      "epoch": 0.136,
      "grad_norm": 84.5,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 20.5474,
      "mean_token_accuracy": 0.6783911313861608,
      "num_tokens": 1677060.0,
      "step": 340
    },
    {
      "entropy": 1.2136471576988697,
      "epoch": 0.14,
      "grad_norm": 96.0,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 19.7364,
      "mean_token_accuracy": 0.6849099691957236,
      "num_tokens": 1725186.0,
      "step": 350
    },
    {
      "entropy": 1.2611235719174148,
      "epoch": 0.144,
      "grad_norm": 96.5,
      "learning_rate": 1.795e-05,
      "loss": 20.2599,
      "mean_token_accuracy": 0.6779890628531575,
      "num_tokens": 1773764.0,
      "step": 360
    },
    {
      "entropy": 1.24961077272892,
      "epoch": 0.148,
      "grad_norm": 93.0,
      "learning_rate": 1.845e-05,
      "loss": 19.9901,
      "mean_token_accuracy": 0.6829751830548048,
      "num_tokens": 1823706.0,
      "step": 370
    },
    {
      "entropy": 1.2856210090219975,
      "epoch": 0.152,
      "grad_norm": 103.5,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 20.5301,
      "mean_token_accuracy": 0.6794548545032739,
      "num_tokens": 1873167.0,
      "step": 380
    },
    {
      "entropy": 1.28184983022511,
      "epoch": 0.156,
      "grad_norm": 87.5,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 20.4718,
      "mean_token_accuracy": 0.6774068776518106,
      "num_tokens": 1921484.0,
      "step": 390
    },
    {
      "entropy": 1.2587715853005648,
      "epoch": 0.16,
      "grad_norm": 90.5,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 20.2765,
      "mean_token_accuracy": 0.6795550292357803,
      "num_tokens": 1970077.0,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_biology_entropy": 1.2585812546014785,
      "eval_biology_loss": 1.2385821342468262,
      "eval_biology_mean_token_accuracy": 0.6883087210059166,
      "eval_biology_num_tokens": 1970077.0,
      "eval_biology_runtime": 40.6484,
      "eval_biology_samples_per_second": 12.301,
      "eval_biology_steps_per_second": 12.301,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_chemistry_entropy": 0.9248176854848862,
      "eval_chemistry_loss": 0.9106165766716003,
      "eval_chemistry_mean_token_accuracy": 0.7573499709963799,
      "eval_chemistry_num_tokens": 1970077.0,
      "eval_chemistry_runtime": 40.4652,
      "eval_chemistry_samples_per_second": 12.356,
      "eval_chemistry_steps_per_second": 12.356,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_math_entropy": 0.7678850111961365,
      "eval_math_loss": 0.897013783454895,
      "eval_math_mean_token_accuracy": 0.7673858489990234,
      "eval_math_num_tokens": 1970077.0,
      "eval_math_runtime": 41.3756,
      "eval_math_samples_per_second": 12.084,
      "eval_math_steps_per_second": 12.084,
      "step": 400
    },
    {
      "epoch": 0.16,
      "eval_physics_entropy": 0.8230015668869018,
      "eval_physics_loss": 0.8317809104919434,
      "eval_physics_mean_token_accuracy": 0.7749597446918488,
      "eval_physics_num_tokens": 1970077.0,
      "eval_physics_runtime": 45.5271,
      "eval_physics_samples_per_second": 10.982,
      "eval_physics_steps_per_second": 10.982,
      "step": 400
    },
    {
      "entropy": 1.2881209932267665,
      "epoch": 0.164,
      "grad_norm": 80.5,
      "learning_rate": 1.9950000000000004e-05,
      "loss": 20.7614,
      "mean_token_accuracy": 0.6718229403719306,
      "num_tokens": 2019053.0,
      "step": 410
    },
    {
      "entropy": 1.2165789466351271,
      "epoch": 0.168,
      "grad_norm": 75.5,
      "learning_rate": 1.9894444444444446e-05,
      "loss": 19.5986,
      "mean_token_accuracy": 0.6869289472699165,
      "num_tokens": 2066815.0,
      "step": 420
    },
    {
      "entropy": 1.2494619188830256,
      "epoch": 0.172,
      "grad_norm": 80.5,
      "learning_rate": 1.983888888888889e-05,
      "loss": 20.2759,
      "mean_token_accuracy": 0.6808055322617292,
      "num_tokens": 2116402.0,
      "step": 430
    },
    {
      "entropy": 1.214642644301057,
      "epoch": 0.176,
      "grad_norm": 81.5,
      "learning_rate": 1.9783333333333334e-05,
      "loss": 19.3769,
      "mean_token_accuracy": 0.6881225043907762,
      "num_tokens": 2168354.0,
      "step": 440
    },
    {
      "entropy": 1.2326878625899553,
      "epoch": 0.18,
      "grad_norm": 76.0,
      "learning_rate": 1.972777777777778e-05,
      "loss": 20.0507,
      "mean_token_accuracy": 0.6823928710073233,
      "num_tokens": 2219158.0,
      "step": 450
    },
    {
      "entropy": 1.2790402939543128,
      "epoch": 0.184,
      "grad_norm": 97.0,
      "learning_rate": 1.9672222222222222e-05,
      "loss": 20.4576,
      "mean_token_accuracy": 0.6788654251024127,
      "num_tokens": 2270617.0,
      "step": 460
    },
    {
      "entropy": 1.2233176242560149,
      "epoch": 0.188,
      "grad_norm": 76.0,
      "learning_rate": 1.9616666666666668e-05,
      "loss": 19.4471,
      "mean_token_accuracy": 0.6852962577715516,
      "num_tokens": 2322034.0,
      "step": 470
    },
    {
      "entropy": 1.284807311743498,
      "epoch": 0.192,
      "grad_norm": 97.0,
      "learning_rate": 1.9561111111111114e-05,
      "loss": 21.2325,
      "mean_token_accuracy": 0.6681901682168245,
      "num_tokens": 2365822.0,
      "step": 480
    },
    {
      "entropy": 1.250083539262414,
      "epoch": 0.196,
      "grad_norm": 93.5,
      "learning_rate": 1.9505555555555556e-05,
      "loss": 19.6203,
      "mean_token_accuracy": 0.6853536464273929,
      "num_tokens": 2413853.0,
      "step": 490
    },
    {
      "entropy": 1.2981292121112347,
      "epoch": 0.2,
      "grad_norm": 85.0,
      "learning_rate": 1.9450000000000002e-05,
      "loss": 20.8012,
      "mean_token_accuracy": 0.6707905715331435,
      "num_tokens": 2459876.0,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_biology_entropy": 1.2485336683392525,
      "eval_biology_loss": 1.2281557321548462,
      "eval_biology_mean_token_accuracy": 0.6897270123958588,
      "eval_biology_num_tokens": 2459876.0,
      "eval_biology_runtime": 38.8945,
      "eval_biology_samples_per_second": 12.855,
      "eval_biology_steps_per_second": 12.855,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_chemistry_entropy": 0.9259556359648704,
      "eval_chemistry_loss": 0.9128974676132202,
      "eval_chemistry_mean_token_accuracy": 0.7565669685602188,
      "eval_chemistry_num_tokens": 2459876.0,
      "eval_chemistry_runtime": 40.516,
      "eval_chemistry_samples_per_second": 12.341,
      "eval_chemistry_steps_per_second": 12.341,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_math_entropy": 0.7720121605396271,
      "eval_math_loss": 0.9020798206329346,
      "eval_math_mean_token_accuracy": 0.7660462435483932,
      "eval_math_num_tokens": 2459876.0,
      "eval_math_runtime": 62.8982,
      "eval_math_samples_per_second": 7.949,
      "eval_math_steps_per_second": 7.949,
      "step": 500
    },
    {
      "epoch": 0.2,
      "eval_physics_entropy": 0.8241469451785087,
      "eval_physics_loss": 0.832719624042511,
      "eval_physics_mean_token_accuracy": 0.7746937532424927,
      "eval_physics_num_tokens": 2459876.0,
      "eval_physics_runtime": 76.8817,
      "eval_physics_samples_per_second": 6.504,
      "eval_physics_steps_per_second": 6.504,
      "step": 500
    },
    {
      "entropy": 1.2163557052612304,
      "epoch": 0.204,
      "grad_norm": 73.5,
      "learning_rate": 1.9394444444444448e-05,
      "loss": 19.9923,
      "mean_token_accuracy": 0.6821251731365919,
      "num_tokens": 2509187.0,
      "step": 510
    },
    {
      "entropy": 1.2736122459173203,
      "epoch": 0.208,
      "grad_norm": 88.5,
      "learning_rate": 1.933888888888889e-05,
      "loss": 20.2436,
      "mean_token_accuracy": 0.6758794913068413,
      "num_tokens": 2558762.0,
      "step": 520
    },
    {
      "entropy": 1.2216465022414922,
      "epoch": 0.212,
      "grad_norm": 71.5,
      "learning_rate": 1.9283333333333332e-05,
      "loss": 19.6407,
      "mean_token_accuracy": 0.6855740923434496,
      "num_tokens": 2608616.0,
      "step": 530
    },
    {
      "entropy": 1.2674344064667822,
      "epoch": 0.216,
      "grad_norm": 86.0,
      "learning_rate": 1.9227777777777778e-05,
      "loss": 20.4517,
      "mean_token_accuracy": 0.6805386714637279,
      "num_tokens": 2654953.0,
      "step": 540
    },
    {
      "entropy": 1.2348025966435672,
      "epoch": 0.22,
      "grad_norm": 79.5,
      "learning_rate": 1.9172222222222224e-05,
      "loss": 19.9027,
      "mean_token_accuracy": 0.6820508353412151,
      "num_tokens": 2704686.0,
      "step": 550
    },
    {
      "entropy": 1.2361498434096574,
      "epoch": 0.224,
      "grad_norm": 86.5,
      "learning_rate": 1.911666666666667e-05,
      "loss": 20.0735,
      "mean_token_accuracy": 0.6846253119409085,
      "num_tokens": 2755347.0,
      "step": 560
    },
    {
      "entropy": 1.2503492299467325,
      "epoch": 0.228,
      "grad_norm": 77.0,
      "learning_rate": 1.9061111111111112e-05,
      "loss": 19.7198,
      "mean_token_accuracy": 0.685746643692255,
      "num_tokens": 2804340.0,
      "step": 570
    },
    {
      "entropy": 1.2075012478977443,
      "epoch": 0.232,
      "grad_norm": 80.0,
      "learning_rate": 1.9005555555555557e-05,
      "loss": 19.6344,
      "mean_token_accuracy": 0.6857940997928381,
      "num_tokens": 2852298.0,
      "step": 580
    },
    {
      "entropy": 1.1989578040316702,
      "epoch": 0.236,
      "grad_norm": 85.5,
      "learning_rate": 1.8950000000000003e-05,
      "loss": 19.3132,
      "mean_token_accuracy": 0.6939592262730002,
      "num_tokens": 2901227.0,
      "step": 590
    },
    {
      "entropy": 1.2394503276795148,
      "epoch": 0.24,
      "grad_norm": 99.5,
      "learning_rate": 1.8894444444444446e-05,
      "loss": 20.0171,
      "mean_token_accuracy": 0.6842290829867125,
      "num_tokens": 2947346.0,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_biology_entropy": 1.1514823293685914,
      "eval_biology_loss": 1.2215782403945923,
      "eval_biology_mean_token_accuracy": 0.6902710503339767,
      "eval_biology_num_tokens": 2947346.0,
      "eval_biology_runtime": 36.8865,
      "eval_biology_samples_per_second": 13.555,
      "eval_biology_steps_per_second": 13.555,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_chemistry_entropy": 0.8571230885386467,
      "eval_chemistry_loss": 0.9023940563201904,
      "eval_chemistry_mean_token_accuracy": 0.7593449745774269,
      "eval_chemistry_num_tokens": 2947346.0,
      "eval_chemistry_runtime": 40.0868,
      "eval_chemistry_samples_per_second": 12.473,
      "eval_chemistry_steps_per_second": 12.473,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_math_entropy": 0.7442211763560772,
      "eval_math_loss": 0.9054449200630188,
      "eval_math_mean_token_accuracy": 0.76559605550766,
      "eval_math_num_tokens": 2947346.0,
      "eval_math_runtime": 41.6093,
      "eval_math_samples_per_second": 12.017,
      "eval_math_steps_per_second": 12.017,
      "step": 600
    },
    {
      "epoch": 0.24,
      "eval_physics_entropy": 0.7813739073872566,
      "eval_physics_loss": 0.8251622319221497,
      "eval_physics_mean_token_accuracy": 0.7764079115390777,
      "eval_physics_num_tokens": 2947346.0,
      "eval_physics_runtime": 45.5977,
      "eval_physics_samples_per_second": 10.965,
      "eval_physics_steps_per_second": 10.965,
      "step": 600
    },
    {
      "entropy": 1.2816234160214663,
      "epoch": 0.244,
      "grad_norm": 82.0,
      "learning_rate": 1.883888888888889e-05,
      "loss": 20.3578,
      "mean_token_accuracy": 0.6772410158067942,
      "num_tokens": 2996175.0,
      "step": 610
    },
    {
      "entropy": 1.2542358499020339,
      "epoch": 0.248,
      "grad_norm": 77.0,
      "learning_rate": 1.8783333333333337e-05,
      "loss": 20.4992,
      "mean_token_accuracy": 0.6766950022429228,
      "num_tokens": 3043128.0,
      "step": 620
    },
    {
      "entropy": 1.2298073736950754,
      "epoch": 0.252,
      "grad_norm": 76.0,
      "learning_rate": 1.872777777777778e-05,
      "loss": 19.7128,
      "mean_token_accuracy": 0.6849291998893022,
      "num_tokens": 3090827.0,
      "step": 630
    },
    {
      "entropy": 1.1652949508279562,
      "epoch": 0.256,
      "grad_norm": 87.0,
      "learning_rate": 1.8672222222222222e-05,
      "loss": 19.0561,
      "mean_token_accuracy": 0.6929412636905908,
      "num_tokens": 3139957.0,
      "step": 640
    },
    {
      "entropy": 1.2646588858217,
      "epoch": 0.26,
      "grad_norm": 84.0,
      "learning_rate": 1.8616666666666667e-05,
      "loss": 20.0112,
      "mean_token_accuracy": 0.6835361944511533,
      "num_tokens": 3183188.0,
      "step": 650
    },
    {
      "entropy": 1.2171090628951788,
      "epoch": 0.264,
      "grad_norm": 81.5,
      "learning_rate": 1.8561111111111113e-05,
      "loss": 19.9192,
      "mean_token_accuracy": 0.6854063520208001,
      "num_tokens": 3232159.0,
      "step": 660
    },
    {
      "entropy": 1.218252230808139,
      "epoch": 0.268,
      "grad_norm": 76.5,
      "learning_rate": 1.8505555555555556e-05,
      "loss": 19.4163,
      "mean_token_accuracy": 0.6893998075276613,
      "num_tokens": 3283908.0,
      "step": 670
    },
    {
      "entropy": 1.1729648176580667,
      "epoch": 0.272,
      "grad_norm": 101.0,
      "learning_rate": 1.845e-05,
      "loss": 19.422,
      "mean_token_accuracy": 0.6887724719941616,
      "num_tokens": 3335951.0,
      "step": 680
    },
    {
      "entropy": 1.2519408775493504,
      "epoch": 0.276,
      "grad_norm": 90.5,
      "learning_rate": 1.8394444444444447e-05,
      "loss": 19.8039,
      "mean_token_accuracy": 0.6823019618168473,
      "num_tokens": 3387755.0,
      "step": 690
    },
    {
      "entropy": 1.195892432704568,
      "epoch": 0.28,
      "grad_norm": 84.0,
      "learning_rate": 1.833888888888889e-05,
      "loss": 19.5225,
      "mean_token_accuracy": 0.685329046845436,
      "num_tokens": 3437213.0,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_biology_entropy": 1.2085156010985374,
      "eval_biology_loss": 1.207029104232788,
      "eval_biology_mean_token_accuracy": 0.6939799420833588,
      "eval_biology_num_tokens": 3437213.0,
      "eval_biology_runtime": 36.7099,
      "eval_biology_samples_per_second": 13.62,
      "eval_biology_steps_per_second": 13.62,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_chemistry_entropy": 0.9005156295895577,
      "eval_chemistry_loss": 0.9021785855293274,
      "eval_chemistry_mean_token_accuracy": 0.7585515556931496,
      "eval_chemistry_num_tokens": 3437213.0,
      "eval_chemistry_runtime": 40.2932,
      "eval_chemistry_samples_per_second": 12.409,
      "eval_chemistry_steps_per_second": 12.409,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_math_entropy": 0.7635650929510593,
      "eval_math_loss": 0.9004127979278564,
      "eval_math_mean_token_accuracy": 0.7664950026273727,
      "eval_math_num_tokens": 3437213.0,
      "eval_math_runtime": 40.3086,
      "eval_math_samples_per_second": 12.404,
      "eval_math_steps_per_second": 12.404,
      "step": 700
    },
    {
      "epoch": 0.28,
      "eval_physics_entropy": 0.8227353655099868,
      "eval_physics_loss": 0.8242940902709961,
      "eval_physics_mean_token_accuracy": 0.7775262135863305,
      "eval_physics_num_tokens": 3437213.0,
      "eval_physics_runtime": 44.9118,
      "eval_physics_samples_per_second": 11.133,
      "eval_physics_steps_per_second": 11.133,
      "step": 700
    },
    {
      "entropy": 1.1728057231754065,
      "epoch": 0.284,
      "grad_norm": 67.5,
      "learning_rate": 1.8283333333333335e-05,
      "loss": 18.654,
      "mean_token_accuracy": 0.6981678586453199,
      "num_tokens": 3489088.0,
      "step": 710
    },
    {
      "entropy": 1.1936262086033822,
      "epoch": 0.288,
      "grad_norm": 72.5,
      "learning_rate": 1.822777777777778e-05,
      "loss": 19.517,
      "mean_token_accuracy": 0.6882296811789275,
      "num_tokens": 3539731.0,
      "step": 720
    },
    {
      "entropy": 1.2277162373065948,
      "epoch": 0.292,
      "grad_norm": 75.5,
      "learning_rate": 1.8172222222222223e-05,
      "loss": 19.4898,
      "mean_token_accuracy": 0.6886593095958233,
      "num_tokens": 3586433.0,
      "step": 730
    },
    {
      "entropy": 1.2317656818777323,
      "epoch": 0.296,
      "grad_norm": 79.0,
      "learning_rate": 1.811666666666667e-05,
      "loss": 20.2521,
      "mean_token_accuracy": 0.6818229380995036,
      "num_tokens": 3633448.0,
      "step": 740
    },
    {
      "entropy": 1.1757729103788734,
      "epoch": 0.3,
      "grad_norm": 75.0,
      "learning_rate": 1.806111111111111e-05,
      "loss": 18.6931,
      "mean_token_accuracy": 0.701389448530972,
      "num_tokens": 3685080.0,
      "step": 750
    },
    {
      "entropy": 1.1586591433733702,
      "epoch": 0.304,
      "grad_norm": 77.5,
      "learning_rate": 1.8005555555555557e-05,
      "loss": 18.9227,
      "mean_token_accuracy": 0.69391386564821,
      "num_tokens": 3733488.0,
      "step": 760
    },
    {
      "entropy": 1.2230437025427818,
      "epoch": 0.308,
      "grad_norm": 76.0,
      "learning_rate": 1.795e-05,
      "loss": 19.675,
      "mean_token_accuracy": 0.6864967104047537,
      "num_tokens": 3779708.0,
      "step": 770
    },
    {
      "entropy": 1.21304402500391,
      "epoch": 0.312,
      "grad_norm": 85.0,
      "learning_rate": 1.7894444444444445e-05,
      "loss": 19.5388,
      "mean_token_accuracy": 0.6890883000567556,
      "num_tokens": 3825212.0,
      "step": 780
    },
    {
      "entropy": 1.2525564279407262,
      "epoch": 0.316,
      "grad_norm": 89.5,
      "learning_rate": 1.783888888888889e-05,
      "loss": 20.2117,
      "mean_token_accuracy": 0.6794506318867206,
      "num_tokens": 3870912.0,
      "step": 790
    },
    {
      "entropy": 1.2013506395742297,
      "epoch": 0.32,
      "grad_norm": 88.5,
      "learning_rate": 1.7783333333333333e-05,
      "loss": 19.2548,
      "mean_token_accuracy": 0.6915859717875719,
      "num_tokens": 3920545.0,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_biology_entropy": 1.1828332220315934,
      "eval_biology_loss": 1.1939783096313477,
      "eval_biology_mean_token_accuracy": 0.6965516614317894,
      "eval_biology_num_tokens": 3920545.0,
      "eval_biology_runtime": 35.8369,
      "eval_biology_samples_per_second": 13.952,
      "eval_biology_steps_per_second": 13.952,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_chemistry_entropy": 0.886673482120037,
      "eval_chemistry_loss": 0.8999966979026794,
      "eval_chemistry_mean_token_accuracy": 0.7598078215122223,
      "eval_chemistry_num_tokens": 3920545.0,
      "eval_chemistry_runtime": 40.2857,
      "eval_chemistry_samples_per_second": 12.411,
      "eval_chemistry_steps_per_second": 12.411,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_math_entropy": 0.7587899960577488,
      "eval_math_loss": 0.8991398215293884,
      "eval_math_mean_token_accuracy": 0.7670052968263626,
      "eval_math_num_tokens": 3920545.0,
      "eval_math_runtime": 41.3295,
      "eval_math_samples_per_second": 12.098,
      "eval_math_steps_per_second": 12.098,
      "step": 800
    },
    {
      "epoch": 0.32,
      "eval_physics_entropy": 0.8086917443871499,
      "eval_physics_loss": 0.820421576499939,
      "eval_physics_mean_token_accuracy": 0.7782125360369683,
      "eval_physics_num_tokens": 3920545.0,
      "eval_physics_runtime": 45.2584,
      "eval_physics_samples_per_second": 11.048,
      "eval_physics_steps_per_second": 11.048,
      "step": 800
    },
    {
      "entropy": 1.2068503752350808,
      "epoch": 0.324,
      "grad_norm": 79.5,
      "learning_rate": 1.772777777777778e-05,
      "loss": 19.4093,
      "mean_token_accuracy": 0.6900258323177695,
      "num_tokens": 3967499.0,
      "step": 810
    },
    {
      "entropy": 1.1701661001890897,
      "epoch": 0.328,
      "grad_norm": 72.5,
      "learning_rate": 1.7672222222222224e-05,
      "loss": 18.743,
      "mean_token_accuracy": 0.6989876503124833,
      "num_tokens": 4016673.0,
      "step": 820
    },
    {
      "entropy": 1.2040905393660069,
      "epoch": 0.332,
      "grad_norm": 80.0,
      "learning_rate": 1.761666666666667e-05,
      "loss": 19.7643,
      "mean_token_accuracy": 0.6862274307757616,
      "num_tokens": 4063859.0,
      "step": 830
    },
    {
      "entropy": 1.1994104048237204,
      "epoch": 0.336,
      "grad_norm": 75.5,
      "learning_rate": 1.7561111111111113e-05,
      "loss": 19.1933,
      "mean_token_accuracy": 0.6893992654979229,
      "num_tokens": 4114077.0,
      "step": 840
    },
    {
      "entropy": 1.21622231900692,
      "epoch": 0.34,
      "grad_norm": 75.5,
      "learning_rate": 1.7505555555555558e-05,
      "loss": 19.6288,
      "mean_token_accuracy": 0.6871093420311809,
      "num_tokens": 4160168.0,
      "step": 850
    },
    {
      "entropy": 1.1868847006931902,
      "epoch": 0.344,
      "grad_norm": 72.5,
      "learning_rate": 1.7450000000000004e-05,
      "loss": 19.3293,
      "mean_token_accuracy": 0.6937917049974203,
      "num_tokens": 4210403.0,
      "step": 860
    },
    {
      "entropy": 1.1898768026381732,
      "epoch": 0.348,
      "grad_norm": 74.5,
      "learning_rate": 1.7394444444444446e-05,
      "loss": 18.7152,
      "mean_token_accuracy": 0.6979909218847752,
      "num_tokens": 4256944.0,
      "step": 870
    },
    {
      "entropy": 1.1888672057539225,
      "epoch": 0.352,
      "grad_norm": 75.5,
      "learning_rate": 1.733888888888889e-05,
      "loss": 19.4446,
      "mean_token_accuracy": 0.6919016648083925,
      "num_tokens": 4306949.0,
      "step": 880
    },
    {
      "entropy": 1.1983365084975959,
      "epoch": 0.356,
      "grad_norm": 69.0,
      "learning_rate": 1.7283333333333334e-05,
      "loss": 19.2777,
      "mean_token_accuracy": 0.6915557283908129,
      "num_tokens": 4355070.0,
      "step": 890
    },
    {
      "entropy": 1.1921971619129181,
      "epoch": 0.36,
      "grad_norm": 63.75,
      "learning_rate": 1.722777777777778e-05,
      "loss": 19.2026,
      "mean_token_accuracy": 0.6915637668222189,
      "num_tokens": 4406104.0,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_biology_entropy": 1.1541215604543686,
      "eval_biology_loss": 1.1821783781051636,
      "eval_biology_mean_token_accuracy": 0.6984301459789276,
      "eval_biology_num_tokens": 4406104.0,
      "eval_biology_runtime": 35.8688,
      "eval_biology_samples_per_second": 13.94,
      "eval_biology_steps_per_second": 13.94,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_chemistry_entropy": 0.8648406760692596,
      "eval_chemistry_loss": 0.9016790390014648,
      "eval_chemistry_mean_token_accuracy": 0.7602673851847649,
      "eval_chemistry_num_tokens": 4406104.0,
      "eval_chemistry_runtime": 40.1126,
      "eval_chemistry_samples_per_second": 12.465,
      "eval_chemistry_steps_per_second": 12.465,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_math_entropy": 0.7475603276491165,
      "eval_math_loss": 0.9146779775619507,
      "eval_math_mean_token_accuracy": 0.7653868557214737,
      "eval_math_num_tokens": 4406104.0,
      "eval_math_runtime": 41.1141,
      "eval_math_samples_per_second": 12.161,
      "eval_math_steps_per_second": 12.161,
      "step": 900
    },
    {
      "epoch": 0.36,
      "eval_physics_entropy": 0.782447283744812,
      "eval_physics_loss": 0.8322422504425049,
      "eval_physics_mean_token_accuracy": 0.7780593295693398,
      "eval_physics_num_tokens": 4406104.0,
      "eval_physics_runtime": 45.2444,
      "eval_physics_samples_per_second": 11.051,
      "eval_physics_steps_per_second": 11.051,
      "step": 900
    },
    {
      "entropy": 1.2114545024931431,
      "epoch": 0.364,
      "grad_norm": 73.0,
      "learning_rate": 1.7172222222222223e-05,
      "loss": 19.4011,
      "mean_token_accuracy": 0.6914434898644686,
      "num_tokens": 4455015.0,
      "step": 910
    },
    {
      "entropy": 1.1504958260804414,
      "epoch": 0.368,
      "grad_norm": 84.5,
      "learning_rate": 1.7116666666666668e-05,
      "loss": 18.8783,
      "mean_token_accuracy": 0.6960504239425063,
      "num_tokens": 4504001.0,
      "step": 920
    },
    {
      "entropy": 1.2334423441439868,
      "epoch": 0.372,
      "grad_norm": 76.5,
      "learning_rate": 1.7061111111111114e-05,
      "loss": 19.3716,
      "mean_token_accuracy": 0.6921723924577237,
      "num_tokens": 4550269.0,
      "step": 930
    },
    {
      "entropy": 1.138117073662579,
      "epoch": 0.376,
      "grad_norm": 81.0,
      "learning_rate": 1.7005555555555556e-05,
      "loss": 18.6873,
      "mean_token_accuracy": 0.6977444339543581,
      "num_tokens": 4597559.0,
      "step": 940
    },
    {
      "entropy": 1.1804296549409627,
      "epoch": 0.38,
      "grad_norm": 84.5,
      "learning_rate": 1.6950000000000002e-05,
      "loss": 18.7748,
      "mean_token_accuracy": 0.6972912590950727,
      "num_tokens": 4646903.0,
      "step": 950
    },
    {
      "entropy": 1.1713591821491718,
      "epoch": 0.384,
      "grad_norm": 71.5,
      "learning_rate": 1.6894444444444448e-05,
      "loss": 19.1636,
      "mean_token_accuracy": 0.6900384306907654,
      "num_tokens": 4693812.0,
      "step": 960
    },
    {
      "entropy": 1.20406583994627,
      "epoch": 0.388,
      "grad_norm": 75.0,
      "learning_rate": 1.683888888888889e-05,
      "loss": 19.1408,
      "mean_token_accuracy": 0.6931487508118153,
      "num_tokens": 4740309.0,
      "step": 970
    },
    {
      "entropy": 1.2038170533254742,
      "epoch": 0.392,
      "grad_norm": 71.5,
      "learning_rate": 1.6783333333333332e-05,
      "loss": 19.1738,
      "mean_token_accuracy": 0.6930541478097438,
      "num_tokens": 4790965.0,
      "step": 980
    },
    {
      "entropy": 1.1741992883384227,
      "epoch": 0.396,
      "grad_norm": 76.0,
      "learning_rate": 1.6727777777777778e-05,
      "loss": 19.1958,
      "mean_token_accuracy": 0.6922040678560734,
      "num_tokens": 4839186.0,
      "step": 990
    },
    {
      "entropy": 1.2012158632278442,
      "epoch": 0.4,
      "grad_norm": 77.5,
      "learning_rate": 1.6672222222222224e-05,
      "loss": 19.1545,
      "mean_token_accuracy": 0.6926527272909879,
      "num_tokens": 4887094.0,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_biology_entropy": 1.153373451292515,
      "eval_biology_loss": 1.1686452627182007,
      "eval_biology_mean_token_accuracy": 0.7013231556415558,
      "eval_biology_num_tokens": 4887094.0,
      "eval_biology_runtime": 36.1843,
      "eval_biology_samples_per_second": 13.818,
      "eval_biology_steps_per_second": 13.818,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_chemistry_entropy": 0.8627612872719764,
      "eval_chemistry_loss": 0.8771724104881287,
      "eval_chemistry_mean_token_accuracy": 0.7635843125581742,
      "eval_chemistry_num_tokens": 4887094.0,
      "eval_chemistry_runtime": 39.65,
      "eval_chemistry_samples_per_second": 12.61,
      "eval_chemistry_steps_per_second": 12.61,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_math_entropy": 0.7435729750990867,
      "eval_math_loss": 0.8909226655960083,
      "eval_math_mean_token_accuracy": 0.7684622532129288,
      "eval_math_num_tokens": 4887094.0,
      "eval_math_runtime": 40.3632,
      "eval_math_samples_per_second": 12.388,
      "eval_math_steps_per_second": 12.388,
      "step": 1000
    },
    {
      "epoch": 0.4,
      "eval_physics_entropy": 0.7918061959147453,
      "eval_physics_loss": 0.8059831261634827,
      "eval_physics_mean_token_accuracy": 0.7804628658294678,
      "eval_physics_num_tokens": 4887094.0,
      "eval_physics_runtime": 44.9291,
      "eval_physics_samples_per_second": 11.129,
      "eval_physics_steps_per_second": 11.129,
      "step": 1000
    }
  ],
  "logging_steps": 10,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 3.277009915508544e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}