{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6156412598800493,
  "global_step": 31000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-07,
      "loss": 10.8568,
      "theoretical_loss": 20.812814784551147,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-06,
      "loss": 10.8727,
      "theoretical_loss": 17.56619808733264,
      "tokens_seen": 131072
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904763e-06,
      "loss": 10.7356,
      "theoretical_loss": 15.939474075840563,
      "tokens_seen": 196608
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-06,
      "loss": 10.7113,
      "theoretical_loss": 14.892313738992565,
      "tokens_seen": 262144
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-06,
      "loss": 10.3498,
      "theoretical_loss": 14.13621392076697,
      "tokens_seen": 327680
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809525e-06,
      "loss": 10.1826,
      "theoretical_loss": 13.552558455554218,
      "tokens_seen": 393216
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-06,
      "loss": 10.0176,
      "theoretical_loss": 13.081805984405184,
      "tokens_seen": 458752
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507936e-06,
      "loss": 9.6515,
      "theoretical_loss": 12.690126608487319,
      "tokens_seen": 524288
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571428e-06,
      "loss": 9.2598,
      "theoretical_loss": 12.356589446877619,
      "tokens_seen": 589824
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-06,
      "loss": 9.1219,
      "theoretical_loss": 12.067409590039073,
      "tokens_seen": 655360
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.0912698412698412e-05,
      "loss": 9.2934,
      "theoretical_loss": 11.813063214105672,
      "tokens_seen": 720896
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 9.0973,
      "theoretical_loss": 11.586716191710725,
      "tokens_seen": 786432
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.2896825396825396e-05,
      "loss": 9.4804,
      "theoretical_loss": 11.383311123190783,
      "tokens_seen": 851968
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.3888888888888888e-05,
      "loss": 9.3208,
      "theoretical_loss": 11.199008685115867,
      "tokens_seen": 917504
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4880952380952381e-05,
      "loss": 9.1864,
      "theoretical_loss": 11.030830900981908,
      "tokens_seen": 983040
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 9.398,
      "theoretical_loss": 10.876425069460945,
      "tokens_seen": 1048576
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6865079365079364e-05,
      "loss": 8.9394,
      "theoretical_loss": 10.73390272306672,
      "tokens_seen": 1114112
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.7857142857142855e-05,
      "loss": 9.0636,
      "theoretical_loss": 10.601726859234274,
      "tokens_seen": 1179648
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.884920634920635e-05,
      "loss": 9.1583,
      "theoretical_loss": 10.478631155360636,
      "tokens_seen": 1245184
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.984126984126984e-05,
      "loss": 9.1124,
      "theoretical_loss": 10.363560926767326,
      "tokens_seen": 1310720
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.0833333333333333e-05,
      "loss": 9.0087,
      "theoretical_loss": 10.25562920390074,
      "tokens_seen": 1376256
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.1825396825396824e-05,
      "loss": 9.0788,
      "theoretical_loss": 10.154083536274014,
      "tokens_seen": 1441792
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2817460317460315e-05,
      "loss": 8.4406,
      "theoretical_loss": 10.058280544736593,
      "tokens_seen": 1507328
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.380952380952381e-05,
      "loss": 8.9754,
      "theoretical_loss": 9.967666161844274,
      "tokens_seen": 1572864
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 12267,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 8.854658126831055,
      "objective/train/theoretical_loss": 9.881760109397105,
      "objective/train/tokens_used": 22098400,
      "theoretical_loss": 9.881760109397105,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.48015873015873e-05,
      "loss": 8.9164,
      "theoretical_loss": 9.881760109397105,
      "tokens_seen": 1638400
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.5793650793650793e-05,
      "loss": 7.9452,
      "theoretical_loss": 9.800143574544556,
      "tokens_seen": 1703936
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6785714285714284e-05,
      "loss": 8.81,
      "theoretical_loss": 9.722449329911441,
      "tokens_seen": 1769472
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.7777777777777776e-05,
      "loss": 9.0741,
      "theoretical_loss": 9.648353742085542,
      "tokens_seen": 1835008
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.876984126984127e-05,
      "loss": 8.8256,
      "theoretical_loss": 9.577570254149634,
      "tokens_seen": 1900544
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 8.3274,
      "theoretical_loss": 9.509844029768848,
      "tokens_seen": 1966080
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.075396825396825e-05,
      "loss": 8.55,
      "theoretical_loss": 9.44494752063593,
      "tokens_seen": 2031616
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 8.0979,
      "theoretical_loss": 9.382676773914453,
      "tokens_seen": 2097152
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.273809523809524e-05,
      "loss": 8.3374,
      "theoretical_loss": 9.322848337237977,
      "tokens_seen": 2162688
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.373015873015873e-05,
      "loss": 8.5177,
      "theoretical_loss": 9.26529664966427,
      "tokens_seen": 2228224
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.472222222222222e-05,
      "loss": 8.6686,
      "theoretical_loss": 9.209871830448751,
      "tokens_seen": 2293760
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.571428571428571e-05,
      "loss": 8.675,
      "theoretical_loss": 9.156437795512288,
      "tokens_seen": 2359296
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.670634920634921e-05,
      "loss": 8.4694,
      "theoretical_loss": 9.104870645417346,
      "tokens_seen": 2424832
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.76984126984127e-05,
      "loss": 8.2407,
      "theoretical_loss": 9.05505727953773,
      "tokens_seen": 2490368
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.8690476190476195e-05,
      "loss": 8.5934,
      "theoretical_loss": 9.006894199647824,
      "tokens_seen": 2555904
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.968253968253968e-05,
      "loss": 7.8491,
      "theoretical_loss": 8.960286472913351,
      "tokens_seen": 2621440
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.067460317460318e-05,
      "loss": 7.8615,
      "theoretical_loss": 8.915146829644607,
      "tokens_seen": 2686976
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.1666666666666665e-05,
      "loss": 7.7589,
      "theoretical_loss": 8.871394875482219,
      "tokens_seen": 2752512
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.265873015873016e-05,
      "loss": 8.0014,
      "theoretical_loss": 8.828956401157495,
      "tokens_seen": 2818048
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.365079365079365e-05,
      "loss": 7.8309,
      "theoretical_loss": 8.787762775782408,
      "tokens_seen": 2883584
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.464285714285714e-05,
      "loss": 8.259,
      "theoretical_loss": 8.747750411915451,
      "tokens_seen": 2949120
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.563492063492063e-05,
      "loss": 8.1107,
      "theoretical_loss": 8.708860292524829,
      "tokens_seen": 3014656
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.6626984126984126e-05,
      "loss": 8.2947,
      "theoretical_loss": 8.671037551512843,
      "tokens_seen": 3080192
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.761904761904762e-05,
      "loss": 7.764,
      "theoretical_loss": 8.63423110073947,
      "tokens_seen": 3145728
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.8611111111111115e-05,
      "loss": 8.1678,
      "theoretical_loss": 8.59839329754032,
      "tokens_seen": 3211264
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 13483,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 8.108365058898926,
      "objective/train/theoretical_loss": 8.563479647615063,
      "objective/train/tokens_used": 23736800,
      "theoretical_loss": 8.563479647615063,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.96031746031746e-05,
      "loss": 7.5801,
      "theoretical_loss": 8.563479647615063,
      "tokens_seen": 3276800
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.05952380952381e-05,
      "loss": 7.8434,
      "theoretical_loss": 8.52944853889911,
      "tokens_seen": 3342336
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.1587301587301586e-05,
      "loss": 7.8492,
      "theoretical_loss": 8.496261002649998,
      "tokens_seen": 3407872
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.257936507936508e-05,
      "loss": 7.8312,
      "theoretical_loss": 8.463880498501183,
      "tokens_seen": 3473408
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.357142857142857e-05,
      "loss": 7.3015,
      "theoretical_loss": 8.432272720676774,
      "tokens_seen": 3538944
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.4563492063492063e-05,
      "loss": 7.9474,
      "theoretical_loss": 8.401405422934712,
      "tokens_seen": 3604480
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.555555555555555e-05,
      "loss": 7.6148,
      "theoretical_loss": 8.371248260124204,
      "tokens_seen": 3670016
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.6547619047619046e-05,
      "loss": 7.5108,
      "theoretical_loss": 8.341772644515071,
      "tokens_seen": 3735552
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.753968253968254e-05,
      "loss": 7.6168,
      "theoretical_loss": 8.312951615289325,
      "tokens_seen": 3801088
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.8531746031746036e-05,
      "loss": 7.8436,
      "theoretical_loss": 8.284759719785177,
      "tokens_seen": 3866624
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.9523809523809524e-05,
      "loss": 7.0651,
      "theoretical_loss": 8.257172905255858,
      "tokens_seen": 3932160
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.051587301587302e-05,
      "loss": 7.4859,
      "theoretical_loss": 8.23016842005411,
      "tokens_seen": 3997696
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.15079365079365e-05,
      "loss": 7.531,
      "theoretical_loss": 8.203724723281965,
      "tokens_seen": 4063232
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.25e-05,
      "loss": 7.5815,
      "theoretical_loss": 8.177821402057042,
      "tokens_seen": 4128768
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.349206349206349e-05,
      "loss": 7.2471,
      "theoretical_loss": 8.152439095643611,
      "tokens_seen": 4194304
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.448412698412699e-05,
      "loss": 7.4488,
      "theoretical_loss": 8.127559425781495,
      "tokens_seen": 4259840
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.547619047619048e-05,
      "loss": 7.4517,
      "theoretical_loss": 8.103164932619705,
      "tokens_seen": 4325376
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.646825396825397e-05,
      "loss": 7.1039,
      "theoretical_loss": 8.079239015726635,
      "tokens_seen": 4390912
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.746031746031745e-05,
      "loss": 7.1279,
      "theoretical_loss": 8.055765879705412,
      "tokens_seen": 4456448
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.845238095238096e-05,
      "loss": 7.3354,
      "theoretical_loss": 8.032730483993003,
      "tokens_seen": 4521984
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.944444444444444e-05,
      "loss": 7.1485,
      "theoretical_loss": 8.01011849646583,
      "tokens_seen": 4587520
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.043650793650793e-05,
      "loss": 6.9851,
      "theoretical_loss": 7.987916250513374,
      "tokens_seen": 4653056
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.142857142857142e-05,
      "loss": 6.9432,
      "theoretical_loss": 7.966110705275796,
      "tokens_seen": 4718592
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.242063492063492e-05,
      "loss": 6.9933,
      "theoretical_loss": 7.944689408771984,
      "tokens_seen": 4784128
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.341269841269842e-05,
      "loss": 7.0325,
      "theoretical_loss": 7.923640463671514,
      "tokens_seen": 4849664
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 14033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.057826519012451,
      "objective/train/theoretical_loss": 7.902952495488062,
      "objective/train/tokens_used": 25375200,
      "theoretical_loss": 7.902952495488062,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.440476190476191e-05,
      "loss": 6.7791,
      "theoretical_loss": 7.902952495488062,
      "tokens_seen": 4915200
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.53968253968254e-05,
      "loss": 6.7978,
      "theoretical_loss": 7.882614622993199,
      "tokens_seen": 4980736
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.63888888888889e-05,
      "loss": 6.8558,
      "theoretical_loss": 7.8626164306686235,
      "tokens_seen": 5046272
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.738095238095239e-05,
      "loss": 6.9141,
      "theoretical_loss": 7.842947943031932,
      "tokens_seen": 5111808
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.837301587301588e-05,
      "loss": 6.9255,
      "theoretical_loss": 7.8235996006863076,
      "tokens_seen": 5177344
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.936507936507937e-05,
      "loss": 6.7895,
      "theoretical_loss": 7.8045622379581605,
      "tokens_seen": 5242880
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.035714285714287e-05,
      "loss": 6.7059,
      "theoretical_loss": 7.785827061999068,
      "tokens_seen": 5308416
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.134920634920635e-05,
      "loss": 6.815,
      "theoretical_loss": 7.767385633239359,
      "tokens_seen": 5373952
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.234126984126984e-05,
      "loss": 6.8445,
      "theoretical_loss": 7.749229847090614,
      "tokens_seen": 5439488
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.333333333333333e-05,
      "loss": 6.7849,
      "theoretical_loss": 7.731351916803314,
      "tokens_seen": 5505024
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.432539682539683e-05,
      "loss": 6.5349,
      "theoretical_loss": 7.7137443573939155,
      "tokens_seen": 5570560
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.531746031746032e-05,
      "loss": 6.5103,
      "theoretical_loss": 7.696399970562929,
      "tokens_seen": 5636096
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.630952380952381e-05,
      "loss": 6.9307,
      "theoretical_loss": 7.679311830532177,
      "tokens_seen": 5701632
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.73015873015873e-05,
      "loss": 6.8115,
      "theoretical_loss": 7.662473270735324,
      "tokens_seen": 5767168
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.82936507936508e-05,
      "loss": 6.6601,
      "theoretical_loss": 7.645877871301275,
      "tokens_seen": 5832704
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.928571428571429e-05,
      "loss": 6.2238,
      "theoretical_loss": 7.6295194472748555,
      "tokens_seen": 5898240
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.027777777777777e-05,
      "loss": 6.6176,
      "theoretical_loss": 7.613392037523692,
      "tokens_seen": 5963776
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.126984126984126e-05,
      "loss": 6.5932,
      "theoretical_loss": 7.597489894284275,
      "tokens_seen": 6029312
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.226190476190476e-05,
      "loss": 6.51,
      "theoretical_loss": 7.581807473303883,
      "tokens_seen": 6094848
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.325396825396825e-05,
      "loss": 6.3768,
      "theoretical_loss": 7.566339424538366,
      "tokens_seen": 6160384
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.424603174603175e-05,
      "loss": 6.5465,
      "theoretical_loss": 7.551080583368945,
      "tokens_seen": 6225920
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.523809523809524e-05,
      "loss": 6.8931,
      "theoretical_loss": 7.536025962303915,
      "tokens_seen": 6291456
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.623015873015874e-05,
      "loss": 6.5266,
      "theoretical_loss": 7.521170743133757,
      "tokens_seen": 6356992
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.722222222222223e-05,
      "loss": 6.4856,
      "theoretical_loss": 7.506510269510493,
      "tokens_seen": 6422528
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.821428571428572e-05,
      "loss": 6.4763,
      "theoretical_loss": 7.492040039924245,
      "tokens_seen": 6488064
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 15357,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.699707984924316,
      "objective/train/theoretical_loss": 7.4777557010520255,
      "objective/train/tokens_used": 27013600,
      "theoretical_loss": 7.4777557010520255,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.92063492063492e-05,
      "loss": 6.6532,
      "theoretical_loss": 7.4777557010520255,
      "tokens_seen": 6553600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010019841269841271,
      "loss": 6.4,
      "theoretical_loss": 7.463653041455457,
      "tokens_seen": 6619136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001011904761904762,
      "loss": 6.2462,
      "theoretical_loss": 7.449727985605911,
      "tokens_seen": 6684672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010218253968253968,
      "loss": 6.6399,
      "theoretical_loss": 7.435976588217015,
      "tokens_seen": 6750208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010317460317460317,
      "loss": 6.1887,
      "theoretical_loss": 7.4223950288659,
      "tokens_seen": 6815744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010416666666666667,
      "loss": 6.4129,
      "theoretical_loss": 7.40897960688587,
      "tokens_seen": 6881280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010515873015873016,
      "loss": 6.2296,
      "theoretical_loss": 7.395726736514341,
      "tokens_seen": 6946816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010615079365079365,
      "loss": 6.3505,
      "theoretical_loss": 7.3826329422810275,
      "tokens_seen": 7012352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010714285714285714,
      "loss": 6.5043,
      "theoretical_loss": 7.369694854622368,
      "tokens_seen": 7077888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010813492063492064,
      "loss": 6.3689,
      "theoretical_loss": 7.3569092057090835,
      "tokens_seen": 7143424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00010912698412698413,
      "loss": 6.5583,
      "theoretical_loss": 7.3442728254746825,
      "tokens_seen": 7208960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011011904761904761,
      "loss": 6.3757,
      "theoretical_loss": 7.3317826378335145,
      "tokens_seen": 7274496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001111111111111111,
      "loss": 6.3164,
      "theoretical_loss": 7.319435657077673,
      "tokens_seen": 7340032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001121031746031746,
      "loss": 6.205,
      "theoretical_loss": 7.307228984442819,
      "tokens_seen": 7405568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011309523809523809,
      "loss": 6.4975,
      "theoretical_loss": 7.29515980483356,
      "tokens_seen": 7471104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011408730158730158,
      "loss": 6.087,
      "theoretical_loss": 7.283225383699648,
      "tokens_seen": 7536640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011507936507936508,
      "loss": 6.0785,
      "theoretical_loss": 7.271423064054828,
      "tokens_seen": 7602176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011607142857142858,
      "loss": 6.4949,
      "theoretical_loss": 7.259750263630618,
      "tokens_seen": 7667712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011706349206349207,
      "loss": 6.2155,
      "theoretical_loss": 7.248204472157866,
      "tokens_seen": 7733248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011805555555555556,
      "loss": 6.0229,
      "theoretical_loss": 7.236783248769257,
      "tokens_seen": 7798784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011904761904761905,
      "loss": 6.6244,
      "theoretical_loss": 7.225484219516492,
      "tokens_seen": 7864320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012003968253968255,
      "loss": 5.9387,
      "theoretical_loss": 7.2143050749961155,
      "tokens_seen": 7929856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012103174603174604,
      "loss": 6.2842,
      "theoretical_loss": 7.2032435680783955,
      "tokens_seen": 7995392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012202380952380953,
      "loss": 6.2047,
      "theoretical_loss": 7.19229751173401,
      "tokens_seen": 8060928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000123015873015873,
      "loss": 6.2571,
      "theoretical_loss": 7.181464776953511,
      "tokens_seen": 8126464
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 16711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.2001729011535645,
      "objective/train/theoretical_loss": 7.1707432907549595,
      "objective/train/tokens_used": 28652000,
      "theoretical_loss": 7.1707432907549595,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001240079365079365,
      "loss": 6.0533,
      "theoretical_loss": 7.1707432907549595,
      "tokens_seen": 8192000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000125,
      "loss": 6.0526,
      "theoretical_loss": 7.160131034275267,
      "tokens_seen": 8257536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001259920634920635,
      "loss": 6.0532,
      "theoretical_loss": 7.149626040941133,
      "tokens_seen": 8323072
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012698412698412698,
      "loss": 5.9719,
      "theoretical_loss": 7.139226394715633,
      "tokens_seen": 8388608
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012797619047619048,
      "loss": 6.0403,
      "theoretical_loss": 7.12893022841679,
      "tokens_seen": 8454144
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012896825396825398,
      "loss": 5.7847,
      "theoretical_loss": 7.118735722104612,
      "tokens_seen": 8519680
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012996031746031748,
      "loss": 5.9094,
      "theoretical_loss": 7.10864110153334,
      "tokens_seen": 8585216
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013095238095238096,
      "loss": 6.0872,
      "theoretical_loss": 7.098644636665766,
      "tokens_seen": 8650752
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013194444444444446,
      "loss": 6.0185,
      "theoretical_loss": 7.088744640246688,
      "tokens_seen": 8716288
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013293650793650793,
      "loss": 6.0768,
      "theoretical_loss": 7.078939466432745,
      "tokens_seen": 8781824
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013392857142857144,
      "loss": 6.3022,
      "theoretical_loss": 7.069227509475962,
      "tokens_seen": 8847360
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001349206349206349,
      "loss": 6.0881,
      "theoretical_loss": 7.059607202458563,
      "tokens_seen": 8912896
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001359126984126984,
      "loss": 5.7381,
      "theoretical_loss": 7.0500770160766635,
      "tokens_seen": 8978432
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001369047619047619,
      "loss": 5.9916,
      "theoretical_loss": 7.04063545747062,
      "tokens_seen": 9043968
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013789682539682541,
      "loss": 6.0983,
      "theoretical_loss": 7.031281069099929,
      "tokens_seen": 9109504
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001388888888888889,
      "loss": 6.0662,
      "theoretical_loss": 7.022012427660673,
      "tokens_seen": 9175040
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001398809523809524,
      "loss": 5.7078,
      "theoretical_loss": 7.012828143043604,
      "tokens_seen": 9240576
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014087301587301586,
      "loss": 6.2455,
      "theoretical_loss": 7.003726857331067,
      "tokens_seen": 9306112
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014186507936507937,
      "loss": 5.9609,
      "theoretical_loss": 6.994707243831053,
      "tokens_seen": 9371648
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014285714285714284,
      "loss": 5.9423,
      "theoretical_loss": 6.985768006146776,
      "tokens_seen": 9437184
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014384920634920634,
      "loss": 6.0329,
      "theoretical_loss": 6.976907877280185,
      "tokens_seen": 9502720
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014484126984126984,
      "loss": 5.7799,
      "theoretical_loss": 6.96812561876801,
      "tokens_seen": 9568256
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014583333333333335,
      "loss": 5.9063,
      "theoretical_loss": 6.95942001984889,
      "tokens_seen": 9633792
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014682539682539685,
      "loss": 5.9317,
      "theoretical_loss": 6.950789896660304,
      "tokens_seen": 9699328
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014781746031746032,
      "loss": 5.8925,
      "theoretical_loss": 6.942234091464025,
      "tokens_seen": 9764864
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 17166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 6.479184627532959,
      "objective/train/theoretical_loss": 6.933751471898896,
      "objective/train/tokens_used": 30290400,
      "theoretical_loss": 6.933751471898896,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014880952380952382,
      "loss": 6.2539,
      "theoretical_loss": 6.933751471898896,
      "tokens_seen": 9830400
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001498015873015873,
      "loss": 5.996,
      "theoretical_loss": 6.925340930259813,
      "tokens_seen": 9895936
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001507936507936508,
      "loss": 5.8197,
      "theoretical_loss": 6.917001382801793,
      "tokens_seen": 9961472
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015178571428571427,
      "loss": 6.0534,
      "theoretical_loss": 6.908731769068142,
      "tokens_seen": 10027008
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001527777777777778,
      "loss": 5.7838,
      "theoretical_loss": 6.900531051241684,
      "tokens_seen": 10092544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015376984126984128,
      "loss": 5.7799,
      "theoretical_loss": 6.892398213518156,
      "tokens_seen": 10158080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015476190476190478,
      "loss": 6.1056,
      "theoretical_loss": 6.884332261500866,
      "tokens_seen": 10223616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015575396825396825,
      "loss": 5.5432,
      "theoretical_loss": 6.876332221615746,
      "tokens_seen": 10289152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015674603174603175,
      "loss": 5.8572,
      "theoretical_loss": 6.8683971405459925,
      "tokens_seen": 10354688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015773809523809523,
      "loss": 6.226,
      "theoretical_loss": 6.860526084685546,
      "tokens_seen": 10420224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015873015873015873,
      "loss": 5.7117,
      "theoretical_loss": 6.852718139610625,
      "tokens_seen": 10485760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001597222222222222,
      "loss": 5.6684,
      "theoretical_loss": 6.844972409568637,
      "tokens_seen": 10551296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016071428571428573,
      "loss": 5.8166,
      "theoretical_loss": 6.8372880169837895,
      "tokens_seen": 10616832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001617063492063492,
      "loss": 5.971,
      "theoretical_loss": 6.829664101978745,
      "tokens_seen": 10682368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001626984126984127,
      "loss": 5.7343,
      "theoretical_loss": 6.8220998219117135,
      "tokens_seen": 10747904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016369047619047618,
      "loss": 5.8061,
      "theoretical_loss": 6.81459435092839,
      "tokens_seen": 10813440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016468253968253969,
      "loss": 6.0029,
      "theoretical_loss": 6.807146879528176,
      "tokens_seen": 10878976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016567460317460316,
      "loss": 5.5367,
      "theoretical_loss": 6.79975661414414,
      "tokens_seen": 10944512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016666666666666666,
      "loss": 5.6256,
      "theoretical_loss": 6.79242277673622,
      "tokens_seen": 11010048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016765873015873016,
      "loss": 5.7868,
      "theoretical_loss": 6.785144604397143,
      "tokens_seen": 11075584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016865079365079366,
      "loss": 5.6046,
      "theoretical_loss": 6.777921348970633,
      "tokens_seen": 11141120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016964285714285717,
      "loss": 5.9753,
      "theoretical_loss": 6.770752276681419,
      "tokens_seen": 11206656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017063492063492064,
      "loss": 5.9022,
      "theoretical_loss": 6.76363666777662,
      "tokens_seen": 11272192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017162698412698414,
      "loss": 5.8921,
      "theoretical_loss": 6.756573816178118,
      "tokens_seen": 11337728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017261904761904762,
      "loss": 5.9722,
      "theoretical_loss": 6.749563029145481,
      "tokens_seen": 11403264
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 18369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.672413349151611,
      "objective/train/theoretical_loss": 6.742603626949085,
      "objective/train/tokens_used": 31928800,
      "theoretical_loss": 6.742603626949085,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017361111111111112,
      "loss": 5.4506,
      "theoretical_loss": 6.742603626949085,
      "tokens_seen": 11468800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001746031746031746,
      "loss": 6.0078,
      "theoretical_loss": 6.735694942553071,
      "tokens_seen": 11534336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001755952380952381,
      "loss": 5.8091,
      "theoretical_loss": 6.728836321307757,
      "tokens_seen": 11599872
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001765873015873016,
      "loss": 5.6113,
      "theoretical_loss": 6.722027120651221,
      "tokens_seen": 11665408
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001775793650793651,
      "loss": 5.6758,
      "theoretical_loss": 6.715266709819684,
      "tokens_seen": 11730944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017857142857142857,
      "loss": 5.6378,
      "theoretical_loss": 6.708554469566408,
      "tokens_seen": 11796480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017956349206349207,
      "loss": 5.7411,
      "theoretical_loss": 6.701889791888819,
      "tokens_seen": 11862016
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018055555555555555,
      "loss": 5.8311,
      "theoretical_loss": 6.695272079763555,
      "tokens_seen": 11927552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018154761904761905,
      "loss": 5.2846,
      "theoretical_loss": 6.688700746889175,
      "tokens_seen": 11993088
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018253968253968252,
      "loss": 5.7687,
      "theoretical_loss": 6.682175217436269,
      "tokens_seen": 12058624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018353174603174602,
      "loss": 5.7687,
      "theoretical_loss": 6.675694925804711,
      "tokens_seen": 12124160
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018452380952380953,
      "loss": 5.6209,
      "theoretical_loss": 6.66925931638781,
      "tokens_seen": 12189696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018551587301587303,
      "loss": 5.9193,
      "theoretical_loss": 6.662867843343154,
      "tokens_seen": 12255232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001865079365079365,
      "loss": 5.6754,
      "theoretical_loss": 6.6565199703698745,
      "tokens_seen": 12320768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001875,
      "loss": 5.8597,
      "theoretical_loss": 6.6502151704921655,
      "tokens_seen": 12386304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001884920634920635,
      "loss": 5.709,
      "theoretical_loss": 6.643952925848826,
      "tokens_seen": 12451840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018948412698412698,
      "loss": 5.7054,
      "theoretical_loss": 6.637732727488622,
      "tokens_seen": 12517376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019047619047619048,
      "loss": 5.5972,
      "theoretical_loss": 6.6315540751713,
      "tokens_seen": 12582912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019146825396825398,
      "loss": 5.5887,
      "theoretical_loss": 6.625416477174044,
      "tokens_seen": 12648448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019246031746031748,
      "loss": 5.7082,
      "theoretical_loss": 6.619319450103218,
      "tokens_seen": 12713984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019345238095238096,
      "loss": 5.5405,
      "theoretical_loss": 6.613262518711206,
      "tokens_seen": 12779520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019444444444444446,
      "loss": 5.5806,
      "theoretical_loss": 6.607245215718208,
      "tokens_seen": 12845056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019543650793650793,
      "loss": 5.4612,
      "theoretical_loss": 6.601267081638806,
      "tokens_seen": 12910592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019642857142857144,
      "loss": 5.6405,
      "theoretical_loss": 6.595327664613176,
      "tokens_seen": 12976128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001974206349206349,
      "loss": 5.3968,
      "theoretical_loss": 6.58942652024278,
      "tokens_seen": 13041664
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 19017,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.398904323577881,
      "objective/train/theoretical_loss": 6.583563211430409,
      "objective/train/tokens_used": 33567200,
      "theoretical_loss": 6.583563211430409,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001984126984126984,
      "loss": 5.5392,
      "theoretical_loss": 6.583563211430409,
      "tokens_seen": 13107200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019940476190476191,
      "loss": 5.3995,
      "theoretical_loss": 6.577737308224426,
      "tokens_seen": 13172736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020039682539682542,
      "loss": 5.5665,
      "theoretical_loss": 6.571948387667094,
      "tokens_seen": 13238272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002013888888888889,
      "loss": 5.5745,
      "theoretical_loss": 6.566196033646859,
      "tokens_seen": 13303808
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002023809523809524,
      "loss": 5.4756,
      "theoretical_loss": 6.560479836754459,
      "tokens_seen": 13369344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020337301587301587,
      "loss": 5.7895,
      "theoretical_loss": 6.554799394142741,
      "tokens_seen": 13434880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020436507936507937,
      "loss": 5.3965,
      "theoretical_loss": 6.549154309390087,
      "tokens_seen": 13500416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020535714285714284,
      "loss": 5.494,
      "theoretical_loss": 6.5435441923673245,
      "tokens_seen": 13565952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020634920634920634,
      "loss": 5.5559,
      "theoretical_loss": 6.537968659108021,
      "tokens_seen": 13631488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020734126984126985,
      "loss": 5.9125,
      "theoretical_loss": 6.532427331682063,
      "tokens_seen": 13697024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020833333333333335,
      "loss": 5.5235,
      "theoretical_loss": 6.5269198380724145,
      "tokens_seen": 13762560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00020932539682539685,
      "loss": 5.6307,
      "theoretical_loss": 6.5214458120549725,
      "tokens_seen": 13828096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021031746031746032,
      "loss": 5.3661,
      "theoretical_loss": 6.516004893081412,
      "tokens_seen": 13893632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021130952380952382,
      "loss": 5.2823,
      "theoretical_loss": 6.510596726164945,
      "tokens_seen": 13959168
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002123015873015873,
      "loss": 5.4241,
      "theoretical_loss": 6.505220961768906,
      "tokens_seen": 14024704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002132936507936508,
      "loss": 5.3679,
      "theoretical_loss": 6.499877255698063,
      "tokens_seen": 14090240
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021428571428571427,
      "loss": 5.2005,
      "theoretical_loss": 6.494565268992613,
      "tokens_seen": 14155776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002152777777777778,
      "loss": 5.6155,
      "theoretical_loss": 6.489284667824741,
      "tokens_seen": 14221312
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021626984126984128,
      "loss": 5.6027,
      "theoretical_loss": 6.484035123397694,
      "tokens_seen": 14286848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021726190476190478,
      "loss": 5.5959,
      "theoretical_loss": 6.478816311847305,
      "tokens_seen": 14352384
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021825396825396825,
      "loss": 5.7842,
      "theoretical_loss": 6.473627914145865,
      "tokens_seen": 14417920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021924603174603176,
      "loss": 5.4174,
      "theoretical_loss": 6.468469616008303,
      "tokens_seen": 14483456
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022023809523809523,
      "loss": 5.4423,
      "theoretical_loss": 6.463341107800611,
      "tokens_seen": 14548992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022123015873015873,
      "loss": 5.2413,
      "theoretical_loss": 6.458242084450426,
      "tokens_seen": 14614528
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002222222222222222,
      "loss": 5.4841,
      "theoretical_loss": 6.453172245359726,
      "tokens_seen": 14680064
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 20187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.4210638999938965,
      "objective/train/theoretical_loss": 6.448131294319589,
      "objective/train/tokens_used": 35205600,
      "theoretical_loss": 6.448131294319589,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022321428571428573,
      "loss": 5.318,
      "theoretical_loss": 6.448131294319589,
      "tokens_seen": 14745600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002242063492063492,
      "loss": 5.5298,
      "theoretical_loss": 6.443118939426935,
      "tokens_seen": 14811136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002251984126984127,
      "loss": 5.5104,
      "theoretical_loss": 6.438134893003209,
      "tokens_seen": 14876672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022619047619047618,
      "loss": 5.4089,
      "theoretical_loss": 6.43317887151496,
      "tokens_seen": 14942208
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022718253968253969,
      "loss": 5.5783,
      "theoretical_loss": 6.428250595496234,
      "tokens_seen": 15007744
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022817460317460316,
      "loss": 5.4029,
      "theoretical_loss": 6.4233497894727805,
      "tokens_seen": 15073280
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022916666666666666,
      "loss": 5.6043,
      "theoretical_loss": 6.418476181887964,
      "tokens_seen": 15138816
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023015873015873016,
      "loss": 5.5123,
      "theoretical_loss": 6.413629505030386,
      "tokens_seen": 15204352
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023115079365079367,
      "loss": 5.5084,
      "theoretical_loss": 6.408809494963135,
      "tokens_seen": 15269888
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023214285714285717,
      "loss": 5.269,
      "theoretical_loss": 6.404015891454651,
      "tokens_seen": 15335424
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023313492063492064,
      "loss": 5.4379,
      "theoretical_loss": 6.399248437911128,
      "tokens_seen": 15400960
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023412698412698414,
      "loss": 5.4346,
      "theoretical_loss": 6.394506881310447,
      "tokens_seen": 15466496
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023511904761904762,
      "loss": 5.7191,
      "theoretical_loss": 6.38979097213757,
      "tokens_seen": 15532032
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023611111111111112,
      "loss": 5.5484,
      "theoretical_loss": 6.385100464321383,
      "tokens_seen": 15597568
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002371031746031746,
      "loss": 5.4379,
      "theoretical_loss": 6.380435115172919,
      "tokens_seen": 15663104
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002380952380952381,
      "loss": 5.4785,
      "theoretical_loss": 6.375794685324962,
      "tokens_seen": 15728640
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002390873015873016,
      "loss": 5.2088,
      "theoretical_loss": 6.371178938672962,
      "tokens_seen": 15794176
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002400793650793651,
      "loss": 5.5076,
      "theoretical_loss": 6.366587642317243,
      "tokens_seen": 15859712
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024107142857142857,
      "loss": 5.6351,
      "theoretical_loss": 6.362020566506475,
      "tokens_seen": 15925248
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024206349206349207,
      "loss": 5.6029,
      "theoretical_loss": 6.3574774845823665,
      "tokens_seen": 15990784
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024305555555555555,
      "loss": 5.1969,
      "theoretical_loss": 6.352958172925549,
      "tokens_seen": 16056320
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024404761904761905,
      "loss": 5.285,
      "theoretical_loss": 6.348462410902625,
      "tokens_seen": 16121856
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024503968253968255,
      "loss": 5.6805,
      "theoretical_loss": 6.343989980814362,
      "tokens_seen": 16187392
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246031746031746,
      "loss": 5.5855,
      "theoretical_loss": 6.339540667844965,
      "tokens_seen": 16252928
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024702380952380955,
      "loss": 5.3559,
      "theoretical_loss": 6.335114260012455,
      "tokens_seen": 16318464
    },
    {
      "epoch": 0.0,
      "objective/train/docs_used": 20737,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.463630199432373,
      "objective/train/theoretical_loss": 6.330710548120079,
      "objective/train/tokens_used": 36844000,
      "theoretical_loss": 6.330710548120079,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248015873015873,
      "loss": 5.515,
      "theoretical_loss": 6.330710548120079,
      "tokens_seen": 16384000
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002490079365079365,
      "loss": 5.3596,
      "theoretical_loss": 6.326329325708746,
      "tokens_seen": 16449536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 5.249,
      "theoretical_loss": 6.321970389010465,
      "tokens_seen": 16515072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002509920634920635,
      "loss": 5.3675,
      "theoretical_loss": 6.317633536902759,
      "tokens_seen": 16580608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000251984126984127,
      "loss": 5.0919,
      "theoretical_loss": 6.313318570864016,
      "tokens_seen": 16646144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025297619047619046,
      "loss": 5.1228,
      "theoretical_loss": 6.30902529492978,
      "tokens_seen": 16711680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025396825396825396,
      "loss": 5.4573,
      "theoretical_loss": 6.304753515649935,
      "tokens_seen": 16777216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025496031746031746,
      "loss": 5.1499,
      "theoretical_loss": 6.30050304204677,
      "tokens_seen": 16842752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025595238095238096,
      "loss": 5.6432,
      "theoretical_loss": 6.296273685573913,
      "tokens_seen": 16908288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002569444444444444,
      "loss": 5.439,
      "theoretical_loss": 6.292065260076094,
      "tokens_seen": 16973824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025793650793650796,
      "loss": 5.2959,
      "theoretical_loss": 6.287877581749726,
      "tokens_seen": 17039360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025892857142857146,
      "loss": 5.3502,
      "theoretical_loss": 6.2837104691042915,
      "tokens_seen": 17104896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025992063492063497,
      "loss": 5.5134,
      "theoretical_loss": 6.279563742924502,
      "tokens_seen": 17170432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002609126984126984,
      "loss": 5.4871,
      "theoretical_loss": 6.275437226233224,
      "tokens_seen": 17235968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002619047619047619,
      "loss": 5.3132,
      "theoretical_loss": 6.271330744255137,
      "tokens_seen": 17301504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002628968253968254,
      "loss": 5.1796,
      "theoretical_loss": 6.267244124381133,
      "tokens_seen": 17367040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002638888888888889,
      "loss": 5.2656,
      "theoretical_loss": 6.2631771961334035,
      "tokens_seen": 17432576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026488095238095237,
      "loss": 5.5896,
      "theoretical_loss": 6.259129791131242,
      "tokens_seen": 17498112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026587301587301587,
      "loss": 5.14,
      "theoretical_loss": 6.255101743057493,
      "tokens_seen": 17563648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026686507936507937,
      "loss": 5.3606,
      "theoretical_loss": 6.251092887625685,
      "tokens_seen": 17629184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026785714285714287,
      "loss": 5.487,
      "theoretical_loss": 6.247103062547796,
      "tokens_seen": 17694720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002688492063492063,
      "loss": 5.2887,
      "theoretical_loss": 6.243132107502647,
      "tokens_seen": 17760256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002698412698412698,
      "loss": 5.4254,
      "theoretical_loss": 6.239179864104911,
      "tokens_seen": 17825792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002708333333333333,
      "loss": 5.4668,
      "theoretical_loss": 6.235246175874727,
      "tokens_seen": 17891328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002718253968253968,
      "loss": 5.3584,
      "theoretical_loss": 6.231330888207894,
      "tokens_seen": 17956864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 21782,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.119491100311279,
      "objective/train/theoretical_loss": 6.227433848346639,
      "objective/train/tokens_used": 38482400,
      "theoretical_loss": 6.227433848346639,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002728174603174603,
      "loss": 5.1069,
      "theoretical_loss": 6.227433848346639,
      "tokens_seen": 18022400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002738095238095238,
      "loss": 5.1568,
      "theoretical_loss": 6.22355490535095,
      "tokens_seen": 18087936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002748015873015873,
      "loss": 5.4409,
      "theoretical_loss": 6.219693910070452,
      "tokens_seen": 18153472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027579365079365083,
      "loss": 5.1939,
      "theoretical_loss": 6.215850715116817,
      "tokens_seen": 18219008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027678571428571433,
      "loss": 5.2274,
      "theoretical_loss": 6.212025174836697,
      "tokens_seen": 18284544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002777777777777778,
      "loss": 5.5299,
      "theoretical_loss": 6.208217145285173,
      "tokens_seen": 18350080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002787698412698413,
      "loss": 5.4441,
      "theoretical_loss": 6.204426484199696,
      "tokens_seen": 18415616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002797619047619048,
      "loss": 5.3135,
      "theoretical_loss": 6.200653050974525,
      "tokens_seen": 18481152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002807539682539683,
      "loss": 5.189,
      "theoretical_loss": 6.196896706635635,
      "tokens_seen": 18546688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028174603174603173,
      "loss": 5.0786,
      "theoretical_loss": 6.1931573138160925,
      "tokens_seen": 18612224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028273809523809523,
      "loss": 5.0434,
      "theoretical_loss": 6.189434736731897,
      "tokens_seen": 18677760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028373015873015873,
      "loss": 5.0132,
      "theoretical_loss": 6.185728841158257,
      "tokens_seen": 18743296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028472222222222223,
      "loss": 5.1159,
      "theoretical_loss": 6.182039494406309,
      "tokens_seen": 18808832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002857142857142857,
      "loss": 5.1786,
      "theoretical_loss": 6.178366565300266,
      "tokens_seen": 18874368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002867063492063492,
      "loss": 5.1947,
      "theoretical_loss": 6.174709924154978,
      "tokens_seen": 18939904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002876984126984127,
      "loss": 5.2396,
      "theoretical_loss": 6.171069442753909,
      "tokens_seen": 19005440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002886904761904762,
      "loss": 5.3947,
      "theoretical_loss": 6.1674449943275045,
      "tokens_seen": 19070976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002896825396825397,
      "loss": 4.8732,
      "theoretical_loss": 6.16383645353196,
      "tokens_seen": 19136512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002906746031746032,
      "loss": 5.3178,
      "theoretical_loss": 6.160243696428367,
      "tokens_seen": 19202048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002916666666666667,
      "loss": 5.1441,
      "theoretical_loss": 6.156666600462238,
      "tokens_seen": 19267584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002926587301587302,
      "loss": 5.2695,
      "theoretical_loss": 6.153105044443393,
      "tokens_seen": 19333120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002936507936507937,
      "loss": 5.3181,
      "theoretical_loss": 6.149558908526206,
      "tokens_seen": 19398656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029464285714285714,
      "loss": 5.2188,
      "theoretical_loss": 6.146028074190217,
      "tokens_seen": 19464192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029563492063492064,
      "loss": 5.1652,
      "theoretical_loss": 6.14251242422106,
      "tokens_seen": 19529728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029662698412698414,
      "loss": 5.3475,
      "theoretical_loss": 6.139011842691756,
      "tokens_seen": 19595264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 22379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.661308765411377,
      "objective/train/theoretical_loss": 6.135526214944321,
      "objective/train/tokens_used": 40120800,
      "theoretical_loss": 6.135526214944321,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029761904761904765,
      "loss": 5.6218,
      "theoretical_loss": 6.135526214944321,
      "tokens_seen": 19660800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002986111111111111,
      "loss": 5.4528,
      "theoretical_loss": 6.1320554275717,
      "tokens_seen": 19726336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002996031746031746,
      "loss": 5.1971,
      "theoretical_loss": 6.128599368400017,
      "tokens_seen": 19791872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003005952380952381,
      "loss": 5.2823,
      "theoretical_loss": 6.125157926471134,
      "tokens_seen": 19857408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003015873015873016,
      "loss": 5.1154,
      "theoretical_loss": 6.121730992025516,
      "tokens_seen": 19922944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030257936507936505,
      "loss": 4.8293,
      "theoretical_loss": 6.118318456485394,
      "tokens_seen": 19988480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030357142857142855,
      "loss": 5.0591,
      "theoretical_loss": 6.114920212438209,
      "tokens_seen": 20054016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030456349206349205,
      "loss": 5.2452,
      "theoretical_loss": 6.111536153620355,
      "tokens_seen": 20119552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003055555555555556,
      "loss": 5.1158,
      "theoretical_loss": 6.108166174901191,
      "tokens_seen": 20185088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030654761904761905,
      "loss": 5.0927,
      "theoretical_loss": 6.104810172267331,
      "tokens_seen": 20250624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030753968253968255,
      "loss": 5.1097,
      "theoretical_loss": 6.101468042807199,
      "tokens_seen": 20316160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030853174603174605,
      "loss": 5.0109,
      "theoretical_loss": 6.098139684695851,
      "tokens_seen": 20381696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00030952380952380956,
      "loss": 5.2,
      "theoretical_loss": 6.094824997180048,
      "tokens_seen": 20447232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000310515873015873,
      "loss": 4.919,
      "theoretical_loss": 6.091523880563589,
      "tokens_seen": 20512768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003115079365079365,
      "loss": 4.7726,
      "theoretical_loss": 6.088236236192881,
      "tokens_seen": 20578304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003125,
      "loss": 5.1579,
      "theoretical_loss": 6.0849619664427586,
      "tokens_seen": 20643840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003134920634920635,
      "loss": 5.2649,
      "theoretical_loss": 6.0817009747025415,
      "tokens_seen": 20709376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000314484126984127,
      "loss": 4.9338,
      "theoretical_loss": 6.07845316536232,
      "tokens_seen": 20774912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031547619047619046,
      "loss": 5.2449,
      "theoretical_loss": 6.075218443799468,
      "tokens_seen": 20840448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031646825396825396,
      "loss": 5.2222,
      "theoretical_loss": 6.071996716365382,
      "tokens_seen": 20905984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031746031746031746,
      "loss": 4.8122,
      "theoretical_loss": 6.068787890372443,
      "tokens_seen": 20971520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031845238095238096,
      "loss": 4.983,
      "theoretical_loss": 6.0655918740811865,
      "tokens_seen": 21037056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003194444444444444,
      "loss": 4.9428,
      "theoretical_loss": 6.062408576687682,
      "tokens_seen": 21102592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032043650793650796,
      "loss": 5.0493,
      "theoretical_loss": 6.059237908311129,
      "tokens_seen": 21168128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032142857142857147,
      "loss": 4.8675,
      "theoretical_loss": 6.056079779981644,
      "tokens_seen": 21233664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 23601,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.199455738067627,
      "objective/train/theoretical_loss": 6.052934103628253,
      "objective/train/tokens_used": 41759200,
      "theoretical_loss": 6.052934103628253,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032242063492063497,
      "loss": 5.2637,
      "theoretical_loss": 6.052934103628253,
      "tokens_seen": 21299200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003234126984126984,
      "loss": 5.4325,
      "theoretical_loss": 6.049800792067078,
      "tokens_seen": 21364736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003244047619047619,
      "loss": 4.8274,
      "theoretical_loss": 6.046679758989709,
      "tokens_seen": 21430272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003253968253968254,
      "loss": 5.2104,
      "theoretical_loss": 6.043570918951775,
      "tokens_seen": 21495808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003263888888888889,
      "loss": 5.1036,
      "theoretical_loss": 6.040474187361681,
      "tokens_seen": 21561344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032738095238095237,
      "loss": 4.8811,
      "theoretical_loss": 6.037389480469548,
      "tokens_seen": 21626880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032837301587301587,
      "loss": 5.0504,
      "theoretical_loss": 6.034316715356304,
      "tokens_seen": 21692416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032936507936507937,
      "loss": 5.0898,
      "theoretical_loss": 6.031255809922974,
      "tokens_seen": 21757952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033035714285714287,
      "loss": 5.0121,
      "theoretical_loss": 6.0282066828801195,
      "tokens_seen": 21823488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003313492063492063,
      "loss": 5.1363,
      "theoretical_loss": 6.0251692537374595,
      "tokens_seen": 21889024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003323412698412698,
      "loss": 5.1503,
      "theoretical_loss": 6.022143442793643,
      "tokens_seen": 21954560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003333333333333333,
      "loss": 5.1066,
      "theoretical_loss": 6.0191291711261945,
      "tokens_seen": 22020096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003343253968253968,
      "loss": 5.0889,
      "theoretical_loss": 6.016126360581609,
      "tokens_seen": 22085632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003353174603174603,
      "loss": 4.9077,
      "theoretical_loss": 6.013134933765605,
      "tokens_seen": 22151168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003363095238095238,
      "loss": 5.2813,
      "theoretical_loss": 6.010154814033528,
      "tokens_seen": 22216704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033730158730158733,
      "loss": 4.6999,
      "theoretical_loss": 6.007185925480902,
      "tokens_seen": 22282240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033829365079365083,
      "loss": 5.0728,
      "theoretical_loss": 6.0042281929341375,
      "tokens_seen": 22347776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033928571428571433,
      "loss": 5.1319,
      "theoretical_loss": 6.001281541941363,
      "tokens_seen": 22413312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003402777777777778,
      "loss": 5.0915,
      "theoretical_loss": 5.998345898763421,
      "tokens_seen": 22478848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003412698412698413,
      "loss": 4.8912,
      "theoretical_loss": 5.995421190364983,
      "tokens_seen": 22544384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003422619047619048,
      "loss": 5.0317,
      "theoretical_loss": 5.992507344405814,
      "tokens_seen": 22609920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003432539682539683,
      "loss": 5.1603,
      "theoretical_loss": 5.9896042892321635,
      "tokens_seen": 22675456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034424603174603173,
      "loss": 4.9822,
      "theoretical_loss": 5.986711953868287,
      "tokens_seen": 22740992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034523809523809523,
      "loss": 5.3948,
      "theoretical_loss": 5.983830268008107,
      "tokens_seen": 22806528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034623015873015873,
      "loss": 5.222,
      "theoretical_loss": 5.980959162006979,
      "tokens_seen": 22872064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 24859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.439995288848877,
      "objective/train/theoretical_loss": 5.978098566873603,
      "objective/train/tokens_used": 43397600,
      "theoretical_loss": 5.978098566873603,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034722222222222224,
      "loss": 4.9666,
      "theoretical_loss": 5.978098566873603,
      "tokens_seen": 22937600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003482142857142857,
      "loss": 4.7873,
      "theoretical_loss": 5.975248414262053,
      "tokens_seen": 23003136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003492063492063492,
      "loss": 5.1429,
      "theoretical_loss": 5.972408636463909,
      "tokens_seen": 23068672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003501984126984127,
      "loss": 4.8387,
      "theoretical_loss": 5.969579166400528,
      "tokens_seen": 23134208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003511904761904762,
      "loss": 5.2741,
      "theoretical_loss": 5.966759937615427,
      "tokens_seen": 23199744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003521825396825397,
      "loss": 4.9993,
      "theoretical_loss": 5.96395088426676,
      "tokens_seen": 23265280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003531746031746032,
      "loss": 4.838,
      "theoretical_loss": 5.961151941119932,
      "tokens_seen": 23330816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003541666666666667,
      "loss": 4.8937,
      "theoretical_loss": 5.95836304354031,
      "tokens_seen": 23396352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003551587301587302,
      "loss": 5.1755,
      "theoretical_loss": 5.9555841274860395,
      "tokens_seen": 23461888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003561507936507937,
      "loss": 5.0192,
      "theoretical_loss": 5.952815129500973,
      "tokens_seen": 23527424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035714285714285714,
      "loss": 5.1528,
      "theoretical_loss": 5.950055986707699,
      "tokens_seen": 23592960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035813492063492064,
      "loss": 5.2473,
      "theoretical_loss": 5.9473066368006755,
      "tokens_seen": 23658496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035912698412698415,
      "loss": 5.1431,
      "theoretical_loss": 5.944567018039454,
      "tokens_seen": 23724032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036011904761904765,
      "loss": 5.2078,
      "theoretical_loss": 5.941837069242023,
      "tokens_seen": 23789568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003611111111111111,
      "loss": 5.2074,
      "theoretical_loss": 5.939116729778224,
      "tokens_seen": 23855104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003621031746031746,
      "loss": 5.045,
      "theoretical_loss": 5.936405939563279,
      "tokens_seen": 23920640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003630952380952381,
      "loss": 5.0272,
      "theoretical_loss": 5.93370463905141,
      "tokens_seen": 23986176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003640873015873016,
      "loss": 5.0628,
      "theoretical_loss": 5.931012769229536,
      "tokens_seen": 24051712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036507936507936505,
      "loss": 4.6127,
      "theoretical_loss": 5.928330271611081,
      "tokens_seen": 24117248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036607142857142855,
      "loss": 4.8408,
      "theoretical_loss": 5.925657088229862,
      "tokens_seen": 24182784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036706349206349205,
      "loss": 4.9603,
      "theoretical_loss": 5.9229931616340545,
      "tokens_seen": 24248320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003680555555555556,
      "loss": 5.0127,
      "theoretical_loss": 5.920338434880263,
      "tokens_seen": 24313856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00036904761904761905,
      "loss": 5.3699,
      "theoretical_loss": 5.9176928515276535,
      "tokens_seen": 24379392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037003968253968255,
      "loss": 4.9974,
      "theoretical_loss": 5.915056355632197,
      "tokens_seen": 24444928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037103174603174606,
      "loss": 4.9348,
      "theoretical_loss": 5.912428891740967,
      "tokens_seen": 24510464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 25411,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.414747714996338,
      "objective/train/theoretical_loss": 5.90981040488653,
      "objective/train/tokens_used": 45036000,
      "theoretical_loss": 5.90981040488653,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037202380952380956,
      "loss": 5.0376,
      "theoretical_loss": 5.90981040488653,
      "tokens_seen": 24576000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000373015873015873,
      "loss": 5.0724,
      "theoretical_loss": 5.907200840581417,
      "tokens_seen": 24641536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003740079365079365,
      "loss": 5.2247,
      "theoretical_loss": 5.904600144812672,
      "tokens_seen": 24707072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000375,
      "loss": 5.4811,
      "theoretical_loss": 5.902008264036468,
      "tokens_seen": 24772608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003759920634920635,
      "loss": 4.9107,
      "theoretical_loss": 5.899425145172803,
      "tokens_seen": 24838144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000376984126984127,
      "loss": 5.0933,
      "theoretical_loss": 5.896850735600281,
      "tokens_seen": 24903680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037797619047619046,
      "loss": 5.1802,
      "theoretical_loss": 5.8942849831509445,
      "tokens_seen": 24969216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037896825396825396,
      "loss": 5.2521,
      "theoretical_loss": 5.891727836105194,
      "tokens_seen": 25034752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037996031746031746,
      "loss": 4.8607,
      "theoretical_loss": 5.889179243186776,
      "tokens_seen": 25100288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038095238095238096,
      "loss": 5.0973,
      "theoretical_loss": 5.886639153557828,
      "tokens_seen": 25165824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003819444444444444,
      "loss": 5.3397,
      "theoretical_loss": 5.8841075168140105,
      "tokens_seen": 25231360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038293650793650797,
      "loss": 4.7777,
      "theoretical_loss": 5.88158428297969,
      "tokens_seen": 25296896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038392857142857147,
      "loss": 4.7219,
      "theoretical_loss": 5.879069402503189,
      "tokens_seen": 25362432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038492063492063497,
      "loss": 5.0299,
      "theoretical_loss": 5.876562826252119,
      "tokens_seen": 25427968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003859126984126984,
      "loss": 4.8369,
      "theoretical_loss": 5.874064505508748,
      "tokens_seen": 25493504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003869047619047619,
      "loss": 5.0017,
      "theoretical_loss": 5.871574391965453,
      "tokens_seen": 25559040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003878968253968254,
      "loss": 4.9705,
      "theoretical_loss": 5.8690924377202265,
      "tokens_seen": 25624576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003888888888888889,
      "loss": 5.034,
      "theoretical_loss": 5.866618595272241,
      "tokens_seen": 25690112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038988095238095237,
      "loss": 4.7181,
      "theoretical_loss": 5.864152817517482,
      "tokens_seen": 25755648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039087301587301587,
      "loss": 4.9562,
      "theoretical_loss": 5.8616950577444245,
      "tokens_seen": 25821184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039186507936507937,
      "loss": 4.9405,
      "theoretical_loss": 5.859245269629783,
      "tokens_seen": 25886720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003928571428571429,
      "loss": 5.099,
      "theoretical_loss": 5.85680340723431,
      "tokens_seen": 25952256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003938492063492063,
      "loss": 4.6503,
      "theoretical_loss": 5.85436942499865,
      "tokens_seen": 26017792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003948412698412698,
      "loss": 4.8672,
      "theoretical_loss": 5.851943277739253,
      "tokens_seen": 26083328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003958333333333333,
      "loss": 5.0201,
      "theoretical_loss": 5.84952492064434,
      "tokens_seen": 26148864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 26668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 5.337137222290039,
      "objective/train/theoretical_loss": 5.847114309269919,
      "objective/train/tokens_used": 46674400,
      "theoretical_loss": 5.847114309269919,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003968253968253968,
      "loss": 5.1138,
      "theoretical_loss": 5.847114309269919,
      "tokens_seen": 26214400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003978174603174603,
      "loss": 4.8239,
      "theoretical_loss": 5.844711399535855,
      "tokens_seen": 26279936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039880952380952383,
      "loss": 4.7304,
      "theoretical_loss": 5.842316147722,
      "tokens_seen": 26345472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039980158730158733,
      "loss": 5.0092,
      "theoretical_loss": 5.839928510464356,
      "tokens_seen": 26411008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040079365079365083,
      "loss": 4.4343,
      "theoretical_loss": 5.837548444751306,
      "tokens_seen": 26476544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040178571428571433,
      "loss": 4.9309,
      "theoretical_loss": 5.835175907919885,
      "tokens_seen": 26542080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004027777777777778,
      "loss": 5.0148,
      "theoretical_loss": 5.832810857652097,
      "tokens_seen": 26607616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004037698412698413,
      "loss": 4.9709,
      "theoretical_loss": 5.830453251971296,
      "tokens_seen": 26673152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004047619047619048,
      "loss": 5.1225,
      "theoretical_loss": 5.8281030492385835,
      "tokens_seen": 26738688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004057539682539683,
      "loss": 4.9439,
      "theoretical_loss": 5.82576020814929,
      "tokens_seen": 26804224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040674603174603173,
      "loss": 5.154,
      "theoretical_loss": 5.823424687729469,
      "tokens_seen": 26869760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040773809523809523,
      "loss": 4.6484,
      "theoretical_loss": 5.821096447332456,
      "tokens_seen": 26935296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040873015873015874,
      "loss": 4.6262,
      "theoretical_loss": 5.818775446635469,
      "tokens_seen": 27000832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040972222222222224,
      "loss": 5.0236,
      "theoretical_loss": 5.816461645636238,
      "tokens_seen": 27066368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004107142857142857,
      "loss": 5.1234,
      "theoretical_loss": 5.814155004649702,
      "tokens_seen": 27131904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004117063492063492,
      "loss": 4.6509,
      "theoretical_loss": 5.811855484304724,
      "tokens_seen": 27197440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004126984126984127,
      "loss": 4.9814,
      "theoretical_loss": 5.809563045540864,
      "tokens_seen": 27262976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004136904761904762,
      "loss": 5.0187,
      "theoretical_loss": 5.8072776496051866,
      "tokens_seen": 27328512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004146825396825397,
      "loss": 4.9255,
      "theoretical_loss": 5.804999258049106,
      "tokens_seen": 27394048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004156746031746032,
      "loss": 5.0265,
      "theoretical_loss": 5.802727832725283,
      "tokens_seen": 27459584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004166666666666667,
      "loss": 4.8827,
      "theoretical_loss": 5.800463335784541,
      "tokens_seen": 27525120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004176587301587302,
      "loss": 5.0962,
      "theoretical_loss": 5.798205729672842,
      "tokens_seen": 27590656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004186507936507937,
      "loss": 5.0409,
      "theoretical_loss": 5.795954977128286,
      "tokens_seen": 27656192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041964285714285714,
      "loss": 4.8332,
      "theoretical_loss": 5.7937110411781525,
      "tokens_seen": 27721728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042063492063492065,
      "loss": 4.813,
      "theoretical_loss": 5.79147388513598,
      "tokens_seen": 27787264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 27295,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.60736083984375,
      "objective/train/theoretical_loss": 5.789243472598683,
      "objective/train/tokens_used": 48312800,
      "theoretical_loss": 5.789243472598683,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042162698412698415,
      "loss": 4.735,
      "theoretical_loss": 5.789243472598683,
      "tokens_seen": 27852800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042261904761904765,
      "loss": 5.0103,
      "theoretical_loss": 5.787019767443696,
      "tokens_seen": 27918336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004236111111111111,
      "loss": 4.9487,
      "theoretical_loss": 5.784802733826166,
      "tokens_seen": 27983872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004246031746031746,
      "loss": 5.0393,
      "theoretical_loss": 5.782592336176171,
      "tokens_seen": 28049408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004255952380952381,
      "loss": 4.8239,
      "theoretical_loss": 5.780388539195972,
      "tokens_seen": 28114944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004265873015873016,
      "loss": 4.8301,
      "theoretical_loss": 5.778191307857307,
      "tokens_seen": 28180480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042757936507936505,
      "loss": 4.8818,
      "theoretical_loss": 5.776000607398707,
      "tokens_seen": 28246016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042857142857142855,
      "loss": 5.0365,
      "theoretical_loss": 5.773816403322854,
      "tokens_seen": 28311552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042956349206349205,
      "loss": 4.8806,
      "theoretical_loss": 5.7716386613939665,
      "tokens_seen": 28377088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004305555555555556,
      "loss": 4.8212,
      "theoretical_loss": 5.769467347635221,
      "tokens_seen": 28442624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043154761904761905,
      "loss": 4.6856,
      "theoretical_loss": 5.767302428326196,
      "tokens_seen": 28508160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043253968253968256,
      "loss": 5.1099,
      "theoretical_loss": 5.765143870000358,
      "tokens_seen": 28573696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043353174603174606,
      "loss": 4.9942,
      "theoretical_loss": 5.762991639442574,
      "tokens_seen": 28639232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043452380952380956,
      "loss": 4.7269,
      "theoretical_loss": 5.7608457036866465,
      "tokens_seen": 28704768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000435515873015873,
      "loss": 4.7079,
      "theoretical_loss": 5.758706030012889,
      "tokens_seen": 28770304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004365079365079365,
      "loss": 5.0535,
      "theoretical_loss": 5.756572585945728,
      "tokens_seen": 28835840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004375,
      "loss": 4.9953,
      "theoretical_loss": 5.754445339251326,
      "tokens_seen": 28901376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004384920634920635,
      "loss": 4.8239,
      "theoretical_loss": 5.752324257935244,
      "tokens_seen": 28966912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000439484126984127,
      "loss": 4.8738,
      "theoretical_loss": 5.750209310240125,
      "tokens_seen": 29032448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044047619047619046,
      "loss": 4.5774,
      "theoretical_loss": 5.7481004646434055,
      "tokens_seen": 29097984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044146825396825396,
      "loss": 4.7821,
      "theoretical_loss": 5.745997689855058,
      "tokens_seen": 29163520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044246031746031746,
      "loss": 4.9197,
      "theoretical_loss": 5.743900954815356,
      "tokens_seen": 29229056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044345238095238096,
      "loss": 4.9178,
      "theoretical_loss": 5.741810228692663,
      "tokens_seen": 29294592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004444444444444444,
      "loss": 4.9811,
      "theoretical_loss": 5.739725480881262,
      "tokens_seen": 29360128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044543650793650797,
      "loss": 4.595,
      "theoretical_loss": 5.737646680999193,
      "tokens_seen": 29425664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 28553,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.193212032318115,
      "objective/train/theoretical_loss": 5.73557379888612,
      "objective/train/tokens_used": 49951200,
      "theoretical_loss": 5.73557379888612,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044642857142857147,
      "loss": 4.6289,
      "theoretical_loss": 5.73557379888612,
      "tokens_seen": 29491200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044742063492063497,
      "loss": 4.5175,
      "theoretical_loss": 5.733506804601236,
      "tokens_seen": 29556736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004484126984126984,
      "loss": 4.97,
      "theoretical_loss": 5.7314456684211725,
      "tokens_seen": 29622272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004494047619047619,
      "loss": 4.9007,
      "theoretical_loss": 5.729390360837952,
      "tokens_seen": 29687808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004503968253968254,
      "loss": 5.0718,
      "theoretical_loss": 5.7273408525569485,
      "tokens_seen": 29753344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004513888888888889,
      "loss": 4.8076,
      "theoretical_loss": 5.725297114494884,
      "tokens_seen": 29818880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045238095238095237,
      "loss": 4.9151,
      "theoretical_loss": 5.7232591177778405,
      "tokens_seen": 29884416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045337301587301587,
      "loss": 4.7611,
      "theoretical_loss": 5.7212268337393,
      "tokens_seen": 29949952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045436507936507937,
      "loss": 4.6533,
      "theoretical_loss": 5.719200233918203,
      "tokens_seen": 30015488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004553571428571429,
      "loss": 4.7689,
      "theoretical_loss": 5.717179290057032,
      "tokens_seen": 30081024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004563492063492063,
      "loss": 4.753,
      "theoretical_loss": 5.715163974099917,
      "tokens_seen": 30146560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004573412698412698,
      "loss": 4.9783,
      "theoretical_loss": 5.713154258190757,
      "tokens_seen": 30212096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004583333333333333,
      "loss": 4.8919,
      "theoretical_loss": 5.711150114671375,
      "tokens_seen": 30277632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004593253968253968,
      "loss": 4.6246,
      "theoretical_loss": 5.709151516079683,
      "tokens_seen": 30343168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046031746031746033,
      "loss": 4.8047,
      "theoretical_loss": 5.707158435147875,
      "tokens_seen": 30408704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046130952380952383,
      "loss": 4.7619,
      "theoretical_loss": 5.705170844800628,
      "tokens_seen": 30474240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046230158730158733,
      "loss": 5.0389,
      "theoretical_loss": 5.703188718153347,
      "tokens_seen": 30539776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046329365079365083,
      "loss": 4.7978,
      "theoretical_loss": 5.7012120285104055,
      "tokens_seen": 30605312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046428571428571433,
      "loss": 4.8182,
      "theoretical_loss": 5.699240749363417,
      "tokens_seen": 30670848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004652777777777778,
      "loss": 5.1046,
      "theoretical_loss": 5.69727485438953,
      "tokens_seen": 30736384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004662698412698413,
      "loss": 4.9301,
      "theoretical_loss": 5.695314317449732,
      "tokens_seen": 30801920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004672619047619048,
      "loss": 4.9513,
      "theoretical_loss": 5.69335911258718,
      "tokens_seen": 30867456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004682539682539683,
      "loss": 4.7506,
      "theoretical_loss": 5.691409214025544,
      "tokens_seen": 30932992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046924603174603173,
      "loss": 4.9203,
      "theoretical_loss": 5.689464596167383,
      "tokens_seen": 30998528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047023809523809523,
      "loss": 4.6588,
      "theoretical_loss": 5.687525233592513,
      "tokens_seen": 31064064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 29210,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.980744361877441,
      "objective/train/theoretical_loss": 5.685591101056422,
      "objective/train/tokens_used": 51589600,
      "theoretical_loss": 5.685591101056422,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047123015873015874,
      "loss": 4.8131,
      "theoretical_loss": 5.685591101056422,
      "tokens_seen": 31129600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047222222222222224,
      "loss": 4.9729,
      "theoretical_loss": 5.683662173488678,
      "tokens_seen": 31195136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004732142857142857,
      "loss": 4.9988,
      "theoretical_loss": 5.681738425991377,
      "tokens_seen": 31260672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004742063492063492,
      "loss": 4.9296,
      "theoretical_loss": 5.679819833837586,
      "tokens_seen": 31326208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004751984126984127,
      "loss": 4.8866,
      "theoretical_loss": 5.677906372469826,
      "tokens_seen": 31391744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004761904761904762,
      "loss": 4.8619,
      "theoretical_loss": 5.675998017498549,
      "tokens_seen": 31457280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004771825396825397,
      "loss": 4.4881,
      "theoretical_loss": 5.674094744700648,
      "tokens_seen": 31522816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004781746031746032,
      "loss": 4.7385,
      "theoretical_loss": 5.672196530017979,
      "tokens_seen": 31588352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004791666666666667,
      "loss": 4.9446,
      "theoretical_loss": 5.670303349555893,
      "tokens_seen": 31653888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004801587301587302,
      "loss": 4.5805,
      "theoretical_loss": 5.668415179581795,
      "tokens_seen": 31719424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004811507936507937,
      "loss": 4.943,
      "theoretical_loss": 5.666531996523711,
      "tokens_seen": 31784960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048214285714285715,
      "loss": 4.9994,
      "theoretical_loss": 5.664653776968867,
      "tokens_seen": 31850496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048313492063492065,
      "loss": 4.9432,
      "theoretical_loss": 5.6627804976622995,
      "tokens_seen": 31916032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048412698412698415,
      "loss": 4.4612,
      "theoretical_loss": 5.660912135505461,
      "tokens_seen": 31981568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048511904761904765,
      "loss": 5.0296,
      "theoretical_loss": 5.659048667554853,
      "tokens_seen": 32047104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004861111111111111,
      "loss": 4.7342,
      "theoretical_loss": 5.657190071020672,
      "tokens_seen": 32112640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004871031746031746,
      "loss": 5.0271,
      "theoretical_loss": 5.65533632326547,
      "tokens_seen": 32178176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004880952380952381,
      "loss": 4.5208,
      "theoretical_loss": 5.6534874018028205,
      "tokens_seen": 32243712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004890873015873016,
      "loss": 4.8883,
      "theoretical_loss": 5.651643284296018,
      "tokens_seen": 32309248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004900793650793651,
      "loss": 4.5918,
      "theoretical_loss": 5.6498039485567695,
      "tokens_seen": 32374784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004910714285714286,
      "loss": 4.8362,
      "theoretical_loss": 5.6479693725439155,
      "tokens_seen": 32440320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492063492063492,
      "loss": 4.4816,
      "theoretical_loss": 5.646139534362161,
      "tokens_seen": 32505856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004930555555555556,
      "loss": 4.7304,
      "theoretical_loss": 5.644314412260815,
      "tokens_seen": 32571392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004940476190476191,
      "loss": 4.7639,
      "theoretical_loss": 5.642493984632544,
      "tokens_seen": 32636928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004950396825396826,
      "loss": 4.486,
      "theoretical_loss": 5.640678230012151,
      "tokens_seen": 32702464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 30192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.205135822296143,
      "objective/train/theoretical_loss": 5.638867127075349,
      "objective/train/tokens_used": 53228000,
      "theoretical_loss": 5.638867127075349,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496031746031746,
      "loss": 4.6383,
      "theoretical_loss": 5.638867127075349,
      "tokens_seen": 32768000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004970238095238095,
      "loss": 4.7665,
      "theoretical_loss": 5.63706065463756,
      "tokens_seen": 32833536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498015873015873,
      "loss": 4.6651,
      "theoretical_loss": 5.635258791652723,
      "tokens_seen": 32899072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990079365079365,
      "loss": 4.6032,
      "theoretical_loss": 5.633461517212115,
      "tokens_seen": 32964608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 4.8611,
      "theoretical_loss": 5.631668810543181,
      "tokens_seen": 33030144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999899699097292,
      "loss": 4.4403,
      "theoretical_loss": 5.629880651008384,
      "tokens_seen": 33095680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999799398194584,
      "loss": 4.6342,
      "theoretical_loss": 5.628097018104057,
      "tokens_seen": 33161216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999699097291876,
      "loss": 4.789,
      "theoretical_loss": 5.626317891459275,
      "tokens_seen": 33226752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999598796389167,
      "loss": 4.4669,
      "theoretical_loss": 5.6245432508347335,
      "tokens_seen": 33292288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499949849548646,
      "loss": 4.5265,
      "theoretical_loss": 5.622773076121648,
      "tokens_seen": 33357824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999398194583751,
      "loss": 4.8668,
      "theoretical_loss": 5.621007347340646,
      "tokens_seen": 33423360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999297893681044,
      "loss": 4.663,
      "theoretical_loss": 5.619246044640693,
      "tokens_seen": 33488896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999197592778335,
      "loss": 4.51,
      "theoretical_loss": 5.617489148298015,
      "tokens_seen": 33554432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999097291875627,
      "loss": 4.4827,
      "theoretical_loss": 5.615736638715033,
      "tokens_seen": 33619968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998996990972919,
      "loss": 4.6552,
      "theoretical_loss": 5.613988496419315,
      "tokens_seen": 33685504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998896690070211,
      "loss": 4.5444,
      "theoretical_loss": 5.6122447020625374,
      "tokens_seen": 33751040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998796389167503,
      "loss": 4.7462,
      "theoretical_loss": 5.610505236419448,
      "tokens_seen": 33816576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998696088264795,
      "loss": 4.6343,
      "theoretical_loss": 5.608770080386854,
      "tokens_seen": 33882112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998595787362087,
      "loss": 4.3194,
      "theoretical_loss": 5.6070392149826045,
      "tokens_seen": 33947648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998495486459378,
      "loss": 4.7376,
      "theoretical_loss": 5.605312621344602,
      "tokens_seen": 34013184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499839518555667,
      "loss": 4.6319,
      "theoretical_loss": 5.603590280729803,
      "tokens_seen": 34078720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998294884653962,
      "loss": 4.3936,
      "theoretical_loss": 5.601872174513245,
      "tokens_seen": 34144256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998194583751254,
      "loss": 4.8045,
      "theoretical_loss": 5.60015828418708,
      "tokens_seen": 34209792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998094282848546,
      "loss": 4.8346,
      "theoretical_loss": 5.59844859135961,
      "tokens_seen": 34275328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997993981945837,
      "loss": 4.4444,
      "theoretical_loss": 5.596743077754337,
      "tokens_seen": 34340864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 30758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.619051933288574,
      "objective/train/theoretical_loss": 5.5950417252090325,
      "objective/train/tokens_used": 54866400,
      "theoretical_loss": 5.5950417252090325,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997893681043129,
      "loss": 4.5684,
      "theoretical_loss": 5.5950417252090325,
      "tokens_seen": 34406400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997793380140421,
      "loss": 4.7403,
      "theoretical_loss": 5.593344515674795,
      "tokens_seen": 34471936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997693079237714,
      "loss": 4.7682,
      "theoretical_loss": 5.591651431215139,
      "tokens_seen": 34537472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997592778335005,
      "loss": 4.9194,
      "theoretical_loss": 5.589962454005072,
      "tokens_seen": 34603008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997492477432298,
      "loss": 4.7926,
      "theoretical_loss": 5.588277566330207,
      "tokens_seen": 34668544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997392176529588,
      "loss": 4.5919,
      "theoretical_loss": 5.586596750585855,
      "tokens_seen": 34734080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997291875626881,
      "loss": 4.8471,
      "theoretical_loss": 5.584919989276146,
      "tokens_seen": 34799616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997191574724173,
      "loss": 4.8823,
      "theoretical_loss": 5.583247265013155,
      "tokens_seen": 34865152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997091273821465,
      "loss": 4.7952,
      "theoretical_loss": 5.581578560516027,
      "tokens_seen": 34930688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996990972918757,
      "loss": 4.8327,
      "theoretical_loss": 5.57991385861013,
      "tokens_seen": 34996224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996890672016048,
      "loss": 4.3745,
      "theoretical_loss": 5.578253142226191,
      "tokens_seen": 35061760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499679037111334,
      "loss": 4.6921,
      "theoretical_loss": 5.576596394399467,
      "tokens_seen": 35127296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996690070210632,
      "loss": 4.6698,
      "theoretical_loss": 5.574943598268902,
      "tokens_seen": 35192832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996589769307924,
      "loss": 4.9374,
      "theoretical_loss": 5.573294737076306,
      "tokens_seen": 35258368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996489468405216,
      "loss": 4.5814,
      "theoretical_loss": 5.571649794165537,
      "tokens_seen": 35323904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996389167502507,
      "loss": 4.6064,
      "theoretical_loss": 5.570008752981688,
      "tokens_seen": 35389440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996288866599799,
      "loss": 4.7137,
      "theoretical_loss": 5.5683715970702945,
      "tokens_seen": 35454976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996188565697091,
      "loss": 4.7919,
      "theoretical_loss": 5.566738310076531,
      "tokens_seen": 35520512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996088264794383,
      "loss": 4.5989,
      "theoretical_loss": 5.5651088757444285,
      "tokens_seen": 35586048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995987963891675,
      "loss": 4.4273,
      "theoretical_loss": 5.5634832779161,
      "tokens_seen": 35651584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995887662988968,
      "loss": 4.6147,
      "theoretical_loss": 5.561861500530965,
      "tokens_seen": 35717120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995787362086258,
      "loss": 4.5803,
      "theoretical_loss": 5.560243527624989,
      "tokens_seen": 35782656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995687061183551,
      "loss": 4.762,
      "theoretical_loss": 5.558629343329924,
      "tokens_seen": 35848192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995586760280842,
      "loss": 4.7117,
      "theoretical_loss": 5.557018931872566,
      "tokens_seen": 35913728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995486459378135,
      "loss": 4.7915,
      "theoretical_loss": 5.555412277574007,
      "tokens_seen": 35979264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 31707,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.547565460205078,
      "objective/train/theoretical_loss": 5.553809364848902,
      "objective/train/tokens_used": 56504800,
      "theoretical_loss": 5.553809364848902,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995386158475427,
      "loss": 4.6425,
      "theoretical_loss": 5.553809364848902,
      "tokens_seen": 36044800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995285857572718,
      "loss": 4.5963,
      "theoretical_loss": 5.552210178204751,
      "tokens_seen": 36110336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499518555667001,
      "loss": 4.6793,
      "theoretical_loss": 5.550614702241162,
      "tokens_seen": 36175872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995085255767302,
      "loss": 4.7561,
      "theoretical_loss": 5.549022921649151,
      "tokens_seen": 36241408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994984954864594,
      "loss": 4.6607,
      "theoretical_loss": 5.54743482121043,
      "tokens_seen": 36306944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994884653961886,
      "loss": 4.6252,
      "theoretical_loss": 5.545850385796712,
      "tokens_seen": 36372480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994784353059178,
      "loss": 4.4331,
      "theoretical_loss": 5.544269600369009,
      "tokens_seen": 36438016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994684052156469,
      "loss": 4.643,
      "theoretical_loss": 5.5426924499769505,
      "tokens_seen": 36503552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994583751253761,
      "loss": 4.7548,
      "theoretical_loss": 5.541118919758107,
      "tokens_seen": 36569088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994483450351053,
      "loss": 4.7337,
      "theoretical_loss": 5.539548994937308,
      "tokens_seen": 36634624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994383149448345,
      "loss": 4.3641,
      "theoretical_loss": 5.537982660825981,
      "tokens_seen": 36700160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994282848545637,
      "loss": 4.5599,
      "theoretical_loss": 5.536419902821491,
      "tokens_seen": 36765696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994182547642928,
      "loss": 4.6875,
      "theoretical_loss": 5.534860706406477,
      "tokens_seen": 36831232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994082246740221,
      "loss": 4.7352,
      "theoretical_loss": 5.533305057148215,
      "tokens_seen": 36896768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993981945837512,
      "loss": 4.6149,
      "theoretical_loss": 5.531752940697968,
      "tokens_seen": 36962304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993881644934805,
      "loss": 4.5067,
      "theoretical_loss": 5.5302043427903484,
      "tokens_seen": 37027840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993781344032096,
      "loss": 4.9474,
      "theoretical_loss": 5.528659249242693,
      "tokens_seen": 37093376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993681043129389,
      "loss": 4.4017,
      "theoretical_loss": 5.527117645954435,
      "tokens_seen": 37158912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499358074222668,
      "loss": 4.5323,
      "theoretical_loss": 5.525579518906484,
      "tokens_seen": 37224448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993480441323972,
      "loss": 4.4802,
      "theoretical_loss": 5.524044854160613,
      "tokens_seen": 37289984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993380140421264,
      "loss": 4.4531,
      "theoretical_loss": 5.522513637858855,
      "tokens_seen": 37355520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993279839518556,
      "loss": 4.6216,
      "theoretical_loss": 5.520985856222893,
      "tokens_seen": 37421056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993179538615848,
      "loss": 4.5407,
      "theoretical_loss": 5.519461495553473,
      "tokens_seen": 37486592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993079237713139,
      "loss": 4.4097,
      "theoretical_loss": 5.5179405422298045,
      "tokens_seen": 37552128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992978936810431,
      "loss": 4.406,
      "theoretical_loss": 5.516422982708983,
      "tokens_seen": 37617664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 32203,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.049692630767822,
      "objective/train/theoretical_loss": 5.514908803525402,
      "objective/train/tokens_used": 58143200,
      "theoretical_loss": 5.514908803525402,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992878635907723,
      "loss": 4.3638,
      "theoretical_loss": 5.514908803525402,
      "tokens_seen": 37683200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992778335005015,
      "loss": 4.6329,
      "theoretical_loss": 5.513397991290185,
      "tokens_seen": 37748736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992678034102307,
      "loss": 4.6538,
      "theoretical_loss": 5.511890532690611,
      "tokens_seen": 37814272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992577733199598,
      "loss": 4.7513,
      "theoretical_loss": 5.510386414489556,
      "tokens_seen": 37879808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499247743229689,
      "loss": 4.6591,
      "theoretical_loss": 5.5088856235249235,
      "tokens_seen": 37945344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992377131394183,
      "loss": 4.4618,
      "theoretical_loss": 5.507388146709105,
      "tokens_seen": 38010880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992276830491475,
      "loss": 4.4231,
      "theoretical_loss": 5.505893971028418,
      "tokens_seen": 38076416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992176529588767,
      "loss": 4.2713,
      "theoretical_loss": 5.504403083542568,
      "tokens_seen": 38141952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992076228686059,
      "loss": 4.4092,
      "theoretical_loss": 5.502915471384111,
      "tokens_seen": 38207488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499197592778335,
      "loss": 4.7235,
      "theoretical_loss": 5.501431121757914,
      "tokens_seen": 38273024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991875626880642,
      "loss": 4.3762,
      "theoretical_loss": 5.49995002194063,
      "tokens_seen": 38338560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991775325977934,
      "loss": 4.6923,
      "theoretical_loss": 5.498472159280171,
      "tokens_seen": 38404096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991675025075226,
      "loss": 4.8392,
      "theoretical_loss": 5.4969975211951905,
      "tokens_seen": 38469632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991574724172518,
      "loss": 4.5345,
      "theoretical_loss": 5.495526095174563,
      "tokens_seen": 38535168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499147442326981,
      "loss": 4.6407,
      "theoretical_loss": 5.494057868776883,
      "tokens_seen": 38600704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991374122367101,
      "loss": 4.5119,
      "theoretical_loss": 5.492592829629946,
      "tokens_seen": 38666240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991273821464393,
      "loss": 4.6456,
      "theoretical_loss": 5.491130965430262,
      "tokens_seen": 38731776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991173520561685,
      "loss": 4.2981,
      "theoretical_loss": 5.489672263942543,
      "tokens_seen": 38797312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991073219658977,
      "loss": 4.7488,
      "theoretical_loss": 5.488216712999222,
      "tokens_seen": 38862848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499097291875627,
      "loss": 4.5654,
      "theoretical_loss": 5.486764300499962,
      "tokens_seen": 38928384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499087261785356,
      "loss": 4.3462,
      "theoretical_loss": 5.4853150144111655,
      "tokens_seen": 38993920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990772316950853,
      "loss": 4.6389,
      "theoretical_loss": 5.4838688427655065,
      "tokens_seen": 39059456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990672016048144,
      "loss": 4.5166,
      "theoretical_loss": 5.482425773661445,
      "tokens_seen": 39124992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990571715145437,
      "loss": 4.5858,
      "theoretical_loss": 5.480985795262759,
      "tokens_seen": 39190528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990471414242729,
      "loss": 4.6098,
      "theoretical_loss": 5.479548895798082,
      "tokens_seen": 39256064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 33134,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.761120319366455,
      "objective/train/theoretical_loss": 5.478115063560433,
      "objective/train/tokens_used": 59781600,
      "theoretical_loss": 5.478115063560433,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499037111334002,
      "loss": 4.4957,
      "theoretical_loss": 5.478115063560433,
      "tokens_seen": 39321600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990270812437312,
      "loss": 4.7046,
      "theoretical_loss": 5.476684286906764,
      "tokens_seen": 39387136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990170511534604,
      "loss": 4.2948,
      "theoretical_loss": 5.475256554257497,
      "tokens_seen": 39452672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990070210631896,
      "loss": 4.5969,
      "theoretical_loss": 5.473831854096085,
      "tokens_seen": 39518208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989969909729188,
      "loss": 4.5511,
      "theoretical_loss": 5.472410174968553,
      "tokens_seen": 39583744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498986960882648,
      "loss": 4.4106,
      "theoretical_loss": 5.470991505483065,
      "tokens_seen": 39649280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989769307923771,
      "loss": 4.623,
      "theoretical_loss": 5.469575834309477,
      "tokens_seen": 39714816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989669007021063,
      "loss": 4.5077,
      "theoretical_loss": 5.468163150178907,
      "tokens_seen": 39780352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989568706118355,
      "loss": 4.475,
      "theoretical_loss": 5.466753441883302,
      "tokens_seen": 39845888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989468405215647,
      "loss": 4.6451,
      "theoretical_loss": 5.465346698275008,
      "tokens_seen": 39911424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989368104312939,
      "loss": 4.3366,
      "theoretical_loss": 5.46394290826635,
      "tokens_seen": 39976960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498926780341023,
      "loss": 4.4816,
      "theoretical_loss": 5.462542060829209,
      "tokens_seen": 40042496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989167502507523,
      "loss": 4.5005,
      "theoretical_loss": 5.461144144994606,
      "tokens_seen": 40108032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989067201604814,
      "loss": 4.5075,
      "theoretical_loss": 5.459749149852287,
      "tokens_seen": 40173568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988966900702107,
      "loss": 4.555,
      "theoretical_loss": 5.4583570645503166,
      "tokens_seen": 40239104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988866599799398,
      "loss": 4.411,
      "theoretical_loss": 5.45696787829467,
      "tokens_seen": 40304640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988766298896691,
      "loss": 4.479,
      "theoretical_loss": 5.455581580348831,
      "tokens_seen": 40370176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988665997993982,
      "loss": 4.424,
      "theoretical_loss": 5.454198160033391,
      "tokens_seen": 40435712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988565697091274,
      "loss": 4.6216,
      "theoretical_loss": 5.452817606725657,
      "tokens_seen": 40501248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988465396188566,
      "loss": 4.5915,
      "theoretical_loss": 5.451439909859255,
      "tokens_seen": 40566784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988365095285858,
      "loss": 4.4225,
      "theoretical_loss": 5.450065058923747,
      "tokens_seen": 40632320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498826479438315,
      "loss": 4.4659,
      "theoretical_loss": 5.448693043464235,
      "tokens_seen": 40697856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988164493480441,
      "loss": 4.7231,
      "theoretical_loss": 5.447323853080991,
      "tokens_seen": 40763392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988064192577733,
      "loss": 4.561,
      "theoretical_loss": 5.445957477429067,
      "tokens_seen": 40828928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987963891675025,
      "loss": 4.4559,
      "theoretical_loss": 5.444593906217927,
      "tokens_seen": 40894464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 33820,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.709114074707031,
      "objective/train/theoretical_loss": 5.443233129211069,
      "objective/train/tokens_used": 61420000,
      "theoretical_loss": 5.443233129211069,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987863590772317,
      "loss": 4.5353,
      "theoretical_loss": 5.443233129211069,
      "tokens_seen": 40960000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987763289869609,
      "loss": 4.5888,
      "theoretical_loss": 5.441875136225658,
      "tokens_seen": 41025536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049876629889669,
      "loss": 4.6609,
      "theoretical_loss": 5.440519917132159,
      "tokens_seen": 41091072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987562688064192,
      "loss": 4.6782,
      "theoretical_loss": 5.439167461853971,
      "tokens_seen": 41156608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987462387161484,
      "loss": 4.3218,
      "theoretical_loss": 5.437817760367073,
      "tokens_seen": 41222144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987362086258777,
      "loss": 4.6102,
      "theoretical_loss": 5.436470802699661,
      "tokens_seen": 41287680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987261785356068,
      "loss": 4.4044,
      "theoretical_loss": 5.43512657893179,
      "tokens_seen": 41353216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987161484453361,
      "loss": 4.5902,
      "theoretical_loss": 5.433785079195035,
      "tokens_seen": 41418752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987061183550651,
      "loss": 4.4395,
      "theoretical_loss": 5.432446293672129,
      "tokens_seen": 41484288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986960882647944,
      "loss": 4.6268,
      "theoretical_loss": 5.43111021259663,
      "tokens_seen": 41549824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986860581745236,
      "loss": 4.7247,
      "theoretical_loss": 5.429776826252565,
      "tokens_seen": 41615360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986760280842528,
      "loss": 4.7844,
      "theoretical_loss": 5.428446124974103,
      "tokens_seen": 41680896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498665997993982,
      "loss": 4.748,
      "theoretical_loss": 5.427118099145207,
      "tokens_seen": 41746432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986559679037111,
      "loss": 4.4521,
      "theoretical_loss": 5.425792739199306,
      "tokens_seen": 41811968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986459378134403,
      "loss": 4.5162,
      "theoretical_loss": 5.424470035618963,
      "tokens_seen": 41877504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986359077231695,
      "loss": 4.5787,
      "theoretical_loss": 5.423149978935547,
      "tokens_seen": 41943040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986258776328987,
      "loss": 4.4447,
      "theoretical_loss": 5.421832559728902,
      "tokens_seen": 42008576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986158475426279,
      "loss": 4.5436,
      "theoretical_loss": 5.420517768627027,
      "tokens_seen": 42074112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498605817452357,
      "loss": 4.5508,
      "theoretical_loss": 5.419205596305758,
      "tokens_seen": 42139648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985957873620862,
      "loss": 4.3851,
      "theoretical_loss": 5.417896033488446,
      "tokens_seen": 42205184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985857572718154,
      "loss": 4.4232,
      "theoretical_loss": 5.416589070945641,
      "tokens_seen": 42270720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985757271815446,
      "loss": 4.4672,
      "theoretical_loss": 5.415284699494782,
      "tokens_seen": 42336256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985656970912738,
      "loss": 4.3494,
      "theoretical_loss": 5.413982909999887,
      "tokens_seen": 42401792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985556670010031,
      "loss": 4.246,
      "theoretical_loss": 5.412683693371241,
      "tokens_seen": 42467328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985456369107321,
      "loss": 4.5265,
      "theoretical_loss": 5.411387040565092,
      "tokens_seen": 42532864
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 34938,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.999885082244873,
      "objective/train/theoretical_loss": 5.410092942583357,
      "objective/train/tokens_used": 63058400,
      "theoretical_loss": 5.410092942583357,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985356068204614,
      "loss": 4.5919,
      "theoretical_loss": 5.410092942583357,
      "tokens_seen": 42598400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985255767301905,
      "loss": 4.603,
      "theoretical_loss": 5.408801390473304,
      "tokens_seen": 42663936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985155466399198,
      "loss": 4.546,
      "theoretical_loss": 5.407512375327272,
      "tokens_seen": 42729472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498505516549649,
      "loss": 4.5326,
      "theoretical_loss": 5.406225888282363,
      "tokens_seen": 42795008
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984954864593782,
      "loss": 4.3889,
      "theoretical_loss": 5.404941920520156,
      "tokens_seen": 42860544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984854563691073,
      "loss": 4.1785,
      "theoretical_loss": 5.403660463266413,
      "tokens_seen": 42926080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984754262788365,
      "loss": 4.5382,
      "theoretical_loss": 5.402381507790792,
      "tokens_seen": 42991616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984653961885657,
      "loss": 4.5693,
      "theoretical_loss": 5.401105045406557,
      "tokens_seen": 43057152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984553660982949,
      "loss": 4.4403,
      "theoretical_loss": 5.399831067470302,
      "tokens_seen": 43122688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984453360080241,
      "loss": 4.7214,
      "theoretical_loss": 5.398559565381662,
      "tokens_seen": 43188224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984353059177532,
      "loss": 4.4442,
      "theoretical_loss": 5.397290530583037,
      "tokens_seen": 43253760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984252758274825,
      "loss": 4.309,
      "theoretical_loss": 5.396023954559315,
      "tokens_seen": 43319296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984152457372116,
      "loss": 4.3474,
      "theoretical_loss": 5.394759828837596,
      "tokens_seen": 43384832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984052156469409,
      "loss": 4.5364,
      "theoretical_loss": 5.393498144986921,
      "tokens_seen": 43450368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049839518555667,
      "loss": 4.647,
      "theoretical_loss": 5.392238894618,
      "tokens_seen": 43515904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983851554663993,
      "loss": 4.4688,
      "theoretical_loss": 5.390982069382944,
      "tokens_seen": 43581440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983751253761284,
      "loss": 4.555,
      "theoretical_loss": 5.389727660974998,
      "tokens_seen": 43646976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983650952858576,
      "loss": 4.2673,
      "theoretical_loss": 5.38847566112828,
      "tokens_seen": 43712512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983550651955868,
      "loss": 4.223,
      "theoretical_loss": 5.387226061617516,
      "tokens_seen": 43778048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498345035105316,
      "loss": 4.5944,
      "theoretical_loss": 5.38597885425778,
      "tokens_seen": 43843584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983350050150452,
      "loss": 4.2395,
      "theoretical_loss": 5.384734030904238,
      "tokens_seen": 43909120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983249749247743,
      "loss": 4.2047,
      "theoretical_loss": 5.383491583451892,
      "tokens_seen": 43974656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983149448345035,
      "loss": 4.6784,
      "theoretical_loss": 5.382251503835324,
      "tokens_seen": 44040192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983049147442327,
      "loss": 4.5094,
      "theoretical_loss": 5.381013784028445,
      "tokens_seen": 44105728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982948846539619,
      "loss": 4.2719,
      "theoretical_loss": 5.379778416044247,
      "tokens_seen": 44171264
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 35534,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.2067718505859375,
      "objective/train/theoretical_loss": 5.378545391934553,
      "objective/train/tokens_used": 64696800,
      "theoretical_loss": 5.378545391934553,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982848545636911,
      "loss": 4.3011,
      "theoretical_loss": 5.378545391934553,
      "tokens_seen": 44236800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982748244734202,
      "loss": 4.3593,
      "theoretical_loss": 5.377314703789773,
      "tokens_seen": 44302336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982647943831494,
      "loss": 4.5546,
      "theoretical_loss": 5.376086343738656,
      "tokens_seen": 44367872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982547642928786,
      "loss": 4.2113,
      "theoretical_loss": 5.3748603039480525,
      "tokens_seen": 44433408
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982447342026079,
      "loss": 4.6548,
      "theoretical_loss": 5.373636576622671,
      "tokens_seen": 44498944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498234704112337,
      "loss": 4.6067,
      "theoretical_loss": 5.372415154004842,
      "tokens_seen": 44564480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982246740220663,
      "loss": 4.5434,
      "theoretical_loss": 5.371196028374278,
      "tokens_seen": 44630016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982146439317953,
      "loss": 4.219,
      "theoretical_loss": 5.369979192047846,
      "tokens_seen": 44695552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982046138415246,
      "loss": 4.4616,
      "theoretical_loss": 5.368764637379322,
      "tokens_seen": 44761088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981945837512538,
      "loss": 4.3927,
      "theoretical_loss": 5.367552356759175,
      "tokens_seen": 44826624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498184553660983,
      "loss": 4.3436,
      "theoretical_loss": 5.366342342614322,
      "tokens_seen": 44892160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981745235707122,
      "loss": 4.3641,
      "theoretical_loss": 5.365134587407918,
      "tokens_seen": 44957696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981644934804413,
      "loss": 4.3878,
      "theoretical_loss": 5.363929083639112,
      "tokens_seen": 45023232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981544633901705,
      "loss": 4.252,
      "theoretical_loss": 5.362725823842838,
      "tokens_seen": 45088768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981444332998997,
      "loss": 4.1657,
      "theoretical_loss": 5.361524800589581,
      "tokens_seen": 45154304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981344032096289,
      "loss": 4.0417,
      "theoretical_loss": 5.360326006485165,
      "tokens_seen": 45219840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981243731193581,
      "loss": 4.6023,
      "theoretical_loss": 5.359129434170529,
      "tokens_seen": 45285376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981143430290873,
      "loss": 4.3602,
      "theoretical_loss": 5.357935076321514,
      "tokens_seen": 45350912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981043129388164,
      "loss": 4.4446,
      "theoretical_loss": 5.35674292564864,
      "tokens_seen": 45416448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980942828485456,
      "loss": 4.4122,
      "theoretical_loss": 5.3555529748969,
      "tokens_seen": 45481984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980842527582748,
      "loss": 4.465,
      "theoretical_loss": 5.354365216845546,
      "tokens_seen": 45547520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498074222668004,
      "loss": 4.1501,
      "theoretical_loss": 5.353179644307868,
      "tokens_seen": 45613056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980641925777333,
      "loss": 4.6183,
      "theoretical_loss": 5.351996250131005,
      "tokens_seen": 45678592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980541624874623,
      "loss": 4.4826,
      "theoretical_loss": 5.350815027195717,
      "tokens_seen": 45744128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980441323971916,
      "loss": 4.0969,
      "theoretical_loss": 5.349635968416189,
      "tokens_seen": 45809664
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 36791,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.983624219894409,
      "objective/train/theoretical_loss": 5.34845906673983,
      "objective/train/tokens_used": 66335200,
      "theoretical_loss": 5.34845906673983,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980341023069207,
      "loss": 4.1114,
      "theoretical_loss": 5.34845906673983,
      "tokens_seen": 45875200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049802407221665,
      "loss": 4.3748,
      "theoretical_loss": 5.347284315147059,
      "tokens_seen": 45940736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980140421263792,
      "loss": 4.3086,
      "theoretical_loss": 5.346111706651114,
      "tokens_seen": 46006272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980040120361084,
      "loss": 4.5521,
      "theoretical_loss": 5.344941234297847,
      "tokens_seen": 46071808
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979939819458375,
      "loss": 4.4492,
      "theoretical_loss": 5.343772891165527,
      "tokens_seen": 46137344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979839518555667,
      "loss": 4.4528,
      "theoretical_loss": 5.34260667036464,
      "tokens_seen": 46202880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979739217652959,
      "loss": 4.6275,
      "theoretical_loss": 5.3414425650377,
      "tokens_seen": 46268416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979638916750251,
      "loss": 4.5955,
      "theoretical_loss": 5.3402805683590495,
      "tokens_seen": 46333952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979538615847543,
      "loss": 4.5274,
      "theoretical_loss": 5.339120673534668,
      "tokens_seen": 46399488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979438314944834,
      "loss": 4.1984,
      "theoretical_loss": 5.337962873801984,
      "tokens_seen": 46465024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979338014042126,
      "loss": 4.4472,
      "theoretical_loss": 5.336807162429681,
      "tokens_seen": 46530560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979237713139418,
      "loss": 4.2688,
      "theoretical_loss": 5.335653532717512,
      "tokens_seen": 46596096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497913741223671,
      "loss": 4.5101,
      "theoretical_loss": 5.33450197799611,
      "tokens_seen": 46661632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979037111334002,
      "loss": 4.4218,
      "theoretical_loss": 5.33335249162681,
      "tokens_seen": 46727168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978936810431293,
      "loss": 4.365,
      "theoretical_loss": 5.332205067001454,
      "tokens_seen": 46792704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978836509528586,
      "loss": 4.4012,
      "theoretical_loss": 5.331059697542216,
      "tokens_seen": 46858240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978736208625877,
      "loss": 4.178,
      "theoretical_loss": 5.329916376701417,
      "tokens_seen": 46923776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497863590772317,
      "loss": 4.404,
      "theoretical_loss": 5.328775097961346,
      "tokens_seen": 46989312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978535606820461,
      "loss": 4.3063,
      "theoretical_loss": 5.327635854834084,
      "tokens_seen": 47054848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978435305917754,
      "loss": 4.3593,
      "theoretical_loss": 5.326498640861321,
      "tokens_seen": 47120384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978335005015045,
      "loss": 4.458,
      "theoretical_loss": 5.325363449614185,
      "tokens_seen": 47185920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978234704112337,
      "loss": 4.4505,
      "theoretical_loss": 5.324230274693065,
      "tokens_seen": 47251456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978134403209629,
      "loss": 4.2784,
      "theoretical_loss": 5.323099109727434,
      "tokens_seen": 47316992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978034102306921,
      "loss": 4.3645,
      "theoretical_loss": 5.321969948375687,
      "tokens_seen": 47382528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977933801404213,
      "loss": 4.3349,
      "theoretical_loss": 5.320842784324954,
      "tokens_seen": 47448064
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 38014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.285874366760254,
      "objective/train/theoretical_loss": 5.319717611290951,
      "objective/train/tokens_used": 67973600,
      "theoretical_loss": 5.319717611290951,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977833500501504,
      "loss": 4.2487,
      "theoretical_loss": 5.319717611290951,
      "tokens_seen": 47513600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977733199598796,
      "loss": 4.227,
      "theoretical_loss": 5.318594423017791,
      "tokens_seen": 47579136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977632898696088,
      "loss": 4.5068,
      "theoretical_loss": 5.317473213277827,
      "tokens_seen": 47644672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497753259779338,
      "loss": 4.4264,
      "theoretical_loss": 5.316353975871486,
      "tokens_seen": 47710208
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977432296890672,
      "loss": 4.4848,
      "theoretical_loss": 5.315236704627106,
      "tokens_seen": 47775744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977331995987965,
      "loss": 4.4288,
      "theoretical_loss": 5.314121393400763,
      "tokens_seen": 47841280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977231695085255,
      "loss": 4.0077,
      "theoretical_loss": 5.313008036076115,
      "tokens_seen": 47906816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977131394182548,
      "loss": 4.3322,
      "theoretical_loss": 5.311896626564246,
      "tokens_seen": 47972352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497703109327984,
      "loss": 4.1105,
      "theoretical_loss": 5.3107871588034925,
      "tokens_seen": 48037888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976930792377132,
      "loss": 4.3437,
      "theoretical_loss": 5.309679626759296,
      "tokens_seen": 48103424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976830491474424,
      "loss": 4.292,
      "theoretical_loss": 5.308574024424041,
      "tokens_seen": 48168960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976730190571715,
      "loss": 4.3868,
      "theoretical_loss": 5.3074703458168955,
      "tokens_seen": 48234496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976629889669007,
      "loss": 4.436,
      "theoretical_loss": 5.3063685849836615,
      "tokens_seen": 48300032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976529588766299,
      "loss": 4.368,
      "theoretical_loss": 5.305268735996615,
      "tokens_seen": 48365568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976429287863591,
      "loss": 4.2524,
      "theoretical_loss": 5.304170792954354,
      "tokens_seen": 48431104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976328986960883,
      "loss": 4.5041,
      "theoretical_loss": 5.303074749981648,
      "tokens_seen": 48496640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976228686058175,
      "loss": 4.412,
      "theoretical_loss": 5.301980601229285,
      "tokens_seen": 48562176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976128385155466,
      "loss": 4.3484,
      "theoretical_loss": 5.300888340873924,
      "tokens_seen": 48627712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976028084252758,
      "loss": 4.4722,
      "theoretical_loss": 5.299797963117941,
      "tokens_seen": 48693248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497592778335005,
      "loss": 4.1945,
      "theoretical_loss": 5.298709462189283,
      "tokens_seen": 48758784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975827482447342,
      "loss": 4.2003,
      "theoretical_loss": 5.297622832341327,
      "tokens_seen": 48824320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975727181544635,
      "loss": 4.0957,
      "theoretical_loss": 5.2965380678527225,
      "tokens_seen": 48889856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975626880641925,
      "loss": 4.2418,
      "theoretical_loss": 5.295455163027258,
      "tokens_seen": 48955392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975526579739218,
      "loss": 4.3033,
      "theoretical_loss": 5.29437411219371,
      "tokens_seen": 49020928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975426278836509,
      "loss": 4.2827,
      "theoretical_loss": 5.293294909705702,
      "tokens_seen": 49086464
    },
    {
      "epoch": 0.01,
      "objective/train/docs_used": 38689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.5455546379089355,
      "objective/train/theoretical_loss": 5.292217549941562,
      "objective/train/tokens_used": 69612000,
      "theoretical_loss": 5.292217549941562,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975325977933802,
      "loss": 4.2752,
      "theoretical_loss": 5.292217549941562,
      "tokens_seen": 49152000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975225677031094,
      "loss": 4.1889,
      "theoretical_loss": 5.291142027304185,
      "tokens_seen": 49217536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975125376128386,
      "loss": 3.9788,
      "theoretical_loss": 5.29006833622089,
      "tokens_seen": 49283072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975025075225677,
      "loss": 4.4666,
      "theoretical_loss": 5.28899647114328,
      "tokens_seen": 49348608
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974924774322969,
      "loss": 4.3282,
      "theoretical_loss": 5.287926426547106,
      "tokens_seen": 49414144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974824473420261,
      "loss": 4.3775,
      "theoretical_loss": 5.2868581969321315,
      "tokens_seen": 49479680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974724172517553,
      "loss": 4.3022,
      "theoretical_loss": 5.285791776821995,
      "tokens_seen": 49545216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974623871614845,
      "loss": 4.559,
      "theoretical_loss": 5.2847271607640725,
      "tokens_seen": 49610752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974523570712136,
      "loss": 4.5333,
      "theoretical_loss": 5.283664343329346,
      "tokens_seen": 49676288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974423269809428,
      "loss": 4.2267,
      "theoretical_loss": 5.282603319112273,
      "tokens_seen": 49741824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497432296890672,
      "loss": 4.3567,
      "theoretical_loss": 5.281544082730649,
      "tokens_seen": 49807360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974222668004012,
      "loss": 4.2938,
      "theoretical_loss": 5.280486628825479,
      "tokens_seen": 49872896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974122367101304,
      "loss": 4.3754,
      "theoretical_loss": 5.279430952060843,
      "tokens_seen": 49938432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974022066198595,
      "loss": 4.3361,
      "theoretical_loss": 5.278377047123777,
      "tokens_seen": 50003968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973921765295888,
      "loss": 4.2186,
      "theoretical_loss": 5.277324908724132,
      "tokens_seen": 50069504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973821464393179,
      "loss": 4.2978,
      "theoretical_loss": 5.276274531594453,
      "tokens_seen": 50135040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973721163490472,
      "loss": 4.3052,
      "theoretical_loss": 5.27522591048985,
      "tokens_seen": 50200576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973620862587763,
      "loss": 4.352,
      "theoretical_loss": 5.274179040187876,
      "tokens_seen": 50266112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973520561685056,
      "loss": 4.0427,
      "theoretical_loss": 5.273133915488394,
      "tokens_seen": 50331648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973420260782347,
      "loss": 4.3078,
      "theoretical_loss": 5.272090531213462,
      "tokens_seen": 50397184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973319959879639,
      "loss": 4.0645,
      "theoretical_loss": 5.271048882207202,
      "tokens_seen": 50462720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973219658976931,
      "loss": 4.6409,
      "theoretical_loss": 5.27000896333568,
      "tokens_seen": 50528256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973119358074223,
      "loss": 4.0496,
      "theoretical_loss": 5.268970769486789,
      "tokens_seen": 50593792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973019057171515,
      "loss": 4.4574,
      "theoretical_loss": 5.267934295570118,
      "tokens_seen": 50659328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972918756268806,
      "loss": 4.2343,
      "theoretical_loss": 5.266899536516842,
      "tokens_seen": 50724864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 39885,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.198443412780762,
      "objective/train/theoretical_loss": 5.265866487279597,
      "objective/train/tokens_used": 71250400,
      "theoretical_loss": 5.265866487279597,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972818455366098,
      "loss": 4.2608,
      "theoretical_loss": 5.265866487279597,
      "tokens_seen": 50790400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497271815446339,
      "loss": 4.2763,
      "theoretical_loss": 5.264835142832364,
      "tokens_seen": 50855936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972617853560682,
      "loss": 4.1906,
      "theoretical_loss": 5.263805498170351,
      "tokens_seen": 50921472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972517552657974,
      "loss": 4.1761,
      "theoretical_loss": 5.26277754830987,
      "tokens_seen": 50987008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972417251755266,
      "loss": 4.2526,
      "theoretical_loss": 5.261751288288236,
      "tokens_seen": 51052544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972316950852557,
      "loss": 4.275,
      "theoretical_loss": 5.260726713163637,
      "tokens_seen": 51118080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972216649949849,
      "loss": 4.3822,
      "theoretical_loss": 5.259703818015023,
      "tokens_seen": 51183616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972116349047142,
      "loss": 4.235,
      "theoretical_loss": 5.258682597941998,
      "tokens_seen": 51249152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972016048144433,
      "loss": 4.1933,
      "theoretical_loss": 5.257663048064704,
      "tokens_seen": 51314688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971915747241726,
      "loss": 4.0819,
      "theoretical_loss": 5.256645163523704,
      "tokens_seen": 51380224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971815446339017,
      "loss": 4.2302,
      "theoretical_loss": 5.255628939479877,
      "tokens_seen": 51445760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971715145436309,
      "loss": 4.1741,
      "theoretical_loss": 5.254614371114306,
      "tokens_seen": 51511296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971614844533601,
      "loss": 4.0105,
      "theoretical_loss": 5.253601453628165,
      "tokens_seen": 51576832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971514543630893,
      "loss": 4.1921,
      "theoretical_loss": 5.2525901822426135,
      "tokens_seen": 51642368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971414242728185,
      "loss": 4.1033,
      "theoretical_loss": 5.251580552198687,
      "tokens_seen": 51707904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313941825477,
      "loss": 4.5223,
      "theoretical_loss": 5.250572558757186,
      "tokens_seen": 51773440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971213640922768,
      "loss": 4.2685,
      "theoretical_loss": 5.2495661971985745,
      "tokens_seen": 51838976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497111334002006,
      "loss": 4.4627,
      "theoretical_loss": 5.248561462822871,
      "tokens_seen": 51904512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971013039117352,
      "loss": 4.2511,
      "theoretical_loss": 5.247558350949539,
      "tokens_seen": 51970048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970912738214644,
      "loss": 4.0554,
      "theoretical_loss": 5.246556856917392,
      "tokens_seen": 52035584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970812437311936,
      "loss": 4.1646,
      "theoretical_loss": 5.24555697608448,
      "tokens_seen": 52101120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970712136409227,
      "loss": 4.2295,
      "theoretical_loss": 5.244558703827986,
      "tokens_seen": 52166656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970611835506519,
      "loss": 4.1183,
      "theoretical_loss": 5.243562035544132,
      "tokens_seen": 52232192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970511534603811,
      "loss": 4.4799,
      "theoretical_loss": 5.24256696664807,
      "tokens_seen": 52297728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970411233701103,
      "loss": 4.1907,
      "theoretical_loss": 5.24157349257378,
      "tokens_seen": 52363264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 40418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.979517936706543,
      "objective/train/theoretical_loss": 5.240581608773973,
      "objective/train/tokens_used": 72888800,
      "theoretical_loss": 5.240581608773973,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970310932798396,
      "loss": 4.0593,
      "theoretical_loss": 5.240581608773973,
      "tokens_seen": 52428800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970210631895686,
      "loss": 4.2029,
      "theoretical_loss": 5.239591310719987,
      "tokens_seen": 52494336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970110330992979,
      "loss": 4.0965,
      "theoretical_loss": 5.238602593901694,
      "tokens_seen": 52559872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970010030090271,
      "loss": 4.0066,
      "theoretical_loss": 5.23761545382739,
      "tokens_seen": 52625408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969909729187563,
      "loss": 4.1287,
      "theoretical_loss": 5.236629886023708,
      "tokens_seen": 52690944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969809428284855,
      "loss": 4.1441,
      "theoretical_loss": 5.235645886035517,
      "tokens_seen": 52756480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969709127382147,
      "loss": 4.2684,
      "theoretical_loss": 5.2346634494258195,
      "tokens_seen": 52822016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969608826479438,
      "loss": 4.1129,
      "theoretical_loss": 5.233682571775664,
      "tokens_seen": 52887552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496950852557673,
      "loss": 4.44,
      "theoretical_loss": 5.232703248684045,
      "tokens_seen": 52953088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969408224674022,
      "loss": 4.3805,
      "theoretical_loss": 5.231725475767806,
      "tokens_seen": 53018624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969307923771314,
      "loss": 4.012,
      "theoretical_loss": 5.2307492486615494,
      "tokens_seen": 53084160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969207622868606,
      "loss": 4.3914,
      "theoretical_loss": 5.229774563017541,
      "tokens_seen": 53149696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969107321965897,
      "loss": 4.2825,
      "theoretical_loss": 5.2288014145056145,
      "tokens_seen": 53215232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496900702106319,
      "loss": 3.9345,
      "theoretical_loss": 5.227829798813083,
      "tokens_seen": 53280768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968906720160481,
      "loss": 4.1139,
      "theoretical_loss": 5.2268597116446465,
      "tokens_seen": 53346304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968806419257774,
      "loss": 4.2213,
      "theoretical_loss": 5.225891148722296,
      "tokens_seen": 53411840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968706118355065,
      "loss": 4.1851,
      "theoretical_loss": 5.224924105785227,
      "tokens_seen": 53477376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968605817452358,
      "loss": 4.1925,
      "theoretical_loss": 5.22395857858975,
      "tokens_seen": 53542912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968505516549649,
      "loss": 3.8839,
      "theoretical_loss": 5.222994562909199,
      "tokens_seen": 53608448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968405215646941,
      "loss": 4.1607,
      "theoretical_loss": 5.222032054533841,
      "tokens_seen": 53673984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968304914744233,
      "loss": 4.2475,
      "theoretical_loss": 5.2210710492707895,
      "tokens_seen": 53739520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968204613841525,
      "loss": 3.881,
      "theoretical_loss": 5.220111542943918,
      "tokens_seen": 53805056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968104312938817,
      "loss": 4.0171,
      "theoretical_loss": 5.21915353139377,
      "tokens_seen": 53870592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968004012036108,
      "loss": 4.0926,
      "theoretical_loss": 5.218197010477477,
      "tokens_seen": 53936128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049679037111334,
      "loss": 4.3292,
      "theoretical_loss": 5.21724197606866,
      "tokens_seen": 54001664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 41741,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.073785305023193,
      "objective/train/theoretical_loss": 5.216288424057362,
      "objective/train/tokens_used": 74527200,
      "theoretical_loss": 5.216288424057362,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967803410230692,
      "loss": 4.2491,
      "theoretical_loss": 5.216288424057362,
      "tokens_seen": 54067200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967703109327984,
      "loss": 4.2446,
      "theoretical_loss": 5.21533635034995,
      "tokens_seen": 54132736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967602808425276,
      "loss": 4.2944,
      "theoretical_loss": 5.214385750869031,
      "tokens_seen": 54198272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967502507522568,
      "loss": 4.1296,
      "theoretical_loss": 5.213436621553377,
      "tokens_seen": 54263808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967402206619859,
      "loss": 4.17,
      "theoretical_loss": 5.212488958357831,
      "tokens_seen": 54329344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967301905717151,
      "loss": 4.0717,
      "theoretical_loss": 5.211542757253229,
      "tokens_seen": 54394880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967201604814444,
      "loss": 4.4436,
      "theoretical_loss": 5.210598014226319,
      "tokens_seen": 54460416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967101303911735,
      "loss": 4.1394,
      "theoretical_loss": 5.209654725279678,
      "tokens_seen": 54525952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967001003009028,
      "loss": 4.4479,
      "theoretical_loss": 5.208712886431626,
      "tokens_seen": 54591488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496690070210632,
      "loss": 3.9774,
      "theoretical_loss": 5.207772493716154,
      "tokens_seen": 54657024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966800401203611,
      "loss": 4.0348,
      "theoretical_loss": 5.206833543182833,
      "tokens_seen": 54722560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966700100300903,
      "loss": 4.1275,
      "theoretical_loss": 5.205896030896748,
      "tokens_seen": 54788096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966599799398195,
      "loss": 4.034,
      "theoretical_loss": 5.2049599529384025,
      "tokens_seen": 54853632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966499498495487,
      "loss": 4.2901,
      "theoretical_loss": 5.204025305403653,
      "tokens_seen": 54919168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966399197592779,
      "loss": 4.0859,
      "theoretical_loss": 5.2030920844036235,
      "tokens_seen": 54984704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496629889669007,
      "loss": 4.2186,
      "theoretical_loss": 5.202160286064628,
      "tokens_seen": 55050240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966198595787362,
      "loss": 4.1753,
      "theoretical_loss": 5.2012299065281,
      "tokens_seen": 55115776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966098294884654,
      "loss": 4.2037,
      "theoretical_loss": 5.2003009419505055,
      "tokens_seen": 55181312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965997993981946,
      "loss": 4.3058,
      "theoretical_loss": 5.199373388503272,
      "tokens_seen": 55246848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965897693079238,
      "loss": 4.3641,
      "theoretical_loss": 5.198447242372716,
      "tokens_seen": 55312384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965797392176529,
      "loss": 4.1662,
      "theoretical_loss": 5.197522499759961,
      "tokens_seen": 55377920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965697091273821,
      "loss": 4.1908,
      "theoretical_loss": 5.196599156880863,
      "tokens_seen": 55443456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965596790371113,
      "loss": 3.9863,
      "theoretical_loss": 5.195677209965942,
      "tokens_seen": 55508992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965496489468405,
      "loss": 4.5222,
      "theoretical_loss": 5.194756655260305,
      "tokens_seen": 55574528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965396188565698,
      "loss": 3.9698,
      "theoretical_loss": 5.193837489023565,
      "tokens_seen": 55640064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 42409,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.305491924285889,
      "objective/train/theoretical_loss": 5.192919707529784,
      "objective/train/tokens_used": 76165600,
      "theoretical_loss": 5.192919707529784,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965295887662988,
      "loss": 4.121,
      "theoretical_loss": 5.192919707529784,
      "tokens_seen": 55705600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965195586760281,
      "loss": 4.108,
      "theoretical_loss": 5.1920033070673846,
      "tokens_seen": 55771136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965095285857573,
      "loss": 3.9647,
      "theoretical_loss": 5.191088283939086,
      "tokens_seen": 55836672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964994984954865,
      "loss": 3.9453,
      "theoretical_loss": 5.190174634461828,
      "tokens_seen": 55902208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964894684052157,
      "loss": 3.9193,
      "theoretical_loss": 5.189262354966708,
      "tokens_seen": 55967744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964794383149449,
      "loss": 4.0925,
      "theoretical_loss": 5.188351441798897,
      "tokens_seen": 56033280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496469408224674,
      "loss": 4.1099,
      "theoretical_loss": 5.187441891317581,
      "tokens_seen": 56098816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964593781344032,
      "loss": 4.0055,
      "theoretical_loss": 5.186533699895888,
      "tokens_seen": 56164352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964493480441324,
      "loss": 4.1547,
      "theoretical_loss": 5.18562686392081,
      "tokens_seen": 56229888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964393179538616,
      "loss": 4.1385,
      "theoretical_loss": 5.184721379793145,
      "tokens_seen": 56295424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964292878635908,
      "loss": 4.1137,
      "theoretical_loss": 5.183817243927425,
      "tokens_seen": 56360960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049641925777332,
      "loss": 3.988,
      "theoretical_loss": 5.182914452751846,
      "tokens_seen": 56426496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964092276830491,
      "loss": 3.9087,
      "theoretical_loss": 5.1820130027081985,
      "tokens_seen": 56492032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963991975927783,
      "loss": 4.1078,
      "theoretical_loss": 5.181112890251804,
      "tokens_seen": 56557568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963891675025075,
      "loss": 3.9657,
      "theoretical_loss": 5.180214111851447,
      "tokens_seen": 56623104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963791374122367,
      "loss": 3.8969,
      "theoretical_loss": 5.179316663989305,
      "tokens_seen": 56688640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963691073219659,
      "loss": 4.0938,
      "theoretical_loss": 5.178420543160889,
      "tokens_seen": 56754176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963590772316951,
      "loss": 4.0717,
      "theoretical_loss": 5.177525745874968,
      "tokens_seen": 56819712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963490471414242,
      "loss": 4.0293,
      "theoretical_loss": 5.176632268653516,
      "tokens_seen": 56885248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963390170511535,
      "loss": 4.0914,
      "theoretical_loss": 5.175740108031633,
      "tokens_seen": 56950784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963289869608827,
      "loss": 4.3386,
      "theoretical_loss": 5.174849260557494,
      "tokens_seen": 57016320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963189568706119,
      "loss": 4.1438,
      "theoretical_loss": 5.173959722792272,
      "tokens_seen": 57081856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496308926780341,
      "loss": 4.1502,
      "theoretical_loss": 5.173071491310086,
      "tokens_seen": 57147392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962988966900702,
      "loss": 3.9246,
      "theoretical_loss": 5.172184562697928,
      "tokens_seen": 57212928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962888665997994,
      "loss": 4.0249,
      "theoretical_loss": 5.171298933555606,
      "tokens_seen": 57278464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 43695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.649152994155884,
      "objective/train/theoretical_loss": 5.170414600495677,
      "objective/train/tokens_used": 77804000,
      "theoretical_loss": 5.170414600495677,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962788365095286,
      "loss": 4.0093,
      "theoretical_loss": 5.170414600495677,
      "tokens_seen": 57344000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962688064192578,
      "loss": 4.1786,
      "theoretical_loss": 5.16953156014339,
      "tokens_seen": 57409536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496258776328987,
      "loss": 4.0607,
      "theoretical_loss": 5.1686498091366175,
      "tokens_seen": 57475072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962487462387161,
      "loss": 3.6266,
      "theoretical_loss": 5.1677693441258,
      "tokens_seen": 57540608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962387161484453,
      "loss": 3.8942,
      "theoretical_loss": 5.16689016177388,
      "tokens_seen": 57606144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962286860581746,
      "loss": 3.9167,
      "theoretical_loss": 5.166012258756243,
      "tokens_seen": 57671680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962186559679037,
      "loss": 4.2399,
      "theoretical_loss": 5.16513563176066,
      "tokens_seen": 57737216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496208625877633,
      "loss": 3.7377,
      "theoretical_loss": 5.164260277487221,
      "tokens_seen": 57802752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961985957873621,
      "loss": 4.0589,
      "theoretical_loss": 5.163386192648282,
      "tokens_seen": 57868288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961885656970913,
      "loss": 4.2436,
      "theoretical_loss": 5.162513373968403,
      "tokens_seen": 57933824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961785356068205,
      "loss": 4.2688,
      "theoretical_loss": 5.1616418181842825,
      "tokens_seen": 57999360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961685055165497,
      "loss": 3.9842,
      "theoretical_loss": 5.160771522044712,
      "tokens_seen": 58064896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961584754262789,
      "loss": 4.202,
      "theoretical_loss": 5.159902482310506,
      "tokens_seen": 58130432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961484453360081,
      "loss": 4.0252,
      "theoretical_loss": 5.15903469575445,
      "tokens_seen": 58195968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961384152457372,
      "loss": 4.0653,
      "theoretical_loss": 5.1581681591612405,
      "tokens_seen": 58261504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961283851554664,
      "loss": 4.1853,
      "theoretical_loss": 5.15730286932743,
      "tokens_seen": 58327040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961183550651956,
      "loss": 4.0287,
      "theoretical_loss": 5.156438823061366,
      "tokens_seen": 58392576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961083249749248,
      "loss": 3.8706,
      "theoretical_loss": 5.15557601718314,
      "tokens_seen": 58458112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496098294884654,
      "loss": 3.4818,
      "theoretical_loss": 5.154714448524525,
      "tokens_seen": 58523648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960882647943831,
      "loss": 4.1847,
      "theoretical_loss": 5.153854113928924,
      "tokens_seen": 58589184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960782347041123,
      "loss": 4.0123,
      "theoretical_loss": 5.152995010251313,
      "tokens_seen": 58654720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960682046138415,
      "loss": 3.8546,
      "theoretical_loss": 5.1521371343581865,
      "tokens_seen": 58720256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960581745235707,
      "loss": 3.9651,
      "theoretical_loss": 5.1512804831275,
      "tokens_seen": 58785792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960481444333,
      "loss": 3.9367,
      "theoretical_loss": 5.1504250534486165,
      "tokens_seen": 58851328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496038114343029,
      "loss": 4.3063,
      "theoretical_loss": 5.149570842222257,
      "tokens_seen": 58916864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 44395,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.063347339630127,
      "objective/train/theoretical_loss": 5.1487178463604355,
      "objective/train/tokens_used": 79442400,
      "theoretical_loss": 5.1487178463604355,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960280842527583,
      "loss": 3.7506,
      "theoretical_loss": 5.1487178463604355,
      "tokens_seen": 58982400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960180541624875,
      "loss": 3.9872,
      "theoretical_loss": 5.147866062786418,
      "tokens_seen": 59047936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960080240722167,
      "loss": 3.9911,
      "theoretical_loss": 5.147015488434661,
      "tokens_seen": 59113472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959979939819459,
      "loss": 3.8278,
      "theoretical_loss": 5.14616612025076,
      "tokens_seen": 59179008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959879638916751,
      "loss": 3.9528,
      "theoretical_loss": 5.145317955191397,
      "tokens_seen": 59244544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959779338014042,
      "loss": 3.7919,
      "theoretical_loss": 5.144470990224288,
      "tokens_seen": 59310080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959679037111334,
      "loss": 3.9634,
      "theoretical_loss": 5.143625222328134,
      "tokens_seen": 59375616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959578736208626,
      "loss": 4.096,
      "theoretical_loss": 5.142780648492563,
      "tokens_seen": 59441152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959478435305918,
      "loss": 3.9749,
      "theoretical_loss": 5.141937265718084,
      "tokens_seen": 59506688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495937813440321,
      "loss": 4.0214,
      "theoretical_loss": 5.141095071016032,
      "tokens_seen": 59572224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959277833500501,
      "loss": 3.9143,
      "theoretical_loss": 5.14025406140852,
      "tokens_seen": 59637760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959177532597793,
      "loss": 4.0642,
      "theoretical_loss": 5.139414233928385,
      "tokens_seen": 59703296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959077231695085,
      "loss": 3.9695,
      "theoretical_loss": 5.138575585619142,
      "tokens_seen": 59768832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958976930792377,
      "loss": 4.1946,
      "theoretical_loss": 5.13773811353493,
      "tokens_seen": 59834368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958876629889669,
      "loss": 4.0784,
      "theoretical_loss": 5.1369018147404635,
      "tokens_seen": 59899904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495877632898696,
      "loss": 4.0826,
      "theoretical_loss": 5.136066686310986,
      "tokens_seen": 59965440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958676028084253,
      "loss": 4.0869,
      "theoretical_loss": 5.135232725332212,
      "tokens_seen": 60030976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958575727181544,
      "loss": 3.9758,
      "theoretical_loss": 5.134399928900292,
      "tokens_seen": 60096512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958475426278837,
      "loss": 4.0519,
      "theoretical_loss": 5.13356829412175,
      "tokens_seen": 60162048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958375125376129,
      "loss": 4.1545,
      "theoretical_loss": 5.132737818113444,
      "tokens_seen": 60227584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958274824473421,
      "loss": 4.2355,
      "theoretical_loss": 5.131908498002513,
      "tokens_seen": 60293120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958174523570712,
      "loss": 3.8189,
      "theoretical_loss": 5.131080330926334,
      "tokens_seen": 60358656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958074222668004,
      "loss": 4.1869,
      "theoretical_loss": 5.130253314032469,
      "tokens_seen": 60424192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957973921765296,
      "loss": 3.728,
      "theoretical_loss": 5.1294274444786225,
      "tokens_seen": 60489728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957873620862588,
      "loss": 4.1881,
      "theoretical_loss": 5.128602719432592,
      "tokens_seen": 60555264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 45571,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.731624603271484,
      "objective/train/theoretical_loss": 5.12777913607222,
      "objective/train/tokens_used": 81080800,
      "theoretical_loss": 5.12777913607222,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495777331995988,
      "loss": 4.3579,
      "theoretical_loss": 5.12777913607222,
      "tokens_seen": 60620800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957673019057172,
      "loss": 3.8709,
      "theoretical_loss": 5.126956691585351,
      "tokens_seen": 60686336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957572718154463,
      "loss": 3.9854,
      "theoretical_loss": 5.1261353831697845,
      "tokens_seen": 60751872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957472417251755,
      "loss": 4.2274,
      "theoretical_loss": 5.125315208033226,
      "tokens_seen": 60817408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957372116349047,
      "loss": 3.8993,
      "theoretical_loss": 5.124496163393244,
      "tokens_seen": 60882944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957271815446339,
      "loss": 3.7806,
      "theoretical_loss": 5.123678246477226,
      "tokens_seen": 60948480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957171514543631,
      "loss": 3.977,
      "theoretical_loss": 5.122861454522329,
      "tokens_seen": 61014016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957071213640923,
      "loss": 3.9739,
      "theoretical_loss": 5.122045784775439,
      "tokens_seen": 61079552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956970912738214,
      "loss": 3.8256,
      "theoretical_loss": 5.1212312344931235,
      "tokens_seen": 61145088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956870611835507,
      "loss": 4.0096,
      "theoretical_loss": 5.120417800941587,
      "tokens_seen": 61210624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956770310932798,
      "loss": 4.0399,
      "theoretical_loss": 5.119605481396629,
      "tokens_seen": 61276160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956670010030091,
      "loss": 3.9718,
      "theoretical_loss": 5.1187942731436,
      "tokens_seen": 61341696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956569709127383,
      "loss": 3.9982,
      "theoretical_loss": 5.117984173477356,
      "tokens_seen": 61407232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956469408224674,
      "loss": 4.0777,
      "theoretical_loss": 5.117175179702215,
      "tokens_seen": 61472768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956369107321966,
      "loss": 3.9865,
      "theoretical_loss": 5.116367289131917,
      "tokens_seen": 61538304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956268806419258,
      "loss": 4.1915,
      "theoretical_loss": 5.115560499089575,
      "tokens_seen": 61603840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495616850551655,
      "loss": 3.8813,
      "theoretical_loss": 5.1147548069076425,
      "tokens_seen": 61669376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956068204613842,
      "loss": 3.97,
      "theoretical_loss": 5.113950209927859,
      "tokens_seen": 61734912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955967903711133,
      "loss": 3.5137,
      "theoretical_loss": 5.113146705501217,
      "tokens_seen": 61800448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955867602808425,
      "loss": 3.7782,
      "theoretical_loss": 5.112344290987915,
      "tokens_seen": 61865984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955767301905717,
      "loss": 4.1055,
      "theoretical_loss": 5.1115429637573175,
      "tokens_seen": 61931520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955667001003009,
      "loss": 3.9588,
      "theoretical_loss": 5.110742721187915,
      "tokens_seen": 61997056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955566700100301,
      "loss": 3.8558,
      "theoretical_loss": 5.109943560667279,
      "tokens_seen": 62062592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955466399197592,
      "loss": 3.8969,
      "theoretical_loss": 5.109145479592028,
      "tokens_seen": 62128128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955366098294884,
      "loss": 4.1078,
      "theoretical_loss": 5.108348475367774,
      "tokens_seen": 62193664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 46598,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.215669631958008,
      "objective/train/theoretical_loss": 5.107552545409097,
      "objective/train/tokens_used": 82719200,
      "theoretical_loss": 5.107552545409097,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955265797392177,
      "loss": 4.056,
      "theoretical_loss": 5.107552545409097,
      "tokens_seen": 62259200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955165496489468,
      "loss": 3.9227,
      "theoretical_loss": 5.106757687139494,
      "tokens_seen": 62324736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955065195586761,
      "loss": 4.2246,
      "theoretical_loss": 5.1059638979913435,
      "tokens_seen": 62390272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954964894684052,
      "loss": 4.0438,
      "theoretical_loss": 5.105171175405864,
      "tokens_seen": 62455808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954864593781344,
      "loss": 4.0275,
      "theoretical_loss": 5.104379516833075,
      "tokens_seen": 62521344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954764292878636,
      "loss": 4.005,
      "theoretical_loss": 5.103588919731757,
      "tokens_seen": 62586880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954663991975928,
      "loss": 3.937,
      "theoretical_loss": 5.1027993815694135,
      "tokens_seen": 62652416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495456369107322,
      "loss": 4.1536,
      "theoretical_loss": 5.10201089982223,
      "tokens_seen": 62717952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954463390170512,
      "loss": 3.971,
      "theoretical_loss": 5.101223471975038,
      "tokens_seen": 62783488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954363089267803,
      "loss": 3.9062,
      "theoretical_loss": 5.100437095521272,
      "tokens_seen": 62849024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954262788365095,
      "loss": 3.8866,
      "theoretical_loss": 5.099651767962936,
      "tokens_seen": 62914560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954162487462387,
      "loss": 3.8535,
      "theoretical_loss": 5.098867486810562,
      "tokens_seen": 62980096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954062186559679,
      "loss": 3.7859,
      "theoretical_loss": 5.098084249583175,
      "tokens_seen": 63045632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953961885656971,
      "loss": 3.9093,
      "theoretical_loss": 5.097302053808251,
      "tokens_seen": 63111168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953861584754263,
      "loss": 3.5672,
      "theoretical_loss": 5.0965208970216835,
      "tokens_seen": 63176704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953761283851555,
      "loss": 3.8846,
      "theoretical_loss": 5.095740776767743,
      "tokens_seen": 63242240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953660982948846,
      "loss": 3.9469,
      "theoretical_loss": 5.0949616905990425,
      "tokens_seen": 63307776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953560682046139,
      "loss": 4.2765,
      "theoretical_loss": 5.094183636076501,
      "tokens_seen": 63373312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953460381143431,
      "loss": 4.015,
      "theoretical_loss": 5.093406610769302,
      "tokens_seen": 63438848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953360080240723,
      "loss": 4.2246,
      "theoretical_loss": 5.092630612254862,
      "tokens_seen": 63504384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953259779338014,
      "loss": 3.9173,
      "theoretical_loss": 5.091855638118792,
      "tokens_seen": 63569920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953159478435306,
      "loss": 4.0249,
      "theoretical_loss": 5.091081685954863,
      "tokens_seen": 63635456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953059177532598,
      "loss": 3.7978,
      "theoretical_loss": 5.090308753364967,
      "tokens_seen": 63700992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495295887662989,
      "loss": 3.8605,
      "theoretical_loss": 5.089536837959084,
      "tokens_seen": 63766528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952858575727182,
      "loss": 3.9657,
      "theoretical_loss": 5.0887659373552445,
      "tokens_seen": 63832064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 47189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3733479976654053,
      "objective/train/theoretical_loss": 5.087996049179497,
      "objective/train/tokens_used": 84357600,
      "theoretical_loss": 5.087996049179497,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952758274824474,
      "loss": 3.6824,
      "theoretical_loss": 5.087996049179497,
      "tokens_seen": 63897600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952657973921765,
      "loss": 3.7677,
      "theoretical_loss": 5.087227171065869,
      "tokens_seen": 63963136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952557673019057,
      "loss": 3.9197,
      "theoretical_loss": 5.086459300656337,
      "tokens_seen": 64028672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952457372116349,
      "loss": 3.927,
      "theoretical_loss": 5.085692435600783,
      "tokens_seen": 64094208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952357071213641,
      "loss": 3.8821,
      "theoretical_loss": 5.084926573556972,
      "tokens_seen": 64159744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952256770310933,
      "loss": 3.7931,
      "theoretical_loss": 5.084161712190507,
      "tokens_seen": 64225280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952156469408225,
      "loss": 3.7647,
      "theoretical_loss": 5.083397849174801,
      "tokens_seen": 64290816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952056168505516,
      "loss": 3.9431,
      "theoretical_loss": 5.082634982191042,
      "tokens_seen": 64356352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951955867602809,
      "loss": 3.9173,
      "theoretical_loss": 5.081873108928154,
      "tokens_seen": 64421888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049518555667001,
      "loss": 4.0204,
      "theoretical_loss": 5.081112227082771,
      "tokens_seen": 64487424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951755265797393,
      "loss": 3.816,
      "theoretical_loss": 5.0803523343591985,
      "tokens_seen": 64552960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951654964894685,
      "loss": 3.8761,
      "theoretical_loss": 5.079593428469382,
      "tokens_seen": 64618496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951554663991976,
      "loss": 4.0756,
      "theoretical_loss": 5.078835507132873,
      "tokens_seen": 64684032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951454363089268,
      "loss": 3.8136,
      "theoretical_loss": 5.078078568076798,
      "tokens_seen": 64749568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495135406218656,
      "loss": 4.0581,
      "theoretical_loss": 5.077322609035821,
      "tokens_seen": 64815104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951253761283852,
      "loss": 3.8215,
      "theoretical_loss": 5.076567627752118,
      "tokens_seen": 64880640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951153460381144,
      "loss": 3.8355,
      "theoretical_loss": 5.075813621975337,
      "tokens_seen": 64946176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951053159478435,
      "loss": 3.9719,
      "theoretical_loss": 5.075060589462572,
      "tokens_seen": 65011712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950952858575727,
      "loss": 3.7383,
      "theoretical_loss": 5.074308527978326,
      "tokens_seen": 65077248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950852557673019,
      "loss": 3.8755,
      "theoretical_loss": 5.073557435294486,
      "tokens_seen": 65142784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950752256770311,
      "loss": 3.6252,
      "theoretical_loss": 5.07280730919028,
      "tokens_seen": 65208320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950651955867603,
      "loss": 3.7529,
      "theoretical_loss": 5.072058147452257,
      "tokens_seen": 65273856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950551654964894,
      "loss": 3.679,
      "theoretical_loss": 5.071309947874248,
      "tokens_seen": 65339392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950451354062186,
      "loss": 3.5894,
      "theoretical_loss": 5.070562708257339,
      "tokens_seen": 65404928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950351053159479,
      "loss": 3.9951,
      "theoretical_loss": 5.069816426409837,
      "tokens_seen": 65470464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 48433,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8971009254455566,
      "objective/train/theoretical_loss": 5.069071100147241,
      "objective/train/tokens_used": 85996000,
      "theoretical_loss": 5.069071100147241,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495025075225677,
      "loss": 3.9133,
      "theoretical_loss": 5.069071100147241,
      "tokens_seen": 65536000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950150451354063,
      "loss": 4.2034,
      "theoretical_loss": 5.068326727292211,
      "tokens_seen": 65601536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950050150451354,
      "loss": 3.7703,
      "theoretical_loss": 5.067583305674537,
      "tokens_seen": 65667072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949949849548646,
      "loss": 3.8678,
      "theoretical_loss": 5.066840833131106,
      "tokens_seen": 65732608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949849548645938,
      "loss": 3.8256,
      "theoretical_loss": 5.066099307505878,
      "tokens_seen": 65798144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494974924774323,
      "loss": 3.6242,
      "theoretical_loss": 5.06535872664985,
      "tokens_seen": 65863680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949648946840522,
      "loss": 3.9033,
      "theoretical_loss": 5.064619088421029,
      "tokens_seen": 65929216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949548645937814,
      "loss": 3.5217,
      "theoretical_loss": 5.0638803906844,
      "tokens_seen": 65994752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949448345035105,
      "loss": 3.8469,
      "theoretical_loss": 5.0631426313119,
      "tokens_seen": 66060288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949348044132397,
      "loss": 3.7472,
      "theoretical_loss": 5.062405808182383,
      "tokens_seen": 66125824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949247743229689,
      "loss": 3.4883,
      "theoretical_loss": 5.061669919181599,
      "tokens_seen": 66191360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949147442326981,
      "loss": 3.8441,
      "theoretical_loss": 5.060934962202155,
      "tokens_seen": 66256896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949047141424273,
      "loss": 3.9786,
      "theoretical_loss": 5.060200935143492,
      "tokens_seen": 66322432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948946840521565,
      "loss": 4.0608,
      "theoretical_loss": 5.059467835911857,
      "tokens_seen": 66387968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948846539618856,
      "loss": 3.7521,
      "theoretical_loss": 5.058735662420268,
      "tokens_seen": 66453504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948746238716148,
      "loss": 3.6167,
      "theoretical_loss": 5.0580044125884935,
      "tokens_seen": 66519040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494864593781344,
      "loss": 3.76,
      "theoretical_loss": 5.057274084343018,
      "tokens_seen": 66584576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948545636910733,
      "loss": 3.8083,
      "theoretical_loss": 5.056544675617017,
      "tokens_seen": 66650112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948445336008024,
      "loss": 3.7601,
      "theoretical_loss": 5.055816184350326,
      "tokens_seen": 66715648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948345035105316,
      "loss": 3.8069,
      "theoretical_loss": 5.055088608489417,
      "tokens_seen": 66781184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948244734202607,
      "loss": 3.3711,
      "theoretical_loss": 5.054361945987363,
      "tokens_seen": 66846720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049481444332999,
      "loss": 3.3429,
      "theoretical_loss": 5.05363619480382,
      "tokens_seen": 66912256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948044132397192,
      "loss": 4.0226,
      "theoretical_loss": 5.052911352904992,
      "tokens_seen": 66977792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947943831494484,
      "loss": 3.7436,
      "theoretical_loss": 5.052187418263609,
      "tokens_seen": 67043328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947843530591776,
      "loss": 3.8333,
      "theoretical_loss": 5.051464388858893,
      "tokens_seen": 67108864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 48877,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.7897393703460693,
      "objective/train/theoretical_loss": 5.050742262676536,
      "objective/train/tokens_used": 87634400,
      "theoretical_loss": 5.050742262676536,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947743229689067,
      "loss": 3.7583,
      "theoretical_loss": 5.050742262676536,
      "tokens_seen": 67174400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947642928786359,
      "loss": 4.0895,
      "theoretical_loss": 5.050021037708673,
      "tokens_seen": 67239936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947542627883651,
      "loss": 3.642,
      "theoretical_loss": 5.049300711953854,
      "tokens_seen": 67305472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947442326980943,
      "loss": 3.7434,
      "theoretical_loss": 5.048581283417015,
      "tokens_seen": 67371008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947342026078235,
      "loss": 3.7182,
      "theoretical_loss": 5.047862750109456,
      "tokens_seen": 67436544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947241725175527,
      "loss": 3.9326,
      "theoretical_loss": 5.047145110048811,
      "tokens_seen": 67502080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947141424272818,
      "loss": 3.638,
      "theoretical_loss": 5.046428361259022,
      "tokens_seen": 67567616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947041123370111,
      "loss": 3.7471,
      "theoretical_loss": 5.045712501770317,
      "tokens_seen": 67633152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946940822467402,
      "loss": 3.699,
      "theoretical_loss": 5.0449975296191765,
      "tokens_seen": 67698688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946840521564695,
      "loss": 3.8026,
      "theoretical_loss": 5.0442834428483145,
      "tokens_seen": 67764224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946740220661987,
      "loss": 3.7185,
      "theoretical_loss": 5.043570239506648,
      "tokens_seen": 67829760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946639919759278,
      "loss": 3.8962,
      "theoretical_loss": 5.042857917649275,
      "tokens_seen": 67895296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494653961885657,
      "loss": 3.7692,
      "theoretical_loss": 5.042146475337448,
      "tokens_seen": 67960832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946439317953862,
      "loss": 3.5141,
      "theoretical_loss": 5.041435910638544,
      "tokens_seen": 68026368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946339017051154,
      "loss": 3.8396,
      "theoretical_loss": 5.040726221626048,
      "tokens_seen": 68091904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946238716148446,
      "loss": 3.8633,
      "theoretical_loss": 5.040017406379521,
      "tokens_seen": 68157440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946138415245737,
      "loss": 3.6866,
      "theoretical_loss": 5.0393094629845745,
      "tokens_seen": 68222976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946038114343029,
      "loss": 3.5706,
      "theoretical_loss": 5.038602389532853,
      "tokens_seen": 68288512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945937813440321,
      "loss": 3.4773,
      "theoretical_loss": 5.037896184122,
      "tokens_seen": 68354048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945837512537613,
      "loss": 3.7606,
      "theoretical_loss": 5.037190844855641,
      "tokens_seen": 68419584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945737211634905,
      "loss": 3.8062,
      "theoretical_loss": 5.036486369843355,
      "tokens_seen": 68485120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945636910732196,
      "loss": 3.6164,
      "theoretical_loss": 5.03578275720065,
      "tokens_seen": 68550656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945536609829488,
      "loss": 3.7046,
      "theoretical_loss": 5.03508000504894,
      "tokens_seen": 68616192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945436308926781,
      "loss": 3.7478,
      "theoretical_loss": 5.034378111515521,
      "tokens_seen": 68681728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945336008024072,
      "loss": 3.663,
      "theoretical_loss": 5.0336770747335455,
      "tokens_seen": 68747264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 49922,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9727299213409424,
      "objective/train/theoretical_loss": 5.032976892842002,
      "objective/train/tokens_used": 89272800,
      "theoretical_loss": 5.032976892842002,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945235707121365,
      "loss": 3.6731,
      "theoretical_loss": 5.032976892842002,
      "tokens_seen": 68812800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945135406218656,
      "loss": 3.8125,
      "theoretical_loss": 5.032277563985687,
      "tokens_seen": 68878336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945035105315948,
      "loss": 3.8008,
      "theoretical_loss": 5.031579086315183,
      "tokens_seen": 68943872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494493480441324,
      "loss": 3.9551,
      "theoretical_loss": 5.030881457986837,
      "tokens_seen": 69009408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944834503510532,
      "loss": 3.843,
      "theoretical_loss": 5.030184677162734,
      "tokens_seen": 69074944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944734202607824,
      "loss": 3.5448,
      "theoretical_loss": 5.029488742010676,
      "tokens_seen": 69140480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944633901705116,
      "loss": 3.8376,
      "theoretical_loss": 5.028793650704154,
      "tokens_seen": 69206016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944533600802407,
      "loss": 3.9125,
      "theoretical_loss": 5.028099401422336,
      "tokens_seen": 69271552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944433299899699,
      "loss": 3.6831,
      "theoretical_loss": 5.0274059923500305,
      "tokens_seen": 69337088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944332998996991,
      "loss": 3.7007,
      "theoretical_loss": 5.026713421677673,
      "tokens_seen": 69402624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944232698094283,
      "loss": 3.6241,
      "theoretical_loss": 5.0260216876013,
      "tokens_seen": 69468160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944132397191575,
      "loss": 3.7006,
      "theoretical_loss": 5.025330788322526,
      "tokens_seen": 69533696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944032096288867,
      "loss": 3.9117,
      "theoretical_loss": 5.024640722048522,
      "tokens_seen": 69599232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943931795386158,
      "loss": 3.4086,
      "theoretical_loss": 5.023951486991994,
      "tokens_seen": 69664768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494383149448345,
      "loss": 3.9368,
      "theoretical_loss": 5.023263081371157,
      "tokens_seen": 69730304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943731193580742,
      "loss": 3.8784,
      "theoretical_loss": 5.022575503409717,
      "tokens_seen": 69795840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943630892678035,
      "loss": 3.941,
      "theoretical_loss": 5.021888751336848,
      "tokens_seen": 69861376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943530591775326,
      "loss": 3.8625,
      "theoretical_loss": 5.02120282338717,
      "tokens_seen": 69926912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943430290872618,
      "loss": 3.6762,
      "theoretical_loss": 5.020517717800723,
      "tokens_seen": 69992448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943329989969909,
      "loss": 3.4889,
      "theoretical_loss": 5.019833432822953,
      "tokens_seen": 70057984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943229689067202,
      "loss": 3.5122,
      "theoretical_loss": 5.019149966704683,
      "tokens_seen": 70123520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943129388164494,
      "loss": 3.6293,
      "theoretical_loss": 5.018467317702097,
      "tokens_seen": 70189056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943029087261786,
      "loss": 3.5085,
      "theoretical_loss": 5.017785484076715,
      "tokens_seen": 70254592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942928786359078,
      "loss": 3.6348,
      "theoretical_loss": 5.017104464095375,
      "tokens_seen": 70320128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828485456369,
      "loss": 3.6907,
      "theoretical_loss": 5.0164242560302075,
      "tokens_seen": 70385664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 51260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.9014105796813965,
      "objective/train/theoretical_loss": 5.015744858158618,
      "objective/train/tokens_used": 90911200,
      "theoretical_loss": 5.015744858158618,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942728184553661,
      "loss": 3.8765,
      "theoretical_loss": 5.015744858158618,
      "tokens_seen": 70451200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942627883650953,
      "loss": 3.8516,
      "theoretical_loss": 5.015066268763264,
      "tokens_seen": 70516736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942527582748245,
      "loss": 3.6519,
      "theoretical_loss": 5.014388486132036,
      "tokens_seen": 70582272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942427281845537,
      "loss": 3.9132,
      "theoretical_loss": 5.0137115085580355,
      "tokens_seen": 70647808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942326980942828,
      "loss": 3.8834,
      "theoretical_loss": 5.013035334339554,
      "tokens_seen": 70713344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494222668004012,
      "loss": 3.8794,
      "theoretical_loss": 5.012359961780053,
      "tokens_seen": 70778880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942126379137412,
      "loss": 3.8483,
      "theoretical_loss": 5.011685389188142,
      "tokens_seen": 70844416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942026078234704,
      "loss": 3.4989,
      "theoretical_loss": 5.011011614877562,
      "tokens_seen": 70909952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941925777331996,
      "loss": 3.7672,
      "theoretical_loss": 5.010338637167162,
      "tokens_seen": 70975488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941825476429289,
      "loss": 3.6359,
      "theoretical_loss": 5.009666454380877,
      "tokens_seen": 71041024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941725175526579,
      "loss": 3.8619,
      "theoretical_loss": 5.008995064847716,
      "tokens_seen": 71106560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941624874623872,
      "loss": 3.8008,
      "theoretical_loss": 5.0083244669017315,
      "tokens_seen": 71172096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941524573721163,
      "loss": 3.9147,
      "theoretical_loss": 5.0076546588820054,
      "tokens_seen": 71237632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941424272818456,
      "loss": 3.3797,
      "theoretical_loss": 5.00698563913263,
      "tokens_seen": 71303168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941323971915748,
      "loss": 3.617,
      "theoretical_loss": 5.006317406002687,
      "tokens_seen": 71368704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941223671013039,
      "loss": 3.5556,
      "theoretical_loss": 5.0056499578462255,
      "tokens_seen": 71434240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941123370110331,
      "loss": 3.4481,
      "theoretical_loss": 5.0049832930222475,
      "tokens_seen": 71499776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941023069207623,
      "loss": 3.826,
      "theoretical_loss": 5.004317409894682,
      "tokens_seen": 71565312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940922768304915,
      "loss": 3.6718,
      "theoretical_loss": 5.003652306832372,
      "tokens_seen": 71630848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940822467402207,
      "loss": 3.8578,
      "theoretical_loss": 5.002987982209052,
      "tokens_seen": 71696384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940722166499498,
      "loss": 3.758,
      "theoretical_loss": 5.00232443440333,
      "tokens_seen": 71761920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494062186559679,
      "loss": 3.4768,
      "theoretical_loss": 5.001661661798666,
      "tokens_seen": 71827456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940521564694082,
      "loss": 3.6442,
      "theoretical_loss": 5.000999662783359,
      "tokens_seen": 71892992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940421263791374,
      "loss": 3.7435,
      "theoretical_loss": 5.0003384357505185,
      "tokens_seen": 71958528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940320962888666,
      "loss": 3.8745,
      "theoretical_loss": 4.999677979098056,
      "tokens_seen": 72024064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 51878,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4065515995025635,
      "objective/train/theoretical_loss": 4.999018291228659,
      "objective/train/tokens_used": 92549600,
      "theoretical_loss": 4.999018291228659,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940220661985958,
      "loss": 3.5808,
      "theoretical_loss": 4.999018291228659,
      "tokens_seen": 72089600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940120361083249,
      "loss": 3.8043,
      "theoretical_loss": 4.998359370549778,
      "tokens_seen": 72155136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940020060180542,
      "loss": 3.649,
      "theoretical_loss": 4.997701215473601,
      "tokens_seen": 72220672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939919759277834,
      "loss": 3.6107,
      "theoretical_loss": 4.997043824417044,
      "tokens_seen": 72286208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939819458375126,
      "loss": 3.6793,
      "theoretical_loss": 4.996387195801723,
      "tokens_seen": 72351744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939719157472418,
      "loss": 3.6529,
      "theoretical_loss": 4.995731328053944,
      "tokens_seen": 72417280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493961885656971,
      "loss": 3.7394,
      "theoretical_loss": 4.995076219604682,
      "tokens_seen": 72482816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939518555667001,
      "loss": 3.6256,
      "theoretical_loss": 4.994421868889559,
      "tokens_seen": 72548352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939418254764293,
      "loss": 3.6209,
      "theoretical_loss": 4.9937682743488345,
      "tokens_seen": 72613888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939317953861585,
      "loss": 3.5004,
      "theoretical_loss": 4.993115434427376,
      "tokens_seen": 72679424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939217652958877,
      "loss": 3.5103,
      "theoretical_loss": 4.992463347574654,
      "tokens_seen": 72744960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939117352056169,
      "loss": 3.7514,
      "theoretical_loss": 4.991812012244717,
      "tokens_seen": 72810496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493901705115346,
      "loss": 3.5319,
      "theoretical_loss": 4.99116142689617,
      "tokens_seen": 72876032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938916750250752,
      "loss": 3.8628,
      "theoretical_loss": 4.990511589992169,
      "tokens_seen": 72941568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938816449348044,
      "loss": 4.0163,
      "theoretical_loss": 4.989862500000392,
      "tokens_seen": 73007104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938716148445337,
      "loss": 3.7264,
      "theoretical_loss": 4.989214155393025,
      "tokens_seen": 73072640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938615847542628,
      "loss": 3.9026,
      "theoretical_loss": 4.988566554646752,
      "tokens_seen": 73138176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493851554663992,
      "loss": 3.4925,
      "theoretical_loss": 4.9879196962427255,
      "tokens_seen": 73203712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938415245737211,
      "loss": 3.5623,
      "theoretical_loss": 4.9872735786665565,
      "tokens_seen": 73269248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938314944834504,
      "loss": 3.415,
      "theoretical_loss": 4.986628200408299,
      "tokens_seen": 73334784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938214643931796,
      "loss": 3.5253,
      "theoretical_loss": 4.985983559962429,
      "tokens_seen": 73400320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938114343029088,
      "loss": 3.5016,
      "theoretical_loss": 4.985339655827827,
      "tokens_seen": 73465856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493801404212638,
      "loss": 3.8156,
      "theoretical_loss": 4.984696486507767,
      "tokens_seen": 73531392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937913741223671,
      "loss": 3.7367,
      "theoretical_loss": 4.984054050509893,
      "tokens_seen": 73596928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937813440320963,
      "loss": 3.6885,
      "theoretical_loss": 4.983412346346206,
      "tokens_seen": 73662464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 53017,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.889392137527466,
      "objective/train/theoretical_loss": 4.982771372533048,
      "objective/train/tokens_used": 94188000,
      "theoretical_loss": 4.982771372533048,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937713139418255,
      "loss": 3.6721,
      "theoretical_loss": 4.982771372533048,
      "tokens_seen": 73728000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937612838515547,
      "loss": 3.9426,
      "theoretical_loss": 4.982131127591083,
      "tokens_seen": 73793536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937512537612839,
      "loss": 3.6135,
      "theoretical_loss": 4.981491610045282,
      "tokens_seen": 73859072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493741223671013,
      "loss": 3.6257,
      "theoretical_loss": 4.980852818424905,
      "tokens_seen": 73924608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937311935807422,
      "loss": 3.6675,
      "theoretical_loss": 4.980214751263491,
      "tokens_seen": 73990144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937211634904714,
      "loss": 3.6453,
      "theoretical_loss": 4.979577407098831,
      "tokens_seen": 74055680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937111334002006,
      "loss": 3.7323,
      "theoretical_loss": 4.978940784472962,
      "tokens_seen": 74121216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937011033099298,
      "loss": 3.4667,
      "theoretical_loss": 4.978304881932145,
      "tokens_seen": 74186752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936910732196591,
      "loss": 3.9063,
      "theoretical_loss": 4.97766969802685,
      "tokens_seen": 74252288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936810431293881,
      "loss": 3.3322,
      "theoretical_loss": 4.977035231311745,
      "tokens_seen": 74317824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936710130391174,
      "loss": 3.6439,
      "theoretical_loss": 4.976401480345672,
      "tokens_seen": 74383360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936609829488465,
      "loss": 3.5552,
      "theoretical_loss": 4.975768443691636,
      "tokens_seen": 74448896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936509528585758,
      "loss": 3.5095,
      "theoretical_loss": 4.97513611991679,
      "tokens_seen": 74514432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493640922768305,
      "loss": 3.7738,
      "theoretical_loss": 4.974504507592419,
      "tokens_seen": 74579968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936308926780341,
      "loss": 3.8278,
      "theoretical_loss": 4.973873605293922,
      "tokens_seen": 74645504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936208625877633,
      "loss": 3.7248,
      "theoretical_loss": 4.973243411600797,
      "tokens_seen": 74711040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936108324974925,
      "loss": 3.8762,
      "theoretical_loss": 4.9726139250966295,
      "tokens_seen": 74776576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936008024072217,
      "loss": 3.6798,
      "theoretical_loss": 4.971985144369073,
      "tokens_seen": 74842112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935907723169509,
      "loss": 3.6282,
      "theoretical_loss": 4.9713570680098345,
      "tokens_seen": 74907648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049358074222668,
      "loss": 3.7215,
      "theoretical_loss": 4.970729694614662,
      "tokens_seen": 74973184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935707121364092,
      "loss": 3.5638,
      "theoretical_loss": 4.970103022783325,
      "tokens_seen": 75038720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935606820461384,
      "loss": 3.5533,
      "theoretical_loss": 4.969477051119606,
      "tokens_seen": 75104256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935506519558676,
      "loss": 3.496,
      "theoretical_loss": 4.968851778231276,
      "tokens_seen": 75169792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935406218655968,
      "loss": 3.6459,
      "theoretical_loss": 4.9682272027300876,
      "tokens_seen": 75235328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493530591775326,
      "loss": 3.5601,
      "theoretical_loss": 4.96760332323176,
      "tokens_seen": 75300864
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 53591,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.8066093921661377,
      "objective/train/theoretical_loss": 4.966980138355957,
      "objective/train/tokens_used": 95826400,
      "theoretical_loss": 4.966980138355957,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935205616850551,
      "loss": 3.7377,
      "theoretical_loss": 4.966980138355957,
      "tokens_seen": 75366400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935105315947844,
      "loss": 3.6051,
      "theoretical_loss": 4.966357646726282,
      "tokens_seen": 75431936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935005015045135,
      "loss": 3.5847,
      "theoretical_loss": 4.965735846970255,
      "tokens_seen": 75497472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934904714142428,
      "loss": 3.7027,
      "theoretical_loss": 4.965114737719302,
      "tokens_seen": 75563008
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934804413239719,
      "loss": 3.648,
      "theoretical_loss": 4.964494317608743,
      "tokens_seen": 75628544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934704112337011,
      "loss": 3.4989,
      "theoretical_loss": 4.963874585277772,
      "tokens_seen": 75694080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934603811434303,
      "loss": 3.8889,
      "theoretical_loss": 4.963255539369445,
      "tokens_seen": 75759616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934503510531595,
      "loss": 3.5961,
      "theoretical_loss": 4.9626371785306675,
      "tokens_seen": 75825152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934403209628887,
      "loss": 3.4261,
      "theoretical_loss": 4.962019501412178,
      "tokens_seen": 75890688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934302908726179,
      "loss": 3.7473,
      "theoretical_loss": 4.961402506668535,
      "tokens_seen": 75956224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493420260782347,
      "loss": 3.4055,
      "theoretical_loss": 4.960786192958103,
      "tokens_seen": 76021760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934102306920762,
      "loss": 3.6034,
      "theoretical_loss": 4.960170558943037,
      "tokens_seen": 76087296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934002006018054,
      "loss": 3.4187,
      "theoretical_loss": 4.959555603289269,
      "tokens_seen": 76152832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933901705115346,
      "loss": 3.6193,
      "theoretical_loss": 4.9589413246664975,
      "tokens_seen": 76218368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933801404212638,
      "loss": 3.794,
      "theoretical_loss": 4.958327721748168,
      "tokens_seen": 76283904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493370110330993,
      "loss": 3.3982,
      "theoretical_loss": 4.957714793211462,
      "tokens_seen": 76349440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933600802407221,
      "loss": 3.3334,
      "theoretical_loss": 4.957102537737284,
      "tokens_seen": 76414976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933500501504513,
      "loss": 3.6421,
      "theoretical_loss": 4.956490954010249,
      "tokens_seen": 76480512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933400200601805,
      "loss": 3.5119,
      "theoretical_loss": 4.9558800407186645,
      "tokens_seen": 76546048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933299899699098,
      "loss": 3.4933,
      "theoretical_loss": 4.95526979655452,
      "tokens_seen": 76611584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933199598796389,
      "loss": 3.7028,
      "theoretical_loss": 4.954660220213473,
      "tokens_seen": 76677120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933099297893682,
      "loss": 3.7165,
      "theoretical_loss": 4.954051310394836,
      "tokens_seen": 76742656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932998996990972,
      "loss": 3.5507,
      "theoretical_loss": 4.953443065801563,
      "tokens_seen": 76808192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932898696088265,
      "loss": 3.4901,
      "theoretical_loss": 4.952835485140236,
      "tokens_seen": 76873728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932798395185557,
      "loss": 3.2798,
      "theoretical_loss": 4.952228567121052,
      "tokens_seen": 76939264
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 54447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6758010387420654,
      "objective/train/theoretical_loss": 4.951622310457807,
      "objective/train/tokens_used": 97464800,
      "theoretical_loss": 4.951622310457807,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932698094282849,
      "loss": 3.4817,
      "theoretical_loss": 4.951622310457807,
      "tokens_seen": 77004800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932597793380141,
      "loss": 3.5561,
      "theoretical_loss": 4.95101671386789,
      "tokens_seen": 77070336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932497492477432,
      "loss": 3.6278,
      "theoretical_loss": 4.950411776072261,
      "tokens_seen": 77135872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932397191574724,
      "loss": 3.7609,
      "theoretical_loss": 4.949807495795445,
      "tokens_seen": 77201408
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932296890672016,
      "loss": 3.6186,
      "theoretical_loss": 4.949203871765516,
      "tokens_seen": 77266944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932196589769308,
      "loss": 3.6535,
      "theoretical_loss": 4.948600902714083,
      "tokens_seen": 77332480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049320962888666,
      "loss": 3.4356,
      "theoretical_loss": 4.947998587376282,
      "tokens_seen": 77398016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931995987963893,
      "loss": 3.7607,
      "theoretical_loss": 4.947396924490757,
      "tokens_seen": 77463552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931895687061183,
      "loss": 3.9334,
      "theoretical_loss": 4.946795912799653,
      "tokens_seen": 77529088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931795386158476,
      "loss": 3.4391,
      "theoretical_loss": 4.946195551048597,
      "tokens_seen": 77594624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931695085255767,
      "loss": 3.4097,
      "theoretical_loss": 4.945595837986693,
      "tokens_seen": 77660160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493159478435306,
      "loss": 3.4903,
      "theoretical_loss": 4.944996772366503,
      "tokens_seen": 77725696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931494483450352,
      "loss": 3.5634,
      "theoretical_loss": 4.944398352944038,
      "tokens_seen": 77791232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931394182547643,
      "loss": 3.5448,
      "theoretical_loss": 4.943800578478745,
      "tokens_seen": 77856768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931293881644935,
      "loss": 3.4788,
      "theoretical_loss": 4.9432034477334925,
      "tokens_seen": 77922304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931193580742227,
      "loss": 3.4963,
      "theoretical_loss": 4.942606959474562,
      "tokens_seen": 77987840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931093279839519,
      "loss": 3.4853,
      "theoretical_loss": 4.942011112471633,
      "tokens_seen": 78053376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930992978936811,
      "loss": 3.3058,
      "theoretical_loss": 4.94141590549777,
      "tokens_seen": 78118912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930892678034102,
      "loss": 3.7623,
      "theoretical_loss": 4.940821337329414,
      "tokens_seen": 78184448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930792377131394,
      "loss": 3.324,
      "theoretical_loss": 4.940227406746368,
      "tokens_seen": 78249984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930692076228686,
      "loss": 3.595,
      "theoretical_loss": 4.939634112531784,
      "tokens_seen": 78315520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930591775325978,
      "loss": 3.6571,
      "theoretical_loss": 4.939041453472152,
      "tokens_seen": 78381056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493049147442327,
      "loss": 3.546,
      "theoretical_loss": 4.9384494283572895,
      "tokens_seen": 78446592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930391173520562,
      "loss": 3.7395,
      "theoretical_loss": 4.937858035980327,
      "tokens_seen": 78512128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930290872617853,
      "loss": 3.5258,
      "theoretical_loss": 4.9372672751377,
      "tokens_seen": 78577664
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 54900,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9137701988220215,
      "objective/train/theoretical_loss": 4.936677144629131,
      "objective/train/tokens_used": 99103200,
      "theoretical_loss": 4.936677144629131,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930190571715146,
      "loss": 3.2102,
      "theoretical_loss": 4.936677144629131,
      "tokens_seen": 78643200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930090270812437,
      "loss": 3.7118,
      "theoretical_loss": 4.936087643257625,
      "tokens_seen": 78708736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492998996990973,
      "loss": 3.5876,
      "theoretical_loss": 4.935498769829453,
      "tokens_seen": 78774272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929889669007021,
      "loss": 3.7919,
      "theoretical_loss": 4.934910523154139,
      "tokens_seen": 78839808
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929789368104313,
      "loss": 3.8801,
      "theoretical_loss": 4.934322902044457,
      "tokens_seen": 78905344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929689067201605,
      "loss": 3.5096,
      "theoretical_loss": 4.933735905316408,
      "tokens_seen": 78970880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929588766298897,
      "loss": 3.4217,
      "theoretical_loss": 4.933149531789217,
      "tokens_seen": 79036416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929488465396189,
      "loss": 3.6953,
      "theoretical_loss": 4.932563780285319,
      "tokens_seen": 79101952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929388164493481,
      "loss": 3.6487,
      "theoretical_loss": 4.931978649630346,
      "tokens_seen": 79167488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929287863590773,
      "loss": 3.6641,
      "theoretical_loss": 4.931394138653117,
      "tokens_seen": 79233024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929187562688064,
      "loss": 3.8094,
      "theoretical_loss": 4.930810246185626,
      "tokens_seen": 79298560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929087261785356,
      "loss": 3.5257,
      "theoretical_loss": 4.930226971063035,
      "tokens_seen": 79364096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928986960882648,
      "loss": 3.4985,
      "theoretical_loss": 4.929644312123655,
      "tokens_seen": 79429632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492888665997994,
      "loss": 3.3258,
      "theoretical_loss": 4.929062268208941,
      "tokens_seen": 79495168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928786359077232,
      "loss": 3.2647,
      "theoretical_loss": 4.92848083816348,
      "tokens_seen": 79560704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928686058174523,
      "loss": 3.8228,
      "theoretical_loss": 4.927900020834979,
      "tokens_seen": 79626240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928585757271815,
      "loss": 3.3943,
      "theoretical_loss": 4.927319815074252,
      "tokens_seen": 79691776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928485456369107,
      "loss": 3.4187,
      "theoretical_loss": 4.926740219735214,
      "tokens_seen": 79757312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.00049283851554664,
      "loss": 3.6324,
      "theoretical_loss": 4.926161233674863,
      "tokens_seen": 79822848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928284854563691,
      "loss": 3.627,
      "theoretical_loss": 4.925582855753279,
      "tokens_seen": 79888384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928184553660984,
      "loss": 3.4882,
      "theoretical_loss": 4.925005084833603,
      "tokens_seen": 79953920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928084252758274,
      "loss": 3.8262,
      "theoretical_loss": 4.924427919782033,
      "tokens_seen": 80019456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927983951855567,
      "loss": 3.5192,
      "theoretical_loss": 4.923851359467811,
      "tokens_seen": 80084992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927883650952859,
      "loss": 3.714,
      "theoretical_loss": 4.923275402763213,
      "tokens_seen": 80150528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927783350050151,
      "loss": 3.4187,
      "theoretical_loss": 4.922700048543536,
      "tokens_seen": 80216064
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 56090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1410176753997803,
      "objective/train/theoretical_loss": 4.922125295687091,
      "objective/train/tokens_used": 100741600,
      "theoretical_loss": 4.922125295687091,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927683049147443,
      "loss": 3.5733,
      "theoretical_loss": 4.922125295687091,
      "tokens_seen": 80281600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927582748244734,
      "loss": 3.396,
      "theoretical_loss": 4.92155114307519,
      "tokens_seen": 80347136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927482447342026,
      "loss": 3.6166,
      "theoretical_loss": 4.9209775895921375,
      "tokens_seen": 80412672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927382146439318,
      "loss": 3.4859,
      "theoretical_loss": 4.920404634125217,
      "tokens_seen": 80478208
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492728184553661,
      "loss": 3.2846,
      "theoretical_loss": 4.919832275564684,
      "tokens_seen": 80543744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927181544633902,
      "loss": 3.653,
      "theoretical_loss": 4.9192605128037545,
      "tokens_seen": 80609280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927081243731193,
      "loss": 3.5658,
      "theoretical_loss": 4.918689344738594,
      "tokens_seen": 80674816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926980942828485,
      "loss": 3.5587,
      "theoretical_loss": 4.918118770268305,
      "tokens_seen": 80740352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926880641925777,
      "loss": 3.3881,
      "theoretical_loss": 4.917548788294924,
      "tokens_seen": 80805888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926780341023069,
      "loss": 3.4548,
      "theoretical_loss": 4.916979397723403,
      "tokens_seen": 80871424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926680040120361,
      "loss": 3.6868,
      "theoretical_loss": 4.916410597461606,
      "tokens_seen": 80936960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926579739217654,
      "loss": 3.6518,
      "theoretical_loss": 4.9158423864202945,
      "tokens_seen": 81002496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926479438314944,
      "loss": 3.4168,
      "theoretical_loss": 4.915274763513119,
      "tokens_seen": 81068032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926379137412237,
      "loss": 3.353,
      "theoretical_loss": 4.914707727656609,
      "tokens_seen": 81133568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926278836509528,
      "loss": 3.3831,
      "theoretical_loss": 4.914141277770164,
      "tokens_seen": 81199104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926178535606821,
      "loss": 3.1851,
      "theoretical_loss": 4.913575412776042,
      "tokens_seen": 81264640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926078234704113,
      "loss": 3.7187,
      "theoretical_loss": 4.91301013159935,
      "tokens_seen": 81330176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925977933801404,
      "loss": 3.5052,
      "theoretical_loss": 4.912445433168036,
      "tokens_seen": 81395712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925877632898696,
      "loss": 3.5131,
      "theoretical_loss": 4.911881316412876,
      "tokens_seen": 81461248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925777331995988,
      "loss": 3.7694,
      "theoretical_loss": 4.911317780267467,
      "tokens_seen": 81526784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492567703109328,
      "loss": 3.3832,
      "theoretical_loss": 4.910754823668215,
      "tokens_seen": 81592320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925576730190572,
      "loss": 3.8546,
      "theoretical_loss": 4.910192445554329,
      "tokens_seen": 81657856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925476429287864,
      "loss": 3.6127,
      "theoretical_loss": 4.909630644867807,
      "tokens_seen": 81723392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925376128385155,
      "loss": 3.3545,
      "theoretical_loss": 4.909069420553429,
      "tokens_seen": 81788928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925275827482447,
      "loss": 3.5589,
      "theoretical_loss": 4.908508771558748,
      "tokens_seen": 81854464
    },
    {
      "epoch": 0.02,
      "objective/train/docs_used": 57069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6827940940856934,
      "objective/train/theoretical_loss": 4.907948696834077,
      "objective/train/tokens_used": 102380000,
      "theoretical_loss": 4.907948696834077,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925175526579739,
      "loss": 3.7195,
      "theoretical_loss": 4.907948696834077,
      "tokens_seen": 81920000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925075225677031,
      "loss": 3.5548,
      "theoretical_loss": 4.907389195332484,
      "tokens_seen": 81985536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924974924774323,
      "loss": 3.807,
      "theoretical_loss": 4.906830266009781,
      "tokens_seen": 82051072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924874623871615,
      "loss": 3.4511,
      "theoretical_loss": 4.90627190782451,
      "tokens_seen": 82116608
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924774322968907,
      "loss": 3.7151,
      "theoretical_loss": 4.905714119737945,
      "tokens_seen": 82182144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924674022066199,
      "loss": 3.5237,
      "theoretical_loss": 4.9051569007140685,
      "tokens_seen": 82247680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924573721163491,
      "loss": 3.332,
      "theoretical_loss": 4.904600249719573,
      "tokens_seen": 82313216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924473420260783,
      "loss": 3.504,
      "theoretical_loss": 4.90404416572385,
      "tokens_seen": 82378752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924373119358075,
      "loss": 3.7408,
      "theoretical_loss": 4.903488647698972,
      "tokens_seen": 82444288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924272818455366,
      "loss": 3.334,
      "theoretical_loss": 4.902933694619698,
      "tokens_seen": 82509824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924172517552658,
      "loss": 3.2703,
      "theoretical_loss": 4.902379305463451,
      "tokens_seen": 82575360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492407221664995,
      "loss": 3.5645,
      "theoretical_loss": 4.901825479210317,
      "tokens_seen": 82640896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923971915747242,
      "loss": 3.3278,
      "theoretical_loss": 4.9012722148430345,
      "tokens_seen": 82706432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923871614844534,
      "loss": 3.6277,
      "theoretical_loss": 4.900719511346983,
      "tokens_seen": 82771968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923771313941825,
      "loss": 3.6241,
      "theoretical_loss": 4.900167367710176,
      "tokens_seen": 82837504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923671013039117,
      "loss": 3.5454,
      "theoretical_loss": 4.899615782923251,
      "tokens_seen": 82903040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923570712136409,
      "loss": 3.6864,
      "theoretical_loss": 4.899064755979464,
      "tokens_seen": 82968576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923470411233702,
      "loss": 3.3855,
      "theoretical_loss": 4.898514285874674,
      "tokens_seen": 83034112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923370110330993,
      "loss": 3.4924,
      "theoretical_loss": 4.897964371607341,
      "tokens_seen": 83099648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923269809428286,
      "loss": 3.4934,
      "theoretical_loss": 4.897415012178515,
      "tokens_seen": 83165184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923169508525576,
      "loss": 3.3992,
      "theoretical_loss": 4.896866206591823,
      "tokens_seen": 83230720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923069207622869,
      "loss": 3.5205,
      "theoretical_loss": 4.896317953853467,
      "tokens_seen": 83296256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922968906720161,
      "loss": 3.624,
      "theoretical_loss": 4.8957702529722145,
      "tokens_seen": 83361792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922868605817453,
      "loss": 3.5903,
      "theoretical_loss": 4.895223102959381,
      "tokens_seen": 83427328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922768304914745,
      "loss": 3.7871,
      "theoretical_loss": 4.894676502828836,
      "tokens_seen": 83492864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 57750,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.804002046585083,
      "objective/train/theoretical_loss": 4.894130451596979,
      "objective/train/tokens_used": 104018400,
      "theoretical_loss": 4.894130451596979,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922668004012036,
      "loss": 3.4882,
      "theoretical_loss": 4.894130451596979,
      "tokens_seen": 83558400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922567703109328,
      "loss": 3.2586,
      "theoretical_loss": 4.893584948282745,
      "tokens_seen": 83623936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492246740220662,
      "loss": 3.5408,
      "theoretical_loss": 4.893039991907587,
      "tokens_seen": 83689472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922367101303912,
      "loss": 3.6106,
      "theoretical_loss": 4.892495581495468,
      "tokens_seen": 83755008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922266800401204,
      "loss": 3.2746,
      "theoretical_loss": 4.8919517160728585,
      "tokens_seen": 83820544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922166499498495,
      "loss": 3.5637,
      "theoretical_loss": 4.891408394668723,
      "tokens_seen": 83886080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922066198595787,
      "loss": 3.5945,
      "theoretical_loss": 4.89086561631451,
      "tokens_seen": 83951616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921965897693079,
      "loss": 3.6297,
      "theoretical_loss": 4.890323380044153,
      "tokens_seen": 84017152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921865596790371,
      "loss": 3.6036,
      "theoretical_loss": 4.889781684894051,
      "tokens_seen": 84082688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921765295887663,
      "loss": 3.7009,
      "theoretical_loss": 4.889240529903066,
      "tokens_seen": 84148224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921664994984956,
      "loss": 3.4759,
      "theoretical_loss": 4.888699914112516,
      "tokens_seen": 84213760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921564694082246,
      "loss": 3.4833,
      "theoretical_loss": 4.888159836566162,
      "tokens_seen": 84279296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921464393179539,
      "loss": 3.7335,
      "theoretical_loss": 4.887620296310205,
      "tokens_seen": 84344832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492136409227683,
      "loss": 3.8101,
      "theoretical_loss": 4.887081292393276,
      "tokens_seen": 84410368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921263791374123,
      "loss": 3.6913,
      "theoretical_loss": 4.886542823866424,
      "tokens_seen": 84475904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921163490471415,
      "loss": 3.5273,
      "theoretical_loss": 4.886004889783116,
      "tokens_seen": 84541440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921063189568706,
      "loss": 3.5718,
      "theoretical_loss": 4.885467489199222,
      "tokens_seen": 84606976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920962888665998,
      "loss": 3.6503,
      "theoretical_loss": 4.884930621173009,
      "tokens_seen": 84672512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492086258776329,
      "loss": 3.5557,
      "theoretical_loss": 4.884394284765135,
      "tokens_seen": 84738048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920762286860582,
      "loss": 3.5114,
      "theoretical_loss": 4.88385847903864,
      "tokens_seen": 84803584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920661985957874,
      "loss": 3.8069,
      "theoretical_loss": 4.883323203058936,
      "tokens_seen": 84869120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920561685055166,
      "loss": 3.529,
      "theoretical_loss": 4.882788455893802,
      "tokens_seen": 84934656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920461384152457,
      "loss": 3.5584,
      "theoretical_loss": 4.882254236613376,
      "tokens_seen": 85000192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920361083249749,
      "loss": 3.7302,
      "theoretical_loss": 4.881720544290144,
      "tokens_seen": 85065728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920260782347041,
      "loss": 3.4352,
      "theoretical_loss": 4.881187377998938,
      "tokens_seen": 85131264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 58904,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6649792194366455,
      "objective/train/theoretical_loss": 4.880654736816922,
      "objective/train/tokens_used": 105656800,
      "theoretical_loss": 4.880654736816922,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920160481444333,
      "loss": 3.7428,
      "theoretical_loss": 4.880654736816922,
      "tokens_seen": 85196800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920060180541625,
      "loss": 3.3637,
      "theoretical_loss": 4.880122619823589,
      "tokens_seen": 85262336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919959879638916,
      "loss": 3.4179,
      "theoretical_loss": 4.87959102610075,
      "tokens_seen": 85327872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919859578736209,
      "loss": 3.3888,
      "theoretical_loss": 4.879059954732529,
      "tokens_seen": 85393408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049197592778335,
      "loss": 3.7427,
      "theoretical_loss": 4.878529404805356,
      "tokens_seen": 85458944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919658976930793,
      "loss": 3.7709,
      "theoretical_loss": 4.8779993754079545,
      "tokens_seen": 85524480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919558676028084,
      "loss": 3.5494,
      "theoretical_loss": 4.877469865631338,
      "tokens_seen": 85590016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919458375125377,
      "loss": 3.5295,
      "theoretical_loss": 4.876940874568804,
      "tokens_seen": 85655552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919358074222668,
      "loss": 3.5878,
      "theoretical_loss": 4.876412401315923,
      "tokens_seen": 85721088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491925777331996,
      "loss": 3.57,
      "theoretical_loss": 4.8758844449705325,
      "tokens_seen": 85786624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919157472417252,
      "loss": 3.5246,
      "theoretical_loss": 4.875357004632728,
      "tokens_seen": 85852160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919057171514544,
      "loss": 3.3963,
      "theoretical_loss": 4.87483007940486,
      "tokens_seen": 85917696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918956870611836,
      "loss": 3.6586,
      "theoretical_loss": 4.874303668391521,
      "tokens_seen": 85983232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918856569709127,
      "loss": 3.6073,
      "theoretical_loss": 4.873777770699542,
      "tokens_seen": 86048768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918756268806419,
      "loss": 3.5692,
      "theoretical_loss": 4.873252385437985,
      "tokens_seen": 86114304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918655967903711,
      "loss": 3.5077,
      "theoretical_loss": 4.872727511718134,
      "tokens_seen": 86179840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918555667001003,
      "loss": 3.6584,
      "theoretical_loss": 4.8722031486534885,
      "tokens_seen": 86245376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918455366098295,
      "loss": 3.4986,
      "theoretical_loss": 4.8716792953597565,
      "tokens_seen": 86310912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918355065195586,
      "loss": 3.7466,
      "theoretical_loss": 4.871155950954847,
      "tokens_seen": 86376448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918254764292878,
      "loss": 3.627,
      "theoretical_loss": 4.870633114558865,
      "tokens_seen": 86441984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491815446339017,
      "loss": 3.5282,
      "theoretical_loss": 4.8701107852941,
      "tokens_seen": 86507520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918054162487463,
      "loss": 3.1841,
      "theoretical_loss": 4.869588962285023,
      "tokens_seen": 86573056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917953861584754,
      "loss": 3.6871,
      "theoretical_loss": 4.869067644658276,
      "tokens_seen": 86638592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917853560682047,
      "loss": 3.6656,
      "theoretical_loss": 4.86854683154267,
      "tokens_seen": 86704128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917753259779337,
      "loss": 3.5156,
      "theoretical_loss": 4.868026522069172,
      "tokens_seen": 86769664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 59520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.652179718017578,
      "objective/train/theoretical_loss": 4.867506715370903,
      "objective/train/tokens_used": 107295200,
      "theoretical_loss": 4.867506715370903,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491765295887663,
      "loss": 3.4928,
      "theoretical_loss": 4.867506715370903,
      "tokens_seen": 86835200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917552657973922,
      "loss": 3.4802,
      "theoretical_loss": 4.866987410583125,
      "tokens_seen": 86900736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917452357071214,
      "loss": 3.5752,
      "theoretical_loss": 4.8664686068432434,
      "tokens_seen": 86966272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917352056168506,
      "loss": 3.5258,
      "theoretical_loss": 4.8659503032907905,
      "tokens_seen": 87031808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917251755265797,
      "loss": 3.7263,
      "theoretical_loss": 4.865432499067425,
      "tokens_seen": 87097344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917151454363089,
      "loss": 3.7067,
      "theoretical_loss": 4.864915193316922,
      "tokens_seen": 87162880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917051153460381,
      "loss": 3.4247,
      "theoretical_loss": 4.864398385185169,
      "tokens_seen": 87228416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916950852557673,
      "loss": 3.5719,
      "theoretical_loss": 4.863882073820154,
      "tokens_seen": 87293952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916850551654965,
      "loss": 3.5349,
      "theoretical_loss": 4.863366258371963,
      "tokens_seen": 87359488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916750250752258,
      "loss": 3.5815,
      "theoretical_loss": 4.862850937992777,
      "tokens_seen": 87425024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916649949849548,
      "loss": 3.5233,
      "theoretical_loss": 4.862336111836852,
      "tokens_seen": 87490560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916549648946841,
      "loss": 3.5181,
      "theoretical_loss": 4.861821779060529,
      "tokens_seen": 87556096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916449348044132,
      "loss": 3.369,
      "theoretical_loss": 4.861307938822215,
      "tokens_seen": 87621632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916349047141425,
      "loss": 3.3532,
      "theoretical_loss": 4.86079459028238,
      "tokens_seen": 87687168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916248746238717,
      "loss": 3.4634,
      "theoretical_loss": 4.860281732603555,
      "tokens_seen": 87752704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916148445336008,
      "loss": 3.64,
      "theoretical_loss": 4.859769364950319,
      "tokens_seen": 87818240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049160481444333,
      "loss": 3.5164,
      "theoretical_loss": 4.859257486489293,
      "tokens_seen": 87883776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915947843530592,
      "loss": 3.5639,
      "theoretical_loss": 4.858746096389139,
      "tokens_seen": 87949312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915847542627884,
      "loss": 3.4632,
      "theoretical_loss": 4.8582351938205495,
      "tokens_seen": 88014848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915747241725176,
      "loss": 3.2569,
      "theoretical_loss": 4.857724777956237,
      "tokens_seen": 88080384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915646940822468,
      "loss": 3.4019,
      "theoretical_loss": 4.857214847970939,
      "tokens_seen": 88145920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915546639919759,
      "loss": 3.266,
      "theoretical_loss": 4.856705403041397,
      "tokens_seen": 88211456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915446339017051,
      "loss": 3.7935,
      "theoretical_loss": 4.8561964423463655,
      "tokens_seen": 88276992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915346038114343,
      "loss": 3.7385,
      "theoretical_loss": 4.8556879650665925,
      "tokens_seen": 88342528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915245737211635,
      "loss": 3.3908,
      "theoretical_loss": 4.855179970384819,
      "tokens_seen": 88408064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 60543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.110377788543701,
      "objective/train/theoretical_loss": 4.8546724574857745,
      "objective/train/tokens_used": 108933600,
      "theoretical_loss": 4.8546724574857745,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915145436308927,
      "loss": 3.4761,
      "theoretical_loss": 4.8546724574857745,
      "tokens_seen": 88473600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915045135406218,
      "loss": 3.442,
      "theoretical_loss": 4.854165425556166,
      "tokens_seen": 88539136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914944834503511,
      "loss": 3.613,
      "theoretical_loss": 4.853658873784675,
      "tokens_seen": 88604672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914844533600802,
      "loss": 3.7501,
      "theoretical_loss": 4.853152801361952,
      "tokens_seen": 88670208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914744232698095,
      "loss": 3.4899,
      "theoretical_loss": 4.852647207480604,
      "tokens_seen": 88735744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914643931795386,
      "loss": 3.6797,
      "theoretical_loss": 4.8521420913352,
      "tokens_seen": 88801280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914543630892679,
      "loss": 3.2736,
      "theoretical_loss": 4.851637452122251,
      "tokens_seen": 88866816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491444332998997,
      "loss": 3.5697,
      "theoretical_loss": 4.851133289040214,
      "tokens_seen": 88932352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343029087262,
      "loss": 3.6628,
      "theoretical_loss": 4.850629601289482,
      "tokens_seen": 88997888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914242728184554,
      "loss": 3.4603,
      "theoretical_loss": 4.850126388072379,
      "tokens_seen": 89063424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914142427281846,
      "loss": 3.5286,
      "theoretical_loss": 4.849623648593151,
      "tokens_seen": 89128960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914042126379138,
      "loss": 3.4864,
      "theoretical_loss": 4.849121382057964,
      "tokens_seen": 89194496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913941825476429,
      "loss": 3.5269,
      "theoretical_loss": 4.848619587674898,
      "tokens_seen": 89260032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913841524573721,
      "loss": 3.4857,
      "theoretical_loss": 4.8481182646539365,
      "tokens_seen": 89325568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913741223671013,
      "loss": 3.6437,
      "theoretical_loss": 4.847617412206963,
      "tokens_seen": 89391104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913640922768305,
      "loss": 3.4963,
      "theoretical_loss": 4.847117029547758,
      "tokens_seen": 89456640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913540621865597,
      "loss": 3.5452,
      "theoretical_loss": 4.846617115891988,
      "tokens_seen": 89522176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913440320962888,
      "loss": 3.2749,
      "theoretical_loss": 4.8461176704572,
      "tokens_seen": 89587712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491334002006018,
      "loss": 3.319,
      "theoretical_loss": 4.845618692462826,
      "tokens_seen": 89653248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913239719157472,
      "loss": 3.4855,
      "theoretical_loss": 4.845120181130158,
      "tokens_seen": 89718784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913139418254765,
      "loss": 3.7812,
      "theoretical_loss": 4.844622135682359,
      "tokens_seen": 89784320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913039117352056,
      "loss": 3.1217,
      "theoretical_loss": 4.84412455534445,
      "tokens_seen": 89849856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912938816449349,
      "loss": 3.3833,
      "theoretical_loss": 4.843627439343306,
      "tokens_seen": 89915392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912838515546639,
      "loss": 3.5363,
      "theoretical_loss": 4.843130786907647,
      "tokens_seen": 89980928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912738214643932,
      "loss": 3.4395,
      "theoretical_loss": 4.842634597268036,
      "tokens_seen": 90046464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 60961,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3933932781219482,
      "objective/train/theoretical_loss": 4.842138869656871,
      "objective/train/tokens_used": 110572000,
      "theoretical_loss": 4.842138869656871,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912637913741224,
      "loss": 3.5361,
      "theoretical_loss": 4.842138869656871,
      "tokens_seen": 90112000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912537612838516,
      "loss": 3.1717,
      "theoretical_loss": 4.841643603308384,
      "tokens_seen": 90177536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912437311935808,
      "loss": 3.6274,
      "theoretical_loss": 4.841148797458627,
      "tokens_seen": 90243072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049123370110331,
      "loss": 3.5958,
      "theoretical_loss": 4.840654451345472,
      "tokens_seen": 90308608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912236710130391,
      "loss": 3.6702,
      "theoretical_loss": 4.840160564208606,
      "tokens_seen": 90374144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912136409227683,
      "loss": 3.4133,
      "theoretical_loss": 4.839667135289522,
      "tokens_seen": 90439680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912036108324975,
      "loss": 3.3868,
      "theoretical_loss": 4.839174163831515,
      "tokens_seen": 90505216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911935807422267,
      "loss": 3.4971,
      "theoretical_loss": 4.838681649079677,
      "tokens_seen": 90570752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911835506519559,
      "loss": 3.581,
      "theoretical_loss": 4.838189590280891,
      "tokens_seen": 90636288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491173520561685,
      "loss": 3.5547,
      "theoretical_loss": 4.837697986683825,
      "tokens_seen": 90701824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911634904714142,
      "loss": 3.6512,
      "theoretical_loss": 4.837206837538925,
      "tokens_seen": 90767360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911534603811434,
      "loss": 3.189,
      "theoretical_loss": 4.836716142098417,
      "tokens_seen": 90832896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911434302908726,
      "loss": 3.4003,
      "theoretical_loss": 4.836225899616288,
      "tokens_seen": 90898432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911334002006019,
      "loss": 3.5343,
      "theoretical_loss": 4.835736109348293,
      "tokens_seen": 90963968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911233701103309,
      "loss": 3.491,
      "theoretical_loss": 4.8352467705519455,
      "tokens_seen": 91029504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911133400200602,
      "loss": 3.543,
      "theoretical_loss": 4.83475788248651,
      "tokens_seen": 91095040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911033099297893,
      "loss": 3.5829,
      "theoretical_loss": 4.834269444412996,
      "tokens_seen": 91160576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910932798395186,
      "loss": 3.5038,
      "theoretical_loss": 4.83378145559416,
      "tokens_seen": 91226112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910832497492478,
      "loss": 3.6003,
      "theoretical_loss": 4.83329391529449,
      "tokens_seen": 91291648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491073219658977,
      "loss": 3.5657,
      "theoretical_loss": 4.832806822780208,
      "tokens_seen": 91357184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910631895687061,
      "loss": 3.3872,
      "theoretical_loss": 4.832320177319261,
      "tokens_seen": 91422720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910531594784353,
      "loss": 3.4167,
      "theoretical_loss": 4.831833978181314,
      "tokens_seen": 91488256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910431293881645,
      "loss": 3.2479,
      "theoretical_loss": 4.831348224637752,
      "tokens_seen": 91553792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910330992978937,
      "loss": 3.5555,
      "theoretical_loss": 4.830862915961665,
      "tokens_seen": 91619328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910230692076229,
      "loss": 3.4961,
      "theoretical_loss": 4.830378051427852,
      "tokens_seen": 91684864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 62148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0938072204589844,
      "objective/train/theoretical_loss": 4.829893630312808,
      "objective/train/tokens_used": 112210400,
      "theoretical_loss": 4.829893630312808,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491013039117352,
      "loss": 3.1482,
      "theoretical_loss": 4.829893630312808,
      "tokens_seen": 91750400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910030090270812,
      "loss": 3.4885,
      "theoretical_loss": 4.829409651894725,
      "tokens_seen": 91815936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909929789368104,
      "loss": 3.5385,
      "theoretical_loss": 4.8289261154534815,
      "tokens_seen": 91881472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909829488465397,
      "loss": 3.2172,
      "theoretical_loss": 4.828443020270642,
      "tokens_seen": 91947008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909729187562688,
      "loss": 3.351,
      "theoretical_loss": 4.827960365629449,
      "tokens_seen": 92012544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909628886659981,
      "loss": 3.6262,
      "theoretical_loss": 4.827478150814821,
      "tokens_seen": 92078080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909528585757272,
      "loss": 3.4536,
      "theoretical_loss": 4.826996375113339,
      "tokens_seen": 92143616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909428284854564,
      "loss": 3.4569,
      "theoretical_loss": 4.826515037813255,
      "tokens_seen": 92209152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909327983951856,
      "loss": 3.7922,
      "theoretical_loss": 4.826034138204474,
      "tokens_seen": 92274688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909227683049148,
      "loss": 3.4368,
      "theoretical_loss": 4.825553675578557,
      "tokens_seen": 92340224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490912738214644,
      "loss": 3.3922,
      "theoretical_loss": 4.825073649228713,
      "tokens_seen": 92405760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909027081243731,
      "loss": 3.5078,
      "theoretical_loss": 4.824594058449794,
      "tokens_seen": 92471296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908926780341023,
      "loss": 3.3106,
      "theoretical_loss": 4.824114902538292,
      "tokens_seen": 92536832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908826479438315,
      "loss": 3.136,
      "theoretical_loss": 4.823636180792329,
      "tokens_seen": 92602368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908726178535607,
      "loss": 3.3286,
      "theoretical_loss": 4.823157892511661,
      "tokens_seen": 92667904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908625877632899,
      "loss": 3.6266,
      "theoretical_loss": 4.82268003699766,
      "tokens_seen": 92733440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490852557673019,
      "loss": 3.5695,
      "theoretical_loss": 4.8222026135533245,
      "tokens_seen": 92798976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908425275827482,
      "loss": 3.487,
      "theoretical_loss": 4.821725621483262,
      "tokens_seen": 92864512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908324974924774,
      "loss": 3.503,
      "theoretical_loss": 4.821249060093692,
      "tokens_seen": 92930048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908224674022067,
      "loss": 3.3758,
      "theoretical_loss": 4.820772928692433,
      "tokens_seen": 92995584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908124373119358,
      "loss": 3.4497,
      "theoretical_loss": 4.820297226588908,
      "tokens_seen": 93061120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908024072216651,
      "loss": 3.6942,
      "theoretical_loss": 4.819821953094133,
      "tokens_seen": 93126656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907923771313941,
      "loss": 3.6533,
      "theoretical_loss": 4.819347107520713,
      "tokens_seen": 93192192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907823470411234,
      "loss": 3.3973,
      "theoretical_loss": 4.818872689182836,
      "tokens_seen": 93257728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907723169508526,
      "loss": 3.2205,
      "theoretical_loss": 4.818398697396274,
      "tokens_seen": 93323264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 62815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5069005489349365,
      "objective/train/theoretical_loss": 4.817925131478374,
      "objective/train/tokens_used": 113848800,
      "theoretical_loss": 4.817925131478374,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907622868605818,
      "loss": 3.5578,
      "theoretical_loss": 4.817925131478374,
      "tokens_seen": 93388800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490752256770311,
      "loss": 3.5471,
      "theoretical_loss": 4.817451990748047,
      "tokens_seen": 93454336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907422266800401,
      "loss": 3.6923,
      "theoretical_loss": 4.81697927452578,
      "tokens_seen": 93519872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907321965897693,
      "loss": 3.5227,
      "theoretical_loss": 4.816506982133613,
      "tokens_seen": 93585408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907221664994985,
      "loss": 3.2295,
      "theoretical_loss": 4.816035112895147,
      "tokens_seen": 93650944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907121364092277,
      "loss": 3.1312,
      "theoretical_loss": 4.815563666135532,
      "tokens_seen": 93716480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907021063189569,
      "loss": 3.6798,
      "theoretical_loss": 4.815092641181467,
      "tokens_seen": 93782016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490692076228686,
      "loss": 3.7669,
      "theoretical_loss": 4.814622037361195,
      "tokens_seen": 93847552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906820461384152,
      "loss": 3.6417,
      "theoretical_loss": 4.814151854004493,
      "tokens_seen": 93913088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906720160481444,
      "loss": 3.2937,
      "theoretical_loss": 4.813682090442674,
      "tokens_seen": 93978624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906619859578736,
      "loss": 3.2833,
      "theoretical_loss": 4.81321274600858,
      "tokens_seen": 94044160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906519558676028,
      "loss": 3.2788,
      "theoretical_loss": 4.812743820036577,
      "tokens_seen": 94109696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906419257773321,
      "loss": 3.5618,
      "theoretical_loss": 4.81227531186255,
      "tokens_seen": 94175232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906318956870611,
      "loss": 3.5329,
      "theoretical_loss": 4.811807220823899,
      "tokens_seen": 94240768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906218655967904,
      "loss": 3.6156,
      "theoretical_loss": 4.811339546259536,
      "tokens_seen": 94306304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906118355065195,
      "loss": 3.3777,
      "theoretical_loss": 4.810872287509877,
      "tokens_seen": 94371840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906018054162488,
      "loss": 3.5883,
      "theoretical_loss": 4.810405443916842,
      "tokens_seen": 94437376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490591775325978,
      "loss": 3.6907,
      "theoretical_loss": 4.809939014823849,
      "tokens_seen": 94502912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905817452357072,
      "loss": 3.6602,
      "theoretical_loss": 4.809472999575805,
      "tokens_seen": 94568448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905717151454363,
      "loss": 3.6322,
      "theoretical_loss": 4.809007397519108,
      "tokens_seen": 94633984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905616850551655,
      "loss": 3.4204,
      "theoretical_loss": 4.80854220800164,
      "tokens_seen": 94699520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905516549648947,
      "loss": 3.5405,
      "theoretical_loss": 4.808077430372762,
      "tokens_seen": 94765056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905416248746239,
      "loss": 3.4647,
      "theoretical_loss": 4.807613063983311,
      "tokens_seen": 94830592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905315947843531,
      "loss": 3.7424,
      "theoretical_loss": 4.807149108185593,
      "tokens_seen": 94896128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905215646940822,
      "loss": 3.3372,
      "theoretical_loss": 4.806685562333383,
      "tokens_seen": 94961664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 64162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0013229846954346,
      "objective/train/theoretical_loss": 4.806222425781916,
      "objective/train/tokens_used": 115487200,
      "theoretical_loss": 4.806222425781916,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905115346038114,
      "loss": 3.4672,
      "theoretical_loss": 4.806222425781916,
      "tokens_seen": 95027200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905015045135406,
      "loss": 3.4124,
      "theoretical_loss": 4.805759697887886,
      "tokens_seen": 95092736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904914744232698,
      "loss": 3.5337,
      "theoretical_loss": 4.80529737800944,
      "tokens_seen": 95158272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490481444332999,
      "loss": 3.1987,
      "theoretical_loss": 4.804835465506176,
      "tokens_seen": 95223808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904714142427281,
      "loss": 3.3264,
      "theoretical_loss": 4.8043739597391335,
      "tokens_seen": 95289344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904613841524574,
      "loss": 3.421,
      "theoretical_loss": 4.803912860070795,
      "tokens_seen": 95354880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904513540621865,
      "loss": 3.5238,
      "theoretical_loss": 4.803452165865082,
      "tokens_seen": 95420416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904413239719158,
      "loss": 3.4263,
      "theoretical_loss": 4.802991876487344,
      "tokens_seen": 95485952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904312938816449,
      "loss": 3.3975,
      "theoretical_loss": 4.802531991304359,
      "tokens_seen": 95551488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904212637913742,
      "loss": 3.5391,
      "theoretical_loss": 4.802072509684331,
      "tokens_seen": 95617024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904112337011033,
      "loss": 3.4524,
      "theoretical_loss": 4.801613430996884,
      "tokens_seen": 95682560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904012036108325,
      "loss": 3.5944,
      "theoretical_loss": 4.801154754613056,
      "tokens_seen": 95748096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903911735205617,
      "loss": 3.2372,
      "theoretical_loss": 4.800696479905296,
      "tokens_seen": 95813632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903811434302909,
      "loss": 3.1815,
      "theoretical_loss": 4.800238606247462,
      "tokens_seen": 95879168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903711133400201,
      "loss": 3.4315,
      "theoretical_loss": 4.799781133014815,
      "tokens_seen": 95944704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903610832497492,
      "loss": 3.2851,
      "theoretical_loss": 4.799324059584013,
      "tokens_seen": 96010240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903510531594784,
      "loss": 3.6757,
      "theoretical_loss": 4.798867385333111,
      "tokens_seen": 96075776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903410230692076,
      "loss": 3.3841,
      "theoretical_loss": 4.798411109641553,
      "tokens_seen": 96141312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903309929789368,
      "loss": 3.6343,
      "theoretical_loss": 4.797955231890175,
      "tokens_seen": 96206848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490320962888666,
      "loss": 3.3026,
      "theoretical_loss": 4.797499751461188,
      "tokens_seen": 96272384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903109327983952,
      "loss": 3.3258,
      "theoretical_loss": 4.797044667738188,
      "tokens_seen": 96337920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903009027081243,
      "loss": 3.4414,
      "theoretical_loss": 4.796589980106143,
      "tokens_seen": 96403456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902908726178535,
      "loss": 3.2512,
      "theoretical_loss": 4.796135687951393,
      "tokens_seen": 96468992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902808425275828,
      "loss": 3.4072,
      "theoretical_loss": 4.795681790661644,
      "tokens_seen": 96534528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902708124373119,
      "loss": 3.4574,
      "theoretical_loss": 4.795228287625964,
      "tokens_seen": 96600064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 64775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.715846538543701,
      "objective/train/theoretical_loss": 4.794775178234782,
      "objective/train/tokens_used": 117125600,
      "theoretical_loss": 4.794775178234782,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902607823470412,
      "loss": 3.622,
      "theoretical_loss": 4.794775178234782,
      "tokens_seen": 96665600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902507522567703,
      "loss": 3.3135,
      "theoretical_loss": 4.794322461879881,
      "tokens_seen": 96731136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902407221664995,
      "loss": 3.3367,
      "theoretical_loss": 4.793870137954395,
      "tokens_seen": 96796672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902306920762287,
      "loss": 3.3923,
      "theoretical_loss": 4.793418205852804,
      "tokens_seen": 96862208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902206619859579,
      "loss": 3.2944,
      "theoretical_loss": 4.7929666649709315,
      "tokens_seen": 96927744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902106318956871,
      "loss": 3.4677,
      "theoretical_loss": 4.7925155147059435,
      "tokens_seen": 96993280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902006018054163,
      "loss": 3.5354,
      "theoretical_loss": 4.792064754456337,
      "tokens_seen": 97058816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901905717151454,
      "loss": 3.4739,
      "theoretical_loss": 4.791614383621943,
      "tokens_seen": 97124352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901805416248746,
      "loss": 3.4712,
      "theoretical_loss": 4.79116440160392,
      "tokens_seen": 97189888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901705115346038,
      "loss": 3.6806,
      "theoretical_loss": 4.790714807804751,
      "tokens_seen": 97255424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490160481444333,
      "loss": 3.2497,
      "theoretical_loss": 4.790265601628235,
      "tokens_seen": 97320960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901504513540623,
      "loss": 3.3692,
      "theoretical_loss": 4.789816782479494,
      "tokens_seen": 97386496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901404212637913,
      "loss": 3.576,
      "theoretical_loss": 4.789368349764956,
      "tokens_seen": 97452032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901303911735206,
      "loss": 3.4743,
      "theoretical_loss": 4.788920302892365,
      "tokens_seen": 97517568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901203610832497,
      "loss": 3.1291,
      "theoretical_loss": 4.7884726412707606,
      "tokens_seen": 97583104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490110330992979,
      "loss": 3.5709,
      "theoretical_loss": 4.788025364310492,
      "tokens_seen": 97648640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901003009027082,
      "loss": 3.3566,
      "theoretical_loss": 4.787578471423203,
      "tokens_seen": 97714176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900902708124374,
      "loss": 3.6399,
      "theoretical_loss": 4.787131962021827,
      "tokens_seen": 97779712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900802407221665,
      "loss": 3.577,
      "theoretical_loss": 4.786685835520595,
      "tokens_seen": 97845248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900702106318957,
      "loss": 3.4237,
      "theoretical_loss": 4.786240091335019,
      "tokens_seen": 97910784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900601805416249,
      "loss": 3.3724,
      "theoretical_loss": 4.785794728881895,
      "tokens_seen": 97976320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900501504513541,
      "loss": 3.3342,
      "theoretical_loss": 4.785349747579299,
      "tokens_seen": 98041856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900401203610833,
      "loss": 3.5862,
      "theoretical_loss": 4.784905146846581,
      "tokens_seen": 98107392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900300902708124,
      "loss": 3.3465,
      "theoretical_loss": 4.784460926104362,
      "tokens_seen": 98172928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900200601805416,
      "loss": 3.3709,
      "theoretical_loss": 4.7840170847745345,
      "tokens_seen": 98238464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 65599,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.469925880432129,
      "objective/train/theoretical_loss": 4.783573622280253,
      "objective/train/tokens_used": 118764000,
      "theoretical_loss": 4.783573622280253,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900100300902708,
      "loss": 3.5123,
      "theoretical_loss": 4.783573622280253,
      "tokens_seen": 98304000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049,
      "loss": 3.1796,
      "theoretical_loss": 4.7831305380459295,
      "tokens_seen": 98369536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899899699097292,
      "loss": 3.5071,
      "theoretical_loss": 4.78268783149724,
      "tokens_seen": 98435072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899799398194583,
      "loss": 3.1618,
      "theoretical_loss": 4.782245502061111,
      "tokens_seen": 98500608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899699097291876,
      "loss": 3.5861,
      "theoretical_loss": 4.781803549165718,
      "tokens_seen": 98566144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899598796389167,
      "loss": 3.7046,
      "theoretical_loss": 4.781361972240484,
      "tokens_seen": 98631680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489949849548646,
      "loss": 3.4213,
      "theoretical_loss": 4.780920770716076,
      "tokens_seen": 98697216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899398194583751,
      "loss": 3.5335,
      "theoretical_loss": 4.780479944024398,
      "tokens_seen": 98762752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899297893681044,
      "loss": 3.3783,
      "theoretical_loss": 4.7800394915985915,
      "tokens_seen": 98828288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899197592778335,
      "loss": 3.5404,
      "theoretical_loss": 4.779599412873031,
      "tokens_seen": 98893824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899097291875627,
      "loss": 3.4117,
      "theoretical_loss": 4.77915970728332,
      "tokens_seen": 98959360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898996990972919,
      "loss": 3.1311,
      "theoretical_loss": 4.7787203742662845,
      "tokens_seen": 99024896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898896690070211,
      "loss": 3.3778,
      "theoretical_loss": 4.778281413259977,
      "tokens_seen": 99090432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898796389167503,
      "loss": 3.3097,
      "theoretical_loss": 4.777842823703665,
      "tokens_seen": 99155968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898696088264794,
      "loss": 3.4696,
      "theoretical_loss": 4.777404605037833,
      "tokens_seen": 99221504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898595787362086,
      "loss": 3.5204,
      "theoretical_loss": 4.7769667567041765,
      "tokens_seen": 99287040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898495486459378,
      "loss": 3.5602,
      "theoretical_loss": 4.776529278145597,
      "tokens_seen": 99352576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489839518555667,
      "loss": 3.4991,
      "theoretical_loss": 4.776092168806207,
      "tokens_seen": 99418112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898294884653962,
      "loss": 3.6413,
      "theoretical_loss": 4.775655428131313,
      "tokens_seen": 99483648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898194583751254,
      "loss": 3.6154,
      "theoretical_loss": 4.775219055567424,
      "tokens_seen": 99549184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898094282848545,
      "loss": 3.4266,
      "theoretical_loss": 4.7747830505622435,
      "tokens_seen": 99614720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897993981945837,
      "loss": 3.2347,
      "theoretical_loss": 4.774347412564666,
      "tokens_seen": 99680256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489789368104313,
      "loss": 3.5374,
      "theoretical_loss": 4.773912141024771,
      "tokens_seen": 99745792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897793380140421,
      "loss": 3.3582,
      "theoretical_loss": 4.773477235393826,
      "tokens_seen": 99811328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897693079237714,
      "loss": 3.4578,
      "theoretical_loss": 4.773042695124279,
      "tokens_seen": 99876864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 66410,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 4.062344551086426,
      "objective/train/theoretical_loss": 4.772608519669756,
      "objective/train/tokens_used": 120402400,
      "theoretical_loss": 4.772608519669756,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897592778335005,
      "loss": 3.5402,
      "theoretical_loss": 4.772608519669756,
      "tokens_seen": 99942400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897492477432297,
      "loss": 3.5992,
      "theoretical_loss": 4.772174708485057,
      "tokens_seen": 100007936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897392176529589,
      "loss": 3.4125,
      "theoretical_loss": 4.771741261026155,
      "tokens_seen": 100073472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897291875626881,
      "loss": 3.3972,
      "theoretical_loss": 4.771308176750187,
      "tokens_seen": 100139008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897191574724173,
      "loss": 3.7219,
      "theoretical_loss": 4.77087545511546,
      "tokens_seen": 100204544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897091273821465,
      "loss": 3.5976,
      "theoretical_loss": 4.77044309558144,
      "tokens_seen": 100270080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896990972918756,
      "loss": 3.5781,
      "theoretical_loss": 4.7700110976087515,
      "tokens_seen": 100335616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896890672016048,
      "loss": 3.4885,
      "theoretical_loss": 4.769579460659173,
      "tokens_seen": 100401152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489679037111334,
      "loss": 3.6787,
      "theoretical_loss": 4.769148184195636,
      "tokens_seen": 100466688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896690070210632,
      "loss": 3.5478,
      "theoretical_loss": 4.768717267682223,
      "tokens_seen": 100532224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896589769307924,
      "loss": 3.4238,
      "theoretical_loss": 4.768286710584157,
      "tokens_seen": 100597760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896489468405215,
      "loss": 3.3217,
      "theoretical_loss": 4.767856512367805,
      "tokens_seen": 100663296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896389167502507,
      "loss": 3.3783,
      "theoretical_loss": 4.767426672500677,
      "tokens_seen": 100728832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896288866599799,
      "loss": 3.2931,
      "theoretical_loss": 4.766997190451413,
      "tokens_seen": 100794368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896188565697091,
      "loss": 3.6434,
      "theoretical_loss": 4.766568065689789,
      "tokens_seen": 100859904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896088264794384,
      "loss": 3.5874,
      "theoretical_loss": 4.766139297686712,
      "tokens_seen": 100925440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895987963891674,
      "loss": 3.4533,
      "theoretical_loss": 4.76571088591421,
      "tokens_seen": 100990976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895887662988967,
      "loss": 3.6816,
      "theoretical_loss": 4.7652828298454395,
      "tokens_seen": 101056512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895787362086259,
      "loss": 3.3644,
      "theoretical_loss": 4.764855128954675,
      "tokens_seen": 101122048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895687061183551,
      "loss": 3.2458,
      "theoretical_loss": 4.764427782717309,
      "tokens_seen": 101187584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895586760280843,
      "loss": 3.3975,
      "theoretical_loss": 4.764000790609848,
      "tokens_seen": 101253120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895486459378135,
      "loss": 3.3918,
      "theoretical_loss": 4.763574152109907,
      "tokens_seen": 101318656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895386158475426,
      "loss": 3.4413,
      "theoretical_loss": 4.763147866696213,
      "tokens_seen": 101384192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895285857572718,
      "loss": 3.154,
      "theoretical_loss": 4.762721933848593,
      "tokens_seen": 101449728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489518555667001,
      "loss": 3.367,
      "theoretical_loss": 4.762296353047979,
      "tokens_seen": 101515264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 66745,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.258596897125244,
      "objective/train/theoretical_loss": 4.7618711237764035,
      "objective/train/tokens_used": 122040800,
      "theoretical_loss": 4.7618711237764035,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895085255767302,
      "loss": 3.3147,
      "theoretical_loss": 4.7618711237764035,
      "tokens_seen": 101580800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894984954864594,
      "loss": 3.2477,
      "theoretical_loss": 4.761446245516988,
      "tokens_seen": 101646336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894884653961885,
      "loss": 3.5315,
      "theoretical_loss": 4.7610217177539536,
      "tokens_seen": 101711872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894784353059178,
      "loss": 3.3191,
      "theoretical_loss": 4.760597539972606,
      "tokens_seen": 101777408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894684052156469,
      "loss": 3.0442,
      "theoretical_loss": 4.760173711659341,
      "tokens_seen": 101842944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894583751253762,
      "loss": 3.4036,
      "theoretical_loss": 4.759750232301635,
      "tokens_seen": 101908480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894483450351053,
      "loss": 3.3851,
      "theoretical_loss": 4.759327101388048,
      "tokens_seen": 101974016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894383149448346,
      "loss": 3.5575,
      "theoretical_loss": 4.7589043184082165,
      "tokens_seen": 102039552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894282848545637,
      "loss": 3.3678,
      "theoretical_loss": 4.75848188285285,
      "tokens_seen": 102105088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894182547642929,
      "loss": 3.5807,
      "theoretical_loss": 4.7580597942137315,
      "tokens_seen": 102170624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894082246740221,
      "loss": 3.4755,
      "theoretical_loss": 4.757638051983713,
      "tokens_seen": 102236160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893981945837513,
      "loss": 3.308,
      "theoretical_loss": 4.757216655656713,
      "tokens_seen": 102301696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893881644934805,
      "loss": 3.4312,
      "theoretical_loss": 4.756795604727708,
      "tokens_seen": 102367232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893781344032096,
      "loss": 3.4421,
      "theoretical_loss": 4.7563748986927425,
      "tokens_seen": 102432768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893681043129388,
      "loss": 3.3071,
      "theoretical_loss": 4.755954537048913,
      "tokens_seen": 102498304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489358074222668,
      "loss": 3.7184,
      "theoretical_loss": 4.755534519294368,
      "tokens_seen": 102563840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893480441323972,
      "loss": 3.6644,
      "theoretical_loss": 4.7551148449283165,
      "tokens_seen": 102629376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893380140421264,
      "loss": 3.3233,
      "theoretical_loss": 4.754695513451004,
      "tokens_seen": 102694912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893279839518556,
      "loss": 3.6128,
      "theoretical_loss": 4.754276524363733,
      "tokens_seen": 102760448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893179538615847,
      "loss": 3.4243,
      "theoretical_loss": 4.75385787716884,
      "tokens_seen": 102825984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893079237713139,
      "loss": 3.6836,
      "theoretical_loss": 4.7534395713697055,
      "tokens_seen": 102891520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892978936810432,
      "loss": 3.6634,
      "theoretical_loss": 4.753021606470748,
      "tokens_seen": 102957056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892878635907723,
      "loss": 3.3886,
      "theoretical_loss": 4.752603981977417,
      "tokens_seen": 103022592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892778335005016,
      "loss": 3.4605,
      "theoretical_loss": 4.752186697396198,
      "tokens_seen": 103088128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892678034102307,
      "loss": 3.3429,
      "theoretical_loss": 4.7517697522345985,
      "tokens_seen": 103153664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 68050,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.391291379928589,
      "objective/train/theoretical_loss": 4.7513531460011595,
      "objective/train/tokens_used": 123679200,
      "theoretical_loss": 4.7513531460011595,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892577733199599,
      "loss": 3.488,
      "theoretical_loss": 4.7513531460011595,
      "tokens_seen": 103219200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892477432296891,
      "loss": 3.4412,
      "theoretical_loss": 4.750936878205438,
      "tokens_seen": 103284736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892377131394183,
      "loss": 3.1551,
      "theoretical_loss": 4.750520948358019,
      "tokens_seen": 103350272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892276830491475,
      "loss": 3.3288,
      "theoretical_loss": 4.750105355970497,
      "tokens_seen": 103415808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892176529588767,
      "loss": 3.3249,
      "theoretical_loss": 4.749690100555487,
      "tokens_seen": 103481344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892076228686058,
      "loss": 3.4856,
      "theoretical_loss": 4.749275181626612,
      "tokens_seen": 103546880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489197592778335,
      "loss": 3.447,
      "theoretical_loss": 4.748860598698509,
      "tokens_seen": 103612416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891875626880642,
      "loss": 3.3712,
      "theoretical_loss": 4.748446351286817,
      "tokens_seen": 103677952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891775325977934,
      "loss": 3.3355,
      "theoretical_loss": 4.748032438908181,
      "tokens_seen": 103743488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891675025075226,
      "loss": 3.6823,
      "theoretical_loss": 4.747618861080247,
      "tokens_seen": 103809024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891574724172517,
      "loss": 3.5451,
      "theoretical_loss": 4.74720561732166,
      "tokens_seen": 103874560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891474423269809,
      "loss": 3.3377,
      "theoretical_loss": 4.746792707152057,
      "tokens_seen": 103940096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891374122367101,
      "loss": 3.2533,
      "theoretical_loss": 4.746380130092074,
      "tokens_seen": 104005632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891273821464393,
      "loss": 3.5114,
      "theoretical_loss": 4.745967885663333,
      "tokens_seen": 104071168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891173520561686,
      "loss": 3.5891,
      "theoretical_loss": 4.745555973388447,
      "tokens_seen": 104136704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891073219658976,
      "loss": 3.5146,
      "theoretical_loss": 4.745144392791008,
      "tokens_seen": 104202240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890972918756269,
      "loss": 3.477,
      "theoretical_loss": 4.744733143395598,
      "tokens_seen": 104267776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890872617853561,
      "loss": 3.5138,
      "theoretical_loss": 4.744322224727773,
      "tokens_seen": 104333312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890772316950853,
      "loss": 3.4684,
      "theoretical_loss": 4.743911636314069,
      "tokens_seen": 104398848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890672016048145,
      "loss": 3.3884,
      "theoretical_loss": 4.743501377681995,
      "tokens_seen": 104464384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890571715145437,
      "loss": 3.2946,
      "theoretical_loss": 4.743091448360034,
      "tokens_seen": 104529920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890471414242728,
      "loss": 3.362,
      "theoretical_loss": 4.742681847877636,
      "tokens_seen": 104595456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489037111334002,
      "loss": 3.1456,
      "theoretical_loss": 4.742272575765218,
      "tokens_seen": 104660992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890270812437312,
      "loss": 3.6056,
      "theoretical_loss": 4.741863631554163,
      "tokens_seen": 104726528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890170511534604,
      "loss": 3.5804,
      "theoretical_loss": 4.741455014776813,
      "tokens_seen": 104792064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 68417,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.796355724334717,
      "objective/train/theoretical_loss": 4.741046724966468,
      "objective/train/tokens_used": 125317600,
      "theoretical_loss": 4.741046724966468,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890070210631896,
      "loss": 3.5708,
      "theoretical_loss": 4.741046724966468,
      "tokens_seen": 104857600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889969909729187,
      "loss": 3.1131,
      "theoretical_loss": 4.74063876165739,
      "tokens_seen": 104923136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889869608826479,
      "loss": 3.4213,
      "theoretical_loss": 4.740231124384789,
      "tokens_seen": 104988672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889769307923771,
      "loss": 3.3977,
      "theoretical_loss": 4.739823812684829,
      "tokens_seen": 105054208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889669007021063,
      "loss": 3.345,
      "theoretical_loss": 4.739416826094621,
      "tokens_seen": 105119744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889568706118355,
      "loss": 3.4305,
      "theoretical_loss": 4.739010164152224,
      "tokens_seen": 105185280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889468405215647,
      "loss": 3.393,
      "theoretical_loss": 4.7386038263966395,
      "tokens_seen": 105250816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889368104312939,
      "loss": 3.4701,
      "theoretical_loss": 4.73819781236781,
      "tokens_seen": 105316352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488926780341023,
      "loss": 3.2626,
      "theoretical_loss": 4.737792121606619,
      "tokens_seen": 105381888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889167502507523,
      "loss": 3.364,
      "theoretical_loss": 4.737386753654883,
      "tokens_seen": 105447424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889067201604815,
      "loss": 3.4215,
      "theoretical_loss": 4.736981708055353,
      "tokens_seen": 105512960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888966900702107,
      "loss": 3.2655,
      "theoretical_loss": 4.7365769843517125,
      "tokens_seen": 105578496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888866599799398,
      "loss": 3.2771,
      "theoretical_loss": 4.736172582088571,
      "tokens_seen": 105644032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488876629889669,
      "loss": 3.3345,
      "theoretical_loss": 4.735768500811469,
      "tokens_seen": 105709568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888665997993982,
      "loss": 3.4764,
      "theoretical_loss": 4.735364740066865,
      "tokens_seen": 105775104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888565697091274,
      "loss": 3.7094,
      "theoretical_loss": 4.734961299402143,
      "tokens_seen": 105840640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888465396188566,
      "loss": 3.4389,
      "theoretical_loss": 4.734558178365605,
      "tokens_seen": 105906176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888365095285858,
      "loss": 3.5066,
      "theoretical_loss": 4.734155376506466,
      "tokens_seen": 105971712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888264794383149,
      "loss": 3.2004,
      "theoretical_loss": 4.733752893374863,
      "tokens_seen": 106037248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888164493480441,
      "loss": 3.9504,
      "theoretical_loss": 4.733350728521835,
      "tokens_seen": 106102784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888064192577733,
      "loss": 3.3078,
      "theoretical_loss": 4.732948881499336,
      "tokens_seen": 106168320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887963891675025,
      "loss": 3.258,
      "theoretical_loss": 4.7325473518602275,
      "tokens_seen": 106233856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887863590772317,
      "loss": 3.4263,
      "theoretical_loss": 4.732146139158271,
      "tokens_seen": 106299392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488776328986961,
      "loss": 3.2168,
      "theoretical_loss": 4.731745242948134,
      "tokens_seen": 106364928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048876629889669,
      "loss": 3.7064,
      "theoretical_loss": 4.731344662785381,
      "tokens_seen": 106430464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 69666,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1660373210906982,
      "objective/train/theoretical_loss": 4.730944398226477,
      "objective/train/tokens_used": 126956000,
      "theoretical_loss": 4.730944398226477,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887562688064193,
      "loss": 3.4624,
      "theoretical_loss": 4.730944398226477,
      "tokens_seen": 106496000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887462387161484,
      "loss": 3.2648,
      "theoretical_loss": 4.730544448828777,
      "tokens_seen": 106561536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887362086258777,
      "loss": 3.3377,
      "theoretical_loss": 4.730144814150533,
      "tokens_seen": 106627072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887261785356069,
      "loss": 3.182,
      "theoretical_loss": 4.729745493750885,
      "tokens_seen": 106692608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488716148445336,
      "loss": 3.5011,
      "theoretical_loss": 4.729346487189863,
      "tokens_seen": 106758144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887061183550652,
      "loss": 3.4765,
      "theoretical_loss": 4.728947794028379,
      "tokens_seen": 106823680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886960882647944,
      "loss": 3.5698,
      "theoretical_loss": 4.728549413828231,
      "tokens_seen": 106889216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886860581745236,
      "loss": 3.5244,
      "theoretical_loss": 4.728151346152098,
      "tokens_seen": 106954752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886760280842528,
      "loss": 3.2724,
      "theoretical_loss": 4.727753590563535,
      "tokens_seen": 107020288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886659979939819,
      "loss": 3.2698,
      "theoretical_loss": 4.727356146626976,
      "tokens_seen": 107085824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886559679037111,
      "loss": 3.3792,
      "theoretical_loss": 4.726959013907727,
      "tokens_seen": 107151360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886459378134403,
      "loss": 3.0486,
      "theoretical_loss": 4.7265621919719685,
      "tokens_seen": 107216896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886359077231695,
      "loss": 3.2127,
      "theoretical_loss": 4.726165680386746,
      "tokens_seen": 107282432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886258776328988,
      "loss": 3.4488,
      "theoretical_loss": 4.725769478719978,
      "tokens_seen": 107347968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886158475426278,
      "loss": 3.6296,
      "theoretical_loss": 4.725373586540442,
      "tokens_seen": 107413504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886058174523571,
      "loss": 3.1497,
      "theoretical_loss": 4.724978003417782,
      "tokens_seen": 107479040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885957873620863,
      "loss": 3.0499,
      "theoretical_loss": 4.724582728922501,
      "tokens_seen": 107544576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885857572718155,
      "loss": 3.3971,
      "theoretical_loss": 4.724187762625959,
      "tokens_seen": 107610112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885757271815447,
      "loss": 3.4614,
      "theoretical_loss": 4.723793104100376,
      "tokens_seen": 107675648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656970912739,
      "loss": 3.2867,
      "theoretical_loss": 4.72339875291882,
      "tokens_seen": 107741184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488555667001003,
      "loss": 3.5243,
      "theoretical_loss": 4.723004708655214,
      "tokens_seen": 107806720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885456369107322,
      "loss": 3.471,
      "theoretical_loss": 4.722610970884331,
      "tokens_seen": 107872256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885356068204614,
      "loss": 3.4896,
      "theoretical_loss": 4.722217539181788,
      "tokens_seen": 107937792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885255767301906,
      "loss": 3.5163,
      "theoretical_loss": 4.7218244131240485,
      "tokens_seen": 108003328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885155466399198,
      "loss": 3.1836,
      "theoretical_loss": 4.721431592288419,
      "tokens_seen": 108068864
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 70233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.18623423576355,
      "objective/train/theoretical_loss": 4.721039076253046,
      "objective/train/tokens_used": 128594400,
      "theoretical_loss": 4.721039076253046,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488505516549649,
      "loss": 3.2707,
      "theoretical_loss": 4.721039076253046,
      "tokens_seen": 108134400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884954864593781,
      "loss": 3.3744,
      "theoretical_loss": 4.720646864596914,
      "tokens_seen": 108199936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884854563691073,
      "loss": 3.2403,
      "theoretical_loss": 4.720254956899844,
      "tokens_seen": 108265472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884754262788365,
      "loss": 3.4192,
      "theoretical_loss": 4.719863352742491,
      "tokens_seen": 108331008
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884653961885657,
      "loss": 3.2221,
      "theoretical_loss": 4.719472051706341,
      "tokens_seen": 108396544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884553660982949,
      "loss": 3.4325,
      "theoretical_loss": 4.71908105337371,
      "tokens_seen": 108462080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884453360080241,
      "loss": 3.3533,
      "theoretical_loss": 4.718690357327743,
      "tokens_seen": 108527616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884353059177532,
      "loss": 3.5112,
      "theoretical_loss": 4.718299963152407,
      "tokens_seen": 108593152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884252758274825,
      "loss": 3.2922,
      "theoretical_loss": 4.7179098704324955,
      "tokens_seen": 108658688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884152457372117,
      "loss": 3.4903,
      "theoretical_loss": 4.717520078753621,
      "tokens_seen": 108724224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884052156469409,
      "loss": 3.5257,
      "theoretical_loss": 4.717130587702217,
      "tokens_seen": 108789760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048839518555667,
      "loss": 3.371,
      "theoretical_loss": 4.716741396865531,
      "tokens_seen": 108855296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883851554663992,
      "loss": 3.388,
      "theoretical_loss": 4.716352505831628,
      "tokens_seen": 108920832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883751253761284,
      "loss": 3.5684,
      "theoretical_loss": 4.715963914189384,
      "tokens_seen": 108986368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883650952858576,
      "loss": 3.3996,
      "theoretical_loss": 4.715575621528487,
      "tokens_seen": 109051904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883550651955868,
      "loss": 3.1784,
      "theoretical_loss": 4.71518762743943,
      "tokens_seen": 109117440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488345035105316,
      "loss": 3.2919,
      "theoretical_loss": 4.714799931513517,
      "tokens_seen": 109182976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883350050150451,
      "loss": 3.1558,
      "theoretical_loss": 4.714412533342855,
      "tokens_seen": 109248512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883249749247743,
      "loss": 3.5333,
      "theoretical_loss": 4.7140254325203506,
      "tokens_seen": 109314048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883149448345035,
      "loss": 3.6226,
      "theoretical_loss": 4.713638628639713,
      "tokens_seen": 109379584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883049147442327,
      "loss": 3.2969,
      "theoretical_loss": 4.713252121295449,
      "tokens_seen": 109445120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882948846539619,
      "loss": 3.6223,
      "theoretical_loss": 4.712865910082863,
      "tokens_seen": 109510656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882848545636911,
      "loss": 3.4854,
      "theoretical_loss": 4.712479994598052,
      "tokens_seen": 109576192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882748244734203,
      "loss": 3.4307,
      "theoretical_loss": 4.7120943744379025,
      "tokens_seen": 109641728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048826479438314946,
      "loss": 3.3526,
      "theoretical_loss": 4.7117090492000955,
      "tokens_seen": 109707264
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 70991,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.7265214920043945,
      "objective/train/theoretical_loss": 4.711324018483099,
      "objective/train/tokens_used": 130232800,
      "theoretical_loss": 4.711324018483099,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048825476429287864,
      "loss": 3.4221,
      "theoretical_loss": 4.711324018483099,
      "tokens_seen": 109772800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882447342026078,
      "loss": 3.156,
      "theoretical_loss": 4.710939281886164,
      "tokens_seen": 109838336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488234704112337,
      "loss": 3.5109,
      "theoretical_loss": 4.71055483900933,
      "tokens_seen": 109903872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822467402206624,
      "loss": 3.6234,
      "theoretical_loss": 4.710170689453415,
      "tokens_seen": 109969408
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048821464393179536,
      "loss": 3.3112,
      "theoretical_loss": 4.709786832820017,
      "tokens_seen": 110034944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882046138415246,
      "loss": 3.3367,
      "theoretical_loss": 4.7094032687115135,
      "tokens_seen": 110100480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881945837512537,
      "loss": 3.3711,
      "theoretical_loss": 4.709019996731058,
      "tokens_seen": 110166016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818455366098296,
      "loss": 3.1866,
      "theoretical_loss": 4.708637016482579,
      "tokens_seen": 110231552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048817452357071214,
      "loss": 3.3009,
      "theoretical_loss": 4.708254327570774,
      "tokens_seen": 110297088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881644934804413,
      "loss": 3.6798,
      "theoretical_loss": 4.7078719296011124,
      "tokens_seen": 110362624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881544633901705,
      "loss": 3.5188,
      "theoretical_loss": 4.7074898221798325,
      "tokens_seen": 110428160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048814443329989974,
      "loss": 3.1,
      "theoretical_loss": 4.707108004913936,
      "tokens_seen": 110493696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048813440320962887,
      "loss": 3.5579,
      "theoretical_loss": 4.706726477411193,
      "tokens_seen": 110559232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881243731193581,
      "loss": 3.3834,
      "theoretical_loss": 4.706345239280133,
      "tokens_seen": 110624768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048811434302908723,
      "loss": 3.3456,
      "theoretical_loss": 4.705964290130046,
      "tokens_seen": 110690304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048810431293881646,
      "loss": 3.3642,
      "theoretical_loss": 4.705583629570983,
      "tokens_seen": 110755840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048809428284854564,
      "loss": 3.2147,
      "theoretical_loss": 4.705203257213746,
      "tokens_seen": 110821376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880842527582748,
      "loss": 3.4359,
      "theoretical_loss": 4.7048231726699,
      "tokens_seen": 110886912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488074222668004,
      "loss": 3.163,
      "theoretical_loss": 4.7044433755517545,
      "tokens_seen": 110952448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880641925777332,
      "loss": 3.2166,
      "theoretical_loss": 4.704063865472374,
      "tokens_seen": 111017984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048805416248746237,
      "loss": 3.2815,
      "theoretical_loss": 4.703684642045571,
      "tokens_seen": 111083520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880441323971916,
      "loss": 3.3855,
      "theoretical_loss": 4.703305704885906,
      "tokens_seen": 111149056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048803410230692073,
      "loss": 3.476,
      "theoretical_loss": 4.7029270536086845,
      "tokens_seen": 111214592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802407221664997,
      "loss": 3.1811,
      "theoretical_loss": 4.702548687829953,
      "tokens_seen": 111280128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048801404212637915,
      "loss": 3.212,
      "theoretical_loss": 4.702170607166503,
      "tokens_seen": 111345664
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 71566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1423094272613525,
      "objective/train/theoretical_loss": 4.701792811235861,
      "objective/train/tokens_used": 131871200,
      "theoretical_loss": 4.701792811235861,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048800401203610833,
      "loss": 3.3106,
      "theoretical_loss": 4.701792811235861,
      "tokens_seen": 111411200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879939819458375,
      "loss": 3.3804,
      "theoretical_loss": 4.701415299656295,
      "tokens_seen": 111476736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879839518555667,
      "loss": 3.3137,
      "theoretical_loss": 4.701038072046808,
      "tokens_seen": 111542272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879739217652959,
      "loss": 3.4184,
      "theoretical_loss": 4.7006611280271375,
      "tokens_seen": 111607808
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879638916750251,
      "loss": 3.5134,
      "theoretical_loss": 4.700284467217749,
      "tokens_seen": 111673344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879538615847543,
      "loss": 3.515,
      "theoretical_loss": 4.699908089239845,
      "tokens_seen": 111738880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048794383149448347,
      "loss": 3.4391,
      "theoretical_loss": 4.6995319937153495,
      "tokens_seen": 111804416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048793380140421265,
      "loss": 3.4928,
      "theoretical_loss": 4.699156180266918,
      "tokens_seen": 111869952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048792377131394183,
      "loss": 3.4204,
      "theoretical_loss": 4.69878064851793,
      "tokens_seen": 111935488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048791374122367107,
      "loss": 3.2864,
      "theoretical_loss": 4.698405398092486,
      "tokens_seen": 112001024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879037111334002,
      "loss": 3.52,
      "theoretical_loss": 4.698030428615411,
      "tokens_seen": 112066560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048789368104312943,
      "loss": 3.2593,
      "theoretical_loss": 4.6976557397122445,
      "tokens_seen": 112132096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048788365095285856,
      "loss": 3.3313,
      "theoretical_loss": 4.697281331009249,
      "tokens_seen": 112197632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878736208625878,
      "loss": 3.1676,
      "theoretical_loss": 4.6969072021333975,
      "tokens_seen": 112263168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048786359077231697,
      "loss": 3.4483,
      "theoretical_loss": 4.696533352712381,
      "tokens_seen": 112328704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048785356068204615,
      "loss": 3.3785,
      "theoretical_loss": 4.696159782374602,
      "tokens_seen": 112394240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048784353059177533,
      "loss": 3.0219,
      "theoretical_loss": 4.695786490749171,
      "tokens_seen": 112459776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048783350050150457,
      "loss": 3.3436,
      "theoretical_loss": 4.6954134774659115,
      "tokens_seen": 112525312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878234704112337,
      "loss": 3.3403,
      "theoretical_loss": 4.695040742155349,
      "tokens_seen": 112590848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048781344032096293,
      "loss": 3.4199,
      "theoretical_loss": 4.694668284448717,
      "tokens_seen": 112656384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048780341023069206,
      "loss": 3.0947,
      "theoretical_loss": 4.694296103977953,
      "tokens_seen": 112721920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877933801404213,
      "loss": 2.9975,
      "theoretical_loss": 4.693924200375694,
      "tokens_seen": 112787456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877833500501505,
      "loss": 3.2263,
      "theoretical_loss": 4.693552573275277,
      "tokens_seen": 112852992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048777331995987966,
      "loss": 3.3911,
      "theoretical_loss": 4.69318122231074,
      "tokens_seen": 112918528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048776328986960884,
      "loss": 3.3579,
      "theoretical_loss": 4.692810147116814,
      "tokens_seen": 112984064
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 72808,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.543510913848877,
      "objective/train/theoretical_loss": 4.692439347328926,
      "objective/train/tokens_used": 133509600,
      "theoretical_loss": 4.692439347328926,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487753259779338,
      "loss": 3.5704,
      "theoretical_loss": 4.692439347328926,
      "tokens_seen": 113049600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877432296890672,
      "loss": 3.2623,
      "theoretical_loss": 4.692068822583197,
      "tokens_seen": 113115136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048773319959879644,
      "loss": 3.7086,
      "theoretical_loss": 4.6916985725164375,
      "tokens_seen": 113180672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048772316950852556,
      "loss": 3.2168,
      "theoretical_loss": 4.691328596766148,
      "tokens_seen": 113246208
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877131394182548,
      "loss": 3.5721,
      "theoretical_loss": 4.690958894970518,
      "tokens_seen": 113311744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877031093279839,
      "loss": 3.5405,
      "theoretical_loss": 4.690589466768422,
      "tokens_seen": 113377280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048769307923771316,
      "loss": 3.4867,
      "theoretical_loss": 4.69022031179942,
      "tokens_seen": 113442816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048768304914744234,
      "loss": 3.3823,
      "theoretical_loss": 4.689851429703753,
      "tokens_seen": 113508352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876730190571715,
      "loss": 3.242,
      "theoretical_loss": 4.689482820122342,
      "tokens_seen": 113573888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876629889669007,
      "loss": 3.6015,
      "theoretical_loss": 4.689114482696793,
      "tokens_seen": 113639424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048765295887662994,
      "loss": 3.4168,
      "theoretical_loss": 4.6887464170693836,
      "tokens_seen": 113704960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048764292878635907,
      "loss": 3.4562,
      "theoretical_loss": 4.6883786228830715,
      "tokens_seen": 113770496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876328986960883,
      "loss": 3.1948,
      "theoretical_loss": 4.688011099781486,
      "tokens_seen": 113836032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048762286860581743,
      "loss": 3.2999,
      "theoretical_loss": 4.687643847408929,
      "tokens_seen": 113901568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048761283851554666,
      "loss": 3.342,
      "theoretical_loss": 4.687276865410377,
      "tokens_seen": 113967104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048760280842527584,
      "loss": 3.4757,
      "theoretical_loss": 4.686910153431472,
      "tokens_seen": 114032640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487592778335005,
      "loss": 3.5307,
      "theoretical_loss": 4.686543711118526,
      "tokens_seen": 114098176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875827482447342,
      "loss": 3.2597,
      "theoretical_loss": 4.686177538118516,
      "tokens_seen": 114163712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875727181544634,
      "loss": 3.4569,
      "theoretical_loss": 4.685811634079084,
      "tokens_seen": 114229248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048756268806419257,
      "loss": 3.3531,
      "theoretical_loss": 4.685445998648532,
      "tokens_seen": 114294784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875526579739218,
      "loss": 3.4034,
      "theoretical_loss": 4.685080631475831,
      "tokens_seen": 114360320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048754262788365093,
      "loss": 3.0558,
      "theoretical_loss": 4.6847155322106016,
      "tokens_seen": 114425856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048753259779338017,
      "loss": 3.4359,
      "theoretical_loss": 4.6843507005031295,
      "tokens_seen": 114491392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048752256770310935,
      "loss": 3.4293,
      "theoretical_loss": 4.683986136004355,
      "tokens_seen": 114556928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048751253761283853,
      "loss": 3.1354,
      "theoretical_loss": 4.6836218383658705,
      "tokens_seen": 114622464
    },
    {
      "epoch": 0.03,
      "objective/train/docs_used": 73349,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8527109622955322,
      "objective/train/theoretical_loss": 4.683257807239925,
      "objective/train/tokens_used": 135148000,
      "theoretical_loss": 4.683257807239925,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875025075225677,
      "loss": 3.1073,
      "theoretical_loss": 4.683257807239925,
      "tokens_seen": 114688000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874924774322969,
      "loss": 3.096,
      "theoretical_loss": 4.682894042279417,
      "tokens_seen": 114753536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048748244734202607,
      "loss": 3.2853,
      "theoretical_loss": 4.682530543137897,
      "tokens_seen": 114819072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874724172517553,
      "loss": 3.6298,
      "theoretical_loss": 4.68216730946956,
      "tokens_seen": 114884608
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048746238716148443,
      "loss": 3.305,
      "theoretical_loss": 4.6818043409292525,
      "tokens_seen": 114950144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048745235707121367,
      "loss": 3.4946,
      "theoretical_loss": 4.681441637172463,
      "tokens_seen": 115015680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874423269809428,
      "loss": 3.3972,
      "theoretical_loss": 4.681079197855325,
      "tokens_seen": 115081216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048743229689067203,
      "loss": 3.1421,
      "theoretical_loss": 4.680717022634612,
      "tokens_seen": 115146752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874222668004012,
      "loss": 3.6661,
      "theoretical_loss": 4.680355111167742,
      "tokens_seen": 115212288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874122367101304,
      "loss": 3.2117,
      "theoretical_loss": 4.679993463112769,
      "tokens_seen": 115277824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874022066198596,
      "loss": 3.4083,
      "theoretical_loss": 4.679632078128385,
      "tokens_seen": 115343360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048739217652958876,
      "loss": 3.2653,
      "theoretical_loss": 4.679270955873918,
      "tokens_seen": 115408896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048738214643931794,
      "loss": 3.4833,
      "theoretical_loss": 4.678910096009329,
      "tokens_seen": 115474432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048737211634904717,
      "loss": 3.3117,
      "theoretical_loss": 4.678549498195213,
      "tokens_seen": 115539968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873620862587763,
      "loss": 3.3255,
      "theoretical_loss": 4.678189162092798,
      "tokens_seen": 115605504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048735205616850553,
      "loss": 3.2552,
      "theoretical_loss": 4.677829087363937,
      "tokens_seen": 115671040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873420260782347,
      "loss": 3.2912,
      "theoretical_loss": 4.677469273671118,
      "tokens_seen": 115736576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873319959879639,
      "loss": 3.1695,
      "theoretical_loss": 4.677109720677448,
      "tokens_seen": 115802112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873219658976931,
      "loss": 3.2226,
      "theoretical_loss": 4.676750428046665,
      "tokens_seen": 115867648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048731193580742226,
      "loss": 3.2803,
      "theoretical_loss": 4.676391395443128,
      "tokens_seen": 115933184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048730190571715144,
      "loss": 3.5568,
      "theoretical_loss": 4.6760326225318165,
      "tokens_seen": 115998720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872918756268807,
      "loss": 3.339,
      "theoretical_loss": 4.675674108978335,
      "tokens_seen": 116064256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872818455366098,
      "loss": 3.1442,
      "theoretical_loss": 4.675315854448903,
      "tokens_seen": 116129792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048727181544633904,
      "loss": 3.1936,
      "theoretical_loss": 4.674957858610361,
      "tokens_seen": 116195328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048726178535606816,
      "loss": 3.2333,
      "theoretical_loss": 4.674600121130163,
      "tokens_seen": 116260864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 74658,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9736719131469727,
      "objective/train/theoretical_loss": 4.674242641676377,
      "objective/train/tokens_used": 136786400,
      "theoretical_loss": 4.674242641676377,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872517552657974,
      "loss": 3.3693,
      "theoretical_loss": 4.674242641676377,
      "tokens_seen": 116326400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872417251755266,
      "loss": 3.3527,
      "theoretical_loss": 4.673885419917689,
      "tokens_seen": 116391936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048723169508525576,
      "loss": 3.5978,
      "theoretical_loss": 4.673528455523392,
      "tokens_seen": 116457472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487221664994985,
      "loss": 3.416,
      "theoretical_loss": 4.673171748163389,
      "tokens_seen": 116523008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872116349047141,
      "loss": 3.3723,
      "theoretical_loss": 4.672815297508194,
      "tokens_seen": 116588544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048720160481444336,
      "loss": 3.1412,
      "theoretical_loss": 4.672459103228929,
      "tokens_seen": 116654080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048719157472417254,
      "loss": 3.6399,
      "theoretical_loss": 4.672103164997319,
      "tokens_seen": 116719616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871815446339017,
      "loss": 3.3651,
      "theoretical_loss": 4.671747482485696,
      "tokens_seen": 116785152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871715145436309,
      "loss": 3.3461,
      "theoretical_loss": 4.671392055366992,
      "tokens_seen": 116850688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048716148445336014,
      "loss": 3.2057,
      "theoretical_loss": 4.671036883314743,
      "tokens_seen": 116916224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048715145436308927,
      "loss": 3.1564,
      "theoretical_loss": 4.670681966003084,
      "tokens_seen": 116981760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871414242728185,
      "loss": 3.3224,
      "theoretical_loss": 4.670327303106748,
      "tokens_seen": 117047296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048713139418254763,
      "loss": 3.5285,
      "theoretical_loss": 4.6699728943010665,
      "tokens_seen": 117112832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048712136409227686,
      "loss": 3.2677,
      "theoretical_loss": 4.669618739261967,
      "tokens_seen": 117178368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048711133400200604,
      "loss": 3.1055,
      "theoretical_loss": 4.669264837665969,
      "tokens_seen": 117243904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871013039117352,
      "loss": 3.3293,
      "theoretical_loss": 4.668911189190185,
      "tokens_seen": 117309440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870912738214644,
      "loss": 3.3489,
      "theoretical_loss": 4.668557793512322,
      "tokens_seen": 117374976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870812437311936,
      "loss": 3.1125,
      "theoretical_loss": 4.668204650310676,
      "tokens_seen": 117440512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048707121364092277,
      "loss": 3.2515,
      "theoretical_loss": 4.667851759264128,
      "tokens_seen": 117506048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487061183550652,
      "loss": 3.3586,
      "theoretical_loss": 4.667499120052151,
      "tokens_seen": 117571584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048705115346038113,
      "loss": 3.2803,
      "theoretical_loss": 4.6671467323548,
      "tokens_seen": 117637120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048704112337011037,
      "loss": 3.3058,
      "theoretical_loss": 4.666794595852718,
      "tokens_seen": 117702656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048703109327983955,
      "loss": 3.2752,
      "theoretical_loss": 4.666442710227129,
      "tokens_seen": 117768192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048702106318956873,
      "loss": 3.3295,
      "theoretical_loss": 4.666091075159839,
      "tokens_seen": 117833728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870110330992979,
      "loss": 3.5682,
      "theoretical_loss": 4.665739690333233,
      "tokens_seen": 117899264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 75891,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9445178508758545,
      "objective/train/theoretical_loss": 4.665388555430277,
      "objective/train/tokens_used": 138424800,
      "theoretical_loss": 4.665388555430277,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870010030090271,
      "loss": 3.4408,
      "theoretical_loss": 4.665388555430277,
      "tokens_seen": 117964800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048699097291875627,
      "loss": 3.42,
      "theoretical_loss": 4.665037670134513,
      "tokens_seen": 118030336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869809428284855,
      "loss": 3.4118,
      "theoretical_loss": 4.66468703413006,
      "tokens_seen": 118095872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048697091273821463,
      "loss": 2.8531,
      "theoretical_loss": 4.664336647101612,
      "tokens_seen": 118161408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048696088264794387,
      "loss": 3.5486,
      "theoretical_loss": 4.663986508734435,
      "tokens_seen": 118226944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486950852557673,
      "loss": 3.3137,
      "theoretical_loss": 4.663636618714368,
      "tokens_seen": 118292480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048694082246740223,
      "loss": 3.2493,
      "theoretical_loss": 4.663286976727821,
      "tokens_seen": 118358016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869307923771314,
      "loss": 3.4941,
      "theoretical_loss": 4.6629375824617725,
      "tokens_seen": 118423552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869207622868606,
      "loss": 3.1022,
      "theoretical_loss": 4.6625884356037695,
      "tokens_seen": 118489088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869107321965898,
      "loss": 3.2497,
      "theoretical_loss": 4.662239535841925,
      "tokens_seen": 118554624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048690070210631896,
      "loss": 3.465,
      "theoretical_loss": 4.661890882864919,
      "tokens_seen": 118620160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048689067201604814,
      "loss": 3.3506,
      "theoretical_loss": 4.661542476361992,
      "tokens_seen": 118685696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688064192577737,
      "loss": 3.2192,
      "theoretical_loss": 4.661194316022952,
      "tokens_seen": 118751232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868706118355065,
      "loss": 3.274,
      "theoretical_loss": 4.660846401538165,
      "tokens_seen": 118816768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686058174523573,
      "loss": 3.1325,
      "theoretical_loss": 4.6604987325985565,
      "tokens_seen": 118882304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868505516549649,
      "loss": 3.241,
      "theoretical_loss": 4.660151308895614,
      "tokens_seen": 118947840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868405215646941,
      "loss": 3.457,
      "theoretical_loss": 4.659804130121378,
      "tokens_seen": 119013376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868304914744233,
      "loss": 3.2732,
      "theoretical_loss": 4.659457195968448,
      "tokens_seen": 119078912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682046138415246,
      "loss": 3.1738,
      "theoretical_loss": 4.659110506129977,
      "tokens_seen": 119144448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048681043129388164,
      "loss": 3.2882,
      "theoretical_loss": 4.6587640602996725,
      "tokens_seen": 119209984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868004012036109,
      "loss": 3.4992,
      "theoretical_loss": 4.658417858171794,
      "tokens_seen": 119275520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048679037111334,
      "loss": 3.2217,
      "theoretical_loss": 4.6580718994411505,
      "tokens_seen": 119341056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678034102306924,
      "loss": 3.3294,
      "theoretical_loss": 4.657726183803101,
      "tokens_seen": 119406592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048677031093279836,
      "loss": 3.0235,
      "theoretical_loss": 4.657380710953554,
      "tokens_seen": 119472128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867602808425276,
      "loss": 3.2772,
      "theoretical_loss": 4.657035480588962,
      "tokens_seen": 119537664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 76546,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.836437225341797,
      "objective/train/theoretical_loss": 4.656690492406327,
      "objective/train/tokens_used": 140063200,
      "theoretical_loss": 4.656690492406327,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867502507522568,
      "loss": 3.395,
      "theoretical_loss": 4.656690492406327,
      "tokens_seen": 119603200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048674022066198596,
      "loss": 3.2332,
      "theoretical_loss": 4.6563457461031925,
      "tokens_seen": 119668736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048673019057171514,
      "loss": 3.3067,
      "theoretical_loss": 4.656001241377647,
      "tokens_seen": 119734272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867201604814443,
      "loss": 3.4375,
      "theoretical_loss": 4.655656977928318,
      "tokens_seen": 119799808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867101303911735,
      "loss": 3.3728,
      "theoretical_loss": 4.655312955454378,
      "tokens_seen": 119865344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048670010030090274,
      "loss": 3.4504,
      "theoretical_loss": 4.654969173655536,
      "tokens_seen": 119930880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048669007021063187,
      "loss": 3.168,
      "theoretical_loss": 4.654625632232037,
      "tokens_seen": 119996416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866800401203611,
      "loss": 3.1419,
      "theoretical_loss": 4.654282330884667,
      "tokens_seen": 120061952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866700100300903,
      "loss": 3.483,
      "theoretical_loss": 4.653939269314744,
      "tokens_seen": 120127488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048665997993981947,
      "loss": 3.1995,
      "theoretical_loss": 4.653596447224125,
      "tokens_seen": 120193024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048664994984954865,
      "loss": 3.0174,
      "theoretical_loss": 4.653253864315193,
      "tokens_seen": 120258560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048663991975927783,
      "loss": 3.2561,
      "theoretical_loss": 4.6529115202908695,
      "tokens_seen": 120324096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486629889669007,
      "loss": 3.3328,
      "theoretical_loss": 4.652569414854604,
      "tokens_seen": 120389632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048661985957873624,
      "loss": 3.2924,
      "theoretical_loss": 4.652227547710373,
      "tokens_seen": 120455168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048660982948846537,
      "loss": 3.2271,
      "theoretical_loss": 4.651885918562684,
      "tokens_seen": 120520704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865997993981946,
      "loss": 3.2564,
      "theoretical_loss": 4.65154452711657,
      "tokens_seen": 120586240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048658976930792373,
      "loss": 3.2625,
      "theoretical_loss": 4.651203373077593,
      "tokens_seen": 120651776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048657973921765297,
      "loss": 3.1227,
      "theoretical_loss": 4.650862456151833,
      "tokens_seen": 120717312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048656970912738215,
      "loss": 3.2546,
      "theoretical_loss": 4.650521776045898,
      "tokens_seen": 120782848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048655967903711133,
      "loss": 3.2324,
      "theoretical_loss": 4.650181332466918,
      "tokens_seen": 120848384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865496489468405,
      "loss": 3.3792,
      "theoretical_loss": 4.64984112512254,
      "tokens_seen": 120913920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048653961885656975,
      "loss": 3.4118,
      "theoretical_loss": 4.6495011537209345,
      "tokens_seen": 120979456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865295887662989,
      "loss": 3.3581,
      "theoretical_loss": 4.649161417970789,
      "tokens_seen": 121044992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865195586760281,
      "loss": 3.4152,
      "theoretical_loss": 4.648821917581308,
      "tokens_seen": 121110528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048650952858575724,
      "loss": 3.4244,
      "theoretical_loss": 4.648482652262211,
      "tokens_seen": 121176064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 77942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9632604122161865,
      "objective/train/theoretical_loss": 4.648143621723734,
      "objective/train/tokens_used": 141701600,
      "theoretical_loss": 4.648143621723734,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048649949849548647,
      "loss": 3.2822,
      "theoretical_loss": 4.648143621723734,
      "tokens_seen": 121241600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048648946840521565,
      "loss": 3.0832,
      "theoretical_loss": 4.6478048256766265,
      "tokens_seen": 121307136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048647943831494483,
      "loss": 3.3079,
      "theoretical_loss": 4.647466263832148,
      "tokens_seen": 121372672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048646940822467407,
      "loss": 3.2723,
      "theoretical_loss": 4.647127935902072,
      "tokens_seen": 121438208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864593781344032,
      "loss": 3.5748,
      "theoretical_loss": 4.646789841598681,
      "tokens_seen": 121503744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048644934804413243,
      "loss": 3.1642,
      "theoretical_loss": 4.646451980634765,
      "tokens_seen": 121569280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864393179538616,
      "loss": 3.1145,
      "theoretical_loss": 4.646114352723625,
      "tokens_seen": 121634816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864292878635908,
      "loss": 3.5071,
      "theoretical_loss": 4.645776957579065,
      "tokens_seen": 121700352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048641925777332,
      "loss": 3.4015,
      "theoretical_loss": 4.645439794915395,
      "tokens_seen": 121765888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048640922768304916,
      "loss": 3.2914,
      "theoretical_loss": 4.6451028644474315,
      "tokens_seen": 121831424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048639919759277834,
      "loss": 3.2374,
      "theoretical_loss": 4.644766165890491,
      "tokens_seen": 121896960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048638916750250757,
      "loss": 3.2349,
      "theoretical_loss": 4.644429698960394,
      "tokens_seen": 121962496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863791374122367,
      "loss": 3.4859,
      "theoretical_loss": 4.644093463373462,
      "tokens_seen": 122028032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048636910732196593,
      "loss": 3.2108,
      "theoretical_loss": 4.643757458846514,
      "tokens_seen": 122093568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863590772316951,
      "loss": 3.2694,
      "theoretical_loss": 4.643421685096866,
      "tokens_seen": 122159104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863490471414243,
      "loss": 3.525,
      "theoretical_loss": 4.643086141842337,
      "tokens_seen": 122224640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863390170511535,
      "loss": 3.4927,
      "theoretical_loss": 4.642750828801238,
      "tokens_seen": 122290176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048632898696088266,
      "loss": 3.4461,
      "theoretical_loss": 4.642415745692375,
      "tokens_seen": 122355712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048631895687061184,
      "loss": 3.3788,
      "theoretical_loss": 4.6420808922350485,
      "tokens_seen": 122421248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863089267803411,
      "loss": 3.3053,
      "theoretical_loss": 4.641746268149052,
      "tokens_seen": 122486784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862988966900702,
      "loss": 3.2416,
      "theoretical_loss": 4.64141187315467,
      "tokens_seen": 122552320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048628886659979944,
      "loss": 3.0169,
      "theoretical_loss": 4.641077706972679,
      "tokens_seen": 122617856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048627883650952857,
      "loss": 3.3138,
      "theoretical_loss": 4.6407437693243425,
      "tokens_seen": 122683392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862688064192578,
      "loss": 3.1577,
      "theoretical_loss": 4.640410059931414,
      "tokens_seen": 122748928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486258776328987,
      "loss": 3.3863,
      "theoretical_loss": 4.640076578516132,
      "tokens_seen": 122814464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 78541,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7341437339782715,
      "objective/train/theoretical_loss": 4.639743324801225,
      "objective/train/tokens_used": 143340000,
      "theoretical_loss": 4.639743324801225,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048624874623871616,
      "loss": 3.1381,
      "theoretical_loss": 4.639743324801225,
      "tokens_seen": 122880000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048623871614844534,
      "loss": 3.3316,
      "theoretical_loss": 4.639410298509901,
      "tokens_seen": 122945536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862286860581745,
      "loss": 3.2356,
      "theoretical_loss": 4.639077499365856,
      "tokens_seen": 123011072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862186559679037,
      "loss": 3.355,
      "theoretical_loss": 4.638744927093268,
      "tokens_seen": 123076608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048620862587763294,
      "loss": 3.3971,
      "theoretical_loss": 4.6384125814167945,
      "tokens_seen": 123142144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048619859578736207,
      "loss": 3.0734,
      "theoretical_loss": 4.638080462061575,
      "tokens_seen": 123207680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861885656970913,
      "loss": 3.1079,
      "theoretical_loss": 4.63774856875323,
      "tokens_seen": 123273216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861785356068205,
      "loss": 3.3234,
      "theoretical_loss": 4.637416901217852,
      "tokens_seen": 123338752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048616850551654967,
      "loss": 3.3152,
      "theoretical_loss": 4.63708545918202,
      "tokens_seen": 123404288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048615847542627885,
      "loss": 3.345,
      "theoretical_loss": 4.636754242372782,
      "tokens_seen": 123469824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048614844533600803,
      "loss": 3.3265,
      "theoretical_loss": 4.636423250517664,
      "tokens_seen": 123535360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861384152457372,
      "loss": 3.3053,
      "theoretical_loss": 4.636092483344665,
      "tokens_seen": 123600896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048612838515546644,
      "loss": 3.3878,
      "theoretical_loss": 4.635761940582256,
      "tokens_seen": 123666432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048611835506519557,
      "loss": 3.3957,
      "theoretical_loss": 4.635431621959384,
      "tokens_seen": 123731968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861083249749248,
      "loss": 3.4768,
      "theoretical_loss": 4.63510152720546,
      "tokens_seen": 123797504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048609829488465393,
      "loss": 3.2736,
      "theoretical_loss": 4.634771656050372,
      "tokens_seen": 123863040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048608826479438317,
      "loss": 3.1239,
      "theoretical_loss": 4.63444200822447,
      "tokens_seen": 123928576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048607823470411235,
      "loss": 3.2963,
      "theoretical_loss": 4.6341125834585775,
      "tokens_seen": 123994112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048606820461384153,
      "loss": 3.4747,
      "theoretical_loss": 4.633783381483978,
      "tokens_seen": 124059648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860581745235707,
      "loss": 3.374,
      "theoretical_loss": 4.633454402032427,
      "tokens_seen": 124125184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048604814443329995,
      "loss": 3.4227,
      "theoretical_loss": 4.63312564483614,
      "tokens_seen": 124190720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860381143430291,
      "loss": 3.6141,
      "theoretical_loss": 4.632797109627798,
      "tokens_seen": 124256256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860280842527583,
      "loss": 3.2804,
      "theoretical_loss": 4.632468796140543,
      "tokens_seen": 124321792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048601805416248744,
      "loss": 3.4935,
      "theoretical_loss": 4.632140704107979,
      "tokens_seen": 124387328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048600802407221667,
      "loss": 3.3112,
      "theoretical_loss": 4.631812833264169,
      "tokens_seen": 124452864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 79709,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.385240316390991,
      "objective/train/theoretical_loss": 4.631485183343639,
      "objective/train/tokens_used": 144978400,
      "theoretical_loss": 4.631485183343639,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048599799398194585,
      "loss": 3.2349,
      "theoretical_loss": 4.631485183343639,
      "tokens_seen": 124518400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048598796389167503,
      "loss": 3.3546,
      "theoretical_loss": 4.631157754081366,
      "tokens_seen": 124583936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859779338014042,
      "loss": 3.088,
      "theoretical_loss": 4.630830545212792,
      "tokens_seen": 124649472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859679037111334,
      "loss": 3.4382,
      "theoretical_loss": 4.63050355647381,
      "tokens_seen": 124715008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859578736208626,
      "loss": 3.3753,
      "theoretical_loss": 4.630176787600771,
      "tokens_seen": 124780544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859478435305918,
      "loss": 3.1749,
      "theoretical_loss": 4.6298502383304765,
      "tokens_seen": 124846080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048593781344032094,
      "loss": 3.1883,
      "theoretical_loss": 4.629523908400184,
      "tokens_seen": 124911616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859277833500502,
      "loss": 3.4012,
      "theoretical_loss": 4.6291977975476035,
      "tokens_seen": 124977152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004859177532597793,
      "loss": 3.1556,
      "theoretical_loss": 4.628871905510893,
      "tokens_seen": 125042688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048590772316950854,
      "loss": 3.5815,
      "theoretical_loss": 4.628546232028661,
      "tokens_seen": 125108224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858976930792377,
      "loss": 3.2682,
      "theoretical_loss": 4.62822077683997,
      "tokens_seen": 125173760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858876629889669,
      "loss": 3.685,
      "theoretical_loss": 4.627895539684323,
      "tokens_seen": 125239296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858776328986961,
      "loss": 3.2897,
      "theoretical_loss": 4.627570520301673,
      "tokens_seen": 125304832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858676028084253,
      "loss": 3.1734,
      "theoretical_loss": 4.627245718432423,
      "tokens_seen": 125370368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048585757271815444,
      "loss": 3.0723,
      "theoretical_loss": 4.626921133817413,
      "tokens_seen": 125435904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858475426278837,
      "loss": 3.1019,
      "theoretical_loss": 4.626596766197936,
      "tokens_seen": 125501440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858375125376128,
      "loss": 3.21,
      "theoretical_loss": 4.62627261531572,
      "tokens_seen": 125566976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048582748244734204,
      "loss": 3.1586,
      "theoretical_loss": 4.625948680912939,
      "tokens_seen": 125632512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858174523570712,
      "loss": 3.1464,
      "theoretical_loss": 4.625624962732207,
      "tokens_seen": 125698048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858074222668004,
      "loss": 3.5956,
      "theoretical_loss": 4.62530146051658,
      "tokens_seen": 125763584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857973921765296,
      "loss": 3.2982,
      "theoretical_loss": 4.624978174009549,
      "tokens_seen": 125829120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048578736208625877,
      "loss": 2.8431,
      "theoretical_loss": 4.624655102955048,
      "tokens_seen": 125894656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577733199598795,
      "loss": 3.2039,
      "theoretical_loss": 4.624332247097442,
      "tokens_seen": 125960192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857673019057172,
      "loss": 3.255,
      "theoretical_loss": 4.624009606181539,
      "tokens_seen": 126025728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857572718154463,
      "loss": 3.3476,
      "theoretical_loss": 4.623687179952577,
      "tokens_seen": 126091264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 80984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0424399375915527,
      "objective/train/theoretical_loss": 4.623364968156229,
      "objective/train/tokens_used": 146616800,
      "theoretical_loss": 4.623364968156229,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048574724172517554,
      "loss": 3.1198,
      "theoretical_loss": 4.623364968156229,
      "tokens_seen": 126156800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573721163490467,
      "loss": 3.2627,
      "theoretical_loss": 4.623042970538604,
      "tokens_seen": 126222336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857271815446339,
      "loss": 3.2619,
      "theoretical_loss": 4.622721186846241,
      "tokens_seen": 126287872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048571715145436314,
      "loss": 3.077,
      "theoretical_loss": 4.62239961682611,
      "tokens_seen": 126353408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048570712136409227,
      "loss": 3.4283,
      "theoretical_loss": 4.622078260225611,
      "tokens_seen": 126418944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856970912738215,
      "loss": 3.1504,
      "theoretical_loss": 4.621757116792576,
      "tokens_seen": 126484480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856870611835507,
      "loss": 3.2759,
      "theoretical_loss": 4.621436186275262,
      "tokens_seen": 126550016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048567703109327987,
      "loss": 3.2942,
      "theoretical_loss": 4.6211154684223565,
      "tokens_seen": 126615552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048566700100300905,
      "loss": 3.462,
      "theoretical_loss": 4.6207949629829725,
      "tokens_seen": 126681088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048565697091273823,
      "loss": 2.9686,
      "theoretical_loss": 4.620474669706647,
      "tokens_seen": 126746624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856469408224674,
      "loss": 3.3596,
      "theoretical_loss": 4.620154588343343,
      "tokens_seen": 126812160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563691073219664,
      "loss": 3.3793,
      "theoretical_loss": 4.619834718643448,
      "tokens_seen": 126877696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048562688064192577,
      "loss": 3.327,
      "theoretical_loss": 4.619515060357771,
      "tokens_seen": 126943232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485616850551655,
      "loss": 3.5938,
      "theoretical_loss": 4.619195613237542,
      "tokens_seen": 127008768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048560682046138413,
      "loss": 3.1359,
      "theoretical_loss": 4.6188763770344154,
      "tokens_seen": 127074304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559679037111337,
      "loss": 3.6939,
      "theoretical_loss": 4.618557351500461,
      "tokens_seen": 127139840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048558676028084255,
      "loss": 3.3604,
      "theoretical_loss": 4.618238536388171,
      "tokens_seen": 127205376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048557673019057173,
      "loss": 3.4903,
      "theoretical_loss": 4.617919931450454,
      "tokens_seen": 127270912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855667001003009,
      "loss": 3.3289,
      "theoretical_loss": 4.617601536440637,
      "tokens_seen": 127336448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048555667001003015,
      "loss": 3.1247,
      "theoretical_loss": 4.617283351112462,
      "tokens_seen": 127401984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855466399197593,
      "loss": 3.5764,
      "theoretical_loss": 4.616965375220088,
      "tokens_seen": 127467520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855366098294885,
      "loss": 3.419,
      "theoretical_loss": 4.616647608518086,
      "tokens_seen": 127533056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048552657973921764,
      "loss": 3.2447,
      "theoretical_loss": 4.616330050761444,
      "tokens_seen": 127598592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048551654964894687,
      "loss": 3.0136,
      "theoretical_loss": 4.616012701705559,
      "tokens_seen": 127664128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048550651955867605,
      "loss": 3.1339,
      "theoretical_loss": 4.615695561106241,
      "tokens_seen": 127729664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 81703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6864616870880127,
      "objective/train/theoretical_loss": 4.6153786287197125,
      "objective/train/tokens_used": 148255200,
      "theoretical_loss": 4.6153786287197125,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048549648946840523,
      "loss": 3.2948,
      "theoretical_loss": 4.6153786287197125,
      "tokens_seen": 127795200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854864593781344,
      "loss": 3.0792,
      "theoretical_loss": 4.615061904302603,
      "tokens_seen": 127860736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854764292878636,
      "loss": 3.1573,
      "theoretical_loss": 4.614745387611955,
      "tokens_seen": 127926272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854663991975928,
      "loss": 3.4941,
      "theoretical_loss": 4.614429078405214,
      "tokens_seen": 127991808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485456369107322,
      "loss": 3.2639,
      "theoretical_loss": 4.6141129764402375,
      "tokens_seen": 128057344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048544633901705114,
      "loss": 3.2058,
      "theoretical_loss": 4.613797081475287,
      "tokens_seen": 128122880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854363089267804,
      "loss": 3.4029,
      "theoretical_loss": 4.613481393269028,
      "tokens_seen": 128188416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854262788365095,
      "loss": 3.5993,
      "theoretical_loss": 4.613165911580533,
      "tokens_seen": 128253952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048541624874623874,
      "loss": 3.1374,
      "theoretical_loss": 4.612850636169279,
      "tokens_seen": 128319488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854062186559679,
      "loss": 3.3721,
      "theoretical_loss": 4.612535566795142,
      "tokens_seen": 128385024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853961885656971,
      "loss": 3.1656,
      "theoretical_loss": 4.612220703218403,
      "tokens_seen": 128450560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853861584754263,
      "loss": 3.4822,
      "theoretical_loss": 4.611906045199744,
      "tokens_seen": 128516096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853761283851555,
      "loss": 3.1366,
      "theoretical_loss": 4.611591592500245,
      "tokens_seen": 128581632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048536609829488464,
      "loss": 3.3858,
      "theoretical_loss": 4.611277344881389,
      "tokens_seen": 128647168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853560682046139,
      "loss": 3.1729,
      "theoretical_loss": 4.610963302105051,
      "tokens_seen": 128712704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485346038114343,
      "loss": 3.2684,
      "theoretical_loss": 4.610649463933511,
      "tokens_seen": 128778240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533600802407224,
      "loss": 3.1187,
      "theoretical_loss": 4.61033583012944,
      "tokens_seen": 128843776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853259779338014,
      "loss": 3.2702,
      "theoretical_loss": 4.610022400455909,
      "tokens_seen": 128909312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853159478435306,
      "loss": 3.1401,
      "theoretical_loss": 4.6097091746763805,
      "tokens_seen": 128974848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853059177532598,
      "loss": 3.0625,
      "theoretical_loss": 4.609396152554714,
      "tokens_seen": 129040384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048529588766298897,
      "loss": 3.4469,
      "theoretical_loss": 4.60908333385516,
      "tokens_seen": 129105920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048528585757271815,
      "loss": 3.2416,
      "theoretical_loss": 4.608770718342361,
      "tokens_seen": 129171456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852758274824474,
      "loss": 3.2241,
      "theoretical_loss": 4.6084583057813555,
      "tokens_seen": 129236992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852657973921765,
      "loss": 3.2384,
      "theoretical_loss": 4.608146095937567,
      "tokens_seen": 129302528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525576730190574,
      "loss": 3.4009,
      "theoretical_loss": 4.607834088576811,
      "tokens_seen": 129368064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 82831,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1748886108398438,
      "objective/train/theoretical_loss": 4.607522283465294,
      "objective/train/tokens_used": 149893600,
      "theoretical_loss": 4.607522283465294,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048524573721163487,
      "loss": 3.2513,
      "theoretical_loss": 4.607522283465294,
      "tokens_seen": 129433600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852357071213641,
      "loss": 3.3024,
      "theoretical_loss": 4.607210680369608,
      "tokens_seen": 129499136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852256770310933,
      "loss": 3.5175,
      "theoretical_loss": 4.606899279056734,
      "tokens_seen": 129564672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521564694082247,
      "loss": 3.1854,
      "theoretical_loss": 4.606588079294038,
      "tokens_seen": 129630208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048520561685055165,
      "loss": 3.1448,
      "theoretical_loss": 4.606277080849273,
      "tokens_seen": 129695744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851955867602809,
      "loss": 3.2784,
      "theoretical_loss": 4.605966283490575,
      "tokens_seen": 129761280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048518555667001,
      "loss": 3.055,
      "theoretical_loss": 4.6056556869864655,
      "tokens_seen": 129826816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048517552657973925,
      "loss": 3.4625,
      "theoretical_loss": 4.60534529110585,
      "tokens_seen": 129892352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851654964894684,
      "loss": 3.2055,
      "theoretical_loss": 4.605035095618014,
      "tokens_seen": 129957888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851554663991976,
      "loss": 3.3952,
      "theoretical_loss": 4.604725100292626,
      "tokens_seen": 130023424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851454363089268,
      "loss": 3.193,
      "theoretical_loss": 4.604415304899735,
      "tokens_seen": 130088960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513540621865597,
      "loss": 3.3349,
      "theoretical_loss": 4.604105709209769,
      "tokens_seen": 130154496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048512537612838515,
      "loss": 3.4078,
      "theoretical_loss": 4.603796312993536,
      "tokens_seen": 130220032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048511534603811433,
      "loss": 3.3374,
      "theoretical_loss": 4.603487116022222,
      "tokens_seen": 130285568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851053159478435,
      "loss": 3.2205,
      "theoretical_loss": 4.60317811806739,
      "tokens_seen": 130351104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048509528585757275,
      "loss": 3.3053,
      "theoretical_loss": 4.60286931890098,
      "tokens_seen": 130416640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850852557673019,
      "loss": 3.3346,
      "theoretical_loss": 4.602560718295308,
      "tokens_seen": 130482176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850752256770311,
      "loss": 3.3864,
      "theoretical_loss": 4.602252316023064,
      "tokens_seen": 130547712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048506519558676024,
      "loss": 3.4327,
      "theoretical_loss": 4.601944111857312,
      "tokens_seen": 130613248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850551654964895,
      "loss": 3.2583,
      "theoretical_loss": 4.601636105571492,
      "tokens_seen": 130678784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048504513540621866,
      "loss": 3.4605,
      "theoretical_loss": 4.601328296939414,
      "tokens_seen": 130744320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503510531594784,
      "loss": 3.2544,
      "theoretical_loss": 4.601020685735259,
      "tokens_seen": 130809856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485025075225677,
      "loss": 3.2836,
      "theoretical_loss": 4.600713271733582,
      "tokens_seen": 130875392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048501504513540625,
      "loss": 3.4586,
      "theoretical_loss": 4.600406054709307,
      "tokens_seen": 130940928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850050150451354,
      "loss": 3.3115,
      "theoretical_loss": 4.600099034437726,
      "tokens_seen": 131006464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 83204,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2750132083892822,
      "objective/train/theoretical_loss": 4.5997922106945,
      "objective/train/tokens_used": 151532000,
      "theoretical_loss": 4.5997922106945,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849949849548646,
      "loss": 3.3318,
      "theoretical_loss": 4.5997922106945,
      "tokens_seen": 131072000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849849548645938,
      "loss": 3.5144,
      "theoretical_loss": 4.5994855832556585,
      "tokens_seen": 131137536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484974924774323,
      "loss": 3.4244,
      "theoretical_loss": 4.5991791518976,
      "tokens_seen": 131203072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849648946840522,
      "loss": 3.2389,
      "theoretical_loss": 4.5988729163970845,
      "tokens_seen": 131268608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048495486459378134,
      "loss": 3.3639,
      "theoretical_loss": 4.59856687653124,
      "tokens_seen": 131334144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849448345035106,
      "loss": 3.2156,
      "theoretical_loss": 4.598261032077561,
      "tokens_seen": 131399680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849348044132397,
      "loss": 3.5008,
      "theoretical_loss": 4.597955382813903,
      "tokens_seen": 131465216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048492477432296894,
      "loss": 3.4594,
      "theoretical_loss": 4.597649928518486,
      "tokens_seen": 131530752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849147442326981,
      "loss": 3.3812,
      "theoretical_loss": 4.597344668969892,
      "tokens_seen": 131596288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849047141424273,
      "loss": 3.3369,
      "theoretical_loss": 4.597039603947064,
      "tokens_seen": 131661824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848946840521565,
      "loss": 3.2768,
      "theoretical_loss": 4.596734733229306,
      "tokens_seen": 131727360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848846539618857,
      "loss": 2.9459,
      "theoretical_loss": 4.596430056596285,
      "tokens_seen": 131792896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048487462387161484,
      "loss": 3.0683,
      "theoretical_loss": 4.596125573828021,
      "tokens_seen": 131858432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848645937813441,
      "loss": 3.3636,
      "theoretical_loss": 4.595821284704899,
      "tokens_seen": 131923968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848545636910732,
      "loss": 3.2567,
      "theoretical_loss": 4.5955171890076585,
      "tokens_seen": 131989504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048484453360080244,
      "loss": 3.5107,
      "theoretical_loss": 4.595213286517394,
      "tokens_seen": 132055040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848345035105316,
      "loss": 3.3222,
      "theoretical_loss": 4.594909577015562,
      "tokens_seen": 132120576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848244734202608,
      "loss": 3.4088,
      "theoretical_loss": 4.594606060283969,
      "tokens_seen": 132186112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048481444332999,
      "loss": 3.3058,
      "theoretical_loss": 4.594302736104778,
      "tokens_seen": 132251648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048480441323971917,
      "loss": 3.3761,
      "theoretical_loss": 4.593999604260507,
      "tokens_seen": 132317184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048479438314944835,
      "loss": 3.115,
      "theoretical_loss": 4.593696664534026,
      "tokens_seen": 132382720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847843530591776,
      "loss": 3.0734,
      "theoretical_loss": 4.593393916708557,
      "tokens_seen": 132448256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847743229689067,
      "loss": 3.4007,
      "theoretical_loss": 4.593091360567676,
      "tokens_seen": 132513792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048476429287863594,
      "loss": 3.4221,
      "theoretical_loss": 4.59278899589531,
      "tokens_seen": 132579328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048475426278836507,
      "loss": 3.3316,
      "theoretical_loss": 4.592486822475731,
      "tokens_seen": 132644864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 84183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4309723377227783,
      "objective/train/theoretical_loss": 4.592184840093566,
      "objective/train/tokens_used": 153170400,
      "theoretical_loss": 4.592184840093566,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847442326980943,
      "loss": 3.1609,
      "theoretical_loss": 4.592184840093566,
      "tokens_seen": 132710400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847342026078235,
      "loss": 3.3487,
      "theoretical_loss": 4.59188304853379,
      "tokens_seen": 132775936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048472417251755267,
      "loss": 3.2624,
      "theoretical_loss": 4.591581447581723,
      "tokens_seen": 132841472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048471414242728185,
      "loss": 3.0568,
      "theoretical_loss": 4.591280037023036,
      "tokens_seen": 132907008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847041123370111,
      "loss": 3.3944,
      "theoretical_loss": 4.590978816643744,
      "tokens_seen": 132972544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846940822467402,
      "loss": 3.375,
      "theoretical_loss": 4.590677786230208,
      "tokens_seen": 133038080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048468405215646945,
      "loss": 3.1915,
      "theoretical_loss": 4.590376945569136,
      "tokens_seen": 133103616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846740220661986,
      "loss": 3.2734,
      "theoretical_loss": 4.590076294447578,
      "tokens_seen": 133169152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846639919759278,
      "loss": 3.1604,
      "theoretical_loss": 4.589775832652929,
      "tokens_seen": 133234688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484653961885657,
      "loss": 3.1837,
      "theoretical_loss": 4.589475559972927,
      "tokens_seen": 133300224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048464393179538617,
      "loss": 3.3031,
      "theoretical_loss": 4.589175476195651,
      "tokens_seen": 133365760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048463390170511535,
      "loss": 3.1795,
      "theoretical_loss": 4.588875581109523,
      "tokens_seen": 133431296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048462387161484453,
      "loss": 3.2039,
      "theoretical_loss": 4.588575874503304,
      "tokens_seen": 133496832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846138415245737,
      "loss": 3.4852,
      "theoretical_loss": 4.588276356166096,
      "tokens_seen": 133562368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048460381143430295,
      "loss": 3.3087,
      "theoretical_loss": 4.587977025887342,
      "tokens_seen": 133627904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845937813440321,
      "loss": 3.1137,
      "theoretical_loss": 4.58767788345682,
      "tokens_seen": 133693440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845837512537613,
      "loss": 3.2587,
      "theoretical_loss": 4.5873789286646485,
      "tokens_seen": 133758976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048457372116349044,
      "loss": 3.3583,
      "theoretical_loss": 4.587080161301283,
      "tokens_seen": 133824512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845636910732197,
      "loss": 3.2645,
      "theoretical_loss": 4.586781581157513,
      "tokens_seen": 133890048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048455366098294886,
      "loss": 3.289,
      "theoretical_loss": 4.58648318802447,
      "tokens_seen": 133955584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048454363089267804,
      "loss": 2.9687,
      "theoretical_loss": 4.586184981693611,
      "tokens_seen": 134021120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845336008024072,
      "loss": 3.1314,
      "theoretical_loss": 4.585886961956737,
      "tokens_seen": 134086656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048452357071213645,
      "loss": 3.6532,
      "theoretical_loss": 4.585589128605976,
      "tokens_seen": 134152192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845135406218656,
      "loss": 3.1767,
      "theoretical_loss": 4.585291481433792,
      "tokens_seen": 134217728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845035105315948,
      "loss": 3.1699,
      "theoretical_loss": 4.584994020232981,
      "tokens_seen": 134283264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 84861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.49385142326355,
      "objective/train/theoretical_loss": 4.58469674479667,
      "objective/train/tokens_used": 154808800,
      "theoretical_loss": 4.58469674479667,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048449348044132394,
      "loss": 3.3437,
      "theoretical_loss": 4.58469674479667,
      "tokens_seen": 134348800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844834503510532,
      "loss": 3.3897,
      "theoretical_loss": 4.584399654918316,
      "tokens_seen": 134414336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048447342026078236,
      "loss": 3.2379,
      "theoretical_loss": 4.584102750391709,
      "tokens_seen": 134479872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048446339017051154,
      "loss": 3.4963,
      "theoretical_loss": 4.583806031010965,
      "tokens_seen": 134545408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844533600802407,
      "loss": 3.4515,
      "theoretical_loss": 4.5835094965705325,
      "tokens_seen": 134610944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844433299899699,
      "loss": 3.323,
      "theoretical_loss": 4.5832131468651855,
      "tokens_seen": 134676480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844332998996991,
      "loss": 3.0751,
      "theoretical_loss": 4.582916981690026,
      "tokens_seen": 134742016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844232698094283,
      "loss": 3.2973,
      "theoretical_loss": 4.582621000840484,
      "tokens_seen": 134807552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048441323971915745,
      "loss": 3.3443,
      "theoretical_loss": 4.582325204112314,
      "tokens_seen": 134873088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844032096288867,
      "loss": 3.2763,
      "theoretical_loss": 4.582029591301595,
      "tokens_seen": 134938624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048439317953861586,
      "loss": 3.3043,
      "theoretical_loss": 4.581734162204734,
      "tokens_seen": 135004160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048438314944834504,
      "loss": 3.2092,
      "theoretical_loss": 4.5814389166184615,
      "tokens_seen": 135069696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843731193580742,
      "loss": 3.2379,
      "theoretical_loss": 4.581143854339828,
      "tokens_seen": 135135232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843630892678034,
      "loss": 3.2659,
      "theoretical_loss": 4.580848975166209,
      "tokens_seen": 135200768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843530591775326,
      "loss": 3.0079,
      "theoretical_loss": 4.580554278895304,
      "tokens_seen": 135266304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843430290872618,
      "loss": 3.2396,
      "theoretical_loss": 4.58025976532513,
      "tokens_seen": 135331840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048433299899699095,
      "loss": 3.2323,
      "theoretical_loss": 4.579965434254028,
      "tokens_seen": 135397376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843229689067202,
      "loss": 3.0602,
      "theoretical_loss": 4.579671285480657,
      "tokens_seen": 135462912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843129388164493,
      "loss": 3.2404,
      "theoretical_loss": 4.579377318803996,
      "tokens_seen": 135528448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048430290872617855,
      "loss": 3.0519,
      "theoretical_loss": 4.579083534023344,
      "tokens_seen": 135593984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048429287863590773,
      "loss": 3.3365,
      "theoretical_loss": 4.578789930938315,
      "tokens_seen": 135659520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842828485456369,
      "loss": 3.2243,
      "theoretical_loss": 4.5784965093488434,
      "tokens_seen": 135725056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842728184553661,
      "loss": 3.1942,
      "theoretical_loss": 4.578203269055179,
      "tokens_seen": 135790592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048426278836509527,
      "loss": 3.4092,
      "theoretical_loss": 4.577910209857889,
      "tokens_seen": 135856128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048425275827482445,
      "loss": 3.1487,
      "theoretical_loss": 4.577617331557854,
      "tokens_seen": 135921664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 86086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4546592235565186,
      "objective/train/theoretical_loss": 4.577324633956271,
      "objective/train/tokens_used": 156447200,
      "theoretical_loss": 4.577324633956271,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842427281845537,
      "loss": 3.3413,
      "theoretical_loss": 4.577324633956271,
      "tokens_seen": 135987200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048423269809428287,
      "loss": 3.4106,
      "theoretical_loss": 4.5770321168546495,
      "tokens_seen": 136052736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048422266800401205,
      "loss": 3.4241,
      "theoretical_loss": 4.576739780054815,
      "tokens_seen": 136118272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842126379137413,
      "loss": 3.254,
      "theoretical_loss": 4.576447623358905,
      "tokens_seen": 136183808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842026078234704,
      "loss": 3.0941,
      "theoretical_loss": 4.576155646569367,
      "tokens_seen": 136249344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048419257773319965,
      "loss": 3.3943,
      "theoretical_loss": 4.575863849488963,
      "tokens_seen": 136314880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841825476429288,
      "loss": 3.2206,
      "theoretical_loss": 4.575572231920763,
      "tokens_seen": 136380416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484172517552658,
      "loss": 3.1229,
      "theoretical_loss": 4.575280793668151,
      "tokens_seen": 136445952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841624874623872,
      "loss": 2.9849,
      "theoretical_loss": 4.574989534534817,
      "tokens_seen": 136511488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048415245737211637,
      "loss": 3.2635,
      "theoretical_loss": 4.574698454324764,
      "tokens_seen": 136577024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048414242728184555,
      "loss": 3.1675,
      "theoretical_loss": 4.574407552842299,
      "tokens_seen": 136642560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048413239719157473,
      "loss": 3.5906,
      "theoretical_loss": 4.57411682989204,
      "tokens_seen": 136708096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841223671013039,
      "loss": 3.2567,
      "theoretical_loss": 4.57382628527891,
      "tokens_seen": 136773632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048411233701103315,
      "loss": 3.5297,
      "theoretical_loss": 4.573535918808141,
      "tokens_seen": 136839168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841023069207623,
      "loss": 3.3214,
      "theoretical_loss": 4.573245730285268,
      "tokens_seen": 136904704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840922768304915,
      "loss": 3.2527,
      "theoretical_loss": 4.572955719516132,
      "tokens_seen": 136970240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048408224674022064,
      "loss": 3.0917,
      "theoretical_loss": 4.572665886306881,
      "tokens_seen": 137035776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840722166499499,
      "loss": 3.0382,
      "theoretical_loss": 4.572376230463965,
      "tokens_seen": 137101312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048406218655967906,
      "loss": 3.3253,
      "theoretical_loss": 4.572086751794135,
      "tokens_seen": 137166848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048405215646940824,
      "loss": 3.4228,
      "theoretical_loss": 4.571797450104451,
      "tokens_seen": 137232384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840421263791374,
      "loss": 3.4935,
      "theoretical_loss": 4.571508325202268,
      "tokens_seen": 137297920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048403209628886665,
      "loss": 3.2935,
      "theoretical_loss": 4.5712193768952485,
      "tokens_seen": 137363456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840220661985958,
      "loss": 3.0949,
      "theoretical_loss": 4.570930604991352,
      "tokens_seen": 137428992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484012036108325,
      "loss": 3.0731,
      "theoretical_loss": 4.57064200929884,
      "tokens_seen": 137494528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048400200601805414,
      "loss": 3.2526,
      "theoretical_loss": 4.570353589626273,
      "tokens_seen": 137560064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 86654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.495647668838501,
      "objective/train/theoretical_loss": 4.570065345782512,
      "objective/train/tokens_used": 158085600,
      "theoretical_loss": 4.570065345782512,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839919759277834,
      "loss": 3.3585,
      "theoretical_loss": 4.570065345782512,
      "tokens_seen": 137625600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048398194583751256,
      "loss": 3.331,
      "theoretical_loss": 4.569777277576713,
      "tokens_seen": 137691136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397191574724174,
      "loss": 3.2636,
      "theoretical_loss": 4.569489384818334,
      "tokens_seen": 137756672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839618856569709,
      "loss": 3.3772,
      "theoretical_loss": 4.569201667317128,
      "tokens_seen": 137822208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839518555667001,
      "loss": 3.1082,
      "theoretical_loss": 4.568914124883145,
      "tokens_seen": 137887744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839418254764293,
      "loss": 3.567,
      "theoretical_loss": 4.568626757326731,
      "tokens_seen": 137953280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839317953861585,
      "loss": 3.3668,
      "theoretical_loss": 4.568339564458528,
      "tokens_seen": 138018816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048392176529588765,
      "loss": 3.2898,
      "theoretical_loss": 4.568052546089472,
      "tokens_seen": 138084352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004839117352056169,
      "loss": 3.2154,
      "theoretical_loss": 4.567765702030792,
      "tokens_seen": 138149888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048390170511534606,
      "loss": 3.3787,
      "theoretical_loss": 4.567479032094015,
      "tokens_seen": 138215424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048389167502507524,
      "loss": 3.2051,
      "theoretical_loss": 4.567192536090956,
      "tokens_seen": 138280960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838816449348044,
      "loss": 3.2675,
      "theoretical_loss": 4.566906213833725,
      "tokens_seen": 138346496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838716148445336,
      "loss": 3.4467,
      "theoretical_loss": 4.566620065134725,
      "tokens_seen": 138412032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838615847542628,
      "loss": 3.3219,
      "theoretical_loss": 4.566334089806647,
      "tokens_seen": 138477568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483851554663992,
      "loss": 3.3479,
      "theoretical_loss": 4.566048287662475,
      "tokens_seen": 138543104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048384152457372115,
      "loss": 3.3473,
      "theoretical_loss": 4.565762658515482,
      "tokens_seen": 138608640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838314944834504,
      "loss": 3.3177,
      "theoretical_loss": 4.565477202179232,
      "tokens_seen": 138674176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838214643931795,
      "loss": 3.2072,
      "theoretical_loss": 4.565191918467576,
      "tokens_seen": 138739712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381143430290875,
      "loss": 3.4187,
      "theoretical_loss": 4.564906807194655,
      "tokens_seen": 138805248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048380140421263793,
      "loss": 2.8605,
      "theoretical_loss": 4.564621868174897,
      "tokens_seen": 138870784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837913741223671,
      "loss": 3.0615,
      "theoretical_loss": 4.564337101223017,
      "tokens_seen": 138936320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837813440320963,
      "loss": 3.3648,
      "theoretical_loss": 4.564052506154015,
      "tokens_seen": 139001856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377131394182547,
      "loss": 3.051,
      "theoretical_loss": 4.563768082783182,
      "tokens_seen": 139067392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048376128385155465,
      "loss": 3.1675,
      "theoretical_loss": 4.563483830926089,
      "tokens_seen": 139132928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837512537612839,
      "loss": 3.1113,
      "theoretical_loss": 4.563199750398592,
      "tokens_seen": 139198464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 87912,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.470696210861206,
      "objective/train/theoretical_loss": 4.562915841016838,
      "objective/train/tokens_used": 159724000,
      "theoretical_loss": 4.562915841016838,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483741223671013,
      "loss": 3.5703,
      "theoretical_loss": 4.562915841016838,
      "tokens_seen": 139264000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373119358074225,
      "loss": 3.2538,
      "theoretical_loss": 4.5626321025972505,
      "tokens_seen": 139329536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048372116349047143,
      "loss": 3.2385,
      "theoretical_loss": 4.562348534956538,
      "tokens_seen": 139395072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837111334002006,
      "loss": 3.4436,
      "theoretical_loss": 4.562065137911691,
      "tokens_seen": 139460608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837011033099298,
      "loss": 3.1403,
      "theoretical_loss": 4.561781911279987,
      "tokens_seen": 139526144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483691073219659,
      "loss": 3.1669,
      "theoretical_loss": 4.5614988548789785,
      "tokens_seen": 139591680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048368104312938816,
      "loss": 3.1656,
      "theoretical_loss": 4.561215968526502,
      "tokens_seen": 139657216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836710130391174,
      "loss": 3.2068,
      "theoretical_loss": 4.560933252040675,
      "tokens_seen": 139722752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836609829488465,
      "loss": 3.2876,
      "theoretical_loss": 4.56065070523989,
      "tokens_seen": 139788288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048365095285857575,
      "loss": 3.7268,
      "theoretical_loss": 4.560368327942827,
      "tokens_seen": 139853824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836409227683049,
      "loss": 3.0936,
      "theoretical_loss": 4.560086119968435,
      "tokens_seen": 139919360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836308926780341,
      "loss": 3.362,
      "theoretical_loss": 4.559804081135949,
      "tokens_seen": 139984896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836208625877633,
      "loss": 3.4023,
      "theoretical_loss": 4.559522211264877,
      "tokens_seen": 140050432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836108324974925,
      "loss": 3.299,
      "theoretical_loss": 4.559240510175006,
      "tokens_seen": 140115968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048360080240722166,
      "loss": 3.5194,
      "theoretical_loss": 4.558958977686398,
      "tokens_seen": 140181504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359077231695084,
      "loss": 3.1333,
      "theoretical_loss": 4.558677613619393,
      "tokens_seen": 140247040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048358074222668,
      "loss": 3.1792,
      "theoretical_loss": 4.558396417794603,
      "tokens_seen": 140312576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357071213640926,
      "loss": 3.2967,
      "theoretical_loss": 4.558115390032916,
      "tokens_seen": 140378112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835606820461384,
      "loss": 3.2453,
      "theoretical_loss": 4.557834530155497,
      "tokens_seen": 140443648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835506519558676,
      "loss": 3.2309,
      "theoretical_loss": 4.5575538379837806,
      "tokens_seen": 140509184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835406218655968,
      "loss": 3.1676,
      "theoretical_loss": 4.557273313339477,
      "tokens_seen": 140574720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483530591775326,
      "loss": 3.3045,
      "theoretical_loss": 4.556992956044569,
      "tokens_seen": 140640256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048352056168505516,
      "loss": 3.4098,
      "theoretical_loss": 4.55671276592131,
      "tokens_seen": 140705792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048351053159478434,
      "loss": 3.2187,
      "theoretical_loss": 4.556432742792225,
      "tokens_seen": 140771328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835005015045135,
      "loss": 3.2732,
      "theoretical_loss": 4.55615288648011,
      "tokens_seen": 140836864
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 88907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.168332576751709,
      "objective/train/theoretical_loss": 4.555873196808033,
      "objective/train/tokens_used": 161362400,
      "theoretical_loss": 4.555873196808033,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349047141424276,
      "loss": 3.2827,
      "theoretical_loss": 4.555873196808033,
      "tokens_seen": 140902400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048348044132397194,
      "loss": 3.0851,
      "theoretical_loss": 4.555593673599329,
      "tokens_seen": 140967936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834704112337011,
      "loss": 3.2459,
      "theoretical_loss": 4.555314316677606,
      "tokens_seen": 141033472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834603811434303,
      "loss": 3.1295,
      "theoretical_loss": 4.555035125866738,
      "tokens_seen": 141099008
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834503510531595,
      "loss": 3.2718,
      "theoretical_loss": 4.5547561009908675,
      "tokens_seen": 141164544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834403209628887,
      "loss": 3.4012,
      "theoretical_loss": 4.554477241874405,
      "tokens_seen": 141230080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343029087261785,
      "loss": 3.2815,
      "theoretical_loss": 4.554198548342029,
      "tokens_seen": 141295616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834202607823471,
      "loss": 3.3752,
      "theoretical_loss": 4.553920020218683,
      "tokens_seen": 141361152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048341023069207626,
      "loss": 3.4304,
      "theoretical_loss": 4.55364165732958,
      "tokens_seen": 141426688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048340020060180544,
      "loss": 3.06,
      "theoretical_loss": 4.553363459500193,
      "tokens_seen": 141492224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833901705115346,
      "loss": 3.4717,
      "theoretical_loss": 4.553085426556265,
      "tokens_seen": 141557760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833801404212638,
      "loss": 3.2863,
      "theoretical_loss": 4.552807558323801,
      "tokens_seen": 141623296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483370110330993,
      "loss": 3.5702,
      "theoretical_loss": 4.552529854629072,
      "tokens_seen": 141688832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833600802407222,
      "loss": 3.3102,
      "theoretical_loss": 4.552252315298611,
      "tokens_seen": 141754368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048335005015045135,
      "loss": 3.2317,
      "theoretical_loss": 4.551974940159212,
      "tokens_seen": 141819904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833400200601806,
      "loss": 3.3219,
      "theoretical_loss": 4.551697729037938,
      "tokens_seen": 141885440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833299899699097,
      "loss": 3.4492,
      "theoretical_loss": 4.551420681762107,
      "tokens_seen": 141950976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048331995987963895,
      "loss": 3.2021,
      "theoretical_loss": 4.551143798159299,
      "tokens_seen": 142016512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048330992978936813,
      "loss": 3.2159,
      "theoretical_loss": 4.550867078057361,
      "tokens_seen": 142082048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832998996990973,
      "loss": 3.2044,
      "theoretical_loss": 4.550590521284393,
      "tokens_seen": 142147584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832898696088265,
      "loss": 3.1647,
      "theoretical_loss": 4.550314127668761,
      "tokens_seen": 142213120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048327983951855567,
      "loss": 3.4173,
      "theoretical_loss": 4.550037897039086,
      "tokens_seen": 142278656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048326980942828485,
      "loss": 3.0888,
      "theoretical_loss": 4.54976182922425,
      "tokens_seen": 142344192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832597793380141,
      "loss": 3.0527,
      "theoretical_loss": 4.549485924053392,
      "tokens_seen": 142409728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832497492477432,
      "loss": 3.3659,
      "theoretical_loss": 4.54921018135591,
      "tokens_seen": 142475264
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 89401,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.601848602294922,
      "objective/train/theoretical_loss": 4.548934600961459,
      "objective/train/tokens_used": 163000800,
      "theoretical_loss": 4.548934600961459,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048323971915747245,
      "loss": 3.4687,
      "theoretical_loss": 4.548934600961459,
      "tokens_seen": 142540800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048322968906720163,
      "loss": 3.2753,
      "theoretical_loss": 4.54865918269995,
      "tokens_seen": 142606336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832196589769308,
      "loss": 3.1228,
      "theoretical_loss": 4.548383926401552,
      "tokens_seen": 142671872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048320962888666,
      "loss": 3.0876,
      "theoretical_loss": 4.548108831896688,
      "tokens_seen": 142737408
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831995987963892,
      "loss": 3.44,
      "theoretical_loss": 4.547833899016037,
      "tokens_seen": 142802944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048318956870611836,
      "loss": 3.0777,
      "theoretical_loss": 4.547559127590535,
      "tokens_seen": 142868480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831795386158476,
      "loss": 3.5091,
      "theoretical_loss": 4.547284517451367,
      "tokens_seen": 142934016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831695085255767,
      "loss": 3.3175,
      "theoretical_loss": 4.547010068429978,
      "tokens_seen": 142999552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048315947843530595,
      "loss": 3.1291,
      "theoretical_loss": 4.546735780358061,
      "tokens_seen": 143065088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831494483450351,
      "loss": 3.1034,
      "theoretical_loss": 4.546461653067565,
      "tokens_seen": 143130624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831394182547643,
      "loss": 3.0849,
      "theoretical_loss": 4.54618768639069,
      "tokens_seen": 143196160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831293881644935,
      "loss": 3.0665,
      "theoretical_loss": 4.545913880159889,
      "tokens_seen": 143261696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831193580742227,
      "loss": 3.2806,
      "theoretical_loss": 4.545640234207866,
      "tokens_seen": 143327232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048310932798395186,
      "loss": 3.3249,
      "theoretical_loss": 4.545366748367574,
      "tokens_seen": 143392768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048309929789368104,
      "loss": 3.4583,
      "theoretical_loss": 4.545093422472218,
      "tokens_seen": 143458304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830892678034102,
      "loss": 3.5278,
      "theoretical_loss": 4.544820256355253,
      "tokens_seen": 143523840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048307923771313946,
      "loss": 3.0959,
      "theoretical_loss": 4.544547249850384,
      "tokens_seen": 143589376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830692076228686,
      "loss": 3.4636,
      "theoretical_loss": 4.544274402791562,
      "tokens_seen": 143654912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830591775325978,
      "loss": 3.1348,
      "theoretical_loss": 4.544001715012989,
      "tokens_seen": 143720448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483049147442327,
      "loss": 3.3713,
      "theoretical_loss": 4.543729186349115,
      "tokens_seen": 143785984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830391173520562,
      "loss": 3.334,
      "theoretical_loss": 4.5434568166346345,
      "tokens_seen": 143851520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048302908726178536,
      "loss": 3.2556,
      "theoretical_loss": 4.543184605704493,
      "tokens_seen": 143917056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048301905717151454,
      "loss": 3.24,
      "theoretical_loss": 4.542912553393879,
      "tokens_seen": 143982592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830090270812437,
      "loss": 3.0543,
      "theoretical_loss": 4.542640659538229,
      "tokens_seen": 144048128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048299899699097296,
      "loss": 3.3306,
      "theoretical_loss": 4.542368923973226,
      "tokens_seen": 144113664
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 90537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.076249361038208,
      "objective/train/theoretical_loss": 4.542097346534794,
      "objective/train/tokens_used": 164639200,
      "theoretical_loss": 4.542097346534794,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829889669007021,
      "loss": 3.2985,
      "theoretical_loss": 4.542097346534794,
      "tokens_seen": 144179200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829789368104313,
      "loss": 3.2073,
      "theoretical_loss": 4.541825927059106,
      "tokens_seen": 144244736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048296890672016045,
      "loss": 3.3592,
      "theoretical_loss": 4.541554665382575,
      "tokens_seen": 144310272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829588766298897,
      "loss": 3.5993,
      "theoretical_loss": 4.541283561341862,
      "tokens_seen": 144375808
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294884653961887,
      "loss": 3.0916,
      "theoretical_loss": 4.541012614773868,
      "tokens_seen": 144441344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048293881644934805,
      "loss": 3.2114,
      "theoretical_loss": 4.540741825515736,
      "tokens_seen": 144506880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292878635907723,
      "loss": 3.4082,
      "theoretical_loss": 4.540471193404855,
      "tokens_seen": 144572416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048291875626880646,
      "loss": 3.1695,
      "theoretical_loss": 4.54020071827885,
      "tokens_seen": 144637952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829087261785356,
      "loss": 3.4008,
      "theoretical_loss": 4.539930399975596,
      "tokens_seen": 144703488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828986960882648,
      "loss": 3.1431,
      "theoretical_loss": 4.539660238333198,
      "tokens_seen": 144769024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048288866599799395,
      "loss": 3.3694,
      "theoretical_loss": 4.53939023319001,
      "tokens_seen": 144834560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828786359077232,
      "loss": 3.3767,
      "theoretical_loss": 4.539120384384621,
      "tokens_seen": 144900096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048286860581745237,
      "loss": 3.5769,
      "theoretical_loss": 4.538850691755861,
      "tokens_seen": 144965632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048285857572718155,
      "loss": 3.3304,
      "theoretical_loss": 4.538581155142799,
      "tokens_seen": 145031168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048284854563691073,
      "loss": 3.1185,
      "theoretical_loss": 4.538311774384743,
      "tokens_seen": 145096704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828385155466399,
      "loss": 3.4686,
      "theoretical_loss": 4.538042549321238,
      "tokens_seen": 145162240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828284854563691,
      "loss": 3.1808,
      "theoretical_loss": 4.537773479792067,
      "tokens_seen": 145227776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048281845536609833,
      "loss": 3.2259,
      "theoretical_loss": 4.537504565637248,
      "tokens_seen": 145293312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048280842527582746,
      "loss": 3.3492,
      "theoretical_loss": 4.537235806697041,
      "tokens_seen": 145358848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827983951855567,
      "loss": 3.4474,
      "theoretical_loss": 4.536967202811935,
      "tokens_seen": 145424384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827883650952858,
      "loss": 3.2921,
      "theoretical_loss": 4.536698753822661,
      "tokens_seen": 145489920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048277833500501505,
      "loss": 3.4976,
      "theoretical_loss": 4.53643045957018,
      "tokens_seen": 145555456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048276830491474423,
      "loss": 3.4497,
      "theoretical_loss": 4.536162319895695,
      "tokens_seen": 145620992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827582748244734,
      "loss": 3.4119,
      "theoretical_loss": 4.535894334640634,
      "tokens_seen": 145686528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827482447342026,
      "loss": 3.0995,
      "theoretical_loss": 4.535626503646666,
      "tokens_seen": 145752064
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 91176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9223389625549316,
      "objective/train/theoretical_loss": 4.535358826755691,
      "objective/train/tokens_used": 166277600,
      "theoretical_loss": 4.535358826755691,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048273821464393183,
      "loss": 3.2178,
      "theoretical_loss": 4.535358826755691,
      "tokens_seen": 145817600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482728184553661,
      "loss": 3.4704,
      "theoretical_loss": 4.535091303809842,
      "tokens_seen": 145883136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827181544633902,
      "loss": 3.2936,
      "theoretical_loss": 4.534823934651484,
      "tokens_seen": 145948672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827081243731194,
      "loss": 3.1424,
      "theoretical_loss": 4.5345567191232155,
      "tokens_seen": 146014208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048269809428284856,
      "loss": 3.3168,
      "theoretical_loss": 4.534289657067865,
      "tokens_seen": 146079744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826880641925778,
      "loss": 2.9011,
      "theoretical_loss": 4.5340227483284945,
      "tokens_seen": 146145280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826780341023069,
      "loss": 3.2725,
      "theoretical_loss": 4.533755992748395,
      "tokens_seen": 146210816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266800401203615,
      "loss": 3.2624,
      "theoretical_loss": 4.533489390171088,
      "tokens_seen": 146276352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826579739217653,
      "loss": 3.4202,
      "theoretical_loss": 4.533222940440325,
      "tokens_seen": 146341888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826479438314945,
      "loss": 3.1386,
      "theoretical_loss": 4.532956643400087,
      "tokens_seen": 146407424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826379137412237,
      "loss": 3.3923,
      "theoretical_loss": 4.532690498894585,
      "tokens_seen": 146472960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826278836509529,
      "loss": 3.4023,
      "theoretical_loss": 4.5324245067682565,
      "tokens_seen": 146538496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048261785356068206,
      "loss": 3.1156,
      "theoretical_loss": 4.532158666865769,
      "tokens_seen": 146604032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048260782347041124,
      "loss": 3.1851,
      "theoretical_loss": 4.5318929790320155,
      "tokens_seen": 146669568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825977933801404,
      "loss": 3.0943,
      "theoretical_loss": 4.53162744311212,
      "tokens_seen": 146735104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258776328986966,
      "loss": 3.392,
      "theoretical_loss": 4.53136205895143,
      "tokens_seen": 146800640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825777331995988,
      "loss": 3.4047,
      "theoretical_loss": 4.531096826395519,
      "tokens_seen": 146866176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482567703109328,
      "loss": 3.2556,
      "theoretical_loss": 4.5308317452901905,
      "tokens_seen": 146931712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825576730190572,
      "loss": 3.1769,
      "theoretical_loss": 4.530566815481469,
      "tokens_seen": 146997248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825476429287864,
      "loss": 3.2533,
      "theoretical_loss": 4.530302036815607,
      "tokens_seen": 147062784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048253761283851556,
      "loss": 3.3405,
      "theoretical_loss": 4.53003740913908,
      "tokens_seen": 147128320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252758274824474,
      "loss": 3.3032,
      "theoretical_loss": 4.529772932298588,
      "tokens_seen": 147193856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825175526579739,
      "loss": 3.4738,
      "theoretical_loss": 4.529508606141057,
      "tokens_seen": 147259392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048250752256770316,
      "loss": 3.2304,
      "theoretical_loss": 4.529244430513632,
      "tokens_seen": 147324928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824974924774323,
      "loss": 3.2967,
      "theoretical_loss": 4.528980405263686,
      "tokens_seen": 147390464
    },
    {
      "epoch": 0.04,
      "objective/train/docs_used": 92532,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4305601119995117,
      "objective/train/theoretical_loss": 4.528716530238812,
      "objective/train/tokens_used": 167916000,
      "theoretical_loss": 4.528716530238812,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824874623871615,
      "loss": 3.4659,
      "theoretical_loss": 4.528716530238812,
      "tokens_seen": 147456000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048247743229689065,
      "loss": 3.2645,
      "theoretical_loss": 4.528452805286824,
      "tokens_seen": 147521536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824674022066199,
      "loss": 3.2269,
      "theoretical_loss": 4.528189230255759,
      "tokens_seen": 147587072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048245737211634907,
      "loss": 3.2869,
      "theoretical_loss": 4.527925804993875,
      "tokens_seen": 147652608
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244734202607825,
      "loss": 3.3874,
      "theoretical_loss": 4.5276625293496515,
      "tokens_seen": 147718144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048243731193580743,
      "loss": 3.1961,
      "theoretical_loss": 4.527399403171788,
      "tokens_seen": 147783680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048242728184553666,
      "loss": 3.3541,
      "theoretical_loss": 4.5271364263092035,
      "tokens_seen": 147849216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824172517552658,
      "loss": 3.2284,
      "theoretical_loss": 4.526873598611038,
      "tokens_seen": 147914752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482407221664995,
      "loss": 3.1859,
      "theoretical_loss": 4.52661091992665,
      "tokens_seen": 147980288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048239719157472415,
      "loss": 3.3306,
      "theoretical_loss": 4.526348390105614,
      "tokens_seen": 148045824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823871614844534,
      "loss": 3.0245,
      "theoretical_loss": 4.526086008997728,
      "tokens_seen": 148111360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048237713139418257,
      "loss": 3.1419,
      "theoretical_loss": 4.525823776453004,
      "tokens_seen": 148176896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048236710130391175,
      "loss": 3.0672,
      "theoretical_loss": 4.525561692321673,
      "tokens_seen": 148242432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048235707121364093,
      "loss": 3.259,
      "theoretical_loss": 4.525299756454182,
      "tokens_seen": 148307968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823470411233701,
      "loss": 3.3995,
      "theoretical_loss": 4.525037968701198,
      "tokens_seen": 148373504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823370110330993,
      "loss": 3.1846,
      "theoretical_loss": 4.5247763289136,
      "tokens_seen": 148439040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048232698094282853,
      "loss": 3.3233,
      "theoretical_loss": 4.524514836942485,
      "tokens_seen": 148504576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048231695085255766,
      "loss": 3.1244,
      "theoretical_loss": 4.524253492639165,
      "tokens_seen": 148570112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823069207622869,
      "loss": 3.1294,
      "theoretical_loss": 4.523992295855169,
      "tokens_seen": 148635648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482296890672016,
      "loss": 3.4438,
      "theoretical_loss": 4.523731246442237,
      "tokens_seen": 148701184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048228686058174525,
      "loss": 3.2735,
      "theoretical_loss": 4.523470344252326,
      "tokens_seen": 148766720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048227683049147443,
      "loss": 3.2357,
      "theoretical_loss": 4.523209589137606,
      "tokens_seen": 148832256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822668004012036,
      "loss": 3.3941,
      "theoretical_loss": 4.522948980950461,
      "tokens_seen": 148897792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822567703109328,
      "loss": 3.0532,
      "theoretical_loss": 4.522688519543487,
      "tokens_seen": 148963328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224674022066203,
      "loss": 3.1432,
      "theoretical_loss": 4.5224282047694935,
      "tokens_seen": 149028864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 93022,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5269510746002197,
      "objective/train/theoretical_loss": 4.522168036481502,
      "objective/train/tokens_used": 169554400,
      "theoretical_loss": 4.522168036481502,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048223671013039116,
      "loss": 3.5101,
      "theoretical_loss": 4.522168036481502,
      "tokens_seen": 149094400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822266800401204,
      "loss": 3.2561,
      "theoretical_loss": 4.521908014532746,
      "tokens_seen": 149159936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822166499498495,
      "loss": 3.3344,
      "theoretical_loss": 4.52164813877667,
      "tokens_seen": 149225472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048220661985957876,
      "loss": 3.2209,
      "theoretical_loss": 4.52138840906693,
      "tokens_seen": 149291008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048219658976930794,
      "loss": 3.1369,
      "theoretical_loss": 4.521128825257392,
      "tokens_seen": 149356544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821865596790371,
      "loss": 3.1751,
      "theoretical_loss": 4.520869387202134,
      "tokens_seen": 149422080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821765295887663,
      "loss": 2.9797,
      "theoretical_loss": 4.52061009475544,
      "tokens_seen": 149487616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821664994984955,
      "loss": 3.196,
      "theoretical_loss": 4.52035094777181,
      "tokens_seen": 149553152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048215646940822466,
      "loss": 3.3068,
      "theoretical_loss": 4.520091946105946,
      "tokens_seen": 149618688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821464393179539,
      "loss": 3.4288,
      "theoretical_loss": 4.519833089612764,
      "tokens_seen": 149684224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482136409227683,
      "loss": 3.2192,
      "theoretical_loss": 4.519574378147384,
      "tokens_seen": 149749760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048212637913741226,
      "loss": 3.291,
      "theoretical_loss": 4.519315811565137,
      "tokens_seen": 149815296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821163490471414,
      "loss": 3.3752,
      "theoretical_loss": 4.51905738972156,
      "tokens_seen": 149880832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821063189568706,
      "loss": 3.6035,
      "theoretical_loss": 4.518799112472401,
      "tokens_seen": 149946368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820962888665998,
      "loss": 3.5239,
      "theoretical_loss": 4.518540979673606,
      "tokens_seen": 150011904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482086258776329,
      "loss": 3.1651,
      "theoretical_loss": 4.518282991181337,
      "tokens_seen": 150077440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048207622868605816,
      "loss": 3.1406,
      "theoretical_loss": 4.518025146851956,
      "tokens_seen": 150142976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820661985957874,
      "loss": 3.2538,
      "theoretical_loss": 4.517767446542033,
      "tokens_seen": 150208512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048205616850551653,
      "loss": 3.1976,
      "theoretical_loss": 4.517509890108343,
      "tokens_seen": 150274048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204613841524576,
      "loss": 3.1097,
      "theoretical_loss": 4.5172524774078635,
      "tokens_seen": 150339584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820361083249749,
      "loss": 3.3081,
      "theoretical_loss": 4.5169952082977805,
      "tokens_seen": 150405120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820260782347041,
      "loss": 3.2403,
      "theoretical_loss": 4.51673808263548,
      "tokens_seen": 150470656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820160481444333,
      "loss": 3.0648,
      "theoretical_loss": 4.516481100278555,
      "tokens_seen": 150536192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820060180541625,
      "loss": 3.1256,
      "theoretical_loss": 4.516224261084799,
      "tokens_seen": 150601728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048199598796389167,
      "loss": 3.0304,
      "theoretical_loss": 4.515967564912211,
      "tokens_seen": 150667264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 94173,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2401325702667236,
      "objective/train/theoretical_loss": 4.515711011618992,
      "objective/train/tokens_used": 171192800,
      "theoretical_loss": 4.515711011618992,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048198595787362085,
      "loss": 3.2269,
      "theoretical_loss": 4.515711011618992,
      "tokens_seen": 150732800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004819759277833501,
      "loss": 3.1091,
      "theoretical_loss": 4.515454601063542,
      "tokens_seen": 150798336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048196589769307927,
      "loss": 3.2182,
      "theoretical_loss": 4.5151983331044665,
      "tokens_seen": 150863872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048195586760280845,
      "loss": 3.2107,
      "theoretical_loss": 4.514942207600572,
      "tokens_seen": 150929408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048194583751253763,
      "loss": 3.1827,
      "theoretical_loss": 4.514686224410863,
      "tokens_seen": 150994944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048193580742226686,
      "loss": 3.3203,
      "theoretical_loss": 4.514430383394549,
      "tokens_seen": 151060480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481925777331996,
      "loss": 3.0874,
      "theoretical_loss": 4.514174684411037,
      "tokens_seen": 151126016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004819157472417252,
      "loss": 3.2352,
      "theoretical_loss": 4.513919127319935,
      "tokens_seen": 151191552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048190571715145435,
      "loss": 3.3179,
      "theoretical_loss": 4.513663711981049,
      "tokens_seen": 151257088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818956870611836,
      "loss": 3.266,
      "theoretical_loss": 4.513408438254386,
      "tokens_seen": 151322624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048188565697091277,
      "loss": 3.1682,
      "theoretical_loss": 4.5131533060001505,
      "tokens_seen": 151388160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048187562688064195,
      "loss": 3.2016,
      "theoretical_loss": 4.512898315078747,
      "tokens_seen": 151453696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048186559679037113,
      "loss": 3.2221,
      "theoretical_loss": 4.512643465350775,
      "tokens_seen": 151519232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818555667001003,
      "loss": 3.394,
      "theoretical_loss": 4.512388756677037,
      "tokens_seen": 151584768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818455366098295,
      "loss": 3.3002,
      "theoretical_loss": 4.5121341889185285,
      "tokens_seen": 151650304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048183550651955873,
      "loss": 3.131,
      "theoretical_loss": 4.511879761936442,
      "tokens_seen": 151715840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048182547642928786,
      "loss": 3.3703,
      "theoretical_loss": 4.511625475592169,
      "tokens_seen": 151781376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818154463390171,
      "loss": 3.2669,
      "theoretical_loss": 4.511371329747295,
      "tokens_seen": 151846912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818054162487462,
      "loss": 3.0424,
      "theoretical_loss": 4.511117324263603,
      "tokens_seen": 151912448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048179538615847545,
      "loss": 3.6565,
      "theoretical_loss": 4.5108634590030725,
      "tokens_seen": 151977984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048178535606820463,
      "loss": 3.411,
      "theoretical_loss": 4.510609733827873,
      "tokens_seen": 152043520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817753259779338,
      "loss": 3.1894,
      "theoretical_loss": 4.5103561486003745,
      "tokens_seen": 152109056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481765295887663,
      "loss": 3.053,
      "theoretical_loss": 4.51010270318314,
      "tokens_seen": 152174592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048175526579739223,
      "loss": 3.1504,
      "theoretical_loss": 4.509849397438925,
      "tokens_seen": 152240128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048174523570712136,
      "loss": 3.2262,
      "theoretical_loss": 4.509596231230679,
      "tokens_seen": 152305664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 94867,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.429643392562866,
      "objective/train/theoretical_loss": 4.509343204421548,
      "objective/train/tokens_used": 172831200,
      "theoretical_loss": 4.509343204421548,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817352056168506,
      "loss": 3.4209,
      "theoretical_loss": 4.509343204421548,
      "tokens_seen": 152371200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817251755265797,
      "loss": 3.2597,
      "theoretical_loss": 4.509090316874865,
      "tokens_seen": 152436736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048171514543630896,
      "loss": 3.2857,
      "theoretical_loss": 4.508837568454162,
      "tokens_seen": 152502272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048170511534603814,
      "loss": 3.5512,
      "theoretical_loss": 4.508584959023159,
      "tokens_seen": 152567808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816950852557673,
      "loss": 3.3878,
      "theoretical_loss": 4.50833248844577,
      "tokens_seen": 152633344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816850551654965,
      "loss": 3.3808,
      "theoretical_loss": 4.5080801565861,
      "tokens_seen": 152698880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816750250752257,
      "loss": 3.1163,
      "theoretical_loss": 4.507827963308447,
      "tokens_seen": 152764416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048166499498495486,
      "loss": 3.2056,
      "theoretical_loss": 4.507575908477296,
      "tokens_seen": 152829952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816549648946841,
      "loss": 3.1235,
      "theoretical_loss": 4.507323991957325,
      "tokens_seen": 152895488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816449348044132,
      "loss": 3.439,
      "theoretical_loss": 4.507072213613402,
      "tokens_seen": 152961024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048163490471414246,
      "loss": 3.2998,
      "theoretical_loss": 4.506820573310586,
      "tokens_seen": 153026560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816248746238716,
      "loss": 3.5354,
      "theoretical_loss": 4.506569070914123,
      "tokens_seen": 153092096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816148445336008,
      "loss": 3.1693,
      "theoretical_loss": 4.506317706289451,
      "tokens_seen": 153157632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048160481444333,
      "loss": 3.3531,
      "theoretical_loss": 4.506066479302193,
      "tokens_seen": 153223168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815947843530592,
      "loss": 3.3859,
      "theoretical_loss": 4.505815389818165,
      "tokens_seen": 153288704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048158475426278837,
      "loss": 3.2548,
      "theoretical_loss": 4.505564437703368,
      "tokens_seen": 153354240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815747241725176,
      "loss": 3.1131,
      "theoretical_loss": 4.5053136228239925,
      "tokens_seen": 153419776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048156469408224673,
      "loss": 3.3206,
      "theoretical_loss": 4.505062945046413,
      "tokens_seen": 153485312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048155466399197596,
      "loss": 3.3988,
      "theoretical_loss": 4.504812404237197,
      "tokens_seen": 153550848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815446339017051,
      "loss": 3.1752,
      "theoretical_loss": 4.504562000263093,
      "tokens_seen": 153616384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815346038114343,
      "loss": 3.1504,
      "theoretical_loss": 4.504311732991039,
      "tokens_seen": 153681920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815245737211635,
      "loss": 3.0018,
      "theoretical_loss": 4.504061602288159,
      "tokens_seen": 153747456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815145436308927,
      "loss": 2.9879,
      "theoretical_loss": 4.503811608021761,
      "tokens_seen": 153812992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048150451354062187,
      "loss": 3.1938,
      "theoretical_loss": 4.50356175005934,
      "tokens_seen": 153878528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048149448345035105,
      "loss": 3.5199,
      "theoretical_loss": 4.503312028268576,
      "tokens_seen": 153944064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 95853,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.367169141769409,
      "objective/train/theoretical_loss": 4.503062442517334,
      "objective/train/tokens_used": 174469600,
      "theoretical_loss": 4.503062442517334,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048148445336008023,
      "loss": 3.257,
      "theoretical_loss": 4.503062442517334,
      "tokens_seen": 154009600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048147442326980947,
      "loss": 3.4149,
      "theoretical_loss": 4.502812992673661,
      "tokens_seen": 154075136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814643931795386,
      "loss": 3.3831,
      "theoretical_loss": 4.50256367860579,
      "tokens_seen": 154140672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048145436308926783,
      "loss": 3.2127,
      "theoretical_loss": 4.502314500182138,
      "tokens_seen": 154206208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048144433299899696,
      "loss": 3.0253,
      "theoretical_loss": 4.502065457271304,
      "tokens_seen": 154271744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814343029087262,
      "loss": 3.522,
      "theoretical_loss": 4.501816549742072,
      "tokens_seen": 154337280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048142427281845537,
      "loss": 3.031,
      "theoretical_loss": 4.501567777463406,
      "tokens_seen": 154402816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048141424272818455,
      "loss": 3.33,
      "theoretical_loss": 4.501319140304456,
      "tokens_seen": 154468352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048140421263791373,
      "loss": 3.0943,
      "theoretical_loss": 4.50107063813455,
      "tokens_seen": 154533888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048139418254764297,
      "loss": 2.9423,
      "theoretical_loss": 4.5008222708232,
      "tokens_seen": 154599424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813841524573721,
      "loss": 3.2762,
      "theoretical_loss": 4.500574038240099,
      "tokens_seen": 154664960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048137412236710133,
      "loss": 3.1341,
      "theoretical_loss": 4.500325940255123,
      "tokens_seen": 154730496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048136409227683046,
      "loss": 3.323,
      "theoretical_loss": 4.500077976738324,
      "tokens_seen": 154796032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813540621865597,
      "loss": 3.2968,
      "theoretical_loss": 4.499830147559939,
      "tokens_seen": 154861568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813440320962889,
      "loss": 3.3287,
      "theoretical_loss": 4.499582452590383,
      "tokens_seen": 154927104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048133400200601806,
      "loss": 3.4478,
      "theoretical_loss": 4.499334891700251,
      "tokens_seen": 154992640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048132397191574724,
      "loss": 3.1801,
      "theoretical_loss": 4.499087464760317,
      "tokens_seen": 155058176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813139418254764,
      "loss": 3.0457,
      "theoretical_loss": 4.498840171641534,
      "tokens_seen": 155123712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813039117352056,
      "loss": 2.9473,
      "theoretical_loss": 4.498593012215036,
      "tokens_seen": 155189248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048129388164493483,
      "loss": 3.1367,
      "theoretical_loss": 4.498345986352133,
      "tokens_seen": 155254784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048128385155466396,
      "loss": 3.2789,
      "theoretical_loss": 4.4980990939243135,
      "tokens_seen": 155320320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812738214643932,
      "loss": 3.265,
      "theoretical_loss": 4.497852334803245,
      "tokens_seen": 155385856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812637913741223,
      "loss": 3.2818,
      "theoretical_loss": 4.497605708860771,
      "tokens_seen": 155451392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048125376128385156,
      "loss": 3.148,
      "theoretical_loss": 4.497359215968914,
      "tokens_seen": 155516928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048124373119358074,
      "loss": 3.1139,
      "theoretical_loss": 4.497112855999871,
      "tokens_seen": 155582464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 96994,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.577995777130127,
      "objective/train/theoretical_loss": 4.496866628826018,
      "objective/train/tokens_used": 176108000,
      "theoretical_loss": 4.496866628826018,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812337011033099,
      "loss": 3.366,
      "theoretical_loss": 4.496866628826018,
      "tokens_seen": 155648000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048122367101303916,
      "loss": 3.2348,
      "theoretical_loss": 4.496620534319904,
      "tokens_seen": 155713536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048121364092276834,
      "loss": 3.1337,
      "theoretical_loss": 4.496374572354256,
      "tokens_seen": 155779072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812036108324975,
      "loss": 3.4091,
      "theoretical_loss": 4.496128742801979,
      "tokens_seen": 155844608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811935807422267,
      "loss": 3.3832,
      "theoretical_loss": 4.495883045536148,
      "tokens_seen": 155910144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811835506519559,
      "loss": 3.1527,
      "theoretical_loss": 4.495637480430018,
      "tokens_seen": 155975680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048117352056168506,
      "loss": 3.3667,
      "theoretical_loss": 4.495392047357015,
      "tokens_seen": 156041216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811634904714143,
      "loss": 3.1819,
      "theoretical_loss": 4.4951467461907395,
      "tokens_seen": 156106752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811534603811434,
      "loss": 3.2472,
      "theoretical_loss": 4.494901576804969,
      "tokens_seen": 156172288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048114343029087266,
      "loss": 3.3371,
      "theoretical_loss": 4.4946565390736515,
      "tokens_seen": 156237824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811334002006018,
      "loss": 3.3338,
      "theoretical_loss": 4.494411632870911,
      "tokens_seen": 156303360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481123370110331,
      "loss": 3.3003,
      "theoretical_loss": 4.494166858071041,
      "tokens_seen": 156368896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811133400200602,
      "loss": 3.0684,
      "theoretical_loss": 4.493922214548512,
      "tokens_seen": 156434432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811033099297894,
      "loss": 3.1874,
      "theoretical_loss": 4.493677702177964,
      "tokens_seen": 156499968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048109327983951857,
      "loss": 3.2209,
      "theoretical_loss": 4.493433320834209,
      "tokens_seen": 156565504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810832497492478,
      "loss": 3.3108,
      "theoretical_loss": 4.493189070392235,
      "tokens_seen": 156631040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048107321965897693,
      "loss": 3.0176,
      "theoretical_loss": 4.492944950727196,
      "tokens_seen": 156696576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048106318956870616,
      "loss": 3.0247,
      "theoretical_loss": 4.492700961714419,
      "tokens_seen": 156762112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810531594784353,
      "loss": 3.1494,
      "theoretical_loss": 4.492457103229405,
      "tokens_seen": 156827648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810431293881645,
      "loss": 3.4656,
      "theoretical_loss": 4.492213375147822,
      "tokens_seen": 156893184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810330992978937,
      "loss": 3.1795,
      "theoretical_loss": 4.491969777345509,
      "tokens_seen": 156958720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810230692076229,
      "loss": 3.2572,
      "theoretical_loss": 4.491726309698475,
      "tokens_seen": 157024256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048101303911735207,
      "loss": 3.074,
      "theoretical_loss": 4.491482972082901,
      "tokens_seen": 157089792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048100300902708125,
      "loss": 3.3692,
      "theoretical_loss": 4.491239764375134,
      "tokens_seen": 157155328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048099297893681043,
      "loss": 3.166,
      "theoretical_loss": 4.490996686451692,
      "tokens_seen": 157220864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 97532,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0817697048187256,
      "objective/train/theoretical_loss": 4.4907537381892615,
      "objective/train/tokens_used": 177746400,
      "theoretical_loss": 4.4907537381892615,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048098294884653967,
      "loss": 3.0902,
      "theoretical_loss": 4.4907537381892615,
      "tokens_seen": 157286400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809729187562688,
      "loss": 3.3092,
      "theoretical_loss": 4.490510919464697,
      "tokens_seen": 157351936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048096288866599803,
      "loss": 3.1099,
      "theoretical_loss": 4.4902682301550225,
      "tokens_seen": 157417472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048095285857572716,
      "loss": 2.9646,
      "theoretical_loss": 4.4900256701374275,
      "tokens_seen": 157483008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809428284854564,
      "loss": 3.1576,
      "theoretical_loss": 4.489783239289271,
      "tokens_seen": 157548544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048093279839518557,
      "loss": 3.1089,
      "theoretical_loss": 4.48954093748808,
      "tokens_seen": 157614080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048092276830491475,
      "loss": 3.4034,
      "theoretical_loss": 4.489298764611546,
      "tokens_seen": 157679616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048091273821464393,
      "loss": 2.9096,
      "theoretical_loss": 4.48905672053753,
      "tokens_seen": 157745152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048090270812437317,
      "loss": 3.1601,
      "theoretical_loss": 4.488814805144055,
      "tokens_seen": 157810688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808926780341023,
      "loss": 3.1648,
      "theoretical_loss": 4.4885730183093155,
      "tokens_seen": 157876224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048088264794383153,
      "loss": 3.3537,
      "theoretical_loss": 4.488331359911669,
      "tokens_seen": 157941760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048087261785356066,
      "loss": 3.1861,
      "theoretical_loss": 4.488089829829638,
      "tokens_seen": 158007296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808625877632899,
      "loss": 3.1146,
      "theoretical_loss": 4.487848427941911,
      "tokens_seen": 158072832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808525576730191,
      "loss": 3.4209,
      "theoretical_loss": 4.487607154127343,
      "tokens_seen": 158138368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084252758274826,
      "loss": 3.4348,
      "theoretical_loss": 4.487366008264949,
      "tokens_seen": 158203904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048083249749247744,
      "loss": 3.3548,
      "theoretical_loss": 4.487124990233915,
      "tokens_seen": 158269440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808224674022066,
      "loss": 3.2445,
      "theoretical_loss": 4.486884099913585,
      "tokens_seen": 158334976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808124373119358,
      "loss": 3.5389,
      "theoretical_loss": 4.486643337183471,
      "tokens_seen": 158400512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048080240722166503,
      "loss": 2.9913,
      "theoretical_loss": 4.486402701923245,
      "tokens_seen": 158466048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048079237713139416,
      "loss": 3.1903,
      "theoretical_loss": 4.486162194012745,
      "tokens_seen": 158531584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807823470411234,
      "loss": 3.3898,
      "theoretical_loss": 4.48592181333197,
      "tokens_seen": 158597120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807723169508525,
      "loss": 3.134,
      "theoretical_loss": 4.485681559761082,
      "tokens_seen": 158662656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048076228686058176,
      "loss": 3.1577,
      "theoretical_loss": 4.485441433180408,
      "tokens_seen": 158728192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048075225677031094,
      "loss": 3.0772,
      "theoretical_loss": 4.485201433470432,
      "tokens_seen": 158793728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807422266800401,
      "loss": 3.3051,
      "theoretical_loss": 4.484961560511804,
      "tokens_seen": 158859264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 98333,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.562363386154175,
      "objective/train/theoretical_loss": 4.484721814185333,
      "objective/train/tokens_used": 179384800,
      "theoretical_loss": 4.484721814185333,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807321965897693,
      "loss": 3.3407,
      "theoretical_loss": 4.484721814185333,
      "tokens_seen": 158924800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072216649949854,
      "loss": 3.439,
      "theoretical_loss": 4.48448219437199,
      "tokens_seen": 158990336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048071213640922766,
      "loss": 3.5068,
      "theoretical_loss": 4.484242700952908,
      "tokens_seen": 159055872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807021063189569,
      "loss": 3.4014,
      "theoretical_loss": 4.48400333380938,
      "tokens_seen": 159121408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480692076228686,
      "loss": 3.3618,
      "theoretical_loss": 4.483764092822857,
      "tokens_seen": 159186944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068204613841526,
      "loss": 3.2863,
      "theoretical_loss": 4.483524977874953,
      "tokens_seen": 159252480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048067201604814444,
      "loss": 3.2132,
      "theoretical_loss": 4.48328598884744,
      "tokens_seen": 159318016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806619859578736,
      "loss": 3.2951,
      "theoretical_loss": 4.483047125622251,
      "tokens_seen": 159383552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806519558676028,
      "loss": 3.3459,
      "theoretical_loss": 4.482808388081477,
      "tokens_seen": 159449088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480641925777332,
      "loss": 3.1496,
      "theoretical_loss": 4.482569776107368,
      "tokens_seen": 159514624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048063189568706117,
      "loss": 3.2453,
      "theoretical_loss": 4.482331289582334,
      "tokens_seen": 159580160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806218655967904,
      "loss": 3.1819,
      "theoretical_loss": 4.482092928388941,
      "tokens_seen": 159645696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048061183550651953,
      "loss": 3.2609,
      "theoretical_loss": 4.4818546924099145,
      "tokens_seen": 159711232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048060180541624877,
      "loss": 3.2747,
      "theoretical_loss": 4.481616581528138,
      "tokens_seen": 159776768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048059177532597795,
      "loss": 3.1228,
      "theoretical_loss": 4.481378595626652,
      "tokens_seen": 159842304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058174523570713,
      "loss": 3.0437,
      "theoretical_loss": 4.481140734588655,
      "tokens_seen": 159907840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805717151454363,
      "loss": 3.3693,
      "theoretical_loss": 4.4809029982975,
      "tokens_seen": 159973376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805616850551655,
      "loss": 3.4456,
      "theoretical_loss": 4.480665386636701,
      "tokens_seen": 160038912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048055165496489467,
      "loss": 3.1342,
      "theoretical_loss": 4.480427899489925,
      "tokens_seen": 160104448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805416248746239,
      "loss": 2.9679,
      "theoretical_loss": 4.480190536740995,
      "tokens_seen": 160169984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048053159478435303,
      "loss": 3.4332,
      "theoretical_loss": 4.479953298273893,
      "tokens_seen": 160235520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052156469408227,
      "loss": 3.2521,
      "theoretical_loss": 4.479716183972753,
      "tokens_seen": 160301056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805115346038114,
      "loss": 3.5047,
      "theoretical_loss": 4.4794791937218665,
      "tokens_seen": 160366592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048050150451354063,
      "loss": 3.3783,
      "theoretical_loss": 4.47924232740568,
      "tokens_seen": 160432128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804914744232698,
      "loss": 3.3839,
      "theoretical_loss": 4.479005584908793,
      "tokens_seen": 160497664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 99019,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4908287525177,
      "objective/train/theoretical_loss": 4.478768966115963,
      "objective/train/tokens_used": 181023200,
      "theoretical_loss": 4.478768966115963,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480481444332999,
      "loss": 3.1208,
      "theoretical_loss": 4.478768966115963,
      "tokens_seen": 160563200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048047141424272823,
      "loss": 3.2379,
      "theoretical_loss": 4.478532470912096,
      "tokens_seen": 160628736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048046138415245736,
      "loss": 3.0995,
      "theoretical_loss": 4.478296099182261,
      "tokens_seen": 160694272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804513540621866,
      "loss": 3.2414,
      "theoretical_loss": 4.47805985081167,
      "tokens_seen": 160759808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048044132397191577,
      "loss": 3.1472,
      "theoretical_loss": 4.477823725685695,
      "tokens_seen": 160825344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048043129388164495,
      "loss": 3.2584,
      "theoretical_loss": 4.477587723689862,
      "tokens_seen": 160890880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042126379137413,
      "loss": 3.2151,
      "theoretical_loss": 4.477351844709846,
      "tokens_seen": 160956416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048041123370110337,
      "loss": 3.464,
      "theoretical_loss": 4.477116088631476,
      "tokens_seen": 161021952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804012036108325,
      "loss": 3.2482,
      "theoretical_loss": 4.476880455340735,
      "tokens_seen": 161087488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048039117352056173,
      "loss": 3.347,
      "theoretical_loss": 4.4766449447237555,
      "tokens_seen": 161153024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038114343029086,
      "loss": 3.2773,
      "theoretical_loss": 4.476409556666825,
      "tokens_seen": 161218560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803711133400201,
      "loss": 3.0332,
      "theoretical_loss": 4.476174291056379,
      "tokens_seen": 161284096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803610832497493,
      "loss": 3.2546,
      "theoretical_loss": 4.475939147779008,
      "tokens_seen": 161349632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048035105315947846,
      "loss": 3.0797,
      "theoretical_loss": 4.47570412672145,
      "tokens_seen": 161415168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034102306920764,
      "loss": 3.1193,
      "theoretical_loss": 4.475469227770596,
      "tokens_seen": 161480704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803309929789368,
      "loss": 3.222,
      "theoretical_loss": 4.475234450813487,
      "tokens_seen": 161546240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480320962888666,
      "loss": 3.3821,
      "theoretical_loss": 4.474999795737315,
      "tokens_seen": 161611776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048031093279839523,
      "loss": 3.17,
      "theoretical_loss": 4.47476526242942,
      "tokens_seen": 161677312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048030090270812436,
      "loss": 3.0239,
      "theoretical_loss": 4.474530850777294,
      "tokens_seen": 161742848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802908726178536,
      "loss": 3.2871,
      "theoretical_loss": 4.474296560668577,
      "tokens_seen": 161808384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802808425275827,
      "loss": 3.0498,
      "theoretical_loss": 4.474062391991058,
      "tokens_seen": 161873920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048027081243731196,
      "loss": 3.0457,
      "theoretical_loss": 4.473828344632677,
      "tokens_seen": 161939456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048026078234704114,
      "loss": 3.114,
      "theoretical_loss": 4.473594418481522,
      "tokens_seen": 162004992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802507522567703,
      "loss": 3.3295,
      "theoretical_loss": 4.473360613425827,
      "tokens_seen": 162070528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802407221664995,
      "loss": 3.046,
      "theoretical_loss": 4.473126929353977,
      "tokens_seen": 162136064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 100000,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.690746545791626,
      "objective/train/theoretical_loss": 4.472893366154503,
      "objective/train/tokens_used": 182661600,
      "theoretical_loss": 4.472893366154503,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048023069207622874,
      "loss": 2.9009,
      "theoretical_loss": 4.472893366154503,
      "tokens_seen": 162201600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048022066198595786,
      "loss": 3.0755,
      "theoretical_loss": 4.472659923716087,
      "tokens_seen": 162267136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802106318956871,
      "loss": 3.0381,
      "theoretical_loss": 4.472426601927554,
      "tokens_seen": 162332672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020060180541623,
      "loss": 3.0383,
      "theoretical_loss": 4.472193400677878,
      "tokens_seen": 162398208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048019057171514546,
      "loss": 3.207,
      "theoretical_loss": 4.471960319856182,
      "tokens_seen": 162463744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048018054162487464,
      "loss": 3.0581,
      "theoretical_loss": 4.471727359351734,
      "tokens_seen": 162529280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801705115346038,
      "loss": 3.323,
      "theoretical_loss": 4.4714945190539455,
      "tokens_seen": 162594816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480160481444333,
      "loss": 3.3514,
      "theoretical_loss": 4.47126179885238,
      "tokens_seen": 162660352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801504513540622,
      "loss": 3.0515,
      "theoretical_loss": 4.471029198636741,
      "tokens_seen": 162725888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014042126379137,
      "loss": 2.9545,
      "theoretical_loss": 4.470796718296884,
      "tokens_seen": 162791424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801303911735206,
      "loss": 3.0233,
      "theoretical_loss": 4.470564357722804,
      "tokens_seen": 162856960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048012036108324973,
      "loss": 3.258,
      "theoretical_loss": 4.470332116804644,
      "tokens_seen": 162922496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048011033099297897,
      "loss": 3.2388,
      "theoretical_loss": 4.470099995432692,
      "tokens_seen": 162988032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048010030090270815,
      "loss": 3.0379,
      "theoretical_loss": 4.469867993497379,
      "tokens_seen": 163053568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048009027081243733,
      "loss": 3.0891,
      "theoretical_loss": 4.469636110889283,
      "tokens_seen": 163119104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800802407221665,
      "loss": 3.2988,
      "theoretical_loss": 4.4694043474991245,
      "tokens_seen": 163184640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800702106318957,
      "loss": 3.0297,
      "theoretical_loss": 4.469172703217767,
      "tokens_seen": 163250176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048006018054162487,
      "loss": 3.0857,
      "theoretical_loss": 4.46894117793622,
      "tokens_seen": 163315712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800501504513541,
      "loss": 3.1744,
      "theoretical_loss": 4.468709771545634,
      "tokens_seen": 163381248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004012036108323,
      "loss": 3.0103,
      "theoretical_loss": 4.4684784839373055,
      "tokens_seen": 163446784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048003009027081247,
      "loss": 3.2331,
      "theoretical_loss": 4.468247315002671,
      "tokens_seen": 163512320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800200601805416,
      "loss": 3.1392,
      "theoretical_loss": 4.468016264633311,
      "tokens_seen": 163577856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048001003009027083,
      "loss": 3.1491,
      "theoretical_loss": 4.4677853327209505,
      "tokens_seen": 163643392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 3.3932,
      "theoretical_loss": 4.467554519157452,
      "tokens_seen": 163708928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799899699097292,
      "loss": 3.0553,
      "theoretical_loss": 4.467323823834824,
      "tokens_seen": 163774464
    },
    {
      "debugging/Self-BLEU-5": 0.4946583130096161,
      "debugging/distinct-1-grams": 0.815790683224271,
      "debugging/distinct-2-grams": 0.9751385384195338,
      "debugging/entropy-1-grams": 5.884314010514224,
      "debugging/entropy-2-grams": 6.564196933680893,
      "debugging/length": 580.8,
      "debugging/num_segments": 10,
      "epoch": 0.05,
      "objective/train/docs_used": 101353,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.959970474243164,
      "objective/train/theoretical_loss": 4.467093246645215,
      "objective/train/tokens_used": 184300000,
      "theoretical_loss": 4.467093246645215,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799799398194584,
      "loss": 3.3758,
      "theoretical_loss": 4.467093246645215,
      "tokens_seen": 163840000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047996990972918756,
      "loss": 3.0913,
      "theoretical_loss": 4.466862787480915,
      "tokens_seen": 163905536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047995987963891674,
      "loss": 3.201,
      "theoretical_loss": 4.466632446234355,
      "tokens_seen": 163971072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047994984954864597,
      "loss": 3.1657,
      "theoretical_loss": 4.466402222798108,
      "tokens_seen": 164036608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799398194583751,
      "loss": 3.0971,
      "theoretical_loss": 4.466172117064889,
      "tokens_seen": 164102144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047992978936810433,
      "loss": 3.0109,
      "theoretical_loss": 4.465942128927549,
      "tokens_seen": 164167680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799197592778335,
      "loss": 3.2315,
      "theoretical_loss": 4.4657122582790825,
      "tokens_seen": 164233216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004799097291875627,
      "loss": 3.2653,
      "theoretical_loss": 4.465482505012624,
      "tokens_seen": 164298752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798996990972919,
      "loss": 3.359,
      "theoretical_loss": 4.465252869021448,
      "tokens_seen": 164364288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047988966900702106,
      "loss": 3.0349,
      "theoretical_loss": 4.465023350198967,
      "tokens_seen": 164429824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047987963891675024,
      "loss": 3.4322,
      "theoretical_loss": 4.464793948438733,
      "tokens_seen": 164495360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798696088264795,
      "loss": 3.1553,
      "theoretical_loss": 4.464564663634439,
      "tokens_seen": 164560896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798595787362086,
      "loss": 3.3483,
      "theoretical_loss": 4.464335495679914,
      "tokens_seen": 164626432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047984954864593784,
      "loss": 3.2043,
      "theoretical_loss": 4.464106444469129,
      "tokens_seen": 164691968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983951855566696,
      "loss": 3.1635,
      "theoretical_loss": 4.463877509896189,
      "tokens_seen": 164757504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798294884653962,
      "loss": 3.298,
      "theoretical_loss": 4.463648691855342,
      "tokens_seen": 164823040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798194583751254,
      "loss": 3.3382,
      "theoretical_loss": 4.46341999024097,
      "tokens_seen": 164888576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047980942828485456,
      "loss": 3.3223,
      "theoretical_loss": 4.463191404947595,
      "tokens_seen": 164954112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047979939819458374,
      "loss": 3.2243,
      "theoretical_loss": 4.462962935869875,
      "tokens_seen": 165019648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797893681043129,
      "loss": 3.1862,
      "theoretical_loss": 4.4627345829026055,
      "tokens_seen": 165085184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797793380140421,
      "loss": 3.2557,
      "theoretical_loss": 4.462506345940719,
      "tokens_seen": 165150720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047976930792377134,
      "loss": 2.9256,
      "theoretical_loss": 4.462278224879286,
      "tokens_seen": 165216256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975927783350047,
      "loss": 2.9158,
      "theoretical_loss": 4.462050219613511,
      "tokens_seen": 165281792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797492477432297,
      "loss": 3.2504,
      "theoretical_loss": 4.461822330038737,
      "tokens_seen": 165347328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797392176529589,
      "loss": 3.042,
      "theoretical_loss": 4.461594556050444,
      "tokens_seen": 165412864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 101896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.066077470779419,
      "objective/train/theoretical_loss": 4.4613668975442415,
      "objective/train/tokens_used": 185938400,
      "theoretical_loss": 4.4613668975442415,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047972918756268807,
      "loss": 3.1283,
      "theoretical_loss": 4.4613668975442415,
      "tokens_seen": 165478400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797191574724173,
      "loss": 3.3684,
      "theoretical_loss": 4.4611393544158835,
      "tokens_seen": 165543936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047970912738214643,
      "loss": 3.3009,
      "theoretical_loss": 4.460911926561252,
      "tokens_seen": 165609472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047969909729187566,
      "loss": 3.4005,
      "theoretical_loss": 4.460684613876367,
      "tokens_seen": 165675008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047968906720160484,
      "loss": 3.1465,
      "theoretical_loss": 4.460457416257386,
      "tokens_seen": 165740544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479679037111334,
      "loss": 2.865,
      "theoretical_loss": 4.460230333600595,
      "tokens_seen": 165806080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796690070210632,
      "loss": 3.2027,
      "theoretical_loss": 4.46000336580242,
      "tokens_seen": 165871616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796589769307924,
      "loss": 3.3585,
      "theoretical_loss": 4.459776512759419,
      "tokens_seen": 165937152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047964894684052157,
      "loss": 3.2687,
      "theoretical_loss": 4.459549774368284,
      "tokens_seen": 166002688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796389167502508,
      "loss": 3.4614,
      "theoretical_loss": 4.45932315052584,
      "tokens_seen": 166068224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047962888665997993,
      "loss": 2.9803,
      "theoretical_loss": 4.459096641129047,
      "tokens_seen": 166133760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047961885656970917,
      "loss": 3.0739,
      "theoretical_loss": 4.458870246074998,
      "tokens_seen": 166199296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047960882647943835,
      "loss": 3.285,
      "theoretical_loss": 4.458643965260917,
      "tokens_seen": 166264832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047959879638916753,
      "loss": 3.1638,
      "theoretical_loss": 4.458417798584164,
      "tokens_seen": 166330368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795887662988967,
      "loss": 2.905,
      "theoretical_loss": 4.45819174594223,
      "tokens_seen": 166395904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795787362086259,
      "loss": 3.2569,
      "theoretical_loss": 4.457965807232739,
      "tokens_seen": 166461440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047956870611835507,
      "loss": 3.299,
      "theoretical_loss": 4.457739982353445,
      "tokens_seen": 166526976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795586760280843,
      "loss": 2.9863,
      "theoretical_loss": 4.457514271202237,
      "tokens_seen": 166592512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047954864593781343,
      "loss": 3.2391,
      "theoretical_loss": 4.457288673677134,
      "tokens_seen": 166658048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953861584754267,
      "loss": 3.3358,
      "theoretical_loss": 4.457063189676287,
      "tokens_seen": 166723584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795285857572718,
      "loss": 3.2532,
      "theoretical_loss": 4.456837819097978,
      "tokens_seen": 166789120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047951855566700103,
      "loss": 3.1618,
      "theoretical_loss": 4.45661256184062,
      "tokens_seen": 166854656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795085255767302,
      "loss": 3.0743,
      "theoretical_loss": 4.456387417802757,
      "tokens_seen": 166920192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794984954864594,
      "loss": 3.3956,
      "theoretical_loss": 4.456162386883065,
      "tokens_seen": 166985728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794884653961886,
      "loss": 3.1311,
      "theoretical_loss": 4.455937468980348,
      "tokens_seen": 167051264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 103097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.102741003036499,
      "objective/train/theoretical_loss": 4.455712663993541,
      "objective/train/tokens_used": 187576800,
      "theoretical_loss": 4.455712663993541,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047947843530591776,
      "loss": 3.2024,
      "theoretical_loss": 4.455712663993541,
      "tokens_seen": 167116800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047946840521564694,
      "loss": 3.4233,
      "theoretical_loss": 4.455487971821709,
      "tokens_seen": 167182336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047945837512537617,
      "loss": 3.1938,
      "theoretical_loss": 4.4552633923640474,
      "tokens_seen": 167247872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794483450351053,
      "loss": 3.1579,
      "theoretical_loss": 4.455038925519882,
      "tokens_seen": 167313408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943831494483453,
      "loss": 3.2525,
      "theoretical_loss": 4.454814571188665,
      "tokens_seen": 167378944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794282848545637,
      "loss": 3.0859,
      "theoretical_loss": 4.454590329269979,
      "tokens_seen": 167444480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794182547642929,
      "loss": 3.4173,
      "theoretical_loss": 4.454366199663537,
      "tokens_seen": 167510016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794082246740221,
      "loss": 3.1989,
      "theoretical_loss": 4.4541421822691785,
      "tokens_seen": 167575552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939819458375126,
      "loss": 3.2261,
      "theoretical_loss": 4.453918276986872,
      "tokens_seen": 167641088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047938816449348044,
      "loss": 3.3857,
      "theoretical_loss": 4.453694483716717,
      "tokens_seen": 167706624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793781344032097,
      "loss": 3.2033,
      "theoretical_loss": 4.453470802358938,
      "tokens_seen": 167772160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793681043129388,
      "loss": 3.3288,
      "theoretical_loss": 4.453247232813885,
      "tokens_seen": 167837696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047935807422266804,
      "loss": 3.3873,
      "theoretical_loss": 4.453023774982041,
      "tokens_seen": 167903232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047934804413239716,
      "loss": 3.027,
      "theoretical_loss": 4.4528004287640135,
      "tokens_seen": 167968768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793380140421264,
      "loss": 2.93,
      "theoretical_loss": 4.452577194060538,
      "tokens_seen": 168034304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793279839518556,
      "loss": 3.338,
      "theoretical_loss": 4.452354070772476,
      "tokens_seen": 168099840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047931795386158476,
      "loss": 3.0985,
      "theoretical_loss": 4.452131058800816,
      "tokens_seen": 168165376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047930792377131394,
      "loss": 3.0258,
      "theoretical_loss": 4.451908158046674,
      "tokens_seen": 168230912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792978936810431,
      "loss": 3.2144,
      "theoretical_loss": 4.451685368411292,
      "tokens_seen": 168296448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792878635907723,
      "loss": 3.1754,
      "theoretical_loss": 4.4514626897960365,
      "tokens_seen": 168361984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047927783350050154,
      "loss": 2.8836,
      "theoretical_loss": 4.451240122102401,
      "tokens_seen": 168427520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047926780341023067,
      "loss": 3.2178,
      "theoretical_loss": 4.451017665232007,
      "tokens_seen": 168493056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792577733199599,
      "loss": 2.9942,
      "theoretical_loss": 4.4507953190865965,
      "tokens_seen": 168558592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792477432296891,
      "loss": 3.2664,
      "theoretical_loss": 4.450573083568042,
      "tokens_seen": 168624128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923771313941827,
      "loss": 2.9989,
      "theoretical_loss": 4.450350958578337,
      "tokens_seen": 168689664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 104347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.790850877761841,
      "objective/train/theoretical_loss": 4.4501289440196015,
      "objective/train/tokens_used": 189215200,
      "theoretical_loss": 4.4501289440196015,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047922768304914745,
      "loss": 3.2164,
      "theoretical_loss": 4.4501289440196015,
      "tokens_seen": 168755200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047921765295887663,
      "loss": 3.3929,
      "theoretical_loss": 4.449907039794081,
      "tokens_seen": 168820736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792076228686058,
      "loss": 3.2061,
      "theoretical_loss": 4.449685245804145,
      "tokens_seen": 168886272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919759277833504,
      "loss": 2.9032,
      "theoretical_loss": 4.449463561952285,
      "tokens_seen": 168951808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047918756268806417,
      "loss": 3.1277,
      "theoretical_loss": 4.44924198814112,
      "tokens_seen": 169017344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791775325977934,
      "loss": 3.2685,
      "theoretical_loss": 4.44902052427339,
      "tokens_seen": 169082880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047916750250752253,
      "loss": 3.0655,
      "theoretical_loss": 4.448799170251961,
      "tokens_seen": 169148416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047915747241725177,
      "loss": 3.3278,
      "theoretical_loss": 4.448577925979821,
      "tokens_seen": 169213952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047914744232698095,
      "loss": 3.2436,
      "theoretical_loss": 4.44835679136008,
      "tokens_seen": 169279488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913741223671013,
      "loss": 3.2687,
      "theoretical_loss": 4.448135766295974,
      "tokens_seen": 169345024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791273821464393,
      "loss": 3.1304,
      "theoretical_loss": 4.447914850690859,
      "tokens_seen": 169410560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047911735205616855,
      "loss": 3.199,
      "theoretical_loss": 4.447694044448216,
      "tokens_seen": 169476096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791073219658977,
      "loss": 3.2522,
      "theoretical_loss": 4.447473347471645,
      "tokens_seen": 169541632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790972918756269,
      "loss": 3.1969,
      "theoretical_loss": 4.447252759664874,
      "tokens_seen": 169607168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047908726178535604,
      "loss": 3.0039,
      "theoretical_loss": 4.4470322809317455,
      "tokens_seen": 169672704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047907723169508527,
      "loss": 3.5408,
      "theoretical_loss": 4.44681191117623,
      "tokens_seen": 169738240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047906720160481445,
      "loss": 3.2186,
      "theoretical_loss": 4.446591650302416,
      "tokens_seen": 169803776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905717151454363,
      "loss": 2.9781,
      "theoretical_loss": 4.446371498214516,
      "tokens_seen": 169869312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790471414242728,
      "loss": 3.4193,
      "theoretical_loss": 4.446151454816862,
      "tokens_seen": 169934848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479037111334002,
      "loss": 3.0771,
      "theoretical_loss": 4.445931520013906,
      "tokens_seen": 170000384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790270812437312,
      "loss": 3.2464,
      "theoretical_loss": 4.445711693710223,
      "tokens_seen": 170065920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790170511534604,
      "loss": 3.2014,
      "theoretical_loss": 4.4454919758105085,
      "tokens_seen": 170131456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047900702106318954,
      "loss": 3.3265,
      "theoretical_loss": 4.445272366219575,
      "tokens_seen": 170196992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789969909729188,
      "loss": 3.073,
      "theoretical_loss": 4.44505286484236,
      "tokens_seen": 170262528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789869608826479,
      "loss": 2.936,
      "theoretical_loss": 4.4448334715839195,
      "tokens_seen": 170328064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 105037,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8210434913635254,
      "objective/train/theoretical_loss": 4.444614186349425,
      "objective/train/tokens_used": 190853600,
      "theoretical_loss": 4.444614186349425,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047897693079237714,
      "loss": 2.9571,
      "theoretical_loss": 4.444614186349425,
      "tokens_seen": 170393600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047896690070210637,
      "loss": 2.9778,
      "theoretical_loss": 4.444395009044175,
      "tokens_seen": 170459136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789568706118355,
      "loss": 3.2967,
      "theoretical_loss": 4.444175939573581,
      "tokens_seen": 170524672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047894684052156473,
      "loss": 2.8109,
      "theoretical_loss": 4.443956977843177,
      "tokens_seen": 170590208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789368104312939,
      "loss": 3.4383,
      "theoretical_loss": 4.4437381237586155,
      "tokens_seen": 170655744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789267803410231,
      "loss": 3.1635,
      "theoretical_loss": 4.443519377225666,
      "tokens_seen": 170721280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789167502507523,
      "loss": 3.1526,
      "theoretical_loss": 4.443300738150221,
      "tokens_seen": 170786816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047890672016048146,
      "loss": 3.3384,
      "theoretical_loss": 4.4430822064382856,
      "tokens_seen": 170852352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047889669007021064,
      "loss": 3.4135,
      "theoretical_loss": 4.442863781995989,
      "tokens_seen": 170917888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788866599799399,
      "loss": 3.0138,
      "theoretical_loss": 4.442645464729573,
      "tokens_seen": 170983424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478876629889669,
      "loss": 3.2596,
      "theoretical_loss": 4.4424272545454,
      "tokens_seen": 171048960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047886659979939824,
      "loss": 3.2555,
      "theoretical_loss": 4.442209151349951,
      "tokens_seen": 171114496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047885656970912736,
      "loss": 3.3042,
      "theoretical_loss": 4.4419911550498234,
      "tokens_seen": 171180032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788465396188566,
      "loss": 3.2061,
      "theoretical_loss": 4.441773265551731,
      "tokens_seen": 171245568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788365095285858,
      "loss": 3.0708,
      "theoretical_loss": 4.441555482762507,
      "tokens_seen": 171311104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047882647943831496,
      "loss": 3.2816,
      "theoretical_loss": 4.441337806589097,
      "tokens_seen": 171376640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047881644934804414,
      "loss": 3.1222,
      "theoretical_loss": 4.441120236938568,
      "tokens_seen": 171442176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788064192577733,
      "loss": 3.263,
      "theoretical_loss": 4.440902773718101,
      "tokens_seen": 171507712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787963891675025,
      "loss": 2.9574,
      "theoretical_loss": 4.440685416834994,
      "tokens_seen": 171573248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047878635907723174,
      "loss": 3.4099,
      "theoretical_loss": 4.440468166196662,
      "tokens_seen": 171638784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047877632898696087,
      "loss": 3.2833,
      "theoretical_loss": 4.440251021710634,
      "tokens_seen": 171704320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787662988966901,
      "loss": 3.2847,
      "theoretical_loss": 4.440033983284556,
      "tokens_seen": 171769856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787562688064193,
      "loss": 3.3134,
      "theoretical_loss": 4.43981705082619,
      "tokens_seen": 171835392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047874623871614847,
      "loss": 3.1022,
      "theoretical_loss": 4.439600224243412,
      "tokens_seen": 171900928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047873620862587765,
      "loss": 3.1256,
      "theoretical_loss": 4.439383503444214,
      "tokens_seen": 171966464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 106163,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.448519468307495,
      "objective/train/theoretical_loss": 4.439166888336702,
      "objective/train/tokens_used": 192492000,
      "theoretical_loss": 4.439166888336702,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047872617853560683,
      "loss": 3.3453,
      "theoretical_loss": 4.439166888336702,
      "tokens_seen": 172032000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478716148445336,
      "loss": 3.0654,
      "theoretical_loss": 4.438950378829098,
      "tokens_seen": 172097536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047870611835506524,
      "loss": 3.3586,
      "theoretical_loss": 4.43873397482974,
      "tokens_seen": 172163072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047869608826479437,
      "loss": 3.1219,
      "theoretical_loss": 4.438517676247075,
      "tokens_seen": 172228608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786860581745236,
      "loss": 3.0914,
      "theoretical_loss": 4.4383014829896705,
      "tokens_seen": 172294144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047867602808425273,
      "loss": 3.1461,
      "theoretical_loss": 4.4380853949662065,
      "tokens_seen": 172359680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047866599799398197,
      "loss": 3.0581,
      "theoretical_loss": 4.4378694120854725,
      "tokens_seen": 172425216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047865596790371115,
      "loss": 3.0474,
      "theoretical_loss": 4.437653534256377,
      "tokens_seen": 172490752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047864593781344033,
      "loss": 3.207,
      "theoretical_loss": 4.437437761387939,
      "tokens_seen": 172556288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786359077231695,
      "loss": 3.1347,
      "theoretical_loss": 4.437222093389293,
      "tokens_seen": 172621824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047862587763289875,
      "loss": 3.0996,
      "theoretical_loss": 4.437006530169684,
      "tokens_seen": 172687360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786158475426279,
      "loss": 3.255,
      "theoretical_loss": 4.436791071638472,
      "tokens_seen": 172752896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786058174523571,
      "loss": 3.3615,
      "theoretical_loss": 4.43657571770513,
      "tokens_seen": 172818432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047859578736208624,
      "loss": 3.3575,
      "theoretical_loss": 4.436360468279242,
      "tokens_seen": 172883968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047858575727181547,
      "loss": 3.1795,
      "theoretical_loss": 4.4361453232705035,
      "tokens_seen": 172949504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047857572718154465,
      "loss": 3.2246,
      "theoretical_loss": 4.435930282588725,
      "tokens_seen": 173015040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047856569709127383,
      "loss": 3.1589,
      "theoretical_loss": 4.435715346143828,
      "tokens_seen": 173080576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478555667001003,
      "loss": 3.1079,
      "theoretical_loss": 4.435500513845845,
      "tokens_seen": 173146112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785456369107322,
      "loss": 3.0886,
      "theoretical_loss": 4.435285785604922,
      "tokens_seen": 173211648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785356068204614,
      "loss": 3.3142,
      "theoretical_loss": 4.435071161331313,
      "tokens_seen": 173277184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785255767301906,
      "loss": 3.3968,
      "theoretical_loss": 4.434856640935388,
      "tokens_seen": 173342720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047851554663991974,
      "loss": 3.1791,
      "theoretical_loss": 4.434642224327625,
      "tokens_seen": 173408256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478505516549649,
      "loss": 3.1941,
      "theoretical_loss": 4.434427911418613,
      "tokens_seen": 173473792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784954864593781,
      "loss": 3.1568,
      "theoretical_loss": 4.434213702119052,
      "tokens_seen": 173539328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047848545636910734,
      "loss": 3.3522,
      "theoretical_loss": 4.433999596339755,
      "tokens_seen": 173604864
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 106755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8909780979156494,
      "objective/train/theoretical_loss": 4.433785593991642,
      "objective/train/tokens_used": 194130400,
      "theoretical_loss": 4.433785593991642,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784754262788365,
      "loss": 3.2649,
      "theoretical_loss": 4.433785593991642,
      "tokens_seen": 173670400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784653961885657,
      "loss": 3.2448,
      "theoretical_loss": 4.433571694985744,
      "tokens_seen": 173735936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784553660982949,
      "loss": 3.2039,
      "theoretical_loss": 4.433357899233204,
      "tokens_seen": 173801472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784453360080241,
      "loss": 3.0269,
      "theoretical_loss": 4.433144206645274,
      "tokens_seen": 173867008
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047843530591775324,
      "loss": 3.0917,
      "theoretical_loss": 4.432930617133313,
      "tokens_seen": 173932544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784252758274825,
      "loss": 3.1783,
      "theoretical_loss": 4.432717130608793,
      "tokens_seen": 173998080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784152457372116,
      "loss": 3.1232,
      "theoretical_loss": 4.432503746983295,
      "tokens_seen": 174063616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047840521564694084,
      "loss": 3.2631,
      "theoretical_loss": 4.432290466168507,
      "tokens_seen": 174129152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047839518555667,
      "loss": 3.2136,
      "theoretical_loss": 4.432077288076228,
      "tokens_seen": 174194688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783851554663992,
      "loss": 3.2796,
      "theoretical_loss": 4.431864212618365,
      "tokens_seen": 174260224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783751253761284,
      "loss": 3.4694,
      "theoretical_loss": 4.431651239706934,
      "tokens_seen": 174325760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047836509528585756,
      "loss": 3.0788,
      "theoretical_loss": 4.431438369254058,
      "tokens_seen": 174391296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047835506519558675,
      "loss": 3.0872,
      "theoretical_loss": 4.43122560117197,
      "tokens_seen": 174456832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478345035105316,
      "loss": 2.9119,
      "theoretical_loss": 4.431012935373012,
      "tokens_seen": 174522368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783350050150451,
      "loss": 3.2154,
      "theoretical_loss": 4.430800371769631,
      "tokens_seen": 174587904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047832497492477434,
      "loss": 3.2817,
      "theoretical_loss": 4.430587910274383,
      "tokens_seen": 174653440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047831494483450347,
      "loss": 3.1007,
      "theoretical_loss": 4.430375550799933,
      "tokens_seen": 174718976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783049147442327,
      "loss": 3.3202,
      "theoretical_loss": 4.430163293259053,
      "tokens_seen": 174784512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782948846539619,
      "loss": 3.2775,
      "theoretical_loss": 4.42995113756462,
      "tokens_seen": 174850048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047828485456369107,
      "loss": 3.4485,
      "theoretical_loss": 4.42973908362962,
      "tokens_seen": 174915584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047827482447342025,
      "loss": 3.1023,
      "theoretical_loss": 4.429527131367147,
      "tokens_seen": 174981120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782647943831495,
      "loss": 3.0822,
      "theoretical_loss": 4.429315280690398,
      "tokens_seen": 175046656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782547642928786,
      "loss": 3.2901,
      "theoretical_loss": 4.4291035315126805,
      "tokens_seen": 175112192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047824473420260785,
      "loss": 3.2442,
      "theoretical_loss": 4.428891883747406,
      "tokens_seen": 175177728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478234704112337,
      "loss": 3.166,
      "theoretical_loss": 4.428680337308094,
      "tokens_seen": 175243264
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 107724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6482720375061035,
      "objective/train/theoretical_loss": 4.428468892108368,
      "objective/train/tokens_used": 195768800,
      "theoretical_loss": 4.428468892108368,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782246740220662,
      "loss": 3.3804,
      "theoretical_loss": 4.428468892108368,
      "tokens_seen": 175308800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047821464393179544,
      "loss": 3.4307,
      "theoretical_loss": 4.428257548061959,
      "tokens_seen": 175374336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047820461384152457,
      "loss": 3.2556,
      "theoretical_loss": 4.428046305082704,
      "tokens_seen": 175439872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781945837512538,
      "loss": 3.3203,
      "theoretical_loss": 4.427835163084543,
      "tokens_seen": 175505408
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047818455366098293,
      "loss": 3.1723,
      "theoretical_loss": 4.427624121981523,
      "tokens_seen": 175570944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047817452357071217,
      "loss": 3.1751,
      "theoretical_loss": 4.427413181687799,
      "tokens_seen": 175636480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047816449348044135,
      "loss": 2.9857,
      "theoretical_loss": 4.427202342117624,
      "tokens_seen": 175702016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047815446339017053,
      "loss": 3.1727,
      "theoretical_loss": 4.426991603185365,
      "tokens_seen": 175767552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781444332998997,
      "loss": 3.121,
      "theoretical_loss": 4.426780964805486,
      "tokens_seen": 175833088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047813440320962895,
      "loss": 3.3494,
      "theoretical_loss": 4.4265704268925585,
      "tokens_seen": 175898624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781243731193581,
      "loss": 3.313,
      "theoretical_loss": 4.426359989361259,
      "tokens_seen": 175964160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781143430290873,
      "loss": 3.3073,
      "theoretical_loss": 4.426149652126369,
      "tokens_seen": 176029696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047810431293881644,
      "loss": 3.0604,
      "theoretical_loss": 4.42593941510277,
      "tokens_seen": 176095232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047809428284854567,
      "loss": 2.7989,
      "theoretical_loss": 4.42572927820545,
      "tokens_seen": 176160768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047808425275827485,
      "loss": 3.2073,
      "theoretical_loss": 4.4255192413495035,
      "tokens_seen": 176226304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047807422266800403,
      "loss": 3.0824,
      "theoretical_loss": 4.425309304450122,
      "tokens_seen": 176291840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780641925777332,
      "loss": 3.1597,
      "theoretical_loss": 4.4250994674226085,
      "tokens_seen": 176357376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780541624874624,
      "loss": 3.1731,
      "theoretical_loss": 4.424889730182361,
      "tokens_seen": 176422912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780441323971916,
      "loss": 3.3307,
      "theoretical_loss": 4.424680092644886,
      "tokens_seen": 176488448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780341023069208,
      "loss": 3.0847,
      "theoretical_loss": 4.42447055472579,
      "tokens_seen": 176553984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047802407221664994,
      "loss": 3.3959,
      "theoretical_loss": 4.4242611163407854,
      "tokens_seen": 176619520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780140421263792,
      "loss": 3.2021,
      "theoretical_loss": 4.424051777405684,
      "tokens_seen": 176685056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780040120361083,
      "loss": 3.0791,
      "theoretical_loss": 4.423842537836402,
      "tokens_seen": 176750592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799398194583754,
      "loss": 3.112,
      "theoretical_loss": 4.423633397548956,
      "tokens_seen": 176816128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779839518555667,
      "loss": 3.0609,
      "theoretical_loss": 4.423424356459467,
      "tokens_seen": 176881664
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 108296,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4872632026672363,
      "objective/train/theoretical_loss": 4.423215414484155,
      "objective/train/tokens_used": 197407200,
      "theoretical_loss": 4.423215414484155,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779739217652959,
      "loss": 3.3284,
      "theoretical_loss": 4.423215414484155,
      "tokens_seen": 176947200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779638916750251,
      "loss": 2.839,
      "theoretical_loss": 4.423006571539346,
      "tokens_seen": 177012736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779538615847543,
      "loss": 3.1684,
      "theoretical_loss": 4.422797827541462,
      "tokens_seen": 177078272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047794383149448344,
      "loss": 3.1424,
      "theoretical_loss": 4.422589182407032,
      "tokens_seen": 177143808
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779338014042127,
      "loss": 3.2199,
      "theoretical_loss": 4.422380636052681,
      "tokens_seen": 177209344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779237713139418,
      "loss": 2.9687,
      "theoretical_loss": 4.422172188395141,
      "tokens_seen": 177274880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047791374122367104,
      "loss": 3.0297,
      "theoretical_loss": 4.421963839351239,
      "tokens_seen": 177340416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779037111334002,
      "loss": 3.1726,
      "theoretical_loss": 4.421755588837907,
      "tokens_seen": 177405952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778936810431294,
      "loss": 3.3507,
      "theoretical_loss": 4.421547436772176,
      "tokens_seen": 177471488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778836509528586,
      "loss": 3.2595,
      "theoretical_loss": 4.421339383071178,
      "tokens_seen": 177537024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047787362086258776,
      "loss": 3.3193,
      "theoretical_loss": 4.421131427652145,
      "tokens_seen": 177602560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047786359077231695,
      "loss": 3.2873,
      "theoretical_loss": 4.420923570432407,
      "tokens_seen": 177668096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778535606820462,
      "loss": 3.1947,
      "theoretical_loss": 4.420715811329398,
      "tokens_seen": 177733632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778435305917753,
      "loss": 3.2383,
      "theoretical_loss": 4.420508150260651,
      "tokens_seen": 177799168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783350050150454,
      "loss": 3.1387,
      "theoretical_loss": 4.420300587143795,
      "tokens_seen": 177864704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047782347041123367,
      "loss": 2.8977,
      "theoretical_loss": 4.420093121896563,
      "tokens_seen": 177930240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778134403209629,
      "loss": 3.1988,
      "theoretical_loss": 4.419885754436784,
      "tokens_seen": 177995776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778034102306921,
      "loss": 3.0306,
      "theoretical_loss": 4.4196784846823896,
      "tokens_seen": 178061312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047779338014042127,
      "loss": 3.1273,
      "theoretical_loss": 4.419471312551408,
      "tokens_seen": 178126848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047778335005015045,
      "loss": 2.9744,
      "theoretical_loss": 4.419264237961966,
      "tokens_seen": 178192384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777733199598797,
      "loss": 3.0912,
      "theoretical_loss": 4.41905726083229,
      "tokens_seen": 178257920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777632898696088,
      "loss": 3.2138,
      "theoretical_loss": 4.418850381080706,
      "tokens_seen": 178323456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775325977933805,
      "loss": 3.2271,
      "theoretical_loss": 4.4186435986256365,
      "tokens_seen": 178388992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777432296890672,
      "loss": 3.0616,
      "theoretical_loss": 4.418436913385604,
      "tokens_seen": 178454528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777331995987964,
      "loss": 3.2644,
      "theoretical_loss": 4.418230325279228,
      "tokens_seen": 178520064
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 109542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0579349994659424,
      "objective/train/theoretical_loss": 4.418023834225226,
      "objective/train/tokens_used": 199045600,
      "theoretical_loss": 4.418023834225226,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777231695085256,
      "loss": 3.1012,
      "theoretical_loss": 4.418023834225226,
      "tokens_seen": 178585600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771313941825477,
      "loss": 3.1625,
      "theoretical_loss": 4.417817440142414,
      "tokens_seen": 178651136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047770310932798395,
      "loss": 3.0349,
      "theoretical_loss": 4.417611142949705,
      "tokens_seen": 178716672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047769307923771313,
      "loss": 3.2227,
      "theoretical_loss": 4.417404942566111,
      "tokens_seen": 178782208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776830491474423,
      "loss": 3.0844,
      "theoretical_loss": 4.417198838910737,
      "tokens_seen": 178847744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047767301905717155,
      "loss": 3.3573,
      "theoretical_loss": 4.4169928319027925,
      "tokens_seen": 178913280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776629889669007,
      "loss": 3.297,
      "theoretical_loss": 4.416786921461577,
      "tokens_seen": 178978816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776529588766299,
      "loss": 3.2404,
      "theoretical_loss": 4.41658110750649,
      "tokens_seen": 179044352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047764292878635904,
      "loss": 3.2353,
      "theoretical_loss": 4.416375389957029,
      "tokens_seen": 179109888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776328986960883,
      "loss": 3.2372,
      "theoretical_loss": 4.416169768732786,
      "tokens_seen": 179175424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047762286860581746,
      "loss": 3.1656,
      "theoretical_loss": 4.415964243753448,
      "tokens_seen": 179240960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761283851554664,
      "loss": 2.8394,
      "theoretical_loss": 4.415758814938803,
      "tokens_seen": 179306496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776028084252758,
      "loss": 3.3786,
      "theoretical_loss": 4.415553482208733,
      "tokens_seen": 179372032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047759277833500505,
      "loss": 3.2626,
      "theoretical_loss": 4.415348245483212,
      "tokens_seen": 179437568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775827482447342,
      "loss": 3.2584,
      "theoretical_loss": 4.415143104682317,
      "tokens_seen": 179503104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775727181544634,
      "loss": 3.2606,
      "theoretical_loss": 4.4149380597262144,
      "tokens_seen": 179568640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047756268806419254,
      "loss": 3.3959,
      "theoretical_loss": 4.41473311053517,
      "tokens_seen": 179634176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775526579739218,
      "loss": 3.1822,
      "theoretical_loss": 4.414528257029543,
      "tokens_seen": 179699712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047754262788365096,
      "loss": 3.3075,
      "theoretical_loss": 4.414323499129791,
      "tokens_seen": 179765248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753259779338014,
      "loss": 3.1764,
      "theoretical_loss": 4.414118836756462,
      "tokens_seen": 179830784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775225677031093,
      "loss": 3.2029,
      "theoretical_loss": 4.4139142698302,
      "tokens_seen": 179896320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775125376128385,
      "loss": 3.0117,
      "theoretical_loss": 4.413709798271748,
      "tokens_seen": 179961856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775025075225677,
      "loss": 3.0247,
      "theoretical_loss": 4.41350542200194,
      "tokens_seen": 180027392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774924774322969,
      "loss": 3.2042,
      "theoretical_loss": 4.413301140941703,
      "tokens_seen": 180092928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047748244734202605,
      "loss": 3.1478,
      "theoretical_loss": 4.413096955012063,
      "tokens_seen": 180158464
    },
    {
      "epoch": 0.05,
      "objective/train/docs_used": 110066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8472869396209717,
      "objective/train/theoretical_loss": 4.412892864134137,
      "objective/train/tokens_used": 200684000,
      "theoretical_loss": 4.412892864134137,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774724172517553,
      "loss": 3.0943,
      "theoretical_loss": 4.412892864134137,
      "tokens_seen": 180224000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774623871614845,
      "loss": 2.8881,
      "theoretical_loss": 4.412688868229136,
      "tokens_seen": 180289536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745235707121364,
      "loss": 3.3796,
      "theoretical_loss": 4.412484967218368,
      "tokens_seen": 180355072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774423269809429,
      "loss": 3.234,
      "theoretical_loss": 4.4122811610232295,
      "tokens_seen": 180420608
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477432296890672,
      "loss": 3.1805,
      "theoretical_loss": 4.412077449565216,
      "tokens_seen": 180486144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047742226680040124,
      "loss": 3.4462,
      "theoretical_loss": 4.411873832765913,
      "tokens_seen": 180551680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774122367101304,
      "loss": 3.1975,
      "theoretical_loss": 4.4116703105470005,
      "tokens_seen": 180617216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774022066198596,
      "loss": 3.1883,
      "theoretical_loss": 4.411466882830252,
      "tokens_seen": 180682752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773921765295888,
      "loss": 3.1599,
      "theoretical_loss": 4.411263549537534,
      "tokens_seen": 180748288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047738214643931797,
      "loss": 3.1214,
      "theoretical_loss": 4.411060310590805,
      "tokens_seen": 180813824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047737211634904715,
      "loss": 3.0375,
      "theoretical_loss": 4.410857165912116,
      "tokens_seen": 180879360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773620862587764,
      "loss": 3.0394,
      "theoretical_loss": 4.410654115423613,
      "tokens_seen": 180944896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773520561685055,
      "loss": 3.0477,
      "theoretical_loss": 4.410451159047533,
      "tokens_seen": 181010432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047734202607823474,
      "loss": 3.2956,
      "theoretical_loss": 4.410248296706204,
      "tokens_seen": 181075968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733199598796387,
      "loss": 3.1011,
      "theoretical_loss": 4.410045528322048,
      "tokens_seen": 181141504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773219658976931,
      "loss": 3.1455,
      "theoretical_loss": 4.409842853817577,
      "tokens_seen": 181207040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773119358074223,
      "loss": 3.1878,
      "theoretical_loss": 4.409640273115399,
      "tokens_seen": 181272576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047730190571715147,
      "loss": 2.8744,
      "theoretical_loss": 4.409437786138211,
      "tokens_seen": 181338112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729187562688065,
      "loss": 3.1718,
      "theoretical_loss": 4.4092353928087995,
      "tokens_seen": 181403648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004772818455366099,
      "loss": 3.1126,
      "theoretical_loss": 4.409033093050046,
      "tokens_seen": 181469184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477271815446339,
      "loss": 3.4133,
      "theoretical_loss": 4.408830886784922,
      "tokens_seen": 181534720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047726178535606825,
      "loss": 3.2217,
      "theoretical_loss": 4.40862877393649,
      "tokens_seen": 181600256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772517552657974,
      "loss": 3.1425,
      "theoretical_loss": 4.408426754427904,
      "tokens_seen": 181665792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772417251755266,
      "loss": 3.009,
      "theoretical_loss": 4.408224828182409,
      "tokens_seen": 181731328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772316950852558,
      "loss": 3.0163,
      "theoretical_loss": 4.40802299512334,
      "tokens_seen": 181796864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 111369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.439749002456665,
      "objective/train/theoretical_loss": 4.407821255174123,
      "objective/train/tokens_used": 202322400,
      "theoretical_loss": 4.407821255174123,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047722166499498497,
      "loss": 3.1176,
      "theoretical_loss": 4.407821255174123,
      "tokens_seen": 181862400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047721163490471415,
      "loss": 3.3191,
      "theoretical_loss": 4.407619608258274,
      "tokens_seen": 181927936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047720160481444333,
      "loss": 3.1351,
      "theoretical_loss": 4.407418054299402,
      "tokens_seen": 181993472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771915747241725,
      "loss": 3.2347,
      "theoretical_loss": 4.407216593221202,
      "tokens_seen": 182059008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047718154463390175,
      "loss": 3.182,
      "theoretical_loss": 4.407015224947463,
      "tokens_seen": 182124544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771715145436309,
      "loss": 3.3172,
      "theoretical_loss": 4.40681394940206,
      "tokens_seen": 182190080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771614844533601,
      "loss": 3.3428,
      "theoretical_loss": 4.406612766508961,
      "tokens_seen": 182255616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047715145436308924,
      "loss": 3.1138,
      "theoretical_loss": 4.4064116761922225,
      "tokens_seen": 182321152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771414242728185,
      "loss": 2.9224,
      "theoretical_loss": 4.406210678375992,
      "tokens_seen": 182386688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713139418254766,
      "loss": 3.0207,
      "theoretical_loss": 4.406009772984502,
      "tokens_seen": 182452224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047712136409227684,
      "loss": 2.9416,
      "theoretical_loss": 4.40580895994208,
      "tokens_seen": 182517760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477111334002006,
      "loss": 3.1591,
      "theoretical_loss": 4.405608239173138,
      "tokens_seen": 182583296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047710130391173525,
      "loss": 3.142,
      "theoretical_loss": 4.405407610602181,
      "tokens_seen": 182648832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770912738214644,
      "loss": 3.2257,
      "theoretical_loss": 4.405207074153797,
      "tokens_seen": 182714368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770812437311936,
      "loss": 3.0585,
      "theoretical_loss": 4.40500662975267,
      "tokens_seen": 182779904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047707121364092274,
      "loss": 3.0379,
      "theoretical_loss": 4.4048062773235666,
      "tokens_seen": 182845440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477061183550652,
      "loss": 3.018,
      "theoretical_loss": 4.404606016791345,
      "tokens_seen": 182910976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047705115346038116,
      "loss": 3.1405,
      "theoretical_loss": 4.404405848080951,
      "tokens_seen": 182976512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047704112337011034,
      "loss": 2.9844,
      "theoretical_loss": 4.404205771117417,
      "tokens_seen": 183042048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770310932798395,
      "loss": 3.0418,
      "theoretical_loss": 4.404005785825866,
      "tokens_seen": 183107584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770210631895687,
      "loss": 3.2314,
      "theoretical_loss": 4.403805892131508,
      "tokens_seen": 183173120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770110330992979,
      "loss": 3.2265,
      "theoretical_loss": 4.40360608995964,
      "tokens_seen": 183238656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770010030090271,
      "loss": 3.0284,
      "theoretical_loss": 4.403406379235646,
      "tokens_seen": 183304192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047699097291875625,
      "loss": 3.0799,
      "theoretical_loss": 4.403206759884999,
      "tokens_seen": 183369728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769809428284855,
      "loss": 3.0723,
      "theoretical_loss": 4.403007231833261,
      "tokens_seen": 183435264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 111925,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.625623941421509,
      "objective/train/theoretical_loss": 4.402807795006074,
      "objective/train/tokens_used": 203960800,
      "theoretical_loss": 4.402807795006074,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047697091273821466,
      "loss": 3.2865,
      "theoretical_loss": 4.402807795006074,
      "tokens_seen": 183500800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047696088264794384,
      "loss": 2.8014,
      "theoretical_loss": 4.402608449329178,
      "tokens_seen": 183566336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476950852557673,
      "loss": 3.007,
      "theoretical_loss": 4.40240919472839,
      "tokens_seen": 183631872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769408224674022,
      "loss": 3.0607,
      "theoretical_loss": 4.40221003112962,
      "tokens_seen": 183697408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769307923771314,
      "loss": 3.393,
      "theoretical_loss": 4.4020109584588605,
      "tokens_seen": 183762944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769207622868606,
      "loss": 3.3108,
      "theoretical_loss": 4.401811976642195,
      "tokens_seen": 183828480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047691073219658975,
      "loss": 2.9578,
      "theoretical_loss": 4.4016130856057885,
      "tokens_seen": 183894016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476900702106319,
      "loss": 3.2152,
      "theoretical_loss": 4.401414285275898,
      "tokens_seen": 183959552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768906720160481,
      "loss": 3.2252,
      "theoretical_loss": 4.401215575578861,
      "tokens_seen": 184025088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047688064192577735,
      "loss": 3.0185,
      "theoretical_loss": 4.4010169564411035,
      "tokens_seen": 184090624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047687061183550653,
      "loss": 3.0891,
      "theoretical_loss": 4.4008184277891385,
      "tokens_seen": 184156160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768605817452357,
      "loss": 3.1172,
      "theoretical_loss": 4.400619989549563,
      "tokens_seen": 184221696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768505516549649,
      "loss": 3.0661,
      "theoretical_loss": 4.400421641649061,
      "tokens_seen": 184287232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047684052156469407,
      "loss": 2.9647,
      "theoretical_loss": 4.400223384014399,
      "tokens_seen": 184352768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047683049147442325,
      "loss": 3.0501,
      "theoretical_loss": 4.400025216572435,
      "tokens_seen": 184418304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768204613841525,
      "loss": 2.9332,
      "theoretical_loss": 4.399827139250106,
      "tokens_seen": 184483840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768104312938816,
      "loss": 3.2431,
      "theoretical_loss": 4.399629151974436,
      "tokens_seen": 184549376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047680040120361085,
      "loss": 2.9531,
      "theoretical_loss": 4.399431254672535,
      "tokens_seen": 184614912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047679037111334003,
      "loss": 3.1576,
      "theoretical_loss": 4.399233447271598,
      "tokens_seen": 184680448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767803410230692,
      "loss": 3.1937,
      "theoretical_loss": 4.399035729698903,
      "tokens_seen": 184745984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767703109327984,
      "loss": 3.0762,
      "theoretical_loss": 4.398838101881815,
      "tokens_seen": 184811520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767602808425276,
      "loss": 3.1412,
      "theoretical_loss": 4.39864056374778,
      "tokens_seen": 184877056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047675025075225676,
      "loss": 3.2723,
      "theoretical_loss": 4.398443115224333,
      "tokens_seen": 184942592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476740220661986,
      "loss": 3.3477,
      "theoretical_loss": 4.398245756239088,
      "tokens_seen": 185008128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767301905717151,
      "loss": 2.7878,
      "theoretical_loss": 4.398048486719748,
      "tokens_seen": 185073664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 113096,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7102208137512207,
      "objective/train/theoretical_loss": 4.397851306594097,
      "objective/train/tokens_used": 205599200,
      "theoretical_loss": 4.397851306594097,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047672016048144435,
      "loss": 2.8609,
      "theoretical_loss": 4.397851306594097,
      "tokens_seen": 185139200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047671013039117353,
      "loss": 3.0009,
      "theoretical_loss": 4.397654215790003,
      "tokens_seen": 185204736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767001003009027,
      "loss": 3.029,
      "theoretical_loss": 4.397457214235419,
      "tokens_seen": 185270272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047669007021063195,
      "loss": 3.292,
      "theoretical_loss": 4.397260301858379,
      "tokens_seen": 185335808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766800401203611,
      "loss": 3.1928,
      "theoretical_loss": 4.397063478587005,
      "tokens_seen": 185401344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766700100300903,
      "loss": 3.3045,
      "theoretical_loss": 4.396866744349498,
      "tokens_seen": 185466880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047665997993981944,
      "loss": 3.1277,
      "theoretical_loss": 4.396670099074143,
      "tokens_seen": 185532416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766499498495487,
      "loss": 3.0829,
      "theoretical_loss": 4.39647354268931,
      "tokens_seen": 185597952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047663991975927786,
      "loss": 3.2087,
      "theoretical_loss": 4.39627707512345,
      "tokens_seen": 185663488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662988966900704,
      "loss": 2.8462,
      "theoretical_loss": 4.3960806963050985,
      "tokens_seen": 185729024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766198595787362,
      "loss": 3.0272,
      "theoretical_loss": 4.395884406162873,
      "tokens_seen": 185794560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047660982948846545,
      "loss": 3.2038,
      "theoretical_loss": 4.395688204625471,
      "tokens_seen": 185860096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765997993981946,
      "loss": 2.8081,
      "theoretical_loss": 4.395492091621677,
      "tokens_seen": 185925632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765897693079238,
      "loss": 3.3035,
      "theoretical_loss": 4.395296067080356,
      "tokens_seen": 185991168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047657973921765294,
      "loss": 3.1633,
      "theoretical_loss": 4.3951001309304525,
      "tokens_seen": 186056704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765697091273822,
      "loss": 3.145,
      "theoretical_loss": 4.394904283100997,
      "tokens_seen": 186122240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047655967903711136,
      "loss": 3.2733,
      "theoretical_loss": 4.394708523521101,
      "tokens_seen": 186187776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654964894684054,
      "loss": 3.1495,
      "theoretical_loss": 4.394512852119956,
      "tokens_seen": 186253312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765396188565697,
      "loss": 3.1897,
      "theoretical_loss": 4.3943172688268355,
      "tokens_seen": 186318848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765295887662989,
      "loss": 3.0105,
      "theoretical_loss": 4.394121773571097,
      "tokens_seen": 186384384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765195586760281,
      "loss": 3.1493,
      "theoretical_loss": 4.3939263662821775,
      "tokens_seen": 186449920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765095285857573,
      "loss": 2.9624,
      "theoretical_loss": 4.393731046889595,
      "tokens_seen": 186515456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047649949849548645,
      "loss": 3.3138,
      "theoretical_loss": 4.3935358153229505,
      "tokens_seen": 186580992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764894684052157,
      "loss": 3.1593,
      "theoretical_loss": 4.3933406715119245,
      "tokens_seen": 186646528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047647943831494486,
      "loss": 3.0466,
      "theoretical_loss": 4.393145615386279,
      "tokens_seen": 186712064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 113525,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.944962978363037,
      "objective/train/theoretical_loss": 4.392950646875857,
      "objective/train/tokens_used": 207237600,
      "theoretical_loss": 4.392950646875857,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047646940822467404,
      "loss": 3.0963,
      "theoretical_loss": 4.392950646875857,
      "tokens_seen": 186777600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764593781344032,
      "loss": 3.0612,
      "theoretical_loss": 4.392755765910581,
      "tokens_seen": 186843136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764493480441324,
      "loss": 3.1482,
      "theoretical_loss": 4.3925609724204575,
      "tokens_seen": 186908672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764393179538616,
      "loss": 3.2602,
      "theoretical_loss": 4.392366266335569,
      "tokens_seen": 186974208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764292878635908,
      "loss": 3.1856,
      "theoretical_loss": 4.392171647586081,
      "tokens_seen": 187039744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047641925777331995,
      "loss": 3.0132,
      "theoretical_loss": 4.39197711610224,
      "tokens_seen": 187105280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764092276830492,
      "loss": 3.152,
      "theoretical_loss": 4.391782671814369,
      "tokens_seen": 187170816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763991975927783,
      "loss": 3.2149,
      "theoretical_loss": 4.391588314652875,
      "tokens_seen": 187236352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047638916750250755,
      "loss": 3.2574,
      "theoretical_loss": 4.3913940445482424,
      "tokens_seen": 187301888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047637913741223673,
      "loss": 3.1899,
      "theoretical_loss": 4.391199861431037,
      "tokens_seen": 187367424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763691073219659,
      "loss": 3.0975,
      "theoretical_loss": 4.3910057652319034,
      "tokens_seen": 187432960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763590772316951,
      "loss": 3.2897,
      "theoretical_loss": 4.390811755881566,
      "tokens_seen": 187498496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047634904714142427,
      "loss": 3.3427,
      "theoretical_loss": 4.390617833310827,
      "tokens_seen": 187564032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047633901705115345,
      "loss": 3.181,
      "theoretical_loss": 4.390423997450571,
      "tokens_seen": 187629568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763289869608827,
      "loss": 3.1201,
      "theoretical_loss": 4.3902302482317594,
      "tokens_seen": 187695104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763189568706118,
      "loss": 3.12,
      "theoretical_loss": 4.390036585585432,
      "tokens_seen": 187760640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047630892678034105,
      "loss": 2.9343,
      "theoretical_loss": 4.3898430094427106,
      "tokens_seen": 187826176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047629889669007023,
      "loss": 3.0305,
      "theoretical_loss": 4.389649519734793,
      "tokens_seen": 187891712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762888665997994,
      "loss": 3.0496,
      "theoretical_loss": 4.3894561163929575,
      "tokens_seen": 187957248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762788365095286,
      "loss": 3.2563,
      "theoretical_loss": 4.389262799348559,
      "tokens_seen": 188022784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762688064192578,
      "loss": 3.2352,
      "theoretical_loss": 4.389069568533033,
      "tokens_seen": 188088320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047625877632898696,
      "loss": 3.1003,
      "theoretical_loss": 4.388876423877893,
      "tokens_seen": 188153856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762487462387162,
      "loss": 3.1905,
      "theoretical_loss": 4.388683365314728,
      "tokens_seen": 188219392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762387161484453,
      "loss": 3.3903,
      "theoretical_loss": 4.388490392775209,
      "tokens_seen": 188284928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047622868605817455,
      "loss": 3.2064,
      "theoretical_loss": 4.3882975061910825,
      "tokens_seen": 188350464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 114652,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.638829469680786,
      "objective/train/theoretical_loss": 4.3881047054941735,
      "objective/train/tokens_used": 208876000,
      "theoretical_loss": 4.3881047054941735,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762186559679037,
      "loss": 3.0819,
      "theoretical_loss": 4.3881047054941735,
      "tokens_seen": 188416000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762086258776329,
      "loss": 2.914,
      "theoretical_loss": 4.387911990616386,
      "tokens_seen": 188481536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761985957873621,
      "loss": 3.038,
      "theoretical_loss": 4.387719361489699,
      "tokens_seen": 188547072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761885656970913,
      "loss": 2.9676,
      "theoretical_loss": 4.3875268180461715,
      "tokens_seen": 188612608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047617853560682046,
      "loss": 3.2529,
      "theoretical_loss": 4.387334360217939,
      "tokens_seen": 188678144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047616850551654964,
      "loss": 3.1968,
      "theoretical_loss": 4.3871419879372135,
      "tokens_seen": 188743680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761584754262788,
      "loss": 3.3051,
      "theoretical_loss": 4.3869497011362855,
      "tokens_seen": 188809216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614844533600806,
      "loss": 3.1021,
      "theoretical_loss": 4.386757499747521,
      "tokens_seen": 188874752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761384152457372,
      "loss": 2.8122,
      "theoretical_loss": 4.386565383703366,
      "tokens_seen": 188940288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761283851554664,
      "loss": 3.308,
      "theoretical_loss": 4.38637335293634,
      "tokens_seen": 189005824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761183550651956,
      "loss": 3.2071,
      "theoretical_loss": 4.3861814073790395,
      "tokens_seen": 189071360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761083249749248,
      "loss": 3.0571,
      "theoretical_loss": 4.38598954696414,
      "tokens_seen": 189136896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047609829488465396,
      "loss": 3.2339,
      "theoretical_loss": 4.385797771624393,
      "tokens_seen": 189202432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047608826479438314,
      "loss": 2.9145,
      "theoretical_loss": 4.385606081292623,
      "tokens_seen": 189267968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760782347041123,
      "loss": 2.9952,
      "theoretical_loss": 4.385414475901737,
      "tokens_seen": 189333504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047606820461384156,
      "loss": 3.4306,
      "theoretical_loss": 4.385222955384711,
      "tokens_seen": 189399040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760581745235707,
      "loss": 3.2663,
      "theoretical_loss": 4.385031519674601,
      "tokens_seen": 189464576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760481444332999,
      "loss": 3.267,
      "theoretical_loss": 4.38484016870454,
      "tokens_seen": 189530112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047603811434302905,
      "loss": 3.1402,
      "theoretical_loss": 4.384648902407734,
      "tokens_seen": 189595648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760280842527583,
      "loss": 3.3339,
      "theoretical_loss": 4.384457720717466,
      "tokens_seen": 189661184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047601805416248746,
      "loss": 3.3079,
      "theoretical_loss": 4.384266623567096,
      "tokens_seen": 189726720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047600802407221665,
      "loss": 3.0988,
      "theoretical_loss": 4.384075610890057,
      "tokens_seen": 189792256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759979939819458,
      "loss": 3.1992,
      "theoretical_loss": 4.3838846826198585,
      "tokens_seen": 189857792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047598796389167506,
      "loss": 3.1131,
      "theoretical_loss": 4.383693838690085,
      "tokens_seen": 189923328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759779338014042,
      "loss": 3.2468,
      "theoretical_loss": 4.383503079034397,
      "tokens_seen": 189988864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 115651,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9877991676330566,
      "objective/train/theoretical_loss": 4.383312403586528,
      "objective/train/tokens_used": 210514400,
      "theoretical_loss": 4.383312403586528,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759679037111334,
      "loss": 3.0685,
      "theoretical_loss": 4.383312403586528,
      "tokens_seen": 190054400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759578736208626,
      "loss": 2.9333,
      "theoretical_loss": 4.38312181228029,
      "tokens_seen": 190119936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759478435305918,
      "loss": 3.366,
      "theoretical_loss": 4.382931305049566,
      "tokens_seen": 190185472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475937813440321,
      "loss": 3.087,
      "theoretical_loss": 4.382740881828316,
      "tokens_seen": 190251008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047592778335005015,
      "loss": 3.0998,
      "theoretical_loss": 4.3825505425505735,
      "tokens_seen": 190316544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759177532597794,
      "loss": 3.172,
      "theoretical_loss": 4.382360287150449,
      "tokens_seen": 190382080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759077231695085,
      "loss": 3.2168,
      "theoretical_loss": 4.382170115562122,
      "tokens_seen": 190447616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047589769307923775,
      "loss": 3.2619,
      "theoretical_loss": 4.381980027719852,
      "tokens_seen": 190513152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047588766298896693,
      "loss": 3.1796,
      "theoretical_loss": 4.3817900235579685,
      "tokens_seen": 190578688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758776328986961,
      "loss": 3.0825,
      "theoretical_loss": 4.381600103010879,
      "tokens_seen": 190644224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758676028084253,
      "loss": 2.9786,
      "theoretical_loss": 4.381410266013061,
      "tokens_seen": 190709760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047585757271815447,
      "loss": 3.2466,
      "theoretical_loss": 4.381220512499068,
      "tokens_seen": 190775296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047584754262788365,
      "loss": 3.0826,
      "theoretical_loss": 4.381030842403527,
      "tokens_seen": 190840832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758375125376129,
      "loss": 3.2852,
      "theoretical_loss": 4.38084125566114,
      "tokens_seen": 190906368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475827482447342,
      "loss": 3.3628,
      "theoretical_loss": 4.380651752206679,
      "tokens_seen": 190971904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047581745235707125,
      "loss": 2.9071,
      "theoretical_loss": 4.380462331974991,
      "tokens_seen": 191037440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047580742226680043,
      "loss": 2.9733,
      "theoretical_loss": 4.380272994900998,
      "tokens_seen": 191102976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757973921765296,
      "loss": 3.2946,
      "theoretical_loss": 4.380083740919693,
      "tokens_seen": 191168512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757873620862588,
      "loss": 3.0393,
      "theoretical_loss": 4.379894569966144,
      "tokens_seen": 191234048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475777331995988,
      "loss": 3.2391,
      "theoretical_loss": 4.37970548197549,
      "tokens_seen": 191299584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047576730190571716,
      "loss": 3.0312,
      "theoretical_loss": 4.3795164768829435,
      "tokens_seen": 191365120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757572718154464,
      "loss": 3.0777,
      "theoretical_loss": 4.37932755462379,
      "tokens_seen": 191430656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757472417251755,
      "loss": 3.3053,
      "theoretical_loss": 4.379138715133389,
      "tokens_seen": 191496192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047573721163490475,
      "loss": 3.0129,
      "theoretical_loss": 4.378949958347171,
      "tokens_seen": 191561728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757271815446339,
      "loss": 3.1993,
      "theoretical_loss": 4.378761284200638,
      "tokens_seen": 191627264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 116251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.686540126800537,
      "objective/train/theoretical_loss": 4.3785726926293655,
      "objective/train/tokens_used": 212152800,
      "theoretical_loss": 4.3785726926293655,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757171514543631,
      "loss": 2.8721,
      "theoretical_loss": 4.3785726926293655,
      "tokens_seen": 191692800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757071213640923,
      "loss": 3.144,
      "theoretical_loss": 4.378384183569002,
      "tokens_seen": 191758336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756970912738215,
      "loss": 3.2052,
      "theoretical_loss": 4.378195756955267,
      "tokens_seen": 191823872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047568706118355066,
      "loss": 3.192,
      "theoretical_loss": 4.378007412723953,
      "tokens_seen": 191889408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047567703109327984,
      "loss": 2.9825,
      "theoretical_loss": 4.377819150810923,
      "tokens_seen": 191954944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475667001003009,
      "loss": 2.9849,
      "theoretical_loss": 4.377630971152113,
      "tokens_seen": 192020480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047565697091273826,
      "loss": 3.2592,
      "theoretical_loss": 4.377442873683529,
      "tokens_seen": 192086016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756469408224674,
      "loss": 3.0667,
      "theoretical_loss": 4.377254858341252,
      "tokens_seen": 192151552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756369107321966,
      "loss": 3.194,
      "theoretical_loss": 4.377066925061431,
      "tokens_seen": 192217088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756268806419258,
      "loss": 3.2189,
      "theoretical_loss": 4.376879073780288,
      "tokens_seen": 192282624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475616850551655,
      "loss": 3.0755,
      "theoretical_loss": 4.376691304434117,
      "tokens_seen": 192348160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047560682046138416,
      "loss": 3.1954,
      "theoretical_loss": 4.376503616959281,
      "tokens_seen": 192413696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047559679037111334,
      "loss": 3.233,
      "theoretical_loss": 4.376316011292214,
      "tokens_seen": 192479232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755867602808425,
      "loss": 3.0876,
      "theoretical_loss": 4.376128487369426,
      "tokens_seen": 192544768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047557673019057176,
      "loss": 2.9403,
      "theoretical_loss": 4.375941045127492,
      "tokens_seen": 192610304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755667001003009,
      "loss": 3.3601,
      "theoretical_loss": 4.375753684503059,
      "tokens_seen": 192675840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755566700100301,
      "loss": 2.9828,
      "theoretical_loss": 4.3755664054328465,
      "tokens_seen": 192741376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047554663991975925,
      "loss": 2.8375,
      "theoretical_loss": 4.375379207853644,
      "tokens_seen": 192806912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755366098294885,
      "loss": 3.323,
      "theoretical_loss": 4.375192091702311,
      "tokens_seen": 192872448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047552657973921766,
      "loss": 3.0783,
      "theoretical_loss": 4.375005056915777,
      "tokens_seen": 192937984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047551654964894685,
      "loss": 2.974,
      "theoretical_loss": 4.374818103431042,
      "tokens_seen": 193003520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047550651955867603,
      "loss": 3.0758,
      "theoretical_loss": 4.374631231185177,
      "tokens_seen": 193069056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047549648946840526,
      "loss": 3.0464,
      "theoretical_loss": 4.374444440115322,
      "tokens_seen": 193134592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754864593781344,
      "loss": 3.1286,
      "theoretical_loss": 4.374257730158687,
      "tokens_seen": 193200128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754764292878636,
      "loss": 3.1677,
      "theoretical_loss": 4.374071101252554,
      "tokens_seen": 193265664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 117431,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7149760723114014,
      "objective/train/theoretical_loss": 4.373884553334271,
      "objective/train/tokens_used": 213791200,
      "theoretical_loss": 4.373884553334271,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047546639919759275,
      "loss": 3.1442,
      "theoretical_loss": 4.373884553334271,
      "tokens_seen": 193331200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475456369107322,
      "loss": 3.1432,
      "theoretical_loss": 4.373698086341258,
      "tokens_seen": 193396736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047544633901705117,
      "loss": 3.2882,
      "theoretical_loss": 4.373511700211004,
      "tokens_seen": 193462272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047543630892678035,
      "loss": 3.1277,
      "theoretical_loss": 4.373325394881069,
      "tokens_seen": 193527808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047542627883650953,
      "loss": 2.8819,
      "theoretical_loss": 4.373139170289079,
      "tokens_seen": 193593344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754162487462387,
      "loss": 3.1914,
      "theoretical_loss": 4.3729530263727305,
      "tokens_seen": 193658880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754062186559679,
      "loss": 3.0863,
      "theoretical_loss": 4.372766963069793,
      "tokens_seen": 193724416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047539618856569713,
      "loss": 3.1611,
      "theoretical_loss": 4.3725809803181,
      "tokens_seen": 193789952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047538615847542625,
      "loss": 3.1407,
      "theoretical_loss": 4.372395078055554,
      "tokens_seen": 193855488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753761283851555,
      "loss": 3.2841,
      "theoretical_loss": 4.372209256220131,
      "tokens_seen": 193921024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753660982948846,
      "loss": 3.2507,
      "theoretical_loss": 4.3720235147498725,
      "tokens_seen": 193986560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047535606820461385,
      "loss": 3.2756,
      "theoretical_loss": 4.371837853582887,
      "tokens_seen": 194052096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047534603811434303,
      "loss": 2.9721,
      "theoretical_loss": 4.3716522726573555,
      "tokens_seen": 194117632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753360080240722,
      "loss": 3.1121,
      "theoretical_loss": 4.371466771911524,
      "tokens_seen": 194183168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753259779338014,
      "loss": 3.07,
      "theoretical_loss": 4.371281351283709,
      "tokens_seen": 194248704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047531594784353063,
      "loss": 3.0433,
      "theoretical_loss": 4.371096010712295,
      "tokens_seen": 194314240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047530591775325976,
      "loss": 3.1014,
      "theoretical_loss": 4.370910750135733,
      "tokens_seen": 194379776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475295887662989,
      "loss": 3.3776,
      "theoretical_loss": 4.370725569492545,
      "tokens_seen": 194445312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752858575727181,
      "loss": 3.2206,
      "theoretical_loss": 4.370540468721318,
      "tokens_seen": 194510848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047527582748244736,
      "loss": 3.1334,
      "theoretical_loss": 4.370355447760708,
      "tokens_seen": 194576384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047526579739217654,
      "loss": 3.1063,
      "theoretical_loss": 4.370170506549439,
      "tokens_seen": 194641920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752557673019057,
      "loss": 3.133,
      "theoretical_loss": 4.369985645026302,
      "tokens_seen": 194707456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752457372116349,
      "loss": 3.1316,
      "theoretical_loss": 4.369800863130157,
      "tokens_seen": 194772992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752357071213641,
      "loss": 3.0326,
      "theoretical_loss": 4.36961616079993,
      "tokens_seen": 194838528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047522567703109326,
      "loss": 3.2395,
      "theoretical_loss": 4.3694315379746165,
      "tokens_seen": 194904064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 118646,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6308515071868896,
      "objective/train/theoretical_loss": 4.369246994593274,
      "objective/train/tokens_used": 215429600,
      "theoretical_loss": 4.369246994593274,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752156469408225,
      "loss": 3.2026,
      "theoretical_loss": 4.369246994593274,
      "tokens_seen": 194969600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752056168505517,
      "loss": 3.152,
      "theoretical_loss": 4.369062530595034,
      "tokens_seen": 195035136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047519558676028086,
      "loss": 2.9739,
      "theoretical_loss": 4.368878145919091,
      "tokens_seen": 195100672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047518555667001004,
      "loss": 3.0244,
      "theoretical_loss": 4.368693840504706,
      "tokens_seen": 195166208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751755265797392,
      "loss": 2.816,
      "theoretical_loss": 4.368509614291209,
      "tokens_seen": 195231744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047516549648946846,
      "loss": 3.0789,
      "theoretical_loss": 4.368325467217998,
      "tokens_seen": 195297280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751554663991976,
      "loss": 3.1642,
      "theoretical_loss": 4.368141399224534,
      "tokens_seen": 195362816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751454363089268,
      "loss": 2.9522,
      "theoretical_loss": 4.367957410250345,
      "tokens_seen": 195428352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475135406218656,
      "loss": 3.0048,
      "theoretical_loss": 4.367773500235029,
      "tokens_seen": 195493888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751253761283852,
      "loss": 3.172,
      "theoretical_loss": 4.367589669118248,
      "tokens_seen": 195559424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047511534603811436,
      "loss": 3.02,
      "theoretical_loss": 4.367405916839729,
      "tokens_seen": 195624960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047510531594784354,
      "loss": 3.0327,
      "theoretical_loss": 4.367222243339268,
      "tokens_seen": 195690496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750952858575727,
      "loss": 3.3324,
      "theoretical_loss": 4.367038648556725,
      "tokens_seen": 195756032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047508525576730196,
      "loss": 3.2043,
      "theoretical_loss": 4.366855132432028,
      "tokens_seen": 195821568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750752256770311,
      "loss": 3.2153,
      "theoretical_loss": 4.366671694905169,
      "tokens_seen": 195887104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750651955867603,
      "loss": 3.2576,
      "theoretical_loss": 4.366488335916207,
      "tokens_seen": 195952640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047505516549648945,
      "loss": 2.9306,
      "theoretical_loss": 4.366305055405267,
      "tokens_seen": 196018176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750451354062187,
      "loss": 3.163,
      "theoretical_loss": 4.366121853312537,
      "tokens_seen": 196083712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047503510531594787,
      "loss": 3.2586,
      "theoretical_loss": 4.365938729578274,
      "tokens_seen": 196149248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047502507522567705,
      "loss": 3.313,
      "theoretical_loss": 4.365755684142799,
      "tokens_seen": 196214784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047501504513540623,
      "loss": 3.2721,
      "theoretical_loss": 4.365572716946499,
      "tokens_seen": 196280320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047500501504513546,
      "loss": 3.3105,
      "theoretical_loss": 4.365389827929825,
      "tokens_seen": 196345856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749949849548646,
      "loss": 3.2141,
      "theoretical_loss": 4.365207017033293,
      "tokens_seen": 196411392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749849548645938,
      "loss": 3.2766,
      "theoretical_loss": 4.365024284197486,
      "tokens_seen": 196476928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047497492477432295,
      "loss": 2.9995,
      "theoretical_loss": 4.364841629363051,
      "tokens_seen": 196542464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 119101,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2888681888580322,
      "objective/train/theoretical_loss": 4.364659052470699,
      "objective/train/tokens_used": 217068000,
      "theoretical_loss": 4.364659052470699,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749648946840522,
      "loss": 3.1609,
      "theoretical_loss": 4.364659052470699,
      "tokens_seen": 196608000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047495486459378137,
      "loss": 3.1296,
      "theoretical_loss": 4.364476553461208,
      "tokens_seen": 196673536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494483450351055,
      "loss": 3.062,
      "theoretical_loss": 4.364294132275419,
      "tokens_seen": 196739072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047493480441323973,
      "loss": 3.1813,
      "theoretical_loss": 4.364111788854237,
      "tokens_seen": 196804608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749247743229689,
      "loss": 3.0742,
      "theoretical_loss": 4.3639295231386335,
      "tokens_seen": 196870144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749147442326981,
      "loss": 3.1893,
      "theoretical_loss": 4.363747335069642,
      "tokens_seen": 196935680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047490471414242733,
      "loss": 3.2753,
      "theoretical_loss": 4.363565224588364,
      "tokens_seen": 197001216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047489468405215646,
      "loss": 3.1104,
      "theoretical_loss": 4.363383191635961,
      "tokens_seen": 197066752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748846539618857,
      "loss": 3.0305,
      "theoretical_loss": 4.363201236153663,
      "tokens_seen": 197132288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748746238716148,
      "loss": 3.0246,
      "theoretical_loss": 4.363019358082759,
      "tokens_seen": 197197824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047486459378134405,
      "loss": 3.1562,
      "theoretical_loss": 4.362837557364608,
      "tokens_seen": 197263360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047485456369107323,
      "loss": 3.1837,
      "theoretical_loss": 4.362655833940627,
      "tokens_seen": 197328896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748445336008024,
      "loss": 3.2525,
      "theoretical_loss": 4.3624741877523,
      "tokens_seen": 197394432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748345035105316,
      "loss": 3.1681,
      "theoretical_loss": 4.362292618741174,
      "tokens_seen": 197459968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047482447342026083,
      "loss": 3.4425,
      "theoretical_loss": 4.362111126848862,
      "tokens_seen": 197525504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047481444332998996,
      "loss": 3.3758,
      "theoretical_loss": 4.3619297120170355,
      "tokens_seen": 197591040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748044132397192,
      "loss": 3.1236,
      "theoretical_loss": 4.361748374187433,
      "tokens_seen": 197656576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747943831494483,
      "loss": 2.9103,
      "theoretical_loss": 4.361567113301858,
      "tokens_seen": 197722112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478435305917756,
      "loss": 3.3328,
      "theoretical_loss": 4.3613859293021715,
      "tokens_seen": 197787648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047477432296890674,
      "loss": 3.2619,
      "theoretical_loss": 4.361204822130304,
      "tokens_seen": 197853184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747642928786359,
      "loss": 2.9688,
      "theoretical_loss": 4.361023791728243,
      "tokens_seen": 197918720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747542627883651,
      "loss": 2.8798,
      "theoretical_loss": 4.360842838038045,
      "tokens_seen": 197984256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747442326980943,
      "loss": 3.164,
      "theoretical_loss": 4.360661961001824,
      "tokens_seen": 198049792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047473420260782346,
      "loss": 3.2726,
      "theoretical_loss": 4.360481160561761,
      "tokens_seen": 198115328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747241725175527,
      "loss": 3.2962,
      "theoretical_loss": 4.360300436660099,
      "tokens_seen": 198180864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 119264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4219841957092285,
      "objective/train/theoretical_loss": 4.36011978923914,
      "objective/train/tokens_used": 218706400,
      "theoretical_loss": 4.36011978923914,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747141424272818,
      "loss": 3.2649,
      "theoretical_loss": 4.36011978923914,
      "tokens_seen": 198246400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470411233701106,
      "loss": 3.1434,
      "theoretical_loss": 4.359939218241253,
      "tokens_seen": 198311936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746940822467402,
      "loss": 3.3527,
      "theoretical_loss": 4.359758723608866,
      "tokens_seen": 198377472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746840521564694,
      "loss": 3.2825,
      "theoretical_loss": 4.359578305284474,
      "tokens_seen": 198443008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746740220661986,
      "loss": 3.3552,
      "theoretical_loss": 4.359397963210629,
      "tokens_seen": 198508544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746639919759278,
      "loss": 3.2728,
      "theoretical_loss": 4.3592176973299495,
      "tokens_seen": 198574080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047465396188565696,
      "loss": 2.9474,
      "theoretical_loss": 4.359037507585111,
      "tokens_seen": 198639616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746439317953862,
      "loss": 2.9234,
      "theoretical_loss": 4.358857393918857,
      "tokens_seen": 198705152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746339017051153,
      "loss": 2.9143,
      "theoretical_loss": 4.358677356273988,
      "tokens_seen": 198770688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047462387161484456,
      "loss": 3.1565,
      "theoretical_loss": 4.358497394593371,
      "tokens_seen": 198836224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746138415245737,
      "loss": 3.118,
      "theoretical_loss": 4.35831750881993,
      "tokens_seen": 198901760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746038114343029,
      "loss": 3.2261,
      "theoretical_loss": 4.358137698896653,
      "tokens_seen": 198967296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745937813440321,
      "loss": 3.1764,
      "theoretical_loss": 4.35795796476659,
      "tokens_seen": 199032832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745837512537613,
      "loss": 3.3871,
      "theoretical_loss": 4.357778306372852,
      "tokens_seen": 199098368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047457372116349047,
      "loss": 3.1078,
      "theoretical_loss": 4.357598723658612,
      "tokens_seen": 199163904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047456369107321965,
      "loss": 3.0791,
      "theoretical_loss": 4.357419216567102,
      "tokens_seen": 199229440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047455366098294883,
      "loss": 3.2517,
      "theoretical_loss": 4.357239785041618,
      "tokens_seen": 199294976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454363089267807,
      "loss": 3.176,
      "theoretical_loss": 4.357060429025516,
      "tokens_seen": 199360512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745336008024072,
      "loss": 3.1573,
      "theoretical_loss": 4.356881148462213,
      "tokens_seen": 199426048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047452357071213643,
      "loss": 2.9391,
      "theoretical_loss": 4.356701943295186,
      "tokens_seen": 199491584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047451354062186555,
      "loss": 3.0178,
      "theoretical_loss": 4.356522813467976,
      "tokens_seen": 199557120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745035105315948,
      "loss": 3.008,
      "theoretical_loss": 4.356343758924181,
      "tokens_seen": 199622656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047449348044132397,
      "loss": 3.088,
      "theoretical_loss": 4.356164779607463,
      "tokens_seen": 199688192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047448345035105315,
      "loss": 3.0246,
      "theoretical_loss": 4.355985875461542,
      "tokens_seen": 199753728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744734202607824,
      "loss": 3.2933,
      "theoretical_loss": 4.3558070464302,
      "tokens_seen": 199819264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 120457,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7158637046813965,
      "objective/train/theoretical_loss": 4.355628292457278,
      "objective/train/tokens_used": 220344800,
      "theoretical_loss": 4.355628292457278,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047446339017051157,
      "loss": 3.2748,
      "theoretical_loss": 4.355628292457278,
      "tokens_seen": 199884800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047445336008024075,
      "loss": 3.3399,
      "theoretical_loss": 4.355449613486681,
      "tokens_seen": 199950336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444332998996993,
      "loss": 2.9307,
      "theoretical_loss": 4.355271009462371,
      "tokens_seen": 200015872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744332998996991,
      "loss": 2.8099,
      "theoretical_loss": 4.355092480328369,
      "tokens_seen": 200081408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744232698094283,
      "loss": 3.2346,
      "theoretical_loss": 4.35491402602876,
      "tokens_seen": 200146944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047441323971915753,
      "loss": 3.1045,
      "theoretical_loss": 4.354735646507687,
      "tokens_seen": 200212480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047440320962888666,
      "loss": 3.1639,
      "theoretical_loss": 4.354557341709354,
      "tokens_seen": 200278016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743931795386159,
      "loss": 3.1702,
      "theoretical_loss": 4.354379111578022,
      "tokens_seen": 200343552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474383149448345,
      "loss": 3.1362,
      "theoretical_loss": 4.3542009560580155,
      "tokens_seen": 200409088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047437311935807425,
      "loss": 3.1808,
      "theoretical_loss": 4.354022875093716,
      "tokens_seen": 200474624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436308926780343,
      "loss": 3.2076,
      "theoretical_loss": 4.353844868629566,
      "tokens_seen": 200540160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743530591775326,
      "loss": 3.1964,
      "theoretical_loss": 4.353666936610068,
      "tokens_seen": 200605696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743430290872618,
      "loss": 3.0344,
      "theoretical_loss": 4.3534890789797815,
      "tokens_seen": 200671232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047433299899699103,
      "loss": 3.0296,
      "theoretical_loss": 4.353311295683328,
      "tokens_seen": 200736768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432296890672016,
      "loss": 3.0211,
      "theoretical_loss": 4.353133586665388,
      "tokens_seen": 200802304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743129388164494,
      "loss": 3.2548,
      "theoretical_loss": 4.352955951870699,
      "tokens_seen": 200867840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743029087261785,
      "loss": 3.09,
      "theoretical_loss": 4.352778391244061,
      "tokens_seen": 200933376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047429287863590776,
      "loss": 2.9689,
      "theoretical_loss": 4.352600904730329,
      "tokens_seen": 200998912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047428284854563694,
      "loss": 3.4101,
      "theoretical_loss": 4.35242349227442,
      "tokens_seen": 201064448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742728184553661,
      "loss": 3.288,
      "theoretical_loss": 4.35224615382131,
      "tokens_seen": 201129984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742627883650953,
      "loss": 3.1146,
      "theoretical_loss": 4.352068889316032,
      "tokens_seen": 201195520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742527582748245,
      "loss": 3.0871,
      "theoretical_loss": 4.351891698703679,
      "tokens_seen": 201261056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047424272818455366,
      "loss": 3.2115,
      "theoretical_loss": 4.351714581929402,
      "tokens_seen": 201326592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742326980942829,
      "loss": 2.8483,
      "theoretical_loss": 4.351537538938409,
      "tokens_seen": 201392128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474222668004012,
      "loss": 3.0209,
      "theoretical_loss": 4.351360569675971,
      "tokens_seen": 201457664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 121711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.200042724609375,
      "objective/train/theoretical_loss": 4.351183674087412,
      "objective/train/tokens_used": 221983200,
      "theoretical_loss": 4.351183674087412,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047421263791374126,
      "loss": 3.0898,
      "theoretical_loss": 4.351183674087412,
      "tokens_seen": 201523200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742026078234704,
      "loss": 3.0786,
      "theoretical_loss": 4.35100685211812,
      "tokens_seen": 201588736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741925777331996,
      "loss": 3.155,
      "theoretical_loss": 4.350830103713535,
      "tokens_seen": 201654272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741825476429288,
      "loss": 3.1987,
      "theoretical_loss": 4.350653428819159,
      "tokens_seen": 201719808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474172517552658,
      "loss": 3.0476,
      "theoretical_loss": 4.350476827380553,
      "tokens_seen": 201785344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416248746238716,
      "loss": 3.157,
      "theoretical_loss": 4.350300299343332,
      "tokens_seen": 201850880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741524573721164,
      "loss": 3.0758,
      "theoretical_loss": 4.350123844653173,
      "tokens_seen": 201916416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741424272818455,
      "loss": 3.2307,
      "theoretical_loss": 4.349947463255808,
      "tokens_seen": 201981952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047413239719157476,
      "loss": 3.1247,
      "theoretical_loss": 4.349771155097029,
      "tokens_seen": 202047488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741223671013039,
      "loss": 3.1268,
      "theoretical_loss": 4.349594920122682,
      "tokens_seen": 202113024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741123370110331,
      "loss": 3.1097,
      "theoretical_loss": 4.349418758278675,
      "tokens_seen": 202178560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741023069207623,
      "loss": 3.139,
      "theoretical_loss": 4.349242669510971,
      "tokens_seen": 202244096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740922768304915,
      "loss": 3.1241,
      "theoretical_loss": 4.3490666537655915,
      "tokens_seen": 202309632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047408224674022067,
      "loss": 3.1542,
      "theoretical_loss": 4.348890710988615,
      "tokens_seen": 202375168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047407221664994985,
      "loss": 3.2895,
      "theoretical_loss": 4.348714841126174,
      "tokens_seen": 202440704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047406218655967903,
      "loss": 3.1676,
      "theoretical_loss": 4.348539044124465,
      "tokens_seen": 202506240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047405215646940827,
      "loss": 2.9643,
      "theoretical_loss": 4.348363319929735,
      "tokens_seen": 202571776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740421263791374,
      "loss": 3.0266,
      "theoretical_loss": 4.348187668488293,
      "tokens_seen": 202637312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047403209628886663,
      "loss": 3.0482,
      "theoretical_loss": 4.3480120897465016,
      "tokens_seen": 202702848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047402206619859575,
      "loss": 3.044,
      "theoretical_loss": 4.347836583650782,
      "tokens_seen": 202768384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474012036108325,
      "loss": 3.1396,
      "theoretical_loss": 4.347661150147611,
      "tokens_seen": 202833920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047400200601805417,
      "loss": 3.2979,
      "theoretical_loss": 4.347485789183523,
      "tokens_seen": 202899456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047399197592778335,
      "loss": 3.1927,
      "theoretical_loss": 4.34731050070511,
      "tokens_seen": 202964992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047398194583751253,
      "loss": 3.0788,
      "theoretical_loss": 4.347135284659017,
      "tokens_seen": 203030528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047397191574724177,
      "loss": 3.0052,
      "theoretical_loss": 4.346960140991949,
      "tokens_seen": 203096064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 122225,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3157236576080322,
      "objective/train/theoretical_loss": 4.346785069650667,
      "objective/train/tokens_used": 223621600,
      "theoretical_loss": 4.346785069650667,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739618856569709,
      "loss": 2.9543,
      "theoretical_loss": 4.346785069650667,
      "tokens_seen": 203161600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047395185556670013,
      "loss": 3.0566,
      "theoretical_loss": 4.346610070581987,
      "tokens_seen": 203227136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047394182547642926,
      "loss": 2.9878,
      "theoretical_loss": 4.34643514373278,
      "tokens_seen": 203292672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739317953861585,
      "loss": 3.0709,
      "theoretical_loss": 4.346260289049978,
      "tokens_seen": 203358208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739217652958877,
      "loss": 3.1681,
      "theoretical_loss": 4.346085506480564,
      "tokens_seen": 203423744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047391173520561686,
      "loss": 3.2437,
      "theoretical_loss": 4.34591079597158,
      "tokens_seen": 203489280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047390170511534604,
      "loss": 3.3256,
      "theoretical_loss": 4.345736157470122,
      "tokens_seen": 203554816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738916750250752,
      "loss": 3.1524,
      "theoretical_loss": 4.345561590923343,
      "tokens_seen": 203620352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738816449348044,
      "loss": 3.3006,
      "theoretical_loss": 4.345387096278453,
      "tokens_seen": 203685888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047387161484453363,
      "loss": 3.088,
      "theoretical_loss": 4.345212673482715,
      "tokens_seen": 203751424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047386158475426276,
      "loss": 3.139,
      "theoretical_loss": 4.345038322483449,
      "tokens_seen": 203816960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473851554663992,
      "loss": 3.1496,
      "theoretical_loss": 4.34486404322803,
      "tokens_seen": 203882496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738415245737211,
      "loss": 3.1446,
      "theoretical_loss": 4.34468983566389,
      "tokens_seen": 203948032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047383149448345036,
      "loss": 3.4082,
      "theoretical_loss": 4.3445156997385155,
      "tokens_seen": 204013568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047382146439317954,
      "loss": 3.1106,
      "theoretical_loss": 4.344341635399447,
      "tokens_seen": 204079104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738114343029087,
      "loss": 2.9751,
      "theoretical_loss": 4.344167642594282,
      "tokens_seen": 204144640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738014042126379,
      "loss": 3.0756,
      "theoretical_loss": 4.343993721270673,
      "tokens_seen": 204210176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047379137412236714,
      "loss": 3.1045,
      "theoretical_loss": 4.343819871376326,
      "tokens_seen": 204275712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047378134403209626,
      "loss": 3.1058,
      "theoretical_loss": 4.343646092859004,
      "tokens_seen": 204341248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737713139418255,
      "loss": 2.8941,
      "theoretical_loss": 4.343472385666525,
      "tokens_seen": 204406784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737612838515546,
      "loss": 3.2849,
      "theoretical_loss": 4.343298749746759,
      "tokens_seen": 204472320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047375125376128386,
      "loss": 3.0801,
      "theoretical_loss": 4.343125185047634,
      "tokens_seen": 204537856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047374122367101304,
      "loss": 3.5135,
      "theoretical_loss": 4.342951691517131,
      "tokens_seen": 204603392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737311935807422,
      "loss": 3.2053,
      "theoretical_loss": 4.342778269103286,
      "tokens_seen": 204668928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047372116349047146,
      "loss": 3.0911,
      "theoretical_loss": 4.342604917754191,
      "tokens_seen": 204734464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 123173,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.030834436416626,
      "objective/train/theoretical_loss": 4.342431637417991,
      "objective/train/tokens_used": 225260000,
      "theoretical_loss": 4.342431637417991,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737111334002006,
      "loss": 3.1185,
      "theoretical_loss": 4.342431637417991,
      "tokens_seen": 204800000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737011033099298,
      "loss": 3.0803,
      "theoretical_loss": 4.342258428042883,
      "tokens_seen": 204865536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473691073219659,
      "loss": 3.2108,
      "theoretical_loss": 4.342085289577124,
      "tokens_seen": 204931072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736810431293882,
      "loss": 3.2812,
      "theoretical_loss": 4.341912221969021,
      "tokens_seen": 204996608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047367101303911736,
      "loss": 3.1871,
      "theoretical_loss": 4.341739225166936,
      "tokens_seen": 205062144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736609829488466,
      "loss": 3.1366,
      "theoretical_loss": 4.341566299119286,
      "tokens_seen": 205127680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736509528585757,
      "loss": 3.1494,
      "theoretical_loss": 4.341393443774542,
      "tokens_seen": 205193216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047364092276830496,
      "loss": 3.0442,
      "theoretical_loss": 4.341220659081228,
      "tokens_seen": 205258752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736308926780341,
      "loss": 3.1534,
      "theoretical_loss": 4.341047944987921,
      "tokens_seen": 205324288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736208625877633,
      "loss": 3.1698,
      "theoretical_loss": 4.340875301443257,
      "tokens_seen": 205389824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736108324974925,
      "loss": 3.1333,
      "theoretical_loss": 4.340702728395918,
      "tokens_seen": 205455360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736008024072217,
      "loss": 3.1996,
      "theoretical_loss": 4.340530225794646,
      "tokens_seen": 205520896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047359077231695087,
      "loss": 3.3751,
      "theoretical_loss": 4.340357793588233,
      "tokens_seen": 205586432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047358074222668005,
      "loss": 3.3156,
      "theoretical_loss": 4.340185431725527,
      "tokens_seen": 205651968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047357071213640923,
      "loss": 3.3006,
      "theoretical_loss": 4.3400131401554285,
      "tokens_seen": 205717504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047356068204613847,
      "loss": 3.1734,
      "theoretical_loss": 4.33984091882689,
      "tokens_seen": 205783040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735506519558676,
      "loss": 2.9499,
      "theoretical_loss": 4.339668767688918,
      "tokens_seen": 205848576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047354062186559683,
      "loss": 2.998,
      "theoretical_loss": 4.339496686690575,
      "tokens_seen": 205914112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047353059177532595,
      "loss": 3.2018,
      "theoretical_loss": 4.339324675780972,
      "tokens_seen": 205979648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735205616850552,
      "loss": 3.0006,
      "theoretical_loss": 4.339152734909277,
      "tokens_seen": 206045184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047351053159478437,
      "loss": 3.1032,
      "theoretical_loss": 4.338980864024709,
      "tokens_seen": 206110720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047350050150451355,
      "loss": 3.1907,
      "theoretical_loss": 4.33880906307654,
      "tokens_seen": 206176256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047349047141424273,
      "loss": 3.1745,
      "theoretical_loss": 4.338637332014096,
      "tokens_seen": 206241792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047348044132397197,
      "loss": 3.1799,
      "theoretical_loss": 4.338465670786754,
      "tokens_seen": 206307328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734704112337011,
      "loss": 3.1274,
      "theoretical_loss": 4.338294079343946,
      "tokens_seen": 206372864
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 123915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9498279094696045,
      "objective/train/theoretical_loss": 4.338122557635156,
      "objective/train/tokens_used": 226898400,
      "theoretical_loss": 4.338122557635156,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047346038114343033,
      "loss": 3.1472,
      "theoretical_loss": 4.338122557635156,
      "tokens_seen": 206438400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047345035105315946,
      "loss": 3.5027,
      "theoretical_loss": 4.3379511056099185,
      "tokens_seen": 206503936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734403209628887,
      "loss": 3.3358,
      "theoretical_loss": 4.337779723217823,
      "tokens_seen": 206569472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734302908726179,
      "loss": 3.3308,
      "theoretical_loss": 4.33760841040851,
      "tokens_seen": 206635008
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047342026078234706,
      "loss": 3.2363,
      "theoretical_loss": 4.337437167131674,
      "tokens_seen": 206700544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047341023069207624,
      "loss": 3.339,
      "theoretical_loss": 4.33726599333706,
      "tokens_seen": 206766080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734002006018054,
      "loss": 3.0416,
      "theoretical_loss": 4.337094888974467,
      "tokens_seen": 206831616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733901705115346,
      "loss": 3.0678,
      "theoretical_loss": 4.336923853993743,
      "tokens_seen": 206897152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047338014042126383,
      "loss": 3.1236,
      "theoretical_loss": 4.336752888344792,
      "tokens_seen": 206962688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047337011033099296,
      "loss": 3.1134,
      "theoretical_loss": 4.336581991977568,
      "tokens_seen": 207028224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733600802407222,
      "loss": 2.9821,
      "theoretical_loss": 4.336411164842078,
      "tokens_seen": 207093760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733500501504513,
      "loss": 3.0005,
      "theoretical_loss": 4.336240406888377,
      "tokens_seen": 207159296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047334002006018056,
      "loss": 3.0117,
      "theoretical_loss": 4.336069718066579,
      "tokens_seen": 207224832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047332998996990974,
      "loss": 2.9441,
      "theoretical_loss": 4.335899098326843,
      "tokens_seen": 207290368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733199598796389,
      "loss": 3.0821,
      "theoretical_loss": 4.335728547619382,
      "tokens_seen": 207355904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733099297893681,
      "loss": 3.1602,
      "theoretical_loss": 4.335558065894464,
      "tokens_seen": 207421440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047329989969909734,
      "loss": 3.2261,
      "theoretical_loss": 4.335387653102402,
      "tokens_seen": 207486976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047328986960882646,
      "loss": 3.2569,
      "theoretical_loss": 4.335217309193567,
      "tokens_seen": 207552512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732798395185557,
      "loss": 3.0878,
      "theoretical_loss": 4.335047034118375,
      "tokens_seen": 207618048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732698094282848,
      "loss": 3.0245,
      "theoretical_loss": 4.334876827827299,
      "tokens_seen": 207683584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325977933801406,
      "loss": 3.0552,
      "theoretical_loss": 4.33470669027086,
      "tokens_seen": 207749120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047324974924774324,
      "loss": 3.0553,
      "theoretical_loss": 4.334536621399632,
      "tokens_seen": 207814656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732397191574724,
      "loss": 3.1419,
      "theoretical_loss": 4.3343666211642375,
      "tokens_seen": 207880192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732296890672016,
      "loss": 2.9697,
      "theoretical_loss": 4.334196689515354,
      "tokens_seen": 207945728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732196589769308,
      "loss": 3.4403,
      "theoretical_loss": 4.334026826403705,
      "tokens_seen": 208011264
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 125279,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2541067600250244,
      "objective/train/theoretical_loss": 4.333857031780069,
      "objective/train/tokens_used": 228536800,
      "theoretical_loss": 4.333857031780069,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047320962888665997,
      "loss": 3.1755,
      "theoretical_loss": 4.333857031780069,
      "tokens_seen": 208076800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731995987963892,
      "loss": 3.3347,
      "theoretical_loss": 4.333687305595276,
      "tokens_seen": 208142336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047318956870611833,
      "loss": 3.1738,
      "theoretical_loss": 4.3335176478002015,
      "tokens_seen": 208207872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047317953861584757,
      "loss": 3.0407,
      "theoretical_loss": 4.333348058345776,
      "tokens_seen": 208273408
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047316950852557675,
      "loss": 3.0055,
      "theoretical_loss": 4.33317853718298,
      "tokens_seen": 208338944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047315947843530593,
      "loss": 3.1805,
      "theoretical_loss": 4.333009084262844,
      "tokens_seen": 208404480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731494483450351,
      "loss": 2.9499,
      "theoretical_loss": 4.33283969953645,
      "tokens_seen": 208470016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731394182547643,
      "loss": 3.0383,
      "theoretical_loss": 4.332670382954928,
      "tokens_seen": 208535552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047312938816449347,
      "loss": 3.1877,
      "theoretical_loss": 4.33250113446946,
      "tokens_seen": 208601088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731193580742227,
      "loss": 3.1835,
      "theoretical_loss": 4.332331954031279,
      "tokens_seen": 208666624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047310932798395183,
      "loss": 3.2001,
      "theoretical_loss": 4.332162841591668,
      "tokens_seen": 208732160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047309929789368107,
      "loss": 3.2484,
      "theoretical_loss": 4.331993797101958,
      "tokens_seen": 208797696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730892678034102,
      "loss": 3.37,
      "theoretical_loss": 4.331824820513534,
      "tokens_seen": 208863232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047307923771313943,
      "loss": 3.1849,
      "theoretical_loss": 4.331655911777826,
      "tokens_seen": 208928768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730692076228686,
      "loss": 3.219,
      "theoretical_loss": 4.331487070846319,
      "tokens_seen": 208994304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730591775325978,
      "loss": 3.3387,
      "theoretical_loss": 4.331318297670545,
      "tokens_seen": 209059840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473049147442327,
      "loss": 3.1956,
      "theoretical_loss": 4.3311495922020855,
      "tokens_seen": 209125376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303911735205615,
      "loss": 3.1449,
      "theoretical_loss": 4.330980954392573,
      "tokens_seen": 209190912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047302908726178534,
      "loss": 2.8264,
      "theoretical_loss": 4.33081238419369,
      "tokens_seen": 209256448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047301905717151457,
      "loss": 3.0265,
      "theoretical_loss": 4.330643881557168,
      "tokens_seen": 209321984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730090270812437,
      "loss": 3.2387,
      "theoretical_loss": 4.330475446434788,
      "tokens_seen": 209387520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047299899699097293,
      "loss": 3.2265,
      "theoretical_loss": 4.330307078778379,
      "tokens_seen": 209453056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729889669007021,
      "loss": 3.2092,
      "theoretical_loss": 4.330138778539823,
      "tokens_seen": 209518592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729789368104313,
      "loss": 3.2283,
      "theoretical_loss": 4.329970545671048,
      "tokens_seen": 209584128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047296890672016053,
      "loss": 3.2865,
      "theoretical_loss": 4.3298023801240335,
      "tokens_seen": 209649664
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 125881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1459879875183105,
      "objective/train/theoretical_loss": 4.329634281850807,
      "objective/train/tokens_used": 230175200,
      "theoretical_loss": 4.329634281850807,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047295887662988966,
      "loss": 3.168,
      "theoretical_loss": 4.329634281850807,
      "tokens_seen": 209715200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729488465396189,
      "loss": 3.1116,
      "theoretical_loss": 4.329466250803446,
      "tokens_seen": 209780736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729388164493481,
      "loss": 3.1933,
      "theoretical_loss": 4.329298286934075,
      "tokens_seen": 209846272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047292878635907726,
      "loss": 2.9711,
      "theoretical_loss": 4.329130390194871,
      "tokens_seen": 209911808
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047291875626880644,
      "loss": 3.0108,
      "theoretical_loss": 4.3289625605380575,
      "tokens_seen": 209977344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729087261785356,
      "loss": 3.1506,
      "theoretical_loss": 4.328794797915908,
      "tokens_seen": 210042880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728986960882648,
      "loss": 3.237,
      "theoretical_loss": 4.328627102280742,
      "tokens_seen": 210108416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047288866599799403,
      "loss": 3.1071,
      "theoretical_loss": 4.328459473584933,
      "tokens_seen": 210173952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047287863590772316,
      "loss": 3.2433,
      "theoretical_loss": 4.328291911780899,
      "tokens_seen": 210239488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728686058174524,
      "loss": 3.207,
      "theoretical_loss": 4.328124416821107,
      "tokens_seen": 210305024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728585757271815,
      "loss": 3.0205,
      "theoretical_loss": 4.327956988658076,
      "tokens_seen": 210370560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047284854563691076,
      "loss": 3.2139,
      "theoretical_loss": 4.3277896272443686,
      "tokens_seen": 210436096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047283851554663994,
      "loss": 3.0787,
      "theoretical_loss": 4.3276223325325995,
      "tokens_seen": 210501632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728284854563691,
      "loss": 3.1398,
      "theoretical_loss": 4.327455104475431,
      "tokens_seen": 210567168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728184553660983,
      "loss": 3.3631,
      "theoretical_loss": 4.327287943025572,
      "tokens_seen": 210632704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047280842527582754,
      "loss": 2.9762,
      "theoretical_loss": 4.327120848135781,
      "tokens_seen": 210698240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047279839518555666,
      "loss": 3.067,
      "theoretical_loss": 4.326953819758867,
      "tokens_seen": 210763776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727883650952859,
      "loss": 3.2263,
      "theoretical_loss": 4.326786857847681,
      "tokens_seen": 210829312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472778335005015,
      "loss": 3.3092,
      "theoretical_loss": 4.326619962355128,
      "tokens_seen": 210894848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047276830491474426,
      "loss": 3.4328,
      "theoretical_loss": 4.326453133234159,
      "tokens_seen": 210960384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047275827482447344,
      "loss": 3.1644,
      "theoretical_loss": 4.326286370437772,
      "tokens_seen": 211025920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727482447342026,
      "loss": 3.1984,
      "theoretical_loss": 4.326119673919013,
      "tokens_seen": 211091456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727382146439318,
      "loss": 3.0863,
      "theoretical_loss": 4.325953043630978,
      "tokens_seen": 211156992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472728184553661,
      "loss": 3.0964,
      "theoretical_loss": 4.3257864795268075,
      "tokens_seen": 211222528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047271815446339017,
      "loss": 3.0099,
      "theoretical_loss": 4.325619981559692,
      "tokens_seen": 211288064
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 127261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0648367404937744,
      "objective/train/theoretical_loss": 4.325453549682869,
      "objective/train/tokens_used": 231813600,
      "theoretical_loss": 4.325453549682869,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727081243731194,
      "loss": 3.116,
      "theoretical_loss": 4.325453549682869,
      "tokens_seen": 211353600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047269809428284853,
      "loss": 3.0499,
      "theoretical_loss": 4.325287183849625,
      "tokens_seen": 211419136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047268806419257777,
      "loss": 3.0799,
      "theoretical_loss": 4.325120884013289,
      "tokens_seen": 211484672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047267803410230695,
      "loss": 3.1633,
      "theoretical_loss": 4.324954650127244,
      "tokens_seen": 211550208
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047266800401203613,
      "loss": 2.8267,
      "theoretical_loss": 4.324788482144916,
      "tokens_seen": 211615744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726579739217653,
      "loss": 2.9392,
      "theoretical_loss": 4.324622380019778,
      "tokens_seen": 211681280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726479438314945,
      "loss": 3.1858,
      "theoretical_loss": 4.3244563437053545,
      "tokens_seen": 211746816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047263791374122367,
      "loss": 3.1446,
      "theoretical_loss": 4.324290373155213,
      "tokens_seen": 211812352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726278836509529,
      "loss": 3.259,
      "theoretical_loss": 4.324124468322971,
      "tokens_seen": 211877888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047261785356068203,
      "loss": 3.195,
      "theoretical_loss": 4.32395862916229,
      "tokens_seen": 211943424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047260782347041127,
      "loss": 3.0971,
      "theoretical_loss": 4.32379285562688,
      "tokens_seen": 212008960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725977933801404,
      "loss": 3.1254,
      "theoretical_loss": 4.323627147670498,
      "tokens_seen": 212074496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047258776328986963,
      "loss": 3.0839,
      "theoretical_loss": 4.323461505246949,
      "tokens_seen": 212140032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725777331995988,
      "loss": 2.8079,
      "theoretical_loss": 4.323295928310084,
      "tokens_seen": 212205568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472567703109328,
      "loss": 3.1229,
      "theoretical_loss": 4.3231304168137985,
      "tokens_seen": 212271104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725576730190572,
      "loss": 3.1019,
      "theoretical_loss": 4.322964970712039,
      "tokens_seen": 212336640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047254764292878636,
      "loss": 3.272,
      "theoretical_loss": 4.322799589958794,
      "tokens_seen": 212402176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047253761283851554,
      "loss": 2.7217,
      "theoretical_loss": 4.3226342745081014,
      "tokens_seen": 212467712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047252758274824477,
      "loss": 3.0247,
      "theoretical_loss": 4.322469024314046,
      "tokens_seen": 212533248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725175526579739,
      "loss": 3.2237,
      "theoretical_loss": 4.322303839330757,
      "tokens_seen": 212598784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047250752256770313,
      "loss": 3.3425,
      "theoretical_loss": 4.322138719512411,
      "tokens_seen": 212664320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724974924774323,
      "loss": 3.2206,
      "theoretical_loss": 4.321973664813231,
      "tokens_seen": 212729856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724874623871615,
      "loss": 3.2825,
      "theoretical_loss": 4.321808675187486,
      "tokens_seen": 212795392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724774322968907,
      "loss": 2.8949,
      "theoretical_loss": 4.321643750589493,
      "tokens_seen": 212860928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047246740220661986,
      "loss": 3.0597,
      "theoretical_loss": 4.32147889097361,
      "tokens_seen": 212926464
    },
    {
      "epoch": 0.06,
      "objective/train/docs_used": 127839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5590968132019043,
      "objective/train/theoretical_loss": 4.321314096294248,
      "objective/train/tokens_used": 233452000,
      "theoretical_loss": 4.321314096294248,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047245737211634904,
      "loss": 2.9514,
      "theoretical_loss": 4.321314096294248,
      "tokens_seen": 212992000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724473420260783,
      "loss": 3.0759,
      "theoretical_loss": 4.321149366505858,
      "tokens_seen": 213057536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724373119358074,
      "loss": 2.9617,
      "theoretical_loss": 4.32098470156294,
      "tokens_seen": 213123072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047242728184553664,
      "loss": 2.9706,
      "theoretical_loss": 4.320820101420041,
      "tokens_seen": 213188608
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047241725175526576,
      "loss": 3.3507,
      "theoretical_loss": 4.320655566031751,
      "tokens_seen": 213254144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472407221664995,
      "loss": 3.0787,
      "theoretical_loss": 4.320491095352706,
      "tokens_seen": 213319680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723971915747242,
      "loss": 3.2185,
      "theoretical_loss": 4.320326689337589,
      "tokens_seen": 213385216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047238716148445336,
      "loss": 3.2852,
      "theoretical_loss": 4.320162347941129,
      "tokens_seen": 213450752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047237713139418254,
      "loss": 3.0982,
      "theoretical_loss": 4.3199980711181,
      "tokens_seen": 213516288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723671013039117,
      "loss": 3.1012,
      "theoretical_loss": 4.319833858823321,
      "tokens_seen": 213581824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723570712136409,
      "loss": 2.945,
      "theoretical_loss": 4.319669711011656,
      "tokens_seen": 213647360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047234704112337014,
      "loss": 3.1732,
      "theoretical_loss": 4.319505627638016,
      "tokens_seen": 213712896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047233701103309927,
      "loss": 3.1215,
      "theoretical_loss": 4.319341608657357,
      "tokens_seen": 213778432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723269809428285,
      "loss": 3.3584,
      "theoretical_loss": 4.319177654024679,
      "tokens_seen": 213843968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723169508525577,
      "loss": 3.0733,
      "theoretical_loss": 4.319013763695029,
      "tokens_seen": 213909504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047230692076228686,
      "loss": 3.2797,
      "theoretical_loss": 4.318849937623497,
      "tokens_seen": 213975040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047229689067201605,
      "loss": 2.9947,
      "theoretical_loss": 4.318686175765221,
      "tokens_seen": 214040576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722868605817452,
      "loss": 3.0984,
      "theoretical_loss": 4.31852247807538,
      "tokens_seen": 214106112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722768304914744,
      "loss": 2.8851,
      "theoretical_loss": 4.318358844509202,
      "tokens_seen": 214171648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047226680040120364,
      "loss": 3.0447,
      "theoretical_loss": 4.318195275021959,
      "tokens_seen": 214237184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047225677031093277,
      "loss": 3.2885,
      "theoretical_loss": 4.318031769568966,
      "tokens_seen": 214302720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472246740220662,
      "loss": 3.0849,
      "theoretical_loss": 4.317868328105585,
      "tokens_seen": 214368256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047223671013039113,
      "loss": 3.052,
      "theoretical_loss": 4.3177049505872205,
      "tokens_seen": 214433792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047222668004012037,
      "loss": 3.0657,
      "theoretical_loss": 4.317541636969324,
      "tokens_seen": 214499328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722166499498496,
      "loss": 2.9079,
      "theoretical_loss": 4.31737838720739,
      "tokens_seen": 214564864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 128852,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.337009906768799,
      "objective/train/theoretical_loss": 4.317215201256958,
      "objective/train/tokens_used": 235090400,
      "theoretical_loss": 4.317215201256958,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047220661985957873,
      "loss": 3.1137,
      "theoretical_loss": 4.317215201256958,
      "tokens_seen": 214630400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047219658976930797,
      "loss": 3.0925,
      "theoretical_loss": 4.317052079073613,
      "tokens_seen": 214695936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047218655967903715,
      "loss": 3.1037,
      "theoretical_loss": 4.316889020612983,
      "tokens_seen": 214761472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047217652958876633,
      "loss": 2.9058,
      "theoretical_loss": 4.316726025830741,
      "tokens_seen": 214827008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721664994984955,
      "loss": 3.1171,
      "theoretical_loss": 4.316563094682604,
      "tokens_seen": 214892544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721564694082247,
      "loss": 3.2248,
      "theoretical_loss": 4.316400227124335,
      "tokens_seen": 214958080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047214643931795387,
      "loss": 3.1588,
      "theoretical_loss": 4.3162374231117395,
      "tokens_seen": 215023616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721364092276831,
      "loss": 3.2238,
      "theoretical_loss": 4.316074682600666,
      "tokens_seen": 215089152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047212637913741223,
      "loss": 3.3115,
      "theoretical_loss": 4.315912005547011,
      "tokens_seen": 215154688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047211634904714147,
      "loss": 2.9668,
      "theoretical_loss": 4.315749391906712,
      "tokens_seen": 215220224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721063189568706,
      "loss": 3.0336,
      "theoretical_loss": 4.31558684163575,
      "tokens_seen": 215285760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047209628886659983,
      "loss": 3.1727,
      "theoretical_loss": 4.315424354690153,
      "tokens_seen": 215351296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472086258776329,
      "loss": 3.0615,
      "theoretical_loss": 4.31526193102599,
      "tokens_seen": 215416832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720762286860582,
      "loss": 2.9988,
      "theoretical_loss": 4.315099570599374,
      "tokens_seen": 215482368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720661985957874,
      "loss": 3.0711,
      "theoretical_loss": 4.314937273366466,
      "tokens_seen": 215547904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047205616850551656,
      "loss": 3.2474,
      "theoretical_loss": 4.314775039283463,
      "tokens_seen": 215613440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047204613841524574,
      "loss": 3.2283,
      "theoretical_loss": 4.314612868306614,
      "tokens_seen": 215678976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047203610832497497,
      "loss": 3.1768,
      "theoretical_loss": 4.314450760392204,
      "tokens_seen": 215744512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720260782347041,
      "loss": 3.23,
      "theoretical_loss": 4.314288715496568,
      "tokens_seen": 215810048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047201604814443333,
      "loss": 3.443,
      "theoretical_loss": 4.314126733576082,
      "tokens_seen": 215875584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720060180541625,
      "loss": 2.9834,
      "theoretical_loss": 4.3139648145871625,
      "tokens_seen": 215941120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719959879638917,
      "loss": 3.3007,
      "theoretical_loss": 4.313802958486274,
      "tokens_seen": 216006656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719859578736209,
      "loss": 3.2257,
      "theoretical_loss": 4.313641165229922,
      "tokens_seen": 216072192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047197592778335006,
      "loss": 2.9371,
      "theoretical_loss": 4.313479434774655,
      "tokens_seen": 216137728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047196589769307924,
      "loss": 3.1597,
      "theoretical_loss": 4.313317767077065,
      "tokens_seen": 216203264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 129467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.151235342025757,
      "objective/train/theoretical_loss": 4.3131561620937875,
      "objective/train/tokens_used": 236728800,
      "theoretical_loss": 4.3131561620937875,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719558676028085,
      "loss": 3.1703,
      "theoretical_loss": 4.3131561620937875,
      "tokens_seen": 216268800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719458375125376,
      "loss": 3.1118,
      "theoretical_loss": 4.312994619781502,
      "tokens_seen": 216334336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193580742226684,
      "loss": 3.2158,
      "theoretical_loss": 4.312833140096929,
      "tokens_seen": 216399872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047192577733199596,
      "loss": 3.1632,
      "theoretical_loss": 4.312671722996835,
      "tokens_seen": 216465408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719157472417252,
      "loss": 3.0144,
      "theoretical_loss": 4.312510368438025,
      "tokens_seen": 216530944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719057171514544,
      "loss": 2.9652,
      "theoretical_loss": 4.31234907637735,
      "tokens_seen": 216596480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047189568706118356,
      "loss": 3.0804,
      "theoretical_loss": 4.312187846771704,
      "tokens_seen": 216662016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047188565697091274,
      "loss": 3.221,
      "theoretical_loss": 4.312026679578022,
      "tokens_seen": 216727552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718756268806419,
      "loss": 3.2019,
      "theoretical_loss": 4.311865574753283,
      "tokens_seen": 216793088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718655967903711,
      "loss": 2.9076,
      "theoretical_loss": 4.311704532254509,
      "tokens_seen": 216858624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047185556670010034,
      "loss": 3.1173,
      "theoretical_loss": 4.311543552038762,
      "tokens_seen": 216924160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047184553660982947,
      "loss": 2.9789,
      "theoretical_loss": 4.31138263406315,
      "tokens_seen": 216989696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718355065195587,
      "loss": 3.0793,
      "theoretical_loss": 4.311221778284821,
      "tokens_seen": 217055232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718254764292879,
      "loss": 3.1944,
      "theoretical_loss": 4.311060984660967,
      "tokens_seen": 217120768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047181544633901706,
      "loss": 2.9635,
      "theoretical_loss": 4.310900253148821,
      "tokens_seen": 217186304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047180541624874625,
      "loss": 3.0845,
      "theoretical_loss": 4.31073958370566,
      "tokens_seen": 217251840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717953861584754,
      "loss": 3.3273,
      "theoretical_loss": 4.3105789762888005,
      "tokens_seen": 217317376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717853560682046,
      "loss": 3.1426,
      "theoretical_loss": 4.310418430855604,
      "tokens_seen": 217382912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177532597793384,
      "loss": 3.1465,
      "theoretical_loss": 4.310257947363474,
      "tokens_seen": 217448448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047176529588766297,
      "loss": 3.1651,
      "theoretical_loss": 4.310097525769854,
      "tokens_seen": 217513984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717552657973922,
      "loss": 3.406,
      "theoretical_loss": 4.309937166032231,
      "tokens_seen": 217579520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047174523570712133,
      "loss": 3.3688,
      "theoretical_loss": 4.309776868108134,
      "tokens_seen": 217645056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173520561685057,
      "loss": 3.2583,
      "theoretical_loss": 4.309616631955134,
      "tokens_seen": 217710592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047172517552657975,
      "loss": 3.2181,
      "theoretical_loss": 4.309456457530843,
      "tokens_seen": 217776128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047171514543630893,
      "loss": 3.0626,
      "theoretical_loss": 4.309296344792916,
      "tokens_seen": 217841664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 130589,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.174252510070801,
      "objective/train/theoretical_loss": 4.309136293699048,
      "objective/train/tokens_used": 238367200,
      "theoretical_loss": 4.309136293699048,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717051153460381,
      "loss": 3.0759,
      "theoretical_loss": 4.309136293699048,
      "tokens_seen": 217907200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169508525576735,
      "loss": 3.0866,
      "theoretical_loss": 4.308976304206978,
      "tokens_seen": 217972736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716850551654965,
      "loss": 3.0519,
      "theoretical_loss": 4.3088163762744855,
      "tokens_seen": 218038272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716750250752257,
      "loss": 2.938,
      "theoretical_loss": 4.308656509859392,
      "tokens_seen": 218103808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047166499498495484,
      "loss": 2.9371,
      "theoretical_loss": 4.308496704919559,
      "tokens_seen": 218169344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047165496489468407,
      "loss": 2.8471,
      "theoretical_loss": 4.308336961412892,
      "tokens_seen": 218234880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047164493480441325,
      "loss": 3.1918,
      "theoretical_loss": 4.308177279297336,
      "tokens_seen": 218300416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163490471414243,
      "loss": 3.0599,
      "theoretical_loss": 4.308017658530879,
      "tokens_seen": 218365952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716248746238716,
      "loss": 3.1856,
      "theoretical_loss": 4.307858099071549,
      "tokens_seen": 218431488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716148445336008,
      "loss": 2.7223,
      "theoretical_loss": 4.307698600877416,
      "tokens_seen": 218497024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047160481444333,
      "loss": 3.1367,
      "theoretical_loss": 4.30753916390659,
      "tokens_seen": 218562560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715947843530592,
      "loss": 3.048,
      "theoretical_loss": 4.307379788117224,
      "tokens_seen": 218628096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047158475426278834,
      "loss": 3.1323,
      "theoretical_loss": 4.3072204734675115,
      "tokens_seen": 218693632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715747241725176,
      "loss": 2.804,
      "theoretical_loss": 4.307061219915687,
      "tokens_seen": 218759168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715646940822467,
      "loss": 3.1227,
      "theoretical_loss": 4.306902027420025,
      "tokens_seen": 218824704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047155466399197594,
      "loss": 3.3008,
      "theoretical_loss": 4.306742895938844,
      "tokens_seen": 218890240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715446339017051,
      "loss": 3.1554,
      "theoretical_loss": 4.306583825430499,
      "tokens_seen": 218955776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715346038114343,
      "loss": 2.8274,
      "theoretical_loss": 4.30642481585339,
      "tokens_seen": 219021312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715245737211635,
      "loss": 3.1366,
      "theoretical_loss": 4.306265867165955,
      "tokens_seen": 219086848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715145436308927,
      "loss": 2.9215,
      "theoretical_loss": 4.306106979326674,
      "tokens_seen": 219152384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047150451354062184,
      "loss": 3.2813,
      "theoretical_loss": 4.305948152294068,
      "tokens_seen": 219217920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714944834503511,
      "loss": 2.9276,
      "theoretical_loss": 4.305789386026698,
      "tokens_seen": 219283456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714844533600802,
      "loss": 3.0313,
      "theoretical_loss": 4.305630680483166,
      "tokens_seen": 219348992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047147442326980944,
      "loss": 3.2763,
      "theoretical_loss": 4.305472035622114,
      "tokens_seen": 219414528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714643931795387,
      "loss": 3.1934,
      "theoretical_loss": 4.305313451402226,
      "tokens_seen": 219480064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 131942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.265974283218384,
      "objective/train/theoretical_loss": 4.305154927782223,
      "objective/train/tokens_used": 240005600,
      "theoretical_loss": 4.305154927782223,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714543630892678,
      "loss": 3.0968,
      "theoretical_loss": 4.305154927782223,
      "tokens_seen": 219545600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047144433299899704,
      "loss": 3.047,
      "theoretical_loss": 4.304996464720872,
      "tokens_seen": 219611136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047143430290872616,
      "loss": 3.1081,
      "theoretical_loss": 4.304838062176974,
      "tokens_seen": 219676672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714242728184554,
      "loss": 3.1186,
      "theoretical_loss": 4.304679720109376,
      "tokens_seen": 219742208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714142427281846,
      "loss": 3.161,
      "theoretical_loss": 4.304521438476961,
      "tokens_seen": 219807744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047140421263791376,
      "loss": 3.122,
      "theoretical_loss": 4.304363217238654,
      "tokens_seen": 219873280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139418254764294,
      "loss": 3.1317,
      "theoretical_loss": 4.304205056353421,
      "tokens_seen": 219938816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713841524573721,
      "loss": 3.0803,
      "theoretical_loss": 4.304046955780267,
      "tokens_seen": 220004352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713741223671013,
      "loss": 3.1409,
      "theoretical_loss": 4.303888915478236,
      "tokens_seen": 220069888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047136409227683054,
      "loss": 3.1707,
      "theoretical_loss": 4.303730935406414,
      "tokens_seen": 220135424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135406218655967,
      "loss": 2.9751,
      "theoretical_loss": 4.303573015523927,
      "tokens_seen": 220200960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713440320962889,
      "loss": 3.177,
      "theoretical_loss": 4.3034151557899385,
      "tokens_seen": 220266496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713340020060181,
      "loss": 3.054,
      "theoretical_loss": 4.303257356163654,
      "tokens_seen": 220332032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047132397191574726,
      "loss": 3.4153,
      "theoretical_loss": 4.303099616604318,
      "tokens_seen": 220397568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047131394182547645,
      "loss": 3.0444,
      "theoretical_loss": 4.302941937071216,
      "tokens_seen": 220463104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047130391173520563,
      "loss": 2.9421,
      "theoretical_loss": 4.3027843175236695,
      "tokens_seen": 220528640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712938816449348,
      "loss": 3.1708,
      "theoretical_loss": 4.302626757921045,
      "tokens_seen": 220594176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047128385155466404,
      "loss": 3.3514,
      "theoretical_loss": 4.302469258222744,
      "tokens_seen": 220659712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047127382146439317,
      "loss": 3.2391,
      "theoretical_loss": 4.30231181838821,
      "tokens_seen": 220725248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712637913741224,
      "loss": 3.122,
      "theoretical_loss": 4.302154438376926,
      "tokens_seen": 220790784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125376128385153,
      "loss": 3.0711,
      "theoretical_loss": 4.301997118148412,
      "tokens_seen": 220856320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047124373119358077,
      "loss": 3.2042,
      "theoretical_loss": 4.301839857662231,
      "tokens_seen": 220921856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047123370110330995,
      "loss": 2.8998,
      "theoretical_loss": 4.301682656877983,
      "tokens_seen": 220987392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047122367101303913,
      "loss": 3.3909,
      "theoretical_loss": 4.301525515755307,
      "tokens_seen": 221052928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712136409227683,
      "loss": 3.0117,
      "theoretical_loss": 4.301368434253883,
      "tokens_seen": 221118464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 132442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3066797256469727,
      "objective/train/theoretical_loss": 4.301211412333429,
      "objective/train/tokens_used": 241644000,
      "theoretical_loss": 4.301211412333429,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047120361083249755,
      "loss": 3.1236,
      "theoretical_loss": 4.301211412333429,
      "tokens_seen": 221184000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711935807422267,
      "loss": 3.0539,
      "theoretical_loss": 4.301054449953702,
      "tokens_seen": 221249536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711835506519559,
      "loss": 3.2493,
      "theoretical_loss": 4.300897547074499,
      "tokens_seen": 221315072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047117352056168504,
      "loss": 3.494,
      "theoretical_loss": 4.300740703655655,
      "tokens_seen": 221380608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047116349047141427,
      "loss": 2.8181,
      "theoretical_loss": 4.300583919657045,
      "tokens_seen": 221446144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115346038114345,
      "loss": 3.1062,
      "theoretical_loss": 4.300427195038582,
      "tokens_seen": 221511680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047114343029087263,
      "loss": 3.2018,
      "theoretical_loss": 4.300270529760219,
      "tokens_seen": 221577216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711334002006018,
      "loss": 2.9325,
      "theoretical_loss": 4.300113923781946,
      "tokens_seen": 221642752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471123370110331,
      "loss": 3.0023,
      "theoretical_loss": 4.2999573770637936,
      "tokens_seen": 221708288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711133400200602,
      "loss": 3.1196,
      "theoretical_loss": 4.299800889565831,
      "tokens_seen": 221773824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711033099297894,
      "loss": 3.0607,
      "theoretical_loss": 4.299644461248165,
      "tokens_seen": 221839360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047109327983951854,
      "loss": 3.2191,
      "theoretical_loss": 4.299488092070942,
      "tokens_seen": 221904896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710832497492478,
      "loss": 3.0538,
      "theoretical_loss": 4.299331781994347,
      "tokens_seen": 221970432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710732196589769,
      "loss": 3.073,
      "theoretical_loss": 4.2991755309786015,
      "tokens_seen": 222035968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047106318956870614,
      "loss": 2.9409,
      "theoretical_loss": 4.299019338983969,
      "tokens_seen": 222101504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710531594784353,
      "loss": 3.0287,
      "theoretical_loss": 4.298863205970751,
      "tokens_seen": 222167040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710431293881645,
      "loss": 3.1599,
      "theoretical_loss": 4.298707131899281,
      "tokens_seen": 222232576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710330992978937,
      "loss": 3.146,
      "theoretical_loss": 4.298551116729941,
      "tokens_seen": 222298112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710230692076229,
      "loss": 3.085,
      "theoretical_loss": 4.298395160423144,
      "tokens_seen": 222363648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047101303911735204,
      "loss": 3.09,
      "theoretical_loss": 4.298239262939344,
      "tokens_seen": 222429184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710030090270813,
      "loss": 3.0419,
      "theoretical_loss": 4.2980834242390324,
      "tokens_seen": 222494720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709929789368104,
      "loss": 3.2707,
      "theoretical_loss": 4.29792764428274,
      "tokens_seen": 222560256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047098294884653964,
      "loss": 3.2489,
      "theoretical_loss": 4.297771923031033,
      "tokens_seen": 222625792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709729187562688,
      "loss": 2.9268,
      "theoretical_loss": 4.297616260444519,
      "tokens_seen": 222691328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470962888665998,
      "loss": 3.1398,
      "theoretical_loss": 4.297460656483842,
      "tokens_seen": 222756864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 133596,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.741494655609131,
      "objective/train/theoretical_loss": 4.297305111109683,
      "objective/train/tokens_used": 243282400,
      "theoretical_loss": 4.297305111109683,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709528585757272,
      "loss": 2.7095,
      "theoretical_loss": 4.297305111109683,
      "tokens_seen": 222822400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047094282848545636,
      "loss": 2.9503,
      "theoretical_loss": 4.297149624282762,
      "tokens_seen": 222887936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047093279839518555,
      "loss": 3.1134,
      "theoretical_loss": 4.296994195963838,
      "tokens_seen": 222953472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709227683049148,
      "loss": 2.8174,
      "theoretical_loss": 4.296838826113706,
      "tokens_seen": 223019008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709127382146439,
      "loss": 2.7694,
      "theoretical_loss": 4.296683514693199,
      "tokens_seen": 223084544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047090270812437314,
      "loss": 3.1462,
      "theoretical_loss": 4.296528261663189,
      "tokens_seen": 223150080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047089267803410227,
      "loss": 3.1611,
      "theoretical_loss": 4.296373066984585,
      "tokens_seen": 223215616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708826479438315,
      "loss": 3.0784,
      "theoretical_loss": 4.296217930618332,
      "tokens_seen": 223281152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708726178535607,
      "loss": 3.2461,
      "theoretical_loss": 4.2960628525254165,
      "tokens_seen": 223346688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047086258776328987,
      "loss": 3.3592,
      "theoretical_loss": 4.295907832666858,
      "tokens_seen": 223412224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047085255767301905,
      "loss": 3.134,
      "theoretical_loss": 4.2957528710037165,
      "tokens_seen": 223477760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708425275827483,
      "loss": 2.793,
      "theoretical_loss": 4.2955979674970886,
      "tokens_seen": 223543296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708324974924774,
      "loss": 3.1087,
      "theoretical_loss": 4.295443122108108,
      "tokens_seen": 223608832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047082246740220665,
      "loss": 3.2745,
      "theoretical_loss": 4.295288334797946,
      "tokens_seen": 223674368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708124373119358,
      "loss": 3.0884,
      "theoretical_loss": 4.295133605527812,
      "tokens_seen": 223739904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470802407221665,
      "loss": 3.0021,
      "theoretical_loss": 4.294978934258951,
      "tokens_seen": 223805440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707923771313942,
      "loss": 2.9157,
      "theoretical_loss": 4.294824320952646,
      "tokens_seen": 223870976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047078234704112337,
      "loss": 3.3173,
      "theoretical_loss": 4.294669765570219,
      "tokens_seen": 223936512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047077231695085255,
      "loss": 3.1127,
      "theoretical_loss": 4.294515268073026,
      "tokens_seen": 224002048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047076228686058173,
      "loss": 2.8344,
      "theoretical_loss": 4.294360828422461,
      "tokens_seen": 224067584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707522567703109,
      "loss": 3.0905,
      "theoretical_loss": 4.294206446579956,
      "tokens_seen": 224133120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047074222668004015,
      "loss": 3.1892,
      "theoretical_loss": 4.29405212250698,
      "tokens_seen": 224198656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707321965897693,
      "loss": 2.8706,
      "theoretical_loss": 4.293897856165039,
      "tokens_seen": 224264192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707221664994985,
      "loss": 3.2766,
      "theoretical_loss": 4.293743647515674,
      "tokens_seen": 224329728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047071213640922775,
      "loss": 3.1984,
      "theoretical_loss": 4.293589496520464,
      "tokens_seen": 224395264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 134273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.291004180908203,
      "objective/train/theoretical_loss": 4.293435403141026,
      "objective/train/tokens_used": 244920800,
      "theoretical_loss": 4.293435403141026,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707021063189569,
      "loss": 3.2134,
      "theoretical_loss": 4.293435403141026,
      "tokens_seen": 224460800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706920762286861,
      "loss": 3.107,
      "theoretical_loss": 4.293281367339013,
      "tokens_seen": 224526336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047068204613841524,
      "loss": 3.2989,
      "theoretical_loss": 4.293127389076113,
      "tokens_seen": 224591872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047067201604814447,
      "loss": 3.4678,
      "theoretical_loss": 4.292973468314053,
      "tokens_seen": 224657408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047066198595787365,
      "loss": 3.0958,
      "theoretical_loss": 4.292819605014596,
      "tokens_seen": 224722944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047065195586760283,
      "loss": 2.9754,
      "theoretical_loss": 4.292665799139541,
      "tokens_seen": 224788480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470641925777332,
      "loss": 3.2233,
      "theoretical_loss": 4.292512050650723,
      "tokens_seen": 224854016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706318956870612,
      "loss": 2.9816,
      "theoretical_loss": 4.292358359510015,
      "tokens_seen": 224919552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706218655967904,
      "loss": 3.2302,
      "theoretical_loss": 4.292204725679326,
      "tokens_seen": 224985088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706118355065196,
      "loss": 3.1335,
      "theoretical_loss": 4.2920511491206,
      "tokens_seen": 225050624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047060180541624874,
      "loss": 3.1664,
      "theoretical_loss": 4.2918976297958205,
      "tokens_seen": 225116160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470591775325978,
      "loss": 3.0454,
      "theoretical_loss": 4.2917441676670025,
      "tokens_seen": 225181696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705817452357071,
      "loss": 3.0739,
      "theoretical_loss": 4.291590762696202,
      "tokens_seen": 225247232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047057171514543634,
      "loss": 3.1721,
      "theoretical_loss": 4.291437414845509,
      "tokens_seen": 225312768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705616850551655,
      "loss": 3.1347,
      "theoretical_loss": 4.291284124077049,
      "tokens_seen": 225378304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705516549648947,
      "loss": 2.9492,
      "theoretical_loss": 4.291130890352985,
      "tokens_seen": 225443840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705416248746239,
      "loss": 3.1897,
      "theoretical_loss": 4.290977713635515,
      "tokens_seen": 225509376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705315947843531,
      "loss": 3.2739,
      "theoretical_loss": 4.2908245938868745,
      "tokens_seen": 225574912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047052156469408224,
      "loss": 3.3183,
      "theoretical_loss": 4.290671531069334,
      "tokens_seen": 225640448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705115346038115,
      "loss": 3.1922,
      "theoretical_loss": 4.290518525145199,
      "tokens_seen": 225705984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705015045135406,
      "loss": 3.282,
      "theoretical_loss": 4.290365576076812,
      "tokens_seen": 225771520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047049147442326984,
      "loss": 3.2065,
      "theoretical_loss": 4.290212683826551,
      "tokens_seen": 225837056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470481444332999,
      "loss": 3.251,
      "theoretical_loss": 4.290059848356833,
      "tokens_seen": 225902592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704714142427282,
      "loss": 2.9908,
      "theoretical_loss": 4.289907069630104,
      "tokens_seen": 225968128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704613841524574,
      "loss": 3.2375,
      "theoretical_loss": 4.28975434760885,
      "tokens_seen": 226033664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 135613,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.803276538848877,
      "objective/train/theoretical_loss": 4.2896016822555945,
      "objective/train/tokens_used": 246559200,
      "theoretical_loss": 4.2896016822555945,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047045135406218656,
      "loss": 3.0062,
      "theoretical_loss": 4.2896016822555945,
      "tokens_seen": 226099200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047044132397191575,
      "loss": 3.1042,
      "theoretical_loss": 4.289449073532892,
      "tokens_seen": 226164736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470431293881645,
      "loss": 3.3344,
      "theoretical_loss": 4.289296521403337,
      "tokens_seen": 226230272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704212637913741,
      "loss": 3.1889,
      "theoretical_loss": 4.289144025829555,
      "tokens_seen": 226295808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047041123370110334,
      "loss": 2.9189,
      "theoretical_loss": 4.288991586774211,
      "tokens_seen": 226361344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047040120361083247,
      "loss": 3.2327,
      "theoretical_loss": 4.288839204200003,
      "tokens_seen": 226426880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703911735205617,
      "loss": 2.9544,
      "theoretical_loss": 4.288686878069666,
      "tokens_seen": 226492416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703811434302909,
      "loss": 3.1703,
      "theoretical_loss": 4.288534608345969,
      "tokens_seen": 226557952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047037111334002007,
      "loss": 3.1722,
      "theoretical_loss": 4.288382394991717,
      "tokens_seen": 226623488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047036108324974925,
      "loss": 3.0562,
      "theoretical_loss": 4.288230237969751,
      "tokens_seen": 226689024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703510531594785,
      "loss": 3.074,
      "theoretical_loss": 4.288078137242946,
      "tokens_seen": 226754560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703410230692076,
      "loss": 2.9752,
      "theoretical_loss": 4.287926092774212,
      "tokens_seen": 226820096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047033099297893685,
      "loss": 2.8926,
      "theoretical_loss": 4.287774104526496,
      "tokens_seen": 226885632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470320962888666,
      "loss": 3.2959,
      "theoretical_loss": 4.287622172462777,
      "tokens_seen": 226951168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703109327983952,
      "loss": 3.1867,
      "theoretical_loss": 4.287470296546073,
      "tokens_seen": 227016704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703009027081244,
      "loss": 3.0586,
      "theoretical_loss": 4.287318476739433,
      "tokens_seen": 227082240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047029087261785357,
      "loss": 3.1358,
      "theoretical_loss": 4.287166713005945,
      "tokens_seen": 227147776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047028084252758275,
      "loss": 3.1048,
      "theoretical_loss": 4.287015005308728,
      "tokens_seen": 227213312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047027081243731193,
      "loss": 3.1529,
      "theoretical_loss": 4.286863353610939,
      "tokens_seen": 227278848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702607823470411,
      "loss": 3.0266,
      "theoretical_loss": 4.286711757875768,
      "tokens_seen": 227344384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047025075225677035,
      "loss": 3.1715,
      "theoretical_loss": 4.28656021806644,
      "tokens_seen": 227409920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702407221664995,
      "loss": 3.0616,
      "theoretical_loss": 4.286408734146216,
      "tokens_seen": 227475456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702306920762287,
      "loss": 3.1768,
      "theoretical_loss": 4.28625730607839,
      "tokens_seen": 227540992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047022066198595784,
      "loss": 3.1572,
      "theoretical_loss": 4.286105933826292,
      "tokens_seen": 227606528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702106318956871,
      "loss": 3.07,
      "theoretical_loss": 4.285954617353285,
      "tokens_seen": 227672064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 136262,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.876788377761841,
      "objective/train/theoretical_loss": 4.28580335662277,
      "objective/train/tokens_used": 248197600,
      "theoretical_loss": 4.28580335662277,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047020060180541626,
      "loss": 3.0746,
      "theoretical_loss": 4.28580335662277,
      "tokens_seen": 227737600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047019057171514544,
      "loss": 3.2485,
      "theoretical_loss": 4.285652151598178,
      "tokens_seen": 227803136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701805416248746,
      "loss": 3.2203,
      "theoretical_loss": 4.285501002242977,
      "tokens_seen": 227868672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047017051153460385,
      "loss": 3.0961,
      "theoretical_loss": 4.285349908520671,
      "tokens_seen": 227934208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470160481444333,
      "loss": 3.1869,
      "theoretical_loss": 4.285198870394796,
      "tokens_seen": 227999744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701504513540622,
      "loss": 3.154,
      "theoretical_loss": 4.28504788782892,
      "tokens_seen": 228065280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047014042126379134,
      "loss": 3.1927,
      "theoretical_loss": 4.284896960786653,
      "tokens_seen": 228130816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701303911735206,
      "loss": 3.2737,
      "theoretical_loss": 4.284746089231631,
      "tokens_seen": 228196352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047012036108324976,
      "loss": 3.299,
      "theoretical_loss": 4.28459527312753,
      "tokens_seen": 228261888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047011033099297894,
      "loss": 2.9585,
      "theoretical_loss": 4.284444512438057,
      "tokens_seen": 228327424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701003009027081,
      "loss": 3.1119,
      "theoretical_loss": 4.2842938071269545,
      "tokens_seen": 228392960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700902708124373,
      "loss": 3.165,
      "theoretical_loss": 4.284143157157999,
      "tokens_seen": 228458496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700802407221665,
      "loss": 3.1586,
      "theoretical_loss": 4.2839925624949995,
      "tokens_seen": 228524032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700702106318957,
      "loss": 3.1651,
      "theoretical_loss": 4.283842023101803,
      "tokens_seen": 228589568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047006018054162484,
      "loss": 3.2229,
      "theoretical_loss": 4.283691538942286,
      "tokens_seen": 228655104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700501504513541,
      "loss": 3.192,
      "theoretical_loss": 4.28354110998036,
      "tokens_seen": 228720640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047004012036108326,
      "loss": 3.03,
      "theoretical_loss": 4.283390736179975,
      "tokens_seen": 228786176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047003009027081244,
      "loss": 3.3334,
      "theoretical_loss": 4.283240417505107,
      "tokens_seen": 228851712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700200601805416,
      "loss": 2.9563,
      "theoretical_loss": 4.2830901539197725,
      "tokens_seen": 228917248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700100300902708,
      "loss": 3.0645,
      "theoretical_loss": 4.282939945388019,
      "tokens_seen": 228982784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047,
      "loss": 3.2184,
      "theoretical_loss": 4.2827897918739275,
      "tokens_seen": 229048320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699899699097292,
      "loss": 3.3057,
      "theoretical_loss": 4.282639693341613,
      "tokens_seen": 229113856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046997993981945835,
      "loss": 3.179,
      "theoretical_loss": 4.282489649755225,
      "tokens_seen": 229179392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699699097291876,
      "loss": 3.0861,
      "theoretical_loss": 4.282339661078946,
      "tokens_seen": 229244928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046995987963891676,
      "loss": 3.2011,
      "theoretical_loss": 4.282189727276991,
      "tokens_seen": 229310464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 137363,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0329909324645996,
      "objective/train/theoretical_loss": 4.282039848313611,
      "objective/train/tokens_used": 249836000,
      "theoretical_loss": 4.282039848313611,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046994984954864595,
      "loss": 3.003,
      "theoretical_loss": 4.282039848313611,
      "tokens_seen": 229376000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699398194583752,
      "loss": 3.2598,
      "theoretical_loss": 4.281890024153089,
      "tokens_seen": 229441536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699297893681043,
      "loss": 3.1051,
      "theoretical_loss": 4.2817402547597405,
      "tokens_seen": 229507072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046991975927783354,
      "loss": 3.2666,
      "theoretical_loss": 4.281590540097916,
      "tokens_seen": 229572608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046990972918756267,
      "loss": 3.1706,
      "theoretical_loss": 4.281440880131999,
      "tokens_seen": 229638144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698996990972919,
      "loss": 2.882,
      "theoretical_loss": 4.2812912748264065,
      "tokens_seen": 229703680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698896690070211,
      "loss": 3.2516,
      "theoretical_loss": 4.281141724145588,
      "tokens_seen": 229769216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046987963891675027,
      "loss": 3.1911,
      "theoretical_loss": 4.280992228054027,
      "tokens_seen": 229834752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046986960882647945,
      "loss": 3.337,
      "theoretical_loss": 4.28084278651624,
      "tokens_seen": 229900288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698595787362087,
      "loss": 3.1908,
      "theoretical_loss": 4.280693399496776,
      "tokens_seen": 229965824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698495486459378,
      "loss": 3.0165,
      "theoretical_loss": 4.280544066960219,
      "tokens_seen": 230031360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046983951855566705,
      "loss": 2.9808,
      "theoretical_loss": 4.280394788871183,
      "tokens_seen": 230096896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698294884653962,
      "loss": 3.0817,
      "theoretical_loss": 4.280245565194319,
      "tokens_seen": 230162432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698194583751254,
      "loss": 3.208,
      "theoretical_loss": 4.280096395894308,
      "tokens_seen": 230227968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698094282848546,
      "loss": 3.1627,
      "theoretical_loss": 4.279947280935864,
      "tokens_seen": 230293504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046979939819458377,
      "loss": 3.0551,
      "theoretical_loss": 4.279798220283736,
      "tokens_seen": 230359040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046978936810431295,
      "loss": 3.0795,
      "theoretical_loss": 4.279649213902704,
      "tokens_seen": 230424576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046977933801404213,
      "loss": 3.3708,
      "theoretical_loss": 4.2795002617575815,
      "tokens_seen": 230490112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697693079237713,
      "loss": 3.1081,
      "theoretical_loss": 4.279351363813215,
      "tokens_seen": 230555648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046975927783350055,
      "loss": 2.8502,
      "theoretical_loss": 4.2792025200344845,
      "tokens_seen": 230621184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697492477432297,
      "loss": 3.1672,
      "theoretical_loss": 4.279053730386301,
      "tokens_seen": 230686720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697392176529589,
      "loss": 3.1429,
      "theoretical_loss": 4.278904994833609,
      "tokens_seen": 230752256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046972918756268804,
      "loss": 3.0444,
      "theoretical_loss": 4.278756313341386,
      "tokens_seen": 230817792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697191574724173,
      "loss": 3.0223,
      "theoretical_loss": 4.278607685874642,
      "tokens_seen": 230883328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046970912738214646,
      "loss": 2.8679,
      "theoretical_loss": 4.278459112398418,
      "tokens_seen": 230948864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 138640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.336690664291382,
      "objective/train/theoretical_loss": 4.278310592877791,
      "objective/train/tokens_used": 251474400,
      "theoretical_loss": 4.278310592877791,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046969909729187564,
      "loss": 2.9576,
      "theoretical_loss": 4.278310592877791,
      "tokens_seen": 231014400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696890672016048,
      "loss": 2.9152,
      "theoretical_loss": 4.2781621272778665,
      "tokens_seen": 231079936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046967903711133405,
      "loss": 3.0227,
      "theoretical_loss": 4.278013715563786,
      "tokens_seen": 231145472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696690070210632,
      "loss": 3.1015,
      "theoretical_loss": 4.27786535770072,
      "tokens_seen": 231211008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696589769307924,
      "loss": 3.1436,
      "theoretical_loss": 4.277717053653875,
      "tokens_seen": 231276544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046964894684052154,
      "loss": 3.424,
      "theoretical_loss": 4.277568803388486,
      "tokens_seen": 231342080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696389167502508,
      "loss": 2.9966,
      "theoretical_loss": 4.277420606869824,
      "tokens_seen": 231407616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046962888665997996,
      "loss": 2.9781,
      "theoretical_loss": 4.27727246406319,
      "tokens_seen": 231473152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046961885656970914,
      "loss": 3.153,
      "theoretical_loss": 4.277124374933917,
      "tokens_seen": 231538688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696088264794383,
      "loss": 3.3026,
      "theoretical_loss": 4.276976339447372,
      "tokens_seen": 231604224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695987963891675,
      "loss": 3.1475,
      "theoretical_loss": 4.276828357568952,
      "tokens_seen": 231669760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695887662988967,
      "loss": 2.9757,
      "theoretical_loss": 4.276680429264088,
      "tokens_seen": 231735296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695787362086259,
      "loss": 3.2429,
      "theoretical_loss": 4.276532554498242,
      "tokens_seen": 231800832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046956870611835505,
      "loss": 2.9495,
      "theoretical_loss": 4.276384733236909,
      "tokens_seen": 231866368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695586760280843,
      "loss": 3.2042,
      "theoretical_loss": 4.276236965445612,
      "tokens_seen": 231931904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046954864593781346,
      "loss": 3.0181,
      "theoretical_loss": 4.276089251089914,
      "tokens_seen": 231997440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046953861584754264,
      "loss": 3.0194,
      "theoretical_loss": 4.275941590135401,
      "tokens_seen": 232062976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695285857572718,
      "loss": 3.1368,
      "theoretical_loss": 4.275793982547698,
      "tokens_seen": 232128512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469518555667001,
      "loss": 3.1842,
      "theoretical_loss": 4.275646428292456,
      "tokens_seen": 232194048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695085255767302,
      "loss": 3.0394,
      "theoretical_loss": 4.275498927335362,
      "tokens_seen": 232259584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694984954864594,
      "loss": 3.2062,
      "theoretical_loss": 4.275351479642134,
      "tokens_seen": 232325120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046948846539618855,
      "loss": 3.2058,
      "theoretical_loss": 4.27520408517852,
      "tokens_seen": 232390656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694784353059178,
      "loss": 2.9021,
      "theoretical_loss": 4.275056743910301,
      "tokens_seen": 232456192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694684052156469,
      "loss": 3.1015,
      "theoretical_loss": 4.2749094558032885,
      "tokens_seen": 232521728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046945837512537615,
      "loss": 3.1834,
      "theoretical_loss": 4.274762220823328,
      "tokens_seen": 232587264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 139202,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1069910526275635,
      "objective/train/theoretical_loss": 4.274615038936293,
      "objective/train/tokens_used": 253112800,
      "theoretical_loss": 4.274615038936293,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694483450351053,
      "loss": 3.0744,
      "theoretical_loss": 4.274615038936293,
      "tokens_seen": 232652800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694383149448345,
      "loss": 3.0827,
      "theoretical_loss": 4.274467910108093,
      "tokens_seen": 232718336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694282848545637,
      "loss": 3.0734,
      "theoretical_loss": 4.274320834304666,
      "tokens_seen": 232783872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046941825476429287,
      "loss": 3.1049,
      "theoretical_loss": 4.27417381149198,
      "tokens_seen": 232849408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046940822467402205,
      "loss": 3.0458,
      "theoretical_loss": 4.274026841636037,
      "tokens_seen": 232914944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693981945837513,
      "loss": 3.2492,
      "theoretical_loss": 4.273879924702872,
      "tokens_seen": 232980480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693881644934804,
      "loss": 3.0141,
      "theoretical_loss": 4.273733060658546,
      "tokens_seen": 233046016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046937813440320965,
      "loss": 2.9252,
      "theoretical_loss": 4.273586249469155,
      "tokens_seen": 233111552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046936810431293883,
      "loss": 3.0968,
      "theoretical_loss": 4.273439491100828,
      "tokens_seen": 233177088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469358074222668,
      "loss": 2.9294,
      "theoretical_loss": 4.273292785519722,
      "tokens_seen": 233242624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693480441323972,
      "loss": 2.9721,
      "theoretical_loss": 4.273146132692023,
      "tokens_seen": 233308160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693380140421264,
      "loss": 3.2223,
      "theoretical_loss": 4.272999532583955,
      "tokens_seen": 233373696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046932798395185555,
      "loss": 3.1393,
      "theoretical_loss": 4.272852985161768,
      "tokens_seen": 233439232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693179538615848,
      "loss": 3.2015,
      "theoretical_loss": 4.272706490391743,
      "tokens_seen": 233504768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693079237713139,
      "loss": 3.0477,
      "theoretical_loss": 4.272560048240195,
      "tokens_seen": 233570304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046929789368104315,
      "loss": 3.0628,
      "theoretical_loss": 4.272413658673468,
      "tokens_seen": 233635840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692878635907723,
      "loss": 3.2822,
      "theoretical_loss": 4.272267321657937,
      "tokens_seen": 233701376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692778335005015,
      "loss": 3.2682,
      "theoretical_loss": 4.272121037160009,
      "tokens_seen": 233766912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692678034102307,
      "loss": 3.1303,
      "theoretical_loss": 4.271974805146121,
      "tokens_seen": 233832448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692577733199599,
      "loss": 3.1491,
      "theoretical_loss": 4.27182862558274,
      "tokens_seen": 233897984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046924774322968906,
      "loss": 3.1131,
      "theoretical_loss": 4.271682498436365,
      "tokens_seen": 233963520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046923771313941824,
      "loss": 3.1811,
      "theoretical_loss": 4.271536423673528,
      "tokens_seen": 234029056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692276830491474,
      "loss": 3.2045,
      "theoretical_loss": 4.271390401260785,
      "tokens_seen": 234094592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046921765295887666,
      "loss": 3.1978,
      "theoretical_loss": 4.271244431164731,
      "tokens_seen": 234160128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046920762286860584,
      "loss": 3.2577,
      "theoretical_loss": 4.271098513351986,
      "tokens_seen": 234225664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 140513,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1093058586120605,
      "objective/train/theoretical_loss": 4.270952647789202,
      "objective/train/tokens_used": 254751200,
      "theoretical_loss": 4.270952647789202,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469197592778335,
      "loss": 3.3018,
      "theoretical_loss": 4.270952647789202,
      "tokens_seen": 234291200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046918756268806425,
      "loss": 3.1679,
      "theoretical_loss": 4.2708068344430625,
      "tokens_seen": 234356736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691775325977934,
      "loss": 3.092,
      "theoretical_loss": 4.270661073280282,
      "tokens_seen": 234422272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691675025075226,
      "loss": 3.2188,
      "theoretical_loss": 4.2705153642676015,
      "tokens_seen": 234487808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046915747241725174,
      "loss": 3.162,
      "theoretical_loss": 4.270369707371799,
      "tokens_seen": 234553344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469147442326981,
      "loss": 3.123,
      "theoretical_loss": 4.270224102559677,
      "tokens_seen": 234618880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046913741223671016,
      "loss": 3.2137,
      "theoretical_loss": 4.270078549798071,
      "tokens_seen": 234684416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046912738214643934,
      "loss": 3.1062,
      "theoretical_loss": 4.269933049053847,
      "tokens_seen": 234749952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691173520561685,
      "loss": 3.2631,
      "theoretical_loss": 4.269787600293903,
      "tokens_seen": 234815488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691073219658977,
      "loss": 3.0201,
      "theoretical_loss": 4.269642203485162,
      "tokens_seen": 234881024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690972918756269,
      "loss": 2.9759,
      "theoretical_loss": 4.269496858594582,
      "tokens_seen": 234946560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690872617853561,
      "loss": 3.1686,
      "theoretical_loss": 4.269351565589151,
      "tokens_seen": 235012096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046907723169508525,
      "loss": 3.1832,
      "theoretical_loss": 4.269206324435885,
      "tokens_seen": 235077632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690672016048145,
      "loss": 3.1976,
      "theoretical_loss": 4.26906113510183,
      "tokens_seen": 235143168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046905717151454366,
      "loss": 3.1711,
      "theoretical_loss": 4.268915997554065,
      "tokens_seen": 235208704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046904714142427284,
      "loss": 3.105,
      "theoretical_loss": 4.268770911759697,
      "tokens_seen": 235274240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469037111334002,
      "loss": 2.9095,
      "theoretical_loss": 4.268625877685863,
      "tokens_seen": 235339776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690270812437312,
      "loss": 3.0525,
      "theoretical_loss": 4.268480895299731,
      "tokens_seen": 235405312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690170511534604,
      "loss": 3.2324,
      "theoretical_loss": 4.268335964568497,
      "tokens_seen": 235470848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690070210631896,
      "loss": 3.1499,
      "theoretical_loss": 4.26819108545939,
      "tokens_seen": 235536384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046899699097291875,
      "loss": 3.1792,
      "theoretical_loss": 4.268046257939666,
      "tokens_seen": 235601920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468986960882648,
      "loss": 3.2761,
      "theoretical_loss": 4.2679014819766135,
      "tokens_seen": 235667456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689769307923771,
      "loss": 3.0686,
      "theoretical_loss": 4.267756757537548,
      "tokens_seen": 235732992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046896690070210635,
      "loss": 3.2238,
      "theoretical_loss": 4.267612084589816,
      "tokens_seen": 235798528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046895687061183553,
      "loss": 3.346,
      "theoretical_loss": 4.267467463100796,
      "tokens_seen": 235864064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 141147,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.236262321472168,
      "objective/train/theoretical_loss": 4.267322893037893,
      "objective/train/tokens_used": 256389600,
      "theoretical_loss": 4.267322893037893,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689468405215647,
      "loss": 3.064,
      "theoretical_loss": 4.267322893037893,
      "tokens_seen": 235929600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689368104312939,
      "loss": 3.0627,
      "theoretical_loss": 4.267178374368543,
      "tokens_seen": 235995136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892678034102307,
      "loss": 3.1095,
      "theoretical_loss": 4.267033907060211,
      "tokens_seen": 236060672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046891675025075225,
      "loss": 3.0446,
      "theoretical_loss": 4.266889491080392,
      "tokens_seen": 236126208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689067201604815,
      "loss": 3.1192,
      "theoretical_loss": 4.2667451263966125,
      "tokens_seen": 236191744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688966900702106,
      "loss": 2.9901,
      "theoretical_loss": 4.266600812976426,
      "tokens_seen": 236257280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046888665997993985,
      "loss": 3.1156,
      "theoretical_loss": 4.2664565507874155,
      "tokens_seen": 236322816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046887662988966903,
      "loss": 3.1053,
      "theoretical_loss": 4.266312339797195,
      "tokens_seen": 236388352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688665997993982,
      "loss": 3.0449,
      "theoretical_loss": 4.266168179973407,
      "tokens_seen": 236453888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688565697091274,
      "loss": 3.2738,
      "theoretical_loss": 4.266024071283725,
      "tokens_seen": 236519424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688465396188566,
      "loss": 2.8719,
      "theoretical_loss": 4.265880013695849,
      "tokens_seen": 236584960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046883650952858575,
      "loss": 3.2135,
      "theoretical_loss": 4.265736007177511,
      "tokens_seen": 236650496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468826479438315,
      "loss": 3.0794,
      "theoretical_loss": 4.265592051696469,
      "tokens_seen": 236716032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688164493480441,
      "loss": 2.9949,
      "theoretical_loss": 4.265448147220515,
      "tokens_seen": 236781568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046880641925777335,
      "loss": 2.8953,
      "theoretical_loss": 4.265304293717467,
      "tokens_seen": 236847104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687963891675025,
      "loss": 2.9984,
      "theoretical_loss": 4.2651604911551715,
      "tokens_seen": 236912640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687863590772317,
      "loss": 3.054,
      "theoretical_loss": 4.265016739501508,
      "tokens_seen": 236978176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687763289869609,
      "loss": 3.052,
      "theoretical_loss": 4.264873038724381,
      "tokens_seen": 237043712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687662988966901,
      "loss": 3.0228,
      "theoretical_loss": 4.264729388791726,
      "tokens_seen": 237109248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046875626880641926,
      "loss": 3.0213,
      "theoretical_loss": 4.264585789671508,
      "tokens_seen": 237174784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046874623871614844,
      "loss": 3.2672,
      "theoretical_loss": 4.26444224133172,
      "tokens_seen": 237240320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687362086258776,
      "loss": 3.0529,
      "theoretical_loss": 4.264298743740384,
      "tokens_seen": 237305856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872617853560686,
      "loss": 2.9217,
      "theoretical_loss": 4.264155296865552,
      "tokens_seen": 237371392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468716148445336,
      "loss": 3.0414,
      "theoretical_loss": 4.264011900675305,
      "tokens_seen": 237436928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687061183550652,
      "loss": 3.0795,
      "theoretical_loss": 4.26386855513775,
      "tokens_seen": 237502464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 142360,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9394404888153076,
      "objective/train/theoretical_loss": 4.263725260221027,
      "objective/train/tokens_used": 258028000,
      "theoretical_loss": 4.263725260221027,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686960882647944,
      "loss": 3.1755,
      "theoretical_loss": 4.263725260221027,
      "tokens_seen": 237568000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686860581745236,
      "loss": 3.2614,
      "theoretical_loss": 4.263582015893302,
      "tokens_seen": 237633536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046867602808425276,
      "loss": 3.2817,
      "theoretical_loss": 4.263438822122771,
      "tokens_seen": 237699072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046866599799398194,
      "loss": 3.1451,
      "theoretical_loss": 4.263295678877658,
      "tokens_seen": 237764608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686559679037111,
      "loss": 3.2993,
      "theoretical_loss": 4.263152586126217,
      "tokens_seen": 237830144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864593781344036,
      "loss": 3.1979,
      "theoretical_loss": 4.263009543836728,
      "tokens_seen": 237895680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686359077231695,
      "loss": 3.1446,
      "theoretical_loss": 4.2628665519775035,
      "tokens_seen": 237961216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686258776328987,
      "loss": 3.3129,
      "theoretical_loss": 4.262723610516881,
      "tokens_seen": 238026752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046861584754262785,
      "loss": 3.0843,
      "theoretical_loss": 4.26258071942323,
      "tokens_seen": 238092288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686058174523571,
      "loss": 3.0462,
      "theoretical_loss": 4.262437878664945,
      "tokens_seen": 238157824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046859578736208626,
      "loss": 3.1094,
      "theoretical_loss": 4.262295088210451,
      "tokens_seen": 238223360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046858575727181545,
      "loss": 3.0836,
      "theoretical_loss": 4.262152348028202,
      "tokens_seen": 238288896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685757271815446,
      "loss": 3.2288,
      "theoretical_loss": 4.26200965808668,
      "tokens_seen": 238354432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656970912738,
      "loss": 3.2874,
      "theoretical_loss": 4.261867018354394,
      "tokens_seen": 238419968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468555667001003,
      "loss": 3.0589,
      "theoretical_loss": 4.261724428799884,
      "tokens_seen": 238485504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685456369107322,
      "loss": 3.0927,
      "theoretical_loss": 4.261581889391715,
      "tokens_seen": 238551040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046853560682046135,
      "loss": 3.2002,
      "theoretical_loss": 4.2614394000984825,
      "tokens_seen": 238616576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685255767301906,
      "loss": 3.1648,
      "theoretical_loss": 4.261296960888811,
      "tokens_seen": 238682112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046851554663991977,
      "loss": 3.3789,
      "theoretical_loss": 4.2611545717313515,
      "tokens_seen": 238747648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046850551654964895,
      "loss": 3.0458,
      "theoretical_loss": 4.261012232594784,
      "tokens_seen": 238813184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046849548645937813,
      "loss": 2.9991,
      "theoretical_loss": 4.260869943447817,
      "tokens_seen": 238878720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684854563691073,
      "loss": 3.0852,
      "theoretical_loss": 4.260727704259187,
      "tokens_seen": 238944256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684754262788365,
      "loss": 3.372,
      "theoretical_loss": 4.260585514997658,
      "tokens_seen": 239009792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046846539618856573,
      "loss": 3.0656,
      "theoretical_loss": 4.260443375632022,
      "tokens_seen": 239075328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684553660982949,
      "loss": 3.048,
      "theoretical_loss": 4.260301286131099,
      "tokens_seen": 239140864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 143562,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0513010025024414,
      "objective/train/theoretical_loss": 4.26015924646374,
      "objective/train/tokens_used": 259666400,
      "theoretical_loss": 4.26015924646374,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684453360080241,
      "loss": 2.9484,
      "theoretical_loss": 4.26015924646374,
      "tokens_seen": 239206400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046843530591775327,
      "loss": 3.4032,
      "theoretical_loss": 4.26001725659882,
      "tokens_seen": 239271936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046842527582748245,
      "loss": 3.1223,
      "theoretical_loss": 4.2598753165052425,
      "tokens_seen": 239337472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684152457372117,
      "loss": 3.1745,
      "theoretical_loss": 4.259733426151941,
      "tokens_seen": 239403008
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684052156469408,
      "loss": 3.0693,
      "theoretical_loss": 4.259591585507876,
      "tokens_seen": 239468544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046839518555667005,
      "loss": 3.2599,
      "theoretical_loss": 4.259449794542036,
      "tokens_seen": 239534080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838515546639923,
      "loss": 3.051,
      "theoretical_loss": 4.259308053223437,
      "tokens_seen": 239599616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683751253761284,
      "loss": 3.2346,
      "theoretical_loss": 4.259166361521121,
      "tokens_seen": 239665152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683650952858576,
      "loss": 3.1637,
      "theoretical_loss": 4.259024719404161,
      "tokens_seen": 239730688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683550651955868,
      "loss": 3.0852,
      "theoretical_loss": 4.258883126841656,
      "tokens_seen": 239796224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834503510531595,
      "loss": 3.0005,
      "theoretical_loss": 4.258741583802734,
      "tokens_seen": 239861760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683350050150452,
      "loss": 3.2479,
      "theoretical_loss": 4.258600090256548,
      "tokens_seen": 239927296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683249749247743,
      "loss": 2.7133,
      "theoretical_loss": 4.258458646172281,
      "tokens_seen": 239992832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046831494483450355,
      "loss": 3.1386,
      "theoretical_loss": 4.258317251519143,
      "tokens_seen": 240058368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683049147442327,
      "loss": 3.2625,
      "theoretical_loss": 4.258175906266371,
      "tokens_seen": 240123904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682948846539619,
      "loss": 3.3216,
      "theoretical_loss": 4.2580346103832305,
      "tokens_seen": 240189440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682848545636911,
      "loss": 3.1082,
      "theoretical_loss": 4.257893363839013,
      "tokens_seen": 240254976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682748244734203,
      "loss": 2.8933,
      "theoretical_loss": 4.257752166603041,
      "tokens_seen": 240320512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046826479438314946,
      "loss": 3.1841,
      "theoretical_loss": 4.257611018644658,
      "tokens_seen": 240386048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046825476429287864,
      "loss": 2.9544,
      "theoretical_loss": 4.257469919933241,
      "tokens_seen": 240451584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682447342026078,
      "loss": 3.0391,
      "theoretical_loss": 4.257328870438193,
      "tokens_seen": 240517120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046823470411233706,
      "loss": 2.95,
      "theoretical_loss": 4.257187870128942,
      "tokens_seen": 240582656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682246740220662,
      "loss": 3.0321,
      "theoretical_loss": 4.257046918974945,
      "tokens_seen": 240648192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682146439317954,
      "loss": 3.1266,
      "theoretical_loss": 4.2569060169456865,
      "tokens_seen": 240713728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682046138415246,
      "loss": 3.1535,
      "theoretical_loss": 4.256765164010678,
      "tokens_seen": 240779264
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 144045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4141147136688232,
      "objective/train/theoretical_loss": 4.256624360139456,
      "objective/train/tokens_used": 261304800,
      "theoretical_loss": 4.256624360139456,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681945837512538,
      "loss": 3.314,
      "theoretical_loss": 4.256624360139456,
      "tokens_seen": 240844800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046818455366098296,
      "loss": 3.0936,
      "theoretical_loss": 4.25648360530159,
      "tokens_seen": 240910336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046817452357071214,
      "loss": 2.9071,
      "theoretical_loss": 4.25634289946667,
      "tokens_seen": 240975872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681644934804413,
      "loss": 3.0465,
      "theoretical_loss": 4.256202242604317,
      "tokens_seen": 241041408
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046815446339017056,
      "loss": 3.2584,
      "theoretical_loss": 4.256061634684178,
      "tokens_seen": 241106944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681444332998997,
      "loss": 3.0502,
      "theoretical_loss": 4.255921075675927,
      "tokens_seen": 241172480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681344032096289,
      "loss": 3.1567,
      "theoretical_loss": 4.255780565549264,
      "tokens_seen": 241238016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046812437311935805,
      "loss": 3.0096,
      "theoretical_loss": 4.255640104273919,
      "tokens_seen": 241303552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681143430290873,
      "loss": 2.7541,
      "theoretical_loss": 4.255499691819647,
      "tokens_seen": 241369088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810431293881646,
      "loss": 3.1162,
      "theoretical_loss": 4.255359328156229,
      "tokens_seen": 241434624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046809428284854565,
      "loss": 3.1105,
      "theoretical_loss": 4.255219013253475,
      "tokens_seen": 241500160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680842527582748,
      "loss": 3.1662,
      "theoretical_loss": 4.255078747081219,
      "tokens_seen": 241565696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468074222668004,
      "loss": 3.0573,
      "theoretical_loss": 4.254938529609326,
      "tokens_seen": 241631232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680641925777332,
      "loss": 3.2336,
      "theoretical_loss": 4.254798360807684,
      "tokens_seen": 241696768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680541624874624,
      "loss": 3.1564,
      "theoretical_loss": 4.254658240646211,
      "tokens_seen": 241762304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046804413239719155,
      "loss": 3.0097,
      "theoretical_loss": 4.254518169094847,
      "tokens_seen": 241827840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680341023069208,
      "loss": 3.0337,
      "theoretical_loss": 4.254378146123563,
      "tokens_seen": 241893376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802407221664997,
      "loss": 2.9549,
      "theoretical_loss": 4.254238171702356,
      "tokens_seen": 241958912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046801404212637915,
      "loss": 2.976,
      "theoretical_loss": 4.2540982458012495,
      "tokens_seen": 242024448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800401203610833,
      "loss": 3.1121,
      "theoretical_loss": 4.253958368390292,
      "tokens_seen": 242089984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679939819458375,
      "loss": 3.1283,
      "theoretical_loss": 4.2538185394395605,
      "tokens_seen": 242155520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679839518555667,
      "loss": 3.2828,
      "theoretical_loss": 4.253678758919156,
      "tokens_seen": 242221056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046797392176529593,
      "loss": 3.1396,
      "theoretical_loss": 4.25353902679921,
      "tokens_seen": 242286592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046796389167502505,
      "loss": 3.2791,
      "theoretical_loss": 4.253399343049878,
      "tokens_seen": 242352128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679538615847543,
      "loss": 3.0178,
      "theoretical_loss": 4.253259707641341,
      "tokens_seen": 242417664
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 145400,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.983632802963257,
      "objective/train/theoretical_loss": 4.253120120543809,
      "objective/train/tokens_used": 262943200,
      "theoretical_loss": 4.253120120543809,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679438314944834,
      "loss": 3.1131,
      "theoretical_loss": 4.253120120543809,
      "tokens_seen": 242483200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046793380140421265,
      "loss": 3.1157,
      "theoretical_loss": 4.252980581727517,
      "tokens_seen": 242548736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046792377131394183,
      "loss": 3.3695,
      "theoretical_loss": 4.2528410911627255,
      "tokens_seen": 242614272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467913741223671,
      "loss": 3.0132,
      "theoretical_loss": 4.2527016488197225,
      "tokens_seen": 242679808
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679037111334002,
      "loss": 3.2259,
      "theoretical_loss": 4.252562254668824,
      "tokens_seen": 242745344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046789368104312943,
      "loss": 3.3286,
      "theoretical_loss": 4.2524229086803675,
      "tokens_seen": 242810880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046788365095285856,
      "loss": 2.9194,
      "theoretical_loss": 4.252283610824723,
      "tokens_seen": 242876416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678736208625878,
      "loss": 3.1186,
      "theoretical_loss": 4.25214436107228,
      "tokens_seen": 242941952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678635907723169,
      "loss": 3.0458,
      "theoretical_loss": 4.25200515939346,
      "tokens_seen": 243007488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046785356068204616,
      "loss": 3.1178,
      "theoretical_loss": 4.251866005758707,
      "tokens_seen": 243073024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046784353059177534,
      "loss": 2.9132,
      "theoretical_loss": 4.251726900138494,
      "tokens_seen": 243138560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678335005015045,
      "loss": 3.1386,
      "theoretical_loss": 4.251587842503317,
      "tokens_seen": 243204096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678234704112337,
      "loss": 3.1516,
      "theoretical_loss": 4.2514488328237,
      "tokens_seen": 243269632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678134403209629,
      "loss": 3.0271,
      "theoretical_loss": 4.251309871070193,
      "tokens_seen": 243335168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046780341023069206,
      "loss": 3.1534,
      "theoretical_loss": 4.25117095721337,
      "tokens_seen": 243400704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677933801404213,
      "loss": 3.2002,
      "theoretical_loss": 4.251032091223834,
      "tokens_seen": 243466240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677833500501504,
      "loss": 3.2106,
      "theoretical_loss": 4.250893273072213,
      "tokens_seen": 243531776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046777331995987966,
      "loss": 3.0027,
      "theoretical_loss": 4.250754502729161,
      "tokens_seen": 243597312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677632898696088,
      "loss": 3.0599,
      "theoretical_loss": 4.250615780165355,
      "tokens_seen": 243662848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467753259779338,
      "loss": 2.7333,
      "theoretical_loss": 4.250477105351502,
      "tokens_seen": 243728384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677432296890672,
      "loss": 3.1958,
      "theoretical_loss": 4.250338478258333,
      "tokens_seen": 243793920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677331995987964,
      "loss": 3.2165,
      "theoretical_loss": 4.250199898856604,
      "tokens_seen": 243859456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046772316950852556,
      "loss": 3.0671,
      "theoretical_loss": 4.250061367117097,
      "tokens_seen": 243924992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677131394182548,
      "loss": 3.0396,
      "theoretical_loss": 4.249922883010623,
      "tokens_seen": 243990528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467703109327984,
      "loss": 3.0198,
      "theoretical_loss": 4.249784446508014,
      "tokens_seen": 244056064
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 146117,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.144657850265503,
      "objective/train/theoretical_loss": 4.24964605758013,
      "objective/train/tokens_used": 264581600,
      "theoretical_loss": 4.24964605758013,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046769307923771316,
      "loss": 3.141,
      "theoretical_loss": 4.24964605758013,
      "tokens_seen": 244121600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046768304914744234,
      "loss": 3.0374,
      "theoretical_loss": 4.249507716197858,
      "tokens_seen": 244187136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676730190571715,
      "loss": 2.9348,
      "theoretical_loss": 4.249369422332107,
      "tokens_seen": 244252672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046766298896690076,
      "loss": 3.2713,
      "theoretical_loss": 4.249231175953814,
      "tokens_seen": 244318208
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676529588766299,
      "loss": 3.0096,
      "theoretical_loss": 4.249092977033943,
      "tokens_seen": 244383744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676429287863591,
      "loss": 3.0521,
      "theoretical_loss": 4.248954825543479,
      "tokens_seen": 244449280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046763289869608825,
      "loss": 3.049,
      "theoretical_loss": 4.248816721453438,
      "tokens_seen": 244514816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676228686058175,
      "loss": 3.0529,
      "theoretical_loss": 4.248678664734857,
      "tokens_seen": 244580352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046761283851554666,
      "loss": 2.8181,
      "theoretical_loss": 4.248540655358801,
      "tokens_seen": 244645888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046760280842527585,
      "loss": 2.9386,
      "theoretical_loss": 4.2484026932963594,
      "tokens_seen": 244711424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467592778335005,
      "loss": 3.1262,
      "theoretical_loss": 4.248264778518647,
      "tokens_seen": 244776960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675827482447342,
      "loss": 3.0133,
      "theoretical_loss": 4.248126910996804,
      "tokens_seen": 244842496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675727181544634,
      "loss": 2.8972,
      "theoretical_loss": 4.2479890907019975,
      "tokens_seen": 244908032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675626880641926,
      "loss": 2.9783,
      "theoretical_loss": 4.2478513176054165,
      "tokens_seen": 244973568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046755265797392175,
      "loss": 3.045,
      "theoretical_loss": 4.247713591678278,
      "tokens_seen": 245039104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467542627883651,
      "loss": 2.8848,
      "theoretical_loss": 4.247575912891824,
      "tokens_seen": 245104640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046753259779338017,
      "loss": 3.0541,
      "theoretical_loss": 4.247438281217321,
      "tokens_seen": 245170176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046752256770310935,
      "loss": 3.0006,
      "theoretical_loss": 4.247300696626062,
      "tokens_seen": 245235712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046751253761283853,
      "loss": 3.1447,
      "theoretical_loss": 4.247163159089362,
      "tokens_seen": 245301248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675025075225677,
      "loss": 3.204,
      "theoretical_loss": 4.247025668578564,
      "tokens_seen": 245366784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674924774322969,
      "loss": 3.2769,
      "theoretical_loss": 4.246888225065036,
      "tokens_seen": 245432320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046748244734202613,
      "loss": 2.7871,
      "theoretical_loss": 4.24675082852017,
      "tokens_seen": 245497856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046747241725175525,
      "loss": 2.8847,
      "theoretical_loss": 4.246613478915384,
      "tokens_seen": 245563392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674623871614845,
      "loss": 3.0322,
      "theoretical_loss": 4.246476176222119,
      "tokens_seen": 245628928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674523570712136,
      "loss": 3.3512,
      "theoretical_loss": 4.246338920411843,
      "tokens_seen": 245694464
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 147054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2560789585113525,
      "objective/train/theoretical_loss": 4.24620171145605,
      "objective/train/tokens_used": 266220000,
      "theoretical_loss": 4.24620171145605,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046744232698094285,
      "loss": 3.1654,
      "theoretical_loss": 4.24620171145605,
      "tokens_seen": 245760000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046743229689067203,
      "loss": 2.945,
      "theoretical_loss": 4.246064549326255,
      "tokens_seen": 245825536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674222668004012,
      "loss": 2.9211,
      "theoretical_loss": 4.245927433994002,
      "tokens_seen": 245891072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674122367101304,
      "loss": 3.1739,
      "theoretical_loss": 4.245790365430857,
      "tokens_seen": 245956608
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046740220661985963,
      "loss": 3.2878,
      "theoretical_loss": 4.245653343608412,
      "tokens_seen": 246022144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046739217652958876,
      "loss": 2.9528,
      "theoretical_loss": 4.245516368498285,
      "tokens_seen": 246087680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467382146439318,
      "loss": 3.1466,
      "theoretical_loss": 4.245379440072116,
      "tokens_seen": 246153216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673721163490471,
      "loss": 2.9806,
      "theoretical_loss": 4.245242558301572,
      "tokens_seen": 246218752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046736208625877636,
      "loss": 2.9725,
      "theoretical_loss": 4.2451057231583444,
      "tokens_seen": 246284288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046735205616850554,
      "loss": 3.115,
      "theoretical_loss": 4.24496893461415,
      "tokens_seen": 246349824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673420260782347,
      "loss": 3.1308,
      "theoretical_loss": 4.244832192640725,
      "tokens_seen": 246415360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673319959879639,
      "loss": 3.1996,
      "theoretical_loss": 4.24469549720984,
      "tokens_seen": 246480896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673219658976931,
      "loss": 3.2534,
      "theoretical_loss": 4.244558848293281,
      "tokens_seen": 246546432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046731193580742226,
      "loss": 2.9928,
      "theoretical_loss": 4.244422245862863,
      "tokens_seen": 246611968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673019057171515,
      "loss": 3.2244,
      "theoretical_loss": 4.244285689890425,
      "tokens_seen": 246677504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672918756268806,
      "loss": 3.103,
      "theoretical_loss": 4.244149180347831,
      "tokens_seen": 246743040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046728184553660986,
      "loss": 3.0084,
      "theoretical_loss": 4.244012717206969,
      "tokens_seen": 246808576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467271815446339,
      "loss": 3.0494,
      "theoretical_loss": 4.24387630043975,
      "tokens_seen": 246874112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672617853560682,
      "loss": 3.2686,
      "theoretical_loss": 4.243739930018112,
      "tokens_seen": 246939648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672517552657974,
      "loss": 3.0173,
      "theoretical_loss": 4.243603605914016,
      "tokens_seen": 247005184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672417251755266,
      "loss": 3.2096,
      "theoretical_loss": 4.243467328099448,
      "tokens_seen": 247070720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046723169508525576,
      "loss": 3.3022,
      "theoretical_loss": 4.2433310965464175,
      "tokens_seen": 247136256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467221664994985,
      "loss": 3.2016,
      "theoretical_loss": 4.24319491122696,
      "tokens_seen": 247201792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672116349047141,
      "loss": 3.0706,
      "theoretical_loss": 4.243058772113131,
      "tokens_seen": 247267328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046720160481444336,
      "loss": 3.0435,
      "theoretical_loss": 4.242922679177017,
      "tokens_seen": 247332864
    },
    {
      "epoch": 0.07,
      "objective/train/docs_used": 147484,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0415401458740234,
      "objective/train/theoretical_loss": 4.2427866323907235,
      "objective/train/tokens_used": 267858400,
      "theoretical_loss": 4.2427866323907235,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671915747241725,
      "loss": 3.1482,
      "theoretical_loss": 4.2427866323907235,
      "tokens_seen": 247398400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671815446339017,
      "loss": 3.0442,
      "theoretical_loss": 4.242650631726383,
      "tokens_seen": 247463936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671715145436309,
      "loss": 3.1308,
      "theoretical_loss": 4.24251467715615,
      "tokens_seen": 247529472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671614844533601,
      "loss": 3.0189,
      "theoretical_loss": 4.2423787686522045,
      "tokens_seen": 247595008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046715145436308927,
      "loss": 3.2769,
      "theoretical_loss": 4.2422429061867515,
      "tokens_seen": 247660544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046714142427281845,
      "loss": 3.1076,
      "theoretical_loss": 4.242107089732018,
      "tokens_seen": 247726080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046713139418254763,
      "loss": 2.9247,
      "theoretical_loss": 4.241971319260257,
      "tokens_seen": 247791616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046712136409227686,
      "loss": 3.0152,
      "theoretical_loss": 4.241835594743743,
      "tokens_seen": 247857152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467111334002006,
      "loss": 3.1742,
      "theoretical_loss": 4.241699916154779,
      "tokens_seen": 247922688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671013039117352,
      "loss": 2.9643,
      "theoretical_loss": 4.241564283465689,
      "tokens_seen": 247988224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046709127382146435,
      "loss": 3.1792,
      "theoretical_loss": 4.241428696648818,
      "tokens_seen": 248053760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670812437311936,
      "loss": 2.9589,
      "theoretical_loss": 4.241293155676542,
      "tokens_seen": 248119296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046707121364092277,
      "loss": 2.9542,
      "theoretical_loss": 4.241157660521255,
      "tokens_seen": 248184832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046706118355065195,
      "loss": 3.0795,
      "theoretical_loss": 4.241022211155379,
      "tokens_seen": 248250368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046705115346038113,
      "loss": 2.8946,
      "theoretical_loss": 4.240886807551356,
      "tokens_seen": 248315904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046704112337011037,
      "loss": 3.1148,
      "theoretical_loss": 4.240751449681655,
      "tokens_seen": 248381440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670310932798395,
      "loss": 3.2406,
      "theoretical_loss": 4.240616137518767,
      "tokens_seen": 248446976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046702106318956873,
      "loss": 3.0567,
      "theoretical_loss": 4.240480871035208,
      "tokens_seen": 248512512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046701103309929786,
      "loss": 3.0778,
      "theoretical_loss": 4.240345650203517,
      "tokens_seen": 248578048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670010030090271,
      "loss": 3.1317,
      "theoretical_loss": 4.240210474996257,
      "tokens_seen": 248643584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669909729187563,
      "loss": 2.8978,
      "theoretical_loss": 4.240075345386014,
      "tokens_seen": 248709120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046698094282848545,
      "loss": 3.1967,
      "theoretical_loss": 4.239940261345399,
      "tokens_seen": 248774656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046697091273821464,
      "loss": 3.2585,
      "theoretical_loss": 4.239805222847047,
      "tokens_seen": 248840192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669608826479438,
      "loss": 3.2483,
      "theoretical_loss": 4.239670229863615,
      "tokens_seen": 248905728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046695085255767305,
      "loss": 3.1135,
      "theoretical_loss": 4.2395352823677825,
      "tokens_seen": 248971264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 148704,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5283594131469727,
      "objective/train/theoretical_loss": 4.239400380332256,
      "objective/train/tokens_used": 269496800,
      "theoretical_loss": 4.239400380332256,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694082246740223,
      "loss": 2.9646,
      "theoretical_loss": 4.239400380332256,
      "tokens_seen": 249036800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669307923771314,
      "loss": 3.1003,
      "theoretical_loss": 4.239265523729765,
      "tokens_seen": 249102336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669207622868606,
      "loss": 3.0558,
      "theoretical_loss": 4.239130712533059,
      "tokens_seen": 249167872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046691073219658983,
      "loss": 3.2011,
      "theoretical_loss": 4.2389959467149145,
      "tokens_seen": 249233408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046690070210631896,
      "loss": 3.1159,
      "theoretical_loss": 4.238861226248131,
      "tokens_seen": 249298944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668906720160482,
      "loss": 2.9774,
      "theoretical_loss": 4.238726551105531,
      "tokens_seen": 249364480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668806419257773,
      "loss": 2.8515,
      "theoretical_loss": 4.2385919212599585,
      "tokens_seen": 249430016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046687061183550656,
      "loss": 3.0703,
      "theoretical_loss": 4.238457336684284,
      "tokens_seen": 249495552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046686058174523574,
      "loss": 3.0331,
      "theoretical_loss": 4.2383227973514,
      "tokens_seen": 249561088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668505516549649,
      "loss": 2.9647,
      "theoretical_loss": 4.238188303234223,
      "tokens_seen": 249626624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668405215646941,
      "loss": 3.195,
      "theoretical_loss": 4.23805385430569,
      "tokens_seen": 249692160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668304914744233,
      "loss": 3.1549,
      "theoretical_loss": 4.237919450538767,
      "tokens_seen": 249757696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682046138415246,
      "loss": 3.257,
      "theoretical_loss": 4.237785091906437,
      "tokens_seen": 249823232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668104312938817,
      "loss": 3.1791,
      "theoretical_loss": 4.23765077838171,
      "tokens_seen": 249888768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668004012036108,
      "loss": 3.0966,
      "theoretical_loss": 4.237516509937617,
      "tokens_seen": 249954304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046679037111334006,
      "loss": 3.0634,
      "theoretical_loss": 4.2373822865472155,
      "tokens_seen": 250019840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667803410230692,
      "loss": 3.2033,
      "theoretical_loss": 4.237248108183583,
      "tokens_seen": 250085376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667703109327984,
      "loss": 3.0081,
      "theoretical_loss": 4.237113974819821,
      "tokens_seen": 250150912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667602808425276,
      "loss": 2.9963,
      "theoretical_loss": 4.236979886429055,
      "tokens_seen": 250216448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667502507522568,
      "loss": 3.0911,
      "theoretical_loss": 4.236845842984433,
      "tokens_seen": 250281984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046674022066198596,
      "loss": 3.1609,
      "theoretical_loss": 4.236711844459125,
      "tokens_seen": 250347520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667301905717152,
      "loss": 2.8868,
      "theoretical_loss": 4.236577890826326,
      "tokens_seen": 250413056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667201604814443,
      "loss": 2.8415,
      "theoretical_loss": 4.236443982059253,
      "tokens_seen": 250478592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046671013039117356,
      "loss": 3.1447,
      "theoretical_loss": 4.236310118131144,
      "tokens_seen": 250544128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667001003009027,
      "loss": 2.6377,
      "theoretical_loss": 4.236176299015265,
      "tokens_seen": 250609664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 149300,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2010979652404785,
      "objective/train/theoretical_loss": 4.2360425246849,
      "objective/train/tokens_used": 271135200,
      "theoretical_loss": 4.2360425246849,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666900702106319,
      "loss": 3.063,
      "theoretical_loss": 4.2360425246849,
      "tokens_seen": 250675200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666800401203611,
      "loss": 3.1872,
      "theoretical_loss": 4.2359087951133585,
      "tokens_seen": 250740736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666700100300903,
      "loss": 3.1641,
      "theoretical_loss": 4.235775110273972,
      "tokens_seen": 250806272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046665997993981947,
      "loss": 3.2843,
      "theoretical_loss": 4.235641470140094,
      "tokens_seen": 250871808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046664994984954865,
      "loss": 3.0174,
      "theoretical_loss": 4.235507874685103,
      "tokens_seen": 250937344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046663991975927783,
      "loss": 2.8743,
      "theoretical_loss": 4.235374323882399,
      "tokens_seen": 251002880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046662988966900706,
      "loss": 3.1391,
      "theoretical_loss": 4.235240817705406,
      "tokens_seen": 251068416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666198595787362,
      "loss": 3.0944,
      "theoretical_loss": 4.235107356127567,
      "tokens_seen": 251133952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046660982948846543,
      "loss": 3.2513,
      "theoretical_loss": 4.234973939122352,
      "tokens_seen": 251199488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046659979939819455,
      "loss": 2.703,
      "theoretical_loss": 4.2348405666632525,
      "tokens_seen": 251265024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665897693079238,
      "loss": 2.8943,
      "theoretical_loss": 4.234707238723782,
      "tokens_seen": 251330560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046657973921765297,
      "loss": 2.98,
      "theoretical_loss": 4.234573955277476,
      "tokens_seen": 251396096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046656970912738215,
      "loss": 3.1948,
      "theoretical_loss": 4.234440716297895,
      "tokens_seen": 251461632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046655967903711133,
      "loss": 3.2021,
      "theoretical_loss": 4.234307521758619,
      "tokens_seen": 251527168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046654964894684057,
      "loss": 3.0329,
      "theoretical_loss": 4.234174371633253,
      "tokens_seen": 251592704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665396188565697,
      "loss": 3.0375,
      "theoretical_loss": 4.234041265895425,
      "tokens_seen": 251658240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046652958876629893,
      "loss": 2.9757,
      "theoretical_loss": 4.233908204518783,
      "tokens_seen": 251723776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046651955867602806,
      "loss": 2.956,
      "theoretical_loss": 4.233775187476998,
      "tokens_seen": 251789312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665095285857573,
      "loss": 3.0873,
      "theoretical_loss": 4.233642214743766,
      "tokens_seen": 251854848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664994984954865,
      "loss": 3.1756,
      "theoretical_loss": 4.2335092862928025,
      "tokens_seen": 251920384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046648946840521565,
      "loss": 3.047,
      "theoretical_loss": 4.233376402097846,
      "tokens_seen": 251985920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046647943831494484,
      "loss": 2.7,
      "theoretical_loss": 4.233243562132661,
      "tokens_seen": 252051456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466469408224674,
      "loss": 2.8238,
      "theoretical_loss": 4.233110766371029,
      "tokens_seen": 252116992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664593781344032,
      "loss": 3.0103,
      "theoretical_loss": 4.232978014786756,
      "tokens_seen": 252182528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046644934804413243,
      "loss": 2.9662,
      "theoretical_loss": 4.2328453073536725,
      "tokens_seen": 252248064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 150610,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.978238582611084,
      "objective/train/theoretical_loss": 4.232712644045627,
      "objective/train/tokens_used": 272773600,
      "theoretical_loss": 4.232712644045627,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046643931795386156,
      "loss": 3.1531,
      "theoretical_loss": 4.232712644045627,
      "tokens_seen": 252313600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664292878635908,
      "loss": 3.2135,
      "theoretical_loss": 4.232580024836494,
      "tokens_seen": 252379136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664192577733199,
      "loss": 3.0233,
      "theoretical_loss": 4.232447449700169,
      "tokens_seen": 252444672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046640922768304916,
      "loss": 2.8671,
      "theoretical_loss": 4.23231491861057,
      "tokens_seen": 252510208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046639919759277834,
      "loss": 3.0974,
      "theoretical_loss": 4.232182431541636,
      "tokens_seen": 252575744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663891675025075,
      "loss": 3.0586,
      "theoretical_loss": 4.232049988467328,
      "tokens_seen": 252641280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663791374122367,
      "loss": 2.9412,
      "theoretical_loss": 4.2319175893616325,
      "tokens_seen": 252706816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046636910732196594,
      "loss": 2.9511,
      "theoretical_loss": 4.231785234198555,
      "tokens_seen": 252772352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046635907723169506,
      "loss": 2.9213,
      "theoretical_loss": 4.231652922952122,
      "tokens_seen": 252837888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663490471414243,
      "loss": 2.9723,
      "theoretical_loss": 4.231520655596388,
      "tokens_seen": 252903424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663390170511534,
      "loss": 3.0252,
      "theoretical_loss": 4.231388432105421,
      "tokens_seen": 252968960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046632898696088266,
      "loss": 3.1199,
      "theoretical_loss": 4.231256252453317,
      "tokens_seen": 253034496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046631895687061184,
      "loss": 3.194,
      "theoretical_loss": 4.231124116614194,
      "tokens_seen": 253100032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466308926780341,
      "loss": 3.1208,
      "theoretical_loss": 4.230992024562189,
      "tokens_seen": 253165568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662988966900702,
      "loss": 2.9426,
      "theoretical_loss": 4.230859976271464,
      "tokens_seen": 253231104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662888665997994,
      "loss": 3.0458,
      "theoretical_loss": 4.2307279717162,
      "tokens_seen": 253296640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046627883650952857,
      "loss": 3.193,
      "theoretical_loss": 4.2305960108706016,
      "tokens_seen": 253362176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662688064192578,
      "loss": 3.1534,
      "theoretical_loss": 4.230464093708894,
      "tokens_seen": 253427712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046625877632898693,
      "loss": 3.1943,
      "theoretical_loss": 4.230332220205328,
      "tokens_seen": 253493248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046624874623871616,
      "loss": 3.0607,
      "theoretical_loss": 4.230200390334172,
      "tokens_seen": 253558784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046623871614844535,
      "loss": 2.8528,
      "theoretical_loss": 4.230068604069717,
      "tokens_seen": 253624320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662286860581745,
      "loss": 2.9088,
      "theoretical_loss": 4.2299368613862764,
      "tokens_seen": 253689856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662186559679037,
      "loss": 3.2573,
      "theoretical_loss": 4.229805162258187,
      "tokens_seen": 253755392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662086258776329,
      "loss": 2.9836,
      "theoretical_loss": 4.229673506659805,
      "tokens_seen": 253820928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661985957873621,
      "loss": 3.1131,
      "theoretical_loss": 4.229541894565508,
      "tokens_seen": 253886464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 151678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8198091983795166,
      "objective/train/theoretical_loss": 4.229410325949698,
      "objective/train/tokens_used": 274412000,
      "theoretical_loss": 4.229410325949698,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661885656970913,
      "loss": 2.9257,
      "theoretical_loss": 4.229410325949698,
      "tokens_seen": 253952000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661785356068205,
      "loss": 3.1713,
      "theoretical_loss": 4.229278800786796,
      "tokens_seen": 254017536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046616850551654967,
      "loss": 2.9401,
      "theoretical_loss": 4.229147319051247,
      "tokens_seen": 254083072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046615847542627885,
      "loss": 2.8864,
      "theoretical_loss": 4.2290158807175136,
      "tokens_seen": 254148608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046614844533600803,
      "loss": 3.1424,
      "theoretical_loss": 4.228884485760086,
      "tokens_seen": 254214144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046613841524573727,
      "loss": 3.2015,
      "theoretical_loss": 4.22875313415347,
      "tokens_seen": 254279680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661283851554664,
      "loss": 3.1367,
      "theoretical_loss": 4.228621825872198,
      "tokens_seen": 254345216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046611835506519563,
      "loss": 2.9269,
      "theoretical_loss": 4.228490560890819,
      "tokens_seen": 254410752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046610832497492475,
      "loss": 3.1191,
      "theoretical_loss": 4.228359339183908,
      "tokens_seen": 254476288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466098294884654,
      "loss": 3.2313,
      "theoretical_loss": 4.228228160726059,
      "tokens_seen": 254541824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046608826479438317,
      "loss": 3.2163,
      "theoretical_loss": 4.228097025491887,
      "tokens_seen": 254607360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046607823470411235,
      "loss": 3.0566,
      "theoretical_loss": 4.22796593345603,
      "tokens_seen": 254672896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046606820461384153,
      "loss": 2.9742,
      "theoretical_loss": 4.227834884593148,
      "tokens_seen": 254738432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046605817452357077,
      "loss": 2.6936,
      "theoretical_loss": 4.22770387887792,
      "tokens_seen": 254803968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660481444332999,
      "loss": 2.989,
      "theoretical_loss": 4.227572916285046,
      "tokens_seen": 254869504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046603811434302913,
      "loss": 3.1684,
      "theoretical_loss": 4.227441996789253,
      "tokens_seen": 254935040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046602808425275826,
      "loss": 3.0604,
      "theoretical_loss": 4.227311120365281,
      "tokens_seen": 255000576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660180541624875,
      "loss": 3.3224,
      "theoretical_loss": 4.227180286987897,
      "tokens_seen": 255066112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660080240722167,
      "loss": 3.2097,
      "theoretical_loss": 4.227049496631889,
      "tokens_seen": 255131648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046599799398194586,
      "loss": 3.1686,
      "theoretical_loss": 4.226918749272063,
      "tokens_seen": 255197184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046598796389167504,
      "loss": 3.0575,
      "theoretical_loss": 4.226788044883248,
      "tokens_seen": 255262720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659779338014042,
      "loss": 3.1225,
      "theoretical_loss": 4.226657383440297,
      "tokens_seen": 255328256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659679037111334,
      "loss": 2.9431,
      "theoretical_loss": 4.226526764918079,
      "tokens_seen": 255393792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595787362086263,
      "loss": 3.1373,
      "theoretical_loss": 4.2263961892914885,
      "tokens_seen": 255459328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046594784353059176,
      "loss": 3.2497,
      "theoretical_loss": 4.226265656535437,
      "tokens_seen": 255524864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 152377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.094141960144043,
      "objective/train/theoretical_loss": 4.226135166624862,
      "objective/train/tokens_used": 276050400,
      "theoretical_loss": 4.226135166624862,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465937813440321,
      "loss": 3.2235,
      "theoretical_loss": 4.226135166624862,
      "tokens_seen": 255590400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659277833500501,
      "loss": 3.158,
      "theoretical_loss": 4.226004719534718,
      "tokens_seen": 255655936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591775325977936,
      "loss": 2.9932,
      "theoretical_loss": 4.225874315239982,
      "tokens_seen": 255721472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046590772316950854,
      "loss": 2.9383,
      "theoretical_loss": 4.225743953715653,
      "tokens_seen": 255787008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658976930792377,
      "loss": 3.0384,
      "theoretical_loss": 4.225613634936749,
      "tokens_seen": 255852544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658876629889669,
      "loss": 3.1105,
      "theoretical_loss": 4.22548335887831,
      "tokens_seen": 255918080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046587763289869614,
      "loss": 3.0846,
      "theoretical_loss": 4.225353125515398,
      "tokens_seen": 255983616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046586760280842526,
      "loss": 3.013,
      "theoretical_loss": 4.225222934823094,
      "tokens_seen": 256049152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658575727181545,
      "loss": 3.1509,
      "theoretical_loss": 4.225092786776502,
      "tokens_seen": 256114688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658475426278836,
      "loss": 2.9779,
      "theoretical_loss": 4.224962681350746,
      "tokens_seen": 256180224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046583751253761286,
      "loss": 2.778,
      "theoretical_loss": 4.224832618520969,
      "tokens_seen": 256245760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046582748244734204,
      "loss": 3.263,
      "theoretical_loss": 4.224702598262336,
      "tokens_seen": 256311296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658174523570712,
      "loss": 2.9598,
      "theoretical_loss": 4.2245726205500365,
      "tokens_seen": 256376832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658074222668004,
      "loss": 3.1954,
      "theoretical_loss": 4.224442685359275,
      "tokens_seen": 256442368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657973921765296,
      "loss": 2.9688,
      "theoretical_loss": 4.22431279266528,
      "tokens_seen": 256507904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046578736208625877,
      "loss": 3.1653,
      "theoretical_loss": 4.2241829424433,
      "tokens_seen": 256573440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465777331995988,
      "loss": 3.0042,
      "theoretical_loss": 4.224053134668605,
      "tokens_seen": 256638976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046576730190571713,
      "loss": 2.9824,
      "theoretical_loss": 4.223923369316485,
      "tokens_seen": 256704512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575727181544636,
      "loss": 2.9607,
      "theoretical_loss": 4.2237936463622505,
      "tokens_seen": 256770048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046574724172517555,
      "loss": 3.0764,
      "theoretical_loss": 4.223663965781233,
      "tokens_seen": 256835584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657372116349047,
      "loss": 3.1061,
      "theoretical_loss": 4.2235343275487836,
      "tokens_seen": 256901120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657271815446339,
      "loss": 2.9893,
      "theoretical_loss": 4.223404731640277,
      "tokens_seen": 256966656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171514543631,
      "loss": 2.9472,
      "theoretical_loss": 4.223275178031105,
      "tokens_seen": 257032192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046570712136409227,
      "loss": 3.0962,
      "theoretical_loss": 4.223145666696682,
      "tokens_seen": 257097728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656970912738215,
      "loss": 3.2783,
      "theoretical_loss": 4.223016197612442,
      "tokens_seen": 257163264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 153004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8444950580596924,
      "objective/train/theoretical_loss": 4.222886770753841,
      "objective/train/tokens_used": 277688800,
      "theoretical_loss": 4.222886770753841,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046568706118355063,
      "loss": 2.8758,
      "theoretical_loss": 4.222886770753841,
      "tokens_seen": 257228800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046567703109327987,
      "loss": 3.2012,
      "theoretical_loss": 4.222757386096354,
      "tokens_seen": 257294336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465667001003009,
      "loss": 3.0649,
      "theoretical_loss": 4.222628043615476,
      "tokens_seen": 257359872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565697091273823,
      "loss": 3.1071,
      "theoretical_loss": 4.2224987432867245,
      "tokens_seen": 257425408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656469408224674,
      "loss": 3.1137,
      "theoretical_loss": 4.222369485085636,
      "tokens_seen": 257490944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656369107321966,
      "loss": 3.1091,
      "theoretical_loss": 4.222240268987768,
      "tokens_seen": 257556480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656268806419258,
      "loss": 2.8566,
      "theoretical_loss": 4.222111094968698,
      "tokens_seen": 257622016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046561685055165495,
      "loss": 3.0556,
      "theoretical_loss": 4.2219819630040245,
      "tokens_seen": 257687552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046560682046138414,
      "loss": 3.1304,
      "theoretical_loss": 4.2218528730693645,
      "tokens_seen": 257753088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046559679037111337,
      "loss": 2.937,
      "theoretical_loss": 4.221723825140359,
      "tokens_seen": 257818624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655867602808425,
      "loss": 2.9502,
      "theoretical_loss": 4.221594819192665,
      "tokens_seen": 257884160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046557673019057173,
      "loss": 3.1884,
      "theoretical_loss": 4.221465855201963,
      "tokens_seen": 257949696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655667001003009,
      "loss": 3.1707,
      "theoretical_loss": 4.221336933143952,
      "tokens_seen": 258015232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655566700100301,
      "loss": 2.9765,
      "theoretical_loss": 4.221208052994352,
      "tokens_seen": 258080768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655466399197593,
      "loss": 3.1372,
      "theoretical_loss": 4.221079214728903,
      "tokens_seen": 258146304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553660982948846,
      "loss": 3.0284,
      "theoretical_loss": 4.2209504183233655,
      "tokens_seen": 258211840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046552657973921764,
      "loss": 3.1995,
      "theoretical_loss": 4.22082166375352,
      "tokens_seen": 258277376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655165496489469,
      "loss": 2.8978,
      "theoretical_loss": 4.220692950995167,
      "tokens_seen": 258342912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465506519558676,
      "loss": 3.0251,
      "theoretical_loss": 4.220564280024126,
      "tokens_seen": 258408448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046549648946840524,
      "loss": 2.9952,
      "theoretical_loss": 4.22043565081624,
      "tokens_seen": 258473984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046548645937813436,
      "loss": 3.1847,
      "theoretical_loss": 4.220307063347369,
      "tokens_seen": 258539520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654764292878636,
      "loss": 2.9241,
      "theoretical_loss": 4.220178517593393,
      "tokens_seen": 258605056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654663991975928,
      "loss": 2.8651,
      "theoretical_loss": 4.220050013530215,
      "tokens_seen": 258670592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046545636910732196,
      "loss": 2.9641,
      "theoretical_loss": 4.219921551133754,
      "tokens_seen": 258736128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654463390170512,
      "loss": 3.1291,
      "theoretical_loss": 4.219793130379952,
      "tokens_seen": 258801664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 154245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.155519485473633,
      "objective/train/theoretical_loss": 4.21966475124477,
      "objective/train/tokens_used": 279327200,
      "theoretical_loss": 4.21966475124477,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654363089267803,
      "loss": 3.232,
      "theoretical_loss": 4.21966475124477,
      "tokens_seen": 258867200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046542627883650956,
      "loss": 3.1265,
      "theoretical_loss": 4.21953641370419,
      "tokens_seen": 258932736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046541624874623874,
      "loss": 2.8509,
      "theoretical_loss": 4.21940811773421,
      "tokens_seen": 258998272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654062186559679,
      "loss": 3.0692,
      "theoretical_loss": 4.219279863310855,
      "tokens_seen": 259063808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653961885656971,
      "loss": 2.9505,
      "theoretical_loss": 4.219151650410162,
      "tokens_seen": 259129344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046538615847542634,
      "loss": 3.049,
      "theoretical_loss": 4.219023479008193,
      "tokens_seen": 259194880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046537612838515546,
      "loss": 3.0422,
      "theoretical_loss": 4.21889534908103,
      "tokens_seen": 259260416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653660982948847,
      "loss": 3.2427,
      "theoretical_loss": 4.218767260604771,
      "tokens_seen": 259325952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653560682046138,
      "loss": 2.8318,
      "theoretical_loss": 4.218639213555537,
      "tokens_seen": 259391488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046534603811434306,
      "loss": 2.817,
      "theoretical_loss": 4.218511207909467,
      "tokens_seen": 259457024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533600802407224,
      "loss": 2.9972,
      "theoretical_loss": 4.218383243642723,
      "tokens_seen": 259522560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653259779338014,
      "loss": 2.8747,
      "theoretical_loss": 4.2182553207314815,
      "tokens_seen": 259588096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653159478435306,
      "loss": 3.0103,
      "theoretical_loss": 4.218127439151944,
      "tokens_seen": 259653632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653059177532598,
      "loss": 3.1678,
      "theoretical_loss": 4.217999598880327,
      "tokens_seen": 259719168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046529588766298897,
      "loss": 3.1776,
      "theoretical_loss": 4.2178717998928725,
      "tokens_seen": 259784704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652858575727182,
      "loss": 3.0578,
      "theoretical_loss": 4.217744042165835,
      "tokens_seen": 259850240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046527582748244733,
      "loss": 3.0386,
      "theoretical_loss": 4.217616325675493,
      "tokens_seen": 259915776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046526579739217656,
      "loss": 2.9572,
      "theoretical_loss": 4.2174886503981455,
      "tokens_seen": 259981312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525576730190575,
      "loss": 2.985,
      "theoretical_loss": 4.217361016310109,
      "tokens_seen": 260046848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652457372116349,
      "loss": 3.0667,
      "theoretical_loss": 4.217233423387719,
      "tokens_seen": 260112384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652357071213641,
      "loss": 3.0605,
      "theoretical_loss": 4.217105871607333,
      "tokens_seen": 260177920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652256770310933,
      "loss": 3.354,
      "theoretical_loss": 4.216978360945325,
      "tokens_seen": 260243456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046521564694082247,
      "loss": 2.9401,
      "theoretical_loss": 4.216850891378092,
      "tokens_seen": 260308992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652056168505517,
      "loss": 3.0811,
      "theoretical_loss": 4.216723462882047,
      "tokens_seen": 260374528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519558676028083,
      "loss": 3.1081,
      "theoretical_loss": 4.2165960754336265,
      "tokens_seen": 260440064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 155441,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7761213779449463,
      "objective/train/theoretical_loss": 4.216468729009281,
      "objective/train/tokens_used": 280965600,
      "theoretical_loss": 4.216468729009281,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046518555667001007,
      "loss": 2.6967,
      "theoretical_loss": 4.216468729009281,
      "tokens_seen": 260505600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651755265797392,
      "loss": 3.038,
      "theoretical_loss": 4.216341423585486,
      "tokens_seen": 260571136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046516549648946843,
      "loss": 3.3396,
      "theoretical_loss": 4.216214159138732,
      "tokens_seen": 260636672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651554663991976,
      "loss": 3.0875,
      "theoretical_loss": 4.216086935645532,
      "tokens_seen": 260702208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651454363089268,
      "loss": 3.1269,
      "theoretical_loss": 4.2159597530824175,
      "tokens_seen": 260767744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465135406218656,
      "loss": 2.9654,
      "theoretical_loss": 4.215832611425937,
      "tokens_seen": 260833280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046512537612838515,
      "loss": 2.9985,
      "theoretical_loss": 4.215705510652662,
      "tokens_seen": 260898816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046511534603811434,
      "loss": 3.0659,
      "theoretical_loss": 4.215578450739182,
      "tokens_seen": 260964352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046510531594784357,
      "loss": 2.859,
      "theoretical_loss": 4.215451431662104,
      "tokens_seen": 261029888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650952858575727,
      "loss": 3.0105,
      "theoretical_loss": 4.215324453398057,
      "tokens_seen": 261095424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046508525576730193,
      "loss": 3.0586,
      "theoretical_loss": 4.215197515923687,
      "tokens_seen": 261160960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650752256770311,
      "loss": 3.0384,
      "theoretical_loss": 4.21507061921566,
      "tokens_seen": 261226496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650651955867603,
      "loss": 2.9461,
      "theoretical_loss": 4.214943763250662,
      "tokens_seen": 261292032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650551654964895,
      "loss": 3.0812,
      "theoretical_loss": 4.214816948005396,
      "tokens_seen": 261357568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046504513540621866,
      "loss": 2.9739,
      "theoretical_loss": 4.214690173456588,
      "tokens_seen": 261423104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046503510531594784,
      "loss": 3.2378,
      "theoretical_loss": 4.214563439580979,
      "tokens_seen": 261488640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650250752256771,
      "loss": 2.9873,
      "theoretical_loss": 4.214436746355331,
      "tokens_seen": 261554176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650150451354062,
      "loss": 3.0108,
      "theoretical_loss": 4.214310093756427,
      "tokens_seen": 261619712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046500501504513544,
      "loss": 2.9422,
      "theoretical_loss": 4.214183481761064,
      "tokens_seen": 261685248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046499498495486456,
      "loss": 3.1189,
      "theoretical_loss": 4.214056910346063,
      "tokens_seen": 261750784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649849548645938,
      "loss": 3.1657,
      "theoretical_loss": 4.2139303794882625,
      "tokens_seen": 261816320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464974924774323,
      "loss": 2.7213,
      "theoretical_loss": 4.213803889164518,
      "tokens_seen": 261881856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046496489468405216,
      "loss": 2.8726,
      "theoretical_loss": 4.213677439351708,
      "tokens_seen": 261947392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046495486459378134,
      "loss": 2.7336,
      "theoretical_loss": 4.213551030026725,
      "tokens_seen": 262012928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649448345035105,
      "loss": 2.9864,
      "theoretical_loss": 4.213424661166485,
      "tokens_seen": 262078464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 156138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.797309398651123,
      "objective/train/theoretical_loss": 4.21329833274792,
      "objective/train/tokens_used": 282604000,
      "theoretical_loss": 4.21329833274792,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649348044132397,
      "loss": 2.5665,
      "theoretical_loss": 4.21329833274792,
      "tokens_seen": 262144000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046492477432296894,
      "loss": 3.0384,
      "theoretical_loss": 4.213172044747982,
      "tokens_seen": 262209536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046491474423269807,
      "loss": 2.8508,
      "theoretical_loss": 4.213045797143643,
      "tokens_seen": 262275072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649047141424273,
      "loss": 2.933,
      "theoretical_loss": 4.212919589911891,
      "tokens_seen": 262340608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648946840521565,
      "loss": 2.8909,
      "theoretical_loss": 4.212793423029735,
      "tokens_seen": 262406144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046488465396188566,
      "loss": 3.0675,
      "theoretical_loss": 4.212667296474204,
      "tokens_seen": 262471680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046487462387161485,
      "loss": 3.0812,
      "theoretical_loss": 4.2125412102223425,
      "tokens_seen": 262537216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464864593781344,
      "loss": 3.2516,
      "theoretical_loss": 4.212415164251215,
      "tokens_seen": 262602752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648545636910732,
      "loss": 2.9727,
      "theoretical_loss": 4.212289158537908,
      "tokens_seen": 262668288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046484453360080244,
      "loss": 3.2213,
      "theoretical_loss": 4.212163193059521,
      "tokens_seen": 262733824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046483450351053157,
      "loss": 3.0169,
      "theoretical_loss": 4.2120372677931766,
      "tokens_seen": 262799360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648244734202608,
      "loss": 3.2831,
      "theoretical_loss": 4.211911382716014,
      "tokens_seen": 262864896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046481444332998993,
      "loss": 2.9297,
      "theoretical_loss": 4.211785537805193,
      "tokens_seen": 262930432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046480441323971917,
      "loss": 3.0726,
      "theoretical_loss": 4.211659733037891,
      "tokens_seen": 262995968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046479438314944835,
      "loss": 3.0694,
      "theoretical_loss": 4.2115339683913025,
      "tokens_seen": 263061504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046478435305917753,
      "loss": 2.8509,
      "theoretical_loss": 4.211408243842643,
      "tokens_seen": 263127040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647743229689067,
      "loss": 3.0094,
      "theoretical_loss": 4.2112825593691445,
      "tokens_seen": 263192576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046476429287863595,
      "loss": 3.1927,
      "theoretical_loss": 4.21115691494806,
      "tokens_seen": 263258112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647542627883651,
      "loss": 3.0788,
      "theoretical_loss": 4.211031310556661,
      "tokens_seen": 263323648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647442326980943,
      "loss": 2.9797,
      "theoretical_loss": 4.210905746172235,
      "tokens_seen": 263389184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046473420260782344,
      "loss": 2.7703,
      "theoretical_loss": 4.210780221772088,
      "tokens_seen": 263454720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046472417251755267,
      "loss": 2.9913,
      "theoretical_loss": 4.210654737333549,
      "tokens_seen": 263520256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647141424272819,
      "loss": 3.0574,
      "theoretical_loss": 4.21052929283396,
      "tokens_seen": 263585792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046470411233701103,
      "loss": 2.9517,
      "theoretical_loss": 4.210403888250685,
      "tokens_seen": 263651328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046469408224674027,
      "loss": 2.9469,
      "theoretical_loss": 4.210278523561106,
      "tokens_seen": 263716864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 157246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.921288251876831,
      "objective/train/theoretical_loss": 4.2101531987426215,
      "objective/train/tokens_used": 284242400,
      "theoretical_loss": 4.2101531987426215,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646840521564694,
      "loss": 3.1574,
      "theoretical_loss": 4.2101531987426215,
      "tokens_seen": 263782400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046467402206619863,
      "loss": 2.9865,
      "theoretical_loss": 4.210027913772651,
      "tokens_seen": 263847936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646639919759278,
      "loss": 2.9807,
      "theoretical_loss": 4.209902668628629,
      "tokens_seen": 263913472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464653961885657,
      "loss": 3.0513,
      "theoretical_loss": 4.209777463288012,
      "tokens_seen": 263979008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646439317953862,
      "loss": 2.8659,
      "theoretical_loss": 4.209652297728274,
      "tokens_seen": 264044544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046463390170511535,
      "loss": 3.1063,
      "theoretical_loss": 4.209527171926906,
      "tokens_seen": 264110080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046462387161484454,
      "loss": 2.9362,
      "theoretical_loss": 4.209402085861418,
      "tokens_seen": 264175616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046461384152457377,
      "loss": 3.1464,
      "theoretical_loss": 4.209277039509338,
      "tokens_seen": 264241152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646038114343029,
      "loss": 3.0352,
      "theoretical_loss": 4.209152032848213,
      "tokens_seen": 264306688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046459378134403213,
      "loss": 2.7659,
      "theoretical_loss": 4.209027065855607,
      "tokens_seen": 264372224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645837512537613,
      "loss": 2.8415,
      "theoretical_loss": 4.208902138509105,
      "tokens_seen": 264437760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645737211634905,
      "loss": 3.2915,
      "theoretical_loss": 4.208777250786305,
      "tokens_seen": 264503296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645636910732197,
      "loss": 3.0093,
      "theoretical_loss": 4.20865240266483,
      "tokens_seen": 264568832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046455366098294886,
      "loss": 3.0984,
      "theoretical_loss": 4.208527594122316,
      "tokens_seen": 264634368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046454363089267804,
      "loss": 3.171,
      "theoretical_loss": 4.208402825136419,
      "tokens_seen": 264699904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645336008024073,
      "loss": 3.2183,
      "theoretical_loss": 4.208278095684813,
      "tokens_seen": 264765440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645235707121364,
      "loss": 3.1366,
      "theoretical_loss": 4.20815340574519,
      "tokens_seen": 264830976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046451354062186564,
      "loss": 2.9705,
      "theoretical_loss": 4.20802875529526,
      "tokens_seen": 264896512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046450351053159476,
      "loss": 3.2497,
      "theoretical_loss": 4.207904144312752,
      "tokens_seen": 264962048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464493480441324,
      "loss": 3.1657,
      "theoretical_loss": 4.2077795727754115,
      "tokens_seen": 265027584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644834503510532,
      "loss": 2.869,
      "theoretical_loss": 4.207655040661004,
      "tokens_seen": 265093120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046447342026078236,
      "loss": 2.8442,
      "theoretical_loss": 4.207530547947311,
      "tokens_seen": 265158656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046446339017051154,
      "loss": 3.1521,
      "theoretical_loss": 4.207406094612133,
      "tokens_seen": 265224192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644533600802407,
      "loss": 2.845,
      "theoretical_loss": 4.207281680633289,
      "tokens_seen": 265289728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644433299899699,
      "loss": 2.7843,
      "theoretical_loss": 4.207157305988615,
      "tokens_seen": 265355264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 157714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.635972738265991,
      "objective/train/theoretical_loss": 4.207032970655965,
      "objective/train/tokens_used": 285880800,
      "theoretical_loss": 4.207032970655965,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046443329989969914,
      "loss": 2.9511,
      "theoretical_loss": 4.207032970655965,
      "tokens_seen": 265420800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046442326980942827,
      "loss": 3.0488,
      "theoretical_loss": 4.206908674613213,
      "tokens_seen": 265486336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644132397191575,
      "loss": 2.7703,
      "theoretical_loss": 4.2067844178382465,
      "tokens_seen": 265551872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644032096288867,
      "loss": 2.9723,
      "theoretical_loss": 4.206660200308975,
      "tokens_seen": 265617408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046439317953861586,
      "loss": 2.896,
      "theoretical_loss": 4.206536022003325,
      "tokens_seen": 265682944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046438314944834505,
      "loss": 3.1176,
      "theoretical_loss": 4.206411882899239,
      "tokens_seen": 265748480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643731193580742,
      "loss": 3.054,
      "theoretical_loss": 4.20628778297468,
      "tokens_seen": 265814016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643630892678034,
      "loss": 3.0583,
      "theoretical_loss": 4.206163722207627,
      "tokens_seen": 265879552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046435305917753264,
      "loss": 2.9522,
      "theoretical_loss": 4.206039700576078,
      "tokens_seen": 265945088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046434302908726177,
      "loss": 3.1155,
      "theoretical_loss": 4.205915718058046,
      "tokens_seen": 266010624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464332998996991,
      "loss": 3.1129,
      "theoretical_loss": 4.205791774631566,
      "tokens_seen": 266076160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046432296890672013,
      "loss": 3.1271,
      "theoretical_loss": 4.205667870274686,
      "tokens_seen": 266141696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046431293881644937,
      "loss": 3.0881,
      "theoretical_loss": 4.2055440049654775,
      "tokens_seen": 266207232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046430290872617855,
      "loss": 3.1299,
      "theoretical_loss": 4.205420178682024,
      "tokens_seen": 266272768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046429287863590773,
      "loss": 2.8421,
      "theoretical_loss": 4.205296391402431,
      "tokens_seen": 266338304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642828485456369,
      "loss": 2.9059,
      "theoretical_loss": 4.205172643104818,
      "tokens_seen": 266403840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046427281845536615,
      "loss": 2.8721,
      "theoretical_loss": 4.205048933767326,
      "tokens_seen": 266469376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642627883650953,
      "loss": 3.1183,
      "theoretical_loss": 4.20492526336811,
      "tokens_seen": 266534912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642527582748245,
      "loss": 2.9405,
      "theoretical_loss": 4.204801631885346,
      "tokens_seen": 266600448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046424272818455364,
      "loss": 2.9166,
      "theoretical_loss": 4.204678039297223,
      "tokens_seen": 266665984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046423269809428287,
      "loss": 3.1439,
      "theoretical_loss": 4.204554485581953,
      "tokens_seen": 266731520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046422266800401205,
      "loss": 3.0142,
      "theoretical_loss": 4.2044309707177625,
      "tokens_seen": 266797056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046421263791374123,
      "loss": 3.166,
      "theoretical_loss": 4.204307494682895,
      "tokens_seen": 266862592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642026078234704,
      "loss": 2.91,
      "theoretical_loss": 4.204184057455613,
      "tokens_seen": 266928128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641925777331996,
      "loss": 2.7895,
      "theoretical_loss": 4.204060659014197,
      "tokens_seen": 266993664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 158889,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5656661987304688,
      "objective/train/theoretical_loss": 4.2039372993369435,
      "objective/train/tokens_used": 287519200,
      "theoretical_loss": 4.2039372993369435,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641825476429288,
      "loss": 2.9939,
      "theoretical_loss": 4.2039372993369435,
      "tokens_seen": 267059200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464172517552658,
      "loss": 2.7803,
      "theoretical_loss": 4.203813978402167,
      "tokens_seen": 267124736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046416248746238714,
      "loss": 3.129,
      "theoretical_loss": 4.203690696188198,
      "tokens_seen": 267190272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641524573721164,
      "loss": 3.0649,
      "theoretical_loss": 4.203567452673388,
      "tokens_seen": 267255808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641424272818455,
      "loss": 3.0718,
      "theoretical_loss": 4.203444247836103,
      "tokens_seen": 267321344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046413239719157474,
      "loss": 3.2325,
      "theoretical_loss": 4.203321081654726,
      "tokens_seen": 267386880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641223671013039,
      "loss": 3.0568,
      "theoretical_loss": 4.203197954107661,
      "tokens_seen": 267452416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641123370110331,
      "loss": 2.8217,
      "theoretical_loss": 4.203074865173325,
      "tokens_seen": 267517952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641023069207623,
      "loss": 2.9454,
      "theoretical_loss": 4.202951814830155,
      "tokens_seen": 267583488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640922768304915,
      "loss": 2.9256,
      "theoretical_loss": 4.202828803056605,
      "tokens_seen": 267649024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046408224674022064,
      "loss": 2.9524,
      "theoretical_loss": 4.202705829831144,
      "tokens_seen": 267714560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640722166499499,
      "loss": 2.9487,
      "theoretical_loss": 4.202582895132262,
      "tokens_seen": 267780096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464062186559679,
      "loss": 2.6154,
      "theoretical_loss": 4.202459998938464,
      "tokens_seen": 267845632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046405215646940824,
      "loss": 3.114,
      "theoretical_loss": 4.202337141228272,
      "tokens_seen": 267911168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640421263791374,
      "loss": 3.0631,
      "theoretical_loss": 4.202214321980227,
      "tokens_seen": 267976704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640320962888666,
      "loss": 2.9271,
      "theoretical_loss": 4.202091541172885,
      "tokens_seen": 268042240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640220661985958,
      "loss": 2.9358,
      "theoretical_loss": 4.201968798784821,
      "tokens_seen": 268107776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046401203610832496,
      "loss": 2.9304,
      "theoretical_loss": 4.201846094794627,
      "tokens_seen": 268173312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046400200601805414,
      "loss": 3.2132,
      "theoretical_loss": 4.201723429180911,
      "tokens_seen": 268238848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639919759277834,
      "loss": 2.8554,
      "theoretical_loss": 4.2016008019222975,
      "tokens_seen": 268304384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639819458375125,
      "loss": 3.0379,
      "theoretical_loss": 4.201478212997431,
      "tokens_seen": 268369920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046397191574724174,
      "loss": 2.8709,
      "theoretical_loss": 4.201355662384971,
      "tokens_seen": 268435456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639618856569709,
      "loss": 3.0275,
      "theoretical_loss": 4.2012331500635955,
      "tokens_seen": 268500992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639518555667001,
      "loss": 3.1757,
      "theoretical_loss": 4.201110676011998,
      "tokens_seen": 268566528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046394182547642934,
      "loss": 3.3162,
      "theoretical_loss": 4.200988240208889,
      "tokens_seen": 268632064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 159274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.012082099914551,
      "objective/train/theoretical_loss": 4.2008658426329974,
      "objective/train/tokens_used": 289157600,
      "theoretical_loss": 4.2008658426329974,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393179538615847,
      "loss": 2.954,
      "theoretical_loss": 4.2008658426329974,
      "tokens_seen": 268697600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639217652958877,
      "loss": 2.9394,
      "theoretical_loss": 4.200743483263068,
      "tokens_seen": 268763136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639117352056169,
      "loss": 2.9931,
      "theoretical_loss": 4.200621162077864,
      "tokens_seen": 268828672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046390170511534606,
      "loss": 3.1958,
      "theoretical_loss": 4.200498879056164,
      "tokens_seen": 268894208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389167502507525,
      "loss": 3.0091,
      "theoretical_loss": 4.200376634176765,
      "tokens_seen": 268959744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638816449348044,
      "loss": 2.9129,
      "theoretical_loss": 4.200254427418478,
      "tokens_seen": 269025280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638716148445336,
      "loss": 3.0744,
      "theoretical_loss": 4.200132258760135,
      "tokens_seen": 269090816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046386158475426284,
      "loss": 3.0598,
      "theoretical_loss": 4.200010128180582,
      "tokens_seen": 269156352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385155466399197,
      "loss": 3.3544,
      "theoretical_loss": 4.1998880356586845,
      "tokens_seen": 269221888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638415245737212,
      "loss": 3.2152,
      "theoretical_loss": 4.1997659811733214,
      "tokens_seen": 269287424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046383149448345033,
      "loss": 3.3844,
      "theoretical_loss": 4.199643964703391,
      "tokens_seen": 269352960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046382146439317957,
      "loss": 3.2429,
      "theoretical_loss": 4.199521986227808,
      "tokens_seen": 269418496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381143430290875,
      "loss": 3.002,
      "theoretical_loss": 4.199400045725502,
      "tokens_seen": 269484032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046380140421263793,
      "loss": 2.9083,
      "theoretical_loss": 4.199278143175423,
      "tokens_seen": 269549568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637913741223671,
      "loss": 3.2151,
      "theoretical_loss": 4.1991562785565355,
      "tokens_seen": 269615104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046378134403209635,
      "loss": 2.7854,
      "theoretical_loss": 4.199034451847821,
      "tokens_seen": 269680640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637713139418255,
      "loss": 3.1079,
      "theoretical_loss": 4.198912663028278,
      "tokens_seen": 269746176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637612838515547,
      "loss": 3.2046,
      "theoretical_loss": 4.19879091207692,
      "tokens_seen": 269811712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375125376128384,
      "loss": 2.8742,
      "theoretical_loss": 4.19866919897278,
      "tokens_seen": 269877248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046374122367101307,
      "loss": 2.8417,
      "theoretical_loss": 4.1985475236949075,
      "tokens_seen": 269942784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046373119358074225,
      "loss": 3.0719,
      "theoretical_loss": 4.198425886222367,
      "tokens_seen": 270008320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046372116349047143,
      "loss": 3.1394,
      "theoretical_loss": 4.198304286534239,
      "tokens_seen": 270073856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637111334002006,
      "loss": 2.8706,
      "theoretical_loss": 4.198182724609624,
      "tokens_seen": 270139392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637011033099298,
      "loss": 2.9891,
      "theoretical_loss": 4.198061200427635,
      "tokens_seen": 270204928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463691073219659,
      "loss": 2.9082,
      "theoretical_loss": 4.197939713967407,
      "tokens_seen": 270270464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 160520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.315969467163086,
      "objective/train/theoretical_loss": 4.197818265208085,
      "objective/train/tokens_used": 290796000,
      "theoretical_loss": 4.197818265208085,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636810431293882,
      "loss": 2.939,
      "theoretical_loss": 4.197818265208085,
      "tokens_seen": 270336000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046367101303911734,
      "loss": 2.9865,
      "theoretical_loss": 4.197696854128836,
      "tokens_seen": 270401536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636609829488466,
      "loss": 2.9617,
      "theoretical_loss": 4.1975754807088395,
      "tokens_seen": 270467072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636509528585757,
      "loss": 3.1313,
      "theoretical_loss": 4.1974541449272955,
      "tokens_seen": 270532608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046364092276830494,
      "loss": 2.8589,
      "theoretical_loss": 4.197332846763418,
      "tokens_seen": 270598144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636308926780341,
      "loss": 3.168,
      "theoretical_loss": 4.197211586196437,
      "tokens_seen": 270663680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636208625877633,
      "loss": 3.1609,
      "theoretical_loss": 4.197090363205602,
      "tokens_seen": 270729216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636108324974925,
      "loss": 2.8238,
      "theoretical_loss": 4.196969177770175,
      "tokens_seen": 270794752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636008024072217,
      "loss": 3.0672,
      "theoretical_loss": 4.196848029869437,
      "tokens_seen": 270860288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359077231695084,
      "loss": 3.1223,
      "theoretical_loss": 4.1967269194826855,
      "tokens_seen": 270925824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635807422266801,
      "loss": 3.1919,
      "theoretical_loss": 4.196605846589234,
      "tokens_seen": 270991360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635707121364092,
      "loss": 3.1143,
      "theoretical_loss": 4.196484811168412,
      "tokens_seen": 271056896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046356068204613844,
      "loss": 2.7825,
      "theoretical_loss": 4.196363813199565,
      "tokens_seen": 271122432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635506519558676,
      "loss": 3.1101,
      "theoretical_loss": 4.196242852662056,
      "tokens_seen": 271187968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635406218655968,
      "loss": 3.1081,
      "theoretical_loss": 4.1961219295352645,
      "tokens_seen": 271253504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463530591775326,
      "loss": 3.1529,
      "theoretical_loss": 4.196001043798585,
      "tokens_seen": 271319040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046352056168505516,
      "loss": 3.1475,
      "theoretical_loss": 4.195880195431428,
      "tokens_seen": 271384576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046351053159478434,
      "loss": 2.9622,
      "theoretical_loss": 4.195759384413224,
      "tokens_seen": 271450112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635005015045136,
      "loss": 2.9907,
      "theoretical_loss": 4.195638610723415,
      "tokens_seen": 271515648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634904714142427,
      "loss": 2.7124,
      "theoretical_loss": 4.195517874341462,
      "tokens_seen": 271581184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046348044132397194,
      "loss": 2.9731,
      "theoretical_loss": 4.195397175246841,
      "tokens_seen": 271646720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046347041123370107,
      "loss": 3.1014,
      "theoretical_loss": 4.195276513419046,
      "tokens_seen": 271712256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634603811434303,
      "loss": 2.8613,
      "theoretical_loss": 4.195155888837585,
      "tokens_seen": 271777792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634503510531595,
      "loss": 3.0449,
      "theoretical_loss": 4.195035301481984,
      "tokens_seen": 271843328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046344032096288867,
      "loss": 2.7986,
      "theoretical_loss": 4.194914751331785,
      "tokens_seen": 271908864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 160942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2946550846099854,
      "objective/train/theoretical_loss": 4.194794238366544,
      "objective/train/tokens_used": 292434400,
      "theoretical_loss": 4.194794238366544,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343029087261785,
      "loss": 3.0722,
      "theoretical_loss": 4.194794238366544,
      "tokens_seen": 271974400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634202607823471,
      "loss": 3.0775,
      "theoretical_loss": 4.194673762565837,
      "tokens_seen": 272039936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634102306920762,
      "loss": 2.9018,
      "theoretical_loss": 4.194553323909252,
      "tokens_seen": 272105472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046340020060180545,
      "loss": 3.1252,
      "theoretical_loss": 4.1944329223763965,
      "tokens_seen": 272171008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046339017051153457,
      "loss": 2.9432,
      "theoretical_loss": 4.194312557946892,
      "tokens_seen": 272236544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633801404212638,
      "loss": 3.1044,
      "theoretical_loss": 4.194192230600376,
      "tokens_seen": 272302080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463370110330993,
      "loss": 3.0371,
      "theoretical_loss": 4.194071940316505,
      "tokens_seen": 272367616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046336008024072217,
      "loss": 3.1066,
      "theoretical_loss": 4.193951687074948,
      "tokens_seen": 272433152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046335005015045135,
      "loss": 2.9569,
      "theoretical_loss": 4.193831470855391,
      "tokens_seen": 272498688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046334002006018053,
      "loss": 3.1706,
      "theoretical_loss": 4.193711291637538,
      "tokens_seen": 272564224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633299899699097,
      "loss": 3.0875,
      "theoretical_loss": 4.1935911494011044,
      "tokens_seen": 272629760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046331995987963895,
      "loss": 3.0274,
      "theoretical_loss": 4.193471044125828,
      "tokens_seen": 272695296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633099297893681,
      "loss": 3.1833,
      "theoretical_loss": 4.193350975791458,
      "tokens_seen": 272760832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632998996990973,
      "loss": 3.0265,
      "theoretical_loss": 4.193230944377762,
      "tokens_seen": 272826368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046328986960882644,
      "loss": 2.999,
      "theoretical_loss": 4.1931109498645185,
      "tokens_seen": 272891904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632798395185557,
      "loss": 2.8691,
      "theoretical_loss": 4.19299099223153,
      "tokens_seen": 272957440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046326980942828485,
      "loss": 3.1468,
      "theoretical_loss": 4.192871071458608,
      "tokens_seen": 273022976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046325977933801404,
      "loss": 3.101,
      "theoretical_loss": 4.192751187525585,
      "tokens_seen": 273088512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632497492477432,
      "loss": 2.8904,
      "theoretical_loss": 4.192631340412304,
      "tokens_seen": 273154048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046323971915747245,
      "loss": 2.8277,
      "theoretical_loss": 4.192511530098629,
      "tokens_seen": 273219584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632296890672016,
      "loss": 3.0174,
      "theoretical_loss": 4.192391756564437,
      "tokens_seen": 273285120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632196589769308,
      "loss": 2.9337,
      "theoretical_loss": 4.19227201978962,
      "tokens_seen": 273350656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046320962888666,
      "loss": 3.2491,
      "theoretical_loss": 4.19215231975409,
      "tokens_seen": 273416192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631995987963892,
      "loss": 3.1248,
      "theoretical_loss": 4.192032656437769,
      "tokens_seen": 273481728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631895687061184,
      "loss": 3.1044,
      "theoretical_loss": 4.191913029820601,
      "tokens_seen": 273547264
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 162274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1903703212738037,
      "objective/train/theoretical_loss": 4.19179343988254,
      "objective/train/tokens_used": 294072800,
      "theoretical_loss": 4.19179343988254,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046317953861584754,
      "loss": 3.1483,
      "theoretical_loss": 4.19179343988254,
      "tokens_seen": 273612800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631695085255768,
      "loss": 2.8653,
      "theoretical_loss": 4.191673886603559,
      "tokens_seen": 273678336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631594784353059,
      "loss": 3.0805,
      "theoretical_loss": 4.191554369963647,
      "tokens_seen": 273743872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046314944834503514,
      "loss": 2.9344,
      "theoretical_loss": 4.191434889942807,
      "tokens_seen": 273809408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631394182547643,
      "loss": 3.0085,
      "theoretical_loss": 4.191315446521058,
      "tokens_seen": 273874944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631293881644935,
      "loss": 2.9289,
      "theoretical_loss": 4.191196039678436,
      "tokens_seen": 273940480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631193580742227,
      "loss": 3.1368,
      "theoretical_loss": 4.191076669394992,
      "tokens_seen": 274006016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631093279839519,
      "loss": 3.1613,
      "theoretical_loss": 4.190957335650791,
      "tokens_seen": 274071552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046309929789368104,
      "loss": 2.8155,
      "theoretical_loss": 4.190838038425916,
      "tokens_seen": 274137088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630892678034103,
      "loss": 3.1902,
      "theoretical_loss": 4.190718777700465,
      "tokens_seen": 274202624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630792377131394,
      "loss": 3.0513,
      "theoretical_loss": 4.190599553454551,
      "tokens_seen": 274268160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046306920762286864,
      "loss": 3.0037,
      "theoretical_loss": 4.190480365668304,
      "tokens_seen": 274333696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630591775325978,
      "loss": 2.8063,
      "theoretical_loss": 4.190361214321866,
      "tokens_seen": 274399232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463049147442327,
      "loss": 2.9742,
      "theoretical_loss": 4.190242099395399,
      "tokens_seen": 274464768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630391173520562,
      "loss": 2.9558,
      "theoretical_loss": 4.190123020869077,
      "tokens_seen": 274530304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046302908726178536,
      "loss": 2.9187,
      "theoretical_loss": 4.190003978723093,
      "tokens_seen": 274595840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046301905717151455,
      "loss": 3.0974,
      "theoretical_loss": 4.189884972937652,
      "tokens_seen": 274661376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630090270812438,
      "loss": 3.1408,
      "theoretical_loss": 4.189766003492977,
      "tokens_seen": 274726912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629989969909729,
      "loss": 3.0551,
      "theoretical_loss": 4.189647070369304,
      "tokens_seen": 274792448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298896690070214,
      "loss": 3.1843,
      "theoretical_loss": 4.189528173546888,
      "tokens_seen": 274857984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046297893681043127,
      "loss": 3.0686,
      "theoretical_loss": 4.189409313005997,
      "tokens_seen": 274923520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629689067201605,
      "loss": 3.0667,
      "theoretical_loss": 4.189290488726913,
      "tokens_seen": 274989056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629588766298897,
      "loss": 3.2895,
      "theoretical_loss": 4.189171700689937,
      "tokens_seen": 275054592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294884653961887,
      "loss": 3.0592,
      "theoretical_loss": 4.189052948875384,
      "tokens_seen": 275120128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046293881644934805,
      "loss": 2.9131,
      "theoretical_loss": 4.188934233263582,
      "tokens_seen": 275185664
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 163440,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.128859519958496,
      "objective/train/theoretical_loss": 4.188815553834879,
      "objective/train/tokens_used": 295711200,
      "theoretical_loss": 4.188815553834879,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629287863590773,
      "loss": 3.0631,
      "theoretical_loss": 4.188815553834879,
      "tokens_seen": 275251200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629187562688064,
      "loss": 2.8185,
      "theoretical_loss": 4.188696910569634,
      "tokens_seen": 275316736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046290872617853565,
      "loss": 3.0597,
      "theoretical_loss": 4.188578303448223,
      "tokens_seen": 275382272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046289869608826477,
      "loss": 3.0644,
      "theoretical_loss": 4.188459732451038,
      "tokens_seen": 275447808
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462888665997994,
      "loss": 3.2541,
      "theoretical_loss": 4.188341197558486,
      "tokens_seen": 275513344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628786359077232,
      "loss": 2.772,
      "theoretical_loss": 4.188222698750988,
      "tokens_seen": 275578880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046286860581745237,
      "loss": 2.945,
      "theoretical_loss": 4.188104236008982,
      "tokens_seen": 275644416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046285857572718155,
      "loss": 2.9857,
      "theoretical_loss": 4.18798580931292,
      "tokens_seen": 275709952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284854563691073,
      "loss": 2.8546,
      "theoretical_loss": 4.18786741864327,
      "tokens_seen": 275775488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628385155466399,
      "loss": 3.1056,
      "theoretical_loss": 4.187749063980516,
      "tokens_seen": 275841024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046282848545636915,
      "loss": 2.8899,
      "theoretical_loss": 4.187630745305154,
      "tokens_seen": 275906560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628184553660983,
      "loss": 2.9852,
      "theoretical_loss": 4.187512462597699,
      "tokens_seen": 275972096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628084252758275,
      "loss": 3.1059,
      "theoretical_loss": 4.187394215838678,
      "tokens_seen": 276037632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046279839518555664,
      "loss": 2.945,
      "theoretical_loss": 4.187276005008637,
      "tokens_seen": 276103168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627883650952859,
      "loss": 3.1584,
      "theoretical_loss": 4.1871578300881325,
      "tokens_seen": 276168704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046277833500501505,
      "loss": 3.1305,
      "theoretical_loss": 4.18703969105774,
      "tokens_seen": 276234240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046276830491474424,
      "loss": 2.8816,
      "theoretical_loss": 4.186921587898048,
      "tokens_seen": 276299776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627582748244734,
      "loss": 2.988,
      "theoretical_loss": 4.1868035205896605,
      "tokens_seen": 276365312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274824473420265,
      "loss": 2.8155,
      "theoretical_loss": 4.186685489113197,
      "tokens_seen": 276430848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627382146439318,
      "loss": 3.1242,
      "theoretical_loss": 4.186567493449292,
      "tokens_seen": 276496384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462728184553661,
      "loss": 3.0243,
      "theoretical_loss": 4.186449533578595,
      "tokens_seen": 276561920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046271815446339014,
      "loss": 2.9596,
      "theoretical_loss": 4.186331609481769,
      "tokens_seen": 276627456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627081243731194,
      "loss": 3.3328,
      "theoretical_loss": 4.186213721139496,
      "tokens_seen": 276692992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046269809428284856,
      "loss": 3.114,
      "theoretical_loss": 4.186095868532469,
      "tokens_seen": 276758528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046268806419257774,
      "loss": 2.8937,
      "theoretical_loss": 4.1859780516413965,
      "tokens_seen": 276824064
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 163970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.389463186264038,
      "objective/train/theoretical_loss": 4.185860270447004,
      "objective/train/tokens_used": 297349600,
      "theoretical_loss": 4.185860270447004,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626780341023069,
      "loss": 3.0341,
      "theoretical_loss": 4.185860270447004,
      "tokens_seen": 276889600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626680040120361,
      "loss": 3.1028,
      "theoretical_loss": 4.185742524930031,
      "tokens_seen": 276955136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626579739217653,
      "loss": 2.9541,
      "theoretical_loss": 4.185624815071232,
      "tokens_seen": 277020672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626479438314945,
      "loss": 2.9788,
      "theoretical_loss": 4.185507140851375,
      "tokens_seen": 277086208
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046263791374122364,
      "loss": 3.0129,
      "theoretical_loss": 4.185389502251246,
      "tokens_seen": 277151744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626278836509529,
      "loss": 2.8961,
      "theoretical_loss": 4.185271899251642,
      "tokens_seen": 277217280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462617853560682,
      "loss": 2.8754,
      "theoretical_loss": 4.185154331833379,
      "tokens_seen": 277282816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260782347041124,
      "loss": 2.9643,
      "theoretical_loss": 4.185036799977285,
      "tokens_seen": 277348352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625977933801404,
      "loss": 3.0866,
      "theoretical_loss": 4.184919303664203,
      "tokens_seen": 277413888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625877632898696,
      "loss": 3.1173,
      "theoretical_loss": 4.184801842874992,
      "tokens_seen": 277479424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625777331995988,
      "loss": 3.0844,
      "theoretical_loss": 4.1846844175905265,
      "tokens_seen": 277544960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462567703109328,
      "loss": 3.0707,
      "theoretical_loss": 4.184567027791693,
      "tokens_seen": 277610496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046255767301905715,
      "loss": 2.9221,
      "theoretical_loss": 4.184449673459395,
      "tokens_seen": 277676032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625476429287864,
      "loss": 3.0805,
      "theoretical_loss": 4.1843323545745505,
      "tokens_seen": 277741568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625376128385155,
      "loss": 2.9624,
      "theoretical_loss": 4.184215071118093,
      "tokens_seen": 277807104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252758274824475,
      "loss": 2.9347,
      "theoretical_loss": 4.184097823070967,
      "tokens_seen": 277872640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625175526579739,
      "loss": 3.0535,
      "theoretical_loss": 4.183980610414137,
      "tokens_seen": 277938176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625075225677031,
      "loss": 2.9498,
      "theoretical_loss": 4.18386343312858,
      "tokens_seen": 278003712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624974924774323,
      "loss": 2.9249,
      "theoretical_loss": 4.183746291195285,
      "tokens_seen": 278069248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046248746238716147,
      "loss": 3.0889,
      "theoretical_loss": 4.18362918459526,
      "tokens_seen": 278134784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046247743229689065,
      "loss": 2.9025,
      "theoretical_loss": 4.183512113309525,
      "tokens_seen": 278200320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624674022066199,
      "loss": 3.1737,
      "theoretical_loss": 4.1833950773191155,
      "tokens_seen": 278265856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046245737211634907,
      "loss": 2.8199,
      "theoretical_loss": 4.183278076605083,
      "tokens_seen": 278331392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244734202607825,
      "loss": 3.1762,
      "theoretical_loss": 4.18316111114849,
      "tokens_seen": 278396928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624373119358075,
      "loss": 3.1915,
      "theoretical_loss": 4.183044180930418,
      "tokens_seen": 278462464
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 165182,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.31074857711792,
      "objective/train/theoretical_loss": 4.182927285931959,
      "objective/train/tokens_used": 298988000,
      "theoretical_loss": 4.182927285931959,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624272818455366,
      "loss": 3.0716,
      "theoretical_loss": 4.182927285931959,
      "tokens_seen": 278528000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046241725175526585,
      "loss": 2.8782,
      "theoretical_loss": 4.182810426134223,
      "tokens_seen": 278593536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462407221664995,
      "loss": 3.1372,
      "theoretical_loss": 4.1826936015183325,
      "tokens_seen": 278659072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623971915747242,
      "loss": 3.0399,
      "theoretical_loss": 4.182576812065426,
      "tokens_seen": 278724608
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623871614844534,
      "loss": 3.1069,
      "theoretical_loss": 4.1824600577566535,
      "tokens_seen": 278790144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046237713139418257,
      "loss": 3.003,
      "theoretical_loss": 4.182343338573185,
      "tokens_seen": 278855680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236710130391175,
      "loss": 2.981,
      "theoretical_loss": 4.1822266544961995,
      "tokens_seen": 278921216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046235707121364093,
      "loss": 2.9963,
      "theoretical_loss": 4.1821100055068925,
      "tokens_seen": 278986752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623470411233701,
      "loss": 2.8282,
      "theoretical_loss": 4.181993391586476,
      "tokens_seen": 279052288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046233701103309935,
      "loss": 2.9174,
      "theoretical_loss": 4.181876812716174,
      "tokens_seen": 279117824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623269809428285,
      "loss": 3.0528,
      "theoretical_loss": 4.181760268877226,
      "tokens_seen": 279183360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623169508525577,
      "loss": 3.0571,
      "theoretical_loss": 4.181643760050885,
      "tokens_seen": 279248896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230692076228684,
      "loss": 3.0749,
      "theoretical_loss": 4.181527286218419,
      "tokens_seen": 279314432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622968906720161,
      "loss": 3.0104,
      "theoretical_loss": 4.181410847361111,
      "tokens_seen": 279379968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046228686058174525,
      "loss": 3.1389,
      "theoretical_loss": 4.181294443460257,
      "tokens_seen": 279445504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046227683049147444,
      "loss": 2.8566,
      "theoretical_loss": 4.18117807449717,
      "tokens_seen": 279511040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622668004012036,
      "loss": 3.1739,
      "theoretical_loss": 4.181061740453174,
      "tokens_seen": 279576576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046225677031093285,
      "loss": 3.1752,
      "theoretical_loss": 4.18094544130961,
      "tokens_seen": 279642112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462246740220662,
      "loss": 2.9331,
      "theoretical_loss": 4.180829177047832,
      "tokens_seen": 279707648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622367101303912,
      "loss": 2.8764,
      "theoretical_loss": 4.180712947649208,
      "tokens_seen": 279773184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046222668004012034,
      "loss": 3.1994,
      "theoretical_loss": 4.180596753095122,
      "tokens_seen": 279838720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622166499498496,
      "loss": 3.2665,
      "theoretical_loss": 4.18048059336697,
      "tokens_seen": 279904256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046220661985957876,
      "loss": 2.9311,
      "theoretical_loss": 4.180364468446165,
      "tokens_seen": 279969792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046219658976930794,
      "loss": 3.0726,
      "theoretical_loss": 4.180248378314133,
      "tokens_seen": 280035328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621865596790371,
      "loss": 3.0656,
      "theoretical_loss": 4.180132322952312,
      "tokens_seen": 280100864
    },
    {
      "epoch": 0.08,
      "objective/train/docs_used": 165816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.640207529067993,
      "objective/train/theoretical_loss": 4.180016302342159,
      "objective/train/tokens_used": 300626400,
      "theoretical_loss": 4.180016302342159,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621765295887663,
      "loss": 2.8359,
      "theoretical_loss": 4.180016302342159,
      "tokens_seen": 280166400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621664994984955,
      "loss": 3.0966,
      "theoretical_loss": 4.179900316465139,
      "tokens_seen": 280231936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621564694082247,
      "loss": 3.0808,
      "theoretical_loss": 4.179784365302739,
      "tokens_seen": 280297472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214643931795384,
      "loss": 2.9101,
      "theoretical_loss": 4.179668448836454,
      "tokens_seen": 280363008
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621364092276831,
      "loss": 3.1295,
      "theoretical_loss": 4.179552567047794,
      "tokens_seen": 280428544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621263791374122,
      "loss": 2.9127,
      "theoretical_loss": 4.179436719918286,
      "tokens_seen": 280494080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046211634904714144,
      "loss": 3.2209,
      "theoretical_loss": 4.179320907429468,
      "tokens_seen": 280559616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621063189568706,
      "loss": 3.0106,
      "theoretical_loss": 4.179205129562895,
      "tokens_seen": 280625152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620962888665998,
      "loss": 3.0269,
      "theoretical_loss": 4.179089386300134,
      "tokens_seen": 280690688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462086258776329,
      "loss": 3.1179,
      "theoretical_loss": 4.178973677622768,
      "tokens_seen": 280756224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620762286860582,
      "loss": 3.278,
      "theoretical_loss": 4.17885800351239,
      "tokens_seen": 280821760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046206619859578735,
      "loss": 2.9086,
      "theoretical_loss": 4.178742363950614,
      "tokens_seen": 280887296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620561685055166,
      "loss": 3.1124,
      "theoretical_loss": 4.178626758919062,
      "tokens_seen": 280952832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620461384152457,
      "loss": 3.0546,
      "theoretical_loss": 4.178511188399372,
      "tokens_seen": 281018368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046203610832497495,
      "loss": 3.1334,
      "theoretical_loss": 4.1783956523731955,
      "tokens_seen": 281083904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620260782347041,
      "loss": 2.9334,
      "theoretical_loss": 4.178280150822202,
      "tokens_seen": 281149440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620160481444333,
      "loss": 2.8178,
      "theoretical_loss": 4.178164683728068,
      "tokens_seen": 281214976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620060180541625,
      "loss": 2.8805,
      "theoretical_loss": 4.17804925107249,
      "tokens_seen": 281280512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046199598796389167,
      "loss": 2.919,
      "theoretical_loss": 4.1779338528371746,
      "tokens_seen": 281346048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046198595787362085,
      "loss": 3.1013,
      "theoretical_loss": 4.177818489003846,
      "tokens_seen": 281411584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619759277833501,
      "loss": 3.1344,
      "theoretical_loss": 4.177703159554238,
      "tokens_seen": 281477120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619658976930792,
      "loss": 3.0054,
      "theoretical_loss": 4.177587864470103,
      "tokens_seen": 281542656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046195586760280845,
      "loss": 2.8447,
      "theoretical_loss": 4.1774726037332055,
      "tokens_seen": 281608192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046194583751253763,
      "loss": 3.0365,
      "theoretical_loss": 4.17735737732532,
      "tokens_seen": 281673728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619358074222668,
      "loss": 3.0124,
      "theoretical_loss": 4.177242185228242,
      "tokens_seen": 281739264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 166963,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.019186019897461,
      "objective/train/theoretical_loss": 4.177127027423776,
      "objective/train/tokens_used": 302264800,
      "theoretical_loss": 4.177127027423776,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461925777331996,
      "loss": 2.9856,
      "theoretical_loss": 4.177127027423776,
      "tokens_seen": 281804800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619157472417252,
      "loss": 3.0215,
      "theoretical_loss": 4.17701190389374,
      "tokens_seen": 281870336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046190571715145435,
      "loss": 2.9971,
      "theoretical_loss": 4.176896814619969,
      "tokens_seen": 281935872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618956870611836,
      "loss": 3.0423,
      "theoretical_loss": 4.176781759584312,
      "tokens_seen": 282001408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618856569709127,
      "loss": 3.0798,
      "theoretical_loss": 4.176666738768627,
      "tokens_seen": 282066944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046187562688064195,
      "loss": 2.8872,
      "theoretical_loss": 4.176551752154791,
      "tokens_seen": 282132480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618655967903711,
      "loss": 3.1228,
      "theoretical_loss": 4.1764367997246925,
      "tokens_seen": 282198016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618555667001003,
      "loss": 3.0256,
      "theoretical_loss": 4.1763218814602325,
      "tokens_seen": 282263552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618455366098295,
      "loss": 3.1382,
      "theoretical_loss": 4.17620699734333,
      "tokens_seen": 282329088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618355065195587,
      "loss": 2.807,
      "theoretical_loss": 4.176092147355915,
      "tokens_seen": 282394624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046182547642928786,
      "loss": 2.9633,
      "theoretical_loss": 4.175977331479929,
      "tokens_seen": 282460160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046181544633901704,
      "loss": 3.0836,
      "theoretical_loss": 4.175862549697332,
      "tokens_seen": 282525696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618054162487462,
      "loss": 3.0762,
      "theoretical_loss": 4.175747801990093,
      "tokens_seen": 282591232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046179538615847545,
      "loss": 3.1594,
      "theoretical_loss": 4.1756330883402,
      "tokens_seen": 282656768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617853560682046,
      "loss": 3.136,
      "theoretical_loss": 4.1755184087296495,
      "tokens_seen": 282722304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617753259779338,
      "loss": 3.0347,
      "theoretical_loss": 4.175403763140456,
      "tokens_seen": 282787840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461765295887663,
      "loss": 2.9652,
      "theoretical_loss": 4.175289151554643,
      "tokens_seen": 282853376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617552657973922,
      "loss": 3.0566,
      "theoretical_loss": 4.175174573954252,
      "tokens_seen": 282918912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046174523570712136,
      "loss": 2.8965,
      "theoretical_loss": 4.175060030321337,
      "tokens_seen": 282984448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046173520561685054,
      "loss": 3.0043,
      "theoretical_loss": 4.174945520637964,
      "tokens_seen": 283049984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617251755265797,
      "loss": 3.124,
      "theoretical_loss": 4.174831044886213,
      "tokens_seen": 283115520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046171514543630896,
      "loss": 3.0059,
      "theoretical_loss": 4.17471660304818,
      "tokens_seen": 283181056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046170511534603814,
      "loss": 3.1171,
      "theoretical_loss": 4.174602195105972,
      "tokens_seen": 283246592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616950852557673,
      "loss": 2.735,
      "theoretical_loss": 4.174487821041711,
      "tokens_seen": 283312128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616850551654965,
      "loss": 2.9887,
      "theoretical_loss": 4.174373480837532,
      "tokens_seen": 283377664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 167624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3214924335479736,
      "objective/train/theoretical_loss": 4.174259174475583,
      "objective/train/tokens_used": 303903200,
      "theoretical_loss": 4.174259174475583,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616750250752257,
      "loss": 3.0628,
      "theoretical_loss": 4.174259174475583,
      "tokens_seen": 283443200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616649949849549,
      "loss": 3.1753,
      "theoretical_loss": 4.174144901938025,
      "tokens_seen": 283508736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046165496489468404,
      "loss": 2.9032,
      "theoretical_loss": 4.174030663207036,
      "tokens_seen": 283574272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616449348044133,
      "loss": 3.0913,
      "theoretical_loss": 4.173916458264804,
      "tokens_seen": 283639808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616349047141424,
      "loss": 3.0559,
      "theoretical_loss": 4.173802287093531,
      "tokens_seen": 283705344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046162487462387164,
      "loss": 2.8405,
      "theoretical_loss": 4.173688149675435,
      "tokens_seen": 283770880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616148445336008,
      "loss": 3.1749,
      "theoretical_loss": 4.173574045992744,
      "tokens_seen": 283836416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046160481444333,
      "loss": 2.9452,
      "theoretical_loss": 4.173459976027702,
      "tokens_seen": 283901952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615947843530592,
      "loss": 3.1586,
      "theoretical_loss": 4.173345939762564,
      "tokens_seen": 283967488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615847542627884,
      "loss": 2.9694,
      "theoretical_loss": 4.173231937179601,
      "tokens_seen": 284033024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046157472417251755,
      "loss": 3.2162,
      "theoretical_loss": 4.173117968261097,
      "tokens_seen": 284098560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615646940822468,
      "loss": 3.1405,
      "theoretical_loss": 4.173004032989347,
      "tokens_seen": 284164096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615546639919759,
      "loss": 3.0749,
      "theoretical_loss": 4.172890131346662,
      "tokens_seen": 284229632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046154463390170515,
      "loss": 3.1627,
      "theoretical_loss": 4.172776263315367,
      "tokens_seen": 284295168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615346038114343,
      "loss": 3.1345,
      "theoretical_loss": 4.172662428877796,
      "tokens_seen": 284360704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615245737211635,
      "loss": 2.9964,
      "theoretical_loss": 4.172548628016302,
      "tokens_seen": 284426240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615145436308927,
      "loss": 3.0997,
      "theoretical_loss": 4.172434860713246,
      "tokens_seen": 284491776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046150451354062187,
      "loss": 3.1053,
      "theoretical_loss": 4.172321126951008,
      "tokens_seen": 284557312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046149448345035105,
      "loss": 3.108,
      "theoretical_loss": 4.172207426711974,
      "tokens_seen": 284622848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614844533600803,
      "loss": 2.8824,
      "theoretical_loss": 4.1720937599785515,
      "tokens_seen": 284688384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614744232698094,
      "loss": 3.2381,
      "theoretical_loss": 4.171980126733155,
      "tokens_seen": 284753920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046146439317953865,
      "loss": 3.1941,
      "theoretical_loss": 4.171866526958215,
      "tokens_seen": 284819456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046145436308926783,
      "loss": 2.7743,
      "theoretical_loss": 4.171752960636176,
      "tokens_seen": 284884992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461444332998997,
      "loss": 2.9867,
      "theoretical_loss": 4.171639427749492,
      "tokens_seen": 284950528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614343029087262,
      "loss": 3.0296,
      "theoretical_loss": 4.171525928280635,
      "tokens_seen": 285016064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 168853,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6973540782928467,
      "objective/train/theoretical_loss": 4.171412462212087,
      "objective/train/tokens_used": 305541600,
      "theoretical_loss": 4.171412462212087,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614242728184554,
      "loss": 2.8165,
      "theoretical_loss": 4.171412462212087,
      "tokens_seen": 285081600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046141424272818455,
      "loss": 3.1382,
      "theoretical_loss": 4.171299029526344,
      "tokens_seen": 285147136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614042126379138,
      "loss": 3.1788,
      "theoretical_loss": 4.171185630205917,
      "tokens_seen": 285212672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613941825476429,
      "loss": 2.9897,
      "theoretical_loss": 4.171072264233327,
      "tokens_seen": 285278208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046138415245737215,
      "loss": 3.0584,
      "theoretical_loss": 4.17095893159111,
      "tokens_seen": 285343744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613741223671013,
      "loss": 3.0978,
      "theoretical_loss": 4.170845632261815,
      "tokens_seen": 285409280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613640922768305,
      "loss": 2.9229,
      "theoretical_loss": 4.1707323662280045,
      "tokens_seen": 285474816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613540621865597,
      "loss": 3.1211,
      "theoretical_loss": 4.170619133472253,
      "tokens_seen": 285540352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613440320962889,
      "loss": 2.9585,
      "theoretical_loss": 4.170505933977149,
      "tokens_seen": 285605888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046133400200601806,
      "loss": 3.1654,
      "theoretical_loss": 4.1703927677252945,
      "tokens_seen": 285671424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046132397191574724,
      "loss": 3.116,
      "theoretical_loss": 4.170279634699304,
      "tokens_seen": 285736960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613139418254764,
      "loss": 3.2018,
      "theoretical_loss": 4.170166534881803,
      "tokens_seen": 285802496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046130391173520566,
      "loss": 3.0549,
      "theoretical_loss": 4.170053468255435,
      "tokens_seen": 285868032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612938816449348,
      "loss": 2.795,
      "theoretical_loss": 4.169940434802851,
      "tokens_seen": 285933568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461283851554664,
      "loss": 3.0738,
      "theoretical_loss": 4.16982743450672,
      "tokens_seen": 285999104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612738214643932,
      "loss": 3.0826,
      "theoretical_loss": 4.169714467349721,
      "tokens_seen": 286064640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612637913741224,
      "loss": 2.9928,
      "theoretical_loss": 4.1696015333145455,
      "tokens_seen": 286130176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046125376128385156,
      "loss": 3.0505,
      "theoretical_loss": 4.169488632383901,
      "tokens_seen": 286195712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046124373119358074,
      "loss": 3.1082,
      "theoretical_loss": 4.169375764540505,
      "tokens_seen": 286261248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612337011033099,
      "loss": 2.9721,
      "theoretical_loss": 4.16926292976709,
      "tokens_seen": 286326784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046122367101303916,
      "loss": 3.1107,
      "theoretical_loss": 4.1691501280464,
      "tokens_seen": 286392320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612136409227683,
      "loss": 3.0903,
      "theoretical_loss": 4.1690373593611945,
      "tokens_seen": 286457856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612036108324975,
      "loss": 2.9085,
      "theoretical_loss": 4.168924623694241,
      "tokens_seen": 286523392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046119358074222665,
      "loss": 2.9324,
      "theoretical_loss": 4.168811921028325,
      "tokens_seen": 286588928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611835506519559,
      "loss": 2.9307,
      "theoretical_loss": 4.168699251346243,
      "tokens_seen": 286654464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 169449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8505656719207764,
      "objective/train/theoretical_loss": 4.168586614630803,
      "objective/train/tokens_used": 307180000,
      "theoretical_loss": 4.168586614630803,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046117352056168506,
      "loss": 3.006,
      "theoretical_loss": 4.168586614630803,
      "tokens_seen": 286720000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046116349047141425,
      "loss": 2.8379,
      "theoretical_loss": 4.168474010864829,
      "tokens_seen": 286785536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611534603811434,
      "loss": 2.8133,
      "theoretical_loss": 4.168361440031154,
      "tokens_seen": 286851072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611434302908726,
      "loss": 2.8832,
      "theoretical_loss": 4.1682489021126266,
      "tokens_seen": 286916608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611334002006018,
      "loss": 3.0822,
      "theoretical_loss": 4.168136397092109,
      "tokens_seen": 286982144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461123370110331,
      "loss": 3.1401,
      "theoretical_loss": 4.168023924952473,
      "tokens_seen": 287047680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046111334002006015,
      "loss": 3.0236,
      "theoretical_loss": 4.167911485676606,
      "tokens_seen": 287113216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611033099297894,
      "loss": 2.9008,
      "theoretical_loss": 4.1677990792474064,
      "tokens_seen": 287178752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046109327983951857,
      "loss": 3.0368,
      "theoretical_loss": 4.167686705647787,
      "tokens_seen": 287244288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046108324974924775,
      "loss": 2.8824,
      "theoretical_loss": 4.167574364860673,
      "tokens_seen": 287309824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046107321965897693,
      "loss": 2.8937,
      "theoretical_loss": 4.167462056869001,
      "tokens_seen": 287375360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610631895687061,
      "loss": 3.037,
      "theoretical_loss": 4.167349781655721,
      "tokens_seen": 287440896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610531594784353,
      "loss": 3.0067,
      "theoretical_loss": 4.167237539203798,
      "tokens_seen": 287506432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610431293881645,
      "loss": 2.8215,
      "theoretical_loss": 4.167125329496207,
      "tokens_seen": 287571968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046103309929789365,
      "loss": 3.092,
      "theoretical_loss": 4.167013152515935,
      "tokens_seen": 287637504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610230692076229,
      "loss": 2.834,
      "theoretical_loss": 4.1669010082459845,
      "tokens_seen": 287703040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461013039117352,
      "loss": 3.0378,
      "theoretical_loss": 4.16678889666937,
      "tokens_seen": 287768576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046100300902708125,
      "loss": 2.9538,
      "theoretical_loss": 4.166676817769118,
      "tokens_seen": 287834112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046099297893681043,
      "loss": 2.936,
      "theoretical_loss": 4.166564771528268,
      "tokens_seen": 287899648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609829488465396,
      "loss": 3.0104,
      "theoretical_loss": 4.16645275792987,
      "tokens_seen": 287965184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609729187562688,
      "loss": 3.0223,
      "theoretical_loss": 4.166340776956991,
      "tokens_seen": 288030720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046096288866599803,
      "loss": 2.9447,
      "theoretical_loss": 4.166228828592708,
      "tokens_seen": 288096256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609528585757272,
      "loss": 2.9425,
      "theoretical_loss": 4.166116912820109,
      "tokens_seen": 288161792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609428284854564,
      "loss": 3.0139,
      "theoretical_loss": 4.166005029622298,
      "tokens_seen": 288227328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609327983951856,
      "loss": 2.7581,
      "theoretical_loss": 4.16589317898239,
      "tokens_seen": 288292864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 170373,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9227163791656494,
      "objective/train/theoretical_loss": 4.165781360883513,
      "objective/train/tokens_used": 308818400,
      "theoretical_loss": 4.165781360883513,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046092276830491475,
      "loss": 3.0162,
      "theoretical_loss": 4.165781360883513,
      "tokens_seen": 288358400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460912738214644,
      "loss": 2.9972,
      "theoretical_loss": 4.165669575308807,
      "tokens_seen": 288423936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609027081243731,
      "loss": 2.9231,
      "theoretical_loss": 4.165557822241425,
      "tokens_seen": 288489472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046089267803410235,
      "loss": 2.9746,
      "theoretical_loss": 4.165446101664531,
      "tokens_seen": 288555008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608826479438315,
      "loss": 3.0331,
      "theoretical_loss": 4.165334413561306,
      "tokens_seen": 288620544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608726178535607,
      "loss": 3.0425,
      "theoretical_loss": 4.165222757914936,
      "tokens_seen": 288686080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608625877632899,
      "loss": 2.9917,
      "theoretical_loss": 4.165111134708629,
      "tokens_seen": 288751616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608525576730191,
      "loss": 3.0351,
      "theoretical_loss": 4.164999543925597,
      "tokens_seen": 288817152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084252758274826,
      "loss": 2.9461,
      "theoretical_loss": 4.164887985549068,
      "tokens_seen": 288882688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046083249749247744,
      "loss": 3.0879,
      "theoretical_loss": 4.164776459562285,
      "tokens_seen": 288948224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608224674022066,
      "loss": 2.9828,
      "theoretical_loss": 4.164664965948499,
      "tokens_seen": 289013760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046081243731193586,
      "loss": 3.1005,
      "theoretical_loss": 4.164553504690976,
      "tokens_seen": 289079296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460802407221665,
      "loss": 3.232,
      "theoretical_loss": 4.164442075772993,
      "tokens_seen": 289144832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607923771313942,
      "loss": 3.0206,
      "theoretical_loss": 4.164330679177841,
      "tokens_seen": 289210368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607823470411234,
      "loss": 2.9927,
      "theoretical_loss": 4.164219314888823,
      "tokens_seen": 289275904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607723169508526,
      "loss": 2.8879,
      "theoretical_loss": 4.164107982889252,
      "tokens_seen": 289341440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046076228686058176,
      "loss": 2.8926,
      "theoretical_loss": 4.163996683162458,
      "tokens_seen": 289406976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046075225677031094,
      "loss": 2.955,
      "theoretical_loss": 4.163885415691779,
      "tokens_seen": 289472512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607422266800401,
      "loss": 2.9791,
      "theoretical_loss": 4.163774180460569,
      "tokens_seen": 289538048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046073219658976936,
      "loss": 2.9538,
      "theoretical_loss": 4.163662977452192,
      "tokens_seen": 289603584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607221664994985,
      "loss": 2.9461,
      "theoretical_loss": 4.163551806650023,
      "tokens_seen": 289669120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607121364092277,
      "loss": 2.8711,
      "theoretical_loss": 4.163440668037453,
      "tokens_seen": 289734656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070210631895685,
      "loss": 3.0029,
      "theoretical_loss": 4.163329561597885,
      "tokens_seen": 289800192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606920762286861,
      "loss": 2.892,
      "theoretical_loss": 4.1632184873147295,
      "tokens_seen": 289865728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046068204613841526,
      "loss": 2.936,
      "theoretical_loss": 4.1631074451714145,
      "tokens_seen": 289931264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 171595,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1895124912261963,
      "objective/train/theoretical_loss": 4.162996435151379,
      "objective/train/tokens_used": 310456800,
      "theoretical_loss": 4.162996435151379,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046067201604814445,
      "loss": 3.0838,
      "theoretical_loss": 4.162996435151379,
      "tokens_seen": 289996800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606619859578736,
      "loss": 2.8576,
      "theoretical_loss": 4.1628854572380725,
      "tokens_seen": 290062336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606519558676028,
      "loss": 2.9221,
      "theoretical_loss": 4.162774511414959,
      "tokens_seen": 290127872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460641925777332,
      "loss": 2.9766,
      "theoretical_loss": 4.162663597665513,
      "tokens_seen": 290193408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606318956870612,
      "loss": 3.1019,
      "theoretical_loss": 4.162552715973222,
      "tokens_seen": 290258944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046062186559679035,
      "loss": 3.0226,
      "theoretical_loss": 4.162441866321586,
      "tokens_seen": 290324480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606118355065196,
      "loss": 3.1766,
      "theoretical_loss": 4.162331048694118,
      "tokens_seen": 290390016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060180541624877,
      "loss": 2.9799,
      "theoretical_loss": 4.162220263074341,
      "tokens_seen": 290455552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046059177532597795,
      "loss": 3.0867,
      "theoretical_loss": 4.1621095094457905,
      "tokens_seen": 290521088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046058174523570713,
      "loss": 3.0066,
      "theoretical_loss": 4.161998787792016,
      "tokens_seen": 290586624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605717151454363,
      "loss": 3.0877,
      "theoretical_loss": 4.161888098096579,
      "tokens_seen": 290652160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605616850551655,
      "loss": 2.8912,
      "theoretical_loss": 4.161777440343051,
      "tokens_seen": 290717696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605516549648947,
      "loss": 3.0279,
      "theoretical_loss": 4.161666814515017,
      "tokens_seen": 290783232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054162487462385,
      "loss": 2.9034,
      "theoretical_loss": 4.161556220596075,
      "tokens_seen": 290848768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605315947843531,
      "loss": 2.9769,
      "theoretical_loss": 4.161445658569833,
      "tokens_seen": 290914304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605215646940822,
      "loss": 3.0383,
      "theoretical_loss": 4.161335128419914,
      "tokens_seen": 290979840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046051153460381145,
      "loss": 2.9355,
      "theoretical_loss": 4.16122463012995,
      "tokens_seen": 291045376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050150451354063,
      "loss": 3.299,
      "theoretical_loss": 4.161114163683587,
      "tokens_seen": 291110912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604914744232698,
      "loss": 2.8822,
      "theoretical_loss": 4.161003729064483,
      "tokens_seen": 291176448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460481444332999,
      "loss": 3.0251,
      "theoretical_loss": 4.160893326256307,
      "tokens_seen": 291241984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046047141424272823,
      "loss": 2.8968,
      "theoretical_loss": 4.160782955242741,
      "tokens_seen": 291307520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046138415245736,
      "loss": 3.0828,
      "theoretical_loss": 4.160672616007479,
      "tokens_seen": 291373056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604513540621866,
      "loss": 3.2492,
      "theoretical_loss": 4.160562308534226,
      "tokens_seen": 291438592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604413239719157,
      "loss": 2.7536,
      "theoretical_loss": 4.160452032806701,
      "tokens_seen": 291504128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046043129388164495,
      "loss": 3.1393,
      "theoretical_loss": 4.160341788808633,
      "tokens_seen": 291569664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 172214,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0195140838623047,
      "objective/train/theoretical_loss": 4.160231576523763,
      "objective/train/tokens_used": 312095200,
      "theoretical_loss": 4.160231576523763,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042126379137414,
      "loss": 2.9725,
      "theoretical_loss": 4.160231576523763,
      "tokens_seen": 291635200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604112337011033,
      "loss": 3.0591,
      "theoretical_loss": 4.160121395935847,
      "tokens_seen": 291700736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604012036108325,
      "loss": 2.6618,
      "theoretical_loss": 4.160011247028649,
      "tokens_seen": 291766272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603911735205617,
      "loss": 2.9955,
      "theoretical_loss": 4.159901129785947,
      "tokens_seen": 291831808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046038114343029086,
      "loss": 2.9759,
      "theoretical_loss": 4.159791044191532,
      "tokens_seen": 291897344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603711133400201,
      "loss": 2.9241,
      "theoretical_loss": 4.1596809902292025,
      "tokens_seen": 291962880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603610832497492,
      "loss": 3.0356,
      "theoretical_loss": 4.159570967882775,
      "tokens_seen": 292028416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046035105315947846,
      "loss": 2.8435,
      "theoretical_loss": 4.1594609771360735,
      "tokens_seen": 292093952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603410230692076,
      "loss": 3.1677,
      "theoretical_loss": 4.159351017972936,
      "tokens_seen": 292159488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603309929789368,
      "loss": 2.7557,
      "theoretical_loss": 4.159241090377211,
      "tokens_seen": 292225024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460320962888666,
      "loss": 2.997,
      "theoretical_loss": 4.15913119433276,
      "tokens_seen": 292290560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603109327983952,
      "loss": 3.1299,
      "theoretical_loss": 4.159021329823457,
      "tokens_seen": 292356096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030090270812436,
      "loss": 2.9866,
      "theoretical_loss": 4.158911496833184,
      "tokens_seen": 292421632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602908726178536,
      "loss": 2.8174,
      "theoretical_loss": 4.15880169534584,
      "tokens_seen": 292487168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602808425275827,
      "loss": 3.0642,
      "theoretical_loss": 4.158691925345334,
      "tokens_seen": 292552704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046027081243731196,
      "loss": 2.9502,
      "theoretical_loss": 4.158582186815584,
      "tokens_seen": 292618240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602607823470411,
      "loss": 2.957,
      "theoretical_loss": 4.158472479740524,
      "tokens_seen": 292683776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602507522567703,
      "loss": 2.9731,
      "theoretical_loss": 4.158362804104097,
      "tokens_seen": 292749312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602407221664995,
      "loss": 3.0535,
      "theoretical_loss": 4.1582531598902595,
      "tokens_seen": 292814848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602306920762287,
      "loss": 2.985,
      "theoretical_loss": 4.158143547082978,
      "tokens_seen": 292880384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046022066198595787,
      "loss": 2.9562,
      "theoretical_loss": 4.158033965666233,
      "tokens_seen": 292945920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046021063189568705,
      "loss": 2.8592,
      "theoretical_loss": 4.157924415624014,
      "tokens_seen": 293011456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602006018054163,
      "loss": 2.9666,
      "theoretical_loss": 4.157814896940324,
      "tokens_seen": 293076992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046019057171514546,
      "loss": 3.2116,
      "theoretical_loss": 4.157705409599179,
      "tokens_seen": 293142528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046018054162487465,
      "loss": 2.9664,
      "theoretical_loss": 4.157595953584604,
      "tokens_seen": 293208064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 173342,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8450427055358887,
      "objective/train/theoretical_loss": 4.1574865288806375,
      "objective/train/tokens_used": 313733600,
      "theoretical_loss": 4.1574865288806375,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601705115346038,
      "loss": 2.9126,
      "theoretical_loss": 4.1574865288806375,
      "tokens_seen": 293273600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460160481444333,
      "loss": 2.7492,
      "theoretical_loss": 4.157377135471328,
      "tokens_seen": 293339136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601504513540622,
      "loss": 2.8848,
      "theoretical_loss": 4.157267773340738,
      "tokens_seen": 293404672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601404212637914,
      "loss": 3.0064,
      "theoretical_loss": 4.15715844247294,
      "tokens_seen": 293470208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046013039117352055,
      "loss": 2.9549,
      "theoretical_loss": 4.157049142852019,
      "tokens_seen": 293535744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601203610832498,
      "loss": 2.9487,
      "theoretical_loss": 4.15693987446207,
      "tokens_seen": 293601280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046011033099297897,
      "loss": 3.0182,
      "theoretical_loss": 4.156830637287203,
      "tokens_seen": 293666816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046010030090270815,
      "loss": 2.9087,
      "theoretical_loss": 4.156721431311535,
      "tokens_seen": 293732352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046009027081243733,
      "loss": 3.1088,
      "theoretical_loss": 4.1566122565192,
      "tokens_seen": 293797888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600802407221665,
      "loss": 3.0616,
      "theoretical_loss": 4.156503112894339,
      "tokens_seen": 293863424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600702106318957,
      "loss": 2.9094,
      "theoretical_loss": 4.156394000421107,
      "tokens_seen": 293928960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046006018054162493,
      "loss": 3.0927,
      "theoretical_loss": 4.15628491908367,
      "tokens_seen": 293994496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046005015045135405,
      "loss": 3.1256,
      "theoretical_loss": 4.156175868866205,
      "tokens_seen": 294060032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600401203610833,
      "loss": 3.0551,
      "theoretical_loss": 4.156066849752902,
      "tokens_seen": 294125568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600300902708124,
      "loss": 3.0325,
      "theoretical_loss": 4.15595786172796,
      "tokens_seen": 294191104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046002006018054165,
      "loss": 3.0107,
      "theoretical_loss": 4.155848904775594,
      "tokens_seen": 294256640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046001003009027083,
      "loss": 2.9292,
      "theoretical_loss": 4.155739978880025,
      "tokens_seen": 294322176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 2.831,
      "theoretical_loss": 4.155631084025489,
      "tokens_seen": 294387712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599899699097292,
      "loss": 3.1475,
      "theoretical_loss": 4.155522220196234,
      "tokens_seen": 294453248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045997993981945843,
      "loss": 2.9871,
      "theoretical_loss": 4.155413387376518,
      "tokens_seen": 294518784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045996990972918756,
      "loss": 3.1932,
      "theoretical_loss": 4.155304585550609,
      "tokens_seen": 294584320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599598796389168,
      "loss": 2.9215,
      "theoretical_loss": 4.15519581470279,
      "tokens_seen": 294649856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599498495486459,
      "loss": 2.9558,
      "theoretical_loss": 4.155087074817353,
      "tokens_seen": 294715392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993981945837515,
      "loss": 3.2962,
      "theoretical_loss": 4.154978365878603,
      "tokens_seen": 294780928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045992978936810434,
      "loss": 3.1586,
      "theoretical_loss": 4.154869687870854,
      "tokens_seen": 294846464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 173989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.6182363033294678,
      "objective/train/theoretical_loss": 4.1547610407784346,
      "objective/train/tokens_used": 315372000,
      "theoretical_loss": 4.1547610407784346,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599197592778335,
      "loss": 3.244,
      "theoretical_loss": 4.1547610407784346,
      "tokens_seen": 294912000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599097291875627,
      "loss": 3.042,
      "theoretical_loss": 4.154652424585683,
      "tokens_seen": 294977536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598996990972919,
      "loss": 2.722,
      "theoretical_loss": 4.154543839276948,
      "tokens_seen": 295043072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045988966900702106,
      "loss": 3.0362,
      "theoretical_loss": 4.154435284836591,
      "tokens_seen": 295108608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598796389167503,
      "loss": 2.8414,
      "theoretical_loss": 4.154326761248987,
      "tokens_seen": 295174144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598696088264794,
      "loss": 2.858,
      "theoretical_loss": 4.154218268498517,
      "tokens_seen": 295239680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045985957873620866,
      "loss": 2.9538,
      "theoretical_loss": 4.154109806569577,
      "tokens_seen": 295305216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598495486459378,
      "loss": 3.0794,
      "theoretical_loss": 4.154001375446576,
      "tokens_seen": 295370752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459839518555667,
      "loss": 3.0819,
      "theoretical_loss": 4.15389297511393,
      "tokens_seen": 295436288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598294884653962,
      "loss": 3.1095,
      "theoretical_loss": 4.153784605556069,
      "tokens_seen": 295501824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598194583751254,
      "loss": 3.2757,
      "theoretical_loss": 4.153676266757434,
      "tokens_seen": 295567360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045980942828485456,
      "loss": 2.7567,
      "theoretical_loss": 4.153567958702476,
      "tokens_seen": 295632896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597993981945838,
      "loss": 3.0827,
      "theoretical_loss": 4.153459681375661,
      "tokens_seen": 295698432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597893681043129,
      "loss": 3.2228,
      "theoretical_loss": 4.15335143476146,
      "tokens_seen": 295763968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045977933801404216,
      "loss": 2.9004,
      "theoretical_loss": 4.153243218844362,
      "tokens_seen": 295829504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597693079237713,
      "loss": 2.9718,
      "theoretical_loss": 4.153135033608863,
      "tokens_seen": 295895040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597592778335005,
      "loss": 2.9733,
      "theoretical_loss": 4.153026879039471,
      "tokens_seen": 295960576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597492477432297,
      "loss": 2.9921,
      "theoretical_loss": 4.152918755120708,
      "tokens_seen": 296026112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597392176529589,
      "loss": 2.9436,
      "theoretical_loss": 4.152810661837102,
      "tokens_seen": 296091648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045972918756268807,
      "loss": 3.1248,
      "theoretical_loss": 4.152702599173197,
      "tokens_seen": 296157184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045971915747241725,
      "loss": 3.1309,
      "theoretical_loss": 4.152594567113546,
      "tokens_seen": 296222720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045970912738214643,
      "loss": 2.9432,
      "theoretical_loss": 4.152486565642714,
      "tokens_seen": 296288256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969909729187566,
      "loss": 3.2128,
      "theoretical_loss": 4.152378594745276,
      "tokens_seen": 296353792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596890672016048,
      "loss": 3.153,
      "theoretical_loss": 4.15227065440582,
      "tokens_seen": 296419328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459679037111334,
      "loss": 3.1355,
      "theoretical_loss": 4.152162744608944,
      "tokens_seen": 296484864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 175144,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.142443895339966,
      "objective/train/theoretical_loss": 4.152054865339256,
      "objective/train/tokens_used": 317010400,
      "theoretical_loss": 4.152054865339256,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045966900702106315,
      "loss": 3.1106,
      "theoretical_loss": 4.152054865339256,
      "tokens_seen": 296550400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596589769307924,
      "loss": 3.0485,
      "theoretical_loss": 4.151947016581378,
      "tokens_seen": 296615936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045964894684052157,
      "loss": 3.0957,
      "theoretical_loss": 4.151839198319942,
      "tokens_seen": 296681472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045963891675025075,
      "loss": 2.7138,
      "theoretical_loss": 4.15173141053959,
      "tokens_seen": 296747008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045962888665997993,
      "loss": 2.8906,
      "theoretical_loss": 4.151623653224975,
      "tokens_seen": 296812544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045961885656970917,
      "loss": 2.9162,
      "theoretical_loss": 4.151515926360764,
      "tokens_seen": 296878080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596088264794383,
      "loss": 2.9349,
      "theoretical_loss": 4.151408229931631,
      "tokens_seen": 296943616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045959879638916753,
      "loss": 3.0348,
      "theoretical_loss": 4.151300563922266,
      "tokens_seen": 297009152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045958876629889666,
      "loss": 3.0585,
      "theoretical_loss": 4.151192928317365,
      "tokens_seen": 297074688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595787362086259,
      "loss": 2.8223,
      "theoretical_loss": 4.151085323101638,
      "tokens_seen": 297140224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595687061183551,
      "loss": 2.8607,
      "theoretical_loss": 4.150977748259805,
      "tokens_seen": 297205760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045955867602808425,
      "loss": 2.8969,
      "theoretical_loss": 4.1508702037765985,
      "tokens_seen": 297271296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045954864593781344,
      "loss": 2.9414,
      "theoretical_loss": 4.150762689636761,
      "tokens_seen": 297336832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595386158475426,
      "loss": 3.0862,
      "theoretical_loss": 4.150655205825045,
      "tokens_seen": 297402368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595285857572718,
      "loss": 3.018,
      "theoretical_loss": 4.150547752326216,
      "tokens_seen": 297467904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045951855566700103,
      "loss": 2.9801,
      "theoretical_loss": 4.15044032912505,
      "tokens_seen": 297533440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045950852557673016,
      "loss": 3.0379,
      "theoretical_loss": 4.150332936206332,
      "tokens_seen": 297598976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594984954864594,
      "loss": 3.249,
      "theoretical_loss": 4.150225573554859,
      "tokens_seen": 297664512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594884653961885,
      "loss": 3.0122,
      "theoretical_loss": 4.150118241155443,
      "tokens_seen": 297730048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045947843530591776,
      "loss": 3.2498,
      "theoretical_loss": 4.1500109389929,
      "tokens_seen": 297795584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045946840521564694,
      "loss": 2.7565,
      "theoretical_loss": 4.1499036670520635,
      "tokens_seen": 297861120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594583751253761,
      "loss": 3.0607,
      "theoretical_loss": 4.149796425317772,
      "tokens_seen": 297926656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045944834503510536,
      "loss": 3.1681,
      "theoretical_loss": 4.149689213774879,
      "tokens_seen": 297992192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045943831494483454,
      "loss": 3.1897,
      "theoretical_loss": 4.149582032408249,
      "tokens_seen": 298057728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594282848545637,
      "loss": 3.1223,
      "theoretical_loss": 4.149474881202755,
      "tokens_seen": 298123264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 176258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.113511323928833,
      "objective/train/theoretical_loss": 4.1493677601432815,
      "objective/train/tokens_used": 318648800,
      "theoretical_loss": 4.1493677601432815,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594182547642929,
      "loss": 3.0056,
      "theoretical_loss": 4.1493677601432815,
      "tokens_seen": 298188800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594082246740221,
      "loss": 3.017,
      "theoretical_loss": 4.149260669214725,
      "tokens_seen": 298254336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045939819458375126,
      "loss": 3.0696,
      "theoretical_loss": 4.149153608401994,
      "tokens_seen": 298319872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593881644934805,
      "loss": 2.9312,
      "theoretical_loss": 4.149046577690003,
      "tokens_seen": 298385408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593781344032096,
      "loss": 2.979,
      "theoretical_loss": 4.148939577063683,
      "tokens_seen": 298450944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045936810431293886,
      "loss": 2.7,
      "theoretical_loss": 4.148832606507974,
      "tokens_seen": 298516480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459358074222668,
      "loss": 2.9176,
      "theoretical_loss": 4.148725666007824,
      "tokens_seen": 298582016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593480441323972,
      "loss": 2.9862,
      "theoretical_loss": 4.1486187555481955,
      "tokens_seen": 298647552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593380140421264,
      "loss": 2.7821,
      "theoretical_loss": 4.14851187511406,
      "tokens_seen": 298713088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593279839518556,
      "loss": 3.0391,
      "theoretical_loss": 4.148405024690402,
      "tokens_seen": 298778624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045931795386158476,
      "loss": 2.9083,
      "theoretical_loss": 4.148298204262213,
      "tokens_seen": 298844160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459307923771314,
      "loss": 2.8753,
      "theoretical_loss": 4.148191413814497,
      "tokens_seen": 298909696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592978936810431,
      "loss": 2.9916,
      "theoretical_loss": 4.148084653332271,
      "tokens_seen": 298975232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045928786359077236,
      "loss": 3.0754,
      "theoretical_loss": 4.14797792280056,
      "tokens_seen": 299040768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592778335005015,
      "loss": 3.0386,
      "theoretical_loss": 4.147871222204401,
      "tokens_seen": 299106304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592678034102307,
      "loss": 2.9645,
      "theoretical_loss": 4.147764551528841,
      "tokens_seen": 299171840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592577733199599,
      "loss": 2.7558,
      "theoretical_loss": 4.147657910758938,
      "tokens_seen": 299237376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592477432296891,
      "loss": 2.7373,
      "theoretical_loss": 4.147551299879762,
      "tokens_seen": 299302912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045923771313941827,
      "loss": 2.8164,
      "theoretical_loss": 4.147444718876391,
      "tokens_seen": 299368448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045922768304914745,
      "loss": 2.8048,
      "theoretical_loss": 4.147338167733917,
      "tokens_seen": 299433984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045921765295887663,
      "loss": 3.0629,
      "theoretical_loss": 4.14723164643744,
      "tokens_seen": 299499520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045920762286860586,
      "loss": 2.948,
      "theoretical_loss": 4.1471251549720725,
      "tokens_seen": 299565056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459197592778335,
      "loss": 2.9864,
      "theoretical_loss": 4.147018693322936,
      "tokens_seen": 299630592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591875626880642,
      "loss": 2.9198,
      "theoretical_loss": 4.146912261475165,
      "tokens_seen": 299696128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045917753259779335,
      "loss": 3.034,
      "theoretical_loss": 4.1468058594139015,
      "tokens_seen": 299761664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 176682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.946582078933716,
      "objective/train/theoretical_loss": 4.146699487124301,
      "objective/train/tokens_used": 320287200,
      "theoretical_loss": 4.146699487124301,
      "tokens_seen": 299827200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591675025075226,
      "loss": 3.0244,
      "theoretical_loss": 4.146699487124301,
      "tokens_seen": 299827200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045915747241725177,
      "loss": 3.0344,
      "theoretical_loss": 4.146593144591527,
      "tokens_seen": 299892736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045914744232698095,
      "loss": 3.1026,
      "theoretical_loss": 4.146486831800758,
      "tokens_seen": 299958272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045913741223671013,
      "loss": 3.005,
      "theoretical_loss": 4.146380548737179,
      "tokens_seen": 300023808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045912738214643937,
      "loss": 3.1644,
      "theoretical_loss": 4.146274295385986,
      "tokens_seen": 300089344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591173520561685,
      "loss": 2.8718,
      "theoretical_loss": 4.146168071732388,
      "tokens_seen": 300154880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045910732196589773,
      "loss": 3.0563,
      "theoretical_loss": 4.146061877761602,
      "tokens_seen": 300220416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045909729187562686,
      "loss": 3.226,
      "theoretical_loss": 4.145955713458856,
      "tokens_seen": 300285952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590872617853561,
      "loss": 3.0286,
      "theoretical_loss": 4.145849578809392,
      "tokens_seen": 300351488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590772316950853,
      "loss": 2.9362,
      "theoretical_loss": 4.145743473798457,
      "tokens_seen": 300417024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045906720160481445,
      "loss": 3.3005,
      "theoretical_loss": 4.145637398411313,
      "tokens_seen": 300482560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045905717151454364,
      "loss": 3.1683,
      "theoretical_loss": 4.145531352633229,
      "tokens_seen": 300548096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590471414242728,
      "loss": 2.8494,
      "theoretical_loss": 4.14542533644949,
      "tokens_seen": 300613632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459037111334002,
      "loss": 2.7638,
      "theoretical_loss": 4.145319349845384,
      "tokens_seen": 300679168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045902708124373123,
      "loss": 3.1229,
      "theoretical_loss": 4.145213392806217,
      "tokens_seen": 300744704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045901705115346036,
      "loss": 2.8249,
      "theoretical_loss": 4.145107465317299,
      "tokens_seen": 300810240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590070210631896,
      "loss": 2.7708,
      "theoretical_loss": 4.1450015673639555,
      "tokens_seen": 300875776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589969909729187,
      "loss": 2.9347,
      "theoretical_loss": 4.144895698931521,
      "tokens_seen": 300941312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045898696088264796,
      "loss": 3.1446,
      "theoretical_loss": 4.144789860005337,
      "tokens_seen": 301006848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045897693079237714,
      "loss": 2.7626,
      "theoretical_loss": 4.14468405057076,
      "tokens_seen": 301072384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589669007021063,
      "loss": 2.9021,
      "theoretical_loss": 4.144578270613157,
      "tokens_seen": 301137920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589568706118355,
      "loss": 2.9562,
      "theoretical_loss": 4.144472520117901,
      "tokens_seen": 301203456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045894684052156474,
      "loss": 3.0293,
      "theoretical_loss": 4.14436679907038,
      "tokens_seen": 301268992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045893681043129386,
      "loss": 2.9459,
      "theoretical_loss": 4.144261107455991,
      "tokens_seen": 301334528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589267803410231,
      "loss": 2.9452,
      "theoretical_loss": 4.1441554452601395,
      "tokens_seen": 301400064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 177896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9508297443389893,
      "objective/train/theoretical_loss": 4.1440498124682446,
      "objective/train/tokens_used": 321925600,
      "theoretical_loss": 4.1440498124682446,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589167502507522,
      "loss": 3.0381,
      "theoretical_loss": 4.1440498124682446,
      "tokens_seen": 301465600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045890672016048146,
      "loss": 3.0254,
      "theoretical_loss": 4.143944209065733,
      "tokens_seen": 301531136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045889669007021064,
      "loss": 2.948,
      "theoretical_loss": 4.143838635038043,
      "tokens_seen": 301596672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588866599799398,
      "loss": 3.035,
      "theoretical_loss": 4.143733090370624,
      "tokens_seen": 301662208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458876629889669,
      "loss": 2.9586,
      "theoretical_loss": 4.143627575048935,
      "tokens_seen": 301727744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588665997993982,
      "loss": 2.904,
      "theoretical_loss": 4.143522089058446,
      "tokens_seen": 301793280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045885656970912737,
      "loss": 2.9791,
      "theoretical_loss": 4.143416632384635,
      "tokens_seen": 301858816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588465396188566,
      "loss": 2.9168,
      "theoretical_loss": 4.143311205012992,
      "tokens_seen": 301924352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045883650952858573,
      "loss": 2.9416,
      "theoretical_loss": 4.14320580692902,
      "tokens_seen": 301989888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045882647943831496,
      "loss": 3.2419,
      "theoretical_loss": 4.143100438118227,
      "tokens_seen": 302055424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045881644934804415,
      "loss": 3.0354,
      "theoretical_loss": 4.142995098566135,
      "tokens_seen": 302120960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588064192577733,
      "loss": 2.891,
      "theoretical_loss": 4.142889788258275,
      "tokens_seen": 302186496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587963891675025,
      "loss": 3.0476,
      "theoretical_loss": 4.142784507180189,
      "tokens_seen": 302252032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587863590772317,
      "loss": 2.9312,
      "theoretical_loss": 4.14267925531743,
      "tokens_seen": 302317568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045877632898696087,
      "loss": 3.0632,
      "theoretical_loss": 4.142574032655558,
      "tokens_seen": 302383104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587662988966901,
      "loss": 3.0175,
      "theoretical_loss": 4.142468839180148,
      "tokens_seen": 302448640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045875626880641923,
      "loss": 3.1759,
      "theoretical_loss": 4.14236367487678,
      "tokens_seen": 302514176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045874623871614847,
      "loss": 3.0104,
      "theoretical_loss": 4.142258539731047,
      "tokens_seen": 302579712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587362086258776,
      "loss": 2.8409,
      "theoretical_loss": 4.142153433728556,
      "tokens_seen": 302645248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045872617853560683,
      "loss": 3.0343,
      "theoretical_loss": 4.142048356854915,
      "tokens_seen": 302710784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458716148445336,
      "loss": 3.1169,
      "theoretical_loss": 4.141943309095752,
      "tokens_seen": 302776320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587061183550652,
      "loss": 2.7164,
      "theoretical_loss": 4.141838290436699,
      "tokens_seen": 302841856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586960882647944,
      "loss": 3.1328,
      "theoretical_loss": 4.1417333008633985,
      "tokens_seen": 302907392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045868605817452355,
      "loss": 2.9251,
      "theoretical_loss": 4.141628340361508,
      "tokens_seen": 302972928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586760280842528,
      "loss": 3.0941,
      "theoretical_loss": 4.141523408916689,
      "tokens_seen": 303038464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 178251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.023979425430298,
      "objective/train/theoretical_loss": 4.141418506514618,
      "objective/train/tokens_used": 323564000,
      "theoretical_loss": 4.141418506514618,
      "tokens_seen": 303104000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045866599799398197,
      "loss": 3.0323,
      "theoretical_loss": 4.141418506514618,
      "tokens_seen": 303104000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045865596790371115,
      "loss": 2.966,
      "theoretical_loss": 4.141313633140978,
      "tokens_seen": 303169536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045864593781344033,
      "loss": 3.0848,
      "theoretical_loss": 4.141208788781466,
      "tokens_seen": 303235072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045863590772316957,
      "loss": 3.0573,
      "theoretical_loss": 4.141103973421784,
      "tokens_seen": 303300608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586258776328987,
      "loss": 3.0507,
      "theoretical_loss": 4.14099918704765,
      "tokens_seen": 303366144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045861584754262793,
      "loss": 2.9278,
      "theoretical_loss": 4.140894429644788,
      "tokens_seen": 303431680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045860581745235706,
      "loss": 3.1005,
      "theoretical_loss": 4.1407897011989325,
      "tokens_seen": 303497216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585957873620863,
      "loss": 3.0779,
      "theoretical_loss": 4.1406850016958305,
      "tokens_seen": 303562752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585857572718155,
      "loss": 2.8597,
      "theoretical_loss": 4.140580331121236,
      "tokens_seen": 303628288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045857572718154465,
      "loss": 2.8306,
      "theoretical_loss": 4.140475689460916,
      "tokens_seen": 303693824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045856569709127384,
      "loss": 3.0461,
      "theoretical_loss": 4.140371076700646,
      "tokens_seen": 303759360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458555667001003,
      "loss": 3.0263,
      "theoretical_loss": 4.140266492826211,
      "tokens_seen": 303824896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585456369107322,
      "loss": 3.0801,
      "theoretical_loss": 4.140161937823407,
      "tokens_seen": 303890432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045853560682046143,
      "loss": 2.9155,
      "theoretical_loss": 4.140057411678041,
      "tokens_seen": 303955968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045852557673019056,
      "loss": 2.8495,
      "theoretical_loss": 4.139952914375927,
      "tokens_seen": 304021504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585155466399198,
      "loss": 2.9827,
      "theoretical_loss": 4.139848445902892,
      "tokens_seen": 304087040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585055165496489,
      "loss": 2.9606,
      "theoretical_loss": 4.139744006244772,
      "tokens_seen": 304152576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045849548645937816,
      "loss": 2.8089,
      "theoretical_loss": 4.139639595387413,
      "tokens_seen": 304218112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045848545636910734,
      "loss": 3.0314,
      "theoretical_loss": 4.139535213316671,
      "tokens_seen": 304283648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584754262788365,
      "loss": 2.972,
      "theoretical_loss": 4.139430860018411,
      "tokens_seen": 304349184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584653961885657,
      "loss": 3.0738,
      "theoretical_loss": 4.13932653547851,
      "tokens_seen": 304414720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045845536609829494,
      "loss": 2.8192,
      "theoretical_loss": 4.139222239682853,
      "tokens_seen": 304480256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045844533600802406,
      "loss": 2.805,
      "theoretical_loss": 4.139117972617337,
      "tokens_seen": 304545792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584353059177533,
      "loss": 2.8678,
      "theoretical_loss": 4.139013734267866,
      "tokens_seen": 304611328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584252758274824,
      "loss": 2.9425,
      "theoretical_loss": 4.138909524620358,
      "tokens_seen": 304676864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 179536,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0331430435180664,
      "objective/train/theoretical_loss": 4.138805343660737,
      "objective/train/tokens_used": 325202400,
      "theoretical_loss": 4.138805343660737,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045841524573721166,
      "loss": 2.9569,
      "theoretical_loss": 4.138805343660737,
      "tokens_seen": 304742400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045840521564694084,
      "loss": 3.0136,
      "theoretical_loss": 4.13870119137494,
      "tokens_seen": 304807936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045839518555667,
      "loss": 2.7669,
      "theoretical_loss": 4.138597067748911,
      "tokens_seen": 304873472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583851554663992,
      "loss": 2.9702,
      "theoretical_loss": 4.138492972768606,
      "tokens_seen": 304939008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583751253761284,
      "loss": 3.1139,
      "theoretical_loss": 4.138388906419991,
      "tokens_seen": 305004544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045836509528585757,
      "loss": 2.9143,
      "theoretical_loss": 4.138284868689041,
      "tokens_seen": 305070080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583550651955868,
      "loss": 2.7962,
      "theoretical_loss": 4.13818085956174,
      "tokens_seen": 305135616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045834503510531593,
      "loss": 3.0717,
      "theoretical_loss": 4.1380768790240845,
      "tokens_seen": 305201152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045833500501504516,
      "loss": 2.8641,
      "theoretical_loss": 4.137972927062078,
      "tokens_seen": 305266688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045832497492477435,
      "loss": 2.9519,
      "theoretical_loss": 4.137869003661737,
      "tokens_seen": 305332224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583149448345035,
      "loss": 3.1127,
      "theoretical_loss": 4.137765108809084,
      "tokens_seen": 305397760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583049147442327,
      "loss": 2.9791,
      "theoretical_loss": 4.1376612424901555,
      "tokens_seen": 305463296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582948846539619,
      "loss": 2.9104,
      "theoretical_loss": 4.137557404690993,
      "tokens_seen": 305528832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045828485456369107,
      "loss": 3.0647,
      "theoretical_loss": 4.137453595397653,
      "tokens_seen": 305594368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582748244734203,
      "loss": 2.9847,
      "theoretical_loss": 4.137349814596198,
      "tokens_seen": 305659904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045826479438314943,
      "loss": 3.0683,
      "theoretical_loss": 4.137246062272702,
      "tokens_seen": 305725440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045825476429287867,
      "loss": 2.9846,
      "theoretical_loss": 4.137142338413248,
      "tokens_seen": 305790976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582447342026078,
      "loss": 2.7466,
      "theoretical_loss": 4.137038643003931,
      "tokens_seen": 305856512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045823470411233703,
      "loss": 2.9283,
      "theoretical_loss": 4.136934976030852,
      "tokens_seen": 305922048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582246740220662,
      "loss": 3.0245,
      "theoretical_loss": 4.136831337480125,
      "tokens_seen": 305987584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582146439317954,
      "loss": 2.8523,
      "theoretical_loss": 4.13672772733787,
      "tokens_seen": 306053120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045820461384152457,
      "loss": 2.8644,
      "theoretical_loss": 4.136624145590222,
      "tokens_seen": 306118656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045819458375125375,
      "loss": 3.1607,
      "theoretical_loss": 4.136520592223322,
      "tokens_seen": 306184192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045818455366098294,
      "loss": 2.9101,
      "theoretical_loss": 4.136417067223322,
      "tokens_seen": 306249728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045817452357071217,
      "loss": 3.1469,
      "theoretical_loss": 4.136313570576383,
      "tokens_seen": 306315264
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 180132,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.187363624572754,
      "objective/train/theoretical_loss": 4.136210102268676,
      "objective/train/tokens_used": 326840800,
      "theoretical_loss": 4.136210102268676,
      "tokens_seen": 306380800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581644934804413,
      "loss": 3.0491,
      "theoretical_loss": 4.136210102268676,
      "tokens_seen": 306380800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045815446339017053,
      "loss": 3.2312,
      "theoretical_loss": 4.136106662286383,
      "tokens_seen": 306446336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581444332998997,
      "loss": 3.1742,
      "theoretical_loss": 4.136003250615693,
      "tokens_seen": 306511872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581344032096289,
      "loss": 2.983,
      "theoretical_loss": 4.135899867242808,
      "tokens_seen": 306577408
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581243731193581,
      "loss": 3.1178,
      "theoretical_loss": 4.135796512153934,
      "tokens_seen": 306642944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045811434302908726,
      "loss": 2.8261,
      "theoretical_loss": 4.135693185335295,
      "tokens_seen": 306708480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045810431293881644,
      "loss": 3.001,
      "theoretical_loss": 4.135589886773119,
      "tokens_seen": 306774016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580942828485457,
      "loss": 2.9546,
      "theoretical_loss": 4.135486616453644,
      "tokens_seen": 306839552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580842527582748,
      "loss": 3.0585,
      "theoretical_loss": 4.135383374363119,
      "tokens_seen": 306905088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045807422266800404,
      "loss": 3.023,
      "theoretical_loss": 4.1352801604878024,
      "tokens_seen": 306970624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045806419257773316,
      "loss": 2.954,
      "theoretical_loss": 4.135176974813961,
      "tokens_seen": 307036160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580541624874624,
      "loss": 3.259,
      "theoretical_loss": 4.135073817327872,
      "tokens_seen": 307101696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580441323971916,
      "loss": 3.053,
      "theoretical_loss": 4.134970688015824,
      "tokens_seen": 307167232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045803410230692076,
      "loss": 2.8994,
      "theoretical_loss": 4.134867586864112,
      "tokens_seen": 307232768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045802407221664994,
      "loss": 3.019,
      "theoretical_loss": 4.134764513859043,
      "tokens_seen": 307298304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580140421263791,
      "loss": 2.7869,
      "theoretical_loss": 4.134661468986932,
      "tokens_seen": 307363840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580040120361083,
      "loss": 2.8573,
      "theoretical_loss": 4.1345584522341055,
      "tokens_seen": 307429376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045799398194583754,
      "loss": 2.8166,
      "theoretical_loss": 4.134455463586898,
      "tokens_seen": 307494912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045798395185556667,
      "loss": 2.6522,
      "theoretical_loss": 4.134352503031652,
      "tokens_seen": 307560448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579739217652959,
      "loss": 3.0032,
      "theoretical_loss": 4.134249570554724,
      "tokens_seen": 307625984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579638916750251,
      "loss": 2.8976,
      "theoretical_loss": 4.134146666142476,
      "tokens_seen": 307691520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045795386158475426,
      "loss": 2.7396,
      "theoretical_loss": 4.134043789781281,
      "tokens_seen": 307757056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579438314944835,
      "loss": 3.0013,
      "theoretical_loss": 4.1339409414575226,
      "tokens_seen": 307822592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579338014042126,
      "loss": 2.9292,
      "theoretical_loss": 4.133838121157592,
      "tokens_seen": 307888128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045792377131394186,
      "loss": 3.2311,
      "theoretical_loss": 4.133735328867892,
      "tokens_seen": 307953664
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 181325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.002859354019165,
      "objective/train/theoretical_loss": 4.133632564574831,
      "objective/train/tokens_used": 328479200,
      "theoretical_loss": 4.133632564574831,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045791374122367104,
      "loss": 2.918,
      "theoretical_loss": 4.133632564574831,
      "tokens_seen": 308019200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579037111334002,
      "loss": 2.8607,
      "theoretical_loss": 4.133529828264831,
      "tokens_seen": 308084736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578936810431294,
      "loss": 2.9817,
      "theoretical_loss": 4.133427119924323,
      "tokens_seen": 308150272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578836509528586,
      "loss": 3.1387,
      "theoretical_loss": 4.133324439539745,
      "tokens_seen": 308215808
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045787362086258777,
      "loss": 3.2184,
      "theoretical_loss": 4.133221787097546,
      "tokens_seen": 308281344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457863590772317,
      "loss": 2.8897,
      "theoretical_loss": 4.133119162584185,
      "tokens_seen": 308346880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045785356068204613,
      "loss": 3.0767,
      "theoretical_loss": 4.1330165659861295,
      "tokens_seen": 308412416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045784353059177536,
      "loss": 2.9042,
      "theoretical_loss": 4.132913997289856,
      "tokens_seen": 308477952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045783350050150455,
      "loss": 2.9545,
      "theoretical_loss": 4.1328114564818526,
      "tokens_seen": 308543488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578234704112337,
      "loss": 2.8544,
      "theoretical_loss": 4.132708943548614,
      "tokens_seen": 308609024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578134403209629,
      "loss": 3.1091,
      "theoretical_loss": 4.132606458476647,
      "tokens_seen": 308674560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578034102306921,
      "loss": 2.9141,
      "theoretical_loss": 4.132504001252465,
      "tokens_seen": 308740096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045779338014042127,
      "loss": 2.9562,
      "theoretical_loss": 4.132401571862593,
      "tokens_seen": 308805632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577833500501505,
      "loss": 2.9368,
      "theoretical_loss": 4.132299170293564,
      "tokens_seen": 308871168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045777331995987963,
      "loss": 2.8221,
      "theoretical_loss": 4.132196796531922,
      "tokens_seen": 308936704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045776328986960887,
      "loss": 2.9433,
      "theoretical_loss": 4.132094450564218,
      "tokens_seen": 309002240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457753259779338,
      "loss": 2.8562,
      "theoretical_loss": 4.1319921323770155,
      "tokens_seen": 309067776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045774322968906723,
      "loss": 3.0209,
      "theoretical_loss": 4.131889841956884,
      "tokens_seen": 309133312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577331995987964,
      "loss": 3.009,
      "theoretical_loss": 4.131787579290405,
      "tokens_seen": 309198848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577231695085256,
      "loss": 2.7863,
      "theoretical_loss": 4.1316853443641675,
      "tokens_seen": 309264384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577131394182548,
      "loss": 3.0576,
      "theoretical_loss": 4.13158313716477,
      "tokens_seen": 309329920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045770310932798395,
      "loss": 2.9071,
      "theoretical_loss": 4.131480957678823,
      "tokens_seen": 309395456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045769307923771314,
      "loss": 2.9949,
      "theoretical_loss": 4.1313788058929415,
      "tokens_seen": 309460992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045768304914744237,
      "loss": 2.8319,
      "theoretical_loss": 4.131276681793754,
      "tokens_seen": 309526528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576730190571715,
      "loss": 2.7952,
      "theoretical_loss": 4.131174585367896,
      "tokens_seen": 309592064
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 182674,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9335567951202393,
      "objective/train/theoretical_loss": 4.131072516602014,
      "objective/train/tokens_used": 330117600,
      "theoretical_loss": 4.131072516602014,
      "tokens_seen": 309657600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045766298896690073,
      "loss": 2.9291,
      "theoretical_loss": 4.131072516602014,
      "tokens_seen": 309657600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576529588766299,
      "loss": 2.8759,
      "theoretical_loss": 4.130970475482761,
      "tokens_seen": 309723136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576429287863591,
      "loss": 2.7693,
      "theoretical_loss": 4.130868461996803,
      "tokens_seen": 309788672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576328986960883,
      "loss": 3.2834,
      "theoretical_loss": 4.1307664761308125,
      "tokens_seen": 309854208
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045762286860581746,
      "loss": 2.7814,
      "theoretical_loss": 4.130664517871471,
      "tokens_seen": 309919744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045761283851554664,
      "loss": 3.364,
      "theoretical_loss": 4.130562587205473,
      "tokens_seen": 309985280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576028084252759,
      "loss": 2.8317,
      "theoretical_loss": 4.130460684119516,
      "tokens_seen": 310050816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457592778335005,
      "loss": 2.9605,
      "theoretical_loss": 4.130358808600313,
      "tokens_seen": 310116352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045758274824473424,
      "loss": 3.1335,
      "theoretical_loss": 4.130256960634582,
      "tokens_seen": 310181888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045757271815446336,
      "loss": 3.0687,
      "theoretical_loss": 4.130155140209052,
      "tokens_seen": 310247424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575626880641926,
      "loss": 2.9797,
      "theoretical_loss": 4.130053347310461,
      "tokens_seen": 310312960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575526579739218,
      "loss": 3.294,
      "theoretical_loss": 4.129951581925555,
      "tokens_seen": 310378496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045754262788365096,
      "loss": 2.9568,
      "theoretical_loss": 4.129849844041091,
      "tokens_seen": 310444032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045753259779338014,
      "loss": 2.941,
      "theoretical_loss": 4.129748133643835,
      "tokens_seen": 310509568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575225677031093,
      "loss": 3.0219,
      "theoretical_loss": 4.129646450720561,
      "tokens_seen": 310575104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575125376128385,
      "loss": 2.9852,
      "theoretical_loss": 4.129544795258052,
      "tokens_seen": 310640640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045750250752256774,
      "loss": 2.8331,
      "theoretical_loss": 4.129443167243101,
      "tokens_seen": 310706176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045749247743229687,
      "loss": 3.061,
      "theoretical_loss": 4.129341566662511,
      "tokens_seen": 310771712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574824473420261,
      "loss": 2.967,
      "theoretical_loss": 4.1292399935030915,
      "tokens_seen": 310837248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574724172517553,
      "loss": 3.0105,
      "theoretical_loss": 4.129138447751664,
      "tokens_seen": 310902784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045746238716148446,
      "loss": 2.9289,
      "theoretical_loss": 4.129036929395058,
      "tokens_seen": 310968320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045745235707121364,
      "loss": 3.0223,
      "theoretical_loss": 4.12893543842011,
      "tokens_seen": 311033856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574423269809428,
      "loss": 2.9751,
      "theoretical_loss": 4.12883397481367,
      "tokens_seen": 311099392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457432296890672,
      "loss": 2.8895,
      "theoretical_loss": 4.128732538562592,
      "tokens_seen": 311164928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045742226680040124,
      "loss": 2.7377,
      "theoretical_loss": 4.128631129653744,
      "tokens_seen": 311230464
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 183229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0134546756744385,
      "objective/train/theoretical_loss": 4.128529748073999,
      "objective/train/tokens_used": 331756000,
      "theoretical_loss": 4.128529748073999,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045741223671013037,
      "loss": 2.6315,
      "theoretical_loss": 4.128529748073999,
      "tokens_seen": 311296000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574022066198596,
      "loss": 3.0727,
      "theoretical_loss": 4.128428393810242,
      "tokens_seen": 311361536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045739217652958873,
      "loss": 3.0927,
      "theoretical_loss": 4.128327066849365,
      "tokens_seen": 311427072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045738214643931797,
      "loss": 2.984,
      "theoretical_loss": 4.12822576717827,
      "tokens_seen": 311492608
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045737211634904715,
      "loss": 3.1219,
      "theoretical_loss": 4.128124494783869,
      "tokens_seen": 311558144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045736208625877633,
      "loss": 2.8423,
      "theoretical_loss": 4.12802324965308,
      "tokens_seen": 311623680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573520561685055,
      "loss": 3.034,
      "theoretical_loss": 4.127922031772834,
      "tokens_seen": 311689216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045734202607823475,
      "loss": 3.0936,
      "theoretical_loss": 4.127820841130068,
      "tokens_seen": 311754752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045733199598796387,
      "loss": 2.8057,
      "theoretical_loss": 4.127719677711729,
      "tokens_seen": 311820288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573219658976931,
      "loss": 2.8598,
      "theoretical_loss": 4.1276185415047735,
      "tokens_seen": 311885824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045731193580742223,
      "loss": 2.8944,
      "theoretical_loss": 4.127517432496166,
      "tokens_seen": 311951360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045730190571715147,
      "loss": 2.9157,
      "theoretical_loss": 4.127416350672879,
      "tokens_seen": 312016896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045729187562688065,
      "loss": 2.938,
      "theoretical_loss": 4.1273152960219,
      "tokens_seen": 312082432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045728184553660983,
      "loss": 3.0826,
      "theoretical_loss": 4.127214268530216,
      "tokens_seen": 312147968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457271815446339,
      "loss": 3.0073,
      "theoretical_loss": 4.12711326818483,
      "tokens_seen": 312213504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572617853560682,
      "loss": 3.0567,
      "theoretical_loss": 4.127012294972753,
      "tokens_seen": 312279040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572517552657974,
      "loss": 3.169,
      "theoretical_loss": 4.126911348881,
      "tokens_seen": 312344576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572417251755266,
      "loss": 2.9469,
      "theoretical_loss": 4.126810429896603,
      "tokens_seen": 312410112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045723169508525574,
      "loss": 2.9813,
      "theoretical_loss": 4.126709538006596,
      "tokens_seen": 312475648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457221664994985,
      "loss": 2.934,
      "theoretical_loss": 4.1266086731980245,
      "tokens_seen": 312541184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572116349047141,
      "loss": 3.2086,
      "theoretical_loss": 4.126507835457945,
      "tokens_seen": 312606720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045720160481444334,
      "loss": 2.7577,
      "theoretical_loss": 4.126407024773418,
      "tokens_seen": 312672256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045719157472417257,
      "loss": 2.9729,
      "theoretical_loss": 4.126306241131519,
      "tokens_seen": 312737792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571815446339017,
      "loss": 2.9627,
      "theoretical_loss": 4.126205484519327,
      "tokens_seen": 312803328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045717151454363093,
      "loss": 3.0685,
      "theoretical_loss": 4.126104754923931,
      "tokens_seen": 312868864
    },
    {
      "epoch": 0.09,
      "objective/train/docs_used": 184180,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2469050884246826,
      "objective/train/theoretical_loss": 4.126004052332432,
      "objective/train/tokens_used": 333394400,
      "theoretical_loss": 4.126004052332432,
      "tokens_seen": 312934400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571614844533601,
      "loss": 3.0363,
      "theoretical_loss": 4.126004052332432,
      "tokens_seen": 312934400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571514543630893,
      "loss": 2.9679,
      "theoretical_loss": 4.125903376731936,
      "tokens_seen": 312999936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571414242728185,
      "loss": 2.8487,
      "theoretical_loss": 4.125802728109562,
      "tokens_seen": 313065472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045713139418254766,
      "loss": 2.8367,
      "theoretical_loss": 4.125702106452433,
      "tokens_seen": 313131008
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045712136409227684,
      "loss": 2.9977,
      "theoretical_loss": 4.125601511747685,
      "tokens_seen": 313196544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571113340020061,
      "loss": 3.0608,
      "theoretical_loss": 4.125500943982459,
      "tokens_seen": 313262080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571013039117352,
      "loss": 2.8297,
      "theoretical_loss": 4.125400403143908,
      "tokens_seen": 313327616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045709127382146444,
      "loss": 3.052,
      "theoretical_loss": 4.125299889219194,
      "tokens_seen": 313393152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045708124373119356,
      "loss": 3.0677,
      "theoretical_loss": 4.125199402195486,
      "tokens_seen": 313458688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570712136409228,
      "loss": 2.9362,
      "theoretical_loss": 4.12509894205996,
      "tokens_seen": 313524224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000457061183550652,
      "loss": 2.9709,
      "theoretical_loss": 4.124998508799806,
      "tokens_seen": 313589760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045705115346038116,
      "loss": 3.04,
      "theoretical_loss": 4.124898102402218,
      "tokens_seen": 313655296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045704112337011034,
      "loss": 3.0373,
      "theoretical_loss": 4.124797722854401,
      "tokens_seen": 313720832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570310932798395,
      "loss": 2.8581,
      "theoretical_loss": 4.12469737014357,
      "tokens_seen": 313786368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570210631895687,
      "loss": 3.0926,
      "theoretical_loss": 4.124597044256945,
      "tokens_seen": 313851904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045701103309929794,
      "loss": 3.0979,
      "theoretical_loss": 4.124496745181759,
      "tokens_seen": 313917440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045700100300902707,
      "loss": 2.9877,
      "theoretical_loss": 4.12439647290525,
      "tokens_seen": 313982976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569909729187563,
      "loss": 3.2337,
      "theoretical_loss": 4.124296227414667,
      "tokens_seen": 314048512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569809428284855,
      "loss": 2.9723,
      "theoretical_loss": 4.124196008697267,
      "tokens_seen": 314114048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045697091273821466,
      "loss": 3.2751,
      "theoretical_loss": 4.124095816740317,
      "tokens_seen": 314179584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045696088264794384,
      "loss": 2.9168,
      "theoretical_loss": 4.123995651531089,
      "tokens_seen": 314245120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456950852557673,
      "loss": 2.9351,
      "theoretical_loss": 4.123895513056869,
      "tokens_seen": 314310656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569408224674022,
      "loss": 3.0289,
      "theoretical_loss": 4.123795401304948,
      "tokens_seen": 314376192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045693079237713144,
      "loss": 2.9276,
      "theoretical_loss": 4.123695316262626,
      "tokens_seen": 314441728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045692076228686057,
      "loss": 3.3239,
      "theoretical_loss": 4.123595257917214,
      "tokens_seen": 314507264
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 184828,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.897738218307495,
      "objective/train/theoretical_loss": 4.123495226256027,
      "objective/train/tokens_used": 335032800,
      "theoretical_loss": 4.123495226256027,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569107321965898,
      "loss": 2.8664,
      "theoretical_loss": 4.123495226256027,
      "tokens_seen": 314572800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045690070210631893,
      "loss": 2.9238,
      "theoretical_loss": 4.123395221266396,
      "tokens_seen": 314638336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045689067201604817,
      "loss": 3.0188,
      "theoretical_loss": 4.123295242935653,
      "tokens_seen": 314703872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045688064192577735,
      "loss": 2.9396,
      "theoretical_loss": 4.123195291251142,
      "tokens_seen": 314769408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045687061183550653,
      "loss": 3.1508,
      "theoretical_loss": 4.123095366200217,
      "tokens_seen": 314834944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568605817452357,
      "loss": 2.9989,
      "theoretical_loss": 4.122995467770239,
      "tokens_seen": 314900480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045685055165496495,
      "loss": 2.9205,
      "theoretical_loss": 4.122895595948577,
      "tokens_seen": 314966016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045684052156469407,
      "loss": 2.8574,
      "theoretical_loss": 4.122795750722612,
      "tokens_seen": 315031552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568304914744233,
      "loss": 2.7756,
      "theoretical_loss": 4.122695932079728,
      "tokens_seen": 315097088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045682046138415243,
      "loss": 3.1906,
      "theoretical_loss": 4.122596140007323,
      "tokens_seen": 315162624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045681043129388167,
      "loss": 2.9296,
      "theoretical_loss": 4.122496374492799,
      "tokens_seen": 315228160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045680040120361085,
      "loss": 2.6739,
      "theoretical_loss": 4.122396635523571,
      "tokens_seen": 315293696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045679037111334003,
      "loss": 2.8686,
      "theoretical_loss": 4.12229692308706,
      "tokens_seen": 315359232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567803410230692,
      "loss": 3.2873,
      "theoretical_loss": 4.122197237170695,
      "tokens_seen": 315424768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567703109327984,
      "loss": 3.0861,
      "theoretical_loss": 4.122097577761917,
      "tokens_seen": 315490304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567602808425276,
      "loss": 3.1712,
      "theoretical_loss": 4.121997944848171,
      "tokens_seen": 315555840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567502507522568,
      "loss": 3.1188,
      "theoretical_loss": 4.121898338416912,
      "tokens_seen": 315621376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045674022066198594,
      "loss": 2.9156,
      "theoretical_loss": 4.121798758455608,
      "tokens_seen": 315686912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567301905717152,
      "loss": 2.9989,
      "theoretical_loss": 4.121699204951728,
      "tokens_seen": 315752448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567201604814443,
      "loss": 2.6808,
      "theoretical_loss": 4.121599677892755,
      "tokens_seen": 315817984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045671013039117354,
      "loss": 2.9749,
      "theoretical_loss": 4.121500177266178,
      "tokens_seen": 315883520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567001003009027,
      "loss": 2.9981,
      "theoretical_loss": 4.121400703059497,
      "tokens_seen": 315949056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566900702106319,
      "loss": 3.0599,
      "theoretical_loss": 4.121301255260217,
      "tokens_seen": 316014592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566800401203611,
      "loss": 2.9343,
      "theoretical_loss": 4.121201833855856,
      "tokens_seen": 316080128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566700100300903,
      "loss": 3.0488,
      "theoretical_loss": 4.121102438833935,
      "tokens_seen": 316145664
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 186091,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.028201103210449,
      "objective/train/theoretical_loss": 4.121003070181987,
      "objective/train/tokens_used": 336671200,
      "theoretical_loss": 4.121003070181987,
      "tokens_seen": 316211200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045665997993981944,
      "loss": 3.0888,
      "theoretical_loss": 4.121003070181987,
      "tokens_seen": 316211200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566499498495487,
      "loss": 2.8606,
      "theoretical_loss": 4.120903727887554,
      "tokens_seen": 316276736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566399197592778,
      "loss": 2.8707,
      "theoretical_loss": 4.120804411938185,
      "tokens_seen": 316342272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045662988966900704,
      "loss": 3.1801,
      "theoretical_loss": 4.120705122321437,
      "tokens_seen": 316407808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566198595787362,
      "loss": 2.9392,
      "theoretical_loss": 4.1206058590248755,
      "tokens_seen": 316473344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566098294884654,
      "loss": 3.0525,
      "theoretical_loss": 4.120506622036077,
      "tokens_seen": 316538880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565997993981946,
      "loss": 2.8932,
      "theoretical_loss": 4.120407411342623,
      "tokens_seen": 316604416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045658976930792376,
      "loss": 3.0008,
      "theoretical_loss": 4.1203082269321065,
      "tokens_seen": 316669952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045657973921765294,
      "loss": 2.9041,
      "theoretical_loss": 4.120209068792127,
      "tokens_seen": 316735488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565697091273822,
      "loss": 2.8144,
      "theoretical_loss": 4.120109936910291,
      "tokens_seen": 316801024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565596790371113,
      "loss": 2.9752,
      "theoretical_loss": 4.120010831274218,
      "tokens_seen": 316866560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045654964894684054,
      "loss": 2.9861,
      "theoretical_loss": 4.119911751871531,
      "tokens_seen": 316932096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045653961885656967,
      "loss": 3.1201,
      "theoretical_loss": 4.1198126986898655,
      "tokens_seen": 316997632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565295887662989,
      "loss": 3.2564,
      "theoretical_loss": 4.119713671716862,
      "tokens_seen": 317063168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565195586760281,
      "loss": 3.0303,
      "theoretical_loss": 4.11961467094017,
      "tokens_seen": 317128704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045650952858575727,
      "loss": 3.0001,
      "theoretical_loss": 4.119515696347451,
      "tokens_seen": 317194240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045649949849548645,
      "loss": 3.1713,
      "theoretical_loss": 4.119416747926369,
      "tokens_seen": 317259776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564894684052157,
      "loss": 2.9097,
      "theoretical_loss": 4.119317825664601,
      "tokens_seen": 317325312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564794383149448,
      "loss": 3.0256,
      "theoretical_loss": 4.119218929549831,
      "tokens_seen": 317390848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045646940822467405,
      "loss": 2.9957,
      "theoretical_loss": 4.119120059569751,
      "tokens_seen": 317456384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045645937813440317,
      "loss": 3.1022,
      "theoretical_loss": 4.11902121571206,
      "tokens_seen": 317521920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564493480441324,
      "loss": 2.8814,
      "theoretical_loss": 4.118922397964468,
      "tokens_seen": 317587456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045643931795386164,
      "loss": 2.8242,
      "theoretical_loss": 4.118823606314692,
      "tokens_seen": 317652992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045642928786359077,
      "loss": 3.1631,
      "theoretical_loss": 4.118724840750456,
      "tokens_seen": 317718528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045641925777332,
      "loss": 3.0526,
      "theoretical_loss": 4.118626101259496,
      "tokens_seen": 317784064
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 187059,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0989761352539062,
      "objective/train/theoretical_loss": 4.118527387829552,
      "objective/train/tokens_used": 338309600,
      "theoretical_loss": 4.118527387829552,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045640922768304913,
      "loss": 3.1599,
      "theoretical_loss": 4.118527387829552,
      "tokens_seen": 317849600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045639919759277837,
      "loss": 2.9083,
      "theoretical_loss": 4.118428700448375,
      "tokens_seen": 317915136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045638916750250755,
      "loss": 2.7942,
      "theoretical_loss": 4.118330039103723,
      "tokens_seen": 317980672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045637913741223673,
      "loss": 2.8139,
      "theoretical_loss": 4.118231403783365,
      "tokens_seen": 318046208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563691073219659,
      "loss": 3.037,
      "theoretical_loss": 4.118132794475073,
      "tokens_seen": 318111744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045635907723169515,
      "loss": 3.0491,
      "theoretical_loss": 4.118034211166632,
      "tokens_seen": 318177280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045634904714142427,
      "loss": 2.8304,
      "theoretical_loss": 4.117935653845834,
      "tokens_seen": 318242816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563390170511535,
      "loss": 2.9257,
      "theoretical_loss": 4.117837122500477,
      "tokens_seen": 318308352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045632898696088263,
      "loss": 3.1308,
      "theoretical_loss": 4.117738617118372,
      "tokens_seen": 318373888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045631895687061187,
      "loss": 2.8832,
      "theoretical_loss": 4.1176401376873315,
      "tokens_seen": 318439424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045630892678034105,
      "loss": 3.0881,
      "theoretical_loss": 4.117541684195183,
      "tokens_seen": 318504960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045629889669007023,
      "loss": 3.0655,
      "theoretical_loss": 4.117443256629758,
      "tokens_seen": 318570496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562888665997994,
      "loss": 3.2203,
      "theoretical_loss": 4.1173448549788985,
      "tokens_seen": 318636032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562788365095286,
      "loss": 2.8932,
      "theoretical_loss": 4.117246479230452,
      "tokens_seen": 318701568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562688064192578,
      "loss": 3.1329,
      "theoretical_loss": 4.117148129372277,
      "tokens_seen": 318767104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456258776328987,
      "loss": 3.0318,
      "theoretical_loss": 4.117049805392239,
      "tokens_seen": 318832640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045624874623871614,
      "loss": 3.0924,
      "theoretical_loss": 4.116951507278211,
      "tokens_seen": 318898176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562387161484454,
      "loss": 3.2415,
      "theoretical_loss": 4.116853235018075,
      "tokens_seen": 318963712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562286860581745,
      "loss": 2.8784,
      "theoretical_loss": 4.116754988599722,
      "tokens_seen": 319029248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045621865596790374,
      "loss": 2.9519,
      "theoretical_loss": 4.11665676801105,
      "tokens_seen": 319094784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562086258776329,
      "loss": 3.0788,
      "theoretical_loss": 4.116558573239965,
      "tokens_seen": 319160320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561985957873621,
      "loss": 2.9773,
      "theoretical_loss": 4.116460404274381,
      "tokens_seen": 319225856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561885656970913,
      "loss": 2.8958,
      "theoretical_loss": 4.116362261102221,
      "tokens_seen": 319291392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561785356068205,
      "loss": 3.1937,
      "theoretical_loss": 4.116264143711416,
      "tokens_seen": 319356928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045616850551654964,
      "loss": 3.0717,
      "theoretical_loss": 4.1161660520899055,
      "tokens_seen": 319422464
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 187800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.174799680709839,
      "objective/train/theoretical_loss": 4.116067986225635,
      "objective/train/tokens_used": 339948000,
      "theoretical_loss": 4.116067986225635,
      "tokens_seen": 319488000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561584754262789,
      "loss": 3.0015,
      "theoretical_loss": 4.116067986225635,
      "tokens_seen": 319488000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456148445336008,
      "loss": 2.8217,
      "theoretical_loss": 4.11596994610656,
      "tokens_seen": 319553536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045613841524573724,
      "loss": 3.3545,
      "theoretical_loss": 4.115871931720645,
      "tokens_seen": 319619072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561283851554664,
      "loss": 2.7344,
      "theoretical_loss": 4.11577394305586,
      "tokens_seen": 319684608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561183550651956,
      "loss": 2.8487,
      "theoretical_loss": 4.115675980100185,
      "tokens_seen": 319750144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561083249749248,
      "loss": 2.9135,
      "theoretical_loss": 4.115578042841607,
      "tokens_seen": 319815680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045609829488465396,
      "loss": 2.9709,
      "theoretical_loss": 4.115480131268122,
      "tokens_seen": 319881216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045608826479438314,
      "loss": 2.9644,
      "theoretical_loss": 4.1153822453677344,
      "tokens_seen": 319946752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560782347041124,
      "loss": 2.9821,
      "theoretical_loss": 4.115284385128454,
      "tokens_seen": 320012288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560682046138415,
      "loss": 3.1831,
      "theoretical_loss": 4.115186550538302,
      "tokens_seen": 320077824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045605817452357074,
      "loss": 3.0413,
      "theoretical_loss": 4.115088741585305,
      "tokens_seen": 320143360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045604814443329987,
      "loss": 2.9108,
      "theoretical_loss": 4.114990958257501,
      "tokens_seen": 320208896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560381143430291,
      "loss": 3.2611,
      "theoretical_loss": 4.11489320054293,
      "tokens_seen": 320274432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560280842527583,
      "loss": 2.6443,
      "theoretical_loss": 4.114795468429648,
      "tokens_seen": 320339968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045601805416248747,
      "loss": 3.2413,
      "theoretical_loss": 4.114697761905712,
      "tokens_seen": 320405504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045600802407221665,
      "loss": 3.2062,
      "theoretical_loss": 4.1146000809591925,
      "tokens_seen": 320471040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559979939819459,
      "loss": 3.2569,
      "theoretical_loss": 4.114502425578163,
      "tokens_seen": 320536576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455987963891675,
      "loss": 3.047,
      "theoretical_loss": 4.1144047957507075,
      "tokens_seen": 320602112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045597793380140425,
      "loss": 2.9388,
      "theoretical_loss": 4.114307191464919,
      "tokens_seen": 320667648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045596790371113337,
      "loss": 3.0027,
      "theoretical_loss": 4.114209612708898,
      "tokens_seen": 320733184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559578736208626,
      "loss": 2.8668,
      "theoretical_loss": 4.1141120594707505,
      "tokens_seen": 320798720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559478435305918,
      "loss": 3.1505,
      "theoretical_loss": 4.114014531738593,
      "tokens_seen": 320864256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045593781344032097,
      "loss": 3.0058,
      "theoretical_loss": 4.113917029500549,
      "tokens_seen": 320929792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045592778335005015,
      "loss": 3.057,
      "theoretical_loss": 4.113819552744752,
      "tokens_seen": 320995328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045591775325977933,
      "loss": 2.9634,
      "theoretical_loss": 4.11372210145934,
      "tokens_seen": 321060864
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 188718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.97586989402771,
      "objective/train/theoretical_loss": 4.11362467563246,
      "objective/train/tokens_used": 341586400,
      "theoretical_loss": 4.11362467563246,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004559077231695085,
      "loss": 3.0639,
      "theoretical_loss": 4.11362467563246,
      "tokens_seen": 321126400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045589769307923775,
      "loss": 3.2388,
      "theoretical_loss": 4.1135272752522685,
      "tokens_seen": 321191936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558876629889669,
      "loss": 3.1461,
      "theoretical_loss": 4.1134299003069295,
      "tokens_seen": 321257472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558776328986961,
      "loss": 3.0283,
      "theoretical_loss": 4.113332550784613,
      "tokens_seen": 321323008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045586760280842524,
      "loss": 2.9909,
      "theoretical_loss": 4.113235226673499,
      "tokens_seen": 321388544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558575727181545,
      "loss": 3.0179,
      "theoretical_loss": 4.113137927961776,
      "tokens_seen": 321454080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045584754262788365,
      "loss": 2.8515,
      "theoretical_loss": 4.113040654637637,
      "tokens_seen": 321519616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045583751253761284,
      "loss": 2.755,
      "theoretical_loss": 4.112943406689286,
      "tokens_seen": 321585152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455827482447342,
      "loss": 2.9629,
      "theoretical_loss": 4.112846184104933,
      "tokens_seen": 321650688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045581745235707125,
      "loss": 2.9968,
      "theoretical_loss": 4.112748986872798,
      "tokens_seen": 321716224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558074222668004,
      "loss": 3.2506,
      "theoretical_loss": 4.112651814981108,
      "tokens_seen": 321781760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557973921765296,
      "loss": 3.103,
      "theoretical_loss": 4.112554668418095,
      "tokens_seen": 321847296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045578736208625874,
      "loss": 3.1271,
      "theoretical_loss": 4.112457547172005,
      "tokens_seen": 321912832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455777331995988,
      "loss": 3.2188,
      "theoretical_loss": 4.112360451231083,
      "tokens_seen": 321978368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045576730190571716,
      "loss": 3.0657,
      "theoretical_loss": 4.112263380583593,
      "tokens_seen": 322043904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045575727181544634,
      "loss": 3.1376,
      "theoretical_loss": 4.112166335217797,
      "tokens_seen": 322109440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557472417251755,
      "loss": 2.8802,
      "theoretical_loss": 4.112069315121969,
      "tokens_seen": 322174976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557372116349047,
      "loss": 3.0532,
      "theoretical_loss": 4.111972320284392,
      "tokens_seen": 322240512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557271815446339,
      "loss": 2.9202,
      "theoretical_loss": 4.111875350693354,
      "tokens_seen": 322306048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557171514543631,
      "loss": 3.1109,
      "theoretical_loss": 4.111778406337153,
      "tokens_seen": 322371584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045570712136409224,
      "loss": 2.8451,
      "theoretical_loss": 4.111681487204094,
      "tokens_seen": 322437120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556970912738215,
      "loss": 2.957,
      "theoretical_loss": 4.111584593282489,
      "tokens_seen": 322502656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556870611835507,
      "loss": 3.2971,
      "theoretical_loss": 4.111487724560659,
      "tokens_seen": 322568192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045567703109327984,
      "loss": 3.227,
      "theoretical_loss": 4.111390881026932,
      "tokens_seen": 322633728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556670010030091,
      "loss": 3.1005,
      "theoretical_loss": 4.111294062669645,
      "tokens_seen": 322699264
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 189140,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8794915676116943,
      "objective/train/theoretical_loss": 4.111197269477141,
      "objective/train/tokens_used": 343224800,
      "theoretical_loss": 4.111197269477141,
      "tokens_seen": 322764800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556569709127382,
      "loss": 2.9512,
      "theoretical_loss": 4.111197269477141,
      "tokens_seen": 322764800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045564694082246744,
      "loss": 3.0384,
      "theoretical_loss": 4.111100501437772,
      "tokens_seen": 322830336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556369107321966,
      "loss": 2.8792,
      "theoretical_loss": 4.111003758539896,
      "tokens_seen": 322895872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556268806419258,
      "loss": 3.1337,
      "theoretical_loss": 4.110907040771883,
      "tokens_seen": 322961408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455616850551655,
      "loss": 3.186,
      "theoretical_loss": 4.110810348122105,
      "tokens_seen": 323026944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045560682046138416,
      "loss": 2.9821,
      "theoretical_loss": 4.110713680578947,
      "tokens_seen": 323092480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045559679037111334,
      "loss": 3.2378,
      "theoretical_loss": 4.110617038130797,
      "tokens_seen": 323158016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555867602808426,
      "loss": 2.6982,
      "theoretical_loss": 4.110520420766054,
      "tokens_seen": 323223552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555767301905717,
      "loss": 3.3153,
      "theoretical_loss": 4.110423828473126,
      "tokens_seen": 323289088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045556670010030094,
      "loss": 3.0101,
      "theoretical_loss": 4.110327261240422,
      "tokens_seen": 323354624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045555667001003007,
      "loss": 3.1061,
      "theoretical_loss": 4.110230719056367,
      "tokens_seen": 323420160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555466399197593,
      "loss": 3.1276,
      "theoretical_loss": 4.110134201909388,
      "tokens_seen": 323485696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555366098294885,
      "loss": 3.2466,
      "theoretical_loss": 4.110037709787923,
      "tokens_seen": 323551232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045552657973921767,
      "loss": 3.347,
      "theoretical_loss": 4.109941242680414,
      "tokens_seen": 323616768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045551654964894685,
      "loss": 3.1481,
      "theoretical_loss": 4.109844800575315,
      "tokens_seen": 323682304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555065195586761,
      "loss": 2.9884,
      "theoretical_loss": 4.109748383461085,
      "tokens_seen": 323747840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554964894684052,
      "loss": 3.1188,
      "theoretical_loss": 4.10965199132619,
      "tokens_seen": 323813376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045548645937813445,
      "loss": 3.2839,
      "theoretical_loss": 4.109555624159107,
      "tokens_seen": 323878912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045547642928786357,
      "loss": 3.3218,
      "theoretical_loss": 4.109459281948317,
      "tokens_seen": 323944448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554663991975928,
      "loss": 2.8989,
      "theoretical_loss": 4.109362964682312,
      "tokens_seen": 324009984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455456369107322,
      "loss": 2.9077,
      "theoretical_loss": 4.1092666723495865,
      "tokens_seen": 324075520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045544633901705117,
      "loss": 3.2878,
      "theoretical_loss": 4.109170404938649,
      "tokens_seen": 324141056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045543630892678035,
      "loss": 3.2166,
      "theoretical_loss": 4.109074162438011,
      "tokens_seen": 324206592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045542627883650953,
      "loss": 2.8978,
      "theoretical_loss": 4.1089779448361945,
      "tokens_seen": 324272128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554162487462387,
      "loss": 3.1213,
      "theoretical_loss": 4.108881752121727,
      "tokens_seen": 324337664
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 190150,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1001152992248535,
      "objective/train/theoretical_loss": 4.108785584283144,
      "objective/train/tokens_used": 344863200,
      "theoretical_loss": 4.108785584283144,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045540621865596795,
      "loss": 2.9022,
      "theoretical_loss": 4.108785584283144,
      "tokens_seen": 324403200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553961885656971,
      "loss": 2.9865,
      "theoretical_loss": 4.108689441308989,
      "tokens_seen": 324468736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553861584754263,
      "loss": 3.112,
      "theoretical_loss": 4.108593323187815,
      "tokens_seen": 324534272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045537612838515544,
      "loss": 3.1477,
      "theoretical_loss": 4.108497229908179,
      "tokens_seen": 324599808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553660982948847,
      "loss": 3.0519,
      "theoretical_loss": 4.108401161458646,
      "tokens_seen": 324665344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045535606820461385,
      "loss": 3.0665,
      "theoretical_loss": 4.1083051178277925,
      "tokens_seen": 324730880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045534603811434304,
      "loss": 2.968,
      "theoretical_loss": 4.108209099004198,
      "tokens_seen": 324796416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553360080240722,
      "loss": 3.1983,
      "theoretical_loss": 4.108113104976453,
      "tokens_seen": 324861952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045532597793380145,
      "loss": 2.8336,
      "theoretical_loss": 4.108017135733152,
      "tokens_seen": 324927488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553159478435306,
      "loss": 3.0941,
      "theoretical_loss": 4.107921191262901,
      "tokens_seen": 324993024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553059177532598,
      "loss": 2.9691,
      "theoretical_loss": 4.10782527155431,
      "tokens_seen": 325058560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045529588766298894,
      "loss": 2.9584,
      "theoretical_loss": 4.107729376595999,
      "tokens_seen": 325124096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552858575727182,
      "loss": 3.253,
      "theoretical_loss": 4.107633506376594,
      "tokens_seen": 325189632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045527582748244736,
      "loss": 3.0396,
      "theoretical_loss": 4.10753766088473,
      "tokens_seen": 325255168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045526579739217654,
      "loss": 3.0754,
      "theoretical_loss": 4.107441840109048,
      "tokens_seen": 325320704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552557673019057,
      "loss": 3.0014,
      "theoretical_loss": 4.107346044038197,
      "tokens_seen": 325386240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552457372116349,
      "loss": 3.0876,
      "theoretical_loss": 4.107250272660834,
      "tokens_seen": 325451776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552357071213641,
      "loss": 2.9613,
      "theoretical_loss": 4.107154525965623,
      "tokens_seen": 325517312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552256770310933,
      "loss": 2.9588,
      "theoretical_loss": 4.107058803941236,
      "tokens_seen": 325582848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045521564694082244,
      "loss": 2.9034,
      "theoretical_loss": 4.106963106576352,
      "tokens_seen": 325648384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552056168505517,
      "loss": 2.8674,
      "theoretical_loss": 4.106867433859656,
      "tokens_seen": 325713920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551955867602808,
      "loss": 2.9357,
      "theoretical_loss": 4.106771785779845,
      "tokens_seen": 325779456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045518555667001004,
      "loss": 2.9561,
      "theoretical_loss": 4.106676162325618,
      "tokens_seen": 325844992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551755265797392,
      "loss": 3.0155,
      "theoretical_loss": 4.106580563485685,
      "tokens_seen": 325910528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551654964894684,
      "loss": 3.0204,
      "theoretical_loss": 4.106484989248762,
      "tokens_seen": 325976064
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 191328,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0341267585754395,
      "objective/train/theoretical_loss": 4.106389439603573,
      "objective/train/tokens_used": 346501600,
      "theoretical_loss": 4.106389439603573,
      "tokens_seen": 326041600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551554663991976,
      "loss": 3.0402,
      "theoretical_loss": 4.106389439603573,
      "tokens_seen": 326041600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551454363089268,
      "loss": 3.1653,
      "theoretical_loss": 4.10629391453885,
      "tokens_seen": 326107136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045513540621865595,
      "loss": 2.9886,
      "theoretical_loss": 4.10619841404333,
      "tokens_seen": 326172672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551253761283852,
      "loss": 3.3446,
      "theoretical_loss": 4.106102938105762,
      "tokens_seen": 326238208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551153460381143,
      "loss": 2.8962,
      "theoretical_loss": 4.106007486714896,
      "tokens_seen": 326303744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045510531594784354,
      "loss": 3.1154,
      "theoretical_loss": 4.105912059859495,
      "tokens_seen": 326369280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550952858575727,
      "loss": 3.048,
      "theoretical_loss": 4.105816657528327,
      "tokens_seen": 326434816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550852557673019,
      "loss": 2.9685,
      "theoretical_loss": 4.105721279710168,
      "tokens_seen": 326500352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550752256770311,
      "loss": 3.1754,
      "theoretical_loss": 4.105625926393801,
      "tokens_seen": 326565888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045506519558676027,
      "loss": 3.2148,
      "theoretical_loss": 4.105530597568015,
      "tokens_seen": 326631424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045505516549648945,
      "loss": 3.1651,
      "theoretical_loss": 4.1054352932216105,
      "tokens_seen": 326696960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550451354062187,
      "loss": 3.0107,
      "theoretical_loss": 4.10534001334339,
      "tokens_seen": 326762496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550351053159478,
      "loss": 3.2755,
      "theoretical_loss": 4.105244757922169,
      "tokens_seen": 326828032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045502507522567705,
      "loss": 2.8913,
      "theoretical_loss": 4.105149526946766,
      "tokens_seen": 326893568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045501504513540623,
      "loss": 2.9583,
      "theoretical_loss": 4.1050543204060075,
      "tokens_seen": 326959104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550050150451354,
      "loss": 3.187,
      "theoretical_loss": 4.10495913828873,
      "tokens_seen": 327024640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549949849548646,
      "loss": 2.9074,
      "theoretical_loss": 4.104863980583774,
      "tokens_seen": 327090176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045498495486459377,
      "loss": 3.144,
      "theoretical_loss": 4.104768847279988,
      "tokens_seen": 327155712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045497492477432295,
      "loss": 3.0299,
      "theoretical_loss": 4.104673738366232,
      "tokens_seen": 327221248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549648946840522,
      "loss": 2.9183,
      "theoretical_loss": 4.104578653831368,
      "tokens_seen": 327286784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045495486459378137,
      "loss": 2.8461,
      "theoretical_loss": 4.104483593664266,
      "tokens_seen": 327352320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045494483450351055,
      "loss": 2.9215,
      "theoretical_loss": 4.104388557853807,
      "tokens_seen": 327417856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045493480441323973,
      "loss": 3.0317,
      "theoretical_loss": 4.104293546388876,
      "tokens_seen": 327483392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549247743229689,
      "loss": 3.0664,
      "theoretical_loss": 4.104198559258364,
      "tokens_seen": 327548928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045491474423269815,
      "loss": 2.9383,
      "theoretical_loss": 4.104103596451175,
      "tokens_seen": 327614464
    },
    {
      "debugging/Self-BLEU-5": 0.2643082570774211,
      "debugging/distinct-1-grams": 0.8218199262737969,
      "debugging/distinct-2-grams": 0.9963369963369964,
      "debugging/entropy-1-grams": 4.917654182369999,
      "debugging/entropy-2-grams": 5.417185152102512,
      "debugging/length": 514.3333333333334,
      "debugging/num_segments": 3,
      "epoch": 0.1,
      "objective/train/docs_used": 191945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.493633270263672,
      "objective/train/theoretical_loss": 4.104008657956216,
      "objective/train/tokens_used": 348140000,
      "theoretical_loss": 4.104008657956216,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549047141424273,
      "loss": 3.1165,
      "theoretical_loss": 4.104008657956216,
      "tokens_seen": 327680000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548946840521565,
      "loss": 2.9939,
      "theoretical_loss": 4.1039137437624,
      "tokens_seen": 327745536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045488465396188564,
      "loss": 2.8305,
      "theoretical_loss": 4.103818853858651,
      "tokens_seen": 327811072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548746238716149,
      "loss": 3.0165,
      "theoretical_loss": 4.103723988233898,
      "tokens_seen": 327876608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045486459378134405,
      "loss": 3.1865,
      "theoretical_loss": 4.103629146877078,
      "tokens_seen": 327942144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045485456369107324,
      "loss": 3.0442,
      "theoretical_loss": 4.103534329777135,
      "tokens_seen": 328007680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548445336008024,
      "loss": 2.9175,
      "theoretical_loss": 4.103439536923021,
      "tokens_seen": 328073216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045483450351053165,
      "loss": 2.9148,
      "theoretical_loss": 4.103344768303694,
      "tokens_seen": 328138752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548244734202608,
      "loss": 3.2821,
      "theoretical_loss": 4.103250023908119,
      "tokens_seen": 328204288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045481444332999,
      "loss": 3.0728,
      "theoretical_loss": 4.103155303725271,
      "tokens_seen": 328269824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045480441323971914,
      "loss": 2.9522,
      "theoretical_loss": 4.103060607744128,
      "tokens_seen": 328335360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547943831494484,
      "loss": 3.0337,
      "theoretical_loss": 4.102965935953679,
      "tokens_seen": 328400896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045478435305917756,
      "loss": 2.7991,
      "theoretical_loss": 4.102871288342918,
      "tokens_seen": 328466432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045477432296890674,
      "loss": 2.9165,
      "theoretical_loss": 4.102776664900848,
      "tokens_seen": 328531968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547642928786359,
      "loss": 3.0264,
      "theoretical_loss": 4.1026820656164755,
      "tokens_seen": 328597504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547542627883651,
      "loss": 2.956,
      "theoretical_loss": 4.102587490478819,
      "tokens_seen": 328663040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547442326980943,
      "loss": 3.0012,
      "theoretical_loss": 4.102492939476901,
      "tokens_seen": 328728576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547342026078235,
      "loss": 3.2309,
      "theoretical_loss": 4.102398412599753,
      "tokens_seen": 328794112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045472417251755264,
      "loss": 2.9335,
      "theoretical_loss": 4.102303909836412,
      "tokens_seen": 328859648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547141424272819,
      "loss": 3.0824,
      "theoretical_loss": 4.102209431175923,
      "tokens_seen": 328925184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454704112337011,
      "loss": 3.0037,
      "theoretical_loss": 4.102114976607338,
      "tokens_seen": 328990720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045469408224674024,
      "loss": 2.9785,
      "theoretical_loss": 4.102020546119716,
      "tokens_seen": 329056256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546840521564694,
      "loss": 3.0101,
      "theoretical_loss": 4.101926139702125,
      "tokens_seen": 329121792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546740220661986,
      "loss": 3.1876,
      "theoretical_loss": 4.101831757343637,
      "tokens_seen": 329187328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546639919759278,
      "loss": 3.0554,
      "theoretical_loss": 4.1017373990333335,
      "tokens_seen": 329252864
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 193205,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.53017520904541,
      "objective/train/theoretical_loss": 4.101643064760301,
      "objective/train/tokens_used": 349778400,
      "theoretical_loss": 4.101643064760301,
      "tokens_seen": 329318400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454653961885657,
      "loss": 2.8743,
      "theoretical_loss": 4.101643064760301,
      "tokens_seen": 329318400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045464393179538615,
      "loss": 2.7826,
      "theoretical_loss": 4.101548754513636,
      "tokens_seen": 329383936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546339017051154,
      "loss": 2.8613,
      "theoretical_loss": 4.101454468282439,
      "tokens_seen": 329449472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546238716148445,
      "loss": 2.9591,
      "theoretical_loss": 4.101360206055821,
      "tokens_seen": 329515008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045461384152457374,
      "loss": 3.0603,
      "theoretical_loss": 4.101265967822897,
      "tokens_seen": 329580544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546038114343029,
      "loss": 2.8043,
      "theoretical_loss": 4.101171753572791,
      "tokens_seen": 329646080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545937813440321,
      "loss": 3.0562,
      "theoretical_loss": 4.101077563294632,
      "tokens_seen": 329711616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545837512537613,
      "loss": 3.1101,
      "theoretical_loss": 4.1009833969775595,
      "tokens_seen": 329777152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045457372116349047,
      "loss": 3.162,
      "theoretical_loss": 4.100889254610717,
      "tokens_seen": 329842688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045456369107321965,
      "loss": 2.9801,
      "theoretical_loss": 4.100795136183256,
      "tokens_seen": 329908224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545536609829489,
      "loss": 2.9315,
      "theoretical_loss": 4.100701041684336,
      "tokens_seen": 329973760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454543630892678,
      "loss": 3.097,
      "theoretical_loss": 4.100606971103122,
      "tokens_seen": 330039296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045453360080240725,
      "loss": 3.0641,
      "theoretical_loss": 4.100512924428788,
      "tokens_seen": 330104832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045452357071213643,
      "loss": 2.9796,
      "theoretical_loss": 4.100418901650513,
      "tokens_seen": 330170368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545135406218656,
      "loss": 2.9823,
      "theoretical_loss": 4.1003249027574835,
      "tokens_seen": 330235904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545035105315948,
      "loss": 3.0358,
      "theoretical_loss": 4.100230927738895,
      "tokens_seen": 330301440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045449348044132397,
      "loss": 3.0782,
      "theoretical_loss": 4.100136976583947,
      "tokens_seen": 330366976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045448345035105315,
      "loss": 3.0677,
      "theoretical_loss": 4.100043049281849,
      "tokens_seen": 330432512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544734202607824,
      "loss": 2.9041,
      "theoretical_loss": 4.099949145821815,
      "tokens_seen": 330498048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544633901705115,
      "loss": 3.1047,
      "theoretical_loss": 4.099855266193067,
      "tokens_seen": 330563584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045445336008024075,
      "loss": 3.2029,
      "theoretical_loss": 4.099761410384835,
      "tokens_seen": 330629120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544433299899699,
      "loss": 2.8177,
      "theoretical_loss": 4.099667578386355,
      "tokens_seen": 330694656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544332998996991,
      "loss": 2.8579,
      "theoretical_loss": 4.099573770186869,
      "tokens_seen": 330760192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544232698094283,
      "loss": 2.9416,
      "theoretical_loss": 4.099479985775629,
      "tokens_seen": 330825728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544132397191575,
      "loss": 2.997,
      "theoretical_loss": 4.09938622514189,
      "tokens_seen": 330891264
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 193823,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6776926517486572,
      "objective/train/theoretical_loss": 4.099292488274917,
      "objective/train/tokens_used": 351416800,
      "theoretical_loss": 4.099292488274917,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045440320962888666,
      "loss": 2.9025,
      "theoretical_loss": 4.099292488274917,
      "tokens_seen": 330956800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045439317953861584,
      "loss": 2.9272,
      "theoretical_loss": 4.0991987751639805,
      "tokens_seen": 331022336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454383149448345,
      "loss": 3.0293,
      "theoretical_loss": 4.099105085798359,
      "tokens_seen": 331087872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045437311935807425,
      "loss": 2.9882,
      "theoretical_loss": 4.099011420167338,
      "tokens_seen": 331153408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543630892678034,
      "loss": 2.9665,
      "theoretical_loss": 4.098917778260208,
      "tokens_seen": 331218944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543530591775326,
      "loss": 3.145,
      "theoretical_loss": 4.098824160066269,
      "tokens_seen": 331284480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543430290872618,
      "loss": 3.0804,
      "theoretical_loss": 4.098730565574826,
      "tokens_seen": 331350016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454332998996991,
      "loss": 2.8699,
      "theoretical_loss": 4.098636994775193,
      "tokens_seen": 331415552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045432296890672016,
      "loss": 2.8012,
      "theoretical_loss": 4.098543447656688,
      "tokens_seen": 331481088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045431293881644934,
      "loss": 3.1938,
      "theoretical_loss": 4.098449924208639,
      "tokens_seen": 331546624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543029087261785,
      "loss": 2.8317,
      "theoretical_loss": 4.098356424420379,
      "tokens_seen": 331612160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045429287863590776,
      "loss": 2.9114,
      "theoretical_loss": 4.098262948281247,
      "tokens_seen": 331677696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542828485456369,
      "loss": 2.9306,
      "theoretical_loss": 4.098169495780592,
      "tokens_seen": 331743232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542728184553661,
      "loss": 3.0543,
      "theoretical_loss": 4.098076066907769,
      "tokens_seen": 331808768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045426278836509525,
      "loss": 2.9948,
      "theoretical_loss": 4.0979826616521375,
      "tokens_seen": 331874304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542527582748245,
      "loss": 3.0341,
      "theoretical_loss": 4.0978892800030655,
      "tokens_seen": 331939840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045424272818455366,
      "loss": 2.7333,
      "theoretical_loss": 4.0977959219499285,
      "tokens_seen": 332005376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045423269809428284,
      "loss": 2.7174,
      "theoretical_loss": 4.097702587482107,
      "tokens_seen": 332070912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454222668004012,
      "loss": 2.9369,
      "theoretical_loss": 4.097609276588993,
      "tokens_seen": 332136448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542126379137412,
      "loss": 2.9834,
      "theoretical_loss": 4.097515989259978,
      "tokens_seen": 332201984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045420260782347044,
      "loss": 2.7845,
      "theoretical_loss": 4.097422725484467,
      "tokens_seen": 332267520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541925777331996,
      "loss": 2.7439,
      "theoretical_loss": 4.097329485251869,
      "tokens_seen": 332333056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541825476429288,
      "loss": 3.1478,
      "theoretical_loss": 4.097236268551598,
      "tokens_seen": 332398592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454172517552658,
      "loss": 2.7357,
      "theoretical_loss": 4.097143075373079,
      "tokens_seen": 332464128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541624874623872,
      "loss": 2.9623,
      "theoretical_loss": 4.097049905705742,
      "tokens_seen": 332529664
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 194956,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.295325756072998,
      "objective/train/theoretical_loss": 4.0969567595390215,
      "objective/train/tokens_used": 353055200,
      "theoretical_loss": 4.0969567595390215,
      "tokens_seen": 332595200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045415245737211635,
      "loss": 3.0697,
      "theoretical_loss": 4.0969567595390215,
      "tokens_seen": 332595200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541424272818456,
      "loss": 2.9866,
      "theoretical_loss": 4.096863636862363,
      "tokens_seen": 332660736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541323971915747,
      "loss": 3.0612,
      "theoretical_loss": 4.096770537665216,
      "tokens_seen": 332726272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045412236710130395,
      "loss": 2.8517,
      "theoretical_loss": 4.096677461937037,
      "tokens_seen": 332791808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541123370110331,
      "loss": 3.2886,
      "theoretical_loss": 4.09658440966729,
      "tokens_seen": 332857344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541023069207623,
      "loss": 3.0414,
      "theoretical_loss": 4.096491380845446,
      "tokens_seen": 332922880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540922768304915,
      "loss": 3.1045,
      "theoretical_loss": 4.096398375460983,
      "tokens_seen": 332988416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045408224674022067,
      "loss": 3.19,
      "theoretical_loss": 4.096305393503385,
      "tokens_seen": 333053952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045407221664994985,
      "loss": 2.9571,
      "theoretical_loss": 4.096212434962142,
      "tokens_seen": 333119488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540621865596791,
      "loss": 2.8609,
      "theoretical_loss": 4.096119499826752,
      "tokens_seen": 333185024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540521564694082,
      "loss": 3.079,
      "theoretical_loss": 4.096026588086722,
      "tokens_seen": 333250560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045404212637913745,
      "loss": 2.9375,
      "theoretical_loss": 4.09593369973156,
      "tokens_seen": 333316096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045403209628886663,
      "loss": 3.0762,
      "theoretical_loss": 4.095840834750787,
      "tokens_seen": 333381632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540220661985958,
      "loss": 3.0485,
      "theoretical_loss": 4.095747993133925,
      "tokens_seen": 333447168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454012036108325,
      "loss": 2.8056,
      "theoretical_loss": 4.095655174870508,
      "tokens_seen": 333512704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045400200601805417,
      "loss": 2.9316,
      "theoretical_loss": 4.095562379950073,
      "tokens_seen": 333578240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045399197592778335,
      "loss": 3.156,
      "theoretical_loss": 4.0954696083621664,
      "tokens_seen": 333643776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539819458375126,
      "loss": 2.8985,
      "theoretical_loss": 4.095376860096339,
      "tokens_seen": 333709312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539719157472417,
      "loss": 2.944,
      "theoretical_loss": 4.095284135142149,
      "tokens_seen": 333774848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045396188565697095,
      "loss": 2.8726,
      "theoretical_loss": 4.095191433489164,
      "tokens_seen": 333840384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539518555667001,
      "loss": 2.8167,
      "theoretical_loss": 4.095098755126954,
      "tokens_seen": 333905920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539418254764293,
      "loss": 2.859,
      "theoretical_loss": 4.095006100045097,
      "tokens_seen": 333971456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539317953861585,
      "loss": 2.9604,
      "theoretical_loss": 4.094913468233182,
      "tokens_seen": 334036992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004539217652958877,
      "loss": 2.9834,
      "theoretical_loss": 4.094820859680797,
      "tokens_seen": 334102528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045391173520561686,
      "loss": 2.9309,
      "theoretical_loss": 4.094728274377543,
      "tokens_seen": 334168064
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 196277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.767504930496216,
      "objective/train/theoretical_loss": 4.094635712313026,
      "objective/train/tokens_used": 354693600,
      "theoretical_loss": 4.094635712313026,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045390170511534604,
      "loss": 3.0447,
      "theoretical_loss": 4.094635712313026,
      "tokens_seen": 334233600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538916750250752,
      "loss": 3.0314,
      "theoretical_loss": 4.094543173476857,
      "tokens_seen": 334299136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045388164493480445,
      "loss": 3.1375,
      "theoretical_loss": 4.094450657858655,
      "tokens_seen": 334364672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538716148445336,
      "loss": 2.9004,
      "theoretical_loss": 4.094358165448046,
      "tokens_seen": 334430208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538615847542628,
      "loss": 2.936,
      "theoretical_loss": 4.094265696234663,
      "tokens_seen": 334495744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453851554663992,
      "loss": 2.9983,
      "theoretical_loss": 4.094173250208143,
      "tokens_seen": 334561280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538415245737212,
      "loss": 3.0602,
      "theoretical_loss": 4.094080827358133,
      "tokens_seen": 334626816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045383149448345036,
      "loss": 2.8065,
      "theoretical_loss": 4.093988427674285,
      "tokens_seen": 334692352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045382146439317954,
      "loss": 2.8173,
      "theoretical_loss": 4.093896051146257,
      "tokens_seen": 334757888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538114343029087,
      "loss": 2.9987,
      "theoretical_loss": 4.0938036977637156,
      "tokens_seen": 334823424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045380140421263796,
      "loss": 2.9723,
      "theoretical_loss": 4.093711367516333,
      "tokens_seen": 334888960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537913741223671,
      "loss": 3.1855,
      "theoretical_loss": 4.093619060393786,
      "tokens_seen": 334954496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537813440320963,
      "loss": 3.0542,
      "theoretical_loss": 4.0935267763857635,
      "tokens_seen": 335020032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045377131394182545,
      "loss": 3.0719,
      "theoretical_loss": 4.093434515481954,
      "tokens_seen": 335085568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537612838515547,
      "loss": 2.7748,
      "theoretical_loss": 4.093342277672058,
      "tokens_seen": 335151104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045375125376128386,
      "loss": 2.9491,
      "theoretical_loss": 4.09325006294578,
      "tokens_seen": 335216640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045374122367101304,
      "loss": 2.9779,
      "theoretical_loss": 4.093157871292833,
      "tokens_seen": 335282176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537311935807422,
      "loss": 2.8597,
      "theoretical_loss": 4.093065702702933,
      "tokens_seen": 335347712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537211634904714,
      "loss": 3.029,
      "theoretical_loss": 4.0929735571658075,
      "tokens_seen": 335413248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537111334002006,
      "loss": 3.136,
      "theoretical_loss": 4.092881434671186,
      "tokens_seen": 335478784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537011033099298,
      "loss": 3.1326,
      "theoretical_loss": 4.09278933520881,
      "tokens_seen": 335544320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045369107321965895,
      "loss": 2.7854,
      "theoretical_loss": 4.092697258768419,
      "tokens_seen": 335609856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536810431293882,
      "loss": 3.006,
      "theoretical_loss": 4.09260520533977,
      "tokens_seen": 335675392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045367101303911737,
      "loss": 3.2209,
      "theoretical_loss": 4.092513174912616,
      "tokens_seen": 335740928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045366098294884655,
      "loss": 3.0421,
      "theoretical_loss": 4.092421167476726,
      "tokens_seen": 335806464
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 196731,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1563823223114014,
      "objective/train/theoretical_loss": 4.0923291830218655,
      "objective/train/tokens_used": 356332000,
      "theoretical_loss": 4.0923291830218655,
      "tokens_seen": 335872000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045365095285857573,
      "loss": 3.0343,
      "theoretical_loss": 4.0923291830218655,
      "tokens_seen": 335872000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536409227683049,
      "loss": 3.1099,
      "theoretical_loss": 4.092237221537817,
      "tokens_seen": 335937536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536308926780341,
      "loss": 3.1611,
      "theoretical_loss": 4.0921452830143625,
      "tokens_seen": 336003072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536208625877633,
      "loss": 3.0936,
      "theoretical_loss": 4.092053367441292,
      "tokens_seen": 336068608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045361083249749245,
      "loss": 3.0695,
      "theoretical_loss": 4.091961474808404,
      "tokens_seen": 336134144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536008024072217,
      "loss": 3.1083,
      "theoretical_loss": 4.091869605105501,
      "tokens_seen": 336199680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535907723169508,
      "loss": 2.9725,
      "theoretical_loss": 4.091777758322394,
      "tokens_seen": 336265216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045358074222668005,
      "loss": 2.7863,
      "theoretical_loss": 4.091685934448899,
      "tokens_seen": 336330752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045357071213640923,
      "loss": 2.7613,
      "theoretical_loss": 4.09159413347484,
      "tokens_seen": 336396288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535606820461384,
      "loss": 2.7013,
      "theoretical_loss": 4.091502355390046,
      "tokens_seen": 336461824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535506519558676,
      "loss": 3.0298,
      "theoretical_loss": 4.091410600184353,
      "tokens_seen": 336527360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045354062186559683,
      "loss": 2.8801,
      "theoretical_loss": 4.091318867847605,
      "tokens_seen": 336592896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045353059177532596,
      "loss": 2.9488,
      "theoretical_loss": 4.091227158369649,
      "tokens_seen": 336658432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535205616850552,
      "loss": 2.8764,
      "theoretical_loss": 4.091135471740342,
      "tokens_seen": 336723968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535105315947843,
      "loss": 2.994,
      "theoretical_loss": 4.091043807949546,
      "tokens_seen": 336789504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045350050150451355,
      "loss": 2.947,
      "theoretical_loss": 4.0909521669871305,
      "tokens_seen": 336855040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045349047141424274,
      "loss": 3.1228,
      "theoretical_loss": 4.0908605488429695,
      "tokens_seen": 336920576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534804413239719,
      "loss": 2.939,
      "theoretical_loss": 4.090768953506944,
      "tokens_seen": 336986112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534704112337011,
      "loss": 2.8694,
      "theoretical_loss": 4.090677380968943,
      "tokens_seen": 337051648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534603811434303,
      "loss": 2.9003,
      "theoretical_loss": 4.09058583121886,
      "tokens_seen": 337117184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534503510531595,
      "loss": 2.9993,
      "theoretical_loss": 4.090494304246597,
      "tokens_seen": 337182720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534403209628887,
      "loss": 3.1168,
      "theoretical_loss": 4.09040280004206,
      "tokens_seen": 337248256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534302908726179,
      "loss": 2.9275,
      "theoretical_loss": 4.090311318595164,
      "tokens_seen": 337313792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045342026078234706,
      "loss": 2.941,
      "theoretical_loss": 4.090219859895829,
      "tokens_seen": 337379328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045341023069207624,
      "loss": 2.8286,
      "theoretical_loss": 4.09012842393398,
      "tokens_seen": 337444864
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 198039,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.190596342086792,
      "objective/train/theoretical_loss": 4.090037010699552,
      "objective/train/tokens_used": 357970400,
      "theoretical_loss": 4.090037010699552,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534002006018054,
      "loss": 3.0192,
      "theoretical_loss": 4.090037010699552,
      "tokens_seen": 337510400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045339017051153465,
      "loss": 2.921,
      "theoretical_loss": 4.089945620182483,
      "tokens_seen": 337575936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533801404212638,
      "loss": 3.1205,
      "theoretical_loss": 4.08985425237272,
      "tokens_seen": 337641472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453370110330993,
      "loss": 2.8657,
      "theoretical_loss": 4.089762907260214,
      "tokens_seen": 337707008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533600802407222,
      "loss": 2.9334,
      "theoretical_loss": 4.0896715848349245,
      "tokens_seen": 337772544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533500501504514,
      "loss": 2.8687,
      "theoretical_loss": 4.089580285086815,
      "tokens_seen": 337838080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045334002006018056,
      "loss": 2.9244,
      "theoretical_loss": 4.0894890080058595,
      "tokens_seen": 337903616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045332998996990974,
      "loss": 2.7939,
      "theoretical_loss": 4.089397753582034,
      "tokens_seen": 337969152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533199598796389,
      "loss": 2.9034,
      "theoretical_loss": 4.089306521805322,
      "tokens_seen": 338034688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045330992978936816,
      "loss": 2.8053,
      "theoretical_loss": 4.089215312665716,
      "tokens_seen": 338100224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532998996990973,
      "loss": 3.0222,
      "theoretical_loss": 4.0891241261532105,
      "tokens_seen": 338165760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532898696088265,
      "loss": 3.2622,
      "theoretical_loss": 4.089032962257811,
      "tokens_seen": 338231296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045327983951855565,
      "loss": 3.0372,
      "theoretical_loss": 4.088941820969525,
      "tokens_seen": 338296832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532698094282849,
      "loss": 2.8771,
      "theoretical_loss": 4.08885070227837,
      "tokens_seen": 338362368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045325977933801406,
      "loss": 2.9213,
      "theoretical_loss": 4.088759606174369,
      "tokens_seen": 338427904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045324974924774324,
      "loss": 2.877,
      "theoretical_loss": 4.088668532647548,
      "tokens_seen": 338493440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532397191574724,
      "loss": 2.9616,
      "theoretical_loss": 4.088577481687944,
      "tokens_seen": 338558976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532296890672016,
      "loss": 2.9701,
      "theoretical_loss": 4.088486453285596,
      "tokens_seen": 338624512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532196589769308,
      "loss": 2.9765,
      "theoretical_loss": 4.088395447430554,
      "tokens_seen": 338690048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045320962888666,
      "loss": 2.9125,
      "theoretical_loss": 4.088304464112872,
      "tokens_seen": 338755584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045319959879638915,
      "loss": 2.7839,
      "theoretical_loss": 4.088213503322607,
      "tokens_seen": 338821120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531895687061184,
      "loss": 2.8615,
      "theoretical_loss": 4.088122565049829,
      "tokens_seen": 338886656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045317953861584757,
      "loss": 3.0402,
      "theoretical_loss": 4.088031649284609,
      "tokens_seen": 338952192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045316950852557675,
      "loss": 3.1794,
      "theoretical_loss": 4.087940756017027,
      "tokens_seen": 339017728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045315947843530593,
      "loss": 2.8781,
      "theoretical_loss": 4.087849885237168,
      "tokens_seen": 339083264
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 198530,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.366729259490967,
      "objective/train/theoretical_loss": 4.0877590369351235,
      "objective/train/tokens_used": 359608800,
      "theoretical_loss": 4.0877590369351235,
      "tokens_seen": 339148800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531494483450351,
      "loss": 3.1109,
      "theoretical_loss": 4.0877590369351235,
      "tokens_seen": 339148800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531394182547643,
      "loss": 2.8525,
      "theoretical_loss": 4.0876682111009925,
      "tokens_seen": 339214336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531293881644935,
      "loss": 3.1438,
      "theoretical_loss": 4.087577407724877,
      "tokens_seen": 339279872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045311935807422265,
      "loss": 2.836,
      "theoretical_loss": 4.08748662679689,
      "tokens_seen": 339345408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531093279839519,
      "loss": 2.8101,
      "theoretical_loss": 4.087395868307146,
      "tokens_seen": 339410944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453099297893681,
      "loss": 3.1357,
      "theoretical_loss": 4.08730513224577,
      "tokens_seen": 339476480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045308926780341025,
      "loss": 3.14,
      "theoretical_loss": 4.08721441860289,
      "tokens_seen": 339542016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045307923771313943,
      "loss": 3.0339,
      "theoretical_loss": 4.087123727368642,
      "tokens_seen": 339607552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530692076228686,
      "loss": 3.1427,
      "theoretical_loss": 4.087033058533167,
      "tokens_seen": 339673088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530591775325978,
      "loss": 3.0171,
      "theoretical_loss": 4.086942412086614,
      "tokens_seen": 339738624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045304914744232703,
      "loss": 3.0101,
      "theoretical_loss": 4.086851788019137,
      "tokens_seen": 339804160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045303911735205616,
      "loss": 3.0004,
      "theoretical_loss": 4.086761186320897,
      "tokens_seen": 339869696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530290872617854,
      "loss": 3.0109,
      "theoretical_loss": 4.086670606982059,
      "tokens_seen": 339935232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530190571715145,
      "loss": 2.7429,
      "theoretical_loss": 4.086580049992798,
      "tokens_seen": 340000768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045300902708124375,
      "loss": 2.9896,
      "theoretical_loss": 4.086489515343292,
      "tokens_seen": 340066304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045299899699097294,
      "loss": 3.0274,
      "theoretical_loss": 4.086399003023725,
      "tokens_seen": 340131840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529889669007021,
      "loss": 3.1022,
      "theoretical_loss": 4.086308513024291,
      "tokens_seen": 340197376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529789368104313,
      "loss": 2.9777,
      "theoretical_loss": 4.086218045335187,
      "tokens_seen": 340262912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529689067201605,
      "loss": 3.1257,
      "theoretical_loss": 4.086127599946616,
      "tokens_seen": 340328448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045295887662988966,
      "loss": 3.0384,
      "theoretical_loss": 4.086037176848789,
      "tokens_seen": 340393984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529488465396189,
      "loss": 3.0784,
      "theoretical_loss": 4.085946776031922,
      "tokens_seen": 340459520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452938816449348,
      "loss": 2.9653,
      "theoretical_loss": 4.085856397486237,
      "tokens_seen": 340525056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045292878635907726,
      "loss": 2.9517,
      "theoretical_loss": 4.085766041201964,
      "tokens_seen": 340590592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529187562688064,
      "loss": 2.742,
      "theoretical_loss": 4.085675707169336,
      "tokens_seen": 340656128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529087261785356,
      "loss": 2.8849,
      "theoretical_loss": 4.085585395378595,
      "tokens_seen": 340721664
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 199795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8782248497009277,
      "objective/train/theoretical_loss": 4.085495105819987,
      "objective/train/tokens_used": 361247200,
      "theoretical_loss": 4.085495105819987,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528986960882648,
      "loss": 2.8611,
      "theoretical_loss": 4.085495105819987,
      "tokens_seen": 340787200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452888665997994,
      "loss": 3.0898,
      "theoretical_loss": 4.085404838483767,
      "tokens_seen": 340852736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045287863590772316,
      "loss": 2.9958,
      "theoretical_loss": 4.085314593360194,
      "tokens_seen": 340918272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528686058174524,
      "loss": 3.0597,
      "theoretical_loss": 4.085224370439533,
      "tokens_seen": 340983808
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528585757271815,
      "loss": 2.7746,
      "theoretical_loss": 4.085134169712055,
      "tokens_seen": 341049344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045284854563691076,
      "loss": 2.9852,
      "theoretical_loss": 4.085043991168039,
      "tokens_seen": 341114880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528385155466399,
      "loss": 3.0912,
      "theoretical_loss": 4.084953834797769,
      "tokens_seen": 341180416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528284854563691,
      "loss": 2.96,
      "theoretical_loss": 4.0848637005915345,
      "tokens_seen": 341245952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528184553660983,
      "loss": 3.2266,
      "theoretical_loss": 4.0847735885396315,
      "tokens_seen": 341311488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528084252758275,
      "loss": 2.9558,
      "theoretical_loss": 4.084683498632363,
      "tokens_seen": 341377024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045279839518555667,
      "loss": 3.0659,
      "theoretical_loss": 4.084593430860037,
      "tokens_seen": 341442560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045278836509528585,
      "loss": 2.8982,
      "theoretical_loss": 4.084503385212968,
      "tokens_seen": 341508096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045277833500501503,
      "loss": 2.7515,
      "theoretical_loss": 4.084413361681476,
      "tokens_seen": 341573632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045276830491474426,
      "loss": 3.1749,
      "theoretical_loss": 4.084323360255889,
      "tokens_seen": 341639168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527582748244734,
      "loss": 2.8382,
      "theoretical_loss": 4.0842333809265385,
      "tokens_seen": 341704704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527482447342026,
      "loss": 2.8088,
      "theoretical_loss": 4.084143423683765,
      "tokens_seen": 341770240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045273821464393175,
      "loss": 2.8347,
      "theoretical_loss": 4.0840534885179105,
      "tokens_seen": 341835776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452728184553661,
      "loss": 3.0138,
      "theoretical_loss": 4.083963575419329,
      "tokens_seen": 341901312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045271815446339017,
      "loss": 2.9717,
      "theoretical_loss": 4.083873684378376,
      "tokens_seen": 341966848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045270812437311935,
      "loss": 2.9964,
      "theoretical_loss": 4.083783815385415,
      "tokens_seen": 342032384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526980942828486,
      "loss": 3.0922,
      "theoretical_loss": 4.083693968430816,
      "tokens_seen": 342097920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045268806419257777,
      "loss": 2.9981,
      "theoretical_loss": 4.083604143504952,
      "tokens_seen": 342163456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045267803410230695,
      "loss": 3.4607,
      "theoretical_loss": 4.083514340598207,
      "tokens_seen": 342228992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045266800401203613,
      "loss": 3.0061,
      "theoretical_loss": 4.083424559700966,
      "tokens_seen": 342294528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526579739217653,
      "loss": 2.9478,
      "theoretical_loss": 4.083334800803624,
      "tokens_seen": 342360064
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 200497,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0360288619995117,
      "objective/train/theoretical_loss": 4.08324506389658,
      "objective/train/tokens_used": 362885600,
      "theoretical_loss": 4.08324506389658,
      "tokens_seen": 342425600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526479438314945,
      "loss": 2.9394,
      "theoretical_loss": 4.08324506389658,
      "tokens_seen": 342425600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526379137412237,
      "loss": 2.9863,
      "theoretical_loss": 4.083155348970238,
      "tokens_seen": 342491136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045262788365095285,
      "loss": 2.9571,
      "theoretical_loss": 4.083065656015012,
      "tokens_seen": 342556672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526178535606821,
      "loss": 3.2024,
      "theoretical_loss": 4.082975985021317,
      "tokens_seen": 342622208
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526078234704112,
      "loss": 3.1355,
      "theoretical_loss": 4.082886335979578,
      "tokens_seen": 342687744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045259779338014045,
      "loss": 2.9816,
      "theoretical_loss": 4.082796708880224,
      "tokens_seen": 342753280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045258776328986963,
      "loss": 3.1751,
      "theoretical_loss": 4.082707103713689,
      "tokens_seen": 342818816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525777331995988,
      "loss": 2.8529,
      "theoretical_loss": 4.082617520470418,
      "tokens_seen": 342884352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452567703109328,
      "loss": 3.0098,
      "theoretical_loss": 4.0825279591408545,
      "tokens_seen": 342949888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045255767301905723,
      "loss": 2.8312,
      "theoretical_loss": 4.082438419715455,
      "tokens_seen": 343015424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045254764292878636,
      "loss": 2.8357,
      "theoretical_loss": 4.082348902184677,
      "tokens_seen": 343080960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525376128385156,
      "loss": 2.9867,
      "theoretical_loss": 4.082259406538987,
      "tokens_seen": 343146496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525275827482447,
      "loss": 2.9535,
      "theoretical_loss": 4.082169932768856,
      "tokens_seen": 343212032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045251755265797395,
      "loss": 2.8995,
      "theoretical_loss": 4.082080480864761,
      "tokens_seen": 343277568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045250752256770314,
      "loss": 3.0748,
      "theoretical_loss": 4.0819910508171855,
      "tokens_seen": 343343104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524974924774323,
      "loss": 2.9816,
      "theoretical_loss": 4.08190164261662,
      "tokens_seen": 343408640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524874623871615,
      "loss": 2.9615,
      "theoretical_loss": 4.081812256253557,
      "tokens_seen": 343474176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524774322968907,
      "loss": 3.0754,
      "theoretical_loss": 4.0817228917185,
      "tokens_seen": 343539712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045246740220661986,
      "loss": 3.277,
      "theoretical_loss": 4.081633549001955,
      "tokens_seen": 343605248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524573721163491,
      "loss": 2.9695,
      "theoretical_loss": 4.081544228094436,
      "tokens_seen": 343670784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524473420260782,
      "loss": 3.0677,
      "theoretical_loss": 4.08145492898646,
      "tokens_seen": 343736320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045243731193580746,
      "loss": 3.0617,
      "theoretical_loss": 4.081365651668555,
      "tokens_seen": 343801856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524272818455366,
      "loss": 2.8337,
      "theoretical_loss": 4.081276396131249,
      "tokens_seen": 343867392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524172517552658,
      "loss": 2.9259,
      "theoretical_loss": 4.0811871623650795,
      "tokens_seen": 343932928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452407221664995,
      "loss": 3.2243,
      "theoretical_loss": 4.08109795036059,
      "tokens_seen": 343998464
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 201708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.375955820083618,
      "objective/train/theoretical_loss": 4.081008760108329,
      "objective/train/tokens_used": 364524000,
      "theoretical_loss": 4.081008760108329,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523971915747242,
      "loss": 3.0911,
      "theoretical_loss": 4.081008760108329,
      "tokens_seen": 344064000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045238716148445336,
      "loss": 2.8323,
      "theoretical_loss": 4.08091959159885,
      "tokens_seen": 344129536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523771313941826,
      "loss": 3.0438,
      "theoretical_loss": 4.0808304448227135,
      "tokens_seen": 344195072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523671013039117,
      "loss": 3.0118,
      "theoretical_loss": 4.080741319770486,
      "tokens_seen": 344260608
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045235707121364096,
      "loss": 3.1077,
      "theoretical_loss": 4.080652216432741,
      "tokens_seen": 344326144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523470411233701,
      "loss": 2.9864,
      "theoretical_loss": 4.0805631348000535,
      "tokens_seen": 344391680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523370110330993,
      "loss": 2.7933,
      "theoretical_loss": 4.080474074863011,
      "tokens_seen": 344457216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523269809428285,
      "loss": 3.0455,
      "theoretical_loss": 4.0803850366122,
      "tokens_seen": 344522752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523169508525577,
      "loss": 3.0973,
      "theoretical_loss": 4.080296020038219,
      "tokens_seen": 344588288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045230692076228687,
      "loss": 2.8627,
      "theoretical_loss": 4.080207025131667,
      "tokens_seen": 344653824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045229689067201605,
      "loss": 2.7906,
      "theoretical_loss": 4.0801180518831535,
      "tokens_seen": 344719360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045228686058174523,
      "loss": 2.7051,
      "theoretical_loss": 4.080029100283291,
      "tokens_seen": 344784896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045227683049147446,
      "loss": 3.1299,
      "theoretical_loss": 4.079940170322698,
      "tokens_seen": 344850432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522668004012036,
      "loss": 2.7122,
      "theoretical_loss": 4.079851261992,
      "tokens_seen": 344915968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522567703109328,
      "loss": 2.7792,
      "theoretical_loss": 4.0797623752818275,
      "tokens_seen": 344981504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045224674022066195,
      "loss": 3.1825,
      "theoretical_loss": 4.079673510182818,
      "tokens_seen": 345047040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522367101303912,
      "loss": 2.8207,
      "theoretical_loss": 4.079584666685612,
      "tokens_seen": 345112576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045222668004012037,
      "loss": 2.9957,
      "theoretical_loss": 4.07949584478086,
      "tokens_seen": 345178112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045221664994984955,
      "loss": 2.8706,
      "theoretical_loss": 4.079407044459216,
      "tokens_seen": 345243648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045220661985957873,
      "loss": 2.7394,
      "theoretical_loss": 4.079318265711338,
      "tokens_seen": 345309184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045219658976930797,
      "loss": 2.7548,
      "theoretical_loss": 4.0792295085278925,
      "tokens_seen": 345374720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521865596790371,
      "loss": 2.9065,
      "theoretical_loss": 4.079140772899552,
      "tokens_seen": 345440256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045217652958876633,
      "loss": 3.0047,
      "theoretical_loss": 4.079052058816993,
      "tokens_seen": 345505792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045216649949849546,
      "loss": 3.1107,
      "theoretical_loss": 4.078963366270899,
      "tokens_seen": 345571328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521564694082247,
      "loss": 2.7583,
      "theoretical_loss": 4.078874695251959,
      "tokens_seen": 345636864
    },
    {
      "epoch": 0.1,
      "objective/train/docs_used": 202216,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1073813438415527,
      "objective/train/theoretical_loss": 4.078786045750867,
      "objective/train/tokens_used": 366162400,
      "theoretical_loss": 4.078786045750867,
      "tokens_seen": 345702400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045214643931795387,
      "loss": 2.8222,
      "theoretical_loss": 4.078786045750867,
      "tokens_seen": 345702400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045213640922768305,
      "loss": 3.0401,
      "theoretical_loss": 4.078697417758325,
      "tokens_seen": 345767936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045212637913741223,
      "loss": 3.0623,
      "theoretical_loss": 4.078608811265037,
      "tokens_seen": 345833472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521163490471414,
      "loss": 2.8066,
      "theoretical_loss": 4.078520226261718,
      "tokens_seen": 345899008
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521063189568706,
      "loss": 2.8332,
      "theoretical_loss": 4.078431662739083,
      "tokens_seen": 345964544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045209628886659983,
      "loss": 3.0442,
      "theoretical_loss": 4.078343120687858,
      "tokens_seen": 346030080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045208625877632896,
      "loss": 3.0793,
      "theoretical_loss": 4.078254600098771,
      "tokens_seen": 346095616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520762286860582,
      "loss": 3.2073,
      "theoretical_loss": 4.078166100962558,
      "tokens_seen": 346161152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520661985957873,
      "loss": 2.9417,
      "theoretical_loss": 4.078077623269959,
      "tokens_seen": 346226688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045205616850551656,
      "loss": 2.9708,
      "theoretical_loss": 4.0779891670117205,
      "tokens_seen": 346292224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045204613841524574,
      "loss": 3.0795,
      "theoretical_loss": 4.077900732178595,
      "tokens_seen": 346357760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520361083249749,
      "loss": 3.1212,
      "theoretical_loss": 4.0778123187613415,
      "tokens_seen": 346423296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520260782347041,
      "loss": 3.0503,
      "theoretical_loss": 4.077723926750724,
      "tokens_seen": 346488832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045201604814443334,
      "loss": 2.8794,
      "theoretical_loss": 4.07763555613751,
      "tokens_seen": 346554368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045200601805416246,
      "loss": 3.0506,
      "theoretical_loss": 4.077547206912477,
      "tokens_seen": 346619904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519959879638917,
      "loss": 3.0215,
      "theoretical_loss": 4.077458879066406,
      "tokens_seen": 346685440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519859578736208,
      "loss": 2.9088,
      "theoretical_loss": 4.077370572590082,
      "tokens_seen": 346750976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045197592778335006,
      "loss": 2.8665,
      "theoretical_loss": 4.077282287474299,
      "tokens_seen": 346816512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045196589769307924,
      "loss": 3.084,
      "theoretical_loss": 4.077194023709854,
      "tokens_seen": 346882048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519558676028084,
      "loss": 2.9599,
      "theoretical_loss": 4.077105781287552,
      "tokens_seen": 346947584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045194583751253766,
      "loss": 2.9752,
      "theoretical_loss": 4.077017560198202,
      "tokens_seen": 347013120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519358074222668,
      "loss": 3.182,
      "theoretical_loss": 4.0769293604326196,
      "tokens_seen": 347078656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451925777331996,
      "loss": 2.9303,
      "theoretical_loss": 4.076841181981624,
      "tokens_seen": 347144192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519157472417252,
      "loss": 3.0146,
      "theoretical_loss": 4.076753024836044,
      "tokens_seen": 347209728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519057171514544,
      "loss": 3.2817,
      "theoretical_loss": 4.076664888986713,
      "tokens_seen": 347275264
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 203448,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4970381259918213,
      "objective/train/theoretical_loss": 4.076576774424465,
      "objective/train/tokens_used": 367800800,
      "theoretical_loss": 4.076576774424465,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045189568706118356,
      "loss": 2.8287,
      "theoretical_loss": 4.076576774424465,
      "tokens_seen": 347340800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518856569709128,
      "loss": 2.8508,
      "theoretical_loss": 4.0764886811401455,
      "tokens_seen": 347406336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518756268806419,
      "loss": 2.837,
      "theoretical_loss": 4.0764006091246054,
      "tokens_seen": 347471872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045186559679037116,
      "loss": 3.1222,
      "theoretical_loss": 4.076312558368698,
      "tokens_seen": 347537408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518555667001003,
      "loss": 2.8597,
      "theoretical_loss": 4.076224528863284,
      "tokens_seen": 347602944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518455366098295,
      "loss": 2.9747,
      "theoretical_loss": 4.0761365205992295,
      "tokens_seen": 347668480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518355065195587,
      "loss": 2.8766,
      "theoretical_loss": 4.076048533567406,
      "tokens_seen": 347734016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518254764292879,
      "loss": 3.0041,
      "theoretical_loss": 4.075960567758694,
      "tokens_seen": 347799552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045181544633901707,
      "loss": 2.9555,
      "theoretical_loss": 4.075872623163973,
      "tokens_seen": 347865088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045180541624874625,
      "loss": 2.7585,
      "theoretical_loss": 4.075784699774134,
      "tokens_seen": 347930624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045179538615847543,
      "loss": 2.9776,
      "theoretical_loss": 4.075696797580071,
      "tokens_seen": 347996160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045178535606820466,
      "loss": 3.0587,
      "theoretical_loss": 4.075608916572684,
      "tokens_seen": 348061696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517753259779338,
      "loss": 2.9124,
      "theoretical_loss": 4.075521056742879,
      "tokens_seen": 348127232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451765295887663,
      "loss": 3.3636,
      "theoretical_loss": 4.0754332180815664,
      "tokens_seen": 348192768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045175526579739215,
      "loss": 3.1258,
      "theoretical_loss": 4.075345400579664,
      "tokens_seen": 348258304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517452357071214,
      "loss": 2.8544,
      "theoretical_loss": 4.075257604228094,
      "tokens_seen": 348323840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045173520561685057,
      "loss": 2.979,
      "theoretical_loss": 4.075169829017785,
      "tokens_seen": 348389376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045172517552657975,
      "loss": 2.9919,
      "theoretical_loss": 4.07508207493967,
      "tokens_seen": 348454912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045171514543630893,
      "loss": 3.0188,
      "theoretical_loss": 4.074994341984689,
      "tokens_seen": 348520448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045170511534603817,
      "loss": 2.7695,
      "theoretical_loss": 4.074906630143786,
      "tokens_seen": 348585984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516950852557673,
      "loss": 3.2483,
      "theoretical_loss": 4.074818939407912,
      "tokens_seen": 348651520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045168505516549653,
      "loss": 3.1331,
      "theoretical_loss": 4.074731269768024,
      "tokens_seen": 348717056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045167502507522566,
      "loss": 2.8828,
      "theoretical_loss": 4.074643621215082,
      "tokens_seen": 348782592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516649949849549,
      "loss": 2.8352,
      "theoretical_loss": 4.074555993740055,
      "tokens_seen": 348848128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045165496489468407,
      "loss": 3.1457,
      "theoretical_loss": 4.074468387333914,
      "tokens_seen": 348913664
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 203975,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.054220199584961,
      "objective/train/theoretical_loss": 4.074380801987639,
      "objective/train/tokens_used": 369439200,
      "theoretical_loss": 4.074380801987639,
      "tokens_seen": 348979200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045164493480441325,
      "loss": 3.0341,
      "theoretical_loss": 4.074380801987639,
      "tokens_seen": 348979200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045163490471414244,
      "loss": 3.2664,
      "theoretical_loss": 4.074293237692212,
      "tokens_seen": 349044736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516248746238716,
      "loss": 3.0151,
      "theoretical_loss": 4.074205694438623,
      "tokens_seen": 349110272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516148445336008,
      "loss": 3.0995,
      "theoretical_loss": 4.074118172217869,
      "tokens_seen": 349175808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045160481444333003,
      "loss": 2.9087,
      "theoretical_loss": 4.074030671020948,
      "tokens_seen": 349241344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045159478435305916,
      "loss": 2.9197,
      "theoretical_loss": 4.073943190838868,
      "tokens_seen": 349306880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515847542627884,
      "loss": 2.9025,
      "theoretical_loss": 4.073855731662639,
      "tokens_seen": 349372416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515747241725175,
      "loss": 3.0998,
      "theoretical_loss": 4.07376829348328,
      "tokens_seen": 349437952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045156469408224676,
      "loss": 3.1268,
      "theoretical_loss": 4.073680876291811,
      "tokens_seen": 349503488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045155466399197594,
      "loss": 2.8534,
      "theoretical_loss": 4.073593480079262,
      "tokens_seen": 349569024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515446339017051,
      "loss": 2.983,
      "theoretical_loss": 4.073506104836667,
      "tokens_seen": 349634560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515346038114343,
      "loss": 3.0149,
      "theoretical_loss": 4.073418750555065,
      "tokens_seen": 349700096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045152457372116354,
      "loss": 3.0564,
      "theoretical_loss": 4.073331417225501,
      "tokens_seen": 349765632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045151454363089266,
      "loss": 2.986,
      "theoretical_loss": 4.0732441048390235,
      "tokens_seen": 349831168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515045135406219,
      "loss": 2.9123,
      "theoretical_loss": 4.07315681338669,
      "tokens_seen": 349896704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451494483450351,
      "loss": 2.8782,
      "theoretical_loss": 4.07306954285956,
      "tokens_seen": 349962240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045148445336008026,
      "loss": 2.8731,
      "theoretical_loss": 4.072982293248703,
      "tokens_seen": 350027776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045147442326980944,
      "loss": 2.9387,
      "theoretical_loss": 4.072895064545189,
      "tokens_seen": 350093312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514643931795386,
      "loss": 2.9963,
      "theoretical_loss": 4.072807856740095,
      "tokens_seen": 350158848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514543630892678,
      "loss": 3.0698,
      "theoretical_loss": 4.072720669824506,
      "tokens_seen": 350224384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451444332998997,
      "loss": 2.9614,
      "theoretical_loss": 4.072633503789511,
      "tokens_seen": 350289920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045143430290872617,
      "loss": 3.2203,
      "theoretical_loss": 4.072546358626202,
      "tokens_seen": 350355456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514242728184554,
      "loss": 3.0721,
      "theoretical_loss": 4.07245923432568,
      "tokens_seen": 350420992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045141424272818453,
      "loss": 2.6751,
      "theoretical_loss": 4.0723721308790495,
      "tokens_seen": 350486528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045140421263791376,
      "loss": 2.6701,
      "theoretical_loss": 4.072285048277421,
      "tokens_seen": 350552064
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 205216,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9308574199676514,
      "objective/train/theoretical_loss": 4.072197986511911,
      "objective/train/tokens_used": 371077600,
      "theoretical_loss": 4.072197986511911,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045139418254764294,
      "loss": 3.0665,
      "theoretical_loss": 4.072197986511911,
      "tokens_seen": 350617600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513841524573721,
      "loss": 3.2069,
      "theoretical_loss": 4.07211094557364,
      "tokens_seen": 350683136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513741223671013,
      "loss": 2.7969,
      "theoretical_loss": 4.072023925453736,
      "tokens_seen": 350748672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513640922768305,
      "loss": 3.0486,
      "theoretical_loss": 4.07193692614333,
      "tokens_seen": 350814208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045135406218655967,
      "loss": 2.8671,
      "theoretical_loss": 4.0718499476335595,
      "tokens_seen": 350879744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513440320962889,
      "loss": 2.9117,
      "theoretical_loss": 4.071762989915569,
      "tokens_seen": 350945280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045133400200601803,
      "loss": 3.066,
      "theoretical_loss": 4.071676052980506,
      "tokens_seen": 351010816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045132397191574727,
      "loss": 3.1371,
      "theoretical_loss": 4.071589136819525,
      "tokens_seen": 351076352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513139418254764,
      "loss": 2.9283,
      "theoretical_loss": 4.071502241423786,
      "tokens_seen": 351141888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045130391173520563,
      "loss": 3.0026,
      "theoretical_loss": 4.071415366784452,
      "tokens_seen": 351207424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512938816449348,
      "loss": 2.7252,
      "theoretical_loss": 4.071328512892694,
      "tokens_seen": 351272960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451283851554664,
      "loss": 2.9749,
      "theoretical_loss": 4.071241679739687,
      "tokens_seen": 351338496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045127382146439317,
      "loss": 2.8857,
      "theoretical_loss": 4.071154867316614,
      "tokens_seen": 351404032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045126379137412235,
      "loss": 3.0538,
      "theoretical_loss": 4.07106807561466,
      "tokens_seen": 351469568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045125376128385153,
      "loss": 3.0197,
      "theoretical_loss": 4.0709813046250165,
      "tokens_seen": 351535104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045124373119358077,
      "loss": 2.8479,
      "theoretical_loss": 4.070894554338881,
      "tokens_seen": 351600640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512337011033099,
      "loss": 2.9532,
      "theoretical_loss": 4.070807824747456,
      "tokens_seen": 351666176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045122367101303913,
      "loss": 2.9436,
      "theoretical_loss": 4.07072111584195,
      "tokens_seen": 351731712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512136409227683,
      "loss": 2.8975,
      "theoretical_loss": 4.070634427613575,
      "tokens_seen": 351797248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512036108324975,
      "loss": 2.8211,
      "theoretical_loss": 4.070547760053551,
      "tokens_seen": 351862784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045119358074222673,
      "loss": 2.8698,
      "theoretical_loss": 4.070461113153101,
      "tokens_seen": 351928320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045118355065195586,
      "loss": 2.8615,
      "theoretical_loss": 4.0703744869034555,
      "tokens_seen": 351993856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511735205616851,
      "loss": 2.9633,
      "theoretical_loss": 4.070287881295848,
      "tokens_seen": 352059392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511634904714143,
      "loss": 2.8897,
      "theoretical_loss": 4.070201296321519,
      "tokens_seen": 352124928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045115346038114345,
      "loss": 2.9223,
      "theoretical_loss": 4.070114731971715,
      "tokens_seen": 352190464
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 206163,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.493535280227661,
      "objective/train/theoretical_loss": 4.070028188237686,
      "objective/train/tokens_used": 372716000,
      "theoretical_loss": 4.070028188237686,
      "tokens_seen": 352256000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045114343029087264,
      "loss": 2.8772,
      "theoretical_loss": 4.070028188237686,
      "tokens_seen": 352256000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511334002006018,
      "loss": 3.067,
      "theoretical_loss": 4.069941665110687,
      "tokens_seen": 352321536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451123370110331,
      "loss": 2.9246,
      "theoretical_loss": 4.06985516258198,
      "tokens_seen": 352387072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045111334002006023,
      "loss": 3.006,
      "theoretical_loss": 4.069768680642833,
      "tokens_seen": 352452608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045110330992978936,
      "loss": 2.9814,
      "theoretical_loss": 4.069682219284518,
      "tokens_seen": 352518144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510932798395186,
      "loss": 3.105,
      "theoretical_loss": 4.069595778498311,
      "tokens_seen": 352583680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510832497492477,
      "loss": 2.8652,
      "theoretical_loss": 4.069509358275495,
      "tokens_seen": 352649216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045107321965897696,
      "loss": 2.854,
      "theoretical_loss": 4.069422958607359,
      "tokens_seen": 352714752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045106318956870614,
      "loss": 2.8667,
      "theoretical_loss": 4.069336579485196,
      "tokens_seen": 352780288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510531594784353,
      "loss": 2.9462,
      "theoretical_loss": 4.0692502209003045,
      "tokens_seen": 352845824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510431293881645,
      "loss": 2.9122,
      "theoretical_loss": 4.0691638828439896,
      "tokens_seen": 352911360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045103309929789374,
      "loss": 2.9916,
      "theoretical_loss": 4.069077565307558,
      "tokens_seen": 352976896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045102306920762286,
      "loss": 2.9217,
      "theoretical_loss": 4.068991268282327,
      "tokens_seen": 353042432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510130391173521,
      "loss": 2.6557,
      "theoretical_loss": 4.068904991759615,
      "tokens_seen": 353107968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510030090270812,
      "loss": 2.8802,
      "theoretical_loss": 4.068818735730747,
      "tokens_seen": 353173504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045099297893681046,
      "loss": 3.0742,
      "theoretical_loss": 4.068732500187053,
      "tokens_seen": 353239040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045098294884653964,
      "loss": 2.9343,
      "theoretical_loss": 4.068646285119871,
      "tokens_seen": 353304576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509729187562688,
      "loss": 2.7554,
      "theoretical_loss": 4.06856009052054,
      "tokens_seen": 353370112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450962888665998,
      "loss": 3.0227,
      "theoretical_loss": 4.068473916380406,
      "tokens_seen": 353435648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509528585757272,
      "loss": 2.8967,
      "theoretical_loss": 4.068387762690822,
      "tokens_seen": 353501184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045094282848545637,
      "loss": 2.9807,
      "theoretical_loss": 4.068301629443143,
      "tokens_seen": 353566720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509327983951856,
      "loss": 2.9679,
      "theoretical_loss": 4.068215516628733,
      "tokens_seen": 353632256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045092276830491473,
      "loss": 2.9093,
      "theoretical_loss": 4.068129424238958,
      "tokens_seen": 353697792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045091273821464396,
      "loss": 2.9453,
      "theoretical_loss": 4.06804335226519,
      "tokens_seen": 353763328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045090270812437314,
      "loss": 2.6639,
      "theoretical_loss": 4.067957300698807,
      "tokens_seen": 353828864
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 206742,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6503775119781494,
      "objective/train/theoretical_loss": 4.067871269531193,
      "objective/train/tokens_used": 374354400,
      "theoretical_loss": 4.067871269531193,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508926780341023,
      "loss": 2.8327,
      "theoretical_loss": 4.067871269531193,
      "tokens_seen": 353894400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508826479438315,
      "loss": 2.6181,
      "theoretical_loss": 4.0677852587537355,
      "tokens_seen": 353959936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508726178535607,
      "loss": 3.0139,
      "theoretical_loss": 4.067699268357828,
      "tokens_seen": 354025472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045086258776328987,
      "loss": 2.9987,
      "theoretical_loss": 4.0676132983348685,
      "tokens_seen": 354091008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508525576730191,
      "loss": 3.1759,
      "theoretical_loss": 4.067527348676262,
      "tokens_seen": 354156544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045084252758274823,
      "loss": 2.9902,
      "theoretical_loss": 4.067441419373417,
      "tokens_seen": 354222080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045083249749247747,
      "loss": 2.9154,
      "theoretical_loss": 4.067355510417749,
      "tokens_seen": 354287616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508224674022066,
      "loss": 2.7929,
      "theoretical_loss": 4.067269621800675,
      "tokens_seen": 354353152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045081243731193583,
      "loss": 2.8677,
      "theoretical_loss": 4.067183753513622,
      "tokens_seen": 354418688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450802407221665,
      "loss": 2.9784,
      "theoretical_loss": 4.067097905548019,
      "tokens_seen": 354484224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507923771313942,
      "loss": 3.0821,
      "theoretical_loss": 4.067012077895301,
      "tokens_seen": 354549760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045078234704112337,
      "loss": 2.7986,
      "theoretical_loss": 4.066926270546909,
      "tokens_seen": 354615296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045077231695085255,
      "loss": 2.842,
      "theoretical_loss": 4.066840483494287,
      "tokens_seen": 354680832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045076228686058173,
      "loss": 2.8722,
      "theoretical_loss": 4.066754716728888,
      "tokens_seen": 354746368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045075225677031097,
      "loss": 2.9914,
      "theoretical_loss": 4.066668970242167,
      "tokens_seen": 354811904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507422266800401,
      "loss": 2.8883,
      "theoretical_loss": 4.066583244025583,
      "tokens_seen": 354877440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045073219658976933,
      "loss": 2.8113,
      "theoretical_loss": 4.066497538070605,
      "tokens_seen": 354942976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507221664994985,
      "loss": 2.591,
      "theoretical_loss": 4.066411852368702,
      "tokens_seen": 355008512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507121364092277,
      "loss": 2.7762,
      "theoretical_loss": 4.066326186911353,
      "tokens_seen": 355074048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507021063189569,
      "loss": 2.7745,
      "theoretical_loss": 4.066240541690038,
      "tokens_seen": 355139584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045069207622868606,
      "loss": 2.7743,
      "theoretical_loss": 4.066154916696245,
      "tokens_seen": 355205120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045068204613841524,
      "loss": 2.9515,
      "theoretical_loss": 4.066069311921465,
      "tokens_seen": 355270656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506720160481445,
      "loss": 2.7834,
      "theoretical_loss": 4.065983727357195,
      "tokens_seen": 355336192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506619859578736,
      "loss": 2.8959,
      "theoretical_loss": 4.065898162994938,
      "tokens_seen": 355401728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045065195586760284,
      "loss": 2.8629,
      "theoretical_loss": 4.065812618826202,
      "tokens_seen": 355467264
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 207983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6451644897460938,
      "objective/train/theoretical_loss": 4.065727094842498,
      "objective/train/tokens_used": 375992800,
      "theoretical_loss": 4.065727094842498,
      "tokens_seen": 355532800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045064192577733196,
      "loss": 2.8682,
      "theoretical_loss": 4.065727094842498,
      "tokens_seen": 355532800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506318956870612,
      "loss": 2.7648,
      "theoretical_loss": 4.0656415910353445,
      "tokens_seen": 355598336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506218655967904,
      "loss": 2.9992,
      "theoretical_loss": 4.065556107396265,
      "tokens_seen": 355663872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045061183550651956,
      "loss": 3.2104,
      "theoretical_loss": 4.065470643916786,
      "tokens_seen": 355729408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045060180541624874,
      "loss": 3.0774,
      "theoretical_loss": 4.065385200588442,
      "tokens_seen": 355794944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505917753259779,
      "loss": 2.9544,
      "theoretical_loss": 4.065299777402771,
      "tokens_seen": 355860480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505817452357071,
      "loss": 2.8752,
      "theoretical_loss": 4.065214374351315,
      "tokens_seen": 355926016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045057171514543634,
      "loss": 2.7962,
      "theoretical_loss": 4.065128991425624,
      "tokens_seen": 355991552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045056168505516547,
      "loss": 2.74,
      "theoretical_loss": 4.06504362861725,
      "tokens_seen": 356057088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505516549648947,
      "loss": 2.8575,
      "theoretical_loss": 4.064958285917752,
      "tokens_seen": 356122624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505416248746239,
      "loss": 2.8417,
      "theoretical_loss": 4.064872963318696,
      "tokens_seen": 356188160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045053159478435306,
      "loss": 3.0414,
      "theoretical_loss": 4.064787660811648,
      "tokens_seen": 356253696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045052156469408224,
      "loss": 2.7343,
      "theoretical_loss": 4.064702378388183,
      "tokens_seen": 356319232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505115346038114,
      "loss": 3.0008,
      "theoretical_loss": 4.064617116039878,
      "tokens_seen": 356384768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505015045135406,
      "loss": 2.9624,
      "theoretical_loss": 4.064531873758321,
      "tokens_seen": 356450304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045049147442326984,
      "loss": 2.9619,
      "theoretical_loss": 4.064446651535098,
      "tokens_seen": 356515840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045048144433299897,
      "loss": 2.7413,
      "theoretical_loss": 4.0643614493618045,
      "tokens_seen": 356581376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504714142427282,
      "loss": 2.9723,
      "theoretical_loss": 4.0642762672300385,
      "tokens_seen": 356646912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045046138415245733,
      "loss": 2.8596,
      "theoretical_loss": 4.064191105131406,
      "tokens_seen": 356712448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045045135406218657,
      "loss": 2.7957,
      "theoretical_loss": 4.064105963057515,
      "tokens_seen": 356777984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504413239719158,
      "loss": 2.7377,
      "theoretical_loss": 4.064020840999981,
      "tokens_seen": 356843520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045043129388164493,
      "loss": 2.8931,
      "theoretical_loss": 4.063935738950423,
      "tokens_seen": 356909056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045042126379137416,
      "loss": 3.1138,
      "theoretical_loss": 4.063850656900465,
      "tokens_seen": 356974592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045041123370110334,
      "loss": 2.9516,
      "theoretical_loss": 4.063765594841737,
      "tokens_seen": 357040128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504012036108325,
      "loss": 2.9924,
      "theoretical_loss": 4.063680552765874,
      "tokens_seen": 357105664
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 208569,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.472182512283325,
      "objective/train/theoretical_loss": 4.063595530664515,
      "objective/train/tokens_used": 377631200,
      "theoretical_loss": 4.063595530664515,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503911735205617,
      "loss": 2.9434,
      "theoretical_loss": 4.063595530664515,
      "tokens_seen": 357171200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503811434302909,
      "loss": 2.7141,
      "theoretical_loss": 4.063510528529306,
      "tokens_seen": 357236736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045037111334002007,
      "loss": 2.9001,
      "theoretical_loss": 4.063425546351894,
      "tokens_seen": 357302272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503610832497493,
      "loss": 2.9435,
      "theoretical_loss": 4.063340584123936,
      "tokens_seen": 357367808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045035105315947843,
      "loss": 3.0876,
      "theoretical_loss": 4.063255641837092,
      "tokens_seen": 357433344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045034102306920767,
      "loss": 2.7518,
      "theoretical_loss": 4.063170719483025,
      "tokens_seen": 357498880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503309929789368,
      "loss": 2.7653,
      "theoretical_loss": 4.063085817053405,
      "tokens_seen": 357564416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045032096288866603,
      "loss": 2.9062,
      "theoretical_loss": 4.063000934539909,
      "tokens_seen": 357629952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503109327983952,
      "loss": 2.7442,
      "theoretical_loss": 4.062916071934215,
      "tokens_seen": 357695488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503009027081244,
      "loss": 3.0829,
      "theoretical_loss": 4.062831229228007,
      "tokens_seen": 357761024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045029087261785357,
      "loss": 2.8354,
      "theoretical_loss": 4.062746406412977,
      "tokens_seen": 357826560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045028084252758275,
      "loss": 2.9435,
      "theoretical_loss": 4.062661603480818,
      "tokens_seen": 357892096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045027081243731193,
      "loss": 2.9386,
      "theoretical_loss": 4.06257682042323,
      "tokens_seen": 357957632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045026078234704117,
      "loss": 2.8088,
      "theoretical_loss": 4.0624920572319185,
      "tokens_seen": 358023168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502507522567703,
      "loss": 2.9035,
      "theoretical_loss": 4.062407313898593,
      "tokens_seen": 358088704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045024072216649953,
      "loss": 2.8364,
      "theoretical_loss": 4.062322590414967,
      "tokens_seen": 358154240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502306920762287,
      "loss": 2.8049,
      "theoretical_loss": 4.0622378867727615,
      "tokens_seen": 358219776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502206619859579,
      "loss": 2.9569,
      "theoretical_loss": 4.062153202963701,
      "tokens_seen": 358285312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502106318956871,
      "loss": 2.8268,
      "theoretical_loss": 4.062068538979515,
      "tokens_seen": 358350848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045020060180541626,
      "loss": 2.8442,
      "theoretical_loss": 4.061983894811938,
      "tokens_seen": 358416384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045019057171514544,
      "loss": 3.1648,
      "theoretical_loss": 4.0618992704527095,
      "tokens_seen": 358481920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501805416248747,
      "loss": 2.9194,
      "theoretical_loss": 4.061814665893573,
      "tokens_seen": 358547456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501705115346038,
      "loss": 2.9655,
      "theoretical_loss": 4.06173008112628,
      "tokens_seen": 358612992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045016048144433304,
      "loss": 2.8605,
      "theoretical_loss": 4.061645516142583,
      "tokens_seen": 358678528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045015045135406216,
      "loss": 2.7901,
      "theoretical_loss": 4.0615609709342415,
      "tokens_seen": 358744064
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 209910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.162785768508911,
      "objective/train/theoretical_loss": 4.061476445493021,
      "objective/train/tokens_used": 379269600,
      "theoretical_loss": 4.061476445493021,
      "tokens_seen": 358809600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501404212637914,
      "loss": 2.6388,
      "theoretical_loss": 4.061476445493021,
      "tokens_seen": 358809600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501303911735206,
      "loss": 2.8801,
      "theoretical_loss": 4.06139193981069,
      "tokens_seen": 358875136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045012036108324976,
      "loss": 3.0025,
      "theoretical_loss": 4.061307453879023,
      "tokens_seen": 358940672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045011033099297894,
      "loss": 2.7736,
      "theoretical_loss": 4.061222987689797,
      "tokens_seen": 359006208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501003009027081,
      "loss": 2.8362,
      "theoretical_loss": 4.061138541234799,
      "tokens_seen": 359071744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500902708124373,
      "loss": 2.6216,
      "theoretical_loss": 4.061054114505816,
      "tokens_seen": 359137280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045008024072216654,
      "loss": 3.0416,
      "theoretical_loss": 4.060969707494642,
      "tokens_seen": 359202816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045007021063189567,
      "loss": 2.8598,
      "theoretical_loss": 4.0608853201930755,
      "tokens_seen": 359268352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500601805416249,
      "loss": 3.0376,
      "theoretical_loss": 4.0608009525929205,
      "tokens_seen": 359333888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500501504513541,
      "loss": 2.9702,
      "theoretical_loss": 4.060716604685986,
      "tokens_seen": 359399424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045004012036108326,
      "loss": 2.911,
      "theoretical_loss": 4.0606322764640845,
      "tokens_seen": 359464960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045003009027081244,
      "loss": 2.6158,
      "theoretical_loss": 4.060547967919034,
      "tokens_seen": 359530496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500200601805416,
      "loss": 2.8483,
      "theoretical_loss": 4.06046367904266,
      "tokens_seen": 359596032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500100300902708,
      "loss": 2.8765,
      "theoretical_loss": 4.060379409826788,
      "tokens_seen": 359661568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045000000000000004,
      "loss": 2.9966,
      "theoretical_loss": 4.0602951602632515,
      "tokens_seen": 359727104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044998996990972917,
      "loss": 2.9513,
      "theoretical_loss": 4.060210930343889,
      "tokens_seen": 359792640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499799398194584,
      "loss": 3.0588,
      "theoretical_loss": 4.060126720060543,
      "tokens_seen": 359858176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044996990972918753,
      "loss": 2.4906,
      "theoretical_loss": 4.060042529405061,
      "tokens_seen": 359923712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044995987963891677,
      "loss": 2.9925,
      "theoretical_loss": 4.059958358369295,
      "tokens_seen": 359989248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044994984954864595,
      "loss": 2.7972,
      "theoretical_loss": 4.0598742069451035,
      "tokens_seen": 360054784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044993981945837513,
      "loss": 3.0398,
      "theoretical_loss": 4.059790075124348,
      "tokens_seen": 360120320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499297893681043,
      "loss": 2.7946,
      "theoretical_loss": 4.059705962898896,
      "tokens_seen": 360185856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044991975927783355,
      "loss": 2.9623,
      "theoretical_loss": 4.059621870260618,
      "tokens_seen": 360251392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044990972918756267,
      "loss": 2.9151,
      "theoretical_loss": 4.059537797201392,
      "tokens_seen": 360316928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498996990972919,
      "loss": 2.7803,
      "theoretical_loss": 4.059453743713099,
      "tokens_seen": 360382464
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 210586,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.735008478164673,
      "objective/train/theoretical_loss": 4.059369709787625,
      "objective/train/tokens_used": 380908000,
      "theoretical_loss": 4.059369709787625,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044988966900702103,
      "loss": 2.9449,
      "theoretical_loss": 4.059369709787625,
      "tokens_seen": 360448000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044987963891675027,
      "loss": 2.8092,
      "theoretical_loss": 4.059285695416863,
      "tokens_seen": 360513536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044986960882647945,
      "loss": 2.7777,
      "theoretical_loss": 4.059201700592707,
      "tokens_seen": 360579072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044985957873620863,
      "loss": 3.0467,
      "theoretical_loss": 4.05911772530706,
      "tokens_seen": 360644608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498495486459378,
      "loss": 3.1316,
      "theoretical_loss": 4.059033769551827,
      "tokens_seen": 360710144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449839518555667,
      "loss": 2.8016,
      "theoretical_loss": 4.058949833318917,
      "tokens_seen": 360775680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498294884653962,
      "loss": 3.0944,
      "theoretical_loss": 4.058865916600248,
      "tokens_seen": 360841216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498194583751254,
      "loss": 2.8891,
      "theoretical_loss": 4.058782019387738,
      "tokens_seen": 360906752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044980942828485454,
      "loss": 2.8118,
      "theoretical_loss": 4.058698141673313,
      "tokens_seen": 360972288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044979939819458377,
      "loss": 2.732,
      "theoretical_loss": 4.058614283448903,
      "tokens_seen": 361037824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497893681043129,
      "loss": 2.844,
      "theoretical_loss": 4.0585304447064425,
      "tokens_seen": 361103360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044977933801404213,
      "loss": 2.8023,
      "theoretical_loss": 4.05844662543787,
      "tokens_seen": 361168896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497693079237713,
      "loss": 2.8618,
      "theoretical_loss": 4.058362825635132,
      "tokens_seen": 361234432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497592778335005,
      "loss": 2.9652,
      "theoretical_loss": 4.058279045290176,
      "tokens_seen": 361299968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497492477432297,
      "loss": 2.7903,
      "theoretical_loss": 4.058195284394955,
      "tokens_seen": 361365504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497392176529589,
      "loss": 2.995,
      "theoretical_loss": 4.058111542941429,
      "tokens_seen": 361431040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044972918756268804,
      "loss": 2.835,
      "theoretical_loss": 4.05802782092156,
      "tokens_seen": 361496576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497191574724173,
      "loss": 2.7238,
      "theoretical_loss": 4.057944118327319,
      "tokens_seen": 361562112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497091273821464,
      "loss": 3.148,
      "theoretical_loss": 4.057860435150675,
      "tokens_seen": 361627648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044969909729187564,
      "loss": 2.6922,
      "theoretical_loss": 4.057776771383609,
      "tokens_seen": 361693184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496890672016049,
      "loss": 2.869,
      "theoretical_loss": 4.057693127018101,
      "tokens_seen": 361758720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449679037111334,
      "loss": 3.0336,
      "theoretical_loss": 4.057609502046141,
      "tokens_seen": 361824256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044966900702106324,
      "loss": 3.0169,
      "theoretical_loss": 4.057525896459719,
      "tokens_seen": 361889792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044965897693079236,
      "loss": 2.8861,
      "theoretical_loss": 4.057442310250831,
      "tokens_seen": 361955328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496489468405216,
      "loss": 2.9571,
      "theoretical_loss": 4.0573587434114815,
      "tokens_seen": 362020864
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 211693,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5061984062194824,
      "objective/train/theoretical_loss": 4.057275195933674,
      "objective/train/tokens_used": 382546400,
      "theoretical_loss": 4.057275195933674,
      "tokens_seen": 362086400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496389167502508,
      "loss": 3.2087,
      "theoretical_loss": 4.057275195933674,
      "tokens_seen": 362086400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044962888665997996,
      "loss": 2.835,
      "theoretical_loss": 4.057191667809421,
      "tokens_seen": 362151936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044961885656970914,
      "loss": 2.8273,
      "theoretical_loss": 4.057108159030737,
      "tokens_seen": 362217472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496088264794383,
      "loss": 2.9131,
      "theoretical_loss": 4.057024669589644,
      "tokens_seen": 362283008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495987963891675,
      "loss": 2.7972,
      "theoretical_loss": 4.056941199478166,
      "tokens_seen": 362348544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044958876629889674,
      "loss": 3.1665,
      "theoretical_loss": 4.0568577486883335,
      "tokens_seen": 362414080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044957873620862587,
      "loss": 2.9409,
      "theoretical_loss": 4.056774317212181,
      "tokens_seen": 362479616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495687061183551,
      "loss": 2.6666,
      "theoretical_loss": 4.056690905041748,
      "tokens_seen": 362545152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495586760280843,
      "loss": 2.8504,
      "theoretical_loss": 4.056607512169078,
      "tokens_seen": 362610688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044954864593781346,
      "loss": 2.8855,
      "theoretical_loss": 4.056524138586221,
      "tokens_seen": 362676224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044953861584754264,
      "loss": 2.654,
      "theoretical_loss": 4.056440784285231,
      "tokens_seen": 362741760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495285857572718,
      "loss": 2.9835,
      "theoretical_loss": 4.056357449258163,
      "tokens_seen": 362807296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449518555667001,
      "loss": 2.9323,
      "theoretical_loss": 4.056274133497084,
      "tokens_seen": 362872832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044950852557673024,
      "loss": 2.8978,
      "theoretical_loss": 4.056190836994059,
      "tokens_seen": 362938368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044949849548645937,
      "loss": 2.6167,
      "theoretical_loss": 4.056107559741161,
      "tokens_seen": 363003904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494884653961886,
      "loss": 2.942,
      "theoretical_loss": 4.056024301730467,
      "tokens_seen": 363069440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044947843530591773,
      "loss": 2.9565,
      "theoretical_loss": 4.05594106295406,
      "tokens_seen": 363134976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044946840521564697,
      "loss": 2.7726,
      "theoretical_loss": 4.055857843404025,
      "tokens_seen": 363200512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044945837512537615,
      "loss": 3.0216,
      "theoretical_loss": 4.0557746430724535,
      "tokens_seen": 363266048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044944834503510533,
      "loss": 3.1954,
      "theoretical_loss": 4.05569146195144,
      "tokens_seen": 363331584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494383149448345,
      "loss": 2.9553,
      "theoretical_loss": 4.055608300033088,
      "tokens_seen": 363397120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044942828485456375,
      "loss": 3.024,
      "theoretical_loss": 4.0555251573095,
      "tokens_seen": 363462656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044941825476429287,
      "loss": 2.7395,
      "theoretical_loss": 4.0554420337727874,
      "tokens_seen": 363528192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494082246740221,
      "loss": 2.8736,
      "theoretical_loss": 4.055358929415063,
      "tokens_seen": 363593728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044939819458375123,
      "loss": 2.8492,
      "theoretical_loss": 4.055275844228448,
      "tokens_seen": 363659264
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 212305,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9765894412994385,
      "objective/train/theoretical_loss": 4.055192778205064,
      "objective/train/tokens_used": 384184800,
      "theoretical_loss": 4.055192778205064,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044938816449348047,
      "loss": 3.0477,
      "theoretical_loss": 4.055192778205064,
      "tokens_seen": 363724800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044937813440320965,
      "loss": 2.7255,
      "theoretical_loss": 4.0551097313370414,
      "tokens_seen": 363790336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044936810431293883,
      "loss": 2.974,
      "theoretical_loss": 4.055026703616512,
      "tokens_seen": 363855872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449358074222668,
      "loss": 2.9851,
      "theoretical_loss": 4.054943695035614,
      "tokens_seen": 363921408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493480441323972,
      "loss": 3.114,
      "theoretical_loss": 4.054860705586489,
      "tokens_seen": 363986944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493380140421264,
      "loss": 2.9801,
      "theoretical_loss": 4.054777735261286,
      "tokens_seen": 364052480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493279839518556,
      "loss": 2.8191,
      "theoretical_loss": 4.054694784052155,
      "tokens_seen": 364118016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044931795386158474,
      "loss": 2.8661,
      "theoretical_loss": 4.054611851951253,
      "tokens_seen": 364183552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044930792377131397,
      "loss": 2.7798,
      "theoretical_loss": 4.054528938950741,
      "tokens_seen": 364249088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492978936810431,
      "loss": 2.9318,
      "theoretical_loss": 4.054446045042783,
      "tokens_seen": 364314624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044928786359077234,
      "loss": 3.0197,
      "theoretical_loss": 4.054363170219551,
      "tokens_seen": 364380160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492778335005015,
      "loss": 2.8279,
      "theoretical_loss": 4.054280314473219,
      "tokens_seen": 364445696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492678034102307,
      "loss": 2.8172,
      "theoretical_loss": 4.054197477795967,
      "tokens_seen": 364511232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492577733199599,
      "loss": 3.0234,
      "theoretical_loss": 4.0541146601799785,
      "tokens_seen": 364576768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492477432296891,
      "loss": 2.675,
      "theoretical_loss": 4.0540318616174424,
      "tokens_seen": 364642304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044923771313941824,
      "loss": 2.6453,
      "theoretical_loss": 4.0539490821005515,
      "tokens_seen": 364707840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492276830491475,
      "loss": 2.7324,
      "theoretical_loss": 4.053866321621504,
      "tokens_seen": 364773376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492176529588766,
      "loss": 2.8083,
      "theoretical_loss": 4.053783580172501,
      "tokens_seen": 364838912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044920762286860584,
      "loss": 2.8838,
      "theoretical_loss": 4.053700857745753,
      "tokens_seen": 364904448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449197592778335,
      "loss": 2.492,
      "theoretical_loss": 4.053618154333467,
      "tokens_seen": 364969984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491875626880642,
      "loss": 2.869,
      "theoretical_loss": 4.053535469927864,
      "tokens_seen": 365035520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491775325977934,
      "loss": 2.8246,
      "theoretical_loss": 4.05345280452116,
      "tokens_seen": 365101056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044916750250752256,
      "loss": 2.7243,
      "theoretical_loss": 4.053370158105584,
      "tokens_seen": 365166592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044915747241725174,
      "loss": 2.5572,
      "theoretical_loss": 4.053287530673364,
      "tokens_seen": 365232128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449147442326981,
      "loss": 2.9422,
      "theoretical_loss": 4.053204922216736,
      "tokens_seen": 365297664
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 213520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0496366024017334,
      "objective/train/theoretical_loss": 4.053122332727937,
      "objective/train/tokens_used": 385823200,
      "theoretical_loss": 4.053122332727937,
      "tokens_seen": 365363200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491374122367101,
      "loss": 2.9336,
      "theoretical_loss": 4.053122332727937,
      "tokens_seen": 365363200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044912738214643934,
      "loss": 2.6247,
      "theoretical_loss": 4.053039762199212,
      "tokens_seen": 365428736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044911735205616847,
      "loss": 2.6713,
      "theoretical_loss": 4.05295721062281,
      "tokens_seen": 365494272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491073219658977,
      "loss": 3.0444,
      "theoretical_loss": 4.0528746779909826,
      "tokens_seen": 365559808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490972918756269,
      "loss": 2.8455,
      "theoretical_loss": 4.052792164295987,
      "tokens_seen": 365625344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044908726178535607,
      "loss": 2.7943,
      "theoretical_loss": 4.052709669530085,
      "tokens_seen": 365690880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044907723169508525,
      "loss": 2.8501,
      "theoretical_loss": 4.052627193685544,
      "tokens_seen": 365756416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490672016048145,
      "loss": 2.8558,
      "theoretical_loss": 4.052544736754634,
      "tokens_seen": 365821952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490571715145436,
      "loss": 2.9659,
      "theoretical_loss": 4.0524622987296315,
      "tokens_seen": 365887488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044904714142427284,
      "loss": 2.8449,
      "theoretical_loss": 4.052379879602816,
      "tokens_seen": 365953024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044903711133400197,
      "loss": 2.8962,
      "theoretical_loss": 4.052297479366471,
      "tokens_seen": 366018560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490270812437312,
      "loss": 2.6586,
      "theoretical_loss": 4.052215098012887,
      "tokens_seen": 366084096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490170511534604,
      "loss": 2.6115,
      "theoretical_loss": 4.052132735534358,
      "tokens_seen": 366149632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044900702106318957,
      "loss": 2.5501,
      "theoretical_loss": 4.0520503919231805,
      "tokens_seen": 366215168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044899699097291875,
      "loss": 2.7229,
      "theoretical_loss": 4.051968067171659,
      "tokens_seen": 366280704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044898696088264793,
      "loss": 2.66,
      "theoretical_loss": 4.051885761272098,
      "tokens_seen": 366346240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489769307923771,
      "loss": 2.8702,
      "theoretical_loss": 4.051803474216813,
      "tokens_seen": 366411776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044896690070210635,
      "loss": 3.0329,
      "theoretical_loss": 4.051721205998117,
      "tokens_seen": 366477312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489568706118355,
      "loss": 3.0362,
      "theoretical_loss": 4.051638956608333,
      "tokens_seen": 366542848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489468405215647,
      "loss": 3.0241,
      "theoretical_loss": 4.0515567260397845,
      "tokens_seen": 366608384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044893681043129395,
      "loss": 3.1542,
      "theoretical_loss": 4.051474514284801,
      "tokens_seen": 366673920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044892678034102307,
      "loss": 2.9845,
      "theoretical_loss": 4.051392321335719,
      "tokens_seen": 366739456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004489167502507523,
      "loss": 2.8097,
      "theoretical_loss": 4.051310147184875,
      "tokens_seen": 366804992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044890672016048143,
      "loss": 2.7092,
      "theoretical_loss": 4.0512279918246135,
      "tokens_seen": 366870528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044889669007021067,
      "loss": 2.7877,
      "theoretical_loss": 4.051145855247281,
      "tokens_seen": 366936064
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 214095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.887453317642212,
      "objective/train/theoretical_loss": 4.051063737445231,
      "objective/train/tokens_used": 387461600,
      "theoretical_loss": 4.051063737445231,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044888665997993985,
      "loss": 3.0162,
      "theoretical_loss": 4.051063737445231,
      "tokens_seen": 367001600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044887662988966903,
      "loss": 2.7889,
      "theoretical_loss": 4.0509816384108195,
      "tokens_seen": 367067136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488665997993982,
      "loss": 2.966,
      "theoretical_loss": 4.050899558136408,
      "tokens_seen": 367132672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488565697091274,
      "loss": 2.8294,
      "theoretical_loss": 4.050817496614361,
      "tokens_seen": 367198208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488465396188566,
      "loss": 2.916,
      "theoretical_loss": 4.050735453837049,
      "tokens_seen": 367263744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488365095285858,
      "loss": 2.8996,
      "theoretical_loss": 4.0506534297968475,
      "tokens_seen": 367329280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044882647943831494,
      "loss": 2.8554,
      "theoretical_loss": 4.050571424486135,
      "tokens_seen": 367394816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488164493480442,
      "loss": 2.7958,
      "theoretical_loss": 4.050489437897295,
      "tokens_seen": 367460352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488064192577733,
      "loss": 2.9325,
      "theoretical_loss": 4.050407470022715,
      "tokens_seen": 367525888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044879638916750254,
      "loss": 2.9501,
      "theoretical_loss": 4.050325520854788,
      "tokens_seen": 367591424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487863590772317,
      "loss": 2.6799,
      "theoretical_loss": 4.050243590385909,
      "tokens_seen": 367656960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487763289869609,
      "loss": 2.9033,
      "theoretical_loss": 4.0501616786084815,
      "tokens_seen": 367722496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487662988966901,
      "loss": 2.8158,
      "theoretical_loss": 4.05007978551491,
      "tokens_seen": 367788032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487562688064193,
      "loss": 2.8058,
      "theoretical_loss": 4.049997911097606,
      "tokens_seen": 367853568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044874623871614844,
      "loss": 2.7876,
      "theoretical_loss": 4.049916055348982,
      "tokens_seen": 367919104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487362086258777,
      "loss": 3.0632,
      "theoretical_loss": 4.049834218261458,
      "tokens_seen": 367984640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487261785356068,
      "loss": 3.0815,
      "theoretical_loss": 4.049752399827458,
      "tokens_seen": 368050176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044871614844533604,
      "loss": 2.7292,
      "theoretical_loss": 4.049670600039409,
      "tokens_seen": 368115712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487061183550652,
      "loss": 2.8603,
      "theoretical_loss": 4.049588818889744,
      "tokens_seen": 368181248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486960882647944,
      "loss": 3.1734,
      "theoretical_loss": 4.049507056370899,
      "tokens_seen": 368246784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486860581745236,
      "loss": 3.0132,
      "theoretical_loss": 4.049425312475315,
      "tokens_seen": 368312320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044867602808425276,
      "loss": 2.9497,
      "theoretical_loss": 4.049343587195439,
      "tokens_seen": 368377856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044866599799398194,
      "loss": 2.9889,
      "theoretical_loss": 4.049261880523719,
      "tokens_seen": 368443392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486559679037112,
      "loss": 2.7858,
      "theoretical_loss": 4.04918019245261,
      "tokens_seen": 368508928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486459378134403,
      "loss": 2.7187,
      "theoretical_loss": 4.049098522974571,
      "tokens_seen": 368574464
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 215091,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5288538932800293,
      "objective/train/theoretical_loss": 4.049016872082065,
      "objective/train/tokens_used": 389100000,
      "theoretical_loss": 4.049016872082065,
      "tokens_seen": 368640000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044863590772316954,
      "loss": 2.8003,
      "theoretical_loss": 4.049016872082065,
      "tokens_seen": 368640000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044862587763289867,
      "loss": 3.0327,
      "theoretical_loss": 4.04893523976756,
      "tokens_seen": 368705536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486158475426279,
      "loss": 2.9907,
      "theoretical_loss": 4.0488536260235275,
      "tokens_seen": 368771072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486058174523571,
      "loss": 2.9698,
      "theoretical_loss": 4.048772030842443,
      "tokens_seen": 368836608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044859578736208627,
      "loss": 3.0078,
      "theoretical_loss": 4.048690454216788,
      "tokens_seen": 368902144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044858575727181545,
      "loss": 2.8499,
      "theoretical_loss": 4.048608896139048,
      "tokens_seen": 368967680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485757271815447,
      "loss": 3.1026,
      "theoretical_loss": 4.048527356601712,
      "tokens_seen": 369033216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485656970912738,
      "loss": 2.9261,
      "theoretical_loss": 4.048445835597272,
      "tokens_seen": 369098752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044855566700100304,
      "loss": 2.9621,
      "theoretical_loss": 4.04836433311823,
      "tokens_seen": 369164288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044854563691073217,
      "loss": 2.8368,
      "theoretical_loss": 4.0482828491570855,
      "tokens_seen": 369229824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485356068204614,
      "loss": 2.7932,
      "theoretical_loss": 4.048201383706347,
      "tokens_seen": 369295360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485255767301906,
      "loss": 2.7461,
      "theoretical_loss": 4.048119936758525,
      "tokens_seen": 369360896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044851554663991977,
      "loss": 2.9926,
      "theoretical_loss": 4.048038508306135,
      "tokens_seen": 369426432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044850551654964895,
      "loss": 2.8286,
      "theoretical_loss": 4.047957098341698,
      "tokens_seen": 369491968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044849548645937813,
      "loss": 2.9703,
      "theoretical_loss": 4.047875706857737,
      "tokens_seen": 369557504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484854563691073,
      "loss": 2.7894,
      "theoretical_loss": 4.047794333846781,
      "tokens_seen": 369623040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044847542627883655,
      "loss": 2.9906,
      "theoretical_loss": 4.047712979301364,
      "tokens_seen": 369688576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484653961885657,
      "loss": 2.8056,
      "theoretical_loss": 4.047631643214023,
      "tokens_seen": 369754112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484553660982949,
      "loss": 2.4839,
      "theoretical_loss": 4.0475503255772995,
      "tokens_seen": 369819648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044844533600802404,
      "loss": 2.7831,
      "theoretical_loss": 4.047469026383739,
      "tokens_seen": 369885184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044843530591775327,
      "loss": 2.9546,
      "theoretical_loss": 4.047387745625892,
      "tokens_seen": 369950720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044842527582748245,
      "loss": 2.7135,
      "theoretical_loss": 4.047306483296315,
      "tokens_seen": 370016256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044841524573721163,
      "loss": 2.861,
      "theoretical_loss": 4.047225239387565,
      "tokens_seen": 370081792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484052156469408,
      "loss": 2.876,
      "theoretical_loss": 4.047144013892206,
      "tokens_seen": 370147328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044839518555667005,
      "loss": 2.8395,
      "theoretical_loss": 4.047062806802805,
      "tokens_seen": 370212864
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 215564,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0313522815704346,
      "objective/train/theoretical_loss": 4.046981618111936,
      "objective/train/tokens_used": 390738400,
      "theoretical_loss": 4.046981618111936,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483851554663992,
      "loss": 2.9527,
      "theoretical_loss": 4.046981618111936,
      "tokens_seen": 370278400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483751253761284,
      "loss": 2.9376,
      "theoretical_loss": 4.046900447812173,
      "tokens_seen": 370343936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044836509528585754,
      "loss": 2.9443,
      "theoretical_loss": 4.046819295896098,
      "tokens_seen": 370409472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483550651955868,
      "loss": 3.0019,
      "theoretical_loss": 4.046738162356295,
      "tokens_seen": 370475008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044834503510531596,
      "loss": 2.8873,
      "theoretical_loss": 4.046657047185354,
      "tokens_seen": 370540544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044833500501504514,
      "loss": 3.11,
      "theoretical_loss": 4.0465759503758685,
      "tokens_seen": 370606080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483249749247743,
      "loss": 2.805,
      "theoretical_loss": 4.046494871920436,
      "tokens_seen": 370671616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483149448345035,
      "loss": 2.853,
      "theoretical_loss": 4.046413811811659,
      "tokens_seen": 370737152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483049147442327,
      "loss": 3.0758,
      "theoretical_loss": 4.0463327700421425,
      "tokens_seen": 370802688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482948846539619,
      "loss": 2.7516,
      "theoretical_loss": 4.046251746604499,
      "tokens_seen": 370868224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044828485456369104,
      "loss": 2.9978,
      "theoretical_loss": 4.046170741491343,
      "tokens_seen": 370933760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482748244734203,
      "loss": 2.7836,
      "theoretical_loss": 4.046089754695292,
      "tokens_seen": 370999296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482647943831494,
      "loss": 2.7176,
      "theoretical_loss": 4.046008786208972,
      "tokens_seen": 371064832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044825476429287864,
      "loss": 2.9292,
      "theoretical_loss": 4.04592783602501,
      "tokens_seen": 371130368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482447342026078,
      "loss": 2.9083,
      "theoretical_loss": 4.045846904136037,
      "tokens_seen": 371195904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448234704112337,
      "loss": 2.6722,
      "theoretical_loss": 4.04576599053469,
      "tokens_seen": 371261440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482246740220662,
      "loss": 3.207,
      "theoretical_loss": 4.045685095213611,
      "tokens_seen": 371326976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482146439317954,
      "loss": 2.7582,
      "theoretical_loss": 4.045604218165442,
      "tokens_seen": 371392512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044820461384152455,
      "loss": 2.6807,
      "theoretical_loss": 4.045523359382835,
      "tokens_seen": 371458048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481945837512538,
      "loss": 3.2028,
      "theoretical_loss": 4.045442518858441,
      "tokens_seen": 371523584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044818455366098296,
      "loss": 2.8334,
      "theoretical_loss": 4.045361696584919,
      "tokens_seen": 371589120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044817452357071214,
      "loss": 2.8413,
      "theoretical_loss": 4.04528089255493,
      "tokens_seen": 371654656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481644934804414,
      "loss": 3.0689,
      "theoretical_loss": 4.045200106761142,
      "tokens_seen": 371720192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481544633901705,
      "loss": 3.1544,
      "theoretical_loss": 4.045119339196222,
      "tokens_seen": 371785728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044814443329989974,
      "loss": 2.6545,
      "theoretical_loss": 4.045038589852847,
      "tokens_seen": 371851264
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 216610,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.039722442626953,
      "objective/train/theoretical_loss": 4.044957858723696,
      "objective/train/tokens_used": 392376800,
      "theoretical_loss": 4.044957858723696,
      "tokens_seen": 371916800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044813440320962887,
      "loss": 2.9195,
      "theoretical_loss": 4.044957858723696,
      "tokens_seen": 371916800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481243731193581,
      "loss": 2.9107,
      "theoretical_loss": 4.0448771458014505,
      "tokens_seen": 371982336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481143430290873,
      "loss": 2.8144,
      "theoretical_loss": 4.044796451078799,
      "tokens_seen": 372047872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044810431293881647,
      "loss": 2.8264,
      "theoretical_loss": 4.044715774548433,
      "tokens_seen": 372113408
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044809428284854565,
      "loss": 2.7547,
      "theoretical_loss": 4.044635116203047,
      "tokens_seen": 372178944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480842527582749,
      "loss": 2.8847,
      "theoretical_loss": 4.0445544760353425,
      "tokens_seen": 372244480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448074222668004,
      "loss": 2.7869,
      "theoretical_loss": 4.044473854038022,
      "tokens_seen": 372310016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044806419257773324,
      "loss": 2.9045,
      "theoretical_loss": 4.044393250203795,
      "tokens_seen": 372375552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044805416248746237,
      "loss": 3.1393,
      "theoretical_loss": 4.044312664525374,
      "tokens_seen": 372441088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480441323971916,
      "loss": 2.9725,
      "theoretical_loss": 4.044232096995476,
      "tokens_seen": 372506624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480341023069208,
      "loss": 2.984,
      "theoretical_loss": 4.044151547606821,
      "tokens_seen": 372572160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044802407221664997,
      "loss": 2.9168,
      "theoretical_loss": 4.0440710163521345,
      "tokens_seen": 372637696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044801404212637915,
      "loss": 2.9977,
      "theoretical_loss": 4.043990503224146,
      "tokens_seen": 372703232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044800401203610833,
      "loss": 3.0303,
      "theoretical_loss": 4.04391000821559,
      "tokens_seen": 372768768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479939819458375,
      "loss": 3.2721,
      "theoretical_loss": 4.0438295313192025,
      "tokens_seen": 372834304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044798395185556675,
      "loss": 2.6246,
      "theoretical_loss": 4.043749072527727,
      "tokens_seen": 372899840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479739217652959,
      "loss": 2.7876,
      "theoretical_loss": 4.043668631833908,
      "tokens_seen": 372965376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479638916750251,
      "loss": 2.6676,
      "theoretical_loss": 4.043588209230498,
      "tokens_seen": 373030912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044795386158475424,
      "loss": 2.7663,
      "theoretical_loss": 4.04350780471025,
      "tokens_seen": 373096448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044794383149448347,
      "loss": 2.9916,
      "theoretical_loss": 4.043427418265923,
      "tokens_seen": 373161984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044793380140421265,
      "loss": 2.8541,
      "theoretical_loss": 4.043347049890279,
      "tokens_seen": 373227520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044792377131394183,
      "loss": 2.8811,
      "theoretical_loss": 4.043266699576086,
      "tokens_seen": 373293056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447913741223671,
      "loss": 2.8581,
      "theoretical_loss": 4.043186367316116,
      "tokens_seen": 373358592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044790371113340025,
      "loss": 2.804,
      "theoretical_loss": 4.043106053103141,
      "tokens_seen": 373424128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478936810431294,
      "loss": 2.9816,
      "theoretical_loss": 4.043025756929945,
      "tokens_seen": 373489664
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 217105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.010817766189575,
      "objective/train/theoretical_loss": 4.042945478789308,
      "objective/train/tokens_used": 394015200,
      "theoretical_loss": 4.042945478789308,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478836509528586,
      "loss": 3.1259,
      "theoretical_loss": 4.042945478789308,
      "tokens_seen": 373555200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044787362086258774,
      "loss": 3.0529,
      "theoretical_loss": 4.042865218674019,
      "tokens_seen": 373620736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447863590772317,
      "loss": 2.7388,
      "theoretical_loss": 4.04278497657687,
      "tokens_seen": 373686272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044785356068204616,
      "loss": 2.8478,
      "theoretical_loss": 4.042704752490658,
      "tokens_seen": 373751808
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044784353059177534,
      "loss": 2.7947,
      "theoretical_loss": 4.042624546408182,
      "tokens_seen": 373817344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478335005015045,
      "loss": 2.8891,
      "theoretical_loss": 4.042544358322247,
      "tokens_seen": 373882880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478234704112337,
      "loss": 2.9725,
      "theoretical_loss": 4.042464188225661,
      "tokens_seen": 373948416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478134403209629,
      "loss": 2.9604,
      "theoretical_loss": 4.042384036111236,
      "tokens_seen": 374013952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478034102306921,
      "loss": 2.8276,
      "theoretical_loss": 4.04230390197179,
      "tokens_seen": 374079488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044779338014042124,
      "loss": 3.1022,
      "theoretical_loss": 4.042223785800144,
      "tokens_seen": 374145024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477833500501505,
      "loss": 3.1569,
      "theoretical_loss": 4.042143687589123,
      "tokens_seen": 374210560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477733199598796,
      "loss": 2.868,
      "theoretical_loss": 4.042063607331555,
      "tokens_seen": 374276096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044776328986960884,
      "loss": 2.8411,
      "theoretical_loss": 4.0419835450202735,
      "tokens_seen": 374341632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447753259779338,
      "loss": 3.0814,
      "theoretical_loss": 4.041903500648116,
      "tokens_seen": 374407168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477432296890672,
      "loss": 2.9514,
      "theoretical_loss": 4.041823474207925,
      "tokens_seen": 374472704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477331995987964,
      "loss": 3.0362,
      "theoretical_loss": 4.041743465692545,
      "tokens_seen": 374538240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477231695085256,
      "loss": 2.8953,
      "theoretical_loss": 4.041663475094826,
      "tokens_seen": 374603776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044771313941825475,
      "loss": 2.8501,
      "theoretical_loss": 4.041583502407622,
      "tokens_seen": 374669312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447703109327984,
      "loss": 3.0705,
      "theoretical_loss": 4.041503547623791,
      "tokens_seen": 374734848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476930792377131,
      "loss": 2.8879,
      "theoretical_loss": 4.041423610736195,
      "tokens_seen": 374800384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044768304914744234,
      "loss": 2.801,
      "theoretical_loss": 4.0413436917376995,
      "tokens_seen": 374865920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476730190571715,
      "loss": 2.9406,
      "theoretical_loss": 4.041263790621175,
      "tokens_seen": 374931456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476629889669007,
      "loss": 2.859,
      "theoretical_loss": 4.041183907379495,
      "tokens_seen": 374996992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476529588766299,
      "loss": 2.9117,
      "theoretical_loss": 4.041104042005539,
      "tokens_seen": 375062528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044764292878635907,
      "loss": 3.1132,
      "theoretical_loss": 4.04102419449219,
      "tokens_seen": 375128064
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 218363,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9804539680480957,
      "objective/train/theoretical_loss": 4.040944364832332,
      "objective/train/tokens_used": 395653600,
      "theoretical_loss": 4.040944364832332,
      "tokens_seen": 375193600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044763289869608825,
      "loss": 2.9225,
      "theoretical_loss": 4.040944364832332,
      "tokens_seen": 375193600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476228686058175,
      "loss": 2.854,
      "theoretical_loss": 4.040864553018858,
      "tokens_seen": 375259136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476128385155466,
      "loss": 2.7824,
      "theoretical_loss": 4.040784759044662,
      "tokens_seen": 375324672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044760280842527585,
      "loss": 2.8757,
      "theoretical_loss": 4.040704982902642,
      "tokens_seen": 375390208
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044759277833500503,
      "loss": 2.9178,
      "theoretical_loss": 4.0406252245857015,
      "tokens_seen": 375455744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475827482447342,
      "loss": 3.1065,
      "theoretical_loss": 4.0405454840867465,
      "tokens_seen": 375521280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475727181544634,
      "loss": 2.7781,
      "theoretical_loss": 4.040465761398689,
      "tokens_seen": 375586816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044756268806419257,
      "loss": 2.9638,
      "theoretical_loss": 4.0403860565144445,
      "tokens_seen": 375652352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044755265797392175,
      "loss": 3.18,
      "theoretical_loss": 4.040306369426929,
      "tokens_seen": 375717888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447542627883651,
      "loss": 2.7766,
      "theoretical_loss": 4.04022670012907,
      "tokens_seen": 375783424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475325977933801,
      "loss": 2.9218,
      "theoretical_loss": 4.0401470486137905,
      "tokens_seen": 375848960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044752256770310935,
      "loss": 3.0006,
      "theoretical_loss": 4.040067414874024,
      "tokens_seen": 375914496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475125376128385,
      "loss": 2.8125,
      "theoretical_loss": 4.039987798902705,
      "tokens_seen": 375980032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475025075225677,
      "loss": 3.1362,
      "theoretical_loss": 4.039908200692773,
      "tokens_seen": 376045568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474924774322969,
      "loss": 2.9677,
      "theoretical_loss": 4.039828620237171,
      "tokens_seen": 376111104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474824473420261,
      "loss": 2.9863,
      "theoretical_loss": 4.039749057528846,
      "tokens_seen": 376176640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044747241725175526,
      "loss": 2.858,
      "theoretical_loss": 4.039669512560751,
      "tokens_seen": 376242176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044746238716148444,
      "loss": 2.9564,
      "theoretical_loss": 4.039589985325839,
      "tokens_seen": 376307712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474523570712136,
      "loss": 2.976,
      "theoretical_loss": 4.039510475817071,
      "tokens_seen": 376373248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044744232698094285,
      "loss": 3.026,
      "theoretical_loss": 4.03943098402741,
      "tokens_seen": 376438784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044743229689067204,
      "loss": 2.7752,
      "theoretical_loss": 4.039351509949824,
      "tokens_seen": 376504320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474222668004012,
      "loss": 2.9442,
      "theoretical_loss": 4.0392720535772835,
      "tokens_seen": 376569856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044741223671013045,
      "loss": 3.1013,
      "theoretical_loss": 4.039192614902763,
      "tokens_seen": 376635392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474022066198596,
      "loss": 2.8631,
      "theoretical_loss": 4.0391131939192455,
      "tokens_seen": 376700928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473921765295888,
      "loss": 2.9903,
      "theoretical_loss": 4.039033790619711,
      "tokens_seen": 376766464
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 218865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9096364974975586,
      "objective/train/theoretical_loss": 4.038954404997148,
      "objective/train/tokens_used": 397292000,
      "theoretical_loss": 4.038954404997148,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044738214643931794,
      "loss": 3.0827,
      "theoretical_loss": 4.038954404997148,
      "tokens_seen": 376832000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473721163490472,
      "loss": 3.2042,
      "theoretical_loss": 4.038875037044549,
      "tokens_seen": 376897536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044736208625877636,
      "loss": 3.0121,
      "theoretical_loss": 4.038795686754908,
      "tokens_seen": 376963072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044735205616850554,
      "loss": 2.8018,
      "theoretical_loss": 4.038716354121226,
      "tokens_seen": 377028608
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473420260782347,
      "loss": 2.963,
      "theoretical_loss": 4.038637039136505,
      "tokens_seen": 377094144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473319959879639,
      "loss": 2.6641,
      "theoretical_loss": 4.0385577417937535,
      "tokens_seen": 377159680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473219658976931,
      "loss": 3.1323,
      "theoretical_loss": 4.0384784620859815,
      "tokens_seen": 377225216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473119358074223,
      "loss": 2.9769,
      "theoretical_loss": 4.038399200006207,
      "tokens_seen": 377290752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044730190571715144,
      "loss": 2.9038,
      "theoretical_loss": 4.038319955547447,
      "tokens_seen": 377356288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472918756268807,
      "loss": 2.8752,
      "theoretical_loss": 4.038240728702725,
      "tokens_seen": 377421824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472818455366098,
      "loss": 3.0864,
      "theoretical_loss": 4.03816151946507,
      "tokens_seen": 377487360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044727181544633904,
      "loss": 2.8703,
      "theoretical_loss": 4.038082327827512,
      "tokens_seen": 377552896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472617853560682,
      "loss": 2.896,
      "theoretical_loss": 4.038003153783086,
      "tokens_seen": 377618432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472517552657974,
      "loss": 3.0818,
      "theoretical_loss": 4.037923997324833,
      "tokens_seen": 377683968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472417251755266,
      "loss": 2.9379,
      "theoretical_loss": 4.037844858445794,
      "tokens_seen": 377749504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472316950852558,
      "loss": 2.9901,
      "theoretical_loss": 4.037765737139018,
      "tokens_seen": 377815040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044722166499498495,
      "loss": 2.9231,
      "theoretical_loss": 4.037686633397555,
      "tokens_seen": 377880576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472116349047142,
      "loss": 2.7279,
      "theoretical_loss": 4.03760754721446,
      "tokens_seen": 377946112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472016048144433,
      "loss": 2.8789,
      "theoretical_loss": 4.037528478582793,
      "tokens_seen": 378011648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044719157472417254,
      "loss": 2.8981,
      "theoretical_loss": 4.037449427495615,
      "tokens_seen": 378077184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471815446339017,
      "loss": 3.0708,
      "theoretical_loss": 4.037370393945996,
      "tokens_seen": 378142720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471715145436309,
      "loss": 2.9477,
      "theoretical_loss": 4.037291377927005,
      "tokens_seen": 378208256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471614844533601,
      "loss": 2.8476,
      "theoretical_loss": 4.0372123794317165,
      "tokens_seen": 378273792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044715145436308927,
      "loss": 3.0008,
      "theoretical_loss": 4.037133398453209,
      "tokens_seen": 378339328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044714142427281845,
      "loss": 2.7895,
      "theoretical_loss": 4.037054434984566,
      "tokens_seen": 378404864
    },
    {
      "epoch": 0.11,
      "objective/train/docs_used": 220023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9392900466918945,
      "objective/train/theoretical_loss": 4.036975489018874,
      "objective/train/tokens_used": 398930400,
      "theoretical_loss": 4.036975489018874,
      "tokens_seen": 378470400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471313941825477,
      "loss": 2.8371,
      "theoretical_loss": 4.036975489018874,
      "tokens_seen": 378470400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471213640922768,
      "loss": 3.062,
      "theoretical_loss": 4.036896560549224,
      "tokens_seen": 378535936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044711133400200605,
      "loss": 2.847,
      "theoretical_loss": 4.03681764956871,
      "tokens_seen": 378601472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044710130391173523,
      "loss": 2.8412,
      "theoretical_loss": 4.036738756070429,
      "tokens_seen": 378667008
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470912738214644,
      "loss": 2.7814,
      "theoretical_loss": 4.036659880047486,
      "tokens_seen": 378732544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470812437311936,
      "loss": 2.9737,
      "theoretical_loss": 4.036581021492984,
      "tokens_seen": 378798080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044707121364092277,
      "loss": 2.8625,
      "theoretical_loss": 4.036502180400036,
      "tokens_seen": 378863616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044706118355065195,
      "loss": 2.8385,
      "theoretical_loss": 4.036423356761754,
      "tokens_seen": 378929152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470511534603812,
      "loss": 2.8222,
      "theoretical_loss": 4.036344550571257,
      "tokens_seen": 378994688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470411233701103,
      "loss": 2.8922,
      "theoretical_loss": 4.0362657618216655,
      "tokens_seen": 379060224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044703109327983955,
      "loss": 2.5815,
      "theoretical_loss": 4.036186990506107,
      "tokens_seen": 379125760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470210631895687,
      "loss": 2.9568,
      "theoretical_loss": 4.036108236617711,
      "tokens_seen": 379191296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470110330992979,
      "loss": 2.9341,
      "theoretical_loss": 4.036029500149609,
      "tokens_seen": 379256832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470010030090271,
      "loss": 2.8396,
      "theoretical_loss": 4.03595078109494,
      "tokens_seen": 379322368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004469909729187563,
      "loss": 3.1731,
      "theoretical_loss": 4.035872079446846,
      "tokens_seen": 379387904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044698094282848546,
      "loss": 2.8303,
      "theoretical_loss": 4.03579339519847,
      "tokens_seen": 379453440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044697091273821464,
      "loss": 2.9496,
      "theoretical_loss": 4.035714728342962,
      "tokens_seen": 379518976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469608826479438,
      "loss": 2.9665,
      "theoretical_loss": 4.035636078873477,
      "tokens_seen": 379584512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044695085255767305,
      "loss": 2.8545,
      "theoretical_loss": 4.035557446783168,
      "tokens_seen": 379650048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469408224674022,
      "loss": 3.0728,
      "theoretical_loss": 4.035478832065198,
      "tokens_seen": 379715584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469307923771314,
      "loss": 2.7782,
      "theoretical_loss": 4.035400234712732,
      "tokens_seen": 379781120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469207622868606,
      "loss": 2.9358,
      "theoretical_loss": 4.035321654718937,
      "tokens_seen": 379846656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469107321965898,
      "loss": 2.9011,
      "theoretical_loss": 4.0352430920769855,
      "tokens_seen": 379912192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044690070210631896,
      "loss": 2.8626,
      "theoretical_loss": 4.035164546780055,
      "tokens_seen": 379977728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044689067201604814,
      "loss": 2.9872,
      "theoretical_loss": 4.035086018821325,
      "tokens_seen": 380043264
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 221040,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.943068742752075,
      "objective/train/theoretical_loss": 4.035007508193978,
      "objective/train/tokens_used": 400568800,
      "theoretical_loss": 4.035007508193978,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468806419257773,
      "loss": 3.0626,
      "theoretical_loss": 4.035007508193978,
      "tokens_seen": 380108800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044687061183550656,
      "loss": 2.7622,
      "theoretical_loss": 4.034929014891202,
      "tokens_seen": 380174336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468605817452357,
      "loss": 2.9553,
      "theoretical_loss": 4.03485053890619,
      "tokens_seen": 380239872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468505516549649,
      "loss": 2.7883,
      "theoretical_loss": 4.034772080232137,
      "tokens_seen": 380305408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044684052156469405,
      "loss": 2.8372,
      "theoretical_loss": 4.034693638862241,
      "tokens_seen": 380370944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468304914744233,
      "loss": 2.9443,
      "theoretical_loss": 4.034615214789707,
      "tokens_seen": 380436480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044682046138415246,
      "loss": 2.7825,
      "theoretical_loss": 4.034536808007741,
      "tokens_seen": 380502016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044681043129388164,
      "loss": 2.8113,
      "theoretical_loss": 4.034458418509553,
      "tokens_seen": 380567552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468004012036108,
      "loss": 2.8747,
      "theoretical_loss": 4.034380046288359,
      "tokens_seen": 380633088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044679037111334,
      "loss": 2.9584,
      "theoretical_loss": 4.034301691337377,
      "tokens_seen": 380698624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467803410230692,
      "loss": 2.9068,
      "theoretical_loss": 4.034223353649828,
      "tokens_seen": 380764160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467703109327984,
      "loss": 2.9041,
      "theoretical_loss": 4.034145033218941,
      "tokens_seen": 380829696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044676028084252755,
      "loss": 2.9763,
      "theoretical_loss": 4.034066730037944,
      "tokens_seen": 380895232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467502507522568,
      "loss": 3.2191,
      "theoretical_loss": 4.0339884441000695,
      "tokens_seen": 380960768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044674022066198597,
      "loss": 3.015,
      "theoretical_loss": 4.0339101753985585,
      "tokens_seen": 381026304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044673019057171515,
      "loss": 3.0623,
      "theoretical_loss": 4.033831923926649,
      "tokens_seen": 381091840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044672016048144433,
      "loss": 2.8022,
      "theoretical_loss": 4.033753689677588,
      "tokens_seen": 381157376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467101303911735,
      "loss": 3.0212,
      "theoretical_loss": 4.033675472644624,
      "tokens_seen": 381222912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467001003009027,
      "loss": 2.8385,
      "theoretical_loss": 4.03359727282101,
      "tokens_seen": 381288448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466900702106319,
      "loss": 2.9087,
      "theoretical_loss": 4.033519090200002,
      "tokens_seen": 381353984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466800401203611,
      "loss": 2.8946,
      "theoretical_loss": 4.033440924774862,
      "tokens_seen": 381419520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466700100300903,
      "loss": 2.7513,
      "theoretical_loss": 4.033362776538853,
      "tokens_seen": 381485056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044665997993981947,
      "loss": 2.809,
      "theoretical_loss": 4.033284645485243,
      "tokens_seen": 381550592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044664994984954865,
      "loss": 2.8946,
      "theoretical_loss": 4.033206531607304,
      "tokens_seen": 381616128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466399197592779,
      "loss": 2.8353,
      "theoretical_loss": 4.033128434898312,
      "tokens_seen": 381681664
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 221564,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5393569469451904,
      "objective/train/theoretical_loss": 4.033050355351545,
      "objective/train/tokens_used": 402207200,
      "theoretical_loss": 4.033050355351545,
      "tokens_seen": 381747200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446629889669007,
      "loss": 2.8139,
      "theoretical_loss": 4.033050355351545,
      "tokens_seen": 381747200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044661985957873625,
      "loss": 2.8216,
      "theoretical_loss": 4.032972292960288,
      "tokens_seen": 381812736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044660982948846543,
      "loss": 2.8422,
      "theoretical_loss": 4.0328942477178265,
      "tokens_seen": 381878272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465997993981946,
      "loss": 2.9652,
      "theoretical_loss": 4.032816219617452,
      "tokens_seen": 381943808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465897693079238,
      "loss": 2.8776,
      "theoretical_loss": 4.032738208652459,
      "tokens_seen": 382009344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044657973921765297,
      "loss": 2.8477,
      "theoretical_loss": 4.032660214816145,
      "tokens_seen": 382074880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044656970912738215,
      "loss": 2.8979,
      "theoretical_loss": 4.0325822381018135,
      "tokens_seen": 382140416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465596790371114,
      "loss": 2.9445,
      "theoretical_loss": 4.03250427850277,
      "tokens_seen": 382205952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465496489468405,
      "loss": 3.0591,
      "theoretical_loss": 4.032426336012321,
      "tokens_seen": 382271488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044653961885656975,
      "loss": 2.8794,
      "theoretical_loss": 4.032348410623785,
      "tokens_seen": 382337024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465295887662989,
      "loss": 3.0788,
      "theoretical_loss": 4.032270502330475,
      "tokens_seen": 382402560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465195586760281,
      "loss": 2.9419,
      "theoretical_loss": 4.032192611125714,
      "tokens_seen": 382468096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465095285857573,
      "loss": 3.056,
      "theoretical_loss": 4.032114737002825,
      "tokens_seen": 382533632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464994984954865,
      "loss": 3.0452,
      "theoretical_loss": 4.032036879955138,
      "tokens_seen": 382599168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044648946840521566,
      "loss": 2.9026,
      "theoretical_loss": 4.031959039975983,
      "tokens_seen": 382664704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044647943831494484,
      "loss": 2.8125,
      "theoretical_loss": 4.031881217058698,
      "tokens_seen": 382730240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446469408224674,
      "loss": 2.6746,
      "theoretical_loss": 4.0318034111966226,
      "tokens_seen": 382795776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044645937813440325,
      "loss": 2.946,
      "theoretical_loss": 4.031725622383099,
      "tokens_seen": 382861312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464493480441324,
      "loss": 3.0812,
      "theoretical_loss": 4.031647850611473,
      "tokens_seen": 382926848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464393179538616,
      "loss": 3.153,
      "theoretical_loss": 4.031570095875098,
      "tokens_seen": 382992384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464292878635908,
      "loss": 3.0693,
      "theoretical_loss": 4.031492358167329,
      "tokens_seen": 383057920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044641925777332,
      "loss": 3.2132,
      "theoretical_loss": 4.031414637481521,
      "tokens_seen": 383123456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044640922768304916,
      "loss": 2.6272,
      "theoretical_loss": 4.031336933811038,
      "tokens_seen": 383188992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044639919759277834,
      "loss": 3.0864,
      "theoretical_loss": 4.031259247149245,
      "tokens_seen": 383254528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463891675025075,
      "loss": 3.0017,
      "theoretical_loss": 4.031181577489512,
      "tokens_seen": 383320064
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 222655,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.689690589904785,
      "objective/train/theoretical_loss": 4.031103924825214,
      "objective/train/tokens_used": 403845600,
      "theoretical_loss": 4.031103924825214,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044637913741223676,
      "loss": 3.0431,
      "theoretical_loss": 4.031103924825214,
      "tokens_seen": 383385600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463691073219659,
      "loss": 2.8683,
      "theoretical_loss": 4.0310262891497235,
      "tokens_seen": 383451136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463590772316951,
      "loss": 3.0141,
      "theoretical_loss": 4.030948670456425,
      "tokens_seen": 383516672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044634904714142425,
      "loss": 2.9239,
      "theoretical_loss": 4.0308710687386995,
      "tokens_seen": 383582208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463390170511535,
      "loss": 2.7962,
      "theoretical_loss": 4.030793483989939,
      "tokens_seen": 383647744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044632898696088266,
      "loss": 2.9929,
      "theoretical_loss": 4.0307159162035315,
      "tokens_seen": 383713280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044631895687061184,
      "loss": 2.9842,
      "theoretical_loss": 4.030638365372874,
      "tokens_seen": 383778816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446308926780341,
      "loss": 2.8692,
      "theoretical_loss": 4.030560831491365,
      "tokens_seen": 383844352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462988966900702,
      "loss": 3.0569,
      "theoretical_loss": 4.030483314552408,
      "tokens_seen": 383909888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462888665997994,
      "loss": 3.2353,
      "theoretical_loss": 4.03040581454941,
      "tokens_seen": 383975424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462788365095286,
      "loss": 3.0781,
      "theoretical_loss": 4.03032833147578,
      "tokens_seen": 384040960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044626880641925775,
      "loss": 2.7578,
      "theoretical_loss": 4.030250865324932,
      "tokens_seen": 384106496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446258776328987,
      "loss": 3.0714,
      "theoretical_loss": 4.030173416090284,
      "tokens_seen": 384172032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044624874623871617,
      "loss": 2.9921,
      "theoretical_loss": 4.030095983765257,
      "tokens_seen": 384237568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044623871614844535,
      "loss": 2.9743,
      "theoretical_loss": 4.030018568343276,
      "tokens_seen": 384303104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044622868605817453,
      "loss": 3.0737,
      "theoretical_loss": 4.02994116981777,
      "tokens_seen": 384368640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462186559679037,
      "loss": 2.7114,
      "theoretical_loss": 4.02986378818217,
      "tokens_seen": 384434176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462086258776329,
      "loss": 2.8272,
      "theoretical_loss": 4.029786423429913,
      "tokens_seen": 384499712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461985957873621,
      "loss": 3.1081,
      "theoretical_loss": 4.029709075554439,
      "tokens_seen": 384565248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044618856569709125,
      "loss": 2.938,
      "theoretical_loss": 4.02963174454919,
      "tokens_seen": 384630784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461785356068205,
      "loss": 2.6041,
      "theoretical_loss": 4.029554430407613,
      "tokens_seen": 384696320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461685055165496,
      "loss": 2.8779,
      "theoretical_loss": 4.0294771331231605,
      "tokens_seen": 384761856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044615847542627885,
      "loss": 2.8709,
      "theoretical_loss": 4.029399852689284,
      "tokens_seen": 384827392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044614844533600803,
      "loss": 3.0417,
      "theoretical_loss": 4.0293225890994435,
      "tokens_seen": 384892928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461384152457372,
      "loss": 2.4958,
      "theoretical_loss": 4.0292453423471,
      "tokens_seen": 384958464
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 222910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9024524688720703,
      "objective/train/theoretical_loss": 4.029168112425719,
      "objective/train/tokens_used": 405484000,
      "theoretical_loss": 4.029168112425719,
      "tokens_seen": 385024000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461283851554664,
      "loss": 3.0308,
      "theoretical_loss": 4.029168112425719,
      "tokens_seen": 385024000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044611835506519563,
      "loss": 2.8817,
      "theoretical_loss": 4.029090899328769,
      "tokens_seen": 385089536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044610832497492476,
      "loss": 2.8394,
      "theoretical_loss": 4.029013703049722,
      "tokens_seen": 385155072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446098294884654,
      "loss": 2.9092,
      "theoretical_loss": 4.028936523582055,
      "tokens_seen": 385220608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460882647943831,
      "loss": 2.9076,
      "theoretical_loss": 4.028859360919247,
      "tokens_seen": 385286144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044607823470411235,
      "loss": 3.1037,
      "theoretical_loss": 4.0287822150547825,
      "tokens_seen": 385351680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044606820461384153,
      "loss": 2.9683,
      "theoretical_loss": 4.028705085982148,
      "tokens_seen": 385417216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460581745235707,
      "loss": 2.8876,
      "theoretical_loss": 4.028627973694833,
      "tokens_seen": 385482752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460481444332999,
      "loss": 2.7867,
      "theoretical_loss": 4.028550878186334,
      "tokens_seen": 385548288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460381143430291,
      "loss": 3.0148,
      "theoretical_loss": 4.028473799450149,
      "tokens_seen": 385613824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044602808425275826,
      "loss": 2.4792,
      "theoretical_loss": 4.028396737479777,
      "tokens_seen": 385679360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460180541624875,
      "loss": 3.0135,
      "theoretical_loss": 4.028319692268725,
      "tokens_seen": 385744896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460080240722166,
      "loss": 3.1208,
      "theoretical_loss": 4.028242663810501,
      "tokens_seen": 385810432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044599799398194586,
      "loss": 2.9895,
      "theoretical_loss": 4.028165652098618,
      "tokens_seen": 385875968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445987963891675,
      "loss": 3.0448,
      "theoretical_loss": 4.028088657126593,
      "tokens_seen": 385941504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459779338014042,
      "loss": 3.0796,
      "theoretical_loss": 4.028011678887944,
      "tokens_seen": 386007040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459679037111334,
      "loss": 3.0512,
      "theoretical_loss": 4.027934717376195,
      "tokens_seen": 386072576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459578736208626,
      "loss": 2.7118,
      "theoretical_loss": 4.027857772584873,
      "tokens_seen": 386138112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044594784353059176,
      "loss": 3.0313,
      "theoretical_loss": 4.027780844507508,
      "tokens_seen": 386203648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445937813440321,
      "loss": 2.834,
      "theoretical_loss": 4.0277039331376345,
      "tokens_seen": 386269184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459277833500502,
      "loss": 2.7997,
      "theoretical_loss": 4.027627038468791,
      "tokens_seen": 386334720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044591775325977936,
      "loss": 2.9824,
      "theoretical_loss": 4.027550160494516,
      "tokens_seen": 386400256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044590772316950854,
      "loss": 3.1209,
      "theoretical_loss": 4.027473299208357,
      "tokens_seen": 386465792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458976930792377,
      "loss": 3.1398,
      "theoretical_loss": 4.027396454603862,
      "tokens_seen": 386531328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044588766298896696,
      "loss": 2.7702,
      "theoretical_loss": 4.027319626674583,
      "tokens_seen": 386596864
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 224330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.956303596496582,
      "objective/train/theoretical_loss": 4.027242815414074,
      "objective/train/tokens_used": 407122400,
      "theoretical_loss": 4.027242815414074,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458776328986961,
      "loss": 3.0091,
      "theoretical_loss": 4.027242815414074,
      "tokens_seen": 386662400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458676028084253,
      "loss": 3.0036,
      "theoretical_loss": 4.027166020815896,
      "tokens_seen": 386727936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044585757271815445,
      "loss": 2.829,
      "theoretical_loss": 4.027089242873612,
      "tokens_seen": 386793472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458475426278837,
      "loss": 3.0081,
      "theoretical_loss": 4.027012481580787,
      "tokens_seen": 386859008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044583751253761286,
      "loss": 2.9494,
      "theoretical_loss": 4.026935736930992,
      "tokens_seen": 386924544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044582748244734204,
      "loss": 2.8553,
      "theoretical_loss": 4.0268590089178,
      "tokens_seen": 386990080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458174523570712,
      "loss": 2.8579,
      "theoretical_loss": 4.026782297534787,
      "tokens_seen": 387055616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458074222668004,
      "loss": 2.9371,
      "theoretical_loss": 4.026705602775536,
      "tokens_seen": 387121152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457973921765296,
      "loss": 2.8851,
      "theoretical_loss": 4.02662892463363,
      "tokens_seen": 387186688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457873620862588,
      "loss": 3.1423,
      "theoretical_loss": 4.026552263102657,
      "tokens_seen": 387252224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044577733199598795,
      "loss": 2.9309,
      "theoretical_loss": 4.026475618176208,
      "tokens_seen": 387317760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457673019057172,
      "loss": 3.0442,
      "theoretical_loss": 4.026398989847877,
      "tokens_seen": 387383296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044575727181544637,
      "loss": 3.0007,
      "theoretical_loss": 4.0263223781112645,
      "tokens_seen": 387448832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044574724172517555,
      "loss": 3.0223,
      "theoretical_loss": 4.026245782959971,
      "tokens_seen": 387514368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044573721163490473,
      "loss": 2.7477,
      "theoretical_loss": 4.026169204387603,
      "tokens_seen": 387579904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457271815446339,
      "loss": 2.8504,
      "theoretical_loss": 4.026092642387769,
      "tokens_seen": 387645440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457171514543631,
      "loss": 2.9407,
      "theoretical_loss": 4.0260160969540815,
      "tokens_seen": 387710976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457071213640923,
      "loss": 3.1619,
      "theoretical_loss": 4.025939568080156,
      "tokens_seen": 387776512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044569709127382145,
      "loss": 2.8439,
      "theoretical_loss": 4.025863055759615,
      "tokens_seen": 387842048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456870611835507,
      "loss": 2.932,
      "theoretical_loss": 4.025786559986079,
      "tokens_seen": 387907584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456770310932798,
      "loss": 2.8177,
      "theoretical_loss": 4.025710080753175,
      "tokens_seen": 387973120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044566700100300905,
      "loss": 2.9714,
      "theoretical_loss": 4.025633618054535,
      "tokens_seen": 388038656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044565697091273823,
      "loss": 3.0176,
      "theoretical_loss": 4.025557171883791,
      "tokens_seen": 388104192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456469408224674,
      "loss": 3.023,
      "theoretical_loss": 4.025480742234581,
      "tokens_seen": 388169728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456369107321966,
      "loss": 2.8862,
      "theoretical_loss": 4.025404329100546,
      "tokens_seen": 388235264
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 225094,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.5328869819641113,
      "objective/train/theoretical_loss": 4.0253279324753315,
      "objective/train/tokens_used": 408760800,
      "theoretical_loss": 4.0253279324753315,
      "tokens_seen": 388300800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044562688064192583,
      "loss": 3.2148,
      "theoretical_loss": 4.0253279324753315,
      "tokens_seen": 388300800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044561685055165496,
      "loss": 2.9028,
      "theoretical_loss": 4.025251552352584,
      "tokens_seen": 388366336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456068204613842,
      "loss": 3.0211,
      "theoretical_loss": 4.025175188725955,
      "tokens_seen": 388431872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455967903711133,
      "loss": 2.894,
      "theoretical_loss": 4.025098841589099,
      "tokens_seen": 388497408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044558676028084255,
      "loss": 2.7945,
      "theoretical_loss": 4.025022510935676,
      "tokens_seen": 388562944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044557673019057173,
      "loss": 2.944,
      "theoretical_loss": 4.024946196759347,
      "tokens_seen": 388628480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455667001003009,
      "loss": 3.0782,
      "theoretical_loss": 4.024869899053778,
      "tokens_seen": 388694016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455566700100301,
      "loss": 3.1571,
      "theoretical_loss": 4.024793617812637,
      "tokens_seen": 388759552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455466399197593,
      "loss": 2.8881,
      "theoretical_loss": 4.024717353029597,
      "tokens_seen": 388825088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044553660982948846,
      "loss": 2.9853,
      "theoretical_loss": 4.024641104698335,
      "tokens_seen": 388890624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455265797392177,
      "loss": 2.8279,
      "theoretical_loss": 4.02456487281253,
      "tokens_seen": 388956160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455165496489468,
      "loss": 2.8868,
      "theoretical_loss": 4.024488657365863,
      "tokens_seen": 389021696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044550651955867606,
      "loss": 2.901,
      "theoretical_loss": 4.024412458352023,
      "tokens_seen": 389087232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454964894684052,
      "loss": 3.0881,
      "theoretical_loss": 4.024336275764698,
      "tokens_seen": 389152768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454864593781344,
      "loss": 2.9645,
      "theoretical_loss": 4.024260109597583,
      "tokens_seen": 389218304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454764292878636,
      "loss": 2.8748,
      "theoretical_loss": 4.024183959844375,
      "tokens_seen": 389283840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454663991975928,
      "loss": 2.8245,
      "theoretical_loss": 4.024107826498772,
      "tokens_seen": 389349376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044545636910732196,
      "loss": 2.966,
      "theoretical_loss": 4.024031709554482,
      "tokens_seen": 389414912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454463390170512,
      "loss": 2.7041,
      "theoretical_loss": 4.023955609005208,
      "tokens_seen": 389480448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454363089267803,
      "loss": 2.8466,
      "theoretical_loss": 4.023879524844663,
      "tokens_seen": 389545984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044542627883650956,
      "loss": 3.0461,
      "theoretical_loss": 4.023803457066561,
      "tokens_seen": 389611520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454162487462387,
      "loss": 3.013,
      "theoretical_loss": 4.023727405664619,
      "tokens_seen": 389677056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454062186559679,
      "loss": 3.1216,
      "theoretical_loss": 4.02365137063256,
      "tokens_seen": 389742592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453961885656971,
      "loss": 2.8899,
      "theoretical_loss": 4.023575351964108,
      "tokens_seen": 389808128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453861584754263,
      "loss": 2.9788,
      "theoretical_loss": 4.02349934965299,
      "tokens_seen": 389873664
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 226325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.121893882751465,
      "objective/train/theoretical_loss": 4.023423363692939,
      "objective/train/tokens_used": 410399200,
      "theoretical_loss": 4.023423363692939,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044537612838515547,
      "loss": 2.988,
      "theoretical_loss": 4.023423363692939,
      "tokens_seen": 389939200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044536609829488465,
      "loss": 2.8528,
      "theoretical_loss": 4.023347394077689,
      "tokens_seen": 390004736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044535606820461383,
      "loss": 3.0006,
      "theoretical_loss": 4.023271440800979,
      "tokens_seen": 390070272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044534603811434306,
      "loss": 2.818,
      "theoretical_loss": 4.023195503856551,
      "tokens_seen": 390135808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453360080240722,
      "loss": 2.8456,
      "theoretical_loss": 4.02311958323815,
      "tokens_seen": 390201344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453259779338014,
      "loss": 3.0276,
      "theoretical_loss": 4.023043678939525,
      "tokens_seen": 390266880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044531594784353055,
      "loss": 2.9954,
      "theoretical_loss": 4.022967790954429,
      "tokens_seen": 390332416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453059177532598,
      "loss": 3.021,
      "theoretical_loss": 4.022891919276617,
      "tokens_seen": 390397952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044529588766298897,
      "loss": 2.8376,
      "theoretical_loss": 4.022816063899848,
      "tokens_seen": 390463488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044528585757271815,
      "loss": 2.9268,
      "theoretical_loss": 4.022740224817886,
      "tokens_seen": 390529024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044527582748244733,
      "loss": 2.9293,
      "theoretical_loss": 4.022664402024495,
      "tokens_seen": 390594560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044526579739217657,
      "loss": 3.2044,
      "theoretical_loss": 4.0225885955134455,
      "tokens_seen": 390660096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452557673019057,
      "loss": 3.0436,
      "theoretical_loss": 4.022512805278511,
      "tokens_seen": 390725632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044524573721163493,
      "loss": 2.7897,
      "theoretical_loss": 4.022437031313467,
      "tokens_seen": 390791168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044523570712136406,
      "loss": 3.1187,
      "theoretical_loss": 4.022361273612093,
      "tokens_seen": 390856704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452256770310933,
      "loss": 3.0885,
      "theoretical_loss": 4.022285532168174,
      "tokens_seen": 390922240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044521564694082247,
      "loss": 3.0223,
      "theoretical_loss": 4.022209806975494,
      "tokens_seen": 390987776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044520561685055165,
      "loss": 2.9709,
      "theoretical_loss": 4.022134098027844,
      "tokens_seen": 391053312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044519558676028083,
      "loss": 2.9389,
      "theoretical_loss": 4.022058405319018,
      "tokens_seen": 391118848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044518555667001,
      "loss": 2.7071,
      "theoretical_loss": 4.021982728842812,
      "tokens_seen": 391184384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044517552657973925,
      "loss": 2.8463,
      "theoretical_loss": 4.021907068593027,
      "tokens_seen": 391249920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044516549648946843,
      "loss": 2.8327,
      "theoretical_loss": 4.021831424563467,
      "tokens_seen": 391315456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451554663991976,
      "loss": 2.7659,
      "theoretical_loss": 4.021755796747938,
      "tokens_seen": 391380992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451454363089268,
      "loss": 2.8349,
      "theoretical_loss": 4.02168018514025,
      "tokens_seen": 391446528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044513540621865603,
      "loss": 2.9836,
      "theoretical_loss": 4.021604589734219,
      "tokens_seen": 391512064
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 226938,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0995724201202393,
      "objective/train/theoretical_loss": 4.021529010523659,
      "objective/train/tokens_used": 412037600,
      "theoretical_loss": 4.021529010523659,
      "tokens_seen": 391577600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044512537612838516,
      "loss": 3.0031,
      "theoretical_loss": 4.021529010523659,
      "tokens_seen": 391577600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451153460381144,
      "loss": 2.8716,
      "theoretical_loss": 4.0214534475023935,
      "tokens_seen": 391643136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451053159478435,
      "loss": 2.902,
      "theoretical_loss": 4.0213779006642465,
      "tokens_seen": 391708672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044509528585757275,
      "loss": 2.9234,
      "theoretical_loss": 4.021302370003044,
      "tokens_seen": 391774208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044508525576730194,
      "loss": 2.9735,
      "theoretical_loss": 4.021226855512616,
      "tokens_seen": 391839744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450752256770311,
      "loss": 2.805,
      "theoretical_loss": 4.021151357186799,
      "tokens_seen": 391905280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450651955867603,
      "loss": 2.8537,
      "theoretical_loss": 4.021075875019429,
      "tokens_seen": 391970816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450551654964895,
      "loss": 2.779,
      "theoretical_loss": 4.021000409004348,
      "tokens_seen": 392036352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044504513540621866,
      "loss": 2.9278,
      "theoretical_loss": 4.0209249591353995,
      "tokens_seen": 392101888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450351053159479,
      "loss": 2.9617,
      "theoretical_loss": 4.020849525406431,
      "tokens_seen": 392167424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445025075225677,
      "loss": 2.8302,
      "theoretical_loss": 4.020774107811294,
      "tokens_seen": 392232960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044501504513540626,
      "loss": 2.8883,
      "theoretical_loss": 4.020698706343844,
      "tokens_seen": 392298496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450050150451354,
      "loss": 2.8845,
      "theoretical_loss": 4.020623320997937,
      "tokens_seen": 392364032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449949849548646,
      "loss": 2.9504,
      "theoretical_loss": 4.020547951767435,
      "tokens_seen": 392429568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449849548645938,
      "loss": 2.6795,
      "theoretical_loss": 4.020472598646203,
      "tokens_seen": 392495104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444974924774323,
      "loss": 2.8708,
      "theoretical_loss": 4.020397261628108,
      "tokens_seen": 392560640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044496489468405216,
      "loss": 2.8069,
      "theoretical_loss": 4.020321940707021,
      "tokens_seen": 392626176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449548645937814,
      "loss": 2.8997,
      "theoretical_loss": 4.020246635876817,
      "tokens_seen": 392691712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449448345035105,
      "loss": 2.9409,
      "theoretical_loss": 4.020171347131376,
      "tokens_seen": 392757248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044493480441323976,
      "loss": 2.8469,
      "theoretical_loss": 4.0200960744645755,
      "tokens_seen": 392822784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449247743229689,
      "loss": 2.9276,
      "theoretical_loss": 4.020020817870303,
      "tokens_seen": 392888320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449147442326981,
      "loss": 3.0255,
      "theoretical_loss": 4.019945577342445,
      "tokens_seen": 392953856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449047141424273,
      "loss": 2.949,
      "theoretical_loss": 4.019870352874894,
      "tokens_seen": 393019392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448946840521565,
      "loss": 2.9662,
      "theoretical_loss": 4.019795144461543,
      "tokens_seen": 393084928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044488465396188567,
      "loss": 2.9866,
      "theoretical_loss": 4.019719952096294,
      "tokens_seen": 393150464
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 228118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.819366931915283,
      "objective/train/theoretical_loss": 4.019644775773043,
      "objective/train/tokens_used": 413676000,
      "theoretical_loss": 4.019644775773043,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044487462387161485,
      "loss": 2.8016,
      "theoretical_loss": 4.019644775773043,
      "tokens_seen": 393216000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044486459378134403,
      "loss": 3.015,
      "theoretical_loss": 4.019569615485698,
      "tokens_seen": 393281536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044485456369107326,
      "loss": 2.7138,
      "theoretical_loss": 4.019494471228166,
      "tokens_seen": 393347072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448445336008024,
      "loss": 3.0278,
      "theoretical_loss": 4.019419342994358,
      "tokens_seen": 393412608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448345035105316,
      "loss": 2.8665,
      "theoretical_loss": 4.01934423077819,
      "tokens_seen": 393478144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044482447342026075,
      "loss": 3.0124,
      "theoretical_loss": 4.019269134573579,
      "tokens_seen": 393543680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044481444332999,
      "loss": 2.9158,
      "theoretical_loss": 4.019194054374447,
      "tokens_seen": 393609216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044480441323971917,
      "loss": 3.0232,
      "theoretical_loss": 4.019118990174718,
      "tokens_seen": 393674752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044479438314944835,
      "loss": 2.8318,
      "theoretical_loss": 4.019043941968319,
      "tokens_seen": 393740288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044478435305917753,
      "loss": 2.8996,
      "theoretical_loss": 4.018968909749184,
      "tokens_seen": 393805824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044477432296890677,
      "loss": 2.8596,
      "theoretical_loss": 4.018893893511246,
      "tokens_seen": 393871360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447642928786359,
      "loss": 3.0738,
      "theoretical_loss": 4.018818893248442,
      "tokens_seen": 393936896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044475426278836513,
      "loss": 2.8301,
      "theoretical_loss": 4.018743908954715,
      "tokens_seen": 394002432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044474423269809426,
      "loss": 2.8219,
      "theoretical_loss": 4.018668940624007,
      "tokens_seen": 394067968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447342026078235,
      "loss": 2.9127,
      "theoretical_loss": 4.018593988250269,
      "tokens_seen": 394133504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044472417251755267,
      "loss": 2.9686,
      "theoretical_loss": 4.01851905182745,
      "tokens_seen": 394199040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044471414242728185,
      "loss": 2.9248,
      "theoretical_loss": 4.018444131349505,
      "tokens_seen": 394264576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044470411233701103,
      "loss": 2.8442,
      "theoretical_loss": 4.018369226810392,
      "tokens_seen": 394330112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446940822467402,
      "loss": 2.9457,
      "theoretical_loss": 4.018294338204071,
      "tokens_seen": 394395648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446840521564694,
      "loss": 2.7568,
      "theoretical_loss": 4.018219465524508,
      "tokens_seen": 394461184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044467402206619863,
      "loss": 2.8902,
      "theoretical_loss": 4.0181446087656685,
      "tokens_seen": 394526720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044466399197592776,
      "loss": 2.8016,
      "theoretical_loss": 4.018069767921524,
      "tokens_seen": 394592256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444653961885657,
      "loss": 3.0541,
      "theoretical_loss": 4.01799494298605,
      "tokens_seen": 394657792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446439317953861,
      "loss": 2.636,
      "theoretical_loss": 4.017920133953223,
      "tokens_seen": 394723328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044463390170511536,
      "loss": 2.9731,
      "theoretical_loss": 4.017845340817024,
      "tokens_seen": 394788864
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 228787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5924181938171387,
      "objective/train/theoretical_loss": 4.017770563571435,
      "objective/train/tokens_used": 415314400,
      "theoretical_loss": 4.017770563571435,
      "tokens_seen": 394854400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044462387161484454,
      "loss": 2.844,
      "theoretical_loss": 4.017770563571435,
      "tokens_seen": 394854400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446138415245737,
      "loss": 2.8228,
      "theoretical_loss": 4.017695802210447,
      "tokens_seen": 394919936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446038114343029,
      "loss": 2.9139,
      "theoretical_loss": 4.017621056728047,
      "tokens_seen": 394985472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044459378134403214,
      "loss": 3.0017,
      "theoretical_loss": 4.017546327118231,
      "tokens_seen": 395051008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044458375125376126,
      "loss": 2.9105,
      "theoretical_loss": 4.017471613374996,
      "tokens_seen": 395116544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445737211634905,
      "loss": 2.7957,
      "theoretical_loss": 4.01739691549234,
      "tokens_seen": 395182080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445636910732196,
      "loss": 2.9505,
      "theoretical_loss": 4.017322233464269,
      "tokens_seen": 395247616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044455366098294886,
      "loss": 2.8694,
      "theoretical_loss": 4.017247567284789,
      "tokens_seen": 395313152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044454363089267804,
      "loss": 2.6045,
      "theoretical_loss": 4.017172916947911,
      "tokens_seen": 395378688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445336008024072,
      "loss": 2.9287,
      "theoretical_loss": 4.017098282447646,
      "tokens_seen": 395444224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445235707121364,
      "loss": 2.9332,
      "theoretical_loss": 4.017023663778014,
      "tokens_seen": 395509760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445135406218656,
      "loss": 2.7736,
      "theoretical_loss": 4.016949060933033,
      "tokens_seen": 395575296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044450351053159477,
      "loss": 2.9836,
      "theoretical_loss": 4.016874473906726,
      "tokens_seen": 395640832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444493480441324,
      "loss": 2.7708,
      "theoretical_loss": 4.016799902693119,
      "tokens_seen": 395706368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044448345035105313,
      "loss": 2.9282,
      "theoretical_loss": 4.016725347286243,
      "tokens_seen": 395771904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044447342026078236,
      "loss": 2.9519,
      "theoretical_loss": 4.01665080768013,
      "tokens_seen": 395837440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444633901705115,
      "loss": 2.6818,
      "theoretical_loss": 4.016576283868816,
      "tokens_seen": 395902976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444533600802407,
      "loss": 2.7103,
      "theoretical_loss": 4.016501775846342,
      "tokens_seen": 395968512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044444332998996996,
      "loss": 2.9564,
      "theoretical_loss": 4.016427283606747,
      "tokens_seen": 396034048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444332998996991,
      "loss": 2.9737,
      "theoretical_loss": 4.01635280714408,
      "tokens_seen": 396099584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444232698094283,
      "loss": 2.8439,
      "theoretical_loss": 4.016278346452389,
      "tokens_seen": 396165120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444132397191575,
      "loss": 2.7579,
      "theoretical_loss": 4.016203901525726,
      "tokens_seen": 396230656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444032096288867,
      "loss": 2.9436,
      "theoretical_loss": 4.0161294723581475,
      "tokens_seen": 396296192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044439317953861587,
      "loss": 2.9978,
      "theoretical_loss": 4.016055058943712,
      "tokens_seen": 396361728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044438314944834505,
      "loss": 2.9448,
      "theoretical_loss": 4.01598066127648,
      "tokens_seen": 396427264
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 229924,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7313523292541504,
      "objective/train/theoretical_loss": 4.015906279350517,
      "objective/train/tokens_used": 416952800,
      "theoretical_loss": 4.015906279350517,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044437311935807423,
      "loss": 2.895,
      "theoretical_loss": 4.015906279350517,
      "tokens_seen": 396492800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044436308926780346,
      "loss": 2.8399,
      "theoretical_loss": 4.015831913159893,
      "tokens_seen": 396558336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443530591775326,
      "loss": 3.0352,
      "theoretical_loss": 4.015757562698678,
      "tokens_seen": 396623872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443430290872618,
      "loss": 2.971,
      "theoretical_loss": 4.015683227960948,
      "tokens_seen": 396689408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044433299899699095,
      "loss": 2.7352,
      "theoretical_loss": 4.015608908940781,
      "tokens_seen": 396754944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443229689067202,
      "loss": 2.7234,
      "theoretical_loss": 4.015534605632256,
      "tokens_seen": 396820480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044431293881644937,
      "loss": 2.9999,
      "theoretical_loss": 4.01546031802946,
      "tokens_seen": 396886016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044430290872617855,
      "loss": 3.0871,
      "theoretical_loss": 4.01538604612648,
      "tokens_seen": 396951552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044429287863590773,
      "loss": 2.8029,
      "theoretical_loss": 4.015311789917406,
      "tokens_seen": 397017088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044428284854563697,
      "loss": 2.9118,
      "theoretical_loss": 4.015237549396333,
      "tokens_seen": 397082624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442728184553661,
      "loss": 2.869,
      "theoretical_loss": 4.015163324557358,
      "tokens_seen": 397148160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044426278836509533,
      "loss": 3.1997,
      "theoretical_loss": 4.015089115394581,
      "tokens_seen": 397213696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044425275827482446,
      "loss": 2.8647,
      "theoretical_loss": 4.015014921902106,
      "tokens_seen": 397279232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442427281845537,
      "loss": 2.787,
      "theoretical_loss": 4.014940744074039,
      "tokens_seen": 397344768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044423269809428287,
      "loss": 2.9288,
      "theoretical_loss": 4.014866581904491,
      "tokens_seen": 397410304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044422266800401205,
      "loss": 2.8941,
      "theoretical_loss": 4.014792435387575,
      "tokens_seen": 397475840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044421263791374123,
      "loss": 3.0604,
      "theoretical_loss": 4.014718304517408,
      "tokens_seen": 397541376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442026078234704,
      "loss": 2.9054,
      "theoretical_loss": 4.0146441892881075,
      "tokens_seen": 397606912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441925777331996,
      "loss": 3.0049,
      "theoretical_loss": 4.014570089693798,
      "tokens_seen": 397672448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044418254764292883,
      "loss": 3.119,
      "theoretical_loss": 4.014496005728605,
      "tokens_seen": 397737984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044417251755265796,
      "loss": 2.8411,
      "theoretical_loss": 4.014421937386658,
      "tokens_seen": 397803520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441624874623872,
      "loss": 3.0238,
      "theoretical_loss": 4.014347884662088,
      "tokens_seen": 397869056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441524573721163,
      "loss": 2.7703,
      "theoretical_loss": 4.014273847549032,
      "tokens_seen": 397934592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044414242728184556,
      "loss": 2.7537,
      "theoretical_loss": 4.014199826041628,
      "tokens_seen": 398000128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044413239719157474,
      "loss": 2.898,
      "theoretical_loss": 4.014125820134016,
      "tokens_seen": 398065664
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 231183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1082088947296143,
      "objective/train/theoretical_loss": 4.014051829820343,
      "objective/train/tokens_used": 418591200,
      "theoretical_loss": 4.014051829820343,
      "tokens_seen": 398131200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441223671013039,
      "loss": 3.0483,
      "theoretical_loss": 4.014051829820343,
      "tokens_seen": 398131200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441123370110331,
      "loss": 2.9731,
      "theoretical_loss": 4.0139778550947565,
      "tokens_seen": 398196736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044410230692076234,
      "loss": 2.8605,
      "theoretical_loss": 4.013903895951408,
      "tokens_seen": 398262272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044409227683049146,
      "loss": 2.9152,
      "theoretical_loss": 4.013829952384452,
      "tokens_seen": 398327808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440822467402207,
      "loss": 3.0129,
      "theoretical_loss": 4.013756024388045,
      "tokens_seen": 398393344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440722166499498,
      "loss": 2.9997,
      "theoretical_loss": 4.013682111956348,
      "tokens_seen": 398458880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044406218655967906,
      "loss": 2.7039,
      "theoretical_loss": 4.013608215083526,
      "tokens_seen": 398524416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044405215646940824,
      "loss": 2.8974,
      "theoretical_loss": 4.0135343337637455,
      "tokens_seen": 398589952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440421263791374,
      "loss": 2.7087,
      "theoretical_loss": 4.013460467991176,
      "tokens_seen": 398655488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440320962888666,
      "loss": 2.8162,
      "theoretical_loss": 4.01338661775999,
      "tokens_seen": 398721024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440220661985958,
      "loss": 3.114,
      "theoretical_loss": 4.013312783064366,
      "tokens_seen": 398786560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044401203610832497,
      "loss": 2.8509,
      "theoretical_loss": 4.013238963898483,
      "tokens_seen": 398852096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440020060180542,
      "loss": 2.6743,
      "theoretical_loss": 4.0131651602565235,
      "tokens_seen": 398917632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044399197592778333,
      "loss": 2.7401,
      "theoretical_loss": 4.013091372132672,
      "tokens_seen": 398983168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044398194583751256,
      "loss": 3.0041,
      "theoretical_loss": 4.01301759952112,
      "tokens_seen": 399048704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439719157472417,
      "loss": 2.8825,
      "theoretical_loss": 4.012943842416057,
      "tokens_seen": 399114240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439618856569709,
      "loss": 2.8614,
      "theoretical_loss": 4.01287010081168,
      "tokens_seen": 399179776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439518555667001,
      "loss": 2.7441,
      "theoretical_loss": 4.012796374702187,
      "tokens_seen": 399245312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439418254764293,
      "loss": 2.7377,
      "theoretical_loss": 4.012722664081779,
      "tokens_seen": 399310848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044393179538615847,
      "loss": 2.8877,
      "theoretical_loss": 4.012648968944661,
      "tokens_seen": 399376384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004439217652958877,
      "loss": 2.9744,
      "theoretical_loss": 4.01257528928504,
      "tokens_seen": 399441920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044391173520561683,
      "loss": 2.6387,
      "theoretical_loss": 4.012501625097128,
      "tokens_seen": 399507456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044390170511534607,
      "loss": 2.8492,
      "theoretical_loss": 4.012427976375139,
      "tokens_seen": 399572992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438916750250752,
      "loss": 2.9283,
      "theoretical_loss": 4.012354343113288,
      "tokens_seen": 399638528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044388164493480443,
      "loss": 2.8775,
      "theoretical_loss": 4.012280725305798,
      "tokens_seen": 399704064
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 231706,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.14528226852417,
      "objective/train/theoretical_loss": 4.01220712294689,
      "objective/train/tokens_used": 420229600,
      "theoretical_loss": 4.01220712294689,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438716148445336,
      "loss": 3.0579,
      "theoretical_loss": 4.01220712294689,
      "tokens_seen": 399769600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438615847542628,
      "loss": 2.7976,
      "theoretical_loss": 4.012133536030791,
      "tokens_seen": 399835136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044385155466399197,
      "loss": 2.9258,
      "theoretical_loss": 4.012059964551731,
      "tokens_seen": 399900672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044384152457372115,
      "loss": 2.7009,
      "theoretical_loss": 4.011986408503942,
      "tokens_seen": 399966208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044383149448345033,
      "loss": 2.7834,
      "theoretical_loss": 4.0119128678816605,
      "tokens_seen": 400031744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044382146439317957,
      "loss": 2.9803,
      "theoretical_loss": 4.011839342679125,
      "tokens_seen": 400097280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438114343029087,
      "loss": 2.6878,
      "theoretical_loss": 4.0117658328905765,
      "tokens_seen": 400162816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044380140421263793,
      "loss": 2.8165,
      "theoretical_loss": 4.011692338510262,
      "tokens_seen": 400228352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437913741223671,
      "loss": 2.6641,
      "theoretical_loss": 4.011618859532426,
      "tokens_seen": 400293888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437813440320963,
      "loss": 2.8163,
      "theoretical_loss": 4.011545395951323,
      "tokens_seen": 400359424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437713139418255,
      "loss": 2.6949,
      "theoretical_loss": 4.011471947761207,
      "tokens_seen": 400424960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044376128385155466,
      "loss": 3.0059,
      "theoretical_loss": 4.011398514956333,
      "tokens_seen": 400490496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044375125376128384,
      "loss": 2.8145,
      "theoretical_loss": 4.011325097530963,
      "tokens_seen": 400556032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044374122367101307,
      "loss": 2.6344,
      "theoretical_loss": 4.011251695479361,
      "tokens_seen": 400621568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437311935807422,
      "loss": 3.0702,
      "theoretical_loss": 4.011178308795792,
      "tokens_seen": 400687104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044372116349047143,
      "loss": 2.8061,
      "theoretical_loss": 4.011104937474527,
      "tokens_seen": 400752640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044371113340020056,
      "loss": 2.6934,
      "theoretical_loss": 4.011031581509838,
      "tokens_seen": 400818176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437011033099298,
      "loss": 2.8448,
      "theoretical_loss": 4.0109582408960005,
      "tokens_seen": 400883712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044369107321965903,
      "loss": 3.0339,
      "theoretical_loss": 4.010884915627294,
      "tokens_seen": 400949248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044368104312938816,
      "loss": 2.77,
      "theoretical_loss": 4.0108116056980005,
      "tokens_seen": 401014784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436710130391174,
      "loss": 2.9177,
      "theoretical_loss": 4.010738311102405,
      "tokens_seen": 401080320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436609829488465,
      "loss": 2.8799,
      "theoretical_loss": 4.010665031834796,
      "tokens_seen": 401145856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044365095285857576,
      "loss": 2.6816,
      "theoretical_loss": 4.010591767889462,
      "tokens_seen": 401211392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044364092276830494,
      "loss": 2.627,
      "theoretical_loss": 4.0105185192607,
      "tokens_seen": 401276928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436308926780341,
      "loss": 2.8835,
      "theoretical_loss": 4.010445285942806,
      "tokens_seen": 401342464
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 232929,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.049757242202759,
      "objective/train/theoretical_loss": 4.0103720679300805,
      "objective/train/tokens_used": 421868000,
      "theoretical_loss": 4.0103720679300805,
      "tokens_seen": 401408000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436208625877633,
      "loss": 2.743,
      "theoretical_loss": 4.0103720679300805,
      "tokens_seen": 401408000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044361083249749254,
      "loss": 2.7237,
      "theoretical_loss": 4.010298865216827,
      "tokens_seen": 401473536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044360080240722166,
      "loss": 2.8752,
      "theoretical_loss": 4.010225677797351,
      "tokens_seen": 401539072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435907723169509,
      "loss": 2.9571,
      "theoretical_loss": 4.010152505665964,
      "tokens_seen": 401604608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044358074222668,
      "loss": 2.823,
      "theoretical_loss": 4.010079348816976,
      "tokens_seen": 401670144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044357071213640926,
      "loss": 2.8259,
      "theoretical_loss": 4.010006207244704,
      "tokens_seen": 401735680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044356068204613844,
      "loss": 2.7683,
      "theoretical_loss": 4.009933080943466,
      "tokens_seen": 401801216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435506519558676,
      "loss": 3.0118,
      "theoretical_loss": 4.009859969907584,
      "tokens_seen": 401866752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435406218655968,
      "loss": 2.8564,
      "theoretical_loss": 4.0097868741313825,
      "tokens_seen": 401932288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443530591775326,
      "loss": 2.7656,
      "theoretical_loss": 4.009713793609189,
      "tokens_seen": 401997824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044352056168505517,
      "loss": 2.6552,
      "theoretical_loss": 4.009640728335333,
      "tokens_seen": 402063360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435105315947844,
      "loss": 2.995,
      "theoretical_loss": 4.009567678304151,
      "tokens_seen": 402128896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044350050150451353,
      "loss": 2.8946,
      "theoretical_loss": 4.009494643509978,
      "tokens_seen": 402194432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044349047141424276,
      "loss": 2.6877,
      "theoretical_loss": 4.009421623947153,
      "tokens_seen": 402259968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434804413239719,
      "loss": 2.9107,
      "theoretical_loss": 4.00934861961002,
      "tokens_seen": 402325504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434704112337011,
      "loss": 3.044,
      "theoretical_loss": 4.009275630492926,
      "tokens_seen": 402391040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434603811434303,
      "loss": 2.9364,
      "theoretical_loss": 4.009202656590217,
      "tokens_seen": 402456576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434503510531595,
      "loss": 2.9931,
      "theoretical_loss": 4.009129697896247,
      "tokens_seen": 402522112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044344032096288867,
      "loss": 2.7369,
      "theoretical_loss": 4.00905675440537,
      "tokens_seen": 402587648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434302908726179,
      "loss": 2.7316,
      "theoretical_loss": 4.008983826111945,
      "tokens_seen": 402653184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044342026078234703,
      "loss": 2.8793,
      "theoretical_loss": 4.008910913010331,
      "tokens_seen": 402718720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044341023069207627,
      "loss": 2.7036,
      "theoretical_loss": 4.008838015094893,
      "tokens_seen": 402784256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434002006018054,
      "loss": 2.7359,
      "theoretical_loss": 4.008765132359999,
      "tokens_seen": 402849792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044339017051153463,
      "loss": 2.8239,
      "theoretical_loss": 4.0086922648000165,
      "tokens_seen": 402915328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004433801404212638,
      "loss": 3.0154,
      "theoretical_loss": 4.00861941240932,
      "tokens_seen": 402980864
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 233431,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7376959323883057,
      "objective/train/theoretical_loss": 4.008546575182286,
      "objective/train/tokens_used": 423506400,
      "theoretical_loss": 4.008546575182286,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443370110330993,
      "loss": 2.7459,
      "theoretical_loss": 4.008546575182286,
      "tokens_seen": 403046400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044336008024072217,
      "loss": 2.8226,
      "theoretical_loss": 4.008473753113292,
      "tokens_seen": 403111936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044335005015045135,
      "loss": 2.8344,
      "theoretical_loss": 4.008400946196721,
      "tokens_seen": 403177472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044334002006018053,
      "loss": 3.1402,
      "theoretical_loss": 4.008328154426956,
      "tokens_seen": 403243008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044332998996990977,
      "loss": 2.8666,
      "theoretical_loss": 4.008255377798388,
      "tokens_seen": 403308544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004433199598796389,
      "loss": 2.8195,
      "theoretical_loss": 4.008182616305406,
      "tokens_seen": 403374080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044330992978936813,
      "loss": 2.707,
      "theoretical_loss": 4.008109869942404,
      "tokens_seen": 403439616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432998996990973,
      "loss": 2.7543,
      "theoretical_loss": 4.008037138703779,
      "tokens_seen": 403505152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432898696088265,
      "loss": 3.1115,
      "theoretical_loss": 4.007964422583931,
      "tokens_seen": 403570688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432798395185557,
      "loss": 2.8065,
      "theoretical_loss": 4.007891721577263,
      "tokens_seen": 403636224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044326980942828486,
      "loss": 2.8883,
      "theoretical_loss": 4.00781903567818,
      "tokens_seen": 403701760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044325977933801404,
      "loss": 2.7961,
      "theoretical_loss": 4.0077463648810925,
      "tokens_seen": 403767296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044324974924774327,
      "loss": 3.0498,
      "theoretical_loss": 4.007673709180411,
      "tokens_seen": 403832832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432397191574724,
      "loss": 2.8936,
      "theoretical_loss": 4.007601068570549,
      "tokens_seen": 403898368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044322968906720163,
      "loss": 2.7508,
      "theoretical_loss": 4.007528443045928,
      "tokens_seen": 403963904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044321965897693076,
      "loss": 3.0503,
      "theoretical_loss": 4.007455832600965,
      "tokens_seen": 404029440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044320962888666,
      "loss": 3.0522,
      "theoretical_loss": 4.007383237230086,
      "tokens_seen": 404094976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431995987963892,
      "loss": 2.5872,
      "theoretical_loss": 4.007310656927716,
      "tokens_seen": 404160512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044318956870611836,
      "loss": 2.891,
      "theoretical_loss": 4.0072380916882855,
      "tokens_seen": 404226048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044317953861584754,
      "loss": 3.0283,
      "theoretical_loss": 4.007165541506227,
      "tokens_seen": 404291584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431695085255767,
      "loss": 2.9584,
      "theoretical_loss": 4.007093006375976,
      "tokens_seen": 404357120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431594784353059,
      "loss": 2.9348,
      "theoretical_loss": 4.007020486291972,
      "tokens_seen": 404422656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044314944834503514,
      "loss": 2.949,
      "theoretical_loss": 4.006947981248654,
      "tokens_seen": 404488192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044313941825476427,
      "loss": 2.7082,
      "theoretical_loss": 4.006875491240468,
      "tokens_seen": 404553728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431293881644935,
      "loss": 2.7423,
      "theoretical_loss": 4.006803016261861,
      "tokens_seen": 404619264
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 234514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.404675245285034,
      "objective/train/theoretical_loss": 4.006730556307284,
      "objective/train/tokens_used": 425144800,
      "theoretical_loss": 4.006730556307284,
      "tokens_seen": 404684800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431193580742227,
      "loss": 3.1039,
      "theoretical_loss": 4.006730556307284,
      "tokens_seen": 404684800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044310932798395186,
      "loss": 2.6791,
      "theoretical_loss": 4.006658111371189,
      "tokens_seen": 404750336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044309929789368104,
      "loss": 2.6862,
      "theoretical_loss": 4.006585681448033,
      "tokens_seen": 404815872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430892678034102,
      "loss": 2.8161,
      "theoretical_loss": 4.006513266532275,
      "tokens_seen": 404881408
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430792377131394,
      "loss": 2.8454,
      "theoretical_loss": 4.006440866618378,
      "tokens_seen": 404946944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044306920762286864,
      "loss": 2.9207,
      "theoretical_loss": 4.006368481700806,
      "tokens_seen": 405012480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044305917753259777,
      "loss": 2.7304,
      "theoretical_loss": 4.0062961117740254,
      "tokens_seen": 405078016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443049147442327,
      "loss": 2.8102,
      "theoretical_loss": 4.00622375683251,
      "tokens_seen": 405143552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044303911735205613,
      "loss": 2.7443,
      "theoretical_loss": 4.006151416870731,
      "tokens_seen": 405209088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044302908726178537,
      "loss": 2.8585,
      "theoretical_loss": 4.0060790918831675,
      "tokens_seen": 405274624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044301905717151455,
      "loss": 2.832,
      "theoretical_loss": 4.006006781864298,
      "tokens_seen": 405340160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044300902708124373,
      "loss": 2.8704,
      "theoretical_loss": 4.005934486808605,
      "tokens_seen": 405405696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429989969909729,
      "loss": 2.8721,
      "theoretical_loss": 4.0058622067105745,
      "tokens_seen": 405471232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429889669007021,
      "loss": 3.0217,
      "theoretical_loss": 4.005789941564695,
      "tokens_seen": 405536768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044297893681043127,
      "loss": 2.781,
      "theoretical_loss": 4.005717691365457,
      "tokens_seen": 405602304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429689067201605,
      "loss": 2.7174,
      "theoretical_loss": 4.005645456107356,
      "tokens_seen": 405667840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044295887662988963,
      "loss": 2.5529,
      "theoretical_loss": 4.005573235784888,
      "tokens_seen": 405733376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044294884653961887,
      "loss": 2.9012,
      "theoretical_loss": 4.005501030392553,
      "tokens_seen": 405798912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429388164493481,
      "loss": 3.0316,
      "theoretical_loss": 4.005428839924855,
      "tokens_seen": 405864448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044292878635907723,
      "loss": 2.766,
      "theoretical_loss": 4.0053566643763014,
      "tokens_seen": 405929984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044291875626880647,
      "loss": 2.9476,
      "theoretical_loss": 4.0052845037413976,
      "tokens_seen": 405995520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429087261785356,
      "loss": 3.0683,
      "theoretical_loss": 4.005212358014658,
      "tokens_seen": 406061056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044289869608826483,
      "loss": 2.8003,
      "theoretical_loss": 4.005140227190596,
      "tokens_seen": 406126592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442888665997994,
      "loss": 2.7452,
      "theoretical_loss": 4.00506811126373,
      "tokens_seen": 406192128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428786359077232,
      "loss": 2.8949,
      "theoretical_loss": 4.00499601022858,
      "tokens_seen": 406257664
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 235119,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8222172260284424,
      "objective/train/theoretical_loss": 4.0049239240796695,
      "objective/train/tokens_used": 426783200,
      "theoretical_loss": 4.0049239240796695,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044286860581745237,
      "loss": 2.999,
      "theoretical_loss": 4.0049239240796695,
      "tokens_seen": 406323200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044285857572718155,
      "loss": 2.8565,
      "theoretical_loss": 4.004851852811523,
      "tokens_seen": 406388736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044284854563691073,
      "loss": 2.8309,
      "theoretical_loss": 4.004779796418673,
      "tokens_seen": 406454272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044283851554663997,
      "loss": 2.713,
      "theoretical_loss": 4.00470775489565,
      "tokens_seen": 406519808
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428284854563691,
      "loss": 2.9481,
      "theoretical_loss": 4.004635728236988,
      "tokens_seen": 406585344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044281845536609833,
      "loss": 2.8521,
      "theoretical_loss": 4.004563716437227,
      "tokens_seen": 406650880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428084252758275,
      "loss": 2.8142,
      "theoretical_loss": 4.004491719490905,
      "tokens_seen": 406716416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427983951855567,
      "loss": 2.6657,
      "theoretical_loss": 4.004419737392568,
      "tokens_seen": 406781952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427883650952859,
      "loss": 2.8134,
      "theoretical_loss": 4.004347770136761,
      "tokens_seen": 406847488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044277833500501506,
      "loss": 2.6711,
      "theoretical_loss": 4.0042758177180335,
      "tokens_seen": 406913024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044276830491474424,
      "loss": 2.8276,
      "theoretical_loss": 4.004203880130939,
      "tokens_seen": 406978560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044275827482447347,
      "loss": 2.881,
      "theoretical_loss": 4.004131957370031,
      "tokens_seen": 407044096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427482447342026,
      "loss": 2.9221,
      "theoretical_loss": 4.004060049429869,
      "tokens_seen": 407109632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044273821464393184,
      "loss": 2.7697,
      "theoretical_loss": 4.003988156305013,
      "tokens_seen": 407175168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044272818455366096,
      "loss": 2.764,
      "theoretical_loss": 4.003916277990028,
      "tokens_seen": 407240704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427181544633902,
      "loss": 2.8386,
      "theoretical_loss": 4.003844414479477,
      "tokens_seen": 407306240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427081243731194,
      "loss": 2.9933,
      "theoretical_loss": 4.003772565767934,
      "tokens_seen": 407371776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044269809428284856,
      "loss": 2.7289,
      "theoretical_loss": 4.003700731849969,
      "tokens_seen": 407437312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044268806419257774,
      "loss": 2.9191,
      "theoretical_loss": 4.003628912720157,
      "tokens_seen": 407502848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426780341023069,
      "loss": 2.6767,
      "theoretical_loss": 4.003557108373077,
      "tokens_seen": 407568384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426680040120361,
      "loss": 2.907,
      "theoretical_loss": 4.00348531880331,
      "tokens_seen": 407633920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044265797392176534,
      "loss": 2.7716,
      "theoretical_loss": 4.003413544005439,
      "tokens_seen": 407699456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044264794383149447,
      "loss": 2.8966,
      "theoretical_loss": 4.003341783974051,
      "tokens_seen": 407764992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426379137412237,
      "loss": 2.8671,
      "theoretical_loss": 4.003270038703736,
      "tokens_seen": 407830528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426278836509529,
      "loss": 2.6621,
      "theoretical_loss": 4.003198308189086,
      "tokens_seen": 407896064
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 236101,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.368303060531616,
      "objective/train/theoretical_loss": 4.003126592424696,
      "objective/train/tokens_used": 428421600,
      "theoretical_loss": 4.003126592424696,
      "tokens_seen": 407961600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044261785356068206,
      "loss": 2.6108,
      "theoretical_loss": 4.003126592424696,
      "tokens_seen": 407961600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044260782347041124,
      "loss": 2.7856,
      "theoretical_loss": 4.003054891405164,
      "tokens_seen": 408027136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425977933801404,
      "loss": 2.846,
      "theoretical_loss": 4.002983205125091,
      "tokens_seen": 408092672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425877632898696,
      "loss": 2.5993,
      "theoretical_loss": 4.00291153357908,
      "tokens_seen": 408158208
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044257773319959884,
      "loss": 2.9444,
      "theoretical_loss": 4.002839876761739,
      "tokens_seen": 408223744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044256770310932797,
      "loss": 2.8485,
      "theoretical_loss": 4.002768234667677,
      "tokens_seen": 408289280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425576730190572,
      "loss": 2.812,
      "theoretical_loss": 4.002696607291506,
      "tokens_seen": 408354816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044254764292878633,
      "loss": 2.9355,
      "theoretical_loss": 4.0026249946278405,
      "tokens_seen": 408420352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044253761283851557,
      "loss": 2.7608,
      "theoretical_loss": 4.0025533966713,
      "tokens_seen": 408485888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044252758274824475,
      "loss": 2.7582,
      "theoretical_loss": 4.002481813416504,
      "tokens_seen": 408551424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044251755265797393,
      "loss": 2.8645,
      "theoretical_loss": 4.002410244858076,
      "tokens_seen": 408616960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425075225677031,
      "loss": 2.9369,
      "theoretical_loss": 4.002338690990644,
      "tokens_seen": 408682496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424974924774323,
      "loss": 2.7085,
      "theoretical_loss": 4.0022671518088355,
      "tokens_seen": 408748032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044248746238716147,
      "loss": 2.7926,
      "theoretical_loss": 4.002195627307283,
      "tokens_seen": 408813568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424774322968907,
      "loss": 2.7285,
      "theoretical_loss": 4.002124117480621,
      "tokens_seen": 408879104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044246740220661983,
      "loss": 2.8286,
      "theoretical_loss": 4.002052622323489,
      "tokens_seen": 408944640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044245737211634907,
      "loss": 3.0013,
      "theoretical_loss": 4.001981141830525,
      "tokens_seen": 409010176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044244734202607825,
      "loss": 2.5402,
      "theoretical_loss": 4.001909675996375,
      "tokens_seen": 409075712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044243731193580743,
      "loss": 2.7802,
      "theoretical_loss": 4.001838224815683,
      "tokens_seen": 409141248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424272818455366,
      "loss": 2.6874,
      "theoretical_loss": 4.001766788283098,
      "tokens_seen": 409206784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424172517552658,
      "loss": 2.8716,
      "theoretical_loss": 4.001695366393275,
      "tokens_seen": 409272320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442407221664995,
      "loss": 2.636,
      "theoretical_loss": 4.001623959140864,
      "tokens_seen": 409337856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423971915747242,
      "loss": 3.065,
      "theoretical_loss": 4.001552566520525,
      "tokens_seen": 409403392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044238716148445334,
      "loss": 2.9232,
      "theoretical_loss": 4.001481188526918,
      "tokens_seen": 409468928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044237713139418257,
      "loss": 2.7951,
      "theoretical_loss": 4.001409825154705,
      "tokens_seen": 409534464
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 237324,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1303606033325195,
      "objective/train/theoretical_loss": 4.001338476398553,
      "objective/train/tokens_used": 430060000,
      "theoretical_loss": 4.001338476398553,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423671013039117,
      "loss": 3.0079,
      "theoretical_loss": 4.001338476398553,
      "tokens_seen": 409600000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044235707121364093,
      "loss": 2.9637,
      "theoretical_loss": 4.0012671422531305,
      "tokens_seen": 409665536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423470411233701,
      "loss": 2.6292,
      "theoretical_loss": 4.001195822713108,
      "tokens_seen": 409731072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423370110330993,
      "loss": 2.883,
      "theoretical_loss": 4.001124517773161,
      "tokens_seen": 409796608
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423269809428285,
      "loss": 2.99,
      "theoretical_loss": 4.001053227427966,
      "tokens_seen": 409862144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423169508525577,
      "loss": 2.7892,
      "theoretical_loss": 4.000981951672202,
      "tokens_seen": 409927680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044230692076228684,
      "loss": 2.9052,
      "theoretical_loss": 4.000910690500552,
      "tokens_seen": 409993216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422968906720161,
      "loss": 3.0222,
      "theoretical_loss": 4.000839443907702,
      "tokens_seen": 410058752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422868605817452,
      "loss": 2.8375,
      "theoretical_loss": 4.000768211888339,
      "tokens_seen": 410124288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044227683049147444,
      "loss": 2.6957,
      "theoretical_loss": 4.000696994437154,
      "tokens_seen": 410189824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422668004012036,
      "loss": 2.7352,
      "theoretical_loss": 4.000625791548843,
      "tokens_seen": 410255360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422567703109328,
      "loss": 2.7709,
      "theoretical_loss": 4.0005546032181,
      "tokens_seen": 410320896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442246740220662,
      "loss": 2.9533,
      "theoretical_loss": 4.000483429439624,
      "tokens_seen": 410386432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044223671013039116,
      "loss": 2.874,
      "theoretical_loss": 4.00041227020812,
      "tokens_seen": 410451968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044222668004012034,
      "loss": 3.1148,
      "theoretical_loss": 4.00034112551829,
      "tokens_seen": 410517504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422166499498496,
      "loss": 3.0062,
      "theoretical_loss": 4.000269995364842,
      "tokens_seen": 410583040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422066198595787,
      "loss": 2.9177,
      "theoretical_loss": 4.000198879742488,
      "tokens_seen": 410648576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044219658976930794,
      "loss": 2.7881,
      "theoretical_loss": 4.000127778645939,
      "tokens_seen": 410714112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421865596790371,
      "loss": 2.7334,
      "theoretical_loss": 4.000056692069911,
      "tokens_seen": 410779648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421765295887663,
      "loss": 2.8035,
      "theoretical_loss": 3.9999856200091255,
      "tokens_seen": 410845184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044216649949849554,
      "loss": 2.8879,
      "theoretical_loss": 3.9999145624583017,
      "tokens_seen": 410910720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044215646940822467,
      "loss": 2.8508,
      "theoretical_loss": 3.9998435194121633,
      "tokens_seen": 410976256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421464393179539,
      "loss": 2.8841,
      "theoretical_loss": 3.999772490865439,
      "tokens_seen": 411041792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421364092276831,
      "loss": 2.7457,
      "theoretical_loss": 3.9997014768128567,
      "tokens_seen": 411107328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044212637913741226,
      "loss": 3.0086,
      "theoretical_loss": 3.9996304772491498,
      "tokens_seen": 411172864
    },
    {
      "epoch": 0.12,
      "objective/train/docs_used": 237892,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8793866634368896,
      "objective/train/theoretical_loss": 3.9995594921690536,
      "objective/train/tokens_used": 431698400,
      "theoretical_loss": 3.9995594921690536,
      "tokens_seen": 411238400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044211634904714144,
      "loss": 2.8024,
      "theoretical_loss": 3.9995594921690536,
      "tokens_seen": 411238400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421063189568706,
      "loss": 2.7261,
      "theoretical_loss": 3.9994885215673057,
      "tokens_seen": 411303936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420962888665998,
      "loss": 2.9451,
      "theoretical_loss": 3.9994175654386472,
      "tokens_seen": 411369472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044208625877632904,
      "loss": 2.7401,
      "theoretical_loss": 3.9993466237778215,
      "tokens_seen": 411435008
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044207622868605817,
      "loss": 2.9415,
      "theoretical_loss": 3.9992756965795744,
      "tokens_seen": 411500544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420661985957874,
      "loss": 2.8241,
      "theoretical_loss": 3.999204783838655,
      "tokens_seen": 411566080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044205616850551653,
      "loss": 2.9024,
      "theoretical_loss": 3.9991338855498153,
      "tokens_seen": 411631616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044204613841524577,
      "loss": 2.9414,
      "theoretical_loss": 3.9990630017078095,
      "tokens_seen": 411697152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044203610832497495,
      "loss": 2.8594,
      "theoretical_loss": 3.9989921323073947,
      "tokens_seen": 411762688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044202607823470413,
      "loss": 2.9609,
      "theoretical_loss": 3.998921277343331,
      "tokens_seen": 411828224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420160481444333,
      "loss": 3.0881,
      "theoretical_loss": 3.9988504368103817,
      "tokens_seen": 411893760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420060180541625,
      "loss": 2.573,
      "theoretical_loss": 3.9987796107033104,
      "tokens_seen": 411959296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044199598796389167,
      "loss": 2.9766,
      "theoretical_loss": 3.9987087990168866,
      "tokens_seen": 412024832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419859578736209,
      "loss": 2.9167,
      "theoretical_loss": 3.998638001745881,
      "tokens_seen": 412090368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044197592778335003,
      "loss": 2.6952,
      "theoretical_loss": 3.998567218885067,
      "tokens_seen": 412155904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044196589769307927,
      "loss": 2.8777,
      "theoretical_loss": 3.9984964504292204,
      "tokens_seen": 412221440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044195586760280845,
      "loss": 2.8639,
      "theoretical_loss": 3.998425696373121,
      "tokens_seen": 412286976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044194583751253763,
      "loss": 2.8543,
      "theoretical_loss": 3.998354956711551,
      "tokens_seen": 412352512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419358074222668,
      "loss": 3.0178,
      "theoretical_loss": 3.9982842314392935,
      "tokens_seen": 412418048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000441925777331996,
      "loss": 2.7787,
      "theoretical_loss": 3.9982135205511358,
      "tokens_seen": 412483584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004419157472417252,
      "loss": 3.1016,
      "theoretical_loss": 3.9981428240418686,
      "tokens_seen": 412549120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004419057171514544,
      "loss": 2.6455,
      "theoretical_loss": 3.998072141906284,
      "tokens_seen": 412614656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044189568706118354,
      "loss": 2.9787,
      "theoretical_loss": 3.998001474139178,
      "tokens_seen": 412680192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044188565697091277,
      "loss": 2.9904,
      "theoretical_loss": 3.997930820735348,
      "tokens_seen": 412745728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418756268806419,
      "loss": 2.9195,
      "theoretical_loss": 3.9978601816895942,
      "tokens_seen": 412811264
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 239244,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.011247396469116,
      "objective/train/theoretical_loss": 3.9977895569967217,
      "objective/train/tokens_used": 433336800,
      "theoretical_loss": 3.9977895569967217,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044186559679037113,
      "loss": 2.8852,
      "theoretical_loss": 3.9977895569967217,
      "tokens_seen": 412876800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418555667001003,
      "loss": 2.8744,
      "theoretical_loss": 3.997718946651535,
      "tokens_seen": 412942336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418455366098295,
      "loss": 2.5262,
      "theoretical_loss": 3.997648350648843,
      "tokens_seen": 413007872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418355065195587,
      "loss": 2.8229,
      "theoretical_loss": 3.997577768983459,
      "tokens_seen": 413073408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418254764292879,
      "loss": 2.7307,
      "theoretical_loss": 3.997507201650196,
      "tokens_seen": 413138944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044181544633901704,
      "loss": 2.87,
      "theoretical_loss": 3.9974366486438706,
      "tokens_seen": 413204480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418054162487463,
      "loss": 2.7706,
      "theoretical_loss": 3.9973661099593034,
      "tokens_seen": 413270016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417953861584754,
      "loss": 2.9375,
      "theoretical_loss": 3.9972955855913157,
      "tokens_seen": 413335552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044178535606820464,
      "loss": 2.6413,
      "theoretical_loss": 3.9972250755347334,
      "tokens_seen": 413401088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417753259779338,
      "loss": 2.741,
      "theoretical_loss": 3.997154579784384,
      "tokens_seen": 413466624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441765295887663,
      "loss": 2.8083,
      "theoretical_loss": 3.997084098335098,
      "tokens_seen": 413532160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417552657973922,
      "loss": 2.9568,
      "theoretical_loss": 3.997013631181708,
      "tokens_seen": 413597696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044174523570712136,
      "loss": 2.906,
      "theoretical_loss": 3.99694317831905,
      "tokens_seen": 413663232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044173520561685054,
      "loss": 2.9089,
      "theoretical_loss": 3.9968727397419634,
      "tokens_seen": 413728768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417251755265798,
      "loss": 3.0527,
      "theoretical_loss": 3.9968023154452883,
      "tokens_seen": 413794304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417151454363089,
      "loss": 2.7741,
      "theoretical_loss": 3.9967319054238684,
      "tokens_seen": 413859840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044170511534603814,
      "loss": 2.9615,
      "theoretical_loss": 3.996661509672551,
      "tokens_seen": 413925376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044169508525576727,
      "loss": 3.0461,
      "theoretical_loss": 3.996591128186185,
      "tokens_seen": 413990912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416850551654965,
      "loss": 2.776,
      "theoretical_loss": 3.996520760959622,
      "tokens_seen": 414056448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416750250752257,
      "loss": 2.8094,
      "theoretical_loss": 3.996450407987716,
      "tokens_seen": 414121984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044166499498495487,
      "loss": 2.9947,
      "theoretical_loss": 3.9963800692653253,
      "tokens_seen": 414187520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044165496489468405,
      "loss": 2.7043,
      "theoretical_loss": 3.9963097447873097,
      "tokens_seen": 414253056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416449348044133,
      "loss": 2.8336,
      "theoretical_loss": 3.9962394345485315,
      "tokens_seen": 414318592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416349047141424,
      "loss": 2.6923,
      "theoretical_loss": 3.9961691385438556,
      "tokens_seen": 414384128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044162487462387164,
      "loss": 2.8808,
      "theoretical_loss": 3.99609885676815,
      "tokens_seen": 414449664
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 239946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0474255084991455,
      "objective/train/theoretical_loss": 3.9960285892162855,
      "objective/train/tokens_used": 434975200,
      "theoretical_loss": 3.9960285892162855,
      "tokens_seen": 414515200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044161484453360077,
      "loss": 2.9599,
      "theoretical_loss": 3.9960285892162855,
      "tokens_seen": 414515200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044160481444333,
      "loss": 2.5837,
      "theoretical_loss": 3.995958335883135,
      "tokens_seen": 414580736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415947843530592,
      "loss": 2.6826,
      "theoretical_loss": 3.995888096763574,
      "tokens_seen": 414646272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044158475426278837,
      "loss": 2.7732,
      "theoretical_loss": 3.9958178718524824,
      "tokens_seen": 414711808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044157472417251755,
      "loss": 2.8478,
      "theoretical_loss": 3.9957476611447396,
      "tokens_seen": 414777344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044156469408224673,
      "loss": 2.7315,
      "theoretical_loss": 3.9956774646352304,
      "tokens_seen": 414842880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415546639919759,
      "loss": 2.796,
      "theoretical_loss": 3.9956072823188418,
      "tokens_seen": 414908416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044154463390170515,
      "loss": 2.9685,
      "theoretical_loss": 3.9955371141904616,
      "tokens_seen": 414973952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415346038114343,
      "loss": 2.9153,
      "theoretical_loss": 3.995466960244982,
      "tokens_seen": 415039488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415245737211635,
      "loss": 2.7571,
      "theoretical_loss": 3.995396820477298,
      "tokens_seen": 415105024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044151454363089264,
      "loss": 2.8539,
      "theoretical_loss": 3.995326694882306,
      "tokens_seen": 415170560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044150451354062187,
      "loss": 2.9172,
      "theoretical_loss": 3.9952565834549056,
      "tokens_seen": 415236096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044149448345035105,
      "loss": 2.9305,
      "theoretical_loss": 3.9951864861899997,
      "tokens_seen": 415301632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044148445336008023,
      "loss": 2.9859,
      "theoretical_loss": 3.995116403082493,
      "tokens_seen": 415367168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414744232698094,
      "loss": 2.755,
      "theoretical_loss": 3.995046334127293,
      "tokens_seen": 415432704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044146439317953865,
      "loss": 2.809,
      "theoretical_loss": 3.99497627931931,
      "tokens_seen": 415498240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414543630892678,
      "loss": 2.9747,
      "theoretical_loss": 3.994906238653457,
      "tokens_seen": 415563776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441444332998997,
      "loss": 2.8364,
      "theoretical_loss": 3.9948362121246497,
      "tokens_seen": 415629312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414343029087262,
      "loss": 2.613,
      "theoretical_loss": 3.9947661997278052,
      "tokens_seen": 415694848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414242728184554,
      "loss": 2.5021,
      "theoretical_loss": 3.9946962014578453,
      "tokens_seen": 415760384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414142427281846,
      "loss": 2.8959,
      "theoretical_loss": 3.994626217309693,
      "tokens_seen": 415825920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044140421263791374,
      "loss": 2.8188,
      "theoretical_loss": 3.994556247278275,
      "tokens_seen": 415891456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044139418254764297,
      "loss": 2.9287,
      "theoretical_loss": 3.994486291358519,
      "tokens_seen": 415956992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413841524573721,
      "loss": 2.9442,
      "theoretical_loss": 3.9944163495453564,
      "tokens_seen": 416022528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044137412236710133,
      "loss": 2.8413,
      "theoretical_loss": 3.994346421833721,
      "tokens_seen": 416088064
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 241045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6522488594055176,
      "objective/train/theoretical_loss": 3.99427650821855,
      "objective/train/tokens_used": 436613600,
      "theoretical_loss": 3.99427650821855,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413640922768305,
      "loss": 2.9212,
      "theoretical_loss": 3.99427650821855,
      "tokens_seen": 416153600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413540621865597,
      "loss": 2.8946,
      "theoretical_loss": 3.994206608694782,
      "tokens_seen": 416219136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413440320962889,
      "loss": 2.7445,
      "theoretical_loss": 3.994136723257358,
      "tokens_seen": 416284672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413340020060181,
      "loss": 2.782,
      "theoretical_loss": 3.994066851901224,
      "tokens_seen": 416350208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044132397191574724,
      "loss": 2.8814,
      "theoretical_loss": 3.993996994621326,
      "tokens_seen": 416415744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413139418254765,
      "loss": 2.82,
      "theoretical_loss": 3.9939271514126125,
      "tokens_seen": 416481280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413039117352056,
      "loss": 2.7661,
      "theoretical_loss": 3.9938573222700375,
      "tokens_seen": 416546816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044129388164493484,
      "loss": 2.9603,
      "theoretical_loss": 3.9937875071885554,
      "tokens_seen": 416612352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441283851554664,
      "loss": 3.1666,
      "theoretical_loss": 3.993717706163122,
      "tokens_seen": 416677888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412738214643932,
      "loss": 2.768,
      "theoretical_loss": 3.9936479191886995,
      "tokens_seen": 416743424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412637913741224,
      "loss": 2.9516,
      "theoretical_loss": 3.9935781462602487,
      "tokens_seen": 416808960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044125376128385156,
      "loss": 2.82,
      "theoretical_loss": 3.9935083873727355,
      "tokens_seen": 416874496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044124373119358074,
      "loss": 2.9864,
      "theoretical_loss": 3.9934386425211277,
      "tokens_seen": 416940032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044123370110331,
      "loss": 2.8075,
      "theoretical_loss": 3.9933689117003954,
      "tokens_seen": 417005568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412236710130391,
      "loss": 2.7716,
      "theoretical_loss": 3.9932991949055117,
      "tokens_seen": 417071104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044121364092276834,
      "loss": 2.6797,
      "theoretical_loss": 3.993229492131452,
      "tokens_seen": 417136640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044120361083249747,
      "loss": 2.8761,
      "theoretical_loss": 3.9931598033731945,
      "tokens_seen": 417202176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411935807422267,
      "loss": 2.9552,
      "theoretical_loss": 3.99309012862572,
      "tokens_seen": 417267712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411835506519559,
      "loss": 2.7448,
      "theoretical_loss": 3.993020467884011,
      "tokens_seen": 417333248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044117352056168507,
      "loss": 2.6875,
      "theoretical_loss": 3.9929508211430553,
      "tokens_seen": 417398784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044116349047141425,
      "loss": 2.7963,
      "theoretical_loss": 3.9928811883978392,
      "tokens_seen": 417464320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411534603811435,
      "loss": 2.988,
      "theoretical_loss": 3.9928115696433553,
      "tokens_seen": 417529856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411434302908726,
      "loss": 2.705,
      "theoretical_loss": 3.992741964874596,
      "tokens_seen": 417595392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044113340020060184,
      "loss": 2.7923,
      "theoretical_loss": 3.9926723740865584,
      "tokens_seen": 417660928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044112337011033097,
      "loss": 2.6621,
      "theoretical_loss": 3.992602797274241,
      "tokens_seen": 417726464
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 241540,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0460522174835205,
      "objective/train/theoretical_loss": 3.9925332344326447,
      "objective/train/tokens_used": 438252000,
      "theoretical_loss": 3.9925332344326447,
      "tokens_seen": 417792000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411133400200602,
      "loss": 2.5784,
      "theoretical_loss": 3.9925332344326447,
      "tokens_seen": 417792000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411033099297894,
      "loss": 3.0353,
      "theoretical_loss": 3.9924636855567743,
      "tokens_seen": 417857536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044109327983951857,
      "loss": 2.7198,
      "theoretical_loss": 3.992394150641636,
      "tokens_seen": 417923072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044108324974924775,
      "loss": 2.956,
      "theoretical_loss": 3.992324629682238,
      "tokens_seen": 417988608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044107321965897693,
      "loss": 2.4868,
      "theoretical_loss": 3.992255122673593,
      "tokens_seen": 418054144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410631895687061,
      "loss": 3.0218,
      "theoretical_loss": 3.9921856296107143,
      "tokens_seen": 418119680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044105315947843535,
      "loss": 2.8273,
      "theoretical_loss": 3.992116150488619,
      "tokens_seen": 418185216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410431293881645,
      "loss": 2.8603,
      "theoretical_loss": 3.992046685302327,
      "tokens_seen": 418250752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410330992978937,
      "loss": 2.6586,
      "theoretical_loss": 3.9919772340468596,
      "tokens_seen": 418316288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044102306920762284,
      "loss": 2.5768,
      "theoretical_loss": 3.9919077967172414,
      "tokens_seen": 418381824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044101303911735207,
      "loss": 3.0208,
      "theoretical_loss": 3.991838373308499,
      "tokens_seen": 418447360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044100300902708125,
      "loss": 3.1297,
      "theoretical_loss": 3.9917689638156624,
      "tokens_seen": 418512896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044099297893681043,
      "loss": 2.701,
      "theoretical_loss": 3.9916995682337633,
      "tokens_seen": 418578432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409829488465396,
      "loss": 2.8845,
      "theoretical_loss": 3.9916301865578365,
      "tokens_seen": 418643968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044097291875626885,
      "loss": 2.8086,
      "theoretical_loss": 3.9915608187829195,
      "tokens_seen": 418709504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440962888665998,
      "loss": 2.9099,
      "theoretical_loss": 3.991491464904052,
      "tokens_seen": 418775040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409528585757272,
      "loss": 2.9136,
      "theoretical_loss": 3.9914221249162756,
      "tokens_seen": 418840576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044094282848545634,
      "loss": 2.8021,
      "theoretical_loss": 3.9913527988146353,
      "tokens_seen": 418906112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409327983951856,
      "loss": 2.8487,
      "theoretical_loss": 3.9912834865941793,
      "tokens_seen": 418971648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044092276830491476,
      "loss": 2.9562,
      "theoretical_loss": 3.9912141882499563,
      "tokens_seen": 419037184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044091273821464394,
      "loss": 2.9475,
      "theoretical_loss": 3.99114490377702,
      "tokens_seen": 419102720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004409027081243731,
      "loss": 2.7868,
      "theoretical_loss": 3.991075633170425,
      "tokens_seen": 419168256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408926780341023,
      "loss": 2.9248,
      "theoretical_loss": 3.991006376425228,
      "tokens_seen": 419233792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408826479438315,
      "loss": 2.8554,
      "theoretical_loss": 3.99093713353649,
      "tokens_seen": 419299328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408726178535607,
      "loss": 2.6956,
      "theoretical_loss": 3.9908679044992725,
      "tokens_seen": 419364864
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 242971,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7546820640563965,
      "objective/train/theoretical_loss": 3.990798689308642,
      "objective/train/tokens_used": 439890400,
      "theoretical_loss": 3.990798689308642,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044086258776328984,
      "loss": 2.9533,
      "theoretical_loss": 3.990798689308642,
      "tokens_seen": 419430400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408525576730191,
      "loss": 3.0452,
      "theoretical_loss": 3.9907294879596655,
      "tokens_seen": 419495936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408425275827482,
      "loss": 2.8135,
      "theoretical_loss": 3.9906603004474137,
      "tokens_seen": 419561472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044083249749247744,
      "loss": 2.7665,
      "theoretical_loss": 3.990591126766958,
      "tokens_seen": 419627008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408224674022066,
      "loss": 2.6103,
      "theoretical_loss": 3.9905219669133745,
      "tokens_seen": 419692544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408124373119358,
      "loss": 3.1082,
      "theoretical_loss": 3.9904528208817407,
      "tokens_seen": 419758080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440802407221665,
      "loss": 2.7595,
      "theoretical_loss": 3.990383688667137,
      "tokens_seen": 419823616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407923771313942,
      "loss": 3.0086,
      "theoretical_loss": 3.9903145702646468,
      "tokens_seen": 419889152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044078234704112335,
      "loss": 2.8238,
      "theoretical_loss": 3.9902454656693545,
      "tokens_seen": 419954688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407723169508526,
      "loss": 2.9094,
      "theoretical_loss": 3.990176374876348,
      "tokens_seen": 420020224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407622868605817,
      "loss": 3.11,
      "theoretical_loss": 3.9901072978807184,
      "tokens_seen": 420085760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044075225677031094,
      "loss": 2.8017,
      "theoretical_loss": 3.9900382346775576,
      "tokens_seen": 420151296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407422266800401,
      "loss": 3.0109,
      "theoretical_loss": 3.9899691852619616,
      "tokens_seen": 420216832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407321965897693,
      "loss": 3.0303,
      "theoretical_loss": 3.9899001496290274,
      "tokens_seen": 420282368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407221664994985,
      "loss": 2.7646,
      "theoretical_loss": 3.989831127773856,
      "tokens_seen": 420347904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044071213640922767,
      "loss": 2.8605,
      "theoretical_loss": 3.989762119691551,
      "tokens_seen": 420413440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044070210631895685,
      "loss": 2.8231,
      "theoretical_loss": 3.9896931253772165,
      "tokens_seen": 420478976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406920762286861,
      "loss": 2.5455,
      "theoretical_loss": 3.989624144825961,
      "tokens_seen": 420544512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044068204613841527,
      "loss": 3.1163,
      "theoretical_loss": 3.9895551780328953,
      "tokens_seen": 420610048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044067201604814445,
      "loss": 2.6076,
      "theoretical_loss": 3.989486224993131,
      "tokens_seen": 420675584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406619859578737,
      "loss": 3.1176,
      "theoretical_loss": 3.9894172857017844,
      "tokens_seen": 420741120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406519558676028,
      "loss": 3.0003,
      "theoretical_loss": 3.989348360153974,
      "tokens_seen": 420806656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044064192577733204,
      "loss": 2.955,
      "theoretical_loss": 3.989279448344818,
      "tokens_seen": 420872192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044063189568706117,
      "loss": 2.8199,
      "theoretical_loss": 3.9892105502694415,
      "tokens_seen": 420937728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406218655967904,
      "loss": 2.9728,
      "theoretical_loss": 3.989141665922969,
      "tokens_seen": 421003264
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 244183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9032411575317383,
      "objective/train/theoretical_loss": 3.989072795300528,
      "objective/train/tokens_used": 441528800,
      "theoretical_loss": 3.989072795300528,
      "tokens_seen": 421068800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406118355065196,
      "loss": 2.8608,
      "theoretical_loss": 3.989072795300528,
      "tokens_seen": 421068800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044060180541624877,
      "loss": 2.9812,
      "theoretical_loss": 3.989003938397249,
      "tokens_seen": 421134336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044059177532597795,
      "loss": 2.9128,
      "theoretical_loss": 3.988935095208266,
      "tokens_seen": 421199872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044058174523570713,
      "loss": 2.8998,
      "theoretical_loss": 3.988866265728712,
      "tokens_seen": 421265408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405717151454363,
      "loss": 2.7796,
      "theoretical_loss": 3.9887974499537266,
      "tokens_seen": 421330944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044056168505516555,
      "loss": 2.9465,
      "theoretical_loss": 3.9887286478784496,
      "tokens_seen": 421396480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405516549648947,
      "loss": 2.6366,
      "theoretical_loss": 3.988659859498023,
      "tokens_seen": 421462016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405416248746239,
      "loss": 2.8122,
      "theoretical_loss": 3.9885910848075934,
      "tokens_seen": 421527552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044053159478435304,
      "loss": 2.9463,
      "theoretical_loss": 3.9885223238023073,
      "tokens_seen": 421593088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044052156469408227,
      "loss": 3.0525,
      "theoretical_loss": 3.988453576477315,
      "tokens_seen": 421658624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044051153460381145,
      "loss": 2.7746,
      "theoretical_loss": 3.9883848428277697,
      "tokens_seen": 421724160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044050150451354063,
      "loss": 2.7614,
      "theoretical_loss": 3.988316122848826,
      "tokens_seen": 421789696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404914744232698,
      "loss": 2.9223,
      "theoretical_loss": 3.988247416535642,
      "tokens_seen": 421855232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044048144433299905,
      "loss": 2.9403,
      "theoretical_loss": 3.9881787238833772,
      "tokens_seen": 421920768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404714142427282,
      "loss": 2.7421,
      "theoretical_loss": 3.9881100448871942,
      "tokens_seen": 421986304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404613841524574,
      "loss": 2.939,
      "theoretical_loss": 3.988041379542258,
      "tokens_seen": 422051840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044045135406218654,
      "loss": 2.831,
      "theoretical_loss": 3.9879727278437365,
      "tokens_seen": 422117376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404413239719158,
      "loss": 2.8991,
      "theoretical_loss": 3.987904089786799,
      "tokens_seen": 422182912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044043129388164496,
      "loss": 2.7653,
      "theoretical_loss": 3.9878354653666177,
      "tokens_seen": 422248448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044042126379137414,
      "loss": 2.9397,
      "theoretical_loss": 3.987766854578368,
      "tokens_seen": 422313984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404112337011033,
      "loss": 3.1411,
      "theoretical_loss": 3.987698257417227,
      "tokens_seen": 422379520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404012036108325,
      "loss": 3.0797,
      "theoretical_loss": 3.9876296738783745,
      "tokens_seen": 422445056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403911735205617,
      "loss": 2.7101,
      "theoretical_loss": 3.987561103956992,
      "tokens_seen": 422510592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403811434302909,
      "loss": 2.8034,
      "theoretical_loss": 3.987492547648265,
      "tokens_seen": 422576128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044037111334002004,
      "loss": 2.8983,
      "theoretical_loss": 3.98742400494738,
      "tokens_seen": 422641664
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 244758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9229331016540527,
      "objective/train/theoretical_loss": 3.987355475849527,
      "objective/train/tokens_used": 443167200,
      "theoretical_loss": 3.987355475849527,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403610832497493,
      "loss": 3.0326,
      "theoretical_loss": 3.987355475849527,
      "tokens_seen": 422707200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403510531594784,
      "loss": 2.7311,
      "theoretical_loss": 3.9872869603498975,
      "tokens_seen": 422772736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044034102306920764,
      "loss": 2.9594,
      "theoretical_loss": 3.987218458443686,
      "tokens_seen": 422838272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403309929789368,
      "loss": 2.9823,
      "theoretical_loss": 3.9871499701260897,
      "tokens_seen": 422903808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440320962888666,
      "loss": 2.9567,
      "theoretical_loss": 3.9870814953923075,
      "tokens_seen": 422969344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403109327983952,
      "loss": 2.9457,
      "theoretical_loss": 3.987013034237541,
      "tokens_seen": 423034880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403009027081244,
      "loss": 2.9932,
      "theoretical_loss": 3.9869445866569944,
      "tokens_seen": 423100416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044029087261785355,
      "loss": 2.9196,
      "theoretical_loss": 3.986876152645875,
      "tokens_seen": 423165952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402808425275828,
      "loss": 2.7942,
      "theoretical_loss": 3.9868077321993916,
      "tokens_seen": 423231488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402708124373119,
      "loss": 2.8662,
      "theoretical_loss": 3.986739325312755,
      "tokens_seen": 423297024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044026078234704114,
      "loss": 2.9557,
      "theoretical_loss": 3.98667093198118,
      "tokens_seen": 423362560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402507522567703,
      "loss": 3.1629,
      "theoretical_loss": 3.986602552199882,
      "tokens_seen": 423428096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402407221664995,
      "loss": 2.9203,
      "theoretical_loss": 3.9865341859640804,
      "tokens_seen": 423493632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402306920762287,
      "loss": 3.0048,
      "theoretical_loss": 3.986465833268996,
      "tokens_seen": 423559168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044022066198595787,
      "loss": 2.7389,
      "theoretical_loss": 3.986397494109853,
      "tokens_seen": 423624704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044021063189568705,
      "loss": 2.9356,
      "theoretical_loss": 3.986329168481877,
      "tokens_seen": 423690240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402006018054163,
      "loss": 2.8892,
      "theoretical_loss": 3.986260856380296,
      "tokens_seen": 423755776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401905717151454,
      "loss": 2.799,
      "theoretical_loss": 3.9861925578003428,
      "tokens_seen": 423821312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044018054162487465,
      "loss": 2.9222,
      "theoretical_loss": 3.986124272737248,
      "tokens_seen": 423886848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044017051153460383,
      "loss": 2.98,
      "theoretical_loss": 3.986056001186249,
      "tokens_seen": 423952384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440160481444333,
      "loss": 2.9878,
      "theoretical_loss": 3.9859877431425836,
      "tokens_seen": 424017920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401504513540622,
      "loss": 2.935,
      "theoretical_loss": 3.9859194986014925,
      "tokens_seen": 424083456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044014042126379137,
      "loss": 2.8919,
      "theoretical_loss": 3.9858512675582185,
      "tokens_seen": 424148992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044013039117352055,
      "loss": 2.856,
      "theoretical_loss": 3.9857830500080067,
      "tokens_seen": 424214528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401203610832498,
      "loss": 2.9074,
      "theoretical_loss": 3.985714845946105,
      "tokens_seen": 424280064
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 245792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5828306674957275,
      "objective/train/theoretical_loss": 3.9856466553677645,
      "objective/train/tokens_used": 444805600,
      "theoretical_loss": 3.9856466553677645,
      "tokens_seen": 424345600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401103309929789,
      "loss": 2.7627,
      "theoretical_loss": 3.9856466553677645,
      "tokens_seen": 424345600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044010030090270815,
      "loss": 2.9771,
      "theoretical_loss": 3.985578478268236,
      "tokens_seen": 424411136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400902708124373,
      "loss": 2.7067,
      "theoretical_loss": 3.985510314642776,
      "tokens_seen": 424476672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400802407221665,
      "loss": 3.1371,
      "theoretical_loss": 3.985442164486641,
      "tokens_seen": 424542208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400702106318957,
      "loss": 2.7251,
      "theoretical_loss": 3.9853740277950918,
      "tokens_seen": 424607744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400601805416249,
      "loss": 2.7574,
      "theoretical_loss": 3.98530590456339,
      "tokens_seen": 424673280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044005015045135406,
      "loss": 2.8466,
      "theoretical_loss": 3.985237794786799,
      "tokens_seen": 424738816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044004012036108324,
      "loss": 3.1828,
      "theoretical_loss": 3.985169698460588,
      "tokens_seen": 424804352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400300902708124,
      "loss": 2.867,
      "theoretical_loss": 3.9851016155800254,
      "tokens_seen": 424869888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044002006018054165,
      "loss": 2.7883,
      "theoretical_loss": 3.9850335461403823,
      "tokens_seen": 424935424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400100300902708,
      "loss": 2.8638,
      "theoretical_loss": 3.984965490136934,
      "tokens_seen": 425000960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044,
      "loss": 2.989,
      "theoretical_loss": 3.9848974475649563,
      "tokens_seen": 425066496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399899699097292,
      "loss": 3.2045,
      "theoretical_loss": 3.9848294184197286,
      "tokens_seen": 425132032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399799398194584,
      "loss": 2.8111,
      "theoretical_loss": 3.984761402696532,
      "tokens_seen": 425197568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043996990972918756,
      "loss": 2.8473,
      "theoretical_loss": 3.9846934003906505,
      "tokens_seen": 425263104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043995987963891674,
      "loss": 2.684,
      "theoretical_loss": 3.98462541149737,
      "tokens_seen": 425328640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399498495486459,
      "loss": 2.6684,
      "theoretical_loss": 3.984557436011978,
      "tokens_seen": 425394176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043993981945837516,
      "loss": 2.8611,
      "theoretical_loss": 3.9844894739297674,
      "tokens_seen": 425459712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043992978936810434,
      "loss": 2.8828,
      "theoretical_loss": 3.98442152524603,
      "tokens_seen": 425525248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399197592778335,
      "loss": 2.7737,
      "theoretical_loss": 3.984353589956062,
      "tokens_seen": 425590784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004399097291875627,
      "loss": 3.0198,
      "theoretical_loss": 3.9842856680551604,
      "tokens_seen": 425656320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398996990972919,
      "loss": 2.966,
      "theoretical_loss": 3.9842177595386277,
      "tokens_seen": 425721856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398896690070211,
      "loss": 2.9014,
      "theoretical_loss": 3.9841498644017643,
      "tokens_seen": 425787392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043987963891675024,
      "loss": 2.9488,
      "theoretical_loss": 3.984081982639877,
      "tokens_seen": 425852928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398696088264795,
      "loss": 2.9806,
      "theoretical_loss": 3.9840141142482723,
      "tokens_seen": 425918464
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 246485,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.934349298477173,
      "objective/train/theoretical_loss": 3.98394625922226,
      "objective/train/tokens_used": 446444000,
      "theoretical_loss": 3.98394625922226,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398595787362086,
      "loss": 2.8034,
      "theoretical_loss": 3.98394625922226,
      "tokens_seen": 425984000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043984954864593784,
      "loss": 2.9262,
      "theoretical_loss": 3.983878417557153,
      "tokens_seen": 426049536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439839518555667,
      "loss": 2.9786,
      "theoretical_loss": 3.9838105892482663,
      "tokens_seen": 426115072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398294884653962,
      "loss": 2.6863,
      "theoretical_loss": 3.9837427742909153,
      "tokens_seen": 426180608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398194583751254,
      "loss": 2.749,
      "theoretical_loss": 3.9836749726804204,
      "tokens_seen": 426246144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398094282848546,
      "loss": 2.8953,
      "theoretical_loss": 3.983607184412103,
      "tokens_seen": 426311680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043979939819458375,
      "loss": 2.841,
      "theoretical_loss": 3.9835394094812875,
      "tokens_seen": 426377216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439789368104313,
      "loss": 2.7696,
      "theoretical_loss": 3.9834716478832997,
      "tokens_seen": 426442752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397793380140421,
      "loss": 2.7188,
      "theoretical_loss": 3.9834038996134686,
      "tokens_seen": 426508288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043976930792377134,
      "loss": 2.717,
      "theoretical_loss": 3.9833361646671257,
      "tokens_seen": 426573824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397592778335005,
      "loss": 2.977,
      "theoretical_loss": 3.9832684430396035,
      "tokens_seen": 426639360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397492477432297,
      "loss": 2.9569,
      "theoretical_loss": 3.983200734726238,
      "tokens_seen": 426704896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397392176529589,
      "loss": 2.6681,
      "theoretical_loss": 3.9831330397223685,
      "tokens_seen": 426770432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043972918756268807,
      "loss": 2.8142,
      "theoretical_loss": 3.9830653580233344,
      "tokens_seen": 426835968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043971915747241725,
      "loss": 2.796,
      "theoretical_loss": 3.9829976896244794,
      "tokens_seen": 426901504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397091273821465,
      "loss": 2.91,
      "theoretical_loss": 3.9829300345211474,
      "tokens_seen": 426967040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396990972918756,
      "loss": 2.8034,
      "theoretical_loss": 3.9828623927086872,
      "tokens_seen": 427032576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043968906720160485,
      "loss": 2.6024,
      "theoretical_loss": 3.9827947641824473,
      "tokens_seen": 427098112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043967903711133403,
      "loss": 2.9885,
      "theoretical_loss": 3.982727148937782,
      "tokens_seen": 427163648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396690070210632,
      "loss": 3.0306,
      "theoretical_loss": 3.982659546970044,
      "tokens_seen": 427229184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396589769307924,
      "loss": 2.9307,
      "theoretical_loss": 3.9825919582745906,
      "tokens_seen": 427294720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043964894684052157,
      "loss": 3.1782,
      "theoretical_loss": 3.9825243828467816,
      "tokens_seen": 427360256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043963891675025075,
      "loss": 2.7458,
      "theoretical_loss": 3.982456820681978,
      "tokens_seen": 427425792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043962888665998,
      "loss": 2.9876,
      "theoretical_loss": 3.982389271775544,
      "tokens_seen": 427491328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396188565697091,
      "loss": 2.8233,
      "theoretical_loss": 3.9823217361228456,
      "tokens_seen": 427556864
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 247615,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7162137031555176,
      "objective/train/theoretical_loss": 3.982254213719252,
      "objective/train/tokens_used": 448082400,
      "theoretical_loss": 3.982254213719252,
      "tokens_seen": 427622400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043960882647943835,
      "loss": 2.8383,
      "theoretical_loss": 3.982254213719252,
      "tokens_seen": 427622400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395987963891675,
      "loss": 2.8669,
      "theoretical_loss": 3.9821867045601325,
      "tokens_seen": 427687936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395887662988967,
      "loss": 2.893,
      "theoretical_loss": 3.9821192086408623,
      "tokens_seen": 427753472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395787362086259,
      "loss": 2.8802,
      "theoretical_loss": 3.982051725956816,
      "tokens_seen": 427819008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395687061183551,
      "loss": 2.587,
      "theoretical_loss": 3.9819842565033703,
      "tokens_seen": 427884544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043955867602808426,
      "loss": 2.9512,
      "theoretical_loss": 3.9819168002759078,
      "tokens_seen": 427950080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043954864593781344,
      "loss": 2.7715,
      "theoretical_loss": 3.9818493572698093,
      "tokens_seen": 428015616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395386158475426,
      "loss": 3.0963,
      "theoretical_loss": 3.9817819274804602,
      "tokens_seen": 428081152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043952858575727185,
      "loss": 2.8617,
      "theoretical_loss": 3.9817145109032475,
      "tokens_seen": 428146688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439518555667001,
      "loss": 2.8287,
      "theoretical_loss": 3.9816471075335604,
      "tokens_seen": 428212224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395085255767302,
      "loss": 2.9118,
      "theoretical_loss": 3.981579717366791,
      "tokens_seen": 428277760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394984954864594,
      "loss": 3.141,
      "theoretical_loss": 3.981512340398333,
      "tokens_seen": 428343296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394884653961886,
      "loss": 2.6428,
      "theoretical_loss": 3.9814449766235835,
      "tokens_seen": 428408832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043947843530591776,
      "loss": 2.9725,
      "theoretical_loss": 3.9813776260379408,
      "tokens_seen": 428474368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043946840521564694,
      "loss": 2.7616,
      "theoretical_loss": 3.9813102886368057,
      "tokens_seen": 428539904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394583751253761,
      "loss": 2.8903,
      "theoretical_loss": 3.9812429644155816,
      "tokens_seen": 428605440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043944834503510536,
      "loss": 3.0138,
      "theoretical_loss": 3.9811756533696743,
      "tokens_seen": 428670976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394383149448345,
      "loss": 2.93,
      "theoretical_loss": 3.9811083554944915,
      "tokens_seen": 428736512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394282848545637,
      "loss": 2.7467,
      "theoretical_loss": 3.9810410707854436,
      "tokens_seen": 428802048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043941825476429285,
      "loss": 3.1358,
      "theoretical_loss": 3.980973799237943,
      "tokens_seen": 428867584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394082246740221,
      "loss": 2.8636,
      "theoretical_loss": 3.9809065408474047,
      "tokens_seen": 428933120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043939819458375126,
      "loss": 3.0446,
      "theoretical_loss": 3.980839295609246,
      "tokens_seen": 428998656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043938816449348044,
      "loss": 2.7096,
      "theoretical_loss": 3.9807720635188857,
      "tokens_seen": 429064192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393781344032096,
      "loss": 2.8974,
      "theoretical_loss": 3.9807048445717457,
      "tokens_seen": 429129728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393681043129388,
      "loss": 3.0659,
      "theoretical_loss": 3.9806376387632505,
      "tokens_seen": 429195264
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 248896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7603349685668945,
      "objective/train/theoretical_loss": 3.9805704460888256,
      "objective/train/tokens_used": 449720800,
      "theoretical_loss": 3.9805704460888256,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439358074222668,
      "loss": 2.9696,
      "theoretical_loss": 3.9805704460888256,
      "tokens_seen": 429260800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393480441323972,
      "loss": 3.04,
      "theoretical_loss": 3.9805032665439004,
      "tokens_seen": 429326336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043933801404212635,
      "loss": 2.7513,
      "theoretical_loss": 3.9804361001239057,
      "tokens_seen": 429391872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393279839518556,
      "loss": 2.8466,
      "theoretical_loss": 3.9803689468242744,
      "tokens_seen": 429457408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043931795386158477,
      "loss": 2.7847,
      "theoretical_loss": 3.9803018066404414,
      "tokens_seen": 429522944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043930792377131395,
      "loss": 2.9597,
      "theoretical_loss": 3.980234679567845,
      "tokens_seen": 429588480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043929789368104313,
      "loss": 2.878,
      "theoretical_loss": 3.980167565601925,
      "tokens_seen": 429654016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392878635907723,
      "loss": 3.0399,
      "theoretical_loss": 3.9801004647381246,
      "tokens_seen": 429719552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392778335005015,
      "loss": 3.145,
      "theoretical_loss": 3.980033376971887,
      "tokens_seen": 429785088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392678034102307,
      "loss": 3.0189,
      "theoretical_loss": 3.97996630229866,
      "tokens_seen": 429850624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043925777331995985,
      "loss": 2.8291,
      "theoretical_loss": 3.9798992407138924,
      "tokens_seen": 429916160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392477432296891,
      "loss": 2.8082,
      "theoretical_loss": 3.9798321922130357,
      "tokens_seen": 429981696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392377131394182,
      "loss": 2.8025,
      "theoretical_loss": 3.979765156791543,
      "tokens_seen": 430047232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043922768304914745,
      "loss": 2.8277,
      "theoretical_loss": 3.979698134444871,
      "tokens_seen": 430112768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043921765295887663,
      "loss": 2.9302,
      "theoretical_loss": 3.979631125168478,
      "tokens_seen": 430178304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392076228686058,
      "loss": 2.7819,
      "theoretical_loss": 3.979564128957824,
      "tokens_seen": 430243840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439197592778335,
      "loss": 2.887,
      "theoretical_loss": 3.979497145808372,
      "tokens_seen": 430309376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043918756268806423,
      "loss": 2.8212,
      "theoretical_loss": 3.979430175715587,
      "tokens_seen": 430374912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391775325977934,
      "loss": 3.0824,
      "theoretical_loss": 3.979363218674936,
      "tokens_seen": 430440448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391675025075226,
      "loss": 2.8863,
      "theoretical_loss": 3.9792962746818894,
      "tokens_seen": 430505984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043915747241725177,
      "loss": 2.9136,
      "theoretical_loss": 3.9792293437319177,
      "tokens_seen": 430571520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043914744232698095,
      "loss": 2.8946,
      "theoretical_loss": 3.979162425820496,
      "tokens_seen": 430637056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391374122367102,
      "loss": 2.8152,
      "theoretical_loss": 3.9790955209431007,
      "tokens_seen": 430702592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391273821464393,
      "loss": 2.5009,
      "theoretical_loss": 3.9790286290952097,
      "tokens_seen": 430768128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043911735205616855,
      "loss": 2.7643,
      "theoretical_loss": 3.978961750272304,
      "tokens_seen": 430833664
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 249330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5802011489868164,
      "objective/train/theoretical_loss": 3.978894884469867,
      "objective/train/tokens_used": 451359200,
      "theoretical_loss": 3.978894884469867,
      "tokens_seen": 430899200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391073219658977,
      "loss": 2.6646,
      "theoretical_loss": 3.978894884469867,
      "tokens_seen": 430899200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390972918756269,
      "loss": 2.8741,
      "theoretical_loss": 3.978828031683384,
      "tokens_seen": 430964736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390872617853561,
      "loss": 3.135,
      "theoretical_loss": 3.978761191908343,
      "tokens_seen": 431030272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390772316950853,
      "loss": 2.7624,
      "theoretical_loss": 3.9786943651402327,
      "tokens_seen": 431095808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043906720160481446,
      "loss": 2.8756,
      "theoretical_loss": 3.9786275513745464,
      "tokens_seen": 431161344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043905717151454364,
      "loss": 2.8513,
      "theoretical_loss": 3.9785607506067775,
      "tokens_seen": 431226880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390471414242728,
      "loss": 2.951,
      "theoretical_loss": 3.978493962832424,
      "tokens_seen": 431292416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043903711133400205,
      "loss": 2.7171,
      "theoretical_loss": 3.9784271880469824,
      "tokens_seen": 431357952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390270812437312,
      "loss": 2.7085,
      "theoretical_loss": 3.978360426245956,
      "tokens_seen": 431423488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390170511534604,
      "loss": 2.8883,
      "theoretical_loss": 3.9782936774248476,
      "tokens_seen": 431489024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390070210631896,
      "loss": 2.7489,
      "theoretical_loss": 3.978226941579162,
      "tokens_seen": 431554560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389969909729188,
      "loss": 2.8769,
      "theoretical_loss": 3.9781602187044074,
      "tokens_seen": 431620096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043898696088264796,
      "loss": 2.9323,
      "theoretical_loss": 3.9780935087960945,
      "tokens_seen": 431685632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043897693079237714,
      "loss": 2.8336,
      "theoretical_loss": 3.978026811849735,
      "tokens_seen": 431751168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389669007021063,
      "loss": 2.4792,
      "theoretical_loss": 3.977960127860843,
      "tokens_seen": 431816704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043895687061183556,
      "loss": 2.7761,
      "theoretical_loss": 3.977893456824936,
      "tokens_seen": 431882240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389468405215647,
      "loss": 2.847,
      "theoretical_loss": 3.9778267987375324,
      "tokens_seen": 431947776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389368104312939,
      "loss": 2.9699,
      "theoretical_loss": 3.9777601535941542,
      "tokens_seen": 432013312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043892678034102305,
      "loss": 2.9322,
      "theoretical_loss": 3.977693521390324,
      "tokens_seen": 432078848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389167502507523,
      "loss": 2.7251,
      "theoretical_loss": 3.9776269021215676,
      "tokens_seen": 432144384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043890672016048146,
      "loss": 2.7903,
      "theoretical_loss": 3.9775602957834133,
      "tokens_seen": 432209920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043889669007021064,
      "loss": 3.0632,
      "theoretical_loss": 3.977493702371391,
      "tokens_seen": 432275456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388866599799398,
      "loss": 2.9212,
      "theoretical_loss": 3.977427121881033,
      "tokens_seen": 432340992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438876629889669,
      "loss": 2.8574,
      "theoretical_loss": 3.977360554307874,
      "tokens_seen": 432406528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388665997993982,
      "loss": 2.6996,
      "theoretical_loss": 3.9772939996474506,
      "tokens_seen": 432472064
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 250327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1851444244384766,
      "objective/train/theoretical_loss": 3.9772274578953013,
      "objective/train/tokens_used": 452997600,
      "theoretical_loss": 3.9772274578953013,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388565697091274,
      "loss": 2.49,
      "theoretical_loss": 3.9772274578953013,
      "tokens_seen": 432537600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043884653961885655,
      "loss": 3.0437,
      "theoretical_loss": 3.977160929046968,
      "tokens_seen": 432603136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388365095285858,
      "loss": 3.0128,
      "theoretical_loss": 3.977094413097994,
      "tokens_seen": 432668672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043882647943831497,
      "loss": 3.064,
      "theoretical_loss": 3.9770279100439256,
      "tokens_seen": 432734208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043881644934804415,
      "loss": 3.0597,
      "theoretical_loss": 3.9769614198803094,
      "tokens_seen": 432799744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043880641925777333,
      "loss": 2.7542,
      "theoretical_loss": 3.976894942602696,
      "tokens_seen": 432865280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387963891675025,
      "loss": 2.9004,
      "theoretical_loss": 3.976828478206638,
      "tokens_seen": 432930816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387863590772317,
      "loss": 2.7999,
      "theoretical_loss": 3.976762026687689,
      "tokens_seen": 432996352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387763289869609,
      "loss": 2.9483,
      "theoretical_loss": 3.9766955880414065,
      "tokens_seen": 433061888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043876629889669005,
      "loss": 2.7133,
      "theoretical_loss": 3.9766291622633494,
      "tokens_seen": 433127424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387562688064193,
      "loss": 2.8788,
      "theoretical_loss": 3.9765627493490783,
      "tokens_seen": 433192960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387462387161484,
      "loss": 2.7892,
      "theoretical_loss": 3.976496349294157,
      "tokens_seen": 433258496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043873620862587765,
      "loss": 2.8578,
      "theoretical_loss": 3.9764299620941506,
      "tokens_seen": 433324032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043872617853560683,
      "loss": 2.86,
      "theoretical_loss": 3.9763635877446273,
      "tokens_seen": 433389568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438716148445336,
      "loss": 2.8879,
      "theoretical_loss": 3.9762972262411562,
      "tokens_seen": 433455104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387061183550652,
      "loss": 2.9078,
      "theoretical_loss": 3.97623087757931,
      "tokens_seen": 433520640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043869608826479443,
      "loss": 2.9007,
      "theoretical_loss": 3.9761645417546627,
      "tokens_seen": 433586176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043868605817452356,
      "loss": 2.7336,
      "theoretical_loss": 3.9760982187627913,
      "tokens_seen": 433651712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386760280842528,
      "loss": 3.0301,
      "theoretical_loss": 3.976031908599274,
      "tokens_seen": 433717248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386659979939819,
      "loss": 2.8723,
      "theoretical_loss": 3.9759656112596913,
      "tokens_seen": 433782784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043865596790371115,
      "loss": 2.9686,
      "theoretical_loss": 3.9758993267396265,
      "tokens_seen": 433848320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043864593781344033,
      "loss": 2.7986,
      "theoretical_loss": 3.9758330550346663,
      "tokens_seen": 433913856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386359077231695,
      "loss": 2.8748,
      "theoretical_loss": 3.975766796140396,
      "tokens_seen": 433979392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386258776328987,
      "loss": 2.8568,
      "theoretical_loss": 3.975700550052406,
      "tokens_seen": 434044928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386158475426279,
      "loss": 2.828,
      "theoretical_loss": 3.9756343167662886,
      "tokens_seen": 434110464
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 251069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.99489426612854,
      "objective/train/theoretical_loss": 3.9755680962776374,
      "objective/train/tokens_used": 454636000,
      "theoretical_loss": 3.9755680962776374,
      "tokens_seen": 434176000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043860581745235706,
      "loss": 2.7617,
      "theoretical_loss": 3.9755680962776374,
      "tokens_seen": 434176000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385957873620863,
      "loss": 3.1528,
      "theoretical_loss": 3.9755018885820483,
      "tokens_seen": 434241536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385857572718154,
      "loss": 2.9258,
      "theoretical_loss": 3.97543569367512,
      "tokens_seen": 434307072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043857572718154466,
      "loss": 2.9792,
      "theoretical_loss": 3.9753695115524534,
      "tokens_seen": 434372608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385656970912738,
      "loss": 2.7644,
      "theoretical_loss": 3.97530334220965,
      "tokens_seen": 434438144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438555667001003,
      "loss": 2.9931,
      "theoretical_loss": 3.975237185642316,
      "tokens_seen": 434503680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385456369107322,
      "loss": 2.948,
      "theoretical_loss": 3.9751710418460577,
      "tokens_seen": 434569216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385356068204614,
      "loss": 3.1,
      "theoretical_loss": 3.9751049108164844,
      "tokens_seen": 434634752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043852557673019056,
      "loss": 2.8339,
      "theoretical_loss": 3.9750387925492072,
      "tokens_seen": 434700288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385155466399198,
      "loss": 2.9765,
      "theoretical_loss": 3.974972687039841,
      "tokens_seen": 434765824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385055165496489,
      "loss": 2.5896,
      "theoretical_loss": 3.9749065942839996,
      "tokens_seen": 434831360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043849548645937816,
      "loss": 2.8618,
      "theoretical_loss": 3.974840514277303,
      "tokens_seen": 434896896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384854563691073,
      "loss": 2.9577,
      "theoretical_loss": 3.974774447015369,
      "tokens_seen": 434962432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384754262788365,
      "loss": 2.9189,
      "theoretical_loss": 3.9747083924938216,
      "tokens_seen": 435027968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384653961885657,
      "loss": 2.779,
      "theoretical_loss": 3.974642350708284,
      "tokens_seen": 435093504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384553660982949,
      "loss": 2.99,
      "theoretical_loss": 3.974576321654384,
      "tokens_seen": 435159040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043844533600802407,
      "loss": 2.6761,
      "theoretical_loss": 3.9745103053277493,
      "tokens_seen": 435224576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043843530591775325,
      "loss": 2.8512,
      "theoretical_loss": 3.974444301724011,
      "tokens_seen": 435290112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384252758274825,
      "loss": 2.8829,
      "theoretical_loss": 3.9743783108388024,
      "tokens_seen": 435355648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043841524573721166,
      "loss": 3.0582,
      "theoretical_loss": 3.9743123326677585,
      "tokens_seen": 435421184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043840521564694084,
      "loss": 3.0055,
      "theoretical_loss": 3.9742463672065167,
      "tokens_seen": 435486720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043839518555667,
      "loss": 2.7174,
      "theoretical_loss": 3.9741804144507165,
      "tokens_seen": 435552256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383851554663992,
      "loss": 2.9559,
      "theoretical_loss": 3.974114474395999,
      "tokens_seen": 435617792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383751253761284,
      "loss": 2.9737,
      "theoretical_loss": 3.9740485470380094,
      "tokens_seen": 435683328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383650952858576,
      "loss": 2.8755,
      "theoretical_loss": 3.973982632372392,
      "tokens_seen": 435748864
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 252128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7361905574798584,
      "objective/train/theoretical_loss": 3.973916730394796,
      "objective/train/tokens_used": 456274400,
      "theoretical_loss": 3.973916730394796,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043835506519558675,
      "loss": 2.7758,
      "theoretical_loss": 3.973916730394796,
      "tokens_seen": 435814400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438345035105316,
      "loss": 2.9473,
      "theoretical_loss": 3.9738508411008713,
      "tokens_seen": 435879936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043833500501504517,
      "loss": 2.8355,
      "theoretical_loss": 3.97378496448627,
      "tokens_seen": 435945472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043832497492477435,
      "loss": 2.9717,
      "theoretical_loss": 3.9737191005466475,
      "tokens_seen": 436011008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043831494483450353,
      "loss": 2.8114,
      "theoretical_loss": 3.973653249277659,
      "tokens_seen": 436076544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383049147442327,
      "loss": 2.9563,
      "theoretical_loss": 3.973587410674965,
      "tokens_seen": 436142080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382948846539619,
      "loss": 3.0945,
      "theoretical_loss": 3.973521584734225,
      "tokens_seen": 436207616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382848545636911,
      "loss": 3.0023,
      "theoretical_loss": 3.9734557714511034,
      "tokens_seen": 436273152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043827482447342025,
      "loss": 3.158,
      "theoretical_loss": 3.9733899708212643,
      "tokens_seen": 436338688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382647943831495,
      "loss": 2.7855,
      "theoretical_loss": 3.973324182840375,
      "tokens_seen": 436404224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382547642928786,
      "loss": 2.7221,
      "theoretical_loss": 3.9732584075041064,
      "tokens_seen": 436469760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043824473420260785,
      "loss": 2.9749,
      "theoretical_loss": 3.9731926448081287,
      "tokens_seen": 436535296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043823470411233703,
      "loss": 2.8126,
      "theoretical_loss": 3.973126894748116,
      "tokens_seen": 436600832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382246740220662,
      "loss": 2.909,
      "theoretical_loss": 3.973061157319745,
      "tokens_seen": 436666368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382146439317954,
      "loss": 2.9275,
      "theoretical_loss": 3.972995432518693,
      "tokens_seen": 436731904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043820461384152463,
      "loss": 2.8421,
      "theoretical_loss": 3.9729297203406393,
      "tokens_seen": 436797440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043819458375125376,
      "loss": 2.7437,
      "theoretical_loss": 3.9728640207812678,
      "tokens_seen": 436862976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438184553660983,
      "loss": 2.7797,
      "theoretical_loss": 3.9727983338362627,
      "tokens_seen": 436928512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381745235707121,
      "loss": 2.9771,
      "theoretical_loss": 3.9727326595013093,
      "tokens_seen": 436994048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043816449348044135,
      "loss": 2.5017,
      "theoretical_loss": 3.972666997772097,
      "tokens_seen": 437059584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043815446339017053,
      "loss": 3.0961,
      "theoretical_loss": 3.972601348644317,
      "tokens_seen": 437125120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381444332998997,
      "loss": 2.7381,
      "theoretical_loss": 3.9725357121136615,
      "tokens_seen": 437190656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381344032096289,
      "loss": 2.9324,
      "theoretical_loss": 3.9724700881758253,
      "tokens_seen": 437256192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381243731193581,
      "loss": 2.7951,
      "theoretical_loss": 3.9724044768265063,
      "tokens_seen": 437321728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043811434302908726,
      "loss": 2.9698,
      "theoretical_loss": 3.9723388780614033,
      "tokens_seen": 437387264
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 252698,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6559126377105713,
      "objective/train/theoretical_loss": 3.9722732918762174,
      "objective/train/tokens_used": 457912800,
      "theoretical_loss": 3.9722732918762174,
      "tokens_seen": 437452800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381043129388165,
      "loss": 2.88,
      "theoretical_loss": 3.9722732918762174,
      "tokens_seen": 437452800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380942828485456,
      "loss": 2.7562,
      "theoretical_loss": 3.9722077182666524,
      "tokens_seen": 437518336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043808425275827486,
      "loss": 2.9245,
      "theoretical_loss": 3.9721421572284137,
      "tokens_seen": 437583872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438074222668004,
      "loss": 2.7057,
      "theoretical_loss": 3.972076608757209,
      "tokens_seen": 437649408
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380641925777332,
      "loss": 3.2024,
      "theoretical_loss": 3.9720110728487485,
      "tokens_seen": 437714944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380541624874624,
      "loss": 2.9399,
      "theoretical_loss": 3.9719455494987432,
      "tokens_seen": 437780480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380441323971916,
      "loss": 2.8494,
      "theoretical_loss": 3.971880038702908,
      "tokens_seen": 437846016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043803410230692076,
      "loss": 3.0082,
      "theoretical_loss": 3.9718145404569585,
      "tokens_seen": 437911552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043802407221665,
      "loss": 2.8472,
      "theoretical_loss": 3.971749054756613,
      "tokens_seen": 437977088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380140421263791,
      "loss": 2.8728,
      "theoretical_loss": 3.971683581597592,
      "tokens_seen": 438042624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043800401203610836,
      "loss": 2.8619,
      "theoretical_loss": 3.9716181209756174,
      "tokens_seen": 438108160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379939819458375,
      "loss": 2.8314,
      "theoretical_loss": 3.9715526728864137,
      "tokens_seen": 438173696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379839518555667,
      "loss": 2.4957,
      "theoretical_loss": 3.971487237325708,
      "tokens_seen": 438239232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379739217652959,
      "loss": 2.8016,
      "theoretical_loss": 3.9714218142892292,
      "tokens_seen": 438304768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379638916750251,
      "loss": 2.8969,
      "theoretical_loss": 3.9713564037727074,
      "tokens_seen": 438370304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043795386158475427,
      "loss": 2.6609,
      "theoretical_loss": 3.9712910057718758,
      "tokens_seen": 438435840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043794383149448345,
      "loss": 3.0509,
      "theoretical_loss": 3.9712256202824694,
      "tokens_seen": 438501376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043793380140421263,
      "loss": 2.8158,
      "theoretical_loss": 3.9711602473002245,
      "tokens_seen": 438566912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043792377131394186,
      "loss": 2.9429,
      "theoretical_loss": 3.9710948868208815,
      "tokens_seen": 438632448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437913741223671,
      "loss": 2.6593,
      "theoretical_loss": 3.971029538840181,
      "tokens_seen": 438697984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004379037111334002,
      "loss": 2.7673,
      "theoretical_loss": 3.9709642033538666,
      "tokens_seen": 438763520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043789368104312935,
      "loss": 2.89,
      "theoretical_loss": 3.970898880357683,
      "tokens_seen": 438829056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378836509528586,
      "loss": 2.8787,
      "theoretical_loss": 3.9708335698473785,
      "tokens_seen": 438894592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043787362086258777,
      "loss": 2.8641,
      "theoretical_loss": 3.9707682718187023,
      "tokens_seen": 438960128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043786359077231695,
      "loss": 2.7491,
      "theoretical_loss": 3.9707029862674066,
      "tokens_seen": 439025664
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 253958,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.985783815383911,
      "objective/train/theoretical_loss": 3.970637713189244,
      "objective/train/tokens_used": 459551200,
      "theoretical_loss": 3.970637713189244,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043785356068204613,
      "loss": 2.9894,
      "theoretical_loss": 3.970637713189244,
      "tokens_seen": 439091200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043784353059177537,
      "loss": 2.6916,
      "theoretical_loss": 3.9705724525799715,
      "tokens_seen": 439156736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378335005015045,
      "loss": 2.6918,
      "theoretical_loss": 3.9705072044353455,
      "tokens_seen": 439222272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043782347041123373,
      "loss": 2.917,
      "theoretical_loss": 3.970441968751128,
      "tokens_seen": 439287808
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043781344032096286,
      "loss": 2.8387,
      "theoretical_loss": 3.9703767455230796,
      "tokens_seen": 439353344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378034102306921,
      "loss": 2.9011,
      "theoretical_loss": 3.970311534746965,
      "tokens_seen": 439418880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043779338014042127,
      "loss": 2.774,
      "theoretical_loss": 3.97024633641855,
      "tokens_seen": 439484416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043778335005015045,
      "loss": 2.9159,
      "theoretical_loss": 3.970181150533603,
      "tokens_seen": 439549952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043777331995987963,
      "loss": 2.7145,
      "theoretical_loss": 3.970115977087895,
      "tokens_seen": 439615488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377632898696088,
      "loss": 2.9115,
      "theoretical_loss": 3.9700508160771966,
      "tokens_seen": 439681024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437753259779338,
      "loss": 2.9576,
      "theoretical_loss": 3.9699856674972844,
      "tokens_seen": 439746560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043774322968906723,
      "loss": 3.0514,
      "theoretical_loss": 3.969920531343934,
      "tokens_seen": 439812096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043773319959879636,
      "loss": 2.7901,
      "theoretical_loss": 3.9698554076129233,
      "tokens_seen": 439877632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377231695085256,
      "loss": 2.6878,
      "theoretical_loss": 3.969790296300034,
      "tokens_seen": 439943168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377131394182547,
      "loss": 2.7006,
      "theoretical_loss": 3.969725197401049,
      "tokens_seen": 440008704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043770310932798396,
      "loss": 2.9952,
      "theoretical_loss": 3.969660110911752,
      "tokens_seen": 440074240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043769307923771314,
      "loss": 3.0273,
      "theoretical_loss": 3.96959503682793,
      "tokens_seen": 440139776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376830491474423,
      "loss": 2.8135,
      "theoretical_loss": 3.969529975145372,
      "tokens_seen": 440205312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043767301905717155,
      "loss": 2.7669,
      "theoretical_loss": 3.9694649258598704,
      "tokens_seen": 440270848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043766298896690073,
      "loss": 2.767,
      "theoretical_loss": 3.9693998889672164,
      "tokens_seen": 440336384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376529588766299,
      "loss": 2.7824,
      "theoretical_loss": 3.969334864463206,
      "tokens_seen": 440401920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376429287863591,
      "loss": 2.6347,
      "theoretical_loss": 3.9692698523436354,
      "tokens_seen": 440467456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376328986960883,
      "loss": 2.8143,
      "theoretical_loss": 3.9692048526043044,
      "tokens_seen": 440532992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043762286860581746,
      "loss": 3.0135,
      "theoretical_loss": 3.969139865241014,
      "tokens_seen": 440598528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376128385155467,
      "loss": 2.7171,
      "theoretical_loss": 3.9690748902495683,
      "tokens_seen": 440664064
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 254624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.297787666320801,
      "objective/train/theoretical_loss": 3.9690099276257715,
      "objective/train/tokens_used": 461189600,
      "theoretical_loss": 3.9690099276257715,
      "tokens_seen": 440729600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376028084252758,
      "loss": 2.9859,
      "theoretical_loss": 3.9690099276257715,
      "tokens_seen": 440729600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043759277833500506,
      "loss": 2.8268,
      "theoretical_loss": 3.9689449773654317,
      "tokens_seen": 440795136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375827482447342,
      "loss": 2.5088,
      "theoretical_loss": 3.9688800394643575,
      "tokens_seen": 440860672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375727181544634,
      "loss": 2.7787,
      "theoretical_loss": 3.968815113918361,
      "tokens_seen": 440926208
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375626880641926,
      "loss": 2.9048,
      "theoretical_loss": 3.968750200723256,
      "tokens_seen": 440991744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375526579739218,
      "loss": 2.8828,
      "theoretical_loss": 3.9686852998748567,
      "tokens_seen": 441057280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043754262788365096,
      "loss": 2.6297,
      "theoretical_loss": 3.968620411368982,
      "tokens_seen": 441122816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375325977933802,
      "loss": 3.0614,
      "theoretical_loss": 3.9685555352014505,
      "tokens_seen": 441188352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375225677031093,
      "loss": 2.958,
      "theoretical_loss": 3.9684906713680848,
      "tokens_seen": 441253888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043751253761283856,
      "loss": 2.84,
      "theoretical_loss": 3.9684258198647075,
      "tokens_seen": 441319424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375025075225677,
      "loss": 2.816,
      "theoretical_loss": 3.9683609806871454,
      "tokens_seen": 441384960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374924774322969,
      "loss": 2.809,
      "theoretical_loss": 3.9682961538312256,
      "tokens_seen": 441450496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374824473420261,
      "loss": 2.9887,
      "theoretical_loss": 3.968231339292778,
      "tokens_seen": 441516032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374724172517553,
      "loss": 2.8036,
      "theoretical_loss": 3.968166537067634,
      "tokens_seen": 441581568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043746238716148447,
      "loss": 2.7846,
      "theoretical_loss": 3.9681017471516284,
      "tokens_seen": 441647104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043745235707121365,
      "loss": 2.7036,
      "theoretical_loss": 3.968036969540596,
      "tokens_seen": 441712640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043744232698094283,
      "loss": 2.7011,
      "theoretical_loss": 3.9679722042303753,
      "tokens_seen": 441778176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043743229689067206,
      "loss": 2.802,
      "theoretical_loss": 3.967907451216806,
      "tokens_seen": 441843712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374222668004012,
      "loss": 2.9564,
      "theoretical_loss": 3.9678427104957303,
      "tokens_seen": 441909248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374122367101304,
      "loss": 2.9014,
      "theoretical_loss": 3.9677779820629917,
      "tokens_seen": 441974784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043740220661985955,
      "loss": 2.8714,
      "theoretical_loss": 3.9677132659144365,
      "tokens_seen": 442040320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373921765295888,
      "loss": 2.8678,
      "theoretical_loss": 3.9676485620459125,
      "tokens_seen": 442105856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043738214643931797,
      "loss": 2.789,
      "theoretical_loss": 3.96758387045327,
      "tokens_seen": 442171392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043737211634904715,
      "loss": 2.6459,
      "theoretical_loss": 3.967519191132361,
      "tokens_seen": 442236928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043736208625877633,
      "loss": 3.0265,
      "theoretical_loss": 3.9674545240790393,
      "tokens_seen": 442302464
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 255794,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5010974407196045,
      "objective/train/theoretical_loss": 3.967389869289161,
      "objective/train/tokens_used": 462828000,
      "theoretical_loss": 3.967389869289161,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043735205616850557,
      "loss": 2.7137,
      "theoretical_loss": 3.967389869289161,
      "tokens_seen": 442368000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373420260782347,
      "loss": 2.6839,
      "theoretical_loss": 3.967325226758584,
      "tokens_seen": 442433536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043733199598796393,
      "loss": 2.985,
      "theoretical_loss": 3.967260596483169,
      "tokens_seen": 442499072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043732196589769306,
      "loss": 2.8362,
      "theoretical_loss": 3.9671959784587774,
      "tokens_seen": 442564608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373119358074223,
      "loss": 2.7015,
      "theoretical_loss": 3.9671313726812736,
      "tokens_seen": 442630144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043730190571715147,
      "loss": 2.957,
      "theoretical_loss": 3.967066779146524,
      "tokens_seen": 442695680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043729187562688065,
      "loss": 2.9622,
      "theoretical_loss": 3.9670021978503964,
      "tokens_seen": 442761216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043728184553660983,
      "loss": 2.7419,
      "theoretical_loss": 3.966937628788761,
      "tokens_seen": 442826752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437271815446339,
      "loss": 2.9077,
      "theoretical_loss": 3.96687307195749,
      "tokens_seen": 442892288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372617853560682,
      "loss": 2.4599,
      "theoretical_loss": 3.966808527352457,
      "tokens_seen": 442957824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043725175526579743,
      "loss": 2.9802,
      "theoretical_loss": 3.966743994969539,
      "tokens_seen": 443023360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043724172517552656,
      "loss": 2.8803,
      "theoretical_loss": 3.9666794748046135,
      "tokens_seen": 443088896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372316950852558,
      "loss": 2.752,
      "theoretical_loss": 3.9666149668535615,
      "tokens_seen": 443154432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372216649949849,
      "loss": 2.9157,
      "theoretical_loss": 3.9665504711122637,
      "tokens_seen": 443219968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043721163490471416,
      "loss": 2.8042,
      "theoretical_loss": 3.9664859875766054,
      "tokens_seen": 443285504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043720160481444334,
      "loss": 2.7464,
      "theoretical_loss": 3.9664215162424723,
      "tokens_seen": 443351040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371915747241725,
      "loss": 2.9279,
      "theoretical_loss": 3.9663570571057525,
      "tokens_seen": 443416576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371815446339017,
      "loss": 2.8611,
      "theoretical_loss": 3.9662926101623364,
      "tokens_seen": 443482112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043717151454363093,
      "loss": 3.013,
      "theoretical_loss": 3.9662281754081157,
      "tokens_seen": 443547648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043716148445336006,
      "loss": 2.9245,
      "theoretical_loss": 3.9661637528389853,
      "tokens_seen": 443613184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371514543630893,
      "loss": 2.8922,
      "theoretical_loss": 3.96609934245084,
      "tokens_seen": 443678720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371414242728184,
      "loss": 2.6798,
      "theoretical_loss": 3.966034944239579,
      "tokens_seen": 443744256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043713139418254766,
      "loss": 2.6617,
      "theoretical_loss": 3.965970558201102,
      "tokens_seen": 443809792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043712136409227684,
      "loss": 2.8185,
      "theoretical_loss": 3.9659061843313106,
      "tokens_seen": 443875328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437111334002006,
      "loss": 2.7108,
      "theoretical_loss": 3.9658418226261096,
      "tokens_seen": 443940864
    },
    {
      "epoch": 0.13,
      "objective/train/docs_used": 257001,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6992599964141846,
      "objective/train/theoretical_loss": 3.9657774730814044,
      "objective/train/tokens_used": 464466400,
      "theoretical_loss": 3.9657774730814044,
      "tokens_seen": 444006400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371013039117352,
      "loss": 2.7786,
      "theoretical_loss": 3.9657774730814044,
      "tokens_seen": 444006400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370912738214644,
      "loss": 3.0641,
      "theoretical_loss": 3.965713135693103,
      "tokens_seen": 444071936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043708124373119356,
      "loss": 2.8226,
      "theoretical_loss": 3.9656488104571155,
      "tokens_seen": 444137472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370712136409228,
      "loss": 2.8475,
      "theoretical_loss": 3.965584497369355,
      "tokens_seen": 444203008
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370611835506519,
      "loss": 2.7757,
      "theoretical_loss": 3.9655201964257336,
      "tokens_seen": 444268544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043705115346038116,
      "loss": 2.6841,
      "theoretical_loss": 3.965455907622168,
      "tokens_seen": 444334080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370411233701103,
      "loss": 2.8601,
      "theoretical_loss": 3.9653916309545765,
      "tokens_seen": 444399616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370310932798395,
      "loss": 2.8329,
      "theoretical_loss": 3.9653273664188786,
      "tokens_seen": 444465152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370210631895687,
      "loss": 2.9373,
      "theoretical_loss": 3.9652631140109955,
      "tokens_seen": 444530688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370110330992979,
      "loss": 2.8243,
      "theoretical_loss": 3.965198873726852,
      "tokens_seen": 444596224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043700100300902707,
      "loss": 3.0494,
      "theoretical_loss": 3.965134645562374,
      "tokens_seen": 444661760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369909729187563,
      "loss": 3.087,
      "theoretical_loss": 3.9650704295134878,
      "tokens_seen": 444727296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043698094282848543,
      "loss": 2.9131,
      "theoretical_loss": 3.9650062255761247,
      "tokens_seen": 444792832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043697091273821467,
      "loss": 2.9083,
      "theoretical_loss": 3.964942033746216,
      "tokens_seen": 444858368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369608826479438,
      "loss": 2.8564,
      "theoretical_loss": 3.9648778540196945,
      "tokens_seen": 444923904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043695085255767303,
      "loss": 2.7252,
      "theoretical_loss": 3.964813686392496,
      "tokens_seen": 444989440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369408224674022,
      "loss": 2.992,
      "theoretical_loss": 3.9647495308605594,
      "tokens_seen": 445054976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369307923771314,
      "loss": 2.9756,
      "theoretical_loss": 3.964685387419823,
      "tokens_seen": 445120512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004369207622868606,
      "loss": 2.8825,
      "theoretical_loss": 3.964621256066229,
      "tokens_seen": 445186048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043691073219658975,
      "loss": 2.804,
      "theoretical_loss": 3.96455713679572,
      "tokens_seen": 445251584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000436900702106319,
      "loss": 2.609,
      "theoretical_loss": 3.9644930296042418,
      "tokens_seen": 445317120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043689067201604817,
      "loss": 2.6154,
      "theoretical_loss": 3.964428934487742,
      "tokens_seen": 445382656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043688064192577735,
      "loss": 2.8911,
      "theoretical_loss": 3.96436485144217,
      "tokens_seen": 445448192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043687061183550653,
      "loss": 2.9299,
      "theoretical_loss": 3.964300780463477,
      "tokens_seen": 445513728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043686058174523577,
      "loss": 2.6714,
      "theoretical_loss": 3.964236721547616,
      "tokens_seen": 445579264
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 257279,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.354438304901123,
      "objective/train/theoretical_loss": 3.964172674690542,
      "objective/train/tokens_used": 466104800,
      "theoretical_loss": 3.964172674690542,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368505516549649,
      "loss": 2.6825,
      "theoretical_loss": 3.964172674690542,
      "tokens_seen": 445644800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043684052156469413,
      "loss": 2.8716,
      "theoretical_loss": 3.9641086398882126,
      "tokens_seen": 445710336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043683049147442326,
      "loss": 2.9858,
      "theoretical_loss": 3.9640446171365866,
      "tokens_seen": 445775872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368204613841525,
      "loss": 2.7256,
      "theoretical_loss": 3.9639806064316256,
      "tokens_seen": 445841408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043681043129388167,
      "loss": 2.6819,
      "theoretical_loss": 3.963916607769292,
      "tokens_seen": 445906944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043680040120361085,
      "loss": 3.0643,
      "theoretical_loss": 3.963852621145551,
      "tokens_seen": 445972480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043679037111334003,
      "loss": 2.9543,
      "theoretical_loss": 3.9637886465563685,
      "tokens_seen": 446038016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367803410230692,
      "loss": 3.0419,
      "theoretical_loss": 3.9637246839977154,
      "tokens_seen": 446103552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367703109327984,
      "loss": 3.1,
      "theoretical_loss": 3.9636607334655602,
      "tokens_seen": 446169088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043676028084252763,
      "loss": 2.8148,
      "theoretical_loss": 3.9635967949558775,
      "tokens_seen": 446234624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043675025075225676,
      "loss": 2.8659,
      "theoretical_loss": 3.963532868464641,
      "tokens_seen": 446300160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436740220661986,
      "loss": 2.8417,
      "theoretical_loss": 3.963468953987827,
      "tokens_seen": 446365696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367301905717151,
      "loss": 2.9253,
      "theoretical_loss": 3.9634050515214145,
      "tokens_seen": 446431232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043672016048144436,
      "loss": 3.0207,
      "theoretical_loss": 3.963341161061384,
      "tokens_seen": 446496768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043671013039117354,
      "loss": 2.7421,
      "theoretical_loss": 3.9632772826037175,
      "tokens_seen": 446562304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367001003009027,
      "loss": 2.7032,
      "theoretical_loss": 3.9632134161444,
      "tokens_seen": 446627840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366900702106319,
      "loss": 2.8559,
      "theoretical_loss": 3.9631495616794172,
      "tokens_seen": 446693376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043668004012036113,
      "loss": 2.7473,
      "theoretical_loss": 3.9630857192047566,
      "tokens_seen": 446758912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043667001003009026,
      "loss": 2.8741,
      "theoretical_loss": 3.96302188871641,
      "tokens_seen": 446824448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366599799398195,
      "loss": 2.9051,
      "theoretical_loss": 3.9629580702103686,
      "tokens_seen": 446889984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366499498495486,
      "loss": 2.9917,
      "theoretical_loss": 3.9628942636826263,
      "tokens_seen": 446955520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043663991975927786,
      "loss": 2.7779,
      "theoretical_loss": 3.9628304691291785,
      "tokens_seen": 447021056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043662988966900704,
      "loss": 3.0163,
      "theoretical_loss": 3.962766686546024,
      "tokens_seen": 447086592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366198595787362,
      "loss": 2.7074,
      "theoretical_loss": 3.962702915929162,
      "tokens_seen": 447152128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366098294884654,
      "loss": 2.8268,
      "theoretical_loss": 3.9626391572745945,
      "tokens_seen": 447217664
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 258449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.09433913230896,
      "objective/train/theoretical_loss": 3.9625754105783244,
      "objective/train/tokens_used": 467743200,
      "theoretical_loss": 3.9625754105783244,
      "tokens_seen": 447283200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365997993981946,
      "loss": 2.9999,
      "theoretical_loss": 3.9625754105783244,
      "tokens_seen": 447283200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043658976930792376,
      "loss": 2.6759,
      "theoretical_loss": 3.9625116758363577,
      "tokens_seen": 447348736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436579739217653,
      "loss": 3.012,
      "theoretical_loss": 3.9624479530447023,
      "tokens_seen": 447414272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043656970912738213,
      "loss": 2.8363,
      "theoretical_loss": 3.962384242199367,
      "tokens_seen": 447479808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043655967903711136,
      "loss": 2.981,
      "theoretical_loss": 3.962320543296363,
      "tokens_seen": 447545344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365496489468405,
      "loss": 2.9546,
      "theoretical_loss": 3.9622568563317038,
      "tokens_seen": 447610880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365396188565697,
      "loss": 2.7905,
      "theoretical_loss": 3.962193181301404,
      "tokens_seen": 447676416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365295887662989,
      "loss": 2.9201,
      "theoretical_loss": 3.9621295182014813,
      "tokens_seen": 447741952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365195586760281,
      "loss": 2.8073,
      "theoretical_loss": 3.9620658670279543,
      "tokens_seen": 447807488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043650952858575727,
      "loss": 3.0315,
      "theoretical_loss": 3.9620022277768436,
      "tokens_seen": 447873024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364994984954865,
      "loss": 2.869,
      "theoretical_loss": 3.9619386004441726,
      "tokens_seen": 447938560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043648946840521563,
      "loss": 2.9296,
      "theoretical_loss": 3.9618749850259647,
      "tokens_seen": 448004096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043647943831494487,
      "loss": 3.0547,
      "theoretical_loss": 3.961811381518248,
      "tokens_seen": 448069632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436469408224674,
      "loss": 3.042,
      "theoretical_loss": 3.96174778991705,
      "tokens_seen": 448135168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043645937813440323,
      "loss": 2.8965,
      "theoretical_loss": 3.961684210218401,
      "tokens_seen": 448200704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364493480441324,
      "loss": 2.8506,
      "theoretical_loss": 3.9616206424183336,
      "tokens_seen": 448266240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364393179538616,
      "loss": 2.8845,
      "theoretical_loss": 3.961557086512882,
      "tokens_seen": 448331776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043642928786359077,
      "loss": 2.824,
      "theoretical_loss": 3.961493542498083,
      "tokens_seen": 448397312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043641925777331995,
      "loss": 2.9255,
      "theoretical_loss": 3.9614300103699733,
      "tokens_seen": 448462848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043640922768304913,
      "loss": 2.9331,
      "theoretical_loss": 3.961366490124593,
      "tokens_seen": 448528384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043639919759277837,
      "loss": 3.0854,
      "theoretical_loss": 3.961302981757984,
      "tokens_seen": 448593920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363891675025075,
      "loss": 2.6079,
      "theoretical_loss": 3.9612394852661907,
      "tokens_seen": 448659456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043637913741223673,
      "loss": 2.8454,
      "theoretical_loss": 3.961176000645258,
      "tokens_seen": 448724992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363691073219659,
      "loss": 3.0051,
      "theoretical_loss": 3.9611125278912334,
      "tokens_seen": 448790528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363590772316951,
      "loss": 2.703,
      "theoretical_loss": 3.9610490670001663,
      "tokens_seen": 448856064
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 259697,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1037912368774414,
      "objective/train/theoretical_loss": 3.9609856179681078,
      "objective/train/tokens_used": 469381600,
      "theoretical_loss": 3.9609856179681078,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363490471414243,
      "loss": 3.0174,
      "theoretical_loss": 3.9609856179681078,
      "tokens_seen": 448921600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043633901705115346,
      "loss": 2.9102,
      "theoretical_loss": 3.9609221807911115,
      "tokens_seen": 448987136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043632898696088264,
      "loss": 2.832,
      "theoretical_loss": 3.960858755465232,
      "tokens_seen": 449052672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043631895687061187,
      "loss": 3.0512,
      "theoretical_loss": 3.9607953419865263,
      "tokens_seen": 449118208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436308926780341,
      "loss": 2.8357,
      "theoretical_loss": 3.960731940351053,
      "tokens_seen": 449183744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043629889669007023,
      "loss": 2.8914,
      "theoretical_loss": 3.9606685505548738,
      "tokens_seen": 449249280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043628886659979936,
      "loss": 2.6328,
      "theoretical_loss": 3.96060517259405,
      "tokens_seen": 449314816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362788365095286,
      "loss": 2.7281,
      "theoretical_loss": 3.9605418064646467,
      "tokens_seen": 449380352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362688064192578,
      "loss": 3.0245,
      "theoretical_loss": 3.9604784521627296,
      "tokens_seen": 449445888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043625877632898696,
      "loss": 3.0669,
      "theoretical_loss": 3.9604151096843676,
      "tokens_seen": 449511424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043624874623871614,
      "loss": 2.9458,
      "theoretical_loss": 3.9603517790256313,
      "tokens_seen": 449576960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362387161484453,
      "loss": 2.9036,
      "theoretical_loss": 3.960288460182591,
      "tokens_seen": 449642496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362286860581745,
      "loss": 3.0075,
      "theoretical_loss": 3.960225153151322,
      "tokens_seen": 449708032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043621865596790374,
      "loss": 2.8153,
      "theoretical_loss": 3.9601618579279,
      "tokens_seen": 449773568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043620862587763286,
      "loss": 2.9289,
      "theoretical_loss": 3.960098574508402,
      "tokens_seen": 449839104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361985957873621,
      "loss": 3.0163,
      "theoretical_loss": 3.9600353028889073,
      "tokens_seen": 449904640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361885656970913,
      "loss": 3.0982,
      "theoretical_loss": 3.9599720430654983,
      "tokens_seen": 449970176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043617853560682046,
      "loss": 2.7447,
      "theoretical_loss": 3.9599087950342566,
      "tokens_seen": 450035712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361685055165497,
      "loss": 2.9742,
      "theoretical_loss": 3.959845558791269,
      "tokens_seen": 450101248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361584754262788,
      "loss": 3.0095,
      "theoretical_loss": 3.9597823343326213,
      "tokens_seen": 450166784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043614844533600806,
      "loss": 2.6936,
      "theoretical_loss": 3.9597191216544028,
      "tokens_seen": 450232320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043613841524573724,
      "loss": 2.9194,
      "theoretical_loss": 3.959655920752704,
      "tokens_seen": 450297856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361283851554664,
      "loss": 2.8928,
      "theoretical_loss": 3.959592731623618,
      "tokens_seen": 450363392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361183550651956,
      "loss": 2.9382,
      "theoretical_loss": 3.9595295542632387,
      "tokens_seen": 450428928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361083249749248,
      "loss": 2.8739,
      "theoretical_loss": 3.9594663886676624,
      "tokens_seen": 450494464
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 260246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8196277618408203,
      "objective/train/theoretical_loss": 3.9594032348329877,
      "objective/train/tokens_used": 471020000,
      "theoretical_loss": 3.9594032348329877,
      "tokens_seen": 450560000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043609829488465397,
      "loss": 2.8036,
      "theoretical_loss": 3.9594032348329877,
      "tokens_seen": 450560000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360882647943832,
      "loss": 2.8566,
      "theoretical_loss": 3.959340092755314,
      "tokens_seen": 450625536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043607823470411233,
      "loss": 2.8465,
      "theoretical_loss": 3.959276962430744,
      "tokens_seen": 450691072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043606820461384156,
      "loss": 2.9252,
      "theoretical_loss": 3.9592138438553808,
      "tokens_seen": 450756608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360581745235707,
      "loss": 2.6681,
      "theoretical_loss": 3.95915073702533,
      "tokens_seen": 450822144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360481444332999,
      "loss": 2.8351,
      "theoretical_loss": 3.959087641936699,
      "tokens_seen": 450887680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360381143430291,
      "loss": 2.6263,
      "theoretical_loss": 3.9590245585855977,
      "tokens_seen": 450953216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360280842527583,
      "loss": 3.0796,
      "theoretical_loss": 3.9589614869681364,
      "tokens_seen": 451018752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043601805416248747,
      "loss": 2.8513,
      "theoretical_loss": 3.9588984270804293,
      "tokens_seen": 451084288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360080240722167,
      "loss": 2.8436,
      "theoretical_loss": 3.9588353789185895,
      "tokens_seen": 451149824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043599799398194583,
      "loss": 2.9357,
      "theoretical_loss": 3.9587723424787358,
      "tokens_seen": 451215360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043598796389167507,
      "loss": 2.5351,
      "theoretical_loss": 3.9587093177569854,
      "tokens_seen": 451280896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004359779338014042,
      "loss": 3.0153,
      "theoretical_loss": 3.958646304749459,
      "tokens_seen": 451346432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043596790371113343,
      "loss": 2.7466,
      "theoretical_loss": 3.9585833034522793,
      "tokens_seen": 451411968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004359578736208626,
      "loss": 2.8312,
      "theoretical_loss": 3.9585203138615697,
      "tokens_seen": 451477504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004359478435305918,
      "loss": 2.7963,
      "theoretical_loss": 3.9584573359734563,
      "tokens_seen": 451543040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043593781344032097,
      "loss": 2.7758,
      "theoretical_loss": 3.958394369784067,
      "tokens_seen": 451608576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043592778335005015,
      "loss": 2.8486,
      "theoretical_loss": 3.9583314152895324,
      "tokens_seen": 451674112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043591775325977933,
      "loss": 2.956,
      "theoretical_loss": 3.9582684724859822,
      "tokens_seen": 451739648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043590772316950857,
      "loss": 2.7166,
      "theoretical_loss": 3.9582055413695514,
      "tokens_seen": 451805184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358976930792377,
      "loss": 3.0765,
      "theoretical_loss": 3.9581426219363745,
      "tokens_seen": 451870720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043588766298896693,
      "loss": 2.679,
      "theoretical_loss": 3.958079714182588,
      "tokens_seen": 451936256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358776328986961,
      "loss": 2.9488,
      "theoretical_loss": 3.958016818104331,
      "tokens_seen": 452001792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358676028084253,
      "loss": 2.8143,
      "theoretical_loss": 3.957953933697745,
      "tokens_seen": 452067328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358575727181545,
      "loss": 3.0067,
      "theoretical_loss": 3.9578910609589713,
      "tokens_seen": 452132864
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 261277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.868929147720337,
      "objective/train/theoretical_loss": 3.957828199884155,
      "objective/train/tokens_used": 472658400,
      "theoretical_loss": 3.957828199884155,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043584754262788366,
      "loss": 2.8863,
      "theoretical_loss": 3.957828199884155,
      "tokens_seen": 452198400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043583751253761284,
      "loss": 2.8311,
      "theoretical_loss": 3.9577653504694426,
      "tokens_seen": 452263936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043582748244734207,
      "loss": 2.8063,
      "theoretical_loss": 3.9577025127109815,
      "tokens_seen": 452329472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358174523570712,
      "loss": 2.7761,
      "theoretical_loss": 3.957639686604921,
      "tokens_seen": 452395008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043580742226680043,
      "loss": 2.8938,
      "theoretical_loss": 3.9575768721474143,
      "tokens_seen": 452460544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043579739217652956,
      "loss": 3.0099,
      "theoretical_loss": 3.9575140693346134,
      "tokens_seen": 452526080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357873620862588,
      "loss": 2.7552,
      "theoretical_loss": 3.957451278162674,
      "tokens_seen": 452591616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435777331995988,
      "loss": 2.6856,
      "theoretical_loss": 3.957388498627754,
      "tokens_seen": 452657152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043576730190571716,
      "loss": 3.014,
      "theoretical_loss": 3.957325730726012,
      "tokens_seen": 452722688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043575727181544634,
      "loss": 2.7964,
      "theoretical_loss": 3.9572629744536085,
      "tokens_seen": 452788224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357472417251755,
      "loss": 2.9826,
      "theoretical_loss": 3.9572002298067064,
      "tokens_seen": 452853760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357372116349047,
      "loss": 2.7544,
      "theoretical_loss": 3.95713749678147,
      "tokens_seen": 452919296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043572718154463394,
      "loss": 2.8092,
      "theoretical_loss": 3.957074775374066,
      "tokens_seen": 452984832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043571715145436306,
      "loss": 2.7585,
      "theoretical_loss": 3.9570120655806615,
      "tokens_seen": 453050368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357071213640923,
      "loss": 2.8345,
      "theoretical_loss": 3.956949367397427,
      "tokens_seen": 453115904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356970912738215,
      "loss": 2.9775,
      "theoretical_loss": 3.956886680820535,
      "tokens_seen": 453181440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043568706118355066,
      "loss": 2.9825,
      "theoretical_loss": 3.9568240058461575,
      "tokens_seen": 453246976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043567703109327984,
      "loss": 2.964,
      "theoretical_loss": 3.956761342470471,
      "tokens_seen": 453312512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435667001003009,
      "loss": 2.8124,
      "theoretical_loss": 3.9566986906896515,
      "tokens_seen": 453378048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356569709127382,
      "loss": 3.0178,
      "theoretical_loss": 3.95663605049988,
      "tokens_seen": 453443584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043564694082246744,
      "loss": 2.9025,
      "theoretical_loss": 3.9565734218973354,
      "tokens_seen": 453509120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043563691073219657,
      "loss": 2.8165,
      "theoretical_loss": 3.9565108048782007,
      "tokens_seen": 453574656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356268806419258,
      "loss": 3.0191,
      "theoretical_loss": 3.9564481994386607,
      "tokens_seen": 453640192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043561685055165493,
      "loss": 2.9861,
      "theoretical_loss": 3.9563856055749014,
      "tokens_seen": 453705728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043560682046138417,
      "loss": 2.9445,
      "theoretical_loss": 3.9563230232831113,
      "tokens_seen": 453771264
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 261864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.966676712036133,
      "objective/train/theoretical_loss": 3.956260452559479,
      "objective/train/tokens_used": 474296800,
      "theoretical_loss": 3.956260452559479,
      "tokens_seen": 453836800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043559679037111335,
      "loss": 2.7952,
      "theoretical_loss": 3.956260452559479,
      "tokens_seen": 453836800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043558676028084253,
      "loss": 3.0193,
      "theoretical_loss": 3.956197893400198,
      "tokens_seen": 453902336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355767301905717,
      "loss": 3.02,
      "theoretical_loss": 3.95613534580146,
      "tokens_seen": 453967872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355667001003009,
      "loss": 2.995,
      "theoretical_loss": 3.956072809759461,
      "tokens_seen": 454033408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043555667001003007,
      "loss": 2.8422,
      "theoretical_loss": 3.9560102852703976,
      "tokens_seen": 454098944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355466399197593,
      "loss": 2.8431,
      "theoretical_loss": 3.955947772330469,
      "tokens_seen": 454164480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043553660982948843,
      "loss": 2.8094,
      "theoretical_loss": 3.9558852709358763,
      "tokens_seen": 454230016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043552657973921767,
      "loss": 2.8377,
      "theoretical_loss": 3.9558227810828215,
      "tokens_seen": 454295552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043551654964894685,
      "loss": 2.6956,
      "theoretical_loss": 3.9557603027675086,
      "tokens_seen": 454361088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043550651955867603,
      "loss": 2.7347,
      "theoretical_loss": 3.955697835986143,
      "tokens_seen": 454426624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354964894684052,
      "loss": 2.8711,
      "theoretical_loss": 3.9556353807349343,
      "tokens_seen": 454492160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354864593781344,
      "loss": 2.9731,
      "theoretical_loss": 3.955572937010091,
      "tokens_seen": 454557696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354764292878636,
      "loss": 2.93,
      "theoretical_loss": 3.9555105048078243,
      "tokens_seen": 454623232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354663991975928,
      "loss": 2.7183,
      "theoretical_loss": 3.9554480841243476,
      "tokens_seen": 454688768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043545636910732194,
      "loss": 2.9553,
      "theoretical_loss": 3.9553856749558762,
      "tokens_seen": 454754304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043544633901705117,
      "loss": 2.8743,
      "theoretical_loss": 3.9553232772986266,
      "tokens_seen": 454819840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354363089267803,
      "loss": 2.999,
      "theoretical_loss": 3.955260891148818,
      "tokens_seen": 454885376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043542627883650953,
      "loss": 2.8236,
      "theoretical_loss": 3.9551985165026693,
      "tokens_seen": 454950912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043541624874623877,
      "loss": 3.0874,
      "theoretical_loss": 3.955136153356404,
      "tokens_seen": 455016448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354062186559679,
      "loss": 2.794,
      "theoretical_loss": 3.955073801706246,
      "tokens_seen": 455081984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043539618856569713,
      "loss": 2.8035,
      "theoretical_loss": 3.9550114615484198,
      "tokens_seen": 455147520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353861584754263,
      "loss": 2.6472,
      "theoretical_loss": 3.954949132879154,
      "tokens_seen": 455213056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353761283851555,
      "loss": 2.7814,
      "theoretical_loss": 3.9548868156946777,
      "tokens_seen": 455278592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353660982948847,
      "loss": 2.9024,
      "theoretical_loss": 3.9548245099912216,
      "tokens_seen": 455344128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043535606820461386,
      "loss": 2.7629,
      "theoretical_loss": 3.954762215765019,
      "tokens_seen": 455409664
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 263055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.755061388015747,
      "objective/train/theoretical_loss": 3.9546999330123036,
      "objective/train/tokens_used": 475935200,
      "theoretical_loss": 3.9546999330123036,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043534603811434304,
      "loss": 2.8091,
      "theoretical_loss": 3.9546999330123036,
      "tokens_seen": 455475200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043533600802407227,
      "loss": 2.7972,
      "theoretical_loss": 3.954637661729312,
      "tokens_seen": 455540736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353259779338014,
      "loss": 2.7967,
      "theoretical_loss": 3.954575401912284,
      "tokens_seen": 455606272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043531594784353063,
      "loss": 2.8888,
      "theoretical_loss": 3.954513153557458,
      "tokens_seen": 455671808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043530591775325976,
      "loss": 2.7848,
      "theoretical_loss": 3.9544509166610755,
      "tokens_seen": 455737344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435295887662989,
      "loss": 2.7125,
      "theoretical_loss": 3.95438869121938,
      "tokens_seen": 455802880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352858575727182,
      "loss": 2.8886,
      "theoretical_loss": 3.9543264772286184,
      "tokens_seen": 455868416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043527582748244736,
      "loss": 2.7964,
      "theoretical_loss": 3.954264274685036,
      "tokens_seen": 455933952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043526579739217654,
      "loss": 3.0038,
      "theoretical_loss": 3.954202083584882,
      "tokens_seen": 455999488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352557673019057,
      "loss": 3.0308,
      "theoretical_loss": 3.954139903924407,
      "tokens_seen": 456065024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352457372116349,
      "loss": 2.8824,
      "theoretical_loss": 3.9540777356998635,
      "tokens_seen": 456130560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043523570712136414,
      "loss": 2.8996,
      "theoretical_loss": 3.954015578907506,
      "tokens_seen": 456196096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043522567703109326,
      "loss": 2.8487,
      "theoretical_loss": 3.953953433543589,
      "tokens_seen": 456261632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352156469408225,
      "loss": 2.7798,
      "theoretical_loss": 3.953891299604371,
      "tokens_seen": 456327168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352056168505517,
      "loss": 2.8852,
      "theoretical_loss": 3.9538291770861123,
      "tokens_seen": 456392704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043519558676028086,
      "loss": 3.0216,
      "theoretical_loss": 3.9537670659850725,
      "tokens_seen": 456458240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043518555667001004,
      "loss": 2.8016,
      "theoretical_loss": 3.953704966297515,
      "tokens_seen": 456523776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351755265797392,
      "loss": 2.7142,
      "theoretical_loss": 3.953642878019705,
      "tokens_seen": 456589312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351654964894684,
      "loss": 2.846,
      "theoretical_loss": 3.953580801147908,
      "tokens_seen": 456654848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043515546639919764,
      "loss": 2.9923,
      "theoretical_loss": 3.9535187356783927,
      "tokens_seen": 456720384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043514543630892677,
      "loss": 2.7513,
      "theoretical_loss": 3.9534566816074292,
      "tokens_seen": 456785920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435135406218656,
      "loss": 2.7321,
      "theoretical_loss": 3.953394638931289,
      "tokens_seen": 456851456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043512537612838513,
      "loss": 2.8101,
      "theoretical_loss": 3.953332607646246,
      "tokens_seen": 456916992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043511534603811437,
      "loss": 2.754,
      "theoretical_loss": 3.953270587748574,
      "tokens_seen": 456982528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043510531594784355,
      "loss": 2.8689,
      "theoretical_loss": 3.9532085792345515,
      "tokens_seen": 457048064
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 264378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8818917274475098,
      "objective/train/theoretical_loss": 3.953146582100456,
      "objective/train/tokens_used": 477573600,
      "theoretical_loss": 3.953146582100456,
      "tokens_seen": 457113600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043509528585757273,
      "loss": 2.9812,
      "theoretical_loss": 3.953146582100456,
      "tokens_seen": 457113600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350852557673019,
      "loss": 2.7279,
      "theoretical_loss": 3.9530845963425696,
      "tokens_seen": 457179136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350752256770311,
      "loss": 2.69,
      "theoretical_loss": 3.9530226219571727,
      "tokens_seen": 457244672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043506519558676027,
      "loss": 2.698,
      "theoretical_loss": 3.9529606589405497,
      "tokens_seen": 457310208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350551654964895,
      "loss": 2.8551,
      "theoretical_loss": 3.952898707288987,
      "tokens_seen": 457375744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043504513540621863,
      "loss": 2.8347,
      "theoretical_loss": 3.9528367669987716,
      "tokens_seen": 457441280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043503510531594787,
      "loss": 2.828,
      "theoretical_loss": 3.952774838066193,
      "tokens_seen": 457506816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043502507522567705,
      "loss": 2.8248,
      "theoretical_loss": 3.9527129204875413,
      "tokens_seen": 457572352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043501504513540623,
      "loss": 3.1028,
      "theoretical_loss": 3.95265101425911,
      "tokens_seen": 457637888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350050150451354,
      "loss": 3.2986,
      "theoretical_loss": 3.952589119377193,
      "tokens_seen": 457703424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349949849548646,
      "loss": 2.7862,
      "theoretical_loss": 3.9525272358380867,
      "tokens_seen": 457768960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349849548645938,
      "loss": 2.91,
      "theoretical_loss": 3.9524653636380895,
      "tokens_seen": 457834496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434974924774323,
      "loss": 2.9778,
      "theoretical_loss": 3.9524035027734996,
      "tokens_seen": 457900032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043496489468405214,
      "loss": 2.9951,
      "theoretical_loss": 3.9523416532406204,
      "tokens_seen": 457965568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043495486459378137,
      "loss": 2.9673,
      "theoretical_loss": 3.9522798150357525,
      "tokens_seen": 458031104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349448345035105,
      "loss": 2.7049,
      "theoretical_loss": 3.952217988155203,
      "tokens_seen": 458096640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043493480441323973,
      "loss": 2.7651,
      "theoretical_loss": 3.952156172595277,
      "tokens_seen": 458162176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349247743229689,
      "loss": 2.9111,
      "theoretical_loss": 3.952094368352284,
      "tokens_seen": 458227712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349147442326981,
      "loss": 3.0533,
      "theoretical_loss": 3.9520325754225327,
      "tokens_seen": 458293248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004349047141424273,
      "loss": 2.8072,
      "theoretical_loss": 3.951970793802336,
      "tokens_seen": 458358784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348946840521565,
      "loss": 2.4948,
      "theoretical_loss": 3.9519090234880068,
      "tokens_seen": 458424320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043488465396188564,
      "loss": 2.8712,
      "theoretical_loss": 3.9518472644758607,
      "tokens_seen": 458489856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348746238716149,
      "loss": 2.9302,
      "theoretical_loss": 3.951785516762214,
      "tokens_seen": 458555392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434864593781344,
      "loss": 2.7907,
      "theoretical_loss": 3.951723780343386,
      "tokens_seen": 458620928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043485456369107324,
      "loss": 2.9902,
      "theoretical_loss": 3.951662055215697,
      "tokens_seen": 458686464
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 264937,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.118872880935669,
      "objective/train/theoretical_loss": 3.951600341375469,
      "objective/train/tokens_used": 479212000,
      "theoretical_loss": 3.951600341375469,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348445336008024,
      "loss": 3.0182,
      "theoretical_loss": 3.951600341375469,
      "tokens_seen": 458752000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348345035105316,
      "loss": 2.8335,
      "theoretical_loss": 3.9515386388190255,
      "tokens_seen": 458817536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348244734202608,
      "loss": 2.8455,
      "theoretical_loss": 3.9514769475426927,
      "tokens_seen": 458883072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043481444332998996,
      "loss": 2.9251,
      "theoretical_loss": 3.9514152675427976,
      "tokens_seen": 458948608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043480441323971914,
      "loss": 2.7918,
      "theoretical_loss": 3.9513535988156696,
      "tokens_seen": 459014144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347943831494484,
      "loss": 2.7582,
      "theoretical_loss": 3.9512919413576384,
      "tokens_seen": 459079680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347843530591775,
      "loss": 2.9296,
      "theoretical_loss": 3.951230295165038,
      "tokens_seen": 459145216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043477432296890674,
      "loss": 2.8319,
      "theoretical_loss": 3.951168660234201,
      "tokens_seen": 459210752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043476429287863587,
      "loss": 2.6616,
      "theoretical_loss": 3.9511070365614644,
      "tokens_seen": 459276288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347542627883651,
      "loss": 2.8487,
      "theoretical_loss": 3.951045424143165,
      "tokens_seen": 459341824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347442326980943,
      "loss": 2.7885,
      "theoretical_loss": 3.9509838229756427,
      "tokens_seen": 459407360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043473420260782346,
      "loss": 2.9401,
      "theoretical_loss": 3.950922233055238,
      "tokens_seen": 459472896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043472417251755265,
      "loss": 2.8083,
      "theoretical_loss": 3.9508606543782943,
      "tokens_seen": 459538432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347141424272819,
      "loss": 2.8683,
      "theoretical_loss": 3.9507990869411556,
      "tokens_seen": 459603968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434704112337011,
      "loss": 2.7853,
      "theoretical_loss": 3.9507375307401684,
      "tokens_seen": 459669504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043469408224674024,
      "loss": 3.0062,
      "theoretical_loss": 3.9506759857716798,
      "tokens_seen": 459735040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346840521564694,
      "loss": 2.7083,
      "theoretical_loss": 3.95061445203204,
      "tokens_seen": 459800576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346740220661986,
      "loss": 3.0644,
      "theoretical_loss": 3.9505529295176,
      "tokens_seen": 459866112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043466399197592784,
      "loss": 2.6074,
      "theoretical_loss": 3.950491418224713,
      "tokens_seen": 459931648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043465396188565697,
      "loss": 2.8359,
      "theoretical_loss": 3.9504299181497338,
      "tokens_seen": 459997184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346439317953862,
      "loss": 2.6802,
      "theoretical_loss": 3.9503684292890187,
      "tokens_seen": 460062720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043463390170511533,
      "loss": 3.035,
      "theoretical_loss": 3.9503069516389253,
      "tokens_seen": 460128256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043462387161484457,
      "loss": 2.9591,
      "theoretical_loss": 3.950245485195814,
      "tokens_seen": 460193792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043461384152457375,
      "loss": 2.8343,
      "theoretical_loss": 3.9501840299560467,
      "tokens_seen": 460259328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043460381143430293,
      "loss": 2.5444,
      "theoretical_loss": 3.950122585915985,
      "tokens_seen": 460324864
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 265992,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.820603370666504,
      "objective/train/theoretical_loss": 3.9500611530719953,
      "objective/train/tokens_used": 480850400,
      "theoretical_loss": 3.9500611530719953,
      "tokens_seen": 460390400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345937813440321,
      "loss": 2.7728,
      "theoretical_loss": 3.9500611530719953,
      "tokens_seen": 460390400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345837512537613,
      "loss": 2.894,
      "theoretical_loss": 3.9499997314204434,
      "tokens_seen": 460455936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043457372116349047,
      "loss": 2.8096,
      "theoretical_loss": 3.949938320957698,
      "tokens_seen": 460521472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345636910732197,
      "loss": 2.4647,
      "theoretical_loss": 3.949876921680129,
      "tokens_seen": 460587008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043455366098294883,
      "loss": 3.0128,
      "theoretical_loss": 3.9498155335841076,
      "tokens_seen": 460652544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043454363089267807,
      "loss": 2.9574,
      "theoretical_loss": 3.949754156666008,
      "tokens_seen": 460718080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043453360080240725,
      "loss": 2.8107,
      "theoretical_loss": 3.949692790922205,
      "tokens_seen": 460783616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043452357071213643,
      "loss": 2.7675,
      "theoretical_loss": 3.9496314363490743,
      "tokens_seen": 460849152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345135406218656,
      "loss": 2.7283,
      "theoretical_loss": 3.949570092942996,
      "tokens_seen": 460914688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345035105315948,
      "loss": 2.7483,
      "theoretical_loss": 3.949508760700349,
      "tokens_seen": 460980224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434493480441324,
      "loss": 2.7427,
      "theoretical_loss": 3.949447439617516,
      "tokens_seen": 461045760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344834503510532,
      "loss": 2.9061,
      "theoretical_loss": 3.94938612969088,
      "tokens_seen": 461111296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043447342026078234,
      "loss": 2.9127,
      "theoretical_loss": 3.949324830916826,
      "tokens_seen": 461176832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043446339017051157,
      "loss": 2.9454,
      "theoretical_loss": 3.949263543291741,
      "tokens_seen": 461242368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344533600802407,
      "loss": 2.848,
      "theoretical_loss": 3.949202266812014,
      "tokens_seen": 461307904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043444332998996993,
      "loss": 3.0778,
      "theoretical_loss": 3.9491410014740347,
      "tokens_seen": 461373440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344332998996991,
      "loss": 2.9834,
      "theoretical_loss": 3.949079747274195,
      "tokens_seen": 461438976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344232698094283,
      "loss": 2.8894,
      "theoretical_loss": 3.9490185042088894,
      "tokens_seen": 461504512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344132397191575,
      "loss": 2.653,
      "theoretical_loss": 3.948957272274512,
      "tokens_seen": 461570048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344032096288867,
      "loss": 2.8366,
      "theoretical_loss": 3.94889605146746,
      "tokens_seen": 461635584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043439317953861584,
      "loss": 2.9733,
      "theoretical_loss": 3.948834841784133,
      "tokens_seen": 461701120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343831494483451,
      "loss": 2.7639,
      "theoretical_loss": 3.9487736432209304,
      "tokens_seen": 461766656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343731193580742,
      "loss": 3.0074,
      "theoretical_loss": 3.9487124557742543,
      "tokens_seen": 461832192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043436308926780344,
      "loss": 3.0052,
      "theoretical_loss": 3.948651279440509,
      "tokens_seen": 461897728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343530591775326,
      "loss": 3.129,
      "theoretical_loss": 3.948590114216098,
      "tokens_seen": 461963264
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 266575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.696946144104004,
      "objective/train/theoretical_loss": 3.9485289600974305,
      "objective/train/tokens_used": 482488800,
      "theoretical_loss": 3.9485289600974305,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343430290872618,
      "loss": 2.6741,
      "theoretical_loss": 3.9485289600974305,
      "tokens_seen": 462028800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434332998996991,
      "loss": 2.8477,
      "theoretical_loss": 3.948467817080915,
      "tokens_seen": 462094336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043432296890672016,
      "loss": 3.0804,
      "theoretical_loss": 3.94840668516296,
      "tokens_seen": 462159872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043431293881644934,
      "loss": 2.9808,
      "theoretical_loss": 3.9483455643399794,
      "tokens_seen": 462225408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343029087261786,
      "loss": 3.0524,
      "theoretical_loss": 3.9482844546083857,
      "tokens_seen": 462290944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342928786359077,
      "loss": 2.8848,
      "theoretical_loss": 3.948223355964595,
      "tokens_seen": 462356480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043428284854563694,
      "loss": 2.9191,
      "theoretical_loss": 3.948162268405024,
      "tokens_seen": 462422016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043427281845536607,
      "loss": 2.7022,
      "theoretical_loss": 3.9481011919260913,
      "tokens_seen": 462487552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342627883650953,
      "loss": 2.759,
      "theoretical_loss": 3.9480401265242175,
      "tokens_seen": 462553088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342527582748245,
      "loss": 2.7993,
      "theoretical_loss": 3.9479790721958246,
      "tokens_seen": 462618624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043424272818455367,
      "loss": 2.7651,
      "theoretical_loss": 3.9479180289373366,
      "tokens_seen": 462684160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043423269809428285,
      "loss": 2.8418,
      "theoretical_loss": 3.9478569967451778,
      "tokens_seen": 462749696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342226680040121,
      "loss": 2.7632,
      "theoretical_loss": 3.947795975615776,
      "tokens_seen": 462815232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342126379137412,
      "loss": 2.9218,
      "theoretical_loss": 3.9477349655455596,
      "tokens_seen": 462880768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043420260782347044,
      "loss": 2.9163,
      "theoretical_loss": 3.9476739665309593,
      "tokens_seen": 462946304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043419257773319957,
      "loss": 2.6965,
      "theoretical_loss": 3.947612978568406,
      "tokens_seen": 463011840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341825476429288,
      "loss": 2.9593,
      "theoretical_loss": 3.9475520016543353,
      "tokens_seen": 463077376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434172517552658,
      "loss": 2.9816,
      "theoretical_loss": 3.9474910357851805,
      "tokens_seen": 463142912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043416248746238717,
      "loss": 2.9265,
      "theoretical_loss": 3.9474300809573797,
      "tokens_seen": 463208448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043415245737211635,
      "loss": 2.7887,
      "theoretical_loss": 3.947369137167371,
      "tokens_seen": 463273984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043414242728184553,
      "loss": 2.7265,
      "theoretical_loss": 3.947308204411595,
      "tokens_seen": 463339520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341323971915747,
      "loss": 2.8877,
      "theoretical_loss": 3.947247282686493,
      "tokens_seen": 463405056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043412236710130395,
      "loss": 2.8725,
      "theoretical_loss": 3.947186371988509,
      "tokens_seen": 463470592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341123370110331,
      "loss": 2.7452,
      "theoretical_loss": 3.947125472314088,
      "tokens_seen": 463536128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341023069207623,
      "loss": 2.7315,
      "theoretical_loss": 3.947064583659678,
      "tokens_seen": 463601664
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 267814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8467938899993896,
      "objective/train/theoretical_loss": 3.9470037060217256,
      "objective/train/tokens_used": 484127200,
      "theoretical_loss": 3.9470037060217256,
      "tokens_seen": 463667200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043409227683049144,
      "loss": 2.9119,
      "theoretical_loss": 3.9470037060217256,
      "tokens_seen": 463667200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043408224674022067,
      "loss": 2.7476,
      "theoretical_loss": 3.946942839396682,
      "tokens_seen": 463732736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043407221664994985,
      "loss": 2.8435,
      "theoretical_loss": 3.946881983780999,
      "tokens_seen": 463798272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043406218655967903,
      "loss": 3.0763,
      "theoretical_loss": 3.9468211391711296,
      "tokens_seen": 463863808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340521564694082,
      "loss": 2.8394,
      "theoretical_loss": 3.9467603055635294,
      "tokens_seen": 463929344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043404212637913745,
      "loss": 2.9131,
      "theoretical_loss": 3.9466994829546542,
      "tokens_seen": 463994880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340320962888666,
      "loss": 3.0057,
      "theoretical_loss": 3.946638671340964,
      "tokens_seen": 464060416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340220661985958,
      "loss": 2.8754,
      "theoretical_loss": 3.946577870718917,
      "tokens_seen": 464125952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043401203610832494,
      "loss": 3.0896,
      "theoretical_loss": 3.9465170810849757,
      "tokens_seen": 464191488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340020060180542,
      "loss": 2.7866,
      "theoretical_loss": 3.946456302435603,
      "tokens_seen": 464257024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043399197592778336,
      "loss": 2.8963,
      "theoretical_loss": 3.946395534767265,
      "tokens_seen": 464322560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043398194583751254,
      "loss": 2.9696,
      "theoretical_loss": 3.946334778076427,
      "tokens_seen": 464388096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339719157472417,
      "loss": 2.8131,
      "theoretical_loss": 3.9462740323595575,
      "tokens_seen": 464453632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339618856569709,
      "loss": 2.8574,
      "theoretical_loss": 3.9462132976131263,
      "tokens_seen": 464519168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339518555667001,
      "loss": 2.7606,
      "theoretical_loss": 3.946152573833605,
      "tokens_seen": 464584704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339418254764293,
      "loss": 3.0352,
      "theoretical_loss": 3.9460918610174662,
      "tokens_seen": 464650240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339317953861585,
      "loss": 2.7562,
      "theoretical_loss": 3.946031159161185,
      "tokens_seen": 464715776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339217652958877,
      "loss": 2.8604,
      "theoretical_loss": 3.9459704682612387,
      "tokens_seen": 464781312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339117352056169,
      "loss": 2.8481,
      "theoretical_loss": 3.945909788314104,
      "tokens_seen": 464846848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043390170511534604,
      "loss": 2.7717,
      "theoretical_loss": 3.945849119316261,
      "tokens_seen": 464912384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338916750250753,
      "loss": 2.7045,
      "theoretical_loss": 3.9457884612641902,
      "tokens_seen": 464977920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338816449348044,
      "loss": 2.9895,
      "theoretical_loss": 3.9457278141543752,
      "tokens_seen": 465043456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043387161484453364,
      "loss": 2.8658,
      "theoretical_loss": 3.9456671779833004,
      "tokens_seen": 465108992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338615847542628,
      "loss": 2.9028,
      "theoretical_loss": 3.945606552747452,
      "tokens_seen": 465174528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433851554663992,
      "loss": 2.954,
      "theoretical_loss": 3.945545938443318,
      "tokens_seen": 465240064
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 268427,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.444042682647705,
      "objective/train/theoretical_loss": 3.945485335067386,
      "objective/train/tokens_used": 485765600,
      "theoretical_loss": 3.945485335067386,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338415245737212,
      "loss": 2.6741,
      "theoretical_loss": 3.945485335067386,
      "tokens_seen": 465305600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043383149448345036,
      "loss": 2.9785,
      "theoretical_loss": 3.945424742616149,
      "tokens_seen": 465371136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043382146439317954,
      "loss": 2.8609,
      "theoretical_loss": 3.945364161086099,
      "tokens_seen": 465436672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338114343029088,
      "loss": 2.82,
      "theoretical_loss": 3.9453035904737304,
      "tokens_seen": 465502208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338014042126379,
      "loss": 2.964,
      "theoretical_loss": 3.9452430307755386,
      "tokens_seen": 465567744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043379137412236714,
      "loss": 2.9767,
      "theoretical_loss": 3.945182481988021,
      "tokens_seen": 465633280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043378134403209627,
      "loss": 2.921,
      "theoretical_loss": 3.945121944107677,
      "tokens_seen": 465698816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337713139418255,
      "loss": 2.8756,
      "theoretical_loss": 3.945061417131007,
      "tokens_seen": 465764352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337612838515547,
      "loss": 2.721,
      "theoretical_loss": 3.945000901054514,
      "tokens_seen": 465829888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043375125376128387,
      "loss": 2.9334,
      "theoretical_loss": 3.944940395874701,
      "tokens_seen": 465895424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043374122367101305,
      "loss": 2.9868,
      "theoretical_loss": 3.944879901588074,
      "tokens_seen": 465960960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337311935807423,
      "loss": 2.7399,
      "theoretical_loss": 3.9448194181911402,
      "tokens_seen": 466026496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337211634904714,
      "loss": 3.0384,
      "theoretical_loss": 3.944758945680408,
      "tokens_seen": 466092032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043371113340020064,
      "loss": 2.8814,
      "theoretical_loss": 3.9446984840523873,
      "tokens_seen": 466157568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043370110330992977,
      "loss": 2.6462,
      "theoretical_loss": 3.9446380333035918,
      "tokens_seen": 466223104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433691073219659,
      "loss": 2.6692,
      "theoretical_loss": 3.944577593430534,
      "tokens_seen": 466288640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336810431293882,
      "loss": 2.8248,
      "theoretical_loss": 3.9445171644297288,
      "tokens_seen": 466354176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043367101303911737,
      "loss": 2.6765,
      "theoretical_loss": 3.9444567462976936,
      "tokens_seen": 466419712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043366098294884655,
      "loss": 2.8589,
      "theoretical_loss": 3.9443963390309458,
      "tokens_seen": 466485248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043365095285857573,
      "loss": 2.9641,
      "theoretical_loss": 3.9443359426260063,
      "tokens_seen": 466550784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336409227683049,
      "loss": 2.8887,
      "theoretical_loss": 3.944275557079397,
      "tokens_seen": 466616320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043363089267803415,
      "loss": 2.9296,
      "theoretical_loss": 3.9442151823876404,
      "tokens_seen": 466681856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336208625877633,
      "loss": 2.7081,
      "theoretical_loss": 3.9441548185472612,
      "tokens_seen": 466747392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336108324974925,
      "loss": 2.7055,
      "theoretical_loss": 3.9440944655547865,
      "tokens_seen": 466812928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043360080240722164,
      "loss": 3.0324,
      "theoretical_loss": 3.944034123406744,
      "tokens_seen": 466878464
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 269429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1784250736236572,
      "objective/train/theoretical_loss": 3.943973792099663,
      "objective/train/tokens_used": 487404000,
      "theoretical_loss": 3.943973792099663,
      "tokens_seen": 466944000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043359077231695087,
      "loss": 2.9881,
      "theoretical_loss": 3.943973792099663,
      "tokens_seen": 466944000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043358074222668005,
      "loss": 2.8874,
      "theoretical_loss": 3.943913471630075,
      "tokens_seen": 467009536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043357071213640923,
      "loss": 2.8577,
      "theoretical_loss": 3.943853161994513,
      "tokens_seen": 467075072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335606820461384,
      "loss": 2.9665,
      "theoretical_loss": 3.943792863189511,
      "tokens_seen": 467140608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043355065195586765,
      "loss": 2.8564,
      "theoretical_loss": 3.9437325752116053,
      "tokens_seen": 467206144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335406218655968,
      "loss": 2.8001,
      "theoretical_loss": 3.943672298057333,
      "tokens_seen": 467271680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433530591775326,
      "loss": 3.0148,
      "theoretical_loss": 3.9436120317232337,
      "tokens_seen": 467337216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043352056168505514,
      "loss": 2.9508,
      "theoretical_loss": 3.9435517762058483,
      "tokens_seen": 467402752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335105315947844,
      "loss": 2.9171,
      "theoretical_loss": 3.9434915315017185,
      "tokens_seen": 467468288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043350050150451356,
      "loss": 2.7347,
      "theoretical_loss": 3.943431297607389,
      "tokens_seen": 467533824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043349047141424274,
      "loss": 2.8175,
      "theoretical_loss": 3.943371074519405,
      "tokens_seen": 467599360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334804413239719,
      "loss": 2.5834,
      "theoretical_loss": 3.943310862234313,
      "tokens_seen": 467664896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334704112337011,
      "loss": 2.8371,
      "theoretical_loss": 3.9432506607486633,
      "tokens_seen": 467730432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334603811434303,
      "loss": 2.6997,
      "theoretical_loss": 3.9431904700590046,
      "tokens_seen": 467795968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334503510531595,
      "loss": 2.9174,
      "theoretical_loss": 3.9431302901618897,
      "tokens_seen": 467861504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043344032096288864,
      "loss": 2.9733,
      "theoretical_loss": 3.943070121053872,
      "tokens_seen": 467927040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334302908726179,
      "loss": 3.0873,
      "theoretical_loss": 3.943009962731506,
      "tokens_seen": 467992576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433420260782347,
      "loss": 2.8206,
      "theoretical_loss": 3.9429498151913487,
      "tokens_seen": 468058112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043341023069207624,
      "loss": 2.6554,
      "theoretical_loss": 3.9428896784299585,
      "tokens_seen": 468123648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334002006018054,
      "loss": 2.8164,
      "theoretical_loss": 3.9428295524438948,
      "tokens_seen": 468189184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333901705115346,
      "loss": 2.8438,
      "theoretical_loss": 3.9427694372297193,
      "tokens_seen": 468254720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333801404212638,
      "loss": 2.8599,
      "theoretical_loss": 3.942709332783995,
      "tokens_seen": 468320256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433370110330993,
      "loss": 2.7072,
      "theoretical_loss": 3.942649239103286,
      "tokens_seen": 468385792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043336008024072215,
      "loss": 2.7251,
      "theoretical_loss": 3.942589156184159,
      "tokens_seen": 468451328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333500501504514,
      "loss": 2.8781,
      "theoretical_loss": 3.942529084023181,
      "tokens_seen": 468516864
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 269989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7550222873687744,
      "objective/train/theoretical_loss": 3.9424690226169212,
      "objective/train/tokens_used": 489042400,
      "theoretical_loss": 3.9424690226169212,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333400200601805,
      "loss": 2.889,
      "theoretical_loss": 3.9424690226169212,
      "tokens_seen": 468582400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043332998996990974,
      "loss": 2.9478,
      "theoretical_loss": 3.9424089719619513,
      "tokens_seen": 468647936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333199598796389,
      "loss": 2.7374,
      "theoretical_loss": 3.942348932054843,
      "tokens_seen": 468713472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333099297893681,
      "loss": 2.8397,
      "theoretical_loss": 3.942288902892171,
      "tokens_seen": 468779008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332998996990973,
      "loss": 2.8605,
      "theoretical_loss": 3.94222888447051,
      "tokens_seen": 468844544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043328986960882647,
      "loss": 2.8271,
      "theoretical_loss": 3.942168876786438,
      "tokens_seen": 468910080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043327983951855565,
      "loss": 2.7673,
      "theoretical_loss": 3.942108879836532,
      "tokens_seen": 468975616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332698094282849,
      "loss": 2.8577,
      "theoretical_loss": 3.942048893617374,
      "tokens_seen": 469041152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433259779338014,
      "loss": 2.8374,
      "theoretical_loss": 3.9419889181255456,
      "tokens_seen": 469106688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043324974924774325,
      "loss": 2.8189,
      "theoretical_loss": 3.941928953357629,
      "tokens_seen": 469172224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043323971915747243,
      "loss": 2.5255,
      "theoretical_loss": 3.941868999310211,
      "tokens_seen": 469237760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332296890672016,
      "loss": 2.6435,
      "theoretical_loss": 3.9418090559798764,
      "tokens_seen": 469303296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332196589769308,
      "loss": 2.8742,
      "theoretical_loss": 3.941749123363214,
      "tokens_seen": 469368832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043320962888665997,
      "loss": 2.7471,
      "theoretical_loss": 3.9416892014568132,
      "tokens_seen": 469434368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043319959879638915,
      "loss": 2.8973,
      "theoretical_loss": 3.941629290257266,
      "tokens_seen": 469499904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331895687061184,
      "loss": 2.7402,
      "theoretical_loss": 3.941569389761164,
      "tokens_seen": 469565440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043317953861584757,
      "loss": 2.7395,
      "theoretical_loss": 3.941509499965102,
      "tokens_seen": 469630976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043316950852557675,
      "loss": 2.9798,
      "theoretical_loss": 3.9414496208656766,
      "tokens_seen": 469696512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043315947843530593,
      "loss": 2.8967,
      "theoretical_loss": 3.941389752459484,
      "tokens_seen": 469762048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331494483450351,
      "loss": 2.9411,
      "theoretical_loss": 3.941329894743124,
      "tokens_seen": 469827584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043313941825476435,
      "loss": 2.9326,
      "theoretical_loss": 3.9412700477131972,
      "tokens_seen": 469893120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331293881644935,
      "loss": 2.8993,
      "theoretical_loss": 3.9412102113663057,
      "tokens_seen": 469958656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331193580742227,
      "loss": 2.8847,
      "theoretical_loss": 3.941150385699052,
      "tokens_seen": 470024192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043310932798395184,
      "loss": 2.7375,
      "theoretical_loss": 3.941090570708043,
      "tokens_seen": 470089728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043309929789368107,
      "loss": 3.0306,
      "theoretical_loss": 3.9410307663898845,
      "tokens_seen": 470155264
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 271247,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.501755475997925,
      "objective/train/theoretical_loss": 3.9409709727411846,
      "objective/train/tokens_used": 490680800,
      "theoretical_loss": 3.9409709727411846,
      "tokens_seen": 470220800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043308926780341025,
      "loss": 2.8002,
      "theoretical_loss": 3.9409709727411846,
      "tokens_seen": 470220800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043307923771313943,
      "loss": 2.8405,
      "theoretical_loss": 3.940911189758554,
      "tokens_seen": 470286336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330692076228686,
      "loss": 2.6275,
      "theoretical_loss": 3.940851417438604,
      "tokens_seen": 470351872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043305917753259785,
      "loss": 2.8477,
      "theoretical_loss": 3.9407916557779465,
      "tokens_seen": 470417408
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433049147442327,
      "loss": 2.8173,
      "theoretical_loss": 3.9407319047731972,
      "tokens_seen": 470482944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330391173520562,
      "loss": 2.7989,
      "theoretical_loss": 3.940672164420972,
      "tokens_seen": 470548480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043302908726178534,
      "loss": 2.9784,
      "theoretical_loss": 3.940612434717888,
      "tokens_seen": 470614016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330190571715146,
      "loss": 3.0426,
      "theoretical_loss": 3.9405527156605644,
      "tokens_seen": 470679552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043300902708124376,
      "loss": 2.8404,
      "theoretical_loss": 3.9404930072456223,
      "tokens_seen": 470745088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043299899699097294,
      "loss": 2.6866,
      "theoretical_loss": 3.9404333094696837,
      "tokens_seen": 470810624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329889669007021,
      "loss": 2.8504,
      "theoretical_loss": 3.940373622329372,
      "tokens_seen": 470876160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329789368104313,
      "loss": 2.8516,
      "theoretical_loss": 3.940313945821314,
      "tokens_seen": 470941696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329689067201605,
      "loss": 2.9968,
      "theoretical_loss": 3.9402542799421347,
      "tokens_seen": 471007232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329588766298897,
      "loss": 2.8706,
      "theoretical_loss": 3.9401946246884636,
      "tokens_seen": 471072768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043294884653961884,
      "loss": 2.8943,
      "theoretical_loss": 3.94013498005693,
      "tokens_seen": 471138304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329388164493481,
      "loss": 2.789,
      "theoretical_loss": 3.9400753460441655,
      "tokens_seen": 471203840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329287863590772,
      "loss": 2.6528,
      "theoretical_loss": 3.940015722646804,
      "tokens_seen": 471269376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043291875626880644,
      "loss": 2.8134,
      "theoretical_loss": 3.9399561098614786,
      "tokens_seen": 471334912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329087261785356,
      "loss": 3.0425,
      "theoretical_loss": 3.939896507684827,
      "tokens_seen": 471400448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328986960882648,
      "loss": 2.8502,
      "theoretical_loss": 3.939836916113485,
      "tokens_seen": 471465984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432888665997994,
      "loss": 2.7366,
      "theoretical_loss": 3.939777335144094,
      "tokens_seen": 471531520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328786359077232,
      "loss": 2.9817,
      "theoretical_loss": 3.9397177647732926,
      "tokens_seen": 471597056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043286860581745235,
      "loss": 2.9623,
      "theoretical_loss": 3.9396582049977242,
      "tokens_seen": 471662592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328585757271816,
      "loss": 2.7969,
      "theoretical_loss": 3.9395986558140317,
      "tokens_seen": 471728128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328485456369107,
      "loss": 2.6316,
      "theoretical_loss": 3.939539117218861,
      "tokens_seen": 471793664
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 272462,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.128762722015381,
      "objective/train/theoretical_loss": 3.9394795892088585,
      "objective/train/tokens_used": 492319200,
      "theoretical_loss": 3.9394795892088585,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043283851554663994,
      "loss": 2.9032,
      "theoretical_loss": 3.9394795892088585,
      "tokens_seen": 471859200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328284854563691,
      "loss": 2.7612,
      "theoretical_loss": 3.939420071780673,
      "tokens_seen": 471924736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328184553660983,
      "loss": 2.8278,
      "theoretical_loss": 3.939360564930954,
      "tokens_seen": 471990272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328084252758275,
      "loss": 2.7656,
      "theoretical_loss": 3.939301068656353,
      "tokens_seen": 472055808
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043279839518555667,
      "loss": 2.8529,
      "theoretical_loss": 3.939241582953523,
      "tokens_seen": 472121344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043278836509528585,
      "loss": 2.8641,
      "theoretical_loss": 3.9391821078191183,
      "tokens_seen": 472186880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327783350050151,
      "loss": 2.9301,
      "theoretical_loss": 3.939122643249795,
      "tokens_seen": 472252416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327683049147442,
      "loss": 2.794,
      "theoretical_loss": 3.9390631892422103,
      "tokens_seen": 472317952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043275827482447345,
      "loss": 2.9563,
      "theoretical_loss": 3.9390037457930234,
      "tokens_seen": 472383488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043274824473420263,
      "loss": 2.6417,
      "theoretical_loss": 3.9389443128988946,
      "tokens_seen": 472449024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327382146439318,
      "loss": 2.9363,
      "theoretical_loss": 3.938884890556486,
      "tokens_seen": 472514560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432728184553661,
      "loss": 2.9786,
      "theoretical_loss": 3.9388254787624617,
      "tokens_seen": 472580096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043271815446339017,
      "loss": 2.8273,
      "theoretical_loss": 3.938766077513486,
      "tokens_seen": 472645632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043270812437311935,
      "loss": 2.8662,
      "theoretical_loss": 3.938706686806226,
      "tokens_seen": 472711168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326980942828486,
      "loss": 2.9246,
      "theoretical_loss": 3.9386473066373497,
      "tokens_seen": 472776704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326880641925777,
      "loss": 2.9862,
      "theoretical_loss": 3.9385879370035264,
      "tokens_seen": 472842240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043267803410230695,
      "loss": 2.6308,
      "theoretical_loss": 3.938528577901427,
      "tokens_seen": 472907776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326680040120361,
      "loss": 2.9238,
      "theoretical_loss": 3.9384692293277253,
      "tokens_seen": 472973312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326579739217653,
      "loss": 2.7366,
      "theoretical_loss": 3.938409891279094,
      "tokens_seen": 473038848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326479438314945,
      "loss": 2.6296,
      "theoretical_loss": 3.93835056375221,
      "tokens_seen": 473104384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326379137412237,
      "loss": 2.8136,
      "theoretical_loss": 3.9382912467437503,
      "tokens_seen": 473169920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043262788365095286,
      "loss": 2.6962,
      "theoretical_loss": 3.938231940250393,
      "tokens_seen": 473235456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043261785356068204,
      "loss": 2.8113,
      "theoretical_loss": 3.9381726442688185,
      "tokens_seen": 473300992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326078234704112,
      "loss": 2.842,
      "theoretical_loss": 3.938113358795709,
      "tokens_seen": 473366528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043259779338014045,
      "loss": 2.8583,
      "theoretical_loss": 3.9380540838277467,
      "tokens_seen": 473432064
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 273042,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9048469066619873,
      "objective/train/theoretical_loss": 3.9379948193616174,
      "objective/train/tokens_used": 493957600,
      "theoretical_loss": 3.9379948193616174,
      "tokens_seen": 473497600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325877632898696,
      "loss": 2.6338,
      "theoretical_loss": 3.9379948193616174,
      "tokens_seen": 473497600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325777331995988,
      "loss": 2.8322,
      "theoretical_loss": 3.9379355653940062,
      "tokens_seen": 473563136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432567703109328,
      "loss": 2.8898,
      "theoretical_loss": 3.9378763219216015,
      "tokens_seen": 473628672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325576730190572,
      "loss": 2.6278,
      "theoretical_loss": 3.937817088941093,
      "tokens_seen": 473694208
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043254764292878636,
      "loss": 2.7849,
      "theoretical_loss": 3.9377578664491706,
      "tokens_seen": 473759744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043253761283851554,
      "loss": 2.7501,
      "theoretical_loss": 3.937698654442527,
      "tokens_seen": 473825280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325275827482447,
      "loss": 2.8944,
      "theoretical_loss": 3.937639452917856,
      "tokens_seen": 473890816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043251755265797396,
      "loss": 3.0332,
      "theoretical_loss": 3.9375802618718523,
      "tokens_seen": 473956352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325075225677031,
      "loss": 2.9932,
      "theoretical_loss": 3.9375210813012127,
      "tokens_seen": 474021888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324974924774323,
      "loss": 2.9509,
      "theoretical_loss": 3.937461911202636,
      "tokens_seen": 474087424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043248746238716145,
      "loss": 2.8843,
      "theoretical_loss": 3.9374027515728214,
      "tokens_seen": 474152960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324774322968907,
      "loss": 3.1177,
      "theoretical_loss": 3.93734360240847,
      "tokens_seen": 474218496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043246740220661986,
      "loss": 2.8071,
      "theoretical_loss": 3.937284463706286,
      "tokens_seen": 474284032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043245737211634904,
      "loss": 2.8188,
      "theoretical_loss": 3.937225335462971,
      "tokens_seen": 474349568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324473420260782,
      "loss": 2.8953,
      "theoretical_loss": 3.937166217675233,
      "tokens_seen": 474415104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324373119358074,
      "loss": 2.719,
      "theoretical_loss": 3.9371071103397783,
      "tokens_seen": 474480640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043242728184553664,
      "loss": 2.6903,
      "theoretical_loss": 3.9370480134533157,
      "tokens_seen": 474546176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324172517552658,
      "loss": 2.7762,
      "theoretical_loss": 3.9369889270125547,
      "tokens_seen": 474611712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432407221664995,
      "loss": 2.5037,
      "theoretical_loss": 3.9369298510142077,
      "tokens_seen": 474677248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323971915747242,
      "loss": 3.0455,
      "theoretical_loss": 3.936870785454988,
      "tokens_seen": 474742784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323871614844534,
      "loss": 2.8792,
      "theoretical_loss": 3.93681173033161,
      "tokens_seen": 474808320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043237713139418255,
      "loss": 2.7244,
      "theoretical_loss": 3.9367526856407897,
      "tokens_seen": 474873856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323671013039118,
      "loss": 2.9051,
      "theoretical_loss": 3.9366936513792448,
      "tokens_seen": 474939392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323570712136409,
      "loss": 2.7048,
      "theoretical_loss": 3.936634627543694,
      "tokens_seen": 475004928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043234704112337014,
      "loss": 2.9354,
      "theoretical_loss": 3.9365756141308594,
      "tokens_seen": 475070464
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 273516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.271859645843506,
      "objective/train/theoretical_loss": 3.9365166111374608,
      "objective/train/tokens_used": 495596000,
      "theoretical_loss": 3.9365166111374608,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323370110330993,
      "loss": 2.8993,
      "theoretical_loss": 3.9365166111374608,
      "tokens_seen": 475136000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323269809428285,
      "loss": 2.8789,
      "theoretical_loss": 3.9364576185602234,
      "tokens_seen": 475201536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323169508525577,
      "loss": 2.9238,
      "theoretical_loss": 3.9363986363958725,
      "tokens_seen": 475267072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043230692076228687,
      "loss": 2.7607,
      "theoretical_loss": 3.936339664641133,
      "tokens_seen": 475332608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043229689067201605,
      "loss": 2.8522,
      "theoretical_loss": 3.9362807032927343,
      "tokens_seen": 475398144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322868605817453,
      "loss": 2.9343,
      "theoretical_loss": 3.9362217523474046,
      "tokens_seen": 475463680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322768304914744,
      "loss": 3.034,
      "theoretical_loss": 3.9361628118018768,
      "tokens_seen": 475529216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043226680040120365,
      "loss": 2.9514,
      "theoretical_loss": 3.936103881652881,
      "tokens_seen": 475594752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043225677031093283,
      "loss": 2.8471,
      "theoretical_loss": 3.936044961897153,
      "tokens_seen": 475660288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432246740220662,
      "loss": 2.8487,
      "theoretical_loss": 3.9359860525314274,
      "tokens_seen": 475725824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322367101303912,
      "loss": 2.7933,
      "theoretical_loss": 3.9359271535524414,
      "tokens_seen": 475791360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043222668004012037,
      "loss": 2.7604,
      "theoretical_loss": 3.935868264956932,
      "tokens_seen": 475856896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043221664994984955,
      "loss": 2.8025,
      "theoretical_loss": 3.9358093867416413,
      "tokens_seen": 475922432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322066198595788,
      "loss": 2.9206,
      "theoretical_loss": 3.935750518903309,
      "tokens_seen": 475987968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321965897693079,
      "loss": 3.1348,
      "theoretical_loss": 3.9356916614386783,
      "tokens_seen": 476053504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043218655967903715,
      "loss": 2.8289,
      "theoretical_loss": 3.9356328143444923,
      "tokens_seen": 476119040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321765295887663,
      "loss": 2.77,
      "theoretical_loss": 3.9355739776174987,
      "tokens_seen": 476184576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321664994984955,
      "loss": 2.9714,
      "theoretical_loss": 3.935515151254444,
      "tokens_seen": 476250112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321564694082247,
      "loss": 3.0004,
      "theoretical_loss": 3.935456335252076,
      "tokens_seen": 476315648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321464393179539,
      "loss": 2.7851,
      "theoretical_loss": 3.9353975296071453,
      "tokens_seen": 476381184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043213640922768306,
      "loss": 2.9986,
      "theoretical_loss": 3.9353387343164035,
      "tokens_seen": 476446720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043212637913741224,
      "loss": 2.8154,
      "theoretical_loss": 3.9352799493766035,
      "tokens_seen": 476512256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321163490471414,
      "loss": 2.9255,
      "theoretical_loss": 3.9352211747845,
      "tokens_seen": 476577792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043210631895687065,
      "loss": 2.792,
      "theoretical_loss": 3.935162410536849,
      "tokens_seen": 476643328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320962888665998,
      "loss": 2.5768,
      "theoretical_loss": 3.9351036566304076,
      "tokens_seen": 476708864
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 274597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.601290225982666,
      "objective/train/theoretical_loss": 3.935044913061936,
      "objective/train/tokens_used": 497234400,
      "theoretical_loss": 3.935044913061936,
      "tokens_seen": 476774400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432086258776329,
      "loss": 2.8793,
      "theoretical_loss": 3.935044913061936,
      "tokens_seen": 476774400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320762286860582,
      "loss": 2.7909,
      "theoretical_loss": 3.9349861798281918,
      "tokens_seen": 476839936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320661985957874,
      "loss": 2.4871,
      "theoretical_loss": 3.934927456925939,
      "tokens_seen": 476905472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043205616850551656,
      "loss": 2.6598,
      "theoretical_loss": 3.934868744351941,
      "tokens_seen": 476971008
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043204613841524574,
      "loss": 2.6697,
      "theoretical_loss": 3.934810042102961,
      "tokens_seen": 477036544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320361083249749,
      "loss": 2.8944,
      "theoretical_loss": 3.934751350175766,
      "tokens_seen": 477102080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043202607823470416,
      "loss": 2.8022,
      "theoretical_loss": 3.9346926685671235,
      "tokens_seen": 477167616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320160481444333,
      "loss": 2.9025,
      "theoretical_loss": 3.934633997273803,
      "tokens_seen": 477233152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004320060180541625,
      "loss": 2.8173,
      "theoretical_loss": 3.934575336292575,
      "tokens_seen": 477298688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043199598796389165,
      "loss": 2.9114,
      "theoretical_loss": 3.9345166856202107,
      "tokens_seen": 477364224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319859578736209,
      "loss": 2.9045,
      "theoretical_loss": 3.9344580452534847,
      "tokens_seen": 477429760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043197592778335006,
      "loss": 2.8471,
      "theoretical_loss": 3.934399415189171,
      "tokens_seen": 477495296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043196589769307924,
      "loss": 2.6342,
      "theoretical_loss": 3.9343407954240464,
      "tokens_seen": 477560832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319558676028084,
      "loss": 2.8732,
      "theoretical_loss": 3.934282185954889,
      "tokens_seen": 477626368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319458375125376,
      "loss": 2.5875,
      "theoretical_loss": 3.934223586778477,
      "tokens_seen": 477691904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319358074222668,
      "loss": 2.9511,
      "theoretical_loss": 3.9341649978915925,
      "tokens_seen": 477757440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000431925777331996,
      "loss": 2.8133,
      "theoretical_loss": 3.9341064192910165,
      "tokens_seen": 477822976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043191574724172515,
      "loss": 2.9208,
      "theoretical_loss": 3.9340478509735326,
      "tokens_seen": 477888512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319057171514544,
      "loss": 2.8644,
      "theoretical_loss": 3.9339892929359266,
      "tokens_seen": 477954048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043189568706118357,
      "loss": 2.9012,
      "theoretical_loss": 3.933930745174985,
      "tokens_seen": 478019584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043188565697091275,
      "loss": 2.5838,
      "theoretical_loss": 3.9338722076874957,
      "tokens_seen": 478085120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043187562688064193,
      "loss": 2.7064,
      "theoretical_loss": 3.933813680470247,
      "tokens_seen": 478150656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318655967903711,
      "loss": 2.9544,
      "theoretical_loss": 3.9337551635200314,
      "tokens_seen": 478216192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318555667001003,
      "loss": 2.7427,
      "theoretical_loss": 3.93369665683364,
      "tokens_seen": 478281728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318455366098295,
      "loss": 2.8325,
      "theoretical_loss": 3.9336381604078667,
      "tokens_seen": 478347264
    },
    {
      "epoch": 0.14,
      "objective/train/docs_used": 275931,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.969980239868164,
      "objective/train/theoretical_loss": 3.933579674239507,
      "objective/train/tokens_used": 498872800,
      "theoretical_loss": 3.933579674239507,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043183550651955865,
      "loss": 2.9991,
      "theoretical_loss": 3.933579674239507,
      "tokens_seen": 478412800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318254764292879,
      "loss": 2.8388,
      "theoretical_loss": 3.933521198325357,
      "tokens_seen": 478478336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431815446339017,
      "loss": 2.872,
      "theoretical_loss": 3.933462732662215,
      "tokens_seen": 478543872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043180541624874625,
      "loss": 2.9391,
      "theoretical_loss": 3.933404277246881,
      "tokens_seen": 478609408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043179538615847543,
      "loss": 2.7832,
      "theoretical_loss": 3.9333458320761547,
      "tokens_seen": 478674944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317853560682046,
      "loss": 2.8224,
      "theoretical_loss": 3.93328739714684,
      "tokens_seen": 478740480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317753259779338,
      "loss": 2.6799,
      "theoretical_loss": 3.9332289724557397,
      "tokens_seen": 478806016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043176529588766303,
      "loss": 2.6965,
      "theoretical_loss": 3.9331705579996585,
      "tokens_seen": 478871552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043175526579739215,
      "loss": 2.9734,
      "theoretical_loss": 3.9331121537754044,
      "tokens_seen": 478937088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317452357071214,
      "loss": 2.726,
      "theoretical_loss": 3.933053759779784,
      "tokens_seen": 479002624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317352056168505,
      "loss": 2.5974,
      "theoretical_loss": 3.9329953760096084,
      "tokens_seen": 479068160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043172517552657975,
      "loss": 2.7127,
      "theoretical_loss": 3.9329370024616876,
      "tokens_seen": 479133696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043171514543630893,
      "loss": 2.7597,
      "theoretical_loss": 3.9328786391328343,
      "tokens_seen": 479199232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317051153460381,
      "loss": 2.9107,
      "theoretical_loss": 3.9328202860198616,
      "tokens_seen": 479264768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316950852557673,
      "loss": 2.9456,
      "theoretical_loss": 3.9327619431195853,
      "tokens_seen": 479330304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316850551654965,
      "loss": 2.769,
      "theoretical_loss": 3.9327036104288227,
      "tokens_seen": 479395840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316750250752257,
      "loss": 3.0487,
      "theoretical_loss": 3.932645287944391,
      "tokens_seen": 479461376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316649949849549,
      "loss": 3.0331,
      "theoretical_loss": 3.932586975663109,
      "tokens_seen": 479526912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316549648946841,
      "loss": 3.109,
      "theoretical_loss": 3.9325286735817997,
      "tokens_seen": 479592448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043164493480441326,
      "loss": 2.7279,
      "theoretical_loss": 3.9324703816972844,
      "tokens_seen": 479657984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043163490471414244,
      "loss": 2.7961,
      "theoretical_loss": 3.9324121000063865,
      "tokens_seen": 479723520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316248746238716,
      "loss": 2.6693,
      "theoretical_loss": 3.9323538285059314,
      "tokens_seen": 479789056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043161484453360085,
      "loss": 2.9624,
      "theoretical_loss": 3.932295567192747,
      "tokens_seen": 479854592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043160481444333,
      "loss": 2.8733,
      "theoretical_loss": 3.932237316063659,
      "tokens_seen": 479920128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315947843530592,
      "loss": 2.8734,
      "theoretical_loss": 3.9321790751154984,
      "tokens_seen": 479985664
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 276348,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0607144832611084,
      "objective/train/theoretical_loss": 3.9321208443450963,
      "objective/train/tokens_used": 500511200,
      "theoretical_loss": 3.9321208443450963,
      "tokens_seen": 480051200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315847542627884,
      "loss": 2.938,
      "theoretical_loss": 3.9321208443450963,
      "tokens_seen": 480051200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315747241725176,
      "loss": 2.5545,
      "theoretical_loss": 3.9320626237492844,
      "tokens_seen": 480116736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043156469408224676,
      "loss": 2.6545,
      "theoretical_loss": 3.9320044133248966,
      "tokens_seen": 480182272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043155466399197594,
      "loss": 2.8061,
      "theoretical_loss": 3.931946213068768,
      "tokens_seen": 480247808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315446339017051,
      "loss": 2.6674,
      "theoretical_loss": 3.931888022977735,
      "tokens_seen": 480313344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043153460381143436,
      "loss": 2.9156,
      "theoretical_loss": 3.9318298430486367,
      "tokens_seen": 480378880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315245737211635,
      "loss": 2.8293,
      "theoretical_loss": 3.931771673278311,
      "tokens_seen": 480444416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315145436308927,
      "loss": 3.0808,
      "theoretical_loss": 3.9317135136635994,
      "tokens_seen": 480509952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043150451354062185,
      "loss": 2.7806,
      "theoretical_loss": 3.931655364201344,
      "tokens_seen": 480575488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314944834503511,
      "loss": 2.9279,
      "theoretical_loss": 3.931597224888389,
      "tokens_seen": 480641024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043148445336008026,
      "loss": 3.0468,
      "theoretical_loss": 3.931539095721578,
      "tokens_seen": 480706560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043147442326980944,
      "loss": 3.0113,
      "theoretical_loss": 3.9314809766977596,
      "tokens_seen": 480772096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314643931795386,
      "loss": 2.9067,
      "theoretical_loss": 3.93142286781378,
      "tokens_seen": 480837632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314543630892678,
      "loss": 2.8236,
      "theoretical_loss": 3.931364769066489,
      "tokens_seen": 480903168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431444332998997,
      "loss": 3.1,
      "theoretical_loss": 3.931306680452737,
      "tokens_seen": 480968704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314343029087262,
      "loss": 3.1625,
      "theoretical_loss": 3.9312486019693766,
      "tokens_seen": 481034240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043142427281845535,
      "loss": 2.8898,
      "theoretical_loss": 3.931190533613261,
      "tokens_seen": 481099776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314142427281846,
      "loss": 2.9089,
      "theoretical_loss": 3.9311324753812453,
      "tokens_seen": 481165312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043140421263791377,
      "loss": 2.8168,
      "theoretical_loss": 3.931074427270186,
      "tokens_seen": 481230848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043139418254764295,
      "loss": 2.9268,
      "theoretical_loss": 3.93101638927694,
      "tokens_seen": 481296384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043138415245737213,
      "loss": 2.8559,
      "theoretical_loss": 3.9309583613983676,
      "tokens_seen": 481361920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313741223671013,
      "loss": 2.8534,
      "theoretical_loss": 3.930900343631328,
      "tokens_seen": 481427456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313640922768305,
      "loss": 2.7964,
      "theoretical_loss": 3.9308423359726845,
      "tokens_seen": 481492992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313540621865597,
      "loss": 2.8632,
      "theoretical_loss": 3.930784338419299,
      "tokens_seen": 481558528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043134403209628885,
      "loss": 2.6951,
      "theoretical_loss": 3.9307263509680372,
      "tokens_seen": 481624064
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 277290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.784728765487671,
      "objective/train/theoretical_loss": 3.930668373615765,
      "objective/train/tokens_used": 502149600,
      "theoretical_loss": 3.930668373615765,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313340020060181,
      "loss": 2.8966,
      "theoretical_loss": 3.930668373615765,
      "tokens_seen": 481689600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313239719157472,
      "loss": 3.0139,
      "theoretical_loss": 3.9306104063593503,
      "tokens_seen": 481755136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043131394182547645,
      "loss": 2.8475,
      "theoretical_loss": 3.9305524491956616,
      "tokens_seen": 481820672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043130391173520563,
      "loss": 2.8799,
      "theoretical_loss": 3.9304945021215687,
      "tokens_seen": 481886208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312938816449348,
      "loss": 2.8447,
      "theoretical_loss": 3.9304365651339443,
      "tokens_seen": 481951744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431283851554664,
      "loss": 3.0417,
      "theoretical_loss": 3.9303786382296604,
      "tokens_seen": 482017280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043127382146439323,
      "loss": 2.7033,
      "theoretical_loss": 3.930320721405593,
      "tokens_seen": 482082816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043126379137412236,
      "loss": 2.8143,
      "theoretical_loss": 3.9302628146586165,
      "tokens_seen": 482148352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312537612838516,
      "loss": 3.1353,
      "theoretical_loss": 3.930204917985609,
      "tokens_seen": 482213888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312437311935807,
      "loss": 2.8771,
      "theoretical_loss": 3.9301470313834495,
      "tokens_seen": 482279424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043123370110330995,
      "loss": 3.1197,
      "theoretical_loss": 3.9300891548490178,
      "tokens_seen": 482344960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043122367101303913,
      "loss": 2.7898,
      "theoretical_loss": 3.9300312883791944,
      "tokens_seen": 482410496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312136409227683,
      "loss": 2.7825,
      "theoretical_loss": 3.929973431970863,
      "tokens_seen": 482476032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312036108324975,
      "loss": 3.0488,
      "theoretical_loss": 3.9299155856209085,
      "tokens_seen": 482541568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311935807422267,
      "loss": 2.6776,
      "theoretical_loss": 3.9298577493262155,
      "tokens_seen": 482607104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043118355065195586,
      "loss": 2.9107,
      "theoretical_loss": 3.9297999230836713,
      "tokens_seen": 482672640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311735205616851,
      "loss": 2.9791,
      "theoretical_loss": 3.9297421068901643,
      "tokens_seen": 482738176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311634904714142,
      "loss": 2.7728,
      "theoretical_loss": 3.9296843007425846,
      "tokens_seen": 482803712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043115346038114346,
      "loss": 2.9737,
      "theoretical_loss": 3.9296265046378225,
      "tokens_seen": 482869248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311434302908726,
      "loss": 2.8517,
      "theoretical_loss": 3.9295687185727717,
      "tokens_seen": 482934784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311334002006018,
      "loss": 2.7383,
      "theoretical_loss": 3.929510942544325,
      "tokens_seen": 483000320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431123370110331,
      "loss": 3.0246,
      "theoretical_loss": 3.9294531765493788,
      "tokens_seen": 483065856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311133400200602,
      "loss": 2.8883,
      "theoretical_loss": 3.9293954205848296,
      "tokens_seen": 483131392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043110330992978936,
      "loss": 2.4493,
      "theoretical_loss": 3.9293376746475746,
      "tokens_seen": 483196928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310932798395186,
      "loss": 3.0786,
      "theoretical_loss": 3.9292799387345143,
      "tokens_seen": 483262464
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 277804,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7938942909240723,
      "objective/train/theoretical_loss": 3.929222212842549,
      "objective/train/tokens_used": 503788000,
      "theoretical_loss": 3.929222212842549,
      "tokens_seen": 483328000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310832497492477,
      "loss": 3.0089,
      "theoretical_loss": 3.929222212842549,
      "tokens_seen": 483328000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043107321965897696,
      "loss": 2.86,
      "theoretical_loss": 3.929164496968581,
      "tokens_seen": 483393536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310631895687061,
      "loss": 2.8599,
      "theoretical_loss": 3.929106791109514,
      "tokens_seen": 483459072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310531594784353,
      "loss": 3.1346,
      "theoretical_loss": 3.9290490952622523,
      "tokens_seen": 483524608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310431293881645,
      "loss": 2.5788,
      "theoretical_loss": 3.9289914094237037,
      "tokens_seen": 483590144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310330992978937,
      "loss": 3.0451,
      "theoretical_loss": 3.9289337335907746,
      "tokens_seen": 483655680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043102306920762286,
      "loss": 2.8503,
      "theoretical_loss": 3.928876067760375,
      "tokens_seen": 483721216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043101303911735205,
      "loss": 2.7226,
      "theoretical_loss": 3.9288184119294147,
      "tokens_seen": 483786752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310030090270812,
      "loss": 2.841,
      "theoretical_loss": 3.928760766094806,
      "tokens_seen": 483852288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043099297893681046,
      "loss": 2.9864,
      "theoretical_loss": 3.928703130253462,
      "tokens_seen": 483917824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309829488465396,
      "loss": 2.8682,
      "theoretical_loss": 3.9286455044022968,
      "tokens_seen": 483983360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309729187562688,
      "loss": 2.7043,
      "theoretical_loss": 3.9285878885382273,
      "tokens_seen": 484048896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043096288866599795,
      "loss": 2.4423,
      "theoretical_loss": 3.9285302826581696,
      "tokens_seen": 484114432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309528585757272,
      "loss": 2.6732,
      "theoretical_loss": 3.9284726867590436,
      "tokens_seen": 484179968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043094282848545637,
      "loss": 2.9397,
      "theoretical_loss": 3.9284151008377686,
      "tokens_seen": 484245504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043093279839518555,
      "loss": 2.6799,
      "theoretical_loss": 3.9283575248912666,
      "tokens_seen": 484311040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309227683049148,
      "loss": 2.4629,
      "theoretical_loss": 3.92829995891646,
      "tokens_seen": 484376576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043091273821464397,
      "loss": 2.809,
      "theoretical_loss": 3.928242402910273,
      "tokens_seen": 484442112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043090270812437315,
      "loss": 2.5356,
      "theoretical_loss": 3.928184856869632,
      "tokens_seen": 484507648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043089267803410233,
      "loss": 2.8043,
      "theoretical_loss": 3.9281273207914618,
      "tokens_seen": 484573184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308826479438315,
      "loss": 2.8214,
      "theoretical_loss": 3.928069794672693,
      "tokens_seen": 484638720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308726178535607,
      "loss": 2.8554,
      "theoretical_loss": 3.928012278510254,
      "tokens_seen": 484704256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308625877632899,
      "loss": 2.7812,
      "theoretical_loss": 3.9279547723010753,
      "tokens_seen": 484769792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043085255767301905,
      "loss": 3.1759,
      "theoretical_loss": 3.9278972760420903,
      "tokens_seen": 484835328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308425275827483,
      "loss": 2.7076,
      "theoretical_loss": 3.927839789730232,
      "tokens_seen": 484900864
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 279062,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1616718769073486,
      "objective/train/theoretical_loss": 3.9277823133624366,
      "objective/train/tokens_used": 505426400,
      "theoretical_loss": 3.9277823133624366,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308324974924774,
      "loss": 2.8253,
      "theoretical_loss": 3.9277823133624366,
      "tokens_seen": 484966400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043082246740220665,
      "loss": 2.749,
      "theoretical_loss": 3.927724846935639,
      "tokens_seen": 485031936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043081243731193583,
      "loss": 2.7862,
      "theoretical_loss": 3.9276673904467776,
      "tokens_seen": 485097472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430802407221665,
      "loss": 2.8595,
      "theoretical_loss": 3.927609943892792,
      "tokens_seen": 485163008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307923771313942,
      "loss": 2.7832,
      "theoretical_loss": 3.9275525072706223,
      "tokens_seen": 485228544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043078234704112343,
      "loss": 2.7835,
      "theoretical_loss": 3.9274950805772098,
      "tokens_seen": 485294080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043077231695085256,
      "loss": 2.7645,
      "theoretical_loss": 3.9274376638094983,
      "tokens_seen": 485359616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307622868605818,
      "loss": 2.8211,
      "theoretical_loss": 3.9273802569644323,
      "tokens_seen": 485425152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307522567703109,
      "loss": 2.6924,
      "theoretical_loss": 3.9273228600389585,
      "tokens_seen": 485490688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043074222668004015,
      "loss": 2.7889,
      "theoretical_loss": 3.9272654730300225,
      "tokens_seen": 485556224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043073219658976933,
      "loss": 3.0397,
      "theoretical_loss": 3.9272080959345734,
      "tokens_seen": 485621760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307221664994985,
      "loss": 2.699,
      "theoretical_loss": 3.927150728749562,
      "tokens_seen": 485687296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307121364092277,
      "loss": 2.729,
      "theoretical_loss": 3.9270933714719396,
      "tokens_seen": 485752832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307021063189569,
      "loss": 2.7437,
      "theoretical_loss": 3.9270360240986575,
      "tokens_seen": 485818368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043069207622868606,
      "loss": 2.8264,
      "theoretical_loss": 3.926978686626671,
      "tokens_seen": 485883904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306820461384153,
      "loss": 2.9605,
      "theoretical_loss": 3.926921359052935,
      "tokens_seen": 485949440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306720160481444,
      "loss": 3.0128,
      "theoretical_loss": 3.9268640413744063,
      "tokens_seen": 486014976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043066198595787366,
      "loss": 2.7019,
      "theoretical_loss": 3.9268067335880428,
      "tokens_seen": 486080512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306519558676028,
      "loss": 2.9512,
      "theoretical_loss": 3.9267494356908044,
      "tokens_seen": 486146048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430641925777332,
      "loss": 2.9346,
      "theoretical_loss": 3.926692147679651,
      "tokens_seen": 486211584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306318956870612,
      "loss": 2.6555,
      "theoretical_loss": 3.926634869551545,
      "tokens_seen": 486277120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306218655967904,
      "loss": 2.9892,
      "theoretical_loss": 3.9265776013034497,
      "tokens_seen": 486342656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043061183550651956,
      "loss": 2.7261,
      "theoretical_loss": 3.9265203429323305,
      "tokens_seen": 486408192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306018054162488,
      "loss": 2.6338,
      "theoretical_loss": 3.926463094435153,
      "tokens_seen": 486473728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305917753259779,
      "loss": 2.7808,
      "theoretical_loss": 3.9264058558088846,
      "tokens_seen": 486539264
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 279686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.470508337020874,
      "objective/train/theoretical_loss": 3.926348627050494,
      "objective/train/tokens_used": 507064800,
      "theoretical_loss": 3.926348627050494,
      "tokens_seen": 486604800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043058174523570716,
      "loss": 2.8973,
      "theoretical_loss": 3.926348627050494,
      "tokens_seen": 486604800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305717151454363,
      "loss": 2.9057,
      "theoretical_loss": 3.9262914081569518,
      "tokens_seen": 486670336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305616850551655,
      "loss": 3.0043,
      "theoretical_loss": 3.9262341991252296,
      "tokens_seen": 486735872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305516549648947,
      "loss": 2.6892,
      "theoretical_loss": 3.9261769999522995,
      "tokens_seen": 486801408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305416248746239,
      "loss": 2.7105,
      "theoretical_loss": 3.926119810635136,
      "tokens_seen": 486866944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043053159478435306,
      "loss": 3.0511,
      "theoretical_loss": 3.9260626311707143,
      "tokens_seen": 486932480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043052156469408225,
      "loss": 2.6944,
      "theoretical_loss": 3.926005461556011,
      "tokens_seen": 486998016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305115346038114,
      "loss": 2.5711,
      "theoretical_loss": 3.9259483017880052,
      "tokens_seen": 487063552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043050150451354066,
      "loss": 2.9071,
      "theoretical_loss": 3.9258911518636763,
      "tokens_seen": 487129088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304914744232698,
      "loss": 2.9919,
      "theoretical_loss": 3.9258340117800037,
      "tokens_seen": 487194624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430481444332999,
      "loss": 2.6231,
      "theoretical_loss": 3.925776881533971,
      "tokens_seen": 487260160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043047141424272815,
      "loss": 3.1136,
      "theoretical_loss": 3.9257197611225614,
      "tokens_seen": 487325696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304613841524574,
      "loss": 2.8313,
      "theoretical_loss": 3.925662650542759,
      "tokens_seen": 487391232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043045135406218657,
      "loss": 2.8077,
      "theoretical_loss": 3.92560554979155,
      "tokens_seen": 487456768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043044132397191575,
      "loss": 2.933,
      "theoretical_loss": 3.925548458865923,
      "tokens_seen": 487522304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043043129388164493,
      "loss": 2.7352,
      "theoretical_loss": 3.925491377762866,
      "tokens_seen": 487587840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043042126379137417,
      "loss": 2.9462,
      "theoretical_loss": 3.925434306479369,
      "tokens_seen": 487653376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304112337011033,
      "loss": 2.8203,
      "theoretical_loss": 3.925377245012424,
      "tokens_seen": 487718912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043040120361083253,
      "loss": 2.6397,
      "theoretical_loss": 3.9253201933590227,
      "tokens_seen": 487784448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043039117352056165,
      "loss": 2.8775,
      "theoretical_loss": 3.92526315151616,
      "tokens_seen": 487849984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303811434302909,
      "loss": 2.7412,
      "theoretical_loss": 3.9252061194808316,
      "tokens_seen": 487915520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043037111334002007,
      "loss": 2.9406,
      "theoretical_loss": 3.925149097250033,
      "tokens_seen": 487981056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043036108324974925,
      "loss": 2.9284,
      "theoretical_loss": 3.925092084820764,
      "tokens_seen": 488046592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043035105315947843,
      "loss": 2.7134,
      "theoretical_loss": 3.9250350821900226,
      "tokens_seen": 488112128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303410230692076,
      "loss": 2.7296,
      "theoretical_loss": 3.92497808935481,
      "tokens_seen": 488177664
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 280726,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.50590443611145,
      "objective/train/theoretical_loss": 3.9249211063121283,
      "objective/train/tokens_used": 508703200,
      "theoretical_loss": 3.9249211063121283,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303309929789368,
      "loss": 3.0183,
      "theoretical_loss": 3.9249211063121283,
      "tokens_seen": 488243200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043032096288866603,
      "loss": 2.6671,
      "theoretical_loss": 3.9248641330589806,
      "tokens_seen": 488308736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043031093279839516,
      "loss": 2.8775,
      "theoretical_loss": 3.924807169592372,
      "tokens_seen": 488374272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303009027081244,
      "loss": 2.9192,
      "theoretical_loss": 3.9247502159093086,
      "tokens_seen": 488439808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302908726178535,
      "loss": 2.6838,
      "theoretical_loss": 3.924693272006796,
      "tokens_seen": 488505344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043028084252758276,
      "loss": 2.9602,
      "theoretical_loss": 3.924636337881845,
      "tokens_seen": 488570880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043027081243731194,
      "loss": 2.9434,
      "theoretical_loss": 3.924579413531465,
      "tokens_seen": 488636416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302607823470411,
      "loss": 2.9598,
      "theoretical_loss": 3.9245224989526664,
      "tokens_seen": 488701952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302507522567703,
      "loss": 2.862,
      "theoretical_loss": 3.924465594142463,
      "tokens_seen": 488767488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043024072216649953,
      "loss": 2.9043,
      "theoretical_loss": 3.9244086990978673,
      "tokens_seen": 488833024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043023069207622866,
      "loss": 2.8912,
      "theoretical_loss": 3.924351813815895,
      "tokens_seen": 488898560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302206619859579,
      "loss": 2.9308,
      "theoretical_loss": 3.9242949382935635,
      "tokens_seen": 488964096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430210631895687,
      "loss": 2.9168,
      "theoretical_loss": 3.9242380725278894,
      "tokens_seen": 489029632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043020060180541626,
      "loss": 2.723,
      "theoretical_loss": 3.9241812165158922,
      "tokens_seen": 489095168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043019057171514544,
      "loss": 2.9341,
      "theoretical_loss": 3.9241243702545923,
      "tokens_seen": 489160704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301805416248746,
      "loss": 2.7376,
      "theoretical_loss": 3.9240675337410122,
      "tokens_seen": 489226240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043017051153460386,
      "loss": 2.9562,
      "theoretical_loss": 3.924010706972174,
      "tokens_seen": 489291776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430160481444333,
      "loss": 2.5616,
      "theoretical_loss": 3.9239538899451025,
      "tokens_seen": 489357312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301504513540622,
      "loss": 2.9733,
      "theoretical_loss": 3.923897082656823,
      "tokens_seen": 489422848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301404212637914,
      "loss": 2.7338,
      "theoretical_loss": 3.9238402851043634,
      "tokens_seen": 489488384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301303911735206,
      "loss": 2.8406,
      "theoretical_loss": 3.9237834972847505,
      "tokens_seen": 489553920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043012036108324976,
      "loss": 2.8607,
      "theoretical_loss": 3.923726719195016,
      "tokens_seen": 489619456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430110330992979,
      "loss": 2.7782,
      "theoretical_loss": 3.9236699508321884,
      "tokens_seen": 489684992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301003009027081,
      "loss": 2.9235,
      "theoretical_loss": 3.9236131921933017,
      "tokens_seen": 489750528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043009027081243736,
      "loss": 2.9848,
      "theoretical_loss": 3.923556443275388,
      "tokens_seen": 489816064
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 281543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0093557834625244,
      "objective/train/theoretical_loss": 3.923499704075483,
      "objective/train/tokens_used": 510341600,
      "theoretical_loss": 3.923499704075483,
      "tokens_seen": 489881600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300802407221665,
      "loss": 2.8803,
      "theoretical_loss": 3.923499704075483,
      "tokens_seen": 489881600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300702106318957,
      "loss": 2.945,
      "theoretical_loss": 3.923442974590623,
      "tokens_seen": 489947136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300601805416249,
      "loss": 2.7447,
      "theoretical_loss": 3.923386254817845,
      "tokens_seen": 490012672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300501504513541,
      "loss": 3.1141,
      "theoretical_loss": 3.9233295447541874,
      "tokens_seen": 490078208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043004012036108326,
      "loss": 2.761,
      "theoretical_loss": 3.92327284439669,
      "tokens_seen": 490143744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043003009027081245,
      "loss": 2.8582,
      "theoretical_loss": 3.9232161537423957,
      "tokens_seen": 490209280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043002006018054163,
      "loss": 2.824,
      "theoretical_loss": 3.9231594727883454,
      "tokens_seen": 490274816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043001003009027086,
      "loss": 2.6088,
      "theoretical_loss": 3.9231028015315834,
      "tokens_seen": 490340352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043,
      "loss": 2.857,
      "theoretical_loss": 3.9230461399691556,
      "tokens_seen": 490405888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299899699097292,
      "loss": 2.7562,
      "theoretical_loss": 3.922989488098107,
      "tokens_seen": 490471424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042997993981945835,
      "loss": 2.9854,
      "theoretical_loss": 3.9229328459154873,
      "tokens_seen": 490536960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299699097291876,
      "loss": 2.8507,
      "theoretical_loss": 3.9228762134183444,
      "tokens_seen": 490602496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042995987963891677,
      "loss": 3.0315,
      "theoretical_loss": 3.922819590603729,
      "tokens_seen": 490668032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042994984954864595,
      "loss": 2.8923,
      "theoretical_loss": 3.922762977468692,
      "tokens_seen": 490733568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042993981945837513,
      "loss": 2.7771,
      "theoretical_loss": 3.9227063740102874,
      "tokens_seen": 490799104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042992978936810437,
      "loss": 3.0867,
      "theoretical_loss": 3.9226497802255684,
      "tokens_seen": 490864640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299197592778335,
      "loss": 2.6551,
      "theoretical_loss": 3.922593196111592,
      "tokens_seen": 490930176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042990972918756273,
      "loss": 2.952,
      "theoretical_loss": 3.9225366216654134,
      "tokens_seen": 490995712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042989969909729185,
      "loss": 2.7502,
      "theoretical_loss": 3.9224800568840914,
      "tokens_seen": 491061248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298896690070211,
      "loss": 2.7318,
      "theoretical_loss": 3.922423501764686,
      "tokens_seen": 491126784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042987963891675027,
      "loss": 3.0758,
      "theoretical_loss": 3.9223669563042565,
      "tokens_seen": 491192320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042986960882647945,
      "loss": 2.9801,
      "theoretical_loss": 3.9223104204998664,
      "tokens_seen": 491257856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042985957873620863,
      "loss": 2.9669,
      "theoretical_loss": 3.922253894348578,
      "tokens_seen": 491323392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298495486459378,
      "loss": 2.9609,
      "theoretical_loss": 3.922197377847456,
      "tokens_seen": 491388928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429839518555667,
      "loss": 2.9549,
      "theoretical_loss": 3.922140870993566,
      "tokens_seen": 491454464
    },
    {
      "debugging/Self-BLEU-5": 0.364302224771738,
      "debugging/distinct-1-grams": 0.7813905547226386,
      "debugging/distinct-2-grams": 0.9746492802208804,
      "debugging/entropy-1-grams": 5.062092644878832,
      "debugging/entropy-2-grams": 5.611156931113771,
      "debugging/length": 539.0,
      "debugging/num_segments": 4,
      "epoch": 0.15,
      "objective/train/docs_used": 282054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8211379051208496,
      "objective/train/theoretical_loss": 3.9220843737839752,
      "objective/train/tokens_used": 511980000,
      "theoretical_loss": 3.9220843737839752,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042982948846539623,
      "loss": 2.85,
      "theoretical_loss": 3.9220843737839752,
      "tokens_seen": 491520000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042981945837512536,
      "loss": 2.6656,
      "theoretical_loss": 3.9220278862157523,
      "tokens_seen": 491585536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298094282848546,
      "loss": 2.8264,
      "theoretical_loss": 3.921971408285967,
      "tokens_seen": 491651072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297993981945837,
      "loss": 2.8791,
      "theoretical_loss": 3.92191493999169,
      "tokens_seen": 491716608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042978936810431296,
      "loss": 2.8162,
      "theoretical_loss": 3.921858481329994,
      "tokens_seen": 491782144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042977933801404214,
      "loss": 3.1575,
      "theoretical_loss": 3.9218020322979514,
      "tokens_seen": 491847680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297693079237713,
      "loss": 2.671,
      "theoretical_loss": 3.9217455928926377,
      "tokens_seen": 491913216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297592778335005,
      "loss": 2.9984,
      "theoretical_loss": 3.9216891631111293,
      "tokens_seen": 491978752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042974924774322973,
      "loss": 2.8436,
      "theoretical_loss": 3.9216327429505027,
      "tokens_seen": 492044288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042973921765295886,
      "loss": 2.7374,
      "theoretical_loss": 3.921576332407837,
      "tokens_seen": 492109824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297291875626881,
      "loss": 3.0636,
      "theoretical_loss": 3.921519931480213,
      "tokens_seen": 492175360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297191574724172,
      "loss": 3.0291,
      "theoretical_loss": 3.92146354016471,
      "tokens_seen": 492240896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042970912738214646,
      "loss": 2.9755,
      "theoretical_loss": 3.921407158458411,
      "tokens_seen": 492306432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042969909729187564,
      "loss": 2.7441,
      "theoretical_loss": 3.9213507863584014,
      "tokens_seen": 492371968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296890672016048,
      "loss": 2.8323,
      "theoretical_loss": 3.921294423861764,
      "tokens_seen": 492437504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429679037111334,
      "loss": 2.925,
      "theoretical_loss": 3.921238070965586,
      "tokens_seen": 492503040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296690070210632,
      "loss": 2.81,
      "theoretical_loss": 3.921181727666955,
      "tokens_seen": 492568576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042965897693079236,
      "loss": 2.8812,
      "theoretical_loss": 3.92112539396296,
      "tokens_seen": 492634112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296489468405216,
      "loss": 3.0883,
      "theoretical_loss": 3.9210690698506907,
      "tokens_seen": 492699648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296389167502507,
      "loss": 3.0465,
      "theoretical_loss": 3.9210127553272383,
      "tokens_seen": 492765184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042962888665997996,
      "loss": 2.9085,
      "theoretical_loss": 3.9209564503896956,
      "tokens_seen": 492830720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296188565697091,
      "loss": 2.6728,
      "theoretical_loss": 3.9209001550351563,
      "tokens_seen": 492896256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296088264794383,
      "loss": 2.7792,
      "theoretical_loss": 3.9208438692607164,
      "tokens_seen": 492961792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295987963891675,
      "loss": 2.9342,
      "theoretical_loss": 3.920787593063471,
      "tokens_seen": 493027328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295887662988967,
      "loss": 3.0404,
      "theoretical_loss": 3.9207313264405186,
      "tokens_seen": 493092864
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 282706,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2857396602630615,
      "objective/train/theoretical_loss": 3.9206750693889583,
      "objective/train/tokens_used": 513618400,
      "theoretical_loss": 3.9206750693889583,
      "tokens_seen": 493158400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042957873620862587,
      "loss": 2.9551,
      "theoretical_loss": 3.9206750693889583,
      "tokens_seen": 493158400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295687061183551,
      "loss": 2.7466,
      "theoretical_loss": 3.9206188219058897,
      "tokens_seen": 493223936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042955867602808423,
      "loss": 2.7163,
      "theoretical_loss": 3.9205625839884144,
      "tokens_seen": 493289472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042954864593781347,
      "loss": 2.7523,
      "theoretical_loss": 3.920506355633636,
      "tokens_seen": 493355008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295386158475426,
      "loss": 2.9156,
      "theoretical_loss": 3.9204501368386575,
      "tokens_seen": 493420544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042952858575727183,
      "loss": 2.8954,
      "theoretical_loss": 3.9203939276005846,
      "tokens_seen": 493486080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429518555667001,
      "loss": 2.885,
      "theoretical_loss": 3.9203377279165235,
      "tokens_seen": 493551616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295085255767302,
      "loss": 2.812,
      "theoretical_loss": 3.920281537783582,
      "tokens_seen": 493617152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042949849548645937,
      "loss": 2.797,
      "theoretical_loss": 3.9202253571988694,
      "tokens_seen": 493682688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042948846539618855,
      "loss": 2.858,
      "theoretical_loss": 3.9201691861594963,
      "tokens_seen": 493748224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042947843530591773,
      "loss": 2.8976,
      "theoretical_loss": 3.9201130246625735,
      "tokens_seen": 493813760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042946840521564697,
      "loss": 2.9483,
      "theoretical_loss": 3.920056872705214,
      "tokens_seen": 493879296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294583751253761,
      "loss": 2.8325,
      "theoretical_loss": 3.9200007302845323,
      "tokens_seen": 493944832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042944834503510533,
      "loss": 2.9839,
      "theoretical_loss": 3.9199445973976434,
      "tokens_seen": 494010368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294383149448345,
      "loss": 2.9204,
      "theoretical_loss": 3.9198884740416644,
      "tokens_seen": 494075904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294282848545637,
      "loss": 2.8143,
      "theoretical_loss": 3.9198323602137126,
      "tokens_seen": 494141440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042941825476429293,
      "loss": 3.0728,
      "theoretical_loss": 3.9197762559109073,
      "tokens_seen": 494206976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042940822467402206,
      "loss": 2.9142,
      "theoretical_loss": 3.919720161130368,
      "tokens_seen": 494272512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293981945837513,
      "loss": 3.0248,
      "theoretical_loss": 3.9196640758692176,
      "tokens_seen": 494338048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042938816449348047,
      "loss": 2.6867,
      "theoretical_loss": 3.9196080001245783,
      "tokens_seen": 494403584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042937813440320965,
      "loss": 2.8787,
      "theoretical_loss": 3.9195519338935743,
      "tokens_seen": 494469120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042936810431293883,
      "loss": 2.6973,
      "theoretical_loss": 3.919495877173331,
      "tokens_seen": 494534656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429358074222668,
      "loss": 2.8302,
      "theoretical_loss": 3.9194398299609743,
      "tokens_seen": 494600192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293480441323972,
      "loss": 3.1014,
      "theoretical_loss": 3.919383792253633,
      "tokens_seen": 494665728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042933801404212643,
      "loss": 2.8187,
      "theoretical_loss": 3.919327764048437,
      "tokens_seen": 494731264
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 283457,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3519198894500732,
      "objective/train/theoretical_loss": 3.919271745342514,
      "objective/train/tokens_used": 515256800,
      "theoretical_loss": 3.919271745342514,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042932798395185556,
      "loss": 2.9657,
      "theoretical_loss": 3.919271745342514,
      "tokens_seen": 494796800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293179538615848,
      "loss": 2.8827,
      "theoretical_loss": 3.9192157361329976,
      "tokens_seen": 494862336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293079237713139,
      "loss": 2.8081,
      "theoretical_loss": 3.9191597364170208,
      "tokens_seen": 494927872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042929789368104316,
      "loss": 2.9318,
      "theoretical_loss": 3.919103746191716,
      "tokens_seen": 494993408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042928786359077234,
      "loss": 2.7932,
      "theoretical_loss": 3.91904776545422,
      "tokens_seen": 495058944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292778335005015,
      "loss": 2.8108,
      "theoretical_loss": 3.918991794201669,
      "tokens_seen": 495124480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292678034102307,
      "loss": 2.684,
      "theoretical_loss": 3.9189358324312007,
      "tokens_seen": 495190016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042925777331995993,
      "loss": 2.6035,
      "theoretical_loss": 3.9188798801399543,
      "tokens_seen": 495255552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042924774322968906,
      "loss": 2.8571,
      "theoretical_loss": 3.9188239373250697,
      "tokens_seen": 495321088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292377131394183,
      "loss": 2.8791,
      "theoretical_loss": 3.9187680039836885,
      "tokens_seen": 495386624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292276830491474,
      "loss": 2.9489,
      "theoretical_loss": 3.918712080112954,
      "tokens_seen": 495452160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042921765295887666,
      "loss": 2.83,
      "theoretical_loss": 3.9186561657100096,
      "tokens_seen": 495517696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042920762286860584,
      "loss": 2.736,
      "theoretical_loss": 3.918600260772001,
      "tokens_seen": 495583232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429197592778335,
      "loss": 2.7423,
      "theoretical_loss": 3.9185443652960745,
      "tokens_seen": 495648768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291875626880642,
      "loss": 3.0763,
      "theoretical_loss": 3.918488479279378,
      "tokens_seen": 495714304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291775325977934,
      "loss": 2.8709,
      "theoretical_loss": 3.9184326027190606,
      "tokens_seen": 495779840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042916750250752256,
      "loss": 2.9463,
      "theoretical_loss": 3.9183767356122714,
      "tokens_seen": 495845376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291574724172518,
      "loss": 3.1239,
      "theoretical_loss": 3.9183208779561634,
      "tokens_seen": 495910912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291474423269809,
      "loss": 2.9989,
      "theoretical_loss": 3.9182650297478885,
      "tokens_seen": 495976448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042913741223671016,
      "loss": 2.8044,
      "theoretical_loss": 3.9182091909846,
      "tokens_seen": 496041984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291273821464393,
      "loss": 2.9178,
      "theoretical_loss": 3.9181533616634545,
      "tokens_seen": 496107520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291173520561685,
      "loss": 3.0983,
      "theoretical_loss": 3.918097541781607,
      "tokens_seen": 496173056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291073219658977,
      "loss": 3.0364,
      "theoretical_loss": 3.9180417313362157,
      "tokens_seen": 496238592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290972918756269,
      "loss": 2.9606,
      "theoretical_loss": 3.9179859303244395,
      "tokens_seen": 496304128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042908726178535607,
      "loss": 2.6999,
      "theoretical_loss": 3.917930138743438,
      "tokens_seen": 496369664
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 284572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4066948890686035,
      "objective/train/theoretical_loss": 3.9178743565903735,
      "objective/train/tokens_used": 516895200,
      "theoretical_loss": 3.9178743565903735,
      "tokens_seen": 496435200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290772316950853,
      "loss": 2.7991,
      "theoretical_loss": 3.9178743565903735,
      "tokens_seen": 496435200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042906720160481443,
      "loss": 2.9311,
      "theoretical_loss": 3.9178185838624078,
      "tokens_seen": 496500736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042905717151454367,
      "loss": 2.6928,
      "theoretical_loss": 3.917762820556705,
      "tokens_seen": 496566272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290471414242728,
      "loss": 2.8094,
      "theoretical_loss": 3.9177070666704292,
      "tokens_seen": 496631808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042903711133400203,
      "loss": 2.7391,
      "theoretical_loss": 3.917651322200747,
      "tokens_seen": 496697344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290270812437312,
      "loss": 2.8844,
      "theoretical_loss": 3.9175955871448265,
      "tokens_seen": 496762880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290170511534604,
      "loss": 2.9062,
      "theoretical_loss": 3.917539861499836,
      "tokens_seen": 496828416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042900702106318957,
      "loss": 2.9001,
      "theoretical_loss": 3.9174841452629456,
      "tokens_seen": 496893952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042899699097291875,
      "loss": 3.0454,
      "theoretical_loss": 3.9174284384313256,
      "tokens_seen": 496959488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042898696088264793,
      "loss": 2.9205,
      "theoretical_loss": 3.917372741002149,
      "tokens_seen": 497025024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042897693079237717,
      "loss": 2.9715,
      "theoretical_loss": 3.917317052972589,
      "tokens_seen": 497090560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289669007021063,
      "loss": 2.823,
      "theoretical_loss": 3.9172613743398204,
      "tokens_seen": 497156096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042895687061183553,
      "loss": 2.7293,
      "theoretical_loss": 3.91720570510102,
      "tokens_seen": 497221632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289468405215647,
      "loss": 2.8304,
      "theoretical_loss": 3.917150045253364,
      "tokens_seen": 497287168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289368104312939,
      "loss": 2.9458,
      "theoretical_loss": 3.917094394794031,
      "tokens_seen": 497352704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289267803410231,
      "loss": 3.1032,
      "theoretical_loss": 3.917038753720201,
      "tokens_seen": 497418240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042891675025075226,
      "loss": 2.8554,
      "theoretical_loss": 3.9169831220290554,
      "tokens_seen": 497483776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042890672016048144,
      "loss": 2.9893,
      "theoretical_loss": 3.916927499717775,
      "tokens_seen": 497549312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042889669007021067,
      "loss": 2.8381,
      "theoretical_loss": 3.916871886783544,
      "tokens_seen": 497614848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288866599799398,
      "loss": 2.9519,
      "theoretical_loss": 3.9168162832235467,
      "tokens_seen": 497680384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042887662988966903,
      "loss": 3.1386,
      "theoretical_loss": 3.9167606890349687,
      "tokens_seen": 497745920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042886659979939816,
      "loss": 2.5596,
      "theoretical_loss": 3.9167051042149974,
      "tokens_seen": 497811456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288565697091274,
      "loss": 2.9973,
      "theoretical_loss": 3.9166495287608205,
      "tokens_seen": 497876992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288465396188566,
      "loss": 2.8093,
      "theoretical_loss": 3.9165939626696273,
      "tokens_seen": 497942528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042883650952858576,
      "loss": 2.8571,
      "theoretical_loss": 3.916538405938609,
      "tokens_seen": 498008064
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 285190,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.950747489929199,
      "objective/train/theoretical_loss": 3.916482858564957,
      "objective/train/tokens_used": 518533600,
      "theoretical_loss": 3.916482858564957,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042882647943831494,
      "loss": 2.9687,
      "theoretical_loss": 3.916482858564957,
      "tokens_seen": 498073600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288164493480441,
      "loss": 2.7534,
      "theoretical_loss": 3.916427320545864,
      "tokens_seen": 498139136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288064192577733,
      "loss": 2.8904,
      "theoretical_loss": 3.9163717918785252,
      "tokens_seen": 498204672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042879638916750254,
      "loss": 2.9187,
      "theoretical_loss": 3.9163162725601355,
      "tokens_seen": 498270208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042878635907723166,
      "loss": 2.6975,
      "theoretical_loss": 3.916260762587892,
      "tokens_seen": 498335744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287763289869609,
      "loss": 2.695,
      "theoretical_loss": 3.916205261958991,
      "tokens_seen": 498401280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287662988966901,
      "loss": 2.7584,
      "theoretical_loss": 3.916149770670633,
      "tokens_seen": 498466816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042875626880641926,
      "loss": 2.8199,
      "theoretical_loss": 3.9160942887200183,
      "tokens_seen": 498532352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042874623871614844,
      "loss": 2.6627,
      "theoretical_loss": 3.916038816104348,
      "tokens_seen": 498597888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287362086258776,
      "loss": 2.9793,
      "theoretical_loss": 3.9159833528208248,
      "tokens_seen": 498663424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287261785356068,
      "loss": 2.6124,
      "theoretical_loss": 3.915927898866653,
      "tokens_seen": 498728960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042871614844533604,
      "loss": 2.8027,
      "theoretical_loss": 3.915872454239037,
      "tokens_seen": 498794496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042870611835506517,
      "loss": 3.0653,
      "theoretical_loss": 3.9158170189351833,
      "tokens_seen": 498860032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286960882647944,
      "loss": 3.1196,
      "theoretical_loss": 3.9157615929523004,
      "tokens_seen": 498925568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042868605817452353,
      "loss": 2.8093,
      "theoretical_loss": 3.9157061762875953,
      "tokens_seen": 498991104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042867602808425276,
      "loss": 2.8654,
      "theoretical_loss": 3.9156507689382796,
      "tokens_seen": 499056640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428665997993982,
      "loss": 2.7376,
      "theoretical_loss": 3.915595370901563,
      "tokens_seen": 499122176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286559679037111,
      "loss": 2.7512,
      "theoretical_loss": 3.915539982174659,
      "tokens_seen": 499187712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042864593781344036,
      "loss": 3.015,
      "theoretical_loss": 3.9154846027547805,
      "tokens_seen": 499253248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286359077231695,
      "loss": 2.6258,
      "theoretical_loss": 3.915429232639142,
      "tokens_seen": 499318784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286258776328987,
      "loss": 2.9637,
      "theoretical_loss": 3.9153738718249604,
      "tokens_seen": 499384320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286158475426279,
      "loss": 2.8947,
      "theoretical_loss": 3.9153185203094516,
      "tokens_seen": 499449856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286058174523571,
      "loss": 3.0437,
      "theoretical_loss": 3.915263178089835,
      "tokens_seen": 499515392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042859578736208627,
      "loss": 3.0086,
      "theoretical_loss": 3.915207845163329,
      "tokens_seen": 499580928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285857572718155,
      "loss": 2.7415,
      "theoretical_loss": 3.915152521527155,
      "tokens_seen": 499646464
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 286367,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.678431272506714,
      "objective/train/theoretical_loss": 3.915097207178535,
      "objective/train/tokens_used": 520172000,
      "theoretical_loss": 3.915097207178535,
      "tokens_seen": 499712000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042857572718154463,
      "loss": 2.8241,
      "theoretical_loss": 3.915097207178535,
      "tokens_seen": 499712000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042856569709127387,
      "loss": 2.8393,
      "theoretical_loss": 3.9150419021146927,
      "tokens_seen": 499777536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428555667001003,
      "loss": 2.9709,
      "theoretical_loss": 3.9149866063328504,
      "tokens_seen": 499843072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042854563691073223,
      "loss": 2.9225,
      "theoretical_loss": 3.914931319830236,
      "tokens_seen": 499908608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285356068204614,
      "loss": 2.8198,
      "theoretical_loss": 3.9148760426040736,
      "tokens_seen": 499974144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285255767301906,
      "loss": 2.7686,
      "theoretical_loss": 3.9148207746515933,
      "tokens_seen": 500039680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042851554663991977,
      "loss": 2.7698,
      "theoretical_loss": 3.9147655159700236,
      "tokens_seen": 500105216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042850551654964895,
      "loss": 2.8607,
      "theoretical_loss": 3.914710266556594,
      "tokens_seen": 500170752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042849548645937813,
      "loss": 2.9713,
      "theoretical_loss": 3.9146550264085374,
      "tokens_seen": 500236288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042848545636910737,
      "loss": 2.6836,
      "theoretical_loss": 3.9145997955230842,
      "tokens_seen": 500301824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284754262788365,
      "loss": 2.6966,
      "theoretical_loss": 3.9145445738974702,
      "tokens_seen": 500367360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042846539618856573,
      "loss": 2.6121,
      "theoretical_loss": 3.91448936152893,
      "tokens_seen": 500432896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284553660982949,
      "loss": 2.7794,
      "theoretical_loss": 3.914434158414699,
      "tokens_seen": 500498432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284453360080241,
      "loss": 3.1444,
      "theoretical_loss": 3.9143789645520153,
      "tokens_seen": 500563968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284353059177533,
      "loss": 2.7484,
      "theoretical_loss": 3.914323779938117,
      "tokens_seen": 500629504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042842527582748246,
      "loss": 2.6557,
      "theoretical_loss": 3.9142686045702444,
      "tokens_seen": 500695040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042841524573721164,
      "loss": 2.847,
      "theoretical_loss": 3.914213438445638,
      "tokens_seen": 500760576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042840521564694087,
      "loss": 3.133,
      "theoretical_loss": 3.9141582815615408,
      "tokens_seen": 500826112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042839518555667,
      "loss": 2.9062,
      "theoretical_loss": 3.9141031339151953,
      "tokens_seen": 500891648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042838515546639923,
      "loss": 2.8711,
      "theoretical_loss": 3.9140479955038456,
      "tokens_seen": 500957184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042837512537612836,
      "loss": 2.7578,
      "theoretical_loss": 3.9139928663247385,
      "tokens_seen": 501022720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283650952858576,
      "loss": 2.7539,
      "theoretical_loss": 3.9139377463751197,
      "tokens_seen": 501088256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283550651955868,
      "loss": 2.7564,
      "theoretical_loss": 3.913882635652238,
      "tokens_seen": 501153792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042834503510531596,
      "loss": 2.7827,
      "theoretical_loss": 3.9138275341533424,
      "tokens_seen": 501219328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042833500501504514,
      "loss": 2.7285,
      "theoretical_loss": 3.913772441875684,
      "tokens_seen": 501284864
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 286935,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.952582359313965,
      "objective/train/theoretical_loss": 3.9137173588165135,
      "objective/train/tokens_used": 521810400,
      "theoretical_loss": 3.9137173588165135,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283249749247743,
      "loss": 2.888,
      "theoretical_loss": 3.9137173588165135,
      "tokens_seen": 501350400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283149448345035,
      "loss": 2.7366,
      "theoretical_loss": 3.9136622849730833,
      "tokens_seen": 501415936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042830491474423274,
      "loss": 2.6463,
      "theoretical_loss": 3.913607220342648,
      "tokens_seen": 501481472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042829488465396186,
      "loss": 2.785,
      "theoretical_loss": 3.9135521649224634,
      "tokens_seen": 501547008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282848545636911,
      "loss": 2.9083,
      "theoretical_loss": 3.9134971187097847,
      "tokens_seen": 501612544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282748244734203,
      "loss": 2.886,
      "theoretical_loss": 3.913442081701869,
      "tokens_seen": 501678080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042826479438314946,
      "loss": 2.6593,
      "theoretical_loss": 3.9133870538959763,
      "tokens_seen": 501743616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042825476429287864,
      "loss": 2.8406,
      "theoretical_loss": 3.9133320352893657,
      "tokens_seen": 501809152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282447342026078,
      "loss": 2.881,
      "theoretical_loss": 3.913277025879298,
      "tokens_seen": 501874688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428234704112337,
      "loss": 2.8322,
      "theoretical_loss": 3.9132220256630363,
      "tokens_seen": 501940224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042822467402206624,
      "loss": 2.919,
      "theoretical_loss": 3.9131670346378424,
      "tokens_seen": 502005760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042821464393179537,
      "loss": 2.8986,
      "theoretical_loss": 3.913112052800982,
      "tokens_seen": 502071296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282046138415246,
      "loss": 3.0142,
      "theoretical_loss": 3.9130570801497204,
      "tokens_seen": 502136832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042819458375125373,
      "loss": 2.8962,
      "theoretical_loss": 3.9130021166813247,
      "tokens_seen": 502202368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042818455366098296,
      "loss": 2.9193,
      "theoretical_loss": 3.9129471623930625,
      "tokens_seen": 502267904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042817452357071215,
      "loss": 2.6178,
      "theoretical_loss": 3.912892217282203,
      "tokens_seen": 502333440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281644934804413,
      "loss": 2.7213,
      "theoretical_loss": 3.912837281346017,
      "tokens_seen": 502398976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281544633901705,
      "loss": 3.0234,
      "theoretical_loss": 3.9127823545817755,
      "tokens_seen": 502464512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281444332998997,
      "loss": 2.8943,
      "theoretical_loss": 3.9127274369867515,
      "tokens_seen": 502530048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042813440320962887,
      "loss": 3.3673,
      "theoretical_loss": 3.912672528558219,
      "tokens_seen": 502595584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281243731193581,
      "loss": 2.9674,
      "theoretical_loss": 3.9126176292934525,
      "tokens_seen": 502661120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042811434302908723,
      "loss": 2.7487,
      "theoretical_loss": 3.9125627391897284,
      "tokens_seen": 502726656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042810431293881647,
      "loss": 2.9542,
      "theoretical_loss": 3.9125078582443247,
      "tokens_seen": 502792192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042809428284854565,
      "loss": 2.7127,
      "theoretical_loss": 3.912452986454519,
      "tokens_seen": 502857728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042808425275827483,
      "loss": 2.9038,
      "theoretical_loss": 3.9123981238175913,
      "tokens_seen": 502923264
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 288241,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.815506935119629,
      "objective/train/theoretical_loss": 3.9123432703308225,
      "objective/train/tokens_used": 523448800,
      "theoretical_loss": 3.9123432703308225,
      "tokens_seen": 502988800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428074222668004,
      "loss": 2.8686,
      "theoretical_loss": 3.9123432703308225,
      "tokens_seen": 502988800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280641925777332,
      "loss": 2.8739,
      "theoretical_loss": 3.9122884259914947,
      "tokens_seen": 503054336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280541624874624,
      "loss": 2.9241,
      "theoretical_loss": 3.912233590796891,
      "tokens_seen": 503119872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280441323971916,
      "loss": 2.8723,
      "theoretical_loss": 3.912178764744296,
      "tokens_seen": 503185408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042803410230692074,
      "loss": 2.7677,
      "theoretical_loss": 3.9121239478309944,
      "tokens_seen": 503250944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042802407221664997,
      "loss": 2.8224,
      "theoretical_loss": 3.9120691400542738,
      "tokens_seen": 503316480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280140421263791,
      "loss": 2.6969,
      "theoretical_loss": 3.912014341411421,
      "tokens_seen": 503382016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042800401203610833,
      "loss": 2.9163,
      "theoretical_loss": 3.9119595518997254,
      "tokens_seen": 503447552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279939819458375,
      "loss": 3.1166,
      "theoretical_loss": 3.911904771516477,
      "tokens_seen": 503513088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279839518555667,
      "loss": 2.8336,
      "theoretical_loss": 3.9118500002589682,
      "tokens_seen": 503578624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279739217652959,
      "loss": 2.8036,
      "theoretical_loss": 3.9117952381244896,
      "tokens_seen": 503644160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279638916750251,
      "loss": 2.8712,
      "theoretical_loss": 3.911740485110337,
      "tokens_seen": 503709696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042795386158475424,
      "loss": 2.8051,
      "theoretical_loss": 3.9116857412138026,
      "tokens_seen": 503775232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279438314944835,
      "loss": 2.9153,
      "theoretical_loss": 3.9116310064321835,
      "tokens_seen": 503840768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279338014042126,
      "loss": 2.7278,
      "theoretical_loss": 3.9115762807627776,
      "tokens_seen": 503906304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042792377131394184,
      "loss": 2.814,
      "theoretical_loss": 3.9115215642028813,
      "tokens_seen": 503971840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042791374122367107,
      "loss": 2.7999,
      "theoretical_loss": 3.911466856749796,
      "tokens_seen": 504037376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004279037111334002,
      "loss": 2.8065,
      "theoretical_loss": 3.911412158400821,
      "tokens_seen": 504102912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042789368104312943,
      "loss": 2.8346,
      "theoretical_loss": 3.9113574691532573,
      "tokens_seen": 504168448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042788365095285856,
      "loss": 3.0613,
      "theoretical_loss": 3.911302789004409,
      "tokens_seen": 504233984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278736208625878,
      "loss": 2.9722,
      "theoretical_loss": 3.9112481179515797,
      "tokens_seen": 504299520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427863590772317,
      "loss": 2.8922,
      "theoretical_loss": 3.9111934559920742,
      "tokens_seen": 504365056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042785356068204616,
      "loss": 2.7941,
      "theoretical_loss": 3.9111388031231993,
      "tokens_seen": 504430592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042784353059177534,
      "loss": 2.9636,
      "theoretical_loss": 3.911084159342261,
      "tokens_seen": 504496128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278335005015045,
      "loss": 2.7491,
      "theoretical_loss": 3.91102952464657,
      "tokens_seen": 504561664
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 289179,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2568519115448,
      "objective/train/theoretical_loss": 3.9109748990334348,
      "objective/train/tokens_used": 525087200,
      "theoretical_loss": 3.9109748990334348,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278234704112337,
      "loss": 2.8859,
      "theoretical_loss": 3.9109748990334348,
      "tokens_seen": 504627200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042781344032096294,
      "loss": 2.7951,
      "theoretical_loss": 3.910920282500166,
      "tokens_seen": 504692736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042780341023069206,
      "loss": 3.0004,
      "theoretical_loss": 3.910865675044076,
      "tokens_seen": 504758272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277933801404213,
      "loss": 2.8993,
      "theoretical_loss": 3.9108110766624775,
      "tokens_seen": 504823808
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277833500501505,
      "loss": 2.6838,
      "theoretical_loss": 3.9107564873526854,
      "tokens_seen": 504889344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042777331995987966,
      "loss": 2.6696,
      "theoretical_loss": 3.9107019071120153,
      "tokens_seen": 504954880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042776328986960884,
      "loss": 2.7526,
      "theoretical_loss": 3.910647335937783,
      "tokens_seen": 505020416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427753259779338,
      "loss": 3.007,
      "theoretical_loss": 3.910592773827306,
      "tokens_seen": 505085952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277432296890672,
      "loss": 2.6543,
      "theoretical_loss": 3.9105382207779043,
      "tokens_seen": 505151488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042773319959879644,
      "loss": 2.9892,
      "theoretical_loss": 3.9104836767868973,
      "tokens_seen": 505217024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042772316950852557,
      "loss": 2.8852,
      "theoretical_loss": 3.9104291418516057,
      "tokens_seen": 505282560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277131394182548,
      "loss": 2.5701,
      "theoretical_loss": 3.910374615969352,
      "tokens_seen": 505348096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042770310932798393,
      "loss": 3.0667,
      "theoretical_loss": 3.9103200991374605,
      "tokens_seen": 505413632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042769307923771316,
      "loss": 2.8631,
      "theoretical_loss": 3.9102655913532542,
      "tokens_seen": 505479168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042768304914744235,
      "loss": 2.7267,
      "theoretical_loss": 3.9102110926140603,
      "tokens_seen": 505544704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042767301905717153,
      "loss": 2.6952,
      "theoretical_loss": 3.910156602917205,
      "tokens_seen": 505610240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276629889669007,
      "loss": 2.6144,
      "theoretical_loss": 3.9101021222600156,
      "tokens_seen": 505675776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276529588766299,
      "loss": 2.9215,
      "theoretical_loss": 3.9100476506398216,
      "tokens_seen": 505741312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042764292878635907,
      "loss": 3.0363,
      "theoretical_loss": 3.909993188053954,
      "tokens_seen": 505806848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276328986960883,
      "loss": 2.9977,
      "theoretical_loss": 3.909938734499743,
      "tokens_seen": 505872384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042762286860581743,
      "loss": 2.7794,
      "theoretical_loss": 3.9098842899745216,
      "tokens_seen": 505937920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042761283851554667,
      "loss": 2.673,
      "theoretical_loss": 3.909829854475624,
      "tokens_seen": 506003456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042760280842527585,
      "loss": 2.8276,
      "theoretical_loss": 3.9097754280003842,
      "tokens_seen": 506068992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042759277833500503,
      "loss": 2.754,
      "theoretical_loss": 3.9097210105461384,
      "tokens_seen": 506134528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275827482447342,
      "loss": 3.0903,
      "theoretical_loss": 3.9096666021102235,
      "tokens_seen": 506200064
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 289852,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1201510429382324,
      "objective/train/theoretical_loss": 3.909612202689977,
      "objective/train/tokens_used": 526725600,
      "theoretical_loss": 3.909612202689977,
      "tokens_seen": 506265600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275727181544634,
      "loss": 2.8416,
      "theoretical_loss": 3.909612202689977,
      "tokens_seen": 506265600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275626880641926,
      "loss": 2.6884,
      "theoretical_loss": 3.90955781228274,
      "tokens_seen": 506331136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275526579739218,
      "loss": 2.7749,
      "theoretical_loss": 3.909503430885851,
      "tokens_seen": 506396672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042754262788365094,
      "loss": 2.8646,
      "theoretical_loss": 3.9094490584966524,
      "tokens_seen": 506462208
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042753259779338017,
      "loss": 2.8293,
      "theoretical_loss": 3.9093946951124874,
      "tokens_seen": 506527744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275225677031093,
      "loss": 2.9254,
      "theoretical_loss": 3.9093403407306986,
      "tokens_seen": 506593280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042751253761283853,
      "loss": 2.8896,
      "theoretical_loss": 3.9092859953486316,
      "tokens_seen": 506658816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275025075225677,
      "loss": 2.9322,
      "theoretical_loss": 3.909231658963633,
      "tokens_seen": 506724352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274924774322969,
      "loss": 2.8241,
      "theoretical_loss": 3.9091773315730487,
      "tokens_seen": 506789888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274824473420261,
      "loss": 2.8595,
      "theoretical_loss": 3.909123013174228,
      "tokens_seen": 506855424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274724172517553,
      "loss": 3.029,
      "theoretical_loss": 3.90906870376452,
      "tokens_seen": 506920960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042746238716148444,
      "loss": 2.889,
      "theoretical_loss": 3.909014403341275,
      "tokens_seen": 506986496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274523570712137,
      "loss": 3.0323,
      "theoretical_loss": 3.9089601119018456,
      "tokens_seen": 507052032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274423269809428,
      "loss": 3.1528,
      "theoretical_loss": 3.908905829443583,
      "tokens_seen": 507117568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042743229689067204,
      "loss": 2.9999,
      "theoretical_loss": 3.9088515559638424,
      "tokens_seen": 507183104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274222668004012,
      "loss": 2.8889,
      "theoretical_loss": 3.9087972914599787,
      "tokens_seen": 507248640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274122367101304,
      "loss": 2.9263,
      "theoretical_loss": 3.908743035929348,
      "tokens_seen": 507314176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274022066198596,
      "loss": 2.697,
      "theoretical_loss": 3.9086887893693065,
      "tokens_seen": 507379712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042739217652958876,
      "loss": 2.9639,
      "theoretical_loss": 3.9086345517772143,
      "tokens_seen": 507445248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042738214643931794,
      "loss": 2.8796,
      "theoretical_loss": 3.90858032315043,
      "tokens_seen": 507510784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273721163490472,
      "loss": 2.7631,
      "theoretical_loss": 3.9085261034863144,
      "tokens_seen": 507576320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273620862587763,
      "loss": 3.0314,
      "theoretical_loss": 3.908471892782229,
      "tokens_seen": 507641856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042735205616850554,
      "loss": 2.9613,
      "theoretical_loss": 3.9084176910355373,
      "tokens_seen": 507707392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042734202607823467,
      "loss": 3.1484,
      "theoretical_loss": 3.9083634982436033,
      "tokens_seen": 507772928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273319959879639,
      "loss": 2.9547,
      "theoretical_loss": 3.908309314403791,
      "tokens_seen": 507838464
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 291045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8749454021453857,
      "objective/train/theoretical_loss": 3.908255139513467,
      "objective/train/tokens_used": 528364000,
      "theoretical_loss": 3.908255139513467,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273219658976931,
      "loss": 3.0369,
      "theoretical_loss": 3.908255139513467,
      "tokens_seen": 507904000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042731193580742226,
      "loss": 2.8929,
      "theoretical_loss": 3.9082009735699996,
      "tokens_seen": 507969536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042730190571715145,
      "loss": 2.7172,
      "theoretical_loss": 3.9081468165707562,
      "tokens_seen": 508035072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272918756268807,
      "loss": 2.8316,
      "theoretical_loss": 3.9080926685131074,
      "tokens_seen": 508100608
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272818455366098,
      "loss": 2.7872,
      "theoretical_loss": 3.908038529394423,
      "tokens_seen": 508166144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042727181544633904,
      "loss": 2.8132,
      "theoretical_loss": 3.907984399212075,
      "tokens_seen": 508231680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042726178535606817,
      "loss": 2.9742,
      "theoretical_loss": 3.9079302779634366,
      "tokens_seen": 508297216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272517552657974,
      "loss": 2.7523,
      "theoretical_loss": 3.907876165645881,
      "tokens_seen": 508362752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272417251755266,
      "loss": 2.9132,
      "theoretical_loss": 3.907822062256784,
      "tokens_seen": 508428288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042723169508525577,
      "loss": 2.6844,
      "theoretical_loss": 3.9077679677935224,
      "tokens_seen": 508493824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042722166499498495,
      "loss": 2.5976,
      "theoretical_loss": 3.9077138822534723,
      "tokens_seen": 508559360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042721163490471413,
      "loss": 2.9316,
      "theoretical_loss": 3.9076598056340126,
      "tokens_seen": 508624896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272016048144433,
      "loss": 3.027,
      "theoretical_loss": 3.9076057379325233,
      "tokens_seen": 508690432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042719157472417255,
      "loss": 2.8114,
      "theoretical_loss": 3.9075516791463842,
      "tokens_seen": 508755968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271815446339017,
      "loss": 2.8752,
      "theoretical_loss": 3.907497629272978,
      "tokens_seen": 508821504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271715145436309,
      "loss": 2.7832,
      "theoretical_loss": 3.907443588309687,
      "tokens_seen": 508887040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271614844533601,
      "loss": 2.9496,
      "theoretical_loss": 3.9073895562538956,
      "tokens_seen": 508952576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042715145436308927,
      "loss": 2.7524,
      "theoretical_loss": 3.907335533102988,
      "tokens_seen": 509018112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271414242728185,
      "loss": 2.7836,
      "theoretical_loss": 3.9072815188543517,
      "tokens_seen": 509083648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042713139418254763,
      "loss": 2.78,
      "theoretical_loss": 3.907227513505373,
      "tokens_seen": 509149184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042712136409227687,
      "loss": 3.1176,
      "theoretical_loss": 3.9071735170534407,
      "tokens_seen": 509214720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042711133400200605,
      "loss": 2.789,
      "theoretical_loss": 3.9071195294959438,
      "tokens_seen": 509280256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042710130391173523,
      "loss": 2.7216,
      "theoretical_loss": 3.9070655508302736,
      "tokens_seen": 509345792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270912738214644,
      "loss": 2.9398,
      "theoretical_loss": 3.9070115810538217,
      "tokens_seen": 509411328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270812437311936,
      "loss": 2.7936,
      "theoretical_loss": 3.9069576201639804,
      "tokens_seen": 509476864
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 291444,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.291428804397583,
      "objective/train/theoretical_loss": 3.9069036681581446,
      "objective/train/tokens_used": 530002400,
      "theoretical_loss": 3.9069036681581446,
      "tokens_seen": 509542400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270712136409228,
      "loss": 2.9279,
      "theoretical_loss": 3.9069036681581446,
      "tokens_seen": 509542400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427061183550652,
      "loss": 3.0056,
      "theoretical_loss": 3.906849725033708,
      "tokens_seen": 509607936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042705115346038114,
      "loss": 2.9895,
      "theoretical_loss": 3.906795790788067,
      "tokens_seen": 509673472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042704112337011037,
      "loss": 2.8832,
      "theoretical_loss": 3.9067418654186197,
      "tokens_seen": 509739008
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270310932798395,
      "loss": 2.8935,
      "theoretical_loss": 3.9066879489227633,
      "tokens_seen": 509804544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042702106318956873,
      "loss": 2.7562,
      "theoretical_loss": 3.906634041297898,
      "tokens_seen": 509870080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270110330992979,
      "loss": 3.0871,
      "theoretical_loss": 3.906580142541424,
      "tokens_seen": 509935616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270010030090271,
      "loss": 2.7015,
      "theoretical_loss": 3.906526252650743,
      "tokens_seen": 510001152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269909729187563,
      "loss": 3.0532,
      "theoretical_loss": 3.9064723716232574,
      "tokens_seen": 510066688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269809428284855,
      "loss": 2.7929,
      "theoretical_loss": 3.9064184994563713,
      "tokens_seen": 510132224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042697091273821464,
      "loss": 2.8344,
      "theoretical_loss": 3.9063646361474893,
      "tokens_seen": 510197760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269608826479439,
      "loss": 2.9352,
      "theoretical_loss": 3.906310781694017,
      "tokens_seen": 510263296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000426950852557673,
      "loss": 2.9629,
      "theoretical_loss": 3.9062569360933623,
      "tokens_seen": 510328832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042694082246740224,
      "loss": 2.9865,
      "theoretical_loss": 3.906203099342933,
      "tokens_seen": 510394368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269307923771314,
      "loss": 2.8716,
      "theoretical_loss": 3.9061492714401385,
      "tokens_seen": 510459904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269207622868606,
      "loss": 2.8544,
      "theoretical_loss": 3.9060954523823885,
      "tokens_seen": 510525440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269107321965898,
      "loss": 3.1565,
      "theoretical_loss": 3.906041642167095,
      "tokens_seen": 510590976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042690070210631896,
      "loss": 2.9906,
      "theoretical_loss": 3.9059878407916706,
      "tokens_seen": 510656512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042689067201604814,
      "loss": 2.8862,
      "theoretical_loss": 3.9059340482535285,
      "tokens_seen": 510722048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268806419257774,
      "loss": 2.815,
      "theoretical_loss": 3.905880264550083,
      "tokens_seen": 510787584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268706118355065,
      "loss": 2.8911,
      "theoretical_loss": 3.905826489678751,
      "tokens_seen": 510853120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042686058174523574,
      "loss": 2.9307,
      "theoretical_loss": 3.9057727236369484,
      "tokens_seen": 510918656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042685055165496487,
      "loss": 2.6768,
      "theoretical_loss": 3.9057189664220937,
      "tokens_seen": 510984192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268405215646941,
      "loss": 2.9439,
      "theoretical_loss": 3.905665218031606,
      "tokens_seen": 511049728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268304914744233,
      "loss": 2.9301,
      "theoretical_loss": 3.9056114784629052,
      "tokens_seen": 511115264
    },
    {
      "epoch": 0.15,
      "objective/train/docs_used": 292775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.923103094100952,
      "objective/train/theoretical_loss": 3.905557747713412,
      "objective/train/tokens_used": 531640800,
      "theoretical_loss": 3.905557747713412,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042682046138415246,
      "loss": 2.9187,
      "theoretical_loss": 3.905557747713412,
      "tokens_seen": 511180800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042681043129388165,
      "loss": 3.001,
      "theoretical_loss": 3.9055040257805493,
      "tokens_seen": 511246336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268004012036109,
      "loss": 2.946,
      "theoretical_loss": 3.9054503126617406,
      "tokens_seen": 511311872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042679037111334,
      "loss": 2.8961,
      "theoretical_loss": 3.90539660835441,
      "tokens_seen": 511377408
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042678034102306924,
      "loss": 2.9013,
      "theoretical_loss": 3.905342912855983,
      "tokens_seen": 511442944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042677031093279837,
      "loss": 2.8939,
      "theoretical_loss": 3.9052892261638865,
      "tokens_seen": 511508480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267602808425276,
      "loss": 3.2111,
      "theoretical_loss": 3.905235548275548,
      "tokens_seen": 511574016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267502507522568,
      "loss": 2.9137,
      "theoretical_loss": 3.9051818791883965,
      "tokens_seen": 511639552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042674022066198597,
      "loss": 2.8786,
      "theoretical_loss": 3.9051282188998613,
      "tokens_seen": 511705088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042673019057171515,
      "loss": 2.8007,
      "theoretical_loss": 3.9050745674073744,
      "tokens_seen": 511770624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042672016048144433,
      "loss": 2.7962,
      "theoretical_loss": 3.9050209247083663,
      "tokens_seen": 511836160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267101303911735,
      "loss": 2.9068,
      "theoretical_loss": 3.9049672908002715,
      "tokens_seen": 511901696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042670010030090275,
      "loss": 3.034,
      "theoretical_loss": 3.9049136656805237,
      "tokens_seen": 511967232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266900702106319,
      "loss": 2.7779,
      "theoretical_loss": 3.904860049346558,
      "tokens_seen": 512032768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266800401203611,
      "loss": 2.7907,
      "theoretical_loss": 3.9048064417958104,
      "tokens_seen": 512098304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042667001003009024,
      "loss": 3.0229,
      "theoretical_loss": 3.9047528430257192,
      "tokens_seen": 512163840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042665997993981947,
      "loss": 2.7588,
      "theoretical_loss": 3.904699253033723,
      "tokens_seen": 512229376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042664994984954865,
      "loss": 2.9785,
      "theoretical_loss": 3.9046456718172595,
      "tokens_seen": 512294912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042663991975927783,
      "loss": 2.7284,
      "theoretical_loss": 3.9045920993737715,
      "tokens_seen": 512360448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426629889669007,
      "loss": 2.9626,
      "theoretical_loss": 3.9045385357006994,
      "tokens_seen": 512425984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042661985957873625,
      "loss": 2.9813,
      "theoretical_loss": 3.9044849807954867,
      "tokens_seen": 512491520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266098294884654,
      "loss": 2.6292,
      "theoretical_loss": 3.904431434655576,
      "tokens_seen": 512557056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265997993981946,
      "loss": 2.8153,
      "theoretical_loss": 3.9043778972784144,
      "tokens_seen": 512622592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042658976930792374,
      "loss": 2.8765,
      "theoretical_loss": 3.904324368661446,
      "tokens_seen": 512688128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426579739217653,
      "loss": 2.8445,
      "theoretical_loss": 3.904270848802119,
      "tokens_seen": 512753664
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 293228,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3358237743377686,
      "objective/train/theoretical_loss": 3.904217337697881,
      "objective/train/tokens_used": 533279200,
      "theoretical_loss": 3.904217337697881,
      "tokens_seen": 512819200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042656970912738216,
      "loss": 2.832,
      "theoretical_loss": 3.904217337697881,
      "tokens_seen": 512819200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042655967903711134,
      "loss": 2.9333,
      "theoretical_loss": 3.904163835346181,
      "tokens_seen": 512884736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265496489468405,
      "loss": 3.0451,
      "theoretical_loss": 3.9041103417444694,
      "tokens_seen": 512950272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265396188565697,
      "loss": 2.9155,
      "theoretical_loss": 3.9040568568901977,
      "tokens_seen": 513015808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265295887662989,
      "loss": 2.9664,
      "theoretical_loss": 3.9040033807808188,
      "tokens_seen": 513081344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265195586760281,
      "loss": 3.024,
      "theoretical_loss": 3.9039499134137854,
      "tokens_seen": 513146880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042650952858575724,
      "loss": 2.8769,
      "theoretical_loss": 3.903896454786552,
      "tokens_seen": 513212416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264994984954865,
      "loss": 2.8886,
      "theoretical_loss": 3.903843004896575,
      "tokens_seen": 513277952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264894684052156,
      "loss": 2.8289,
      "theoretical_loss": 3.9037895637413103,
      "tokens_seen": 513343488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042647943831494484,
      "loss": 2.8079,
      "theoretical_loss": 3.903736131318216,
      "tokens_seen": 513409024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426469408224674,
      "loss": 2.8159,
      "theoretical_loss": 3.9036827076247516,
      "tokens_seen": 513474560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264593781344032,
      "loss": 2.9007,
      "theoretical_loss": 3.9036292926583753,
      "tokens_seen": 513540096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264493480441324,
      "loss": 2.8079,
      "theoretical_loss": 3.903575886416549,
      "tokens_seen": 513605632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264393179538616,
      "loss": 2.7109,
      "theoretical_loss": 3.9035224888967344,
      "tokens_seen": 513671168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042642928786359075,
      "loss": 2.8299,
      "theoretical_loss": 3.9034691000963955,
      "tokens_seen": 513736704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042641925777332,
      "loss": 2.8933,
      "theoretical_loss": 3.9034157200129953,
      "tokens_seen": 513802240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042640922768304916,
      "loss": 2.8528,
      "theoretical_loss": 3.9033623486439994,
      "tokens_seen": 513867776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042639919759277834,
      "loss": 3.0509,
      "theoretical_loss": 3.9033089859868744,
      "tokens_seen": 513933312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263891675025076,
      "loss": 2.8992,
      "theoretical_loss": 3.9032556320390874,
      "tokens_seen": 513998848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263791374122367,
      "loss": 2.7701,
      "theoretical_loss": 3.9032022867981055,
      "tokens_seen": 514064384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042636910732196594,
      "loss": 2.9763,
      "theoretical_loss": 3.9031489502614,
      "tokens_seen": 514129920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042635907723169507,
      "loss": 2.6713,
      "theoretical_loss": 3.9030956224264406,
      "tokens_seen": 514195456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263490471414243,
      "loss": 2.7494,
      "theoretical_loss": 3.903042303290699,
      "tokens_seen": 514260992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263390170511535,
      "loss": 2.9207,
      "theoretical_loss": 3.9029889928516472,
      "tokens_seen": 514326528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042632898696088266,
      "loss": 2.804,
      "theoretical_loss": 3.902935691106759,
      "tokens_seen": 514392064
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 294365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.787262201309204,
      "objective/train/theoretical_loss": 3.90288239805351,
      "objective/train/tokens_used": 534917600,
      "theoretical_loss": 3.90288239805351,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042631895687061185,
      "loss": 3.0647,
      "theoretical_loss": 3.90288239805351,
      "tokens_seen": 514457600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263089267803411,
      "loss": 2.6791,
      "theoretical_loss": 3.902829113689375,
      "tokens_seen": 514523136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262988966900702,
      "loss": 2.7555,
      "theoretical_loss": 3.902775838011831,
      "tokens_seen": 514588672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042628886659979944,
      "loss": 2.9535,
      "theoretical_loss": 3.902722571018356,
      "tokens_seen": 514654208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042627883650952857,
      "loss": 2.8113,
      "theoretical_loss": 3.902669312706429,
      "tokens_seen": 514719744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262688064192578,
      "loss": 2.8344,
      "theoretical_loss": 3.90261606307353,
      "tokens_seen": 514785280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426258776328987,
      "loss": 3.0159,
      "theoretical_loss": 3.90256282211714,
      "tokens_seen": 514850816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042624874623871617,
      "loss": 2.8466,
      "theoretical_loss": 3.9025095898347404,
      "tokens_seen": 514916352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042623871614844535,
      "loss": 2.6467,
      "theoretical_loss": 3.9024563662238156,
      "tokens_seen": 514981888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042622868605817453,
      "loss": 2.5978,
      "theoretical_loss": 3.902403151281848,
      "tokens_seen": 515047424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262186559679037,
      "loss": 2.5738,
      "theoretical_loss": 3.902349945006325,
      "tokens_seen": 515112960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042620862587763295,
      "loss": 3.0073,
      "theoretical_loss": 3.9022967473947308,
      "tokens_seen": 515178496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261985957873621,
      "loss": 2.8529,
      "theoretical_loss": 3.9022435584445545,
      "tokens_seen": 515244032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261885656970913,
      "loss": 2.8852,
      "theoretical_loss": 3.902190378153283,
      "tokens_seen": 515309568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042617853560682044,
      "loss": 2.8072,
      "theoretical_loss": 3.9021372065184066,
      "tokens_seen": 515375104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042616850551654967,
      "loss": 2.724,
      "theoretical_loss": 3.902084043537416,
      "tokens_seen": 515440640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042615847542627885,
      "loss": 2.8464,
      "theoretical_loss": 3.902030889207802,
      "tokens_seen": 515506176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042614844533600803,
      "loss": 2.5744,
      "theoretical_loss": 3.9019777435270573,
      "tokens_seen": 515571712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261384152457372,
      "loss": 2.8335,
      "theoretical_loss": 3.901924606492675,
      "tokens_seen": 515637248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042612838515546645,
      "loss": 3.0876,
      "theoretical_loss": 3.9018714781021515,
      "tokens_seen": 515702784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261183550651956,
      "loss": 2.8044,
      "theoretical_loss": 3.9018183583529806,
      "tokens_seen": 515768320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261083249749248,
      "loss": 2.9733,
      "theoretical_loss": 3.9017652472426603,
      "tokens_seen": 515833856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042609829488465394,
      "loss": 2.9147,
      "theoretical_loss": 3.9017121447686876,
      "tokens_seen": 515899392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260882647943832,
      "loss": 2.93,
      "theoretical_loss": 3.9016590509285614,
      "tokens_seen": 515964928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042607823470411236,
      "loss": 2.8142,
      "theoretical_loss": 3.901605965719782,
      "tokens_seen": 516030464
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 295169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.053830623626709,
      "objective/train/theoretical_loss": 3.9015528891398494,
      "objective/train/tokens_used": 536556000,
      "theoretical_loss": 3.9015528891398494,
      "tokens_seen": 516096000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042606820461384154,
      "loss": 3.1288,
      "theoretical_loss": 3.9015528891398494,
      "tokens_seen": 516096000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260581745235707,
      "loss": 2.694,
      "theoretical_loss": 3.9014998211862673,
      "tokens_seen": 516161536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260481444332999,
      "loss": 3.0028,
      "theoretical_loss": 3.901446761856537,
      "tokens_seen": 516227072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260381143430291,
      "loss": 2.7803,
      "theoretical_loss": 3.901393711148163,
      "tokens_seen": 516292608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260280842527583,
      "loss": 2.8191,
      "theoretical_loss": 3.901340669058651,
      "tokens_seen": 516358144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042601805416248744,
      "loss": 3.0177,
      "theoretical_loss": 3.9012876355855064,
      "tokens_seen": 516423680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260080240722167,
      "loss": 2.7097,
      "theoretical_loss": 3.9012346107262363,
      "tokens_seen": 516489216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259979939819458,
      "loss": 2.9959,
      "theoretical_loss": 3.9011815944783494,
      "tokens_seen": 516554752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042598796389167504,
      "loss": 3.0318,
      "theoretical_loss": 3.9011285868393544,
      "tokens_seen": 516620288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259779338014042,
      "loss": 2.9474,
      "theoretical_loss": 3.901075587806762,
      "tokens_seen": 516685824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259679037111334,
      "loss": 3.0822,
      "theoretical_loss": 3.901022597378083,
      "tokens_seen": 516751360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259578736208626,
      "loss": 2.8803,
      "theoretical_loss": 3.90096961555083,
      "tokens_seen": 516816896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259478435305918,
      "loss": 2.8745,
      "theoretical_loss": 3.9009166423225166,
      "tokens_seen": 516882432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042593781344032095,
      "loss": 3.0311,
      "theoretical_loss": 3.9008636776906567,
      "tokens_seen": 516947968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259277833500502,
      "loss": 2.9673,
      "theoretical_loss": 3.9008107216527663,
      "tokens_seen": 517013504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004259177532597793,
      "loss": 2.9108,
      "theoretical_loss": 3.9007577742063613,
      "tokens_seen": 517079040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042590772316950854,
      "loss": 2.7388,
      "theoretical_loss": 3.900704835348959,
      "tokens_seen": 517144576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258976930792377,
      "loss": 2.991,
      "theoretical_loss": 3.900651905078079,
      "tokens_seen": 517210112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258876629889669,
      "loss": 2.7754,
      "theoretical_loss": 3.90059898339124,
      "tokens_seen": 517275648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258776328986961,
      "loss": 2.8331,
      "theoretical_loss": 3.9005460702859627,
      "tokens_seen": 517341184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042586760280842527,
      "loss": 2.9162,
      "theoretical_loss": 3.9004931657597686,
      "tokens_seen": 517406720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042585757271815445,
      "loss": 2.8646,
      "theoretical_loss": 3.900440269810181,
      "tokens_seen": 517472256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258475426278837,
      "loss": 3.0237,
      "theoretical_loss": 3.900387382434723,
      "tokens_seen": 517537792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258375125376128,
      "loss": 2.8647,
      "theoretical_loss": 3.900334503630919,
      "tokens_seen": 517603328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042582748244734205,
      "loss": 2.8821,
      "theoretical_loss": 3.9002816333962955,
      "tokens_seen": 517668864
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 295690,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6655638217926025,
      "objective/train/theoretical_loss": 3.9002287717283783,
      "objective/train/tokens_used": 538194400,
      "theoretical_loss": 3.9002287717283783,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258174523570712,
      "loss": 2.6781,
      "theoretical_loss": 3.9002287717283783,
      "tokens_seen": 517734400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258074222668004,
      "loss": 2.737,
      "theoretical_loss": 3.9001759186246967,
      "tokens_seen": 517799936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257973921765296,
      "loss": 2.8423,
      "theoretical_loss": 3.900123074082778,
      "tokens_seen": 517865472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042578736208625877,
      "loss": 3.0461,
      "theoretical_loss": 3.9000702381001533,
      "tokens_seen": 517931008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042577733199598795,
      "loss": 3.0473,
      "theoretical_loss": 3.900017410674352,
      "tokens_seen": 517996544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257673019057172,
      "loss": 2.764,
      "theoretical_loss": 3.899964591802907,
      "tokens_seen": 518062080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257572718154463,
      "loss": 3.0369,
      "theoretical_loss": 3.8999117814833513,
      "tokens_seen": 518127616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042574724172517555,
      "loss": 2.8756,
      "theoretical_loss": 3.899858979713218,
      "tokens_seen": 518193152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257372116349047,
      "loss": 3.0317,
      "theoretical_loss": 3.899806186490043,
      "tokens_seen": 518258688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257271815446339,
      "loss": 2.885,
      "theoretical_loss": 3.8997534018113615,
      "tokens_seen": 518324224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257171514543631,
      "loss": 2.8826,
      "theoretical_loss": 3.8997006256747113,
      "tokens_seen": 518389760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257071213640923,
      "loss": 2.9002,
      "theoretical_loss": 3.8996478580776293,
      "tokens_seen": 518455296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042569709127382145,
      "loss": 2.8571,
      "theoretical_loss": 3.899595099017656,
      "tokens_seen": 518520832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042568706118355064,
      "loss": 2.8265,
      "theoretical_loss": 3.8995423484923304,
      "tokens_seen": 518586368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256770310932798,
      "loss": 2.7872,
      "theoretical_loss": 3.899489606499193,
      "tokens_seen": 518651904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042566700100300905,
      "loss": 2.8207,
      "theoretical_loss": 3.8994368730357873,
      "tokens_seen": 518717440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042565697091273823,
      "loss": 2.8399,
      "theoretical_loss": 3.8993841480996565,
      "tokens_seen": 518782976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256469408224674,
      "loss": 2.8967,
      "theoretical_loss": 3.899331431688344,
      "tokens_seen": 518848512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042563691073219665,
      "loss": 2.8206,
      "theoretical_loss": 3.8992787237993944,
      "tokens_seen": 518914048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256268806419258,
      "loss": 2.6858,
      "theoretical_loss": 3.8992260244303547,
      "tokens_seen": 518979584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425616850551655,
      "loss": 2.8494,
      "theoretical_loss": 3.899173333578772,
      "tokens_seen": 519045120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042560682046138414,
      "loss": 2.7067,
      "theoretical_loss": 3.8991206512421943,
      "tokens_seen": 519110656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255967903711134,
      "loss": 2.8115,
      "theoretical_loss": 3.8990679774181713,
      "tokens_seen": 519176192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042558676028084256,
      "loss": 2.8671,
      "theoretical_loss": 3.899015312104253,
      "tokens_seen": 519241728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042557673019057174,
      "loss": 2.787,
      "theoretical_loss": 3.89896265529799,
      "tokens_seen": 519307264
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 296800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2833645343780518,
      "objective/train/theoretical_loss": 3.8989100069969354,
      "objective/train/tokens_used": 539832800,
      "theoretical_loss": 3.8989100069969354,
      "tokens_seen": 519372800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255667001003009,
      "loss": 2.8438,
      "theoretical_loss": 3.8989100069969354,
      "tokens_seen": 519372800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255566700100301,
      "loss": 2.7813,
      "theoretical_loss": 3.8988573671986417,
      "tokens_seen": 519438336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255466399197593,
      "loss": 2.6913,
      "theoretical_loss": 3.898804735900664,
      "tokens_seen": 519503872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255366098294885,
      "loss": 2.9325,
      "theoretical_loss": 3.8987521131005574,
      "tokens_seen": 519569408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042552657973921764,
      "loss": 3.0525,
      "theoretical_loss": 3.898699498795878,
      "tokens_seen": 519634944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255165496489469,
      "loss": 3.1233,
      "theoretical_loss": 3.898646892984183,
      "tokens_seen": 519700480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425506519558676,
      "loss": 2.9536,
      "theoretical_loss": 3.8985942956630306,
      "tokens_seen": 519766016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042549648946840524,
      "loss": 2.8703,
      "theoretical_loss": 3.898541706829981,
      "tokens_seen": 519831552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254864593781344,
      "loss": 3.0978,
      "theoretical_loss": 3.898489126482594,
      "tokens_seen": 519897088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254764292878636,
      "loss": 3.0565,
      "theoretical_loss": 3.8984365546184305,
      "tokens_seen": 519962624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254663991975928,
      "loss": 3.0101,
      "theoretical_loss": 3.898383991235053,
      "tokens_seen": 520028160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425456369107322,
      "loss": 3.0552,
      "theoretical_loss": 3.898331436330026,
      "tokens_seen": 520093696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042544633901705115,
      "loss": 2.6484,
      "theoretical_loss": 3.8982788899009124,
      "tokens_seen": 520159232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254363089267804,
      "loss": 2.9835,
      "theoretical_loss": 3.8982263519452793,
      "tokens_seen": 520224768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254262788365095,
      "loss": 3.0324,
      "theoretical_loss": 3.8981738224606914,
      "tokens_seen": 520290304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042541624874623874,
      "loss": 2.9913,
      "theoretical_loss": 3.898121301444716,
      "tokens_seen": 520355840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254062186559679,
      "loss": 2.8293,
      "theoretical_loss": 3.898068788894923,
      "tokens_seen": 520421376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253961885656971,
      "loss": 3.0361,
      "theoretical_loss": 3.898016284808881,
      "tokens_seen": 520486912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253861584754263,
      "loss": 2.9718,
      "theoretical_loss": 3.8979637891841605,
      "tokens_seen": 520552448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042537612838515547,
      "loss": 2.8781,
      "theoretical_loss": 3.8979113020183327,
      "tokens_seen": 520617984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042536609829488465,
      "loss": 2.9975,
      "theoretical_loss": 3.8978588233089706,
      "tokens_seen": 520683520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253560682046139,
      "loss": 2.7801,
      "theoretical_loss": 3.897806353053647,
      "tokens_seen": 520749056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425346038114343,
      "loss": 2.9144,
      "theoretical_loss": 3.897753891249936,
      "tokens_seen": 520814592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042533600802407225,
      "loss": 3.0736,
      "theoretical_loss": 3.897701437895414,
      "tokens_seen": 520880128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253259779338014,
      "loss": 2.9123,
      "theoretical_loss": 3.8976489929876568,
      "tokens_seen": 520945664
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 297883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9455604553222656,
      "objective/train/theoretical_loss": 3.8975965565242423,
      "objective/train/tokens_used": 541471200,
      "theoretical_loss": 3.8975965565242423,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253159478435306,
      "loss": 2.724,
      "theoretical_loss": 3.8975965565242423,
      "tokens_seen": 521011200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253059177532598,
      "loss": 3.0535,
      "theoretical_loss": 3.897544128502749,
      "tokens_seen": 521076736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042529588766298897,
      "loss": 2.9999,
      "theoretical_loss": 3.897491708920755,
      "tokens_seen": 521142272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042528585757271815,
      "loss": 2.9009,
      "theoretical_loss": 3.897439297775842,
      "tokens_seen": 521207808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252758274824474,
      "loss": 2.8293,
      "theoretical_loss": 3.8973868950655914,
      "tokens_seen": 521273344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252657973921765,
      "loss": 2.9852,
      "theoretical_loss": 3.8973345007875846,
      "tokens_seen": 521338880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042525576730190575,
      "loss": 2.9392,
      "theoretical_loss": 3.897282114939406,
      "tokens_seen": 521404416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252457372116349,
      "loss": 2.8227,
      "theoretical_loss": 3.89722973751864,
      "tokens_seen": 521469952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252357071213641,
      "loss": 2.7058,
      "theoretical_loss": 3.897177368522872,
      "tokens_seen": 521535488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252256770310933,
      "loss": 2.9889,
      "theoretical_loss": 3.897125007949687,
      "tokens_seen": 521601024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252156469408225,
      "loss": 2.935,
      "theoretical_loss": 3.8970726557966744,
      "tokens_seen": 521666560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042520561685055165,
      "loss": 2.9667,
      "theoretical_loss": 3.897020312061421,
      "tokens_seen": 521732096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042519558676028084,
      "loss": 2.941,
      "theoretical_loss": 3.896967976741517,
      "tokens_seen": 521797632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042518555667001,
      "loss": 2.9092,
      "theoretical_loss": 3.896915649834553,
      "tokens_seen": 521863168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042517552657973925,
      "loss": 2.6848,
      "theoretical_loss": 3.8968633313381202,
      "tokens_seen": 521928704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251654964894684,
      "loss": 2.7817,
      "theoretical_loss": 3.8968110212498104,
      "tokens_seen": 521994240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251554663991976,
      "loss": 2.7736,
      "theoretical_loss": 3.8967587195672175,
      "tokens_seen": 522059776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251454363089268,
      "loss": 2.9141,
      "theoretical_loss": 3.8967064262879356,
      "tokens_seen": 522125312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425135406218656,
      "loss": 2.6049,
      "theoretical_loss": 3.8966541414095595,
      "tokens_seen": 522190848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042512537612838516,
      "loss": 2.6158,
      "theoretical_loss": 3.8966018649296865,
      "tokens_seen": 522256384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042511534603811434,
      "loss": 2.7402,
      "theoretical_loss": 3.8965495968459134,
      "tokens_seen": 522321920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251053159478435,
      "loss": 3.1019,
      "theoretical_loss": 3.896497337155839,
      "tokens_seen": 522387456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042509528585757276,
      "loss": 2.9382,
      "theoretical_loss": 3.896445085857062,
      "tokens_seen": 522452992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250852557673019,
      "loss": 2.9426,
      "theoretical_loss": 3.896392842947183,
      "tokens_seen": 522518528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250752256770311,
      "loss": 2.6713,
      "theoretical_loss": 3.8963406084238024,
      "tokens_seen": 522584064
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 298609,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9465911388397217,
      "objective/train/theoretical_loss": 3.8962883822845233,
      "objective/train/tokens_used": 543109600,
      "theoretical_loss": 3.8962883822845233,
      "tokens_seen": 522649600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042506519558676024,
      "loss": 2.7065,
      "theoretical_loss": 3.8962883822845233,
      "tokens_seen": 522649600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250551654964895,
      "loss": 2.9139,
      "theoretical_loss": 3.8962361645269494,
      "tokens_seen": 522715136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042504513540621866,
      "loss": 2.9572,
      "theoretical_loss": 3.8961839551486843,
      "tokens_seen": 522780672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042503510531594784,
      "loss": 2.8489,
      "theoretical_loss": 3.8961317541473326,
      "tokens_seen": 522846208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425025075225677,
      "loss": 3.0263,
      "theoretical_loss": 3.8960795615205015,
      "tokens_seen": 522911744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250150451354062,
      "loss": 3.0704,
      "theoretical_loss": 3.896027377265798,
      "tokens_seen": 522977280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250050150451354,
      "loss": 2.9052,
      "theoretical_loss": 3.8959752013808293,
      "tokens_seen": 523042816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249949849548646,
      "loss": 2.916,
      "theoretical_loss": 3.895923033863206,
      "tokens_seen": 523108352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042498495486459375,
      "loss": 2.955,
      "theoretical_loss": 3.8958708747105373,
      "tokens_seen": 523173888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424974924774323,
      "loss": 2.8302,
      "theoretical_loss": 3.895818723920434,
      "tokens_seen": 523239424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042496489468405216,
      "loss": 2.7975,
      "theoretical_loss": 3.895766581490509,
      "tokens_seen": 523304960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042495486459378135,
      "loss": 2.8087,
      "theoretical_loss": 3.895714447418375,
      "tokens_seen": 523370496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249448345035105,
      "loss": 2.9109,
      "theoretical_loss": 3.8956623217016464,
      "tokens_seen": 523436032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249348044132397,
      "loss": 2.8087,
      "theoretical_loss": 3.895610204337938,
      "tokens_seen": 523501568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042492477432296894,
      "loss": 2.8582,
      "theoretical_loss": 3.8955580953248656,
      "tokens_seen": 523567104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249147442326981,
      "loss": 2.8843,
      "theoretical_loss": 3.895505994660046,
      "tokens_seen": 523632640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249047141424273,
      "loss": 3.1563,
      "theoretical_loss": 3.8954539023410972,
      "tokens_seen": 523698176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248946840521565,
      "loss": 2.9491,
      "theoretical_loss": 3.895401818365639,
      "tokens_seen": 523763712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042488465396188567,
      "loss": 2.9519,
      "theoretical_loss": 3.89534974273129,
      "tokens_seen": 523829248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042487462387161485,
      "loss": 2.5973,
      "theoretical_loss": 3.895297675435673,
      "tokens_seen": 523894784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248645937813441,
      "loss": 2.8273,
      "theoretical_loss": 3.8952456164764078,
      "tokens_seen": 523960320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248545636910732,
      "loss": 2.7711,
      "theoretical_loss": 3.8951935658511183,
      "tokens_seen": 524025856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042484453360080245,
      "loss": 2.9623,
      "theoretical_loss": 3.895141523557428,
      "tokens_seen": 524091392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248345035105316,
      "loss": 2.7556,
      "theoretical_loss": 3.895089489592962,
      "tokens_seen": 524156928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248244734202608,
      "loss": 2.907,
      "theoretical_loss": 3.8950374639553456,
      "tokens_seen": 524222464
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 299683,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.230189085006714,
      "objective/train/theoretical_loss": 3.894985446642206,
      "objective/train/tokens_used": 544748000,
      "theoretical_loss": 3.894985446642206,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042481444332999,
      "loss": 2.674,
      "theoretical_loss": 3.894985446642206,
      "tokens_seen": 524288000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042480441323971917,
      "loss": 2.7178,
      "theoretical_loss": 3.8949334376511713,
      "tokens_seen": 524353536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042479438314944835,
      "loss": 2.6602,
      "theoretical_loss": 3.894881436979869,
      "tokens_seen": 524419072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247843530591776,
      "loss": 2.881,
      "theoretical_loss": 3.8948294446259295,
      "tokens_seen": 524484608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247743229689067,
      "loss": 3.029,
      "theoretical_loss": 3.894777460586983,
      "tokens_seen": 524550144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042476429287863595,
      "loss": 3.0115,
      "theoretical_loss": 3.894725484860662,
      "tokens_seen": 524615680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247542627883651,
      "loss": 2.9568,
      "theoretical_loss": 3.8946735174445983,
      "tokens_seen": 524681216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247442326980943,
      "loss": 2.6509,
      "theoretical_loss": 3.894621558336426,
      "tokens_seen": 524746752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247342026078235,
      "loss": 2.7833,
      "theoretical_loss": 3.8945696075337786,
      "tokens_seen": 524812288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247241725175527,
      "loss": 2.8974,
      "theoretical_loss": 3.8945176650342925,
      "tokens_seen": 524877824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042471414242728186,
      "loss": 3.0308,
      "theoretical_loss": 3.894465730835604,
      "tokens_seen": 524943360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042470411233701104,
      "loss": 2.8243,
      "theoretical_loss": 3.8944138049353505,
      "tokens_seen": 525008896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246940822467402,
      "loss": 2.8344,
      "theoretical_loss": 3.8943618873311703,
      "tokens_seen": 525074432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042468405215646945,
      "loss": 2.6421,
      "theoretical_loss": 3.8943099780207024,
      "tokens_seen": 525139968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246740220661986,
      "loss": 3.2505,
      "theoretical_loss": 3.894258077001587,
      "tokens_seen": 525205504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246639919759278,
      "loss": 2.9114,
      "theoretical_loss": 3.8942061842714666,
      "tokens_seen": 525271040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424653961885657,
      "loss": 2.9455,
      "theoretical_loss": 3.8941542998279823,
      "tokens_seen": 525336576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246439317953862,
      "loss": 2.7369,
      "theoretical_loss": 3.894102423668778,
      "tokens_seen": 525402112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042463390170511536,
      "loss": 2.879,
      "theoretical_loss": 3.8940505557914973,
      "tokens_seen": 525467648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042462387161484454,
      "loss": 2.9338,
      "theoretical_loss": 3.893998696193785,
      "tokens_seen": 525533184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246138415245737,
      "loss": 2.7185,
      "theoretical_loss": 3.8939468448732892,
      "tokens_seen": 525598720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042460381143430296,
      "loss": 2.723,
      "theoretical_loss": 3.893895001827654,
      "tokens_seen": 525664256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245937813440321,
      "loss": 2.8829,
      "theoretical_loss": 3.89384316705453,
      "tokens_seen": 525729792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245837512537613,
      "loss": 2.8745,
      "theoretical_loss": 3.893791340551565,
      "tokens_seen": 525795328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042457372116349044,
      "loss": 2.9526,
      "theoretical_loss": 3.8937395223164093,
      "tokens_seen": 525860864
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 300593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.640450954437256,
      "objective/train/theoretical_loss": 3.893687712346713,
      "objective/train/tokens_used": 546386400,
      "theoretical_loss": 3.893687712346713,
      "tokens_seen": 525926400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245636910732197,
      "loss": 3.066,
      "theoretical_loss": 3.893687712346713,
      "tokens_seen": 525926400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042455366098294886,
      "loss": 2.814,
      "theoretical_loss": 3.8936359106401293,
      "tokens_seen": 525991936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042454363089267804,
      "loss": 2.7216,
      "theoretical_loss": 3.8935841171943095,
      "tokens_seen": 526057472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245336008024072,
      "loss": 2.7412,
      "theoretical_loss": 3.893532332006909,
      "tokens_seen": 526123008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245235707121364,
      "loss": 2.9448,
      "theoretical_loss": 3.8934805550755813,
      "tokens_seen": 526188544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245135406218656,
      "loss": 2.7285,
      "theoretical_loss": 3.893428786397983,
      "tokens_seen": 526254080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245035105315948,
      "loss": 2.7658,
      "theoretical_loss": 3.8933770259717697,
      "tokens_seen": 526319616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042449348044132395,
      "loss": 3.0131,
      "theoretical_loss": 3.8933252737946002,
      "tokens_seen": 526385152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244834503510532,
      "loss": 2.7621,
      "theoretical_loss": 3.893273529864132,
      "tokens_seen": 526450688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042447342026078236,
      "loss": 2.893,
      "theoretical_loss": 3.893221794178025,
      "tokens_seen": 526516224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042446339017051155,
      "loss": 2.8413,
      "theoretical_loss": 3.893170066733941,
      "tokens_seen": 526581760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244533600802407,
      "loss": 2.6966,
      "theoretical_loss": 3.8931183475295397,
      "tokens_seen": 526647296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244433299899699,
      "loss": 2.6659,
      "theoretical_loss": 3.8930666365624838,
      "tokens_seen": 526712832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244332998996991,
      "loss": 2.8686,
      "theoretical_loss": 3.893014933830437,
      "tokens_seen": 526778368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244232698094283,
      "loss": 2.7931,
      "theoretical_loss": 3.8929632393310634,
      "tokens_seen": 526843904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042441323971915745,
      "loss": 2.9316,
      "theoretical_loss": 3.892911553062029,
      "tokens_seen": 526909440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244032096288867,
      "loss": 2.7593,
      "theoretical_loss": 3.892859875020999,
      "tokens_seen": 526974976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243931795386158,
      "loss": 2.6133,
      "theoretical_loss": 3.8928082052056405,
      "tokens_seen": 527040512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042438314944834505,
      "loss": 3.0805,
      "theoretical_loss": 3.892756543613623,
      "tokens_seen": 527106048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042437311935807423,
      "loss": 3.0741,
      "theoretical_loss": 3.892704890242614,
      "tokens_seen": 527171584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243630892678034,
      "loss": 2.9281,
      "theoretical_loss": 3.892653245090284,
      "tokens_seen": 527237120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243530591775326,
      "loss": 2.9436,
      "theoretical_loss": 3.8926016081543047,
      "tokens_seen": 527302656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243430290872618,
      "loss": 2.8254,
      "theoretical_loss": 3.892549979432347,
      "tokens_seen": 527368192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042433299899699095,
      "loss": 2.8296,
      "theoretical_loss": 3.8924983589220843,
      "tokens_seen": 527433728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243229689067202,
      "loss": 2.8305,
      "theoretical_loss": 3.8924467466211903,
      "tokens_seen": 527499264
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 301147,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.089402198791504,
      "objective/train/theoretical_loss": 3.8923951425273398,
      "objective/train/tokens_used": 548024800,
      "theoretical_loss": 3.8923951425273398,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243129388164493,
      "loss": 2.883,
      "theoretical_loss": 3.8923951425273398,
      "tokens_seen": 527564800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042430290872617855,
      "loss": 2.8607,
      "theoretical_loss": 3.8923435466382084,
      "tokens_seen": 527630336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042429287863590773,
      "loss": 2.9292,
      "theoretical_loss": 3.892291958951473,
      "tokens_seen": 527695872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242828485456369,
      "loss": 2.8227,
      "theoretical_loss": 3.892240379464811,
      "tokens_seen": 527761408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242728184553661,
      "loss": 2.941,
      "theoretical_loss": 3.8921888081759004,
      "tokens_seen": 527826944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242627883650953,
      "loss": 2.7021,
      "theoretical_loss": 3.892137245082422,
      "tokens_seen": 527892480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042425275827482446,
      "loss": 2.98,
      "theoretical_loss": 3.8920856901820553,
      "tokens_seen": 527958016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242427281845537,
      "loss": 2.8898,
      "theoretical_loss": 3.8920341434724817,
      "tokens_seen": 528023552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242326980942828,
      "loss": 2.8186,
      "theoretical_loss": 3.8919826049513837,
      "tokens_seen": 528089088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042422266800401206,
      "loss": 2.9792,
      "theoretical_loss": 3.891931074616444,
      "tokens_seen": 528154624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242126379137412,
      "loss": 3.1397,
      "theoretical_loss": 3.8918795524653484,
      "tokens_seen": 528220160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242026078234704,
      "loss": 2.5991,
      "theoretical_loss": 3.89182803849578,
      "tokens_seen": 528285696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241925777331996,
      "loss": 2.9587,
      "theoretical_loss": 3.8917765327054266,
      "tokens_seen": 528351232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241825476429288,
      "loss": 2.8182,
      "theoretical_loss": 3.8917250350919748,
      "tokens_seen": 528416768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424172517552658,
      "loss": 2.8705,
      "theoretical_loss": 3.891673545653112,
      "tokens_seen": 528482304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241624874623872,
      "loss": 2.6681,
      "theoretical_loss": 3.891622064386527,
      "tokens_seen": 528547840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241524573721164,
      "loss": 2.9959,
      "theoretical_loss": 3.89157059128991,
      "tokens_seen": 528613376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042414242728184556,
      "loss": 2.9152,
      "theoretical_loss": 3.8915191263609525,
      "tokens_seen": 528678912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042413239719157474,
      "loss": 2.9309,
      "theoretical_loss": 3.8914676695973456,
      "tokens_seen": 528744448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241223671013039,
      "loss": 2.8499,
      "theoretical_loss": 3.8914162209967813,
      "tokens_seen": 528809984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042411233701103316,
      "loss": 2.915,
      "theoretical_loss": 3.8913647805569545,
      "tokens_seen": 528875520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241023069207623,
      "loss": 2.9174,
      "theoretical_loss": 3.891313348275559,
      "tokens_seen": 528941056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240922768304915,
      "loss": 2.9057,
      "theoretical_loss": 3.8912619241502906,
      "tokens_seen": 529006592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042408224674022065,
      "loss": 2.674,
      "theoretical_loss": 3.891210508178845,
      "tokens_seen": 529072128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240722166499499,
      "loss": 2.7977,
      "theoretical_loss": 3.8911591003589203,
      "tokens_seen": 529137664
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 301749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2091262340545654,
      "objective/train/theoretical_loss": 3.891107700688215,
      "objective/train/tokens_used": 549663200,
      "theoretical_loss": 3.891107700688215,
      "tokens_seen": 529203200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042406218655967906,
      "loss": 2.9569,
      "theoretical_loss": 3.891107700688215,
      "tokens_seen": 529203200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042405215646940824,
      "loss": 2.9637,
      "theoretical_loss": 3.8910563091644277,
      "tokens_seen": 529268736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240421263791374,
      "loss": 2.8253,
      "theoretical_loss": 3.891004925785259,
      "tokens_seen": 529334272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240320962888666,
      "loss": 2.7514,
      "theoretical_loss": 3.890953550548409,
      "tokens_seen": 529399808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240220661985958,
      "loss": 2.8764,
      "theoretical_loss": 3.8909021834515816,
      "tokens_seen": 529465344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424012036108325,
      "loss": 2.8901,
      "theoretical_loss": 3.8908508244924778,
      "tokens_seen": 529530880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042400200601805415,
      "loss": 2.7382,
      "theoretical_loss": 3.8907994736688027,
      "tokens_seen": 529596416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239919759277834,
      "loss": 2.7764,
      "theoretical_loss": 3.8907481309782606,
      "tokens_seen": 529661952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042398194583751256,
      "loss": 3.0124,
      "theoretical_loss": 3.890696796418558,
      "tokens_seen": 529727488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042397191574724175,
      "loss": 2.912,
      "theoretical_loss": 3.8906454699874002,
      "tokens_seen": 529793024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239618856569709,
      "loss": 2.8759,
      "theoretical_loss": 3.890594151682496,
      "tokens_seen": 529858560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239518555667001,
      "loss": 2.779,
      "theoretical_loss": 3.8905428415015537,
      "tokens_seen": 529924096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239418254764293,
      "loss": 2.8285,
      "theoretical_loss": 3.8904915394422828,
      "tokens_seen": 529989632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239317953861585,
      "loss": 2.7266,
      "theoretical_loss": 3.8904402455023934,
      "tokens_seen": 530055168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042392176529588765,
      "loss": 2.9637,
      "theoretical_loss": 3.890388959679597,
      "tokens_seen": 530120704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004239117352056169,
      "loss": 2.7824,
      "theoretical_loss": 3.8903376819716056,
      "tokens_seen": 530186240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423901705115346,
      "loss": 2.9065,
      "theoretical_loss": 3.890286412376133,
      "tokens_seen": 530251776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042389167502507525,
      "loss": 2.7732,
      "theoretical_loss": 3.8902351508908923,
      "tokens_seen": 530317312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042388164493480443,
      "loss": 2.7168,
      "theoretical_loss": 3.8901838975136,
      "tokens_seen": 530382848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238716148445336,
      "loss": 2.9481,
      "theoretical_loss": 3.890132652241971,
      "tokens_seen": 530448384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238615847542628,
      "loss": 2.4908,
      "theoretical_loss": 3.890081415073723,
      "tokens_seen": 530513920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423851554663992,
      "loss": 3.0909,
      "theoretical_loss": 3.8900301860065727,
      "tokens_seen": 530579456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042384152457372115,
      "loss": 2.7948,
      "theoretical_loss": 3.8899789650382397,
      "tokens_seen": 530644992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238314944834504,
      "loss": 2.9124,
      "theoretical_loss": 3.889927752166443,
      "tokens_seen": 530710528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238214643931795,
      "loss": 2.7639,
      "theoretical_loss": 3.889876547388904,
      "tokens_seen": 530776064
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 302930,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.119241714477539,
      "objective/train/theoretical_loss": 3.889825350703344,
      "objective/train/tokens_used": 551301600,
      "theoretical_loss": 3.889825350703344,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042381143430290875,
      "loss": 2.8456,
      "theoretical_loss": 3.889825350703344,
      "tokens_seen": 530841600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042380140421263793,
      "loss": 3.1018,
      "theoretical_loss": 3.889774162107485,
      "tokens_seen": 530907136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237913741223671,
      "loss": 2.637,
      "theoretical_loss": 3.889722981599051,
      "tokens_seen": 530972672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237813440320963,
      "loss": 2.8671,
      "theoretical_loss": 3.8896718091757654,
      "tokens_seen": 531038208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237713139418255,
      "loss": 2.8487,
      "theoretical_loss": 3.889620644835354,
      "tokens_seen": 531103744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042376128385155466,
      "loss": 2.7057,
      "theoretical_loss": 3.889569488575543,
      "tokens_seen": 531169280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237512537612839,
      "loss": 2.6746,
      "theoretical_loss": 3.8895183403940594,
      "tokens_seen": 531234816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423741223671013,
      "loss": 2.564,
      "theoretical_loss": 3.889467200288631,
      "tokens_seen": 531300352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042373119358074226,
      "loss": 2.684,
      "theoretical_loss": 3.889416068256986,
      "tokens_seen": 531365888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237211634904714,
      "loss": 2.7755,
      "theoretical_loss": 3.8893649442968554,
      "tokens_seen": 531431424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237111334002006,
      "loss": 2.9413,
      "theoretical_loss": 3.8893138284059696,
      "tokens_seen": 531496960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237011033099298,
      "loss": 3.0095,
      "theoretical_loss": 3.8892627205820602,
      "tokens_seen": 531562496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423691073219659,
      "loss": 2.9443,
      "theoretical_loss": 3.8892116208228593,
      "tokens_seen": 531628032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042368104312938816,
      "loss": 2.7001,
      "theoretical_loss": 3.8891605291261007,
      "tokens_seen": 531693568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236710130391174,
      "loss": 2.7396,
      "theoretical_loss": 3.889109445489519,
      "tokens_seen": 531759104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236609829488465,
      "loss": 2.8817,
      "theoretical_loss": 3.8890583699108494,
      "tokens_seen": 531824640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042365095285857576,
      "loss": 3.0576,
      "theoretical_loss": 3.889007302387828,
      "tokens_seen": 531890176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236409227683049,
      "loss": 2.9742,
      "theoretical_loss": 3.8889562429181916,
      "tokens_seen": 531955712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236308926780341,
      "loss": 2.8414,
      "theoretical_loss": 3.8889051914996786,
      "tokens_seen": 532021248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236208625877633,
      "loss": 2.9494,
      "theoretical_loss": 3.888854148130028,
      "tokens_seen": 532086784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236108324974925,
      "loss": 2.9868,
      "theoretical_loss": 3.88880311280698,
      "tokens_seen": 532152320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042360080240722166,
      "loss": 2.7983,
      "theoretical_loss": 3.8887520855282744,
      "tokens_seen": 532217856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042359077231695085,
      "loss": 2.9949,
      "theoretical_loss": 3.888701066291654,
      "tokens_seen": 532283392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042358074222668,
      "loss": 2.7991,
      "theoretical_loss": 3.888650055094861,
      "tokens_seen": 532348928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042357071213640926,
      "loss": 2.8885,
      "theoretical_loss": 3.888599051935638,
      "tokens_seen": 532414464
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 303466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.733795404434204,
      "objective/train/theoretical_loss": 3.8885480568117314,
      "objective/train/tokens_used": 552940000,
      "theoretical_loss": 3.8885480568117314,
      "tokens_seen": 532480000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235606820461384,
      "loss": 2.8752,
      "theoretical_loss": 3.8885480568117314,
      "tokens_seen": 532480000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235506519558676,
      "loss": 2.9894,
      "theoretical_loss": 3.888497069720885,
      "tokens_seen": 532545536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042354062186559675,
      "loss": 2.7836,
      "theoretical_loss": 3.888446090660845,
      "tokens_seen": 532611072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423530591775326,
      "loss": 2.7741,
      "theoretical_loss": 3.88839511962936,
      "tokens_seen": 532676608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042352056168505517,
      "loss": 3.0174,
      "theoretical_loss": 3.8883441566241763,
      "tokens_seen": 532742144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042351053159478435,
      "loss": 2.9035,
      "theoretical_loss": 3.8882932016430436,
      "tokens_seen": 532807680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042350050150451353,
      "loss": 2.735,
      "theoretical_loss": 3.8882422546837123,
      "tokens_seen": 532873216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042349047141424276,
      "loss": 2.9197,
      "theoretical_loss": 3.888191315743933,
      "tokens_seen": 532938752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234804413239719,
      "loss": 2.8019,
      "theoretical_loss": 3.888140384821457,
      "tokens_seen": 533004288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042347041123370113,
      "loss": 2.7563,
      "theoretical_loss": 3.8880894619140367,
      "tokens_seen": 533069824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042346038114343025,
      "loss": 3.0713,
      "theoretical_loss": 3.888038547019427,
      "tokens_seen": 533135360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234503510531595,
      "loss": 2.9346,
      "theoretical_loss": 3.88798764013538,
      "tokens_seen": 533200896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042344032096288867,
      "loss": 2.9696,
      "theoretical_loss": 3.8879367412596535,
      "tokens_seen": 533266432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042343029087261785,
      "loss": 2.8591,
      "theoretical_loss": 3.887885850390002,
      "tokens_seen": 533331968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234202607823471,
      "loss": 2.7303,
      "theoretical_loss": 3.8878349675241832,
      "tokens_seen": 533397504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234102306920762,
      "loss": 2.936,
      "theoretical_loss": 3.8877840926599556,
      "tokens_seen": 533463040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042340020060180545,
      "loss": 2.6803,
      "theoretical_loss": 3.8877332257950776,
      "tokens_seen": 533528576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042339017051153463,
      "loss": 2.8323,
      "theoretical_loss": 3.887682366927309,
      "tokens_seen": 533594112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233801404212638,
      "loss": 2.9108,
      "theoretical_loss": 3.8876315160544106,
      "tokens_seen": 533659648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423370110330993,
      "loss": 2.9117,
      "theoretical_loss": 3.8875806731741447,
      "tokens_seen": 533725184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233600802407222,
      "loss": 2.8684,
      "theoretical_loss": 3.887529838284273,
      "tokens_seen": 533790720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042335005015045135,
      "loss": 2.7934,
      "theoretical_loss": 3.887479011382559,
      "tokens_seen": 533856256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233400200601806,
      "loss": 2.6989,
      "theoretical_loss": 3.8874281924667677,
      "tokens_seen": 533921792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233299899699097,
      "loss": 2.6725,
      "theoretical_loss": 3.887377381534664,
      "tokens_seen": 533987328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042331995987963895,
      "loss": 2.764,
      "theoretical_loss": 3.8873265785840134,
      "tokens_seen": 534052864
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 304686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.990126848220825,
      "objective/train/theoretical_loss": 3.8872757836125844,
      "objective/train/tokens_used": 554578400,
      "theoretical_loss": 3.8872757836125844,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042330992978936813,
      "loss": 2.8816,
      "theoretical_loss": 3.8872757836125844,
      "tokens_seen": 534118400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232998996990973,
      "loss": 2.7513,
      "theoretical_loss": 3.8872249966181442,
      "tokens_seen": 534183936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232898696088265,
      "loss": 2.9891,
      "theoretical_loss": 3.887174217598461,
      "tokens_seen": 534249472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232798395185557,
      "loss": 2.7367,
      "theoretical_loss": 3.8871234465513056,
      "tokens_seen": 534315008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042326980942828486,
      "loss": 3.0058,
      "theoretical_loss": 3.8870726834744485,
      "tokens_seen": 534380544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232597793380141,
      "loss": 2.7603,
      "theoretical_loss": 3.8870219283656597,
      "tokens_seen": 534446080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232497492477432,
      "loss": 2.8634,
      "theoretical_loss": 3.886971181222714,
      "tokens_seen": 534511616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042323971915747246,
      "loss": 2.8499,
      "theoretical_loss": 3.886920442043383,
      "tokens_seen": 534577152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232296890672016,
      "loss": 2.9109,
      "theoretical_loss": 3.886869710825442,
      "tokens_seen": 534642688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232196589769308,
      "loss": 2.884,
      "theoretical_loss": 3.886818987566665,
      "tokens_seen": 534708224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042320962888666,
      "loss": 2.5804,
      "theoretical_loss": 3.8867682722648293,
      "tokens_seen": 534773760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231995987963892,
      "loss": 2.7504,
      "theoretical_loss": 3.886717564917711,
      "tokens_seen": 534839296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042318956870611836,
      "loss": 2.9655,
      "theoretical_loss": 3.886666865523088,
      "tokens_seen": 534904832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231795386158476,
      "loss": 2.7419,
      "theoretical_loss": 3.886616174078739,
      "tokens_seen": 534970368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231695085255767,
      "loss": 2.7549,
      "theoretical_loss": 3.886565490582444,
      "tokens_seen": 535035904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042315947843530596,
      "loss": 2.9635,
      "theoretical_loss": 3.8865148150319824,
      "tokens_seen": 535101440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231494483450351,
      "loss": 3.0665,
      "theoretical_loss": 3.886464147425137,
      "tokens_seen": 535166976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231394182547643,
      "loss": 2.8279,
      "theoretical_loss": 3.8864134877596896,
      "tokens_seen": 535232512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231293881644935,
      "loss": 2.7782,
      "theoretical_loss": 3.886362836033423,
      "tokens_seen": 535298048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231193580742227,
      "loss": 2.3214,
      "theoretical_loss": 3.8863121922441213,
      "tokens_seen": 535363584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042310932798395186,
      "loss": 2.8828,
      "theoretical_loss": 3.8862615563895693,
      "tokens_seen": 535429120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042309929789368105,
      "loss": 2.8513,
      "theoretical_loss": 3.886210928467553,
      "tokens_seen": 535494656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230892678034102,
      "loss": 2.9486,
      "theoretical_loss": 3.8861603084758594,
      "tokens_seen": 535560192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042307923771313946,
      "loss": 2.845,
      "theoretical_loss": 3.8861096964122757,
      "tokens_seen": 535625728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230692076228686,
      "loss": 2.7738,
      "theoretical_loss": 3.886059092274591,
      "tokens_seen": 535691264
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 305312,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0122928619384766,
      "objective/train/theoretical_loss": 3.886008496060594,
      "objective/train/tokens_used": 556216800,
      "theoretical_loss": 3.886008496060594,
      "tokens_seen": 535756800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230591775325978,
      "loss": 2.9321,
      "theoretical_loss": 3.886008496060594,
      "tokens_seen": 535756800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042304914744232695,
      "loss": 2.8908,
      "theoretical_loss": 3.8859579077680753,
      "tokens_seen": 535822336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230391173520562,
      "loss": 2.6939,
      "theoretical_loss": 3.8859073273948255,
      "tokens_seen": 535887872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042302908726178537,
      "loss": 2.8025,
      "theoretical_loss": 3.8858567549386374,
      "tokens_seen": 535953408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042301905717151455,
      "loss": 2.9458,
      "theoretical_loss": 3.885806190397304,
      "tokens_seen": 536018944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042300902708124373,
      "loss": 2.6714,
      "theoretical_loss": 3.8857556337686177,
      "tokens_seen": 536084480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042299899699097297,
      "loss": 2.8947,
      "theoretical_loss": 3.8857050850503754,
      "tokens_seen": 536150016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229889669007021,
      "loss": 2.9313,
      "theoretical_loss": 3.8856545442403707,
      "tokens_seen": 536215552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042297893681043133,
      "loss": 2.8646,
      "theoretical_loss": 3.885604011336401,
      "tokens_seen": 536281088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042296890672016045,
      "loss": 2.829,
      "theoretical_loss": 3.8855534863362635,
      "tokens_seen": 536346624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229588766298897,
      "loss": 2.9151,
      "theoretical_loss": 3.8855029692377565,
      "tokens_seen": 536412160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042294884653961887,
      "loss": 3.0059,
      "theoretical_loss": 3.885452460038679,
      "tokens_seen": 536477696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042293881644934805,
      "loss": 2.7475,
      "theoretical_loss": 3.8854019587368307,
      "tokens_seen": 536543232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042292878635907723,
      "loss": 2.8329,
      "theoretical_loss": 3.8853514653300136,
      "tokens_seen": 536608768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229187562688064,
      "loss": 2.7283,
      "theoretical_loss": 3.8853009798160274,
      "tokens_seen": 536674304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229087261785356,
      "loss": 2.659,
      "theoretical_loss": 3.885250502192677,
      "tokens_seen": 536739840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042289869608826483,
      "loss": 2.8879,
      "theoretical_loss": 3.8852000324577647,
      "tokens_seen": 536805376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042288866599799396,
      "loss": 2.8903,
      "theoretical_loss": 3.8851495706090944,
      "tokens_seen": 536870912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228786359077232,
      "loss": 2.5484,
      "theoretical_loss": 3.8850991166444726,
      "tokens_seen": 536936448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228686058174523,
      "loss": 2.8039,
      "theoretical_loss": 3.885048670561705,
      "tokens_seen": 537001984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042285857572718155,
      "loss": 2.8749,
      "theoretical_loss": 3.8849982323585985,
      "tokens_seen": 537067520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042284854563691074,
      "loss": 2.8962,
      "theoretical_loss": 3.8849478020329604,
      "tokens_seen": 537133056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228385155466399,
      "loss": 2.9466,
      "theoretical_loss": 3.8848973795826005,
      "tokens_seen": 537198592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228284854563691,
      "loss": 2.796,
      "theoretical_loss": 3.884846965005328,
      "tokens_seen": 537264128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042281845536609833,
      "loss": 2.6911,
      "theoretical_loss": 3.8847965582989534,
      "tokens_seen": 537329664
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 306391,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.048434257507324,
      "objective/train/theoretical_loss": 3.8847461594612884,
      "objective/train/tokens_used": 557855200,
      "theoretical_loss": 3.8847461594612884,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042280842527582746,
      "loss": 2.8548,
      "theoretical_loss": 3.8847461594612884,
      "tokens_seen": 537395200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227983951855567,
      "loss": 2.8251,
      "theoretical_loss": 3.884695768490145,
      "tokens_seen": 537460736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227883650952858,
      "loss": 2.8225,
      "theoretical_loss": 3.8846453853833367,
      "tokens_seen": 537526272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042277833500501506,
      "loss": 2.9452,
      "theoretical_loss": 3.8845950101386766,
      "tokens_seen": 537591808
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042276830491474424,
      "loss": 2.7989,
      "theoretical_loss": 3.8845446427539807,
      "tokens_seen": 537657344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227582748244734,
      "loss": 2.6692,
      "theoretical_loss": 3.884494283227064,
      "tokens_seen": 537722880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227482447342026,
      "loss": 2.9536,
      "theoretical_loss": 3.884443931555744,
      "tokens_seen": 537788416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227382146439318,
      "loss": 3.0258,
      "theoretical_loss": 3.884393587737838,
      "tokens_seen": 537853952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042272818455366096,
      "loss": 2.7838,
      "theoretical_loss": 3.8843432517711634,
      "tokens_seen": 537919488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227181544633902,
      "loss": 2.7952,
      "theoretical_loss": 3.8842929236535406,
      "tokens_seen": 537985024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227081243731193,
      "loss": 3.0073,
      "theoretical_loss": 3.8842426033827895,
      "tokens_seen": 538050560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042269809428284856,
      "loss": 2.7916,
      "theoretical_loss": 3.8841922909567312,
      "tokens_seen": 538116096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226880641925777,
      "loss": 2.8229,
      "theoretical_loss": 3.8841419863731863,
      "tokens_seen": 538181632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226780341023069,
      "loss": 2.6835,
      "theoretical_loss": 3.8840916896299795,
      "tokens_seen": 538247168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042266800401203616,
      "loss": 2.8953,
      "theoretical_loss": 3.8840414007249335,
      "tokens_seen": 538312704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226579739217653,
      "loss": 2.7475,
      "theoretical_loss": 3.8839911196558727,
      "tokens_seen": 538378240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226479438314945,
      "loss": 2.9781,
      "theoretical_loss": 3.8839408464206224,
      "tokens_seen": 538443776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226379137412237,
      "loss": 3.0243,
      "theoretical_loss": 3.8838905810170097,
      "tokens_seen": 538509312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226278836509529,
      "loss": 3.0065,
      "theoretical_loss": 3.8838403234428602,
      "tokens_seen": 538574848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042261785356068206,
      "loss": 2.9446,
      "theoretical_loss": 3.8837900736960034,
      "tokens_seen": 538640384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042260782347041125,
      "loss": 2.8828,
      "theoretical_loss": 3.8837398317742666,
      "tokens_seen": 538705920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225977933801404,
      "loss": 2.9098,
      "theoretical_loss": 3.883689597675481,
      "tokens_seen": 538771456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042258776328986966,
      "loss": 2.9336,
      "theoretical_loss": 3.8836393713974764,
      "tokens_seen": 538836992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225777331995988,
      "loss": 3.1058,
      "theoretical_loss": 3.883589152938084,
      "tokens_seen": 538902528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422567703109328,
      "loss": 2.8419,
      "theoretical_loss": 3.8835389422951367,
      "tokens_seen": 538968064
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 307598,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.633847236633301,
      "objective/train/theoretical_loss": 3.883488739466467,
      "objective/train/tokens_used": 559493600,
      "theoretical_loss": 3.883488739466467,
      "tokens_seen": 539033600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042255767301905715,
      "loss": 2.788,
      "theoretical_loss": 3.883488739466467,
      "tokens_seen": 539033600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225476429287864,
      "loss": 2.76,
      "theoretical_loss": 3.8834385444499095,
      "tokens_seen": 539099136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042253761283851557,
      "loss": 2.8685,
      "theoretical_loss": 3.883388357243299,
      "tokens_seen": 539164672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042252758274824475,
      "loss": 2.7135,
      "theoretical_loss": 3.883338177844471,
      "tokens_seen": 539230208
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042251755265797393,
      "loss": 2.9524,
      "theoretical_loss": 3.883288006251262,
      "tokens_seen": 539295744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042250752256770317,
      "loss": 2.871,
      "theoretical_loss": 3.88323784246151,
      "tokens_seen": 539361280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224974924774323,
      "loss": 2.7594,
      "theoretical_loss": 3.883187686473053,
      "tokens_seen": 539426816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042248746238716153,
      "loss": 2.8279,
      "theoretical_loss": 3.8831375382837305,
      "tokens_seen": 539492352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042247743229689065,
      "loss": 2.7441,
      "theoretical_loss": 3.883087397891382,
      "tokens_seen": 539557888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224674022066199,
      "loss": 3.0017,
      "theoretical_loss": 3.883037265293849,
      "tokens_seen": 539623424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042245737211634907,
      "loss": 2.7708,
      "theoretical_loss": 3.8829871404889724,
      "tokens_seen": 539688960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042244734202607825,
      "loss": 2.9023,
      "theoretical_loss": 3.8829370234745957,
      "tokens_seen": 539754496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042243731193580743,
      "loss": 2.7448,
      "theoretical_loss": 3.8828869142485622,
      "tokens_seen": 539820032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224272818455366,
      "loss": 3.064,
      "theoretical_loss": 3.8828368128087165,
      "tokens_seen": 539885568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224172517552658,
      "loss": 2.834,
      "theoretical_loss": 3.882786719152903,
      "tokens_seen": 539951104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042240722166499503,
      "loss": 2.7914,
      "theoretical_loss": 3.8827366332789683,
      "tokens_seen": 540016640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042239719157472416,
      "loss": 2.7571,
      "theoretical_loss": 3.882686555184759,
      "tokens_seen": 540082176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223871614844534,
      "loss": 2.812,
      "theoretical_loss": 3.882636484868124,
      "tokens_seen": 540147712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223771313941825,
      "loss": 3.0633,
      "theoretical_loss": 3.882586422326911,
      "tokens_seen": 540213248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042236710130391176,
      "loss": 2.8002,
      "theoretical_loss": 3.882536367558969,
      "tokens_seen": 540278784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042235707121364094,
      "loss": 2.7051,
      "theoretical_loss": 3.8824863205621494,
      "tokens_seen": 540344320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223470411233701,
      "loss": 2.7578,
      "theoretical_loss": 3.882436281334303,
      "tokens_seen": 540409856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223370110330993,
      "loss": 2.8951,
      "theoretical_loss": 3.8823862498732815,
      "tokens_seen": 540475392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042232698094282853,
      "loss": 2.7626,
      "theoretical_loss": 3.882336226176939,
      "tokens_seen": 540540928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042231695085255766,
      "loss": 2.8586,
      "theoretical_loss": 3.8822862102431275,
      "tokens_seen": 540606464
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 308194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8666329383850098,
      "objective/train/theoretical_loss": 3.882236202069703,
      "objective/train/tokens_used": 561132000,
      "theoretical_loss": 3.882236202069703,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223069207622869,
      "loss": 2.8035,
      "theoretical_loss": 3.882236202069703,
      "tokens_seen": 540672000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422296890672016,
      "loss": 2.8564,
      "theoretical_loss": 3.8821862016545206,
      "tokens_seen": 540737536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042228686058174526,
      "loss": 2.959,
      "theoretical_loss": 3.8821362089954365,
      "tokens_seen": 540803072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042227683049147444,
      "loss": 2.827,
      "theoretical_loss": 3.8820862240903082,
      "tokens_seen": 540868608
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222668004012036,
      "loss": 2.8683,
      "theoretical_loss": 3.882036246936993,
      "tokens_seen": 540934144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222567703109328,
      "loss": 2.9829,
      "theoretical_loss": 3.8819862775333505,
      "tokens_seen": 540999680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422246740220662,
      "loss": 2.9701,
      "theoretical_loss": 3.88193631587724,
      "tokens_seen": 541065216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042223671013039116,
      "loss": 2.8898,
      "theoretical_loss": 3.8818863619665223,
      "tokens_seen": 541130752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222266800401204,
      "loss": 2.7672,
      "theoretical_loss": 3.881836415799059,
      "tokens_seen": 541196288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222166499498495,
      "loss": 2.9484,
      "theoretical_loss": 3.881786477372712,
      "tokens_seen": 541261824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042220661985957876,
      "loss": 2.8886,
      "theoretical_loss": 3.881736546685345,
      "tokens_seen": 541327360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221965897693079,
      "loss": 3.0252,
      "theoretical_loss": 3.8816866237348218,
      "tokens_seen": 541392896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221865596790371,
      "loss": 2.8781,
      "theoretical_loss": 3.8816367085190064,
      "tokens_seen": 541458432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221765295887663,
      "loss": 2.9667,
      "theoretical_loss": 3.8815868010357657,
      "tokens_seen": 541523968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221664994984955,
      "loss": 2.7153,
      "theoretical_loss": 3.8815369012829657,
      "tokens_seen": 541589504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042215646940822467,
      "loss": 2.5785,
      "theoretical_loss": 3.881487009258473,
      "tokens_seen": 541655040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221464393179539,
      "loss": 2.8231,
      "theoretical_loss": 3.881437124960157,
      "tokens_seen": 541720576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042213640922768303,
      "loss": 2.7915,
      "theoretical_loss": 3.8813872483858867,
      "tokens_seen": 541786112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042212637913741226,
      "loss": 2.8312,
      "theoretical_loss": 3.881337379533531,
      "tokens_seen": 541851648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221163490471414,
      "loss": 2.7949,
      "theoretical_loss": 3.8812875184009616,
      "tokens_seen": 541917184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221063189568706,
      "loss": 2.8074,
      "theoretical_loss": 3.88123766498605,
      "tokens_seen": 541982720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220962888665998,
      "loss": 2.7497,
      "theoretical_loss": 3.881187819286668,
      "tokens_seen": 542048256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422086258776329,
      "loss": 2.9072,
      "theoretical_loss": 3.8811379813006894,
      "tokens_seen": 542113792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042207622868605817,
      "loss": 2.9012,
      "theoretical_loss": 3.8810881510259883,
      "tokens_seen": 542179328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042206619859578735,
      "loss": 2.6692,
      "theoretical_loss": 3.8810383284604395,
      "tokens_seen": 542244864
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 309537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.619420289993286,
      "objective/train/theoretical_loss": 3.8809885136019195,
      "objective/train/tokens_used": 562770400,
      "theoretical_loss": 3.8809885136019195,
      "tokens_seen": 542310400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042205616850551653,
      "loss": 2.9674,
      "theoretical_loss": 3.8809885136019195,
      "tokens_seen": 542310400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042204613841524577,
      "loss": 2.7469,
      "theoretical_loss": 3.880938706448304,
      "tokens_seen": 542375936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220361083249749,
      "loss": 2.8841,
      "theoretical_loss": 3.8808889069974706,
      "tokens_seen": 542441472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042202607823470413,
      "loss": 2.8074,
      "theoretical_loss": 3.880839115247298,
      "tokens_seen": 542507008
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220160481444333,
      "loss": 2.7575,
      "theoretical_loss": 3.880789331195665,
      "tokens_seen": 542572544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220060180541625,
      "loss": 2.8394,
      "theoretical_loss": 3.8807395548404524,
      "tokens_seen": 542638080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004219959879638917,
      "loss": 2.7962,
      "theoretical_loss": 3.8806897861795404,
      "tokens_seen": 542703616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042198595787362085,
      "loss": 2.9498,
      "theoretical_loss": 3.880640025210811,
      "tokens_seen": 542769152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042197592778335004,
      "loss": 2.7269,
      "theoretical_loss": 3.880590271932147,
      "tokens_seen": 542834688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042196589769307927,
      "loss": 2.8028,
      "theoretical_loss": 3.8805405263414308,
      "tokens_seen": 542900224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004219558676028084,
      "loss": 2.8841,
      "theoretical_loss": 3.8804907884365476,
      "tokens_seen": 542965760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042194583751253763,
      "loss": 3.0371,
      "theoretical_loss": 3.8804410582153817,
      "tokens_seen": 543031296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042193580742226676,
      "loss": 2.912,
      "theoretical_loss": 3.8803913356758195,
      "tokens_seen": 543096832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421925777331996,
      "loss": 2.7219,
      "theoretical_loss": 3.8803416208157477,
      "tokens_seen": 543162368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042191574724172523,
      "loss": 2.6821,
      "theoretical_loss": 3.8802919136330534,
      "tokens_seen": 543227904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042190571715145436,
      "loss": 2.8521,
      "theoretical_loss": 3.8802422141256256,
      "tokens_seen": 543293440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218956870611836,
      "loss": 2.7739,
      "theoretical_loss": 3.8801925222913534,
      "tokens_seen": 543358976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218856569709127,
      "loss": 2.9423,
      "theoretical_loss": 3.880142838128126,
      "tokens_seen": 543424512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042187562688064196,
      "loss": 2.7782,
      "theoretical_loss": 3.880093161633836,
      "tokens_seen": 543490048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042186559679037114,
      "loss": 2.7181,
      "theoretical_loss": 3.8800434928063736,
      "tokens_seen": 543555584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218555667001003,
      "loss": 2.8917,
      "theoretical_loss": 3.879993831643632,
      "tokens_seen": 543621120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218455366098295,
      "loss": 3.0724,
      "theoretical_loss": 3.879944178143505,
      "tokens_seen": 543686656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042183550651955873,
      "loss": 2.8608,
      "theoretical_loss": 3.879894532303886,
      "tokens_seen": 543752192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042182547642928786,
      "loss": 2.8151,
      "theoretical_loss": 3.8798448941226704,
      "tokens_seen": 543817728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218154463390171,
      "loss": 2.7074,
      "theoretical_loss": 3.879795263597754,
      "tokens_seen": 543883264
    },
    {
      "epoch": 0.16,
      "objective/train/docs_used": 310176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.820920705795288,
      "objective/train/theoretical_loss": 3.8797456407270343,
      "objective/train/tokens_used": 564408800,
      "theoretical_loss": 3.8797456407270343,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218054162487462,
      "loss": 2.8936,
      "theoretical_loss": 3.8797456407270343,
      "tokens_seen": 543948800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042179538615847546,
      "loss": 2.8416,
      "theoretical_loss": 3.879696025508408,
      "tokens_seen": 544014336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042178535606820464,
      "loss": 2.9673,
      "theoretical_loss": 3.8796464179397736,
      "tokens_seen": 544079872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217753259779338,
      "loss": 3.0778,
      "theoretical_loss": 3.8795968180190306,
      "tokens_seen": 544145408
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421765295887663,
      "loss": 2.9471,
      "theoretical_loss": 3.8795472257440786,
      "tokens_seen": 544210944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217552657973922,
      "loss": 2.7651,
      "theoretical_loss": 3.8794976411128186,
      "tokens_seen": 544276480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042174523570712136,
      "loss": 2.761,
      "theoretical_loss": 3.8794480641231535,
      "tokens_seen": 544342016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217352056168506,
      "loss": 3.0898,
      "theoretical_loss": 3.879398494772984,
      "tokens_seen": 544407552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217251755265797,
      "loss": 2.7585,
      "theoretical_loss": 3.879348933060215,
      "tokens_seen": 544473088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042171514543630896,
      "loss": 2.9004,
      "theoretical_loss": 3.8792993789827497,
      "tokens_seen": 544538624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004217051153460381,
      "loss": 2.7262,
      "theoretical_loss": 3.879249832538493,
      "tokens_seen": 544604160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216950852557673,
      "loss": 2.8847,
      "theoretical_loss": 3.879200293725352,
      "tokens_seen": 544669696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216850551654965,
      "loss": 2.6465,
      "theoretical_loss": 3.879150762541232,
      "tokens_seen": 544735232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216750250752257,
      "loss": 2.953,
      "theoretical_loss": 3.8791012389840414,
      "tokens_seen": 544800768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042166499498495487,
      "loss": 2.6372,
      "theoretical_loss": 3.879051723051688,
      "tokens_seen": 544866304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216549648946841,
      "loss": 2.8553,
      "theoretical_loss": 3.8790022147420817,
      "tokens_seen": 544931840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042164493480441323,
      "loss": 2.861,
      "theoretical_loss": 3.8789527140531312,
      "tokens_seen": 544997376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042163490471414246,
      "loss": 2.9117,
      "theoretical_loss": 3.8789032209827488,
      "tokens_seen": 545062912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216248746238716,
      "loss": 2.7233,
      "theoretical_loss": 3.8788537355288444,
      "tokens_seen": 545128448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216148445336008,
      "loss": 2.8419,
      "theoretical_loss": 3.8788042576893322,
      "tokens_seen": 545193984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042160481444333,
      "loss": 2.7416,
      "theoretical_loss": 3.8787547874621247,
      "tokens_seen": 545259520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215947843530592,
      "loss": 2.9282,
      "theoretical_loss": 3.8787053248451357,
      "tokens_seen": 545325056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042158475426278837,
      "loss": 2.8425,
      "theoretical_loss": 3.8786558698362805,
      "tokens_seen": 545390592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042157472417251755,
      "loss": 2.9664,
      "theoretical_loss": 3.8786064224334744,
      "tokens_seen": 545456128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042156469408224673,
      "loss": 2.6686,
      "theoretical_loss": 3.8785569826346347,
      "tokens_seen": 545521664
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 311361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7344307899475098,
      "objective/train/theoretical_loss": 3.8785075504376785,
      "objective/train/tokens_used": 566047200,
      "theoretical_loss": 3.8785075504376785,
      "tokens_seen": 545587200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042155466399197597,
      "loss": 2.9008,
      "theoretical_loss": 3.8785075504376785,
      "tokens_seen": 545587200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215446339017051,
      "loss": 2.8395,
      "theoretical_loss": 3.878458125840524,
      "tokens_seen": 545652736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042153460381143433,
      "loss": 2.852,
      "theoretical_loss": 3.87840870884109,
      "tokens_seen": 545718272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215245737211635,
      "loss": 2.828,
      "theoretical_loss": 3.878359299437296,
      "tokens_seen": 545783808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215145436308927,
      "loss": 2.9685,
      "theoretical_loss": 3.878309897627064,
      "tokens_seen": 545849344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215045135406219,
      "loss": 2.9763,
      "theoretical_loss": 3.878260503408314,
      "tokens_seen": 545914880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042149448345035105,
      "loss": 2.9497,
      "theoretical_loss": 3.878211116778969,
      "tokens_seen": 545980416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042148445336008024,
      "loss": 2.9101,
      "theoretical_loss": 3.878161737736952,
      "tokens_seen": 546045952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042147442326980947,
      "loss": 2.5154,
      "theoretical_loss": 3.8781123662801873,
      "tokens_seen": 546111488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214643931795386,
      "loss": 2.8239,
      "theoretical_loss": 3.878063002406599,
      "tokens_seen": 546177024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042145436308926783,
      "loss": 2.9453,
      "theoretical_loss": 3.878013646114113,
      "tokens_seen": 546242560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042144433299899696,
      "loss": 2.692,
      "theoretical_loss": 3.8779642974006556,
      "tokens_seen": 546308096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214343029087262,
      "loss": 2.7597,
      "theoretical_loss": 3.877914956264154,
      "tokens_seen": 546373632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214242728184554,
      "loss": 2.8578,
      "theoretical_loss": 3.8778656227025365,
      "tokens_seen": 546439168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042141424272818456,
      "loss": 2.9577,
      "theoretical_loss": 3.877816296713732,
      "tokens_seen": 546504704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042140421263791374,
      "loss": 2.9289,
      "theoretical_loss": 3.8777669782956696,
      "tokens_seen": 546570240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213941825476429,
      "loss": 2.9022,
      "theoretical_loss": 3.8777176674462797,
      "tokens_seen": 546635776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213841524573721,
      "loss": 2.8726,
      "theoretical_loss": 3.877668364163494,
      "tokens_seen": 546701312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042137412236710134,
      "loss": 2.6827,
      "theoretical_loss": 3.8776190684452447,
      "tokens_seen": 546766848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042136409227683046,
      "loss": 2.8975,
      "theoretical_loss": 3.8775697802894644,
      "tokens_seen": 546832384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213540621865597,
      "loss": 2.8235,
      "theoretical_loss": 3.8775204996940866,
      "tokens_seen": 546897920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213440320962889,
      "loss": 2.7518,
      "theoretical_loss": 3.877471226657047,
      "tokens_seen": 546963456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042133400200601806,
      "loss": 2.9066,
      "theoretical_loss": 3.877421961176279,
      "tokens_seen": 547028992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042132397191574724,
      "loss": 3.0313,
      "theoretical_loss": 3.87737270324972,
      "tokens_seen": 547094528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213139418254764,
      "loss": 3.0107,
      "theoretical_loss": 3.8773234528753076,
      "tokens_seen": 547160064
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 312508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.020892381668091,
      "objective/train/theoretical_loss": 3.8772742100509774,
      "objective/train/tokens_used": 567685600,
      "theoretical_loss": 3.8772742100509774,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213039117352056,
      "loss": 2.8083,
      "theoretical_loss": 3.8772742100509774,
      "tokens_seen": 547225600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042129388164493484,
      "loss": 2.5294,
      "theoretical_loss": 3.8772249747746703,
      "tokens_seen": 547291136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042128385155466397,
      "loss": 3.0742,
      "theoretical_loss": 3.8771757470443244,
      "tokens_seen": 547356672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212738214643932,
      "loss": 2.6558,
      "theoretical_loss": 3.87712652685788,
      "tokens_seen": 547422208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042126379137412233,
      "loss": 3.0363,
      "theoretical_loss": 3.8770773142132784,
      "tokens_seen": 547487744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042125376128385156,
      "loss": 2.8451,
      "theoretical_loss": 3.8770281091084615,
      "tokens_seen": 547553280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042124373119358075,
      "loss": 2.7166,
      "theoretical_loss": 3.8769789115413715,
      "tokens_seen": 547618816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212337011033099,
      "loss": 2.7842,
      "theoretical_loss": 3.8769297215099523,
      "tokens_seen": 547684352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212236710130391,
      "loss": 3.1148,
      "theoretical_loss": 3.8768805390121477,
      "tokens_seen": 547749888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212136409227683,
      "loss": 2.8194,
      "theoretical_loss": 3.876831364045903,
      "tokens_seen": 547815424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042120361083249747,
      "loss": 2.7916,
      "theoretical_loss": 3.876782196609164,
      "tokens_seen": 547880960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211935807422267,
      "loss": 2.4156,
      "theoretical_loss": 3.8767330366998776,
      "tokens_seen": 547946496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042118355065195583,
      "loss": 2.9464,
      "theoretical_loss": 3.8766838843159905,
      "tokens_seen": 548012032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042117352056168507,
      "loss": 2.8205,
      "theoretical_loss": 3.8766347394554517,
      "tokens_seen": 548077568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211634904714143,
      "loss": 2.9376,
      "theoretical_loss": 3.87658560211621,
      "tokens_seen": 548143104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042115346038114343,
      "loss": 2.8582,
      "theoretical_loss": 3.8765364722962152,
      "tokens_seen": 548208640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042114343029087266,
      "loss": 2.9846,
      "theoretical_loss": 3.8764873499934183,
      "tokens_seen": 548274176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211334002006018,
      "loss": 2.7357,
      "theoretical_loss": 3.8764382352057707,
      "tokens_seen": 548339712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042112337011033103,
      "loss": 2.911,
      "theoretical_loss": 3.8763891279312244,
      "tokens_seen": 548405248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211133400200602,
      "loss": 2.8353,
      "theoretical_loss": 3.876340028167732,
      "tokens_seen": 548470784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211033099297894,
      "loss": 2.8009,
      "theoretical_loss": 3.876290935913249,
      "tokens_seen": 548536320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042109327983951857,
      "loss": 2.7628,
      "theoretical_loss": 3.876241851165729,
      "tokens_seen": 548601856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042108324974924775,
      "loss": 2.74,
      "theoretical_loss": 3.876192773923127,
      "tokens_seen": 548667392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042107321965897693,
      "loss": 3.0434,
      "theoretical_loss": 3.8761437041834004,
      "tokens_seen": 548732928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042106318956870617,
      "loss": 2.8916,
      "theoretical_loss": 3.8760946419445057,
      "tokens_seen": 548798464
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 312943,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1393747329711914,
      "objective/train/theoretical_loss": 3.876045587204401,
      "objective/train/tokens_used": 569324000,
      "theoretical_loss": 3.876045587204401,
      "tokens_seen": 548864000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210531594784353,
      "loss": 2.7268,
      "theoretical_loss": 3.876045587204401,
      "tokens_seen": 548864000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042104312938816453,
      "loss": 2.7706,
      "theoretical_loss": 3.875996539961045,
      "tokens_seen": 548929536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210330992978937,
      "loss": 2.8024,
      "theoretical_loss": 3.875947500212397,
      "tokens_seen": 548995072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210230692076229,
      "loss": 2.8611,
      "theoretical_loss": 3.8758984679564175,
      "tokens_seen": 549060608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210130391173521,
      "loss": 2.5877,
      "theoretical_loss": 3.875849443191067,
      "tokens_seen": 549126144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042100300902708125,
      "loss": 2.7126,
      "theoretical_loss": 3.875800425914308,
      "tokens_seen": 549191680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042099297893681044,
      "loss": 2.8382,
      "theoretical_loss": 3.8757514161241033,
      "tokens_seen": 549257216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042098294884653967,
      "loss": 2.8309,
      "theoretical_loss": 3.8757024138184164,
      "tokens_seen": 549322752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209729187562688,
      "loss": 2.7394,
      "theoretical_loss": 3.8756534189952108,
      "tokens_seen": 549388288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042096288866599803,
      "loss": 2.8002,
      "theoretical_loss": 3.8756044316524525,
      "tokens_seen": 549453824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042095285857572716,
      "loss": 2.8171,
      "theoretical_loss": 3.875555451788107,
      "tokens_seen": 549519360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209428284854564,
      "loss": 2.8046,
      "theoretical_loss": 3.8755064794001406,
      "tokens_seen": 549584896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209327983951856,
      "loss": 2.8639,
      "theoretical_loss": 3.875457514486521,
      "tokens_seen": 549650432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042092276830491476,
      "loss": 2.8338,
      "theoretical_loss": 3.8754085570452164,
      "tokens_seen": 549715968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042091273821464394,
      "loss": 2.9081,
      "theoretical_loss": 3.8753596070741967,
      "tokens_seen": 549781504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209027081243731,
      "loss": 2.5443,
      "theoretical_loss": 3.8753106645714306,
      "tokens_seen": 549847040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208926780341023,
      "loss": 2.587,
      "theoretical_loss": 3.8752617295348895,
      "tokens_seen": 549912576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042088264794383154,
      "loss": 2.7818,
      "theoretical_loss": 3.8752128019625443,
      "tokens_seen": 549978112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042087261785356066,
      "loss": 2.8385,
      "theoretical_loss": 3.8751638818523677,
      "tokens_seen": 550043648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208625877632899,
      "loss": 2.7451,
      "theoretical_loss": 3.875114969202332,
      "tokens_seen": 550109184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208525576730191,
      "loss": 2.7683,
      "theoretical_loss": 3.8750660640104124,
      "tokens_seen": 550174720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042084252758274826,
      "loss": 2.8129,
      "theoretical_loss": 3.875017166274582,
      "tokens_seen": 550240256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042083249749247744,
      "loss": 3.0056,
      "theoretical_loss": 3.8749682759928166,
      "tokens_seen": 550305792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208224674022066,
      "loss": 2.8207,
      "theoretical_loss": 3.8749193931630925,
      "tokens_seen": 550371328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208124373119358,
      "loss": 2.9711,
      "theoretical_loss": 3.8748705177833873,
      "tokens_seen": 550436864
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 314102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.005920886993408,
      "objective/train/theoretical_loss": 3.874821649851678,
      "objective/train/tokens_used": 570962400,
      "theoretical_loss": 3.874821649851678,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042080240722166504,
      "loss": 3.0135,
      "theoretical_loss": 3.874821649851678,
      "tokens_seen": 550502400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042079237713139417,
      "loss": 2.6148,
      "theoretical_loss": 3.8747727893659434,
      "tokens_seen": 550567936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207823470411234,
      "loss": 2.7024,
      "theoretical_loss": 3.8747239363241626,
      "tokens_seen": 550633472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042077231695085253,
      "loss": 2.698,
      "theoretical_loss": 3.8746750907243164,
      "tokens_seen": 550699008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042076228686058176,
      "loss": 3.0277,
      "theoretical_loss": 3.8746262525643846,
      "tokens_seen": 550764544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042075225677031095,
      "loss": 2.8225,
      "theoretical_loss": 3.8745774218423508,
      "tokens_seen": 550830080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207422266800401,
      "loss": 2.9363,
      "theoretical_loss": 3.874528598556195,
      "tokens_seen": 550895616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207321965897693,
      "loss": 2.7584,
      "theoretical_loss": 3.8744797827039026,
      "tokens_seen": 550961152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207221664994985,
      "loss": 2.7688,
      "theoretical_loss": 3.874430974283456,
      "tokens_seen": 551026688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042071213640922767,
      "loss": 2.9367,
      "theoretical_loss": 3.874382173292842,
      "tokens_seen": 551092224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207021063189569,
      "loss": 2.7892,
      "theoretical_loss": 3.874333379730044,
      "tokens_seen": 551157760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042069207622868603,
      "loss": 2.8285,
      "theoretical_loss": 3.8742845935930506,
      "tokens_seen": 551223296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042068204613841527,
      "loss": 2.6297,
      "theoretical_loss": 3.8742358148798473,
      "tokens_seen": 551288832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042067201604814445,
      "loss": 2.8855,
      "theoretical_loss": 3.8741870435884227,
      "tokens_seen": 551354368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042066198595787363,
      "loss": 2.7368,
      "theoretical_loss": 3.874138279716766,
      "tokens_seen": 551419904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206519558676028,
      "loss": 2.7036,
      "theoretical_loss": 3.8740895232628665,
      "tokens_seen": 551485440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420641925777332,
      "loss": 2.9885,
      "theoretical_loss": 3.874040774224714,
      "tokens_seen": 551550976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206318956870612,
      "loss": 2.836,
      "theoretical_loss": 3.8739920326003006,
      "tokens_seen": 551616512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206218655967904,
      "loss": 2.8656,
      "theoretical_loss": 3.8739432983876174,
      "tokens_seen": 551682048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042061183550651954,
      "loss": 2.766,
      "theoretical_loss": 3.8738945715846578,
      "tokens_seen": 551747584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042060180541624877,
      "loss": 2.8527,
      "theoretical_loss": 3.8738458521894144,
      "tokens_seen": 551813120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205917753259779,
      "loss": 2.8719,
      "theoretical_loss": 3.8737971401998825,
      "tokens_seen": 551878656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042058174523570713,
      "loss": 2.7869,
      "theoretical_loss": 3.8737484356140564,
      "tokens_seen": 551944192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205717151454363,
      "loss": 2.759,
      "theoretical_loss": 3.873699738429932,
      "tokens_seen": 552009728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205616850551655,
      "loss": 2.8348,
      "theoretical_loss": 3.873651048645506,
      "tokens_seen": 552075264
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 314596,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6391680240631104,
      "objective/train/theoretical_loss": 3.873602366258776,
      "objective/train/tokens_used": 572600800,
      "theoretical_loss": 3.873602366258776,
      "tokens_seen": 552140800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205516549648947,
      "loss": 2.7018,
      "theoretical_loss": 3.873602366258776,
      "tokens_seen": 552140800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205416248746239,
      "loss": 3.0718,
      "theoretical_loss": 3.87355369126774,
      "tokens_seen": 552206336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042053159478435304,
      "loss": 2.9741,
      "theoretical_loss": 3.873505023670397,
      "tokens_seen": 552271872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205215646940823,
      "loss": 2.6878,
      "theoretical_loss": 3.8734563634647463,
      "tokens_seen": 552337408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205115346038114,
      "loss": 2.7524,
      "theoretical_loss": 3.8734077106487894,
      "tokens_seen": 552402944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042050150451354064,
      "loss": 2.8489,
      "theoretical_loss": 3.8733590652205265,
      "tokens_seen": 552468480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204914744232698,
      "loss": 2.8195,
      "theoretical_loss": 3.8733104271779606,
      "tokens_seen": 552534016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420481444332999,
      "loss": 2.7318,
      "theoretical_loss": 3.8732617965190936,
      "tokens_seen": 552599552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204714142427282,
      "loss": 2.7373,
      "theoretical_loss": 3.873213173241929,
      "tokens_seen": 552665088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042046138415245736,
      "loss": 2.7681,
      "theoretical_loss": 3.8731645573444724,
      "tokens_seen": 552730624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042045135406218654,
      "loss": 2.801,
      "theoretical_loss": 3.8731159488247284,
      "tokens_seen": 552796160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204413239719158,
      "loss": 2.7685,
      "theoretical_loss": 3.8730673476807023,
      "tokens_seen": 552861696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204312938816449,
      "loss": 2.8712,
      "theoretical_loss": 3.8730187539104017,
      "tokens_seen": 552927232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042042126379137414,
      "loss": 2.7277,
      "theoretical_loss": 3.8729701675118338,
      "tokens_seen": 552992768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204112337011033,
      "loss": 2.8187,
      "theoretical_loss": 3.8729215884830066,
      "tokens_seen": 553058304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204012036108325,
      "loss": 2.6754,
      "theoretical_loss": 3.8728730168219294,
      "tokens_seen": 553123840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042039117352056174,
      "loss": 2.6005,
      "theoretical_loss": 3.872824452526612,
      "tokens_seen": 553189376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042038114343029086,
      "loss": 2.6217,
      "theoretical_loss": 3.8727758955950646,
      "tokens_seen": 553254912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203711133400201,
      "loss": 2.9321,
      "theoretical_loss": 3.8727273460252993,
      "tokens_seen": 553320448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203610832497493,
      "loss": 2.6787,
      "theoretical_loss": 3.8726788038153277,
      "tokens_seen": 553385984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042035105315947846,
      "loss": 2.8041,
      "theoretical_loss": 3.872630268963163,
      "tokens_seen": 553451520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042034102306920764,
      "loss": 2.811,
      "theoretical_loss": 3.872581741466818,
      "tokens_seen": 553517056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203309929789368,
      "loss": 2.6364,
      "theoretical_loss": 3.8725332213243084,
      "tokens_seen": 553582592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420320962888666,
      "loss": 2.6177,
      "theoretical_loss": 3.8724847085336487,
      "tokens_seen": 553648128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042031093279839524,
      "loss": 2.7531,
      "theoretical_loss": 3.872436203092855,
      "tokens_seen": 553713664
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 315339,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5412378311157227,
      "objective/train/theoretical_loss": 3.8723877049999444,
      "objective/train/tokens_used": 574239200,
      "theoretical_loss": 3.8723877049999444,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042030090270812437,
      "loss": 2.4705,
      "theoretical_loss": 3.8723877049999444,
      "tokens_seen": 553779200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202908726178536,
      "loss": 2.6838,
      "theoretical_loss": 3.872339214252934,
      "tokens_seen": 553844736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042028084252758273,
      "loss": 2.6182,
      "theoretical_loss": 3.8722907308498415,
      "tokens_seen": 553910272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042027081243731196,
      "loss": 2.8419,
      "theoretical_loss": 3.8722422547886874,
      "tokens_seen": 553975808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042026078234704115,
      "loss": 2.7933,
      "theoretical_loss": 3.8721937860674904,
      "tokens_seen": 554041344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202507522567703,
      "loss": 2.9367,
      "theoretical_loss": 3.872145324684272,
      "tokens_seen": 554106880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202407221664995,
      "loss": 2.5143,
      "theoretical_loss": 3.8720968706370527,
      "tokens_seen": 554172416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202306920762287,
      "loss": 2.8308,
      "theoretical_loss": 3.8720484239238555,
      "tokens_seen": 554237952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042022066198595787,
      "loss": 2.707,
      "theoretical_loss": 3.8719999845427022,
      "tokens_seen": 554303488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202106318956871,
      "loss": 2.4634,
      "theoretical_loss": 3.871951552491617,
      "tokens_seen": 554369024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042020060180541623,
      "loss": 2.6944,
      "theoretical_loss": 3.871903127768625,
      "tokens_seen": 554434560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042019057171514547,
      "loss": 2.8056,
      "theoretical_loss": 3.8718547103717507,
      "tokens_seen": 554500096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042018054162487465,
      "loss": 2.7418,
      "theoretical_loss": 3.87180630029902,
      "tokens_seen": 554565632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042017051153460383,
      "loss": 2.9632,
      "theoretical_loss": 3.87175789754846,
      "tokens_seen": 554631168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420160481444333,
      "loss": 2.5859,
      "theoretical_loss": 3.8717095021180983,
      "tokens_seen": 554696704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201504513540622,
      "loss": 2.8359,
      "theoretical_loss": 3.8716611140059625,
      "tokens_seen": 554762240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201404212637914,
      "loss": 2.6759,
      "theoretical_loss": 3.871612733210082,
      "tokens_seen": 554827776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201303911735206,
      "loss": 2.7389,
      "theoretical_loss": 3.8715643597284863,
      "tokens_seen": 554893312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042012036108324974,
      "loss": 2.7195,
      "theoretical_loss": 3.871515993559207,
      "tokens_seen": 554958848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042011033099297897,
      "loss": 2.814,
      "theoretical_loss": 3.871467634700274,
      "tokens_seen": 555024384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201003009027081,
      "loss": 2.5207,
      "theoretical_loss": 3.8714192831497205,
      "tokens_seen": 555089920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042009027081243733,
      "loss": 2.6039,
      "theoretical_loss": 3.871370938905579,
      "tokens_seen": 555155456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200802407221665,
      "loss": 3.1536,
      "theoretical_loss": 3.8713226019658826,
      "tokens_seen": 555220992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200702106318957,
      "loss": 2.6002,
      "theoretical_loss": 3.871274272328666,
      "tokens_seen": 555286528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200601805416249,
      "loss": 2.8405,
      "theoretical_loss": 3.871225949991965,
      "tokens_seen": 555352064
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 316316,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.795095682144165,
      "objective/train/theoretical_loss": 3.8711776349538143,
      "objective/train/tokens_used": 575877600,
      "theoretical_loss": 3.8711776349538143,
      "tokens_seen": 555417600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200501504513541,
      "loss": 2.7872,
      "theoretical_loss": 3.8711776349538143,
      "tokens_seen": 555417600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042004012036108324,
      "loss": 2.8702,
      "theoretical_loss": 3.8711293272122513,
      "tokens_seen": 555483136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200300902708125,
      "loss": 2.9276,
      "theoretical_loss": 3.871081026765313,
      "tokens_seen": 555548672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200200601805416,
      "loss": 2.7231,
      "theoretical_loss": 3.871032733611038,
      "tokens_seen": 555614208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042001003009027084,
      "loss": 2.8501,
      "theoretical_loss": 3.870984447747465,
      "tokens_seen": 555679744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042,
      "loss": 2.8278,
      "theoretical_loss": 3.8709361691726336,
      "tokens_seen": 555745280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199899699097292,
      "loss": 2.8936,
      "theoretical_loss": 3.870887897884584,
      "tokens_seen": 555810816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199799398194584,
      "loss": 2.6572,
      "theoretical_loss": 3.870839633881358,
      "tokens_seen": 555876352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041996990972918756,
      "loss": 2.9405,
      "theoretical_loss": 3.8707913771609976,
      "tokens_seen": 555941888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041995987963891674,
      "loss": 2.4867,
      "theoretical_loss": 3.8707431277215454,
      "tokens_seen": 556007424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419949849548646,
      "loss": 2.8431,
      "theoretical_loss": 3.8706948855610435,
      "tokens_seen": 556072960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199398194583751,
      "loss": 2.947,
      "theoretical_loss": 3.8706466506775383,
      "tokens_seen": 556138496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041992978936810434,
      "loss": 3.0076,
      "theoretical_loss": 3.8705984230690733,
      "tokens_seen": 556204032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041991975927783347,
      "loss": 2.8488,
      "theoretical_loss": 3.870550202733695,
      "tokens_seen": 556269568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199097291875627,
      "loss": 2.6245,
      "theoretical_loss": 3.8705019896694495,
      "tokens_seen": 556335104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198996990972919,
      "loss": 2.8031,
      "theoretical_loss": 3.8704537838743835,
      "tokens_seen": 556400640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041988966900702106,
      "loss": 2.6974,
      "theoretical_loss": 3.870405585346546,
      "tokens_seen": 556466176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041987963891675025,
      "loss": 2.6182,
      "theoretical_loss": 3.8703573940839857,
      "tokens_seen": 556531712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198696088264795,
      "loss": 2.6669,
      "theoretical_loss": 3.8703092100847516,
      "tokens_seen": 556597248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198595787362086,
      "loss": 2.8824,
      "theoretical_loss": 3.8702610333468943,
      "tokens_seen": 556662784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041984954864593784,
      "loss": 2.8379,
      "theoretical_loss": 3.8702128638684643,
      "tokens_seen": 556728320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041983951855566697,
      "loss": 2.602,
      "theoretical_loss": 3.870164701647514,
      "tokens_seen": 556793856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198294884653962,
      "loss": 2.84,
      "theoretical_loss": 3.8701165466820955,
      "tokens_seen": 556859392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198194583751254,
      "loss": 2.7329,
      "theoretical_loss": 3.870068398970262,
      "tokens_seen": 556924928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041980942828485457,
      "loss": 2.7621,
      "theoretical_loss": 3.870020258510068,
      "tokens_seen": 556990464
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 317467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9780025482177734,
      "objective/train/theoretical_loss": 3.869972125299568,
      "objective/train/tokens_used": 577516000,
      "theoretical_loss": 3.869972125299568,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041979939819458375,
      "loss": 2.8299,
      "theoretical_loss": 3.869972125299568,
      "tokens_seen": 557056000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041978936810431293,
      "loss": 2.7222,
      "theoretical_loss": 3.8699239993368177,
      "tokens_seen": 557121536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197793380140421,
      "loss": 2.6026,
      "theoretical_loss": 3.869875880619873,
      "tokens_seen": 557187072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041976930792377135,
      "loss": 2.8492,
      "theoretical_loss": 3.869827769146791,
      "tokens_seen": 557252608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041975927783350047,
      "loss": 2.6935,
      "theoretical_loss": 3.8697796649156295,
      "tokens_seen": 557318144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197492477432297,
      "loss": 2.5823,
      "theoretical_loss": 3.8697315679244477,
      "tokens_seen": 557383680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041973921765295883,
      "loss": 2.7691,
      "theoretical_loss": 3.8696834781713036,
      "tokens_seen": 557449216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041972918756268807,
      "loss": 2.6545,
      "theoretical_loss": 3.8696353956542584,
      "tokens_seen": 557514752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041971915747241725,
      "loss": 2.6819,
      "theoretical_loss": 3.869587320371372,
      "tokens_seen": 557580288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041970912738214643,
      "loss": 2.7852,
      "theoretical_loss": 3.869539252320707,
      "tokens_seen": 557645824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196990972918756,
      "loss": 2.6624,
      "theoretical_loss": 3.8694911915003245,
      "tokens_seen": 557711360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041968906720160485,
      "loss": 2.8304,
      "theoretical_loss": 3.869443137908288,
      "tokens_seen": 557776896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419679037111334,
      "loss": 2.9432,
      "theoretical_loss": 3.869395091542661,
      "tokens_seen": 557842432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196690070210632,
      "loss": 2.6705,
      "theoretical_loss": 3.8693470524015083,
      "tokens_seen": 557907968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196589769307924,
      "loss": 2.9572,
      "theoretical_loss": 3.869299020482895,
      "tokens_seen": 557973504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196489468405216,
      "loss": 2.5857,
      "theoretical_loss": 3.8692509957848875,
      "tokens_seen": 558039040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196389167502508,
      "loss": 2.8012,
      "theoretical_loss": 3.8692029783055517,
      "tokens_seen": 558104576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041962888665997994,
      "loss": 2.8241,
      "theoretical_loss": 3.869154968042956,
      "tokens_seen": 558170112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041961885656970917,
      "loss": 2.8948,
      "theoretical_loss": 3.8691069649951677,
      "tokens_seen": 558235648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196088264794383,
      "loss": 2.827,
      "theoretical_loss": 3.869058969160257,
      "tokens_seen": 558301184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041959879638916753,
      "loss": 2.8092,
      "theoretical_loss": 3.869010980536292,
      "tokens_seen": 558366720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195887662988967,
      "loss": 2.8753,
      "theoretical_loss": 3.868962999121344,
      "tokens_seen": 558432256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195787362086259,
      "loss": 2.8558,
      "theoretical_loss": 3.868915024913485,
      "tokens_seen": 558497792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195687061183551,
      "loss": 2.8519,
      "theoretical_loss": 3.8688670579107853,
      "tokens_seen": 558563328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195586760280843,
      "loss": 2.9895,
      "theoretical_loss": 3.8688190981113193,
      "tokens_seen": 558628864
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 318032,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9913525581359863,
      "objective/train/theoretical_loss": 3.868771145513159,
      "objective/train/tokens_used": 579154400,
      "theoretical_loss": 3.868771145513159,
      "tokens_seen": 558694400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041954864593781344,
      "loss": 2.6942,
      "theoretical_loss": 3.868771145513159,
      "tokens_seen": 558694400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195386158475427,
      "loss": 2.7493,
      "theoretical_loss": 3.8687232001143794,
      "tokens_seen": 558759936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195285857572718,
      "loss": 2.9329,
      "theoretical_loss": 3.8686752619130553,
      "tokens_seen": 558825472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041951855566700104,
      "loss": 2.8755,
      "theoretical_loss": 3.8686273309072616,
      "tokens_seen": 558891008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195085255767302,
      "loss": 2.7809,
      "theoretical_loss": 3.8685794070950754,
      "tokens_seen": 558956544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194984954864594,
      "loss": 3.0322,
      "theoretical_loss": 3.8685314904745742,
      "tokens_seen": 559022080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194884653961886,
      "loss": 2.6098,
      "theoretical_loss": 3.8684835810438347,
      "tokens_seen": 559087616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041947843530591776,
      "loss": 2.7776,
      "theoretical_loss": 3.868435678800937,
      "tokens_seen": 559153152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041946840521564694,
      "loss": 2.7414,
      "theoretical_loss": 3.868387783743959,
      "tokens_seen": 559218688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194583751253762,
      "loss": 2.883,
      "theoretical_loss": 3.8683398958709816,
      "tokens_seen": 559284224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194483450351053,
      "loss": 2.6664,
      "theoretical_loss": 3.868292015180085,
      "tokens_seen": 559349760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041943831494483454,
      "loss": 2.7937,
      "theoretical_loss": 3.8682441416693516,
      "tokens_seen": 559415296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041942828485456367,
      "loss": 2.8014,
      "theoretical_loss": 3.868196275336863,
      "tokens_seen": 559480832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194182547642929,
      "loss": 2.8912,
      "theoretical_loss": 3.8681484161807025,
      "tokens_seen": 559546368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194082246740221,
      "loss": 2.8199,
      "theoretical_loss": 3.8681005641989543,
      "tokens_seen": 559611904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041939819458375126,
      "loss": 2.792,
      "theoretical_loss": 3.8680527193897016,
      "tokens_seen": 559677440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041938816449348045,
      "loss": 2.8252,
      "theoretical_loss": 3.8680048817510313,
      "tokens_seen": 559742976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193781344032097,
      "loss": 2.9691,
      "theoretical_loss": 3.867957051281028,
      "tokens_seen": 559808512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193681043129388,
      "loss": 2.7795,
      "theoretical_loss": 3.867909227977779,
      "tokens_seen": 559874048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041935807422266804,
      "loss": 2.7819,
      "theoretical_loss": 3.8678614118393724,
      "tokens_seen": 559939584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041934804413239717,
      "loss": 2.6689,
      "theoretical_loss": 3.8678136028638948,
      "tokens_seen": 560005120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193380140421264,
      "loss": 2.8671,
      "theoretical_loss": 3.867765801049437,
      "tokens_seen": 560070656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193279839518556,
      "loss": 2.8617,
      "theoretical_loss": 3.8677180063940866,
      "tokens_seen": 560136192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041931795386158477,
      "loss": 2.8101,
      "theoretical_loss": 3.8676702188959355,
      "tokens_seen": 560201728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041930792377131395,
      "loss": 2.8684,
      "theoretical_loss": 3.8676224385530746,
      "tokens_seen": 560267264
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 319260,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.651021718978882,
      "objective/train/theoretical_loss": 3.867574665363595,
      "objective/train/tokens_used": 580792800,
      "theoretical_loss": 3.867574665363595,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041929789368104313,
      "loss": 2.8879,
      "theoretical_loss": 3.867574665363595,
      "tokens_seen": 560332800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192878635907723,
      "loss": 2.8662,
      "theoretical_loss": 3.8675268993255902,
      "tokens_seen": 560398336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041927783350050155,
      "loss": 2.8046,
      "theoretical_loss": 3.867479140437153,
      "tokens_seen": 560463872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041926780341023067,
      "loss": 2.7195,
      "theoretical_loss": 3.867431388696377,
      "tokens_seen": 560529408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192577733199599,
      "loss": 2.6808,
      "theoretical_loss": 3.867383644101358,
      "tokens_seen": 560594944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041924774322968904,
      "loss": 2.8332,
      "theoretical_loss": 3.8673359066501902,
      "tokens_seen": 560660480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041923771313941827,
      "loss": 2.884,
      "theoretical_loss": 3.8672881763409714,
      "tokens_seen": 560726016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041922768304914745,
      "loss": 2.8302,
      "theoretical_loss": 3.8672404531717977,
      "tokens_seen": 560791552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041921765295887663,
      "loss": 2.9092,
      "theoretical_loss": 3.8671927371407664,
      "tokens_seen": 560857088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192076228686058,
      "loss": 2.8494,
      "theoretical_loss": 3.8671450282459765,
      "tokens_seen": 560922624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041919759277833505,
      "loss": 2.8229,
      "theoretical_loss": 3.867097326485527,
      "tokens_seen": 560988160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191875626880642,
      "loss": 2.8452,
      "theoretical_loss": 3.867049631857518,
      "tokens_seen": 561053696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191775325977934,
      "loss": 2.8199,
      "theoretical_loss": 3.86700194436005,
      "tokens_seen": 561119232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041916750250752254,
      "loss": 2.9338,
      "theoretical_loss": 3.8669542639912238,
      "tokens_seen": 561184768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191574724172518,
      "loss": 3.0957,
      "theoretical_loss": 3.866906590749142,
      "tokens_seen": 561250304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041914744232698095,
      "loss": 2.7999,
      "theoretical_loss": 3.8668589246319067,
      "tokens_seen": 561315840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041913741223671014,
      "loss": 2.9401,
      "theoretical_loss": 3.8668112656376223,
      "tokens_seen": 561381376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191273821464393,
      "loss": 2.7745,
      "theoretical_loss": 3.8667636137643937,
      "tokens_seen": 561446912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191173520561685,
      "loss": 2.9357,
      "theoretical_loss": 3.8667159690103237,
      "tokens_seen": 561512448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191073219658977,
      "loss": 2.6896,
      "theoretical_loss": 3.8666683313735195,
      "tokens_seen": 561577984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190972918756269,
      "loss": 2.7651,
      "theoretical_loss": 3.8666207008520868,
      "tokens_seen": 561643520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041908726178535604,
      "loss": 2.8477,
      "theoretical_loss": 3.8665730774441336,
      "tokens_seen": 561709056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190772316950853,
      "loss": 3.0061,
      "theoretical_loss": 3.866525461147767,
      "tokens_seen": 561774592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190672016048144,
      "loss": 2.7424,
      "theoretical_loss": 3.866477851961096,
      "tokens_seen": 561840128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041905717151454364,
      "loss": 2.9924,
      "theoretical_loss": 3.8664302498822294,
      "tokens_seen": 561905664
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 319818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6618897914886475,
      "objective/train/theoretical_loss": 3.8663826549092777,
      "objective/train/tokens_used": 582431200,
      "theoretical_loss": 3.8663826549092777,
      "tokens_seen": 561971200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190471414242728,
      "loss": 3.0017,
      "theoretical_loss": 3.8663826549092777,
      "tokens_seen": 561971200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419037111334002,
      "loss": 2.7453,
      "theoretical_loss": 3.8663350670403513,
      "tokens_seen": 562036736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190270812437312,
      "loss": 2.7955,
      "theoretical_loss": 3.8662874862735626,
      "tokens_seen": 562102272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190170511534604,
      "loss": 2.7434,
      "theoretical_loss": 3.8662399126070226,
      "tokens_seen": 562167808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041900702106318954,
      "loss": 2.8103,
      "theoretical_loss": 3.866192346038845,
      "tokens_seen": 562233344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189969909729188,
      "loss": 2.7913,
      "theoretical_loss": 3.866144786567143,
      "tokens_seen": 562298880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189869608826479,
      "loss": 2.6216,
      "theoretical_loss": 3.8660972341900313,
      "tokens_seen": 562364416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041897693079237714,
      "loss": 2.7139,
      "theoretical_loss": 3.8660496889056244,
      "tokens_seen": 562429952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189669007021063,
      "loss": 3.0612,
      "theoretical_loss": 3.8660021507120392,
      "tokens_seen": 562495488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189568706118355,
      "loss": 2.7447,
      "theoretical_loss": 3.8659546196073906,
      "tokens_seen": 562561024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189468405215647,
      "loss": 2.624,
      "theoretical_loss": 3.8659070955897974,
      "tokens_seen": 562626560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041893681043129387,
      "loss": 2.5809,
      "theoretical_loss": 3.865859578657377,
      "tokens_seen": 562692096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041892678034102305,
      "loss": 2.6602,
      "theoretical_loss": 3.865812068808248,
      "tokens_seen": 562757632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189167502507523,
      "loss": 2.7531,
      "theoretical_loss": 3.86576456604053,
      "tokens_seen": 562823168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041890672016048146,
      "loss": 2.8401,
      "theoretical_loss": 3.8657170703523427,
      "tokens_seen": 562888704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041889669007021065,
      "loss": 2.7188,
      "theoretical_loss": 3.865669581741807,
      "tokens_seen": 562954240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188866599799399,
      "loss": 2.7441,
      "theoretical_loss": 3.865622100207045,
      "tokens_seen": 563019776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418876629889669,
      "loss": 2.7755,
      "theoretical_loss": 3.8655746257461785,
      "tokens_seen": 563085312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041886659979939824,
      "loss": 2.8061,
      "theoretical_loss": 3.8655271583573305,
      "tokens_seen": 563150848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041885656970912737,
      "loss": 2.8223,
      "theoretical_loss": 3.8654796980386252,
      "tokens_seen": 563216384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188465396188566,
      "loss": 2.7483,
      "theoretical_loss": 3.8654322447881864,
      "tokens_seen": 563281920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188365095285858,
      "loss": 3.0129,
      "theoretical_loss": 3.8653847986041394,
      "tokens_seen": 563347456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041882647943831497,
      "loss": 2.9452,
      "theoretical_loss": 3.8653373594846103,
      "tokens_seen": 563412992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041881644934804415,
      "loss": 2.8245,
      "theoretical_loss": 3.8652899274277255,
      "tokens_seen": 563478528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041880641925777333,
      "loss": 3.1132,
      "theoretical_loss": 3.865242502431612,
      "tokens_seen": 563544064
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 321068,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.633208751678467,
      "objective/train/theoretical_loss": 3.865195084494398,
      "objective/train/tokens_used": 584069600,
      "theoretical_loss": 3.865195084494398,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187963891675025,
      "loss": 2.6685,
      "theoretical_loss": 3.865195084494398,
      "tokens_seen": 563609600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041878635907723175,
      "loss": 2.5112,
      "theoretical_loss": 3.8651476736142127,
      "tokens_seen": 563675136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187763289869609,
      "loss": 2.7993,
      "theoretical_loss": 3.8651002697891848,
      "tokens_seen": 563740672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187662988966901,
      "loss": 2.7206,
      "theoretical_loss": 3.8650528730174454,
      "tokens_seen": 563806208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041875626880641924,
      "loss": 2.8497,
      "theoretical_loss": 3.8650054832971237,
      "tokens_seen": 563871744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041874623871614847,
      "loss": 2.6861,
      "theoretical_loss": 3.8649581006263527,
      "tokens_seen": 563937280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041873620862587765,
      "loss": 2.8536,
      "theoretical_loss": 3.8649107250032646,
      "tokens_seen": 564002816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041872617853560683,
      "loss": 2.6021,
      "theoretical_loss": 3.8648633564259915,
      "tokens_seen": 564068352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418716148445336,
      "loss": 2.8927,
      "theoretical_loss": 3.864815994892668,
      "tokens_seen": 564133888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041870611835506525,
      "loss": 2.6526,
      "theoretical_loss": 3.8647686404014276,
      "tokens_seen": 564199424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186960882647944,
      "loss": 2.9926,
      "theoretical_loss": 3.8647212929504065,
      "tokens_seen": 564264960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186860581745236,
      "loss": 2.8074,
      "theoretical_loss": 3.8646739525377396,
      "tokens_seen": 564330496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041867602808425274,
      "loss": 2.8183,
      "theoretical_loss": 3.864626619161564,
      "tokens_seen": 564396032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418665997993982,
      "loss": 3.0095,
      "theoretical_loss": 3.8645792928200167,
      "tokens_seen": 564461568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041865596790371115,
      "loss": 2.9241,
      "theoretical_loss": 3.8645319735112356,
      "tokens_seen": 564527104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041864593781344034,
      "loss": 2.8379,
      "theoretical_loss": 3.8644846612333597,
      "tokens_seen": 564592640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186359077231695,
      "loss": 2.9107,
      "theoretical_loss": 3.864437355984528,
      "tokens_seen": 564658176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186258776328987,
      "loss": 2.6939,
      "theoretical_loss": 3.864390057762881,
      "tokens_seen": 564723712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186158475426279,
      "loss": 2.601,
      "theoretical_loss": 3.8643427665665593,
      "tokens_seen": 564789248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186058174523571,
      "loss": 2.9561,
      "theoretical_loss": 3.864295482393704,
      "tokens_seen": 564854784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041859578736208624,
      "loss": 2.6425,
      "theoretical_loss": 3.8642482052424576,
      "tokens_seen": 564920320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185857572718155,
      "loss": 2.8036,
      "theoretical_loss": 3.864200935110963,
      "tokens_seen": 564985856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185757271815446,
      "loss": 2.6034,
      "theoretical_loss": 3.8641536719973644,
      "tokens_seen": 565051392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041856569709127384,
      "loss": 2.6046,
      "theoretical_loss": 3.8641064158998057,
      "tokens_seen": 565116928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418555667001003,
      "loss": 2.9108,
      "theoretical_loss": 3.864059166816431,
      "tokens_seen": 565182464
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 321728,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3612844944000244,
      "objective/train/theoretical_loss": 3.8640119247453875,
      "objective/train/tokens_used": 585708000,
      "theoretical_loss": 3.8640119247453875,
      "tokens_seen": 565248000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185456369107322,
      "loss": 2.8488,
      "theoretical_loss": 3.8640119247453875,
      "tokens_seen": 565248000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185356068204614,
      "loss": 2.7459,
      "theoretical_loss": 3.863964689684821,
      "tokens_seen": 565313536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185255767301906,
      "loss": 2.7467,
      "theoretical_loss": 3.8639174616328784,
      "tokens_seen": 565379072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041851554663991974,
      "loss": 2.5563,
      "theoretical_loss": 3.863870240587708,
      "tokens_seen": 565444608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418505516549649,
      "loss": 2.7328,
      "theoretical_loss": 3.8638230265474585,
      "tokens_seen": 565510144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184954864593781,
      "loss": 2.6208,
      "theoretical_loss": 3.863775819510278,
      "tokens_seen": 565575680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041848545636910734,
      "loss": 2.6409,
      "theoretical_loss": 3.863728619474318,
      "tokens_seen": 565641216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184754262788365,
      "loss": 2.8141,
      "theoretical_loss": 3.8636814264377284,
      "tokens_seen": 565706752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184653961885657,
      "loss": 2.6808,
      "theoretical_loss": 3.8636342403986603,
      "tokens_seen": 565772288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184553660982949,
      "loss": 2.676,
      "theoretical_loss": 3.8635870613552665,
      "tokens_seen": 565837824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041844533600802407,
      "loss": 2.9051,
      "theoretical_loss": 3.863539889305699,
      "tokens_seen": 565903360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041843530591775325,
      "loss": 2.8133,
      "theoretical_loss": 3.863492724248112,
      "tokens_seen": 565968896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184252758274825,
      "loss": 2.7888,
      "theoretical_loss": 3.8634455661806593,
      "tokens_seen": 566034432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184152457372116,
      "loss": 2.9215,
      "theoretical_loss": 3.8633984151014955,
      "tokens_seen": 566099968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041840521564694085,
      "loss": 2.8133,
      "theoretical_loss": 3.863351271008777,
      "tokens_seen": 566165504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041839518555666997,
      "loss": 2.8181,
      "theoretical_loss": 3.863304133900659,
      "tokens_seen": 566231040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183851554663992,
      "loss": 2.6397,
      "theoretical_loss": 3.863257003775299,
      "tokens_seen": 566296576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183751253761284,
      "loss": 3.028,
      "theoretical_loss": 3.8632098806308557,
      "tokens_seen": 566362112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041836509528585757,
      "loss": 2.5726,
      "theoretical_loss": 3.8631627644654856,
      "tokens_seen": 566427648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041835506519558675,
      "loss": 2.7655,
      "theoretical_loss": 3.8631156552773493,
      "tokens_seen": 566493184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418345035105316,
      "loss": 2.7821,
      "theoretical_loss": 3.8630685530646054,
      "tokens_seen": 566558720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183350050150451,
      "loss": 2.9222,
      "theoretical_loss": 3.8630214578254156,
      "tokens_seen": 566624256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041832497492477435,
      "loss": 2.7699,
      "theoretical_loss": 3.8629743695579393,
      "tokens_seen": 566689792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183149448345035,
      "loss": 2.8905,
      "theoretical_loss": 3.86292728826034,
      "tokens_seen": 566755328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183049147442327,
      "loss": 2.8477,
      "theoretical_loss": 3.8628802139307803,
      "tokens_seen": 566820864
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 322867,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6280651092529297,
      "objective/train/theoretical_loss": 3.8628331465674224,
      "objective/train/tokens_used": 587346400,
      "theoretical_loss": 3.8628331465674224,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182948846539619,
      "loss": 2.8288,
      "theoretical_loss": 3.8628331465674224,
      "tokens_seen": 566886400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182848545636911,
      "loss": 2.683,
      "theoretical_loss": 3.8627860861684304,
      "tokens_seen": 566951936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041827482447342025,
      "loss": 2.6811,
      "theoretical_loss": 3.8627390327319695,
      "tokens_seen": 567017472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041826479438314944,
      "loss": 2.9301,
      "theoretical_loss": 3.8626919862562055,
      "tokens_seen": 567083008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182547642928786,
      "loss": 2.6402,
      "theoretical_loss": 3.8626449467393034,
      "tokens_seen": 567148544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041824473420260785,
      "loss": 2.9405,
      "theoretical_loss": 3.86259791417943,
      "tokens_seen": 567214080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418234704112337,
      "loss": 2.5639,
      "theoretical_loss": 3.862550888574753,
      "tokens_seen": 567279616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182246740220662,
      "loss": 2.9047,
      "theoretical_loss": 3.862503869923441,
      "tokens_seen": 567345152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182146439317954,
      "loss": 2.5505,
      "theoretical_loss": 3.8624568582236622,
      "tokens_seen": 567410688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182046138415246,
      "loss": 2.625,
      "theoretical_loss": 3.8624098534735865,
      "tokens_seen": 567476224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041819458375125376,
      "loss": 2.9179,
      "theoretical_loss": 3.8623628556713836,
      "tokens_seen": 567541760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041818455366098294,
      "loss": 2.8786,
      "theoretical_loss": 3.8623158648152254,
      "tokens_seen": 567607296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181745235707121,
      "loss": 2.6793,
      "theoretical_loss": 3.862268880903282,
      "tokens_seen": 567672832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041816449348044136,
      "loss": 2.6765,
      "theoretical_loss": 3.8622219039337273,
      "tokens_seen": 567738368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041815446339017054,
      "loss": 3.0911,
      "theoretical_loss": 3.8621749339047335,
      "tokens_seen": 567803904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181444332998997,
      "loss": 2.6648,
      "theoretical_loss": 3.862127970814474,
      "tokens_seen": 567869440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181344032096289,
      "loss": 2.6876,
      "theoretical_loss": 3.8620810146611233,
      "tokens_seen": 567934976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181243731193581,
      "loss": 3.0197,
      "theoretical_loss": 3.862034065442857,
      "tokens_seen": 568000512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181143430290873,
      "loss": 2.8192,
      "theoretical_loss": 3.8619871231578498,
      "tokens_seen": 568066048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041810431293881644,
      "loss": 2.8597,
      "theoretical_loss": 3.8619401878042794,
      "tokens_seen": 568131584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180942828485457,
      "loss": 2.7689,
      "theoretical_loss": 3.861893259380322,
      "tokens_seen": 568197120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180842527582748,
      "loss": 2.9388,
      "theoretical_loss": 3.8618463378841557,
      "tokens_seen": 568262656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041807422266800404,
      "loss": 2.7452,
      "theoretical_loss": 3.861799423313959,
      "tokens_seen": 568328192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180641925777332,
      "loss": 2.7699,
      "theoretical_loss": 3.861752515667911,
      "tokens_seen": 568393728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180541624874624,
      "loss": 2.7727,
      "theoretical_loss": 3.861705614944192,
      "tokens_seen": 568459264
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 324005,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.827101230621338,
      "objective/train/theoretical_loss": 3.8616587211409823,
      "objective/train/tokens_used": 588984800,
      "theoretical_loss": 3.8616587211409823,
      "tokens_seen": 568524800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180441323971916,
      "loss": 2.78,
      "theoretical_loss": 3.8616587211409823,
      "tokens_seen": 568524800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180341023069208,
      "loss": 2.9464,
      "theoretical_loss": 3.861611834256463,
      "tokens_seen": 568590336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041802407221664994,
      "loss": 2.7669,
      "theoretical_loss": 3.8615649542888164,
      "tokens_seen": 568655872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180140421263792,
      "loss": 2.9159,
      "theoretical_loss": 3.861518081236224,
      "tokens_seen": 568721408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180040120361083,
      "loss": 2.6733,
      "theoretical_loss": 3.861471215096871,
      "tokens_seen": 568786944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041799398194583754,
      "loss": 2.6195,
      "theoretical_loss": 3.86142435586894,
      "tokens_seen": 568852480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179839518555667,
      "loss": 2.5565,
      "theoretical_loss": 3.8613775035506164,
      "tokens_seen": 568918016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179739217652959,
      "loss": 2.9737,
      "theoretical_loss": 3.8613306581400852,
      "tokens_seen": 568983552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179638916750251,
      "loss": 2.9827,
      "theoretical_loss": 3.861283819635532,
      "tokens_seen": 569049088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041795386158475427,
      "loss": 2.5987,
      "theoretical_loss": 3.861236988035145,
      "tokens_seen": 569114624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041794383149448345,
      "loss": 2.8994,
      "theoretical_loss": 3.86119016333711,
      "tokens_seen": 569180160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179338014042127,
      "loss": 2.9604,
      "theoretical_loss": 3.861143345539616,
      "tokens_seen": 569245696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179237713139418,
      "loss": 2.5996,
      "theoretical_loss": 3.861096534640852,
      "tokens_seen": 569311232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041791374122367105,
      "loss": 2.7902,
      "theoretical_loss": 3.8610497306390066,
      "tokens_seen": 569376768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041790371113340017,
      "loss": 2.6739,
      "theoretical_loss": 3.861002933532271,
      "tokens_seen": 569442304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178936810431294,
      "loss": 2.668,
      "theoretical_loss": 3.860956143318835,
      "tokens_seen": 569507840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178836509528586,
      "loss": 2.9281,
      "theoretical_loss": 3.8609093599968913,
      "tokens_seen": 569573376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041787362086258777,
      "loss": 2.9559,
      "theoretical_loss": 3.8608625835646313,
      "tokens_seen": 569638912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041786359077231695,
      "loss": 2.6292,
      "theoretical_loss": 3.860815814020248,
      "tokens_seen": 569704448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178535606820462,
      "loss": 2.8555,
      "theoretical_loss": 3.860769051361935,
      "tokens_seen": 569769984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178435305917753,
      "loss": 2.6426,
      "theoretical_loss": 3.8607222955878866,
      "tokens_seen": 569835520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041783350050150455,
      "loss": 2.7501,
      "theoretical_loss": 3.860675546696298,
      "tokens_seen": 569901056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178234704112337,
      "loss": 2.9662,
      "theoretical_loss": 3.8606288046853647,
      "tokens_seen": 569966592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178134403209629,
      "loss": 2.7383,
      "theoretical_loss": 3.8605820695532826,
      "tokens_seen": 570032128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178034102306921,
      "loss": 2.6567,
      "theoretical_loss": 3.860535341298249,
      "tokens_seen": 570097664
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 324627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.945570468902588,
      "objective/train/theoretical_loss": 3.860488619918462,
      "objective/train/tokens_used": 590623200,
      "theoretical_loss": 3.860488619918462,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177933801404213,
      "loss": 2.8017,
      "theoretical_loss": 3.860488619918462,
      "tokens_seen": 570163200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041778335005015045,
      "loss": 3.0509,
      "theoretical_loss": 3.8604419054121193,
      "tokens_seen": 570228736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041777331995987964,
      "loss": 2.8242,
      "theoretical_loss": 3.86039519777742,
      "tokens_seen": 570294272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177632898696088,
      "loss": 2.59,
      "theoretical_loss": 3.8603484970125637,
      "tokens_seen": 570359808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041775325977933805,
      "loss": 2.8078,
      "theoretical_loss": 3.8603018031157514,
      "tokens_seen": 570425344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177432296890672,
      "loss": 3.0908,
      "theoretical_loss": 3.8602551160851837,
      "tokens_seen": 570490880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177331995987964,
      "loss": 2.8318,
      "theoretical_loss": 3.8602084359190627,
      "tokens_seen": 570556416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177231695085256,
      "loss": 2.9725,
      "theoretical_loss": 3.86016176261559,
      "tokens_seen": 570621952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177131394182548,
      "loss": 2.67,
      "theoretical_loss": 3.8601150961729696,
      "tokens_seen": 570687488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041770310932798396,
      "loss": 2.6319,
      "theoretical_loss": 3.8600684365894047,
      "tokens_seen": 570753024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041769307923771314,
      "loss": 2.8908,
      "theoretical_loss": 3.8600217838631,
      "tokens_seen": 570818560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176830491474423,
      "loss": 2.7601,
      "theoretical_loss": 3.85997513799226,
      "tokens_seen": 570884096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041767301905717156,
      "loss": 2.9639,
      "theoretical_loss": 3.859928498975092,
      "tokens_seen": 570949632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176629889669007,
      "loss": 2.9724,
      "theoretical_loss": 3.859881866809801,
      "tokens_seen": 571015168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176529588766299,
      "loss": 2.8134,
      "theoretical_loss": 3.859835241494595,
      "tokens_seen": 571080704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041764292878635904,
      "loss": 2.7331,
      "theoretical_loss": 3.859788623027681,
      "tokens_seen": 571146240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176328986960883,
      "loss": 2.6944,
      "theoretical_loss": 3.859742011407268,
      "tokens_seen": 571211776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041762286860581746,
      "loss": 2.8846,
      "theoretical_loss": 3.859695406631565,
      "tokens_seen": 571277312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041761283851554664,
      "loss": 2.7068,
      "theoretical_loss": 3.8596488086987826,
      "tokens_seen": 571342848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176028084252758,
      "loss": 2.7232,
      "theoretical_loss": 3.8596022176071303,
      "tokens_seen": 571408384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417592778335005,
      "loss": 2.7892,
      "theoretical_loss": 3.8595556333548195,
      "tokens_seen": 571473920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175827482447342,
      "loss": 2.4485,
      "theoretical_loss": 3.859509055940063,
      "tokens_seen": 571539456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175727181544634,
      "loss": 2.5149,
      "theoretical_loss": 3.859462485361072,
      "tokens_seen": 571604992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041756268806419255,
      "loss": 2.6744,
      "theoretical_loss": 3.85941592161606,
      "tokens_seen": 571670528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175526579739218,
      "loss": 2.7721,
      "theoretical_loss": 3.859369364703242,
      "tokens_seen": 571736064
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 325757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.774052619934082,
      "objective/train/theoretical_loss": 3.8593228146208314,
      "objective/train/tokens_used": 592261600,
      "theoretical_loss": 3.8593228146208314,
      "tokens_seen": 571801600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041754262788365096,
      "loss": 2.8318,
      "theoretical_loss": 3.8593228146208314,
      "tokens_seen": 571801600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041753259779338015,
      "loss": 2.966,
      "theoretical_loss": 3.859276271367044,
      "tokens_seen": 571867136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175225677031093,
      "loss": 2.7908,
      "theoretical_loss": 3.859229734940095,
      "tokens_seen": 571932672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175125376128385,
      "loss": 2.5661,
      "theoretical_loss": 3.8591832053382014,
      "tokens_seen": 571998208
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175025075225677,
      "loss": 2.7081,
      "theoretical_loss": 3.8591366825595808,
      "tokens_seen": 572063744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174924774322969,
      "loss": 2.621,
      "theoretical_loss": 3.85909016660245,
      "tokens_seen": 572129280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041748244734202605,
      "loss": 2.8749,
      "theoretical_loss": 3.8590436574650293,
      "tokens_seen": 572194816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174724172517553,
      "loss": 2.705,
      "theoretical_loss": 3.858997155145537,
      "tokens_seen": 572260352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174623871614844,
      "loss": 2.7483,
      "theoretical_loss": 3.858950659642192,
      "tokens_seen": 572325888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041745235707121365,
      "loss": 2.7732,
      "theoretical_loss": 3.8589041709532164,
      "tokens_seen": 572391424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041744232698094283,
      "loss": 2.5971,
      "theoretical_loss": 3.8588576890768307,
      "tokens_seen": 572456960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417432296890672,
      "loss": 2.672,
      "theoretical_loss": 3.858811214011258,
      "tokens_seen": 572522496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174222668004012,
      "loss": 2.8971,
      "theoretical_loss": 3.8587647457547183,
      "tokens_seen": 572588032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041741223671013037,
      "loss": 2.6925,
      "theoretical_loss": 3.8587182843054375,
      "tokens_seen": 572653568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174022066198596,
      "loss": 2.6789,
      "theoretical_loss": 3.8586718296616382,
      "tokens_seen": 572719104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173921765295888,
      "loss": 2.8008,
      "theoretical_loss": 3.858625381821545,
      "tokens_seen": 572784640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041738214643931797,
      "loss": 2.7738,
      "theoretical_loss": 3.858578940783384,
      "tokens_seen": 572850176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041737211634904715,
      "loss": 2.721,
      "theoretical_loss": 3.85853250654538,
      "tokens_seen": 572915712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173620862587764,
      "loss": 2.7751,
      "theoretical_loss": 3.8584860791057602,
      "tokens_seen": 572981248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173520561685055,
      "loss": 2.5138,
      "theoretical_loss": 3.8584396584627516,
      "tokens_seen": 573046784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041734202607823475,
      "loss": 2.7566,
      "theoretical_loss": 3.8583932446145828,
      "tokens_seen": 573112320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173319959879639,
      "loss": 2.808,
      "theoretical_loss": 3.8583468375594814,
      "tokens_seen": 573177856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173219658976931,
      "loss": 2.7426,
      "theoretical_loss": 3.858300437295677,
      "tokens_seen": 573243392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173119358074223,
      "loss": 2.7867,
      "theoretical_loss": 3.8582540438214,
      "tokens_seen": 573308928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173019057171515,
      "loss": 2.8751,
      "theoretical_loss": 3.85820765713488,
      "tokens_seen": 573374464
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 326467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8871238231658936,
      "objective/train/theoretical_loss": 3.858161277234349,
      "objective/train/tokens_used": 593900000,
      "theoretical_loss": 3.858161277234349,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041729187562688065,
      "loss": 2.8817,
      "theoretical_loss": 3.858161277234349,
      "tokens_seen": 573440000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041728184553660984,
      "loss": 2.7829,
      "theoretical_loss": 3.8581149041180387,
      "tokens_seen": 573505536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417271815446339,
      "loss": 2.7302,
      "theoretical_loss": 3.858068537784182,
      "tokens_seen": 573571072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041726178535606825,
      "loss": 2.8625,
      "theoretical_loss": 3.8580221782310113,
      "tokens_seen": 573636608
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172517552657974,
      "loss": 2.8321,
      "theoretical_loss": 3.8579758254567613,
      "tokens_seen": 573702144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172417251755266,
      "loss": 2.7675,
      "theoretical_loss": 3.8579294794596657,
      "tokens_seen": 573767680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172316950852558,
      "loss": 2.8152,
      "theoretical_loss": 3.85788314023796,
      "tokens_seen": 573833216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417221664994985,
      "loss": 2.8103,
      "theoretical_loss": 3.857836807789881,
      "tokens_seen": 573898752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041721163490471416,
      "loss": 2.833,
      "theoretical_loss": 3.857790482113664,
      "tokens_seen": 573964288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041720160481444334,
      "loss": 2.886,
      "theoretical_loss": 3.8577441632075473,
      "tokens_seen": 574029824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171915747241725,
      "loss": 2.9433,
      "theoretical_loss": 3.8576978510697675,
      "tokens_seen": 574095360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041718154463390176,
      "loss": 2.8847,
      "theoretical_loss": 3.8576515456985634,
      "tokens_seen": 574160896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171715145436309,
      "loss": 2.6894,
      "theoretical_loss": 3.857605247092175,
      "tokens_seen": 574226432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171614844533601,
      "loss": 2.6527,
      "theoretical_loss": 3.857558955248842,
      "tokens_seen": 574291968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041715145436308924,
      "loss": 2.6139,
      "theoretical_loss": 3.8575126701668037,
      "tokens_seen": 574357504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171414242728185,
      "loss": 2.9821,
      "theoretical_loss": 3.8574663918443024,
      "tokens_seen": 574423040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041713139418254766,
      "loss": 2.7302,
      "theoretical_loss": 3.857420120279579,
      "tokens_seen": 574488576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041712136409227684,
      "loss": 2.6908,
      "theoretical_loss": 3.857373855470877,
      "tokens_seen": 574554112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417111334002006,
      "loss": 2.8465,
      "theoretical_loss": 3.8573275974164387,
      "tokens_seen": 574619648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171013039117352,
      "loss": 2.6284,
      "theoretical_loss": 3.8572813461145086,
      "tokens_seen": 574685184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170912738214644,
      "loss": 2.7787,
      "theoretical_loss": 3.8572351015633304,
      "tokens_seen": 574750720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170812437311936,
      "loss": 2.6762,
      "theoretical_loss": 3.8571888637611496,
      "tokens_seen": 574816256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041707121364092275,
      "loss": 2.9144,
      "theoretical_loss": 3.857142632706212,
      "tokens_seen": 574881792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417061183550652,
      "loss": 2.8601,
      "theoretical_loss": 3.857096408396764,
      "tokens_seen": 574947328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041705115346038116,
      "loss": 2.6951,
      "theoretical_loss": 3.8570501908310515,
      "tokens_seen": 575012864
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 327672,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.893551826477051,
      "objective/train/theoretical_loss": 3.857003980007324,
      "objective/train/tokens_used": 595538400,
      "theoretical_loss": 3.857003980007324,
      "tokens_seen": 575078400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041704112337011035,
      "loss": 2.8592,
      "theoretical_loss": 3.857003980007324,
      "tokens_seen": 575078400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170310932798395,
      "loss": 2.7133,
      "theoretical_loss": 3.8569577759238287,
      "tokens_seen": 575143936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170210631895687,
      "loss": 2.7563,
      "theoretical_loss": 3.8569115785788153,
      "tokens_seen": 575209472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170110330992979,
      "loss": 2.9113,
      "theoretical_loss": 3.856865387970533,
      "tokens_seen": 575275008
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170010030090271,
      "loss": 2.8067,
      "theoretical_loss": 3.8568192040972322,
      "tokens_seen": 575340544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041699097291875625,
      "loss": 2.5334,
      "theoretical_loss": 3.856773026957164,
      "tokens_seen": 575406080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169809428284855,
      "loss": 2.765,
      "theoretical_loss": 3.8567268565485806,
      "tokens_seen": 575471616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169709127382146,
      "loss": 2.7038,
      "theoretical_loss": 3.856680692869733,
      "tokens_seen": 575537152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041696088264794385,
      "loss": 2.8842,
      "theoretical_loss": 3.856634535918875,
      "tokens_seen": 575602688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041695085255767303,
      "loss": 2.7644,
      "theoretical_loss": 3.85658838569426,
      "tokens_seen": 575668224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169408224674022,
      "loss": 2.7948,
      "theoretical_loss": 3.856542242194142,
      "tokens_seen": 575733760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169307923771314,
      "loss": 2.7855,
      "theoretical_loss": 3.8564961054167766,
      "tokens_seen": 575799296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169207622868606,
      "loss": 2.8138,
      "theoretical_loss": 3.8564499753604187,
      "tokens_seen": 575864832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041691073219658975,
      "loss": 2.8032,
      "theoretical_loss": 3.856403852023325,
      "tokens_seen": 575930368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000416900702106319,
      "loss": 2.9167,
      "theoretical_loss": 3.8563577354037513,
      "tokens_seen": 575995904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168906720160481,
      "loss": 3.0506,
      "theoretical_loss": 3.8563116254999565,
      "tokens_seen": 576061440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041688064192577735,
      "loss": 3.0837,
      "theoretical_loss": 3.856265522310198,
      "tokens_seen": 576126976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041687061183550653,
      "loss": 2.6939,
      "theoretical_loss": 3.8562194258327347,
      "tokens_seen": 576192512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168605817452357,
      "loss": 2.6713,
      "theoretical_loss": 3.856173336065826,
      "tokens_seen": 576258048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168505516549649,
      "loss": 2.8135,
      "theoretical_loss": 3.856127253007732,
      "tokens_seen": 576323584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168405215646941,
      "loss": 2.6818,
      "theoretical_loss": 3.8560811766567133,
      "tokens_seen": 576389120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041683049147442326,
      "loss": 2.9436,
      "theoretical_loss": 3.856035107011032,
      "tokens_seen": 576454656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168204613841525,
      "loss": 2.7948,
      "theoretical_loss": 3.8559890440689495,
      "tokens_seen": 576520192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168104312938816,
      "loss": 2.7917,
      "theoretical_loss": 3.855942987828728,
      "tokens_seen": 576585728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041680040120361085,
      "loss": 2.6469,
      "theoretical_loss": 3.855896938288632,
      "tokens_seen": 576651264
    },
    {
      "epoch": 0.17,
      "objective/train/docs_used": 328706,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.591416597366333,
      "objective/train/theoretical_loss": 3.855850895446925,
      "objective/train/tokens_used": 597176800,
      "theoretical_loss": 3.855850895446925,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041679037111334,
      "loss": 2.5328,
      "theoretical_loss": 3.855850895446925,
      "tokens_seen": 576716800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167803410230692,
      "loss": 2.7952,
      "theoretical_loss": 3.8558048593018714,
      "tokens_seen": 576782336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167703109327984,
      "loss": 2.6321,
      "theoretical_loss": 3.855758829851737,
      "tokens_seen": 576847872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167602808425276,
      "loss": 2.7992,
      "theoretical_loss": 3.855712807094787,
      "tokens_seen": 576913408
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041675025075225676,
      "loss": 2.8481,
      "theoretical_loss": 3.8556667910292886,
      "tokens_seen": 576978944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000416740220661986,
      "loss": 2.6563,
      "theoretical_loss": 3.855620781653509,
      "tokens_seen": 577044480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167301905717151,
      "loss": 2.8306,
      "theoretical_loss": 3.8555747789657158,
      "tokens_seen": 577110016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041672016048144436,
      "loss": 2.8782,
      "theoretical_loss": 3.855528782964177,
      "tokens_seen": 577175552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167101303911735,
      "loss": 2.6459,
      "theoretical_loss": 3.8554827936471625,
      "tokens_seen": 577241088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167001003009027,
      "loss": 2.6478,
      "theoretical_loss": 3.855436811012942,
      "tokens_seen": 577306624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166900702106319,
      "loss": 3.059,
      "theoretical_loss": 3.8553908350597865,
      "tokens_seen": 577372160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166800401203611,
      "loss": 2.7494,
      "theoretical_loss": 3.855344865785966,
      "tokens_seen": 577437696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041667001003009026,
      "loss": 2.936,
      "theoretical_loss": 3.8552989031897527,
      "tokens_seen": 577503232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041665997993981944,
      "loss": 2.4975,
      "theoretical_loss": 3.855252947269419,
      "tokens_seen": 577568768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166499498495487,
      "loss": 2.5769,
      "theoretical_loss": 3.8552069980232377,
      "tokens_seen": 577634304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041663991975927786,
      "loss": 2.8171,
      "theoretical_loss": 3.855161055449483,
      "tokens_seen": 577699840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041662988966900704,
      "loss": 2.7957,
      "theoretical_loss": 3.8551151195464284,
      "tokens_seen": 577765376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166198595787362,
      "loss": 2.8499,
      "theoretical_loss": 3.8550691903123493,
      "tokens_seen": 577830912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166098294884654,
      "loss": 2.7152,
      "theoretical_loss": 3.855023267745522,
      "tokens_seen": 577896448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165997993981946,
      "loss": 2.7041,
      "theoretical_loss": 3.854977351844221,
      "tokens_seen": 577961984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165897693079238,
      "loss": 2.7119,
      "theoretical_loss": 3.8549314426067243,
      "tokens_seen": 578027520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041657973921765295,
      "loss": 2.9118,
      "theoretical_loss": 3.85488554003131,
      "tokens_seen": 578093056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165697091273822,
      "loss": 2.7783,
      "theoretical_loss": 3.854839644116255,
      "tokens_seen": 578158592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041655967903711136,
      "loss": 2.8074,
      "theoretical_loss": 3.8547937548598386,
      "tokens_seen": 578224128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041654964894684055,
      "loss": 2.9678,
      "theoretical_loss": 3.8547478722603405,
      "tokens_seen": 578289664
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 329258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3565433025360107,
      "objective/train/theoretical_loss": 3.8547019963160407,
      "objective/train/tokens_used": 598815200,
      "theoretical_loss": 3.8547019963160407,
      "tokens_seen": 578355200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165396188565697,
      "loss": 2.7488,
      "theoretical_loss": 3.8547019963160407,
      "tokens_seen": 578355200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165295887662989,
      "loss": 2.957,
      "theoretical_loss": 3.854656127025219,
      "tokens_seen": 578420736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165195586760281,
      "loss": 2.5597,
      "theoretical_loss": 3.854610264386158,
      "tokens_seen": 578486272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165095285857573,
      "loss": 2.8326,
      "theoretical_loss": 3.854564408397139,
      "tokens_seen": 578551808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041649949849548645,
      "loss": 2.8244,
      "theoretical_loss": 3.854518559056445,
      "tokens_seen": 578617344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164894684052157,
      "loss": 2.7684,
      "theoretical_loss": 3.8544727163623587,
      "tokens_seen": 578682880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164794383149448,
      "loss": 2.878,
      "theoretical_loss": 3.8544268803131643,
      "tokens_seen": 578748416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041646940822467405,
      "loss": 2.7852,
      "theoretical_loss": 3.8543810509071466,
      "tokens_seen": 578813952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041645937813440323,
      "loss": 2.7624,
      "theoretical_loss": 3.8543352281425904,
      "tokens_seen": 578879488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164493480441324,
      "loss": 2.6575,
      "theoretical_loss": 3.8542894120177817,
      "tokens_seen": 578945024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164393179538616,
      "loss": 2.8381,
      "theoretical_loss": 3.8542436025310067,
      "tokens_seen": 579010560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164292878635908,
      "loss": 2.6767,
      "theoretical_loss": 3.854197799680553,
      "tokens_seen": 579076096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041641925777331995,
      "loss": 2.9494,
      "theoretical_loss": 3.854152003464708,
      "tokens_seen": 579141632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164092276830492,
      "loss": 2.8045,
      "theoretical_loss": 3.85410621388176,
      "tokens_seen": 579207168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163991975927783,
      "loss": 2.8555,
      "theoretical_loss": 3.854060430929998,
      "tokens_seen": 579272704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041638916750250755,
      "loss": 3.0175,
      "theoretical_loss": 3.854014654607712,
      "tokens_seen": 579338240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041637913741223673,
      "loss": 2.8226,
      "theoretical_loss": 3.8539688849131917,
      "tokens_seen": 579403776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163691073219659,
      "loss": 3.0263,
      "theoretical_loss": 3.853923121844728,
      "tokens_seen": 579469312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163590772316951,
      "loss": 2.6421,
      "theoretical_loss": 3.8538773654006127,
      "tokens_seen": 579534848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163490471414243,
      "loss": 2.8617,
      "theoretical_loss": 3.8538316155791383,
      "tokens_seen": 579600384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041633901705115346,
      "loss": 2.6148,
      "theoretical_loss": 3.853785872378597,
      "tokens_seen": 579665920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163289869608827,
      "loss": 2.7086,
      "theoretical_loss": 3.853740135797282,
      "tokens_seen": 579731456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163189568706118,
      "loss": 2.9555,
      "theoretical_loss": 3.8536944058334885,
      "tokens_seen": 579796992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041630892678034105,
      "loss": 2.782,
      "theoretical_loss": 3.85364868248551,
      "tokens_seen": 579862528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162988966900702,
      "loss": 3.079,
      "theoretical_loss": 3.8536029657516426,
      "tokens_seen": 579928064
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 329612,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.266913890838623,
      "objective/train/theoretical_loss": 3.8535572556301823,
      "objective/train/tokens_used": 600453600,
      "theoretical_loss": 3.8535572556301823,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162888665997994,
      "loss": 2.9089,
      "theoretical_loss": 3.8535572556301823,
      "tokens_seen": 579993600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162788365095286,
      "loss": 2.8912,
      "theoretical_loss": 3.8535115521194245,
      "tokens_seen": 580059136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162688064192578,
      "loss": 2.857,
      "theoretical_loss": 3.8534658552176677,
      "tokens_seen": 580124672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041625877632898696,
      "loss": 2.8663,
      "theoretical_loss": 3.853420164923209,
      "tokens_seen": 580190208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162487462387162,
      "loss": 2.6089,
      "theoretical_loss": 3.853374481234348,
      "tokens_seen": 580255744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162387161484453,
      "loss": 3.0777,
      "theoretical_loss": 3.8533288041493825,
      "tokens_seen": 580321280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041622868605817456,
      "loss": 2.7666,
      "theoretical_loss": 3.853283133666613,
      "tokens_seen": 580386816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162186559679037,
      "loss": 2.8469,
      "theoretical_loss": 3.8532374697843395,
      "tokens_seen": 580452352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162086258776329,
      "loss": 3.1069,
      "theoretical_loss": 3.853191812500863,
      "tokens_seen": 580517888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161985957873621,
      "loss": 2.5315,
      "theoretical_loss": 3.853146161814485,
      "tokens_seen": 580583424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161885656970913,
      "loss": 3.1341,
      "theoretical_loss": 3.853100517723509,
      "tokens_seen": 580648960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041617853560682046,
      "loss": 2.8452,
      "theoretical_loss": 3.853054880226236,
      "tokens_seen": 580714496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041616850551654964,
      "loss": 2.642,
      "theoretical_loss": 3.8530092493209707,
      "tokens_seen": 580780032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161584754262788,
      "loss": 2.9064,
      "theoretical_loss": 3.8529636250060166,
      "tokens_seen": 580845568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041614844533600806,
      "loss": 2.8612,
      "theoretical_loss": 3.85291800727968,
      "tokens_seen": 580911104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161384152457372,
      "loss": 3.071,
      "theoretical_loss": 3.8528723961402638,
      "tokens_seen": 580976640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161283851554664,
      "loss": 2.9736,
      "theoretical_loss": 3.852826791586076,
      "tokens_seen": 581042176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041611835506519555,
      "loss": 2.659,
      "theoretical_loss": 3.8527811936154217,
      "tokens_seen": 581107712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161083249749248,
      "loss": 2.744,
      "theoretical_loss": 3.8527356022266104,
      "tokens_seen": 581173248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041609829488465397,
      "loss": 2.8711,
      "theoretical_loss": 3.852690017417948,
      "tokens_seen": 581238784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041608826479438315,
      "loss": 3.0193,
      "theoretical_loss": 3.8526444391877437,
      "tokens_seen": 581304320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041607823470411233,
      "loss": 2.7875,
      "theoretical_loss": 3.852598867534307,
      "tokens_seen": 581369856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041606820461384156,
      "loss": 2.6953,
      "theoretical_loss": 3.852553302455947,
      "tokens_seen": 581435392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160581745235707,
      "loss": 3.0129,
      "theoretical_loss": 3.8525077439509747,
      "tokens_seen": 581500928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160481444332999,
      "loss": 2.7702,
      "theoretical_loss": 3.8524621920177005,
      "tokens_seen": 581566464
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 330311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.916626453399658,
      "objective/train/theoretical_loss": 3.852416646654437,
      "objective/train/tokens_used": 602092000,
      "theoretical_loss": 3.852416646654437,
      "tokens_seen": 581632000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041603811434302905,
      "loss": 2.8318,
      "theoretical_loss": 3.852416646654437,
      "tokens_seen": 581632000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160280842527583,
      "loss": 2.5899,
      "theoretical_loss": 3.8523711078594953,
      "tokens_seen": 581697536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041601805416248747,
      "loss": 2.9073,
      "theoretical_loss": 3.8523255756311894,
      "tokens_seen": 581763072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041600802407221665,
      "loss": 2.9442,
      "theoretical_loss": 3.852280049967832,
      "tokens_seen": 581828608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041599799398194583,
      "loss": 2.9509,
      "theoretical_loss": 3.8522345308677375,
      "tokens_seen": 581894144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415987963891675,
      "loss": 2.5574,
      "theoretical_loss": 3.8521890183292213,
      "tokens_seen": 581959680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159779338014042,
      "loss": 2.918,
      "theoretical_loss": 3.8521435123505974,
      "tokens_seen": 582025216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041596790371113343,
      "loss": 2.8101,
      "theoretical_loss": 3.8520980129301834,
      "tokens_seen": 582090752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041595787362086256,
      "loss": 2.611,
      "theoretical_loss": 3.852052520066295,
      "tokens_seen": 582156288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159478435305918,
      "loss": 2.7954,
      "theoretical_loss": 3.85200703375725,
      "tokens_seen": 582221824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159378134403209,
      "loss": 2.8191,
      "theoretical_loss": 3.851961554001365,
      "tokens_seen": 582287360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041592778335005015,
      "loss": 2.8662,
      "theoretical_loss": 3.85191608079696,
      "tokens_seen": 582352896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041591775325977934,
      "loss": 2.8765,
      "theoretical_loss": 3.8518706141423538,
      "tokens_seen": 582418432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159077231695085,
      "loss": 2.9141,
      "theoretical_loss": 3.851825154035865,
      "tokens_seen": 582483968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041589769307923775,
      "loss": 2.9011,
      "theoretical_loss": 3.8517797004758156,
      "tokens_seen": 582549504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041588766298896693,
      "loss": 2.7864,
      "theoretical_loss": 3.8517342534605254,
      "tokens_seen": 582615040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158776328986961,
      "loss": 2.6935,
      "theoretical_loss": 3.851688812988317,
      "tokens_seen": 582680576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158676028084253,
      "loss": 2.9189,
      "theoretical_loss": 3.8516433790575118,
      "tokens_seen": 582746112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158575727181545,
      "loss": 3.0231,
      "theoretical_loss": 3.8515979516664327,
      "tokens_seen": 582811648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041584754262788366,
      "loss": 2.7779,
      "theoretical_loss": 3.851552530813403,
      "tokens_seen": 582877184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158375125376129,
      "loss": 2.7526,
      "theoretical_loss": 3.8515071164967476,
      "tokens_seen": 582942720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415827482447342,
      "loss": 2.8509,
      "theoretical_loss": 3.8514617087147904,
      "tokens_seen": 583008256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041581745235707126,
      "loss": 2.9512,
      "theoretical_loss": 3.8514163074658576,
      "tokens_seen": 583073792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158074222668004,
      "loss": 2.8851,
      "theoretical_loss": 3.851370912748274,
      "tokens_seen": 583139328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157973921765296,
      "loss": 2.9043,
      "theoretical_loss": 3.851325524560367,
      "tokens_seen": 583204864
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 331192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3419594764709473,
      "objective/train/theoretical_loss": 3.851280142900463,
      "objective/train/tokens_used": 603730400,
      "theoretical_loss": 3.851280142900463,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157873620862588,
      "loss": 2.7418,
      "theoretical_loss": 3.851280142900463,
      "tokens_seen": 583270400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415777331995988,
      "loss": 2.8718,
      "theoretical_loss": 3.8512347677668908,
      "tokens_seen": 583335936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041576730190571716,
      "loss": 2.7871,
      "theoretical_loss": 3.851189399157978,
      "tokens_seen": 583401472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157572718154464,
      "loss": 2.6968,
      "theoretical_loss": 3.8511440370720535,
      "tokens_seen": 583467008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157472417251755,
      "loss": 2.8935,
      "theoretical_loss": 3.851098681507448,
      "tokens_seen": 583532544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041573721163490476,
      "loss": 2.6811,
      "theoretical_loss": 3.8510533324624907,
      "tokens_seen": 583598080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157271815446339,
      "loss": 2.7665,
      "theoretical_loss": 3.8510079899355123,
      "tokens_seen": 583663616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157171514543631,
      "loss": 3.0478,
      "theoretical_loss": 3.8509626539248454,
      "tokens_seen": 583729152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157071213640923,
      "loss": 2.8275,
      "theoretical_loss": 3.8509173244288206,
      "tokens_seen": 583794688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156970912738215,
      "loss": 2.7585,
      "theoretical_loss": 3.8508720014457722,
      "tokens_seen": 583860224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041568706118355066,
      "loss": 2.8001,
      "theoretical_loss": 3.8508266849740327,
      "tokens_seen": 583925760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041567703109327984,
      "loss": 2.9887,
      "theoretical_loss": 3.8507813750119353,
      "tokens_seen": 583991296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415667001003009,
      "loss": 2.8654,
      "theoretical_loss": 3.8507360715578156,
      "tokens_seen": 584056832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041565697091273826,
      "loss": 2.9364,
      "theoretical_loss": 3.8506907746100087,
      "tokens_seen": 584122368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156469408224674,
      "loss": 2.8208,
      "theoretical_loss": 3.85064548416685,
      "tokens_seen": 584187904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156369107321966,
      "loss": 3.1207,
      "theoretical_loss": 3.850600200226676,
      "tokens_seen": 584253440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041562688064192575,
      "loss": 2.9591,
      "theoretical_loss": 3.8505549227878237,
      "tokens_seen": 584318976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415616850551655,
      "loss": 2.8618,
      "theoretical_loss": 3.8505096518486304,
      "tokens_seen": 584384512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041560682046138417,
      "loss": 2.8048,
      "theoretical_loss": 3.8504643874074347,
      "tokens_seen": 584450048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041559679037111335,
      "loss": 2.7938,
      "theoretical_loss": 3.8504191294625754,
      "tokens_seen": 584515584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041558676028084253,
      "loss": 3.0779,
      "theoretical_loss": 3.850373878012392,
      "tokens_seen": 584581120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041557673019057176,
      "loss": 2.727,
      "theoretical_loss": 3.850328633055224,
      "tokens_seen": 584646656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155667001003009,
      "loss": 3.0438,
      "theoretical_loss": 3.8502833945894124,
      "tokens_seen": 584712192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155566700100301,
      "loss": 2.9217,
      "theoretical_loss": 3.850238162613299,
      "tokens_seen": 584777728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041554663991975925,
      "loss": 2.9176,
      "theoretical_loss": 3.850192937125225,
      "tokens_seen": 584843264
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 331815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.829080104827881,
      "objective/train/theoretical_loss": 3.850147718123533,
      "objective/train/tokens_used": 605368800,
      "theoretical_loss": 3.850147718123533,
      "tokens_seen": 584908800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155366098294885,
      "loss": 2.7848,
      "theoretical_loss": 3.850147718123533,
      "tokens_seen": 584908800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041552657973921767,
      "loss": 2.877,
      "theoretical_loss": 3.850102505606566,
      "tokens_seen": 584974336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041551654964894685,
      "loss": 2.7364,
      "theoretical_loss": 3.8500572995726676,
      "tokens_seen": 585039872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041550651955867603,
      "loss": 2.891,
      "theoretical_loss": 3.850012100020183,
      "tokens_seen": 585105408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154964894684052,
      "loss": 2.779,
      "theoretical_loss": 3.849966906947456,
      "tokens_seen": 585170944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154864593781344,
      "loss": 2.9868,
      "theoretical_loss": 3.8499217203528335,
      "tokens_seen": 585236480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041547642928786363,
      "loss": 2.7509,
      "theoretical_loss": 3.8498765402346597,
      "tokens_seen": 585302016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041546639919759276,
      "loss": 2.851,
      "theoretical_loss": 3.8498313665912827,
      "tokens_seen": 585367552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415456369107322,
      "loss": 2.8037,
      "theoretical_loss": 3.84978619942105,
      "tokens_seen": 585433088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154463390170511,
      "loss": 2.8823,
      "theoretical_loss": 3.8497410387223088,
      "tokens_seen": 585498624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041543630892678035,
      "loss": 3.0639,
      "theoretical_loss": 3.849695884493408,
      "tokens_seen": 585564160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041542627883650954,
      "loss": 2.5332,
      "theoretical_loss": 3.849650736732696,
      "tokens_seen": 585629696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154162487462387,
      "loss": 2.8753,
      "theoretical_loss": 3.849605595438524,
      "tokens_seen": 585695232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154062186559679,
      "loss": 3.111,
      "theoretical_loss": 3.8495604606092417,
      "tokens_seen": 585760768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041539618856569713,
      "loss": 2.9358,
      "theoretical_loss": 3.8495153322432003,
      "tokens_seen": 585826304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041538615847542626,
      "loss": 2.7838,
      "theoretical_loss": 3.849470210338751,
      "tokens_seen": 585891840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153761283851555,
      "loss": 2.8008,
      "theoretical_loss": 3.849425094894246,
      "tokens_seen": 585957376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153660982948846,
      "loss": 2.8761,
      "theoretical_loss": 3.8493799859080378,
      "tokens_seen": 586022912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041535606820461386,
      "loss": 2.9351,
      "theoretical_loss": 3.8493348833784813,
      "tokens_seen": 586088448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041534603811434304,
      "loss": 2.8079,
      "theoretical_loss": 3.849289787303929,
      "tokens_seen": 586153984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153360080240722,
      "loss": 2.8135,
      "theoretical_loss": 3.8492446976827357,
      "tokens_seen": 586219520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153259779338014,
      "loss": 2.7149,
      "theoretical_loss": 3.8491996145132568,
      "tokens_seen": 586285056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153159478435306,
      "loss": 2.7679,
      "theoretical_loss": 3.8491545377938485,
      "tokens_seen": 586350592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041530591775325976,
      "loss": 2.5723,
      "theoretical_loss": 3.849109467522867,
      "tokens_seen": 586416128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415295887662989,
      "loss": 2.7364,
      "theoretical_loss": 3.8490644036986694,
      "tokens_seen": 586481664
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 333131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1692469120025635,
      "objective/train/theoretical_loss": 3.8490193463196127,
      "objective/train/tokens_used": 607007200,
      "theoretical_loss": 3.8490193463196127,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152858575727181,
      "loss": 2.6856,
      "theoretical_loss": 3.8490193463196127,
      "tokens_seen": 586547200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041527582748244736,
      "loss": 2.8743,
      "theoretical_loss": 3.848974295384056,
      "tokens_seen": 586612736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152657973921765,
      "loss": 2.8334,
      "theoretical_loss": 3.8489292508903574,
      "tokens_seen": 586678272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152557673019057,
      "loss": 2.7919,
      "theoretical_loss": 3.848884212836877,
      "tokens_seen": 586743808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152457372116349,
      "loss": 2.8642,
      "theoretical_loss": 3.848839181221974,
      "tokens_seen": 586809344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152357071213641,
      "loss": 2.7809,
      "theoretical_loss": 3.84879415604401,
      "tokens_seen": 586874880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041522567703109327,
      "loss": 2.907,
      "theoretical_loss": 3.8487491373013456,
      "tokens_seen": 586940416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152156469408225,
      "loss": 2.6461,
      "theoretical_loss": 3.848704124992343,
      "tokens_seen": 587005952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041520561685055163,
      "loss": 2.59,
      "theoretical_loss": 3.8486591191153643,
      "tokens_seen": 587071488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041519558676028086,
      "loss": 2.8257,
      "theoretical_loss": 3.848614119668772,
      "tokens_seen": 587137024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041518555667001,
      "loss": 2.8721,
      "theoretical_loss": 3.8485691266509314,
      "tokens_seen": 587202560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151755265797392,
      "loss": 3.1213,
      "theoretical_loss": 3.8485241400602055,
      "tokens_seen": 587268096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151654964894684,
      "loss": 2.6248,
      "theoretical_loss": 3.848479159894959,
      "tokens_seen": 587333632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151554663991976,
      "loss": 2.8182,
      "theoretical_loss": 3.8484341861535576,
      "tokens_seen": 587399168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151454363089268,
      "loss": 2.7422,
      "theoretical_loss": 3.8483892188343676,
      "tokens_seen": 587464704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041513540621865595,
      "loss": 2.9843,
      "theoretical_loss": 3.8483442579357554,
      "tokens_seen": 587530240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151253761283852,
      "loss": 2.7539,
      "theoretical_loss": 3.8482993034560877,
      "tokens_seen": 587595776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041511534603811437,
      "loss": 2.876,
      "theoretical_loss": 3.848254355393734,
      "tokens_seen": 587661312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041510531594784355,
      "loss": 2.7914,
      "theoretical_loss": 3.8482094137470604,
      "tokens_seen": 587726848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041509528585757273,
      "loss": 2.9117,
      "theoretical_loss": 3.8481644785144375,
      "tokens_seen": 587792384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041508525576730196,
      "loss": 2.741,
      "theoretical_loss": 3.8481195496942346,
      "tokens_seen": 587857920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150752256770311,
      "loss": 2.6959,
      "theoretical_loss": 3.848074627284822,
      "tokens_seen": 587923456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150651955867603,
      "loss": 2.7976,
      "theoretical_loss": 3.84802971128457,
      "tokens_seen": 587988992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041505516549648945,
      "loss": 2.8536,
      "theoretical_loss": 3.84798480169185,
      "tokens_seen": 588054528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150451354062187,
      "loss": 2.9974,
      "theoretical_loss": 3.8479398985050346,
      "tokens_seen": 588120064
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 334325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.72865891456604,
      "objective/train/theoretical_loss": 3.847895001722496,
      "objective/train/tokens_used": 608645600,
      "theoretical_loss": 3.847895001722496,
      "tokens_seen": 588185600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041503510531594787,
      "loss": 2.8432,
      "theoretical_loss": 3.847895001722496,
      "tokens_seen": 588185600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041502507522567705,
      "loss": 2.6422,
      "theoretical_loss": 3.8478501113426073,
      "tokens_seen": 588251136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041501504513540623,
      "loss": 2.5214,
      "theoretical_loss": 3.847805227363743,
      "tokens_seen": 588316672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150050150451354,
      "loss": 2.6265,
      "theoretical_loss": 3.8477603497842763,
      "tokens_seen": 588382208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149949849548646,
      "loss": 2.6646,
      "theoretical_loss": 3.8477154786025825,
      "tokens_seen": 588447744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041498495486459383,
      "loss": 2.7582,
      "theoretical_loss": 3.847670613817038,
      "tokens_seen": 588513280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041497492477432296,
      "loss": 2.8478,
      "theoretical_loss": 3.847625755426018,
      "tokens_seen": 588578816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149648946840522,
      "loss": 2.6564,
      "theoretical_loss": 3.8475809034278994,
      "tokens_seen": 588644352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149548645937813,
      "loss": 2.4809,
      "theoretical_loss": 3.8475360578210602,
      "tokens_seen": 588709888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041494483450351055,
      "loss": 2.8239,
      "theoretical_loss": 3.847491218603878,
      "tokens_seen": 588775424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041493480441323974,
      "loss": 2.6325,
      "theoretical_loss": 3.847446385774731,
      "tokens_seen": 588840960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149247743229689,
      "loss": 2.8157,
      "theoretical_loss": 3.847401559331998,
      "tokens_seen": 588906496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149147442326981,
      "loss": 2.8234,
      "theoretical_loss": 3.8473567392740593,
      "tokens_seen": 588972032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041490471414242733,
      "loss": 2.6742,
      "theoretical_loss": 3.8473119255992954,
      "tokens_seen": 589037568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041489468405215646,
      "loss": 2.9159,
      "theoretical_loss": 3.8472671183060867,
      "tokens_seen": 589103104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148846539618857,
      "loss": 2.8963,
      "theoretical_loss": 3.8472223173928146,
      "tokens_seen": 589168640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148746238716148,
      "loss": 2.7451,
      "theoretical_loss": 3.8471775228578617,
      "tokens_seen": 589234176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041486459378134406,
      "loss": 2.7525,
      "theoretical_loss": 3.8471327346996107,
      "tokens_seen": 589299712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041485456369107324,
      "loss": 2.73,
      "theoretical_loss": 3.8470879529164437,
      "tokens_seen": 589365248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148445336008024,
      "loss": 2.8554,
      "theoretical_loss": 3.847043177506746,
      "tokens_seen": 589430784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148345035105316,
      "loss": 2.9111,
      "theoretical_loss": 3.8469984084689015,
      "tokens_seen": 589496320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148244734202608,
      "loss": 2.7169,
      "theoretical_loss": 3.8469536458012943,
      "tokens_seen": 589561856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041481444332998996,
      "loss": 2.8186,
      "theoretical_loss": 3.8469088895023114,
      "tokens_seen": 589627392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148044132397192,
      "loss": 2.8273,
      "theoretical_loss": 3.8468641395703385,
      "tokens_seen": 589692928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147943831494483,
      "loss": 2.6757,
      "theoretical_loss": 3.8468193960037618,
      "tokens_seen": 589758464
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 334849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.178940534591675,
      "objective/train/theoretical_loss": 3.8467746588009692,
      "objective/train/tokens_used": 610284000,
      "theoretical_loss": 3.8467746588009692,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041478435305917756,
      "loss": 2.6907,
      "theoretical_loss": 3.8467746588009692,
      "tokens_seen": 589824000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147743229689067,
      "loss": 2.798,
      "theoretical_loss": 3.846729927960349,
      "tokens_seen": 589889536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147642928786359,
      "loss": 2.9737,
      "theoretical_loss": 3.846685203480289,
      "tokens_seen": 589955072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147542627883651,
      "loss": 2.6387,
      "theoretical_loss": 3.8466404853591785,
      "tokens_seen": 590020608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147442326980943,
      "loss": 2.5792,
      "theoretical_loss": 3.8465957735954075,
      "tokens_seen": 590086144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041473420260782347,
      "loss": 3.0323,
      "theoretical_loss": 3.8465510681873667,
      "tokens_seen": 590151680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147241725175527,
      "loss": 2.8036,
      "theoretical_loss": 3.8465063691334453,
      "tokens_seen": 590217216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041471414242728183,
      "loss": 2.8123,
      "theoretical_loss": 3.846461676432037,
      "tokens_seen": 590282752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041470411233701106,
      "loss": 2.9491,
      "theoretical_loss": 3.8464169900815324,
      "tokens_seen": 590348288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146940822467402,
      "loss": 2.687,
      "theoretical_loss": 3.8463723100803247,
      "tokens_seen": 590413824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146840521564694,
      "loss": 2.708,
      "theoretical_loss": 3.846327636426807,
      "tokens_seen": 590479360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146740220661986,
      "loss": 2.7274,
      "theoretical_loss": 3.8462829691193727,
      "tokens_seen": 590544896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146639919759278,
      "loss": 2.809,
      "theoretical_loss": 3.8462383081564173,
      "tokens_seen": 590610432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041465396188565697,
      "loss": 2.8318,
      "theoretical_loss": 3.846193653536335,
      "tokens_seen": 590675968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041464393179538615,
      "loss": 2.8892,
      "theoretical_loss": 3.846149005257521,
      "tokens_seen": 590741504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041463390170511533,
      "loss": 2.6609,
      "theoretical_loss": 3.8461043633183722,
      "tokens_seen": 590807040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041462387161484457,
      "loss": 2.5279,
      "theoretical_loss": 3.8460597277172845,
      "tokens_seen": 590872576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146138415245737,
      "loss": 2.7456,
      "theoretical_loss": 3.846015098452656,
      "tokens_seen": 590938112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041460381143430293,
      "loss": 2.9341,
      "theoretical_loss": 3.8459704755228854,
      "tokens_seen": 591003648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145937813440321,
      "loss": 2.9464,
      "theoretical_loss": 3.8459258589263694,
      "tokens_seen": 591069184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145837512537613,
      "loss": 2.7979,
      "theoretical_loss": 3.8458812486615077,
      "tokens_seen": 591134720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145737211634905,
      "loss": 2.7745,
      "theoretical_loss": 3.8458366447267003,
      "tokens_seen": 591200256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041456369107321965,
      "loss": 2.7959,
      "theoretical_loss": 3.845792047120347,
      "tokens_seen": 591265792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041455366098294884,
      "loss": 2.8272,
      "theoretical_loss": 3.845747455840849,
      "tokens_seen": 591331328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041454363089267807,
      "loss": 2.9054,
      "theoretical_loss": 3.8457028708866074,
      "tokens_seen": 591396864
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 335803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.751352071762085,
      "objective/train/theoretical_loss": 3.845658292256025,
      "objective/train/tokens_used": 611922400,
      "theoretical_loss": 3.845658292256025,
      "tokens_seen": 591462400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145336008024072,
      "loss": 2.7783,
      "theoretical_loss": 3.845658292256025,
      "tokens_seen": 591462400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041452357071213643,
      "loss": 2.7238,
      "theoretical_loss": 3.8456137199475027,
      "tokens_seen": 591527936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041451354062186556,
      "loss": 2.6259,
      "theoretical_loss": 3.845569153959445,
      "tokens_seen": 591593472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145035105315948,
      "loss": 2.8755,
      "theoretical_loss": 3.8455245942902554,
      "tokens_seen": 591659008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414493480441324,
      "loss": 2.9484,
      "theoretical_loss": 3.8454800409383374,
      "tokens_seen": 591724544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041448345035105316,
      "loss": 2.7721,
      "theoretical_loss": 3.845435493902097,
      "tokens_seen": 591790080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041447342026078234,
      "loss": 2.8939,
      "theoretical_loss": 3.8453909531799395,
      "tokens_seen": 591855616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144633901705115,
      "loss": 2.9146,
      "theoretical_loss": 3.84534641877027,
      "tokens_seen": 591921152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144533600802407,
      "loss": 2.7202,
      "theoretical_loss": 3.845301890671496,
      "tokens_seen": 591986688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041444332998996994,
      "loss": 2.8713,
      "theoretical_loss": 3.845257368882024,
      "tokens_seen": 592052224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041443329989969906,
      "loss": 3.1036,
      "theoretical_loss": 3.8452128534002625,
      "tokens_seen": 592117760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144232698094283,
      "loss": 2.7427,
      "theoretical_loss": 3.8451683442246196,
      "tokens_seen": 592183296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041441323971915753,
      "loss": 2.8624,
      "theoretical_loss": 3.8451238413535043,
      "tokens_seen": 592248832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041440320962888666,
      "loss": 2.5677,
      "theoretical_loss": 3.8450793447853258,
      "tokens_seen": 592314368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143931795386159,
      "loss": 2.8878,
      "theoretical_loss": 3.8450348545184942,
      "tokens_seen": 592379904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414383149448345,
      "loss": 2.5672,
      "theoretical_loss": 3.84499037055142,
      "tokens_seen": 592445440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041437311935807426,
      "loss": 2.7789,
      "theoretical_loss": 3.8449458928825155,
      "tokens_seen": 592510976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041436308926780344,
      "loss": 2.9836,
      "theoretical_loss": 3.8449014215101913,
      "tokens_seen": 592576512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143530591775326,
      "loss": 2.8149,
      "theoretical_loss": 3.84485695643286,
      "tokens_seen": 592642048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143430290872618,
      "loss": 2.8165,
      "theoretical_loss": 3.8448124976489346,
      "tokens_seen": 592707584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414332998996991,
      "loss": 3.0223,
      "theoretical_loss": 3.8447680451568296,
      "tokens_seen": 592773120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041432296890672016,
      "loss": 2.7377,
      "theoretical_loss": 3.844723598954958,
      "tokens_seen": 592838656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143129388164494,
      "loss": 2.8208,
      "theoretical_loss": 3.8446791590417346,
      "tokens_seen": 592904192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143029087261785,
      "loss": 2.7563,
      "theoretical_loss": 3.8446347254155753,
      "tokens_seen": 592969728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041429287863590776,
      "loss": 2.9717,
      "theoretical_loss": 3.8445902980748956,
      "tokens_seen": 593035264
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 336288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.919393539428711,
      "objective/train/theoretical_loss": 3.8445458770181116,
      "objective/train/tokens_used": 613560800,
      "theoretical_loss": 3.8445458770181116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142828485456369,
      "loss": 2.8637,
      "theoretical_loss": 3.8445458770181116,
      "tokens_seen": 593100800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142728184553661,
      "loss": 2.6485,
      "theoretical_loss": 3.844501462243641,
      "tokens_seen": 593166336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142627883650953,
      "loss": 2.8594,
      "theoretical_loss": 3.8444570537499,
      "tokens_seen": 593231872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142527582748245,
      "loss": 2.4827,
      "theoretical_loss": 3.8444126515353076,
      "tokens_seen": 593297408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041424272818455367,
      "loss": 2.7269,
      "theoretical_loss": 3.8443682555982837,
      "tokens_seen": 593362944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142326980942829,
      "loss": 2.8108,
      "theoretical_loss": 3.8443238659372456,
      "tokens_seen": 593428480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041422266800401203,
      "loss": 2.8032,
      "theoretical_loss": 3.844279482550614,
      "tokens_seen": 593494016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041421263791374126,
      "loss": 2.8399,
      "theoretical_loss": 3.84423510543681,
      "tokens_seen": 593559552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142026078234704,
      "loss": 2.7256,
      "theoretical_loss": 3.8441907345942528,
      "tokens_seen": 593625088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141925777331996,
      "loss": 2.891,
      "theoretical_loss": 3.844146370021366,
      "tokens_seen": 593690624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141825476429288,
      "loss": 2.8849,
      "theoretical_loss": 3.8441020117165703,
      "tokens_seen": 593756160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414172517552658,
      "loss": 2.8343,
      "theoretical_loss": 3.8440576596782887,
      "tokens_seen": 593821696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041416248746238717,
      "loss": 2.9734,
      "theoretical_loss": 3.844013313904945,
      "tokens_seen": 593887232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041415245737211635,
      "loss": 2.7278,
      "theoretical_loss": 3.8439689743949628,
      "tokens_seen": 593952768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041414242728184553,
      "loss": 2.6113,
      "theoretical_loss": 3.8439246411467662,
      "tokens_seen": 594018304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041413239719157477,
      "loss": 2.7153,
      "theoretical_loss": 3.8438803141587807,
      "tokens_seen": 594083840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141223671013039,
      "loss": 2.9235,
      "theoretical_loss": 3.8438359934294315,
      "tokens_seen": 594149376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041411233701103313,
      "loss": 2.7539,
      "theoretical_loss": 3.843791678957145,
      "tokens_seen": 594214912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141023069207623,
      "loss": 2.8045,
      "theoretical_loss": 3.8437473707403482,
      "tokens_seen": 594280448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140922768304915,
      "loss": 2.9262,
      "theoretical_loss": 3.843703068777467,
      "tokens_seen": 594345984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140822467402207,
      "loss": 2.9616,
      "theoretical_loss": 3.843658773066931,
      "tokens_seen": 594411520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041407221664994985,
      "loss": 2.8237,
      "theoretical_loss": 3.8436144836071673,
      "tokens_seen": 594477056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041406218655967904,
      "loss": 2.9555,
      "theoretical_loss": 3.8435702003966052,
      "tokens_seen": 594542592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041405215646940827,
      "loss": 2.7086,
      "theoretical_loss": 3.843525923433675,
      "tokens_seen": 594608128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140421263791374,
      "loss": 2.8386,
      "theoretical_loss": 3.8434816527168056,
      "tokens_seen": 594673664
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 337523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4961068630218506,
      "objective/train/theoretical_loss": 3.8434373882444284,
      "objective/train/tokens_used": 615199200,
      "theoretical_loss": 3.8434373882444284,
      "tokens_seen": 594739200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041403209628886663,
      "loss": 2.7317,
      "theoretical_loss": 3.8434373882444284,
      "tokens_seen": 594739200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041402206619859576,
      "loss": 2.6545,
      "theoretical_loss": 3.8433931300149746,
      "tokens_seen": 594804736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414012036108325,
      "loss": 2.952,
      "theoretical_loss": 3.843348878026876,
      "tokens_seen": 594870272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140020060180542,
      "loss": 2.8919,
      "theoretical_loss": 3.843304632278565,
      "tokens_seen": 594935808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041399197592778336,
      "loss": 2.9386,
      "theoretical_loss": 3.8432603927684745,
      "tokens_seen": 595001344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041398194583751254,
      "loss": 2.533,
      "theoretical_loss": 3.8432161594950376,
      "tokens_seen": 595066880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139719157472417,
      "loss": 2.7428,
      "theoretical_loss": 3.8431719324566895,
      "tokens_seen": 595132416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139618856569709,
      "loss": 2.6911,
      "theoretical_loss": 3.8431277116518636,
      "tokens_seen": 595197952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041395185556670014,
      "loss": 2.8322,
      "theoretical_loss": 3.8430834970789953,
      "tokens_seen": 595263488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041394182547642926,
      "loss": 2.8112,
      "theoretical_loss": 3.843039288736521,
      "tokens_seen": 595329024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139317953861585,
      "loss": 2.7172,
      "theoretical_loss": 3.842995086622877,
      "tokens_seen": 595394560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139217652958877,
      "loss": 2.713,
      "theoretical_loss": 3.8429508907364998,
      "tokens_seen": 595460096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041391173520561686,
      "loss": 2.8304,
      "theoretical_loss": 3.842906701075827,
      "tokens_seen": 595525632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041390170511534604,
      "loss": 2.942,
      "theoretical_loss": 3.8428625176392965,
      "tokens_seen": 595591168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138916750250752,
      "loss": 2.8326,
      "theoretical_loss": 3.8428183404253473,
      "tokens_seen": 595656704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138816449348044,
      "loss": 2.8756,
      "theoretical_loss": 3.8427741694324182,
      "tokens_seen": 595722240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041387161484453364,
      "loss": 2.6363,
      "theoretical_loss": 3.8427300046589488,
      "tokens_seen": 595787776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041386158475426277,
      "loss": 2.7768,
      "theoretical_loss": 3.8426858461033797,
      "tokens_seen": 595853312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413851554663992,
      "loss": 2.6984,
      "theoretical_loss": 3.8426416937641514,
      "tokens_seen": 595918848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041384152457372113,
      "loss": 2.5968,
      "theoretical_loss": 3.842597547639706,
      "tokens_seen": 595984384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041383149448345036,
      "loss": 2.9943,
      "theoretical_loss": 3.842553407728485,
      "tokens_seen": 596049920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041382146439317954,
      "loss": 2.6488,
      "theoretical_loss": 3.842509274028931,
      "tokens_seen": 596115456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138114343029087,
      "loss": 2.6736,
      "theoretical_loss": 3.8424651465394866,
      "tokens_seen": 596180992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138014042126379,
      "loss": 2.6599,
      "theoretical_loss": 3.842421025258596,
      "tokens_seen": 596246528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137913741223671,
      "loss": 2.826,
      "theoretical_loss": 3.8423769101847034,
      "tokens_seen": 596312064
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 338188,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9216253757476807,
      "objective/train/theoretical_loss": 3.842332801316254,
      "objective/train/tokens_used": 616837600,
      "theoretical_loss": 3.842332801316254,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041378134403209627,
      "loss": 2.8479,
      "theoretical_loss": 3.842332801316254,
      "tokens_seen": 596377600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137713139418255,
      "loss": 2.7266,
      "theoretical_loss": 3.842288698651692,
      "tokens_seen": 596443136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041376128385155463,
      "loss": 2.689,
      "theoretical_loss": 3.8422446021894645,
      "tokens_seen": 596508672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041375125376128387,
      "loss": 2.9293,
      "theoretical_loss": 3.842200511928017,
      "tokens_seen": 596574208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041374122367101305,
      "loss": 2.5976,
      "theoretical_loss": 3.842156427865797,
      "tokens_seen": 596639744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041373119358074223,
      "loss": 2.9682,
      "theoretical_loss": 3.842112350001252,
      "tokens_seen": 596705280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137211634904714,
      "loss": 2.8541,
      "theoretical_loss": 3.84206827833283,
      "tokens_seen": 596770816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137111334002006,
      "loss": 2.7331,
      "theoretical_loss": 3.84202421285898,
      "tokens_seen": 596836352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041370110330992977,
      "loss": 2.9806,
      "theoretical_loss": 3.8419801535781515,
      "tokens_seen": 596901888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413691073219659,
      "loss": 2.6249,
      "theoretical_loss": 3.841936100488793,
      "tokens_seen": 596967424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041368104312938813,
      "loss": 2.7589,
      "theoretical_loss": 3.8418920535893566,
      "tokens_seen": 597032960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041367101303911737,
      "loss": 2.6685,
      "theoretical_loss": 3.841848012878292,
      "tokens_seen": 597098496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041366098294884655,
      "loss": 2.8725,
      "theoretical_loss": 3.8418039783540507,
      "tokens_seen": 597164032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041365095285857573,
      "loss": 2.767,
      "theoretical_loss": 3.841759950015086,
      "tokens_seen": 597229568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041364092276830497,
      "loss": 2.6848,
      "theoretical_loss": 3.841715927859849,
      "tokens_seen": 597295104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136308926780341,
      "loss": 2.7853,
      "theoretical_loss": 3.8416719118867944,
      "tokens_seen": 597360640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041362086258776333,
      "loss": 2.6599,
      "theoretical_loss": 3.841627902094374,
      "tokens_seen": 597426176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136108324974925,
      "loss": 2.919,
      "theoretical_loss": 3.8415838984810433,
      "tokens_seen": 597491712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136008024072217,
      "loss": 2.8007,
      "theoretical_loss": 3.841539901045257,
      "tokens_seen": 597557248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135907723169509,
      "loss": 2.7802,
      "theoretical_loss": 3.841495909785471,
      "tokens_seen": 597622784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041358074222668005,
      "loss": 2.7021,
      "theoretical_loss": 3.8414519247001397,
      "tokens_seen": 597688320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041357071213640924,
      "loss": 2.6627,
      "theoretical_loss": 3.8414079457877213,
      "tokens_seen": 597753856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041356068204613847,
      "loss": 2.7435,
      "theoretical_loss": 3.8413639730466715,
      "tokens_seen": 597819392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135506519558676,
      "loss": 2.9956,
      "theoretical_loss": 3.8413200064754487,
      "tokens_seen": 597884928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041354062186559683,
      "loss": 2.6928,
      "theoretical_loss": 3.841276046072511,
      "tokens_seen": 597950464
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 339419,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.796257972717285,
      "objective/train/theoretical_loss": 3.841232091836317,
      "objective/train/tokens_used": 618476000,
      "theoretical_loss": 3.841232091836317,
      "tokens_seen": 598016000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041353059177532596,
      "loss": 2.649,
      "theoretical_loss": 3.841232091836317,
      "tokens_seen": 598016000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135205616850552,
      "loss": 2.9916,
      "theoretical_loss": 3.8411881437653252,
      "tokens_seen": 598081536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135105315947844,
      "loss": 2.767,
      "theoretical_loss": 3.841144201857997,
      "tokens_seen": 598147072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041350050150451356,
      "loss": 2.6961,
      "theoretical_loss": 3.8411002661127913,
      "tokens_seen": 598212608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041349047141424274,
      "loss": 2.5988,
      "theoretical_loss": 3.8410563365281702,
      "tokens_seen": 598278144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134804413239719,
      "loss": 2.4411,
      "theoretical_loss": 3.8410124131025944,
      "tokens_seen": 598343680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134704112337011,
      "loss": 2.9326,
      "theoretical_loss": 3.8409684958345265,
      "tokens_seen": 598409216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041346038114343034,
      "loss": 2.7858,
      "theoretical_loss": 3.840924584722428,
      "tokens_seen": 598474752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041345035105315946,
      "loss": 2.6159,
      "theoretical_loss": 3.840880679764763,
      "tokens_seen": 598540288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134403209628887,
      "loss": 2.9933,
      "theoretical_loss": 3.8408367809599957,
      "tokens_seen": 598605824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134302908726179,
      "loss": 2.634,
      "theoretical_loss": 3.8407928883065887,
      "tokens_seen": 598671360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041342026078234706,
      "loss": 2.5301,
      "theoretical_loss": 3.840749001803008,
      "tokens_seen": 598736896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041341023069207624,
      "loss": 2.6363,
      "theoretical_loss": 3.8407051214477193,
      "tokens_seen": 598802432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134002006018054,
      "loss": 2.7874,
      "theoretical_loss": 3.8406612472391872,
      "tokens_seen": 598867968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133901705115346,
      "loss": 2.9248,
      "theoretical_loss": 3.8406173791758786,
      "tokens_seen": 598933504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041338014042126384,
      "loss": 2.8438,
      "theoretical_loss": 3.840573517256261,
      "tokens_seen": 598999040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041337011033099297,
      "loss": 2.9863,
      "theoretical_loss": 3.8405296614788016,
      "tokens_seen": 599064576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133600802407222,
      "loss": 2.4437,
      "theoretical_loss": 3.8404858118419685,
      "tokens_seen": 599130112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041335005015045133,
      "loss": 2.6434,
      "theoretical_loss": 3.84044196834423,
      "tokens_seen": 599195648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041334002006018056,
      "loss": 2.7596,
      "theoretical_loss": 3.840398130984056,
      "tokens_seen": 599261184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041332998996990975,
      "loss": 2.8734,
      "theoretical_loss": 3.840354299759916,
      "tokens_seen": 599326720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133199598796389,
      "loss": 2.8495,
      "theoretical_loss": 3.84031047467028,
      "tokens_seen": 599392256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133099297893681,
      "loss": 2.6736,
      "theoretical_loss": 3.840266655713619,
      "tokens_seen": 599457792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132998996990973,
      "loss": 3.0788,
      "theoretical_loss": 3.840222842888405,
      "tokens_seen": 599523328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041328986960882647,
      "loss": 3.0841,
      "theoretical_loss": 3.840179036193109,
      "tokens_seen": 599588864
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 340168,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.811565399169922,
      "objective/train/theoretical_loss": 3.840135235626204,
      "objective/train/tokens_used": 620114400,
      "theoretical_loss": 3.840135235626204,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132798395185557,
      "loss": 2.6611,
      "theoretical_loss": 3.840135235626204,
      "tokens_seen": 599654400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041326980942828483,
      "loss": 2.7923,
      "theoretical_loss": 3.8400914411861624,
      "tokens_seen": 599719936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041325977933801407,
      "loss": 2.8316,
      "theoretical_loss": 3.8400476528714584,
      "tokens_seen": 599785472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041324974924774325,
      "loss": 2.8943,
      "theoretical_loss": 3.840003870680566,
      "tokens_seen": 599851008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041323971915747243,
      "loss": 2.6886,
      "theoretical_loss": 3.8399600946119605,
      "tokens_seen": 599916544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132296890672016,
      "loss": 2.8353,
      "theoretical_loss": 3.8399163246641157,
      "tokens_seen": 599982080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132196589769308,
      "loss": 3.022,
      "theoretical_loss": 3.839872560835509,
      "tokens_seen": 600047616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041320962888665997,
      "loss": 2.9232,
      "theoretical_loss": 3.839828803124615,
      "tokens_seen": 600113152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131995987963892,
      "loss": 3.0362,
      "theoretical_loss": 3.839785051529912,
      "tokens_seen": 600178688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041318956870611833,
      "loss": 2.633,
      "theoretical_loss": 3.839741306049876,
      "tokens_seen": 600244224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041317953861584757,
      "loss": 2.9153,
      "theoretical_loss": 3.839697566682987,
      "tokens_seen": 600309760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131695085255767,
      "loss": 2.8834,
      "theoretical_loss": 3.8396538334277217,
      "tokens_seen": 600375296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041315947843530593,
      "loss": 2.8856,
      "theoretical_loss": 3.839610106282559,
      "tokens_seen": 600440832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131494483450351,
      "loss": 2.9453,
      "theoretical_loss": 3.83956638524598,
      "tokens_seen": 600506368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131394182547643,
      "loss": 2.6424,
      "theoretical_loss": 3.8395226703164638,
      "tokens_seen": 600571904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131293881644935,
      "loss": 3.0177,
      "theoretical_loss": 3.8394789614924916,
      "tokens_seen": 600637440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131193580742227,
      "loss": 2.7729,
      "theoretical_loss": 3.8394352587725438,
      "tokens_seen": 600702976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041310932798395184,
      "loss": 2.9922,
      "theoretical_loss": 3.839391562155103,
      "tokens_seen": 600768512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130992978936811,
      "loss": 2.7626,
      "theoretical_loss": 3.8393478716386507,
      "tokens_seen": 600834048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130892678034102,
      "loss": 2.7941,
      "theoretical_loss": 3.8393041872216704,
      "tokens_seen": 600899584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041307923771313944,
      "loss": 2.4599,
      "theoretical_loss": 3.8392605089026457,
      "tokens_seen": 600965120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130692076228686,
      "loss": 2.6901,
      "theoretical_loss": 3.83921683668006,
      "tokens_seen": 601030656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130591775325978,
      "loss": 2.7902,
      "theoretical_loss": 3.8391731705523977,
      "tokens_seen": 601096192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413049147442327,
      "loss": 2.6181,
      "theoretical_loss": 3.8391295105181436,
      "tokens_seen": 601161728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041303911735205616,
      "loss": 2.7678,
      "theoretical_loss": 3.8390858565757844,
      "tokens_seen": 601227264
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 341333,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.842754364013672,
      "objective/train/theoretical_loss": 3.8390422087238045,
      "objective/train/tokens_used": 621752800,
      "theoretical_loss": 3.8390422087238045,
      "tokens_seen": 601292800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041302908726178534,
      "loss": 2.7321,
      "theoretical_loss": 3.8390422087238045,
      "tokens_seen": 601292800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130190571715146,
      "loss": 2.9798,
      "theoretical_loss": 3.8389985669606927,
      "tokens_seen": 601358336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130090270812437,
      "loss": 2.5315,
      "theoretical_loss": 3.8389549312849343,
      "tokens_seen": 601423872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041299899699097294,
      "loss": 2.8848,
      "theoretical_loss": 3.838911301695018,
      "tokens_seen": 601489408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041298896690070207,
      "loss": 2.8289,
      "theoretical_loss": 3.8388676781894313,
      "tokens_seen": 601554944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129789368104313,
      "loss": 2.8317,
      "theoretical_loss": 3.8388240607666635,
      "tokens_seen": 601620480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129689067201605,
      "loss": 2.7689,
      "theoretical_loss": 3.838780449425204,
      "tokens_seen": 601686016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041295887662988966,
      "loss": 2.5929,
      "theoretical_loss": 3.8387368441635425,
      "tokens_seen": 601751552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041294884653961884,
      "loss": 2.9393,
      "theoretical_loss": 3.8386932449801696,
      "tokens_seen": 601817088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129388164493481,
      "loss": 2.8339,
      "theoretical_loss": 3.8386496518735767,
      "tokens_seen": 601882624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129287863590772,
      "loss": 2.7379,
      "theoretical_loss": 3.8386060648422538,
      "tokens_seen": 601948160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041291875626880644,
      "loss": 2.9387,
      "theoretical_loss": 3.8385624838846946,
      "tokens_seen": 602013696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129087261785356,
      "loss": 2.8846,
      "theoretical_loss": 3.838518908999391,
      "tokens_seen": 602079232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128986960882648,
      "loss": 2.9549,
      "theoretical_loss": 3.8384753401848353,
      "tokens_seen": 602144768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041288866599799404,
      "loss": 2.8541,
      "theoretical_loss": 3.8384317774395225,
      "tokens_seen": 602210304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041287863590772317,
      "loss": 2.7912,
      "theoretical_loss": 3.8383882207619457,
      "tokens_seen": 602275840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128686058174524,
      "loss": 2.635,
      "theoretical_loss": 3.838344670150601,
      "tokens_seen": 602341376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041285857572718153,
      "loss": 2.9703,
      "theoretical_loss": 3.838301125603982,
      "tokens_seen": 602406912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041284854563691076,
      "loss": 2.989,
      "theoretical_loss": 3.8382575871205855,
      "tokens_seen": 602472448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041283851554663995,
      "loss": 2.6135,
      "theoretical_loss": 3.8382140546989074,
      "tokens_seen": 602537984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128284854563691,
      "loss": 2.7771,
      "theoretical_loss": 3.8381705283374457,
      "tokens_seen": 602603520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128184553660983,
      "loss": 2.6952,
      "theoretical_loss": 3.8381270080346956,
      "tokens_seen": 602669056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128084252758275,
      "loss": 2.5534,
      "theoretical_loss": 3.8380834937891573,
      "tokens_seen": 602734592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041279839518555667,
      "loss": 2.8246,
      "theoretical_loss": 3.8380399855993286,
      "tokens_seen": 602800128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127883650952859,
      "loss": 2.5559,
      "theoretical_loss": 3.8379964834637073,
      "tokens_seen": 602865664
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 342479,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2375457286834717,
      "objective/train/theoretical_loss": 3.8379529873807945,
      "objective/train/tokens_used": 623391200,
      "theoretical_loss": 3.8379529873807945,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041277833500501503,
      "loss": 2.7622,
      "theoretical_loss": 3.8379529873807945,
      "tokens_seen": 602931200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041276830491474427,
      "loss": 2.7673,
      "theoretical_loss": 3.8379094973490897,
      "tokens_seen": 602996736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041275827482447345,
      "loss": 3.0164,
      "theoretical_loss": 3.837866013367093,
      "tokens_seen": 603062272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041274824473420263,
      "loss": 2.8215,
      "theoretical_loss": 3.8378225354333066,
      "tokens_seen": 603127808
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127382146439318,
      "loss": 2.6788,
      "theoretical_loss": 3.8377790635462308,
      "tokens_seen": 603193344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412728184553661,
      "loss": 2.8211,
      "theoretical_loss": 3.83773559770437,
      "tokens_seen": 603258880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041271815446339017,
      "loss": 2.9036,
      "theoretical_loss": 3.8376921379062248,
      "tokens_seen": 603324416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127081243731194,
      "loss": 2.5925,
      "theoretical_loss": 3.8376486841502992,
      "tokens_seen": 603389952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041269809428284854,
      "loss": 2.7658,
      "theoretical_loss": 3.837605236435097,
      "tokens_seen": 603455488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041268806419257777,
      "loss": 2.5781,
      "theoretical_loss": 3.8375617947591234,
      "tokens_seen": 603521024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126780341023069,
      "loss": 2.856,
      "theoretical_loss": 3.8375183591208826,
      "tokens_seen": 603586560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041266800401203613,
      "loss": 2.5308,
      "theoretical_loss": 3.837474929518879,
      "tokens_seen": 603652096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126579739217653,
      "loss": 2.6879,
      "theoretical_loss": 3.8374315059516206,
      "tokens_seen": 603717632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126479438314945,
      "loss": 2.8421,
      "theoretical_loss": 3.8373880884176126,
      "tokens_seen": 603783168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126379137412237,
      "loss": 2.881,
      "theoretical_loss": 3.837344676915362,
      "tokens_seen": 603848704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126278836509529,
      "loss": 2.3681,
      "theoretical_loss": 3.8373012714433767,
      "tokens_seen": 603914240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041261785356068204,
      "loss": 2.6518,
      "theoretical_loss": 3.8372578720001647,
      "tokens_seen": 603979776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126078234704113,
      "loss": 2.8608,
      "theoretical_loss": 3.8372144785842344,
      "tokens_seen": 604045312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125977933801404,
      "loss": 2.5857,
      "theoretical_loss": 3.8371710911940955,
      "tokens_seen": 604110848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041258776328986964,
      "loss": 2.3453,
      "theoretical_loss": 3.8371277098282572,
      "tokens_seen": 604176384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125777331995988,
      "loss": 2.8187,
      "theoretical_loss": 3.8370843344852297,
      "tokens_seen": 604241920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412567703109328,
      "loss": 2.6687,
      "theoretical_loss": 3.8370409651635238,
      "tokens_seen": 604307456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125576730190572,
      "loss": 2.7338,
      "theoretical_loss": 3.8369976018616505,
      "tokens_seen": 604372992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041254764292878636,
      "loss": 2.8385,
      "theoretical_loss": 3.8369542445781226,
      "tokens_seen": 604438528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041253761283851554,
      "loss": 2.6823,
      "theoretical_loss": 3.836910893311451,
      "tokens_seen": 604504064
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 342997,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.627711534500122,
      "objective/train/theoretical_loss": 3.83686754806015,
      "objective/train/tokens_used": 625029600,
      "theoretical_loss": 3.83686754806015,
      "tokens_seen": 604569600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125275827482448,
      "loss": 2.6497,
      "theoretical_loss": 3.83686754806015,
      "tokens_seen": 604569600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125175526579739,
      "loss": 2.9453,
      "theoretical_loss": 3.8368242088227316,
      "tokens_seen": 604635136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041250752256770314,
      "loss": 2.9436,
      "theoretical_loss": 3.836780875597711,
      "tokens_seen": 604700672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041249749247743227,
      "loss": 2.6017,
      "theoretical_loss": 3.836737548383601,
      "tokens_seen": 604766208
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124874623871615,
      "loss": 2.9184,
      "theoretical_loss": 3.836694227178919,
      "tokens_seen": 604831744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124774322968907,
      "loss": 2.8143,
      "theoretical_loss": 3.8366509119821774,
      "tokens_seen": 604897280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041246740220661986,
      "loss": 2.9792,
      "theoretical_loss": 3.836607602791895,
      "tokens_seen": 604962816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041245737211634904,
      "loss": 2.7049,
      "theoretical_loss": 3.8365642996065867,
      "tokens_seen": 605028352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124473420260783,
      "loss": 2.4616,
      "theoretical_loss": 3.83652100242477,
      "tokens_seen": 605093888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124373119358074,
      "loss": 2.7036,
      "theoretical_loss": 3.836477711244963,
      "tokens_seen": 605159424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041242728184553664,
      "loss": 2.7757,
      "theoretical_loss": 3.836434426065683,
      "tokens_seen": 605224960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041241725175526577,
      "loss": 2.8633,
      "theoretical_loss": 3.836391146885449,
      "tokens_seen": 605290496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412407221664995,
      "loss": 2.8372,
      "theoretical_loss": 3.8363478737027803,
      "tokens_seen": 605356032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123971915747242,
      "loss": 2.6962,
      "theoretical_loss": 3.836304606516197,
      "tokens_seen": 605421568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041238716148445337,
      "loss": 2.8052,
      "theoretical_loss": 3.8362613453242185,
      "tokens_seen": 605487104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041237713139418255,
      "loss": 2.6351,
      "theoretical_loss": 3.8362180901253655,
      "tokens_seen": 605552640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041236710130391173,
      "loss": 2.7867,
      "theoretical_loss": 3.83617484091816,
      "tokens_seen": 605618176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123570712136409,
      "loss": 2.6174,
      "theoretical_loss": 3.8361315977011237,
      "tokens_seen": 605683712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041234704112337015,
      "loss": 2.62,
      "theoretical_loss": 3.836088360472779,
      "tokens_seen": 605749248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041233701103309927,
      "loss": 2.7599,
      "theoretical_loss": 3.8360451292316475,
      "tokens_seen": 605814784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123269809428285,
      "loss": 2.6258,
      "theoretical_loss": 3.8360019039762543,
      "tokens_seen": 605880320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041231695085255763,
      "loss": 2.8677,
      "theoretical_loss": 3.835958684705122,
      "tokens_seen": 605945856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041230692076228687,
      "loss": 2.9551,
      "theoretical_loss": 3.835915471416776,
      "tokens_seen": 606011392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041229689067201605,
      "loss": 2.7614,
      "theoretical_loss": 3.8358722641097405,
      "tokens_seen": 606076928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041228686058174523,
      "loss": 2.7635,
      "theoretical_loss": 3.8358290627825413,
      "tokens_seen": 606142464
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 344327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1000027656555176,
      "objective/train/theoretical_loss": 3.835785867433705,
      "objective/train/tokens_used": 626668000,
      "theoretical_loss": 3.835785867433705,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122768304914744,
      "loss": 2.8939,
      "theoretical_loss": 3.835785867433705,
      "tokens_seen": 606208000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041226680040120365,
      "loss": 2.8328,
      "theoretical_loss": 3.8357426780617567,
      "tokens_seen": 606273536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122567703109328,
      "loss": 2.9931,
      "theoretical_loss": 3.8356994946652248,
      "tokens_seen": 606339072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412246740220662,
      "loss": 2.5709,
      "theoretical_loss": 3.835656317242636,
      "tokens_seen": 606404608
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041223671013039114,
      "loss": 2.7535,
      "theoretical_loss": 3.835613145792519,
      "tokens_seen": 606470144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122266800401204,
      "loss": 2.6616,
      "theoretical_loss": 3.8355699803134016,
      "tokens_seen": 606535680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041221664994984955,
      "loss": 2.766,
      "theoretical_loss": 3.8355268208038136,
      "tokens_seen": 606601216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041220661985957874,
      "loss": 2.7006,
      "theoretical_loss": 3.8354836672622845,
      "tokens_seen": 606666752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121965897693079,
      "loss": 2.8494,
      "theoretical_loss": 3.835440519687345,
      "tokens_seen": 606732288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121865596790371,
      "loss": 2.797,
      "theoretical_loss": 3.8353973780775243,
      "tokens_seen": 606797824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121765295887663,
      "loss": 2.8522,
      "theoretical_loss": 3.8353542424313547,
      "tokens_seen": 606863360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121664994984955,
      "loss": 2.9709,
      "theoretical_loss": 3.835311112747368,
      "tokens_seen": 606928896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121564694082247,
      "loss": 2.7044,
      "theoretical_loss": 3.8352679890240964,
      "tokens_seen": 606994432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121464393179539,
      "loss": 2.9497,
      "theoretical_loss": 3.8352248712600723,
      "tokens_seen": 607059968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121364092276831,
      "loss": 2.8229,
      "theoretical_loss": 3.835181759453829,
      "tokens_seen": 607125504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041212637913741224,
      "loss": 2.7877,
      "theoretical_loss": 3.8351386536039005,
      "tokens_seen": 607191040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121163490471415,
      "loss": 2.8895,
      "theoretical_loss": 3.8350955537088214,
      "tokens_seen": 607256576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121063189568706,
      "loss": 2.7407,
      "theoretical_loss": 3.835052459767126,
      "tokens_seen": 607322112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041209628886659984,
      "loss": 2.8907,
      "theoretical_loss": 3.8350093717773497,
      "tokens_seen": 607387648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412086258776329,
      "loss": 2.6943,
      "theoretical_loss": 3.8349662897380288,
      "tokens_seen": 607453184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120762286860582,
      "loss": 2.8111,
      "theoretical_loss": 3.8349232136476994,
      "tokens_seen": 607518720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120661985957874,
      "loss": 2.6154,
      "theoretical_loss": 3.8348801435048987,
      "tokens_seen": 607584256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041205616850551656,
      "loss": 2.7947,
      "theoretical_loss": 3.834837079308164,
      "tokens_seen": 607649792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041204613841524574,
      "loss": 2.4331,
      "theoretical_loss": 3.834794021056033,
      "tokens_seen": 607715328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412036108324975,
      "loss": 2.5009,
      "theoretical_loss": 3.8347509687470445,
      "tokens_seen": 607780864
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 345027,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9385602474212646,
      "objective/train/theoretical_loss": 3.8347079223797373,
      "objective/train/tokens_used": 628306400,
      "theoretical_loss": 3.8347079223797373,
      "tokens_seen": 607846400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120260782347041,
      "loss": 2.7218,
      "theoretical_loss": 3.8347079223797373,
      "tokens_seen": 607846400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041201604814443334,
      "loss": 2.4772,
      "theoretical_loss": 3.8346648819526505,
      "tokens_seen": 607911936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041200601805416247,
      "loss": 2.6235,
      "theoretical_loss": 3.834621847464325,
      "tokens_seen": 607977472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119959879638917,
      "loss": 2.7243,
      "theoretical_loss": 3.834578818913301,
      "tokens_seen": 608043008
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119859578736209,
      "loss": 3.0051,
      "theoretical_loss": 3.8345357962981197,
      "tokens_seen": 608108544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041197592778335006,
      "loss": 2.7735,
      "theoretical_loss": 3.834492779617322,
      "tokens_seen": 608174080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041196589769307924,
      "loss": 2.628,
      "theoretical_loss": 3.834449768869451,
      "tokens_seen": 608239616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119558676028085,
      "loss": 2.6384,
      "theoretical_loss": 3.834406764053048,
      "tokens_seen": 608305152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119458375125376,
      "loss": 2.5853,
      "theoretical_loss": 3.834363765166657,
      "tokens_seen": 608370688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041193580742226684,
      "loss": 2.8413,
      "theoretical_loss": 3.834320772208822,
      "tokens_seen": 608436224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041192577733199597,
      "loss": 2.8341,
      "theoretical_loss": 3.8342777851780863,
      "tokens_seen": 608501760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119157472417252,
      "loss": 2.5505,
      "theoretical_loss": 3.8342348040729943,
      "tokens_seen": 608567296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119057171514544,
      "loss": 2.6599,
      "theoretical_loss": 3.8341918288920924,
      "tokens_seen": 608632832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041189568706118357,
      "loss": 2.7973,
      "theoretical_loss": 3.8341488596339257,
      "tokens_seen": 608698368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041188565697091275,
      "loss": 2.799,
      "theoretical_loss": 3.8341058962970402,
      "tokens_seen": 608763904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041187562688064193,
      "loss": 2.8924,
      "theoretical_loss": 3.834062938879982,
      "tokens_seen": 608829440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118655967903711,
      "loss": 2.6892,
      "theoretical_loss": 3.8340199873813,
      "tokens_seen": 608894976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041185556670010035,
      "loss": 2.8457,
      "theoretical_loss": 3.8339770417995407,
      "tokens_seen": 608960512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041184553660982947,
      "loss": 2.9086,
      "theoretical_loss": 3.8339341021332527,
      "tokens_seen": 609026048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118355065195587,
      "loss": 2.7818,
      "theoretical_loss": 3.833891168380984,
      "tokens_seen": 609091584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041182547642928783,
      "loss": 2.6026,
      "theoretical_loss": 3.833848240541286,
      "tokens_seen": 609157120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041181544633901707,
      "loss": 2.7824,
      "theoretical_loss": 3.833805318612706,
      "tokens_seen": 609222656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041180541624874625,
      "loss": 2.6473,
      "theoretical_loss": 3.8337624025937957,
      "tokens_seen": 609288192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041179538615847543,
      "loss": 2.7189,
      "theoretical_loss": 3.8337194924831053,
      "tokens_seen": 609353728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117853560682046,
      "loss": 2.7765,
      "theoretical_loss": 3.833676588279187,
      "tokens_seen": 609419264
    },
    {
      "epoch": 0.18,
      "objective/train/docs_used": 346111,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7798197269439697,
      "objective/train/theoretical_loss": 3.8336336899805916,
      "objective/train/tokens_used": 629944800,
      "theoretical_loss": 3.8336336899805916,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041177532597793385,
      "loss": 2.5951,
      "theoretical_loss": 3.8336336899805916,
      "tokens_seen": 609484800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000411765295887663,
      "loss": 2.6398,
      "theoretical_loss": 3.833590797585872,
      "tokens_seen": 609550336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117552657973922,
      "loss": 2.7884,
      "theoretical_loss": 3.833547911093581,
      "tokens_seen": 609615872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041174523570712134,
      "loss": 2.7703,
      "theoretical_loss": 3.8335050305022715,
      "tokens_seen": 609681408
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117352056168506,
      "loss": 2.696,
      "theoretical_loss": 3.833462155810498,
      "tokens_seen": 609746944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041172517552657975,
      "loss": 2.8528,
      "theoretical_loss": 3.833419287016815,
      "tokens_seen": 609812480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041171514543630894,
      "loss": 2.6289,
      "theoretical_loss": 3.8333764241197765,
      "tokens_seen": 609878016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117051153460381,
      "loss": 2.6666,
      "theoretical_loss": 3.833333567117939,
      "tokens_seen": 609943552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116950852557673,
      "loss": 2.97,
      "theoretical_loss": 3.8332907160098575,
      "tokens_seen": 610009088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116850551654965,
      "loss": 2.8593,
      "theoretical_loss": 3.8332478707940885,
      "tokens_seen": 610074624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116750250752257,
      "loss": 2.6773,
      "theoretical_loss": 3.8332050314691894,
      "tokens_seen": 610140160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041166499498495484,
      "loss": 2.8794,
      "theoretical_loss": 3.833162198033718,
      "tokens_seen": 610205696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116549648946841,
      "loss": 2.9646,
      "theoretical_loss": 3.833119370486231,
      "tokens_seen": 610271232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116449348044132,
      "loss": 2.7601,
      "theoretical_loss": 3.8330765488252876,
      "tokens_seen": 610336768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041163490471414244,
      "loss": 2.6683,
      "theoretical_loss": 3.8330337330494473,
      "tokens_seen": 610402304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116248746238716,
      "loss": 2.8002,
      "theoretical_loss": 3.8329909231572685,
      "tokens_seen": 610467840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004116148445336008,
      "loss": 2.7484,
      "theoretical_loss": 3.832948119147312,
      "tokens_seen": 610533376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041160481444333,
      "loss": 2.6991,
      "theoretical_loss": 3.8329053210181376,
      "tokens_seen": 610598912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115947843530592,
      "loss": 2.7888,
      "theoretical_loss": 3.8328625287683065,
      "tokens_seen": 610664448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041158475426278834,
      "loss": 2.8296,
      "theoretical_loss": 3.8328197423963806,
      "tokens_seen": 610729984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115747241725176,
      "loss": 2.712,
      "theoretical_loss": 3.8327769619009215,
      "tokens_seen": 610795520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115646940822467,
      "loss": 2.8347,
      "theoretical_loss": 3.832734187280492,
      "tokens_seen": 610861056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041155466399197594,
      "loss": 2.9096,
      "theoretical_loss": 3.832691418533655,
      "tokens_seen": 610926592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115446339017051,
      "loss": 2.4744,
      "theoretical_loss": 3.832648655658973,
      "tokens_seen": 610992128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115346038114343,
      "loss": 2.8143,
      "theoretical_loss": 3.8326058986550127,
      "tokens_seen": 611057664
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 346814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.812352180480957,
      "objective/train/theoretical_loss": 3.832563147520336,
      "objective/train/tokens_used": 631583200,
      "theoretical_loss": 3.832563147520336,
      "tokens_seen": 611123200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115245737211635,
      "loss": 2.8388,
      "theoretical_loss": 3.832563147520336,
      "tokens_seen": 611123200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041151454363089267,
      "loss": 2.8298,
      "theoretical_loss": 3.8325204022535084,
      "tokens_seen": 611188736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041150451354062185,
      "loss": 2.8495,
      "theoretical_loss": 3.832477662853096,
      "tokens_seen": 611254272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114944834503511,
      "loss": 2.829,
      "theoretical_loss": 3.8324349293176656,
      "tokens_seen": 611319808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114844533600802,
      "loss": 2.6977,
      "theoretical_loss": 3.8323922016457823,
      "tokens_seen": 611385344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041147442326980944,
      "loss": 2.774,
      "theoretical_loss": 3.8323494798360134,
      "tokens_seen": 611450880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041146439317953857,
      "loss": 2.7721,
      "theoretical_loss": 3.832306763886927,
      "tokens_seen": 611516416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114543630892678,
      "loss": 2.8789,
      "theoretical_loss": 3.8322640537970916,
      "tokens_seen": 611581952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000411444332998997,
      "loss": 2.794,
      "theoretical_loss": 3.8322213495650743,
      "tokens_seen": 611647488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041143430290872617,
      "loss": 2.6454,
      "theoretical_loss": 3.8321786511894453,
      "tokens_seen": 611713024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041142427281845535,
      "loss": 2.9449,
      "theoretical_loss": 3.8321359586687738,
      "tokens_seen": 611778560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114142427281846,
      "loss": 2.8482,
      "theoretical_loss": 3.83209327200163,
      "tokens_seen": 611844096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041140421263791377,
      "loss": 2.7537,
      "theoretical_loss": 3.832050591186584,
      "tokens_seen": 611909632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041139418254764295,
      "loss": 2.6385,
      "theoretical_loss": 3.832007916222208,
      "tokens_seen": 611975168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041138415245737213,
      "loss": 2.9391,
      "theoretical_loss": 3.8319652471070724,
      "tokens_seen": 612040704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113741223671013,
      "loss": 2.7014,
      "theoretical_loss": 3.8319225838397495,
      "tokens_seen": 612106240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041136409227683055,
      "loss": 2.8633,
      "theoretical_loss": 3.8318799264188126,
      "tokens_seen": 612171776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041135406218655967,
      "loss": 2.7242,
      "theoretical_loss": 3.831837274842834,
      "tokens_seen": 612237312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113440320962889,
      "loss": 2.6845,
      "theoretical_loss": 3.8317946291103873,
      "tokens_seen": 612302848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041133400200601803,
      "loss": 2.723,
      "theoretical_loss": 3.8317519892200473,
      "tokens_seen": 612368384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041132397191574727,
      "loss": 2.8871,
      "theoretical_loss": 3.8317093551703882,
      "tokens_seen": 612433920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041131394182547645,
      "loss": 2.9039,
      "theoretical_loss": 3.8316667269599844,
      "tokens_seen": 612499456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041130391173520563,
      "loss": 2.7024,
      "theoretical_loss": 3.831624104587413,
      "tokens_seen": 612564992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112938816449348,
      "loss": 2.6651,
      "theoretical_loss": 3.831581488051248,
      "tokens_seen": 612630528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041128385155466405,
      "loss": 2.7755,
      "theoretical_loss": 3.8315388773500683,
      "tokens_seen": 612696064
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 348068,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.487290859222412,
      "objective/train/theoretical_loss": 3.8314962724824495,
      "objective/train/tokens_used": 633221600,
      "theoretical_loss": 3.8314962724824495,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112738214643932,
      "loss": 2.6193,
      "theoretical_loss": 3.8314962724824495,
      "tokens_seen": 612761600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112637913741224,
      "loss": 2.9404,
      "theoretical_loss": 3.831453673446969,
      "tokens_seen": 612827136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041125376128385154,
      "loss": 2.6889,
      "theoretical_loss": 3.831411080242206,
      "tokens_seen": 612892672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112437311935808,
      "loss": 2.8746,
      "theoretical_loss": 3.831368492866738,
      "tokens_seen": 612958208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041123370110330995,
      "loss": 2.7783,
      "theoretical_loss": 3.831325911319145,
      "tokens_seen": 613023744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041122367101303914,
      "loss": 2.529,
      "theoretical_loss": 3.8312833355980063,
      "tokens_seen": 613089280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112136409227683,
      "loss": 2.8101,
      "theoretical_loss": 3.831240765701901,
      "tokens_seen": 613154816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112036108324975,
      "loss": 2.5207,
      "theoretical_loss": 3.8311982016294115,
      "tokens_seen": 613220352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111935807422267,
      "loss": 2.7324,
      "theoretical_loss": 3.831155643379117,
      "tokens_seen": 613285888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111835506519559,
      "loss": 2.7735,
      "theoretical_loss": 3.831113090949601,
      "tokens_seen": 613351424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041117352056168504,
      "loss": 2.91,
      "theoretical_loss": 3.8310705443394433,
      "tokens_seen": 613416960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111634904714143,
      "loss": 3.0031,
      "theoretical_loss": 3.8310280035472286,
      "tokens_seen": 613482496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111534603811434,
      "loss": 2.77,
      "theoretical_loss": 3.830985468571538,
      "tokens_seen": 613548032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041114343029087264,
      "loss": 2.6051,
      "theoretical_loss": 3.8309429394109573,
      "tokens_seen": 613613568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111334002006018,
      "loss": 2.6809,
      "theoretical_loss": 3.8309004160640683,
      "tokens_seen": 613679104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000411123370110331,
      "loss": 2.5764,
      "theoretical_loss": 3.830857898529457,
      "tokens_seen": 613744640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111133400200602,
      "loss": 2.8588,
      "theoretical_loss": 3.8308153868057078,
      "tokens_seen": 613810176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111033099297894,
      "loss": 2.6019,
      "theoretical_loss": 3.8307728808914066,
      "tokens_seen": 613875712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041109327983951854,
      "loss": 2.6243,
      "theoretical_loss": 3.8307303807851394,
      "tokens_seen": 613941248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110832497492478,
      "loss": 2.9228,
      "theoretical_loss": 3.8306878864854923,
      "tokens_seen": 614006784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110732196589769,
      "loss": 2.6788,
      "theoretical_loss": 3.8306453979910526,
      "tokens_seen": 614072320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041106318956870614,
      "loss": 2.7896,
      "theoretical_loss": 3.830602915300408,
      "tokens_seen": 614137856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110531594784353,
      "loss": 2.8886,
      "theoretical_loss": 3.8305604384121468,
      "tokens_seen": 614203392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110431293881645,
      "loss": 2.6726,
      "theoretical_loss": 3.830517967324856,
      "tokens_seen": 614268928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110330992978937,
      "loss": 2.7219,
      "theoretical_loss": 3.8304755020371264,
      "tokens_seen": 614334464
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 348575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.63563871383667,
      "objective/train/theoretical_loss": 3.830433042547546,
      "objective/train/tokens_used": 634860000,
      "theoretical_loss": 3.830433042547546,
      "tokens_seen": 614400000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041102306920762287,
      "loss": 2.8635,
      "theoretical_loss": 3.830433042547546,
      "tokens_seen": 614400000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041101303911735205,
      "loss": 2.7305,
      "theoretical_loss": 3.8303905888547067,
      "tokens_seen": 614465536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110030090270813,
      "loss": 2.823,
      "theoretical_loss": 3.830348140957197,
      "tokens_seen": 614531072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109929789368104,
      "loss": 2.6875,
      "theoretical_loss": 3.830305698853609,
      "tokens_seen": 614596608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041098294884653965,
      "loss": 2.8067,
      "theoretical_loss": 3.830263262542534,
      "tokens_seen": 614662144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041097291875626877,
      "loss": 2.755,
      "theoretical_loss": 3.8302208320225635,
      "tokens_seen": 614727680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410962888665998,
      "loss": 2.7622,
      "theoretical_loss": 3.8301784072922898,
      "tokens_seen": 614793216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109528585757272,
      "loss": 2.7521,
      "theoretical_loss": 3.8301359883503068,
      "tokens_seen": 614858752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041094282848545637,
      "loss": 2.606,
      "theoretical_loss": 3.8300935751952077,
      "tokens_seen": 614924288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041093279839518555,
      "loss": 2.7874,
      "theoretical_loss": 3.8300511678255855,
      "tokens_seen": 614989824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109227683049148,
      "loss": 2.6015,
      "theoretical_loss": 3.830008766240036,
      "tokens_seen": 615055360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109127382146439,
      "loss": 2.6941,
      "theoretical_loss": 3.8299663704371527,
      "tokens_seen": 615120896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041090270812437315,
      "loss": 2.5843,
      "theoretical_loss": 3.829923980415532,
      "tokens_seen": 615186432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108926780341023,
      "loss": 2.8617,
      "theoretical_loss": 3.8298815961737693,
      "tokens_seen": 615251968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108826479438315,
      "loss": 2.7504,
      "theoretical_loss": 3.829839217710462,
      "tokens_seen": 615317504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108726178535607,
      "loss": 2.6976,
      "theoretical_loss": 3.8297968450242053,
      "tokens_seen": 615383040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041086258776328987,
      "loss": 2.874,
      "theoretical_loss": 3.829754478113597,
      "tokens_seen": 615448576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041085255767301905,
      "loss": 2.6128,
      "theoretical_loss": 3.829712116977236,
      "tokens_seen": 615514112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041084252758274823,
      "loss": 2.6613,
      "theoretical_loss": 3.82966976161372,
      "tokens_seen": 615579648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108324974924774,
      "loss": 2.7108,
      "theoretical_loss": 3.8296274120216474,
      "tokens_seen": 615645184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041082246740220665,
      "loss": 2.5394,
      "theoretical_loss": 3.829585068199618,
      "tokens_seen": 615710720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108124373119358,
      "loss": 2.7524,
      "theoretical_loss": 3.8295427301462324,
      "tokens_seen": 615776256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410802407221665,
      "loss": 2.8075,
      "theoretical_loss": 3.8295003978600892,
      "tokens_seen": 615841792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107923771313942,
      "loss": 2.7175,
      "theoretical_loss": 3.82945807133979,
      "tokens_seen": 615907328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107823470411234,
      "loss": 2.7832,
      "theoretical_loss": 3.8294157505839364,
      "tokens_seen": 615972864
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 349861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8283627033233643,
      "objective/train/theoretical_loss": 3.8293734355911297,
      "objective/train/tokens_used": 636498400,
      "theoretical_loss": 3.8293734355911297,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041077231695085256,
      "loss": 2.578,
      "theoretical_loss": 3.8293734355911297,
      "tokens_seen": 616038400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041076228686058174,
      "loss": 2.726,
      "theoretical_loss": 3.829331126359972,
      "tokens_seen": 616103936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107522567703109,
      "loss": 2.8416,
      "theoretical_loss": 3.8292888228890662,
      "tokens_seen": 616169472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041074222668004015,
      "loss": 2.7834,
      "theoretical_loss": 3.8292465251770165,
      "tokens_seen": 616235008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107321965897693,
      "loss": 2.5378,
      "theoretical_loss": 3.829204233222425,
      "tokens_seen": 616300544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107221664994985,
      "loss": 2.7511,
      "theoretical_loss": 3.829161947023897,
      "tokens_seen": 616366080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041071213640922764,
      "loss": 2.5905,
      "theoretical_loss": 3.829119666580037,
      "tokens_seen": 616431616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107021063189569,
      "loss": 2.8605,
      "theoretical_loss": 3.82907739188945,
      "tokens_seen": 616497152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041069207622868606,
      "loss": 2.9454,
      "theoretical_loss": 3.8290351229507418,
      "tokens_seen": 616562688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041068204613841524,
      "loss": 2.8061,
      "theoretical_loss": 3.8289928597625185,
      "tokens_seen": 616628224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106720160481444,
      "loss": 2.7851,
      "theoretical_loss": 3.828950602323386,
      "tokens_seen": 616693760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106619859578736,
      "loss": 2.7758,
      "theoretical_loss": 3.828908350631953,
      "tokens_seen": 616759296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041065195586760284,
      "loss": 2.8322,
      "theoretical_loss": 3.828866104686825,
      "tokens_seen": 616824832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410641925777332,
      "loss": 2.7427,
      "theoretical_loss": 3.828823864486612,
      "tokens_seen": 616890368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106318956870612,
      "loss": 2.854,
      "theoretical_loss": 3.828781630029922,
      "tokens_seen": 616955904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106218655967904,
      "loss": 2.7491,
      "theoretical_loss": 3.8287394013153637,
      "tokens_seen": 617021440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106118355065196,
      "loss": 2.5018,
      "theoretical_loss": 3.828697178341547,
      "tokens_seen": 617086976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041060180541624874,
      "loss": 2.7982,
      "theoretical_loss": 3.8286549611070813,
      "tokens_seen": 617152512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410591775325978,
      "loss": 2.7268,
      "theoretical_loss": 3.8286127496105777,
      "tokens_seen": 617218048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105817452357071,
      "loss": 2.8512,
      "theoretical_loss": 3.828570543850647,
      "tokens_seen": 617283584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041057171514543634,
      "loss": 2.7221,
      "theoretical_loss": 3.8285283438259015,
      "tokens_seen": 617349120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105616850551655,
      "loss": 2.7407,
      "theoretical_loss": 3.8284861495349514,
      "tokens_seen": 617414656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105516549648947,
      "loss": 2.6172,
      "theoretical_loss": 3.828443960976411,
      "tokens_seen": 617480192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105416248746239,
      "loss": 2.8896,
      "theoretical_loss": 3.828401778148892,
      "tokens_seen": 617545728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041053159478435307,
      "loss": 2.5109,
      "theoretical_loss": 3.828359601051008,
      "tokens_seen": 617611264
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 350619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7856407165527344,
      "objective/train/theoretical_loss": 3.8283174296813733,
      "objective/train/tokens_used": 638136800,
      "theoretical_loss": 3.8283174296813733,
      "tokens_seen": 617676800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041052156469408225,
      "loss": 2.6453,
      "theoretical_loss": 3.8283174296813733,
      "tokens_seen": 617676800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105115346038115,
      "loss": 2.6044,
      "theoretical_loss": 3.828275264038602,
      "tokens_seen": 617742336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105015045135406,
      "loss": 2.5576,
      "theoretical_loss": 3.828233104121309,
      "tokens_seen": 617807872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041049147442326985,
      "loss": 2.6203,
      "theoretical_loss": 3.8281909499281097,
      "tokens_seen": 617873408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041048144433299897,
      "loss": 2.6868,
      "theoretical_loss": 3.82814880145762,
      "tokens_seen": 617938944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104714142427282,
      "loss": 2.7666,
      "theoretical_loss": 3.828106658708456,
      "tokens_seen": 618004480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104613841524574,
      "loss": 2.8267,
      "theoretical_loss": 3.8280645216792344,
      "tokens_seen": 618070016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041045135406218657,
      "loss": 2.9681,
      "theoretical_loss": 3.828022390368572,
      "tokens_seen": 618135552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041044132397191575,
      "loss": 2.6423,
      "theoretical_loss": 3.8279802647750882,
      "tokens_seen": 618201088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410431293881645,
      "loss": 2.4776,
      "theoretical_loss": 3.8279381448973995,
      "tokens_seen": 618266624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104212637913741,
      "loss": 2.5911,
      "theoretical_loss": 3.827896030734126,
      "tokens_seen": 618332160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041041123370110335,
      "loss": 2.4597,
      "theoretical_loss": 3.8278539222838854,
      "tokens_seen": 618397696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104012036108325,
      "loss": 2.8388,
      "theoretical_loss": 3.827811819545299,
      "tokens_seen": 618463232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103911735205617,
      "loss": 2.7457,
      "theoretical_loss": 3.827769722516985,
      "tokens_seen": 618528768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103811434302909,
      "loss": 2.8415,
      "theoretical_loss": 3.8277276311975656,
      "tokens_seen": 618594304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041037111334002007,
      "loss": 2.8539,
      "theoretical_loss": 3.8276855455856618,
      "tokens_seen": 618659840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041036108324974925,
      "loss": 2.7644,
      "theoretical_loss": 3.8276434656798948,
      "tokens_seen": 618725376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041035105315947844,
      "loss": 2.6033,
      "theoretical_loss": 3.8276013914788862,
      "tokens_seen": 618790912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103410230692076,
      "loss": 2.9456,
      "theoretical_loss": 3.827559322981259,
      "tokens_seen": 618856448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041033099297893685,
      "loss": 2.5836,
      "theoretical_loss": 3.827517260185637,
      "tokens_seen": 618921984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410320962888666,
      "loss": 2.8173,
      "theoretical_loss": 3.827475203090642,
      "tokens_seen": 618987520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103109327983952,
      "loss": 2.898,
      "theoretical_loss": 3.827433151694899,
      "tokens_seen": 619053056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103009027081244,
      "loss": 2.7016,
      "theoretical_loss": 3.827391105997033,
      "tokens_seen": 619118592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102908726178536,
      "loss": 2.7453,
      "theoretical_loss": 3.8273490659956675,
      "tokens_seen": 619184128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041028084252758276,
      "loss": 2.6333,
      "theoretical_loss": 3.827307031689429,
      "tokens_seen": 619249664
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 351768,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.970876693725586,
      "objective/train/theoretical_loss": 3.8272650030769433,
      "objective/train/tokens_used": 639775200,
      "theoretical_loss": 3.8272650030769433,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041027081243731194,
      "loss": 2.797,
      "theoretical_loss": 3.8272650030769433,
      "tokens_seen": 619315200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102607823470411,
      "loss": 2.6878,
      "theoretical_loss": 3.8272229801568365,
      "tokens_seen": 619380736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041025075225677035,
      "loss": 2.6138,
      "theoretical_loss": 3.827180962927735,
      "tokens_seen": 619446272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102407221664995,
      "loss": 2.7821,
      "theoretical_loss": 3.827138951388267,
      "tokens_seen": 619511808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102306920762287,
      "loss": 2.8132,
      "theoretical_loss": 3.8270969455370594,
      "tokens_seen": 619577344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041022066198595784,
      "loss": 2.7769,
      "theoretical_loss": 3.8270549453727414,
      "tokens_seen": 619642880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102106318956871,
      "loss": 2.6755,
      "theoretical_loss": 3.8270129508939412,
      "tokens_seen": 619708416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041020060180541626,
      "loss": 2.593,
      "theoretical_loss": 3.8269709620992884,
      "tokens_seen": 619773952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041019057171514544,
      "loss": 2.873,
      "theoretical_loss": 3.826928978987412,
      "tokens_seen": 619839488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101805416248746,
      "loss": 2.8805,
      "theoretical_loss": 3.8268870015569427,
      "tokens_seen": 619905024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101705115346038,
      "loss": 2.6737,
      "theoretical_loss": 3.826845029806511,
      "tokens_seen": 619970560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410160481444333,
      "loss": 2.5944,
      "theoretical_loss": 3.826803063734748,
      "tokens_seen": 620036096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101504513540622,
      "loss": 2.7064,
      "theoretical_loss": 3.826761103340286,
      "tokens_seen": 620101632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041014042126379135,
      "loss": 2.834,
      "theoretical_loss": 3.8267191486217556,
      "tokens_seen": 620167168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101303911735206,
      "loss": 2.5143,
      "theoretical_loss": 3.8266771995777904,
      "tokens_seen": 620232704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041012036108324976,
      "loss": 2.864,
      "theoretical_loss": 3.826635256207023,
      "tokens_seen": 620298240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041011033099297894,
      "loss": 2.5869,
      "theoretical_loss": 3.8265933185080874,
      "tokens_seen": 620363776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101003009027081,
      "loss": 2.7903,
      "theoretical_loss": 3.8265513864796166,
      "tokens_seen": 620429312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100902708124373,
      "loss": 2.7977,
      "theoretical_loss": 3.8265094601202465,
      "tokens_seen": 620494848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100802407221665,
      "loss": 2.7521,
      "theoretical_loss": 3.8264675394286103,
      "tokens_seen": 620560384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100702106318957,
      "loss": 3.0977,
      "theoretical_loss": 3.8264256244033445,
      "tokens_seen": 620625920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041006018054162485,
      "loss": 2.8243,
      "theoretical_loss": 3.826383715043085,
      "tokens_seen": 620691456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100501504513541,
      "loss": 2.8642,
      "theoretical_loss": 3.826341811346468,
      "tokens_seen": 620756992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100401203610832,
      "loss": 2.6281,
      "theoretical_loss": 3.8262999133121296,
      "tokens_seen": 620822528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041003009027081245,
      "loss": 2.935,
      "theoretical_loss": 3.826258020938708,
      "tokens_seen": 620888064
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 352593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.924837827682495,
      "objective/train/theoretical_loss": 3.82621613422484,
      "objective/train/tokens_used": 641413600,
      "theoretical_loss": 3.82621613422484,
      "tokens_seen": 620953600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041002006018054163,
      "loss": 2.7235,
      "theoretical_loss": 3.82621613422484,
      "tokens_seen": 620953600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100100300902708,
      "loss": 2.9794,
      "theoretical_loss": 3.8261742531691647,
      "tokens_seen": 621019136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041,
      "loss": 2.6264,
      "theoretical_loss": 3.8261323777703207,
      "tokens_seen": 621084672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040998996990972917,
      "loss": 2.7152,
      "theoretical_loss": 3.8260905080269465,
      "tokens_seen": 621150208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040997993981945835,
      "loss": 2.9183,
      "theoretical_loss": 3.8260486439376824,
      "tokens_seen": 621215744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099699097291876,
      "loss": 2.8273,
      "theoretical_loss": 3.826006785501168,
      "tokens_seen": 621281280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099598796389167,
      "loss": 2.8176,
      "theoretical_loss": 3.8259649327160448,
      "tokens_seen": 621346816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040994984954864595,
      "loss": 2.6598,
      "theoretical_loss": 3.8259230855809525,
      "tokens_seen": 621412352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040993981945837513,
      "loss": 2.8564,
      "theoretical_loss": 3.825881244094533,
      "tokens_seen": 621477888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099297893681043,
      "loss": 2.8814,
      "theoretical_loss": 3.825839408255429,
      "tokens_seen": 621543424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099197592778335,
      "loss": 2.97,
      "theoretical_loss": 3.825797578062282,
      "tokens_seen": 621608960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099097291875627,
      "loss": 2.5886,
      "theoretical_loss": 3.825755753513736,
      "tokens_seen": 621674496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098996990972919,
      "loss": 2.8166,
      "theoretical_loss": 3.8257139346084332,
      "tokens_seen": 621740032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098896690070211,
      "loss": 2.6562,
      "theoretical_loss": 3.8256721213450184,
      "tokens_seen": 621805568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098796389167503,
      "loss": 2.6982,
      "theoretical_loss": 3.8256303137221357,
      "tokens_seen": 621871104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040986960882647945,
      "loss": 2.6747,
      "theoretical_loss": 3.825588511738429,
      "tokens_seen": 621936640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040985957873620864,
      "loss": 2.7269,
      "theoretical_loss": 3.8255467153925453,
      "tokens_seen": 622002176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098495486459378,
      "loss": 2.8566,
      "theoretical_loss": 3.8255049246831287,
      "tokens_seen": 622067712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040983951855566705,
      "loss": 2.934,
      "theoretical_loss": 3.825463139608827,
      "tokens_seen": 622133248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098294884653962,
      "loss": 2.9967,
      "theoretical_loss": 3.825421360168285,
      "tokens_seen": 622198784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098194583751254,
      "loss": 2.5096,
      "theoretical_loss": 3.8253795863601505,
      "tokens_seen": 622264320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098094282848546,
      "loss": 2.7312,
      "theoretical_loss": 3.825337818183072,
      "tokens_seen": 622329856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097993981945838,
      "loss": 2.5197,
      "theoretical_loss": 3.8252960556356967,
      "tokens_seen": 622395392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040978936810431296,
      "loss": 2.8355,
      "theoretical_loss": 3.8252542987166738,
      "tokens_seen": 622460928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040977933801404214,
      "loss": 2.6231,
      "theoretical_loss": 3.8252125474246514,
      "tokens_seen": 622526464
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 353199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.111541748046875,
      "objective/train/theoretical_loss": 3.8251708017582793,
      "objective/train/tokens_used": 643052000,
      "theoretical_loss": 3.8251708017582793,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097693079237713,
      "loss": 2.7705,
      "theoretical_loss": 3.8251708017582793,
      "tokens_seen": 622592000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040975927783350055,
      "loss": 2.8652,
      "theoretical_loss": 3.825129061716208,
      "tokens_seen": 622657536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097492477432297,
      "loss": 2.6621,
      "theoretical_loss": 3.825087327297087,
      "tokens_seen": 622723072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097392176529589,
      "loss": 2.7638,
      "theoretical_loss": 3.8250455984995684,
      "tokens_seen": 622788608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040972918756268804,
      "loss": 2.8562,
      "theoretical_loss": 3.825003875322302,
      "tokens_seen": 622854144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097191574724173,
      "loss": 2.762,
      "theoretical_loss": 3.8249621577639408,
      "tokens_seen": 622919680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040970912738214646,
      "loss": 2.8821,
      "theoretical_loss": 3.8249204458231363,
      "tokens_seen": 622985216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040969909729187564,
      "loss": 2.8805,
      "theoretical_loss": 3.8248787394985415,
      "tokens_seen": 623050752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096890672016048,
      "loss": 2.7553,
      "theoretical_loss": 3.82483703878881,
      "tokens_seen": 623116288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409679037111334,
      "loss": 2.7981,
      "theoretical_loss": 3.824795343692595,
      "tokens_seen": 623181824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096690070210632,
      "loss": 2.5148,
      "theoretical_loss": 3.8247536542085507,
      "tokens_seen": 623247360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096589769307924,
      "loss": 2.7058,
      "theoretical_loss": 3.824711970335332,
      "tokens_seen": 623312896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040964894684052155,
      "loss": 2.8072,
      "theoretical_loss": 3.824670292071593,
      "tokens_seen": 623378432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096389167502508,
      "loss": 2.7979,
      "theoretical_loss": 3.82462861941599,
      "tokens_seen": 623443968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040962888665997996,
      "loss": 2.6498,
      "theoretical_loss": 3.8245869523671794,
      "tokens_seen": 623509504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040961885656970914,
      "loss": 2.844,
      "theoretical_loss": 3.8245452909238167,
      "tokens_seen": 623575040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096088264794383,
      "loss": 2.9147,
      "theoretical_loss": 3.824503635084559,
      "tokens_seen": 623640576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095987963891675,
      "loss": 2.7143,
      "theoretical_loss": 3.8244619848480648,
      "tokens_seen": 623706112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095887662988967,
      "loss": 2.8396,
      "theoretical_loss": 3.82442034021299,
      "tokens_seen": 623771648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095787362086259,
      "loss": 2.7964,
      "theoretical_loss": 3.824378701177994,
      "tokens_seen": 623837184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040956870611835505,
      "loss": 2.6739,
      "theoretical_loss": 3.8243370677417357,
      "tokens_seen": 623902720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095586760280843,
      "loss": 2.6736,
      "theoretical_loss": 3.8242954399028743,
      "tokens_seen": 623968256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095486459378134,
      "loss": 2.6681,
      "theoretical_loss": 3.824253817660069,
      "tokens_seen": 624033792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040953861584754265,
      "loss": 2.8982,
      "theoretical_loss": 3.82421220101198,
      "tokens_seen": 624099328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040952858575727183,
      "loss": 2.914,
      "theoretical_loss": 3.8241705899572684,
      "tokens_seen": 624164864
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 354273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4973878860473633,
      "objective/train/theoretical_loss": 3.8241289844945947,
      "objective/train/tokens_used": 644690400,
      "theoretical_loss": 3.8241289844945947,
      "tokens_seen": 624230400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409518555667001,
      "loss": 2.747,
      "theoretical_loss": 3.8241289844945947,
      "tokens_seen": 624230400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095085255767302,
      "loss": 2.6438,
      "theoretical_loss": 3.8240873846226204,
      "tokens_seen": 624295936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040949849548645937,
      "loss": 2.6574,
      "theoretical_loss": 3.8240457903400085,
      "tokens_seen": 624361472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040948846539618855,
      "loss": 2.902,
      "theoretical_loss": 3.82400420164542,
      "tokens_seen": 624427008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094784353059178,
      "loss": 2.9509,
      "theoretical_loss": 3.823962618537519,
      "tokens_seen": 624492544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094684052156469,
      "loss": 2.9318,
      "theoretical_loss": 3.8239210410149678,
      "tokens_seen": 624558080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040945837512537615,
      "loss": 2.6767,
      "theoretical_loss": 3.823879469076431,
      "tokens_seen": 624623616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040944834503510533,
      "loss": 2.8474,
      "theoretical_loss": 3.8238379027205727,
      "tokens_seen": 624689152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094383149448345,
      "loss": 2.7381,
      "theoretical_loss": 3.823796341946058,
      "tokens_seen": 624754688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094282848545637,
      "loss": 2.8622,
      "theoretical_loss": 3.823754786751551,
      "tokens_seen": 624820224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094182547642929,
      "loss": 2.8514,
      "theoretical_loss": 3.8237132371357188,
      "tokens_seen": 624885760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040940822467402206,
      "loss": 2.8813,
      "theoretical_loss": 3.823671693097226,
      "tokens_seen": 624951296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093981945837513,
      "loss": 2.614,
      "theoretical_loss": 3.8236301546347407,
      "tokens_seen": 625016832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093881644934804,
      "loss": 2.7085,
      "theoretical_loss": 3.8235886217469286,
      "tokens_seen": 625082368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040937813440320965,
      "loss": 2.7141,
      "theoretical_loss": 3.8235470944324588,
      "tokens_seen": 625147904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093681043129388,
      "loss": 2.7833,
      "theoretical_loss": 3.823505572689997,
      "tokens_seen": 625213440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409358074222668,
      "loss": 2.8488,
      "theoretical_loss": 3.8234640565182136,
      "tokens_seen": 625278976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093480441323972,
      "loss": 2.8908,
      "theoretical_loss": 3.8234225459157773,
      "tokens_seen": 625344512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093380140421264,
      "loss": 2.879,
      "theoretical_loss": 3.823381040881356,
      "tokens_seen": 625410048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040932798395185556,
      "loss": 2.656,
      "theoretical_loss": 3.8233395414136204,
      "tokens_seen": 625475584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093179538615848,
      "loss": 2.5753,
      "theoretical_loss": 3.8232980475112415,
      "tokens_seen": 625541120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093079237713139,
      "loss": 2.5492,
      "theoretical_loss": 3.8232565591728886,
      "tokens_seen": 625606656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040929789368104316,
      "loss": 2.7042,
      "theoretical_loss": 3.8232150763972337,
      "tokens_seen": 625672192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092878635907723,
      "loss": 2.7082,
      "theoretical_loss": 3.8231735991829483,
      "tokens_seen": 625737728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092778335005015,
      "loss": 2.8102,
      "theoretical_loss": 3.8231321275287042,
      "tokens_seen": 625803264
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 354855,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.032952308654785,
      "objective/train/theoretical_loss": 3.8230906614331737,
      "objective/train/tokens_used": 646328800,
      "theoretical_loss": 3.8230906614331737,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092678034102307,
      "loss": 2.8881,
      "theoretical_loss": 3.8230906614331737,
      "tokens_seen": 625868800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092577733199599,
      "loss": 2.8983,
      "theoretical_loss": 3.823049200895031,
      "tokens_seen": 625934336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040924774322968906,
      "loss": 2.857,
      "theoretical_loss": 3.823007745912948,
      "tokens_seen": 625999872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040923771313941824,
      "loss": 3.0152,
      "theoretical_loss": 3.8229662964855997,
      "tokens_seen": 626065408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092276830491474,
      "loss": 2.9206,
      "theoretical_loss": 3.82292485261166,
      "tokens_seen": 626130944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040921765295887666,
      "loss": 2.6842,
      "theoretical_loss": 3.822883414289804,
      "tokens_seen": 626196480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092076228686058,
      "loss": 2.9224,
      "theoretical_loss": 3.8228419815187062,
      "tokens_seen": 626262016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409197592778335,
      "loss": 2.6684,
      "theoretical_loss": 3.822800554297043,
      "tokens_seen": 626327552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040918756268806415,
      "loss": 2.9722,
      "theoretical_loss": 3.8227591326234904,
      "tokens_seen": 626393088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091775325977934,
      "loss": 2.6933,
      "theoretical_loss": 3.822717716496725,
      "tokens_seen": 626458624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040916750250752257,
      "loss": 2.9309,
      "theoretical_loss": 3.822676305915424,
      "tokens_seen": 626524160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040915747241725175,
      "loss": 2.9623,
      "theoretical_loss": 3.822634900878265,
      "tokens_seen": 626589696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409147442326981,
      "loss": 2.7984,
      "theoretical_loss": 3.8225935013839254,
      "tokens_seen": 626655232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040913741223671016,
      "loss": 2.7617,
      "theoretical_loss": 3.8225521074310844,
      "tokens_seen": 626720768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040912738214643934,
      "loss": 2.9621,
      "theoretical_loss": 3.82251071901842,
      "tokens_seen": 626786304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091173520561685,
      "loss": 2.7483,
      "theoretical_loss": 3.8224693361446134,
      "tokens_seen": 626851840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091073219658977,
      "loss": 2.4055,
      "theoretical_loss": 3.822427958808342,
      "tokens_seen": 626917376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090972918756269,
      "loss": 2.9295,
      "theoretical_loss": 3.822386587008287,
      "tokens_seen": 626982912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090872617853561,
      "loss": 2.787,
      "theoretical_loss": 3.82234522074313,
      "tokens_seen": 627048448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040907723169508525,
      "loss": 2.6379,
      "theoretical_loss": 3.8223038600115506,
      "tokens_seen": 627113984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090672016048145,
      "loss": 2.8665,
      "theoretical_loss": 3.8222625048122314,
      "tokens_seen": 627179520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090571715145436,
      "loss": 2.844,
      "theoretical_loss": 3.8222211551438545,
      "tokens_seen": 627245056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040904714142427285,
      "loss": 2.5677,
      "theoretical_loss": 3.8221798110051024,
      "tokens_seen": 627310592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040903711133400203,
      "loss": 2.8402,
      "theoretical_loss": 3.8221384723946574,
      "tokens_seen": 627376128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090270812437312,
      "loss": 2.7906,
      "theoretical_loss": 3.8220971393112033,
      "tokens_seen": 627441664
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 356151,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0492939949035645,
      "objective/train/theoretical_loss": 3.8220558117534242,
      "objective/train/tokens_used": 647967200,
      "theoretical_loss": 3.8220558117534242,
      "tokens_seen": 627507200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090170511534604,
      "loss": 2.7642,
      "theoretical_loss": 3.8220558117534242,
      "tokens_seen": 627507200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040900702106318957,
      "loss": 2.7001,
      "theoretical_loss": 3.8220144897200035,
      "tokens_seen": 627572736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040899699097291875,
      "loss": 3.154,
      "theoretical_loss": 3.821973173209628,
      "tokens_seen": 627638272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408986960882648,
      "loss": 2.9608,
      "theoretical_loss": 3.82193186222098,
      "tokens_seen": 627703808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089769307923771,
      "loss": 2.7454,
      "theoretical_loss": 3.821890556752748,
      "tokens_seen": 627769344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040896690070210635,
      "loss": 2.8077,
      "theoretical_loss": 3.821849256803617,
      "tokens_seen": 627834880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040895687061183553,
      "loss": 2.6292,
      "theoretical_loss": 3.8218079623722723,
      "tokens_seen": 627900416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089468405215647,
      "loss": 2.6008,
      "theoretical_loss": 3.821766673457403,
      "tokens_seen": 627965952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089368104312939,
      "loss": 2.8147,
      "theoretical_loss": 3.8217253900576953,
      "tokens_seen": 628031488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089267803410231,
      "loss": 2.9192,
      "theoretical_loss": 3.821684112171837,
      "tokens_seen": 628097024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040891675025075226,
      "loss": 3.0209,
      "theoretical_loss": 3.8216428397985176,
      "tokens_seen": 628162560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089067201604815,
      "loss": 2.6749,
      "theoretical_loss": 3.8216015729364248,
      "tokens_seen": 628228096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088966900702106,
      "loss": 2.847,
      "theoretical_loss": 3.821560311584248,
      "tokens_seen": 628293632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040888665997993985,
      "loss": 2.6515,
      "theoretical_loss": 3.8215190557406773,
      "tokens_seen": 628359168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408876629889669,
      "loss": 2.6926,
      "theoretical_loss": 3.821477805404403,
      "tokens_seen": 628424704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088665997993982,
      "loss": 2.9088,
      "theoretical_loss": 3.8214365605741154,
      "tokens_seen": 628490240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088565697091274,
      "loss": 2.7691,
      "theoretical_loss": 3.8213953212485055,
      "tokens_seen": 628555776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088465396188566,
      "loss": 2.8476,
      "theoretical_loss": 3.821354087426265,
      "tokens_seen": 628621312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040883650952858576,
      "loss": 2.6684,
      "theoretical_loss": 3.821312859106085,
      "tokens_seen": 628686848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408826479438315,
      "loss": 2.7181,
      "theoretical_loss": 3.8212716362866597,
      "tokens_seen": 628752384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088164493480441,
      "loss": 2.5384,
      "theoretical_loss": 3.8212304189666804,
      "tokens_seen": 628817920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040880641925777336,
      "loss": 2.6511,
      "theoretical_loss": 3.8211892071448403,
      "tokens_seen": 628883456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087963891675025,
      "loss": 2.672,
      "theoretical_loss": 3.8211480008198344,
      "tokens_seen": 628948992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087863590772317,
      "loss": 2.6401,
      "theoretical_loss": 3.821106799990356,
      "tokens_seen": 629014528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087763289869609,
      "loss": 2.6822,
      "theoretical_loss": 3.8210656046550997,
      "tokens_seen": 629080064
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 356825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.88836932182312,
      "objective/train/theoretical_loss": 3.821024414812761,
      "objective/train/tokens_used": 649605600,
      "theoretical_loss": 3.821024414812761,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087662988966901,
      "loss": 2.7571,
      "theoretical_loss": 3.821024414812761,
      "tokens_seen": 629145600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040875626880641926,
      "loss": 2.8849,
      "theoretical_loss": 3.8209832304620353,
      "tokens_seen": 629211136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040874623871614844,
      "loss": 2.8463,
      "theoretical_loss": 3.8209420516016177,
      "tokens_seen": 629276672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087362086258776,
      "loss": 2.5645,
      "theoretical_loss": 3.820900878230206,
      "tokens_seen": 629342208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040872617853560686,
      "loss": 2.9021,
      "theoretical_loss": 3.8208597103464963,
      "tokens_seen": 629407744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408716148445336,
      "loss": 2.8184,
      "theoretical_loss": 3.820818547949186,
      "tokens_seen": 629473280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087061183550652,
      "loss": 2.9188,
      "theoretical_loss": 3.820777391036973,
      "tokens_seen": 629538816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040869608826479435,
      "loss": 3.0061,
      "theoretical_loss": 3.8207362396085554,
      "tokens_seen": 629604352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086860581745236,
      "loss": 2.9378,
      "theoretical_loss": 3.820695093662631,
      "tokens_seen": 629669888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040867602808425277,
      "loss": 2.6738,
      "theoretical_loss": 3.8206539531979002,
      "tokens_seen": 629735424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040866599799398195,
      "loss": 2.6967,
      "theoretical_loss": 3.8206128182130623,
      "tokens_seen": 629800960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040865596790371113,
      "loss": 2.6859,
      "theoretical_loss": 3.8205716887068166,
      "tokens_seen": 629866496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040864593781344036,
      "loss": 2.8031,
      "theoretical_loss": 3.8205305646778633,
      "tokens_seen": 629932032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086359077231695,
      "loss": 2.8133,
      "theoretical_loss": 3.820489446124905,
      "tokens_seen": 629997568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086258776328987,
      "loss": 2.8721,
      "theoretical_loss": 3.8204483330466408,
      "tokens_seen": 630063104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040861584754262785,
      "loss": 3.052,
      "theoretical_loss": 3.820407225441774,
      "tokens_seen": 630128640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086058174523571,
      "loss": 3.0175,
      "theoretical_loss": 3.820366123309006,
      "tokens_seen": 630194176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040859578736208627,
      "loss": 2.6538,
      "theoretical_loss": 3.82032502664704,
      "tokens_seen": 630259712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040858575727181545,
      "loss": 2.8017,
      "theoretical_loss": 3.8202839354545786,
      "tokens_seen": 630325248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040857572718154463,
      "loss": 2.8723,
      "theoretical_loss": 3.8202428497303247,
      "tokens_seen": 630390784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085656970912738,
      "loss": 2.8212,
      "theoretical_loss": 3.820201769472984,
      "tokens_seen": 630456320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408555667001003,
      "loss": 2.9297,
      "theoretical_loss": 3.8201606946812596,
      "tokens_seen": 630521856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040854563691073223,
      "loss": 2.9594,
      "theoretical_loss": 3.820119625353856,
      "tokens_seen": 630587392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040853560682046136,
      "loss": 2.7654,
      "theoretical_loss": 3.8200785614894794,
      "tokens_seen": 630652928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085255767301906,
      "loss": 2.7078,
      "theoretical_loss": 3.820037503086836,
      "tokens_seen": 630718464
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 357395,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.229945182800293,
      "objective/train/theoretical_loss": 3.81999645014463,
      "objective/train/tokens_used": 651244000,
      "theoretical_loss": 3.81999645014463,
      "tokens_seen": 630784000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085155466399197,
      "loss": 2.9155,
      "theoretical_loss": 3.81999645014463,
      "tokens_seen": 630784000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040850551654964895,
      "loss": 2.7578,
      "theoretical_loss": 3.81995540266157,
      "tokens_seen": 630849536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040849548645937814,
      "loss": 2.7972,
      "theoretical_loss": 3.819914360636362,
      "tokens_seen": 630915072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084854563691073,
      "loss": 2.9271,
      "theoretical_loss": 3.8198733240677134,
      "tokens_seen": 630980608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084754262788365,
      "loss": 2.8088,
      "theoretical_loss": 3.8198322929543327,
      "tokens_seen": 631046144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040846539618856573,
      "loss": 2.8873,
      "theoretical_loss": 3.819791267294928,
      "tokens_seen": 631111680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040845536609829486,
      "loss": 2.764,
      "theoretical_loss": 3.8197502470882077,
      "tokens_seen": 631177216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084453360080241,
      "loss": 2.6695,
      "theoretical_loss": 3.8197092323328814,
      "tokens_seen": 631242752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084353059177532,
      "loss": 3.0271,
      "theoretical_loss": 3.8196682230276595,
      "tokens_seen": 631308288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040842527582748246,
      "loss": 2.8736,
      "theoretical_loss": 3.8196272191712506,
      "tokens_seen": 631373824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040841524573721164,
      "loss": 2.8502,
      "theoretical_loss": 3.8195862207623668,
      "tokens_seen": 631439360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084052156469408,
      "loss": 2.6863,
      "theoretical_loss": 3.8195452277997184,
      "tokens_seen": 631504896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040839518555667005,
      "loss": 2.6358,
      "theoretical_loss": 3.819504240282016,
      "tokens_seen": 631570432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083851554663992,
      "loss": 2.8913,
      "theoretical_loss": 3.819463258207973,
      "tokens_seen": 631635968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083751253761284,
      "loss": 2.8457,
      "theoretical_loss": 3.8194222815763004,
      "tokens_seen": 631701504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083650952858576,
      "loss": 3.0551,
      "theoretical_loss": 3.8193813103857126,
      "tokens_seen": 631767040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083550651955868,
      "loss": 2.7654,
      "theoretical_loss": 3.8193403446349206,
      "tokens_seen": 631832576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040834503510531596,
      "loss": 2.8818,
      "theoretical_loss": 3.8192993843226395,
      "tokens_seen": 631898112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083350050150452,
      "loss": 2.9639,
      "theoretical_loss": 3.8192584294475833,
      "tokens_seen": 631963648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083249749247743,
      "loss": 2.812,
      "theoretical_loss": 3.8192174800084664,
      "tokens_seen": 632029184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040831494483450356,
      "loss": 2.7267,
      "theoretical_loss": 3.819176536004003,
      "tokens_seen": 632094720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083049147442327,
      "loss": 2.8316,
      "theoretical_loss": 3.8191355974329095,
      "tokens_seen": 632160256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082948846539619,
      "loss": 2.7239,
      "theoretical_loss": 3.8190946642939014,
      "tokens_seen": 632225792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082848545636911,
      "loss": 2.6075,
      "theoretical_loss": 3.819053736585694,
      "tokens_seen": 632291328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082748244734203,
      "loss": 2.7152,
      "theoretical_loss": 3.8190128143070057,
      "tokens_seen": 632356864
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 358555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.62202525138855,
      "objective/train/theoretical_loss": 3.818971897456552,
      "objective/train/tokens_used": 652882400,
      "theoretical_loss": 3.818971897456552,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040826479438314946,
      "loss": 2.9621,
      "theoretical_loss": 3.818971897456552,
      "tokens_seen": 632422400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040825476429287864,
      "loss": 2.7414,
      "theoretical_loss": 3.818930986033052,
      "tokens_seen": 632487936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082447342026078,
      "loss": 2.7112,
      "theoretical_loss": 3.8188900800352226,
      "tokens_seen": 632553472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040823470411233706,
      "loss": 2.9688,
      "theoretical_loss": 3.8188491794617825,
      "tokens_seen": 632619008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082246740220662,
      "loss": 2.9359,
      "theoretical_loss": 3.81880828431145,
      "tokens_seen": 632684544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082146439317954,
      "loss": 2.7537,
      "theoretical_loss": 3.8187673945829457,
      "tokens_seen": 632750080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040820461384152455,
      "loss": 2.583,
      "theoretical_loss": 3.8187265102749883,
      "tokens_seen": 632815616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081945837512538,
      "loss": 2.8925,
      "theoretical_loss": 3.8186856313862982,
      "tokens_seen": 632881152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040818455366098297,
      "loss": 2.487,
      "theoretical_loss": 3.818644757915597,
      "tokens_seen": 632946688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040817452357071215,
      "loss": 2.7202,
      "theoretical_loss": 3.8186038898616035,
      "tokens_seen": 633012224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040816449348044133,
      "loss": 2.8124,
      "theoretical_loss": 3.8185630272230413,
      "tokens_seen": 633077760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040815446339017056,
      "loss": 2.66,
      "theoretical_loss": 3.8185221699986314,
      "tokens_seen": 633143296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081444332998997,
      "loss": 2.8485,
      "theoretical_loss": 3.818481318187096,
      "tokens_seen": 633208832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081344032096289,
      "loss": 2.8472,
      "theoretical_loss": 3.8184404717871585,
      "tokens_seen": 633274368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040812437311935805,
      "loss": 2.766,
      "theoretical_loss": 3.8183996307975416,
      "tokens_seen": 633339904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081143430290873,
      "loss": 2.5663,
      "theoretical_loss": 3.8183587952169686,
      "tokens_seen": 633405440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040810431293881647,
      "loss": 2.7984,
      "theoretical_loss": 3.818317965044165,
      "tokens_seen": 633470976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040809428284854565,
      "loss": 2.9312,
      "theoretical_loss": 3.818277140277854,
      "tokens_seen": 633536512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040808425275827483,
      "loss": 2.6212,
      "theoretical_loss": 3.8182363209167605,
      "tokens_seen": 633602048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408074222668004,
      "loss": 2.8187,
      "theoretical_loss": 3.8181955069596105,
      "tokens_seen": 633667584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080641925777332,
      "loss": 2.7484,
      "theoretical_loss": 3.81815469840513,
      "tokens_seen": 633733120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040805416248746243,
      "loss": 2.8055,
      "theoretical_loss": 3.8181138952520444,
      "tokens_seen": 633798656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040804413239719156,
      "loss": 2.805,
      "theoretical_loss": 3.8180730974990813,
      "tokens_seen": 633864192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080341023069208,
      "loss": 2.8455,
      "theoretical_loss": 3.8180323051449667,
      "tokens_seen": 633929728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080240722166499,
      "loss": 2.824,
      "theoretical_loss": 3.817991518188429,
      "tokens_seen": 633995264
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 359090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7171945571899414,
      "objective/train/theoretical_loss": 3.8179507366281964,
      "objective/train/tokens_used": 654520800,
      "theoretical_loss": 3.8179507366281964,
      "tokens_seen": 634060800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040801404212637915,
      "loss": 2.8011,
      "theoretical_loss": 3.8179507366281964,
      "tokens_seen": 634060800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040800401203610834,
      "loss": 2.6352,
      "theoretical_loss": 3.817909960462996,
      "tokens_seen": 634126336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079939819458375,
      "loss": 2.7008,
      "theoretical_loss": 3.817869189691558,
      "tokens_seen": 634191872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079839518555667,
      "loss": 2.4876,
      "theoretical_loss": 3.8178284243126113,
      "tokens_seen": 634257408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040797392176529593,
      "loss": 2.5169,
      "theoretical_loss": 3.8177876643248845,
      "tokens_seen": 634322944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040796389167502506,
      "loss": 2.5934,
      "theoretical_loss": 3.8177469097271093,
      "tokens_seen": 634388480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079538615847543,
      "loss": 2.5588,
      "theoretical_loss": 3.817706160518015,
      "tokens_seen": 634454016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079438314944834,
      "loss": 2.6794,
      "theoretical_loss": 3.8176654166963337,
      "tokens_seen": 634519552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040793380140421266,
      "loss": 2.5933,
      "theoretical_loss": 3.8176246782607963,
      "tokens_seen": 634585088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040792377131394184,
      "loss": 2.6201,
      "theoretical_loss": 3.817583945210134,
      "tokens_seen": 634650624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407913741223671,
      "loss": 2.9393,
      "theoretical_loss": 3.8175432175430797,
      "tokens_seen": 634716160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079037111334002,
      "loss": 2.8609,
      "theoretical_loss": 3.817502495258366,
      "tokens_seen": 634781696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078936810431294,
      "loss": 2.7891,
      "theoretical_loss": 3.817461778354727,
      "tokens_seen": 634847232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040788365095285856,
      "loss": 2.9239,
      "theoretical_loss": 3.8174210668308945,
      "tokens_seen": 634912768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078736208625878,
      "loss": 2.7264,
      "theoretical_loss": 3.8173803606856036,
      "tokens_seen": 634978304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078635907723169,
      "loss": 2.7941,
      "theoretical_loss": 3.8173396599175886,
      "tokens_seen": 635043840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040785356068204616,
      "loss": 2.8307,
      "theoretical_loss": 3.817298964525584,
      "tokens_seen": 635109376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078435305917753,
      "loss": 2.8384,
      "theoretical_loss": 3.817258274508325,
      "tokens_seen": 635174912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078335005015045,
      "loss": 2.7362,
      "theoretical_loss": 3.8172175898645477,
      "tokens_seen": 635240448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078234704112337,
      "loss": 2.5734,
      "theoretical_loss": 3.817176910592988,
      "tokens_seen": 635305984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078134403209629,
      "loss": 2.732,
      "theoretical_loss": 3.8171362366923827,
      "tokens_seen": 635371520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040780341023069207,
      "loss": 2.6684,
      "theoretical_loss": 3.817095568161468,
      "tokens_seen": 635437056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077933801404213,
      "loss": 2.8199,
      "theoretical_loss": 3.817054904998983,
      "tokens_seen": 635502592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040778335005015043,
      "loss": 2.601,
      "theoretical_loss": 3.817014247203664,
      "tokens_seen": 635568128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040777331995987966,
      "loss": 2.6932,
      "theoretical_loss": 3.8169735947742494,
      "tokens_seen": 635633664
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 360366,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.813488483428955,
      "objective/train/theoretical_loss": 3.8169329477094784,
      "objective/train/tokens_used": 656159200,
      "theoretical_loss": 3.8169329477094784,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077632898696088,
      "loss": 2.7224,
      "theoretical_loss": 3.8169329477094784,
      "tokens_seen": 635699200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407753259779338,
      "loss": 2.7715,
      "theoretical_loss": 3.81689230600809,
      "tokens_seen": 635764736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077432296890672,
      "loss": 2.6614,
      "theoretical_loss": 3.8168516696688233,
      "tokens_seen": 635830272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077331995987964,
      "loss": 2.7141,
      "theoretical_loss": 3.8168110386904193,
      "tokens_seen": 635895808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040772316950852557,
      "loss": 2.8513,
      "theoretical_loss": 3.8167704130716174,
      "tokens_seen": 635961344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040771313941825475,
      "loss": 2.8096,
      "theoretical_loss": 3.8167297928111585,
      "tokens_seen": 636026880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040770310932798393,
      "loss": 2.5866,
      "theoretical_loss": 3.8166891779077847,
      "tokens_seen": 636092416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040769307923771317,
      "loss": 3.1336,
      "theoretical_loss": 3.816648568360236,
      "tokens_seen": 636157952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076830491474423,
      "loss": 2.6937,
      "theoretical_loss": 3.8166079641672566,
      "tokens_seen": 636223488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040767301905717153,
      "loss": 2.7533,
      "theoretical_loss": 3.8165673653275873,
      "tokens_seen": 636289024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076629889669007,
      "loss": 2.8758,
      "theoretical_loss": 3.816526771839972,
      "tokens_seen": 636354560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076529588766299,
      "loss": 2.6713,
      "theoretical_loss": 3.816486183703154,
      "tokens_seen": 636420096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076429287863591,
      "loss": 3.0028,
      "theoretical_loss": 3.8164456009158765,
      "tokens_seen": 636485632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040763289869608825,
      "loss": 2.6476,
      "theoretical_loss": 3.816405023476884,
      "tokens_seen": 636551168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076228686058175,
      "loss": 3.0361,
      "theoretical_loss": 3.8163644513849215,
      "tokens_seen": 636616704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040761283851554667,
      "loss": 2.4477,
      "theoretical_loss": 3.816323884638734,
      "tokens_seen": 636682240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040760280842527585,
      "loss": 2.7836,
      "theoretical_loss": 3.8162833232370668,
      "tokens_seen": 636747776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040759277833500503,
      "loss": 2.7696,
      "theoretical_loss": 3.8162427671786654,
      "tokens_seen": 636813312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075827482447342,
      "loss": 2.6237,
      "theoretical_loss": 3.8162022164622766,
      "tokens_seen": 636878848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075727181544634,
      "loss": 2.7036,
      "theoretical_loss": 3.8161616710866473,
      "tokens_seen": 636944384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040756268806419263,
      "loss": 2.7311,
      "theoretical_loss": 3.8161211310505245,
      "tokens_seen": 637009920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040755265797392176,
      "loss": 2.8929,
      "theoretical_loss": 3.8160805963526556,
      "tokens_seen": 637075456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407542627883651,
      "loss": 2.9721,
      "theoretical_loss": 3.816040066991789,
      "tokens_seen": 637140992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075325977933801,
      "loss": 2.9736,
      "theoretical_loss": 3.815999542966673,
      "tokens_seen": 637206528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040752256770310935,
      "loss": 2.5887,
      "theoretical_loss": 3.8159590242760566,
      "tokens_seen": 637272064
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 360972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.695178270339966,
      "objective/train/theoretical_loss": 3.8159185109186886,
      "objective/train/tokens_used": 657797600,
      "theoretical_loss": 3.8159185109186886,
      "tokens_seen": 637337600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040751253761283854,
      "loss": 2.7188,
      "theoretical_loss": 3.8159185109186886,
      "tokens_seen": 637337600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075025075225677,
      "loss": 2.8754,
      "theoretical_loss": 3.815878002893319,
      "tokens_seen": 637403136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074924774322969,
      "loss": 2.8012,
      "theoretical_loss": 3.8158375001986986,
      "tokens_seen": 637468672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040748244734202613,
      "loss": 2.7555,
      "theoretical_loss": 3.815797002833577,
      "tokens_seen": 637534208
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040747241725175526,
      "loss": 2.6793,
      "theoretical_loss": 3.815756510796705,
      "tokens_seen": 637599744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074623871614845,
      "loss": 2.8207,
      "theoretical_loss": 3.8157160240868353,
      "tokens_seen": 637665280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074523570712136,
      "loss": 2.9315,
      "theoretical_loss": 3.815675542702719,
      "tokens_seen": 637730816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040744232698094286,
      "loss": 2.6283,
      "theoretical_loss": 3.8156350666431074,
      "tokens_seen": 637796352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040743229689067204,
      "loss": 2.8794,
      "theoretical_loss": 3.815594595906755,
      "tokens_seen": 637861888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074222668004012,
      "loss": 2.7645,
      "theoretical_loss": 3.815554130492414,
      "tokens_seen": 637927424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074122367101304,
      "loss": 2.61,
      "theoretical_loss": 3.815513670398837,
      "tokens_seen": 637992960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074022066198596,
      "loss": 2.7538,
      "theoretical_loss": 3.8154732156247793,
      "tokens_seen": 638058496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040739217652958876,
      "loss": 2.7017,
      "theoretical_loss": 3.8154327661689944,
      "tokens_seen": 638124032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407382146439318,
      "loss": 2.7127,
      "theoretical_loss": 3.815392322030238,
      "tokens_seen": 638189568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073721163490471,
      "loss": 2.6691,
      "theoretical_loss": 3.8153518832072644,
      "tokens_seen": 638255104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040736208625877636,
      "loss": 2.5609,
      "theoretical_loss": 3.8153114496988296,
      "tokens_seen": 638320640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073520561685055,
      "loss": 2.9655,
      "theoretical_loss": 3.8152710215036896,
      "tokens_seen": 638386176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073420260782347,
      "loss": 2.825,
      "theoretical_loss": 3.8152305986206,
      "tokens_seen": 638451712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073319959879639,
      "loss": 2.7397,
      "theoretical_loss": 3.815190181048319,
      "tokens_seen": 638517248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073219658976931,
      "loss": 2.8857,
      "theoretical_loss": 3.815149768785603,
      "tokens_seen": 638582784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040731193580742227,
      "loss": 2.8068,
      "theoretical_loss": 3.81510936183121,
      "tokens_seen": 638648320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073019057171515,
      "loss": 2.6231,
      "theoretical_loss": 3.8150689601838983,
      "tokens_seen": 638713856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040729187562688063,
      "loss": 2.9175,
      "theoretical_loss": 3.815028563842426,
      "tokens_seen": 638779392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040728184553660986,
      "loss": 3.0054,
      "theoretical_loss": 3.8149881728055526,
      "tokens_seen": 638844928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407271815446339,
      "loss": 2.8376,
      "theoretical_loss": 3.814947787072037,
      "tokens_seen": 638910464
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 361973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6586904525756836,
      "objective/train/theoretical_loss": 3.814907406640639,
      "objective/train/tokens_used": 659436000,
      "theoretical_loss": 3.814907406640639,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072617853560682,
      "loss": 2.7195,
      "theoretical_loss": 3.814907406640639,
      "tokens_seen": 638976000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072517552657974,
      "loss": 2.8213,
      "theoretical_loss": 3.8148670315101185,
      "tokens_seen": 639041536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072417251755266,
      "loss": 2.882,
      "theoretical_loss": 3.8148266616792377,
      "tokens_seen": 639107072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040723169508525577,
      "loss": 2.8047,
      "theoretical_loss": 3.814786297146756,
      "tokens_seen": 639172608
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040722166499498495,
      "loss": 2.7604,
      "theoretical_loss": 3.8147459379114346,
      "tokens_seen": 639238144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040721163490471413,
      "loss": 2.8299,
      "theoretical_loss": 3.8147055839720365,
      "tokens_seen": 639303680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040720160481444337,
      "loss": 2.827,
      "theoretical_loss": 3.8146652353273236,
      "tokens_seen": 639369216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071915747241725,
      "loss": 3.0141,
      "theoretical_loss": 3.8146248919760586,
      "tokens_seen": 639434752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040718154463390173,
      "loss": 2.6655,
      "theoretical_loss": 3.8145845539170047,
      "tokens_seen": 639500288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071715145436309,
      "loss": 2.9756,
      "theoretical_loss": 3.8145442211489256,
      "tokens_seen": 639565824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071614844533601,
      "loss": 2.6459,
      "theoretical_loss": 3.8145038936705844,
      "tokens_seen": 639631360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040715145436308927,
      "loss": 2.7277,
      "theoretical_loss": 3.814463571480746,
      "tokens_seen": 639696896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040714142427281845,
      "loss": 2.6967,
      "theoretical_loss": 3.814423254578176,
      "tokens_seen": 639762432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040713139418254763,
      "loss": 2.9031,
      "theoretical_loss": 3.8143829429616383,
      "tokens_seen": 639827968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040712136409227687,
      "loss": 2.8567,
      "theoretical_loss": 3.814342636629899,
      "tokens_seen": 639893504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407111334002006,
      "loss": 2.6368,
      "theoretical_loss": 3.814302335581724,
      "tokens_seen": 639959040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040710130391173523,
      "loss": 2.7911,
      "theoretical_loss": 3.81426203981588,
      "tokens_seen": 640024576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040709127382146436,
      "loss": 2.9786,
      "theoretical_loss": 3.814221749331134,
      "tokens_seen": 640090112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070812437311936,
      "loss": 2.7101,
      "theoretical_loss": 3.8141814641262526,
      "tokens_seen": 640155648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070712136409228,
      "loss": 2.7521,
      "theoretical_loss": 3.814141184200004,
      "tokens_seen": 640221184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040706118355065196,
      "loss": 2.8041,
      "theoretical_loss": 3.814100909551156,
      "tokens_seen": 640286720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040705115346038114,
      "loss": 2.8164,
      "theoretical_loss": 3.814060640178478,
      "tokens_seen": 640352256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070411233701103,
      "loss": 2.9017,
      "theoretical_loss": 3.8140203760807374,
      "tokens_seen": 640417792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070310932798395,
      "loss": 2.7351,
      "theoretical_loss": 3.8139801172567047,
      "tokens_seen": 640483328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040702106318956874,
      "loss": 2.7543,
      "theoretical_loss": 3.8139398637051496,
      "tokens_seen": 640548864
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 362406,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6734325885772705,
      "objective/train/theoretical_loss": 3.813899615424841,
      "objective/train/tokens_used": 661074400,
      "theoretical_loss": 3.813899615424841,
      "tokens_seen": 640614400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040701103309929786,
      "loss": 2.9096,
      "theoretical_loss": 3.813899615424841,
      "tokens_seen": 640614400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070010030090271,
      "loss": 2.8807,
      "theoretical_loss": 3.8138593724145515,
      "tokens_seen": 640679936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069909729187563,
      "loss": 2.6374,
      "theoretical_loss": 3.81381913467305,
      "tokens_seen": 640745472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040698094282848546,
      "loss": 2.6867,
      "theoretical_loss": 3.8137789021991093,
      "tokens_seen": 640811008
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040697091273821464,
      "loss": 2.7622,
      "theoretical_loss": 3.813738674991501,
      "tokens_seen": 640876544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069608826479438,
      "loss": 2.7671,
      "theoretical_loss": 3.813698453048997,
      "tokens_seen": 640942080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406950852557673,
      "loss": 2.7282,
      "theoretical_loss": 3.81365823637037,
      "tokens_seen": 641007616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040694082246740224,
      "loss": 2.8956,
      "theoretical_loss": 3.813618024954393,
      "tokens_seen": 641073152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040693079237713137,
      "loss": 2.8595,
      "theoretical_loss": 3.81357781879984,
      "tokens_seen": 641138688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069207622868606,
      "loss": 2.797,
      "theoretical_loss": 3.8135376179054843,
      "tokens_seen": 641204224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040691073219658973,
      "loss": 2.9285,
      "theoretical_loss": 3.8134974222700997,
      "tokens_seen": 641269760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040690070210631896,
      "loss": 2.9507,
      "theoretical_loss": 3.813457231892462,
      "tokens_seen": 641335296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068906720160482,
      "loss": 2.9627,
      "theoretical_loss": 3.8134170467713453,
      "tokens_seen": 641400832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068806419257773,
      "loss": 2.7184,
      "theoretical_loss": 3.813376866905526,
      "tokens_seen": 641466368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040687061183550656,
      "loss": 2.9675,
      "theoretical_loss": 3.81333669229378,
      "tokens_seen": 641531904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068605817452357,
      "loss": 2.7346,
      "theoretical_loss": 3.813296522934883,
      "tokens_seen": 641597440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068505516549649,
      "loss": 2.7329,
      "theoretical_loss": 3.813256358827611,
      "tokens_seen": 641662976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068405215646941,
      "loss": 2.6317,
      "theoretical_loss": 3.8132161999707437,
      "tokens_seen": 641728512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068304914744233,
      "loss": 2.8005,
      "theoretical_loss": 3.813176046363056,
      "tokens_seen": 641794048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040682046138415247,
      "loss": 2.8122,
      "theoretical_loss": 3.8131358980033276,
      "tokens_seen": 641859584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068104312938817,
      "loss": 2.8557,
      "theoretical_loss": 3.8130957548903357,
      "tokens_seen": 641925120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040680040120361083,
      "loss": 2.8691,
      "theoretical_loss": 3.81305561702286,
      "tokens_seen": 641990656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040679037111334006,
      "loss": 2.6403,
      "theoretical_loss": 3.8130154843996795,
      "tokens_seen": 642056192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067803410230692,
      "loss": 2.8516,
      "theoretical_loss": 3.812975357019573,
      "tokens_seen": 642121728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067703109327984,
      "loss": 3.0307,
      "theoretical_loss": 3.812935234881322,
      "tokens_seen": 642187264
    },
    {
      "epoch": 0.19,
      "objective/train/docs_used": 363472,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8334643840789795,
      "objective/train/theoretical_loss": 3.8128951179837056,
      "objective/train/tokens_used": 662712800,
      "theoretical_loss": 3.8128951179837056,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067602808425276,
      "loss": 2.7628,
      "theoretical_loss": 3.8128951179837056,
      "tokens_seen": 642252800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067502507522568,
      "loss": 2.6521,
      "theoretical_loss": 3.812855006325505,
      "tokens_seen": 642318336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040674022066198597,
      "loss": 2.7719,
      "theoretical_loss": 3.8128148999055016,
      "tokens_seen": 642383872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040673019057171515,
      "loss": 2.9862,
      "theoretical_loss": 3.8127747987224767,
      "tokens_seen": 642449408
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040672016048144433,
      "loss": 2.7238,
      "theoretical_loss": 3.812734702775213,
      "tokens_seen": 642514944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040671013039117357,
      "loss": 2.7985,
      "theoretical_loss": 3.8126946120624927,
      "tokens_seen": 642580480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067001003009027,
      "loss": 2.9992,
      "theoretical_loss": 3.8126545265830982,
      "tokens_seen": 642646016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040669007021063193,
      "loss": 2.7378,
      "theoretical_loss": 3.8126144463358136,
      "tokens_seen": 642711552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066800401203611,
      "loss": 2.7071,
      "theoretical_loss": 3.812574371319422,
      "tokens_seen": 642777088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066700100300903,
      "loss": 2.7973,
      "theoretical_loss": 3.812534301532707,
      "tokens_seen": 642842624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040665997993981947,
      "loss": 2.8874,
      "theoretical_loss": 3.8124942369744543,
      "tokens_seen": 642908160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040664994984954865,
      "loss": 2.7893,
      "theoretical_loss": 3.812454177643448,
      "tokens_seen": 642973696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040663991975927783,
      "loss": 2.7979,
      "theoretical_loss": 3.812414123538474,
      "tokens_seen": 643039232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040662988966900707,
      "loss": 2.8078,
      "theoretical_loss": 3.812374074658317,
      "tokens_seen": 643104768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066198595787362,
      "loss": 2.8899,
      "theoretical_loss": 3.812334031001764,
      "tokens_seen": 643170304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040660982948846543,
      "loss": 2.6655,
      "theoretical_loss": 3.812293992567602,
      "tokens_seen": 643235840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040659979939819456,
      "loss": 2.7146,
      "theoretical_loss": 3.8122539593546163,
      "tokens_seen": 643301376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004065897693079238,
      "loss": 2.9341,
      "theoretical_loss": 3.8122139313615953,
      "tokens_seen": 643366912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406579739217653,
      "loss": 2.856,
      "theoretical_loss": 3.8121739085873267,
      "tokens_seen": 643432448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040656970912738216,
      "loss": 2.7503,
      "theoretical_loss": 3.812133891030598,
      "tokens_seen": 643497984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040655967903711134,
      "loss": 2.8333,
      "theoretical_loss": 3.812093878690199,
      "tokens_seen": 643563520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065496489468405,
      "loss": 2.6505,
      "theoretical_loss": 3.8120538715649177,
      "tokens_seen": 643629056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065396188565697,
      "loss": 2.6991,
      "theoretical_loss": 3.8120138696535437,
      "tokens_seen": 643694592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040652958876629894,
      "loss": 2.8435,
      "theoretical_loss": 3.8119738729548667,
      "tokens_seen": 643760128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040651955867602806,
      "loss": 2.9121,
      "theoretical_loss": 3.811933881467677,
      "tokens_seen": 643825664
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 364407,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.769031047821045,
      "objective/train/theoretical_loss": 3.8118938951907646,
      "objective/train/tokens_used": 664351200,
      "theoretical_loss": 3.8118938951907646,
      "tokens_seen": 643891200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065095285857573,
      "loss": 2.9188,
      "theoretical_loss": 3.8118938951907646,
      "tokens_seen": 643891200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064994984954865,
      "loss": 3.0245,
      "theoretical_loss": 3.8118539141229215,
      "tokens_seen": 643956736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040648946840521566,
      "loss": 2.8493,
      "theoretical_loss": 3.811813938262939,
      "tokens_seen": 644022272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040647943831494484,
      "loss": 2.7306,
      "theoretical_loss": 3.811773967609608,
      "tokens_seen": 644087808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406469408224674,
      "loss": 2.5553,
      "theoretical_loss": 3.811734002161721,
      "tokens_seen": 644153344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064593781344032,
      "loss": 2.7385,
      "theoretical_loss": 3.8116940419180705,
      "tokens_seen": 644218880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040644934804413244,
      "loss": 2.7768,
      "theoretical_loss": 3.81165408687745,
      "tokens_seen": 644284416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040643931795386157,
      "loss": 2.9223,
      "theoretical_loss": 3.8116141370386525,
      "tokens_seen": 644349952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064292878635908,
      "loss": 2.5762,
      "theoretical_loss": 3.811574192400472,
      "tokens_seen": 644415488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040641925777331993,
      "loss": 2.6792,
      "theoretical_loss": 3.8115342529617022,
      "tokens_seen": 644481024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040640922768304916,
      "loss": 2.7865,
      "theoretical_loss": 3.811494318721139,
      "tokens_seen": 644546560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040639919759277834,
      "loss": 2.9013,
      "theoretical_loss": 3.811454389677576,
      "tokens_seen": 644612096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063891675025075,
      "loss": 2.5841,
      "theoretical_loss": 3.811414465829809,
      "tokens_seen": 644677632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063791374122367,
      "loss": 2.7939,
      "theoretical_loss": 3.8113745471766336,
      "tokens_seen": 644743168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063691073219659,
      "loss": 3.0081,
      "theoretical_loss": 3.811334633716847,
      "tokens_seen": 644808704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040635907723169507,
      "loss": 2.5528,
      "theoretical_loss": 3.811294725449245,
      "tokens_seen": 644874240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063490471414243,
      "loss": 2.7421,
      "theoretical_loss": 3.811254822372624,
      "tokens_seen": 644939776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040633901705115343,
      "loss": 2.7564,
      "theoretical_loss": 3.811214924485783,
      "tokens_seen": 645005312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040632898696088267,
      "loss": 2.8748,
      "theoretical_loss": 3.811175031787519,
      "tokens_seen": 645070848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040631895687061185,
      "loss": 2.7806,
      "theoretical_loss": 3.81113514427663,
      "tokens_seen": 645136384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040630892678034103,
      "loss": 2.9969,
      "theoretical_loss": 3.8110952619519143,
      "tokens_seen": 645201920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062988966900702,
      "loss": 2.7653,
      "theoretical_loss": 3.811055384812172,
      "tokens_seen": 645267456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062888665997994,
      "loss": 2.8918,
      "theoretical_loss": 3.8110155128562018,
      "tokens_seen": 645332992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040627883650952857,
      "loss": 2.7826,
      "theoretical_loss": 3.8109756460828037,
      "tokens_seen": 645398528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062688064192578,
      "loss": 2.9599,
      "theoretical_loss": 3.810935784490778,
      "tokens_seen": 645464064
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 365032,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9448306560516357,
      "objective/train/theoretical_loss": 3.8108959280789243,
      "objective/train/tokens_used": 665989600,
      "theoretical_loss": 3.8108959280789243,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040625877632898693,
      "loss": 2.8312,
      "theoretical_loss": 3.8108959280789243,
      "tokens_seen": 645529600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040624874623871617,
      "loss": 2.8855,
      "theoretical_loss": 3.8108560768460453,
      "tokens_seen": 645595136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062387161484453,
      "loss": 2.8467,
      "theoretical_loss": 3.810816230790942,
      "tokens_seen": 645660672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040622868605817453,
      "loss": 2.6674,
      "theoretical_loss": 3.8107763899124145,
      "tokens_seen": 645726208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062186559679037,
      "loss": 2.8258,
      "theoretical_loss": 3.8107365542092673,
      "tokens_seen": 645791744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062086258776329,
      "loss": 2.9253,
      "theoretical_loss": 3.8106967236803015,
      "tokens_seen": 645857280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061985957873621,
      "loss": 2.7386,
      "theoretical_loss": 3.8106568983243205,
      "tokens_seen": 645922816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040618856569709126,
      "loss": 2.7275,
      "theoretical_loss": 3.8106170781401287,
      "tokens_seen": 645988352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040617853560682044,
      "loss": 2.8876,
      "theoretical_loss": 3.8105772631265284,
      "tokens_seen": 646053888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040616850551654967,
      "loss": 2.9051,
      "theoretical_loss": 3.8105374532823246,
      "tokens_seen": 646119424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061584754262788,
      "loss": 2.9001,
      "theoretical_loss": 3.8104976486063213,
      "tokens_seen": 646184960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040614844533600804,
      "loss": 2.9,
      "theoretical_loss": 3.8104578490973244,
      "tokens_seen": 646250496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040613841524573727,
      "loss": 2.8382,
      "theoretical_loss": 3.810418054754139,
      "tokens_seen": 646316032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061283851554664,
      "loss": 2.8047,
      "theoretical_loss": 3.81037826557557,
      "tokens_seen": 646381568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040611835506519563,
      "loss": 2.7919,
      "theoretical_loss": 3.810338481560425,
      "tokens_seen": 646447104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040610832497492476,
      "loss": 2.9111,
      "theoretical_loss": 3.8102987027075095,
      "tokens_seen": 646512640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406098294884654,
      "loss": 2.8586,
      "theoretical_loss": 3.810258929015631,
      "tokens_seen": 646578176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060882647943832,
      "loss": 2.678,
      "theoretical_loss": 3.8102191604835967,
      "tokens_seen": 646643712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040607823470411236,
      "loss": 2.781,
      "theoretical_loss": 3.8101793971102147,
      "tokens_seen": 646709248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040606820461384154,
      "loss": 2.914,
      "theoretical_loss": 3.8101396388942925,
      "tokens_seen": 646774784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060581745235707,
      "loss": 2.7807,
      "theoretical_loss": 3.810099885834639,
      "tokens_seen": 646840320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060481444332999,
      "loss": 2.8564,
      "theoretical_loss": 3.810060137930064,
      "tokens_seen": 646905856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040603811434302914,
      "loss": 2.7308,
      "theoretical_loss": 3.810020395179375,
      "tokens_seen": 646971392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040602808425275826,
      "loss": 2.898,
      "theoretical_loss": 3.8099806575813835,
      "tokens_seen": 647036928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060180541624875,
      "loss": 2.8808,
      "theoretical_loss": 3.809940925134899,
      "tokens_seen": 647102464
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 365530,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.693211317062378,
      "objective/train/theoretical_loss": 3.809901197838732,
      "objective/train/tokens_used": 667628000,
      "theoretical_loss": 3.809901197838732,
      "tokens_seen": 647168000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060080240722167,
      "loss": 2.7837,
      "theoretical_loss": 3.809901197838732,
      "tokens_seen": 647168000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040599799398194586,
      "loss": 3.0024,
      "theoretical_loss": 3.8098614756916933,
      "tokens_seen": 647233536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040598796389167504,
      "loss": 2.8533,
      "theoretical_loss": 3.8098217586925944,
      "tokens_seen": 647299072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059779338014042,
      "loss": 2.7603,
      "theoretical_loss": 3.8097820468402475,
      "tokens_seen": 647364608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059679037111334,
      "loss": 2.6107,
      "theoretical_loss": 3.8097423401334636,
      "tokens_seen": 647430144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040595787362086264,
      "loss": 2.7351,
      "theoretical_loss": 3.809702638571056,
      "tokens_seen": 647495680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040594784353059177,
      "loss": 2.6653,
      "theoretical_loss": 3.8096629421518378,
      "tokens_seen": 647561216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405937813440321,
      "loss": 3.0185,
      "theoretical_loss": 3.809623250874622,
      "tokens_seen": 647626752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040592778335005013,
      "loss": 2.7992,
      "theoretical_loss": 3.8095835647382223,
      "tokens_seen": 647692288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040591775325977936,
      "loss": 2.6663,
      "theoretical_loss": 3.809543883741453,
      "tokens_seen": 647757824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040590772316950854,
      "loss": 2.6354,
      "theoretical_loss": 3.809504207883128,
      "tokens_seen": 647823360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058976930792377,
      "loss": 2.7836,
      "theoretical_loss": 3.8094645371620626,
      "tokens_seen": 647888896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058876629889669,
      "loss": 2.7201,
      "theoretical_loss": 3.809424871577072,
      "tokens_seen": 647954432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058776328986961,
      "loss": 2.7792,
      "theoretical_loss": 3.8093852111269726,
      "tokens_seen": 648019968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040586760280842527,
      "loss": 2.7333,
      "theoretical_loss": 3.809345555810579,
      "tokens_seen": 648085504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058575727181545,
      "loss": 2.7829,
      "theoretical_loss": 3.8093059056267085,
      "tokens_seen": 648151040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040584754262788363,
      "loss": 2.8198,
      "theoretical_loss": 3.8092662605741783,
      "tokens_seen": 648216576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040583751253761287,
      "loss": 2.7442,
      "theoretical_loss": 3.8092266206518044,
      "tokens_seen": 648282112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040582748244734205,
      "loss": 2.7689,
      "theoretical_loss": 3.809186985858406,
      "tokens_seen": 648347648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040581745235707123,
      "loss": 2.9267,
      "theoretical_loss": 3.8091473561928,
      "tokens_seen": 648413184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058074222668004,
      "loss": 2.9537,
      "theoretical_loss": 3.8091077316538056,
      "tokens_seen": 648478720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057973921765296,
      "loss": 2.9389,
      "theoretical_loss": 3.809068112240241,
      "tokens_seen": 648544256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040578736208625877,
      "loss": 2.7254,
      "theoretical_loss": 3.8090284979509255,
      "tokens_seen": 648609792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405777331995988,
      "loss": 2.8984,
      "theoretical_loss": 3.8089888887846786,
      "tokens_seen": 648675328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040576730190571713,
      "loss": 2.8705,
      "theoretical_loss": 3.808949284740321,
      "tokens_seen": 648740864
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 366782,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.094092845916748,
      "objective/train/theoretical_loss": 3.8089096858166718,
      "objective/train/tokens_used": 669266400,
      "theoretical_loss": 3.8089096858166718,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040575727181544637,
      "loss": 2.7799,
      "theoretical_loss": 3.8089096858166718,
      "tokens_seen": 648806400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057472417251755,
      "loss": 2.9153,
      "theoretical_loss": 3.808870092012553,
      "tokens_seen": 648871936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040573721163490473,
      "loss": 2.8236,
      "theoretical_loss": 3.808830503326785,
      "tokens_seen": 648937472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057271815446339,
      "loss": 2.5253,
      "theoretical_loss": 3.8087909197581897,
      "tokens_seen": 649003008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057171514543631,
      "loss": 3.0232,
      "theoretical_loss": 3.8087513413055887,
      "tokens_seen": 649068544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057071213640923,
      "loss": 2.9203,
      "theoretical_loss": 3.8087117679678046,
      "tokens_seen": 649134080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040569709127382146,
      "loss": 2.7979,
      "theoretical_loss": 3.8086721997436603,
      "tokens_seen": 649199616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040568706118355064,
      "loss": 2.9185,
      "theoretical_loss": 3.8086326366319785,
      "tokens_seen": 649265152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056770310932799,
      "loss": 2.5927,
      "theoretical_loss": 3.808593078631583,
      "tokens_seen": 649330688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405667001003009,
      "loss": 2.7122,
      "theoretical_loss": 3.808553525741297,
      "tokens_seen": 649396224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040565697091273824,
      "loss": 2.9182,
      "theoretical_loss": 3.808513977959946,
      "tokens_seen": 649461760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056469408224674,
      "loss": 2.899,
      "theoretical_loss": 3.8084744352863535,
      "tokens_seen": 649527296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056369107321966,
      "loss": 2.7249,
      "theoretical_loss": 3.808434897719345,
      "tokens_seen": 649592832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056268806419258,
      "loss": 2.9955,
      "theoretical_loss": 3.8083953652577462,
      "tokens_seen": 649658368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040561685055165496,
      "loss": 2.591,
      "theoretical_loss": 3.8083558379003826,
      "tokens_seen": 649723904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040560682046138414,
      "loss": 2.8997,
      "theoretical_loss": 3.8083163156460804,
      "tokens_seen": 649789440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055967903711134,
      "loss": 2.8928,
      "theoretical_loss": 3.808276798493666,
      "tokens_seen": 649854976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055867602808425,
      "loss": 2.8441,
      "theoretical_loss": 3.808237286441967,
      "tokens_seen": 649920512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040557673019057174,
      "loss": 2.9768,
      "theoretical_loss": 3.8081977794898103,
      "tokens_seen": 649986048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040556670010030087,
      "loss": 2.5429,
      "theoretical_loss": 3.808158277636024,
      "tokens_seen": 650051584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055566700100301,
      "loss": 2.8073,
      "theoretical_loss": 3.808118780879436,
      "tokens_seen": 650117120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055466399197593,
      "loss": 2.8952,
      "theoretical_loss": 3.8080792892188753,
      "tokens_seen": 650182656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040553660982948846,
      "loss": 2.7382,
      "theoretical_loss": 3.8080398026531697,
      "tokens_seen": 650248192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040552657973921764,
      "loss": 2.7826,
      "theoretical_loss": 3.80800032118115,
      "tokens_seen": 650313728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055165496489469,
      "loss": 2.7656,
      "theoretical_loss": 3.807960844801644,
      "tokens_seen": 650379264
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 367488,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.001251459121704,
      "objective/train/theoretical_loss": 3.8079213735134845,
      "objective/train/tokens_used": 670904800,
      "theoretical_loss": 3.8079213735134845,
      "tokens_seen": 650444800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405506519558676,
      "loss": 2.835,
      "theoretical_loss": 3.8079213735134845,
      "tokens_seen": 650444800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040549648946840524,
      "loss": 2.6539,
      "theoretical_loss": 3.8078819073154997,
      "tokens_seen": 650510336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040548645937813437,
      "loss": 2.7046,
      "theoretical_loss": 3.807842446206521,
      "tokens_seen": 650575872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054764292878636,
      "loss": 2.928,
      "theoretical_loss": 3.8078029901853805,
      "tokens_seen": 650641408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054663991975928,
      "loss": 2.6514,
      "theoretical_loss": 3.8077635392509084,
      "tokens_seen": 650706944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040545636910732197,
      "loss": 3.1162,
      "theoretical_loss": 3.807724093401938,
      "tokens_seen": 650772480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040544633901705115,
      "loss": 2.582,
      "theoretical_loss": 3.8076846526373007,
      "tokens_seen": 650838016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040543630892678033,
      "loss": 2.9161,
      "theoretical_loss": 3.8076452169558306,
      "tokens_seen": 650903552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054262788365095,
      "loss": 2.8318,
      "theoretical_loss": 3.8076057863563593,
      "tokens_seen": 650969088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040541624874623874,
      "loss": 2.7556,
      "theoretical_loss": 3.807566360837722,
      "tokens_seen": 651034624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040540621865596787,
      "loss": 2.7484,
      "theoretical_loss": 3.807526940398751,
      "tokens_seen": 651100160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053961885656971,
      "loss": 2.7967,
      "theoretical_loss": 3.807487525038282,
      "tokens_seen": 651165696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053861584754263,
      "loss": 2.9202,
      "theoretical_loss": 3.8074481147551484,
      "tokens_seen": 651231232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040537612838515547,
      "loss": 2.7285,
      "theoretical_loss": 3.807408709548187,
      "tokens_seen": 651296768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053660982948847,
      "loss": 2.8267,
      "theoretical_loss": 3.8073693094162318,
      "tokens_seen": 651362304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040535606820461383,
      "loss": 2.7598,
      "theoretical_loss": 3.807329914358119,
      "tokens_seen": 651427840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040534603811434307,
      "loss": 2.7406,
      "theoretical_loss": 3.807290524372686,
      "tokens_seen": 651493376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040533600802407225,
      "loss": 2.5667,
      "theoretical_loss": 3.807251139458767,
      "tokens_seen": 651558912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040532597793380143,
      "loss": 2.6472,
      "theoretical_loss": 3.8072117596152015,
      "tokens_seen": 651624448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053159478435306,
      "loss": 2.6189,
      "theoretical_loss": 3.807172384840826,
      "tokens_seen": 651689984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053059177532598,
      "loss": 2.5579,
      "theoretical_loss": 3.807133015134478,
      "tokens_seen": 651755520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040529588766298897,
      "loss": 2.6739,
      "theoretical_loss": 3.807093650494996,
      "tokens_seen": 651821056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052858575727182,
      "loss": 2.8714,
      "theoretical_loss": 3.8070542909212186,
      "tokens_seen": 651886592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040527582748244733,
      "loss": 2.9979,
      "theoretical_loss": 3.807014936411984,
      "tokens_seen": 651952128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040526579739217657,
      "loss": 2.6012,
      "theoretical_loss": 3.806975586966133,
      "tokens_seen": 652017664
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 368763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9794094562530518,
      "objective/train/theoretical_loss": 3.8069362425825037,
      "objective/train/tokens_used": 672543200,
      "theoretical_loss": 3.8069362425825037,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052557673019057,
      "loss": 2.7193,
      "theoretical_loss": 3.8069362425825037,
      "tokens_seen": 652083200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040524573721163493,
      "loss": 2.9792,
      "theoretical_loss": 3.806896903259937,
      "tokens_seen": 652148736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052357071213641,
      "loss": 2.8212,
      "theoretical_loss": 3.8068575689972737,
      "tokens_seen": 652214272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052256770310933,
      "loss": 2.8342,
      "theoretical_loss": 3.806818239793354,
      "tokens_seen": 652279808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052156469408225,
      "loss": 2.4792,
      "theoretical_loss": 3.8067789156470195,
      "tokens_seen": 652345344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040520561685055166,
      "loss": 3.0418,
      "theoretical_loss": 3.8067395965571116,
      "tokens_seen": 652410880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040519558676028084,
      "loss": 2.7229,
      "theoretical_loss": 3.8067002825224723,
      "tokens_seen": 652476416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051855566700101,
      "loss": 2.9192,
      "theoretical_loss": 3.806660973541944,
      "tokens_seen": 652541952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051755265797392,
      "loss": 2.7193,
      "theoretical_loss": 3.80662166961437,
      "tokens_seen": 652607488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040516549648946844,
      "loss": 2.719,
      "theoretical_loss": 3.8065823707385924,
      "tokens_seen": 652673024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051554663991976,
      "loss": 2.6547,
      "theoretical_loss": 3.8065430769134556,
      "tokens_seen": 652738560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051454363089268,
      "loss": 2.7879,
      "theoretical_loss": 3.8065037881378028,
      "tokens_seen": 652804096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405135406218656,
      "loss": 2.9661,
      "theoretical_loss": 3.8064645044104792,
      "tokens_seen": 652869632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040512537612838516,
      "loss": 2.6646,
      "theoretical_loss": 3.8064252257303286,
      "tokens_seen": 652935168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040511534603811434,
      "loss": 2.6163,
      "theoretical_loss": 3.8063859520961967,
      "tokens_seen": 653000704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051053159478436,
      "loss": 2.9179,
      "theoretical_loss": 3.8063466835069284,
      "tokens_seen": 653066240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050952858575727,
      "loss": 2.8113,
      "theoretical_loss": 3.8063074199613696,
      "tokens_seen": 653131776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040508525576730194,
      "loss": 2.7895,
      "theoretical_loss": 3.806268161458367,
      "tokens_seen": 653197312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040507522567703107,
      "loss": 2.6098,
      "theoretical_loss": 3.8062289079967666,
      "tokens_seen": 653262848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050651955867603,
      "loss": 2.6069,
      "theoretical_loss": 3.8061896595754154,
      "tokens_seen": 653328384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050551654964895,
      "loss": 2.7038,
      "theoretical_loss": 3.8061504161931605,
      "tokens_seen": 653393920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040504513540621866,
      "loss": 2.5877,
      "theoretical_loss": 3.8061111778488503,
      "tokens_seen": 653459456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040503510531594784,
      "loss": 2.7855,
      "theoretical_loss": 3.8060719445413325,
      "tokens_seen": 653524992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050250752256771,
      "loss": 2.6148,
      "theoretical_loss": 3.8060327162694563,
      "tokens_seen": 653590528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050150451354062,
      "loss": 2.8388,
      "theoretical_loss": 3.8059934930320685,
      "tokens_seen": 653656064
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 369902,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6009209156036377,
      "objective/train/theoretical_loss": 3.8059542748280206,
      "objective/train/tokens_used": 674181600,
      "theoretical_loss": 3.8059542748280206,
      "tokens_seen": 653721600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040500501504513544,
      "loss": 2.5162,
      "theoretical_loss": 3.8059542748280206,
      "tokens_seen": 653721600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040499498495486457,
      "loss": 2.681,
      "theoretical_loss": 3.805915061656161,
      "tokens_seen": 653787136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049849548645938,
      "loss": 2.7516,
      "theoretical_loss": 3.8058758535153396,
      "tokens_seen": 653852672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404974924774323,
      "loss": 2.6446,
      "theoretical_loss": 3.805836650404407,
      "tokens_seen": 653918208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040496489468405217,
      "loss": 2.659,
      "theoretical_loss": 3.8057974523222144,
      "tokens_seen": 653983744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040495486459378135,
      "loss": 2.7514,
      "theoretical_loss": 3.8057582592676127,
      "tokens_seen": 654049280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040494483450351053,
      "loss": 2.8106,
      "theoretical_loss": 3.8057190712394524,
      "tokens_seen": 654114816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049348044132397,
      "loss": 2.5477,
      "theoretical_loss": 3.805679888236587,
      "tokens_seen": 654180352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040492477432296894,
      "loss": 2.889,
      "theoretical_loss": 3.805640710257867,
      "tokens_seen": 654245888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040491474423269807,
      "loss": 2.886,
      "theoretical_loss": 3.805601537302147,
      "tokens_seen": 654311424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049047141424273,
      "loss": 2.6806,
      "theoretical_loss": 3.8055623693682783,
      "tokens_seen": 654376960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040489468405215643,
      "loss": 2.8927,
      "theoretical_loss": 3.8055232064551148,
      "tokens_seen": 654442496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040488465396188567,
      "loss": 3.0574,
      "theoretical_loss": 3.8054840485615102,
      "tokens_seen": 654508032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040487462387161485,
      "loss": 2.7444,
      "theoretical_loss": 3.8054448956863194,
      "tokens_seen": 654573568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040486459378134403,
      "loss": 2.9624,
      "theoretical_loss": 3.8054057478283956,
      "tokens_seen": 654639104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048545636910732,
      "loss": 2.9824,
      "theoretical_loss": 3.805366604986595,
      "tokens_seen": 654704640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040484453360080245,
      "loss": 2.9506,
      "theoretical_loss": 3.8053274671597714,
      "tokens_seen": 654770176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048345035105316,
      "loss": 2.8364,
      "theoretical_loss": 3.8052883343467814,
      "tokens_seen": 654835712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048244734202608,
      "loss": 2.8075,
      "theoretical_loss": 3.805249206546481,
      "tokens_seen": 654901248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040481444332998994,
      "loss": 2.7262,
      "theoretical_loss": 3.8052100837577263,
      "tokens_seen": 654966784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040480441323971917,
      "loss": 2.7897,
      "theoretical_loss": 3.805170965979374,
      "tokens_seen": 655032320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040479438314944835,
      "loss": 2.6729,
      "theoretical_loss": 3.805131853210282,
      "tokens_seen": 655097856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040478435305917753,
      "loss": 2.7982,
      "theoretical_loss": 3.8050927454493064,
      "tokens_seen": 655163392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047743229689067,
      "loss": 2.6755,
      "theoretical_loss": 3.8050536426953063,
      "tokens_seen": 655228928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047642928786359,
      "loss": 2.6891,
      "theoretical_loss": 3.8050145449471398,
      "tokens_seen": 655294464
    },
    {
      "debugging/Self-BLEU-5": NaN,
      "debugging/distinct-1-grams": 0.0,
      "debugging/distinct-2-grams": 0.0,
      "debugging/entropy-1-grams": 0.0,
      "debugging/entropy-2-grams": 0.0,
      "debugging/length": NaN,
      "debugging/num_segments": 0,
      "debugging/score": NaN,
      "debugging/score_std": NaN,
      "epoch": 0.2,
      "objective/train/docs_used": 370281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.031588315963745,
      "objective/train/theoretical_loss": 3.8049754522036645,
      "objective/train/tokens_used": 675820000,
      "theoretical_loss": 3.8049754522036645,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047542627883651,
      "loss": 2.844,
      "theoretical_loss": 3.8049754522036645,
      "tokens_seen": 655360000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047442326980943,
      "loss": 2.7745,
      "theoretical_loss": 3.8049363644637406,
      "tokens_seen": 655425536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040473420260782344,
      "loss": 2.6569,
      "theoretical_loss": 3.8048972817262268,
      "tokens_seen": 655491072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047241725175527,
      "loss": 3.1308,
      "theoretical_loss": 3.8048582039899834,
      "tokens_seen": 655556608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047141424272818,
      "loss": 2.8553,
      "theoretical_loss": 3.8048191312538697,
      "tokens_seen": 655622144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040470411233701104,
      "loss": 2.5904,
      "theoretical_loss": 3.8047800635167466,
      "tokens_seen": 655687680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046940822467402,
      "loss": 2.6663,
      "theoretical_loss": 3.804741000777476,
      "tokens_seen": 655753216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046840521564694,
      "loss": 3.0068,
      "theoretical_loss": 3.8047019430349174,
      "tokens_seen": 655818752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046740220661986,
      "loss": 2.9107,
      "theoretical_loss": 3.8046628902879336,
      "tokens_seen": 655884288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046639919759278,
      "loss": 2.843,
      "theoretical_loss": 3.8046238425353858,
      "tokens_seen": 655949824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404653961885657,
      "loss": 2.7753,
      "theoretical_loss": 3.804584799776137,
      "tokens_seen": 656015360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046439317953862,
      "loss": 2.8462,
      "theoretical_loss": 3.80454576200905,
      "tokens_seen": 656080896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040463390170511536,
      "loss": 2.6241,
      "theoretical_loss": 3.8045067292329877,
      "tokens_seen": 656146432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040462387161484454,
      "loss": 2.8776,
      "theoretical_loss": 3.8044677014468133,
      "tokens_seen": 656211968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046138415245738,
      "loss": 2.8266,
      "theoretical_loss": 3.804428678649391,
      "tokens_seen": 656277504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046038114343029,
      "loss": 2.913,
      "theoretical_loss": 3.8043896608395844,
      "tokens_seen": 656343040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040459378134403214,
      "loss": 2.7933,
      "theoretical_loss": 3.804350648016259,
      "tokens_seen": 656408576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040458375125376127,
      "loss": 2.8499,
      "theoretical_loss": 3.8043116401782795,
      "tokens_seen": 656474112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045737211634905,
      "loss": 2.7959,
      "theoretical_loss": 3.8042726373245115,
      "tokens_seen": 656539648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045636910732197,
      "loss": 2.9439,
      "theoretical_loss": 3.8042336394538196,
      "tokens_seen": 656605184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040455366098294886,
      "loss": 2.9103,
      "theoretical_loss": 3.8041946465650707,
      "tokens_seen": 656670720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040454363089267804,
      "loss": 2.5819,
      "theoretical_loss": 3.8041556586571312,
      "tokens_seen": 656736256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045336008024073,
      "loss": 2.8263,
      "theoretical_loss": 3.804116675728868,
      "tokens_seen": 656801792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045235707121364,
      "loss": 2.9633,
      "theoretical_loss": 3.8040776977791486,
      "tokens_seen": 656867328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040451354062186564,
      "loss": 2.9724,
      "theoretical_loss": 3.8040387248068397,
      "tokens_seen": 656932864
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 371094,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9736223220825195,
      "objective/train/theoretical_loss": 3.8039997568108097,
      "objective/train/tokens_used": 677458400,
      "theoretical_loss": 3.8039997568108097,
      "tokens_seen": 656998400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040450351053159477,
      "loss": 2.8304,
      "theoretical_loss": 3.8039997568108097,
      "tokens_seen": 656998400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404493480441324,
      "loss": 2.9945,
      "theoretical_loss": 3.803960793789927,
      "tokens_seen": 657063936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044834503510532,
      "loss": 2.7311,
      "theoretical_loss": 3.8039218357430604,
      "tokens_seen": 657129472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040447342026078237,
      "loss": 2.7709,
      "theoretical_loss": 3.8038828826690785,
      "tokens_seen": 657195008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040446339017051155,
      "loss": 2.8639,
      "theoretical_loss": 3.803843934566851,
      "tokens_seen": 657260544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040445336008024073,
      "loss": 2.6372,
      "theoretical_loss": 3.8038049914352476,
      "tokens_seen": 657326080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044433299899699,
      "loss": 2.75,
      "theoretical_loss": 3.803766053273139,
      "tokens_seen": 657391616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040443329989969915,
      "loss": 3.0652,
      "theoretical_loss": 3.8037271200793947,
      "tokens_seen": 657457152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040442326980942827,
      "loss": 2.8485,
      "theoretical_loss": 3.803688191852886,
      "tokens_seen": 657522688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044132397191575,
      "loss": 2.6473,
      "theoretical_loss": 3.803649268592485,
      "tokens_seen": 657588224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040440320962888663,
      "loss": 2.7618,
      "theoretical_loss": 3.8036103502970624,
      "tokens_seen": 657653760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040439317953861587,
      "loss": 2.9614,
      "theoretical_loss": 3.8035714369654903,
      "tokens_seen": 657719296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040438314944834505,
      "loss": 2.8396,
      "theoretical_loss": 3.8035325285966413,
      "tokens_seen": 657784832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040437311935807423,
      "loss": 2.9647,
      "theoretical_loss": 3.8034936251893883,
      "tokens_seen": 657850368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043630892678034,
      "loss": 2.844,
      "theoretical_loss": 3.8034547267426038,
      "tokens_seen": 657915904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040435305917753265,
      "loss": 2.7157,
      "theoretical_loss": 3.803415833255162,
      "tokens_seen": 657981440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043430290872618,
      "loss": 2.8558,
      "theoretical_loss": 3.803376944725936,
      "tokens_seen": 658046976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404332998996991,
      "loss": 2.6925,
      "theoretical_loss": 3.803338061153801,
      "tokens_seen": 658112512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040432296890672014,
      "loss": 2.8953,
      "theoretical_loss": 3.8032991825376303,
      "tokens_seen": 658178048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040431293881644937,
      "loss": 2.8478,
      "theoretical_loss": 3.8032603088763,
      "tokens_seen": 658243584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040430290872617855,
      "loss": 2.6926,
      "theoretical_loss": 3.803221440168685,
      "tokens_seen": 658309120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040429287863590773,
      "loss": 2.7789,
      "theoretical_loss": 3.8031825764136604,
      "tokens_seen": 658374656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042828485456369,
      "loss": 2.8289,
      "theoretical_loss": 3.8031437176101033,
      "tokens_seen": 658440192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042728184553661,
      "loss": 2.8906,
      "theoretical_loss": 3.80310486375689,
      "tokens_seen": 658505728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042627883650953,
      "loss": 2.972,
      "theoretical_loss": 3.803066014852896,
      "tokens_seen": 658571264
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 371720,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1109814643859863,
      "objective/train/theoretical_loss": 3.8030271708970003,
      "objective/train/tokens_used": 679096800,
      "theoretical_loss": 3.8030271708970003,
      "tokens_seen": 658636800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042527582748245,
      "loss": 2.7763,
      "theoretical_loss": 3.8030271708970003,
      "tokens_seen": 658636800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040424272818455364,
      "loss": 2.7485,
      "theoretical_loss": 3.802988331888079,
      "tokens_seen": 658702336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042326980942829,
      "loss": 3.001,
      "theoretical_loss": 3.8029494978250105,
      "tokens_seen": 658767872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404222668004012,
      "loss": 2.9512,
      "theoretical_loss": 3.802910668706673,
      "tokens_seen": 658833408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040421263791374124,
      "loss": 2.7941,
      "theoretical_loss": 3.802871844531945,
      "tokens_seen": 658898944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042026078234704,
      "loss": 2.8382,
      "theoretical_loss": 3.802833025299706,
      "tokens_seen": 658964480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041925777331996,
      "loss": 3.1284,
      "theoretical_loss": 3.802794211008835,
      "tokens_seen": 659030016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041825476429288,
      "loss": 2.8907,
      "theoretical_loss": 3.802755401658211,
      "tokens_seen": 659095552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404172517552658,
      "loss": 2.5927,
      "theoretical_loss": 3.802716597246715,
      "tokens_seen": 659161088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040416248746238714,
      "loss": 2.8932,
      "theoretical_loss": 3.802677797773227,
      "tokens_seen": 659226624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041524573721164,
      "loss": 2.766,
      "theoretical_loss": 3.802639003236629,
      "tokens_seen": 659292160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041424272818455,
      "loss": 2.8984,
      "theoretical_loss": 3.8026002136358006,
      "tokens_seen": 659357696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040413239719157474,
      "loss": 2.9041,
      "theoretical_loss": 3.8025614289696232,
      "tokens_seen": 659423232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041223671013039,
      "loss": 2.748,
      "theoretical_loss": 3.8025226492369804,
      "tokens_seen": 659488768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041123370110331,
      "loss": 2.8719,
      "theoretical_loss": 3.802483874436753,
      "tokens_seen": 659554304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041023069207623,
      "loss": 2.6527,
      "theoretical_loss": 3.8024451045678243,
      "tokens_seen": 659619840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040409227683049147,
      "loss": 2.8325,
      "theoretical_loss": 3.8024063396290773,
      "tokens_seen": 659685376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040408224674022065,
      "loss": 2.7487,
      "theoretical_loss": 3.8023675796193945,
      "tokens_seen": 659750912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040722166499499,
      "loss": 2.8749,
      "theoretical_loss": 3.8023288245376614,
      "tokens_seen": 659816448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404062186559679,
      "loss": 2.8572,
      "theoretical_loss": 3.8022900743827606,
      "tokens_seen": 659881984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040405215646940824,
      "loss": 2.9801,
      "theoretical_loss": 3.8022513291535764,
      "tokens_seen": 659947520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040404212637913737,
      "loss": 2.92,
      "theoretical_loss": 3.8022125888489944,
      "tokens_seen": 660013056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040320962888666,
      "loss": 2.8742,
      "theoretical_loss": 3.8021738534678997,
      "tokens_seen": 660078592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040220661985958,
      "loss": 2.8415,
      "theoretical_loss": 3.8021351230091773,
      "tokens_seen": 660144128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040401203610832497,
      "loss": 3.0197,
      "theoretical_loss": 3.802096397471714,
      "tokens_seen": 660209664
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 373016,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.935034990310669,
      "objective/train/theoretical_loss": 3.802057676854395,
      "objective/train/tokens_used": 680735200,
      "theoretical_loss": 3.802057676854395,
      "tokens_seen": 660275200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040400200601805415,
      "loss": 2.6929,
      "theoretical_loss": 3.802057676854395,
      "tokens_seen": 660275200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039919759277834,
      "loss": 2.8014,
      "theoretical_loss": 3.802018961156108,
      "tokens_seen": 660340736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039819458375125,
      "loss": 3.053,
      "theoretical_loss": 3.801980250375739,
      "tokens_seen": 660406272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040397191574724175,
      "loss": 2.898,
      "theoretical_loss": 3.801941544512176,
      "tokens_seen": 660471808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039618856569709,
      "loss": 2.8489,
      "theoretical_loss": 3.8019028435643065,
      "tokens_seen": 660537344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039518555667001,
      "loss": 2.8731,
      "theoretical_loss": 3.8018641475310186,
      "tokens_seen": 660602880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039418254764293,
      "loss": 2.7511,
      "theoretical_loss": 3.801825456411201,
      "tokens_seen": 660668416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040393179538615847,
      "loss": 2.8653,
      "theoretical_loss": 3.8017867702037424,
      "tokens_seen": 660733952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040392176529588765,
      "loss": 2.764,
      "theoretical_loss": 3.801748088907531,
      "tokens_seen": 660799488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040391173520561683,
      "loss": 2.763,
      "theoretical_loss": 3.8017094125214577,
      "tokens_seen": 660865024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040390170511534607,
      "loss": 2.9756,
      "theoretical_loss": 3.8016707410444117,
      "tokens_seen": 660930560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040389167502507525,
      "loss": 2.8566,
      "theoretical_loss": 3.8016320744752834,
      "tokens_seen": 660996096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040388164493480443,
      "loss": 2.8749,
      "theoretical_loss": 3.8015934128129634,
      "tokens_seen": 661061632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038716148445336,
      "loss": 2.8313,
      "theoretical_loss": 3.8015547560563423,
      "tokens_seen": 661127168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040386158475426285,
      "loss": 2.8092,
      "theoretical_loss": 3.801516104204312,
      "tokens_seen": 661192704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403851554663992,
      "loss": 2.9406,
      "theoretical_loss": 3.801477457255764,
      "tokens_seen": 661258240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038415245737212,
      "loss": 2.7659,
      "theoretical_loss": 3.8014388152095906,
      "tokens_seen": 661323776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040383149448345034,
      "loss": 2.9872,
      "theoretical_loss": 3.8014001780646836,
      "tokens_seen": 661389312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040382146439317957,
      "loss": 2.9488,
      "theoretical_loss": 3.801361545819936,
      "tokens_seen": 661454848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040381143430290875,
      "loss": 2.9806,
      "theoretical_loss": 3.8013229184742414,
      "tokens_seen": 661520384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040380140421263794,
      "loss": 2.9426,
      "theoretical_loss": 3.801284296026492,
      "tokens_seen": 661585920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037913741223671,
      "loss": 2.6688,
      "theoretical_loss": 3.8012456784755835,
      "tokens_seen": 661651456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037813440320963,
      "loss": 2.8687,
      "theoretical_loss": 3.801207065820409,
      "tokens_seen": 661716992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037713139418255,
      "loss": 2.9271,
      "theoretical_loss": 3.8011684580598626,
      "tokens_seen": 661782528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037612838515547,
      "loss": 2.7706,
      "theoretical_loss": 3.8011298551928405,
      "tokens_seen": 661848064
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 374068,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0581512451171875,
      "objective/train/theoretical_loss": 3.801091257218237,
      "objective/train/tokens_used": 682373600,
      "theoretical_loss": 3.801091257218237,
      "tokens_seen": 661913600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040375125376128384,
      "loss": 2.7388,
      "theoretical_loss": 3.801091257218237,
      "tokens_seen": 661913600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037412236710131,
      "loss": 2.5943,
      "theoretical_loss": 3.801052664134948,
      "tokens_seen": 661979136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037311935807422,
      "loss": 2.9245,
      "theoretical_loss": 3.80101407594187,
      "tokens_seen": 662044672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040372116349047144,
      "loss": 2.7558,
      "theoretical_loss": 3.8009754926378982,
      "tokens_seen": 662110208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037111334002006,
      "loss": 2.9066,
      "theoretical_loss": 3.8009369142219303,
      "tokens_seen": 662175744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037011033099298,
      "loss": 2.6965,
      "theoretical_loss": 3.8008983406928634,
      "tokens_seen": 662241280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403691073219659,
      "loss": 2.8863,
      "theoretical_loss": 3.800859772049595,
      "tokens_seen": 662306816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036810431293882,
      "loss": 2.7285,
      "theoretical_loss": 3.8008212082910218,
      "tokens_seen": 662372352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040367101303911734,
      "loss": 2.8133,
      "theoretical_loss": 3.8007826494160426,
      "tokens_seen": 662437888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036609829488466,
      "loss": 2.7862,
      "theoretical_loss": 3.8007440954235565,
      "tokens_seen": 662503424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036509528585757,
      "loss": 2.635,
      "theoretical_loss": 3.8007055463124617,
      "tokens_seen": 662568960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040364092276830494,
      "loss": 2.6339,
      "theoretical_loss": 3.8006670020816578,
      "tokens_seen": 662634496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036308926780341,
      "loss": 2.8162,
      "theoretical_loss": 3.800628462730044,
      "tokens_seen": 662700032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036208625877633,
      "loss": 2.8348,
      "theoretical_loss": 3.8005899282565205,
      "tokens_seen": 662765568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036108324974925,
      "loss": 2.8832,
      "theoretical_loss": 3.8005513986599873,
      "tokens_seen": 662831104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040360080240722167,
      "loss": 2.8673,
      "theoretical_loss": 3.8005128739393452,
      "tokens_seen": 662896640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040359077231695085,
      "loss": 3.0328,
      "theoretical_loss": 3.8004743540934953,
      "tokens_seen": 662962176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035807422266801,
      "loss": 2.6331,
      "theoretical_loss": 3.8004358391213393,
      "tokens_seen": 663027712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035707121364092,
      "loss": 2.6823,
      "theoretical_loss": 3.800397329021778,
      "tokens_seen": 663093248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040356068204613844,
      "loss": 2.7886,
      "theoretical_loss": 3.8003588237937143,
      "tokens_seen": 663158784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040355065195586757,
      "loss": 3.0434,
      "theoretical_loss": 3.8003203234360505,
      "tokens_seen": 663224320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035406218655968,
      "loss": 2.8486,
      "theoretical_loss": 3.800281827947689,
      "tokens_seen": 663289856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403530591775326,
      "loss": 2.8891,
      "theoretical_loss": 3.800243337327533,
      "tokens_seen": 663355392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040352056168505517,
      "loss": 2.98,
      "theoretical_loss": 3.8002048515744864,
      "tokens_seen": 663420928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040351053159478435,
      "loss": 3.0469,
      "theoretical_loss": 3.800166370687453,
      "tokens_seen": 663486464
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 374525,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.942661762237549,
      "objective/train/theoretical_loss": 3.8001278946653363,
      "objective/train/tokens_used": 684012000,
      "theoretical_loss": 3.8001278946653363,
      "tokens_seen": 663552000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035005015045136,
      "loss": 2.7616,
      "theoretical_loss": 3.8001278946653363,
      "tokens_seen": 663552000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034904714142427,
      "loss": 2.6917,
      "theoretical_loss": 3.8000894235070417,
      "tokens_seen": 663617536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040348044132397195,
      "loss": 2.8898,
      "theoretical_loss": 3.800050957211474,
      "tokens_seen": 663683072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034704112337011,
      "loss": 2.6352,
      "theoretical_loss": 3.800012495777538,
      "tokens_seen": 663748608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034603811434303,
      "loss": 2.9331,
      "theoretical_loss": 3.7999740392041392,
      "tokens_seen": 663814144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034503510531595,
      "loss": 2.8874,
      "theoretical_loss": 3.799935587490185,
      "tokens_seen": 663879680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040344032096288867,
      "loss": 3.0548,
      "theoretical_loss": 3.7998971406345796,
      "tokens_seen": 663945216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040343029087261785,
      "loss": 2.5845,
      "theoretical_loss": 3.7998586986362315,
      "tokens_seen": 664010752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040342026078234703,
      "loss": 2.8137,
      "theoretical_loss": 3.799820261494047,
      "tokens_seen": 664076288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034102306920762,
      "loss": 2.8299,
      "theoretical_loss": 3.7997818292069327,
      "tokens_seen": 664141824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040340020060180545,
      "loss": 2.8644,
      "theoretical_loss": 3.7997434017737985,
      "tokens_seen": 664207360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033901705115346,
      "loss": 2.6154,
      "theoretical_loss": 3.7997049791935504,
      "tokens_seen": 664272896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033801404212638,
      "loss": 2.739,
      "theoretical_loss": 3.7996665614650977,
      "tokens_seen": 664338432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403370110330993,
      "loss": 2.6332,
      "theoretical_loss": 3.799628148587349,
      "tokens_seen": 664403968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033600802407222,
      "loss": 2.6486,
      "theoretical_loss": 3.799589740559214,
      "tokens_seen": 664469504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040335005015045136,
      "loss": 2.8741,
      "theoretical_loss": 3.7995513373796017,
      "tokens_seen": 664535040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040334002006018054,
      "loss": 3.0168,
      "theoretical_loss": 3.799512939047422,
      "tokens_seen": 664600576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033299899699097,
      "loss": 2.946,
      "theoretical_loss": 3.7994745455615857,
      "tokens_seen": 664666112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040331995987963895,
      "loss": 2.8055,
      "theoretical_loss": 3.7994361569210024,
      "tokens_seen": 664731648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033099297893681,
      "loss": 2.9599,
      "theoretical_loss": 3.799397773124584,
      "tokens_seen": 664797184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032998996990973,
      "loss": 2.7457,
      "theoretical_loss": 3.7993593941712405,
      "tokens_seen": 664862720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040328986960882644,
      "loss": 2.9906,
      "theoretical_loss": 3.7993210200598853,
      "tokens_seen": 664928256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032798395185557,
      "loss": 2.7725,
      "theoretical_loss": 3.7992826507894284,
      "tokens_seen": 664993792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040326980942828486,
      "loss": 2.564,
      "theoretical_loss": 3.7992442863587836,
      "tokens_seen": 665059328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040325977933801404,
      "loss": 2.7479,
      "theoretical_loss": 3.7992059267668634,
      "tokens_seen": 665124864
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 375614,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8573155403137207,
      "objective/train/theoretical_loss": 3.79916757201258,
      "objective/train/tokens_used": 685650400,
      "theoretical_loss": 3.79916757201258,
      "tokens_seen": 665190400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032497492477432,
      "loss": 2.8858,
      "theoretical_loss": 3.79916757201258,
      "tokens_seen": 665190400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032397191574724,
      "loss": 3.0415,
      "theoretical_loss": 3.799129222094848,
      "tokens_seen": 665255936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032296890672016,
      "loss": 3.0551,
      "theoretical_loss": 3.7990908770125804,
      "tokens_seen": 665321472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032196589769308,
      "loss": 2.8895,
      "theoretical_loss": 3.799052536764691,
      "tokens_seen": 665387008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040320962888665995,
      "loss": 2.7187,
      "theoretical_loss": 3.7990142013500945,
      "tokens_seen": 665452544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031995987963892,
      "loss": 2.8924,
      "theoretical_loss": 3.7989758707677064,
      "tokens_seen": 665518080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040318956870611836,
      "loss": 2.8798,
      "theoretical_loss": 3.7989375450164413,
      "tokens_seen": 665583616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040317953861584754,
      "loss": 2.915,
      "theoretical_loss": 3.798899224095214,
      "tokens_seen": 665649152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031695085255767,
      "loss": 2.8973,
      "theoretical_loss": 3.798860908002941,
      "tokens_seen": 665714688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031594784353059,
      "loss": 2.725,
      "theoretical_loss": 3.798822596738539,
      "tokens_seen": 665780224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040314944834503514,
      "loss": 2.9635,
      "theoretical_loss": 3.7987842903009237,
      "tokens_seen": 665845760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031394182547643,
      "loss": 2.8181,
      "theoretical_loss": 3.7987459886890127,
      "tokens_seen": 665911296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031293881644935,
      "loss": 2.9355,
      "theoretical_loss": 3.798707691901722,
      "tokens_seen": 665976832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031193580742227,
      "loss": 2.6325,
      "theoretical_loss": 3.798669399937971,
      "tokens_seen": 666042368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040310932798395187,
      "loss": 2.7554,
      "theoretical_loss": 3.7986311127966763,
      "tokens_seen": 666107904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040309929789368105,
      "loss": 2.917,
      "theoretical_loss": 3.798592830476757,
      "tokens_seen": 666173440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030892678034103,
      "loss": 2.6412,
      "theoretical_loss": 3.79855455297713,
      "tokens_seen": 666238976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030792377131394,
      "loss": 2.8717,
      "theoretical_loss": 3.7985162802967167,
      "tokens_seen": 666304512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040306920762286864,
      "loss": 2.7658,
      "theoretical_loss": 3.798478012434435,
      "tokens_seen": 666370048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040305917753259777,
      "loss": 3.1829,
      "theoretical_loss": 3.7984397493892055,
      "tokens_seen": 666435584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403049147442327,
      "loss": 2.9596,
      "theoretical_loss": 3.798401491159947,
      "tokens_seen": 666501120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030391173520562,
      "loss": 2.9012,
      "theoretical_loss": 3.7983632377455807,
      "tokens_seen": 666566656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040302908726178537,
      "loss": 2.7809,
      "theoretical_loss": 3.7983249891450273,
      "tokens_seen": 666632192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040301905717151455,
      "loss": 2.761,
      "theoretical_loss": 3.7982867453572076,
      "tokens_seen": 666697728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030090270812438,
      "loss": 2.7542,
      "theoretical_loss": 3.798248506381043,
      "tokens_seen": 666763264
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 376311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.005213975906372,
      "objective/train/theoretical_loss": 3.7982102722154556,
      "objective/train/tokens_used": 687288800,
      "theoretical_loss": 3.7982102722154556,
      "tokens_seen": 666828800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029989969909729,
      "loss": 2.8448,
      "theoretical_loss": 3.7982102722154556,
      "tokens_seen": 666828800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040298896690070215,
      "loss": 2.6927,
      "theoretical_loss": 3.7981720428593677,
      "tokens_seen": 666894336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029789368104313,
      "loss": 2.7377,
      "theoretical_loss": 3.798133818311701,
      "tokens_seen": 666959872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029689067201605,
      "loss": 2.5689,
      "theoretical_loss": 3.798095598571379,
      "tokens_seen": 667025408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029588766298897,
      "loss": 2.8091,
      "theoretical_loss": 3.798057383637324,
      "tokens_seen": 667090944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040294884653961887,
      "loss": 2.9873,
      "theoretical_loss": 3.798019173508461,
      "tokens_seen": 667156480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040293881644934805,
      "loss": 2.8216,
      "theoretical_loss": 3.7979809681837127,
      "tokens_seen": 667222016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040292878635907723,
      "loss": 2.8597,
      "theoretical_loss": 3.7979427676620032,
      "tokens_seen": 667287552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029187562688064,
      "loss": 2.8129,
      "theoretical_loss": 3.797904571942258,
      "tokens_seen": 667353088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040290872617853565,
      "loss": 2.7172,
      "theoretical_loss": 3.797866381023401,
      "tokens_seen": 667418624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028986960882648,
      "loss": 2.6435,
      "theoretical_loss": 3.7978281949043575,
      "tokens_seen": 667484160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402888665997994,
      "loss": 2.6663,
      "theoretical_loss": 3.797790013584054,
      "tokens_seen": 667549696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028786359077232,
      "loss": 2.9507,
      "theoretical_loss": 3.7977518370614156,
      "tokens_seen": 667615232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028686058174524,
      "loss": 2.8432,
      "theoretical_loss": 3.797713665335369,
      "tokens_seen": 667680768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040285857572718156,
      "loss": 2.7833,
      "theoretical_loss": 3.797675498404841,
      "tokens_seen": 667746304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040284854563691074,
      "loss": 2.8119,
      "theoretical_loss": 3.7976373362687577,
      "tokens_seen": 667811840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028385155466399,
      "loss": 2.8527,
      "theoretical_loss": 3.7975991789260473,
      "tokens_seen": 667877376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040282848545636915,
      "loss": 2.8808,
      "theoretical_loss": 3.7975610263756368,
      "tokens_seen": 667942912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028184553660983,
      "loss": 2.6857,
      "theoretical_loss": 3.797522878616455,
      "tokens_seen": 668008448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028084252758275,
      "loss": 2.9767,
      "theoretical_loss": 3.7974847356474295,
      "tokens_seen": 668073984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040279839518555664,
      "loss": 2.8602,
      "theoretical_loss": 3.797446597467489,
      "tokens_seen": 668139520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027883650952859,
      "loss": 2.8874,
      "theoretical_loss": 3.797408464075563,
      "tokens_seen": 668205056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040277833500501506,
      "loss": 2.6017,
      "theoretical_loss": 3.797370335470581,
      "tokens_seen": 668270592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040276830491474424,
      "loss": 2.5136,
      "theoretical_loss": 3.7973322116514723,
      "tokens_seen": 668336128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027582748244734,
      "loss": 2.704,
      "theoretical_loss": 3.797294092617167,
      "tokens_seen": 668401664
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 377619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8243908882141113,
      "objective/train/theoretical_loss": 3.7972559783665965,
      "objective/train/tokens_used": 688927200,
      "theoretical_loss": 3.7972559783665965,
      "tokens_seen": 668467200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027482447342026,
      "loss": 2.5602,
      "theoretical_loss": 3.7972559783665965,
      "tokens_seen": 668467200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027382146439318,
      "loss": 2.8683,
      "theoretical_loss": 3.7972178688986897,
      "tokens_seen": 668532736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402728184553661,
      "loss": 2.5877,
      "theoretical_loss": 3.7971797642123786,
      "tokens_seen": 668598272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040271815446339015,
      "loss": 2.9583,
      "theoretical_loss": 3.7971416643065954,
      "tokens_seen": 668663808
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027081243731194,
      "loss": 2.7348,
      "theoretical_loss": 3.797103569180271,
      "tokens_seen": 668729344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040269809428284856,
      "loss": 2.7532,
      "theoretical_loss": 3.7970654788323377,
      "tokens_seen": 668794880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040268806419257774,
      "loss": 2.9621,
      "theoretical_loss": 3.797027393261728,
      "tokens_seen": 668860416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026780341023069,
      "loss": 2.8518,
      "theoretical_loss": 3.7969893124673746,
      "tokens_seen": 668925952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026680040120361,
      "loss": 2.7747,
      "theoretical_loss": 3.7969512364482116,
      "tokens_seen": 668991488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026579739217653,
      "loss": 2.7672,
      "theoretical_loss": 3.796913165203171,
      "tokens_seen": 669057024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026479438314945,
      "loss": 2.896,
      "theoretical_loss": 3.7968750987311877,
      "tokens_seen": 669122560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040263791374122365,
      "loss": 2.9705,
      "theoretical_loss": 3.796837037031196,
      "tokens_seen": 669188096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040263791374122365,
      "loss": 3.0305,
      "theoretical_loss": 3.796798980102129,
      "tokens_seen": 669253632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026278836509529,
      "loss": 2.8599,
      "theoretical_loss": 3.7967609279429233,
      "tokens_seen": 669319168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402617853560682,
      "loss": 2.8396,
      "theoretical_loss": 3.7967228805525135,
      "tokens_seen": 669384704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040260782347041125,
      "loss": 2.7468,
      "theoretical_loss": 3.796684837929835,
      "tokens_seen": 669450240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040259779338014043,
      "loss": 2.9226,
      "theoretical_loss": 3.796646800073824,
      "tokens_seen": 669515776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025877632898696,
      "loss": 2.8581,
      "theoretical_loss": 3.796608766983416,
      "tokens_seen": 669581312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025777331995988,
      "loss": 2.9392,
      "theoretical_loss": 3.7965707386575485,
      "tokens_seen": 669646848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040256770310932797,
      "loss": 2.8257,
      "theoretical_loss": 3.796532715095158,
      "tokens_seen": 669712384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040255767301905715,
      "loss": 2.7459,
      "theoretical_loss": 3.7964946962951815,
      "tokens_seen": 669777920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025476429287864,
      "loss": 3.0115,
      "theoretical_loss": 3.7964566822565575,
      "tokens_seen": 669843456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025376128385155,
      "loss": 2.7659,
      "theoretical_loss": 3.796418672978223,
      "tokens_seen": 669908992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040252758274824475,
      "loss": 2.7765,
      "theoretical_loss": 3.7963806684591166,
      "tokens_seen": 669974528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040251755265797393,
      "loss": 2.9028,
      "theoretical_loss": 3.7963426686981774,
      "tokens_seen": 670040064
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 378282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.932396650314331,
      "objective/train/theoretical_loss": 3.796304673694344,
      "objective/train/tokens_used": 690565600,
      "theoretical_loss": 3.796304673694344,
      "tokens_seen": 670105600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025075225677031,
      "loss": 2.9771,
      "theoretical_loss": 3.796304673694344,
      "tokens_seen": 670105600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024974924774323,
      "loss": 2.7668,
      "theoretical_loss": 3.7962666834465555,
      "tokens_seen": 670171136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024874623871615,
      "loss": 2.634,
      "theoretical_loss": 3.796228697953752,
      "tokens_seen": 670236672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040247743229689066,
      "loss": 2.9584,
      "theoretical_loss": 3.7961907172148726,
      "tokens_seen": 670302208
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024674022066199,
      "loss": 2.8543,
      "theoretical_loss": 3.7961527412288585,
      "tokens_seen": 670367744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402457372116349,
      "loss": 3.0343,
      "theoretical_loss": 3.7961147699946505,
      "tokens_seen": 670433280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040244734202607825,
      "loss": 2.8191,
      "theoretical_loss": 3.796076803511189,
      "tokens_seen": 670498816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024373119358074,
      "loss": 2.7031,
      "theoretical_loss": 3.7960388417774156,
      "tokens_seen": 670564352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024272818455366,
      "loss": 3.0333,
      "theoretical_loss": 3.796000884792272,
      "tokens_seen": 670629888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024172517552658,
      "loss": 2.9055,
      "theoretical_loss": 3.7959629325547004,
      "tokens_seen": 670695424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402407221664995,
      "loss": 3.193,
      "theoretical_loss": 3.7959249850636425,
      "tokens_seen": 670760960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023971915747242,
      "loss": 2.7088,
      "theoretical_loss": 3.795887042318042,
      "tokens_seen": 670826496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023871614844534,
      "loss": 2.6832,
      "theoretical_loss": 3.7958491043168414,
      "tokens_seen": 670892032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023771313941826,
      "loss": 2.6732,
      "theoretical_loss": 3.7958111710589844,
      "tokens_seen": 670957568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040236710130391176,
      "loss": 2.7021,
      "theoretical_loss": 3.7957732425434143,
      "tokens_seen": 671023104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040235707121364094,
      "loss": 2.4503,
      "theoretical_loss": 3.7957353187690748,
      "tokens_seen": 671088640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023470411233701,
      "loss": 2.8226,
      "theoretical_loss": 3.7956973997349115,
      "tokens_seen": 671154176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040233701103309935,
      "loss": 3.0277,
      "theoretical_loss": 3.7956594854398684,
      "tokens_seen": 671219712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023269809428285,
      "loss": 2.7509,
      "theoretical_loss": 3.7956215758828904,
      "tokens_seen": 671285248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023169508525577,
      "loss": 2.7952,
      "theoretical_loss": 3.795583671062923,
      "tokens_seen": 671350784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040230692076228684,
      "loss": 2.8198,
      "theoretical_loss": 3.7955457709789124,
      "tokens_seen": 671416320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022968906720161,
      "loss": 2.8985,
      "theoretical_loss": 3.795507875629805,
      "tokens_seen": 671481856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040228686058174526,
      "loss": 2.5066,
      "theoretical_loss": 3.7954699850145457,
      "tokens_seen": 671547392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040227683049147444,
      "loss": 2.6576,
      "theoretical_loss": 3.795432099132083,
      "tokens_seen": 671612928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022668004012036,
      "loss": 2.6896,
      "theoretical_loss": 3.7953942179813627,
      "tokens_seen": 671678464
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 379317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.70872163772583,
      "objective/train/theoretical_loss": 3.7953563415613325,
      "objective/train/tokens_used": 692204000,
      "theoretical_loss": 3.7953563415613325,
      "tokens_seen": 671744000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022567703109328,
      "loss": 2.8489,
      "theoretical_loss": 3.7953563415613325,
      "tokens_seen": 671744000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402246740220662,
      "loss": 2.6866,
      "theoretical_loss": 3.795318469870941,
      "tokens_seen": 671809536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022367101303912,
      "loss": 3.0096,
      "theoretical_loss": 3.795280602909135,
      "tokens_seen": 671875072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040222668004012035,
      "loss": 2.771,
      "theoretical_loss": 3.7952427406748646,
      "tokens_seen": 671940608
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022166499498496,
      "loss": 2.5575,
      "theoretical_loss": 3.7952048831670773,
      "tokens_seen": 672006144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040220661985957876,
      "loss": 2.7143,
      "theoretical_loss": 3.795167030384722,
      "tokens_seen": 672071680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040219658976930794,
      "loss": 2.7934,
      "theoretical_loss": 3.7951291823267495,
      "tokens_seen": 672137216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021865596790371,
      "loss": 2.986,
      "theoretical_loss": 3.7950913389921084,
      "tokens_seen": 672202752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021765295887663,
      "loss": 2.8392,
      "theoretical_loss": 3.7950535003797494,
      "tokens_seen": 672268288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021664994984955,
      "loss": 2.7923,
      "theoretical_loss": 3.795015666488623,
      "tokens_seen": 672333824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021564694082247,
      "loss": 2.7625,
      "theoretical_loss": 3.7949778373176795,
      "tokens_seen": 672399360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040214643931795385,
      "loss": 2.8883,
      "theoretical_loss": 3.7949400128658706,
      "tokens_seen": 672464896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021364092276831,
      "loss": 2.8509,
      "theoretical_loss": 3.7949021931321476,
      "tokens_seen": 672530432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021263791374122,
      "loss": 2.662,
      "theoretical_loss": 3.794864378115462,
      "tokens_seen": 672595968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040211634904714145,
      "loss": 2.6493,
      "theoretical_loss": 3.7948265678147663,
      "tokens_seen": 672661504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040210631895687063,
      "loss": 2.8812,
      "theoretical_loss": 3.7947887622290133,
      "tokens_seen": 672727040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020962888665998,
      "loss": 2.9279,
      "theoretical_loss": 3.7947509613571553,
      "tokens_seen": 672792576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402086258776329,
      "loss": 2.8662,
      "theoretical_loss": 3.7947131651981447,
      "tokens_seen": 672858112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040207622868605817,
      "loss": 2.9209,
      "theoretical_loss": 3.7946753737509367,
      "tokens_seen": 672923648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040206619859578735,
      "loss": 2.86,
      "theoretical_loss": 3.794637587014484,
      "tokens_seen": 672989184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020561685055166,
      "loss": 2.9311,
      "theoretical_loss": 3.7945998049877403,
      "tokens_seen": 673054720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020461384152457,
      "loss": 2.7814,
      "theoretical_loss": 3.794562027669661,
      "tokens_seen": 673120256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040203610832497495,
      "loss": 3.022,
      "theoretical_loss": 3.7945242550592013,
      "tokens_seen": 673185792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040202607823470413,
      "loss": 2.8168,
      "theoretical_loss": 3.7944864871553152,
      "tokens_seen": 673251328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020160481444333,
      "loss": 2.8123,
      "theoretical_loss": 3.794448723956959,
      "tokens_seen": 673316864
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 379824,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8459346294403076,
      "objective/train/theoretical_loss": 3.794410965463088,
      "objective/train/tokens_used": 693842400,
      "theoretical_loss": 3.794410965463088,
      "tokens_seen": 673382400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020060180541625,
      "loss": 2.9922,
      "theoretical_loss": 3.794410965463088,
      "tokens_seen": 673382400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019959879638917,
      "loss": 2.5647,
      "theoretical_loss": 3.7943732116726583,
      "tokens_seen": 673447936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040198595787362086,
      "loss": 2.6541,
      "theoretical_loss": 3.794335462584627,
      "tokens_seen": 673513472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019759277833501,
      "loss": 2.7915,
      "theoretical_loss": 3.7942977181979503,
      "tokens_seen": 673579008
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019658976930792,
      "loss": 2.7506,
      "theoretical_loss": 3.794259978511586,
      "tokens_seen": 673644544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040195586760280845,
      "loss": 2.8794,
      "theoretical_loss": 3.794222243524491,
      "tokens_seen": 673710080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019458375125376,
      "loss": 2.9101,
      "theoretical_loss": 3.7941845132356233,
      "tokens_seen": 673775616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019358074222668,
      "loss": 3.0697,
      "theoretical_loss": 3.7941467876439408,
      "tokens_seen": 673841152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401925777331996,
      "loss": 3.0465,
      "theoretical_loss": 3.7941090667484025,
      "tokens_seen": 673906688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019157472417252,
      "loss": 2.7341,
      "theoretical_loss": 3.794071350547967,
      "tokens_seen": 673972224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040190571715145436,
      "loss": 2.8905,
      "theoretical_loss": 3.794033639041593,
      "tokens_seen": 674037760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018956870611836,
      "loss": 2.7838,
      "theoretical_loss": 3.793995932228241,
      "tokens_seen": 674103296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018856569709127,
      "loss": 2.8153,
      "theoretical_loss": 3.7939582301068695,
      "tokens_seen": 674168832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040187562688064196,
      "loss": 2.647,
      "theoretical_loss": 3.79392053267644,
      "tokens_seen": 674234368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018655967903711,
      "loss": 2.6909,
      "theoretical_loss": 3.7938828399359115,
      "tokens_seen": 674299904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018555667001003,
      "loss": 3.1,
      "theoretical_loss": 3.7938451518842458,
      "tokens_seen": 674365440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018455366098295,
      "loss": 2.9233,
      "theoretical_loss": 3.793807468520404,
      "tokens_seen": 674430976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018355065195587,
      "loss": 2.8234,
      "theoretical_loss": 3.7937697898433473,
      "tokens_seen": 674496512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040182547642928786,
      "loss": 2.998,
      "theoretical_loss": 3.7937321158520376,
      "tokens_seen": 674562048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040181544633901704,
      "loss": 2.9544,
      "theoretical_loss": 3.7936944465454374,
      "tokens_seen": 674627584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018054162487462,
      "loss": 2.7967,
      "theoretical_loss": 3.7936567819225084,
      "tokens_seen": 674693120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040179538615847546,
      "loss": 2.8876,
      "theoretical_loss": 3.7936191219822133,
      "tokens_seen": 674758656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017853560682046,
      "loss": 2.907,
      "theoretical_loss": 3.7935814667235164,
      "tokens_seen": 674824192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017753259779338,
      "loss": 2.9187,
      "theoretical_loss": 3.79354381614538,
      "tokens_seen": 674889728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040176529588766295,
      "loss": 2.9948,
      "theoretical_loss": 3.793506170246768,
      "tokens_seen": 674955264
    },
    {
      "epoch": 0.2,
      "objective/train/docs_used": 381014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8223540782928467,
      "objective/train/theoretical_loss": 3.7934685290266454,
      "objective/train/tokens_used": 695480800,
      "theoretical_loss": 3.7934685290266454,
      "tokens_seen": 675020800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040176529588766295,
      "loss": 2.8171,
      "theoretical_loss": 3.7934685290266454,
      "tokens_seen": 675020800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017552657973922,
      "loss": 2.9283,
      "theoretical_loss": 3.7934308924839755,
      "tokens_seen": 675086336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040174523570712137,
      "loss": 2.9153,
      "theoretical_loss": 3.793393260617724,
      "tokens_seen": 675151872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040173520561685055,
      "loss": 2.7576,
      "theoretical_loss": 3.7933556334268554,
      "tokens_seen": 675217408
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040172517552657973,
      "loss": 2.7632,
      "theoretical_loss": 3.793318010910335,
      "tokens_seen": 675282944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040171514543630896,
      "loss": 2.6694,
      "theoretical_loss": 3.7932803930671297,
      "tokens_seen": 675348480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017051153460381,
      "loss": 2.9775,
      "theoretical_loss": 3.7932427798962043,
      "tokens_seen": 675414016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016950852557673,
      "loss": 2.9839,
      "theoretical_loss": 3.793205171396526,
      "tokens_seen": 675479552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040168505516549645,
      "loss": 2.7652,
      "theoretical_loss": 3.7931675675670604,
      "tokens_seen": 675545088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016750250752257,
      "loss": 2.7479,
      "theoretical_loss": 3.7931299684067756,
      "tokens_seen": 675610624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040166499498495487,
      "loss": 2.564,
      "theoretical_loss": 3.7930923739146394,
      "tokens_seen": 675676160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040165496489468405,
      "loss": 2.5691,
      "theoretical_loss": 3.7930547840896187,
      "tokens_seen": 675741696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016449348044133,
      "loss": 2.9134,
      "theoretical_loss": 3.7930171989306816,
      "tokens_seen": 675807232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016349047141424,
      "loss": 2.7036,
      "theoretical_loss": 3.7929796184367968,
      "tokens_seen": 675872768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040162487462387165,
      "loss": 2.6151,
      "theoretical_loss": 3.7929420426069327,
      "tokens_seen": 675938304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040161484453360083,
      "loss": 2.7893,
      "theoretical_loss": 3.7929044714400586,
      "tokens_seen": 676003840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040160481444333,
      "loss": 2.6284,
      "theoretical_loss": 3.792866904935144,
      "tokens_seen": 676069376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015947843530592,
      "loss": 2.7396,
      "theoretical_loss": 3.792829343091158,
      "tokens_seen": 676134912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040158475426278837,
      "loss": 2.8118,
      "theoretical_loss": 3.7927917859070712,
      "tokens_seen": 676200448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040157472417251755,
      "loss": 2.774,
      "theoretical_loss": 3.792754233381854,
      "tokens_seen": 676265984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015646940822468,
      "loss": 3.0124,
      "theoretical_loss": 3.7927166855144763,
      "tokens_seen": 676331520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015546639919759,
      "loss": 2.5057,
      "theoretical_loss": 3.7926791423039097,
      "tokens_seen": 676397056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040154463390170515,
      "loss": 2.8492,
      "theoretical_loss": 3.792641603749126,
      "tokens_seen": 676462592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040153460381143433,
      "loss": 2.8639,
      "theoretical_loss": 3.7926040698490957,
      "tokens_seen": 676528128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004015245737211635,
      "loss": 3.042,
      "theoretical_loss": 3.7925665406027917,
      "tokens_seen": 676593664
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 382043,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.4060447216033936,
      "objective/train/theoretical_loss": 3.792529016009186,
      "objective/train/tokens_used": 697119200,
      "theoretical_loss": 3.792529016009186,
      "tokens_seen": 676659200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004015145436308927,
      "loss": 2.8706,
      "theoretical_loss": 3.792529016009186,
      "tokens_seen": 676659200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004015045135406219,
      "loss": 2.7768,
      "theoretical_loss": 3.7924914960672513,
      "tokens_seen": 676724736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040149448345035106,
      "loss": 2.817,
      "theoretical_loss": 3.792453980775961,
      "tokens_seen": 676790272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014844533600803,
      "loss": 2.7802,
      "theoretical_loss": 3.7924164701342873,
      "tokens_seen": 676855808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014744232698094,
      "loss": 2.9134,
      "theoretical_loss": 3.7923789641412045,
      "tokens_seen": 676921344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040146439317953865,
      "loss": 2.8138,
      "theoretical_loss": 3.7923414627956866,
      "tokens_seen": 676986880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014543630892678,
      "loss": 2.6232,
      "theoretical_loss": 3.792303966096708,
      "tokens_seen": 677052416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401444332998997,
      "loss": 2.9834,
      "theoretical_loss": 3.792266474043243,
      "tokens_seen": 677117952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014343029087262,
      "loss": 2.8613,
      "theoretical_loss": 3.7922289866342664,
      "tokens_seen": 677183488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014242728184554,
      "loss": 2.7659,
      "theoretical_loss": 3.7921915038687537,
      "tokens_seen": 677249024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040141424272818456,
      "loss": 2.8298,
      "theoretical_loss": 3.7921540257456803,
      "tokens_seen": 677314560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014042126379138,
      "loss": 2.7942,
      "theoretical_loss": 3.7921165522640226,
      "tokens_seen": 677380096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013941825476429,
      "loss": 2.4966,
      "theoretical_loss": 3.792079083422756,
      "tokens_seen": 677445632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040138415245737216,
      "loss": 2.7792,
      "theoretical_loss": 3.7920416192208575,
      "tokens_seen": 677511168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013741223671013,
      "loss": 2.6478,
      "theoretical_loss": 3.7920041596573038,
      "tokens_seen": 677576704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013640922768305,
      "loss": 2.6984,
      "theoretical_loss": 3.7919667047310726,
      "tokens_seen": 677642240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013540621865597,
      "loss": 2.6497,
      "theoretical_loss": 3.7919292544411407,
      "tokens_seen": 677707776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013440320962889,
      "loss": 2.7696,
      "theoretical_loss": 3.791891808786487,
      "tokens_seen": 677773312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040133400200601806,
      "loss": 2.8313,
      "theoretical_loss": 3.791854367766088,
      "tokens_seen": 677838848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040132397191574724,
      "loss": 2.9008,
      "theoretical_loss": 3.7918169313789236,
      "tokens_seen": 677904384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013139418254764,
      "loss": 2.7586,
      "theoretical_loss": 3.7917794996239724,
      "tokens_seen": 677969920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040130391173520566,
      "loss": 2.8737,
      "theoretical_loss": 3.791742072500213,
      "tokens_seen": 678035456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012938816449348,
      "loss": 2.9405,
      "theoretical_loss": 3.7917046500066256,
      "tokens_seen": 678100992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401283851554664,
      "loss": 2.7992,
      "theoretical_loss": 3.7916672321421894,
      "tokens_seen": 678166528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040127382146439315,
      "loss": 2.8993,
      "theoretical_loss": 3.791629818905884,
      "tokens_seen": 678232064
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 382695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.529493808746338,
      "objective/train/theoretical_loss": 3.7915924102966914,
      "objective/train/tokens_used": 698757600,
      "theoretical_loss": 3.7915924102966914,
      "tokens_seen": 678297600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012637913741224,
      "loss": 2.7103,
      "theoretical_loss": 3.7915924102966914,
      "tokens_seen": 678297600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040125376128385157,
      "loss": 2.852,
      "theoretical_loss": 3.7915550063135908,
      "tokens_seen": 678363136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040124373119358075,
      "loss": 3.0078,
      "theoretical_loss": 3.7915176069555643,
      "tokens_seen": 678428672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040123370110330993,
      "loss": 2.9411,
      "theoretical_loss": 3.791480212221593,
      "tokens_seen": 678494208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040122367101303916,
      "loss": 2.9414,
      "theoretical_loss": 3.7914428221106586,
      "tokens_seen": 678559744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012136409227683,
      "loss": 2.8493,
      "theoretical_loss": 3.791405436621743,
      "tokens_seen": 678625280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012036108324975,
      "loss": 2.5601,
      "theoretical_loss": 3.7913680557538285,
      "tokens_seen": 678690816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040119358074222665,
      "loss": 2.6255,
      "theoretical_loss": 3.791330679505898,
      "tokens_seen": 678756352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011835506519559,
      "loss": 2.9282,
      "theoretical_loss": 3.791293307876935,
      "tokens_seen": 678821888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040117352056168507,
      "loss": 2.784,
      "theoretical_loss": 3.7912559408659217,
      "tokens_seen": 678887424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040116349047141425,
      "loss": 2.7854,
      "theoretical_loss": 3.7912185784718435,
      "tokens_seen": 678952960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040115346038114343,
      "loss": 2.6421,
      "theoretical_loss": 3.7911812206936824,
      "tokens_seen": 679018496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011434302908726,
      "loss": 2.9031,
      "theoretical_loss": 3.7911438675304234,
      "tokens_seen": 679084032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011334002006018,
      "loss": 2.8137,
      "theoretical_loss": 3.791106518981052,
      "tokens_seen": 679149568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040112337011033103,
      "loss": 3.0802,
      "theoretical_loss": 3.7910691750445515,
      "tokens_seen": 679215104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040111334002006016,
      "loss": 2.664,
      "theoretical_loss": 3.7910318357199086,
      "tokens_seen": 679280640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011033099297894,
      "loss": 2.9432,
      "theoretical_loss": 3.790994501006109,
      "tokens_seen": 679346176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010932798395185,
      "loss": 2.9391,
      "theoretical_loss": 3.7909571709021375,
      "tokens_seen": 679411712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040108324974924775,
      "loss": 2.8087,
      "theoretical_loss": 3.7909198454069806,
      "tokens_seen": 679477248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040107321965897693,
      "loss": 2.7766,
      "theoretical_loss": 3.790882524519626,
      "tokens_seen": 679542784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010631895687061,
      "loss": 2.8428,
      "theoretical_loss": 3.7908452082390585,
      "tokens_seen": 679608320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010531594784353,
      "loss": 2.7802,
      "theoretical_loss": 3.7908078965642673,
      "tokens_seen": 679673856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040104312938816453,
      "loss": 2.659,
      "theoretical_loss": 3.7907705894942385,
      "tokens_seen": 679739392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040103309929789366,
      "loss": 2.6857,
      "theoretical_loss": 3.790733287027961,
      "tokens_seen": 679804928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010230692076229,
      "loss": 2.7388,
      "theoretical_loss": 3.790695989164422,
      "tokens_seen": 679870464
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 383786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7080867290496826,
      "objective/train/theoretical_loss": 3.7906586959026107,
      "objective/train/tokens_used": 700396000,
      "theoretical_loss": 3.7906586959026107,
      "tokens_seen": 679936000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401013039117352,
      "loss": 2.8253,
      "theoretical_loss": 3.7906586959026107,
      "tokens_seen": 679936000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040100300902708126,
      "loss": 3.1717,
      "theoretical_loss": 3.7906214072415154,
      "tokens_seen": 680001536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040099297893681044,
      "loss": 2.9097,
      "theoretical_loss": 3.7905841231801256,
      "tokens_seen": 680067072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009829488465396,
      "loss": 2.8925,
      "theoretical_loss": 3.790546843717431,
      "tokens_seen": 680132608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009729187562688,
      "loss": 2.6975,
      "theoretical_loss": 3.79050956885242,
      "tokens_seen": 680198144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400962888665998,
      "loss": 2.6732,
      "theoretical_loss": 3.790472298584084,
      "tokens_seen": 680263680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040095285857572716,
      "loss": 2.906,
      "theoretical_loss": 3.7904350329114127,
      "tokens_seen": 680329216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009428284854564,
      "loss": 2.7083,
      "theoretical_loss": 3.7903977718333977,
      "tokens_seen": 680394752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009327983951855,
      "loss": 2.8424,
      "theoretical_loss": 3.790360515349029,
      "tokens_seen": 680460288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040092276830491476,
      "loss": 2.9744,
      "theoretical_loss": 3.7903232634572985,
      "tokens_seen": 680525824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009127382146439,
      "loss": 2.8977,
      "theoretical_loss": 3.7902860161571974,
      "tokens_seen": 680591360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009027081243731,
      "loss": 2.7069,
      "theoretical_loss": 3.7902487734477184,
      "tokens_seen": 680656896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040089267803410236,
      "loss": 2.8793,
      "theoretical_loss": 3.790211535327853,
      "tokens_seen": 680722432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008826479438315,
      "loss": 2.6679,
      "theoretical_loss": 3.7901743017965943,
      "tokens_seen": 680787968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008726178535607,
      "loss": 2.8633,
      "theoretical_loss": 3.7901370728529353,
      "tokens_seen": 680853504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008625877632899,
      "loss": 2.8139,
      "theoretical_loss": 3.790099848495869,
      "tokens_seen": 680919040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008525576730191,
      "loss": 2.8153,
      "theoretical_loss": 3.790062628724389,
      "tokens_seen": 680984576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040084252758274826,
      "loss": 2.7675,
      "theoretical_loss": 3.790025413537489,
      "tokens_seen": 681050112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040083249749247744,
      "loss": 2.9715,
      "theoretical_loss": 3.789988202934164,
      "tokens_seen": 681115648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008224674022066,
      "loss": 2.936,
      "theoretical_loss": 3.789950996913408,
      "tokens_seen": 681181184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040081243731193586,
      "loss": 2.655,
      "theoretical_loss": 3.7899137954742153,
      "tokens_seen": 681246720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400802407221665,
      "loss": 3.0181,
      "theoretical_loss": 3.7898765986155816,
      "tokens_seen": 681312256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007923771313942,
      "loss": 2.9044,
      "theoretical_loss": 3.789839406336503,
      "tokens_seen": 681377792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040078234704112335,
      "loss": 2.8241,
      "theoretical_loss": 3.789802218635974,
      "tokens_seen": 681443328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007723169508526,
      "loss": 2.7947,
      "theoretical_loss": 3.7897650355129917,
      "tokens_seen": 681508864
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 384361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9259161949157715,
      "objective/train/theoretical_loss": 3.789727856966552,
      "objective/train/tokens_used": 702034400,
      "theoretical_loss": 3.789727856966552,
      "tokens_seen": 681574400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040076228686058177,
      "loss": 2.7609,
      "theoretical_loss": 3.789727856966552,
      "tokens_seen": 681574400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040075225677031095,
      "loss": 2.7081,
      "theoretical_loss": 3.7896906829956523,
      "tokens_seen": 681639936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040074222668004013,
      "loss": 2.9535,
      "theoretical_loss": 3.789653513599289,
      "tokens_seen": 681705472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040073219658976936,
      "loss": 2.7223,
      "theoretical_loss": 3.7896163487764594,
      "tokens_seen": 681771008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007221664994985,
      "loss": 2.8175,
      "theoretical_loss": 3.7895791885261616,
      "tokens_seen": 681836544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007121364092277,
      "loss": 2.9393,
      "theoretical_loss": 3.789542032847393,
      "tokens_seen": 681902080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040070210631895685,
      "loss": 2.6259,
      "theoretical_loss": 3.7895048817391532,
      "tokens_seen": 681967616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006920762286861,
      "loss": 2.7234,
      "theoretical_loss": 3.78946773520044,
      "tokens_seen": 682033152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040068204613841527,
      "loss": 2.882,
      "theoretical_loss": 3.789430593230252,
      "tokens_seen": 682098688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040067201604814445,
      "loss": 2.5952,
      "theoretical_loss": 3.7893934558275886,
      "tokens_seen": 682164224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040066198595787363,
      "loss": 2.914,
      "theoretical_loss": 3.78935632299145,
      "tokens_seen": 682229760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006519558676028,
      "loss": 2.7516,
      "theoretical_loss": 3.789319194720836,
      "tokens_seen": 682295296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400641925777332,
      "loss": 2.6412,
      "theoretical_loss": 3.789282071014746,
      "tokens_seen": 682360832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040063189568706123,
      "loss": 2.8659,
      "theoretical_loss": 3.789244951872181,
      "tokens_seen": 682426368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040062186559679036,
      "loss": 2.6498,
      "theoretical_loss": 3.7892078372921425,
      "tokens_seen": 682491904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006118355065196,
      "loss": 2.8014,
      "theoretical_loss": 3.789170727273631,
      "tokens_seen": 682557440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006018054162487,
      "loss": 2.7229,
      "theoretical_loss": 3.789133621815648,
      "tokens_seen": 682622976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040059177532597795,
      "loss": 2.7762,
      "theoretical_loss": 3.7890965209171954,
      "tokens_seen": 682688512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040058174523570713,
      "loss": 2.7451,
      "theoretical_loss": 3.7890594245772746,
      "tokens_seen": 682754048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005717151454363,
      "loss": 2.9726,
      "theoretical_loss": 3.7890223327948895,
      "tokens_seen": 682819584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005616850551655,
      "loss": 2.6521,
      "theoretical_loss": 3.788985245569042,
      "tokens_seen": 682885120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040055165496489473,
      "loss": 2.8672,
      "theoretical_loss": 3.788948162898735,
      "tokens_seen": 682950656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040054162487462386,
      "loss": 2.7924,
      "theoretical_loss": 3.7889110847829715,
      "tokens_seen": 683016192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005315947843531,
      "loss": 2.8288,
      "theoretical_loss": 3.788874011220756,
      "tokens_seen": 683081728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005215646940822,
      "loss": 2.8745,
      "theoretical_loss": 3.788836942211092,
      "tokens_seen": 683147264
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 385388,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.051872491836548,
      "objective/train/theoretical_loss": 3.7887998777529845,
      "objective/train/tokens_used": 703672800,
      "theoretical_loss": 3.7887998777529845,
      "tokens_seen": 683212800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040051153460381146,
      "loss": 2.8854,
      "theoretical_loss": 3.7887998777529845,
      "tokens_seen": 683212800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040050150451354064,
      "loss": 3.0064,
      "theoretical_loss": 3.7887628178454373,
      "tokens_seen": 683278336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004914744232698,
      "loss": 2.973,
      "theoretical_loss": 3.7887257624874557,
      "tokens_seen": 683343872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400481444332999,
      "loss": 2.9739,
      "theoretical_loss": 3.7886887116780446,
      "tokens_seen": 683409408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004714142427282,
      "loss": 2.628,
      "theoretical_loss": 3.78865166541621,
      "tokens_seen": 683474944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040046138415245736,
      "loss": 2.7112,
      "theoretical_loss": 3.7886146237009575,
      "tokens_seen": 683540480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004513540621866,
      "loss": 2.8486,
      "theoretical_loss": 3.788577586531294,
      "tokens_seen": 683606016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004413239719157,
      "loss": 2.8971,
      "theoretical_loss": 3.7885405539062242,
      "tokens_seen": 683671552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040043129388164496,
      "loss": 2.8651,
      "theoretical_loss": 3.7885035258247566,
      "tokens_seen": 683737088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004212637913741,
      "loss": 2.8794,
      "theoretical_loss": 3.788466502285898,
      "tokens_seen": 683802624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004112337011033,
      "loss": 2.769,
      "theoretical_loss": 3.7884294832886547,
      "tokens_seen": 683868160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004012036108325,
      "loss": 2.9099,
      "theoretical_loss": 3.788392468832036,
      "tokens_seen": 683933696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003911735205617,
      "loss": 2.7388,
      "theoretical_loss": 3.788355458915049,
      "tokens_seen": 683999232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040038114343029087,
      "loss": 3.0831,
      "theoretical_loss": 3.7883184535367027,
      "tokens_seen": 684064768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003711133400201,
      "loss": 2.6536,
      "theoretical_loss": 3.7882814526960047,
      "tokens_seen": 684130304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040036108324974923,
      "loss": 2.7046,
      "theoretical_loss": 3.7882444563919657,
      "tokens_seen": 684195840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040035105315947846,
      "loss": 2.6015,
      "theoretical_loss": 3.7882074646235933,
      "tokens_seen": 684261376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003410230692076,
      "loss": 2.7814,
      "theoretical_loss": 3.7881704773898974,
      "tokens_seen": 684326912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003309929789368,
      "loss": 2.8598,
      "theoretical_loss": 3.7881334946898892,
      "tokens_seen": 684392448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400320962888666,
      "loss": 2.7818,
      "theoretical_loss": 3.7880965165225775,
      "tokens_seen": 684457984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003109327983952,
      "loss": 2.5295,
      "theoretical_loss": 3.7880595428869737,
      "tokens_seen": 684523520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040030090270812437,
      "loss": 2.7813,
      "theoretical_loss": 3.7880225737820883,
      "tokens_seen": 684589056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040029087261785355,
      "loss": 2.7439,
      "theoretical_loss": 3.787985609206932,
      "tokens_seen": 684654592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040028084252758273,
      "loss": 2.8324,
      "theoretical_loss": 3.7879486491605174,
      "tokens_seen": 684720128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040027081243731197,
      "loss": 2.7086,
      "theoretical_loss": 3.7879116936418553,
      "tokens_seen": 684785664
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 386107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4937210083007812,
      "objective/train/theoretical_loss": 3.787874742649958,
      "objective/train/tokens_used": 705311200,
      "theoretical_loss": 3.787874742649958,
      "tokens_seen": 684851200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002607823470411,
      "loss": 2.6333,
      "theoretical_loss": 3.787874742649958,
      "tokens_seen": 684851200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040025075225677033,
      "loss": 2.9428,
      "theoretical_loss": 3.787837796183838,
      "tokens_seen": 684916736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040024072216649946,
      "loss": 2.7115,
      "theoretical_loss": 3.7878008542425086,
      "tokens_seen": 684982272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002306920762287,
      "loss": 2.7586,
      "theoretical_loss": 3.7877639168249817,
      "tokens_seen": 685047808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040022066198595787,
      "loss": 2.7163,
      "theoretical_loss": 3.7877269839302716,
      "tokens_seen": 685113344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040021063189568705,
      "loss": 2.6753,
      "theoretical_loss": 3.787690055557391,
      "tokens_seen": 685178880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040020060180541623,
      "loss": 2.8251,
      "theoretical_loss": 3.7876531317053552,
      "tokens_seen": 685244416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040019057171514547,
      "loss": 2.8668,
      "theoretical_loss": 3.7876162123731776,
      "tokens_seen": 685309952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001805416248746,
      "loss": 2.9311,
      "theoretical_loss": 3.787579297559872,
      "tokens_seen": 685375488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040017051153460383,
      "loss": 2.853,
      "theoretical_loss": 3.7875423872644545,
      "tokens_seen": 685441024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040016048144433296,
      "loss": 2.8186,
      "theoretical_loss": 3.7875054814859404,
      "tokens_seen": 685506560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001504513540622,
      "loss": 2.6879,
      "theoretical_loss": 3.787468580223344,
      "tokens_seen": 685572096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040014042126379143,
      "loss": 2.7141,
      "theoretical_loss": 3.787431683475682,
      "tokens_seen": 685637632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040013039117352056,
      "loss": 2.6609,
      "theoretical_loss": 3.7873947912419705,
      "tokens_seen": 685703168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001203610832498,
      "loss": 2.8451,
      "theoretical_loss": 3.7873579035212255,
      "tokens_seen": 685768704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001103309929789,
      "loss": 2.7973,
      "theoretical_loss": 3.787321020312464,
      "tokens_seen": 685834240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040010030090270815,
      "loss": 3.0104,
      "theoretical_loss": 3.7872841416147036,
      "tokens_seen": 685899776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040009027081243733,
      "loss": 2.9192,
      "theoretical_loss": 3.7872472674269604,
      "tokens_seen": 685965312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000802407221665,
      "loss": 2.7897,
      "theoretical_loss": 3.787210397748253,
      "tokens_seen": 686030848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000702106318957,
      "loss": 2.7675,
      "theoretical_loss": 3.7871735325775986,
      "tokens_seen": 686096384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040006018054162493,
      "loss": 2.7428,
      "theoretical_loss": 3.7871366719140163,
      "tokens_seen": 686161920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040005015045135406,
      "loss": 2.7018,
      "theoretical_loss": 3.787099815756524,
      "tokens_seen": 686227456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000401203610833,
      "loss": 2.7264,
      "theoretical_loss": 3.787062964104141,
      "tokens_seen": 686292992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000300902708124,
      "loss": 2.7911,
      "theoretical_loss": 3.7870261169558868,
      "tokens_seen": 686358528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040002006018054166,
      "loss": 2.7299,
      "theoretical_loss": 3.78698927431078,
      "tokens_seen": 686424064
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 387264,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.96657395362854,
      "objective/train/theoretical_loss": 3.786952436167841,
      "objective/train/tokens_used": 706949600,
      "theoretical_loss": 3.786952436167841,
      "tokens_seen": 686489600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040001003009027084,
      "loss": 2.6869,
      "theoretical_loss": 3.786952436167841,
      "tokens_seen": 686489600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004,
      "loss": 2.8198,
      "theoretical_loss": 3.78691560252609,
      "tokens_seen": 686555136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999899699097292,
      "loss": 2.5833,
      "theoretical_loss": 3.7868787733845473,
      "tokens_seen": 686620672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999799398194584,
      "loss": 2.6918,
      "theoretical_loss": 3.786841948742233,
      "tokens_seen": 686686208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039996990972918756,
      "loss": 2.823,
      "theoretical_loss": 3.7868051285981696,
      "tokens_seen": 686751744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999598796389168,
      "loss": 2.8408,
      "theoretical_loss": 3.7867683129513767,
      "tokens_seen": 686817280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999498495486459,
      "loss": 2.7538,
      "theoretical_loss": 3.786731501800877,
      "tokens_seen": 686882816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039993981945837516,
      "loss": 2.8252,
      "theoretical_loss": 3.7866946951456923,
      "tokens_seen": 686948352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999297893681043,
      "loss": 2.6463,
      "theoretical_loss": 3.786657892984845,
      "tokens_seen": 687013888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999197592778335,
      "loss": 2.8026,
      "theoretical_loss": 3.786621095317357,
      "tokens_seen": 687079424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999097291875627,
      "loss": 2.8123,
      "theoretical_loss": 3.7865843021422516,
      "tokens_seen": 687144960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998996990972919,
      "loss": 2.5796,
      "theoretical_loss": 3.7865475134585522,
      "tokens_seen": 687210496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039988966900702107,
      "loss": 3.0825,
      "theoretical_loss": 3.7865107292652818,
      "tokens_seen": 687276032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998796389167503,
      "loss": 2.7263,
      "theoretical_loss": 3.786473949561465,
      "tokens_seen": 687341568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039986960882647943,
      "loss": 2.6653,
      "theoretical_loss": 3.7864371743461254,
      "tokens_seen": 687407104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039985957873620866,
      "loss": 2.7328,
      "theoretical_loss": 3.786400403618287,
      "tokens_seen": 687472640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998495486459378,
      "loss": 2.5849,
      "theoretical_loss": 3.7863636373769745,
      "tokens_seen": 687538176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399839518555667,
      "loss": 2.6915,
      "theoretical_loss": 3.7863268756212136,
      "tokens_seen": 687603712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998294884653962,
      "loss": 2.8043,
      "theoretical_loss": 3.7862901183500295,
      "tokens_seen": 687669248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998194583751254,
      "loss": 2.9892,
      "theoretical_loss": 3.7862533655624473,
      "tokens_seen": 687734784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039980942828485457,
      "loss": 2.7182,
      "theoretical_loss": 3.786216617257493,
      "tokens_seen": 687800320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039979939819458375,
      "loss": 2.7629,
      "theoretical_loss": 3.786179873434193,
      "tokens_seen": 687865856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039978936810431293,
      "loss": 2.4144,
      "theoretical_loss": 3.7861431340915743,
      "tokens_seen": 687931392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039977933801404217,
      "loss": 2.7272,
      "theoretical_loss": 3.7861063992286637,
      "tokens_seen": 687996928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997693079237713,
      "loss": 2.8178,
      "theoretical_loss": 3.786069668844487,
      "tokens_seen": 688062464
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 387908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.436974287033081,
      "objective/train/theoretical_loss": 3.786032942938073,
      "objective/train/tokens_used": 708588000,
      "theoretical_loss": 3.786032942938073,
      "tokens_seen": 688128000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039975927783350053,
      "loss": 2.6594,
      "theoretical_loss": 3.786032942938073,
      "tokens_seen": 688128000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039974924774322966,
      "loss": 2.7045,
      "theoretical_loss": 3.785996221508449,
      "tokens_seen": 688193536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997392176529589,
      "loss": 2.6022,
      "theoretical_loss": 3.7859595045546426,
      "tokens_seen": 688259072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039972918756268807,
      "loss": 2.8448,
      "theoretical_loss": 3.7859227920756835,
      "tokens_seen": 688324608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039971915747241725,
      "loss": 2.8839,
      "theoretical_loss": 3.785886084070599,
      "tokens_seen": 688390144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039970912738214643,
      "loss": 2.8146,
      "theoretical_loss": 3.7858493805384184,
      "tokens_seen": 688455680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039969909729187567,
      "loss": 2.8004,
      "theoretical_loss": 3.7858126814781707,
      "tokens_seen": 688521216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996890672016048,
      "loss": 2.7867,
      "theoretical_loss": 3.7857759868888863,
      "tokens_seen": 688586752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039967903711133403,
      "loss": 3.0736,
      "theoretical_loss": 3.785739296769595,
      "tokens_seen": 688652288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039966900702106316,
      "loss": 2.7904,
      "theoretical_loss": 3.7857026111193264,
      "tokens_seen": 688717824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996589769307924,
      "loss": 2.7473,
      "theoretical_loss": 3.785665929937111,
      "tokens_seen": 688783360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996489468405216,
      "loss": 2.8171,
      "theoretical_loss": 3.7856292532219795,
      "tokens_seen": 688848896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039963891675025076,
      "loss": 2.9703,
      "theoretical_loss": 3.7855925809729634,
      "tokens_seen": 688914432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039962888665997994,
      "loss": 2.3776,
      "theoretical_loss": 3.7855559131890946,
      "tokens_seen": 688979968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996188565697091,
      "loss": 2.8586,
      "theoretical_loss": 3.7855192498694032,
      "tokens_seen": 689045504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996088264794383,
      "loss": 2.7614,
      "theoretical_loss": 3.785482591012922,
      "tokens_seen": 689111040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039959879638916754,
      "loss": 2.5907,
      "theoretical_loss": 3.785445936618684,
      "tokens_seen": 689176576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039958876629889666,
      "loss": 2.6694,
      "theoretical_loss": 3.785409286685721,
      "tokens_seen": 689242112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995787362086259,
      "loss": 2.7639,
      "theoretical_loss": 3.785372641213066,
      "tokens_seen": 689307648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995687061183551,
      "loss": 2.7779,
      "theoretical_loss": 3.7853360001997522,
      "tokens_seen": 689373184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039955867602808426,
      "loss": 2.5631,
      "theoretical_loss": 3.7852993636448127,
      "tokens_seen": 689438720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039954864593781344,
      "loss": 2.7811,
      "theoretical_loss": 3.7852627315472818,
      "tokens_seen": 689504256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995386158475426,
      "loss": 2.7612,
      "theoretical_loss": 3.785226103906194,
      "tokens_seen": 689569792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995285857572718,
      "loss": 2.7939,
      "theoretical_loss": 3.7851894807205833,
      "tokens_seen": 689635328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039951855566700104,
      "loss": 2.8221,
      "theoretical_loss": 3.7851528619894843,
      "tokens_seen": 689700864
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 389134,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.952434778213501,
      "objective/train/theoretical_loss": 3.7851162477119313,
      "objective/train/tokens_used": 710226400,
      "theoretical_loss": 3.7851162477119313,
      "tokens_seen": 689766400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039950852557673017,
      "loss": 2.9159,
      "theoretical_loss": 3.7851162477119313,
      "tokens_seen": 689766400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994984954864594,
      "loss": 2.7995,
      "theoretical_loss": 3.7850796378869607,
      "tokens_seen": 689831936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039948846539618853,
      "loss": 2.595,
      "theoretical_loss": 3.7850430325136077,
      "tokens_seen": 689897472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039947843530591776,
      "loss": 2.771,
      "theoretical_loss": 3.7850064315909084,
      "tokens_seen": 689963008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039946840521564694,
      "loss": 2.78,
      "theoretical_loss": 3.7849698351178986,
      "tokens_seen": 690028544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994583751253761,
      "loss": 2.7472,
      "theoretical_loss": 3.784933243093615,
      "tokens_seen": 690094080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994483450351053,
      "loss": 2.8655,
      "theoretical_loss": 3.7848966555170946,
      "tokens_seen": 690159616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994383149448345,
      "loss": 2.6616,
      "theoretical_loss": 3.7848600723873744,
      "tokens_seen": 690225152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039942828485456367,
      "loss": 2.6761,
      "theoretical_loss": 3.7848234937034917,
      "tokens_seen": 690290688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994182547642929,
      "loss": 2.7662,
      "theoretical_loss": 3.784786919464484,
      "tokens_seen": 690356224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039940822467402203,
      "loss": 2.9131,
      "theoretical_loss": 3.7847503496693893,
      "tokens_seen": 690421760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039939819458375127,
      "loss": 2.8552,
      "theoretical_loss": 3.7847137843172467,
      "tokens_seen": 690487296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993881644934805,
      "loss": 2.823,
      "theoretical_loss": 3.784677223407094,
      "tokens_seen": 690552832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039937813440320963,
      "loss": 2.7278,
      "theoretical_loss": 3.78464066693797,
      "tokens_seen": 690618368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039936810431293886,
      "loss": 3.1157,
      "theoretical_loss": 3.7846041149089142,
      "tokens_seen": 690683904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399358074222668,
      "loss": 2.9657,
      "theoretical_loss": 3.784567567318967,
      "tokens_seen": 690749440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993480441323972,
      "loss": 2.7648,
      "theoretical_loss": 3.7845310241671664,
      "tokens_seen": 690814976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993380140421264,
      "loss": 2.6964,
      "theoretical_loss": 3.7844944854525537,
      "tokens_seen": 690880512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993279839518556,
      "loss": 2.8395,
      "theoretical_loss": 3.7844579511741694,
      "tokens_seen": 690946048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039931795386158477,
      "loss": 2.7431,
      "theoretical_loss": 3.784421421331053,
      "tokens_seen": 691011584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039930792377131395,
      "loss": 2.8044,
      "theoretical_loss": 3.784384895922247,
      "tokens_seen": 691077120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039929789368104313,
      "loss": 2.6512,
      "theoretical_loss": 3.7843483749467914,
      "tokens_seen": 691142656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039928786359077237,
      "loss": 2.7691,
      "theoretical_loss": 3.7843118584037287,
      "tokens_seen": 691208192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992778335005015,
      "loss": 2.7728,
      "theoretical_loss": 3.7842753462921004,
      "tokens_seen": 691273728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039926780341023073,
      "loss": 2.6623,
      "theoretical_loss": 3.784238838610949,
      "tokens_seen": 691339264
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 389824,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.667067050933838,
      "objective/train/theoretical_loss": 3.784202335359316,
      "objective/train/tokens_used": 711864800,
      "theoretical_loss": 3.784202335359316,
      "tokens_seen": 691404800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039925777331995986,
      "loss": 2.8217,
      "theoretical_loss": 3.784202335359316,
      "tokens_seen": 691404800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992477432296891,
      "loss": 2.6805,
      "theoretical_loss": 3.7841658365362454,
      "tokens_seen": 691470336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039923771313941827,
      "loss": 2.8809,
      "theoretical_loss": 3.78412934214078,
      "tokens_seen": 691535872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039922768304914745,
      "loss": 2.7089,
      "theoretical_loss": 3.784092852171962,
      "tokens_seen": 691601408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039921765295887663,
      "loss": 2.7421,
      "theoretical_loss": 3.784056366628837,
      "tokens_seen": 691666944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039920762286860587,
      "loss": 2.7233,
      "theoretical_loss": 3.7840198855104474,
      "tokens_seen": 691732480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399197592778335,
      "loss": 2.5454,
      "theoretical_loss": 3.783983408815838,
      "tokens_seen": 691798016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039918756268806423,
      "loss": 2.7803,
      "theoretical_loss": 3.7839469365440537,
      "tokens_seen": 691863552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039917753259779336,
      "loss": 2.9658,
      "theoretical_loss": 3.783910468694139,
      "tokens_seen": 691929088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991675025075226,
      "loss": 2.8263,
      "theoretical_loss": 3.7838740052651394,
      "tokens_seen": 691994624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991574724172518,
      "loss": 2.8775,
      "theoretical_loss": 3.783837546256099,
      "tokens_seen": 692060160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039914744232698096,
      "loss": 2.5907,
      "theoretical_loss": 3.7838010916660654,
      "tokens_seen": 692125696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039913741223671014,
      "loss": 2.9517,
      "theoretical_loss": 3.783764641494084,
      "tokens_seen": 692191232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991273821464393,
      "loss": 2.7124,
      "theoretical_loss": 3.783728195739201,
      "tokens_seen": 692256768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991173520561685,
      "loss": 2.6897,
      "theoretical_loss": 3.7836917544004627,
      "tokens_seen": 692322304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039910732196589774,
      "loss": 2.8147,
      "theoretical_loss": 3.7836553174769163,
      "tokens_seen": 692387840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039909729187562686,
      "loss": 2.8122,
      "theoretical_loss": 3.783618884967609,
      "tokens_seen": 692453376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990872617853561,
      "loss": 2.7954,
      "theoretical_loss": 3.783582456871589,
      "tokens_seen": 692518912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990772316950853,
      "loss": 2.767,
      "theoretical_loss": 3.7835460331879034,
      "tokens_seen": 692584448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039906720160481446,
      "loss": 2.5768,
      "theoretical_loss": 3.7835096139155997,
      "tokens_seen": 692649984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039905717151454364,
      "loss": 3.0041,
      "theoretical_loss": 3.783473199053728,
      "tokens_seen": 692715520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990471414242728,
      "loss": 2.7621,
      "theoretical_loss": 3.7834367886013354,
      "tokens_seen": 692781056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399037111334002,
      "loss": 2.6121,
      "theoretical_loss": 3.7834003825574722,
      "tokens_seen": 692846592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039902708124373124,
      "loss": 2.6698,
      "theoretical_loss": 3.7833639809211865,
      "tokens_seen": 692912128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039901705115346037,
      "loss": 2.7172,
      "theoretical_loss": 3.7833275836915288,
      "tokens_seen": 692977664
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 391100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.815354585647583,
      "objective/train/theoretical_loss": 3.7832911908675486,
      "objective/train/tokens_used": 713503200,
      "theoretical_loss": 3.7832911908675486,
      "tokens_seen": 693043200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990070210631896,
      "loss": 2.8474,
      "theoretical_loss": 3.7832911908675486,
      "tokens_seen": 693043200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039899699097291873,
      "loss": 2.557,
      "theoretical_loss": 3.7832548024482957,
      "tokens_seen": 693108736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039898696088264796,
      "loss": 2.4426,
      "theoretical_loss": 3.783218418432822,
      "tokens_seen": 693174272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039897693079237714,
      "loss": 3.0426,
      "theoretical_loss": 3.7831820388201765,
      "tokens_seen": 693239808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989669007021063,
      "loss": 2.754,
      "theoretical_loss": 3.783145663609411,
      "tokens_seen": 693305344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989568706118355,
      "loss": 2.7348,
      "theoretical_loss": 3.783109292799577,
      "tokens_seen": 693370880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989468405215647,
      "loss": 2.7635,
      "theoretical_loss": 3.7830729263897265,
      "tokens_seen": 693436416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039893681043129387,
      "loss": 2.5699,
      "theoretical_loss": 3.783036564378911,
      "tokens_seen": 693501952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989267803410231,
      "loss": 2.8455,
      "theoretical_loss": 3.7830002067661814,
      "tokens_seen": 693567488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039891675025075223,
      "loss": 2.8716,
      "theoretical_loss": 3.782963853550593,
      "tokens_seen": 693633024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039890672016048147,
      "loss": 2.7498,
      "theoretical_loss": 3.7829275047311968,
      "tokens_seen": 693698560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039889669007021065,
      "loss": 2.733,
      "theoretical_loss": 3.782891160307046,
      "tokens_seen": 693764096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039888665997993983,
      "loss": 2.7969,
      "theoretical_loss": 3.782854820277195,
      "tokens_seen": 693829632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398876629889669,
      "loss": 2.7611,
      "theoretical_loss": 3.7828184846406963,
      "tokens_seen": 693895168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988665997993982,
      "loss": 2.7715,
      "theoretical_loss": 3.7827821533966044,
      "tokens_seen": 693960704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039885656970912737,
      "loss": 2.778,
      "theoretical_loss": 3.782745826543974,
      "tokens_seen": 694026240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988465396188566,
      "loss": 2.9197,
      "theoretical_loss": 3.7827095040818586,
      "tokens_seen": 694091776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039883650952858573,
      "loss": 2.6453,
      "theoretical_loss": 3.7826731860093146,
      "tokens_seen": 694157312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039882647943831497,
      "loss": 2.7944,
      "theoretical_loss": 3.782636872325396,
      "tokens_seen": 694222848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988164493480441,
      "loss": 2.958,
      "theoretical_loss": 3.782600563029159,
      "tokens_seen": 694288384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039880641925777333,
      "loss": 2.7868,
      "theoretical_loss": 3.782564258119659,
      "tokens_seen": 694353920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987963891675025,
      "loss": 2.7221,
      "theoretical_loss": 3.7825279575959514,
      "tokens_seen": 694419456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987863590772317,
      "loss": 2.7764,
      "theoretical_loss": 3.782491661457094,
      "tokens_seen": 694484992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987763289869609,
      "loss": 2.6213,
      "theoretical_loss": 3.7824553697021415,
      "tokens_seen": 694550528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039876629889669006,
      "loss": 2.5601,
      "theoretical_loss": 3.782419082330153,
      "tokens_seen": 694616064
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 392291,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8680379390716553,
      "objective/train/theoretical_loss": 3.7823827993401844,
      "objective/train/tokens_used": 715141600,
      "theoretical_loss": 3.7823827993401844,
      "tokens_seen": 694681600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039875626880641924,
      "loss": 2.7737,
      "theoretical_loss": 3.7823827993401844,
      "tokens_seen": 694681600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039874623871614847,
      "loss": 2.7944,
      "theoretical_loss": 3.782346520731293,
      "tokens_seen": 694747136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987362086258776,
      "loss": 2.8212,
      "theoretical_loss": 3.7823102465025373,
      "tokens_seen": 694812672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039872617853560683,
      "loss": 2.728,
      "theoretical_loss": 3.7822739766529754,
      "tokens_seen": 694878208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398716148445336,
      "loss": 2.824,
      "theoretical_loss": 3.7822377111816654,
      "tokens_seen": 694943744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987061183550652,
      "loss": 2.633,
      "theoretical_loss": 3.7822014500876664,
      "tokens_seen": 695009280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986960882647944,
      "loss": 2.7526,
      "theoretical_loss": 3.7821651933700364,
      "tokens_seen": 695074816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039868605817452356,
      "loss": 2.6716,
      "theoretical_loss": 3.7821289410278354,
      "tokens_seen": 695140352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039867602808425274,
      "loss": 2.8782,
      "theoretical_loss": 3.7820926930601226,
      "tokens_seen": 695205888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398665997993982,
      "loss": 2.7649,
      "theoretical_loss": 3.7820564494659585,
      "tokens_seen": 695271424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986559679037111,
      "loss": 2.6169,
      "theoretical_loss": 3.7820202102444025,
      "tokens_seen": 695336960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039864593781344034,
      "loss": 2.7008,
      "theoretical_loss": 3.7819839753945153,
      "tokens_seen": 695402496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986359077231695,
      "loss": 2.9026,
      "theoretical_loss": 3.7819477449153576,
      "tokens_seen": 695468032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986258776328987,
      "loss": 2.7479,
      "theoretical_loss": 3.78191151880599,
      "tokens_seen": 695533568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039861584754262794,
      "loss": 2.8503,
      "theoretical_loss": 3.7818752970654748,
      "tokens_seen": 695599104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039860581745235706,
      "loss": 2.7765,
      "theoretical_loss": 3.7818390796928725,
      "tokens_seen": 695664640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985957873620863,
      "loss": 2.7862,
      "theoretical_loss": 3.781802866687246,
      "tokens_seen": 695730176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985857572718155,
      "loss": 2.9017,
      "theoretical_loss": 3.7817666580476565,
      "tokens_seen": 695795712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039857572718154466,
      "loss": 2.6179,
      "theoretical_loss": 3.781730453773167,
      "tokens_seen": 695861248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039856569709127384,
      "loss": 2.6699,
      "theoretical_loss": 3.78169425386284,
      "tokens_seen": 695926784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398555667001003,
      "loss": 2.706,
      "theoretical_loss": 3.781658058315738,
      "tokens_seen": 695992320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985456369107322,
      "loss": 2.9331,
      "theoretical_loss": 3.7816218671309256,
      "tokens_seen": 696057856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039853560682046144,
      "loss": 2.619,
      "theoretical_loss": 3.781585680307466,
      "tokens_seen": 696123392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039852557673019057,
      "loss": 2.5891,
      "theoretical_loss": 3.7815494978444226,
      "tokens_seen": 696188928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985155466399198,
      "loss": 2.785,
      "theoretical_loss": 3.78151331974086,
      "tokens_seen": 696254464
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 392983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6465673446655273,
      "objective/train/theoretical_loss": 3.781477145995842,
      "objective/train/tokens_used": 716780000,
      "theoretical_loss": 3.781477145995842,
      "tokens_seen": 696320000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039850551654964893,
      "loss": 2.5574,
      "theoretical_loss": 3.781477145995842,
      "tokens_seen": 696320000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039849548645937816,
      "loss": 2.7248,
      "theoretical_loss": 3.781440976608434,
      "tokens_seen": 696385536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039848545636910734,
      "loss": 2.9657,
      "theoretical_loss": 3.7814048115777013,
      "tokens_seen": 696451072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984754262788365,
      "loss": 2.9466,
      "theoretical_loss": 3.7813686509027082,
      "tokens_seen": 696516608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984653961885657,
      "loss": 2.8556,
      "theoretical_loss": 3.7813324945825215,
      "tokens_seen": 696582144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984553660982949,
      "loss": 2.7166,
      "theoretical_loss": 3.7812963426162067,
      "tokens_seen": 696647680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039844533600802407,
      "loss": 2.8797,
      "theoretical_loss": 3.7812601950028295,
      "tokens_seen": 696713216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984353059177533,
      "loss": 2.6726,
      "theoretical_loss": 3.7812240517414573,
      "tokens_seen": 696778752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039842527582748243,
      "loss": 2.865,
      "theoretical_loss": 3.7811879128311556,
      "tokens_seen": 696844288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039841524573721167,
      "loss": 2.7871,
      "theoretical_loss": 3.781151778270993,
      "tokens_seen": 696909824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039840521564694085,
      "loss": 2.7724,
      "theoretical_loss": 3.7811156480600356,
      "tokens_seen": 696975360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039839518555667003,
      "loss": 2.6619,
      "theoretical_loss": 3.781079522197352,
      "tokens_seen": 697040896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983851554663992,
      "loss": 2.6867,
      "theoretical_loss": 3.7810434006820097,
      "tokens_seen": 697106432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983751253761284,
      "loss": 2.7643,
      "theoretical_loss": 3.7810072835130764,
      "tokens_seen": 697171968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039836509528585757,
      "loss": 2.6857,
      "theoretical_loss": 3.7809711706896216,
      "tokens_seen": 697237504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983550651955868,
      "loss": 2.7306,
      "theoretical_loss": 3.7809350622107134,
      "tokens_seen": 697303040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039834503510531593,
      "loss": 2.6711,
      "theoretical_loss": 3.7808989580754213,
      "tokens_seen": 697368576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039833500501504517,
      "loss": 2.8561,
      "theoretical_loss": 3.780862858282814,
      "tokens_seen": 697434112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983249749247743,
      "loss": 2.7013,
      "theoretical_loss": 3.780826762831962,
      "tokens_seen": 697499648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039831494483450353,
      "loss": 2.7573,
      "theoretical_loss": 3.7807906717219346,
      "tokens_seen": 697565184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983049147442327,
      "loss": 2.8347,
      "theoretical_loss": 3.7807545849518025,
      "tokens_seen": 697630720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982948846539619,
      "loss": 2.6078,
      "theoretical_loss": 3.780718502520636,
      "tokens_seen": 697696256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982848545636911,
      "loss": 2.6766,
      "theoretical_loss": 3.780682424427506,
      "tokens_seen": 697761792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039827482447342026,
      "loss": 2.6125,
      "theoretical_loss": 3.780646350671484,
      "tokens_seen": 697827328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039826479438314944,
      "loss": 2.8044,
      "theoretical_loss": 3.78061028125164,
      "tokens_seen": 697892864
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 394024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.852461338043213,
      "objective/train/theoretical_loss": 3.7805742161670466,
      "objective/train/tokens_used": 718418400,
      "theoretical_loss": 3.7805742161670466,
      "tokens_seen": 697958400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039825476429287867,
      "loss": 2.8367,
      "theoretical_loss": 3.7805742161670466,
      "tokens_seen": 697958400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982447342026078,
      "loss": 2.7577,
      "theoretical_loss": 3.7805381554167763,
      "tokens_seen": 698023936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039823470411233703,
      "loss": 2.5646,
      "theoretical_loss": 3.7805020989999,
      "tokens_seen": 698089472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982246740220662,
      "loss": 2.7665,
      "theoretical_loss": 3.7804660469154907,
      "tokens_seen": 698155008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982146439317954,
      "loss": 2.8406,
      "theoretical_loss": 3.7804299991626222,
      "tokens_seen": 698220544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982046138415246,
      "loss": 2.6331,
      "theoretical_loss": 3.780393955740367,
      "tokens_seen": 698286080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039819458375125376,
      "loss": 2.8337,
      "theoretical_loss": 3.7803579166477976,
      "tokens_seen": 698351616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039818455366098294,
      "loss": 2.6492,
      "theoretical_loss": 3.7803218818839883,
      "tokens_seen": 698417152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981745235707122,
      "loss": 2.7657,
      "theoretical_loss": 3.7802858514480135,
      "tokens_seen": 698482688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981644934804413,
      "loss": 2.8308,
      "theoretical_loss": 3.780249825338947,
      "tokens_seen": 698548224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039815446339017054,
      "loss": 2.6012,
      "theoretical_loss": 3.7802138035558634,
      "tokens_seen": 698613760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039814443329989966,
      "loss": 2.8382,
      "theoretical_loss": 3.7801777860978367,
      "tokens_seen": 698679296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981344032096289,
      "loss": 2.8067,
      "theoretical_loss": 3.780141772963944,
      "tokens_seen": 698744832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981243731193581,
      "loss": 2.8721,
      "theoretical_loss": 3.7801057641532587,
      "tokens_seen": 698810368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039811434302908726,
      "loss": 2.538,
      "theoretical_loss": 3.7800697596648565,
      "tokens_seen": 698875904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039810431293881644,
      "loss": 2.9022,
      "theoretical_loss": 3.7800337594978153,
      "tokens_seen": 698941440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980942828485457,
      "loss": 2.7558,
      "theoretical_loss": 3.7799977636512088,
      "tokens_seen": 699006976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980842527582748,
      "loss": 2.975,
      "theoretical_loss": 3.779961772124115,
      "tokens_seen": 699072512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039807422266800404,
      "loss": 2.7963,
      "theoretical_loss": 3.7799257849156107,
      "tokens_seen": 699138048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039806419257773317,
      "loss": 2.838,
      "theoretical_loss": 3.7798898020247726,
      "tokens_seen": 699203584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980541624874624,
      "loss": 2.5871,
      "theoretical_loss": 3.779853823450678,
      "tokens_seen": 699269120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980441323971916,
      "loss": 2.8418,
      "theoretical_loss": 3.779817849192405,
      "tokens_seen": 699334656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039803410230692077,
      "loss": 2.5972,
      "theoretical_loss": 3.7797818792490303,
      "tokens_seen": 699400192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039802407221664995,
      "loss": 2.8504,
      "theoretical_loss": 3.7797459136196334,
      "tokens_seen": 699465728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039801404212637913,
      "loss": 2.5989,
      "theoretical_loss": 3.7797099523032927,
      "tokens_seen": 699531264
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 394602,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6187872886657715,
      "objective/train/theoretical_loss": 3.7796739952990865,
      "objective/train/tokens_used": 720056800,
      "theoretical_loss": 3.7796739952990865,
      "tokens_seen": 699596800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980040120361083,
      "loss": 2.5868,
      "theoretical_loss": 3.7796739952990865,
      "tokens_seen": 699596800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039799398194583754,
      "loss": 2.6251,
      "theoretical_loss": 3.779638042606094,
      "tokens_seen": 699662336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039798395185556667,
      "loss": 2.7199,
      "theoretical_loss": 3.779602094223394,
      "tokens_seen": 699727872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979739217652959,
      "loss": 2.6618,
      "theoretical_loss": 3.7795661501500675,
      "tokens_seen": 699793408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039796389167502503,
      "loss": 2.7668,
      "theoretical_loss": 3.779530210385193,
      "tokens_seen": 699858944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039795386158475427,
      "loss": 2.4999,
      "theoretical_loss": 3.7794942749278517,
      "tokens_seen": 699924480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039794383149448345,
      "loss": 2.9616,
      "theoretical_loss": 3.7794583437771236,
      "tokens_seen": 699990016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039793380140421263,
      "loss": 2.9088,
      "theoretical_loss": 3.779422416932089,
      "tokens_seen": 700055552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979237713139418,
      "loss": 2.8126,
      "theoretical_loss": 3.77938649439183,
      "tokens_seen": 700121088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039791374122367105,
      "loss": 2.6539,
      "theoretical_loss": 3.779350576155427,
      "tokens_seen": 700186624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979037111334002,
      "loss": 2.8109,
      "theoretical_loss": 3.779314662221962,
      "tokens_seen": 700252160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978936810431294,
      "loss": 2.6834,
      "theoretical_loss": 3.7792787525905167,
      "tokens_seen": 700317696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978836509528586,
      "loss": 2.8797,
      "theoretical_loss": 3.779242847260173,
      "tokens_seen": 700383232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039787362086258777,
      "loss": 2.7285,
      "theoretical_loss": 3.7792069462300146,
      "tokens_seen": 700448768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397863590772317,
      "loss": 2.6344,
      "theoretical_loss": 3.779171049499123,
      "tokens_seen": 700514304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039785356068204613,
      "loss": 2.7949,
      "theoretical_loss": 3.779135157066581,
      "tokens_seen": 700579840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039784353059177537,
      "loss": 2.7319,
      "theoretical_loss": 3.779099268931473,
      "tokens_seen": 700645376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978335005015045,
      "loss": 2.836,
      "theoretical_loss": 3.779063385092882,
      "tokens_seen": 700710912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039782347041123373,
      "loss": 2.7745,
      "theoretical_loss": 3.779027505549891,
      "tokens_seen": 700776448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978134403209629,
      "loss": 2.8877,
      "theoretical_loss": 3.778991630301586,
      "tokens_seen": 700841984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978034102306921,
      "loss": 2.7129,
      "theoretical_loss": 3.77895575934705,
      "tokens_seen": 700907520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977933801404213,
      "loss": 2.9092,
      "theoretical_loss": 3.7789198926853684,
      "tokens_seen": 700973056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039778335005015046,
      "loss": 2.7402,
      "theoretical_loss": 3.7788840303156253,
      "tokens_seen": 701038592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039777331995987964,
      "loss": 2.8797,
      "theoretical_loss": 3.7788481722369065,
      "tokens_seen": 701104128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039776328986960887,
      "loss": 2.8667,
      "theoretical_loss": 3.7788123184482982,
      "tokens_seen": 701169664
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 395841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9108428955078125,
      "objective/train/theoretical_loss": 3.7787764689488847,
      "objective/train/tokens_used": 721695200,
      "theoretical_loss": 3.7787764689488847,
      "tokens_seen": 701235200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397753259779338,
      "loss": 2.8756,
      "theoretical_loss": 3.7787764689488847,
      "tokens_seen": 701235200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039774322968906723,
      "loss": 2.907,
      "theoretical_loss": 3.778740623737754,
      "tokens_seen": 701300736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977331995987964,
      "loss": 2.5534,
      "theoretical_loss": 3.778704782813991,
      "tokens_seen": 701366272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977231695085256,
      "loss": 2.6829,
      "theoretical_loss": 3.7786689461766825,
      "tokens_seen": 701431808
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977131394182548,
      "loss": 2.5951,
      "theoretical_loss": 3.778633113824916,
      "tokens_seen": 701497344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039770310932798396,
      "loss": 2.9051,
      "theoretical_loss": 3.778597285757778,
      "tokens_seen": 701562880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039769307923771314,
      "loss": 2.7553,
      "theoretical_loss": 3.7785614619743573,
      "tokens_seen": 701628416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976830491474424,
      "loss": 2.5643,
      "theoretical_loss": 3.77852564247374,
      "tokens_seen": 701693952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976730190571715,
      "loss": 2.9458,
      "theoretical_loss": 3.778489827255015,
      "tokens_seen": 701759488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039766298896690074,
      "loss": 2.8461,
      "theoretical_loss": 3.778454016317271,
      "tokens_seen": 701825024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039765295887662987,
      "loss": 2.7984,
      "theoretical_loss": 3.7784182096595957,
      "tokens_seen": 701890560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976429287863591,
      "loss": 2.8552,
      "theoretical_loss": 3.7783824072810788,
      "tokens_seen": 701956096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976328986960883,
      "loss": 2.9651,
      "theoretical_loss": 3.7783466091808093,
      "tokens_seen": 702021632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039762286860581746,
      "loss": 2.8818,
      "theoretical_loss": 3.7783108153578766,
      "tokens_seen": 702087168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039761283851554664,
      "loss": 2.8327,
      "theoretical_loss": 3.7782750258113698,
      "tokens_seen": 702152704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976028084252759,
      "loss": 2.5367,
      "theoretical_loss": 3.7782392405403797,
      "tokens_seen": 702218240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397592778335005,
      "loss": 2.8229,
      "theoretical_loss": 3.7782034595439966,
      "tokens_seen": 702283776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039758274824473424,
      "loss": 2.5208,
      "theoretical_loss": 3.778167682821311,
      "tokens_seen": 702349312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039757271815446337,
      "loss": 2.695,
      "theoretical_loss": 3.778131910371413,
      "tokens_seen": 702414848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975626880641926,
      "loss": 2.6583,
      "theoretical_loss": 3.7780961421933945,
      "tokens_seen": 702480384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975526579739218,
      "loss": 2.774,
      "theoretical_loss": 3.7780603782863467,
      "tokens_seen": 702545920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039754262788365097,
      "loss": 3.0157,
      "theoretical_loss": 3.778024618649361,
      "tokens_seen": 702611456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039753259779338015,
      "loss": 2.8291,
      "theoretical_loss": 3.7779888632815295,
      "tokens_seen": 702676992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039752256770310933,
      "loss": 2.859,
      "theoretical_loss": 3.7779531121819447,
      "tokens_seen": 702742528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975125376128385,
      "loss": 2.8686,
      "theoretical_loss": 3.7779173653496994,
      "tokens_seen": 702808064
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 397024,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.391489028930664,
      "objective/train/theoretical_loss": 3.7778816227838856,
      "objective/train/tokens_used": 723333600,
      "theoretical_loss": 3.7778816227838856,
      "tokens_seen": 702873600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039750250752256774,
      "loss": 2.806,
      "theoretical_loss": 3.7778816227838856,
      "tokens_seen": 702873600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039749247743229687,
      "loss": 2.8732,
      "theoretical_loss": 3.7778458844835963,
      "tokens_seen": 702939136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974824473420261,
      "loss": 2.7415,
      "theoretical_loss": 3.777810150447926,
      "tokens_seen": 703004672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039747241725175523,
      "loss": 2.728,
      "theoretical_loss": 3.777774420675967,
      "tokens_seen": 703070208
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039746238716148447,
      "loss": 2.6557,
      "theoretical_loss": 3.7777386951668133,
      "tokens_seen": 703135744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039745235707121365,
      "loss": 2.6884,
      "theoretical_loss": 3.7777029739195602,
      "tokens_seen": 703201280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039744232698094283,
      "loss": 2.8055,
      "theoretical_loss": 3.7776672569333014,
      "tokens_seen": 703266816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397432296890672,
      "loss": 2.6844,
      "theoretical_loss": 3.7776315442071313,
      "tokens_seen": 703332352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039742226680040125,
      "loss": 2.7397,
      "theoretical_loss": 3.7775958357401453,
      "tokens_seen": 703397888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974122367101304,
      "loss": 2.8919,
      "theoretical_loss": 3.7775601315314393,
      "tokens_seen": 703463424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974022066198596,
      "loss": 2.5523,
      "theoretical_loss": 3.7775244315801078,
      "tokens_seen": 703528960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039739217652958874,
      "loss": 2.6017,
      "theoretical_loss": 3.777488735885247,
      "tokens_seen": 703594496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039738214643931797,
      "loss": 2.8401,
      "theoretical_loss": 3.777453044445953,
      "tokens_seen": 703660032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039737211634904715,
      "loss": 2.7133,
      "theoretical_loss": 3.7774173572613226,
      "tokens_seen": 703725568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039736208625877633,
      "loss": 2.6392,
      "theoretical_loss": 3.777381674330452,
      "tokens_seen": 703791104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973520561685055,
      "loss": 2.8618,
      "theoretical_loss": 3.7773459956524382,
      "tokens_seen": 703856640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973420260782347,
      "loss": 2.7136,
      "theoretical_loss": 3.7773103212263783,
      "tokens_seen": 703922176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973319959879639,
      "loss": 2.7602,
      "theoretical_loss": 3.7772746510513704,
      "tokens_seen": 703987712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973219658976931,
      "loss": 2.7088,
      "theoretical_loss": 3.7772389851265116,
      "tokens_seen": 704053248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039731193580742224,
      "loss": 2.6738,
      "theoretical_loss": 3.7772033234509,
      "tokens_seen": 704118784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973019057171515,
      "loss": 2.5981,
      "theoretical_loss": 3.777167666023634,
      "tokens_seen": 704184320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972918756268806,
      "loss": 2.9311,
      "theoretical_loss": 3.777132012843812,
      "tokens_seen": 704249856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039728184553660984,
      "loss": 2.7248,
      "theoretical_loss": 3.777096363910534,
      "tokens_seen": 704315392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397271815446339,
      "loss": 2.8256,
      "theoretical_loss": 3.7770607192228978,
      "tokens_seen": 704380928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972617853560682,
      "loss": 2.8975,
      "theoretical_loss": 3.777025078780003,
      "tokens_seen": 704446464
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 397723,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0135793685913086,
      "objective/train/theoretical_loss": 3.77698944258095,
      "objective/train/tokens_used": 724972000,
      "theoretical_loss": 3.77698944258095,
      "tokens_seen": 704512000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972517552657974,
      "loss": 2.7687,
      "theoretical_loss": 3.77698944258095,
      "tokens_seen": 704512000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972417251755266,
      "loss": 2.8214,
      "theoretical_loss": 3.776953810624838,
      "tokens_seen": 704577536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039723169508525574,
      "loss": 2.4458,
      "theoretical_loss": 3.7769181829107676,
      "tokens_seen": 704643072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397221664994985,
      "loss": 3.0101,
      "theoretical_loss": 3.7768825594378397,
      "tokens_seen": 704708608
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972116349047141,
      "loss": 2.8727,
      "theoretical_loss": 3.7768469402051545,
      "tokens_seen": 704774144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039720160481444334,
      "loss": 2.7502,
      "theoretical_loss": 3.7768113252118134,
      "tokens_seen": 704839680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971915747241725,
      "loss": 2.8048,
      "theoretical_loss": 3.7767757144569174,
      "tokens_seen": 704905216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971815446339017,
      "loss": 2.6821,
      "theoretical_loss": 3.7767401079395677,
      "tokens_seen": 704970752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971715145436309,
      "loss": 2.7107,
      "theoretical_loss": 3.7767045056588677,
      "tokens_seen": 705036288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039716148445336007,
      "loss": 2.9334,
      "theoretical_loss": 3.776668907613918,
      "tokens_seen": 705101824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039715145436308925,
      "loss": 2.7057,
      "theoretical_loss": 3.7766333138038224,
      "tokens_seen": 705167360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971414242728185,
      "loss": 2.6521,
      "theoretical_loss": 3.776597724227682,
      "tokens_seen": 705232896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039713139418254766,
      "loss": 2.5163,
      "theoretical_loss": 3.776562138884601,
      "tokens_seen": 705298432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039712136409227684,
      "loss": 2.9374,
      "theoretical_loss": 3.7765265577736824,
      "tokens_seen": 705363968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971113340020061,
      "loss": 2.6889,
      "theoretical_loss": 3.7764909808940295,
      "tokens_seen": 705429504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971013039117352,
      "loss": 2.7154,
      "theoretical_loss": 3.776455408244746,
      "tokens_seen": 705495040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039709127382146444,
      "loss": 2.7799,
      "theoretical_loss": 3.7764198398249365,
      "tokens_seen": 705560576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039708124373119357,
      "loss": 2.8133,
      "theoretical_loss": 3.7763842756337054,
      "tokens_seen": 705626112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003970712136409228,
      "loss": 2.6184,
      "theoretical_loss": 3.776348715670156,
      "tokens_seen": 705691648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397061183550652,
      "loss": 2.6634,
      "theoretical_loss": 3.7763131599333946,
      "tokens_seen": 705757184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039705115346038117,
      "loss": 2.6501,
      "theoretical_loss": 3.776277608422526,
      "tokens_seen": 705822720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039704112337011035,
      "loss": 2.7362,
      "theoretical_loss": 3.7762420611366556,
      "tokens_seen": 705888256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039703109327983953,
      "loss": 2.9123,
      "theoretical_loss": 3.776206518074889,
      "tokens_seen": 705953792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003970210631895687,
      "loss": 2.6798,
      "theoretical_loss": 3.776170979236332,
      "tokens_seen": 706019328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039701103309929794,
      "loss": 2.7245,
      "theoretical_loss": 3.7761354446200914,
      "tokens_seen": 706084864
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 398840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.317671775817871,
      "objective/train/theoretical_loss": 3.776099914225273,
      "objective/train/tokens_used": 726610400,
      "theoretical_loss": 3.776099914225273,
      "tokens_seen": 706150400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039700100300902707,
      "loss": 2.7278,
      "theoretical_loss": 3.776099914225273,
      "tokens_seen": 706150400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969909729187563,
      "loss": 2.8361,
      "theoretical_loss": 3.7760643880509845,
      "tokens_seen": 706215936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039698094282848543,
      "loss": 2.8228,
      "theoretical_loss": 3.776028866096332,
      "tokens_seen": 706281472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039697091273821467,
      "loss": 2.8389,
      "theoretical_loss": 3.7759933483604238,
      "tokens_seen": 706347008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039696088264794385,
      "loss": 2.7686,
      "theoretical_loss": 3.775957834842367,
      "tokens_seen": 706412544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039695085255767303,
      "loss": 2.8636,
      "theoretical_loss": 3.775922325541269,
      "tokens_seen": 706478080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969408224674022,
      "loss": 2.8,
      "theoretical_loss": 3.775886820456239,
      "tokens_seen": 706543616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039693079237713145,
      "loss": 2.8553,
      "theoretical_loss": 3.7758513195863843,
      "tokens_seen": 706609152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969207622868606,
      "loss": 2.7078,
      "theoretical_loss": 3.7758158229308143,
      "tokens_seen": 706674688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003969107321965898,
      "loss": 2.6724,
      "theoretical_loss": 3.775780330488638,
      "tokens_seen": 706740224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039690070210631894,
      "loss": 2.797,
      "theoretical_loss": 3.7757448422589643,
      "tokens_seen": 706805760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039689067201604817,
      "loss": 2.8585,
      "theoretical_loss": 3.7757093582409036,
      "tokens_seen": 706871296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039688064192577735,
      "loss": 2.5493,
      "theoretical_loss": 3.7756738784335644,
      "tokens_seen": 706936832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039687061183550653,
      "loss": 2.6593,
      "theoretical_loss": 3.775638402836057,
      "tokens_seen": 707002368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968605817452357,
      "loss": 2.8549,
      "theoretical_loss": 3.7756029314474926,
      "tokens_seen": 707067904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968505516549649,
      "loss": 2.6769,
      "theoretical_loss": 3.775567464266981,
      "tokens_seen": 707133440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968405215646941,
      "loss": 2.8655,
      "theoretical_loss": 3.7755320012936338,
      "tokens_seen": 707198976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968304914744233,
      "loss": 2.7365,
      "theoretical_loss": 3.775496542526561,
      "tokens_seen": 707264512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039682046138415244,
      "loss": 3.1246,
      "theoretical_loss": 3.7754610879648745,
      "tokens_seen": 707330048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968104312938817,
      "loss": 2.9093,
      "theoretical_loss": 3.775425637607687,
      "tokens_seen": 707395584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968004012036108,
      "loss": 2.7906,
      "theoretical_loss": 3.775390191454109,
      "tokens_seen": 707461120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039679037111334004,
      "loss": 2.6881,
      "theoretical_loss": 3.7753547495032533,
      "tokens_seen": 707526656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967803410230692,
      "loss": 2.6145,
      "theoretical_loss": 3.775319311754232,
      "tokens_seen": 707592192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967703109327984,
      "loss": 2.8557,
      "theoretical_loss": 3.775283878206159,
      "tokens_seen": 707657728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967602808425276,
      "loss": 2.8326,
      "theoretical_loss": 3.7752484488581457,
      "tokens_seen": 707723264
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 399432,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6968586444854736,
      "objective/train/theoretical_loss": 3.7752130237093064,
      "objective/train/tokens_used": 728248800,
      "theoretical_loss": 3.7752130237093064,
      "tokens_seen": 707788800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967502507522568,
      "loss": 2.6011,
      "theoretical_loss": 3.7752130237093064,
      "tokens_seen": 707788800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039674022066198594,
      "loss": 2.7763,
      "theoretical_loss": 3.775177602758755,
      "tokens_seen": 707854336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967301905717152,
      "loss": 2.9726,
      "theoretical_loss": 3.7751421860056045,
      "tokens_seen": 707919872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967201604814443,
      "loss": 2.7084,
      "theoretical_loss": 3.775106773448969,
      "tokens_seen": 707985408
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039671013039117354,
      "loss": 2.8755,
      "theoretical_loss": 3.7750713650879635,
      "tokens_seen": 708050944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967001003009027,
      "loss": 2.7319,
      "theoretical_loss": 3.775035960921702,
      "tokens_seen": 708116480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966900702106319,
      "loss": 2.816,
      "theoretical_loss": 3.7750005609493003,
      "tokens_seen": 708182016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966800401203611,
      "loss": 2.6943,
      "theoretical_loss": 3.774965165169873,
      "tokens_seen": 708247552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039667001003009027,
      "loss": 2.7305,
      "theoretical_loss": 3.774929773582535,
      "tokens_seen": 708313088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039665997993981945,
      "loss": 2.8117,
      "theoretical_loss": 3.7748943861864026,
      "tokens_seen": 708378624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966499498495487,
      "loss": 2.7627,
      "theoretical_loss": 3.774859002980592,
      "tokens_seen": 708444160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966399197592778,
      "loss": 2.8163,
      "theoretical_loss": 3.774823623964219,
      "tokens_seen": 708509696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039662988966900704,
      "loss": 2.7808,
      "theoretical_loss": 3.7747882491364004,
      "tokens_seen": 708575232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039661985957873617,
      "loss": 2.7834,
      "theoretical_loss": 3.774752878496253,
      "tokens_seen": 708640768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966098294884654,
      "loss": 2.867,
      "theoretical_loss": 3.7747175120428933,
      "tokens_seen": 708706304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965997993981946,
      "loss": 2.8166,
      "theoretical_loss": 3.774682149775439,
      "tokens_seen": 708771840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039658976930792377,
      "loss": 2.8267,
      "theoretical_loss": 3.7746467916930078,
      "tokens_seen": 708837376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039657973921765295,
      "loss": 2.5516,
      "theoretical_loss": 3.774611437794718,
      "tokens_seen": 708902912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965697091273822,
      "loss": 2.8087,
      "theoretical_loss": 3.7745760880796864,
      "tokens_seen": 708968448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965596790371113,
      "loss": 2.8472,
      "theoretical_loss": 3.7745407425470328,
      "tokens_seen": 709033984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039654964894684055,
      "loss": 2.6325,
      "theoretical_loss": 3.774505401195875,
      "tokens_seen": 709099520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965396188565697,
      "loss": 2.6698,
      "theoretical_loss": 3.774470064025332,
      "tokens_seen": 709165056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965295887662989,
      "loss": 2.4007,
      "theoretical_loss": 3.774434731034524,
      "tokens_seen": 709230592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965195586760281,
      "loss": 2.8729,
      "theoretical_loss": 3.7743994022225684,
      "tokens_seen": 709296128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039650952858575727,
      "loss": 2.6295,
      "theoretical_loss": 3.774364077588587,
      "tokens_seen": 709361664
    },
    {
      "epoch": 0.21,
      "objective/train/docs_used": 400468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9617061614990234,
      "objective/train/theoretical_loss": 3.7743287571316984,
      "objective/train/tokens_used": 729887200,
      "theoretical_loss": 3.7743287571316984,
      "tokens_seen": 709427200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039649949849548645,
      "loss": 2.8305,
      "theoretical_loss": 3.7743287571316984,
      "tokens_seen": 709427200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039648946840521563,
      "loss": 2.9433,
      "theoretical_loss": 3.774293440851024,
      "tokens_seen": 709492736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964794383149448,
      "loss": 2.7475,
      "theoretical_loss": 3.774258128745683,
      "tokens_seen": 709558272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039646940822467405,
      "loss": 2.7497,
      "theoretical_loss": 3.7742228208147974,
      "tokens_seen": 709623808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964593781344032,
      "loss": 2.5815,
      "theoretical_loss": 3.7741875170574875,
      "tokens_seen": 709689344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964493480441324,
      "loss": 2.7236,
      "theoretical_loss": 3.774152217472875,
      "tokens_seen": 709754880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964393179538616,
      "loss": 2.6442,
      "theoretical_loss": 3.774116922060081,
      "tokens_seen": 709820416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964292878635908,
      "loss": 2.8988,
      "theoretical_loss": 3.7740816308182286,
      "tokens_seen": 709885952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039641925777331996,
      "loss": 2.9272,
      "theoretical_loss": 3.7740463437464387,
      "tokens_seen": 709951488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039640922768304914,
      "loss": 2.836,
      "theoretical_loss": 3.7740110608438338,
      "tokens_seen": 710017024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963991975927783,
      "loss": 2.9205,
      "theoretical_loss": 3.7739757821095363,
      "tokens_seen": 710082560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039638916750250755,
      "loss": 2.7784,
      "theoretical_loss": 3.7739405075426706,
      "tokens_seen": 710148096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039637913741223673,
      "loss": 2.7215,
      "theoretical_loss": 3.7739052371423583,
      "tokens_seen": 710213632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963691073219659,
      "loss": 2.76,
      "theoretical_loss": 3.7738699709077235,
      "tokens_seen": 710279168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963590772316951,
      "loss": 2.7615,
      "theoretical_loss": 3.77383470883789,
      "tokens_seen": 710344704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963490471414243,
      "loss": 2.6549,
      "theoretical_loss": 3.773799450931982,
      "tokens_seen": 710410240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963390170511535,
      "loss": 2.7139,
      "theoretical_loss": 3.7737641971891227,
      "tokens_seen": 710475776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039632898696088264,
      "loss": 2.8427,
      "theoretical_loss": 3.7737289476084372,
      "tokens_seen": 710541312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963189568706119,
      "loss": 2.8954,
      "theoretical_loss": 3.773693702189051,
      "tokens_seen": 710606848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396308926780341,
      "loss": 2.6769,
      "theoretical_loss": 3.773658460930088,
      "tokens_seen": 710672384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039629889669007024,
      "loss": 2.7509,
      "theoretical_loss": 3.7736232238306746,
      "tokens_seen": 710737920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962888665997994,
      "loss": 2.7514,
      "theoretical_loss": 3.7735879908899355,
      "tokens_seen": 710803456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962788365095286,
      "loss": 2.7883,
      "theoretical_loss": 3.7735527621069966,
      "tokens_seen": 710868992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962688064192578,
      "loss": 2.935,
      "theoretical_loss": 3.773517537480984,
      "tokens_seen": 710934528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396258776328987,
      "loss": 2.6263,
      "theoretical_loss": 3.773482317011025,
      "tokens_seen": 711000064
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 401088,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.054889440536499,
      "objective/train/theoretical_loss": 3.773447100696245,
      "objective/train/tokens_used": 731525600,
      "theoretical_loss": 3.773447100696245,
      "tokens_seen": 711065600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039624874623871614,
      "loss": 2.9868,
      "theoretical_loss": 3.773447100696245,
      "tokens_seen": 711065600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962387161484454,
      "loss": 2.7579,
      "theoretical_loss": 3.7734118885357715,
      "tokens_seen": 711131136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962286860581745,
      "loss": 2.8857,
      "theoretical_loss": 3.7733766805287314,
      "tokens_seen": 711196672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039621865596790374,
      "loss": 2.9905,
      "theoretical_loss": 3.7733414766742523,
      "tokens_seen": 711262208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962086258776329,
      "loss": 2.584,
      "theoretical_loss": 3.7733062769714616,
      "tokens_seen": 711327744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961985957873621,
      "loss": 2.625,
      "theoretical_loss": 3.7732710814194883,
      "tokens_seen": 711393280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961885656970913,
      "loss": 2.8167,
      "theoretical_loss": 3.7732358900174585,
      "tokens_seen": 711458816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039617853560682047,
      "loss": 2.9579,
      "theoretical_loss": 3.7732007027645027,
      "tokens_seen": 711524352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039616850551654965,
      "loss": 2.876,
      "theoretical_loss": 3.7731655196597496,
      "tokens_seen": 711589888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961584754262789,
      "loss": 2.757,
      "theoretical_loss": 3.7731303407023264,
      "tokens_seen": 711655424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396148445336008,
      "loss": 2.8455,
      "theoretical_loss": 3.773095165891364,
      "tokens_seen": 711720960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039613841524573724,
      "loss": 2.6882,
      "theoretical_loss": 3.7730599952259913,
      "tokens_seen": 711786496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039612838515546637,
      "loss": 2.7801,
      "theoretical_loss": 3.7730248287053376,
      "tokens_seen": 711852032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961183550651956,
      "loss": 2.8345,
      "theoretical_loss": 3.772989666328534,
      "tokens_seen": 711917568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961083249749248,
      "loss": 2.8088,
      "theoretical_loss": 3.7729545080947102,
      "tokens_seen": 711983104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039609829488465397,
      "loss": 2.4923,
      "theoretical_loss": 3.772919354002997,
      "tokens_seen": 712048640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039608826479438315,
      "loss": 2.6372,
      "theoretical_loss": 3.772884204052525,
      "tokens_seen": 712114176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960782347041124,
      "loss": 2.7558,
      "theoretical_loss": 3.7728490582424254,
      "tokens_seen": 712179712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960682046138415,
      "loss": 2.7638,
      "theoretical_loss": 3.7728139165718297,
      "tokens_seen": 712245248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039605817452357075,
      "loss": 2.875,
      "theoretical_loss": 3.7727787790398697,
      "tokens_seen": 712310784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960481444332999,
      "loss": 2.9478,
      "theoretical_loss": 3.7727436456456767,
      "tokens_seen": 712376320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960381143430291,
      "loss": 2.7149,
      "theoretical_loss": 3.772708516388383,
      "tokens_seen": 712441856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960280842527583,
      "loss": 2.8476,
      "theoretical_loss": 3.772673391267121,
      "tokens_seen": 712507392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039601805416248747,
      "loss": 2.7814,
      "theoretical_loss": 3.772638270281024,
      "tokens_seen": 712572928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039600802407221665,
      "loss": 2.7704,
      "theoretical_loss": 3.7726031534292246,
      "tokens_seen": 712638464
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 402378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6918859481811523,
      "objective/train/theoretical_loss": 3.7725680407108557,
      "objective/train/tokens_used": 733164000,
      "theoretical_loss": 3.7725680407108557,
      "tokens_seen": 712704000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039599799398194583,
      "loss": 2.7392,
      "theoretical_loss": 3.7725680407108557,
      "tokens_seen": 712704000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395987963891675,
      "loss": 2.6279,
      "theoretical_loss": 3.7725329321250505,
      "tokens_seen": 712769536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039597793380140425,
      "loss": 2.7209,
      "theoretical_loss": 3.7724978276709433,
      "tokens_seen": 712835072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959679037111334,
      "loss": 2.7042,
      "theoretical_loss": 3.772462727347668,
      "tokens_seen": 712900608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959578736208626,
      "loss": 2.6791,
      "theoretical_loss": 3.7724276311543585,
      "tokens_seen": 712966144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959478435305918,
      "loss": 2.8125,
      "theoretical_loss": 3.7723925390901503,
      "tokens_seen": 713031680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395937813440321,
      "loss": 2.7226,
      "theoretical_loss": 3.7723574511541766,
      "tokens_seen": 713097216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039592778335005016,
      "loss": 2.6501,
      "theoretical_loss": 3.7723223673455735,
      "tokens_seen": 713162752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039591775325977934,
      "loss": 2.8047,
      "theoretical_loss": 3.772287287663476,
      "tokens_seen": 713228288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959077231695085,
      "loss": 2.7207,
      "theoretical_loss": 3.77225221210702,
      "tokens_seen": 713293824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039589769307923775,
      "loss": 2.7146,
      "theoretical_loss": 3.7722171406753406,
      "tokens_seen": 713359360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958876629889669,
      "loss": 2.713,
      "theoretical_loss": 3.772182073367574,
      "tokens_seen": 713424896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958776328986961,
      "loss": 2.8498,
      "theoretical_loss": 3.7721470101828576,
      "tokens_seen": 713490432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039586760280842524,
      "loss": 2.3899,
      "theoretical_loss": 3.7721119511203263,
      "tokens_seen": 713555968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958575727181545,
      "loss": 2.5299,
      "theoretical_loss": 3.7720768961791182,
      "tokens_seen": 713621504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039584754262788366,
      "loss": 2.6941,
      "theoretical_loss": 3.7720418453583697,
      "tokens_seen": 713687040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039583751253761284,
      "loss": 2.6472,
      "theoretical_loss": 3.7720067986572188,
      "tokens_seen": 713752576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395827482447342,
      "loss": 2.7839,
      "theoretical_loss": 3.7719717560748025,
      "tokens_seen": 713818112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958174523570712,
      "loss": 2.7576,
      "theoretical_loss": 3.771936717610259,
      "tokens_seen": 713883648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958074222668004,
      "loss": 2.722,
      "theoretical_loss": 3.7719016832627266,
      "tokens_seen": 713949184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957973921765296,
      "loss": 2.44,
      "theoretical_loss": 3.7718666530313434,
      "tokens_seen": 714014720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039578736208625875,
      "loss": 2.7879,
      "theoretical_loss": 3.7718316269152483,
      "tokens_seen": 714080256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395777331995988,
      "loss": 3.0367,
      "theoretical_loss": 3.77179660491358,
      "tokens_seen": 714145792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039576730190571716,
      "loss": 2.8812,
      "theoretical_loss": 3.7717615870254777,
      "tokens_seen": 714211328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039575727181544634,
      "loss": 2.9063,
      "theoretical_loss": 3.7717265732500804,
      "tokens_seen": 714276864
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 402960,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7978687286376953,
      "objective/train/theoretical_loss": 3.771691563586529,
      "objective/train/tokens_used": 734802400,
      "theoretical_loss": 3.771691563586529,
      "tokens_seen": 714342400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957472417251755,
      "loss": 2.6231,
      "theoretical_loss": 3.771691563586529,
      "tokens_seen": 714342400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957372116349047,
      "loss": 2.7291,
      "theoretical_loss": 3.7716565580339623,
      "tokens_seen": 714407936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957271815446339,
      "loss": 2.7152,
      "theoretical_loss": 3.771621556591522,
      "tokens_seen": 714473472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957171514543631,
      "loss": 2.7035,
      "theoretical_loss": 3.7715865592583464,
      "tokens_seen": 714539008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039570712136409225,
      "loss": 2.6655,
      "theoretical_loss": 3.771551566033578,
      "tokens_seen": 714604544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956970912738215,
      "loss": 2.6651,
      "theoretical_loss": 3.771516576916357,
      "tokens_seen": 714670080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956870611835506,
      "loss": 2.6674,
      "theoretical_loss": 3.7714815919058244,
      "tokens_seen": 714735616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039567703109327985,
      "loss": 2.7265,
      "theoretical_loss": 3.771446611001123,
      "tokens_seen": 714801152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039566700100300903,
      "loss": 2.6466,
      "theoretical_loss": 3.7714116342013932,
      "tokens_seen": 714866688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956569709127382,
      "loss": 2.8245,
      "theoretical_loss": 3.7713766615057778,
      "tokens_seen": 714932224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956469408224674,
      "loss": 2.7489,
      "theoretical_loss": 3.7713416929134187,
      "tokens_seen": 714997760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039563691073219657,
      "loss": 2.6609,
      "theoretical_loss": 3.7713067284234585,
      "tokens_seen": 715063296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956268806419258,
      "loss": 2.6133,
      "theoretical_loss": 3.7712717680350405,
      "tokens_seen": 715128832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395616850551655,
      "loss": 2.7701,
      "theoretical_loss": 3.7712368117473067,
      "tokens_seen": 715194368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039560682046138417,
      "loss": 2.5693,
      "theoretical_loss": 3.771201859559401,
      "tokens_seen": 715259904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039559679037111335,
      "loss": 2.9091,
      "theoretical_loss": 3.771166911470468,
      "tokens_seen": 715325440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955867602808426,
      "loss": 2.6383,
      "theoretical_loss": 3.7711319674796497,
      "tokens_seen": 715390976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955767301905717,
      "loss": 2.8301,
      "theoretical_loss": 3.771097027586091,
      "tokens_seen": 715456512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039556670010030095,
      "loss": 2.8646,
      "theoretical_loss": 3.7710620917889366,
      "tokens_seen": 715522048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955566700100301,
      "loss": 2.664,
      "theoretical_loss": 3.77102716008733,
      "tokens_seen": 715587584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955466399197593,
      "loss": 2.5999,
      "theoretical_loss": 3.7709922324804177,
      "tokens_seen": 715653120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955366098294885,
      "loss": 2.9666,
      "theoretical_loss": 3.7709573089673443,
      "tokens_seen": 715718656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039552657973921767,
      "loss": 2.6747,
      "theoretical_loss": 3.7709223895472537,
      "tokens_seen": 715784192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039551654964894685,
      "loss": 2.4762,
      "theoretical_loss": 3.7708874742192933,
      "tokens_seen": 715849728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039550651955867603,
      "loss": 2.8351,
      "theoretical_loss": 3.7708525629826077,
      "tokens_seen": 715915264
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 403589,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7969202995300293,
      "objective/train/theoretical_loss": 3.7708176558363444,
      "objective/train/tokens_used": 736440800,
      "theoretical_loss": 3.7708176558363444,
      "tokens_seen": 715980800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954964894684052,
      "loss": 2.5087,
      "theoretical_loss": 3.7708176558363444,
      "tokens_seen": 715980800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039548645937813445,
      "loss": 2.7577,
      "theoretical_loss": 3.770782752779649,
      "tokens_seen": 716046336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954764292878636,
      "loss": 2.8457,
      "theoretical_loss": 3.7707478538116677,
      "tokens_seen": 716111872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954663991975928,
      "loss": 2.5495,
      "theoretical_loss": 3.7707129589315485,
      "tokens_seen": 716177408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395456369107322,
      "loss": 2.7968,
      "theoretical_loss": 3.7706780681384373,
      "tokens_seen": 716242944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954463390170512,
      "loss": 2.5126,
      "theoretical_loss": 3.770643181431483,
      "tokens_seen": 716308480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039543630892678036,
      "loss": 2.7764,
      "theoretical_loss": 3.770608298809832,
      "tokens_seen": 716374016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039542627883650954,
      "loss": 2.7939,
      "theoretical_loss": 3.770573420272633,
      "tokens_seen": 716439552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954162487462387,
      "loss": 2.6007,
      "theoretical_loss": 3.7705385458190337,
      "tokens_seen": 716505088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039540621865596795,
      "loss": 2.7519,
      "theoretical_loss": 3.7705036754481824,
      "tokens_seen": 716570624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953961885656971,
      "loss": 2.9651,
      "theoretical_loss": 3.7704688091592287,
      "tokens_seen": 716636160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953861584754263,
      "loss": 2.511,
      "theoretical_loss": 3.7704339469513206,
      "tokens_seen": 716701696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039537612838515544,
      "loss": 2.6907,
      "theoretical_loss": 3.770399088823608,
      "tokens_seen": 716767232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953660982948847,
      "loss": 2.5684,
      "theoretical_loss": 3.77036423477524,
      "tokens_seen": 716832768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039535606820461386,
      "loss": 2.6212,
      "theoretical_loss": 3.7703293848053665,
      "tokens_seen": 716898304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039534603811434304,
      "loss": 2.618,
      "theoretical_loss": 3.770294538913137,
      "tokens_seen": 716963840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953360080240722,
      "loss": 2.6721,
      "theoretical_loss": 3.7702596970977025,
      "tokens_seen": 717029376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953259779338014,
      "loss": 2.7419,
      "theoretical_loss": 3.770224859358213,
      "tokens_seen": 717094912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953159478435306,
      "loss": 2.7789,
      "theoretical_loss": 3.770190025693819,
      "tokens_seen": 717160448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953059177532598,
      "loss": 2.6623,
      "theoretical_loss": 3.7701551961036723,
      "tokens_seen": 717225984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039529588766298895,
      "loss": 2.9295,
      "theoretical_loss": 3.7701203705869233,
      "tokens_seen": 717291520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952858575727182,
      "loss": 2.7038,
      "theoretical_loss": 3.770085549142724,
      "tokens_seen": 717357056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039527582748244736,
      "loss": 2.759,
      "theoretical_loss": 3.770050731770226,
      "tokens_seen": 717422592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039526579739217654,
      "loss": 2.4058,
      "theoretical_loss": 3.770015918468582,
      "tokens_seen": 717488128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952557673019057,
      "loss": 2.7957,
      "theoretical_loss": 3.7699811092369426,
      "tokens_seen": 717553664
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 404678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7875523567199707,
      "objective/train/theoretical_loss": 3.7699463040744616,
      "objective/train/tokens_used": 738079200,
      "theoretical_loss": 3.7699463040744616,
      "tokens_seen": 717619200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952457372116349,
      "loss": 2.6614,
      "theoretical_loss": 3.7699463040744616,
      "tokens_seen": 717619200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952357071213641,
      "loss": 2.672,
      "theoretical_loss": 3.7699115029802917,
      "tokens_seen": 717684736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952256770310933,
      "loss": 2.7551,
      "theoretical_loss": 3.769876705953586,
      "tokens_seen": 717750272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039521564694082245,
      "loss": 2.6962,
      "theoretical_loss": 3.769841912993497,
      "tokens_seen": 717815808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952056168505517,
      "loss": 2.7611,
      "theoretical_loss": 3.769807124099179,
      "tokens_seen": 717881344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951955867602808,
      "loss": 2.8147,
      "theoretical_loss": 3.7697723392697853,
      "tokens_seen": 717946880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039518555667001005,
      "loss": 2.7761,
      "theoretical_loss": 3.7697375585044703,
      "tokens_seen": 718012416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039517552657973923,
      "loss": 2.8367,
      "theoretical_loss": 3.769702781802388,
      "tokens_seen": 718077952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951654964894684,
      "loss": 2.6044,
      "theoretical_loss": 3.769668009162693,
      "tokens_seen": 718143488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951554663991976,
      "loss": 2.5959,
      "theoretical_loss": 3.7696332405845405,
      "tokens_seen": 718209024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039514543630892677,
      "loss": 2.517,
      "theoretical_loss": 3.7695984760670855,
      "tokens_seen": 718274560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039513540621865595,
      "loss": 2.7682,
      "theoretical_loss": 3.7695637156094826,
      "tokens_seen": 718340096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951253761283852,
      "loss": 2.4989,
      "theoretical_loss": 3.769528959210888,
      "tokens_seen": 718405632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951153460381143,
      "loss": 2.647,
      "theoretical_loss": 3.7694942068704576,
      "tokens_seen": 718471168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039510531594784355,
      "loss": 2.735,
      "theoretical_loss": 3.7694594585873467,
      "tokens_seen": 718536704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039509528585757273,
      "loss": 2.6373,
      "theoretical_loss": 3.7694247143607127,
      "tokens_seen": 718602240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950852557673019,
      "loss": 2.8542,
      "theoretical_loss": 3.7693899741897106,
      "tokens_seen": 718667776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950752256770311,
      "loss": 2.843,
      "theoretical_loss": 3.7693552380734987,
      "tokens_seen": 718733312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950651955867603,
      "loss": 2.7613,
      "theoretical_loss": 3.7693205060112343,
      "tokens_seen": 718798848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039505516549648946,
      "loss": 2.7541,
      "theoretical_loss": 3.769285778002073,
      "tokens_seen": 718864384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950451354062187,
      "loss": 2.5525,
      "theoretical_loss": 3.769251054045174,
      "tokens_seen": 718929920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950351053159478,
      "loss": 2.7406,
      "theoretical_loss": 3.769216334139694,
      "tokens_seen": 718995456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039502507522567705,
      "loss": 2.6661,
      "theoretical_loss": 3.769181618284792,
      "tokens_seen": 719060992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950150451354062,
      "loss": 2.4692,
      "theoretical_loss": 3.7691469064796257,
      "tokens_seen": 719126528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950050150451354,
      "loss": 2.5154,
      "theoretical_loss": 3.769112198723354,
      "tokens_seen": 719192064
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 405859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0279910564422607,
      "objective/train/theoretical_loss": 3.769077495015136,
      "objective/train/tokens_used": 739717600,
      "theoretical_loss": 3.769077495015136,
      "tokens_seen": 719257600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949949849548646,
      "loss": 2.6503,
      "theoretical_loss": 3.769077495015136,
      "tokens_seen": 719257600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949849548645938,
      "loss": 2.8353,
      "theoretical_loss": 3.76904279535413,
      "tokens_seen": 719323136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039497492477432296,
      "loss": 2.6663,
      "theoretical_loss": 3.7690080997394952,
      "tokens_seen": 719388672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949648946840522,
      "loss": 2.7235,
      "theoretical_loss": 3.7689734081703925,
      "tokens_seen": 719454208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949548645937813,
      "loss": 2.5129,
      "theoretical_loss": 3.7689387206459806,
      "tokens_seen": 719519744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039494483450351056,
      "loss": 2.8694,
      "theoretical_loss": 3.7689040371654206,
      "tokens_seen": 719585280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949348044132397,
      "loss": 2.801,
      "theoretical_loss": 3.768869357727872,
      "tokens_seen": 719650816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949247743229689,
      "loss": 2.8021,
      "theoretical_loss": 3.7688346823324954,
      "tokens_seen": 719716352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949147442326981,
      "loss": 2.8085,
      "theoretical_loss": 3.7688000109784516,
      "tokens_seen": 719781888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949047141424273,
      "loss": 2.6279,
      "theoretical_loss": 3.768765343664902,
      "tokens_seen": 719847424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948946840521565,
      "loss": 2.6426,
      "theoretical_loss": 3.7687306803910086,
      "tokens_seen": 719912960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039488465396188564,
      "loss": 2.7365,
      "theoretical_loss": 3.7686960211559315,
      "tokens_seen": 719978496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948746238716149,
      "loss": 2.8506,
      "theoretical_loss": 3.7686613659588337,
      "tokens_seen": 720044032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039486459378134406,
      "loss": 2.9218,
      "theoretical_loss": 3.7686267147988772,
      "tokens_seen": 720109568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039485456369107324,
      "loss": 2.7985,
      "theoretical_loss": 3.7685920676752236,
      "tokens_seen": 720175104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948445336008024,
      "loss": 2.7395,
      "theoretical_loss": 3.768557424587036,
      "tokens_seen": 720240640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948345035105316,
      "loss": 2.9131,
      "theoretical_loss": 3.7685227855334777,
      "tokens_seen": 720306176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948244734202608,
      "loss": 2.7174,
      "theoretical_loss": 3.7684881505137113,
      "tokens_seen": 720371712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039481444332999,
      "loss": 3.0072,
      "theoretical_loss": 3.7684535195268998,
      "tokens_seen": 720437248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039480441323971915,
      "loss": 2.7996,
      "theoretical_loss": 3.7684188925722073,
      "tokens_seen": 720502784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947943831494484,
      "loss": 2.8799,
      "theoretical_loss": 3.7683842696487977,
      "tokens_seen": 720568320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039478435305917756,
      "loss": 2.6279,
      "theoretical_loss": 3.7683496507558347,
      "tokens_seen": 720633856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039477432296890674,
      "loss": 3.06,
      "theoretical_loss": 3.7683150358924826,
      "tokens_seen": 720699392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947642928786359,
      "loss": 3.0075,
      "theoretical_loss": 3.7682804250579065,
      "tokens_seen": 720764928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947542627883651,
      "loss": 2.9534,
      "theoretical_loss": 3.768245818251271,
      "tokens_seen": 720830464
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 406138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.268052577972412,
      "objective/train/theoretical_loss": 3.768211215471741,
      "objective/train/tokens_used": 741356000,
      "theoretical_loss": 3.768211215471741,
      "tokens_seen": 720896000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947442326980943,
      "loss": 2.7847,
      "theoretical_loss": 3.768211215471741,
      "tokens_seen": 720896000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947342026078235,
      "loss": 2.6443,
      "theoretical_loss": 3.768176616718482,
      "tokens_seen": 720961536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039472417251755265,
      "loss": 2.5376,
      "theoretical_loss": 3.7681420219906596,
      "tokens_seen": 721027072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947141424272819,
      "loss": 2.579,
      "theoretical_loss": 3.76810743128744,
      "tokens_seen": 721092608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394704112337011,
      "loss": 2.7805,
      "theoretical_loss": 3.768072844607988,
      "tokens_seen": 721158144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039469408224674025,
      "loss": 2.6879,
      "theoretical_loss": 3.768038261951472,
      "tokens_seen": 721223680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039468405215646943,
      "loss": 2.6076,
      "theoretical_loss": 3.7680036833170565,
      "tokens_seen": 721289216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946740220661986,
      "loss": 2.699,
      "theoretical_loss": 3.7679691087039098,
      "tokens_seen": 721354752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946639919759278,
      "loss": 2.7964,
      "theoretical_loss": 3.7679345381111986,
      "tokens_seen": 721420288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039465396188565697,
      "loss": 2.5689,
      "theoretical_loss": 3.76789997153809,
      "tokens_seen": 721485824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039464393179538615,
      "loss": 2.7629,
      "theoretical_loss": 3.767865408983751,
      "tokens_seen": 721551360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946339017051154,
      "loss": 2.6272,
      "theoretical_loss": 3.7678308504473508,
      "tokens_seen": 721616896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946238716148445,
      "loss": 2.581,
      "theoretical_loss": 3.767796295928056,
      "tokens_seen": 721682432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039461384152457375,
      "loss": 2.7804,
      "theoretical_loss": 3.767761745425036,
      "tokens_seen": 721747968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039460381143430293,
      "loss": 2.8805,
      "theoretical_loss": 3.767727198937459,
      "tokens_seen": 721813504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945937813440321,
      "loss": 2.8968,
      "theoretical_loss": 3.767692656464494,
      "tokens_seen": 721879040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945837512537613,
      "loss": 2.6535,
      "theoretical_loss": 3.76765811800531,
      "tokens_seen": 721944576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945737211634905,
      "loss": 2.8241,
      "theoretical_loss": 3.767623583559076,
      "tokens_seen": 722010112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039456369107321966,
      "loss": 2.9034,
      "theoretical_loss": 3.767589053124962,
      "tokens_seen": 722075648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945536609829489,
      "loss": 3.0686,
      "theoretical_loss": 3.767554526702137,
      "tokens_seen": 722141184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394543630892678,
      "loss": 2.8172,
      "theoretical_loss": 3.7675200042897723,
      "tokens_seen": 722206720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039453360080240725,
      "loss": 2.8621,
      "theoretical_loss": 3.767485485887037,
      "tokens_seen": 722272256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945235707121364,
      "loss": 2.8103,
      "theoretical_loss": 3.7674509714931026,
      "tokens_seen": 722337792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945135406218656,
      "loss": 2.893,
      "theoretical_loss": 3.767416461107139,
      "tokens_seen": 722403328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945035105315948,
      "loss": 2.8334,
      "theoretical_loss": 3.767381954728318,
      "tokens_seen": 722468864
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 407490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3906686305999756,
      "objective/train/theoretical_loss": 3.7673474523558106,
      "objective/train/tokens_used": 742994400,
      "theoretical_loss": 3.7673474523558106,
      "tokens_seen": 722534400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394493480441324,
      "loss": 2.661,
      "theoretical_loss": 3.7673474523558106,
      "tokens_seen": 722534400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039448345035105316,
      "loss": 2.7209,
      "theoretical_loss": 3.767312953988788,
      "tokens_seen": 722599936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944734202607824,
      "loss": 2.5782,
      "theoretical_loss": 3.767278459626423,
      "tokens_seen": 722665472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944633901705115,
      "loss": 2.589,
      "theoretical_loss": 3.767243969267886,
      "tokens_seen": 722731008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039445336008024076,
      "loss": 2.7244,
      "theoretical_loss": 3.7672094829123512,
      "tokens_seen": 722796544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944433299899699,
      "loss": 2.6615,
      "theoretical_loss": 3.7671750005589892,
      "tokens_seen": 722862080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944332998996991,
      "loss": 3.0289,
      "theoretical_loss": 3.767140522206974,
      "tokens_seen": 722927616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944232698094283,
      "loss": 2.6406,
      "theoretical_loss": 3.7671060478554783,
      "tokens_seen": 722993152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944132397191575,
      "loss": 2.746,
      "theoretical_loss": 3.7670715775036756,
      "tokens_seen": 723058688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039440320962888666,
      "loss": 2.9204,
      "theoretical_loss": 3.7670371111507386,
      "tokens_seen": 723124224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039439317953861584,
      "loss": 2.5796,
      "theoretical_loss": 3.7670026487958417,
      "tokens_seen": 723189760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394383149448345,
      "loss": 2.6218,
      "theoretical_loss": 3.766968190438159,
      "tokens_seen": 723255296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039437311935807426,
      "loss": 2.7064,
      "theoretical_loss": 3.766933736076864,
      "tokens_seen": 723320832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943630892678034,
      "loss": 2.7234,
      "theoretical_loss": 3.766899285711132,
      "tokens_seen": 723386368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943530591775326,
      "loss": 3.1493,
      "theoretical_loss": 3.766864839340138,
      "tokens_seen": 723451904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039434302908726175,
      "loss": 2.6272,
      "theoretical_loss": 3.7668303969630554,
      "tokens_seen": 723517440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394332998996991,
      "loss": 2.9986,
      "theoretical_loss": 3.7667959585790607,
      "tokens_seen": 723582976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039432296890672017,
      "loss": 2.8482,
      "theoretical_loss": 3.76676152418733,
      "tokens_seen": 723648512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039431293881644935,
      "loss": 2.808,
      "theoretical_loss": 3.7667270937870363,
      "tokens_seen": 723714048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039430290872617853,
      "loss": 2.7588,
      "theoretical_loss": 3.7666926673773586,
      "tokens_seen": 723779584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039429287863590776,
      "loss": 2.6425,
      "theoretical_loss": 3.7666582449574717,
      "tokens_seen": 723845120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942828485456369,
      "loss": 2.6043,
      "theoretical_loss": 3.766623826526552,
      "tokens_seen": 723910656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942728184553661,
      "loss": 2.3893,
      "theoretical_loss": 3.766589412083776,
      "tokens_seen": 723976192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039426278836509525,
      "loss": 2.8363,
      "theoretical_loss": 3.766555001628322,
      "tokens_seen": 724041728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942527582748245,
      "loss": 2.7142,
      "theoretical_loss": 3.766520595159365,
      "tokens_seen": 724107264
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 408057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.163313150405884,
      "objective/train/theoretical_loss": 3.766486192676084,
      "objective/train/tokens_used": 744632800,
      "theoretical_loss": 3.766486192676084,
      "tokens_seen": 724172800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039424272818455367,
      "loss": 2.7706,
      "theoretical_loss": 3.766486192676084,
      "tokens_seen": 724172800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039423269809428285,
      "loss": 2.7204,
      "theoretical_loss": 3.766451794177656,
      "tokens_seen": 724238336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039422266800401203,
      "loss": 2.8968,
      "theoretical_loss": 3.7664173996632595,
      "tokens_seen": 724303872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942126379137412,
      "loss": 2.7789,
      "theoretical_loss": 3.7663830091320722,
      "tokens_seen": 724369408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942026078234704,
      "loss": 2.8649,
      "theoretical_loss": 3.7663486225832727,
      "tokens_seen": 724434944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039419257773319963,
      "loss": 3.0287,
      "theoretical_loss": 3.7663142400160403,
      "tokens_seen": 724500480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039418254764292876,
      "loss": 2.9706,
      "theoretical_loss": 3.7662798614295516,
      "tokens_seen": 724566016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394172517552658,
      "loss": 2.5185,
      "theoretical_loss": 3.7662454868229887,
      "tokens_seen": 724631552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941624874623871,
      "loss": 2.6261,
      "theoretical_loss": 3.766211116195528,
      "tokens_seen": 724697088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039415245737211635,
      "loss": 2.7275,
      "theoretical_loss": 3.766176749546352,
      "tokens_seen": 724762624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941424272818456,
      "loss": 2.8693,
      "theoretical_loss": 3.7661423868746384,
      "tokens_seen": 724828160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941323971915747,
      "loss": 2.6889,
      "theoretical_loss": 3.766108028179568,
      "tokens_seen": 724893696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039412236710130395,
      "loss": 2.7965,
      "theoretical_loss": 3.7660736734603217,
      "tokens_seen": 724959232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039411233701103313,
      "loss": 2.7244,
      "theoretical_loss": 3.7660393227160798,
      "tokens_seen": 725024768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941023069207623,
      "loss": 2.8347,
      "theoretical_loss": 3.7660049759460223,
      "tokens_seen": 725090304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940922768304915,
      "loss": 2.7925,
      "theoretical_loss": 3.7659706331493314,
      "tokens_seen": 725155840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940822467402207,
      "loss": 2.8973,
      "theoretical_loss": 3.7659362943251877,
      "tokens_seen": 725221376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039407221664994986,
      "loss": 2.4894,
      "theoretical_loss": 3.765901959472773,
      "tokens_seen": 725286912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940621865596791,
      "loss": 2.7852,
      "theoretical_loss": 3.7658676285912693,
      "tokens_seen": 725352448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940521564694082,
      "loss": 2.8938,
      "theoretical_loss": 3.765833301679858,
      "tokens_seen": 725417984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039404212637913745,
      "loss": 2.7822,
      "theoretical_loss": 3.765798978737722,
      "tokens_seen": 725483520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940320962888666,
      "loss": 2.7152,
      "theoretical_loss": 3.765764659764044,
      "tokens_seen": 725549056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940220661985958,
      "loss": 2.7457,
      "theoretical_loss": 3.7657303447580066,
      "tokens_seen": 725614592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394012036108325,
      "loss": 2.7628,
      "theoretical_loss": 3.765696033718792,
      "tokens_seen": 725680128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940020060180542,
      "loss": 2.8026,
      "theoretical_loss": 3.7656617266455843,
      "tokens_seen": 725745664
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 409133,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.649945020675659,
      "objective/train/theoretical_loss": 3.765627423537567,
      "objective/train/tokens_used": 746271200,
      "theoretical_loss": 3.765627423537567,
      "tokens_seen": 725811200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039399197592778336,
      "loss": 2.6373,
      "theoretical_loss": 3.765627423537567,
      "tokens_seen": 725811200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939819458375126,
      "loss": 2.9122,
      "theoretical_loss": 3.765593124393923,
      "tokens_seen": 725876736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939719157472417,
      "loss": 2.8708,
      "theoretical_loss": 3.7655588292138376,
      "tokens_seen": 725942272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039396188565697096,
      "loss": 2.7742,
      "theoretical_loss": 3.7655245379964937,
      "tokens_seen": 726007808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939518555667001,
      "loss": 2.6562,
      "theoretical_loss": 3.765490250741077,
      "tokens_seen": 726073344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939418254764293,
      "loss": 2.6231,
      "theoretical_loss": 3.765455967446771,
      "tokens_seen": 726138880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939317953861585,
      "loss": 2.758,
      "theoretical_loss": 3.7654216881127613,
      "tokens_seen": 726204416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939217652958877,
      "loss": 2.7845,
      "theoretical_loss": 3.7653874127382334,
      "tokens_seen": 726269952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039391173520561686,
      "loss": 2.8462,
      "theoretical_loss": 3.765353141322372,
      "tokens_seen": 726335488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039390170511534604,
      "loss": 2.6643,
      "theoretical_loss": 3.7653188738643637,
      "tokens_seen": 726401024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938916750250752,
      "loss": 2.9446,
      "theoretical_loss": 3.765284610363393,
      "tokens_seen": 726466560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039388164493480446,
      "loss": 2.9237,
      "theoretical_loss": 3.765250350818647,
      "tokens_seen": 726532096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938716148445336,
      "loss": 2.5889,
      "theoretical_loss": 3.7652160952293126,
      "tokens_seen": 726597632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938615847542628,
      "loss": 2.7943,
      "theoretical_loss": 3.765181843594575,
      "tokens_seen": 726663168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039385155466399195,
      "loss": 2.8595,
      "theoretical_loss": 3.7651475959136222,
      "tokens_seen": 726728704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938415245737212,
      "loss": 2.7223,
      "theoretical_loss": 3.765113352185641,
      "tokens_seen": 726794240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039383149448345037,
      "loss": 2.4638,
      "theoretical_loss": 3.7650791124098193,
      "tokens_seen": 726859776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039382146439317955,
      "loss": 2.8602,
      "theoretical_loss": 3.7650448765853435,
      "tokens_seen": 726925312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039381143430290873,
      "loss": 2.732,
      "theoretical_loss": 3.765010644711402,
      "tokens_seen": 726990848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039380140421263796,
      "loss": 2.8741,
      "theoretical_loss": 3.7649764167871824,
      "tokens_seen": 727056384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937913741223671,
      "loss": 2.7782,
      "theoretical_loss": 3.7649421928118745,
      "tokens_seen": 727121920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937813440320963,
      "loss": 2.6986,
      "theoretical_loss": 3.7649079727846653,
      "tokens_seen": 727187456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039377131394182545,
      "loss": 3.0349,
      "theoretical_loss": 3.764873756704745,
      "tokens_seen": 727252992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937612838515547,
      "loss": 2.8043,
      "theoretical_loss": 3.764839544571301,
      "tokens_seen": 727318528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039375125376128387,
      "loss": 2.7541,
      "theoretical_loss": 3.7648053363835237,
      "tokens_seen": 727384064
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 409693,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.393568992614746,
      "objective/train/theoretical_loss": 3.764771132140602,
      "objective/train/tokens_used": 747909600,
      "theoretical_loss": 3.764771132140602,
      "tokens_seen": 727449600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039374122367101305,
      "loss": 2.8103,
      "theoretical_loss": 3.764771132140602,
      "tokens_seen": 727449600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039373119358074223,
      "loss": 2.768,
      "theoretical_loss": 3.7647369318417265,
      "tokens_seen": 727515136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937211634904714,
      "loss": 2.853,
      "theoretical_loss": 3.764702735486086,
      "tokens_seen": 727580672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937111334002006,
      "loss": 2.7399,
      "theoretical_loss": 3.764668543072872,
      "tokens_seen": 727646208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039370110330992983,
      "loss": 2.8607,
      "theoretical_loss": 3.764634354601274,
      "tokens_seen": 727711744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039369107321965896,
      "loss": 2.9141,
      "theoretical_loss": 3.7646001700704836,
      "tokens_seen": 727777280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936810431293882,
      "loss": 2.9176,
      "theoretical_loss": 3.7645659894796903,
      "tokens_seen": 727842816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936710130391173,
      "loss": 2.7348,
      "theoretical_loss": 3.7645318128280865,
      "tokens_seen": 727908352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039366098294884655,
      "loss": 2.9764,
      "theoretical_loss": 3.7644976401148638,
      "tokens_seen": 727973888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039365095285857573,
      "loss": 2.6341,
      "theoretical_loss": 3.764463471339213,
      "tokens_seen": 728039424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936409227683049,
      "loss": 2.823,
      "theoretical_loss": 3.7644293065003263,
      "tokens_seen": 728104960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936308926780341,
      "loss": 2.6593,
      "theoretical_loss": 3.764395145597396,
      "tokens_seen": 728170496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039362086258776333,
      "loss": 2.7037,
      "theoretical_loss": 3.7643609886296145,
      "tokens_seen": 728236032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039361083249749246,
      "loss": 2.6955,
      "theoretical_loss": 3.764326835596174,
      "tokens_seen": 728301568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936008024072217,
      "loss": 2.6841,
      "theoretical_loss": 3.7642926864962685,
      "tokens_seen": 728367104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935907723169508,
      "loss": 2.8496,
      "theoretical_loss": 3.76425854132909,
      "tokens_seen": 728432640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039358074222668006,
      "loss": 2.7119,
      "theoretical_loss": 3.7642244000938314,
      "tokens_seen": 728498176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039357071213640924,
      "loss": 2.7208,
      "theoretical_loss": 3.764190262789688,
      "tokens_seen": 728563712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935606820461384,
      "loss": 2.5211,
      "theoretical_loss": 3.764156129415852,
      "tokens_seen": 728629248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935506519558676,
      "loss": 2.6906,
      "theoretical_loss": 3.7641219999715174,
      "tokens_seen": 728694784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935406218655968,
      "loss": 2.8027,
      "theoretical_loss": 3.76408787445588,
      "tokens_seen": 728760320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039353059177532596,
      "loss": 2.7496,
      "theoretical_loss": 3.7640537528681333,
      "tokens_seen": 728825856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935205616850552,
      "loss": 2.7888,
      "theoretical_loss": 3.7640196352074726,
      "tokens_seen": 728891392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935105315947843,
      "loss": 2.8002,
      "theoretical_loss": 3.7639855214730917,
      "tokens_seen": 728956928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039350050150451356,
      "loss": 2.8055,
      "theoretical_loss": 3.763951411664187,
      "tokens_seen": 729022464
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 410670,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.842893123626709,
      "objective/train/theoretical_loss": 3.7639173057799535,
      "objective/train/tokens_used": 749548000,
      "theoretical_loss": 3.7639173057799535,
      "tokens_seen": 729088000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934904714142427,
      "loss": 2.8283,
      "theoretical_loss": 3.7639173057799535,
      "tokens_seen": 729088000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934804413239719,
      "loss": 2.8309,
      "theoretical_loss": 3.7638832038195877,
      "tokens_seen": 729153536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934704112337011,
      "loss": 2.8713,
      "theoretical_loss": 3.763849105782284,
      "tokens_seen": 729219072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934603811434303,
      "loss": 2.6372,
      "theoretical_loss": 3.7638150116672398,
      "tokens_seen": 729284608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039345035105315946,
      "loss": 2.6612,
      "theoretical_loss": 3.763780921473651,
      "tokens_seen": 729350144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934403209628887,
      "loss": 2.7433,
      "theoretical_loss": 3.7637468352007146,
      "tokens_seen": 729415680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039343029087261783,
      "loss": 2.681,
      "theoretical_loss": 3.7637127528476277,
      "tokens_seen": 729481216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039342026078234706,
      "loss": 2.8188,
      "theoretical_loss": 3.763678674413586,
      "tokens_seen": 729546752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934102306920762,
      "loss": 2.8761,
      "theoretical_loss": 3.7636445998977885,
      "tokens_seen": 729612288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934002006018054,
      "loss": 2.7117,
      "theoretical_loss": 3.7636105292994326,
      "tokens_seen": 729677824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039339017051153466,
      "loss": 3.0125,
      "theoretical_loss": 3.7635764626177153,
      "tokens_seen": 729743360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933801404212638,
      "loss": 2.9,
      "theoretical_loss": 3.7635423998518354,
      "tokens_seen": 729808896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393370110330993,
      "loss": 3.011,
      "theoretical_loss": 3.763508341000991,
      "tokens_seen": 729874432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039336008024072215,
      "loss": 2.765,
      "theoretical_loss": 3.76347428606438,
      "tokens_seen": 729939968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933500501504514,
      "loss": 2.8786,
      "theoretical_loss": 3.7634402350412026,
      "tokens_seen": 730005504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039334002006018057,
      "loss": 2.7195,
      "theoretical_loss": 3.7634061879306566,
      "tokens_seen": 730071040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039332998996990975,
      "loss": 2.8264,
      "theoretical_loss": 3.7633721447319415,
      "tokens_seen": 730136576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039331995987963893,
      "loss": 2.8929,
      "theoretical_loss": 3.763338105444257,
      "tokens_seen": 730202112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039330992978936816,
      "loss": 2.9521,
      "theoretical_loss": 3.763304070066803,
      "tokens_seen": 730267648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932998996990973,
      "loss": 2.9408,
      "theoretical_loss": 3.7632700385987787,
      "tokens_seen": 730333184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932898696088265,
      "loss": 2.7265,
      "theoretical_loss": 3.7632360110393854,
      "tokens_seen": 730398720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039327983951855565,
      "loss": 2.6464,
      "theoretical_loss": 3.7632019873878226,
      "tokens_seen": 730464256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932698094282849,
      "loss": 2.8734,
      "theoretical_loss": 3.763167967643292,
      "tokens_seen": 730529792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039325977933801407,
      "loss": 2.8426,
      "theoretical_loss": 3.763133951804993,
      "tokens_seen": 730595328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039324974924774325,
      "loss": 2.6467,
      "theoretical_loss": 3.763099939872128,
      "tokens_seen": 730660864
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 411160,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.812192678451538,
      "objective/train/theoretical_loss": 3.763065931843898,
      "objective/train/tokens_used": 751186400,
      "theoretical_loss": 3.763065931843898,
      "tokens_seen": 730726400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039323971915747243,
      "loss": 2.8428,
      "theoretical_loss": 3.763065931843898,
      "tokens_seen": 730726400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932296890672016,
      "loss": 2.6838,
      "theoretical_loss": 3.7630319277195046,
      "tokens_seen": 730791936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932196589769308,
      "loss": 3.172,
      "theoretical_loss": 3.7629979274981498,
      "tokens_seen": 730857472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039320962888666003,
      "loss": 2.6505,
      "theoretical_loss": 3.7629639311790353,
      "tokens_seen": 730923008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039319959879638916,
      "loss": 2.8953,
      "theoretical_loss": 3.762929938761363,
      "tokens_seen": 730988544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931895687061184,
      "loss": 2.8217,
      "theoretical_loss": 3.762895950244337,
      "tokens_seen": 731054080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931795386158475,
      "loss": 2.6984,
      "theoretical_loss": 3.7628619656271587,
      "tokens_seen": 731119616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039316950852557675,
      "loss": 2.958,
      "theoretical_loss": 3.7628279849090314,
      "tokens_seen": 731185152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039315947843530593,
      "loss": 2.8194,
      "theoretical_loss": 3.762794008089158,
      "tokens_seen": 731250688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931494483450351,
      "loss": 2.8385,
      "theoretical_loss": 3.7627600351667434,
      "tokens_seen": 731316224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931394182547643,
      "loss": 2.7625,
      "theoretical_loss": 3.7627260661409903,
      "tokens_seen": 731381760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039312938816449353,
      "loss": 2.675,
      "theoretical_loss": 3.762692101011103,
      "tokens_seen": 731447296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039311935807422266,
      "loss": 2.8194,
      "theoretical_loss": 3.762658139776285,
      "tokens_seen": 731512832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931093279839519,
      "loss": 2.9804,
      "theoretical_loss": 3.762624182435741,
      "tokens_seen": 731578368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393099297893681,
      "loss": 2.8233,
      "theoretical_loss": 3.7625902289886763,
      "tokens_seen": 731643904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039308926780341026,
      "loss": 3.0054,
      "theoretical_loss": 3.762556279434295,
      "tokens_seen": 731709440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039307923771313944,
      "loss": 2.9656,
      "theoretical_loss": 3.7625223337718023,
      "tokens_seen": 731774976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930692076228686,
      "loss": 2.7612,
      "theoretical_loss": 3.762488392000404,
      "tokens_seen": 731840512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930591775325978,
      "loss": 2.9062,
      "theoretical_loss": 3.7624544541193057,
      "tokens_seen": 731906048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393049147442327,
      "loss": 2.5917,
      "theoretical_loss": 3.762420520127713,
      "tokens_seen": 731971584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039303911735205616,
      "loss": 2.8431,
      "theoretical_loss": 3.762386590024832,
      "tokens_seen": 732037120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930290872617854,
      "loss": 2.8048,
      "theoretical_loss": 3.7623526638098683,
      "tokens_seen": 732102656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930190571715145,
      "loss": 2.6144,
      "theoretical_loss": 3.76231874148203,
      "tokens_seen": 732168192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039300902708124376,
      "loss": 2.6214,
      "theoretical_loss": 3.762284823040522,
      "tokens_seen": 732233728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929989969909729,
      "loss": 2.866,
      "theoretical_loss": 3.762250908484553,
      "tokens_seen": 732299264
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 412414,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.900200366973877,
      "objective/train/theoretical_loss": 3.7622169978133293,
      "objective/train/tokens_used": 752824800,
      "theoretical_loss": 3.7622169978133293,
      "tokens_seen": 732364800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929889669007021,
      "loss": 3.019,
      "theoretical_loss": 3.7622169978133293,
      "tokens_seen": 732364800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929789368104313,
      "loss": 2.8027,
      "theoretical_loss": 3.762183091026058,
      "tokens_seen": 732430336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929689067201605,
      "loss": 2.7653,
      "theoretical_loss": 3.7621491881219478,
      "tokens_seen": 732495872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039295887662988967,
      "loss": 2.5643,
      "theoretical_loss": 3.7621152891002057,
      "tokens_seen": 732561408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929488465396189,
      "loss": 2.5249,
      "theoretical_loss": 3.7620813939600404,
      "tokens_seen": 732626944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039293881644934803,
      "loss": 2.7563,
      "theoretical_loss": 3.76204750270066,
      "tokens_seen": 732692480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039292878635907726,
      "loss": 2.6964,
      "theoretical_loss": 3.7620136153212735,
      "tokens_seen": 732758016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929187562688064,
      "loss": 2.9321,
      "theoretical_loss": 3.7619797318210892,
      "tokens_seen": 732823552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929087261785356,
      "loss": 2.9943,
      "theoretical_loss": 3.7619458521993168,
      "tokens_seen": 732889088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928986960882648,
      "loss": 3.0453,
      "theoretical_loss": 3.761911976455165,
      "tokens_seen": 732954624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392888665997994,
      "loss": 2.7816,
      "theoretical_loss": 3.7618781045878436,
      "tokens_seen": 733020160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039287863590772317,
      "loss": 2.9523,
      "theoretical_loss": 3.7618442365965628,
      "tokens_seen": 733085696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039286860581745235,
      "loss": 2.7939,
      "theoretical_loss": 3.7618103724805314,
      "tokens_seen": 733151232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039285857572718153,
      "loss": 2.8391,
      "theoretical_loss": 3.7617765122389613,
      "tokens_seen": 733216768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039284854563691077,
      "loss": 2.8972,
      "theoretical_loss": 3.7617426558710614,
      "tokens_seen": 733282304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928385155466399,
      "loss": 2.7244,
      "theoretical_loss": 3.7617088033760435,
      "tokens_seen": 733347840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039282848545636913,
      "loss": 2.5487,
      "theoretical_loss": 3.7616749547531185,
      "tokens_seen": 733413376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039281845536609825,
      "loss": 2.4826,
      "theoretical_loss": 3.7616411100014964,
      "tokens_seen": 733478912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928084252758275,
      "loss": 2.7962,
      "theoretical_loss": 3.7616072691203897,
      "tokens_seen": 733544448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039279839518555667,
      "loss": 2.9535,
      "theoretical_loss": 3.76157343210901,
      "tokens_seen": 733609984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039278836509528585,
      "loss": 2.9491,
      "theoretical_loss": 3.7615395989665688,
      "tokens_seen": 733675520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039277833500501503,
      "loss": 2.9444,
      "theoretical_loss": 3.7615057696922785,
      "tokens_seen": 733741056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039276830491474427,
      "loss": 2.8252,
      "theoretical_loss": 3.7614719442853506,
      "tokens_seen": 733806592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927582748244734,
      "loss": 2.7134,
      "theoretical_loss": 3.7614381227449982,
      "tokens_seen": 733872128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039274824473420263,
      "loss": 2.8441,
      "theoretical_loss": 3.7614043050704344,
      "tokens_seen": 733937664
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 413470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.670003652572632,
      "objective/train/theoretical_loss": 3.7613704912608723,
      "objective/train/tokens_used": 754463200,
      "theoretical_loss": 3.7613704912608723,
      "tokens_seen": 734003200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039273821464393176,
      "loss": 2.9001,
      "theoretical_loss": 3.7613704912608723,
      "tokens_seen": 734003200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392728184553661,
      "loss": 2.7337,
      "theoretical_loss": 3.761336681315524,
      "tokens_seen": 734068736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927181544633902,
      "loss": 2.5806,
      "theoretical_loss": 3.7613028752336044,
      "tokens_seen": 734134272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039270812437311936,
      "loss": 2.8466,
      "theoretical_loss": 3.761269073014326,
      "tokens_seen": 734199808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039269809428284854,
      "loss": 2.7271,
      "theoretical_loss": 3.7612352746569035,
      "tokens_seen": 734265344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926880641925777,
      "loss": 2.8845,
      "theoretical_loss": 3.761201480160551,
      "tokens_seen": 734330880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926780341023069,
      "loss": 2.6129,
      "theoretical_loss": 3.7611676895244828,
      "tokens_seen": 734396416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039266800401203613,
      "loss": 2.646,
      "theoretical_loss": 3.761133902747913,
      "tokens_seen": 734461952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039265797392176526,
      "loss": 2.6668,
      "theoretical_loss": 3.7611001198300578,
      "tokens_seen": 734527488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926479438314945,
      "loss": 2.625,
      "theoretical_loss": 3.7610663407701304,
      "tokens_seen": 734593024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039263791374122373,
      "loss": 2.9871,
      "theoretical_loss": 3.7610325655673478,
      "tokens_seen": 734658560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039262788365095286,
      "loss": 2.8132,
      "theoretical_loss": 3.7609987942209244,
      "tokens_seen": 734724096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926178535606821,
      "loss": 2.7216,
      "theoretical_loss": 3.7609650267300765,
      "tokens_seen": 734789632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926078234704112,
      "loss": 2.7455,
      "theoretical_loss": 3.7609312630940197,
      "tokens_seen": 734855168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039259779338014046,
      "loss": 2.6653,
      "theoretical_loss": 3.7608975033119707,
      "tokens_seen": 734920704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039258776328986964,
      "loss": 2.6471,
      "theoretical_loss": 3.760863747383146,
      "tokens_seen": 734986240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925777331995988,
      "loss": 2.8097,
      "theoretical_loss": 3.7608299953067625,
      "tokens_seen": 735051776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392567703109328,
      "loss": 2.8716,
      "theoretical_loss": 3.760796247082036,
      "tokens_seen": 735117312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925576730190572,
      "loss": 2.8082,
      "theoretical_loss": 3.760762502708185,
      "tokens_seen": 735182848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039254764292878636,
      "loss": 2.6588,
      "theoretical_loss": 3.760728762184425,
      "tokens_seen": 735248384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925376128385156,
      "loss": 2.731,
      "theoretical_loss": 3.7606950255099765,
      "tokens_seen": 735313920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925275827482447,
      "loss": 2.8463,
      "theoretical_loss": 3.7606612926840546,
      "tokens_seen": 735379456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039251755265797396,
      "loss": 2.6797,
      "theoretical_loss": 3.760627563705879,
      "tokens_seen": 735444992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925075225677031,
      "loss": 2.7702,
      "theoretical_loss": 3.7605938385746667,
      "tokens_seen": 735510528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924974924774323,
      "loss": 2.8931,
      "theoretical_loss": 3.7605601172896366,
      "tokens_seen": 735576064
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 413970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.987671375274658,
      "objective/train/theoretical_loss": 3.7605263998500087,
      "objective/train/tokens_used": 756101600,
      "theoretical_loss": 3.7605263998500087,
      "tokens_seen": 735641600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924874623871615,
      "loss": 2.9723,
      "theoretical_loss": 3.7605263998500087,
      "tokens_seen": 735641600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924774322968907,
      "loss": 2.759,
      "theoretical_loss": 3.7604926862550005,
      "tokens_seen": 735707136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039246740220661987,
      "loss": 3.0025,
      "theoretical_loss": 3.7604589765038323,
      "tokens_seen": 735772672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924573721163491,
      "loss": 2.7319,
      "theoretical_loss": 3.7604252705957224,
      "tokens_seen": 735838208
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039244734202607823,
      "loss": 2.8027,
      "theoretical_loss": 3.7603915685298914,
      "tokens_seen": 735903744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039243731193580746,
      "loss": 2.7283,
      "theoretical_loss": 3.7603578703055587,
      "tokens_seen": 735969280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924272818455366,
      "loss": 2.699,
      "theoretical_loss": 3.7603241759219443,
      "tokens_seen": 736034816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924172517552658,
      "loss": 2.9509,
      "theoretical_loss": 3.7602904853782695,
      "tokens_seen": 736100352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392407221664995,
      "loss": 2.468,
      "theoretical_loss": 3.7602567986737534,
      "tokens_seen": 736165888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923971915747242,
      "loss": 2.6287,
      "theoretical_loss": 3.760223115807618,
      "tokens_seen": 736231424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039238716148445337,
      "loss": 2.8342,
      "theoretical_loss": 3.7601894367790836,
      "tokens_seen": 736296960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039237713139418255,
      "loss": 2.7653,
      "theoretical_loss": 3.760155761587372,
      "tokens_seen": 736362496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039236710130391173,
      "loss": 3.1294,
      "theoretical_loss": 3.760122090231704,
      "tokens_seen": 736428032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039235707121364097,
      "loss": 2.7462,
      "theoretical_loss": 3.760088422711302,
      "tokens_seen": 736493568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923470411233701,
      "loss": 2.7666,
      "theoretical_loss": 3.7600547590253877,
      "tokens_seen": 736559104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039233701103309933,
      "loss": 2.5778,
      "theoretical_loss": 3.7600210991731826,
      "tokens_seen": 736624640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039232698094282846,
      "loss": 2.6942,
      "theoretical_loss": 3.75998744315391,
      "tokens_seen": 736690176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923169508525577,
      "loss": 2.8405,
      "theoretical_loss": 3.759953790966793,
      "tokens_seen": 736755712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039230692076228687,
      "loss": 2.7331,
      "theoretical_loss": 3.759920142611053,
      "tokens_seen": 736821248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039229689067201605,
      "loss": 2.8218,
      "theoretical_loss": 3.7598864980859132,
      "tokens_seen": 736886784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039228686058174523,
      "loss": 2.6316,
      "theoretical_loss": 3.759852857390598,
      "tokens_seen": 736952320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039227683049147447,
      "loss": 2.8765,
      "theoretical_loss": 3.75981922052433,
      "tokens_seen": 737017856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922668004012036,
      "loss": 2.7292,
      "theoretical_loss": 3.7597855874863333,
      "tokens_seen": 737083392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039225677031093283,
      "loss": 2.7402,
      "theoretical_loss": 3.7597519582758325,
      "tokens_seen": 737148928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039224674022066196,
      "loss": 2.6244,
      "theoretical_loss": 3.7597183328920507,
      "tokens_seen": 737214464
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 414896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8144378662109375,
      "objective/train/theoretical_loss": 3.759684711334213,
      "objective/train/tokens_used": 757740000,
      "theoretical_loss": 3.759684711334213,
      "tokens_seen": 737280000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922367101303912,
      "loss": 2.7107,
      "theoretical_loss": 3.759684711334213,
      "tokens_seen": 737280000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922266800401204,
      "loss": 2.8554,
      "theoretical_loss": 3.7596510936015433,
      "tokens_seen": 737345536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039221664994984956,
      "loss": 2.7267,
      "theoretical_loss": 3.759617479693267,
      "tokens_seen": 737411072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039220661985957874,
      "loss": 2.9452,
      "theoretical_loss": 3.75958386960861,
      "tokens_seen": 737476608
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921965897693079,
      "loss": 2.7452,
      "theoretical_loss": 3.759550263346796,
      "tokens_seen": 737542144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921865596790371,
      "loss": 3.0845,
      "theoretical_loss": 3.759516660907052,
      "tokens_seen": 737607680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039217652958876633,
      "loss": 2.5286,
      "theoretical_loss": 3.759483062288603,
      "tokens_seen": 737673216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039216649949849546,
      "loss": 2.8046,
      "theoretical_loss": 3.759449467490675,
      "tokens_seen": 737738752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921564694082247,
      "loss": 2.9167,
      "theoretical_loss": 3.7594158765124943,
      "tokens_seen": 737804288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921464393179539,
      "loss": 2.9723,
      "theoretical_loss": 3.759382289353288,
      "tokens_seen": 737869824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039213640922768306,
      "loss": 2.7637,
      "theoretical_loss": 3.759348706012282,
      "tokens_seen": 737935360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039212637913741224,
      "loss": 2.8318,
      "theoretical_loss": 3.759315126488704,
      "tokens_seen": 738000896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921163490471414,
      "loss": 2.885,
      "theoretical_loss": 3.75928155078178,
      "tokens_seen": 738066432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921063189568706,
      "loss": 2.8873,
      "theoretical_loss": 3.7592479788907376,
      "tokens_seen": 738131968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039209628886659984,
      "loss": 2.8126,
      "theoretical_loss": 3.7592144108148053,
      "tokens_seen": 738197504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039208625877632896,
      "loss": 2.7385,
      "theoretical_loss": 3.7591808465532104,
      "tokens_seen": 738263040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920762286860582,
      "loss": 2.6291,
      "theoretical_loss": 3.759147286105181,
      "tokens_seen": 738328576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920661985957873,
      "loss": 2.9438,
      "theoretical_loss": 3.7591137294699446,
      "tokens_seen": 738394112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039205616850551656,
      "loss": 2.8742,
      "theoretical_loss": 3.759080176646731,
      "tokens_seen": 738459648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039204613841524574,
      "loss": 2.5495,
      "theoretical_loss": 3.759046627634768,
      "tokens_seen": 738525184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920361083249749,
      "loss": 2.5249,
      "theoretical_loss": 3.759013082433285,
      "tokens_seen": 738590720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920260782347041,
      "loss": 2.7102,
      "theoretical_loss": 3.7589795410415103,
      "tokens_seen": 738656256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920160481444333,
      "loss": 2.9134,
      "theoretical_loss": 3.758946003458675,
      "tokens_seen": 738721792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039200601805416247,
      "loss": 2.8764,
      "theoretical_loss": 3.758912469684007,
      "tokens_seen": 738787328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919959879638917,
      "loss": 2.7331,
      "theoretical_loss": 3.7588789397167366,
      "tokens_seen": 738852864
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 415468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.898322820663452,
      "objective/train/theoretical_loss": 3.7588454135560943,
      "objective/train/tokens_used": 759378400,
      "theoretical_loss": 3.7588454135560943,
      "tokens_seen": 738918400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039198595787362083,
      "loss": 2.7734,
      "theoretical_loss": 3.7588454135560943,
      "tokens_seen": 738918400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039197592778335007,
      "loss": 2.7841,
      "theoretical_loss": 3.7588118912013098,
      "tokens_seen": 738983936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039196589769307925,
      "loss": 2.8882,
      "theoretical_loss": 3.758778372651614,
      "tokens_seen": 739049472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039195586760280843,
      "loss": 2.8642,
      "theoretical_loss": 3.7587448579062377,
      "tokens_seen": 739115008
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919458375125376,
      "loss": 2.7614,
      "theoretical_loss": 3.758711346964411,
      "tokens_seen": 739180544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919358074222668,
      "loss": 2.8015,
      "theoretical_loss": 3.7586778398253666,
      "tokens_seen": 739246080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039192577733199597,
      "loss": 2.893,
      "theoretical_loss": 3.7586443364883344,
      "tokens_seen": 739311616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919157472417252,
      "loss": 2.9149,
      "theoretical_loss": 3.7586108369525473,
      "tokens_seen": 739377152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039190571715145433,
      "loss": 2.5778,
      "theoretical_loss": 3.7585773412172356,
      "tokens_seen": 739442688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039189568706118357,
      "loss": 2.9246,
      "theoretical_loss": 3.758543849281633,
      "tokens_seen": 739508224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039188565697091275,
      "loss": 2.844,
      "theoretical_loss": 3.758510361144971,
      "tokens_seen": 739573760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039187562688064193,
      "loss": 2.7873,
      "theoretical_loss": 3.7584768768064816,
      "tokens_seen": 739639296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039186559679037117,
      "loss": 2.9233,
      "theoretical_loss": 3.758443396265399,
      "tokens_seen": 739704832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918555667001003,
      "loss": 2.7354,
      "theoretical_loss": 3.758409919520955,
      "tokens_seen": 739770368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039184553660982953,
      "loss": 2.8651,
      "theoretical_loss": 3.7583764465723823,
      "tokens_seen": 739835904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039183550651955866,
      "loss": 2.791,
      "theoretical_loss": 3.758342977418916,
      "tokens_seen": 739901440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918254764292879,
      "loss": 2.7292,
      "theoretical_loss": 3.758309512059788,
      "tokens_seen": 739966976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039181544633901707,
      "loss": 2.9291,
      "theoretical_loss": 3.758276050494233,
      "tokens_seen": 740032512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039180541624874625,
      "loss": 2.8555,
      "theoretical_loss": 3.7582425927214853,
      "tokens_seen": 740098048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039179538615847543,
      "loss": 2.8255,
      "theoretical_loss": 3.7582091387407788,
      "tokens_seen": 740163584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039178535606820467,
      "loss": 2.6274,
      "theoretical_loss": 3.758175688551348,
      "tokens_seen": 740229120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917753259779338,
      "loss": 2.7021,
      "theoretical_loss": 3.758142242152428,
      "tokens_seen": 740294656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039176529588766303,
      "loss": 2.7977,
      "theoretical_loss": 3.758108799543254,
      "tokens_seen": 740360192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039175526579739216,
      "loss": 3.0315,
      "theoretical_loss": 3.7580753607230593,
      "tokens_seen": 740425728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917452357071214,
      "loss": 2.8211,
      "theoretical_loss": 3.7580419256910815,
      "tokens_seen": 740491264
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 416744,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.713137626647949,
      "objective/train/theoretical_loss": 3.7580084944465555,
      "objective/train/tokens_used": 761016800,
      "theoretical_loss": 3.7580084944465555,
      "tokens_seen": 740556800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917352056168506,
      "loss": 2.8892,
      "theoretical_loss": 3.7580084944465555,
      "tokens_seen": 740556800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039172517552657976,
      "loss": 2.933,
      "theoretical_loss": 3.757975066988717,
      "tokens_seen": 740622336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039171514543630894,
      "loss": 2.7738,
      "theoretical_loss": 3.7579416433168022,
      "tokens_seen": 740687872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917051153460381,
      "loss": 3.0515,
      "theoretical_loss": 3.757908223430047,
      "tokens_seen": 740753408
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916950852557673,
      "loss": 2.7393,
      "theoretical_loss": 3.7578748073276884,
      "tokens_seen": 740818944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039168505516549653,
      "loss": 2.5972,
      "theoretical_loss": 3.7578413950089633,
      "tokens_seen": 740884480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039167502507522566,
      "loss": 2.8358,
      "theoretical_loss": 3.757807986473108,
      "tokens_seen": 740950016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916649949849549,
      "loss": 2.6884,
      "theoretical_loss": 3.75777458171936,
      "tokens_seen": 741015552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916549648946841,
      "loss": 2.9793,
      "theoretical_loss": 3.757741180746957,
      "tokens_seen": 741081088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039164493480441326,
      "loss": 2.938,
      "theoretical_loss": 3.7577077835551362,
      "tokens_seen": 741146624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039163490471414244,
      "loss": 2.772,
      "theoretical_loss": 3.7576743901431353,
      "tokens_seen": 741212160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916248746238716,
      "loss": 2.7569,
      "theoretical_loss": 3.7576410005101932,
      "tokens_seen": 741277696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916148445336008,
      "loss": 2.74,
      "theoretical_loss": 3.757607614655547,
      "tokens_seen": 741343232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039160481444333004,
      "loss": 2.8583,
      "theoretical_loss": 3.7575742325784365,
      "tokens_seen": 741408768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039159478435305916,
      "loss": 2.9713,
      "theoretical_loss": 3.7575408542780995,
      "tokens_seen": 741474304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915847542627884,
      "loss": 2.6112,
      "theoretical_loss": 3.757507479753775,
      "tokens_seen": 741539840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915747241725175,
      "loss": 2.6381,
      "theoretical_loss": 3.7574741090047024,
      "tokens_seen": 741605376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039156469408224676,
      "loss": 2.6271,
      "theoretical_loss": 3.7574407420301217,
      "tokens_seen": 741670912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039155466399197594,
      "loss": 2.7429,
      "theoretical_loss": 3.7574073788292712,
      "tokens_seen": 741736448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915446339017051,
      "loss": 2.7142,
      "theoretical_loss": 3.7573740194013916,
      "tokens_seen": 741801984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915346038114343,
      "loss": 2.7973,
      "theoretical_loss": 3.7573406637457225,
      "tokens_seen": 741867520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915245737211635,
      "loss": 2.6915,
      "theoretical_loss": 3.757307311861505,
      "tokens_seen": 741933056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039151454363089267,
      "loss": 2.9268,
      "theoretical_loss": 3.757273963747979,
      "tokens_seen": 741998592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915045135406219,
      "loss": 2.5838,
      "theoretical_loss": 3.7572406194043846,
      "tokens_seen": 742064128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039149448345035103,
      "loss": 2.7691,
      "theoretical_loss": 3.7572072788299637,
      "tokens_seen": 742129664
    },
    {
      "epoch": 0.22,
      "objective/train/docs_used": 417361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1539878845214844,
      "objective/train/theoretical_loss": 3.7571739420239574,
      "objective/train/tokens_used": 762655200,
      "theoretical_loss": 3.7571739420239574,
      "tokens_seen": 742195200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039148445336008027,
      "loss": 2.8494,
      "theoretical_loss": 3.7571739420239574,
      "tokens_seen": 742195200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039147442326980945,
      "loss": 2.7746,
      "theoretical_loss": 3.757140608985606,
      "tokens_seen": 742260736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039146439317953863,
      "loss": 2.8907,
      "theoretical_loss": 3.7571072797141523,
      "tokens_seen": 742326272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914543630892678,
      "loss": 2.8257,
      "theoretical_loss": 3.7570739542088374,
      "tokens_seen": 742391808
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000391444332998997,
      "loss": 2.9461,
      "theoretical_loss": 3.757040632468904,
      "tokens_seen": 742457344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039143430290872617,
      "loss": 2.7556,
      "theoretical_loss": 3.7570073144935936,
      "tokens_seen": 742522880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003914242728184554,
      "loss": 2.8871,
      "theoretical_loss": 3.7569740002821495,
      "tokens_seen": 742588416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039141424272818453,
      "loss": 2.8353,
      "theoretical_loss": 3.7569406898338134,
      "tokens_seen": 742653952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039140421263791377,
      "loss": 2.7706,
      "theoretical_loss": 3.756907383147829,
      "tokens_seen": 742719488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913941825476429,
      "loss": 2.7584,
      "theoretical_loss": 3.756874080223439,
      "tokens_seen": 742785024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039138415245737213,
      "loss": 2.802,
      "theoretical_loss": 3.7568407810598865,
      "tokens_seen": 742850560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913741223671013,
      "loss": 2.6823,
      "theoretical_loss": 3.756807485656416,
      "tokens_seen": 742916096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913640922768305,
      "loss": 2.8226,
      "theoretical_loss": 3.7567741940122703,
      "tokens_seen": 742981632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913540621865597,
      "loss": 2.7535,
      "theoretical_loss": 3.7567409061266943,
      "tokens_seen": 743047168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039134403209628886,
      "loss": 2.5167,
      "theoretical_loss": 3.756707621998931,
      "tokens_seen": 743112704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039133400200601804,
      "loss": 2.7885,
      "theoretical_loss": 3.7566743416282264,
      "tokens_seen": 743178240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039132397191574727,
      "loss": 2.6626,
      "theoretical_loss": 3.7566410650138238,
      "tokens_seen": 743243776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913139418254764,
      "loss": 2.6599,
      "theoretical_loss": 3.756607792154969,
      "tokens_seen": 743309312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039130391173520563,
      "loss": 2.755,
      "theoretical_loss": 3.7565745230509062,
      "tokens_seen": 743374848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912938816449348,
      "loss": 2.7896,
      "theoretical_loss": 3.7565412577008814,
      "tokens_seen": 743440384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391283851554664,
      "loss": 2.5392,
      "theoretical_loss": 3.7565079961041397,
      "tokens_seen": 743505920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912738214643932,
      "loss": 2.7874,
      "theoretical_loss": 3.7564747382599273,
      "tokens_seen": 743571456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039126379137412236,
      "loss": 2.5402,
      "theoretical_loss": 3.75644148416749,
      "tokens_seen": 743636992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039125376128385154,
      "loss": 2.6678,
      "theoretical_loss": 3.7564082338260736,
      "tokens_seen": 743702528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912437311935808,
      "loss": 2.7886,
      "theoretical_loss": 3.756374987234925,
      "tokens_seen": 743768064
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 418521,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7649760246276855,
      "objective/train/theoretical_loss": 3.7563417443932905,
      "objective/train/tokens_used": 764293600,
      "theoretical_loss": 3.7563417443932905,
      "tokens_seen": 743833600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912337011033099,
      "loss": 2.9361,
      "theoretical_loss": 3.7563417443932905,
      "tokens_seen": 743833600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039122367101303914,
      "loss": 2.9609,
      "theoretical_loss": 3.7563085053004173,
      "tokens_seen": 743899136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039121364092276826,
      "loss": 2.7547,
      "theoretical_loss": 3.7562752699555517,
      "tokens_seen": 743964672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912036108324975,
      "loss": 2.5979,
      "theoretical_loss": 3.756242038357942,
      "tokens_seen": 744030208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911935807422267,
      "loss": 2.9392,
      "theoretical_loss": 3.756208810506835,
      "tokens_seen": 744095744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039118355065195586,
      "loss": 2.6047,
      "theoretical_loss": 3.7561755864014787,
      "tokens_seen": 744161280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039117352056168504,
      "loss": 2.8347,
      "theoretical_loss": 3.756142366041121,
      "tokens_seen": 744226816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911634904714143,
      "loss": 2.7851,
      "theoretical_loss": 3.7561091494250096,
      "tokens_seen": 744292352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911534603811434,
      "loss": 2.7388,
      "theoretical_loss": 3.756075936552394,
      "tokens_seen": 744357888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039114343029087264,
      "loss": 2.7189,
      "theoretical_loss": 3.756042727422521,
      "tokens_seen": 744423424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911334002006018,
      "loss": 2.6181,
      "theoretical_loss": 3.756009522034641,
      "tokens_seen": 744488960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391123370110331,
      "loss": 2.7899,
      "theoretical_loss": 3.755976320388002,
      "tokens_seen": 744554496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039111334002006024,
      "loss": 2.7496,
      "theoretical_loss": 3.755943122481854,
      "tokens_seen": 744620032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039110330992978936,
      "loss": 2.9665,
      "theoretical_loss": 3.755909928315446,
      "tokens_seen": 744685568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910932798395186,
      "loss": 2.9362,
      "theoretical_loss": 3.7558767378880273,
      "tokens_seen": 744751104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039108324974924773,
      "loss": 2.5904,
      "theoretical_loss": 3.755843551198849,
      "tokens_seen": 744816640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039107321965897696,
      "loss": 2.6235,
      "theoretical_loss": 3.7558103682471593,
      "tokens_seen": 744882176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039106318956870614,
      "loss": 2.6325,
      "theoretical_loss": 3.7557771890322105,
      "tokens_seen": 744947712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910531594784353,
      "loss": 2.7591,
      "theoretical_loss": 3.755744013553252,
      "tokens_seen": 745013248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910431293881645,
      "loss": 2.6799,
      "theoretical_loss": 3.755710841809534,
      "tokens_seen": 745078784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910330992978937,
      "loss": 2.7825,
      "theoretical_loss": 3.755677673800309,
      "tokens_seen": 745144320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039102306920762287,
      "loss": 2.7031,
      "theoretical_loss": 3.755644509524827,
      "tokens_seen": 745209856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910130391173521,
      "loss": 2.6609,
      "theoretical_loss": 3.75561134898234,
      "tokens_seen": 745275392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039100300902708123,
      "loss": 2.5443,
      "theoretical_loss": 3.7555781921720994,
      "tokens_seen": 745340928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039099297893681047,
      "loss": 2.7118,
      "theoretical_loss": 3.7555450390933567,
      "tokens_seen": 745406464
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 419167,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.160205125808716,
      "objective/train/theoretical_loss": 3.7555118897453648,
      "objective/train/tokens_used": 765932000,
      "theoretical_loss": 3.7555118897453648,
      "tokens_seen": 745472000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039098294884653965,
      "loss": 2.8914,
      "theoretical_loss": 3.7555118897453648,
      "tokens_seen": 745472000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039097291875626883,
      "loss": 2.6525,
      "theoretical_loss": 3.755478744127375,
      "tokens_seen": 745537536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390962888665998,
      "loss": 2.771,
      "theoretical_loss": 3.75544560223864,
      "tokens_seen": 745603072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909528585757272,
      "loss": 2.7758,
      "theoretical_loss": 3.755412464078413,
      "tokens_seen": 745668608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039094282848545637,
      "loss": 2.7658,
      "theoretical_loss": 3.755379329645946,
      "tokens_seen": 745734144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909327983951856,
      "loss": 2.8785,
      "theoretical_loss": 3.7553461989404933,
      "tokens_seen": 745799680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039092276830491473,
      "loss": 2.7476,
      "theoretical_loss": 3.7553130719613073,
      "tokens_seen": 745865216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039091273821464397,
      "loss": 2.5613,
      "theoretical_loss": 3.755279948707642,
      "tokens_seen": 745930752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909027081243731,
      "loss": 2.7278,
      "theoretical_loss": 3.7552468291787506,
      "tokens_seen": 745996288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039089267803410233,
      "loss": 2.4587,
      "theoretical_loss": 3.7552137133738874,
      "tokens_seen": 746061824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908826479438315,
      "loss": 2.873,
      "theoretical_loss": 3.755180601292307,
      "tokens_seen": 746127360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908726178535607,
      "loss": 2.8608,
      "theoretical_loss": 3.7551474929332636,
      "tokens_seen": 746192896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908625877632899,
      "loss": 2.9343,
      "theoretical_loss": 3.755114388296011,
      "tokens_seen": 746258432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039085255767301906,
      "loss": 2.5345,
      "theoretical_loss": 3.7550812873798054,
      "tokens_seen": 746323968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039084252758274824,
      "loss": 2.5794,
      "theoretical_loss": 3.755048190183901,
      "tokens_seen": 746389504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039083249749247747,
      "loss": 2.5169,
      "theoretical_loss": 3.755015096707553,
      "tokens_seen": 746455040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908224674022066,
      "loss": 2.7485,
      "theoretical_loss": 3.754982006950017,
      "tokens_seen": 746520576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039081243731193583,
      "loss": 2.6492,
      "theoretical_loss": 3.7549489209105493,
      "tokens_seen": 746586112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390802407221665,
      "loss": 2.4441,
      "theoretical_loss": 3.754915838588405,
      "tokens_seen": 746651648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907923771313942,
      "loss": 2.6574,
      "theoretical_loss": 3.7548827599828405,
      "tokens_seen": 746717184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907823470411234,
      "loss": 2.8731,
      "theoretical_loss": 3.754849685093112,
      "tokens_seen": 746782720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039077231695085256,
      "loss": 2.6081,
      "theoretical_loss": 3.754816613918476,
      "tokens_seen": 746848256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039076228686058174,
      "loss": 2.744,
      "theoretical_loss": 3.7547835464581896,
      "tokens_seen": 746913792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390752256770311,
      "loss": 2.9731,
      "theoretical_loss": 3.7547504827115104,
      "tokens_seen": 746979328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907422266800401,
      "loss": 2.9428,
      "theoretical_loss": 3.754717422677694,
      "tokens_seen": 747044864
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 420047,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9569199085235596,
      "objective/train/theoretical_loss": 3.754684366355999,
      "objective/train/tokens_used": 767570400,
      "theoretical_loss": 3.754684366355999,
      "tokens_seen": 747110400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039073219658976934,
      "loss": 2.6868,
      "theoretical_loss": 3.754684366355999,
      "tokens_seen": 747110400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039072216649949846,
      "loss": 2.8481,
      "theoretical_loss": 3.7546513137456823,
      "tokens_seen": 747175936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907121364092277,
      "loss": 2.6763,
      "theoretical_loss": 3.7546182648460023,
      "tokens_seen": 747241472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907021063189569,
      "loss": 2.541,
      "theoretical_loss": 3.754585219656217,
      "tokens_seen": 747307008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039069207622868606,
      "loss": 2.811,
      "theoretical_loss": 3.754552178175584,
      "tokens_seen": 747372544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039068204613841524,
      "loss": 2.8739,
      "theoretical_loss": 3.7545191404033624,
      "tokens_seen": 747438080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906720160481445,
      "loss": 2.7239,
      "theoretical_loss": 3.7544861063388106,
      "tokens_seen": 747503616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906619859578736,
      "loss": 2.7788,
      "theoretical_loss": 3.7544530759811883,
      "tokens_seen": 747569152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039065195586760284,
      "loss": 2.7612,
      "theoretical_loss": 3.7544200493297533,
      "tokens_seen": 747634688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039064192577733197,
      "loss": 2.4776,
      "theoretical_loss": 3.7543870263837658,
      "tokens_seen": 747700224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906318956870612,
      "loss": 2.6039,
      "theoretical_loss": 3.754354007142485,
      "tokens_seen": 747765760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906218655967904,
      "loss": 2.8779,
      "theoretical_loss": 3.754320991605171,
      "tokens_seen": 747831296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039061183550651957,
      "loss": 2.6714,
      "theoretical_loss": 3.754287979771083,
      "tokens_seen": 747896832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039060180541624875,
      "loss": 2.8675,
      "theoretical_loss": 3.754254971639482,
      "tokens_seen": 747962368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039059177532597793,
      "loss": 2.8031,
      "theoretical_loss": 3.754221967209628,
      "tokens_seen": 748027904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905817452357071,
      "loss": 2.5851,
      "theoretical_loss": 3.754188966480782,
      "tokens_seen": 748093440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039057171514543634,
      "loss": 2.6848,
      "theoretical_loss": 3.754155969452204,
      "tokens_seen": 748158976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039056168505516547,
      "loss": 2.8457,
      "theoretical_loss": 3.7541229761231563,
      "tokens_seen": 748224512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905516549648947,
      "loss": 2.8551,
      "theoretical_loss": 3.7540899864928985,
      "tokens_seen": 748290048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039054162487462383,
      "loss": 2.9909,
      "theoretical_loss": 3.7540570005606932,
      "tokens_seen": 748355584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039053159478435307,
      "loss": 2.5867,
      "theoretical_loss": 3.754024018325802,
      "tokens_seen": 748421120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039052156469408225,
      "loss": 2.7651,
      "theoretical_loss": 3.7539910397874863,
      "tokens_seen": 748486656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039051153460381143,
      "loss": 2.7933,
      "theoretical_loss": 3.7539580649450084,
      "tokens_seen": 748552192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905015045135406,
      "loss": 2.9215,
      "theoretical_loss": 3.7539250937976307,
      "tokens_seen": 748617728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039049147442326985,
      "loss": 2.6618,
      "theoretical_loss": 3.7538921263446157,
      "tokens_seen": 748683264
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 421176,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7754266262054443,
      "objective/train/theoretical_loss": 3.753859162585226,
      "objective/train/tokens_used": 769208800,
      "theoretical_loss": 3.753859162585226,
      "tokens_seen": 748748800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390481444332999,
      "loss": 2.8535,
      "theoretical_loss": 3.753859162585226,
      "tokens_seen": 748748800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904714142427282,
      "loss": 2.8332,
      "theoretical_loss": 3.7538262025187246,
      "tokens_seen": 748814336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039046138415245734,
      "loss": 2.6704,
      "theoretical_loss": 3.753793246144374,
      "tokens_seen": 748879872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039045135406218657,
      "loss": 2.8584,
      "theoretical_loss": 3.753760293461439,
      "tokens_seen": 748945408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039044132397191575,
      "loss": 2.7273,
      "theoretical_loss": 3.7537273444691825,
      "tokens_seen": 749010944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039043129388164493,
      "loss": 3.0141,
      "theoretical_loss": 3.753694399166867,
      "tokens_seen": 749076480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904212637913741,
      "loss": 2.5281,
      "theoretical_loss": 3.7536614575537586,
      "tokens_seen": 749142016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904112337011033,
      "loss": 2.7591,
      "theoretical_loss": 3.75362851962912,
      "tokens_seen": 749207552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904012036108325,
      "loss": 2.7426,
      "theoretical_loss": 3.7535955853922167,
      "tokens_seen": 749273088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903911735205617,
      "loss": 2.9189,
      "theoretical_loss": 3.7535626548423124,
      "tokens_seen": 749338624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903811434302909,
      "loss": 2.7107,
      "theoretical_loss": 3.7535297279786723,
      "tokens_seen": 749404160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903711133400201,
      "loss": 2.8125,
      "theoretical_loss": 3.7534968048005615,
      "tokens_seen": 749469696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039036108324974926,
      "loss": 2.8113,
      "theoretical_loss": 3.7534638853072453,
      "tokens_seen": 749535232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039035105315947844,
      "loss": 2.7766,
      "theoretical_loss": 3.7534309694979884,
      "tokens_seen": 749600768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039034102306920767,
      "loss": 2.8065,
      "theoretical_loss": 3.753398057372058,
      "tokens_seen": 749666304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903309929789368,
      "loss": 2.6713,
      "theoretical_loss": 3.7533651489287188,
      "tokens_seen": 749731840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039032096288866603,
      "loss": 2.8039,
      "theoretical_loss": 3.753332244167237,
      "tokens_seen": 749797376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903109327983952,
      "loss": 2.7021,
      "theoretical_loss": 3.753299343086879,
      "tokens_seen": 749862912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903009027081244,
      "loss": 2.6358,
      "theoretical_loss": 3.753266445686912,
      "tokens_seen": 749928448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902908726178536,
      "loss": 2.5634,
      "theoretical_loss": 3.753233551966602,
      "tokens_seen": 749993984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039028084252758276,
      "loss": 2.717,
      "theoretical_loss": 3.7532006619252156,
      "tokens_seen": 750059520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039027081243731194,
      "loss": 2.8146,
      "theoretical_loss": 3.753167775562021,
      "tokens_seen": 750125056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902607823470412,
      "loss": 2.799,
      "theoretical_loss": 3.753134892876285,
      "tokens_seen": 750190592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902507522567703,
      "loss": 2.648,
      "theoretical_loss": 3.753102013867275,
      "tokens_seen": 750256128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039024072216649954,
      "loss": 2.6237,
      "theoretical_loss": 3.753069138534259,
      "tokens_seen": 750321664
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 421843,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.485619068145752,
      "objective/train/theoretical_loss": 3.753036266876505,
      "objective/train/tokens_used": 770847200,
      "theoretical_loss": 3.753036266876505,
      "tokens_seen": 750387200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039023069207622866,
      "loss": 2.5361,
      "theoretical_loss": 3.753036266876505,
      "tokens_seen": 750387200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902206619859579,
      "loss": 2.828,
      "theoretical_loss": 3.753003398893281,
      "tokens_seen": 750452736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902106318956871,
      "loss": 2.6835,
      "theoretical_loss": 3.7529705345838558,
      "tokens_seen": 750518272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039020060180541626,
      "loss": 2.8157,
      "theoretical_loss": 3.752937673947498,
      "tokens_seen": 750583808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039019057171514544,
      "loss": 2.6794,
      "theoretical_loss": 3.7529048169834756,
      "tokens_seen": 750649344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901805416248747,
      "loss": 2.8133,
      "theoretical_loss": 3.7528719636910584,
      "tokens_seen": 750714880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901705115346038,
      "loss": 2.6176,
      "theoretical_loss": 3.7528391140695154,
      "tokens_seen": 750780416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039016048144433304,
      "loss": 2.9973,
      "theoretical_loss": 3.7528062681181167,
      "tokens_seen": 750845952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039015045135406217,
      "loss": 2.8111,
      "theoretical_loss": 3.752773425836131,
      "tokens_seen": 750911488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901404212637914,
      "loss": 2.8781,
      "theoretical_loss": 3.7527405872228288,
      "tokens_seen": 750977024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901303911735206,
      "loss": 2.8973,
      "theoretical_loss": 3.75270775227748,
      "tokens_seen": 751042560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039012036108324977,
      "loss": 2.7338,
      "theoretical_loss": 3.7526749209993544,
      "tokens_seen": 751108096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039011033099297895,
      "loss": 2.7779,
      "theoretical_loss": 3.752642093387723,
      "tokens_seen": 751173632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039010030090270813,
      "loss": 2.7332,
      "theoretical_loss": 3.752609269441857,
      "tokens_seen": 751239168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900902708124373,
      "loss": 2.6754,
      "theoretical_loss": 3.7525764491610265,
      "tokens_seen": 751304704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039008024072216654,
      "loss": 2.7638,
      "theoretical_loss": 3.7525436325445027,
      "tokens_seen": 751370240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039007021063189567,
      "loss": 2.9505,
      "theoretical_loss": 3.7525108195915573,
      "tokens_seen": 751435776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900601805416249,
      "loss": 2.5367,
      "theoretical_loss": 3.752478010301462,
      "tokens_seen": 751501312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039005015045135403,
      "loss": 2.679,
      "theoretical_loss": 3.7524452046734877,
      "tokens_seen": 751566848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039004012036108327,
      "loss": 2.8081,
      "theoretical_loss": 3.752412402706907,
      "tokens_seen": 751632384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039003009027081245,
      "loss": 2.8126,
      "theoretical_loss": 3.7523796044009923,
      "tokens_seen": 751697920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039002006018054163,
      "loss": 2.6605,
      "theoretical_loss": 3.7523468097550152,
      "tokens_seen": 751763456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900100300902708,
      "loss": 2.6298,
      "theoretical_loss": 3.7523140187682493,
      "tokens_seen": 751828992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039000000000000005,
      "loss": 2.7272,
      "theoretical_loss": 3.7522812314399663,
      "tokens_seen": 751894528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899899699097292,
      "loss": 2.9884,
      "theoretical_loss": 3.7522484477694396,
      "tokens_seen": 751960064
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 423030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.662358283996582,
      "objective/train/theoretical_loss": 3.7522156677559426,
      "objective/train/tokens_used": 772485600,
      "theoretical_loss": 3.7522156677559426,
      "tokens_seen": 752025600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899799398194584,
      "loss": 2.7389,
      "theoretical_loss": 3.7522156677559426,
      "tokens_seen": 752025600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038996990972918754,
      "loss": 2.9452,
      "theoretical_loss": 3.752182891398749,
      "tokens_seen": 752091136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038995987963891677,
      "loss": 2.9664,
      "theoretical_loss": 3.752150118697132,
      "tokens_seen": 752156672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038994984954864595,
      "loss": 2.7687,
      "theoretical_loss": 3.752117349650365,
      "tokens_seen": 752222208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038993981945837513,
      "loss": 2.9713,
      "theoretical_loss": 3.752084584257723,
      "tokens_seen": 752287744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899297893681043,
      "loss": 2.8341,
      "theoretical_loss": 3.75205182251848,
      "tokens_seen": 752353280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899197592778335,
      "loss": 2.7872,
      "theoretical_loss": 3.7520190644319094,
      "tokens_seen": 752418816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899097291875627,
      "loss": 2.5177,
      "theoretical_loss": 3.751986309997287,
      "tokens_seen": 752484352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898996990972919,
      "loss": 3.0072,
      "theoretical_loss": 3.751953559213888,
      "tokens_seen": 752549888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038988966900702104,
      "loss": 2.7622,
      "theoretical_loss": 3.7519208120809857,
      "tokens_seen": 752615424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898796389167503,
      "loss": 2.8546,
      "theoretical_loss": 3.7518880685978573,
      "tokens_seen": 752680960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898696088264794,
      "loss": 2.6212,
      "theoretical_loss": 3.751855328763777,
      "tokens_seen": 752746496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038985957873620864,
      "loss": 2.4672,
      "theoretical_loss": 3.751822592578022,
      "tokens_seen": 752812032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898495486459378,
      "loss": 2.624,
      "theoretical_loss": 3.7517898600398665,
      "tokens_seen": 752877568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389839518555667,
      "loss": 2.7594,
      "theoretical_loss": 3.7517571311485867,
      "tokens_seen": 752943104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898294884653962,
      "loss": 2.8784,
      "theoretical_loss": 3.7517244059034605,
      "tokens_seen": 753008640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898194583751254,
      "loss": 2.57,
      "theoretical_loss": 3.7516916843037627,
      "tokens_seen": 753074176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038980942828485454,
      "loss": 2.69,
      "theoretical_loss": 3.751658966348771,
      "tokens_seen": 753139712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897993981945838,
      "loss": 2.7796,
      "theoretical_loss": 3.751626252037762,
      "tokens_seen": 753205248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897893681043129,
      "loss": 2.8479,
      "theoretical_loss": 3.7515935413700134,
      "tokens_seen": 753270784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038977933801404214,
      "loss": 2.803,
      "theoretical_loss": 3.7515608343448017,
      "tokens_seen": 753336320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897693079237713,
      "loss": 2.6135,
      "theoretical_loss": 3.751528130961405,
      "tokens_seen": 753401856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897592778335005,
      "loss": 2.8919,
      "theoretical_loss": 3.7514954312191007,
      "tokens_seen": 753467392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897492477432297,
      "loss": 2.759,
      "theoretical_loss": 3.751462735117167,
      "tokens_seen": 753532928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038973921765295886,
      "loss": 2.639,
      "theoretical_loss": 3.7514300426548823,
      "tokens_seen": 753598464
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 424346,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4410223960876465,
      "objective/train/theoretical_loss": 3.751397353831524,
      "objective/train/tokens_used": 774124000,
      "theoretical_loss": 3.751397353831524,
      "tokens_seen": 753664000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038972918756268805,
      "loss": 2.5345,
      "theoretical_loss": 3.751397353831524,
      "tokens_seen": 753664000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897191574724173,
      "loss": 2.8886,
      "theoretical_loss": 3.751364668646372,
      "tokens_seen": 753729536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897091273821464,
      "loss": 2.5767,
      "theoretical_loss": 3.7513319870987045,
      "tokens_seen": 753795072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038969909729187564,
      "loss": 2.6723,
      "theoretical_loss": 3.7512993091878,
      "tokens_seen": 753860608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038968906720160477,
      "loss": 2.8139,
      "theoretical_loss": 3.751266634912939,
      "tokens_seen": 753926144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389679037111334,
      "loss": 2.673,
      "theoretical_loss": 3.7512339642734,
      "tokens_seen": 753991680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896690070210632,
      "loss": 2.4749,
      "theoretical_loss": 3.7512012972684623,
      "tokens_seen": 754057216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038965897693079237,
      "loss": 2.8392,
      "theoretical_loss": 3.751168633897406,
      "tokens_seen": 754122752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038964894684052155,
      "loss": 2.6861,
      "theoretical_loss": 3.751135974159511,
      "tokens_seen": 754188288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896389167502508,
      "loss": 2.6624,
      "theoretical_loss": 3.7511033180540583,
      "tokens_seen": 754253824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038962888665997997,
      "loss": 2.7724,
      "theoretical_loss": 3.7510706655803276,
      "tokens_seen": 754319360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038961885656970915,
      "loss": 2.9919,
      "theoretical_loss": 3.7510380167375996,
      "tokens_seen": 754384896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038960882647943833,
      "loss": 2.6663,
      "theoretical_loss": 3.7510053715251557,
      "tokens_seen": 754450432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895987963891675,
      "loss": 2.778,
      "theoretical_loss": 3.750972729942276,
      "tokens_seen": 754515968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038958876629889674,
      "loss": 2.795,
      "theoretical_loss": 3.7509400919882427,
      "tokens_seen": 754581504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038957873620862587,
      "loss": 2.6879,
      "theoretical_loss": 3.7509074576623362,
      "tokens_seen": 754647040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895687061183551,
      "loss": 2.8031,
      "theoretical_loss": 3.7508748269638397,
      "tokens_seen": 754712576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038955867602808423,
      "loss": 2.7347,
      "theoretical_loss": 3.7508421998920336,
      "tokens_seen": 754778112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038954864593781347,
      "loss": 2.8483,
      "theoretical_loss": 3.7508095764462004,
      "tokens_seen": 754843648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038953861584754265,
      "loss": 2.8203,
      "theoretical_loss": 3.750776956625623,
      "tokens_seen": 754909184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038952858575727183,
      "loss": 2.6643,
      "theoretical_loss": 3.7507443404295833,
      "tokens_seen": 754974720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389518555667001,
      "loss": 2.8899,
      "theoretical_loss": 3.7507117278573636,
      "tokens_seen": 755040256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038950852557673025,
      "loss": 2.8196,
      "theoretical_loss": 3.7506791189082476,
      "tokens_seen": 755105792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894984954864594,
      "loss": 2.7476,
      "theoretical_loss": 3.750646513581518,
      "tokens_seen": 755171328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894884653961886,
      "loss": 2.5687,
      "theoretical_loss": 3.750613911876458,
      "tokens_seen": 755236864
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 425020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5594170093536377,
      "objective/train/theoretical_loss": 3.750581313792351,
      "objective/train/tokens_used": 775762400,
      "theoretical_loss": 3.750581313792351,
      "tokens_seen": 755302400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038947843530591774,
      "loss": 2.419,
      "theoretical_loss": 3.750581313792351,
      "tokens_seen": 755302400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038946840521564697,
      "loss": 2.659,
      "theoretical_loss": 3.7505487193284814,
      "tokens_seen": 755367936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038945837512537615,
      "loss": 2.7292,
      "theoretical_loss": 3.750516128484133,
      "tokens_seen": 755433472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038944834503510533,
      "loss": 2.8103,
      "theoretical_loss": 3.7504835412585886,
      "tokens_seen": 755499008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894383149448345,
      "loss": 2.7047,
      "theoretical_loss": 3.7504509576511342,
      "tokens_seen": 755564544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894282848545637,
      "loss": 2.6846,
      "theoretical_loss": 3.7504183776610533,
      "tokens_seen": 755630080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894182547642929,
      "loss": 2.7696,
      "theoretical_loss": 3.7503858012876305,
      "tokens_seen": 755695616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894082246740221,
      "loss": 2.746,
      "theoretical_loss": 3.7503532285301517,
      "tokens_seen": 755761152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038939819458375124,
      "loss": 2.7698,
      "theoretical_loss": 3.7503206593879015,
      "tokens_seen": 755826688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893881644934805,
      "loss": 2.661,
      "theoretical_loss": 3.7502880938601653,
      "tokens_seen": 755892224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893781344032096,
      "loss": 2.5849,
      "theoretical_loss": 3.750255531946228,
      "tokens_seen": 755957760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038936810431293884,
      "loss": 2.7654,
      "theoretical_loss": 3.7502229736453763,
      "tokens_seen": 756023296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389358074222668,
      "loss": 2.8642,
      "theoretical_loss": 3.7501904189568958,
      "tokens_seen": 756088832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893480441323972,
      "loss": 2.7808,
      "theoretical_loss": 3.7501578678800724,
      "tokens_seen": 756154368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893380140421264,
      "loss": 2.4493,
      "theoretical_loss": 3.750125320414192,
      "tokens_seen": 756219904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893279839518556,
      "loss": 2.7259,
      "theoretical_loss": 3.7500927765585423,
      "tokens_seen": 756285440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038931795386158474,
      "loss": 2.7156,
      "theoretical_loss": 3.75006023631241,
      "tokens_seen": 756350976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389307923771314,
      "loss": 2.7651,
      "theoretical_loss": 3.7500276996750808,
      "tokens_seen": 756416512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892978936810431,
      "loss": 2.979,
      "theoretical_loss": 3.749995166645843,
      "tokens_seen": 756482048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038928786359077234,
      "loss": 3.0294,
      "theoretical_loss": 3.749962637223984,
      "tokens_seen": 756547584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892778335005015,
      "loss": 2.7881,
      "theoretical_loss": 3.74993011140879,
      "tokens_seen": 756613120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892678034102307,
      "loss": 2.6948,
      "theoretical_loss": 3.7498975891995503,
      "tokens_seen": 756678656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892577733199599,
      "loss": 2.7056,
      "theoretical_loss": 3.7498650705955523,
      "tokens_seen": 756744192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038924774322968906,
      "loss": 2.6511,
      "theoretical_loss": 3.7498325555960843,
      "tokens_seen": 756809728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038923771313941825,
      "loss": 2.7346,
      "theoretical_loss": 3.749800044200434,
      "tokens_seen": 756875264
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 426245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.891552448272705,
      "objective/train/theoretical_loss": 3.749767536407891,
      "objective/train/tokens_used": 777400800,
      "theoretical_loss": 3.749767536407891,
      "tokens_seen": 756940800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892276830491475,
      "loss": 2.7255,
      "theoretical_loss": 3.749767536407891,
      "tokens_seen": 756940800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892176529588766,
      "loss": 2.753,
      "theoretical_loss": 3.7497350322177434,
      "tokens_seen": 757006336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038920762286860584,
      "loss": 2.7117,
      "theoretical_loss": 3.7497025316292802,
      "tokens_seen": 757071872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038919759277833497,
      "loss": 2.6464,
      "theoretical_loss": 3.749670034641791,
      "tokens_seen": 757137408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891875626880642,
      "loss": 2.8426,
      "theoretical_loss": 3.7496375412545646,
      "tokens_seen": 757202944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891775325977934,
      "loss": 2.6978,
      "theoretical_loss": 3.749605051466891,
      "tokens_seen": 757268480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038916750250752257,
      "loss": 2.936,
      "theoretical_loss": 3.74957256527806,
      "tokens_seen": 757334016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038915747241725175,
      "loss": 2.8052,
      "theoretical_loss": 3.7495400826873615,
      "tokens_seen": 757399552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389147442326981,
      "loss": 2.7019,
      "theoretical_loss": 3.7495076036940853,
      "tokens_seen": 757465088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891374122367101,
      "loss": 2.8646,
      "theoretical_loss": 3.749475128297523,
      "tokens_seen": 757530624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038912738214643935,
      "loss": 2.6208,
      "theoretical_loss": 3.7494426564969636,
      "tokens_seen": 757596160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891173520561685,
      "loss": 2.8859,
      "theoretical_loss": 3.749410188291699,
      "tokens_seen": 757661696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891073219658977,
      "loss": 2.7943,
      "theoretical_loss": 3.74937772368102,
      "tokens_seen": 757727232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890972918756269,
      "loss": 2.7624,
      "theoretical_loss": 3.7493452626642174,
      "tokens_seen": 757792768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038908726178535607,
      "loss": 2.7132,
      "theoretical_loss": 3.7493128052405824,
      "tokens_seen": 757858304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038907723169508525,
      "loss": 2.7807,
      "theoretical_loss": 3.7492803514094075,
      "tokens_seen": 757923840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038906720160481443,
      "loss": 2.6621,
      "theoretical_loss": 3.7492479011699835,
      "tokens_seen": 757989376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890571715145436,
      "loss": 2.8097,
      "theoretical_loss": 3.749215454521604,
      "tokens_seen": 758054912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038904714142427285,
      "loss": 2.7598,
      "theoretical_loss": 3.7491830114635594,
      "tokens_seen": 758120448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389037111334002,
      "loss": 2.876,
      "theoretical_loss": 3.749150571995143,
      "tokens_seen": 758185984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890270812437312,
      "loss": 2.8083,
      "theoretical_loss": 3.749118136115647,
      "tokens_seen": 758251520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890170511534604,
      "loss": 2.6312,
      "theoretical_loss": 3.749085703824364,
      "tokens_seen": 758317056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890070210631896,
      "loss": 2.7658,
      "theoretical_loss": 3.7490532751205885,
      "tokens_seen": 758382592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038899699097291876,
      "loss": 2.7504,
      "theoretical_loss": 3.749020850003612,
      "tokens_seen": 758448128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038898696088264794,
      "loss": 2.6115,
      "theoretical_loss": 3.7489884284727286,
      "tokens_seen": 758513664
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 426718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.676023244857788,
      "objective/train/theoretical_loss": 3.748956010527232,
      "objective/train/tokens_used": 779039200,
      "theoretical_loss": 3.748956010527232,
      "tokens_seen": 758579200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889769307923771,
      "loss": 2.7646,
      "theoretical_loss": 3.748956010527232,
      "tokens_seen": 758579200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038896690070210635,
      "loss": 2.8833,
      "theoretical_loss": 3.7489235961664154,
      "tokens_seen": 758644736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889568706118355,
      "loss": 2.7432,
      "theoretical_loss": 3.748891185389574,
      "tokens_seen": 758710272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889468405215647,
      "loss": 2.5897,
      "theoretical_loss": 3.7488587781960008,
      "tokens_seen": 758775808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038893681043129384,
      "loss": 2.9097,
      "theoretical_loss": 3.748826374584991,
      "tokens_seen": 758841344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889267803410231,
      "loss": 2.9009,
      "theoretical_loss": 3.748793974555838,
      "tokens_seen": 758906880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038891675025075226,
      "loss": 2.6654,
      "theoretical_loss": 3.748761578107838,
      "tokens_seen": 758972416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038890672016048144,
      "loss": 2.9037,
      "theoretical_loss": 3.7487291852402853,
      "tokens_seen": 759037952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888966900702106,
      "loss": 2.5612,
      "theoretical_loss": 3.7486967959524753,
      "tokens_seen": 759103488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888866599799398,
      "loss": 2.5881,
      "theoretical_loss": 3.7486644102437037,
      "tokens_seen": 759169024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038887662988966904,
      "loss": 2.9656,
      "theoretical_loss": 3.7486320281132657,
      "tokens_seen": 759234560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888665997993982,
      "loss": 2.801,
      "theoretical_loss": 3.748599649560457,
      "tokens_seen": 759300096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888565697091274,
      "loss": 2.7532,
      "theoretical_loss": 3.7485672745845733,
      "tokens_seen": 759365632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888465396188566,
      "loss": 2.4411,
      "theoretical_loss": 3.748534903184912,
      "tokens_seen": 759431168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888365095285858,
      "loss": 2.7694,
      "theoretical_loss": 3.7485025353607684,
      "tokens_seen": 759496704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038882647943831494,
      "loss": 2.6058,
      "theoretical_loss": 3.7484701711114394,
      "tokens_seen": 759562240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888164493480442,
      "loss": 2.8315,
      "theoretical_loss": 3.748437810436222,
      "tokens_seen": 759627776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888064192577733,
      "loss": 2.9115,
      "theoretical_loss": 3.748405453334413,
      "tokens_seen": 759693312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038879638916750254,
      "loss": 2.6568,
      "theoretical_loss": 3.7483730998053097,
      "tokens_seen": 759758848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887863590772317,
      "loss": 2.7921,
      "theoretical_loss": 3.748340749848209,
      "tokens_seen": 759824384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887763289869609,
      "loss": 2.8153,
      "theoretical_loss": 3.74830840346241,
      "tokens_seen": 759889920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887662988966901,
      "loss": 2.6874,
      "theoretical_loss": 3.7482760606472088,
      "tokens_seen": 759955456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038875626880641927,
      "loss": 2.949,
      "theoretical_loss": 3.748243721401904,
      "tokens_seen": 760020992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038874623871614845,
      "loss": 2.771,
      "theoretical_loss": 3.748211385725793,
      "tokens_seen": 760086528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887362086258777,
      "loss": 2.7016,
      "theoretical_loss": 3.748179053618176,
      "tokens_seen": 760152064
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 427515,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.012765645980835,
      "objective/train/theoretical_loss": 3.7481467250783504,
      "objective/train/tokens_used": 780677600,
      "theoretical_loss": 3.7481467250783504,
      "tokens_seen": 760217600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887261785356068,
      "loss": 3.079,
      "theoretical_loss": 3.7481467250783504,
      "tokens_seen": 760217600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038871614844533604,
      "loss": 2.777,
      "theoretical_loss": 3.748114400105615,
      "tokens_seen": 760283136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038870611835506517,
      "loss": 2.8811,
      "theoretical_loss": 3.748082078699269,
      "tokens_seen": 760348672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886960882647944,
      "loss": 2.8731,
      "theoretical_loss": 3.7480497608586116,
      "tokens_seen": 760414208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886860581745236,
      "loss": 2.8526,
      "theoretical_loss": 3.748017446582942,
      "tokens_seen": 760479744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038867602808425277,
      "loss": 2.9608,
      "theoretical_loss": 3.74798513587156,
      "tokens_seen": 760545280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038866599799398195,
      "loss": 2.7168,
      "theoretical_loss": 3.747952828723765,
      "tokens_seen": 760610816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886559679037112,
      "loss": 2.7147,
      "theoretical_loss": 3.7479205251388574,
      "tokens_seen": 760676352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886459378134403,
      "loss": 2.8606,
      "theoretical_loss": 3.747888225116137,
      "tokens_seen": 760741888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038863590772316955,
      "loss": 2.9467,
      "theoretical_loss": 3.7478559286549045,
      "tokens_seen": 760807424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886258776328987,
      "loss": 2.7865,
      "theoretical_loss": 3.74782363575446,
      "tokens_seen": 760872960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886158475426279,
      "loss": 2.7565,
      "theoretical_loss": 3.747791346414105,
      "tokens_seen": 760938496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886058174523571,
      "loss": 2.7405,
      "theoretical_loss": 3.74775906063314,
      "tokens_seen": 761004032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038859578736208627,
      "loss": 2.7645,
      "theoretical_loss": 3.747726778410866,
      "tokens_seen": 761069568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038858575727181545,
      "loss": 2.9492,
      "theoretical_loss": 3.7476944997465846,
      "tokens_seen": 761135104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038857572718154463,
      "loss": 2.8445,
      "theoretical_loss": 3.747662224639597,
      "tokens_seen": 761200640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885656970912738,
      "loss": 2.7768,
      "theoretical_loss": 3.7476299530892057,
      "tokens_seen": 761266176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038855566700100305,
      "loss": 3.0213,
      "theoretical_loss": 3.7475976850947124,
      "tokens_seen": 761331712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885456369107322,
      "loss": 2.9531,
      "theoretical_loss": 3.7475654206554188,
      "tokens_seen": 761397248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885356068204614,
      "loss": 2.9068,
      "theoretical_loss": 3.7475331597706267,
      "tokens_seen": 761462784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885255767301906,
      "loss": 2.7472,
      "theoretical_loss": 3.74750090243964,
      "tokens_seen": 761528320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885155466399198,
      "loss": 2.8358,
      "theoretical_loss": 3.747468648661761,
      "tokens_seen": 761593856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038850551654964896,
      "loss": 3.0352,
      "theoretical_loss": 3.7474363984362924,
      "tokens_seen": 761659392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038849548645937814,
      "loss": 2.7928,
      "theoretical_loss": 3.7474041517625376,
      "tokens_seen": 761724928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884854563691073,
      "loss": 2.9632,
      "theoretical_loss": 3.7473719086397996,
      "tokens_seen": 761790464
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 428729,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0259807109832764,
      "objective/train/theoretical_loss": 3.747339669067382,
      "objective/train/tokens_used": 782316000,
      "theoretical_loss": 3.747339669067382,
      "tokens_seen": 761856000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038847542627883655,
      "loss": 3.03,
      "theoretical_loss": 3.747339669067382,
      "tokens_seen": 761856000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884653961885657,
      "loss": 2.7308,
      "theoretical_loss": 3.747307433044589,
      "tokens_seen": 761921536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884553660982949,
      "loss": 2.7257,
      "theoretical_loss": 3.7472752005707233,
      "tokens_seen": 761987072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038844533600802404,
      "loss": 2.8627,
      "theoretical_loss": 3.7472429716450897,
      "tokens_seen": 762052608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884353059177533,
      "loss": 2.8747,
      "theoretical_loss": 3.747210746266993,
      "tokens_seen": 762118144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038842527582748246,
      "loss": 2.8288,
      "theoretical_loss": 3.747178524435738,
      "tokens_seen": 762183680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038841524573721164,
      "loss": 2.7417,
      "theoretical_loss": 3.747146306150628,
      "tokens_seen": 762249216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884052156469408,
      "loss": 2.8478,
      "theoretical_loss": 3.7471140914109684,
      "tokens_seen": 762314752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038839518555667,
      "loss": 2.6023,
      "theoretical_loss": 3.7470818802160646,
      "tokens_seen": 762380288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883851554663992,
      "loss": 2.88,
      "theoretical_loss": 3.747049672565222,
      "tokens_seen": 762445824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883751253761284,
      "loss": 2.9135,
      "theoretical_loss": 3.747017468457746,
      "tokens_seen": 762511360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038836509528585755,
      "loss": 3.1032,
      "theoretical_loss": 3.7469852678929416,
      "tokens_seen": 762576896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883550651955868,
      "loss": 2.8413,
      "theoretical_loss": 3.7469530708701155,
      "tokens_seen": 762642432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038834503510531596,
      "loss": 2.6678,
      "theoretical_loss": 3.7469208773885736,
      "tokens_seen": 762707968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038833500501504514,
      "loss": 2.764,
      "theoretical_loss": 3.7468886874476217,
      "tokens_seen": 762773504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883249749247743,
      "loss": 2.7965,
      "theoretical_loss": 3.746856501046567,
      "tokens_seen": 762839040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883149448345035,
      "loss": 2.6913,
      "theoretical_loss": 3.746824318184715,
      "tokens_seen": 762904576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883049147442327,
      "loss": 2.9067,
      "theoretical_loss": 3.7467921388613745,
      "tokens_seen": 762970112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882948846539619,
      "loss": 2.7015,
      "theoretical_loss": 3.746759963075851,
      "tokens_seen": 763035648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038828485456369105,
      "loss": 2.962,
      "theoretical_loss": 3.7467277908274514,
      "tokens_seen": 763101184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882748244734203,
      "loss": 3.0969,
      "theoretical_loss": 3.7466956221154843,
      "tokens_seen": 763166720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882647943831494,
      "loss": 2.8899,
      "theoretical_loss": 3.746663456939257,
      "tokens_seen": 763232256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038825476429287865,
      "loss": 3.1046,
      "theoretical_loss": 3.7466312952980774,
      "tokens_seen": 763297792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038824473420260783,
      "loss": 2.7427,
      "theoretical_loss": 3.746599137191253,
      "tokens_seen": 763363328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388234704112337,
      "loss": 2.9909,
      "theoretical_loss": 3.7465669826180923,
      "tokens_seen": 763428864
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 429385,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7105305194854736,
      "objective/train/theoretical_loss": 3.746534831577904,
      "objective/train/tokens_used": 783954400,
      "theoretical_loss": 3.746534831577904,
      "tokens_seen": 763494400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882246740220662,
      "loss": 2.7427,
      "theoretical_loss": 3.746534831577904,
      "tokens_seen": 763494400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038821464393179537,
      "loss": 2.726,
      "theoretical_loss": 3.7465026840699958,
      "tokens_seen": 763559936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038820461384152455,
      "loss": 2.7427,
      "theoretical_loss": 3.7464705400936777,
      "tokens_seen": 763625472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881945837512538,
      "loss": 2.7322,
      "theoretical_loss": 3.746438399648258,
      "tokens_seen": 763691008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881845536609829,
      "loss": 2.996,
      "theoretical_loss": 3.7464062627330463,
      "tokens_seen": 763756544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038817452357071215,
      "loss": 2.5489,
      "theoretical_loss": 3.7463741293473514,
      "tokens_seen": 763822080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038816449348044133,
      "loss": 2.7436,
      "theoretical_loss": 3.7463419994904834,
      "tokens_seen": 763887616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881544633901705,
      "loss": 2.758,
      "theoretical_loss": 3.746309873161752,
      "tokens_seen": 763953152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881444332998997,
      "loss": 2.8826,
      "theoretical_loss": 3.7462777503604667,
      "tokens_seen": 764018688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881344032096289,
      "loss": 2.7736,
      "theoretical_loss": 3.746245631085938,
      "tokens_seen": 764084224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881243731193581,
      "loss": 2.6549,
      "theoretical_loss": 3.7462135153374767,
      "tokens_seen": 764149760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881143430290873,
      "loss": 2.7209,
      "theoretical_loss": 3.7461814031143925,
      "tokens_seen": 764215296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038810431293881647,
      "loss": 2.7706,
      "theoretical_loss": 3.746149294415997,
      "tokens_seen": 764280832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038809428284854565,
      "loss": 2.7119,
      "theoretical_loss": 3.7461171892416,
      "tokens_seen": 764346368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038808425275827483,
      "loss": 2.4525,
      "theoretical_loss": 3.746085087590514,
      "tokens_seen": 764411904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388074222668004,
      "loss": 2.7598,
      "theoretical_loss": 3.7460529894620493,
      "tokens_seen": 764477440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038806419257773325,
      "loss": 2.6697,
      "theoretical_loss": 3.7460208948555183,
      "tokens_seen": 764542976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880541624874624,
      "loss": 2.7454,
      "theoretical_loss": 3.7459888037702314,
      "tokens_seen": 764608512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880441323971916,
      "loss": 2.8053,
      "theoretical_loss": 3.7459567162055016,
      "tokens_seen": 764674048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880341023069208,
      "loss": 2.5966,
      "theoretical_loss": 3.7459246321606408,
      "tokens_seen": 764739584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038802407221665,
      "loss": 2.8628,
      "theoretical_loss": 3.7458925516349613,
      "tokens_seen": 764805120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038801404212637916,
      "loss": 2.7437,
      "theoretical_loss": 3.745860474627775,
      "tokens_seen": 764870656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038800401203610834,
      "loss": 2.7495,
      "theoretical_loss": 3.7458284011383953,
      "tokens_seen": 764936192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879939819458375,
      "loss": 2.5588,
      "theoretical_loss": 3.745796331166135,
      "tokens_seen": 765001728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038798395185556675,
      "loss": 2.6211,
      "theoretical_loss": 3.745764264710307,
      "tokens_seen": 765067264
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 430652,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8052632808685303,
      "objective/train/theoretical_loss": 3.7457322017702244,
      "objective/train/tokens_used": 785592800,
      "theoretical_loss": 3.7457322017702244,
      "tokens_seen": 765132800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879739217652959,
      "loss": 2.8717,
      "theoretical_loss": 3.7457322017702244,
      "tokens_seen": 765132800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879638916750251,
      "loss": 2.8154,
      "theoretical_loss": 3.7457001423452008,
      "tokens_seen": 765198336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038795386158475424,
      "loss": 2.8259,
      "theoretical_loss": 3.7456680864345495,
      "tokens_seen": 765263872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879438314944835,
      "loss": 2.955,
      "theoretical_loss": 3.745636034037585,
      "tokens_seen": 765329408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038793380140421266,
      "loss": 2.7045,
      "theoretical_loss": 3.745603985153621,
      "tokens_seen": 765394944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038792377131394184,
      "loss": 2.7434,
      "theoretical_loss": 3.745571939781972,
      "tokens_seen": 765460480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387913741223671,
      "loss": 2.6924,
      "theoretical_loss": 3.7455398979219514,
      "tokens_seen": 765526016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879037111334002,
      "loss": 2.7036,
      "theoretical_loss": 3.745507859572875,
      "tokens_seen": 765591552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878936810431294,
      "loss": 2.8034,
      "theoretical_loss": 3.745475824734057,
      "tokens_seen": 765657088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878836509528586,
      "loss": 2.9501,
      "theoretical_loss": 3.745443793404813,
      "tokens_seen": 765722624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038787362086258775,
      "loss": 2.675,
      "theoretical_loss": 3.745411765584457,
      "tokens_seen": 765788160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387863590772317,
      "loss": 2.8542,
      "theoretical_loss": 3.7453797412723056,
      "tokens_seen": 765853696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038785356068204616,
      "loss": 2.7184,
      "theoretical_loss": 3.7453477204676737,
      "tokens_seen": 765919232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038784353059177534,
      "loss": 2.8863,
      "theoretical_loss": 3.7453157031698776,
      "tokens_seen": 765984768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878335005015045,
      "loss": 2.6537,
      "theoretical_loss": 3.745283689378232,
      "tokens_seen": 766050304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878234704112337,
      "loss": 2.8882,
      "theoretical_loss": 3.7452516790920543,
      "tokens_seen": 766115840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878134403209629,
      "loss": 2.7967,
      "theoretical_loss": 3.745219672310661,
      "tokens_seen": 766181376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878034102306921,
      "loss": 2.8745,
      "theoretical_loss": 3.745187669033368,
      "tokens_seen": 766246912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038779338014042125,
      "loss": 2.7758,
      "theoretical_loss": 3.745155669259491,
      "tokens_seen": 766312448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877833500501505,
      "loss": 2.8668,
      "theoretical_loss": 3.745123672988349,
      "tokens_seen": 766377984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877733199598796,
      "loss": 2.8476,
      "theoretical_loss": 3.745091680219258,
      "tokens_seen": 766443520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038776328986960885,
      "loss": 2.8759,
      "theoretical_loss": 3.7450596909515355,
      "tokens_seen": 766509056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038775325977933803,
      "loss": 3.0193,
      "theoretical_loss": 3.7450277051844987,
      "tokens_seen": 766574592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877432296890672,
      "loss": 2.7882,
      "theoretical_loss": 3.7449957229174653,
      "tokens_seen": 766640128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877331995987964,
      "loss": 2.9495,
      "theoretical_loss": 3.7449637441497536,
      "tokens_seen": 766705664
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 431055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6919522285461426,
      "objective/train/theoretical_loss": 3.744931768880681,
      "objective/train/tokens_used": 787231200,
      "theoretical_loss": 3.744931768880681,
      "tokens_seen": 766771200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038772316950852557,
      "loss": 2.92,
      "theoretical_loss": 3.744931768880681,
      "tokens_seen": 766771200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038771313941825475,
      "loss": 2.7832,
      "theoretical_loss": 3.744899797109567,
      "tokens_seen": 766836736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387703109327984,
      "loss": 2.8811,
      "theoretical_loss": 3.7448678288357287,
      "tokens_seen": 766902272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876930792377131,
      "loss": 2.921,
      "theoretical_loss": 3.7448358640584853,
      "tokens_seen": 766967808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038768304914744235,
      "loss": 2.9364,
      "theoretical_loss": 3.7448039027771554,
      "tokens_seen": 767033344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038767301905717153,
      "loss": 2.5438,
      "theoretical_loss": 3.744771944991058,
      "tokens_seen": 767098880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876629889669007,
      "loss": 2.7965,
      "theoretical_loss": 3.7447399906995127,
      "tokens_seen": 767164416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876529588766299,
      "loss": 2.8438,
      "theoretical_loss": 3.7447080399018393,
      "tokens_seen": 767229952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876429287863591,
      "loss": 2.9344,
      "theoretical_loss": 3.7446760925973557,
      "tokens_seen": 767295488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038763289869608826,
      "loss": 2.7249,
      "theoretical_loss": 3.744644148785383,
      "tokens_seen": 767361024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876228686058175,
      "loss": 2.8412,
      "theoretical_loss": 3.7446122084652416,
      "tokens_seen": 767426560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876128385155466,
      "loss": 2.777,
      "theoretical_loss": 3.7445802716362504,
      "tokens_seen": 767492096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038760280842527585,
      "loss": 2.8745,
      "theoretical_loss": 3.7445483382977303,
      "tokens_seen": 767557632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387592778335005,
      "loss": 2.7448,
      "theoretical_loss": 3.744516408449002,
      "tokens_seen": 767623168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875827482447342,
      "loss": 2.9832,
      "theoretical_loss": 3.744484482089386,
      "tokens_seen": 767688704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875727181544634,
      "loss": 2.8086,
      "theoretical_loss": 3.744452559218203,
      "tokens_seen": 767754240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875626880641926,
      "loss": 2.6926,
      "theoretical_loss": 3.7444206398347752,
      "tokens_seen": 767819776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038755265797392176,
      "loss": 2.8013,
      "theoretical_loss": 3.744388723938423,
      "tokens_seen": 767885312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387542627883651,
      "loss": 2.776,
      "theoretical_loss": 3.7443568115284673,
      "tokens_seen": 767950848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875325977933801,
      "loss": 2.659,
      "theoretical_loss": 3.744324902604231,
      "tokens_seen": 768016384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038752256770310936,
      "loss": 2.7347,
      "theoretical_loss": 3.7442929971650356,
      "tokens_seen": 768081920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875125376128385,
      "loss": 2.7669,
      "theoretical_loss": 3.744261095210203,
      "tokens_seen": 768147456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875025075225677,
      "loss": 2.8019,
      "theoretical_loss": 3.744229196739055,
      "tokens_seen": 768212992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874924774322969,
      "loss": 2.8041,
      "theoretical_loss": 3.7441973017509147,
      "tokens_seen": 768278528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874824473420261,
      "loss": 2.8703,
      "theoretical_loss": 3.7441654102451047,
      "tokens_seen": 768344064
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 432209,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0975613594055176,
      "objective/train/theoretical_loss": 3.744133522220947,
      "objective/train/tokens_used": 788869600,
      "theoretical_loss": 3.744133522220947,
      "tokens_seen": 768409600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038747241725175526,
      "loss": 2.8819,
      "theoretical_loss": 3.744133522220947,
      "tokens_seen": 768409600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038746238716148444,
      "loss": 2.7959,
      "theoretical_loss": 3.7441016376777663,
      "tokens_seen": 768475136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874523570712136,
      "loss": 2.9047,
      "theoretical_loss": 3.744069756614884,
      "tokens_seen": 768540672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038744232698094286,
      "loss": 2.8923,
      "theoretical_loss": 3.744037879031625,
      "tokens_seen": 768606208
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387432296890672,
      "loss": 2.9207,
      "theoretical_loss": 3.7440060049273116,
      "tokens_seen": 768671744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874222668004012,
      "loss": 2.703,
      "theoretical_loss": 3.7439741343012676,
      "tokens_seen": 768737280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038741223671013035,
      "loss": 2.7158,
      "theoretical_loss": 3.7439422671528177,
      "tokens_seen": 768802816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874022066198596,
      "loss": 2.7988,
      "theoretical_loss": 3.7439104034812862,
      "tokens_seen": 768868352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038739217652958876,
      "loss": 2.9013,
      "theoretical_loss": 3.743878543285997,
      "tokens_seen": 768933888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038738214643931795,
      "loss": 2.932,
      "theoretical_loss": 3.7438466865662745,
      "tokens_seen": 768999424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873721163490472,
      "loss": 2.9536,
      "theoretical_loss": 3.7438148333214434,
      "tokens_seen": 769064960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038736208625877636,
      "loss": 2.6854,
      "theoretical_loss": 3.7437829835508287,
      "tokens_seen": 769130496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038735205616850554,
      "loss": 2.749,
      "theoretical_loss": 3.743751137253756,
      "tokens_seen": 769196032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873420260782347,
      "loss": 2.4222,
      "theoretical_loss": 3.7437192944295496,
      "tokens_seen": 769261568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873319959879639,
      "loss": 2.9377,
      "theoretical_loss": 3.743687455077536,
      "tokens_seen": 769327104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873219658976931,
      "loss": 2.5288,
      "theoretical_loss": 3.74365561919704,
      "tokens_seen": 769392640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873119358074223,
      "loss": 2.9269,
      "theoretical_loss": 3.7436237867873876,
      "tokens_seen": 769458176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038730190571715145,
      "loss": 3.2117,
      "theoretical_loss": 3.743591957847905,
      "tokens_seen": 769523712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872918756268807,
      "loss": 2.8716,
      "theoretical_loss": 3.7435601323779184,
      "tokens_seen": 769589248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872818455366098,
      "loss": 2.9267,
      "theoretical_loss": 3.743528310376754,
      "tokens_seen": 769654784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038727181544633905,
      "loss": 2.8562,
      "theoretical_loss": 3.7434964918437394,
      "tokens_seen": 769720320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038726178535606823,
      "loss": 2.617,
      "theoretical_loss": 3.7434646767782,
      "tokens_seen": 769785856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872517552657974,
      "loss": 2.8305,
      "theoretical_loss": 3.7434328651794635,
      "tokens_seen": 769851392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872417251755266,
      "loss": 2.8065,
      "theoretical_loss": 3.743401057046857,
      "tokens_seen": 769916928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038723169508525577,
      "loss": 2.7927,
      "theoretical_loss": 3.7433692523797077,
      "tokens_seen": 769982464
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 432880,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1428420543670654,
      "objective/train/theoretical_loss": 3.743337451177343,
      "objective/train/tokens_used": 790508000,
      "theoretical_loss": 3.743337451177343,
      "tokens_seen": 770048000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038722166499498495,
      "loss": 3.0053,
      "theoretical_loss": 3.743337451177343,
      "tokens_seen": 770048000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872116349047142,
      "loss": 2.8568,
      "theoretical_loss": 3.743305653439091,
      "tokens_seen": 770113536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872016048144433,
      "loss": 2.941,
      "theoretical_loss": 3.7432738591642796,
      "tokens_seen": 770179072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038719157472417255,
      "loss": 2.8203,
      "theoretical_loss": 3.7432420683522363,
      "tokens_seen": 770244608
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038718154463390173,
      "loss": 2.8672,
      "theoretical_loss": 3.7432102810022903,
      "tokens_seen": 770310144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871715145436309,
      "loss": 2.8655,
      "theoretical_loss": 3.743178497113769,
      "tokens_seen": 770375680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871614844533601,
      "loss": 2.7875,
      "theoretical_loss": 3.7431467166860015,
      "tokens_seen": 770441216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871514543630893,
      "loss": 2.7909,
      "theoretical_loss": 3.7431149397183177,
      "tokens_seen": 770506752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038714142427281846,
      "loss": 2.8959,
      "theoretical_loss": 3.743083166210045,
      "tokens_seen": 770572288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871313941825477,
      "loss": 2.7932,
      "theoretical_loss": 3.7430513961605136,
      "tokens_seen": 770637824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871213640922768,
      "loss": 2.814,
      "theoretical_loss": 3.7430196295690523,
      "tokens_seen": 770703360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038711133400200605,
      "loss": 3.0201,
      "theoretical_loss": 3.7429878664349916,
      "tokens_seen": 770768896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871013039117352,
      "loss": 2.8408,
      "theoretical_loss": 3.74295610675766,
      "tokens_seen": 770834432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870912738214644,
      "loss": 2.6841,
      "theoretical_loss": 3.7429243505363887,
      "tokens_seen": 770899968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870812437311936,
      "loss": 2.7229,
      "theoretical_loss": 3.742892597770507,
      "tokens_seen": 770965504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870712136409228,
      "loss": 2.8088,
      "theoretical_loss": 3.7428608484593457,
      "tokens_seen": 771031040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038706118355065196,
      "loss": 2.6387,
      "theoretical_loss": 3.7428291026022356,
      "tokens_seen": 771096576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870511534603812,
      "loss": 2.7468,
      "theoretical_loss": 3.742797360198507,
      "tokens_seen": 771162112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870411233701103,
      "loss": 2.7474,
      "theoretical_loss": 3.7427656212474902,
      "tokens_seen": 771227648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038703109327983956,
      "loss": 2.8616,
      "theoretical_loss": 3.7427338857485175,
      "tokens_seen": 771293184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870210631895687,
      "loss": 2.6703,
      "theoretical_loss": 3.742702153700919,
      "tokens_seen": 771358720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870110330992979,
      "loss": 2.7043,
      "theoretical_loss": 3.742670425104027,
      "tokens_seen": 771424256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870010030090271,
      "loss": 2.797,
      "theoretical_loss": 3.7426386999571735,
      "tokens_seen": 771489792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869909729187563,
      "loss": 2.8079,
      "theoretical_loss": 3.7426069782596887,
      "tokens_seen": 771555328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038698094282848546,
      "loss": 2.8836,
      "theoretical_loss": 3.7425752600109066,
      "tokens_seen": 771620864
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 434052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.293273448944092,
      "objective/train/theoretical_loss": 3.742543545210158,
      "objective/train/tokens_used": 792146400,
      "theoretical_loss": 3.742543545210158,
      "tokens_seen": 771686400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038697091273821464,
      "loss": 2.6631,
      "theoretical_loss": 3.742543545210158,
      "tokens_seen": 771686400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869608826479438,
      "loss": 2.892,
      "theoretical_loss": 3.7425118338567755,
      "tokens_seen": 771751936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038695085255767306,
      "loss": 2.8685,
      "theoretical_loss": 3.7424801259500926,
      "tokens_seen": 771817472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869408224674022,
      "loss": 2.8693,
      "theoretical_loss": 3.7424484214894407,
      "tokens_seen": 771883008
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869307923771314,
      "loss": 2.8568,
      "theoretical_loss": 3.7424167204741536,
      "tokens_seen": 771948544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038692076228686055,
      "loss": 2.937,
      "theoretical_loss": 3.7423850229035645,
      "tokens_seen": 772014080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869107321965898,
      "loss": 2.6878,
      "theoretical_loss": 3.742353328777006,
      "tokens_seen": 772079616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038690070210631896,
      "loss": 2.7514,
      "theoretical_loss": 3.742321638093813,
      "tokens_seen": 772145152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038689067201604815,
      "loss": 2.8053,
      "theoretical_loss": 3.7422899508533174,
      "tokens_seen": 772210688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038688064192577733,
      "loss": 2.6689,
      "theoretical_loss": 3.7422582670548543,
      "tokens_seen": 772276224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038687061183550656,
      "loss": 2.6903,
      "theoretical_loss": 3.7422265866977575,
      "tokens_seen": 772341760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868605817452357,
      "loss": 2.7236,
      "theoretical_loss": 3.742194909781361,
      "tokens_seen": 772407296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868505516549649,
      "loss": 2.9353,
      "theoretical_loss": 3.7421632363049993,
      "tokens_seen": 772472832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038684052156469405,
      "loss": 2.823,
      "theoretical_loss": 3.7421315662680072,
      "tokens_seen": 772538368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868304914744233,
      "loss": 2.7551,
      "theoretical_loss": 3.7420998996697197,
      "tokens_seen": 772603904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038682046138415247,
      "loss": 2.9628,
      "theoretical_loss": 3.7420682365094713,
      "tokens_seen": 772669440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038681043129388165,
      "loss": 2.7756,
      "theoretical_loss": 3.7420365767865977,
      "tokens_seen": 772734976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038680040120361083,
      "loss": 2.9,
      "theoretical_loss": 3.7420049205004338,
      "tokens_seen": 772800512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038679037111334,
      "loss": 2.8719,
      "theoretical_loss": 3.741973267650315,
      "tokens_seen": 772866048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867803410230692,
      "loss": 2.9457,
      "theoretical_loss": 3.741941618235577,
      "tokens_seen": 772931584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038677031093279843,
      "loss": 2.8025,
      "theoretical_loss": 3.7419099722555567,
      "tokens_seen": 772997120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038676028084252755,
      "loss": 2.6723,
      "theoretical_loss": 3.7418783297095892,
      "tokens_seen": 773062656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867502507522568,
      "loss": 2.8905,
      "theoretical_loss": 3.741846690597012,
      "tokens_seen": 773128192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867402206619859,
      "loss": 2.7707,
      "theoretical_loss": 3.7418150549171596,
      "tokens_seen": 773193728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038673019057171515,
      "loss": 2.873,
      "theoretical_loss": 3.74178342266937,
      "tokens_seen": 773259264
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 434666,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.016434907913208,
      "objective/train/theoretical_loss": 3.74175179385298,
      "objective/train/tokens_used": 793784800,
      "theoretical_loss": 3.74175179385298,
      "tokens_seen": 773324800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038672016048144433,
      "loss": 2.8806,
      "theoretical_loss": 3.74175179385298,
      "tokens_seen": 773324800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867101303911735,
      "loss": 2.7735,
      "theoretical_loss": 3.7417201684673262,
      "tokens_seen": 773390336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867001003009027,
      "loss": 2.8635,
      "theoretical_loss": 3.741688546511746,
      "tokens_seen": 773455872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038669007021063193,
      "loss": 2.7011,
      "theoretical_loss": 3.7416569279855767,
      "tokens_seen": 773521408
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038668004012036106,
      "loss": 2.6612,
      "theoretical_loss": 3.7416253128881563,
      "tokens_seen": 773586944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866700100300903,
      "loss": 3.141,
      "theoretical_loss": 3.7415937012188216,
      "tokens_seen": 773652480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866599799398194,
      "loss": 2.8049,
      "theoretical_loss": 3.7415620929769116,
      "tokens_seen": 773718016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038664994984954866,
      "loss": 2.7794,
      "theoretical_loss": 3.7415304881617644,
      "tokens_seen": 773783552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038663991975927784,
      "loss": 2.9484,
      "theoretical_loss": 3.7414988867727166,
      "tokens_seen": 773849088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386629889669007,
      "loss": 2.8271,
      "theoretical_loss": 3.741467288809109,
      "tokens_seen": 773914624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038661985957873625,
      "loss": 2.9336,
      "theoretical_loss": 3.7414356942702787,
      "tokens_seen": 773980160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866098294884654,
      "loss": 2.8063,
      "theoretical_loss": 3.7414041031555656,
      "tokens_seen": 774045696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865997993981946,
      "loss": 2.7585,
      "theoretical_loss": 3.7413725154643083,
      "tokens_seen": 774111232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865897693079238,
      "loss": 2.7239,
      "theoretical_loss": 3.741340931195846,
      "tokens_seen": 774176768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386579739217653,
      "loss": 2.6134,
      "theoretical_loss": 3.7413093503495176,
      "tokens_seen": 774242304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038656970912738216,
      "loss": 2.9152,
      "theoretical_loss": 3.741277772924663,
      "tokens_seen": 774307840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865596790371114,
      "loss": 2.7841,
      "theoretical_loss": 3.7412461989206225,
      "tokens_seen": 774373376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865496489468405,
      "loss": 2.7092,
      "theoretical_loss": 3.741214628336736,
      "tokens_seen": 774438912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038653961885656976,
      "loss": 2.724,
      "theoretical_loss": 3.741183061172343,
      "tokens_seen": 774504448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865295887662989,
      "loss": 2.8625,
      "theoretical_loss": 3.7411514974267845,
      "tokens_seen": 774569984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865195586760281,
      "loss": 3.0894,
      "theoretical_loss": 3.7411199370994,
      "tokens_seen": 774635520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865095285857573,
      "loss": 2.8484,
      "theoretical_loss": 3.741088380189532,
      "tokens_seen": 774701056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864994984954865,
      "loss": 2.7995,
      "theoretical_loss": 3.7410568266965196,
      "tokens_seen": 774766592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038648946840521566,
      "loss": 2.8275,
      "theoretical_loss": 3.7410252766197045,
      "tokens_seen": 774832128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038647943831494484,
      "loss": 2.7769,
      "theoretical_loss": 3.740993729958428,
      "tokens_seen": 774897664
    },
    {
      "epoch": 0.23,
      "objective/train/docs_used": 435763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.750361680984497,
      "objective/train/theoretical_loss": 3.7409621867120313,
      "objective/train/tokens_used": 795423200,
      "theoretical_loss": 3.7409621867120313,
      "tokens_seen": 774963200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386469408224674,
      "loss": 2.6144,
      "theoretical_loss": 3.7409621867120313,
      "tokens_seen": 774963200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038645937813440326,
      "loss": 2.5263,
      "theoretical_loss": 3.7409306468798564,
      "tokens_seen": 775028736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864493480441324,
      "loss": 2.7634,
      "theoretical_loss": 3.7408991104612452,
      "tokens_seen": 775094272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864393179538616,
      "loss": 2.7988,
      "theoretical_loss": 3.740867577455539,
      "tokens_seen": 775159808
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038642928786359075,
      "loss": 2.9711,
      "theoretical_loss": 3.7408360478620803,
      "tokens_seen": 775225344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038641925777332,
      "loss": 2.7774,
      "theoretical_loss": 3.7408045216802117,
      "tokens_seen": 775290880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038640922768304917,
      "loss": 2.8231,
      "theoretical_loss": 3.740772998909275,
      "tokens_seen": 775356416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038639919759277835,
      "loss": 2.9609,
      "theoretical_loss": 3.740741479548613,
      "tokens_seen": 775421952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038638916750250753,
      "loss": 2.8501,
      "theoretical_loss": 3.74070996359757,
      "tokens_seen": 775487488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038637913741223676,
      "loss": 2.7855,
      "theoretical_loss": 3.7406784510554876,
      "tokens_seen": 775553024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863691073219659,
      "loss": 2.5942,
      "theoretical_loss": 3.7406469419217094,
      "tokens_seen": 775618560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863590772316951,
      "loss": 2.7587,
      "theoretical_loss": 3.7406154361955792,
      "tokens_seen": 775684096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038634904714142425,
      "loss": 2.4622,
      "theoretical_loss": 3.7405839338764397,
      "tokens_seen": 775749632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863390170511535,
      "loss": 2.7932,
      "theoretical_loss": 3.740552434963636,
      "tokens_seen": 775815168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038632898696088267,
      "loss": 2.8275,
      "theoretical_loss": 3.740520939456511,
      "tokens_seen": 775880704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038631895687061185,
      "loss": 2.9642,
      "theoretical_loss": 3.740489447354409,
      "tokens_seen": 775946240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038630892678034103,
      "loss": 2.7347,
      "theoretical_loss": 3.7404579586566746,
      "tokens_seen": 776011776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862988966900702,
      "loss": 2.6956,
      "theoretical_loss": 3.7404264733626524,
      "tokens_seen": 776077312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862888665997994,
      "loss": 3.0187,
      "theoretical_loss": 3.740394991471687,
      "tokens_seen": 776142848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038627883650952863,
      "loss": 2.956,
      "theoretical_loss": 3.740363512983123,
      "tokens_seen": 776208384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038626880641925775,
      "loss": 2.7489,
      "theoretical_loss": 3.740332037896306,
      "tokens_seen": 776273920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386258776328987,
      "loss": 2.8328,
      "theoretical_loss": 3.740300566210581,
      "tokens_seen": 776339456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862487462387161,
      "loss": 3.0334,
      "theoretical_loss": 3.7402690979252937,
      "tokens_seen": 776404992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038623871614844535,
      "loss": 2.8642,
      "theoretical_loss": 3.740237633039789,
      "tokens_seen": 776470528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038622868605817453,
      "loss": 2.7713,
      "theoretical_loss": 3.7402061715534134,
      "tokens_seen": 776536064
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 436694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.752096652984619,
      "objective/train/theoretical_loss": 3.740174713465512,
      "objective/train/tokens_used": 797061600,
      "theoretical_loss": 3.740174713465512,
      "tokens_seen": 776601600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862186559679037,
      "loss": 2.7836,
      "theoretical_loss": 3.740174713465512,
      "tokens_seen": 776601600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862086258776329,
      "loss": 2.7319,
      "theoretical_loss": 3.7401432587754324,
      "tokens_seen": 776667136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038619859578736213,
      "loss": 2.9838,
      "theoretical_loss": 3.7401118074825197,
      "tokens_seen": 776732672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038618856569709126,
      "loss": 2.8541,
      "theoretical_loss": 3.7400803595861203,
      "tokens_seen": 776798208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861785356068205,
      "loss": 2.7572,
      "theoretical_loss": 3.740048915085582,
      "tokens_seen": 776863744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861685055165496,
      "loss": 3.0645,
      "theoretical_loss": 3.7400174739802505,
      "tokens_seen": 776929280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038615847542627886,
      "loss": 2.8094,
      "theoretical_loss": 3.7399860362694737,
      "tokens_seen": 776994816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038614844533600804,
      "loss": 2.7469,
      "theoretical_loss": 3.7399546019525984,
      "tokens_seen": 777060352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861384152457372,
      "loss": 2.7164,
      "theoretical_loss": 3.7399231710289724,
      "tokens_seen": 777125888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861283851554664,
      "loss": 2.7987,
      "theoretical_loss": 3.7398917434979424,
      "tokens_seen": 777191424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861183550651956,
      "loss": 2.9797,
      "theoretical_loss": 3.7398603193588578,
      "tokens_seen": 777256960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038610832497492476,
      "loss": 2.8151,
      "theoretical_loss": 3.739828898611065,
      "tokens_seen": 777322496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386098294884654,
      "loss": 2.8161,
      "theoretical_loss": 3.7397974812539125,
      "tokens_seen": 777388032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860882647943831,
      "loss": 2.7239,
      "theoretical_loss": 3.739766067286749,
      "tokens_seen": 777453568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038607823470411236,
      "loss": 2.7941,
      "theoretical_loss": 3.739734656708923,
      "tokens_seen": 777519104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860682046138415,
      "loss": 2.7307,
      "theoretical_loss": 3.739703249519783,
      "tokens_seen": 777584640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860581745235707,
      "loss": 2.7405,
      "theoretical_loss": 3.7396718457186777,
      "tokens_seen": 777650176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860481444332999,
      "loss": 2.8985,
      "theoretical_loss": 3.739640445304957,
      "tokens_seen": 777715712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860381143430291,
      "loss": 2.7029,
      "theoretical_loss": 3.7396090482779685,
      "tokens_seen": 777781248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038602808425275826,
      "loss": 2.7713,
      "theoretical_loss": 3.7395776546370634,
      "tokens_seen": 777846784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860180541624875,
      "loss": 3.0061,
      "theoretical_loss": 3.7395462643815893,
      "tokens_seen": 777912320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860080240722166,
      "loss": 2.7808,
      "theoretical_loss": 3.739514877510898,
      "tokens_seen": 777977856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038599799398194586,
      "loss": 2.9553,
      "theoretical_loss": 3.739483494024338,
      "tokens_seen": 778043392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000385987963891675,
      "loss": 2.9095,
      "theoretical_loss": 3.73945211392126,
      "tokens_seen": 778108928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859779338014042,
      "loss": 2.862,
      "theoretical_loss": 3.739420737201015,
      "tokens_seen": 778174464
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 437315,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8093600273132324,
      "objective/train/theoretical_loss": 3.7393893638629514,
      "objective/train/tokens_used": 798700000,
      "theoretical_loss": 3.7393893638629514,
      "tokens_seen": 778240000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859679037111334,
      "loss": 2.887,
      "theoretical_loss": 3.7393893638629514,
      "tokens_seen": 778240000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859578736208626,
      "loss": 2.8717,
      "theoretical_loss": 3.739357993906422,
      "tokens_seen": 778305536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038594784353059177,
      "loss": 2.6532,
      "theoretical_loss": 3.739326627330777,
      "tokens_seen": 778371072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038593781344032095,
      "loss": 2.8611,
      "theoretical_loss": 3.7392952641353663,
      "tokens_seen": 778436608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038592778335005013,
      "loss": 2.6005,
      "theoretical_loss": 3.739263904319543,
      "tokens_seen": 778502144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038591775325977937,
      "loss": 2.7977,
      "theoretical_loss": 3.739232547882657,
      "tokens_seen": 778567680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003859077231695085,
      "loss": 2.921,
      "theoretical_loss": 3.73920119482406,
      "tokens_seen": 778633216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038589769307923773,
      "loss": 2.6604,
      "theoretical_loss": 3.7391698451431044,
      "tokens_seen": 778698752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038588766298896685,
      "loss": 2.8638,
      "theoretical_loss": 3.7391384988391416,
      "tokens_seen": 778764288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858776328986961,
      "loss": 3.0366,
      "theoretical_loss": 3.7391071559115243,
      "tokens_seen": 778829824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858676028084253,
      "loss": 2.8517,
      "theoretical_loss": 3.7390758163596036,
      "tokens_seen": 778895360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038585757271815445,
      "loss": 2.8778,
      "theoretical_loss": 3.739044480182733,
      "tokens_seen": 778960896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858475426278837,
      "loss": 2.49,
      "theoretical_loss": 3.7390131473802652,
      "tokens_seen": 779026432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038583751253761287,
      "loss": 2.9749,
      "theoretical_loss": 3.7389818179515517,
      "tokens_seen": 779091968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038582748244734205,
      "loss": 2.7421,
      "theoretical_loss": 3.738950491895947,
      "tokens_seen": 779157504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038581745235707123,
      "loss": 2.663,
      "theoretical_loss": 3.738919169212804,
      "tokens_seen": 779223040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858074222668004,
      "loss": 2.7339,
      "theoretical_loss": 3.7388878499014746,
      "tokens_seen": 779288576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857973921765296,
      "loss": 2.6934,
      "theoretical_loss": 3.7388565339613145,
      "tokens_seen": 779354112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038578736208625883,
      "loss": 2.7529,
      "theoretical_loss": 3.738825221391676,
      "tokens_seen": 779419648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038577733199598796,
      "loss": 2.9809,
      "theoretical_loss": 3.738793912191913,
      "tokens_seen": 779485184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857673019057172,
      "loss": 2.6457,
      "theoretical_loss": 3.7387626063613797,
      "tokens_seen": 779550720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857572718154463,
      "loss": 2.8474,
      "theoretical_loss": 3.7387313038994305,
      "tokens_seen": 779616256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038574724172517555,
      "loss": 2.6877,
      "theoretical_loss": 3.7387000048054198,
      "tokens_seen": 779681792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038573721163490473,
      "loss": 2.9636,
      "theoretical_loss": 3.738668709078702,
      "tokens_seen": 779747328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857271815446339,
      "loss": 2.7967,
      "theoretical_loss": 3.7386374167186327,
      "tokens_seen": 779812864
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 438429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.841843843460083,
      "objective/train/theoretical_loss": 3.7386061277245655,
      "objective/train/tokens_used": 800338400,
      "theoretical_loss": 3.7386061277245655,
      "tokens_seen": 779878400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857171514543631,
      "loss": 3.0358,
      "theoretical_loss": 3.7386061277245655,
      "tokens_seen": 779878400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038570712136409233,
      "loss": 3.0418,
      "theoretical_loss": 3.7385748420958564,
      "tokens_seen": 779943936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038569709127382146,
      "loss": 2.6862,
      "theoretical_loss": 3.7385435598318604,
      "tokens_seen": 780009472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856870611835507,
      "loss": 2.7656,
      "theoretical_loss": 3.738512280931933,
      "tokens_seen": 780075008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856770310932798,
      "loss": 2.7398,
      "theoretical_loss": 3.73848100539543,
      "tokens_seen": 780140544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038566700100300906,
      "loss": 2.5016,
      "theoretical_loss": 3.7384497332217075,
      "tokens_seen": 780206080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038565697091273824,
      "loss": 2.6233,
      "theoretical_loss": 3.7384184644101204,
      "tokens_seen": 780271616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856469408224674,
      "loss": 2.7823,
      "theoretical_loss": 3.7383871989600257,
      "tokens_seen": 780337152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856369107321966,
      "loss": 2.7999,
      "theoretical_loss": 3.7383559368707804,
      "tokens_seen": 780402688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856268806419258,
      "loss": 2.6636,
      "theoretical_loss": 3.7383246781417396,
      "tokens_seen": 780468224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038561685055165496,
      "loss": 2.7589,
      "theoretical_loss": 3.738293422772261,
      "tokens_seen": 780533760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856068204613842,
      "loss": 2.6067,
      "theoretical_loss": 3.738262170761701,
      "tokens_seen": 780599296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855967903711133,
      "loss": 3.0154,
      "theoretical_loss": 3.738230922109417,
      "tokens_seen": 780664832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038558676028084256,
      "loss": 2.6455,
      "theoretical_loss": 3.7381996768147663,
      "tokens_seen": 780730368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855767301905717,
      "loss": 2.9596,
      "theoretical_loss": 3.738168434877106,
      "tokens_seen": 780795904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855667001003009,
      "loss": 2.868,
      "theoretical_loss": 3.738137196295794,
      "tokens_seen": 780861440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855566700100301,
      "loss": 2.8184,
      "theoretical_loss": 3.738105961070188,
      "tokens_seen": 780926976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855466399197593,
      "loss": 2.8756,
      "theoretical_loss": 3.7380747291996457,
      "tokens_seen": 780992512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038553660982948846,
      "loss": 2.7564,
      "theoretical_loss": 3.738043500683526,
      "tokens_seen": 781058048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855265797392177,
      "loss": 2.8796,
      "theoretical_loss": 3.7380122755211858,
      "tokens_seen": 781123584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855165496489468,
      "loss": 2.8602,
      "theoretical_loss": 3.7379810537119846,
      "tokens_seen": 781189120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038550651955867606,
      "loss": 2.7505,
      "theoretical_loss": 3.7379498352552814,
      "tokens_seen": 781254656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854964894684052,
      "loss": 2.3087,
      "theoretical_loss": 3.7379186201504346,
      "tokens_seen": 781320192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854864593781344,
      "loss": 2.7144,
      "theoretical_loss": 3.737887408396803,
      "tokens_seen": 781385728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854764292878636,
      "loss": 2.4728,
      "theoretical_loss": 3.737856199993746,
      "tokens_seen": 781451264
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 439045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9254188537597656,
      "objective/train/theoretical_loss": 3.737824994940622,
      "objective/train/tokens_used": 801976800,
      "theoretical_loss": 3.737824994940622,
      "tokens_seen": 781516800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854663991975928,
      "loss": 2.8778,
      "theoretical_loss": 3.737824994940622,
      "tokens_seen": 781516800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038545636910732197,
      "loss": 2.9068,
      "theoretical_loss": 3.7377937932367926,
      "tokens_seen": 781582336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038544633901705115,
      "loss": 2.5584,
      "theoretical_loss": 3.737762594881616,
      "tokens_seen": 781647872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038543630892678033,
      "loss": 2.6853,
      "theoretical_loss": 3.7377313998744524,
      "tokens_seen": 781713408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038542627883650957,
      "loss": 2.6201,
      "theoretical_loss": 3.7377002082146618,
      "tokens_seen": 781778944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854162487462387,
      "loss": 2.8869,
      "theoretical_loss": 3.737669019901605,
      "tokens_seen": 781844480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038540621865596793,
      "loss": 2.7393,
      "theoretical_loss": 3.7376378349346417,
      "tokens_seen": 781910016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038539618856569705,
      "loss": 2.868,
      "theoretical_loss": 3.737606653313133,
      "tokens_seen": 781975552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853861584754263,
      "loss": 2.9731,
      "theoretical_loss": 3.737575475036439,
      "tokens_seen": 782041088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038537612838515547,
      "loss": 2.7581,
      "theoretical_loss": 3.737544300103922,
      "tokens_seen": 782106624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038536609829488465,
      "loss": 2.7906,
      "theoretical_loss": 3.7375131285149417,
      "tokens_seen": 782172160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038535606820461383,
      "loss": 2.7958,
      "theoretical_loss": 3.73748196026886,
      "tokens_seen": 782237696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038534603811434307,
      "loss": 2.7489,
      "theoretical_loss": 3.7374507953650387,
      "tokens_seen": 782303232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853360080240722,
      "loss": 2.8457,
      "theoretical_loss": 3.7374196338028387,
      "tokens_seen": 782368768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038532597793380143,
      "loss": 2.6944,
      "theoretical_loss": 3.7373884755816222,
      "tokens_seen": 782434304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038531594784353056,
      "loss": 2.95,
      "theoretical_loss": 3.7373573207007516,
      "tokens_seen": 782499840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853059177532598,
      "loss": 2.8403,
      "theoretical_loss": 3.737326169159589,
      "tokens_seen": 782565376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000385295887662989,
      "loss": 2.7555,
      "theoretical_loss": 3.7372950209574958,
      "tokens_seen": 782630912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038528585757271816,
      "loss": 2.7885,
      "theoretical_loss": 3.7372638760938353,
      "tokens_seen": 782696448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038527582748244734,
      "loss": 2.7648,
      "theoretical_loss": 3.7372327345679706,
      "tokens_seen": 782761984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852657973921765,
      "loss": 3.0226,
      "theoretical_loss": 3.737201596379263,
      "tokens_seen": 782827520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852557673019057,
      "loss": 2.7361,
      "theoretical_loss": 3.7371704615270778,
      "tokens_seen": 782893056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038524573721163493,
      "loss": 2.8869,
      "theoretical_loss": 3.7371393300107765,
      "tokens_seen": 782958592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038523570712136406,
      "loss": 2.7412,
      "theoretical_loss": 3.7371082018297233,
      "tokens_seen": 783024128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852256770310933,
      "loss": 2.7908,
      "theoretical_loss": 3.7370770769832813,
      "tokens_seen": 783089664
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 440183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6950714588165283,
      "objective/train/theoretical_loss": 3.7370459554708146,
      "objective/train/tokens_used": 803615200,
      "theoretical_loss": 3.7370459554708146,
      "tokens_seen": 783155200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852156469408225,
      "loss": 2.6978,
      "theoretical_loss": 3.7370459554708146,
      "tokens_seen": 783155200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038520561685055166,
      "loss": 2.9029,
      "theoretical_loss": 3.737014837291687,
      "tokens_seen": 783220736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038519558676028084,
      "loss": 2.8232,
      "theoretical_loss": 3.736983722445263,
      "tokens_seen": 783286272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038518555667001,
      "loss": 2.7811,
      "theoretical_loss": 3.7369526109309064,
      "tokens_seen": 783351808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851755265797392,
      "loss": 2.783,
      "theoretical_loss": 3.7369215027479816,
      "tokens_seen": 783417344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038516549648946844,
      "loss": 2.7497,
      "theoretical_loss": 3.736890397895853,
      "tokens_seen": 783482880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038515546639919756,
      "loss": 2.7636,
      "theoretical_loss": 3.7368592963738863,
      "tokens_seen": 783548416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851454363089268,
      "loss": 2.6221,
      "theoretical_loss": 3.736828198181446,
      "tokens_seen": 783613952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851354062186559,
      "loss": 3.0129,
      "theoretical_loss": 3.736797103317897,
      "tokens_seen": 783679488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038512537612838516,
      "loss": 2.7719,
      "theoretical_loss": 3.7367660117826054,
      "tokens_seen": 783745024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851153460381144,
      "loss": 2.8759,
      "theoretical_loss": 3.736734923574935,
      "tokens_seen": 783810560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851053159478435,
      "loss": 2.8192,
      "theoretical_loss": 3.736703838694254,
      "tokens_seen": 783876096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038509528585757276,
      "loss": 2.7093,
      "theoretical_loss": 3.7366727571399254,
      "tokens_seen": 783941632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850852557673019,
      "loss": 2.8386,
      "theoretical_loss": 3.736641678911318,
      "tokens_seen": 784007168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850752256770311,
      "loss": 2.8519,
      "theoretical_loss": 3.7366106040077955,
      "tokens_seen": 784072704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850651955867603,
      "loss": 2.9498,
      "theoretical_loss": 3.736579532428726,
      "tokens_seen": 784138240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850551654964895,
      "loss": 2.9198,
      "theoretical_loss": 3.7365484641734756,
      "tokens_seen": 784203776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038504513540621866,
      "loss": 2.7531,
      "theoretical_loss": 3.7365173992414107,
      "tokens_seen": 784269312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850351053159479,
      "loss": 2.6193,
      "theoretical_loss": 3.7364863376318986,
      "tokens_seen": 784334848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000385025075225677,
      "loss": 3.0243,
      "theoretical_loss": 3.736455279344306,
      "tokens_seen": 784400384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038501504513540626,
      "loss": 2.92,
      "theoretical_loss": 3.736424224378,
      "tokens_seen": 784465920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850050150451354,
      "loss": 2.9541,
      "theoretical_loss": 3.736393172732349,
      "tokens_seen": 784531456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849949849548646,
      "loss": 2.8379,
      "theoretical_loss": 3.7363621244067193,
      "tokens_seen": 784596992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849849548645938,
      "loss": 2.7373,
      "theoretical_loss": 3.736331079400479,
      "tokens_seen": 784662528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384974924774323,
      "loss": 2.803,
      "theoretical_loss": 3.7363000377129962,
      "tokens_seen": 784728064
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 440843,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.821906566619873,
      "objective/train/theoretical_loss": 3.7362689993436398,
      "objective/train/tokens_used": 805253600,
      "theoretical_loss": 3.7362689993436398,
      "tokens_seen": 784793600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038496489468405217,
      "loss": 2.7818,
      "theoretical_loss": 3.7362689993436398,
      "tokens_seen": 784793600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038495486459378135,
      "loss": 2.7273,
      "theoretical_loss": 3.7362379642917767,
      "tokens_seen": 784859136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038494483450351053,
      "loss": 2.6269,
      "theoretical_loss": 3.736206932556776,
      "tokens_seen": 784924672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038493480441323977,
      "loss": 2.8092,
      "theoretical_loss": 3.7361759041380065,
      "tokens_seen": 784990208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849247743229689,
      "loss": 2.8882,
      "theoretical_loss": 3.736144879034837,
      "tokens_seen": 785055744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038491474423269813,
      "loss": 2.5399,
      "theoretical_loss": 3.736113857246636,
      "tokens_seen": 785121280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038490471414242725,
      "loss": 2.6503,
      "theoretical_loss": 3.7360828387727727,
      "tokens_seen": 785186816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848946840521565,
      "loss": 2.7638,
      "theoretical_loss": 3.736051823612616,
      "tokens_seen": 785252352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038488465396188567,
      "loss": 2.6412,
      "theoretical_loss": 3.736020811765537,
      "tokens_seen": 785317888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038487462387161485,
      "loss": 2.8907,
      "theoretical_loss": 3.735989803230904,
      "tokens_seen": 785383424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038486459378134403,
      "loss": 2.7288,
      "theoretical_loss": 3.7359587980080873,
      "tokens_seen": 785448960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038485456369107327,
      "loss": 2.9444,
      "theoretical_loss": 3.7359277960964565,
      "tokens_seen": 785514496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848445336008024,
      "loss": 2.9635,
      "theoretical_loss": 3.735896797495382,
      "tokens_seen": 785580032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038483450351053163,
      "loss": 2.8273,
      "theoretical_loss": 3.735865802204234,
      "tokens_seen": 785645568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038482447342026076,
      "loss": 2.6847,
      "theoretical_loss": 3.735834810222384,
      "tokens_seen": 785711104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038481444332999,
      "loss": 2.6525,
      "theoretical_loss": 3.735803821549201,
      "tokens_seen": 785776640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848044132397192,
      "loss": 2.6116,
      "theoretical_loss": 3.735772836184057,
      "tokens_seen": 785842176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038479438314944836,
      "loss": 2.8336,
      "theoretical_loss": 3.7357418541263234,
      "tokens_seen": 785907712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038478435305917754,
      "loss": 2.9349,
      "theoretical_loss": 3.7357108753753705,
      "tokens_seen": 785973248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847743229689067,
      "loss": 2.7884,
      "theoretical_loss": 3.7356798999305694,
      "tokens_seen": 786038784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847642928786359,
      "loss": 2.5847,
      "theoretical_loss": 3.7356489277912925,
      "tokens_seen": 786104320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038475426278836513,
      "loss": 2.9119,
      "theoretical_loss": 3.7356179589569116,
      "tokens_seen": 786169856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038474423269809426,
      "loss": 2.5322,
      "theoretical_loss": 3.735586993426798,
      "tokens_seen": 786235392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847342026078235,
      "loss": 2.5874,
      "theoretical_loss": 3.735556031200324,
      "tokens_seen": 786300928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847241725175527,
      "loss": 2.9919,
      "theoretical_loss": 3.7355250722768614,
      "tokens_seen": 786366464
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 442027,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.940617561340332,
      "objective/train/theoretical_loss": 3.735494116655784,
      "objective/train/tokens_used": 806892000,
      "theoretical_loss": 3.735494116655784,
      "tokens_seen": 786432000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038471414242728186,
      "loss": 2.9008,
      "theoretical_loss": 3.735494116655784,
      "tokens_seen": 786432000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038470411233701104,
      "loss": 2.8244,
      "theoretical_loss": 3.735463164336462,
      "tokens_seen": 786497536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846940822467402,
      "loss": 2.8714,
      "theoretical_loss": 3.7354322153182706,
      "tokens_seen": 786563072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846840521564694,
      "loss": 2.9422,
      "theoretical_loss": 3.7354012696005814,
      "tokens_seen": 786628608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038467402206619864,
      "loss": 2.7046,
      "theoretical_loss": 3.735370327182768,
      "tokens_seen": 786694144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038466399197592776,
      "loss": 2.5513,
      "theoretical_loss": 3.7353393880642036,
      "tokens_seen": 786759680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384653961885657,
      "loss": 2.9961,
      "theoretical_loss": 3.7353084522442614,
      "tokens_seen": 786825216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846439317953861,
      "loss": 2.7466,
      "theoretical_loss": 3.735277519722315,
      "tokens_seen": 786890752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038463390170511536,
      "loss": 2.7936,
      "theoretical_loss": 3.735246590497738,
      "tokens_seen": 786956288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038462387161484454,
      "loss": 2.6091,
      "theoretical_loss": 3.735215664569905,
      "tokens_seen": 787021824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846138415245737,
      "loss": 2.5691,
      "theoretical_loss": 3.7351847419381894,
      "tokens_seen": 787087360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846038114343029,
      "loss": 2.775,
      "theoretical_loss": 3.7351538226019665,
      "tokens_seen": 787152896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845937813440321,
      "loss": 2.9211,
      "theoretical_loss": 3.7351229065606093,
      "tokens_seen": 787218432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038458375125376127,
      "loss": 2.8669,
      "theoretical_loss": 3.735091993813494,
      "tokens_seen": 787283968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845737211634905,
      "loss": 2.7544,
      "theoretical_loss": 3.7350610843599945,
      "tokens_seen": 787349504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038456369107321963,
      "loss": 2.7132,
      "theoretical_loss": 3.7350301781994855,
      "tokens_seen": 787415040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038455366098294886,
      "loss": 2.8449,
      "theoretical_loss": 3.734999275331343,
      "tokens_seen": 787480576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038454363089267805,
      "loss": 2.8362,
      "theoretical_loss": 3.7349683757549412,
      "tokens_seen": 787546112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038453360080240723,
      "loss": 2.607,
      "theoretical_loss": 3.7349374794696573,
      "tokens_seen": 787611648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845235707121364,
      "loss": 2.7922,
      "theoretical_loss": 3.734906586474865,
      "tokens_seen": 787677184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845135406218656,
      "loss": 2.7689,
      "theoretical_loss": 3.734875696769942,
      "tokens_seen": 787742720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038450351053159477,
      "loss": 2.8311,
      "theoretical_loss": 3.7348448103542626,
      "tokens_seen": 787808256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384493480441324,
      "loss": 2.7133,
      "theoretical_loss": 3.734813927227204,
      "tokens_seen": 787873792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038448345035105313,
      "loss": 2.7541,
      "theoretical_loss": 3.734783047388142,
      "tokens_seen": 787939328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038447342026078237,
      "loss": 2.7583,
      "theoretical_loss": 3.7347521708364537,
      "tokens_seen": 788004864
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 442587,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.145890474319458,
      "objective/train/theoretical_loss": 3.7347212975715154,
      "objective/train/tokens_used": 808530400,
      "theoretical_loss": 3.7347212975715154,
      "tokens_seen": 788070400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844633901705115,
      "loss": 2.9654,
      "theoretical_loss": 3.7347212975715154,
      "tokens_seen": 788070400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038445336008024073,
      "loss": 2.7115,
      "theoretical_loss": 3.734690427592704,
      "tokens_seen": 788135936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844433299899699,
      "loss": 2.3644,
      "theoretical_loss": 3.7346595608993964,
      "tokens_seen": 788201472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844332998996991,
      "loss": 2.6955,
      "theoretical_loss": 3.73462869749097,
      "tokens_seen": 788267008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844232698094283,
      "loss": 2.8928,
      "theoretical_loss": 3.7345978373668025,
      "tokens_seen": 788332544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038441323971915745,
      "loss": 2.5706,
      "theoretical_loss": 3.73456698052627,
      "tokens_seen": 788398080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038440320962888664,
      "loss": 2.6816,
      "theoretical_loss": 3.7345361269687523,
      "tokens_seen": 788463616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038439317953861587,
      "loss": 2.6384,
      "theoretical_loss": 3.7345052766936253,
      "tokens_seen": 788529152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038438314944834505,
      "loss": 2.5129,
      "theoretical_loss": 3.7344744297002688,
      "tokens_seen": 788594688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038437311935807423,
      "loss": 2.6681,
      "theoretical_loss": 3.7344435859880596,
      "tokens_seen": 788660224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038436308926780347,
      "loss": 2.7354,
      "theoretical_loss": 3.7344127455563765,
      "tokens_seen": 788725760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843530591775326,
      "loss": 2.5316,
      "theoretical_loss": 3.7343819084045977,
      "tokens_seen": 788791296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038434302908726183,
      "loss": 2.7936,
      "theoretical_loss": 3.734351074532103,
      "tokens_seen": 788856832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038433299899699096,
      "loss": 2.7816,
      "theoretical_loss": 3.7343202439382703,
      "tokens_seen": 788922368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843229689067202,
      "loss": 2.5337,
      "theoretical_loss": 3.734289416622479,
      "tokens_seen": 788987904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843129388164494,
      "loss": 2.655,
      "theoretical_loss": 3.7342585925841085,
      "tokens_seen": 789053440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038430290872617856,
      "loss": 2.6479,
      "theoretical_loss": 3.734227771822538,
      "tokens_seen": 789118976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038429287863590774,
      "loss": 2.5965,
      "theoretical_loss": 3.7341969543371474,
      "tokens_seen": 789184512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842828485456369,
      "loss": 2.8551,
      "theoretical_loss": 3.7341661401273156,
      "tokens_seen": 789250048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842728184553661,
      "loss": 2.9743,
      "theoretical_loss": 3.734135329192423,
      "tokens_seen": 789315584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038426278836509533,
      "loss": 2.6725,
      "theoretical_loss": 3.73410452153185,
      "tokens_seen": 789381120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038425275827482446,
      "loss": 2.8066,
      "theoretical_loss": 3.734073717144976,
      "tokens_seen": 789446656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842427281845537,
      "loss": 2.9044,
      "theoretical_loss": 3.734042916031182,
      "tokens_seen": 789512192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842326980942829,
      "loss": 2.7009,
      "theoretical_loss": 3.734012118189849,
      "tokens_seen": 789577728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038422266800401206,
      "loss": 2.6561,
      "theoretical_loss": 3.7339813236203567,
      "tokens_seen": 789643264
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 443777,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3760197162628174,
      "objective/train/theoretical_loss": 3.733950532322087,
      "objective/train/tokens_used": 810168800,
      "theoretical_loss": 3.733950532322087,
      "tokens_seen": 789708800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038421263791374124,
      "loss": 3.0686,
      "theoretical_loss": 3.733950532322087,
      "tokens_seen": 789708800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842026078234704,
      "loss": 2.6925,
      "theoretical_loss": 3.73391974429442,
      "tokens_seen": 789774336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841925777331996,
      "loss": 2.6252,
      "theoretical_loss": 3.7338889595367375,
      "tokens_seen": 789839872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038418254764292884,
      "loss": 2.6106,
      "theoretical_loss": 3.733858178048421,
      "tokens_seen": 789905408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038417251755265796,
      "loss": 2.6877,
      "theoretical_loss": 3.7338273998288516,
      "tokens_seen": 789970944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841624874623872,
      "loss": 2.6514,
      "theoretical_loss": 3.733796624877412,
      "tokens_seen": 790036480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841524573721163,
      "loss": 2.515,
      "theoretical_loss": 3.7337658531934834,
      "tokens_seen": 790102016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038414242728184556,
      "loss": 2.8697,
      "theoretical_loss": 3.733735084776448,
      "tokens_seen": 790167552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038413239719157474,
      "loss": 2.6622,
      "theoretical_loss": 3.733704319625688,
      "tokens_seen": 790233088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841223671013039,
      "loss": 2.8917,
      "theoretical_loss": 3.7336735577405857,
      "tokens_seen": 790298624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841123370110331,
      "loss": 2.5297,
      "theoretical_loss": 3.733642799120524,
      "tokens_seen": 790364160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841023069207623,
      "loss": 2.7386,
      "theoretical_loss": 3.733612043764886,
      "tokens_seen": 790429696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038409227683049147,
      "loss": 2.8363,
      "theoretical_loss": 3.7335812916730537,
      "tokens_seen": 790495232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840822467402207,
      "loss": 2.917,
      "theoretical_loss": 3.733550542844411,
      "tokens_seen": 790560768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038407221664994983,
      "loss": 2.7862,
      "theoretical_loss": 3.7335197972783405,
      "tokens_seen": 790626304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038406218655967907,
      "loss": 2.6633,
      "theoretical_loss": 3.7334890549742257,
      "tokens_seen": 790691840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038405215646940825,
      "loss": 2.7952,
      "theoretical_loss": 3.7334583159314514,
      "tokens_seen": 790757376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038404212637913743,
      "loss": 2.7748,
      "theoretical_loss": 3.7334275801494,
      "tokens_seen": 790822912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840320962888666,
      "loss": 2.6345,
      "theoretical_loss": 3.733396847627456,
      "tokens_seen": 790888448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840220661985958,
      "loss": 2.6611,
      "theoretical_loss": 3.7333661183650033,
      "tokens_seen": 790953984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038401203610832497,
      "loss": 2.7081,
      "theoretical_loss": 3.7333353923614263,
      "tokens_seen": 791019520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840020060180542,
      "loss": 2.5762,
      "theoretical_loss": 3.7333046696161087,
      "tokens_seen": 791085056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038399197592778333,
      "loss": 2.9757,
      "theoretical_loss": 3.733273950128437,
      "tokens_seen": 791150592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038398194583751257,
      "loss": 2.7254,
      "theoretical_loss": 3.733243233897794,
      "tokens_seen": 791216128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839719157472417,
      "loss": 2.7849,
      "theoretical_loss": 3.7332125209235656,
      "tokens_seen": 791281664
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 444379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.697193145751953,
      "objective/train/theoretical_loss": 3.7331818112051365,
      "objective/train/tokens_used": 811807200,
      "theoretical_loss": 3.7331818112051365,
      "tokens_seen": 791347200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038396188565697093,
      "loss": 2.9443,
      "theoretical_loss": 3.7331818112051365,
      "tokens_seen": 791347200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839518555667001,
      "loss": 2.7015,
      "theoretical_loss": 3.7331511047418924,
      "tokens_seen": 791412736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839418254764293,
      "loss": 2.8082,
      "theoretical_loss": 3.7331204015332187,
      "tokens_seen": 791478272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003839317953861585,
      "loss": 2.6834,
      "theoretical_loss": 3.7330897015785003,
      "tokens_seen": 791543808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038392176529588765,
      "loss": 2.7401,
      "theoretical_loss": 3.733059004877124,
      "tokens_seen": 791609344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038391173520561684,
      "loss": 2.4514,
      "theoretical_loss": 3.733028311428475,
      "tokens_seen": 791674880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038390170511534607,
      "loss": 2.6914,
      "theoretical_loss": 3.73299762123194,
      "tokens_seen": 791740416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838916750250752,
      "loss": 2.843,
      "theoretical_loss": 3.732966934286904,
      "tokens_seen": 791805952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038388164493480443,
      "loss": 2.693,
      "theoretical_loss": 3.7329362505927555,
      "tokens_seen": 791871488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838716148445336,
      "loss": 2.7452,
      "theoretical_loss": 3.73290557014888,
      "tokens_seen": 791937024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838615847542628,
      "loss": 2.4392,
      "theoretical_loss": 3.732874892954663,
      "tokens_seen": 792002560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383851554663992,
      "loss": 2.6381,
      "theoretical_loss": 3.7328442190094937,
      "tokens_seen": 792068096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038384152457372116,
      "loss": 2.7508,
      "theoretical_loss": 3.732813548312758,
      "tokens_seen": 792133632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038383149448345034,
      "loss": 2.6207,
      "theoretical_loss": 3.7327828808638435,
      "tokens_seen": 792199168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838214643931796,
      "loss": 2.7213,
      "theoretical_loss": 3.732752216662137,
      "tokens_seen": 792264704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838114343029087,
      "loss": 2.6121,
      "theoretical_loss": 3.732721555707027,
      "tokens_seen": 792330240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038380140421263794,
      "loss": 2.937,
      "theoretical_loss": 3.7326908979979008,
      "tokens_seen": 792395776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038379137412236706,
      "loss": 2.624,
      "theoretical_loss": 3.7326602435341467,
      "tokens_seen": 792461312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837813440320963,
      "loss": 2.928,
      "theoretical_loss": 3.732629592315152,
      "tokens_seen": 792526848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837713139418255,
      "loss": 2.8103,
      "theoretical_loss": 3.732598944340306,
      "tokens_seen": 792592384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038376128385155466,
      "loss": 2.6622,
      "theoretical_loss": 3.732568299608996,
      "tokens_seen": 792657920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038375125376128384,
      "loss": 2.7235,
      "theoretical_loss": 3.732537658120612,
      "tokens_seen": 792723456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837412236710131,
      "loss": 2.7727,
      "theoretical_loss": 3.7325070198745416,
      "tokens_seen": 792788992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837311935807422,
      "loss": 2.8658,
      "theoretical_loss": 3.732476384870174,
      "tokens_seen": 792854528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038372116349047144,
      "loss": 2.6207,
      "theoretical_loss": 3.732445753106899,
      "tokens_seen": 792920064
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 445600,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.453542947769165,
      "objective/train/theoretical_loss": 3.7324151245841044,
      "objective/train/tokens_used": 813445600,
      "theoretical_loss": 3.7324151245841044,
      "tokens_seen": 792985600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038371113340020057,
      "loss": 2.8558,
      "theoretical_loss": 3.7324151245841044,
      "tokens_seen": 792985600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837011033099298,
      "loss": 2.7573,
      "theoretical_loss": 3.7323844993011814,
      "tokens_seen": 793051136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383691073219659,
      "loss": 2.7434,
      "theoretical_loss": 3.732353877257518,
      "tokens_seen": 793116672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038368104312938816,
      "loss": 2.7893,
      "theoretical_loss": 3.732323258452505,
      "tokens_seen": 793182208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038367101303911735,
      "loss": 2.8657,
      "theoretical_loss": 3.7322926428855316,
      "tokens_seen": 793247744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836609829488465,
      "loss": 2.8876,
      "theoretical_loss": 3.7322620305559893,
      "tokens_seen": 793313280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836509528585757,
      "loss": 2.7263,
      "theoretical_loss": 3.732231421463266,
      "tokens_seen": 793378816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038364092276830494,
      "loss": 2.7745,
      "theoretical_loss": 3.7322008156067543,
      "tokens_seen": 793444352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836308926780341,
      "loss": 2.8643,
      "theoretical_loss": 3.732170212985844,
      "tokens_seen": 793509888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836208625877633,
      "loss": 2.6559,
      "theoretical_loss": 3.7321396135999256,
      "tokens_seen": 793575424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836108324974925,
      "loss": 2.6391,
      "theoretical_loss": 3.73210901744839,
      "tokens_seen": 793640960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038360080240722167,
      "loss": 2.76,
      "theoretical_loss": 3.732078424530629,
      "tokens_seen": 793706496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835907723169509,
      "loss": 2.8954,
      "theoretical_loss": 3.732047834846033,
      "tokens_seen": 793772032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038358074222668003,
      "loss": 2.6983,
      "theoretical_loss": 3.732017248393994,
      "tokens_seen": 793837568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038357071213640927,
      "loss": 2.628,
      "theoretical_loss": 3.7319866651739035,
      "tokens_seen": 793903104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038356068204613845,
      "loss": 2.7902,
      "theoretical_loss": 3.7319560851851525,
      "tokens_seen": 793968640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038355065195586763,
      "loss": 2.6634,
      "theoretical_loss": 3.7319255084271337,
      "tokens_seen": 794034176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835406218655968,
      "loss": 2.8909,
      "theoretical_loss": 3.731894934899239,
      "tokens_seen": 794099712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383530591775326,
      "loss": 2.7643,
      "theoretical_loss": 3.7318643646008605,
      "tokens_seen": 794165248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038352056168505517,
      "loss": 2.7051,
      "theoretical_loss": 3.731833797531391,
      "tokens_seen": 794230784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835105315947844,
      "loss": 2.8871,
      "theoretical_loss": 3.731803233690222,
      "tokens_seen": 794296320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038350050150451353,
      "loss": 2.9249,
      "theoretical_loss": 3.7317726730767475,
      "tokens_seen": 794361856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038349047141424277,
      "loss": 2.7298,
      "theoretical_loss": 3.73174211569036,
      "tokens_seen": 794427392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834804413239719,
      "loss": 2.8595,
      "theoretical_loss": 3.731711561530453,
      "tokens_seen": 794492928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038347041123370113,
      "loss": 2.9234,
      "theoretical_loss": 3.731681010596418,
      "tokens_seen": 794558464
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 446700,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2344651222229004,
      "objective/train/theoretical_loss": 3.73165046288765,
      "objective/train/tokens_used": 815084000,
      "theoretical_loss": 3.73165046288765,
      "tokens_seen": 794624000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834603811434303,
      "loss": 2.5192,
      "theoretical_loss": 3.73165046288765,
      "tokens_seen": 794624000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834503510531595,
      "loss": 2.7426,
      "theoretical_loss": 3.731619918403542,
      "tokens_seen": 794689536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834403209628887,
      "loss": 2.6839,
      "theoretical_loss": 3.731589377143488,
      "tokens_seen": 794755072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038343029087261786,
      "loss": 2.638,
      "theoretical_loss": 3.7315588391068815,
      "tokens_seen": 794820608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038342026078234704,
      "loss": 2.7918,
      "theoretical_loss": 3.731528304293117,
      "tokens_seen": 794886144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038341023069207627,
      "loss": 2.7242,
      "theoretical_loss": 3.731497772701588,
      "tokens_seen": 794951680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834002006018054,
      "loss": 2.7139,
      "theoretical_loss": 3.7314672443316894,
      "tokens_seen": 795017216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038339017051153463,
      "loss": 2.8092,
      "theoretical_loss": 3.7314367191828155,
      "tokens_seen": 795082752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833801404212638,
      "loss": 2.3607,
      "theoretical_loss": 3.731406197254361,
      "tokens_seen": 795148288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383370110330993,
      "loss": 2.7424,
      "theoretical_loss": 3.7313756785457217,
      "tokens_seen": 795213824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833600802407222,
      "loss": 2.669,
      "theoretical_loss": 3.7313451630562913,
      "tokens_seen": 795279360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038335005015045136,
      "loss": 2.8487,
      "theoretical_loss": 3.731314650785465,
      "tokens_seen": 795344896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038334002006018054,
      "loss": 2.7453,
      "theoretical_loss": 3.7312841417326394,
      "tokens_seen": 795410432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833299899699098,
      "loss": 2.7604,
      "theoretical_loss": 3.7312536358972084,
      "tokens_seen": 795475968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833199598796389,
      "loss": 2.5242,
      "theoretical_loss": 3.7312231332785695,
      "tokens_seen": 795541504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038330992978936814,
      "loss": 2.9914,
      "theoretical_loss": 3.731192633876117,
      "tokens_seen": 795607040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038329989969909726,
      "loss": 2.7639,
      "theoretical_loss": 3.7311621376892474,
      "tokens_seen": 795672576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832898696088265,
      "loss": 2.9645,
      "theoretical_loss": 3.7311316447173573,
      "tokens_seen": 795738112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832798395185557,
      "loss": 2.8367,
      "theoretical_loss": 3.731101154959842,
      "tokens_seen": 795803648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038326980942828486,
      "loss": 2.8634,
      "theoretical_loss": 3.7310706684160992,
      "tokens_seen": 795869184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038325977933801404,
      "loss": 2.6604,
      "theoretical_loss": 3.731040185085525,
      "tokens_seen": 795934720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832497492477433,
      "loss": 2.6438,
      "theoretical_loss": 3.731009704967516,
      "tokens_seen": 796000256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832397191574724,
      "loss": 2.8332,
      "theoretical_loss": 3.7309792280614698,
      "tokens_seen": 796065792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038322968906720164,
      "loss": 2.6339,
      "theoretical_loss": 3.730948754366783,
      "tokens_seen": 796131328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038321965897693077,
      "loss": 2.6891,
      "theoretical_loss": 3.730918283882853,
      "tokens_seen": 796196864
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 447326,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5912489891052246,
      "objective/train/theoretical_loss": 3.730887816609077,
      "objective/train/tokens_used": 816722400,
      "theoretical_loss": 3.730887816609077,
      "tokens_seen": 796262400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038320962888666,
      "loss": 2.8906,
      "theoretical_loss": 3.730887816609077,
      "tokens_seen": 796262400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831995987963892,
      "loss": 2.6539,
      "theoretical_loss": 3.730857352544853,
      "tokens_seen": 796327936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038318956870611836,
      "loss": 2.6463,
      "theoretical_loss": 3.7308268916895795,
      "tokens_seen": 796393472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038317953861584755,
      "loss": 3.0121,
      "theoretical_loss": 3.730796434042653,
      "tokens_seen": 796459008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831695085255767,
      "loss": 2.5218,
      "theoretical_loss": 3.730765979603473,
      "tokens_seen": 796524544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831594784353059,
      "loss": 2.6623,
      "theoretical_loss": 3.7307355283714365,
      "tokens_seen": 796590080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038314944834503514,
      "loss": 2.9648,
      "theoretical_loss": 3.7307050803459427,
      "tokens_seen": 796655616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038313941825476427,
      "loss": 2.9332,
      "theoretical_loss": 3.73067463552639,
      "tokens_seen": 796721152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831293881644935,
      "loss": 2.6942,
      "theoretical_loss": 3.7306441939121773,
      "tokens_seen": 796786688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038311935807422263,
      "loss": 2.7156,
      "theoretical_loss": 3.7306137555027035,
      "tokens_seen": 796852224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038310932798395187,
      "loss": 2.6772,
      "theoretical_loss": 3.7305833202973675,
      "tokens_seen": 796917760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038309929789368105,
      "loss": 2.8115,
      "theoretical_loss": 3.730552888295569,
      "tokens_seen": 796983296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038308926780341023,
      "loss": 2.8397,
      "theoretical_loss": 3.730522459496707,
      "tokens_seen": 797048832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830792377131394,
      "loss": 2.7367,
      "theoretical_loss": 3.7304920339001812,
      "tokens_seen": 797114368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038306920762286865,
      "loss": 2.7863,
      "theoretical_loss": 3.7304616115053917,
      "tokens_seen": 797179904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830591775325978,
      "loss": 2.8075,
      "theoretical_loss": 3.7304311923117375,
      "tokens_seen": 797245440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383049147442327,
      "loss": 2.631,
      "theoretical_loss": 3.7304007763186195,
      "tokens_seen": 797310976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038303911735205614,
      "loss": 2.5869,
      "theoretical_loss": 3.730370363525438,
      "tokens_seen": 797376512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038302908726178537,
      "loss": 2.8324,
      "theoretical_loss": 3.7303399539315927,
      "tokens_seen": 797442048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038301905717151455,
      "loss": 2.586,
      "theoretical_loss": 3.7303095475364842,
      "tokens_seen": 797507584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038300902708124373,
      "loss": 2.7626,
      "theoretical_loss": 3.730279144339514,
      "tokens_seen": 797573120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829989969909729,
      "loss": 2.8478,
      "theoretical_loss": 3.7302487443400825,
      "tokens_seen": 797638656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829889669007021,
      "loss": 2.798,
      "theoretical_loss": 3.7302183475375905,
      "tokens_seen": 797704192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829789368104313,
      "loss": 2.6434,
      "theoretical_loss": 3.7301879539314395,
      "tokens_seen": 797769728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829689067201605,
      "loss": 2.6643,
      "theoretical_loss": 3.730157563521031,
      "tokens_seen": 797835264
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 447858,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1472742557525635,
      "objective/train/theoretical_loss": 3.730127176305765,
      "objective/train/tokens_used": 818360800,
      "theoretical_loss": 3.730127176305765,
      "tokens_seen": 797900800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038295887662988964,
      "loss": 2.845,
      "theoretical_loss": 3.730127176305765,
      "tokens_seen": 797900800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829488465396189,
      "loss": 2.7808,
      "theoretical_loss": 3.730096792285046,
      "tokens_seen": 797966336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382938816449348,
      "loss": 2.6188,
      "theoretical_loss": 3.7300664114582736,
      "tokens_seen": 798031872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038292878635907724,
      "loss": 2.8079,
      "theoretical_loss": 3.7300360338248506,
      "tokens_seen": 798097408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829187562688064,
      "loss": 2.9146,
      "theoretical_loss": 3.7300056593841795,
      "tokens_seen": 798162944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829087261785356,
      "loss": 2.9053,
      "theoretical_loss": 3.7299752881356616,
      "tokens_seen": 798228480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828986960882648,
      "loss": 2.5042,
      "theoretical_loss": 3.7299449200787,
      "tokens_seen": 798294016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382888665997994,
      "loss": 2.6706,
      "theoretical_loss": 3.729914555212698,
      "tokens_seen": 798359552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828786359077232,
      "loss": 2.3704,
      "theoretical_loss": 3.7298841935370572,
      "tokens_seen": 798425088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828686058174524,
      "loss": 2.6598,
      "theoretical_loss": 3.729853835051181,
      "tokens_seen": 798490624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038285857572718156,
      "loss": 2.8312,
      "theoretical_loss": 3.729823479754473,
      "tokens_seen": 798556160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038284854563691074,
      "loss": 2.8285,
      "theoretical_loss": 3.729793127646336,
      "tokens_seen": 798621696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038283851554664,
      "loss": 2.6438,
      "theoretical_loss": 3.7297627787261733,
      "tokens_seen": 798687232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828284854563691,
      "loss": 2.6833,
      "theoretical_loss": 3.729732432993389,
      "tokens_seen": 798752768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038281845536609834,
      "loss": 2.7391,
      "theoretical_loss": 3.729702090447387,
      "tokens_seen": 798818304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038280842527582746,
      "loss": 2.4969,
      "theoretical_loss": 3.7296717510875705,
      "tokens_seen": 798883840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827983951855567,
      "loss": 2.7266,
      "theoretical_loss": 3.7296414149133437,
      "tokens_seen": 798949376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827883650952859,
      "loss": 2.7095,
      "theoretical_loss": 3.7296110819241117,
      "tokens_seen": 799014912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038277833500501506,
      "loss": 2.6529,
      "theoretical_loss": 3.729580752119278,
      "tokens_seen": 799080448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038276830491474424,
      "loss": 2.774,
      "theoretical_loss": 3.729550425498248,
      "tokens_seen": 799145984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827582748244735,
      "loss": 2.671,
      "theoretical_loss": 3.729520102060425,
      "tokens_seen": 799211520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827482447342026,
      "loss": 2.6735,
      "theoretical_loss": 3.7294897818052157,
      "tokens_seen": 799277056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038273821464393184,
      "loss": 2.6403,
      "theoretical_loss": 3.729459464732024,
      "tokens_seen": 799342592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038272818455366097,
      "loss": 2.7248,
      "theoretical_loss": 3.7294291508402555,
      "tokens_seen": 799408128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827181544633902,
      "loss": 2.765,
      "theoretical_loss": 3.729398840129315,
      "tokens_seen": 799473664
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 448758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5953433513641357,
      "objective/train/theoretical_loss": 3.729368532598609,
      "objective/train/tokens_used": 819999200,
      "theoretical_loss": 3.729368532598609,
      "tokens_seen": 799539200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827081243731194,
      "loss": 2.7885,
      "theoretical_loss": 3.729368532598609,
      "tokens_seen": 799539200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038269809428284856,
      "loss": 2.8063,
      "theoretical_loss": 3.729338228247543,
      "tokens_seen": 799604736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038268806419257775,
      "loss": 2.7712,
      "theoretical_loss": 3.7293079270755216,
      "tokens_seen": 799670272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826780341023069,
      "loss": 2.814,
      "theoretical_loss": 3.7292776290819525,
      "tokens_seen": 799735808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826680040120361,
      "loss": 2.8649,
      "theoretical_loss": 3.729247334266241,
      "tokens_seen": 799801344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038265797392176534,
      "loss": 2.8075,
      "theoretical_loss": 3.7292170426277935,
      "tokens_seen": 799866880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038264794383149447,
      "loss": 2.6499,
      "theoretical_loss": 3.7291867541660166,
      "tokens_seen": 799932416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826379137412237,
      "loss": 2.7997,
      "theoretical_loss": 3.729156468880317,
      "tokens_seen": 799997952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038262788365095283,
      "loss": 2.6938,
      "theoretical_loss": 3.729126186770101,
      "tokens_seen": 800063488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038261785356068207,
      "loss": 2.9072,
      "theoretical_loss": 3.7290959078347763,
      "tokens_seen": 800129024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038260782347041125,
      "loss": 2.6993,
      "theoretical_loss": 3.72906563207375,
      "tokens_seen": 800194560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038259779338014043,
      "loss": 2.5006,
      "theoretical_loss": 3.7290353594864287,
      "tokens_seen": 800260096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825877632898696,
      "loss": 3.0636,
      "theoretical_loss": 3.7290050900722203,
      "tokens_seen": 800325632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038257773319959885,
      "loss": 2.8181,
      "theoretical_loss": 3.7289748238305327,
      "tokens_seen": 800391168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382567703109328,
      "loss": 2.6539,
      "theoretical_loss": 3.7289445607607727,
      "tokens_seen": 800456704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825576730190572,
      "loss": 2.9976,
      "theoretical_loss": 3.728914300862349,
      "tokens_seen": 800522240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038254764292878634,
      "loss": 2.7945,
      "theoretical_loss": 3.72888404413467,
      "tokens_seen": 800587776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038253761283851557,
      "loss": 2.7775,
      "theoretical_loss": 3.7288537905771433,
      "tokens_seen": 800653312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038252758274824475,
      "loss": 2.5832,
      "theoretical_loss": 3.728823540189177,
      "tokens_seen": 800718848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038251755265797393,
      "loss": 2.8424,
      "theoretical_loss": 3.728793292970181,
      "tokens_seen": 800784384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825075225677031,
      "loss": 3.0617,
      "theoretical_loss": 3.7287630489195625,
      "tokens_seen": 800849920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824974924774323,
      "loss": 2.771,
      "theoretical_loss": 3.728732808036731,
      "tokens_seen": 800915456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824874623871615,
      "loss": 2.8826,
      "theoretical_loss": 3.7287025703210954,
      "tokens_seen": 800980992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824774322968907,
      "loss": 2.9558,
      "theoretical_loss": 3.7286723357720657,
      "tokens_seen": 801046528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038246740220661984,
      "loss": 2.7418,
      "theoretical_loss": 3.72864210438905,
      "tokens_seen": 801112064
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 449233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.920238733291626,
      "objective/train/theoretical_loss": 3.728611876171459,
      "objective/train/tokens_used": 821637600,
      "theoretical_loss": 3.728611876171459,
      "tokens_seen": 801177600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824573721163491,
      "loss": 2.7952,
      "theoretical_loss": 3.728611876171459,
      "tokens_seen": 801177600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824473420260782,
      "loss": 2.6668,
      "theoretical_loss": 3.7285816511187013,
      "tokens_seen": 801243136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038243731193580744,
      "loss": 2.7591,
      "theoretical_loss": 3.7285514292301873,
      "tokens_seen": 801308672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824272818455366,
      "loss": 2.7631,
      "theoretical_loss": 3.728521210505327,
      "tokens_seen": 801374208
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824172517552658,
      "loss": 2.7571,
      "theoretical_loss": 3.7284909949435305,
      "tokens_seen": 801439744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382407221664995,
      "loss": 2.655,
      "theoretical_loss": 3.728460782544208,
      "tokens_seen": 801505280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823971915747242,
      "loss": 2.884,
      "theoretical_loss": 3.7284305733067704,
      "tokens_seen": 801570816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038238716148445334,
      "loss": 2.7354,
      "theoretical_loss": 3.728400367230627,
      "tokens_seen": 801636352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823771313941826,
      "loss": 2.7962,
      "theoretical_loss": 3.7283701643151903,
      "tokens_seen": 801701888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823671013039117,
      "loss": 2.7577,
      "theoretical_loss": 3.7283399645598703,
      "tokens_seen": 801767424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038235707121364094,
      "loss": 2.929,
      "theoretical_loss": 3.728309767964078,
      "tokens_seen": 801832960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823470411233701,
      "loss": 2.6252,
      "theoretical_loss": 3.728279574527225,
      "tokens_seen": 801898496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823370110330993,
      "loss": 2.6991,
      "theoretical_loss": 3.7282493842487225,
      "tokens_seen": 801964032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823269809428285,
      "loss": 2.6825,
      "theoretical_loss": 3.7282191971279826,
      "tokens_seen": 802029568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038231695085255766,
      "loss": 2.6514,
      "theoretical_loss": 3.7281890131644158,
      "tokens_seen": 802095104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038230692076228685,
      "loss": 2.7606,
      "theoretical_loss": 3.728158832357435,
      "tokens_seen": 802160640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822968906720161,
      "loss": 2.7181,
      "theoretical_loss": 3.7281286547064525,
      "tokens_seen": 802226176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822868605817452,
      "loss": 2.7173,
      "theoretical_loss": 3.7280984802108796,
      "tokens_seen": 802291712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038227683049147444,
      "loss": 2.7012,
      "theoretical_loss": 3.7280683088701294,
      "tokens_seen": 802357248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038226680040120357,
      "loss": 2.6665,
      "theoretical_loss": 3.728038140683614,
      "tokens_seen": 802422784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822567703109328,
      "loss": 2.7573,
      "theoretical_loss": 3.728007975650746,
      "tokens_seen": 802488320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382246740220662,
      "loss": 2.8618,
      "theoretical_loss": 3.727977813770938,
      "tokens_seen": 802553856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038223671013039117,
      "loss": 2.6473,
      "theoretical_loss": 3.7279476550436037,
      "tokens_seen": 802619392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038222668004012035,
      "loss": 2.6769,
      "theoretical_loss": 3.727917499468156,
      "tokens_seen": 802684928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822166499498496,
      "loss": 2.8372,
      "theoretical_loss": 3.7278873470440077,
      "tokens_seen": 802750464
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 450645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0350584983825684,
      "objective/train/theoretical_loss": 3.7278571977705734,
      "objective/train/tokens_used": 823276000,
      "theoretical_loss": 3.7278571977705734,
      "tokens_seen": 802816000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822066198595787,
      "loss": 2.6396,
      "theoretical_loss": 3.7278571977705734,
      "tokens_seen": 802816000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038219658976930795,
      "loss": 2.7545,
      "theoretical_loss": 3.727827051647265,
      "tokens_seen": 802881536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821865596790371,
      "loss": 2.7913,
      "theoretical_loss": 3.727796908673498,
      "tokens_seen": 802947072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821765295887663,
      "loss": 2.7719,
      "theoretical_loss": 3.727766768848685,
      "tokens_seen": 803012608
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821664994984955,
      "loss": 2.6693,
      "theoretical_loss": 3.727736632172241,
      "tokens_seen": 803078144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038215646940822467,
      "loss": 2.6396,
      "theoretical_loss": 3.7277064986435793,
      "tokens_seen": 803143680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038214643931795385,
      "loss": 2.7354,
      "theoretical_loss": 3.7276763682621157,
      "tokens_seen": 803209216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038213640922768303,
      "loss": 2.6416,
      "theoretical_loss": 3.7276462410272635,
      "tokens_seen": 803274752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038212637913741227,
      "loss": 2.9243,
      "theoretical_loss": 3.727616116938438,
      "tokens_seen": 803340288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038211634904714145,
      "loss": 2.6483,
      "theoretical_loss": 3.7275859959950535,
      "tokens_seen": 803405824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038210631895687063,
      "loss": 2.7773,
      "theoretical_loss": 3.727555878196526,
      "tokens_seen": 803471360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820962888665998,
      "loss": 2.7851,
      "theoretical_loss": 3.7275257635422694,
      "tokens_seen": 803536896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038208625877632905,
      "loss": 2.5762,
      "theoretical_loss": 3.7274956520317,
      "tokens_seen": 803602432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820762286860582,
      "loss": 2.8452,
      "theoretical_loss": 3.727465543664233,
      "tokens_seen": 803667968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820661985957874,
      "loss": 2.7465,
      "theoretical_loss": 3.7274354384392847,
      "tokens_seen": 803733504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038205616850551654,
      "loss": 2.6667,
      "theoretical_loss": 3.7274053363562696,
      "tokens_seen": 803799040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038204613841524577,
      "loss": 2.8064,
      "theoretical_loss": 3.7273752374146043,
      "tokens_seen": 803864576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038203610832497495,
      "loss": 2.7668,
      "theoretical_loss": 3.7273451416137053,
      "tokens_seen": 803930112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038202607823470413,
      "loss": 2.7471,
      "theoretical_loss": 3.7273150489529883,
      "tokens_seen": 803995648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820160481444333,
      "loss": 2.863,
      "theoretical_loss": 3.72728495943187,
      "tokens_seen": 804061184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820060180541625,
      "loss": 2.7216,
      "theoretical_loss": 3.727254873049767,
      "tokens_seen": 804126720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819959879638917,
      "loss": 2.6481,
      "theoretical_loss": 3.7272247898060966,
      "tokens_seen": 804192256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819859578736209,
      "loss": 2.6998,
      "theoretical_loss": 3.7271947097002744,
      "tokens_seen": 804257792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038197592778335004,
      "loss": 2.9825,
      "theoretical_loss": 3.727164632731718,
      "tokens_seen": 804323328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819658976930793,
      "loss": 3.0266,
      "theoretical_loss": 3.727134558899845,
      "tokens_seen": 804388864
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 451555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6607892513275146,
      "objective/train/theoretical_loss": 3.7271044882040725,
      "objective/train/tokens_used": 824914400,
      "theoretical_loss": 3.7271044882040725,
      "tokens_seen": 804454400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819558676028084,
      "loss": 2.6691,
      "theoretical_loss": 3.7271044882040725,
      "tokens_seen": 804454400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038194583751253764,
      "loss": 2.9954,
      "theoretical_loss": 3.7270744206438184,
      "tokens_seen": 804519936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819358074222668,
      "loss": 3.0286,
      "theoretical_loss": 3.7270443562184994,
      "tokens_seen": 804585472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381925777331996,
      "loss": 2.8963,
      "theoretical_loss": 3.7270142949275344,
      "tokens_seen": 804651008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819157472417252,
      "loss": 2.7464,
      "theoretical_loss": 3.726984236770341,
      "tokens_seen": 804716544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003819057171514544,
      "loss": 2.6961,
      "theoretical_loss": 3.7269541817463363,
      "tokens_seen": 804782080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038189568706118354,
      "loss": 2.8959,
      "theoretical_loss": 3.7269241298549405,
      "tokens_seen": 804847616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818856569709128,
      "loss": 2.8578,
      "theoretical_loss": 3.726894081095571,
      "tokens_seen": 804913152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818756268806419,
      "loss": 2.7054,
      "theoretical_loss": 3.7268640354676466,
      "tokens_seen": 804978688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038186559679037114,
      "loss": 2.7952,
      "theoretical_loss": 3.726833992970586,
      "tokens_seen": 805044224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818555667001003,
      "loss": 2.8643,
      "theoretical_loss": 3.726803953603808,
      "tokens_seen": 805109760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818455366098295,
      "loss": 2.9446,
      "theoretical_loss": 3.7267739173667316,
      "tokens_seen": 805175296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818355065195587,
      "loss": 3.0333,
      "theoretical_loss": 3.7267438842587763,
      "tokens_seen": 805240832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038182547642928786,
      "loss": 2.7721,
      "theoretical_loss": 3.726713854279361,
      "tokens_seen": 805306368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038181544633901705,
      "loss": 2.7606,
      "theoretical_loss": 3.7266838274279066,
      "tokens_seen": 805371904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818054162487463,
      "loss": 2.8632,
      "theoretical_loss": 3.7266538037038313,
      "tokens_seen": 805437440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817953861584754,
      "loss": 2.8549,
      "theoretical_loss": 3.7266237831065556,
      "tokens_seen": 805502976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038178535606820464,
      "loss": 2.9827,
      "theoretical_loss": 3.7265937656354993,
      "tokens_seen": 805568512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038177532597793377,
      "loss": 2.5919,
      "theoretical_loss": 3.7265637512900827,
      "tokens_seen": 805634048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381765295887663,
      "loss": 2.5312,
      "theoretical_loss": 3.7265337400697263,
      "tokens_seen": 805699584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817552657973922,
      "loss": 2.6245,
      "theoretical_loss": 3.7265037319738497,
      "tokens_seen": 805765120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038174523570712137,
      "loss": 2.6506,
      "theoretical_loss": 3.7264737270018746,
      "tokens_seen": 805830656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038173520561685055,
      "loss": 2.7696,
      "theoretical_loss": 3.7264437251532216,
      "tokens_seen": 805896192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817251755265798,
      "loss": 2.8109,
      "theoretical_loss": 3.7264137264273107,
      "tokens_seen": 805961728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817151454363089,
      "loss": 2.7733,
      "theoretical_loss": 3.726383730823564,
      "tokens_seen": 806027264
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 452218,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7736008167266846,
      "objective/train/theoretical_loss": 3.7263537383414023,
      "objective/train/tokens_used": 826552800,
      "theoretical_loss": 3.7263537383414023,
      "tokens_seen": 806092800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038170511534603815,
      "loss": 2.6347,
      "theoretical_loss": 3.7263537383414023,
      "tokens_seen": 806092800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816950852557673,
      "loss": 2.6228,
      "theoretical_loss": 3.7263237489802474,
      "tokens_seen": 806158336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816850551654965,
      "loss": 2.7033,
      "theoretical_loss": 3.72629376273952,
      "tokens_seen": 806223872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816750250752257,
      "loss": 2.6791,
      "theoretical_loss": 3.726263779618643,
      "tokens_seen": 806289408
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038166499498495487,
      "loss": 2.7004,
      "theoretical_loss": 3.726233799617037,
      "tokens_seen": 806354944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038165496489468405,
      "loss": 2.6775,
      "theoretical_loss": 3.7262038227341248,
      "tokens_seen": 806420480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038164493480441323,
      "loss": 2.7478,
      "theoretical_loss": 3.726173848969329,
      "tokens_seen": 806486016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816349047141424,
      "loss": 2.7054,
      "theoretical_loss": 3.72614387832207,
      "tokens_seen": 806551552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038162487462387165,
      "loss": 2.9783,
      "theoretical_loss": 3.7261139107917725,
      "tokens_seen": 806617088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816148445336008,
      "loss": 2.7914,
      "theoretical_loss": 3.726083946377858,
      "tokens_seen": 806682624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038160481444333,
      "loss": 2.67,
      "theoretical_loss": 3.726053985079749,
      "tokens_seen": 806748160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038159478435305914,
      "loss": 2.6837,
      "theoretical_loss": 3.7260240268968694,
      "tokens_seen": 806813696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815847542627884,
      "loss": 2.6676,
      "theoretical_loss": 3.7259940718286417,
      "tokens_seen": 806879232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038157472417251756,
      "loss": 2.7729,
      "theoretical_loss": 3.7259641198744893,
      "tokens_seen": 806944768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038156469408224674,
      "loss": 2.7676,
      "theoretical_loss": 3.7259341710338356,
      "tokens_seen": 807010304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815546639919759,
      "loss": 2.7686,
      "theoretical_loss": 3.7259042253061043,
      "tokens_seen": 807075840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038154463390170515,
      "loss": 2.8106,
      "theoretical_loss": 3.725874282690718,
      "tokens_seen": 807141376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815346038114343,
      "loss": 2.7067,
      "theoretical_loss": 3.7258443431871022,
      "tokens_seen": 807206912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815245737211635,
      "loss": 2.9003,
      "theoretical_loss": 3.7258144067946803,
      "tokens_seen": 807272448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038151454363089264,
      "loss": 2.8686,
      "theoretical_loss": 3.725784473512876,
      "tokens_seen": 807337984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815045135406219,
      "loss": 2.6977,
      "theoretical_loss": 3.7257545433411137,
      "tokens_seen": 807403520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038149448345035106,
      "loss": 2.5566,
      "theoretical_loss": 3.7257246162788187,
      "tokens_seen": 807469056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038148445336008024,
      "loss": 2.6849,
      "theoretical_loss": 3.7256946923254146,
      "tokens_seen": 807534592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814744232698094,
      "loss": 2.8655,
      "theoretical_loss": 3.725664771480327,
      "tokens_seen": 807600128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814643931795386,
      "loss": 2.8026,
      "theoretical_loss": 3.7256348537429798,
      "tokens_seen": 807665664
    },
    {
      "epoch": 0.24,
      "objective/train/docs_used": 453465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5026791095733643,
      "objective/train/theoretical_loss": 3.7256049391127988,
      "objective/train/tokens_used": 828191200,
      "theoretical_loss": 3.7256049391127988,
      "tokens_seen": 807731200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814543630892678,
      "loss": 2.7585,
      "theoretical_loss": 3.7256049391127988,
      "tokens_seen": 807731200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000381444332998997,
      "loss": 2.5317,
      "theoretical_loss": 3.7255750275892097,
      "tokens_seen": 807796736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038143430290872614,
      "loss": 2.5894,
      "theoretical_loss": 3.725545119171637,
      "tokens_seen": 807862272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814242728184554,
      "loss": 2.8322,
      "theoretical_loss": 3.7255152138595067,
      "tokens_seen": 807927808
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038141424272818456,
      "loss": 2.89,
      "theoretical_loss": 3.7254853116522444,
      "tokens_seen": 807993344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038140421263791374,
      "loss": 2.7496,
      "theoretical_loss": 3.7254554125492754,
      "tokens_seen": 808058880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813941825476429,
      "loss": 2.8516,
      "theoretical_loss": 3.7254255165500267,
      "tokens_seen": 808124416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813841524573721,
      "loss": 2.6526,
      "theoretical_loss": 3.725395623653924,
      "tokens_seen": 808189952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038137412236710134,
      "loss": 2.6249,
      "theoretical_loss": 3.7253657338603934,
      "tokens_seen": 808255488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813640922768305,
      "loss": 2.8297,
      "theoretical_loss": 3.7253358471688616,
      "tokens_seen": 808321024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813540621865597,
      "loss": 2.6693,
      "theoretical_loss": 3.725305963578755,
      "tokens_seen": 808386560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813440320962889,
      "loss": 2.7322,
      "theoretical_loss": 3.725276083089501,
      "tokens_seen": 808452096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038133400200601806,
      "loss": 2.7031,
      "theoretical_loss": 3.725246205700526,
      "tokens_seen": 808517632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038132397191574725,
      "loss": 2.6764,
      "theoretical_loss": 3.7252163314112567,
      "tokens_seen": 808583168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003813139418254765,
      "loss": 2.6776,
      "theoretical_loss": 3.7251864602211207,
      "tokens_seen": 808648704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003813039117352056,
      "loss": 2.4349,
      "theoretical_loss": 3.7251565921295455,
      "tokens_seen": 808714240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038129388164493484,
      "loss": 2.6752,
      "theoretical_loss": 3.725126727135959,
      "tokens_seen": 808779776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038128385155466397,
      "loss": 2.7746,
      "theoretical_loss": 3.7250968652397876,
      "tokens_seen": 808845312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812738214643932,
      "loss": 2.8142,
      "theoretical_loss": 3.7250670064404603,
      "tokens_seen": 808910848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812637913741224,
      "loss": 2.9367,
      "theoretical_loss": 3.725037150737405,
      "tokens_seen": 808976384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038125376128385157,
      "loss": 2.7664,
      "theoretical_loss": 3.725007298130049,
      "tokens_seen": 809041920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038124373119358075,
      "loss": 2.6014,
      "theoretical_loss": 3.724977448617821,
      "tokens_seen": 809107456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038123370110331,
      "loss": 2.67,
      "theoretical_loss": 3.72494760220015,
      "tokens_seen": 809172992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812236710130391,
      "loss": 2.9132,
      "theoretical_loss": 3.724917758876464,
      "tokens_seen": 809238528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038121364092276835,
      "loss": 2.8343,
      "theoretical_loss": 3.724887918646192,
      "tokens_seen": 809304064
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 454216,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.918653726577759,
      "objective/train/theoretical_loss": 3.7248580815087626,
      "objective/train/tokens_used": 829829600,
      "theoretical_loss": 3.7248580815087626,
      "tokens_seen": 809369600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812036108324975,
      "loss": 2.5661,
      "theoretical_loss": 3.7248580815087626,
      "tokens_seen": 809369600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811935807422267,
      "loss": 2.5607,
      "theoretical_loss": 3.724828247463605,
      "tokens_seen": 809435136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811835506519559,
      "loss": 2.6057,
      "theoretical_loss": 3.7247984165101484,
      "tokens_seen": 809500672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038117352056168507,
      "loss": 2.5209,
      "theoretical_loss": 3.724768588647822,
      "tokens_seen": 809566208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038116349047141425,
      "loss": 2.658,
      "theoretical_loss": 3.724738763876056,
      "tokens_seen": 809631744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038115346038114343,
      "loss": 2.7838,
      "theoretical_loss": 3.724708942194279,
      "tokens_seen": 809697280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811434302908726,
      "loss": 2.648,
      "theoretical_loss": 3.7246791236019217,
      "tokens_seen": 809762816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038113340020060185,
      "loss": 2.7635,
      "theoretical_loss": 3.7246493080984133,
      "tokens_seen": 809828352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000381123370110331,
      "loss": 2.8069,
      "theoretical_loss": 3.724619495683185,
      "tokens_seen": 809893888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811133400200602,
      "loss": 2.812,
      "theoretical_loss": 3.724589686355665,
      "tokens_seen": 809959424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038110330992978934,
      "loss": 2.706,
      "theoretical_loss": 3.7245598801152857,
      "tokens_seen": 810024960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810932798395186,
      "loss": 2.8072,
      "theoretical_loss": 3.7245300769614773,
      "tokens_seen": 810090496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038108324974924776,
      "loss": 2.7348,
      "theoretical_loss": 3.7245002768936697,
      "tokens_seen": 810156032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038107321965897694,
      "loss": 2.8901,
      "theoretical_loss": 3.724470479911295,
      "tokens_seen": 810221568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810631895687061,
      "loss": 3.0223,
      "theoretical_loss": 3.724440686013782,
      "tokens_seen": 810287104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038105315947843535,
      "loss": 2.6307,
      "theoretical_loss": 3.724410895200565,
      "tokens_seen": 810352640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810431293881645,
      "loss": 2.8834,
      "theoretical_loss": 3.724381107471072,
      "tokens_seen": 810418176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810330992978937,
      "loss": 2.4979,
      "theoretical_loss": 3.724351322824737,
      "tokens_seen": 810483712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038102306920762284,
      "loss": 2.7276,
      "theoretical_loss": 3.724321541260991,
      "tokens_seen": 810549248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810130391173521,
      "loss": 2.6296,
      "theoretical_loss": 3.7242917627792647,
      "tokens_seen": 810614784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038100300902708126,
      "loss": 2.4256,
      "theoretical_loss": 3.7242619873789913,
      "tokens_seen": 810680320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038099297893681044,
      "loss": 2.7176,
      "theoretical_loss": 3.724232215059603,
      "tokens_seen": 810745856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809829488465396,
      "loss": 2.7171,
      "theoretical_loss": 3.72420244582053,
      "tokens_seen": 810811392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809729187562688,
      "loss": 2.6742,
      "theoretical_loss": 3.7241726796612067,
      "tokens_seen": 810876928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380962888665998,
      "loss": 2.5437,
      "theoretical_loss": 3.724142916581065,
      "tokens_seen": 810942464
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 455558,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9195406436920166,
      "objective/train/theoretical_loss": 3.724113156579537,
      "objective/train/tokens_used": 831468000,
      "theoretical_loss": 3.724113156579537,
      "tokens_seen": 811008000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809528585757272,
      "loss": 2.731,
      "theoretical_loss": 3.724113156579537,
      "tokens_seen": 811008000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038094282848545635,
      "loss": 2.6566,
      "theoretical_loss": 3.7240833996560565,
      "tokens_seen": 811073536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809327983951856,
      "loss": 2.8637,
      "theoretical_loss": 3.7240536458100557,
      "tokens_seen": 811139072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038092276830491476,
      "loss": 2.8568,
      "theoretical_loss": 3.7240238950409683,
      "tokens_seen": 811204608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038091273821464394,
      "loss": 2.5492,
      "theoretical_loss": 3.723994147348227,
      "tokens_seen": 811270144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809027081243731,
      "loss": 2.6746,
      "theoretical_loss": 3.723964402731266,
      "tokens_seen": 811335680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808926780341023,
      "loss": 2.7153,
      "theoretical_loss": 3.723934661189518,
      "tokens_seen": 811401216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808826479438315,
      "loss": 2.8033,
      "theoretical_loss": 3.7239049227224172,
      "tokens_seen": 811466752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808726178535607,
      "loss": 2.917,
      "theoretical_loss": 3.723875187329398,
      "tokens_seen": 811532288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038086258776328985,
      "loss": 2.8844,
      "theoretical_loss": 3.7238454550098927,
      "tokens_seen": 811597824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808525576730191,
      "loss": 2.4595,
      "theoretical_loss": 3.7238157257633366,
      "tokens_seen": 811663360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808425275827482,
      "loss": 2.9098,
      "theoretical_loss": 3.7237859995891647,
      "tokens_seen": 811728896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038083249749247745,
      "loss": 2.7129,
      "theoretical_loss": 3.72375627648681,
      "tokens_seen": 811794432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808224674022066,
      "loss": 2.668,
      "theoretical_loss": 3.7237265564557083,
      "tokens_seen": 811859968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808124373119358,
      "loss": 2.6803,
      "theoretical_loss": 3.723696839495294,
      "tokens_seen": 811925504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380802407221665,
      "loss": 2.7145,
      "theoretical_loss": 3.723667125605002,
      "tokens_seen": 811991040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038079237713139417,
      "loss": 2.7764,
      "theoretical_loss": 3.723637414784267,
      "tokens_seen": 812056576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038078234704112335,
      "loss": 2.6353,
      "theoretical_loss": 3.723607707032525,
      "tokens_seen": 812122112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807723169508526,
      "loss": 2.8169,
      "theoretical_loss": 3.7235780023492104,
      "tokens_seen": 812187648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807622868605817,
      "loss": 2.9368,
      "theoretical_loss": 3.7235483007337593,
      "tokens_seen": 812253184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038075225677031095,
      "loss": 2.8544,
      "theoretical_loss": 3.7235186021856075,
      "tokens_seen": 812318720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038074222668004013,
      "loss": 2.5776,
      "theoretical_loss": 3.7234889067041905,
      "tokens_seen": 812384256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807321965897693,
      "loss": 2.8329,
      "theoretical_loss": 3.723459214288944,
      "tokens_seen": 812449792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807221664994985,
      "loss": 2.7954,
      "theoretical_loss": 3.723429524939305,
      "tokens_seen": 812515328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807121364092277,
      "loss": 2.5979,
      "theoretical_loss": 3.723399838654709,
      "tokens_seen": 812580864
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 456229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.552311897277832,
      "objective/train/theoretical_loss": 3.7233701554345924,
      "objective/train/tokens_used": 833106400,
      "theoretical_loss": 3.7233701554345924,
      "tokens_seen": 812646400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038070210631895685,
      "loss": 2.6729,
      "theoretical_loss": 3.7233701554345924,
      "tokens_seen": 812646400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806920762286861,
      "loss": 2.6202,
      "theoretical_loss": 3.723340475278392,
      "tokens_seen": 812711936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806820461384152,
      "loss": 2.6352,
      "theoretical_loss": 3.7233107981855453,
      "tokens_seen": 812777472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038067201604814445,
      "loss": 2.9071,
      "theoretical_loss": 3.7232811241554877,
      "tokens_seen": 812843008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806619859578736,
      "loss": 2.8939,
      "theoretical_loss": 3.7232514531876566,
      "tokens_seen": 812908544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806519558676028,
      "loss": 2.5899,
      "theoretical_loss": 3.72322178528149,
      "tokens_seen": 812974080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380641925777332,
      "loss": 2.7315,
      "theoretical_loss": 3.7231921204364244,
      "tokens_seen": 813039616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806318956870612,
      "loss": 2.8286,
      "theoretical_loss": 3.7231624586518977,
      "tokens_seen": 813105152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806218655967904,
      "loss": 2.5656,
      "theoretical_loss": 3.723132799927347,
      "tokens_seen": 813170688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038061183550651954,
      "loss": 2.7309,
      "theoretical_loss": 3.7231031442622107,
      "tokens_seen": 813236224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806018054162488,
      "loss": 2.7168,
      "theoretical_loss": 3.723073491655926,
      "tokens_seen": 813301760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038059177532597796,
      "loss": 2.8353,
      "theoretical_loss": 3.7230438421079315,
      "tokens_seen": 813367296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038058174523570714,
      "loss": 2.542,
      "theoretical_loss": 3.7230141956176652,
      "tokens_seen": 813432832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805717151454363,
      "loss": 2.7511,
      "theoretical_loss": 3.722984552184565,
      "tokens_seen": 813498368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038056168505516555,
      "loss": 2.5704,
      "theoretical_loss": 3.7229549118080705,
      "tokens_seen": 813563904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805516549648947,
      "loss": 2.5106,
      "theoretical_loss": 3.7229252744876193,
      "tokens_seen": 813629440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805416248746239,
      "loss": 2.8,
      "theoretical_loss": 3.7228956402226503,
      "tokens_seen": 813694976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038053159478435304,
      "loss": 2.5614,
      "theoretical_loss": 3.722866009012603,
      "tokens_seen": 813760512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805215646940823,
      "loss": 2.7477,
      "theoretical_loss": 3.7228363808569163,
      "tokens_seen": 813826048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038051153460381146,
      "loss": 2.7359,
      "theoretical_loss": 3.722806755755029,
      "tokens_seen": 813891584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038050150451354064,
      "loss": 2.8882,
      "theoretical_loss": 3.722777133706381,
      "tokens_seen": 813957120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804914744232698,
      "loss": 2.7194,
      "theoretical_loss": 3.7227475147104117,
      "tokens_seen": 814022656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380481444332999,
      "loss": 2.6075,
      "theoretical_loss": 3.7227178987665606,
      "tokens_seen": 814088192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804714142427282,
      "loss": 2.7222,
      "theoretical_loss": 3.722688285874267,
      "tokens_seen": 814153728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804613841524574,
      "loss": 2.6867,
      "theoretical_loss": 3.7226586760329727,
      "tokens_seen": 814219264
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 457261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5911662578582764,
      "objective/train/theoretical_loss": 3.7226290692421165,
      "objective/train/tokens_used": 834744800,
      "theoretical_loss": 3.7226290692421165,
      "tokens_seen": 814284800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038045135406218655,
      "loss": 2.6262,
      "theoretical_loss": 3.7226290692421165,
      "tokens_seen": 814284800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804413239719158,
      "loss": 2.5049,
      "theoretical_loss": 3.722599465501138,
      "tokens_seen": 814350336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038043129388164496,
      "loss": 2.7759,
      "theoretical_loss": 3.722569864809479,
      "tokens_seen": 814415872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038042126379137414,
      "loss": 2.6387,
      "theoretical_loss": 3.722540267166579,
      "tokens_seen": 814481408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804112337011033,
      "loss": 2.712,
      "theoretical_loss": 3.7225106725718797,
      "tokens_seen": 814546944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804012036108325,
      "loss": 2.7145,
      "theoretical_loss": 3.722481081024821,
      "tokens_seen": 814612480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803911735205617,
      "loss": 2.8807,
      "theoretical_loss": 3.7224514925248444,
      "tokens_seen": 814678016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803811434302909,
      "loss": 2.6498,
      "theoretical_loss": 3.7224219070713915,
      "tokens_seen": 814743552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038037111334002005,
      "loss": 2.8192,
      "theoretical_loss": 3.722392324663903,
      "tokens_seen": 814809088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803610832497493,
      "loss": 2.6205,
      "theoretical_loss": 3.7223627453018207,
      "tokens_seen": 814874624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803510531594784,
      "loss": 3.0639,
      "theoretical_loss": 3.7223331689845853,
      "tokens_seen": 814940160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038034102306920765,
      "loss": 2.5076,
      "theoretical_loss": 3.7223035957116393,
      "tokens_seen": 815005696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038033099297893683,
      "loss": 2.4231,
      "theoretical_loss": 3.722274025482425,
      "tokens_seen": 815071232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380320962888666,
      "loss": 2.5898,
      "theoretical_loss": 3.722244458296384,
      "tokens_seen": 815136768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803109327983952,
      "loss": 2.756,
      "theoretical_loss": 3.722214894152958,
      "tokens_seen": 815202304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038030090270812437,
      "loss": 2.8386,
      "theoretical_loss": 3.72218533305159,
      "tokens_seen": 815267840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038029087261785355,
      "loss": 2.6173,
      "theoretical_loss": 3.7221557749917222,
      "tokens_seen": 815333376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802808425275828,
      "loss": 2.808,
      "theoretical_loss": 3.722126219972797,
      "tokens_seen": 815398912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802708124373119,
      "loss": 2.7507,
      "theoretical_loss": 3.7220966679942578,
      "tokens_seen": 815464448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038026078234704115,
      "loss": 2.5634,
      "theoretical_loss": 3.722067119055547,
      "tokens_seen": 815529984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038025075225677033,
      "loss": 2.7806,
      "theoretical_loss": 3.7220375731561086,
      "tokens_seen": 815595520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802407221664995,
      "loss": 2.6871,
      "theoretical_loss": 3.722008030295384,
      "tokens_seen": 815661056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802306920762287,
      "loss": 2.5132,
      "theoretical_loss": 3.7219784904728184,
      "tokens_seen": 815726592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802206619859579,
      "loss": 2.883,
      "theoretical_loss": 3.7219489536878543,
      "tokens_seen": 815792128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038021063189568705,
      "loss": 2.8185,
      "theoretical_loss": 3.7219194199399355,
      "tokens_seen": 815857664
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 457884,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9456984996795654,
      "objective/train/theoretical_loss": 3.721889889228506,
      "objective/train/tokens_used": 836383200,
      "theoretical_loss": 3.721889889228506,
      "tokens_seen": 815923200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802006018054163,
      "loss": 2.5694,
      "theoretical_loss": 3.721889889228506,
      "tokens_seen": 815923200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801905717151454,
      "loss": 2.4299,
      "theoretical_loss": 3.7218603615530093,
      "tokens_seen": 815988736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038018054162487465,
      "loss": 2.7284,
      "theoretical_loss": 3.7218308369128903,
      "tokens_seen": 816054272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801705115346038,
      "loss": 2.5405,
      "theoretical_loss": 3.7218013153075926,
      "tokens_seen": 816119808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380160481444333,
      "loss": 2.6405,
      "theoretical_loss": 3.7217717967365607,
      "tokens_seen": 816185344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801504513540622,
      "loss": 2.5513,
      "theoretical_loss": 3.7217422811992393,
      "tokens_seen": 816250880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801404212637914,
      "loss": 2.761,
      "theoretical_loss": 3.721712768695073,
      "tokens_seen": 816316416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038013039117352056,
      "loss": 2.5803,
      "theoretical_loss": 3.7216832592235067,
      "tokens_seen": 816381952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038012036108324974,
      "loss": 2.6744,
      "theoretical_loss": 3.7216537527839852,
      "tokens_seen": 816447488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801103309929789,
      "loss": 2.6138,
      "theoretical_loss": 3.7216242493759535,
      "tokens_seen": 816513024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038010030090270816,
      "loss": 2.4472,
      "theoretical_loss": 3.7215947489988572,
      "tokens_seen": 816578560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800902708124373,
      "loss": 2.4756,
      "theoretical_loss": 3.721565251652142,
      "tokens_seen": 816644096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800802407221665,
      "loss": 2.5205,
      "theoretical_loss": 3.7215357573352525,
      "tokens_seen": 816709632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800702106318957,
      "loss": 2.7038,
      "theoretical_loss": 3.7215062660476352,
      "tokens_seen": 816775168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800601805416249,
      "loss": 2.499,
      "theoretical_loss": 3.7214767777887356,
      "tokens_seen": 816840704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038005015045135406,
      "loss": 2.6442,
      "theoretical_loss": 3.7214472925579996,
      "tokens_seen": 816906240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038004012036108324,
      "loss": 2.792,
      "theoretical_loss": 3.7214178103548736,
      "tokens_seen": 816971776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800300902708124,
      "loss": 2.8231,
      "theoretical_loss": 3.7213883311788036,
      "tokens_seen": 817037312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038002006018054166,
      "loss": 2.6325,
      "theoretical_loss": 3.721358855029237,
      "tokens_seen": 817102848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800100300902708,
      "loss": 2.4452,
      "theoretical_loss": 3.721329381905619,
      "tokens_seen": 817168384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038,
      "loss": 2.7884,
      "theoretical_loss": 3.721299911807397,
      "tokens_seen": 817233920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037998996990972915,
      "loss": 2.6552,
      "theoretical_loss": 3.721270444734018,
      "tokens_seen": 817299456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799799398194584,
      "loss": 2.8386,
      "theoretical_loss": 3.7212409806849287,
      "tokens_seen": 817364992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037996990972918756,
      "loss": 2.5461,
      "theoretical_loss": 3.7212115196595765,
      "tokens_seen": 817430528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037995987963891675,
      "loss": 2.6774,
      "theoretical_loss": 3.7211820616574087,
      "tokens_seen": 817496064
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 458869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.828730344772339,
      "objective/train/theoretical_loss": 3.7211526066778724,
      "objective/train/tokens_used": 838021600,
      "theoretical_loss": 3.7211526066778724,
      "tokens_seen": 817561600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799498495486459,
      "loss": 2.8248,
      "theoretical_loss": 3.7211526066778724,
      "tokens_seen": 817561600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037993981945837516,
      "loss": 2.6308,
      "theoretical_loss": 3.7211231547204155,
      "tokens_seen": 817627136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799297893681043,
      "loss": 2.7227,
      "theoretical_loss": 3.7210937057844857,
      "tokens_seen": 817692672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003799197592778335,
      "loss": 2.4619,
      "theoretical_loss": 3.7210642598695305,
      "tokens_seen": 817758208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037990972918756265,
      "loss": 2.5265,
      "theoretical_loss": 3.7210348169749987,
      "tokens_seen": 817823744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798996990972919,
      "loss": 2.7976,
      "theoretical_loss": 3.721005377100338,
      "tokens_seen": 817889280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037988966900702107,
      "loss": 2.5068,
      "theoretical_loss": 3.7209759402449967,
      "tokens_seen": 817954816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037987963891675025,
      "loss": 2.94,
      "theoretical_loss": 3.7209465064084233,
      "tokens_seen": 818020352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798696088264795,
      "loss": 2.5955,
      "theoretical_loss": 3.7209170755900667,
      "tokens_seen": 818085888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798595787362086,
      "loss": 2.8317,
      "theoretical_loss": 3.7208876477893758,
      "tokens_seen": 818151424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037984954864593785,
      "loss": 2.6954,
      "theoretical_loss": 3.7208582230057985,
      "tokens_seen": 818216960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037983951855566703,
      "loss": 2.8328,
      "theoretical_loss": 3.7208288012387847,
      "tokens_seen": 818282496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798294884653962,
      "loss": 2.5533,
      "theoretical_loss": 3.720799382487783,
      "tokens_seen": 818348032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798194583751254,
      "loss": 2.7027,
      "theoretical_loss": 3.7207699667522434,
      "tokens_seen": 818413568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037980942828485457,
      "loss": 2.4817,
      "theoretical_loss": 3.7207405540316154,
      "tokens_seen": 818479104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037979939819458375,
      "loss": 2.6257,
      "theoretical_loss": 3.7207111443253478,
      "tokens_seen": 818544640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379789368104313,
      "loss": 2.8737,
      "theoretical_loss": 3.7206817376328907,
      "tokens_seen": 818610176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797793380140421,
      "loss": 2.6199,
      "theoretical_loss": 3.7206523339536943,
      "tokens_seen": 818675712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037976930792377135,
      "loss": 2.5395,
      "theoretical_loss": 3.720622933287209,
      "tokens_seen": 818741248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037975927783350053,
      "loss": 2.6246,
      "theoretical_loss": 3.7205935356328848,
      "tokens_seen": 818806784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797492477432297,
      "loss": 2.797,
      "theoretical_loss": 3.720564140990171,
      "tokens_seen": 818872320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797392176529589,
      "loss": 2.7855,
      "theoretical_loss": 3.720534749358519,
      "tokens_seen": 818937856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797291875626881,
      "loss": 2.7592,
      "theoretical_loss": 3.7205053607373797,
      "tokens_seen": 819003392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037971915747241725,
      "loss": 2.515,
      "theoretical_loss": 3.720475975126203,
      "tokens_seen": 819068928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797091273821465,
      "loss": 2.4283,
      "theoretical_loss": 3.72044659252444,
      "tokens_seen": 819134464
    },
    {
      "debugging/Self-BLEU-5": 0.4343450632679149,
      "debugging/distinct-1-grams": 0.7759737985147822,
      "debugging/distinct-2-grams": 0.9682119066950529,
      "debugging/entropy-1-grams": 5.4098389736484584,
      "debugging/entropy-2-grams": 6.131371694974163,
      "debugging/length": 546.8333333333334,
      "debugging/num_segments": 6,
      "epoch": 0.25,
      "objective/train/docs_used": 459733,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.788059711456299,
      "objective/train/theoretical_loss": 3.720417212931543,
      "objective/train/tokens_used": 839660000,
      "theoretical_loss": 3.720417212931543,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796990972918756,
      "loss": 2.7732,
      "theoretical_loss": 3.720417212931543,
      "tokens_seen": 819200000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037968906720160485,
      "loss": 2.6896,
      "theoretical_loss": 3.720387836346962,
      "tokens_seen": 819265536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379679037111334,
      "loss": 2.5718,
      "theoretical_loss": 3.7203584627701485,
      "tokens_seen": 819331072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796690070210632,
      "loss": 2.7729,
      "theoretical_loss": 3.7203290922005543,
      "tokens_seen": 819396608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796589769307924,
      "loss": 2.7949,
      "theoretical_loss": 3.7202997246376306,
      "tokens_seen": 819462144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796489468405216,
      "loss": 2.7875,
      "theoretical_loss": 3.72027036008083,
      "tokens_seen": 819527680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037963891675025076,
      "loss": 2.5432,
      "theoretical_loss": 3.7202409985296034,
      "tokens_seen": 819593216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037962888665997994,
      "loss": 2.8756,
      "theoretical_loss": 3.7202116399834035,
      "tokens_seen": 819658752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796188565697091,
      "loss": 2.6533,
      "theoretical_loss": 3.720182284441682,
      "tokens_seen": 819724288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037960882647943836,
      "loss": 2.8107,
      "theoretical_loss": 3.7201529319038924,
      "tokens_seen": 819789824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795987963891675,
      "loss": 2.8196,
      "theoretical_loss": 3.720123582369486,
      "tokens_seen": 819855360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795887662988967,
      "loss": 2.8017,
      "theoretical_loss": 3.720094235837916,
      "tokens_seen": 819920896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795787362086259,
      "loss": 2.7922,
      "theoretical_loss": 3.7200648923086357,
      "tokens_seen": 819986432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795687061183551,
      "loss": 2.5626,
      "theoretical_loss": 3.7200355517810966,
      "tokens_seen": 820051968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037955867602808426,
      "loss": 2.4739,
      "theoretical_loss": 3.7200062142547528,
      "tokens_seen": 820117504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037954864593781344,
      "loss": 2.7152,
      "theoretical_loss": 3.7199768797290576,
      "tokens_seen": 820183040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795386158475426,
      "loss": 2.8596,
      "theoretical_loss": 3.7199475482034643,
      "tokens_seen": 820248576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037952858575727186,
      "loss": 2.817,
      "theoretical_loss": 3.7199182196774254,
      "tokens_seen": 820314112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379518555667001,
      "loss": 2.5966,
      "theoretical_loss": 3.719888894150396,
      "tokens_seen": 820379648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795085255767302,
      "loss": 2.8667,
      "theoretical_loss": 3.7198595716218295,
      "tokens_seen": 820445184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037949849548645935,
      "loss": 2.5299,
      "theoretical_loss": 3.719830252091179,
      "tokens_seen": 820510720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794884653961886,
      "loss": 2.4508,
      "theoretical_loss": 3.7198009355578994,
      "tokens_seen": 820576256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037947843530591776,
      "loss": 2.7271,
      "theoretical_loss": 3.7197716220214443,
      "tokens_seen": 820641792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037946840521564695,
      "loss": 2.6313,
      "theoretical_loss": 3.7197423114812693,
      "tokens_seen": 820707328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794583751253761,
      "loss": 2.8888,
      "theoretical_loss": 3.7197130039368274,
      "tokens_seen": 820772864
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 460372,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5753989219665527,
      "objective/train/theoretical_loss": 3.7196836993875744,
      "objective/train/tokens_used": 841298400,
      "theoretical_loss": 3.7196836993875744,
      "tokens_seen": 820838400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037944834503510536,
      "loss": 2.6251,
      "theoretical_loss": 3.7196836993875744,
      "tokens_seen": 820838400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794383149448345,
      "loss": 2.6649,
      "theoretical_loss": 3.719654397832964,
      "tokens_seen": 820903936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794282848545637,
      "loss": 2.6335,
      "theoretical_loss": 3.7196250992724527,
      "tokens_seen": 820969472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037941825476429285,
      "loss": 2.7872,
      "theoretical_loss": 3.719595803705494,
      "tokens_seen": 821035008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794082246740221,
      "loss": 2.7039,
      "theoretical_loss": 3.7195665111315437,
      "tokens_seen": 821100544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037939819458375127,
      "loss": 2.5931,
      "theoretical_loss": 3.7195372215500573,
      "tokens_seen": 821166080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037938816449348045,
      "loss": 2.4618,
      "theoretical_loss": 3.7195079349604905,
      "tokens_seen": 821231616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037937813440320963,
      "loss": 2.5012,
      "theoretical_loss": 3.719478651362299,
      "tokens_seen": 821297152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793681043129388,
      "loss": 2.6347,
      "theoretical_loss": 3.719449370754938,
      "tokens_seen": 821362688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379358074222668,
      "loss": 2.5941,
      "theoretical_loss": 3.7194200931378636,
      "tokens_seen": 821428224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037934804413239723,
      "loss": 2.6462,
      "theoretical_loss": 3.719390818510532,
      "tokens_seen": 821493760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037933801404212635,
      "loss": 2.7512,
      "theoretical_loss": 3.7193615468724,
      "tokens_seen": 821559296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793279839518556,
      "loss": 2.7186,
      "theoretical_loss": 3.7193322782229226,
      "tokens_seen": 821624832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793179538615847,
      "loss": 2.7558,
      "theoretical_loss": 3.719303012561557,
      "tokens_seen": 821690368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037930792377131395,
      "loss": 2.676,
      "theoretical_loss": 3.719273749887761,
      "tokens_seen": 821755904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037929789368104313,
      "loss": 2.5643,
      "theoretical_loss": 3.71924449020099,
      "tokens_seen": 821821440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792878635907723,
      "loss": 2.5705,
      "theoretical_loss": 3.7192152335007007,
      "tokens_seen": 821886976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792778335005015,
      "loss": 2.8495,
      "theoretical_loss": 3.719185979786351,
      "tokens_seen": 821952512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037926780341023073,
      "loss": 2.8069,
      "theoretical_loss": 3.719156729057398,
      "tokens_seen": 822018048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037925777331995986,
      "loss": 2.6097,
      "theoretical_loss": 3.719127481313299,
      "tokens_seen": 822083584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792477432296891,
      "loss": 2.5657,
      "theoretical_loss": 3.719098236553511,
      "tokens_seen": 822149120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792377131394182,
      "loss": 2.7598,
      "theoretical_loss": 3.7190689947774924,
      "tokens_seen": 822214656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037922768304914746,
      "loss": 2.8197,
      "theoretical_loss": 3.7190397559847006,
      "tokens_seen": 822280192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037921765295887664,
      "loss": 2.7267,
      "theoretical_loss": 3.7190105201745935,
      "tokens_seen": 822345728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792076228686058,
      "loss": 2.8076,
      "theoretical_loss": 3.7189812873466295,
      "tokens_seen": 822411264
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 461421,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8889474868774414,
      "objective/train/theoretical_loss": 3.7189520575002666,
      "objective/train/tokens_used": 842936800,
      "theoretical_loss": 3.7189520575002666,
      "tokens_seen": 822476800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379197592778335,
      "loss": 2.8099,
      "theoretical_loss": 3.7189520575002666,
      "tokens_seen": 822476800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791875626880642,
      "loss": 2.6691,
      "theoretical_loss": 3.7189228306349627,
      "tokens_seen": 822542336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037917753259779336,
      "loss": 2.6943,
      "theoretical_loss": 3.7188936067501768,
      "tokens_seen": 822607872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791675025075226,
      "loss": 2.8235,
      "theoretical_loss": 3.718864385845367,
      "tokens_seen": 822673408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791574724172517,
      "loss": 2.7914,
      "theoretical_loss": 3.7188351679199934,
      "tokens_seen": 822738944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037914744232698096,
      "loss": 2.7442,
      "theoretical_loss": 3.7188059529735136,
      "tokens_seen": 822804480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791374122367101,
      "loss": 2.7754,
      "theoretical_loss": 3.7187767410053874,
      "tokens_seen": 822870016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791273821464393,
      "loss": 2.5717,
      "theoretical_loss": 3.7187475320150734,
      "tokens_seen": 822935552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037911735205616856,
      "loss": 2.7173,
      "theoretical_loss": 3.7187183260020307,
      "tokens_seen": 823001088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791073219658977,
      "loss": 2.8607,
      "theoretical_loss": 3.7186891229657197,
      "tokens_seen": 823066624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790972918756269,
      "loss": 2.7677,
      "theoretical_loss": 3.7186599229056,
      "tokens_seen": 823132160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790872617853561,
      "loss": 2.8683,
      "theoretical_loss": 3.718630725821131,
      "tokens_seen": 823197696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790772316950853,
      "loss": 2.6781,
      "theoretical_loss": 3.718601531711772,
      "tokens_seen": 823263232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037906720160481446,
      "loss": 2.7821,
      "theoretical_loss": 3.718572340576984,
      "tokens_seen": 823328768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037905717151454364,
      "loss": 2.2804,
      "theoretical_loss": 3.7185431524162267,
      "tokens_seen": 823394304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790471414242728,
      "loss": 2.5982,
      "theoretical_loss": 3.7185139672289607,
      "tokens_seen": 823459840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037903711133400206,
      "loss": 2.7199,
      "theoretical_loss": 3.718484785014646,
      "tokens_seen": 823525376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790270812437312,
      "loss": 2.814,
      "theoretical_loss": 3.7184556057727436,
      "tokens_seen": 823590912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790170511534604,
      "loss": 2.4827,
      "theoretical_loss": 3.7184264295027143,
      "tokens_seen": 823656448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037900702106318955,
      "loss": 2.7661,
      "theoretical_loss": 3.718397256204019,
      "tokens_seen": 823721984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789969909729188,
      "loss": 2.8633,
      "theoretical_loss": 3.7183680858761186,
      "tokens_seen": 823787520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037898696088264796,
      "loss": 2.8532,
      "theoretical_loss": 3.718338918518474,
      "tokens_seen": 823853056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037897693079237715,
      "loss": 2.5904,
      "theoretical_loss": 3.718309754130547,
      "tokens_seen": 823918592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789669007021063,
      "loss": 2.6954,
      "theoretical_loss": 3.718280592711799,
      "tokens_seen": 823984128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037895687061183556,
      "loss": 2.9142,
      "theoretical_loss": 3.7182514342616906,
      "tokens_seen": 824049664
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 462012,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.056985855102539,
      "objective/train/theoretical_loss": 3.7182222787796846,
      "objective/train/tokens_used": 844575200,
      "theoretical_loss": 3.7182222787796846,
      "tokens_seen": 824115200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789468405215647,
      "loss": 2.7643,
      "theoretical_loss": 3.7182222787796846,
      "tokens_seen": 824115200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789368104312939,
      "loss": 2.7109,
      "theoretical_loss": 3.718193126265243,
      "tokens_seen": 824180736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037892678034102305,
      "loss": 2.6935,
      "theoretical_loss": 3.718163976717828,
      "tokens_seen": 824246272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789167502507523,
      "loss": 2.5794,
      "theoretical_loss": 3.7181348301369006,
      "tokens_seen": 824311808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037890672016048147,
      "loss": 2.6364,
      "theoretical_loss": 3.7181056865219233,
      "tokens_seen": 824377344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037889669007021065,
      "loss": 2.678,
      "theoretical_loss": 3.7180765458723597,
      "tokens_seen": 824442880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037888665997993983,
      "loss": 2.6142,
      "theoretical_loss": 3.718047408187671,
      "tokens_seen": 824508416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378876629889669,
      "loss": 2.6126,
      "theoretical_loss": 3.7180182734673206,
      "tokens_seen": 824573952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788665997993982,
      "loss": 2.3591,
      "theoretical_loss": 3.7179891417107718,
      "tokens_seen": 824639488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037885656970912743,
      "loss": 2.6444,
      "theoretical_loss": 3.717960012917487,
      "tokens_seen": 824705024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037884653961885655,
      "loss": 2.842,
      "theoretical_loss": 3.717930887086929,
      "tokens_seen": 824770560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788365095285858,
      "loss": 2.8211,
      "theoretical_loss": 3.7179017642185617,
      "tokens_seen": 824836096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788264794383149,
      "loss": 2.7662,
      "theoretical_loss": 3.717872644311848,
      "tokens_seen": 824901632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037881644934804415,
      "loss": 2.6669,
      "theoretical_loss": 3.7178435273662522,
      "tokens_seen": 824967168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037880641925777333,
      "loss": 2.953,
      "theoretical_loss": 3.7178144133812374,
      "tokens_seen": 825032704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787963891675025,
      "loss": 2.6334,
      "theoretical_loss": 3.7177853023562673,
      "tokens_seen": 825098240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787863590772317,
      "loss": 2.847,
      "theoretical_loss": 3.7177561942908066,
      "tokens_seen": 825163776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037877632898696093,
      "loss": 2.6949,
      "theoretical_loss": 3.717727089184319,
      "tokens_seen": 825229312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037876629889669006,
      "loss": 2.6026,
      "theoretical_loss": 3.717697987036269,
      "tokens_seen": 825294848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787562688064193,
      "loss": 2.9019,
      "theoretical_loss": 3.71766888784612,
      "tokens_seen": 825360384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787462387161484,
      "loss": 2.733,
      "theoretical_loss": 3.7176397916133377,
      "tokens_seen": 825425920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037873620862587766,
      "loss": 2.5797,
      "theoretical_loss": 3.7176106983373858,
      "tokens_seen": 825491456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037872617853560684,
      "loss": 2.6358,
      "theoretical_loss": 3.7175816080177304,
      "tokens_seen": 825556992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378716148445336,
      "loss": 2.93,
      "theoretical_loss": 3.717552520653835,
      "tokens_seen": 825622528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787061183550652,
      "loss": 2.7954,
      "theoretical_loss": 3.717523436245166,
      "tokens_seen": 825688064
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 463076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.083082914352417,
      "objective/train/theoretical_loss": 3.717494354791188,
      "objective/train/tokens_used": 846213600,
      "theoretical_loss": 3.717494354791188,
      "tokens_seen": 825753600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786960882647944,
      "loss": 3.0909,
      "theoretical_loss": 3.717494354791188,
      "tokens_seen": 825753600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037868605817452356,
      "loss": 3.0866,
      "theoretical_loss": 3.7174652762913656,
      "tokens_seen": 825819136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786760280842528,
      "loss": 2.7781,
      "theoretical_loss": 3.7174362007451656,
      "tokens_seen": 825884672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786659979939819,
      "loss": 2.5382,
      "theoretical_loss": 3.717407128152053,
      "tokens_seen": 825950208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037865596790371116,
      "loss": 2.5681,
      "theoretical_loss": 3.7173780585114944,
      "tokens_seen": 826015744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786459378134403,
      "loss": 2.6316,
      "theoretical_loss": 3.7173489918229548,
      "tokens_seen": 826081280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786359077231695,
      "loss": 2.6745,
      "theoretical_loss": 3.7173199280859,
      "tokens_seen": 826146816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786258776328987,
      "loss": 2.7971,
      "theoretical_loss": 3.717290867299797,
      "tokens_seen": 826212352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786158475426279,
      "loss": 2.6546,
      "theoretical_loss": 3.717261809464112,
      "tokens_seen": 826277888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037860581745235706,
      "loss": 2.5943,
      "theoretical_loss": 3.717232754578311,
      "tokens_seen": 826343424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785957873620863,
      "loss": 2.8337,
      "theoretical_loss": 3.717203702641861,
      "tokens_seen": 826408960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785857572718154,
      "loss": 2.8984,
      "theoretical_loss": 3.717174653654229,
      "tokens_seen": 826474496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037857572718154466,
      "loss": 2.8819,
      "theoretical_loss": 3.7171456076148814,
      "tokens_seen": 826540032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785656970912738,
      "loss": 2.6567,
      "theoretical_loss": 3.7171165645232858,
      "tokens_seen": 826605568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378555667001003,
      "loss": 2.8761,
      "theoretical_loss": 3.7170875243789085,
      "tokens_seen": 826671104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785456369107322,
      "loss": 2.8024,
      "theoretical_loss": 3.717058487181218,
      "tokens_seen": 826736640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785356068204614,
      "loss": 2.7838,
      "theoretical_loss": 3.7170294529296806,
      "tokens_seen": 826802176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037852557673019057,
      "loss": 2.7053,
      "theoretical_loss": 3.7170004216237644,
      "tokens_seen": 826867712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037851554663991975,
      "loss": 2.8404,
      "theoretical_loss": 3.716971393262937,
      "tokens_seen": 826933248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037850551654964893,
      "loss": 2.7194,
      "theoretical_loss": 3.7169423678466664,
      "tokens_seen": 826998784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037849548645937816,
      "loss": 2.8596,
      "theoretical_loss": 3.7169133453744205,
      "tokens_seen": 827064320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784854563691073,
      "loss": 2.8185,
      "theoretical_loss": 3.716884325845668,
      "tokens_seen": 827129856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784754262788365,
      "loss": 2.5357,
      "theoretical_loss": 3.716855309259877,
      "tokens_seen": 827195392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037846539618856565,
      "loss": 2.6472,
      "theoretical_loss": 3.7168262956165146,
      "tokens_seen": 827260928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784553660982949,
      "loss": 2.6843,
      "theoretical_loss": 3.7167972849150512,
      "tokens_seen": 827326464
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 464152,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.622303009033203,
      "objective/train/theoretical_loss": 3.716768277154954,
      "objective/train/tokens_used": 847852000,
      "theoretical_loss": 3.716768277154954,
      "tokens_seen": 827392000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037844533600802407,
      "loss": 2.6362,
      "theoretical_loss": 3.716768277154954,
      "tokens_seen": 827392000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037843530591775325,
      "loss": 2.7811,
      "theoretical_loss": 3.716739272335693,
      "tokens_seen": 827457536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037842527582748243,
      "loss": 2.7028,
      "theoretical_loss": 3.7167102704567365,
      "tokens_seen": 827523072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037841524573721167,
      "loss": 2.8274,
      "theoretical_loss": 3.7166812715175537,
      "tokens_seen": 827588608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784052156469408,
      "loss": 2.7487,
      "theoretical_loss": 3.716652275517614,
      "tokens_seen": 827654144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037839518555667003,
      "loss": 2.7647,
      "theoretical_loss": 3.716623282456387,
      "tokens_seen": 827719680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037838515546639916,
      "loss": 2.9639,
      "theoretical_loss": 3.7165942923333417,
      "tokens_seen": 827785216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783751253761284,
      "loss": 2.8212,
      "theoretical_loss": 3.716565305147948,
      "tokens_seen": 827850752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037836509528585763,
      "loss": 2.4917,
      "theoretical_loss": 3.7165363208996753,
      "tokens_seen": 827916288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037835506519558675,
      "loss": 2.9519,
      "theoretical_loss": 3.716507339587995,
      "tokens_seen": 827981824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378345035105316,
      "loss": 2.9141,
      "theoretical_loss": 3.716478361212375,
      "tokens_seen": 828047360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783350050150451,
      "loss": 2.7782,
      "theoretical_loss": 3.716449385772288,
      "tokens_seen": 828112896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037832497492477435,
      "loss": 2.6236,
      "theoretical_loss": 3.716420413267201,
      "tokens_seen": 828178432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037831494483450353,
      "loss": 2.834,
      "theoretical_loss": 3.7163914436965877,
      "tokens_seen": 828243968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783049147442327,
      "loss": 2.597,
      "theoretical_loss": 3.7163624770599175,
      "tokens_seen": 828309504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782948846539619,
      "loss": 2.7529,
      "theoretical_loss": 3.716333513356661,
      "tokens_seen": 828375040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037828485456369113,
      "loss": 2.3636,
      "theoretical_loss": 3.7163045525862897,
      "tokens_seen": 828440576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037827482447342026,
      "loss": 2.833,
      "theoretical_loss": 3.7162755947482733,
      "tokens_seen": 828506112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782647943831495,
      "loss": 2.7211,
      "theoretical_loss": 3.716246639842084,
      "tokens_seen": 828571648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782547642928786,
      "loss": 2.7162,
      "theoretical_loss": 3.7162176878671938,
      "tokens_seen": 828637184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037824473420260786,
      "loss": 2.7256,
      "theoretical_loss": 3.7161887388230728,
      "tokens_seen": 828702720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037823470411233704,
      "loss": 2.9198,
      "theoretical_loss": 3.7161597927091927,
      "tokens_seen": 828768256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782246740220662,
      "loss": 2.8702,
      "theoretical_loss": 3.716130849525026,
      "tokens_seen": 828833792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782146439317954,
      "loss": 2.7121,
      "theoretical_loss": 3.7161019092700442,
      "tokens_seen": 828899328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782046138415246,
      "loss": 2.5721,
      "theoretical_loss": 3.716072971943719,
      "tokens_seen": 828964864
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 464802,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.640942096710205,
      "objective/train/theoretical_loss": 3.716044037545523,
      "objective/train/tokens_used": 849490400,
      "theoretical_loss": 3.716044037545523,
      "tokens_seen": 829030400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037819458375125376,
      "loss": 2.6406,
      "theoretical_loss": 3.716044037545523,
      "tokens_seen": 829030400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378184553660983,
      "loss": 2.7612,
      "theoretical_loss": 3.716015106074928,
      "tokens_seen": 829095936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781745235707121,
      "loss": 2.5001,
      "theoretical_loss": 3.715986177531407,
      "tokens_seen": 829161472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037816449348044136,
      "loss": 2.8508,
      "theoretical_loss": 3.715957251914432,
      "tokens_seen": 829227008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781544633901705,
      "loss": 2.9323,
      "theoretical_loss": 3.715928329223476,
      "tokens_seen": 829292544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781444332998997,
      "loss": 2.7472,
      "theoretical_loss": 3.715899409458012,
      "tokens_seen": 829358080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781344032096289,
      "loss": 2.8221,
      "theoretical_loss": 3.7158704926175123,
      "tokens_seen": 829423616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781243731193581,
      "loss": 2.7649,
      "theoretical_loss": 3.7158415787014505,
      "tokens_seen": 829489152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037811434302908726,
      "loss": 2.7626,
      "theoretical_loss": 3.7158126677092995,
      "tokens_seen": 829554688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781043129388165,
      "loss": 2.777,
      "theoretical_loss": 3.7157837596405328,
      "tokens_seen": 829620224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780942828485456,
      "loss": 2.8599,
      "theoretical_loss": 3.7157548544946244,
      "tokens_seen": 829685760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037808425275827486,
      "loss": 2.7063,
      "theoretical_loss": 3.7157259522710473,
      "tokens_seen": 829751296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378074222668004,
      "loss": 2.7576,
      "theoretical_loss": 3.7156970529692757,
      "tokens_seen": 829816832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780641925777332,
      "loss": 2.6135,
      "theoretical_loss": 3.7156681565887824,
      "tokens_seen": 829882368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780541624874624,
      "loss": 2.6237,
      "theoretical_loss": 3.715639263129043,
      "tokens_seen": 829947904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780441323971916,
      "loss": 2.6227,
      "theoretical_loss": 3.7156103725895306,
      "tokens_seen": 830013440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037803410230692077,
      "loss": 2.5027,
      "theoretical_loss": 3.71558148496972,
      "tokens_seen": 830078976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037802407221664995,
      "loss": 2.6977,
      "theoretical_loss": 3.7155526002690857,
      "tokens_seen": 830144512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037801404212637913,
      "loss": 2.5946,
      "theoretical_loss": 3.715523718487102,
      "tokens_seen": 830210048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037800401203610836,
      "loss": 2.7186,
      "theoretical_loss": 3.7154948396232443,
      "tokens_seen": 830275584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003779939819458375,
      "loss": 2.7893,
      "theoretical_loss": 3.7154659636769862,
      "tokens_seen": 830341120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037798395185556673,
      "loss": 2.7367,
      "theoretical_loss": 3.7154370906478036,
      "tokens_seen": 830406656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037797392176529585,
      "loss": 2.6869,
      "theoretical_loss": 3.7154082205351715,
      "tokens_seen": 830472192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003779638916750251,
      "loss": 2.8705,
      "theoretical_loss": 3.715379353338565,
      "tokens_seen": 830537728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037795386158475427,
      "loss": 2.7076,
      "theoretical_loss": 3.7153504890574602,
      "tokens_seen": 830603264
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 465950,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4607279300689697,
      "objective/train/theoretical_loss": 3.7153216276913312,
      "objective/train/tokens_used": 851128800,
      "theoretical_loss": 3.7153216276913312,
      "tokens_seen": 830668800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037794383149448345,
      "loss": 2.4907,
      "theoretical_loss": 3.7153216276913312,
      "tokens_seen": 830668800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037793380140421263,
      "loss": 2.7686,
      "theoretical_loss": 3.7152927692396553,
      "tokens_seen": 830734336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037792377131394187,
      "loss": 2.9923,
      "theoretical_loss": 3.7152639137019072,
      "tokens_seen": 830799872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377913741223671,
      "loss": 2.7894,
      "theoretical_loss": 3.7152350610775633,
      "tokens_seen": 830865408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037790371113340023,
      "loss": 2.3556,
      "theoretical_loss": 3.7152062113660995,
      "tokens_seen": 830930944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037789368104312936,
      "loss": 2.7755,
      "theoretical_loss": 3.7151773645669923,
      "tokens_seen": 830996480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778836509528586,
      "loss": 2.7287,
      "theoretical_loss": 3.7151485206797177,
      "tokens_seen": 831062016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778736208625878,
      "loss": 2.8398,
      "theoretical_loss": 3.7151196797037525,
      "tokens_seen": 831127552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037786359077231695,
      "loss": 2.686,
      "theoretical_loss": 3.715090841638573,
      "tokens_seen": 831193088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037785356068204614,
      "loss": 2.6514,
      "theoretical_loss": 3.715062006483656,
      "tokens_seen": 831258624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778435305917753,
      "loss": 2.8311,
      "theoretical_loss": 3.7150331742384783,
      "tokens_seen": 831324160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778335005015045,
      "loss": 2.819,
      "theoretical_loss": 3.7150043449025176,
      "tokens_seen": 831389696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037782347041123373,
      "loss": 2.8208,
      "theoretical_loss": 3.71497551847525,
      "tokens_seen": 831455232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037781344032096286,
      "loss": 2.6147,
      "theoretical_loss": 3.7149466949561547,
      "tokens_seen": 831520768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778034102306921,
      "loss": 2.6827,
      "theoretical_loss": 3.7149178743447067,
      "tokens_seen": 831586304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777933801404213,
      "loss": 2.7768,
      "theoretical_loss": 3.714889056640385,
      "tokens_seen": 831651840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037778335005015046,
      "loss": 2.4861,
      "theoretical_loss": 3.7148602418426675,
      "tokens_seen": 831717376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037777331995987964,
      "loss": 2.6438,
      "theoretical_loss": 3.7148314299510314,
      "tokens_seen": 831782912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777632898696088,
      "loss": 2.8952,
      "theoretical_loss": 3.7148026209649543,
      "tokens_seen": 831848448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377753259779338,
      "loss": 2.7831,
      "theoretical_loss": 3.714773814883915,
      "tokens_seen": 831913984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037774322968906724,
      "loss": 2.751,
      "theoretical_loss": 3.714745011707392,
      "tokens_seen": 831979520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037773319959879636,
      "loss": 2.5419,
      "theoretical_loss": 3.714716211434863,
      "tokens_seen": 832045056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777231695085256,
      "loss": 2.954,
      "theoretical_loss": 3.714687414065807,
      "tokens_seen": 832110592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777131394182547,
      "loss": 2.6081,
      "theoretical_loss": 3.7146586195997022,
      "tokens_seen": 832176128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037770310932798396,
      "loss": 2.8164,
      "theoretical_loss": 3.7146298280360277,
      "tokens_seen": 832241664
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 466904,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.74540638923645,
      "objective/train/theoretical_loss": 3.714601039374263,
      "objective/train/tokens_used": 852767200,
      "theoretical_loss": 3.714601039374263,
      "tokens_seen": 832307200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037769307923771314,
      "loss": 2.7948,
      "theoretical_loss": 3.714601039374263,
      "tokens_seen": 832307200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776830491474423,
      "loss": 2.7842,
      "theoretical_loss": 3.714572253613886,
      "tokens_seen": 832372736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776730190571715,
      "loss": 2.7081,
      "theoretical_loss": 3.714543470754376,
      "tokens_seen": 832438272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776629889669007,
      "loss": 2.7968,
      "theoretical_loss": 3.714514690795213,
      "tokens_seen": 832503808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037765295887662987,
      "loss": 2.92,
      "theoretical_loss": 3.714485913735876,
      "tokens_seen": 832569344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776429287863591,
      "loss": 2.6305,
      "theoretical_loss": 3.7144571395758454,
      "tokens_seen": 832634880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037763289869608823,
      "loss": 2.8311,
      "theoretical_loss": 3.7144283683145995,
      "tokens_seen": 832700416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037762286860581746,
      "loss": 2.9982,
      "theoretical_loss": 3.7143995999516193,
      "tokens_seen": 832765952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776128385155467,
      "loss": 2.915,
      "theoretical_loss": 3.7143708344863846,
      "tokens_seen": 832831488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776028084252758,
      "loss": 2.6645,
      "theoretical_loss": 3.7143420719183755,
      "tokens_seen": 832897024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037759277833500506,
      "loss": 2.8094,
      "theoretical_loss": 3.7143133122470715,
      "tokens_seen": 832962560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775827482447342,
      "loss": 2.8851,
      "theoretical_loss": 3.7142845554719544,
      "tokens_seen": 833028096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775727181544634,
      "loss": 2.6911,
      "theoretical_loss": 3.714255801592503,
      "tokens_seen": 833093632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775626880641926,
      "loss": 2.7356,
      "theoretical_loss": 3.7142270506081996,
      "tokens_seen": 833159168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775526579739218,
      "loss": 2.9705,
      "theoretical_loss": 3.7141983025185246,
      "tokens_seen": 833224704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037754262788365097,
      "loss": 2.901,
      "theoretical_loss": 3.714169557322958,
      "tokens_seen": 833290240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037753259779338015,
      "loss": 2.9581,
      "theoretical_loss": 3.714140815020982,
      "tokens_seen": 833355776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037752256770310933,
      "loss": 2.7036,
      "theoretical_loss": 3.7141120756120767,
      "tokens_seen": 833421312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037751253761283857,
      "loss": 2.7045,
      "theoretical_loss": 3.714083339095725,
      "tokens_seen": 833486848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775025075225677,
      "loss": 2.8971,
      "theoretical_loss": 3.7140546054714068,
      "tokens_seen": 833552384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037749247743229693,
      "loss": 2.6651,
      "theoretical_loss": 3.7140258747386046,
      "tokens_seen": 833617920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037748244734202605,
      "loss": 2.7434,
      "theoretical_loss": 3.7139971468967996,
      "tokens_seen": 833683456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774724172517553,
      "loss": 2.8242,
      "theoretical_loss": 3.713968421945474,
      "tokens_seen": 833748992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037746238716148447,
      "loss": 2.7712,
      "theoretical_loss": 3.7139396998841097,
      "tokens_seen": 833814528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037745235707121365,
      "loss": 2.8285,
      "theoretical_loss": 3.713910980712189,
      "tokens_seen": 833880064
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 467427,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3416881561279297,
      "objective/train/theoretical_loss": 3.713882264429194,
      "objective/train/tokens_used": 854405600,
      "theoretical_loss": 3.713882264429194,
      "tokens_seen": 833945600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037744232698094283,
      "loss": 2.7444,
      "theoretical_loss": 3.713882264429194,
      "tokens_seen": 833945600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037743229689067207,
      "loss": 2.7785,
      "theoretical_loss": 3.713853551034607,
      "tokens_seen": 834011136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774222668004012,
      "loss": 2.7879,
      "theoretical_loss": 3.713824840527911,
      "tokens_seen": 834076672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037741223671013043,
      "loss": 2.8718,
      "theoretical_loss": 3.7137961329085885,
      "tokens_seen": 834142208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037740220661985956,
      "loss": 2.618,
      "theoretical_loss": 3.713767428176122,
      "tokens_seen": 834207744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773921765295888,
      "loss": 2.6378,
      "theoretical_loss": 3.713738726329994,
      "tokens_seen": 834273280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377382146439318,
      "loss": 2.7106,
      "theoretical_loss": 3.7137100273696886,
      "tokens_seen": 834338816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037737211634904715,
      "loss": 2.7905,
      "theoretical_loss": 3.713681331294689,
      "tokens_seen": 834404352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037736208625877634,
      "loss": 2.8341,
      "theoretical_loss": 3.7136526381044774,
      "tokens_seen": 834469888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773520561685055,
      "loss": 3.0653,
      "theoretical_loss": 3.7136239477985384,
      "tokens_seen": 834535424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773420260782347,
      "loss": 2.6993,
      "theoretical_loss": 3.713595260376355,
      "tokens_seen": 834600960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037733199598796393,
      "loss": 2.6912,
      "theoretical_loss": 3.713566575837411,
      "tokens_seen": 834666496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037732196589769306,
      "loss": 2.887,
      "theoretical_loss": 3.713537894181191,
      "tokens_seen": 834732032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773119358074223,
      "loss": 2.8895,
      "theoretical_loss": 3.7135092154071776,
      "tokens_seen": 834797568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773019057171515,
      "loss": 2.737,
      "theoretical_loss": 3.7134805395148556,
      "tokens_seen": 834863104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037729187562688066,
      "loss": 2.5348,
      "theoretical_loss": 3.71345186650371,
      "tokens_seen": 834928640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037728184553660984,
      "loss": 2.673,
      "theoretical_loss": 3.7134231963732245,
      "tokens_seen": 834994176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377271815446339,
      "loss": 2.5236,
      "theoretical_loss": 3.7133945291228834,
      "tokens_seen": 835059712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772617853560682,
      "loss": 2.8603,
      "theoretical_loss": 3.713365864752172,
      "tokens_seen": 835125248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037725175526579744,
      "loss": 2.7683,
      "theoretical_loss": 3.7133372032605747,
      "tokens_seen": 835190784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037724172517552656,
      "loss": 2.822,
      "theoretical_loss": 3.7133085446475764,
      "tokens_seen": 835256320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772316950852558,
      "loss": 2.6931,
      "theoretical_loss": 3.7132798889126626,
      "tokens_seen": 835321856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772216649949849,
      "loss": 2.5958,
      "theoretical_loss": 3.713251236055317,
      "tokens_seen": 835387392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037721163490471416,
      "loss": 2.7673,
      "theoretical_loss": 3.7132225860750276,
      "tokens_seen": 835452928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037720160481444334,
      "loss": 2.8958,
      "theoretical_loss": 3.713193938971278,
      "tokens_seen": 835518464
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 468837,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6179723739624023,
      "objective/train/theoretical_loss": 3.7131652947435536,
      "objective/train/tokens_used": 856044000,
      "theoretical_loss": 3.7131652947435536,
      "tokens_seen": 835584000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771915747241725,
      "loss": 2.513,
      "theoretical_loss": 3.7131652947435536,
      "tokens_seen": 835584000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771815446339017,
      "loss": 2.6243,
      "theoretical_loss": 3.7131366533913406,
      "tokens_seen": 835649536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771715145436309,
      "loss": 2.6509,
      "theoretical_loss": 3.713108014914125,
      "tokens_seen": 835715072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037716148445336007,
      "loss": 2.594,
      "theoretical_loss": 3.7130793793113934,
      "tokens_seen": 835780608
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771514543630893,
      "loss": 2.697,
      "theoretical_loss": 3.7130507465826303,
      "tokens_seen": 835846144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037714142427281843,
      "loss": 2.6427,
      "theoretical_loss": 3.713022116727324,
      "tokens_seen": 835911680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037713139418254766,
      "loss": 2.9435,
      "theoretical_loss": 3.7129934897449592,
      "tokens_seen": 835977216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037712136409227685,
      "loss": 2.7768,
      "theoretical_loss": 3.7129648656350227,
      "tokens_seen": 836042752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377111334002006,
      "loss": 2.7805,
      "theoretical_loss": 3.7129362443970013,
      "tokens_seen": 836108288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771013039117352,
      "loss": 2.5428,
      "theoretical_loss": 3.7129076260303826,
      "tokens_seen": 836173824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770912738214644,
      "loss": 2.8572,
      "theoretical_loss": 3.7128790105346523,
      "tokens_seen": 836239360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037708124373119357,
      "loss": 2.6759,
      "theoretical_loss": 3.7128503979092984,
      "tokens_seen": 836304896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770712136409228,
      "loss": 2.9712,
      "theoretical_loss": 3.7128217881538075,
      "tokens_seen": 836370432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037706118355065193,
      "loss": 2.9666,
      "theoretical_loss": 3.712793181267667,
      "tokens_seen": 836435968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037705115346038117,
      "loss": 2.6751,
      "theoretical_loss": 3.7127645772503644,
      "tokens_seen": 836501504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770411233701103,
      "loss": 2.9037,
      "theoretical_loss": 3.7127359761013876,
      "tokens_seen": 836567040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037703109327983953,
      "loss": 2.8934,
      "theoretical_loss": 3.7127073778202235,
      "tokens_seen": 836632576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770210631895687,
      "loss": 2.7859,
      "theoretical_loss": 3.7126787824063605,
      "tokens_seen": 836698112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770110330992979,
      "loss": 2.5454,
      "theoretical_loss": 3.7126501898592865,
      "tokens_seen": 836763648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770010030090271,
      "loss": 2.836,
      "theoretical_loss": 3.71262160017849,
      "tokens_seen": 836829184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037699097291875625,
      "loss": 2.8422,
      "theoretical_loss": 3.7125930133634584,
      "tokens_seen": 836894720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037698094282848544,
      "loss": 2.8695,
      "theoretical_loss": 3.712564429413681,
      "tokens_seen": 836960256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037697091273821467,
      "loss": 2.9816,
      "theoretical_loss": 3.7125358483286455,
      "tokens_seen": 837025792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769608826479438,
      "loss": 2.7577,
      "theoretical_loss": 3.71250727010784,
      "tokens_seen": 837091328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037695085255767303,
      "loss": 2.8082,
      "theoretical_loss": 3.7124786947507546,
      "tokens_seen": 837156864
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 470129,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7207224369049072,
      "objective/train/theoretical_loss": 3.712450122256878,
      "objective/train/tokens_used": 857682400,
      "theoretical_loss": 3.712450122256878,
      "tokens_seen": 837222400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769408224674022,
      "loss": 2.7488,
      "theoretical_loss": 3.712450122256878,
      "tokens_seen": 837222400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769307923771314,
      "loss": 2.6886,
      "theoretical_loss": 3.712421552625699,
      "tokens_seen": 837287936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769207622868606,
      "loss": 2.4888,
      "theoretical_loss": 3.712392985856706,
      "tokens_seen": 837353472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037691073219658976,
      "loss": 2.9448,
      "theoretical_loss": 3.712364421949389,
      "tokens_seen": 837419008
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037690070210631894,
      "loss": 2.7908,
      "theoretical_loss": 3.712335860903238,
      "tokens_seen": 837484544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768906720160482,
      "loss": 2.7475,
      "theoretical_loss": 3.7123073027177407,
      "tokens_seen": 837550080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768806419257773,
      "loss": 2.8197,
      "theoretical_loss": 3.712278747392389,
      "tokens_seen": 837615616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037687061183550654,
      "loss": 2.3862,
      "theoretical_loss": 3.712250194926671,
      "tokens_seen": 837681152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768605817452357,
      "loss": 2.7735,
      "theoretical_loss": 3.712221645320077,
      "tokens_seen": 837746688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768505516549649,
      "loss": 2.7545,
      "theoretical_loss": 3.712193098572098,
      "tokens_seen": 837812224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037684052156469413,
      "loss": 2.5102,
      "theoretical_loss": 3.7121645546822233,
      "tokens_seen": 837877760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037683049147442326,
      "loss": 2.8537,
      "theoretical_loss": 3.7121360136499435,
      "tokens_seen": 837943296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768204613841525,
      "loss": 2.5985,
      "theoretical_loss": 3.712107475474749,
      "tokens_seen": 838008832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768104312938817,
      "loss": 2.97,
      "theoretical_loss": 3.7120789401561303,
      "tokens_seen": 838074368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037680040120361086,
      "loss": 2.7995,
      "theoretical_loss": 3.712050407693578,
      "tokens_seen": 838139904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037679037111334004,
      "loss": 3.0605,
      "theoretical_loss": 3.712021878086584,
      "tokens_seen": 838205440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767803410230692,
      "loss": 2.8332,
      "theoretical_loss": 3.7119933513346375,
      "tokens_seen": 838270976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767703109327984,
      "loss": 2.9444,
      "theoretical_loss": 3.7119648274372308,
      "tokens_seen": 838336512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037676028084252764,
      "loss": 3.0157,
      "theoretical_loss": 3.711936306393855,
      "tokens_seen": 838402048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037675025075225676,
      "loss": 2.6819,
      "theoretical_loss": 3.7119077882040017,
      "tokens_seen": 838467584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376740220661986,
      "loss": 2.7764,
      "theoretical_loss": 3.711879272867162,
      "tokens_seen": 838533120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767301905717151,
      "loss": 2.7516,
      "theoretical_loss": 3.711850760382827,
      "tokens_seen": 838598656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037672016048144436,
      "loss": 3.0291,
      "theoretical_loss": 3.71182225075049,
      "tokens_seen": 838664192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037671013039117354,
      "loss": 2.9408,
      "theoretical_loss": 3.7117937439696416,
      "tokens_seen": 838729728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767001003009027,
      "loss": 2.6546,
      "theoretical_loss": 3.7117652400397736,
      "tokens_seen": 838795264
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 470653,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.006452798843384,
      "objective/train/theoretical_loss": 3.7117367389603793,
      "objective/train/tokens_used": 859320800,
      "theoretical_loss": 3.7117367389603793,
      "tokens_seen": 838860800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766900702106319,
      "loss": 2.886,
      "theoretical_loss": 3.7117367389603793,
      "tokens_seen": 838860800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766800401203611,
      "loss": 2.6274,
      "theoretical_loss": 3.7117082407309505,
      "tokens_seen": 838926336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037667001003009027,
      "loss": 2.8262,
      "theoretical_loss": 3.7116797453509793,
      "tokens_seen": 838991872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766599799398195,
      "loss": 2.8575,
      "theoretical_loss": 3.711651252819959,
      "tokens_seen": 839057408
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037664994984954863,
      "loss": 2.7635,
      "theoretical_loss": 3.711622763137381,
      "tokens_seen": 839122944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037663991975927786,
      "loss": 2.7222,
      "theoretical_loss": 3.7115942763027396,
      "tokens_seen": 839188480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037662988966900705,
      "loss": 2.79,
      "theoretical_loss": 3.7115657923155267,
      "tokens_seen": 839254016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766198595787362,
      "loss": 2.621,
      "theoretical_loss": 3.711537311175236,
      "tokens_seen": 839319552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766098294884654,
      "loss": 2.7895,
      "theoretical_loss": 3.7115088328813597,
      "tokens_seen": 839385088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765997993981946,
      "loss": 2.7553,
      "theoretical_loss": 3.7114803574333926,
      "tokens_seen": 839450624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037658976930792377,
      "loss": 2.6999,
      "theoretical_loss": 3.711451884830827,
      "tokens_seen": 839516160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376579739217653,
      "loss": 2.827,
      "theoretical_loss": 3.7114234150731567,
      "tokens_seen": 839581696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037656970912738213,
      "loss": 2.5286,
      "theoretical_loss": 3.711394948159876,
      "tokens_seen": 839647232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037655967903711137,
      "loss": 2.7859,
      "theoretical_loss": 3.711366484090478,
      "tokens_seen": 839712768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765496489468405,
      "loss": 2.8513,
      "theoretical_loss": 3.7113380228644566,
      "tokens_seen": 839778304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037653961885656973,
      "loss": 2.6668,
      "theoretical_loss": 3.711309564481307,
      "tokens_seen": 839843840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765295887662989,
      "loss": 2.8288,
      "theoretical_loss": 3.711281108940522,
      "tokens_seen": 839909376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765195586760281,
      "loss": 2.907,
      "theoretical_loss": 3.711252656241597,
      "tokens_seen": 839974912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765095285857573,
      "loss": 2.6672,
      "theoretical_loss": 3.711224206384026,
      "tokens_seen": 840040448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037649949849548645,
      "loss": 2.5705,
      "theoretical_loss": 3.711195759367304,
      "tokens_seen": 840105984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037648946840521564,
      "loss": 2.5694,
      "theoretical_loss": 3.7111673151909255,
      "tokens_seen": 840171520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037647943831494487,
      "loss": 2.9223,
      "theoretical_loss": 3.7111388738543853,
      "tokens_seen": 840237056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376469408224674,
      "loss": 2.7859,
      "theoretical_loss": 3.7111104353571784,
      "tokens_seen": 840302592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037645937813440323,
      "loss": 2.8041,
      "theoretical_loss": 3.7110819996988003,
      "tokens_seen": 840368128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764493480441324,
      "loss": 2.6228,
      "theoretical_loss": 3.711053566878746,
      "tokens_seen": 840433664
    },
    {
      "epoch": 0.25,
      "objective/train/docs_used": 471790,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.994243860244751,
      "objective/train/theoretical_loss": 3.711025136896511,
      "objective/train/tokens_used": 860959200,
      "theoretical_loss": 3.711025136896511,
      "tokens_seen": 840499200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764393179538616,
      "loss": 2.864,
      "theoretical_loss": 3.711025136896511,
      "tokens_seen": 840499200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764292878635908,
      "loss": 2.5993,
      "theoretical_loss": 3.7109967097515906,
      "tokens_seen": 840564736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037641925777331996,
      "loss": 2.894,
      "theoretical_loss": 3.7109682854434802,
      "tokens_seen": 840630272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037640922768304914,
      "loss": 2.7505,
      "theoretical_loss": 3.710939863971676,
      "tokens_seen": 840695808
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763991975927784,
      "loss": 2.8284,
      "theoretical_loss": 3.7109114453356744,
      "tokens_seen": 840761344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763891675025075,
      "loss": 2.9597,
      "theoretical_loss": 3.710883029534971,
      "tokens_seen": 840826880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037637913741223674,
      "loss": 2.5393,
      "theoretical_loss": 3.710854616569061,
      "tokens_seen": 840892416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037636910732196586,
      "loss": 2.6253,
      "theoretical_loss": 3.710826206437442,
      "tokens_seen": 840957952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763590772316951,
      "loss": 2.9871,
      "theoretical_loss": 3.7107977991396104,
      "tokens_seen": 841023488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763490471414243,
      "loss": 2.5868,
      "theoretical_loss": 3.7107693946750624,
      "tokens_seen": 841089024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037633901705115346,
      "loss": 2.8583,
      "theoretical_loss": 3.710740993043294,
      "tokens_seen": 841154560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037632898696088264,
      "loss": 2.642,
      "theoretical_loss": 3.7107125942438035,
      "tokens_seen": 841220096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763189568706119,
      "loss": 3.005,
      "theoretical_loss": 3.710684198276086,
      "tokens_seen": 841285632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376308926780341,
      "loss": 2.6299,
      "theoretical_loss": 3.7106558051396403,
      "tokens_seen": 841351168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037629889669007024,
      "loss": 2.8208,
      "theoretical_loss": 3.7106274148339624,
      "tokens_seen": 841416704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037628886659979937,
      "loss": 2.6842,
      "theoretical_loss": 3.7105990273585503,
      "tokens_seen": 841482240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762788365095286,
      "loss": 2.8769,
      "theoretical_loss": 3.7105706427129013,
      "tokens_seen": 841547776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762688064192578,
      "loss": 2.8247,
      "theoretical_loss": 3.710542260896513,
      "tokens_seen": 841613312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037625877632898696,
      "loss": 2.8089,
      "theoretical_loss": 3.710513881908883,
      "tokens_seen": 841678848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037624874623871615,
      "loss": 2.6378,
      "theoretical_loss": 3.710485505749509,
      "tokens_seen": 841744384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762387161484453,
      "loss": 2.6468,
      "theoretical_loss": 3.7104571324178894,
      "tokens_seen": 841809920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762286860581745,
      "loss": 2.7883,
      "theoretical_loss": 3.710428761913522,
      "tokens_seen": 841875456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037621865596790374,
      "loss": 2.6884,
      "theoretical_loss": 3.710400394235905,
      "tokens_seen": 841940992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037620862587763287,
      "loss": 2.7462,
      "theoretical_loss": 3.710372029384537,
      "tokens_seen": 842006528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761985957873621,
      "loss": 2.7431,
      "theoretical_loss": 3.7103436673589156,
      "tokens_seen": 842072064
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 472419,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.929213762283325,
      "objective/train/theoretical_loss": 3.710315308158541,
      "objective/train/tokens_used": 862597600,
      "theoretical_loss": 3.710315308158541,
      "tokens_seen": 842137600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037618856569709123,
      "loss": 2.922,
      "theoretical_loss": 3.710315308158541,
      "tokens_seen": 842137600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037617853560682047,
      "loss": 2.7697,
      "theoretical_loss": 3.7102869517829102,
      "tokens_seen": 842203136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037616850551654965,
      "loss": 2.5601,
      "theoretical_loss": 3.7102585982315235,
      "tokens_seen": 842268672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037615847542627883,
      "loss": 2.746,
      "theoretical_loss": 3.7102302475038793,
      "tokens_seen": 842334208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000376148445336008,
      "loss": 2.8692,
      "theoretical_loss": 3.7102018995994763,
      "tokens_seen": 842399744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037613841524573725,
      "loss": 2.585,
      "theoretical_loss": 3.710173554517815,
      "tokens_seen": 842465280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037612838515546637,
      "loss": 2.4941,
      "theoretical_loss": 3.710145212258393,
      "tokens_seen": 842530816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761183550651956,
      "loss": 2.7491,
      "theoretical_loss": 3.710116872820711,
      "tokens_seen": 842596352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761083249749248,
      "loss": 2.7298,
      "theoretical_loss": 3.7100885362042684,
      "tokens_seen": 842661888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037609829488465397,
      "loss": 3.0135,
      "theoretical_loss": 3.710060202408565,
      "tokens_seen": 842727424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760882647943832,
      "loss": 2.7064,
      "theoretical_loss": 3.7100318714331006,
      "tokens_seen": 842792960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037607823470411233,
      "loss": 2.8233,
      "theoretical_loss": 3.710003543277376,
      "tokens_seen": 842858496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037606820461384157,
      "loss": 2.6252,
      "theoretical_loss": 3.7099752179408894,
      "tokens_seen": 842924032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760581745235707,
      "loss": 2.6879,
      "theoretical_loss": 3.709946895423143,
      "tokens_seen": 842989568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037604814443329993,
      "loss": 2.7986,
      "theoretical_loss": 3.709918575723636,
      "tokens_seen": 843055104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760381143430291,
      "loss": 2.8934,
      "theoretical_loss": 3.7098902588418694,
      "tokens_seen": 843120640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760280842527583,
      "loss": 2.7243,
      "theoretical_loss": 3.7098619447773444,
      "tokens_seen": 843186176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760180541624875,
      "loss": 2.7749,
      "theoretical_loss": 3.7098336335295605,
      "tokens_seen": 843251712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037600802407221665,
      "loss": 2.7942,
      "theoretical_loss": 3.70980532509802,
      "tokens_seen": 843317248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037599799398194584,
      "loss": 2.636,
      "theoretical_loss": 3.7097770194822224,
      "tokens_seen": 843382784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037598796389167507,
      "loss": 2.5375,
      "theoretical_loss": 3.70974871668167,
      "tokens_seen": 843448320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759779338014042,
      "loss": 2.8475,
      "theoretical_loss": 3.709720416695864,
      "tokens_seen": 843513856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037596790371113343,
      "loss": 2.6591,
      "theoretical_loss": 3.709692119524305,
      "tokens_seen": 843579392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759578736208626,
      "loss": 2.4604,
      "theoretical_loss": 3.709663825166496,
      "tokens_seen": 843644928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759478435305918,
      "loss": 2.8007,
      "theoretical_loss": 3.7096355336219373,
      "tokens_seen": 843710464
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 473620,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1842246055603027,
      "objective/train/theoretical_loss": 3.709607244890131,
      "objective/train/tokens_used": 864236000,
      "theoretical_loss": 3.709607244890131,
      "tokens_seen": 843776000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375937813440321,
      "loss": 2.5554,
      "theoretical_loss": 3.709607244890131,
      "tokens_seen": 843776000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037592778335005016,
      "loss": 2.854,
      "theoretical_loss": 3.70957895897058,
      "tokens_seen": 843841536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037591775325977934,
      "loss": 2.844,
      "theoretical_loss": 3.7095506758627845,
      "tokens_seen": 843907072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759077231695086,
      "loss": 2.656,
      "theoretical_loss": 3.7095223955662484,
      "tokens_seen": 843972608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758976930792377,
      "loss": 2.69,
      "theoretical_loss": 3.709494118080473,
      "tokens_seen": 844038144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037588766298896694,
      "loss": 2.6778,
      "theoretical_loss": 3.7094658434049617,
      "tokens_seen": 844103680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037587763289869606,
      "loss": 2.9431,
      "theoretical_loss": 3.709437571539216,
      "tokens_seen": 844169216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758676028084253,
      "loss": 2.8346,
      "theoretical_loss": 3.709409302482739,
      "tokens_seen": 844234752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758575727181545,
      "loss": 2.7605,
      "theoretical_loss": 3.7093810362350332,
      "tokens_seen": 844300288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037584754262788366,
      "loss": 2.5804,
      "theoretical_loss": 3.7093527727956026,
      "tokens_seen": 844365824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037583751253761284,
      "loss": 2.5707,
      "theoretical_loss": 3.709324512163949,
      "tokens_seen": 844431360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758274824473421,
      "loss": 2.9522,
      "theoretical_loss": 3.709296254339576,
      "tokens_seen": 844496896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758174523570712,
      "loss": 2.804,
      "theoretical_loss": 3.7092679993219875,
      "tokens_seen": 844562432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037580742226680044,
      "loss": 2.6918,
      "theoretical_loss": 3.709239747110686,
      "tokens_seen": 844627968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037579739217652957,
      "loss": 2.7089,
      "theoretical_loss": 3.7092114977051756,
      "tokens_seen": 844693504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757873620862588,
      "loss": 2.8012,
      "theoretical_loss": 3.7091832511049603,
      "tokens_seen": 844759040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375777331995988,
      "loss": 2.5886,
      "theoretical_loss": 3.709155007309543,
      "tokens_seen": 844824576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037576730190571716,
      "loss": 2.7937,
      "theoretical_loss": 3.7091267663184286,
      "tokens_seen": 844890112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037575727181544635,
      "loss": 2.6963,
      "theoretical_loss": 3.7090985281311206,
      "tokens_seen": 844955648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757472417251755,
      "loss": 2.7533,
      "theoretical_loss": 3.7090702927471235,
      "tokens_seen": 845021184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757372116349047,
      "loss": 2.7116,
      "theoretical_loss": 3.7090420601659417,
      "tokens_seen": 845086720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037572718154463394,
      "loss": 2.975,
      "theoretical_loss": 3.7090138303870788,
      "tokens_seen": 845152256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037571715145436307,
      "loss": 2.8198,
      "theoretical_loss": 3.7089856034100404,
      "tokens_seen": 845217792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757071213640923,
      "loss": 2.7096,
      "theoretical_loss": 3.708957379234331,
      "tokens_seen": 845283328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037569709127382143,
      "loss": 2.7556,
      "theoretical_loss": 3.7089291578594548,
      "tokens_seen": 845348864
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 474273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4147276878356934,
      "objective/train/theoretical_loss": 3.7089009392849173,
      "objective/train/tokens_used": 865874400,
      "theoretical_loss": 3.7089009392849173,
      "tokens_seen": 845414400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037568706118355067,
      "loss": 2.8156,
      "theoretical_loss": 3.7089009392849173,
      "tokens_seen": 845414400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037567703109327985,
      "loss": 2.8181,
      "theoretical_loss": 3.7088727235102237,
      "tokens_seen": 845479936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037566700100300903,
      "loss": 2.7168,
      "theoretical_loss": 3.7088445105348793,
      "tokens_seen": 845545472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756569709127382,
      "loss": 2.7358,
      "theoretical_loss": 3.7088163003583885,
      "tokens_seen": 845611008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037564694082246745,
      "loss": 2.6648,
      "theoretical_loss": 3.7087880929802575,
      "tokens_seen": 845676544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756369107321966,
      "loss": 2.7797,
      "theoretical_loss": 3.708759888399992,
      "tokens_seen": 845742080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756268806419258,
      "loss": 2.7489,
      "theoretical_loss": 3.708731686617097,
      "tokens_seen": 845807616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037561685055165494,
      "loss": 2.6411,
      "theoretical_loss": 3.7087034876310794,
      "tokens_seen": 845873152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037560682046138417,
      "loss": 2.5334,
      "theoretical_loss": 3.7086752914414443,
      "tokens_seen": 845938688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037559679037111335,
      "loss": 2.5147,
      "theoretical_loss": 3.708647098047698,
      "tokens_seen": 846004224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037558676028084253,
      "loss": 2.771,
      "theoretical_loss": 3.708618907449347,
      "tokens_seen": 846069760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755767301905717,
      "loss": 2.8255,
      "theoretical_loss": 3.7085907196458976,
      "tokens_seen": 846135296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755667001003009,
      "loss": 2.8221,
      "theoretical_loss": 3.7085625346368554,
      "tokens_seen": 846200832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755566700100301,
      "loss": 2.6886,
      "theoretical_loss": 3.7085343524217285,
      "tokens_seen": 846266368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755466399197593,
      "loss": 2.5909,
      "theoretical_loss": 3.708506173000022,
      "tokens_seen": 846331904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037553660982948844,
      "loss": 2.6621,
      "theoretical_loss": 3.7084779963712435,
      "tokens_seen": 846397440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755265797392177,
      "loss": 2.4955,
      "theoretical_loss": 3.7084498225349005,
      "tokens_seen": 846462976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755165496489468,
      "loss": 2.7909,
      "theoretical_loss": 3.7084216514904993,
      "tokens_seen": 846528512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037550651955867604,
      "loss": 2.7867,
      "theoretical_loss": 3.7083934832375474,
      "tokens_seen": 846594048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754964894684052,
      "loss": 2.63,
      "theoretical_loss": 3.7083653177755522,
      "tokens_seen": 846659584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754864593781344,
      "loss": 2.647,
      "theoretical_loss": 3.708337155104021,
      "tokens_seen": 846725120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754764292878636,
      "loss": 2.5783,
      "theoretical_loss": 3.7083089952224615,
      "tokens_seen": 846790656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754663991975928,
      "loss": 2.7086,
      "theoretical_loss": 3.7082808381303813,
      "tokens_seen": 846856192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037545636910732194,
      "loss": 2.8197,
      "theoretical_loss": 3.7082526838272885,
      "tokens_seen": 846921728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754463390170512,
      "loss": 2.6207,
      "theoretical_loss": 3.7082245323126903,
      "tokens_seen": 846987264
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 475572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3531529903411865,
      "objective/train/theoretical_loss": 3.7081963835860963,
      "objective/train/tokens_used": 867512800,
      "theoretical_loss": 3.7081963835860963,
      "tokens_seen": 847052800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754363089267803,
      "loss": 2.3995,
      "theoretical_loss": 3.7081963835860963,
      "tokens_seen": 847052800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037542627883650954,
      "loss": 2.7983,
      "theoretical_loss": 3.7081682376470133,
      "tokens_seen": 847118336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754162487462387,
      "loss": 2.8877,
      "theoretical_loss": 3.7081400944949503,
      "tokens_seen": 847183872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754062186559679,
      "loss": 2.7212,
      "theoretical_loss": 3.708111954129415,
      "tokens_seen": 847249408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753961885656971,
      "loss": 2.747,
      "theoretical_loss": 3.7080838165499177,
      "tokens_seen": 847314944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037538615847542626,
      "loss": 2.8743,
      "theoretical_loss": 3.7080556817559653,
      "tokens_seen": 847380480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037537612838515544,
      "loss": 2.7958,
      "theoretical_loss": 3.7080275497470674,
      "tokens_seen": 847446016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753660982948847,
      "loss": 2.5247,
      "theoretical_loss": 3.707999420522733,
      "tokens_seen": 847511552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037535606820461386,
      "loss": 2.8099,
      "theoretical_loss": 3.707971294082471,
      "tokens_seen": 847577088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037534603811434304,
      "loss": 2.8774,
      "theoretical_loss": 3.7079431704257906,
      "tokens_seen": 847642624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753360080240723,
      "loss": 2.8221,
      "theoretical_loss": 3.7079150495522013,
      "tokens_seen": 847708160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753259779338014,
      "loss": 2.7504,
      "theoretical_loss": 3.7078869314612124,
      "tokens_seen": 847773696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037531594784353064,
      "loss": 2.5578,
      "theoretical_loss": 3.707858816152334,
      "tokens_seen": 847839232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037530591775325977,
      "loss": 2.8172,
      "theoretical_loss": 3.707830703625075,
      "tokens_seen": 847904768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375295887662989,
      "loss": 2.7628,
      "theoretical_loss": 3.7078025938789456,
      "tokens_seen": 847970304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752858575727182,
      "loss": 2.4287,
      "theoretical_loss": 3.707774486913456,
      "tokens_seen": 848035840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037527582748244736,
      "loss": 2.8295,
      "theoretical_loss": 3.707746382728116,
      "tokens_seen": 848101376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037526579739217655,
      "loss": 2.6956,
      "theoretical_loss": 3.7077182813224354,
      "tokens_seen": 848166912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752557673019057,
      "loss": 2.5746,
      "theoretical_loss": 3.7076901826959254,
      "tokens_seen": 848232448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752457372116349,
      "loss": 2.4742,
      "theoretical_loss": 3.7076620868480963,
      "tokens_seen": 848297984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037523570712136414,
      "loss": 2.6048,
      "theoretical_loss": 3.707633993778458,
      "tokens_seen": 848363520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037522567703109327,
      "loss": 2.8559,
      "theoretical_loss": 3.707605903486522,
      "tokens_seen": 848429056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752156469408225,
      "loss": 2.4246,
      "theoretical_loss": 3.7075778159717983,
      "tokens_seen": 848494592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037520561685055163,
      "loss": 2.7434,
      "theoretical_loss": 3.7075497312337986,
      "tokens_seen": 848560128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037519558676028087,
      "loss": 2.717,
      "theoretical_loss": 3.7075216492720338,
      "tokens_seen": 848625664
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 476007,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1800570487976074,
      "objective/train/theoretical_loss": 3.7074935700860143,
      "objective/train/tokens_used": 869151200,
      "theoretical_loss": 3.7074935700860143,
      "tokens_seen": 848691200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037518555667001005,
      "loss": 2.9501,
      "theoretical_loss": 3.7074935700860143,
      "tokens_seen": 848691200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037517552657973923,
      "loss": 2.5482,
      "theoretical_loss": 3.7074654936752527,
      "tokens_seen": 848756736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751654964894684,
      "loss": 2.6684,
      "theoretical_loss": 3.70743742003926,
      "tokens_seen": 848822272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037515546639919765,
      "loss": 2.66,
      "theoretical_loss": 3.7074093491775475,
      "tokens_seen": 848887808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751454363089268,
      "loss": 2.5614,
      "theoretical_loss": 3.7073812810896265,
      "tokens_seen": 848953344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375135406218656,
      "loss": 2.7405,
      "theoretical_loss": 3.7073532157750098,
      "tokens_seen": 849018880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037512537612838514,
      "loss": 2.7963,
      "theoretical_loss": 3.7073251532332088,
      "tokens_seen": 849084416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037511534603811437,
      "loss": 2.7726,
      "theoretical_loss": 3.707297093463735,
      "tokens_seen": 849149952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037510531594784355,
      "loss": 2.526,
      "theoretical_loss": 3.7072690364661023,
      "tokens_seen": 849215488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037509528585757273,
      "loss": 2.7597,
      "theoretical_loss": 3.707240982239821,
      "tokens_seen": 849281024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750852557673019,
      "loss": 2.6971,
      "theoretical_loss": 3.7072129307844053,
      "tokens_seen": 849346560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750752256770311,
      "loss": 2.8605,
      "theoretical_loss": 3.707184882099366,
      "tokens_seen": 849412096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750651955867603,
      "loss": 2.5104,
      "theoretical_loss": 3.7071568361842173,
      "tokens_seen": 849477632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750551654964895,
      "loss": 2.5936,
      "theoretical_loss": 3.7071287930384713,
      "tokens_seen": 849543168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037504513540621864,
      "loss": 2.5493,
      "theoretical_loss": 3.7071007526616406,
      "tokens_seen": 849608704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750351053159479,
      "loss": 2.6672,
      "theoretical_loss": 3.707072715053239,
      "tokens_seen": 849674240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375025075225677,
      "loss": 2.8005,
      "theoretical_loss": 3.707044680212779,
      "tokens_seen": 849739776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037501504513540624,
      "loss": 2.7663,
      "theoretical_loss": 3.707016648139775,
      "tokens_seen": 849805312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750050150451354,
      "loss": 2.8586,
      "theoretical_loss": 3.7069886188337393,
      "tokens_seen": 849870848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749949849548646,
      "loss": 2.801,
      "theoretical_loss": 3.7069605922941857,
      "tokens_seen": 849936384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749849548645938,
      "loss": 2.865,
      "theoretical_loss": 3.706932568520628,
      "tokens_seen": 850001920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374974924774323,
      "loss": 2.7011,
      "theoretical_loss": 3.70690454751258,
      "tokens_seen": 850067456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037496489468405214,
      "loss": 2.7869,
      "theoretical_loss": 3.7068765292695556,
      "tokens_seen": 850132992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749548645937814,
      "loss": 2.889,
      "theoretical_loss": 3.7068485137910683,
      "tokens_seen": 850198528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749448345035105,
      "loss": 2.6505,
      "theoretical_loss": 3.7068205010766335,
      "tokens_seen": 850264064
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 477422,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5388436317443848,
      "objective/train/theoretical_loss": 3.7067924911257646,
      "objective/train/tokens_used": 870789600,
      "theoretical_loss": 3.7067924911257646,
      "tokens_seen": 850329600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037493480441323974,
      "loss": 2.5794,
      "theoretical_loss": 3.7067924911257646,
      "tokens_seen": 850329600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749247743229689,
      "loss": 2.5744,
      "theoretical_loss": 3.7067644839379756,
      "tokens_seen": 850395136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749147442326981,
      "loss": 2.7938,
      "theoretical_loss": 3.706736479512782,
      "tokens_seen": 850460672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749047141424273,
      "loss": 2.7906,
      "theoretical_loss": 3.7067084778496984,
      "tokens_seen": 850526208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037489468405215646,
      "loss": 2.7479,
      "theoretical_loss": 3.7066804789482384,
      "tokens_seen": 850591744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037488465396188564,
      "loss": 2.5192,
      "theoretical_loss": 3.706652482807918,
      "tokens_seen": 850657280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748746238716149,
      "loss": 2.5515,
      "theoretical_loss": 3.7066244894282523,
      "tokens_seen": 850722816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374864593781344,
      "loss": 2.8724,
      "theoretical_loss": 3.7065964988087554,
      "tokens_seen": 850788352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037485456369107324,
      "loss": 2.6995,
      "theoretical_loss": 3.7065685109489435,
      "tokens_seen": 850853888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037484453360080237,
      "loss": 2.5975,
      "theoretical_loss": 3.7065405258483315,
      "tokens_seen": 850919424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748345035105316,
      "loss": 2.9188,
      "theoretical_loss": 3.7065125435064354,
      "tokens_seen": 850984960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748244734202608,
      "loss": 2.7731,
      "theoretical_loss": 3.7064845639227704,
      "tokens_seen": 851050496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037481444332998997,
      "loss": 2.7823,
      "theoretical_loss": 3.706456587096852,
      "tokens_seen": 851116032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037480441323971915,
      "loss": 2.5889,
      "theoretical_loss": 3.706428613028197,
      "tokens_seen": 851181568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747943831494484,
      "loss": 2.7635,
      "theoretical_loss": 3.706400641716321,
      "tokens_seen": 851247104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747843530591775,
      "loss": 2.6144,
      "theoretical_loss": 3.7063726731607396,
      "tokens_seen": 851312640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037477432296890675,
      "loss": 2.663,
      "theoretical_loss": 3.7063447073609694,
      "tokens_seen": 851378176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037476429287863587,
      "loss": 2.6883,
      "theoretical_loss": 3.706316744316527,
      "tokens_seen": 851443712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747542627883651,
      "loss": 2.6976,
      "theoretical_loss": 3.7062887840269276,
      "tokens_seen": 851509248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747442326980943,
      "loss": 2.778,
      "theoretical_loss": 3.7062608264916896,
      "tokens_seen": 851574784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037473420260782347,
      "loss": 2.7051,
      "theoretical_loss": 3.7062328717103292,
      "tokens_seen": 851640320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037472417251755265,
      "loss": 2.8306,
      "theoretical_loss": 3.7062049196823628,
      "tokens_seen": 851705856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037471414242728183,
      "loss": 2.4378,
      "theoretical_loss": 3.7061769704073075,
      "tokens_seen": 851771392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374704112337011,
      "loss": 2.5944,
      "theoretical_loss": 3.706149023884681,
      "tokens_seen": 851836928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037469408224674025,
      "loss": 2.783,
      "theoretical_loss": 3.7061210801139994,
      "tokens_seen": 851902464
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 478724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.294400453567505,
      "objective/train/theoretical_loss": 3.706093139094781,
      "objective/train/tokens_used": 872428000,
      "theoretical_loss": 3.706093139094781,
      "tokens_seen": 851968000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003746840521564694,
      "loss": 2.5393,
      "theoretical_loss": 3.706093139094781,
      "tokens_seen": 851968000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003746740220661986,
      "loss": 2.7337,
      "theoretical_loss": 3.706065200826543,
      "tokens_seen": 852033536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037466399197592774,
      "loss": 2.5102,
      "theoretical_loss": 3.7060372653088027,
      "tokens_seen": 852099072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374653961885657,
      "loss": 2.5369,
      "theoretical_loss": 3.706009332541078,
      "tokens_seen": 852164608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037464393179538615,
      "loss": 2.7483,
      "theoretical_loss": 3.7059814025228865,
      "tokens_seen": 852230144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037463390170511534,
      "loss": 2.7406,
      "theoretical_loss": 3.7059534752537466,
      "tokens_seen": 852295680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037462387161484457,
      "loss": 2.7608,
      "theoretical_loss": 3.705925550733176,
      "tokens_seen": 852361216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037461384152457375,
      "loss": 2.7804,
      "theoretical_loss": 3.7058976289606935,
      "tokens_seen": 852426752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037460381143430293,
      "loss": 2.7932,
      "theoretical_loss": 3.7058697099358167,
      "tokens_seen": 852492288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745937813440321,
      "loss": 2.7504,
      "theoretical_loss": 3.705841793658064,
      "tokens_seen": 852557824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745837512537613,
      "loss": 2.8511,
      "theoretical_loss": 3.7058138801269545,
      "tokens_seen": 852623360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745737211634905,
      "loss": 2.6997,
      "theoretical_loss": 3.7057859693420063,
      "tokens_seen": 852688896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745636910732197,
      "loss": 2.8532,
      "theoretical_loss": 3.7057580613027383,
      "tokens_seen": 852754432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037455366098294884,
      "loss": 2.5094,
      "theoretical_loss": 3.70573015600867,
      "tokens_seen": 852819968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745436308926781,
      "loss": 2.8309,
      "theoretical_loss": 3.70570225345932,
      "tokens_seen": 852885504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745336008024072,
      "loss": 2.7093,
      "theoretical_loss": 3.7056743536542074,
      "tokens_seen": 852951040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037452357071213644,
      "loss": 2.8218,
      "theoretical_loss": 3.7056464565928517,
      "tokens_seen": 853016576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745135406218656,
      "loss": 2.5999,
      "theoretical_loss": 3.705618562274772,
      "tokens_seen": 853082112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745035105315948,
      "loss": 2.562,
      "theoretical_loss": 3.7055906706994883,
      "tokens_seen": 853147648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374493480441324,
      "loss": 2.7136,
      "theoretical_loss": 3.7055627818665196,
      "tokens_seen": 853213184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744834503510532,
      "loss": 2.7561,
      "theoretical_loss": 3.7055348957753864,
      "tokens_seen": 853278720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037447342026078234,
      "loss": 2.7161,
      "theoretical_loss": 3.705507012425608,
      "tokens_seen": 853344256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744633901705116,
      "loss": 2.7196,
      "theoretical_loss": 3.7054791318167046,
      "tokens_seen": 853409792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744533600802407,
      "loss": 2.8581,
      "theoretical_loss": 3.7054512539481967,
      "tokens_seen": 853475328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037444332998996994,
      "loss": 2.7128,
      "theoretical_loss": 3.705423378819604,
      "tokens_seen": 853540864
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 479323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.415107488632202,
      "objective/train/theoretical_loss": 3.7053955064304467,
      "objective/train/tokens_used": 874066400,
      "theoretical_loss": 3.7053955064304467,
      "tokens_seen": 853606400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744332998996991,
      "loss": 2.7135,
      "theoretical_loss": 3.7053955064304467,
      "tokens_seen": 853606400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744232698094283,
      "loss": 2.5923,
      "theoretical_loss": 3.705367636780246,
      "tokens_seen": 853671936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744132397191575,
      "loss": 2.6997,
      "theoretical_loss": 3.705339769868522,
      "tokens_seen": 853737472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037440320962888666,
      "loss": 2.777,
      "theoretical_loss": 3.7053119056947956,
      "tokens_seen": 853803008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037439317953861585,
      "loss": 2.5342,
      "theoretical_loss": 3.7052840442585877,
      "tokens_seen": 853868544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743831494483451,
      "loss": 2.7932,
      "theoretical_loss": 3.7052561855594193,
      "tokens_seen": 853934080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743731193580742,
      "loss": 2.5486,
      "theoretical_loss": 3.705228329596811,
      "tokens_seen": 853999616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037436308926780344,
      "loss": 2.8331,
      "theoretical_loss": 3.705200476370285,
      "tokens_seen": 854065152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037435305917753257,
      "loss": 2.6312,
      "theoretical_loss": 3.7051726258793622,
      "tokens_seen": 854130688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743430290872618,
      "loss": 2.6545,
      "theoretical_loss": 3.7051447781235636,
      "tokens_seen": 854196224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374332998996991,
      "loss": 2.6804,
      "theoretical_loss": 3.7051169331024107,
      "tokens_seen": 854261760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037432296890672017,
      "loss": 2.646,
      "theoretical_loss": 3.7050890908154264,
      "tokens_seen": 854327296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037431293881644935,
      "loss": 2.5474,
      "theoretical_loss": 3.7050612512621313,
      "tokens_seen": 854392832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743029087261786,
      "loss": 2.6389,
      "theoretical_loss": 3.7050334144420476,
      "tokens_seen": 854458368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742928786359077,
      "loss": 2.6941,
      "theoretical_loss": 3.7050055803546975,
      "tokens_seen": 854523904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037428284854563695,
      "loss": 2.6424,
      "theoretical_loss": 3.7049777489996036,
      "tokens_seen": 854589440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037427281845536607,
      "loss": 2.6623,
      "theoretical_loss": 3.7049499203762872,
      "tokens_seen": 854654976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742627883650953,
      "loss": 2.6297,
      "theoretical_loss": 3.704922094484272,
      "tokens_seen": 854720512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742527582748245,
      "loss": 2.774,
      "theoretical_loss": 3.7048942713230795,
      "tokens_seen": 854786048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037424272818455367,
      "loss": 2.73,
      "theoretical_loss": 3.7048664508922324,
      "tokens_seen": 854851584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037423269809428285,
      "loss": 2.7748,
      "theoretical_loss": 3.704838633191254,
      "tokens_seen": 854917120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037422266800401203,
      "loss": 2.6596,
      "theoretical_loss": 3.7048108182196673,
      "tokens_seen": 854982656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742126379137412,
      "loss": 2.7619,
      "theoretical_loss": 3.704783005976995,
      "tokens_seen": 855048192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037420260782347045,
      "loss": 2.7137,
      "theoretical_loss": 3.7047551964627594,
      "tokens_seen": 855113728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741925777331996,
      "loss": 2.8362,
      "theoretical_loss": 3.704727389676486,
      "tokens_seen": 855179264
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 479893,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.8994100093841553,
      "objective/train/theoretical_loss": 3.7046995856176954,
      "objective/train/tokens_used": 875704800,
      "theoretical_loss": 3.7046995856176954,
      "tokens_seen": 855244800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741825476429288,
      "loss": 2.3327,
      "theoretical_loss": 3.7046995856176954,
      "tokens_seen": 855244800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037417251755265794,
      "loss": 2.7426,
      "theoretical_loss": 3.704671784285913,
      "tokens_seen": 855310336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741624874623872,
      "loss": 2.7,
      "theoretical_loss": 3.704643985680662,
      "tokens_seen": 855375872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037415245737211635,
      "loss": 2.7171,
      "theoretical_loss": 3.704616189801466,
      "tokens_seen": 855441408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037414242728184554,
      "loss": 2.7728,
      "theoretical_loss": 3.704588396647848,
      "tokens_seen": 855506944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741323971915747,
      "loss": 2.7179,
      "theoretical_loss": 3.704560606219334,
      "tokens_seen": 855572480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037412236710130395,
      "loss": 2.4159,
      "theoretical_loss": 3.7045328185154465,
      "tokens_seen": 855638016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741123370110331,
      "loss": 2.6816,
      "theoretical_loss": 3.7045050335357104,
      "tokens_seen": 855703552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741023069207623,
      "loss": 2.7457,
      "theoretical_loss": 3.7044772512796493,
      "tokens_seen": 855769088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037409227683049144,
      "loss": 2.4778,
      "theoretical_loss": 3.7044494717467886,
      "tokens_seen": 855834624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740822467402207,
      "loss": 2.759,
      "theoretical_loss": 3.704421694936652,
      "tokens_seen": 855900160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037407221664994986,
      "loss": 2.7427,
      "theoretical_loss": 3.704393920848765,
      "tokens_seen": 855965696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037406218655967904,
      "loss": 2.8548,
      "theoretical_loss": 3.704366149482652,
      "tokens_seen": 856031232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740521564694082,
      "loss": 2.6829,
      "theoretical_loss": 3.7043383808378376,
      "tokens_seen": 856096768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740421263791374,
      "loss": 2.4935,
      "theoretical_loss": 3.7043106149138474,
      "tokens_seen": 856162304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740320962888666,
      "loss": 2.4234,
      "theoretical_loss": 3.7042828517102064,
      "tokens_seen": 856227840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740220661985958,
      "loss": 2.6602,
      "theoretical_loss": 3.7042550912264405,
      "tokens_seen": 856293376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037401203610832494,
      "loss": 2.6378,
      "theoretical_loss": 3.7042273334620734,
      "tokens_seen": 856358912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740020060180542,
      "loss": 2.5173,
      "theoretical_loss": 3.7041995784166324,
      "tokens_seen": 856424448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037399197592778336,
      "loss": 2.8294,
      "theoretical_loss": 3.7041718260896426,
      "tokens_seen": 856489984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037398194583751254,
      "loss": 2.7361,
      "theoretical_loss": 3.7041440764806293,
      "tokens_seen": 856555520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739719157472417,
      "loss": 2.9064,
      "theoretical_loss": 3.7041163295891186,
      "tokens_seen": 856621056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739618856569709,
      "loss": 2.7333,
      "theoretical_loss": 3.7040885854146364,
      "tokens_seen": 856686592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739518555667001,
      "loss": 2.7232,
      "theoretical_loss": 3.7040608439567095,
      "tokens_seen": 856752128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739418254764293,
      "loss": 2.6206,
      "theoretical_loss": 3.7040331052148634,
      "tokens_seen": 856817664
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 480416,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0554239749908447,
      "objective/train/theoretical_loss": 3.704005369188625,
      "objective/train/tokens_used": 877343200,
      "theoretical_loss": 3.704005369188625,
      "tokens_seen": 856883200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037393179538615845,
      "loss": 2.8683,
      "theoretical_loss": 3.704005369188625,
      "tokens_seen": 856883200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739217652958877,
      "loss": 2.8753,
      "theoretical_loss": 3.70397763587752,
      "tokens_seen": 856948736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739117352056168,
      "loss": 2.4591,
      "theoretical_loss": 3.7039499052810765,
      "tokens_seen": 857014272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037390170511534605,
      "loss": 2.7519,
      "theoretical_loss": 3.703922177398819,
      "tokens_seen": 857079808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738916750250752,
      "loss": 2.7931,
      "theoretical_loss": 3.7038944522302764,
      "tokens_seen": 857145344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738816449348044,
      "loss": 2.6563,
      "theoretical_loss": 3.7038667297749743,
      "tokens_seen": 857210880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037387161484453364,
      "loss": 2.6392,
      "theoretical_loss": 3.703839010032441,
      "tokens_seen": 857276416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037386158475426277,
      "loss": 2.6964,
      "theoretical_loss": 3.703811293002202,
      "tokens_seen": 857341952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373851554663992,
      "loss": 2.5881,
      "theoretical_loss": 3.7037835786837867,
      "tokens_seen": 857407488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738415245737212,
      "loss": 2.697,
      "theoretical_loss": 3.7037558670767208,
      "tokens_seen": 857473024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037383149448345037,
      "loss": 2.611,
      "theoretical_loss": 3.7037281581805326,
      "tokens_seen": 857538560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037382146439317955,
      "loss": 2.5995,
      "theoretical_loss": 3.7037004519947496,
      "tokens_seen": 857604096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738114343029088,
      "loss": 2.3132,
      "theoretical_loss": 3.7036727485188994,
      "tokens_seen": 857669632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738014042126379,
      "loss": 2.7532,
      "theoretical_loss": 3.7036450477525102,
      "tokens_seen": 857735168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037379137412236715,
      "loss": 2.7108,
      "theoretical_loss": 3.7036173496951106,
      "tokens_seen": 857800704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037378134403209627,
      "loss": 2.9879,
      "theoretical_loss": 3.703589654346227,
      "tokens_seen": 857866240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737713139418255,
      "loss": 2.9101,
      "theoretical_loss": 3.7035619617053896,
      "tokens_seen": 857931776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737612838515547,
      "loss": 2.6775,
      "theoretical_loss": 3.703534271772125,
      "tokens_seen": 857997312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037375125376128387,
      "loss": 2.6034,
      "theoretical_loss": 3.7035065845459636,
      "tokens_seen": 858062848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037374122367101305,
      "loss": 2.6527,
      "theoretical_loss": 3.7034789000264325,
      "tokens_seen": 858128384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037373119358074223,
      "loss": 2.7135,
      "theoretical_loss": 3.7034512182130612,
      "tokens_seen": 858193920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737211634904714,
      "loss": 2.5494,
      "theoretical_loss": 3.703423539105378,
      "tokens_seen": 858259456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037371113340020065,
      "loss": 2.7046,
      "theoretical_loss": 3.703395862702912,
      "tokens_seen": 858324992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737011033099298,
      "loss": 2.5938,
      "theoretical_loss": 3.703368189005193,
      "tokens_seen": 858390528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373691073219659,
      "loss": 2.5671,
      "theoretical_loss": 3.70334051801175,
      "tokens_seen": 858456064
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 481568,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5986602306365967,
      "objective/train/theoretical_loss": 3.703312849722111,
      "objective/train/tokens_used": 878981600,
      "theoretical_loss": 3.703312849722111,
      "tokens_seen": 858521600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037368104312938814,
      "loss": 2.8,
      "theoretical_loss": 3.703312849722111,
      "tokens_seen": 858521600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736710130391174,
      "loss": 2.6659,
      "theoretical_loss": 3.7032851841358063,
      "tokens_seen": 858587136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037366098294884655,
      "loss": 2.431,
      "theoretical_loss": 3.7032575212523664,
      "tokens_seen": 858652672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037365095285857574,
      "loss": 2.788,
      "theoretical_loss": 3.7032298610713195,
      "tokens_seen": 858718208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736409227683049,
      "loss": 2.5209,
      "theoretical_loss": 3.7032022035921965,
      "tokens_seen": 858783744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037363089267803415,
      "loss": 2.6643,
      "theoretical_loss": 3.7031745488145265,
      "tokens_seen": 858849280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736208625877633,
      "loss": 2.8994,
      "theoretical_loss": 3.70314689673784,
      "tokens_seen": 858914816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736108324974925,
      "loss": 2.9014,
      "theoretical_loss": 3.703119247361667,
      "tokens_seen": 858980352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037360080240722164,
      "loss": 2.8771,
      "theoretical_loss": 3.7030916006855374,
      "tokens_seen": 859045888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735907723169509,
      "loss": 2.8042,
      "theoretical_loss": 3.7030639567089825,
      "tokens_seen": 859111424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037358074222668006,
      "loss": 2.6583,
      "theoretical_loss": 3.7030363154315316,
      "tokens_seen": 859176960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037357071213640924,
      "loss": 2.6478,
      "theoretical_loss": 3.7030086768527166,
      "tokens_seen": 859242496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735606820461384,
      "loss": 2.5504,
      "theoretical_loss": 3.702981040972067,
      "tokens_seen": 859308032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735506519558676,
      "loss": 2.7879,
      "theoretical_loss": 3.7029534077891144,
      "tokens_seen": 859373568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735406218655968,
      "loss": 2.9783,
      "theoretical_loss": 3.70292577730339,
      "tokens_seen": 859439104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373530591775326,
      "loss": 2.9209,
      "theoretical_loss": 3.702898149514424,
      "tokens_seen": 859504640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037352056168505514,
      "loss": 2.592,
      "theoretical_loss": 3.702870524421748,
      "tokens_seen": 859570176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735105315947844,
      "loss": 2.816,
      "theoretical_loss": 3.7028429020248934,
      "tokens_seen": 859635712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037350050150451356,
      "loss": 2.4548,
      "theoretical_loss": 3.7028152823233924,
      "tokens_seen": 859701248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037349047141424274,
      "loss": 2.8427,
      "theoretical_loss": 3.702787665316775,
      "tokens_seen": 859766784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734804413239719,
      "loss": 2.6741,
      "theoretical_loss": 3.7027600510045735,
      "tokens_seen": 859832320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734704112337011,
      "loss": 2.844,
      "theoretical_loss": 3.7027324393863204,
      "tokens_seen": 859897856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734603811434303,
      "loss": 2.8788,
      "theoretical_loss": 3.702704830461547,
      "tokens_seen": 859963392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734503510531595,
      "loss": 2.5995,
      "theoretical_loss": 3.702677224229785,
      "tokens_seen": 860028928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037344032096288865,
      "loss": 2.7775,
      "theoretical_loss": 3.702649620690567,
      "tokens_seen": 860094464
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 482749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.050727367401123,
      "objective/train/theoretical_loss": 3.7026220198434254,
      "objective/train/tokens_used": 880620000,
      "theoretical_loss": 3.7026220198434254,
      "tokens_seen": 860160000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734302908726179,
      "loss": 2.7565,
      "theoretical_loss": 3.7026220198434254,
      "tokens_seen": 860160000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373420260782347,
      "loss": 2.5725,
      "theoretical_loss": 3.702594421687892,
      "tokens_seen": 860225536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037341023069207625,
      "loss": 2.5902,
      "theoretical_loss": 3.7025668262234994,
      "tokens_seen": 860291072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734002006018054,
      "loss": 2.8756,
      "theoretical_loss": 3.702539233449781,
      "tokens_seen": 860356608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733901705115346,
      "loss": 2.4553,
      "theoretical_loss": 3.702511643366268,
      "tokens_seen": 860422144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733801404212638,
      "loss": 2.7468,
      "theoretical_loss": 3.702484055972495,
      "tokens_seen": 860487680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037337011033099297,
      "loss": 2.7126,
      "theoretical_loss": 3.702456471267994,
      "tokens_seen": 860553216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037336008024072215,
      "loss": 2.7523,
      "theoretical_loss": 3.7024288892522983,
      "tokens_seen": 860618752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733500501504514,
      "loss": 2.6144,
      "theoretical_loss": 3.702401309924941,
      "tokens_seen": 860684288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733400200601805,
      "loss": 2.5013,
      "theoretical_loss": 3.702373733285455,
      "tokens_seen": 860749824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037332998996990975,
      "loss": 2.5324,
      "theoretical_loss": 3.7023461593333744,
      "tokens_seen": 860815360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037331995987963893,
      "loss": 2.6501,
      "theoretical_loss": 3.7023185880682323,
      "tokens_seen": 860880896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733099297893681,
      "loss": 2.8881,
      "theoretical_loss": 3.702291019489562,
      "tokens_seen": 860946432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732998996990973,
      "loss": 2.7107,
      "theoretical_loss": 3.7022634535968986,
      "tokens_seen": 861011968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732898696088265,
      "loss": 2.8248,
      "theoretical_loss": 3.7022358903897747,
      "tokens_seen": 861077504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037327983951855565,
      "loss": 2.6673,
      "theoretical_loss": 3.702208329867725,
      "tokens_seen": 861143040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732698094282849,
      "loss": 2.7935,
      "theoretical_loss": 3.7021807720302835,
      "tokens_seen": 861208576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373259779338014,
      "loss": 2.6674,
      "theoretical_loss": 3.702153216876984,
      "tokens_seen": 861274112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037324974924774325,
      "loss": 2.7553,
      "theoretical_loss": 3.702125664407361,
      "tokens_seen": 861339648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732397191574724,
      "loss": 2.7629,
      "theoretical_loss": 3.7020981146209495,
      "tokens_seen": 861405184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732296890672016,
      "loss": 2.6363,
      "theoretical_loss": 3.7020705675172834,
      "tokens_seen": 861470720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732196589769308,
      "loss": 2.9832,
      "theoretical_loss": 3.702043023095898,
      "tokens_seen": 861536256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037320962888666,
      "loss": 2.7581,
      "theoretical_loss": 3.7020154813563275,
      "tokens_seen": 861601792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037319959879638916,
      "loss": 2.6392,
      "theoretical_loss": 3.7019879422981075,
      "tokens_seen": 861667328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037318956870611834,
      "loss": 2.7417,
      "theoretical_loss": 3.7019604059207727,
      "tokens_seen": 861732864
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 483414,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.466357707977295,
      "objective/train/theoretical_loss": 3.701932872223858,
      "objective/train/tokens_used": 882258400,
      "theoretical_loss": 3.701932872223858,
      "tokens_seen": 861798400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731795386158475,
      "loss": 2.7092,
      "theoretical_loss": 3.701932872223858,
      "tokens_seen": 861798400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037316950852557675,
      "loss": 2.5774,
      "theoretical_loss": 3.701905341206899,
      "tokens_seen": 861863936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731594784353059,
      "loss": 2.5797,
      "theoretical_loss": 3.701877812869431,
      "tokens_seen": 861929472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731494483450351,
      "loss": 2.4721,
      "theoretical_loss": 3.70185028721099,
      "tokens_seen": 861995008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731394182547643,
      "loss": 2.6781,
      "theoretical_loss": 3.7018227642311112,
      "tokens_seen": 862060544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731293881644935,
      "loss": 2.6883,
      "theoretical_loss": 3.70179524392933,
      "tokens_seen": 862126080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731193580742227,
      "loss": 2.7031,
      "theoretical_loss": 3.7017677263051825,
      "tokens_seen": 862191616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037310932798395184,
      "loss": 2.7407,
      "theoretical_loss": 3.701740211358205,
      "tokens_seen": 862257152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730992978936811,
      "loss": 2.8633,
      "theoretical_loss": 3.7017126990879334,
      "tokens_seen": 862322688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037308926780341026,
      "loss": 2.856,
      "theoretical_loss": 3.7016851894939036,
      "tokens_seen": 862388224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037307923771313944,
      "loss": 2.7041,
      "theoretical_loss": 3.701657682575653,
      "tokens_seen": 862453760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730692076228686,
      "loss": 2.8978,
      "theoretical_loss": 3.7016301783327163,
      "tokens_seen": 862519296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730591775325978,
      "loss": 2.5655,
      "theoretical_loss": 3.7016026767646313,
      "tokens_seen": 862584832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373049147442327,
      "loss": 2.4688,
      "theoretical_loss": 3.7015751778709345,
      "tokens_seen": 862650368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730391173520562,
      "loss": 2.8267,
      "theoretical_loss": 3.701547681651162,
      "tokens_seen": 862715904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037302908726178534,
      "loss": 2.7697,
      "theoretical_loss": 3.701520188104852,
      "tokens_seen": 862781440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730190571715146,
      "loss": 2.7066,
      "theoretical_loss": 3.70149269723154,
      "tokens_seen": 862846976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037300902708124376,
      "loss": 2.8466,
      "theoretical_loss": 3.7014652090307645,
      "tokens_seen": 862912512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037299899699097294,
      "loss": 2.5227,
      "theoretical_loss": 3.701437723502062,
      "tokens_seen": 862978048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729889669007021,
      "loss": 2.7204,
      "theoretical_loss": 3.70141024064497,
      "tokens_seen": 863043584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729789368104313,
      "loss": 2.5038,
      "theoretical_loss": 3.7013827604590253,
      "tokens_seen": 863109120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729689067201605,
      "loss": 2.7124,
      "theoretical_loss": 3.7013552829437675,
      "tokens_seen": 863174656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729588766298897,
      "loss": 2.8171,
      "theoretical_loss": 3.701327808098732,
      "tokens_seen": 863240192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037294884653961885,
      "loss": 2.9852,
      "theoretical_loss": 3.7013003359234578,
      "tokens_seen": 863305728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729388164493481,
      "loss": 2.6671,
      "theoretical_loss": 3.701272866417482,
      "tokens_seen": 863371264
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 484154,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.915800094604492,
      "objective/train/theoretical_loss": 3.701245399580344,
      "objective/train/tokens_used": 883896800,
      "theoretical_loss": 3.701245399580344,
      "tokens_seen": 863436800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729287863590772,
      "loss": 2.7206,
      "theoretical_loss": 3.701245399580344,
      "tokens_seen": 863436800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037291875626880645,
      "loss": 2.7457,
      "theoretical_loss": 3.7012179354115813,
      "tokens_seen": 863502336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729087261785356,
      "loss": 2.8516,
      "theoretical_loss": 3.701190473910732,
      "tokens_seen": 863567872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728986960882648,
      "loss": 2.6364,
      "theoretical_loss": 3.7011630150773343,
      "tokens_seen": 863633408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372888665997994,
      "loss": 2.7432,
      "theoretical_loss": 3.7011355589109276,
      "tokens_seen": 863698944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037287863590772317,
      "loss": 2.6735,
      "theoretical_loss": 3.7011081054110493,
      "tokens_seen": 863764480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037286860581745235,
      "loss": 2.675,
      "theoretical_loss": 3.7010806545772397,
      "tokens_seen": 863830016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728585757271816,
      "loss": 2.8821,
      "theoretical_loss": 3.7010532064090356,
      "tokens_seen": 863895552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728485456369107,
      "loss": 2.5917,
      "theoretical_loss": 3.701025760905978,
      "tokens_seen": 863961088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037283851554663995,
      "loss": 2.761,
      "theoretical_loss": 3.7009983180676045,
      "tokens_seen": 864026624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037282848545636913,
      "loss": 2.8932,
      "theoretical_loss": 3.7009708778934556,
      "tokens_seen": 864092160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728184553660983,
      "loss": 2.8616,
      "theoretical_loss": 3.7009434403830697,
      "tokens_seen": 864157696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728084252758275,
      "loss": 2.7249,
      "theoretical_loss": 3.700916005535986,
      "tokens_seen": 864223232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727983951855567,
      "loss": 2.6967,
      "theoretical_loss": 3.7008885733517447,
      "tokens_seen": 864288768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037278836509528585,
      "loss": 2.6364,
      "theoretical_loss": 3.7008611438298855,
      "tokens_seen": 864354304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727783350050151,
      "loss": 2.8586,
      "theoretical_loss": 3.7008337169699477,
      "tokens_seen": 864419840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727683049147442,
      "loss": 2.7462,
      "theoretical_loss": 3.700806292771472,
      "tokens_seen": 864485376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037275827482447345,
      "loss": 2.6446,
      "theoretical_loss": 3.7007788712339966,
      "tokens_seen": 864550912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727482447342026,
      "loss": 2.58,
      "theoretical_loss": 3.700751452357064,
      "tokens_seen": 864616448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727382146439318,
      "loss": 2.6979,
      "theoretical_loss": 3.7007240361402127,
      "tokens_seen": 864681984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372728184553661,
      "loss": 2.7701,
      "theoretical_loss": 3.700696622582984,
      "tokens_seen": 864747520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727181544633902,
      "loss": 2.7526,
      "theoretical_loss": 3.7006692116849176,
      "tokens_seen": 864813056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037270812437311936,
      "loss": 2.7688,
      "theoretical_loss": 3.7006418034455546,
      "tokens_seen": 864878592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037269809428284854,
      "loss": 2.9445,
      "theoretical_loss": 3.7006143978644355,
      "tokens_seen": 864944128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726880641925777,
      "loss": 2.7108,
      "theoretical_loss": 3.7005869949411014,
      "tokens_seen": 865009664
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 484565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.951662063598633,
      "objective/train/theoretical_loss": 3.7005595946750924,
      "objective/train/tokens_used": 885535200,
      "theoretical_loss": 3.7005595946750924,
      "tokens_seen": 865075200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037267803410230696,
      "loss": 2.7365,
      "theoretical_loss": 3.7005595946750924,
      "tokens_seen": 865075200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726680040120361,
      "loss": 2.5957,
      "theoretical_loss": 3.70053219706595,
      "tokens_seen": 865140736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726579739217653,
      "loss": 2.8784,
      "theoretical_loss": 3.7005048021132163,
      "tokens_seen": 865206272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726479438314945,
      "loss": 2.6853,
      "theoretical_loss": 3.700477409816431,
      "tokens_seen": 865271808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726379137412237,
      "loss": 3.0309,
      "theoretical_loss": 3.7004500201751362,
      "tokens_seen": 865337344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037262788365095286,
      "loss": 2.9844,
      "theoretical_loss": 3.7004226331888734,
      "tokens_seen": 865402880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037261785356068204,
      "loss": 2.457,
      "theoretical_loss": 3.700395248857184,
      "tokens_seen": 865468416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726078234704112,
      "loss": 2.9597,
      "theoretical_loss": 3.7003678671796103,
      "tokens_seen": 865533952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037259779338014046,
      "loss": 2.7733,
      "theoretical_loss": 3.7003404881556934,
      "tokens_seen": 865599488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725877632898696,
      "loss": 2.6146,
      "theoretical_loss": 3.7003131117849755,
      "tokens_seen": 865665024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725777331995988,
      "loss": 2.8385,
      "theoretical_loss": 3.7002857380669987,
      "tokens_seen": 865730560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037256770310932795,
      "loss": 2.5251,
      "theoretical_loss": 3.7002583670013047,
      "tokens_seen": 865796096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725576730190572,
      "loss": 2.7264,
      "theoretical_loss": 3.700230998587437,
      "tokens_seen": 865861632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037254764292878636,
      "loss": 2.718,
      "theoretical_loss": 3.7002036328249366,
      "tokens_seen": 865927168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037253761283851554,
      "loss": 2.9428,
      "theoretical_loss": 3.7001762697133467,
      "tokens_seen": 865992704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725275827482447,
      "loss": 2.7134,
      "theoretical_loss": 3.70014890925221,
      "tokens_seen": 866058240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037251755265797396,
      "loss": 2.866,
      "theoretical_loss": 3.700121551441069,
      "tokens_seen": 866123776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725075225677031,
      "loss": 2.5316,
      "theoretical_loss": 3.7000941962794673,
      "tokens_seen": 866189312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724974924774323,
      "loss": 2.955,
      "theoretical_loss": 3.7000668437669466,
      "tokens_seen": 866254848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037248746238716145,
      "loss": 2.7666,
      "theoretical_loss": 3.7000394939030503,
      "tokens_seen": 866320384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724774322968907,
      "loss": 2.803,
      "theoretical_loss": 3.7000121466873224,
      "tokens_seen": 866385920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037246740220661987,
      "loss": 2.553,
      "theoretical_loss": 3.6999848021193054,
      "tokens_seen": 866451456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037245737211634905,
      "loss": 2.7715,
      "theoretical_loss": 3.6999574601985428,
      "tokens_seen": 866516992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037244734202607823,
      "loss": 2.645,
      "theoretical_loss": 3.699930120924579,
      "tokens_seen": 866582528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724373119358074,
      "loss": 2.8014,
      "theoretical_loss": 3.699902784296956,
      "tokens_seen": 866648064
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 485783,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.643681049346924,
      "objective/train/theoretical_loss": 3.6998754503152194,
      "objective/train/tokens_used": 887173600,
      "theoretical_loss": 3.6998754503152194,
      "tokens_seen": 866713600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724272818455366,
      "loss": 2.8837,
      "theoretical_loss": 3.6998754503152194,
      "tokens_seen": 866713600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724172517552658,
      "loss": 2.6677,
      "theoretical_loss": 3.6998481189789114,
      "tokens_seen": 866779136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037240722166499495,
      "loss": 2.7465,
      "theoretical_loss": 3.6998207902875775,
      "tokens_seen": 866844672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723971915747242,
      "loss": 2.8192,
      "theoretical_loss": 3.6997934642407606,
      "tokens_seen": 866910208
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723871614844533,
      "loss": 2.8605,
      "theoretical_loss": 3.6997661408380056,
      "tokens_seen": 866975744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037237713139418255,
      "loss": 2.7537,
      "theoretical_loss": 3.6997388200788563,
      "tokens_seen": 867041280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723671013039118,
      "loss": 2.7422,
      "theoretical_loss": 3.6997115019628577,
      "tokens_seen": 867106816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723570712136409,
      "loss": 2.7609,
      "theoretical_loss": 3.699684186489554,
      "tokens_seen": 867172352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037234704112337015,
      "loss": 2.7326,
      "theoretical_loss": 3.6996568736584896,
      "tokens_seen": 867237888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037233701103309933,
      "loss": 2.6956,
      "theoretical_loss": 3.69962956346921,
      "tokens_seen": 867303424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723269809428285,
      "loss": 2.6659,
      "theoretical_loss": 3.69960225592126,
      "tokens_seen": 867368960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723169508525577,
      "loss": 2.8015,
      "theoretical_loss": 3.699574951014183,
      "tokens_seen": 867434496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723069207622869,
      "loss": 2.6047,
      "theoretical_loss": 3.6995476487475267,
      "tokens_seen": 867500032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037229689067201605,
      "loss": 2.6313,
      "theoretical_loss": 3.6995203491208346,
      "tokens_seen": 867565568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722868605817453,
      "loss": 2.7724,
      "theoretical_loss": 3.6994930521336524,
      "tokens_seen": 867631104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722768304914744,
      "loss": 2.7818,
      "theoretical_loss": 3.699465757785526,
      "tokens_seen": 867696640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037226680040120365,
      "loss": 2.6688,
      "theoretical_loss": 3.699438466076,
      "tokens_seen": 867762176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722567703109328,
      "loss": 2.6619,
      "theoretical_loss": 3.699411177004621,
      "tokens_seen": 867827712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372246740220662,
      "loss": 2.7528,
      "theoretical_loss": 3.6993838905709344,
      "tokens_seen": 867893248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722367101303912,
      "loss": 2.6117,
      "theoretical_loss": 3.6993566067744856,
      "tokens_seen": 867958784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722266800401204,
      "loss": 2.5744,
      "theoretical_loss": 3.699329325614822,
      "tokens_seen": 868024320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037221664994984956,
      "loss": 2.6486,
      "theoretical_loss": 3.6993020470914884,
      "tokens_seen": 868089856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037220661985957874,
      "loss": 2.8173,
      "theoretical_loss": 3.6992747712040317,
      "tokens_seen": 868155392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721965897693079,
      "loss": 2.6862,
      "theoretical_loss": 3.699247497951998,
      "tokens_seen": 868220928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037218655967903716,
      "loss": 2.7351,
      "theoretical_loss": 3.699220227334934,
      "tokens_seen": 868286464
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 486350,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.766869068145752,
      "objective/train/theoretical_loss": 3.699192959352386,
      "objective/train/tokens_used": 888812000,
      "theoretical_loss": 3.699192959352386,
      "tokens_seen": 868352000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721765295887663,
      "loss": 2.5503,
      "theoretical_loss": 3.699192959352386,
      "tokens_seen": 868352000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721664994984955,
      "loss": 3.0311,
      "theoretical_loss": 3.6991656940039004,
      "tokens_seen": 868417536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721564694082247,
      "loss": 2.6552,
      "theoretical_loss": 3.6991384312890245,
      "tokens_seen": 868483072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721464393179539,
      "loss": 2.4007,
      "theoretical_loss": 3.6991111712073055,
      "tokens_seen": 868548608
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037213640922768306,
      "loss": 2.809,
      "theoretical_loss": 3.69908391375829,
      "tokens_seen": 868614144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037212637913741224,
      "loss": 2.7626,
      "theoretical_loss": 3.699056658941525,
      "tokens_seen": 868679680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721163490471414,
      "loss": 2.519,
      "theoretical_loss": 3.6990294067565577,
      "tokens_seen": 868745216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037210631895687066,
      "loss": 2.9509,
      "theoretical_loss": 3.699002157202936,
      "tokens_seen": 868810752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720962888665998,
      "loss": 2.7953,
      "theoretical_loss": 3.6989749102802065,
      "tokens_seen": 868876288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372086258776329,
      "loss": 2.6473,
      "theoretical_loss": 3.6989476659879177,
      "tokens_seen": 868941824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037207622868605815,
      "loss": 2.6791,
      "theoretical_loss": 3.6989204243256166,
      "tokens_seen": 869007360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720661985957874,
      "loss": 2.9017,
      "theoretical_loss": 3.698893185292852,
      "tokens_seen": 869072896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037205616850551656,
      "loss": 2.7375,
      "theoretical_loss": 3.6988659488891704,
      "tokens_seen": 869138432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037204613841524575,
      "loss": 3.008,
      "theoretical_loss": 3.698838715114121,
      "tokens_seen": 869203968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720361083249749,
      "loss": 2.8997,
      "theoretical_loss": 3.6988114839672512,
      "tokens_seen": 869269504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037202607823470416,
      "loss": 2.8079,
      "theoretical_loss": 3.6987842554481096,
      "tokens_seen": 869335040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720160481444333,
      "loss": 2.7345,
      "theoretical_loss": 3.6987570295562446,
      "tokens_seen": 869400576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720060180541625,
      "loss": 2.7938,
      "theoretical_loss": 3.6987298062912046,
      "tokens_seen": 869466112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037199598796389165,
      "loss": 2.7915,
      "theoretical_loss": 3.6987025856525384,
      "tokens_seen": 869531648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719859578736209,
      "loss": 2.8572,
      "theoretical_loss": 3.6986753676397948,
      "tokens_seen": 869597184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037197592778335007,
      "loss": 2.687,
      "theoretical_loss": 3.698648152252521,
      "tokens_seen": 869662720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037196589769307925,
      "loss": 2.7066,
      "theoretical_loss": 3.6986209394902687,
      "tokens_seen": 869728256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037195586760280843,
      "loss": 2.6519,
      "theoretical_loss": 3.6985937293525843,
      "tokens_seen": 869793792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719458375125376,
      "loss": 2.6958,
      "theoretical_loss": 3.6985665218390187,
      "tokens_seen": 869859328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719358074222668,
      "loss": 2.6778,
      "theoretical_loss": 3.6985393169491205,
      "tokens_seen": 869924864
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 487478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4885756969451904,
      "objective/train/theoretical_loss": 3.698512114682439,
      "objective/train/tokens_used": 890450400,
      "theoretical_loss": 3.698512114682439,
      "tokens_seen": 869990400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371925777331996,
      "loss": 2.6768,
      "theoretical_loss": 3.698512114682439,
      "tokens_seen": 869990400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037191574724172515,
      "loss": 2.6099,
      "theoretical_loss": 3.698484915038524,
      "tokens_seen": 870055936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719057171514544,
      "loss": 2.6162,
      "theoretical_loss": 3.698457718016925,
      "tokens_seen": 870121472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718956870611835,
      "loss": 3.0378,
      "theoretical_loss": 3.6984305236171915,
      "tokens_seen": 870187008
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037188565697091275,
      "loss": 2.4252,
      "theoretical_loss": 3.698403331838873,
      "tokens_seen": 870252544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037187562688064193,
      "loss": 2.5673,
      "theoretical_loss": 3.6983761426815205,
      "tokens_seen": 870318080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718655967903711,
      "loss": 2.865,
      "theoretical_loss": 3.6983489561446827,
      "tokens_seen": 870383616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718555667001003,
      "loss": 2.7507,
      "theoretical_loss": 3.698321772227911,
      "tokens_seen": 870449152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037184553660982953,
      "loss": 2.5034,
      "theoretical_loss": 3.6982945909307547,
      "tokens_seen": 870514688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037183550651955866,
      "loss": 2.5666,
      "theoretical_loss": 3.698267412252765,
      "tokens_seen": 870580224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718254764292879,
      "loss": 2.7432,
      "theoretical_loss": 3.6982402361934916,
      "tokens_seen": 870645760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371815446339017,
      "loss": 2.6505,
      "theoretical_loss": 3.698213062752486,
      "tokens_seen": 870711296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037180541624874625,
      "loss": 2.8117,
      "theoretical_loss": 3.698185891929297,
      "tokens_seen": 870776832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037179538615847544,
      "loss": 2.7201,
      "theoretical_loss": 3.698158723723478,
      "tokens_seen": 870842368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717853560682046,
      "loss": 2.5757,
      "theoretical_loss": 3.6981315581345786,
      "tokens_seen": 870907904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717753259779338,
      "loss": 2.8679,
      "theoretical_loss": 3.698104395162149,
      "tokens_seen": 870973440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371765295887663,
      "loss": 2.5399,
      "theoretical_loss": 3.6980772348057425,
      "tokens_seen": 871038976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037175526579739216,
      "loss": 2.7885,
      "theoretical_loss": 3.6980500770649085,
      "tokens_seen": 871104512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717452357071214,
      "loss": 2.707,
      "theoretical_loss": 3.6980229219391987,
      "tokens_seen": 871170048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717352056168505,
      "loss": 2.6914,
      "theoretical_loss": 3.6979957694281653,
      "tokens_seen": 871235584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037172517552657976,
      "loss": 2.7263,
      "theoretical_loss": 3.697968619531359,
      "tokens_seen": 871301120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717151454363089,
      "loss": 2.5678,
      "theoretical_loss": 3.697941472248332,
      "tokens_seen": 871366656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717051153460381,
      "loss": 2.5981,
      "theoretical_loss": 3.6979143275786366,
      "tokens_seen": 871432192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716950852557673,
      "loss": 2.7973,
      "theoretical_loss": 3.697887185521824,
      "tokens_seen": 871497728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716850551654965,
      "loss": 2.897,
      "theoretical_loss": 3.6978600460774462,
      "tokens_seen": 871563264
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 487872,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4374098777770996,
      "objective/train/theoretical_loss": 3.6978329092450557,
      "objective/train/tokens_used": 892088800,
      "theoretical_loss": 3.6978329092450557,
      "tokens_seen": 871628800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037167502507522566,
      "loss": 2.6415,
      "theoretical_loss": 3.6978329092450557,
      "tokens_seen": 871628800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716649949849549,
      "loss": 2.8264,
      "theoretical_loss": 3.6978057750242046,
      "tokens_seen": 871694336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371654964894684,
      "loss": 2.8088,
      "theoretical_loss": 3.697778643414445,
      "tokens_seen": 871759872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037164493480441326,
      "loss": 2.6045,
      "theoretical_loss": 3.69775151441533,
      "tokens_seen": 871825408
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716349047141424,
      "loss": 2.806,
      "theoretical_loss": 3.697724388026412,
      "tokens_seen": 871890944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716248746238716,
      "loss": 2.804,
      "theoretical_loss": 3.6976972642472425,
      "tokens_seen": 871956480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037161484453360086,
      "loss": 2.7078,
      "theoretical_loss": 3.697670143077376,
      "tokens_seen": 872022016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037160481444333,
      "loss": 2.6623,
      "theoretical_loss": 3.697643024516365,
      "tokens_seen": 872087552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715947843530592,
      "loss": 2.8471,
      "theoretical_loss": 3.697615908563762,
      "tokens_seen": 872153088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037158475426278835,
      "loss": 2.5388,
      "theoretical_loss": 3.6975887952191204,
      "tokens_seen": 872218624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715747241725176,
      "loss": 2.7085,
      "theoretical_loss": 3.697561684481993,
      "tokens_seen": 872284160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037156469408224676,
      "loss": 2.8703,
      "theoretical_loss": 3.6975345763519343,
      "tokens_seen": 872349696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037155466399197595,
      "loss": 2.6017,
      "theoretical_loss": 3.697507470828497,
      "tokens_seen": 872415232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715446339017051,
      "loss": 2.5009,
      "theoretical_loss": 3.6974803679112345,
      "tokens_seen": 872480768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037153460381143436,
      "loss": 2.5021,
      "theoretical_loss": 3.697453267599701,
      "tokens_seen": 872546304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715245737211635,
      "loss": 3.0077,
      "theoretical_loss": 3.6974261698934496,
      "tokens_seen": 872611840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715145436308927,
      "loss": 2.8439,
      "theoretical_loss": 3.6973990747920347,
      "tokens_seen": 872677376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037150451354062185,
      "loss": 2.7113,
      "theoretical_loss": 3.69737198229501,
      "tokens_seen": 872742912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714944834503511,
      "loss": 2.5746,
      "theoretical_loss": 3.6973448924019308,
      "tokens_seen": 872808448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037148445336008027,
      "loss": 2.7075,
      "theoretical_loss": 3.6973178051123496,
      "tokens_seen": 872873984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037147442326980945,
      "loss": 2.5569,
      "theoretical_loss": 3.697290720425822,
      "tokens_seen": 872939520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037146439317953863,
      "loss": 2.6775,
      "theoretical_loss": 3.6972636383419015,
      "tokens_seen": 873005056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714543630892678,
      "loss": 2.7166,
      "theoretical_loss": 3.6972365588601432,
      "tokens_seen": 873070592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371444332998997,
      "loss": 2.6978,
      "theoretical_loss": 3.697209481980102,
      "tokens_seen": 873136128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037143430290872623,
      "loss": 2.7577,
      "theoretical_loss": 3.697182407701332,
      "tokens_seen": 873201664
    },
    {
      "epoch": 0.26,
      "objective/train/docs_used": 488898,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9824957847595215,
      "objective/train/theoretical_loss": 3.697155336023389,
      "objective/train/tokens_used": 893727200,
      "theoretical_loss": 3.697155336023389,
      "tokens_seen": 873267200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037142427281845535,
      "loss": 2.7698,
      "theoretical_loss": 3.697155336023389,
      "tokens_seen": 873267200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714142427281846,
      "loss": 2.9909,
      "theoretical_loss": 3.6971282669458274,
      "tokens_seen": 873332736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714042126379137,
      "loss": 2.651,
      "theoretical_loss": 3.6971012004682025,
      "tokens_seen": 873398272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037139418254764295,
      "loss": 2.7109,
      "theoretical_loss": 3.697074136590069,
      "tokens_seen": 873463808
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037138415245737213,
      "loss": 2.8332,
      "theoretical_loss": 3.6970470753109828,
      "tokens_seen": 873529344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713741223671013,
      "loss": 2.7787,
      "theoretical_loss": 3.6970200166304994,
      "tokens_seen": 873594880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713640922768305,
      "loss": 2.9767,
      "theoretical_loss": 3.696992960548174,
      "tokens_seen": 873660416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037135406218655973,
      "loss": 2.8765,
      "theoretical_loss": 3.6969659070635625,
      "tokens_seen": 873725952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037134403209628886,
      "loss": 2.7108,
      "theoretical_loss": 3.696938856176221,
      "tokens_seen": 873791488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713340020060181,
      "loss": 2.7555,
      "theoretical_loss": 3.6969118078857046,
      "tokens_seen": 873857024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713239719157472,
      "loss": 2.7623,
      "theoretical_loss": 3.6968847621915693,
      "tokens_seen": 873922560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037131394182547645,
      "loss": 2.7329,
      "theoretical_loss": 3.696857719093372,
      "tokens_seen": 873988096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037130391173520564,
      "loss": 2.6343,
      "theoretical_loss": 3.6968306785906684,
      "tokens_seen": 874053632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712938816449348,
      "loss": 2.5664,
      "theoretical_loss": 3.6968036406830147,
      "tokens_seen": 874119168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371283851554664,
      "loss": 2.7419,
      "theoretical_loss": 3.6967766053699673,
      "tokens_seen": 874184704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712738214643932,
      "loss": 2.6855,
      "theoretical_loss": 3.6967495726510835,
      "tokens_seen": 874250240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037126379137412236,
      "loss": 2.8358,
      "theoretical_loss": 3.6967225425259187,
      "tokens_seen": 874315776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712537612838516,
      "loss": 2.7927,
      "theoretical_loss": 3.6966955149940306,
      "tokens_seen": 874381312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712437311935807,
      "loss": 2.6235,
      "theoretical_loss": 3.696668490054976,
      "tokens_seen": 874446848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037123370110330996,
      "loss": 2.5521,
      "theoretical_loss": 3.6966414677083113,
      "tokens_seen": 874512384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003712236710130391,
      "loss": 2.936,
      "theoretical_loss": 3.696614447953594,
      "tokens_seen": 874577920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003712136409227683,
      "loss": 2.5652,
      "theoretical_loss": 3.696587430790381,
      "tokens_seen": 874643456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003712036108324975,
      "loss": 2.7471,
      "theoretical_loss": 3.6965604162182304,
      "tokens_seen": 874708992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711935807422267,
      "loss": 2.6699,
      "theoretical_loss": 3.696533404236699,
      "tokens_seen": 874774528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037118355065195586,
      "loss": 2.8556,
      "theoretical_loss": 3.6965063948453434,
      "tokens_seen": 874840064
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 489972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.883964776992798,
      "objective/train/theoretical_loss": 3.6964793880437226,
      "objective/train/tokens_used": 895365600,
      "theoretical_loss": 3.6964793880437226,
      "tokens_seen": 874905600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711735205616851,
      "loss": 2.8779,
      "theoretical_loss": 3.6964793880437226,
      "tokens_seen": 874905600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711634904714142,
      "loss": 2.854,
      "theoretical_loss": 3.696452383831394,
      "tokens_seen": 874971136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037115346038114346,
      "loss": 2.901,
      "theoretical_loss": 3.6964253822079147,
      "tokens_seen": 875036672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711434302908726,
      "loss": 3.0085,
      "theoretical_loss": 3.696398383172844,
      "tokens_seen": 875102208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711334002006018,
      "loss": 2.5525,
      "theoretical_loss": 3.696371386725739,
      "tokens_seen": 875167744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000371123370110331,
      "loss": 2.8206,
      "theoretical_loss": 3.696344392866158,
      "tokens_seen": 875233280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711133400200602,
      "loss": 2.9085,
      "theoretical_loss": 3.6963174015936593,
      "tokens_seen": 875298816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037110330992978937,
      "loss": 2.7336,
      "theoretical_loss": 3.6962904129078016,
      "tokens_seen": 875364352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037109327983951855,
      "loss": 2.5856,
      "theoretical_loss": 3.6962634268081427,
      "tokens_seen": 875429888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037108324974924773,
      "loss": 2.5817,
      "theoretical_loss": 3.6962364432942416,
      "tokens_seen": 875495424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037107321965897696,
      "loss": 2.808,
      "theoretical_loss": 3.6962094623656574,
      "tokens_seen": 875560960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710631895687061,
      "loss": 2.8817,
      "theoretical_loss": 3.696182484021948,
      "tokens_seen": 875626496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710531594784353,
      "loss": 2.7093,
      "theoretical_loss": 3.696155508262673,
      "tokens_seen": 875692032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037104312938816445,
      "loss": 2.671,
      "theoretical_loss": 3.6961285350873916,
      "tokens_seen": 875757568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710330992978937,
      "loss": 2.6938,
      "theoretical_loss": 3.6961015644956623,
      "tokens_seen": 875823104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037102306920762287,
      "loss": 2.7216,
      "theoretical_loss": 3.696074596487045,
      "tokens_seen": 875888640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037101303911735205,
      "loss": 2.5135,
      "theoretical_loss": 3.6960476310610986,
      "tokens_seen": 875954176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037100300902708123,
      "loss": 2.663,
      "theoretical_loss": 3.696020668217382,
      "tokens_seen": 876019712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037099297893681047,
      "loss": 2.7203,
      "theoretical_loss": 3.6959937079554566,
      "tokens_seen": 876085248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709829488465396,
      "loss": 2.7627,
      "theoretical_loss": 3.69596675027488,
      "tokens_seen": 876150784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037097291875626883,
      "loss": 2.8865,
      "theoretical_loss": 3.6959397951752133,
      "tokens_seen": 876216320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037096288866599796,
      "loss": 2.6588,
      "theoretical_loss": 3.695912842656016,
      "tokens_seen": 876281856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709528585757272,
      "loss": 2.8049,
      "theoretical_loss": 3.695885892716848,
      "tokens_seen": 876347392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709428284854564,
      "loss": 2.7291,
      "theoretical_loss": 3.6958589453572697,
      "tokens_seen": 876412928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037093279839518555,
      "loss": 2.826,
      "theoretical_loss": 3.695832000576841,
      "tokens_seen": 876478464
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 490667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.061556816101074,
      "objective/train/theoretical_loss": 3.6958050583751225,
      "objective/train/tokens_used": 897004000,
      "theoretical_loss": 3.6958050583751225,
      "tokens_seen": 876544000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037092276830491474,
      "loss": 2.8808,
      "theoretical_loss": 3.6958050583751225,
      "tokens_seen": 876544000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709127382146439,
      "loss": 2.7035,
      "theoretical_loss": 3.6957781187516745,
      "tokens_seen": 876609536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709027081243731,
      "loss": 2.8415,
      "theoretical_loss": 3.6957511817060578,
      "tokens_seen": 876675072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037089267803410233,
      "loss": 2.6895,
      "theoretical_loss": 3.695724247237832,
      "tokens_seen": 876740608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037088264794383146,
      "loss": 2.6108,
      "theoretical_loss": 3.6956973153465595,
      "tokens_seen": 876806144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708726178535607,
      "loss": 2.8491,
      "theoretical_loss": 3.6956703860318,
      "tokens_seen": 876871680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037086258776328993,
      "loss": 2.8227,
      "theoretical_loss": 3.695643459293115,
      "tokens_seen": 876937216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037085255767301906,
      "loss": 2.6601,
      "theoretical_loss": 3.695616535130065,
      "tokens_seen": 877002752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708425275827483,
      "loss": 2.7217,
      "theoretical_loss": 3.6955896135422117,
      "tokens_seen": 877068288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708324974924774,
      "loss": 2.8116,
      "theoretical_loss": 3.6955626945291167,
      "tokens_seen": 877133824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037082246740220665,
      "loss": 2.6787,
      "theoretical_loss": 3.6955357780903406,
      "tokens_seen": 877199360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037081243731193584,
      "loss": 2.5873,
      "theoretical_loss": 3.6955088642254452,
      "tokens_seen": 877264896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370802407221665,
      "loss": 2.8596,
      "theoretical_loss": 3.6954819529339926,
      "tokens_seen": 877330432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707923771313942,
      "loss": 2.6279,
      "theoretical_loss": 3.695455044215544,
      "tokens_seen": 877395968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707823470411234,
      "loss": 2.6978,
      "theoretical_loss": 3.695428138069661,
      "tokens_seen": 877461504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037077231695085256,
      "loss": 2.6799,
      "theoretical_loss": 3.6954012344959066,
      "tokens_seen": 877527040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707622868605818,
      "loss": 2.7125,
      "theoretical_loss": 3.6953743334938416,
      "tokens_seen": 877592576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707522567703109,
      "loss": 2.7077,
      "theoretical_loss": 3.6953474350630287,
      "tokens_seen": 877658112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037074222668004016,
      "loss": 2.81,
      "theoretical_loss": 3.6953205392030304,
      "tokens_seen": 877723648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707321965897693,
      "loss": 2.6282,
      "theoretical_loss": 3.6952936459134085,
      "tokens_seen": 877789184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707221664994985,
      "loss": 2.9963,
      "theoretical_loss": 3.6952667551937264,
      "tokens_seen": 877854720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707121364092277,
      "loss": 2.6417,
      "theoretical_loss": 3.695239867043546,
      "tokens_seen": 877920256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707021063189569,
      "loss": 2.889,
      "theoretical_loss": 3.6952129814624297,
      "tokens_seen": 877985792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037069207622868606,
      "loss": 2.7111,
      "theoretical_loss": 3.695186098449941,
      "tokens_seen": 878051328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706820461384153,
      "loss": 2.7687,
      "theoretical_loss": 3.6951592180056423,
      "tokens_seen": 878116864
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 491863,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.594583511352539,
      "objective/train/theoretical_loss": 3.6951323401290974,
      "objective/train/tokens_used": 898642400,
      "theoretical_loss": 3.6951323401290974,
      "tokens_seen": 878182400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706720160481444,
      "loss": 2.6835,
      "theoretical_loss": 3.6951323401290974,
      "tokens_seen": 878182400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037066198595787366,
      "loss": 2.8534,
      "theoretical_loss": 3.695105464819868,
      "tokens_seen": 878247936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706519558676028,
      "loss": 2.758,
      "theoretical_loss": 3.6950785920775187,
      "tokens_seen": 878313472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370641925777332,
      "loss": 2.9426,
      "theoretical_loss": 3.6950517219016117,
      "tokens_seen": 878379008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706318956870612,
      "loss": 2.8975,
      "theoretical_loss": 3.695024854291712,
      "tokens_seen": 878444544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706218655967904,
      "loss": 2.549,
      "theoretical_loss": 3.6949979892473817,
      "tokens_seen": 878510080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037061183550651957,
      "loss": 2.9225,
      "theoretical_loss": 3.694971126768185,
      "tokens_seen": 878575616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037060180541624875,
      "loss": 2.5897,
      "theoretical_loss": 3.6949442668536854,
      "tokens_seen": 878641152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037059177532597793,
      "loss": 2.8333,
      "theoretical_loss": 3.694917409503447,
      "tokens_seen": 878706688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037058174523570716,
      "loss": 3.032,
      "theoretical_loss": 3.6948905547170336,
      "tokens_seen": 878772224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705717151454363,
      "loss": 2.8523,
      "theoretical_loss": 3.6948637024940094,
      "tokens_seen": 878837760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705616850551655,
      "loss": 2.8115,
      "theoretical_loss": 3.694836852833939,
      "tokens_seen": 878903296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037055165496489465,
      "loss": 2.618,
      "theoretical_loss": 3.694810005736386,
      "tokens_seen": 878968832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705416248746239,
      "loss": 2.7893,
      "theoretical_loss": 3.694783161200916,
      "tokens_seen": 879034368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037053159478435307,
      "loss": 2.8638,
      "theoretical_loss": 3.6947563192270914,
      "tokens_seen": 879099904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037052156469408225,
      "loss": 2.5287,
      "theoretical_loss": 3.6947294798144785,
      "tokens_seen": 879165440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037051153460381143,
      "loss": 2.7751,
      "theoretical_loss": 3.6947026429626417,
      "tokens_seen": 879230976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037050150451354067,
      "loss": 2.7179,
      "theoretical_loss": 3.6946758086711453,
      "tokens_seen": 879296512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704914744232698,
      "loss": 2.8058,
      "theoretical_loss": 3.694648976939555,
      "tokens_seen": 879362048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037048144433299903,
      "loss": 2.7717,
      "theoretical_loss": 3.694622147767435,
      "tokens_seen": 879427584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037047141424272816,
      "loss": 2.6822,
      "theoretical_loss": 3.694595321154351,
      "tokens_seen": 879493120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704613841524574,
      "loss": 2.5861,
      "theoretical_loss": 3.6945684970998682,
      "tokens_seen": 879558656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704513540621866,
      "loss": 2.6915,
      "theoretical_loss": 3.694541675603552,
      "tokens_seen": 879624192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037044132397191575,
      "loss": 2.7335,
      "theoretical_loss": 3.6945148566649677,
      "tokens_seen": 879689728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037043129388164494,
      "loss": 2.8133,
      "theoretical_loss": 3.694488040283681,
      "tokens_seen": 879755264
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 492132,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4907939434051514,
      "objective/train/theoretical_loss": 3.6944612264592576,
      "objective/train/tokens_used": 900280800,
      "theoretical_loss": 3.6944612264592576,
      "tokens_seen": 879820800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704212637913741,
      "loss": 2.7163,
      "theoretical_loss": 3.6944612264592576,
      "tokens_seen": 879820800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704112337011033,
      "loss": 2.9041,
      "theoretical_loss": 3.694434415191263,
      "tokens_seen": 879886336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037040120361083253,
      "loss": 2.8063,
      "theoretical_loss": 3.6944076064792633,
      "tokens_seen": 879951872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037039117352056166,
      "loss": 2.7982,
      "theoretical_loss": 3.6943808003228247,
      "tokens_seen": 880017408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703811434302909,
      "loss": 2.8089,
      "theoretical_loss": 3.694353996721513,
      "tokens_seen": 880082944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703711133400201,
      "loss": 2.5928,
      "theoretical_loss": 3.694327195674894,
      "tokens_seen": 880148480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037036108324974926,
      "loss": 2.8074,
      "theoretical_loss": 3.694300397182535,
      "tokens_seen": 880214016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037035105315947844,
      "loss": 2.8316,
      "theoretical_loss": 3.6942736012440016,
      "tokens_seen": 880279552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703410230692076,
      "loss": 2.7329,
      "theoretical_loss": 3.6942468078588604,
      "tokens_seen": 880345088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703309929789368,
      "loss": 2.9809,
      "theoretical_loss": 3.6942200170266792,
      "tokens_seen": 880410624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037032096288866604,
      "loss": 2.8692,
      "theoretical_loss": 3.694193228747023,
      "tokens_seen": 880476160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037031093279839516,
      "loss": 2.7912,
      "theoretical_loss": 3.6941664430194594,
      "tokens_seen": 880541696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703009027081244,
      "loss": 2.738,
      "theoretical_loss": 3.6941396598435556,
      "tokens_seen": 880607232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702908726178535,
      "loss": 2.8439,
      "theoretical_loss": 3.6941128792188787,
      "tokens_seen": 880672768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037028084252758276,
      "loss": 2.624,
      "theoretical_loss": 3.6940861011449955,
      "tokens_seen": 880738304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037027081243731194,
      "loss": 2.9418,
      "theoretical_loss": 3.694059325621473,
      "tokens_seen": 880803840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702607823470411,
      "loss": 2.5756,
      "theoretical_loss": 3.6940325526478794,
      "tokens_seen": 880869376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702507522567703,
      "loss": 2.7575,
      "theoretical_loss": 3.6940057822237815,
      "tokens_seen": 880934912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702407221664995,
      "loss": 2.9176,
      "theoretical_loss": 3.6939790143487476,
      "tokens_seen": 881000448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037023069207622867,
      "loss": 2.6867,
      "theoretical_loss": 3.6939522490223444,
      "tokens_seen": 881065984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702206619859579,
      "loss": 2.8547,
      "theoretical_loss": 3.6939254862441397,
      "tokens_seen": 881131520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037021063189568703,
      "loss": 2.504,
      "theoretical_loss": 3.6938987260137024,
      "tokens_seen": 881197056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037020060180541626,
      "loss": 2.8788,
      "theoretical_loss": 3.6938719683306003,
      "tokens_seen": 881262592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037019057171514544,
      "loss": 2.5329,
      "theoretical_loss": 3.6938452131944004,
      "tokens_seen": 881328128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701805416248746,
      "loss": 2.8602,
      "theoretical_loss": 3.6938184606046724,
      "tokens_seen": 881393664
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 493018,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3380496501922607,
      "objective/train/theoretical_loss": 3.6937917105609834,
      "objective/train/tokens_used": 901919200,
      "theoretical_loss": 3.6937917105609834,
      "tokens_seen": 881459200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701705115346038,
      "loss": 2.6497,
      "theoretical_loss": 3.6937917105609834,
      "tokens_seen": 881459200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370160481444333,
      "loss": 2.9965,
      "theoretical_loss": 3.6937649630629026,
      "tokens_seen": 881524736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037015045135406217,
      "loss": 2.7815,
      "theoretical_loss": 3.6937382181099983,
      "tokens_seen": 881590272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701404212637914,
      "loss": 2.5983,
      "theoretical_loss": 3.6937114757018388,
      "tokens_seen": 881655808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037013039117352053,
      "loss": 2.9515,
      "theoretical_loss": 3.6936847358379934,
      "tokens_seen": 881721344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037012036108324977,
      "loss": 2.7363,
      "theoretical_loss": 3.693657998518031,
      "tokens_seen": 881786880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037011033099297895,
      "loss": 2.8001,
      "theoretical_loss": 3.6936312637415196,
      "tokens_seen": 881852416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037010030090270813,
      "loss": 2.7913,
      "theoretical_loss": 3.6936045315080293,
      "tokens_seen": 881917952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037009027081243736,
      "loss": 2.5676,
      "theoretical_loss": 3.6935778018171286,
      "tokens_seen": 881983488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700802407221665,
      "loss": 2.8248,
      "theoretical_loss": 3.6935510746683873,
      "tokens_seen": 882049024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700702106318957,
      "loss": 2.7201,
      "theoretical_loss": 3.693524350061374,
      "tokens_seen": 882114560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037006018054162485,
      "loss": 2.7187,
      "theoretical_loss": 3.693497627995659,
      "tokens_seen": 882180096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700501504513541,
      "loss": 2.9054,
      "theoretical_loss": 3.693470908470812,
      "tokens_seen": 882245632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037004012036108327,
      "loss": 2.4963,
      "theoretical_loss": 3.6934441914864014,
      "tokens_seen": 882311168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037003009027081245,
      "loss": 2.6246,
      "theoretical_loss": 3.693417477041998,
      "tokens_seen": 882376704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037002006018054163,
      "loss": 2.7817,
      "theoretical_loss": 3.693390765137172,
      "tokens_seen": 882442240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037001003009027087,
      "loss": 2.7273,
      "theoretical_loss": 3.693364055771492,
      "tokens_seen": 882507776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037,
      "loss": 2.8035,
      "theoretical_loss": 3.69333734894453,
      "tokens_seen": 882573312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036998996990972923,
      "loss": 2.9529,
      "theoretical_loss": 3.693310644655855,
      "tokens_seen": 882638848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036997993981945836,
      "loss": 2.8309,
      "theoretical_loss": 3.693283942905037,
      "tokens_seen": 882704384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699699097291876,
      "loss": 3.0268,
      "theoretical_loss": 3.6932572436916473,
      "tokens_seen": 882769920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699598796389168,
      "loss": 2.9085,
      "theoretical_loss": 3.6932305470152564,
      "tokens_seen": 882835456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036994984954864595,
      "loss": 2.8015,
      "theoretical_loss": 3.693203852875434,
      "tokens_seen": 882900992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036993981945837514,
      "loss": 2.5191,
      "theoretical_loss": 3.693177161271752,
      "tokens_seen": 882966528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699297893681043,
      "loss": 2.8193,
      "theoretical_loss": 3.6931504722037802,
      "tokens_seen": 883032064
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 493587,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.08097767829895,
      "objective/train/theoretical_loss": 3.6931237856710903,
      "objective/train/tokens_used": 903557600,
      "theoretical_loss": 3.6931237856710903,
      "tokens_seen": 883097600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003699197592778335,
      "loss": 2.993,
      "theoretical_loss": 3.6931237856710903,
      "tokens_seen": 883097600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036990972918756273,
      "loss": 2.6583,
      "theoretical_loss": 3.693097101673253,
      "tokens_seen": 883163136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036989969909729186,
      "loss": 2.7407,
      "theoretical_loss": 3.6930704202098394,
      "tokens_seen": 883228672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698896690070211,
      "loss": 2.965,
      "theoretical_loss": 3.6930437412804213,
      "tokens_seen": 883294208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698796389167503,
      "loss": 2.8227,
      "theoretical_loss": 3.693017064884569,
      "tokens_seen": 883359744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036986960882647946,
      "loss": 2.8113,
      "theoretical_loss": 3.692990391021855,
      "tokens_seen": 883425280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036985957873620864,
      "loss": 2.77,
      "theoretical_loss": 3.69296371969185,
      "tokens_seen": 883490816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698495486459378,
      "loss": 2.7267,
      "theoretical_loss": 3.692937050894127,
      "tokens_seen": 883556352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369839518555667,
      "loss": 2.5705,
      "theoretical_loss": 3.692910384628256,
      "tokens_seen": 883621888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036982948846539624,
      "loss": 2.9653,
      "theoretical_loss": 3.69288372089381,
      "tokens_seen": 883687424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036981945837512536,
      "loss": 2.7199,
      "theoretical_loss": 3.6928570596903607,
      "tokens_seen": 883752960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698094282848546,
      "loss": 2.7355,
      "theoretical_loss": 3.6928304010174804,
      "tokens_seen": 883818496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697993981945837,
      "loss": 2.5483,
      "theoretical_loss": 3.692803744874741,
      "tokens_seen": 883884032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036978936810431296,
      "loss": 2.8324,
      "theoretical_loss": 3.692777091261715,
      "tokens_seen": 883949568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036977933801404214,
      "loss": 2.5926,
      "theoretical_loss": 3.692750440177975,
      "tokens_seen": 884015104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697693079237713,
      "loss": 2.809,
      "theoretical_loss": 3.6927237916230933,
      "tokens_seen": 884080640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697592778335005,
      "loss": 2.8743,
      "theoretical_loss": 3.692697145596642,
      "tokens_seen": 884146176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697492477432297,
      "loss": 2.6491,
      "theoretical_loss": 3.6926705020981943,
      "tokens_seen": 884211712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036973921765295887,
      "loss": 2.8035,
      "theoretical_loss": 3.692643861127323,
      "tokens_seen": 884277248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697291875626881,
      "loss": 2.9265,
      "theoretical_loss": 3.692617222683601,
      "tokens_seen": 884342784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036971915747241723,
      "loss": 2.7762,
      "theoretical_loss": 3.6925905867666016,
      "tokens_seen": 884408320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036970912738214646,
      "loss": 2.8945,
      "theoretical_loss": 3.6925639533758976,
      "tokens_seen": 884473856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036969909729187565,
      "loss": 2.9008,
      "theoretical_loss": 3.692537322511062,
      "tokens_seen": 884539392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696890672016048,
      "loss": 2.7927,
      "theoretical_loss": 3.6925106941716686,
      "tokens_seen": 884604928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369679037111334,
      "loss": 2.5308,
      "theoretical_loss": 3.692484068357291,
      "tokens_seen": 884670464
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 494919,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.518890857696533,
      "objective/train/theoretical_loss": 3.692457445067501,
      "objective/train/tokens_used": 905196000,
      "theoretical_loss": 3.692457445067501,
      "tokens_seen": 884736000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696690070210632,
      "loss": 2.6387,
      "theoretical_loss": 3.692457445067501,
      "tokens_seen": 884736000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036965897693079237,
      "loss": 2.8342,
      "theoretical_loss": 3.692430824301875,
      "tokens_seen": 884801536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696489468405216,
      "loss": 2.6288,
      "theoretical_loss": 3.692404206059985,
      "tokens_seen": 884867072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036963891675025073,
      "loss": 2.7104,
      "theoretical_loss": 3.692377590341405,
      "tokens_seen": 884932608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036962888665997997,
      "loss": 2.8244,
      "theoretical_loss": 3.6923509771457095,
      "tokens_seen": 884998144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696188565697091,
      "loss": 2.8814,
      "theoretical_loss": 3.692324366472472,
      "tokens_seen": 885063680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036960882647943833,
      "loss": 2.9496,
      "theoretical_loss": 3.692297758321267,
      "tokens_seen": 885129216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695987963891675,
      "loss": 2.5732,
      "theoretical_loss": 3.692271152691669,
      "tokens_seen": 885194752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695887662988967,
      "loss": 2.6226,
      "theoretical_loss": 3.6922445495832514,
      "tokens_seen": 885260288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036957873620862587,
      "loss": 2.8412,
      "theoretical_loss": 3.69221794899559,
      "tokens_seen": 885325824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036956870611835505,
      "loss": 2.6286,
      "theoretical_loss": 3.6921913509282582,
      "tokens_seen": 885391360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036955867602808424,
      "loss": 2.7941,
      "theoretical_loss": 3.692164755380831,
      "tokens_seen": 885456896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036954864593781347,
      "loss": 2.7921,
      "theoretical_loss": 3.692138162352884,
      "tokens_seen": 885522432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695386158475426,
      "loss": 2.793,
      "theoretical_loss": 3.6921115718439905,
      "tokens_seen": 885587968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036952858575727183,
      "loss": 2.7073,
      "theoretical_loss": 3.6920849838537273,
      "tokens_seen": 885653504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369518555667001,
      "loss": 2.8973,
      "theoretical_loss": 3.6920583983816684,
      "tokens_seen": 885719040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695085255767302,
      "loss": 2.797,
      "theoretical_loss": 3.6920318154273892,
      "tokens_seen": 885784576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694984954864594,
      "loss": 2.8963,
      "theoretical_loss": 3.6920052349904653,
      "tokens_seen": 885850112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036948846539618856,
      "loss": 2.6427,
      "theoretical_loss": 3.691978657070471,
      "tokens_seen": 885915648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036947843530591774,
      "loss": 2.7269,
      "theoretical_loss": 3.691952081666983,
      "tokens_seen": 885981184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369468405215647,
      "loss": 2.745,
      "theoretical_loss": 3.691925508779577,
      "tokens_seen": 886046720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694583751253761,
      "loss": 2.5784,
      "theoretical_loss": 3.691898938407828,
      "tokens_seen": 886112256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036944834503510534,
      "loss": 2.7445,
      "theoretical_loss": 3.691872370551312,
      "tokens_seen": 886177792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036943831494483446,
      "loss": 2.6901,
      "theoretical_loss": 3.6918458052096046,
      "tokens_seen": 886243328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694282848545637,
      "loss": 2.8246,
      "theoretical_loss": 3.691819242382283,
      "tokens_seen": 886308864
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 496242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7018520832061768,
      "objective/train/theoretical_loss": 3.6917926820689217,
      "objective/train/tokens_used": 906834400,
      "theoretical_loss": 3.6917926820689217,
      "tokens_seen": 886374400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694182547642929,
      "loss": 2.6236,
      "theoretical_loss": 3.6917926820689217,
      "tokens_seen": 886374400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036940822467402206,
      "loss": 2.6371,
      "theoretical_loss": 3.6917661242690984,
      "tokens_seen": 886439936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036939819458375124,
      "loss": 2.8313,
      "theoretical_loss": 3.691739568982388,
      "tokens_seen": 886505472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693881644934805,
      "loss": 2.8547,
      "theoretical_loss": 3.691713016208368,
      "tokens_seen": 886571008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693781344032096,
      "loss": 2.9052,
      "theoretical_loss": 3.691686465946615,
      "tokens_seen": 886636544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036936810431293884,
      "loss": 2.8262,
      "theoretical_loss": 3.691659918196705,
      "tokens_seen": 886702080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369358074222668,
      "loss": 2.5036,
      "theoretical_loss": 3.691633372958215,
      "tokens_seen": 886767616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693480441323972,
      "loss": 2.7461,
      "theoretical_loss": 3.6916068302307217,
      "tokens_seen": 886833152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036933801404212644,
      "loss": 2.6719,
      "theoretical_loss": 3.6915802900138024,
      "tokens_seen": 886898688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036932798395185556,
      "loss": 2.8203,
      "theoretical_loss": 3.691553752307034,
      "tokens_seen": 886964224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693179538615848,
      "loss": 2.7727,
      "theoretical_loss": 3.6915272171099938,
      "tokens_seen": 887029760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693079237713139,
      "loss": 2.5439,
      "theoretical_loss": 3.691500684422258,
      "tokens_seen": 887095296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036929789368104316,
      "loss": 2.6267,
      "theoretical_loss": 3.691474154243405,
      "tokens_seen": 887160832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036928786359077234,
      "loss": 2.8903,
      "theoretical_loss": 3.6914476265730123,
      "tokens_seen": 887226368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692778335005015,
      "loss": 2.8947,
      "theoretical_loss": 3.6914211014106573,
      "tokens_seen": 887291904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692678034102307,
      "loss": 3.0139,
      "theoretical_loss": 3.6913945787559177,
      "tokens_seen": 887357440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692577733199599,
      "loss": 2.8974,
      "theoretical_loss": 3.691368058608371,
      "tokens_seen": 887422976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036924774322968907,
      "loss": 2.7185,
      "theoretical_loss": 3.6913415409675947,
      "tokens_seen": 887488512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692377131394183,
      "loss": 2.7825,
      "theoretical_loss": 3.691315025833168,
      "tokens_seen": 887554048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036922768304914743,
      "loss": 2.737,
      "theoretical_loss": 3.6912885132046673,
      "tokens_seen": 887619584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036921765295887666,
      "loss": 2.7198,
      "theoretical_loss": 3.6912620030816723,
      "tokens_seen": 887685120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036920762286860585,
      "loss": 2.7995,
      "theoretical_loss": 3.6912354954637605,
      "tokens_seen": 887750656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369197592778335,
      "loss": 2.732,
      "theoretical_loss": 3.6912089903505105,
      "tokens_seen": 887816192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691875626880642,
      "loss": 2.8471,
      "theoretical_loss": 3.691182487741501,
      "tokens_seen": 887881728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691775325977934,
      "loss": 2.6629,
      "theoretical_loss": 3.69115598763631,
      "tokens_seen": 887947264
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 496686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.581552267074585,
      "objective/train/theoretical_loss": 3.6911294900345166,
      "objective/train/tokens_used": 908472800,
      "theoretical_loss": 3.6911294900345166,
      "tokens_seen": 888012800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036916750250752257,
      "loss": 2.5457,
      "theoretical_loss": 3.6911294900345166,
      "tokens_seen": 888012800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691574724172518,
      "loss": 2.9441,
      "theoretical_loss": 3.691102994935699,
      "tokens_seen": 888078336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036914744232698093,
      "loss": 2.9934,
      "theoretical_loss": 3.6910765023394374,
      "tokens_seen": 888143872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036913741223671017,
      "loss": 2.813,
      "theoretical_loss": 3.6910500122453094,
      "tokens_seen": 888209408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691273821464393,
      "loss": 2.7331,
      "theoretical_loss": 3.691023524652895,
      "tokens_seen": 888274944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036911735205616853,
      "loss": 2.5818,
      "theoretical_loss": 3.6909970395617724,
      "tokens_seen": 888340480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691073219658977,
      "loss": 2.7829,
      "theoretical_loss": 3.690970556971522,
      "tokens_seen": 888406016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690972918756269,
      "loss": 2.603,
      "theoretical_loss": 3.6909440768817223,
      "tokens_seen": 888471552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690872617853561,
      "loss": 2.7331,
      "theoretical_loss": 3.6909175992919536,
      "tokens_seen": 888537088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036907723169508525,
      "loss": 2.5585,
      "theoretical_loss": 3.690891124201795,
      "tokens_seen": 888602624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036906720160481444,
      "loss": 2.7386,
      "theoretical_loss": 3.690864651610826,
      "tokens_seen": 888668160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036905717151454367,
      "loss": 2.7889,
      "theoretical_loss": 3.690838181518627,
      "tokens_seen": 888733696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690471414242728,
      "loss": 2.9249,
      "theoretical_loss": 3.690811713924778,
      "tokens_seen": 888799232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036903711133400203,
      "loss": 2.835,
      "theoretical_loss": 3.690785248828858,
      "tokens_seen": 888864768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690270812437312,
      "loss": 2.5885,
      "theoretical_loss": 3.690758786230448,
      "tokens_seen": 888930304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690170511534604,
      "loss": 2.8022,
      "theoretical_loss": 3.690732326129128,
      "tokens_seen": 888995840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690070210631896,
      "loss": 2.7309,
      "theoretical_loss": 3.690705868524478,
      "tokens_seen": 889061376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036899699097291876,
      "loss": 2.8972,
      "theoretical_loss": 3.6906794134160785,
      "tokens_seen": 889126912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036898696088264794,
      "loss": 2.7127,
      "theoretical_loss": 3.6906529608035106,
      "tokens_seen": 889192448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689769307923772,
      "loss": 2.949,
      "theoretical_loss": 3.6906265106863545,
      "tokens_seen": 889257984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689669007021063,
      "loss": 2.6056,
      "theoretical_loss": 3.690600063064191,
      "tokens_seen": 889323520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036895687061183554,
      "loss": 2.7401,
      "theoretical_loss": 3.6905736179366,
      "tokens_seen": 889389056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036894684052156466,
      "loss": 2.8579,
      "theoretical_loss": 3.690547175303164,
      "tokens_seen": 889454592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689368104312939,
      "loss": 2.7001,
      "theoretical_loss": 3.690520735163463,
      "tokens_seen": 889520128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689267803410231,
      "loss": 2.6667,
      "theoretical_loss": 3.6904942975170787,
      "tokens_seen": 889585664
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 497720,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7590861320495605,
      "objective/train/theoretical_loss": 3.6904678623635916,
      "objective/train/tokens_used": 910111200,
      "theoretical_loss": 3.6904678623635916,
      "tokens_seen": 889651200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036891675025075226,
      "loss": 2.8205,
      "theoretical_loss": 3.6904678623635916,
      "tokens_seen": 889651200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036890672016048144,
      "loss": 2.5425,
      "theoretical_loss": 3.6904414297025836,
      "tokens_seen": 889716736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688966900702107,
      "loss": 2.9111,
      "theoretical_loss": 3.690414999533636,
      "tokens_seen": 889782272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688866599799398,
      "loss": 2.8207,
      "theoretical_loss": 3.69038857185633,
      "tokens_seen": 889847808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036887662988966904,
      "loss": 2.5862,
      "theoretical_loss": 3.6903621466702474,
      "tokens_seen": 889913344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036886659979939817,
      "loss": 2.7536,
      "theoretical_loss": 3.6903357239749703,
      "tokens_seen": 889978880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688565697091274,
      "loss": 2.7551,
      "theoretical_loss": 3.6903093037700803,
      "tokens_seen": 890044416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688465396188566,
      "loss": 2.7463,
      "theoretical_loss": 3.690282886055159,
      "tokens_seen": 890109952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036883650952858576,
      "loss": 2.6848,
      "theoretical_loss": 3.690256470829789,
      "tokens_seen": 890175488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036882647943831494,
      "loss": 2.8369,
      "theoretical_loss": 3.690230058093552,
      "tokens_seen": 890241024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688164493480441,
      "loss": 2.7004,
      "theoretical_loss": 3.6902036478460305,
      "tokens_seen": 890306560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688064192577733,
      "loss": 2.6664,
      "theoretical_loss": 3.6901772400868063,
      "tokens_seen": 890372096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036879638916750254,
      "loss": 2.6107,
      "theoretical_loss": 3.6901508348154626,
      "tokens_seen": 890437632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036878635907723167,
      "loss": 2.6259,
      "theoretical_loss": 3.6901244320315816,
      "tokens_seen": 890503168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687763289869609,
      "loss": 2.6725,
      "theoretical_loss": 3.6900980317347454,
      "tokens_seen": 890568704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036876629889669003,
      "loss": 2.6895,
      "theoretical_loss": 3.6900716339245383,
      "tokens_seen": 890634240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036875626880641927,
      "loss": 2.7007,
      "theoretical_loss": 3.6900452386005407,
      "tokens_seen": 890699776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036874623871614845,
      "loss": 2.7271,
      "theoretical_loss": 3.6900188457623377,
      "tokens_seen": 890765312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036873620862587763,
      "loss": 2.7241,
      "theoretical_loss": 3.689992455409512,
      "tokens_seen": 890830848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687261785356068,
      "loss": 2.5983,
      "theoretical_loss": 3.6899660675416457,
      "tokens_seen": 890896384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036871614844533605,
      "loss": 2.9601,
      "theoretical_loss": 3.6899396821583226,
      "tokens_seen": 890961920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036870611835506517,
      "loss": 2.7535,
      "theoretical_loss": 3.6899132992591266,
      "tokens_seen": 891027456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686960882647944,
      "loss": 2.5921,
      "theoretical_loss": 3.68988691884364,
      "tokens_seen": 891092992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036868605817452353,
      "loss": 2.887,
      "theoretical_loss": 3.6898605409114476,
      "tokens_seen": 891158528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036867602808425277,
      "loss": 2.6801,
      "theoretical_loss": 3.6898341654621314,
      "tokens_seen": 891224064
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 498750,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7770748138427734,
      "objective/train/theoretical_loss": 3.6898077924952775,
      "objective/train/tokens_used": 911749600,
      "theoretical_loss": 3.6898077924952775,
      "tokens_seen": 891289600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036866599799398195,
      "loss": 2.663,
      "theoretical_loss": 3.6898077924952775,
      "tokens_seen": 891289600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036865596790371113,
      "loss": 2.883,
      "theoretical_loss": 3.6897814220104674,
      "tokens_seen": 891355136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686459378134403,
      "loss": 2.7317,
      "theoretical_loss": 3.689755054007286,
      "tokens_seen": 891420672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686359077231695,
      "loss": 3.0149,
      "theoretical_loss": 3.6897286884853173,
      "tokens_seen": 891486208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686258776328987,
      "loss": 2.7615,
      "theoretical_loss": 3.6897023254441463,
      "tokens_seen": 891551744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686158475426279,
      "loss": 2.6727,
      "theoretical_loss": 3.689675964883356,
      "tokens_seen": 891617280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686058174523571,
      "loss": 2.8897,
      "theoretical_loss": 3.689649606802531,
      "tokens_seen": 891682816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685957873620863,
      "loss": 2.7298,
      "theoretical_loss": 3.6896232512012563,
      "tokens_seen": 891748352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036858575727181545,
      "loss": 2.6903,
      "theoretical_loss": 3.6895968980791163,
      "tokens_seen": 891813888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036857572718154464,
      "loss": 2.8455,
      "theoretical_loss": 3.689570547435695,
      "tokens_seen": 891879424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036856569709127387,
      "loss": 2.9195,
      "theoretical_loss": 3.6895441992705775,
      "tokens_seen": 891944960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368555667001003,
      "loss": 2.9814,
      "theoretical_loss": 3.689517853583349,
      "tokens_seen": 892010496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036854563691073223,
      "loss": 2.5849,
      "theoretical_loss": 3.6894915103735944,
      "tokens_seen": 892076032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685356068204614,
      "loss": 2.847,
      "theoretical_loss": 3.689465169640898,
      "tokens_seen": 892141568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685255767301906,
      "loss": 2.8258,
      "theoretical_loss": 3.6894388313848454,
      "tokens_seen": 892207104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685155466399198,
      "loss": 2.6168,
      "theoretical_loss": 3.6894124956050223,
      "tokens_seen": 892272640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036850551654964896,
      "loss": 2.7964,
      "theoretical_loss": 3.6893861623010133,
      "tokens_seen": 892338176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036849548645937814,
      "loss": 2.6983,
      "theoretical_loss": 3.6893598314724043,
      "tokens_seen": 892403712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684854563691074,
      "loss": 2.502,
      "theoretical_loss": 3.689333503118781,
      "tokens_seen": 892469248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684754262788365,
      "loss": 2.8063,
      "theoretical_loss": 3.6893071772397286,
      "tokens_seen": 892534784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036846539618856574,
      "loss": 2.6416,
      "theoretical_loss": 3.689280853834833,
      "tokens_seen": 892600320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036845536609829486,
      "loss": 2.7694,
      "theoretical_loss": 3.6892545329036803,
      "tokens_seen": 892665856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684453360080241,
      "loss": 2.3705,
      "theoretical_loss": 3.6892282144458557,
      "tokens_seen": 892731392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684353059177533,
      "loss": 2.7649,
      "theoretical_loss": 3.6892018984609463,
      "tokens_seen": 892796928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036842527582748246,
      "loss": 2.8509,
      "theoretical_loss": 3.6891755849485373,
      "tokens_seen": 892862464
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 499509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5587472915649414,
      "objective/train/theoretical_loss": 3.6891492739082157,
      "objective/train/tokens_used": 913388000,
      "theoretical_loss": 3.6891492739082157,
      "tokens_seen": 892928000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036841524573721164,
      "loss": 2.8054,
      "theoretical_loss": 3.6891492739082157,
      "tokens_seen": 892928000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684052156469409,
      "loss": 2.5245,
      "theoretical_loss": 3.6891229653395667,
      "tokens_seen": 892993536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036839518555667,
      "loss": 2.39,
      "theoretical_loss": 3.6890966592421783,
      "tokens_seen": 893059072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036838515546639924,
      "loss": 2.5772,
      "theoretical_loss": 3.689070355615636,
      "tokens_seen": 893124608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036837512537612837,
      "loss": 2.8144,
      "theoretical_loss": 3.6890440544595267,
      "tokens_seen": 893190144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683650952858576,
      "loss": 2.7312,
      "theoretical_loss": 3.6890177557734374,
      "tokens_seen": 893255680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683550651955868,
      "loss": 2.8735,
      "theoretical_loss": 3.6889914595569544,
      "tokens_seen": 893321216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036834503510531596,
      "loss": 2.6178,
      "theoretical_loss": 3.688965165809665,
      "tokens_seen": 893386752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036833500501504514,
      "loss": 2.7219,
      "theoretical_loss": 3.688938874531156,
      "tokens_seen": 893452288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683249749247743,
      "loss": 2.7928,
      "theoretical_loss": 3.688912585721015,
      "tokens_seen": 893517824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683149448345035,
      "loss": 2.6863,
      "theoretical_loss": 3.6888862993788285,
      "tokens_seen": 893583360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036830491474423274,
      "loss": 2.8111,
      "theoretical_loss": 3.6888600155041846,
      "tokens_seen": 893648896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036829488465396187,
      "loss": 2.7794,
      "theoretical_loss": 3.6888337340966704,
      "tokens_seen": 893714432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682848545636911,
      "loss": 2.7854,
      "theoretical_loss": 3.688807455155873,
      "tokens_seen": 893779968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036827482447342023,
      "loss": 2.8295,
      "theoretical_loss": 3.688781178681381,
      "tokens_seen": 893845504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036826479438314947,
      "loss": 2.7785,
      "theoretical_loss": 3.6887549046727814,
      "tokens_seen": 893911040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036825476429287865,
      "loss": 2.7202,
      "theoretical_loss": 3.6887286331296627,
      "tokens_seen": 893976576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036824473420260783,
      "loss": 2.6292,
      "theoretical_loss": 3.6887023640516117,
      "tokens_seen": 894042112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368234704112337,
      "loss": 2.5048,
      "theoretical_loss": 3.688676097438217,
      "tokens_seen": 894107648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036822467402206625,
      "loss": 2.7371,
      "theoretical_loss": 3.6886498332890674,
      "tokens_seen": 894173184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036821464393179537,
      "loss": 2.9452,
      "theoretical_loss": 3.68862357160375,
      "tokens_seen": 894238720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682046138415246,
      "loss": 2.7559,
      "theoretical_loss": 3.6885973123818543,
      "tokens_seen": 894304256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036819458375125373,
      "loss": 2.7215,
      "theoretical_loss": 3.688571055622967,
      "tokens_seen": 894369792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036818455366098297,
      "loss": 2.6896,
      "theoretical_loss": 3.688544801326679,
      "tokens_seen": 894435328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036817452357071215,
      "loss": 2.7963,
      "theoretical_loss": 3.688518549492577,
      "tokens_seen": 894500864
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 500468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6406912803649902,
      "objective/train/theoretical_loss": 3.6884923001202505,
      "objective/train/tokens_used": 915026400,
      "theoretical_loss": 3.6884923001202505,
      "tokens_seen": 894566400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036816449348044133,
      "loss": 2.8942,
      "theoretical_loss": 3.6884923001202505,
      "tokens_seen": 894566400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681544633901705,
      "loss": 2.6235,
      "theoretical_loss": 3.6884660532092886,
      "tokens_seen": 894631936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681444332998997,
      "loss": 2.8042,
      "theoretical_loss": 3.6884398087592793,
      "tokens_seen": 894697472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681344032096289,
      "loss": 2.82,
      "theoretical_loss": 3.688413566769812,
      "tokens_seen": 894763008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681243731193581,
      "loss": 2.9533,
      "theoretical_loss": 3.6883873272404766,
      "tokens_seen": 894828544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036811434302908724,
      "loss": 2.6947,
      "theoretical_loss": 3.6883610901708614,
      "tokens_seen": 894894080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681043129388165,
      "loss": 2.8405,
      "theoretical_loss": 3.688334855560556,
      "tokens_seen": 894959616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680942828485456,
      "loss": 2.8924,
      "theoretical_loss": 3.6883086234091493,
      "tokens_seen": 895025152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036808425275827484,
      "loss": 2.7066,
      "theoretical_loss": 3.688282393716232,
      "tokens_seen": 895090688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368074222668004,
      "loss": 2.8829,
      "theoretical_loss": 3.688256166481393,
      "tokens_seen": 895156224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680641925777332,
      "loss": 2.8899,
      "theoretical_loss": 3.688229941704222,
      "tokens_seen": 895221760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680541624874624,
      "loss": 2.8019,
      "theoretical_loss": 3.6882037193843082,
      "tokens_seen": 895287296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680441323971916,
      "loss": 2.4944,
      "theoretical_loss": 3.6881774995212426,
      "tokens_seen": 895352832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036803410230692074,
      "loss": 2.4534,
      "theoretical_loss": 3.6881512821146147,
      "tokens_seen": 895418368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036802407221665,
      "loss": 2.7225,
      "theoretical_loss": 3.688125067164015,
      "tokens_seen": 895483904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680140421263791,
      "loss": 2.8461,
      "theoretical_loss": 3.6880988546690325,
      "tokens_seen": 895549440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036800401203610834,
      "loss": 3.0693,
      "theoretical_loss": 3.6880726446292593,
      "tokens_seen": 895614976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679939819458375,
      "loss": 2.7929,
      "theoretical_loss": 3.6880464370442843,
      "tokens_seen": 895680512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679839518555667,
      "loss": 2.7852,
      "theoretical_loss": 3.6880202319136988,
      "tokens_seen": 895746048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679739217652959,
      "loss": 2.7983,
      "theoretical_loss": 3.6879940292370934,
      "tokens_seen": 895811584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036796389167502506,
      "loss": 2.7209,
      "theoretical_loss": 3.687967829014058,
      "tokens_seen": 895877120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036795386158475424,
      "loss": 2.8093,
      "theoretical_loss": 3.6879416312441844,
      "tokens_seen": 895942656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679438314944835,
      "loss": 2.8308,
      "theoretical_loss": 3.6879154359270627,
      "tokens_seen": 896008192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679338014042126,
      "loss": 2.6901,
      "theoretical_loss": 3.6878892430622843,
      "tokens_seen": 896073728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036792377131394184,
      "loss": 2.8194,
      "theoretical_loss": 3.6878630526494405,
      "tokens_seen": 896139264
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 501526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.16762113571167,
      "objective/train/theoretical_loss": 3.687836864688122,
      "objective/train/tokens_used": 916664800,
      "theoretical_loss": 3.687836864688122,
      "tokens_seen": 896204800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036791374122367097,
      "loss": 3.0439,
      "theoretical_loss": 3.687836864688122,
      "tokens_seen": 896204800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003679037111334002,
      "loss": 2.8419,
      "theoretical_loss": 3.68781067917792,
      "tokens_seen": 896270336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678936810431294,
      "loss": 3.0514,
      "theoretical_loss": 3.687784496118426,
      "tokens_seen": 896335872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036788365095285857,
      "loss": 2.8178,
      "theoretical_loss": 3.687758315509232,
      "tokens_seen": 896401408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036787362086258775,
      "loss": 2.7582,
      "theoretical_loss": 3.687732137349929,
      "tokens_seen": 896466944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367863590772317,
      "loss": 3.0367,
      "theoretical_loss": 3.687705961640109,
      "tokens_seen": 896532480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036785356068204616,
      "loss": 2.7561,
      "theoretical_loss": 3.687679788379364,
      "tokens_seen": 896598016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036784353059177535,
      "loss": 2.6126,
      "theoretical_loss": 3.6876536175672854,
      "tokens_seen": 896663552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678335005015045,
      "loss": 2.9616,
      "theoretical_loss": 3.687627449203465,
      "tokens_seen": 896729088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678234704112337,
      "loss": 2.7894,
      "theoretical_loss": 3.6876012832874956,
      "tokens_seen": 896794624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036781344032096294,
      "loss": 2.7923,
      "theoretical_loss": 3.6875751198189683,
      "tokens_seen": 896860160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036780341023069207,
      "loss": 2.7743,
      "theoretical_loss": 3.687548958797476,
      "tokens_seen": 896925696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677933801404213,
      "loss": 2.873,
      "theoretical_loss": 3.6875228002226113,
      "tokens_seen": 896991232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036778335005015043,
      "loss": 2.7466,
      "theoretical_loss": 3.6874966440939665,
      "tokens_seen": 897056768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036777331995987967,
      "loss": 2.9035,
      "theoretical_loss": 3.6874704904111337,
      "tokens_seen": 897122304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036776328986960885,
      "loss": 2.9417,
      "theoretical_loss": 3.687444339173706,
      "tokens_seen": 897187840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036775325977933803,
      "loss": 2.7169,
      "theoretical_loss": 3.687418190381276,
      "tokens_seen": 897253376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677432296890672,
      "loss": 2.6652,
      "theoretical_loss": 3.6873920440334373,
      "tokens_seen": 897318912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036773319959879645,
      "loss": 2.402,
      "theoretical_loss": 3.687365900129781,
      "tokens_seen": 897384448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036772316950852557,
      "loss": 2.8106,
      "theoretical_loss": 3.6873397586699026,
      "tokens_seen": 897449984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677131394182548,
      "loss": 2.7959,
      "theoretical_loss": 3.687313619653393,
      "tokens_seen": 897515520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036770310932798393,
      "loss": 2.6755,
      "theoretical_loss": 3.6872874830798463,
      "tokens_seen": 897581056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036769307923771317,
      "loss": 2.8384,
      "theoretical_loss": 3.6872613489488564,
      "tokens_seen": 897646592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036768304914744235,
      "loss": 2.7293,
      "theoretical_loss": 3.6872352172600156,
      "tokens_seen": 897712128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036767301905717153,
      "loss": 2.5484,
      "theoretical_loss": 3.6872090880129185,
      "tokens_seen": 897777664
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 502221,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.841933012008667,
      "objective/train/theoretical_loss": 3.6871829612071583,
      "objective/train/tokens_used": 918303200,
      "theoretical_loss": 3.6871829612071583,
      "tokens_seen": 897843200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676629889669007,
      "loss": 2.6498,
      "theoretical_loss": 3.6871829612071583,
      "tokens_seen": 897843200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676529588766299,
      "loss": 2.8582,
      "theoretical_loss": 3.6871568368423286,
      "tokens_seen": 897908736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676429287863591,
      "loss": 2.7102,
      "theoretical_loss": 3.6871307149180232,
      "tokens_seen": 897974272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676328986960883,
      "loss": 2.7657,
      "theoretical_loss": 3.687104595433836,
      "tokens_seen": 898039808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036762286860581744,
      "loss": 2.7004,
      "theoretical_loss": 3.6870784783893615,
      "tokens_seen": 898105344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676128385155467,
      "loss": 2.7002,
      "theoretical_loss": 3.6870523637841934,
      "tokens_seen": 898170880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676028084252758,
      "loss": 2.6307,
      "theoretical_loss": 3.687026251617926,
      "tokens_seen": 898236416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036759277833500504,
      "loss": 2.877,
      "theoretical_loss": 3.6870001418901532,
      "tokens_seen": 898301952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675827482447342,
      "loss": 2.9289,
      "theoretical_loss": 3.6869740346004702,
      "tokens_seen": 898367488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675727181544634,
      "loss": 2.7154,
      "theoretical_loss": 3.686947929748471,
      "tokens_seen": 898433024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675626880641926,
      "loss": 3.0352,
      "theoretical_loss": 3.6869218273337503,
      "tokens_seen": 898498560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675526579739218,
      "loss": 2.8265,
      "theoretical_loss": 3.6868957273559024,
      "tokens_seen": 898564096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036754262788365094,
      "loss": 2.6395,
      "theoretical_loss": 3.686869629814523,
      "tokens_seen": 898629632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675325977933802,
      "loss": 2.8067,
      "theoretical_loss": 3.6868435347092063,
      "tokens_seen": 898695168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675225677031093,
      "loss": 2.7041,
      "theoretical_loss": 3.686817442039547,
      "tokens_seen": 898760704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036751253761283854,
      "loss": 2.8153,
      "theoretical_loss": 3.6867913518051414,
      "tokens_seen": 898826240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675025075225677,
      "loss": 2.6424,
      "theoretical_loss": 3.6867652640055835,
      "tokens_seen": 898891776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674924774322969,
      "loss": 2.8134,
      "theoretical_loss": 3.686739178640469,
      "tokens_seen": 898957312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674824473420261,
      "loss": 2.6906,
      "theoretical_loss": 3.6867130957093934,
      "tokens_seen": 899022848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036747241725175526,
      "loss": 2.707,
      "theoretical_loss": 3.6866870152119517,
      "tokens_seen": 899088384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036746238716148444,
      "loss": 2.9336,
      "theoretical_loss": 3.6866609371477397,
      "tokens_seen": 899153920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674523570712137,
      "loss": 2.5698,
      "theoretical_loss": 3.6866348615163536,
      "tokens_seen": 899219456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674423269809428,
      "loss": 2.7786,
      "theoretical_loss": 3.686608788317389,
      "tokens_seen": 899284992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036743229689067204,
      "loss": 2.795,
      "theoretical_loss": 3.686582717550441,
      "tokens_seen": 899350528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036742226680040117,
      "loss": 2.8089,
      "theoretical_loss": 3.6865566492151056,
      "tokens_seen": 899416064
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 503170,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.086702585220337,
      "objective/train/theoretical_loss": 3.6865305833109794,
      "objective/train/tokens_used": 919941600,
      "theoretical_loss": 3.6865305833109794,
      "tokens_seen": 899481600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674122367101304,
      "loss": 2.8526,
      "theoretical_loss": 3.6865305833109794,
      "tokens_seen": 899481600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674022066198596,
      "loss": 2.7103,
      "theoretical_loss": 3.6865045198376594,
      "tokens_seen": 899547136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036739217652958877,
      "loss": 2.5794,
      "theoretical_loss": 3.6864784587947397,
      "tokens_seen": 899612672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036738214643931795,
      "loss": 2.8274,
      "theoretical_loss": 3.6864524001818184,
      "tokens_seen": 899678208
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673721163490472,
      "loss": 2.7288,
      "theoretical_loss": 3.686426343998491,
      "tokens_seen": 899743744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673620862587763,
      "loss": 2.7591,
      "theoretical_loss": 3.6864002902443547,
      "tokens_seen": 899809280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036735205616850555,
      "loss": 2.5409,
      "theoretical_loss": 3.6863742389190053,
      "tokens_seen": 899874816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036734202607823467,
      "loss": 2.87,
      "theoretical_loss": 3.6863481900220405,
      "tokens_seen": 899940352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673319959879639,
      "loss": 2.6941,
      "theoretical_loss": 3.6863221435530567,
      "tokens_seen": 900005888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673219658976931,
      "loss": 3.0026,
      "theoretical_loss": 3.6862960995116505,
      "tokens_seen": 900071424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036731193580742227,
      "loss": 2.6116,
      "theoretical_loss": 3.6862700578974192,
      "tokens_seen": 900136960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036730190571715145,
      "loss": 2.8755,
      "theoretical_loss": 3.68624401870996,
      "tokens_seen": 900202496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036729187562688063,
      "loss": 2.4658,
      "theoretical_loss": 3.68621798194887,
      "tokens_seen": 900268032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672818455366098,
      "loss": 2.8991,
      "theoretical_loss": 3.686191947613746,
      "tokens_seen": 900333568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036727181544633905,
      "loss": 2.972,
      "theoretical_loss": 3.686165915704186,
      "tokens_seen": 900399104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672617853560682,
      "loss": 2.9544,
      "theoretical_loss": 3.686139886219788,
      "tokens_seen": 900464640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672517552657974,
      "loss": 2.7917,
      "theoretical_loss": 3.6861138591601486,
      "tokens_seen": 900530176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036724172517552654,
      "loss": 2.7219,
      "theoretical_loss": 3.6860878345248658,
      "tokens_seen": 900595712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036723169508525577,
      "loss": 2.7,
      "theoretical_loss": 3.686061812313538,
      "tokens_seen": 900661248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036722166499498495,
      "loss": 2.9898,
      "theoretical_loss": 3.686035792525762,
      "tokens_seen": 900726784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036721163490471414,
      "loss": 2.7812,
      "theoretical_loss": 3.6860097751611356,
      "tokens_seen": 900792320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672016048144433,
      "loss": 2.7424,
      "theoretical_loss": 3.6859837602192584,
      "tokens_seen": 900857856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036719157472417255,
      "loss": 2.9703,
      "theoretical_loss": 3.6859577476997276,
      "tokens_seen": 900923392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671815446339017,
      "loss": 2.8666,
      "theoretical_loss": 3.6859317376021417,
      "tokens_seen": 900988928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671715145436309,
      "loss": 2.7763,
      "theoretical_loss": 3.6859057299260987,
      "tokens_seen": 901054464
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 504328,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588376045227051,
      "objective/train/theoretical_loss": 3.6858797246711976,
      "objective/train/tokens_used": 921580000,
      "theoretical_loss": 3.6858797246711976,
      "tokens_seen": 901120000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036716148445336004,
      "loss": 2.7931,
      "theoretical_loss": 3.6858797246711976,
      "tokens_seen": 901120000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671514543630893,
      "loss": 2.7644,
      "theoretical_loss": 3.685853721837036,
      "tokens_seen": 901185536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036714142427281846,
      "loss": 2.8663,
      "theoretical_loss": 3.6858277214232142,
      "tokens_seen": 901251072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036713139418254764,
      "loss": 2.8726,
      "theoretical_loss": 3.6858017234293294,
      "tokens_seen": 901316608
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671213640922768,
      "loss": 2.8011,
      "theoretical_loss": 3.6857757278549803,
      "tokens_seen": 901382144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367111334002006,
      "loss": 2.9276,
      "theoretical_loss": 3.6857497346997676,
      "tokens_seen": 901447680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036710130391173524,
      "loss": 2.8506,
      "theoretical_loss": 3.6857237439632886,
      "tokens_seen": 901513216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670912738214644,
      "loss": 2.8113,
      "theoretical_loss": 3.685697755645143,
      "tokens_seen": 901578752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670812437311936,
      "loss": 2.7378,
      "theoretical_loss": 3.6856717697449306,
      "tokens_seen": 901644288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670712136409228,
      "loss": 2.7982,
      "theoretical_loss": 3.68564578626225,
      "tokens_seen": 901709824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367061183550652,
      "loss": 2.3642,
      "theoretical_loss": 3.6856198051967004,
      "tokens_seen": 901775360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036705115346038114,
      "loss": 2.837,
      "theoretical_loss": 3.6855938265478825,
      "tokens_seen": 901840896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670411233701104,
      "loss": 2.7009,
      "theoretical_loss": 3.6855678503153944,
      "tokens_seen": 901906432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670310932798395,
      "loss": 2.8224,
      "theoretical_loss": 3.685541876498837,
      "tokens_seen": 901971968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036702106318956874,
      "loss": 2.8333,
      "theoretical_loss": 3.685515905097809,
      "tokens_seen": 902037504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670110330992979,
      "loss": 2.8069,
      "theoretical_loss": 3.685489936111911,
      "tokens_seen": 902103040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670010030090271,
      "loss": 2.8903,
      "theoretical_loss": 3.685463969540743,
      "tokens_seen": 902168576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669909729187563,
      "loss": 2.6957,
      "theoretical_loss": 3.685438005383905,
      "tokens_seen": 902234112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036698094282848546,
      "loss": 2.7104,
      "theoretical_loss": 3.685412043640997,
      "tokens_seen": 902299648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036697091273821464,
      "loss": 2.7687,
      "theoretical_loss": 3.6853860843116193,
      "tokens_seen": 902365184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669608826479439,
      "loss": 2.8506,
      "theoretical_loss": 3.6853601273953727,
      "tokens_seen": 902430720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366950852557673,
      "loss": 2.6235,
      "theoretical_loss": 3.6853341728918574,
      "tokens_seen": 902496256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036694082246740224,
      "loss": 2.8101,
      "theoretical_loss": 3.6853082208006733,
      "tokens_seen": 902561792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036693079237713137,
      "loss": 2.9473,
      "theoretical_loss": 3.685282271121422,
      "tokens_seen": 902627328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669207622868606,
      "loss": 2.7178,
      "theoretical_loss": 3.6852563238537037,
      "tokens_seen": 902692864
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 504856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7679882049560547,
      "objective/train/theoretical_loss": 3.6852303789971192,
      "objective/train/tokens_used": 923218400,
      "theoretical_loss": 3.6852303789971192,
      "tokens_seen": 902758400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669107321965898,
      "loss": 2.8659,
      "theoretical_loss": 3.6852303789971192,
      "tokens_seen": 902758400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036690070210631897,
      "loss": 2.9552,
      "theoretical_loss": 3.68520443655127,
      "tokens_seen": 902823936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036689067201604815,
      "loss": 2.63,
      "theoretical_loss": 3.6851784965157566,
      "tokens_seen": 902889472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668806419257774,
      "loss": 2.665,
      "theoretical_loss": 3.68515255889018,
      "tokens_seen": 902955008
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668706118355065,
      "loss": 2.7384,
      "theoretical_loss": 3.685126623674142,
      "tokens_seen": 903020544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036686058174523575,
      "loss": 2.8029,
      "theoretical_loss": 3.6851006908672437,
      "tokens_seen": 903086080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036685055165496487,
      "loss": 2.7919,
      "theoretical_loss": 3.685074760469086,
      "tokens_seen": 903151616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668405215646941,
      "loss": 2.74,
      "theoretical_loss": 3.6850488324792714,
      "tokens_seen": 903217152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668304914744233,
      "loss": 2.8418,
      "theoretical_loss": 3.6850229068974008,
      "tokens_seen": 903282688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036682046138415247,
      "loss": 2.7863,
      "theoretical_loss": 3.6849969837230754,
      "tokens_seen": 903348224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036681043129388165,
      "loss": 2.8803,
      "theoretical_loss": 3.6849710629558983,
      "tokens_seen": 903413760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036680040120361083,
      "loss": 2.9019,
      "theoretical_loss": 3.6849451445954706,
      "tokens_seen": 903479296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036679037111334,
      "loss": 2.836,
      "theoretical_loss": 3.684919228641394,
      "tokens_seen": 903544832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036678034102306925,
      "loss": 2.7521,
      "theoretical_loss": 3.6848933150932712,
      "tokens_seen": 903610368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667703109327984,
      "loss": 2.5911,
      "theoretical_loss": 3.684867403950704,
      "tokens_seen": 903675904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667602808425276,
      "loss": 2.6519,
      "theoretical_loss": 3.684841495213295,
      "tokens_seen": 903741440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036675025075225674,
      "loss": 2.8198,
      "theoretical_loss": 3.684815588880646,
      "tokens_seen": 903806976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366740220661986,
      "loss": 2.5777,
      "theoretical_loss": 3.68478968495236,
      "tokens_seen": 903872512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036673019057171515,
      "loss": 2.875,
      "theoretical_loss": 3.684763783428039,
      "tokens_seen": 903938048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036672016048144434,
      "loss": 2.7931,
      "theoretical_loss": 3.684737884307286,
      "tokens_seen": 904003584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667101303911735,
      "loss": 2.6516,
      "theoretical_loss": 3.6847119875897043,
      "tokens_seen": 904069120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036670010030090275,
      "loss": 2.8678,
      "theoretical_loss": 3.684686093274896,
      "tokens_seen": 904134656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666900702106319,
      "loss": 2.6876,
      "theoretical_loss": 3.6846602013624636,
      "tokens_seen": 904200192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666800401203611,
      "loss": 2.5903,
      "theoretical_loss": 3.684634311852011,
      "tokens_seen": 904265728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036667001003009024,
      "loss": 2.8271,
      "theoretical_loss": 3.684608424743141,
      "tokens_seen": 904331264
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 505908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.720822811126709,
      "objective/train/theoretical_loss": 3.684582540035456,
      "objective/train/tokens_used": 924856800,
      "theoretical_loss": 3.684582540035456,
      "tokens_seen": 904396800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666599799398195,
      "loss": 2.7747,
      "theoretical_loss": 3.684582540035456,
      "tokens_seen": 904396800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036664994984954866,
      "loss": 2.5759,
      "theoretical_loss": 3.684556657728561,
      "tokens_seen": 904462336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036663991975927784,
      "loss": 2.8172,
      "theoretical_loss": 3.684530777822059,
      "tokens_seen": 904527872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366629889669007,
      "loss": 2.6142,
      "theoretical_loss": 3.684504900315552,
      "tokens_seen": 904593408
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666198595787362,
      "loss": 2.8247,
      "theoretical_loss": 3.6844790252086446,
      "tokens_seen": 904658944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666098294884654,
      "loss": 2.7764,
      "theoretical_loss": 3.6844531525009403,
      "tokens_seen": 904724480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665997993981946,
      "loss": 2.8979,
      "theoretical_loss": 3.6844272821920434,
      "tokens_seen": 904790016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036658976930792374,
      "loss": 2.7468,
      "theoretical_loss": 3.6844014142815578,
      "tokens_seen": 904855552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366579739217653,
      "loss": 2.9656,
      "theoretical_loss": 3.6843755487690864,
      "tokens_seen": 904921088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036656970912738216,
      "loss": 2.6941,
      "theoretical_loss": 3.684349685654234,
      "tokens_seen": 904986624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036655967903711134,
      "loss": 2.6184,
      "theoretical_loss": 3.6843238249366044,
      "tokens_seen": 905052160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665496489468405,
      "loss": 2.659,
      "theoretical_loss": 3.6842979666158024,
      "tokens_seen": 905117696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665396188565697,
      "loss": 2.9313,
      "theoretical_loss": 3.6842721106914316,
      "tokens_seen": 905183232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665295887662989,
      "loss": 2.6039,
      "theoretical_loss": 3.6842462571630974,
      "tokens_seen": 905248768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665195586760281,
      "loss": 3.0197,
      "theoretical_loss": 3.6842204060304033,
      "tokens_seen": 905314304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036650952858575725,
      "loss": 2.7065,
      "theoretical_loss": 3.6841945572929546,
      "tokens_seen": 905379840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664994984954865,
      "loss": 2.6915,
      "theoretical_loss": 3.684168710950356,
      "tokens_seen": 905445376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664894684052156,
      "loss": 2.6842,
      "theoretical_loss": 3.684142867002212,
      "tokens_seen": 905510912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036647943831494484,
      "loss": 2.6965,
      "theoretical_loss": 3.684117025448127,
      "tokens_seen": 905576448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366469408224674,
      "loss": 2.9644,
      "theoretical_loss": 3.684091186287707,
      "tokens_seen": 905641984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664593781344032,
      "loss": 2.8037,
      "theoretical_loss": 3.684065349520557,
      "tokens_seen": 905707520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664493480441324,
      "loss": 2.5095,
      "theoretical_loss": 3.6840395151462815,
      "tokens_seen": 905773056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036643931795386157,
      "loss": 2.6812,
      "theoretical_loss": 3.684013683164486,
      "tokens_seen": 905838592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036642928786359075,
      "loss": 2.852,
      "theoretical_loss": 3.6839878535747763,
      "tokens_seen": 905904128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036641925777332,
      "loss": 2.6574,
      "theoretical_loss": 3.683962026376758,
      "tokens_seen": 905969664
    },
    {
      "epoch": 0.27,
      "objective/train/docs_used": 506404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0108041763305664,
      "objective/train/theoretical_loss": 3.6839362015700354,
      "objective/train/tokens_used": 926495200,
      "theoretical_loss": 3.6839362015700354,
      "tokens_seen": 906035200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664092276830491,
      "loss": 2.5565,
      "theoretical_loss": 3.6839362015700354,
      "tokens_seen": 906035200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036639919759277835,
      "loss": 2.7534,
      "theoretical_loss": 3.683910379154215,
      "tokens_seen": 906100736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036638916750250753,
      "loss": 2.4499,
      "theoretical_loss": 3.683884559128903,
      "tokens_seen": 906166272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663791374122367,
      "loss": 2.7014,
      "theoretical_loss": 3.6838587414937045,
      "tokens_seen": 906231808
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663691073219659,
      "loss": 2.6931,
      "theoretical_loss": 3.6838329262482254,
      "tokens_seen": 906297344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036635907723169507,
      "loss": 2.9389,
      "theoretical_loss": 3.6838071133920725,
      "tokens_seen": 906362880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663490471414243,
      "loss": 2.7749,
      "theoretical_loss": 3.683781302924851,
      "tokens_seen": 906428416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663390170511535,
      "loss": 2.7101,
      "theoretical_loss": 3.6837554948461677,
      "tokens_seen": 906493952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036632898696088267,
      "loss": 2.6792,
      "theoretical_loss": 3.683729689155629,
      "tokens_seen": 906559488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036631895687061185,
      "loss": 2.7912,
      "theoretical_loss": 3.6837038858528404,
      "tokens_seen": 906625024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036630892678034103,
      "loss": 2.728,
      "theoretical_loss": 3.6836780849374096,
      "tokens_seen": 906690560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662988966900702,
      "loss": 2.6709,
      "theoretical_loss": 3.6836522864089423,
      "tokens_seen": 906756096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036628886659979945,
      "loss": 2.8336,
      "theoretical_loss": 3.683626490267046,
      "tokens_seen": 906821632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662788365095286,
      "loss": 2.7861,
      "theoretical_loss": 3.683600696511326,
      "tokens_seen": 906887168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662688064192578,
      "loss": 2.8822,
      "theoretical_loss": 3.6835749051413904,
      "tokens_seen": 906952704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036625877632898694,
      "loss": 2.8046,
      "theoretical_loss": 3.683549116156846,
      "tokens_seen": 907018240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662487462387162,
      "loss": 2.6543,
      "theoretical_loss": 3.6835233295573,
      "tokens_seen": 907083776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036623871614844535,
      "loss": 2.9645,
      "theoretical_loss": 3.6834975453423593,
      "tokens_seen": 907149312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036622868605817454,
      "loss": 2.7094,
      "theoretical_loss": 3.6834717635116307,
      "tokens_seen": 907214848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662186559679037,
      "loss": 2.7114,
      "theoretical_loss": 3.683445984064722,
      "tokens_seen": 907280384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036620862587763295,
      "loss": 2.6491,
      "theoretical_loss": 3.6834202070012405,
      "tokens_seen": 907345920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003661985957873621,
      "loss": 2.8925,
      "theoretical_loss": 3.6833944323207937,
      "tokens_seen": 907411456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003661885656970913,
      "loss": 2.9666,
      "theoretical_loss": 3.68336866002299,
      "tokens_seen": 907476992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036617853560682044,
      "loss": 2.6134,
      "theoretical_loss": 3.6833428901074354,
      "tokens_seen": 907542528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661685055165497,
      "loss": 2.7143,
      "theoretical_loss": 3.683317122573739,
      "tokens_seen": 907608064
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 507582,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5680878162384033,
      "objective/train/theoretical_loss": 3.683291357421508,
      "objective/train/tokens_used": 928133600,
      "theoretical_loss": 3.683291357421508,
      "tokens_seen": 907673600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036615847542627886,
      "loss": 2.7327,
      "theoretical_loss": 3.683291357421508,
      "tokens_seen": 907673600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036614844533600804,
      "loss": 2.8497,
      "theoretical_loss": 3.683265594650351,
      "tokens_seen": 907739136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661384152457372,
      "loss": 2.6495,
      "theoretical_loss": 3.683239834259876,
      "tokens_seen": 907804672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661283851554664,
      "loss": 2.9766,
      "theoretical_loss": 3.6832140762496905,
      "tokens_seen": 907870208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661183550651956,
      "loss": 2.7795,
      "theoretical_loss": 3.683188320619404,
      "tokens_seen": 907935744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661083249749248,
      "loss": 2.6712,
      "theoretical_loss": 3.6831625673686235,
      "tokens_seen": 908001280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036609829488465394,
      "loss": 2.7726,
      "theoretical_loss": 3.6831368164969582,
      "tokens_seen": 908066816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660882647943832,
      "loss": 2.7,
      "theoretical_loss": 3.683111068004016,
      "tokens_seen": 908132352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036607823470411236,
      "loss": 2.6533,
      "theoretical_loss": 3.6830853218894064,
      "tokens_seen": 908197888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036606820461384154,
      "loss": 2.6621,
      "theoretical_loss": 3.6830595781527373,
      "tokens_seen": 908263424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660581745235707,
      "loss": 2.8009,
      "theoretical_loss": 3.6830338367936184,
      "tokens_seen": 908328960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660481444332999,
      "loss": 2.6356,
      "theoretical_loss": 3.683008097811658,
      "tokens_seen": 908394496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660381143430291,
      "loss": 2.8399,
      "theoretical_loss": 3.6829823612064647,
      "tokens_seen": 908460032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660280842527583,
      "loss": 2.8992,
      "theoretical_loss": 3.6829566269776484,
      "tokens_seen": 908525568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036601805416248745,
      "loss": 2.7555,
      "theoretical_loss": 3.6829308951248185,
      "tokens_seen": 908591104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660080240722167,
      "loss": 2.9406,
      "theoretical_loss": 3.682905165647583,
      "tokens_seen": 908656640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659979939819458,
      "loss": 2.7128,
      "theoretical_loss": 3.6828794385455526,
      "tokens_seen": 908722176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036598796389167504,
      "loss": 2.6451,
      "theoretical_loss": 3.6828537138183357,
      "tokens_seen": 908787712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659779338014042,
      "loss": 2.8677,
      "theoretical_loss": 3.682827991465543,
      "tokens_seen": 908853248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659679037111334,
      "loss": 2.7044,
      "theoretical_loss": 3.682802271486783,
      "tokens_seen": 908918784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659578736208626,
      "loss": 2.7429,
      "theoretical_loss": 3.6827765538816655,
      "tokens_seen": 908984320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036594784353059177,
      "loss": 2.8765,
      "theoretical_loss": 3.6827508386498016,
      "tokens_seen": 909049856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036593781344032095,
      "loss": 2.6742,
      "theoretical_loss": 3.6827251257907996,
      "tokens_seen": 909115392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659277833500502,
      "loss": 2.651,
      "theoretical_loss": 3.682699415304271,
      "tokens_seen": 909180928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003659177532597793,
      "loss": 2.6706,
      "theoretical_loss": 3.682673707189825,
      "tokens_seen": 909246464
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 508185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.757549524307251,
      "objective/train/theoretical_loss": 3.6826480014470717,
      "objective/train/tokens_used": 929772000,
      "theoretical_loss": 3.6826480014470717,
      "tokens_seen": 909312000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036590772316950855,
      "loss": 2.6325,
      "theoretical_loss": 3.6826480014470717,
      "tokens_seen": 909312000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036589769307923773,
      "loss": 2.5434,
      "theoretical_loss": 3.6826222980756222,
      "tokens_seen": 909377536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658876629889669,
      "loss": 2.9387,
      "theoretical_loss": 3.6825965970750856,
      "tokens_seen": 909443072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658776328986961,
      "loss": 2.8113,
      "theoretical_loss": 3.682570898445074,
      "tokens_seen": 909508608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036586760280842527,
      "loss": 2.8491,
      "theoretical_loss": 3.6825452021851963,
      "tokens_seen": 909574144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036585757271815445,
      "loss": 2.6226,
      "theoretical_loss": 3.682519508295065,
      "tokens_seen": 909639680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658475426278837,
      "loss": 2.4749,
      "theoretical_loss": 3.6824938167742887,
      "tokens_seen": 909705216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658375125376128,
      "loss": 2.7062,
      "theoretical_loss": 3.6824681276224798,
      "tokens_seen": 909770752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036582748244734205,
      "loss": 2.5519,
      "theoretical_loss": 3.682442440839249,
      "tokens_seen": 909836288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658174523570712,
      "loss": 2.7983,
      "theoretical_loss": 3.6824167564242067,
      "tokens_seen": 909901824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658074222668004,
      "loss": 2.7716,
      "theoretical_loss": 3.682391074376965,
      "tokens_seen": 909967360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657973921765296,
      "loss": 2.7686,
      "theoretical_loss": 3.6823653946971344,
      "tokens_seen": 910032896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657873620862588,
      "loss": 2.6568,
      "theoretical_loss": 3.682339717384326,
      "tokens_seen": 910098432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036577733199598796,
      "loss": 2.7357,
      "theoretical_loss": 3.6823140424381524,
      "tokens_seen": 910163968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036576730190571714,
      "loss": 2.7161,
      "theoretical_loss": 3.682288369858223,
      "tokens_seen": 910229504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657572718154463,
      "loss": 2.7993,
      "theoretical_loss": 3.6822626996441516,
      "tokens_seen": 910295040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036574724172517555,
      "loss": 2.9071,
      "theoretical_loss": 3.6822370317955486,
      "tokens_seen": 910360576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657372116349047,
      "loss": 2.5976,
      "theoretical_loss": 3.682211366312026,
      "tokens_seen": 910426112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657271815446339,
      "loss": 2.7127,
      "theoretical_loss": 3.6821857031931957,
      "tokens_seen": 910491648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657171514543631,
      "loss": 2.7244,
      "theoretical_loss": 3.68216004243867,
      "tokens_seen": 910557184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657071213640923,
      "loss": 2.768,
      "theoretical_loss": 3.6821343840480605,
      "tokens_seen": 910622720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036569709127382146,
      "loss": 2.6111,
      "theoretical_loss": 3.682108728020979,
      "tokens_seen": 910688256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036568706118355064,
      "loss": 2.8181,
      "theoretical_loss": 3.6820830743570383,
      "tokens_seen": 910753792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656770310932798,
      "loss": 2.7025,
      "theoretical_loss": 3.6820574230558503,
      "tokens_seen": 910819328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036566700100300906,
      "loss": 2.7488,
      "theoretical_loss": 3.682031774117028,
      "tokens_seen": 910884864
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 509502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5731163024902344,
      "objective/train/theoretical_loss": 3.682006127540184,
      "objective/train/tokens_used": 931410400,
      "theoretical_loss": 3.682006127540184,
      "tokens_seen": 910950400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656569709127382,
      "loss": 2.7661,
      "theoretical_loss": 3.682006127540184,
      "tokens_seen": 910950400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656469408224674,
      "loss": 2.9085,
      "theoretical_loss": 3.6819804833249297,
      "tokens_seen": 911015936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036563691073219655,
      "loss": 2.7918,
      "theoretical_loss": 3.681954841470879,
      "tokens_seen": 911081472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656268806419258,
      "loss": 2.6731,
      "theoretical_loss": 3.6819292019776437,
      "tokens_seen": 911147008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036561685055165496,
      "loss": 2.7863,
      "theoretical_loss": 3.6819035648448373,
      "tokens_seen": 911212544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036560682046138414,
      "loss": 2.8253,
      "theoretical_loss": 3.6818779300720728,
      "tokens_seen": 911278080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655967903711134,
      "loss": 2.7498,
      "theoretical_loss": 3.6818522976589625,
      "tokens_seen": 911343616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036558676028084256,
      "loss": 2.6159,
      "theoretical_loss": 3.6818266676051206,
      "tokens_seen": 911409152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036557673019057174,
      "loss": 2.7005,
      "theoretical_loss": 3.6818010399101597,
      "tokens_seen": 911474688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655667001003009,
      "loss": 2.6145,
      "theoretical_loss": 3.6817754145736927,
      "tokens_seen": 911540224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655566700100301,
      "loss": 2.7677,
      "theoretical_loss": 3.681749791595334,
      "tokens_seen": 911605760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655466399197593,
      "loss": 2.9102,
      "theoretical_loss": 3.6817241709746966,
      "tokens_seen": 911671296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655366098294885,
      "loss": 2.8622,
      "theoretical_loss": 3.681698552711394,
      "tokens_seen": 911736832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036552657973921765,
      "loss": 2.7773,
      "theoretical_loss": 3.6816729368050396,
      "tokens_seen": 911802368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655165496489469,
      "loss": 2.6455,
      "theoretical_loss": 3.6816473232552482,
      "tokens_seen": 911867904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365506519558676,
      "loss": 2.699,
      "theoretical_loss": 3.6816217120616326,
      "tokens_seen": 911933440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036549648946840525,
      "loss": 2.8395,
      "theoretical_loss": 3.681596103223807,
      "tokens_seen": 911998976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654864593781344,
      "loss": 2.7838,
      "theoretical_loss": 3.6815704967413865,
      "tokens_seen": 912064512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654764292878636,
      "loss": 2.7302,
      "theoretical_loss": 3.6815448926139833,
      "tokens_seen": 912130048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654663991975928,
      "loss": 2.8579,
      "theoretical_loss": 3.6815192908412135,
      "tokens_seen": 912195584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036545636910732197,
      "loss": 2.7153,
      "theoretical_loss": 3.6814936914226903,
      "tokens_seen": 912261120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036544633901705115,
      "loss": 2.8546,
      "theoretical_loss": 3.6814680943580287,
      "tokens_seen": 912326656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654363089267804,
      "loss": 2.5904,
      "theoretical_loss": 3.681442499646842,
      "tokens_seen": 912392192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654262788365095,
      "loss": 2.6852,
      "theoretical_loss": 3.6814169072887464,
      "tokens_seen": 912457728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036541624874623875,
      "loss": 2.6125,
      "theoretical_loss": 3.6813913172833557,
      "tokens_seen": 912523264
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 510138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.649203062057495,
      "objective/train/theoretical_loss": 3.6813657296302846,
      "objective/train/tokens_used": 933048800,
      "theoretical_loss": 3.6813657296302846,
      "tokens_seen": 912588800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036540621865596793,
      "loss": 2.6317,
      "theoretical_loss": 3.6813657296302846,
      "tokens_seen": 912588800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653961885656971,
      "loss": 2.6273,
      "theoretical_loss": 3.6813401443291482,
      "tokens_seen": 912654336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653861584754263,
      "loss": 2.8201,
      "theoretical_loss": 3.681314561379562,
      "tokens_seen": 912719872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036537612838515547,
      "loss": 2.8762,
      "theoretical_loss": 3.68128898078114,
      "tokens_seen": 912785408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036536609829488465,
      "loss": 2.7939,
      "theoretical_loss": 3.6812634025334985,
      "tokens_seen": 912850944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653560682046139,
      "loss": 2.7877,
      "theoretical_loss": 3.6812378266362513,
      "tokens_seen": 912916480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365346038114343,
      "loss": 2.7212,
      "theoretical_loss": 3.6812122530890146,
      "tokens_seen": 912982016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036533600802407225,
      "loss": 2.683,
      "theoretical_loss": 3.681186681891404,
      "tokens_seen": 913047552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653259779338014,
      "loss": 2.7768,
      "theoretical_loss": 3.6811611130430344,
      "tokens_seen": 913113088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653159478435306,
      "loss": 2.6558,
      "theoretical_loss": 3.681135546543522,
      "tokens_seen": 913178624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653059177532598,
      "loss": 2.6317,
      "theoretical_loss": 3.6811099823924813,
      "tokens_seen": 913244160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365295887662989,
      "loss": 2.7281,
      "theoretical_loss": 3.6810844205895297,
      "tokens_seen": 913309696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036528585757271816,
      "loss": 2.6293,
      "theoretical_loss": 3.6810588611342823,
      "tokens_seen": 913375232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036527582748244734,
      "loss": 2.7732,
      "theoretical_loss": 3.6810333040263545,
      "tokens_seen": 913440768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652657973921765,
      "loss": 2.7447,
      "theoretical_loss": 3.6810077492653632,
      "tokens_seen": 913506304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036525576730190575,
      "loss": 2.8119,
      "theoretical_loss": 3.680982196850924,
      "tokens_seen": 913571840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652457372116349,
      "loss": 2.6916,
      "theoretical_loss": 3.6809566467826533,
      "tokens_seen": 913637376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652357071213641,
      "loss": 2.8739,
      "theoretical_loss": 3.680931099060168,
      "tokens_seen": 913702912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652256770310933,
      "loss": 2.6952,
      "theoretical_loss": 3.680905553683083,
      "tokens_seen": 913768448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652156469408225,
      "loss": 2.7815,
      "theoretical_loss": 3.680880010651016,
      "tokens_seen": 913833984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036520561685055166,
      "loss": 2.8346,
      "theoretical_loss": 3.6808544699635837,
      "tokens_seen": 913899520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036519558676028084,
      "loss": 2.5171,
      "theoretical_loss": 3.680828931620402,
      "tokens_seen": 913965056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036518555667001,
      "loss": 2.6756,
      "theoretical_loss": 3.680803395621088,
      "tokens_seen": 914030592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036517552657973926,
      "loss": 2.5853,
      "theoretical_loss": 3.680777861965259,
      "tokens_seen": 914096128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651654964894684,
      "loss": 2.5823,
      "theoretical_loss": 3.680752330652531,
      "tokens_seen": 914161664
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 511255,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8368189334869385,
      "objective/train/theoretical_loss": 3.680726801682522,
      "objective/train/tokens_used": 934687200,
      "theoretical_loss": 3.680726801682522,
      "tokens_seen": 914227200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651554663991976,
      "loss": 2.7186,
      "theoretical_loss": 3.680726801682522,
      "tokens_seen": 914227200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036514543630892675,
      "loss": 2.6055,
      "theoretical_loss": 3.6807012750548482,
      "tokens_seen": 914292736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365135406218656,
      "loss": 2.6287,
      "theoretical_loss": 3.6806757507691277,
      "tokens_seen": 914358272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036512537612838516,
      "loss": 2.552,
      "theoretical_loss": 3.6806502288249767,
      "tokens_seen": 914423808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036511534603811434,
      "loss": 2.6986,
      "theoretical_loss": 3.6806247092220143,
      "tokens_seen": 914489344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651053159478435,
      "loss": 2.8004,
      "theoretical_loss": 3.6805991919598564,
      "tokens_seen": 914554880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036509528585757276,
      "loss": 2.5273,
      "theoretical_loss": 3.680573677038121,
      "tokens_seen": 914620416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650852557673019,
      "loss": 2.7229,
      "theoretical_loss": 3.680548164456426,
      "tokens_seen": 914685952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650752256770311,
      "loss": 2.4541,
      "theoretical_loss": 3.680522654214389,
      "tokens_seen": 914751488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036506519558676025,
      "loss": 2.6284,
      "theoretical_loss": 3.680497146311628,
      "tokens_seen": 914817024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650551654964895,
      "loss": 2.7057,
      "theoretical_loss": 3.6804716407477605,
      "tokens_seen": 914882560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036504513540621867,
      "loss": 2.8252,
      "theoretical_loss": 3.6804461375224053,
      "tokens_seen": 914948096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036503510531594785,
      "loss": 2.7718,
      "theoretical_loss": 3.68042063663518,
      "tokens_seen": 915013632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036502507522567703,
      "loss": 2.7386,
      "theoretical_loss": 3.6803951380857027,
      "tokens_seen": 915079168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650150451354062,
      "loss": 2.4951,
      "theoretical_loss": 3.6803696418735914,
      "tokens_seen": 915144704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650050150451354,
      "loss": 2.8558,
      "theoretical_loss": 3.6803441479984653,
      "tokens_seen": 915210240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649949849548646,
      "loss": 2.743,
      "theoretical_loss": 3.6803186564599426,
      "tokens_seen": 915275776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649949849548646,
      "loss": 2.5659,
      "theoretical_loss": 3.6802931672576413,
      "tokens_seen": 915341312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036498495486459375,
      "loss": 2.7712,
      "theoretical_loss": 3.680267680391181,
      "tokens_seen": 915406848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364974924774323,
      "loss": 2.741,
      "theoretical_loss": 3.6802421958601794,
      "tokens_seen": 915472384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649648946840521,
      "loss": 2.5163,
      "theoretical_loss": 3.6802167136642563,
      "tokens_seen": 915537920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036495486459378135,
      "loss": 2.8221,
      "theoretical_loss": 3.68019123380303,
      "tokens_seen": 915603456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036494483450351053,
      "loss": 2.6905,
      "theoretical_loss": 3.6801657562761196,
      "tokens_seen": 915668992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649348044132397,
      "loss": 2.8177,
      "theoretical_loss": 3.6801402810831445,
      "tokens_seen": 915734528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649247743229689,
      "loss": 2.8277,
      "theoretical_loss": 3.6801148082237236,
      "tokens_seen": 915800064
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 511864,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.596713066101074,
      "objective/train/theoretical_loss": 3.680089337697476,
      "objective/train/tokens_used": 936325600,
      "theoretical_loss": 3.680089337697476,
      "tokens_seen": 915865600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036491474423269813,
      "loss": 2.5618,
      "theoretical_loss": 3.680089337697476,
      "tokens_seen": 915865600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036490471414242726,
      "loss": 2.7874,
      "theoretical_loss": 3.6800638695040213,
      "tokens_seen": 915931136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648946840521565,
      "loss": 2.8756,
      "theoretical_loss": 3.680038403642979,
      "tokens_seen": 915996672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648846539618856,
      "loss": 2.9633,
      "theoretical_loss": 3.6800129401139685,
      "tokens_seen": 916062208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036487462387161485,
      "loss": 2.4687,
      "theoretical_loss": 3.67998747891661,
      "tokens_seen": 916127744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648645937813441,
      "loss": 2.6901,
      "theoretical_loss": 3.6799620200505223,
      "tokens_seen": 916193280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648545636910732,
      "loss": 2.4379,
      "theoretical_loss": 3.6799365635153256,
      "tokens_seen": 916258816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036484453360080245,
      "loss": 2.813,
      "theoretical_loss": 3.6799111093106402,
      "tokens_seen": 916324352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648345035105316,
      "loss": 2.7068,
      "theoretical_loss": 3.6798856574360856,
      "tokens_seen": 916389888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648244734202608,
      "loss": 2.9076,
      "theoretical_loss": 3.6798602078912817,
      "tokens_seen": 916455424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036481444332999,
      "loss": 2.7046,
      "theoretical_loss": 3.679834760675849,
      "tokens_seen": 916520960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648044132397192,
      "loss": 2.6643,
      "theoretical_loss": 3.679809315789408,
      "tokens_seen": 916586496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036479438314944836,
      "loss": 2.9033,
      "theoretical_loss": 3.679783873231579,
      "tokens_seen": 916652032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036478435305917754,
      "loss": 2.7993,
      "theoretical_loss": 3.679758433001982,
      "tokens_seen": 916717568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647743229689067,
      "loss": 2.7358,
      "theoretical_loss": 3.6797329951002373,
      "tokens_seen": 916783104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036476429287863595,
      "loss": 2.7013,
      "theoretical_loss": 3.6797075595259665,
      "tokens_seen": 916848640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647542627883651,
      "loss": 2.4434,
      "theoretical_loss": 3.6796821262787898,
      "tokens_seen": 916914176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647442326980943,
      "loss": 2.5695,
      "theoretical_loss": 3.679656695358328,
      "tokens_seen": 916979712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647342026078235,
      "loss": 2.8252,
      "theoretical_loss": 3.679631266764201,
      "tokens_seen": 917045248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647241725175527,
      "loss": 2.8018,
      "theoretical_loss": 3.6796058404960315,
      "tokens_seen": 917110784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036471414242728186,
      "loss": 2.804,
      "theoretical_loss": 3.6795804165534394,
      "tokens_seen": 917176320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036470411233701104,
      "loss": 2.869,
      "theoretical_loss": 3.6795549949360464,
      "tokens_seen": 917241856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646940822467402,
      "loss": 2.7698,
      "theoretical_loss": 3.6795295756434734,
      "tokens_seen": 917307392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036468405215646946,
      "loss": 2.5769,
      "theoretical_loss": 3.679504158675342,
      "tokens_seen": 917372928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646740220661986,
      "loss": 2.6722,
      "theoretical_loss": 3.6794787440312735,
      "tokens_seen": 917438464
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 513081,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8393032550811768,
      "objective/train/theoretical_loss": 3.679453331710889,
      "objective/train/tokens_used": 937964000,
      "theoretical_loss": 3.679453331710889,
      "tokens_seen": 917504000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646639919759278,
      "loss": 2.6943,
      "theoretical_loss": 3.679453331710889,
      "tokens_seen": 917504000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036465396188565695,
      "loss": 3.0041,
      "theoretical_loss": 3.6794279217138106,
      "tokens_seen": 917569536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646439317953862,
      "loss": 2.7193,
      "theoretical_loss": 3.6794025140396602,
      "tokens_seen": 917635072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036463390170511536,
      "loss": 2.8523,
      "theoretical_loss": 3.6793771086880587,
      "tokens_seen": 917700608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036462387161484454,
      "loss": 2.6422,
      "theoretical_loss": 3.6793517056586285,
      "tokens_seen": 917766144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646138415245737,
      "loss": 2.8937,
      "theoretical_loss": 3.6793263049509917,
      "tokens_seen": 917831680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036460381143430296,
      "loss": 2.6182,
      "theoretical_loss": 3.67930090656477,
      "tokens_seen": 917897216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645937813440321,
      "loss": 2.6326,
      "theoretical_loss": 3.6792755104995853,
      "tokens_seen": 917962752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645837512537613,
      "loss": 2.6397,
      "theoretical_loss": 3.6792501167550604,
      "tokens_seen": 918028288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036457372116349045,
      "loss": 2.5801,
      "theoretical_loss": 3.679224725330818,
      "tokens_seen": 918093824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645636910732197,
      "loss": 2.7532,
      "theoretical_loss": 3.679199336226479,
      "tokens_seen": 918159360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036455366098294887,
      "loss": 2.7627,
      "theoretical_loss": 3.679173949441667,
      "tokens_seen": 918224896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036454363089267805,
      "loss": 2.7622,
      "theoretical_loss": 3.6791485649760043,
      "tokens_seen": 918290432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036453360080240723,
      "loss": 2.7045,
      "theoretical_loss": 3.6791231828291133,
      "tokens_seen": 918355968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645235707121364,
      "loss": 2.7075,
      "theoretical_loss": 3.679097803000617,
      "tokens_seen": 918421504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645135406218656,
      "loss": 2.8868,
      "theoretical_loss": 3.6790724254901384,
      "tokens_seen": 918487040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645035105315948,
      "loss": 2.8046,
      "theoretical_loss": 3.6790470502973,
      "tokens_seen": 918552576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036449348044132395,
      "loss": 2.6042,
      "theoretical_loss": 3.679021677421725,
      "tokens_seen": 918618112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644834503510532,
      "loss": 2.8111,
      "theoretical_loss": 3.678996306863037,
      "tokens_seen": 918683648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644734202607823,
      "loss": 2.713,
      "theoretical_loss": 3.678970938620858,
      "tokens_seen": 918749184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036446339017051155,
      "loss": 2.6038,
      "theoretical_loss": 3.678945572694812,
      "tokens_seen": 918814720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036445336008024073,
      "loss": 2.6464,
      "theoretical_loss": 3.6789202090845223,
      "tokens_seen": 918880256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644433299899699,
      "loss": 2.7903,
      "theoretical_loss": 3.6788948477896124,
      "tokens_seen": 918945792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644332998996991,
      "loss": 2.7264,
      "theoretical_loss": 3.678869488809706,
      "tokens_seen": 919011328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036442326980942833,
      "loss": 2.5116,
      "theoretical_loss": 3.6788441321444267,
      "tokens_seen": 919076864
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 514023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6861374378204346,
      "objective/train/theoretical_loss": 3.6788187777933974,
      "objective/train/tokens_used": 939602400,
      "theoretical_loss": 3.6788187777933974,
      "tokens_seen": 919142400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036441323971915746,
      "loss": 2.5727,
      "theoretical_loss": 3.6788187777933974,
      "tokens_seen": 919142400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644032096288867,
      "loss": 2.7196,
      "theoretical_loss": 3.678793425756243,
      "tokens_seen": 919207936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643931795386158,
      "loss": 2.6527,
      "theoretical_loss": 3.6787680760325863,
      "tokens_seen": 919273472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036438314944834505,
      "loss": 2.7031,
      "theoretical_loss": 3.678742728622052,
      "tokens_seen": 919339008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036437311935807424,
      "loss": 2.7691,
      "theoretical_loss": 3.6787173835242646,
      "tokens_seen": 919404544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643630892678034,
      "loss": 2.6177,
      "theoretical_loss": 3.6786920407388473,
      "tokens_seen": 919470080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643530591775326,
      "loss": 2.7784,
      "theoretical_loss": 3.678666700265425,
      "tokens_seen": 919535616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643430290872618,
      "loss": 2.586,
      "theoretical_loss": 3.678641362103621,
      "tokens_seen": 919601152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036433299899699096,
      "loss": 2.5905,
      "theoretical_loss": 3.678616026253061,
      "tokens_seen": 919666688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643229689067202,
      "loss": 2.6251,
      "theoretical_loss": 3.6785906927133687,
      "tokens_seen": 919732224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643129388164493,
      "loss": 2.748,
      "theoretical_loss": 3.678565361484169,
      "tokens_seen": 919797760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036430290872617856,
      "loss": 2.8521,
      "theoretical_loss": 3.6785400325650865,
      "tokens_seen": 919863296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642928786359077,
      "loss": 2.71,
      "theoretical_loss": 3.678514705955746,
      "tokens_seen": 919928832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642828485456369,
      "loss": 2.6943,
      "theoretical_loss": 3.678489381655772,
      "tokens_seen": 919994368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642728184553661,
      "loss": 2.9214,
      "theoretical_loss": 3.6784640596647904,
      "tokens_seen": 920059904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642627883650953,
      "loss": 2.8038,
      "theoretical_loss": 3.6784387399824254,
      "tokens_seen": 920125440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036425275827482446,
      "loss": 2.9118,
      "theoretical_loss": 3.678413422608302,
      "tokens_seen": 920190976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642427281845537,
      "loss": 2.4233,
      "theoretical_loss": 3.6783881075420455,
      "tokens_seen": 920256512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642326980942828,
      "loss": 2.7865,
      "theoretical_loss": 3.6783627947832818,
      "tokens_seen": 920322048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036422266800401206,
      "loss": 2.6493,
      "theoretical_loss": 3.678337484331636,
      "tokens_seen": 920387584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642126379137412,
      "loss": 2.8862,
      "theoretical_loss": 3.6783121761867332,
      "tokens_seen": 920453120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642026078234704,
      "loss": 2.6877,
      "theoretical_loss": 3.678286870348199,
      "tokens_seen": 920518656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641925777331996,
      "loss": 2.7539,
      "theoretical_loss": 3.6782615668156593,
      "tokens_seen": 920584192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641825476429288,
      "loss": 2.793,
      "theoretical_loss": 3.67823626558874,
      "tokens_seen": 920649728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036417251755265797,
      "loss": 2.5625,
      "theoretical_loss": 3.6782109666670664,
      "tokens_seen": 920715264
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 514680,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3389620780944824,
      "objective/train/theoretical_loss": 3.6781856700502646,
      "objective/train/tokens_used": 941240800,
      "theoretical_loss": 3.6781856700502646,
      "tokens_seen": 920780800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036416248746238715,
      "loss": 2.6741,
      "theoretical_loss": 3.6781856700502646,
      "tokens_seen": 920780800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036415245737211633,
      "loss": 2.5103,
      "theoretical_loss": 3.678160375737961,
      "tokens_seen": 920846336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036414242728184556,
      "loss": 2.7652,
      "theoretical_loss": 3.678135083729781,
      "tokens_seen": 920911872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641323971915747,
      "loss": 2.6444,
      "theoretical_loss": 3.6781097940253513,
      "tokens_seen": 920977408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641223671013039,
      "loss": 2.6945,
      "theoretical_loss": 3.6780845066242978,
      "tokens_seen": 921042944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036411233701103316,
      "loss": 2.9332,
      "theoretical_loss": 3.678059221526247,
      "tokens_seen": 921108480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641023069207623,
      "loss": 2.614,
      "theoretical_loss": 3.678033938730825,
      "tokens_seen": 921174016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640922768304915,
      "loss": 2.7244,
      "theoretical_loss": 3.6780086582376588,
      "tokens_seen": 921239552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036408224674022065,
      "loss": 2.7352,
      "theoretical_loss": 3.677983380046375,
      "tokens_seen": 921305088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640722166499499,
      "loss": 2.4935,
      "theoretical_loss": 3.6779581041566,
      "tokens_seen": 921370624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036406218655967907,
      "loss": 2.4393,
      "theoretical_loss": 3.6779328305679604,
      "tokens_seen": 921436160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036405215646940825,
      "loss": 2.6316,
      "theoretical_loss": 3.6779075592800834,
      "tokens_seen": 921501696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036404212637913743,
      "loss": 2.7714,
      "theoretical_loss": 3.677882290292596,
      "tokens_seen": 921567232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640320962888666,
      "loss": 2.6272,
      "theoretical_loss": 3.677857023605125,
      "tokens_seen": 921632768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640220661985958,
      "loss": 2.7397,
      "theoretical_loss": 3.6778317592172973,
      "tokens_seen": 921698304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364012036108325,
      "loss": 2.6956,
      "theoretical_loss": 3.677806497128741,
      "tokens_seen": 921763840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036400200601805415,
      "loss": 2.8998,
      "theoretical_loss": 3.677781237339083,
      "tokens_seen": 921829376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639919759277834,
      "loss": 2.5165,
      "theoretical_loss": 3.6777559798479498,
      "tokens_seen": 921894912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639819458375125,
      "loss": 2.685,
      "theoretical_loss": 3.67773072465497,
      "tokens_seen": 921960448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036397191574724175,
      "loss": 2.6519,
      "theoretical_loss": 3.677705471759771,
      "tokens_seen": 922025984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036396188565697093,
      "loss": 2.6,
      "theoretical_loss": 3.6776802211619795,
      "tokens_seen": 922091520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639518555667001,
      "loss": 2.7074,
      "theoretical_loss": 3.677654972861224,
      "tokens_seen": 922157056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639418254764293,
      "loss": 2.7473,
      "theoretical_loss": 3.6776297268571323,
      "tokens_seen": 922222592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036393179538615853,
      "loss": 2.8253,
      "theoretical_loss": 3.6776044831493326,
      "tokens_seen": 922288128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036392176529588766,
      "loss": 2.7029,
      "theoretical_loss": 3.677579241737452,
      "tokens_seen": 922353664
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 515999,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7671639919281006,
      "objective/train/theoretical_loss": 3.677554002621119,
      "objective/train/tokens_used": 942879200,
      "theoretical_loss": 3.677554002621119,
      "tokens_seen": 922419200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003639117352056169,
      "loss": 2.6393,
      "theoretical_loss": 3.677554002621119,
      "tokens_seen": 922419200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363901705115346,
      "loss": 2.7322,
      "theoretical_loss": 3.6775287657999627,
      "tokens_seen": 922484736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036389167502507525,
      "loss": 2.7465,
      "theoretical_loss": 3.6775035312736097,
      "tokens_seen": 922550272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036388164493480444,
      "loss": 2.7354,
      "theoretical_loss": 3.67747829904169,
      "tokens_seen": 922615808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638716148445336,
      "loss": 2.7393,
      "theoretical_loss": 3.67745306910383,
      "tokens_seen": 922681344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638615847542628,
      "loss": 2.6674,
      "theoretical_loss": 3.67742784145966,
      "tokens_seen": 922746880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363851554663992,
      "loss": 2.9504,
      "theoretical_loss": 3.6774026161088083,
      "tokens_seen": 922812416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036384152457372116,
      "loss": 2.645,
      "theoretical_loss": 3.6773773930509024,
      "tokens_seen": 922877952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638314944834504,
      "loss": 2.6187,
      "theoretical_loss": 3.6773521722855724,
      "tokens_seen": 922943488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638214643931795,
      "loss": 2.6137,
      "theoretical_loss": 3.6773269538124467,
      "tokens_seen": 923009024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036381143430290876,
      "loss": 2.5773,
      "theoretical_loss": 3.677301737631154,
      "tokens_seen": 923074560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638014042126379,
      "loss": 2.6507,
      "theoretical_loss": 3.677276523741323,
      "tokens_seen": 923140096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637913741223671,
      "loss": 2.8147,
      "theoretical_loss": 3.6772513121425847,
      "tokens_seen": 923205632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637813440320963,
      "loss": 2.9722,
      "theoretical_loss": 3.677226102834566,
      "tokens_seen": 923271168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637713139418255,
      "loss": 2.8759,
      "theoretical_loss": 3.677200895816897,
      "tokens_seen": 923336704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036376128385155466,
      "loss": 2.895,
      "theoretical_loss": 3.6771756910892077,
      "tokens_seen": 923402240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637512537612839,
      "loss": 2.7822,
      "theoretical_loss": 3.677150488651127,
      "tokens_seen": 923467776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363741223671013,
      "loss": 2.752,
      "theoretical_loss": 3.6771252885022845,
      "tokens_seen": 923533312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036373119358074226,
      "loss": 2.7301,
      "theoretical_loss": 3.6771000906423095,
      "tokens_seen": 923598848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637211634904714,
      "loss": 2.8236,
      "theoretical_loss": 3.6770748950708323,
      "tokens_seen": 923664384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637111334002006,
      "loss": 2.8326,
      "theoretical_loss": 3.6770497017874826,
      "tokens_seen": 923729920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637011033099298,
      "loss": 2.6637,
      "theoretical_loss": 3.67702451079189,
      "tokens_seen": 923795456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363691073219659,
      "loss": 2.5562,
      "theoretical_loss": 3.6769993220836845,
      "tokens_seen": 923860992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036368104312938817,
      "loss": 2.7717,
      "theoretical_loss": 3.6769741356624968,
      "tokens_seen": 923926528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036367101303911735,
      "loss": 2.9733,
      "theoretical_loss": 3.676948951527956,
      "tokens_seen": 923992064
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 516502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.229719638824463,
      "objective/train/theoretical_loss": 3.6769237696796933,
      "objective/train/tokens_used": 944517600,
      "theoretical_loss": 3.6769237696796933,
      "tokens_seen": 924057600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036366098294884653,
      "loss": 2.9243,
      "theoretical_loss": 3.6769237696796933,
      "tokens_seen": 924057600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036365095285857576,
      "loss": 2.5783,
      "theoretical_loss": 3.676898590117338,
      "tokens_seen": 924123136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636409227683049,
      "loss": 2.8002,
      "theoretical_loss": 3.6768734128405214,
      "tokens_seen": 924188672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636308926780341,
      "loss": 2.5052,
      "theoretical_loss": 3.6768482378488745,
      "tokens_seen": 924254208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036362086258776325,
      "loss": 2.8707,
      "theoretical_loss": 3.6768230651420257,
      "tokens_seen": 924319744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636108324974925,
      "loss": 2.6332,
      "theoretical_loss": 3.6767978947196083,
      "tokens_seen": 924385280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036360080240722167,
      "loss": 2.8033,
      "theoretical_loss": 3.676772726581251,
      "tokens_seen": 924450816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036359077231695085,
      "loss": 2.6453,
      "theoretical_loss": 3.676747560726586,
      "tokens_seen": 924516352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036358074222668003,
      "loss": 2.7782,
      "theoretical_loss": 3.6767223971552427,
      "tokens_seen": 924581888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036357071213640927,
      "loss": 2.6504,
      "theoretical_loss": 3.676697235866854,
      "tokens_seen": 924647424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635606820461384,
      "loss": 2.8365,
      "theoretical_loss": 3.67667207686105,
      "tokens_seen": 924712960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036355065195586763,
      "loss": 2.769,
      "theoretical_loss": 3.676646920137461,
      "tokens_seen": 924778496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036354062186559676,
      "loss": 2.5838,
      "theoretical_loss": 3.67662176569572,
      "tokens_seen": 924844032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363530591775326,
      "loss": 2.4492,
      "theoretical_loss": 3.6765966135354575,
      "tokens_seen": 924909568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036352056168505517,
      "loss": 2.735,
      "theoretical_loss": 3.676571463656305,
      "tokens_seen": 924975104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036351053159478435,
      "loss": 2.6674,
      "theoretical_loss": 3.6765463160578937,
      "tokens_seen": 925040640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036350050150451353,
      "loss": 2.4657,
      "theoretical_loss": 3.676521170739856,
      "tokens_seen": 925106176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634904714142427,
      "loss": 2.6863,
      "theoretical_loss": 3.676496027701823,
      "tokens_seen": 925171712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634804413239719,
      "loss": 2.806,
      "theoretical_loss": 3.6764708869434264,
      "tokens_seen": 925237248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036347041123370113,
      "loss": 2.6572,
      "theoretical_loss": 3.676445748464298,
      "tokens_seen": 925302784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036346038114343026,
      "loss": 2.612,
      "theoretical_loss": 3.67642061226407,
      "tokens_seen": 925368320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634503510531595,
      "loss": 2.6244,
      "theoretical_loss": 3.676395478342375,
      "tokens_seen": 925433856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634403209628886,
      "loss": 2.7324,
      "theoretical_loss": 3.676370346698844,
      "tokens_seen": 925499392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036343029087261786,
      "loss": 2.6495,
      "theoretical_loss": 3.67634521733311,
      "tokens_seen": 925564928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036342026078234704,
      "loss": 2.6261,
      "theoretical_loss": 3.6763200902448054,
      "tokens_seen": 925630464
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 517571,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.828040361404419,
      "objective/train/theoretical_loss": 3.676294965433562,
      "objective/train/tokens_used": 946156000,
      "theoretical_loss": 3.676294965433562,
      "tokens_seen": 925696000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634102306920762,
      "loss": 2.7145,
      "theoretical_loss": 3.676294965433562,
      "tokens_seen": 925696000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634002006018054,
      "loss": 2.8367,
      "theoretical_loss": 3.6762698428990124,
      "tokens_seen": 925761536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036339017051153464,
      "loss": 2.6012,
      "theoretical_loss": 3.6762447226407895,
      "tokens_seen": 925827072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036338014042126376,
      "loss": 2.668,
      "theoretical_loss": 3.6762196046585256,
      "tokens_seen": 925892608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363370110330993,
      "loss": 2.8896,
      "theoretical_loss": 3.676194488951853,
      "tokens_seen": 925958144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633600802407222,
      "loss": 2.8035,
      "theoretical_loss": 3.6761693755204057,
      "tokens_seen": 926023680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036335005015045136,
      "loss": 2.8345,
      "theoretical_loss": 3.676144264363815,
      "tokens_seen": 926089216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633400200601806,
      "loss": 2.8302,
      "theoretical_loss": 3.6761191554817154,
      "tokens_seen": 926154752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633299899699097,
      "loss": 2.5149,
      "theoretical_loss": 3.6760940488737397,
      "tokens_seen": 926220288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036331995987963896,
      "loss": 2.4549,
      "theoretical_loss": 3.6760689445395203,
      "tokens_seen": 926285824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633099297893681,
      "loss": 2.7747,
      "theoretical_loss": 3.6760438424786903,
      "tokens_seen": 926351360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632998996990973,
      "loss": 2.7876,
      "theoretical_loss": 3.676018742690884,
      "tokens_seen": 926416896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632898696088265,
      "loss": 2.7019,
      "theoretical_loss": 3.6759936451757342,
      "tokens_seen": 926482432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632798395185557,
      "loss": 2.8405,
      "theoretical_loss": 3.675968549932875,
      "tokens_seen": 926547968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036326980942828486,
      "loss": 2.6827,
      "theoretical_loss": 3.6759434569619387,
      "tokens_seen": 926613504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632597793380141,
      "loss": 2.4907,
      "theoretical_loss": 3.67591836626256,
      "tokens_seen": 926679040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632497492477432,
      "loss": 2.8732,
      "theoretical_loss": 3.6758932778343727,
      "tokens_seen": 926744576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036323971915747246,
      "loss": 2.6314,
      "theoretical_loss": 3.6758681916770097,
      "tokens_seen": 926810112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632296890672016,
      "loss": 2.7107,
      "theoretical_loss": 3.675843107790106,
      "tokens_seen": 926875648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632196589769308,
      "loss": 2.7786,
      "theoretical_loss": 3.675818026173295,
      "tokens_seen": 926941184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036320962888666,
      "loss": 2.6493,
      "theoretical_loss": 3.6757929468262107,
      "tokens_seen": 927006720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631995987963892,
      "loss": 2.5168,
      "theoretical_loss": 3.6757678697484875,
      "tokens_seen": 927072256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036318956870611837,
      "loss": 2.7334,
      "theoretical_loss": 3.67574279493976,
      "tokens_seen": 927137792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036317953861584755,
      "loss": 2.6639,
      "theoretical_loss": 3.6757177223996615,
      "tokens_seen": 927203328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036316950852557673,
      "loss": 2.7426,
      "theoretical_loss": 3.675692652127827,
      "tokens_seen": 927268864
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 518862,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0217199325561523,
      "objective/train/theoretical_loss": 3.6756675841238913,
      "objective/train/tokens_used": 947794400,
      "theoretical_loss": 3.6756675841238913,
      "tokens_seen": 927334400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036315947843530596,
      "loss": 2.8993,
      "theoretical_loss": 3.6756675841238913,
      "tokens_seen": 927334400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631494483450351,
      "loss": 2.385,
      "theoretical_loss": 3.6756425183874892,
      "tokens_seen": 927399936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631394182547643,
      "loss": 2.7425,
      "theoretical_loss": 3.675617454918254,
      "tokens_seen": 927465472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036312938816449345,
      "loss": 2.7258,
      "theoretical_loss": 3.675592393715822,
      "tokens_seen": 927531008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631193580742227,
      "loss": 2.6855,
      "theoretical_loss": 3.675567334779827,
      "tokens_seen": 927596544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036310932798395187,
      "loss": 3.0311,
      "theoretical_loss": 3.675542278109905,
      "tokens_seen": 927662080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036309929789368105,
      "loss": 2.6216,
      "theoretical_loss": 3.6755172237056897,
      "tokens_seen": 927727616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036308926780341023,
      "loss": 2.2776,
      "theoretical_loss": 3.675492171566817,
      "tokens_seen": 927793152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036307923771313947,
      "loss": 2.6105,
      "theoretical_loss": 3.675467121692922,
      "tokens_seen": 927858688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003630692076228686,
      "loss": 2.8852,
      "theoretical_loss": 3.67544207408364,
      "tokens_seen": 927924224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036305917753259783,
      "loss": 2.6106,
      "theoretical_loss": 3.675417028738606,
      "tokens_seen": 927989760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036304914744232696,
      "loss": 2.7,
      "theoretical_loss": 3.675391985657456,
      "tokens_seen": 928055296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003630391173520562,
      "loss": 2.6264,
      "theoretical_loss": 3.675366944839825,
      "tokens_seen": 928120832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036302908726178537,
      "loss": 2.6397,
      "theoretical_loss": 3.6753419062853485,
      "tokens_seen": 928186368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036301905717151455,
      "loss": 2.5242,
      "theoretical_loss": 3.6753168699936625,
      "tokens_seen": 928251904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036300902708124374,
      "loss": 2.6,
      "theoretical_loss": 3.675291835964403,
      "tokens_seen": 928317440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629989969909729,
      "loss": 2.7545,
      "theoretical_loss": 3.675266804197206,
      "tokens_seen": 928382976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629889669007021,
      "loss": 2.8854,
      "theoretical_loss": 3.6752417746917065,
      "tokens_seen": 928448512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036297893681043133,
      "loss": 2.6897,
      "theoretical_loss": 3.675216747447541,
      "tokens_seen": 928514048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036296890672016046,
      "loss": 2.7338,
      "theoretical_loss": 3.675191722464346,
      "tokens_seen": 928579584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629588766298897,
      "loss": 2.6159,
      "theoretical_loss": 3.6751666997417574,
      "tokens_seen": 928645120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629488465396188,
      "loss": 2.8166,
      "theoretical_loss": 3.6751416792794114,
      "tokens_seen": 928710656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036293881644934806,
      "loss": 2.6078,
      "theoretical_loss": 3.6751166610769443,
      "tokens_seen": 928776192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036292878635907724,
      "loss": 2.8892,
      "theoretical_loss": 3.6750916451339926,
      "tokens_seen": 928841728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629187562688064,
      "loss": 2.4984,
      "theoretical_loss": 3.6750666314501927,
      "tokens_seen": 928907264
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 519491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.917207717895508,
      "objective/train/theoretical_loss": 3.675041620025182,
      "objective/train/tokens_used": 949432800,
      "theoretical_loss": 3.675041620025182,
      "tokens_seen": 928972800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629087261785356,
      "loss": 2.6886,
      "theoretical_loss": 3.675041620025182,
      "tokens_seen": 928972800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036289869608826484,
      "loss": 2.6166,
      "theoretical_loss": 3.675016610858596,
      "tokens_seen": 929038336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036288866599799396,
      "loss": 2.7116,
      "theoretical_loss": 3.6749916039500725,
      "tokens_seen": 929103872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628786359077232,
      "loss": 2.5936,
      "theoretical_loss": 3.674966599299248,
      "tokens_seen": 929169408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628686058174523,
      "loss": 2.6567,
      "theoretical_loss": 3.674941596905759,
      "tokens_seen": 929234944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036285857572718156,
      "loss": 2.7593,
      "theoretical_loss": 3.6749165967692434,
      "tokens_seen": 929300480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036284854563691074,
      "loss": 2.6485,
      "theoretical_loss": 3.674891598889338,
      "tokens_seen": 929366016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628385155466399,
      "loss": 2.7085,
      "theoretical_loss": 3.674866603265679,
      "tokens_seen": 929431552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628284854563691,
      "loss": 2.5491,
      "theoretical_loss": 3.674841609897905,
      "tokens_seen": 929497088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628184553660983,
      "loss": 2.6976,
      "theoretical_loss": 3.674816618785653,
      "tokens_seen": 929562624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036280842527582747,
      "loss": 2.5553,
      "theoretical_loss": 3.67479162992856,
      "tokens_seen": 929628160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627983951855567,
      "loss": 2.8656,
      "theoretical_loss": 3.6747666433262642,
      "tokens_seen": 929693696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036278836509528583,
      "loss": 2.6386,
      "theoretical_loss": 3.674741658978403,
      "tokens_seen": 929759232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036277833500501506,
      "loss": 2.6119,
      "theoretical_loss": 3.6747166768846142,
      "tokens_seen": 929824768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036276830491474424,
      "loss": 2.6403,
      "theoretical_loss": 3.674691697044535,
      "tokens_seen": 929890304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627582748244734,
      "loss": 2.7474,
      "theoretical_loss": 3.6746667194578038,
      "tokens_seen": 929955840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627482447342026,
      "loss": 2.6571,
      "theoretical_loss": 3.6746417441240586,
      "tokens_seen": 930021376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627382146439318,
      "loss": 2.82,
      "theoretical_loss": 3.674616771042937,
      "tokens_seen": 930086912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036272818455366097,
      "loss": 2.8842,
      "theoretical_loss": 3.6745918002140776,
      "tokens_seen": 930152448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627181544633902,
      "loss": 2.7982,
      "theoretical_loss": 3.6745668316371183,
      "tokens_seen": 930217984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036270812437311933,
      "loss": 2.6886,
      "theoretical_loss": 3.6745418653116975,
      "tokens_seen": 930283520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036269809428284857,
      "loss": 2.7248,
      "theoretical_loss": 3.6745169012374537,
      "tokens_seen": 930349056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626880641925777,
      "loss": 2.6175,
      "theoretical_loss": 3.6744919394140254,
      "tokens_seen": 930414592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036267803410230693,
      "loss": 2.8798,
      "theoretical_loss": 3.6744669798410508,
      "tokens_seen": 930480128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626680040120361,
      "loss": 2.7445,
      "theoretical_loss": 3.6744420225181686,
      "tokens_seen": 930545664
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 520743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.064671277999878,
      "objective/train/theoretical_loss": 3.6744170674450176,
      "objective/train/tokens_used": 951071200,
      "theoretical_loss": 3.6744170674450176,
      "tokens_seen": 930611200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626579739217653,
      "loss": 2.841,
      "theoretical_loss": 3.6744170674450176,
      "tokens_seen": 930611200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036264794383149447,
      "loss": 2.5725,
      "theoretical_loss": 3.6743921146212366,
      "tokens_seen": 930676736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036263791374122365,
      "loss": 2.7547,
      "theoretical_loss": 3.6743671640464646,
      "tokens_seen": 930742272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036262788365095283,
      "loss": 2.7484,
      "theoretical_loss": 3.6743422157203405,
      "tokens_seen": 930807808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036261785356068207,
      "loss": 2.5509,
      "theoretical_loss": 3.6743172696425033,
      "tokens_seen": 930873344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036260782347041125,
      "loss": 2.5975,
      "theoretical_loss": 3.6742923258125915,
      "tokens_seen": 930938880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036259779338014043,
      "loss": 2.8421,
      "theoretical_loss": 3.674267384230246,
      "tokens_seen": 931004416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036258776328986967,
      "loss": 2.928,
      "theoretical_loss": 3.6742424448951043,
      "tokens_seen": 931069952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625777331995988,
      "loss": 2.7928,
      "theoretical_loss": 3.6742175078068064,
      "tokens_seen": 931135488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036256770310932803,
      "loss": 2.6882,
      "theoretical_loss": 3.674192572964992,
      "tokens_seen": 931201024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036255767301905716,
      "loss": 2.8744,
      "theoretical_loss": 3.6741676403693004,
      "tokens_seen": 931266560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625476429287864,
      "loss": 2.51,
      "theoretical_loss": 3.6741427100193715,
      "tokens_seen": 931332096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625376128385156,
      "loss": 2.8063,
      "theoretical_loss": 3.6741177819148447,
      "tokens_seen": 931397632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036252758274824475,
      "loss": 2.8343,
      "theoretical_loss": 3.67409285605536,
      "tokens_seen": 931463168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036251755265797394,
      "loss": 2.771,
      "theoretical_loss": 3.674067932440557,
      "tokens_seen": 931528704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625075225677031,
      "loss": 2.5716,
      "theoretical_loss": 3.674043011070076,
      "tokens_seen": 931594240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624974924774323,
      "loss": 2.7609,
      "theoretical_loss": 3.674018091943557,
      "tokens_seen": 931659776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036248746238716153,
      "loss": 2.7397,
      "theoretical_loss": 3.67399317506064,
      "tokens_seen": 931725312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036247743229689066,
      "loss": 2.6042,
      "theoretical_loss": 3.673968260420965,
      "tokens_seen": 931790848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624674022066199,
      "loss": 2.5878,
      "theoretical_loss": 3.6739433480241726,
      "tokens_seen": 931856384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362457372116349,
      "loss": 2.8429,
      "theoretical_loss": 3.6739184378699026,
      "tokens_seen": 931921920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036244734202607826,
      "loss": 2.447,
      "theoretical_loss": 3.6738935299577964,
      "tokens_seen": 931987456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036243731193580744,
      "loss": 2.7397,
      "theoretical_loss": 3.6738686242874943,
      "tokens_seen": 932052992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624272818455366,
      "loss": 2.5438,
      "theoretical_loss": 3.6738437208586365,
      "tokens_seen": 932118528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624172517552658,
      "loss": 2.8547,
      "theoretical_loss": 3.6738188196708634,
      "tokens_seen": 932184064
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 521434,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9683523178100586,
      "objective/train/theoretical_loss": 3.673793920723817,
      "objective/train/tokens_used": 952709600,
      "theoretical_loss": 3.673793920723817,
      "tokens_seen": 932249600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036240722166499504,
      "loss": 2.8287,
      "theoretical_loss": 3.673793920723817,
      "tokens_seen": 932249600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036239719157472416,
      "loss": 2.4303,
      "theoretical_loss": 3.673769024017137,
      "tokens_seen": 932315136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623871614844534,
      "loss": 2.56,
      "theoretical_loss": 3.6737441295504647,
      "tokens_seen": 932380672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623771313941825,
      "loss": 2.6744,
      "theoretical_loss": 3.6737192373234415,
      "tokens_seen": 932446208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036236710130391176,
      "loss": 2.8192,
      "theoretical_loss": 3.6736943473357084,
      "tokens_seen": 932511744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036235707121364094,
      "loss": 2.8758,
      "theoretical_loss": 3.673669459586906,
      "tokens_seen": 932577280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623470411233701,
      "loss": 2.6161,
      "theoretical_loss": 3.6736445740766763,
      "tokens_seen": 932642816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623370110330993,
      "loss": 2.5468,
      "theoretical_loss": 3.673619690804661,
      "tokens_seen": 932708352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623269809428285,
      "loss": 2.8673,
      "theoretical_loss": 3.6735948097705,
      "tokens_seen": 932773888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036231695085255767,
      "loss": 2.694,
      "theoretical_loss": 3.6735699309738363,
      "tokens_seen": 932839424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623069207622869,
      "loss": 2.5771,
      "theoretical_loss": 3.6735450544143107,
      "tokens_seen": 932904960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036229689067201603,
      "loss": 2.638,
      "theoretical_loss": 3.6735201800915656,
      "tokens_seen": 932970496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036228686058174526,
      "loss": 2.8063,
      "theoretical_loss": 3.6734953080052417,
      "tokens_seen": 933036032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036227683049147444,
      "loss": 2.7244,
      "theoretical_loss": 3.673470438154982,
      "tokens_seen": 933101568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622668004012036,
      "loss": 2.8428,
      "theoretical_loss": 3.6734455705404283,
      "tokens_seen": 933167104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622567703109328,
      "loss": 2.7789,
      "theoretical_loss": 3.6734207051612215,
      "tokens_seen": 933232640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362246740220662,
      "loss": 2.8115,
      "theoretical_loss": 3.673395842017005,
      "tokens_seen": 933298176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036223671013039117,
      "loss": 2.5103,
      "theoretical_loss": 3.6733709811074204,
      "tokens_seen": 933363712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622266800401204,
      "loss": 2.604,
      "theoretical_loss": 3.6733461224321102,
      "tokens_seen": 933429248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036221664994984953,
      "loss": 2.9288,
      "theoretical_loss": 3.673321265990717,
      "tokens_seen": 933494784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036220661985957877,
      "loss": 2.6005,
      "theoretical_loss": 3.673296411782882,
      "tokens_seen": 933560320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621965897693079,
      "loss": 2.5925,
      "theoretical_loss": 3.6732715598082493,
      "tokens_seen": 933625856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036218655967903713,
      "loss": 2.6181,
      "theoretical_loss": 3.6732467100664605,
      "tokens_seen": 933691392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621765295887663,
      "loss": 2.6131,
      "theoretical_loss": 3.673221862557159,
      "tokens_seen": 933756928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621664994984955,
      "loss": 2.5553,
      "theoretical_loss": 3.6731970172799864,
      "tokens_seen": 933822464
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 522330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3445885181427,
      "objective/train/theoretical_loss": 3.673172174234587,
      "objective/train/tokens_used": 954348000,
      "theoretical_loss": 3.673172174234587,
      "tokens_seen": 933888000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036215646940822467,
      "loss": 2.617,
      "theoretical_loss": 3.673172174234587,
      "tokens_seen": 933888000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036214643931795385,
      "loss": 2.6809,
      "theoretical_loss": 3.673147333420603,
      "tokens_seen": 933953536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036213640922768303,
      "loss": 2.634,
      "theoretical_loss": 3.673122494837677,
      "tokens_seen": 934019072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036212637913741227,
      "loss": 2.6592,
      "theoretical_loss": 3.6730976584854527,
      "tokens_seen": 934084608
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621163490471414,
      "loss": 2.7614,
      "theoretical_loss": 3.6730728243635733,
      "tokens_seen": 934150144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036210631895687063,
      "loss": 2.7846,
      "theoretical_loss": 3.673047992471682,
      "tokens_seen": 934215680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620962888665998,
      "loss": 2.5949,
      "theoretical_loss": 3.6730231628094216,
      "tokens_seen": 934281216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362086258776329,
      "loss": 2.7675,
      "theoretical_loss": 3.672998335376436,
      "tokens_seen": 934346752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620762286860582,
      "loss": 2.8264,
      "theoretical_loss": 3.672973510172369,
      "tokens_seen": 934412288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036206619859578736,
      "loss": 2.6534,
      "theoretical_loss": 3.6729486871968637,
      "tokens_seen": 934477824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036205616850551654,
      "loss": 2.724,
      "theoretical_loss": 3.672923866449564,
      "tokens_seen": 934543360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620461384152458,
      "loss": 2.7642,
      "theoretical_loss": 3.6728990479301133,
      "tokens_seen": 934608896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620361083249749,
      "loss": 2.8175,
      "theoretical_loss": 3.672874231638156,
      "tokens_seen": 934674432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036202607823470414,
      "loss": 2.6981,
      "theoretical_loss": 3.672849417573336,
      "tokens_seen": 934739968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036201604814443326,
      "loss": 2.4681,
      "theoretical_loss": 3.672824605735297,
      "tokens_seen": 934805504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620060180541625,
      "loss": 2.7621,
      "theoretical_loss": 3.6727997961236825,
      "tokens_seen": 934871040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619959879638917,
      "loss": 2.7309,
      "theoretical_loss": 3.6727749887381376,
      "tokens_seen": 934936576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036198595787362086,
      "loss": 2.7253,
      "theoretical_loss": 3.672750183578306,
      "tokens_seen": 935002112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036197592778335004,
      "loss": 2.8102,
      "theoretical_loss": 3.6727253806438327,
      "tokens_seen": 935067648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619658976930792,
      "loss": 2.5928,
      "theoretical_loss": 3.6727005799343613,
      "tokens_seen": 935133184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619558676028084,
      "loss": 2.6811,
      "theoretical_loss": 3.672675781449537,
      "tokens_seen": 935198720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036194583751253764,
      "loss": 2.7066,
      "theoretical_loss": 3.6726509851890032,
      "tokens_seen": 935264256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036193580742226677,
      "loss": 2.7786,
      "theoretical_loss": 3.672626191152406,
      "tokens_seen": 935329792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361925777331996,
      "loss": 2.6312,
      "theoretical_loss": 3.67260139933939,
      "tokens_seen": 935395328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619157472417252,
      "loss": 2.7573,
      "theoretical_loss": 3.6725766097495987,
      "tokens_seen": 935460864
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 523607,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.482036590576172,
      "objective/train/theoretical_loss": 3.672551822382678,
      "objective/train/tokens_used": 955986400,
      "theoretical_loss": 3.672551822382678,
      "tokens_seen": 935526400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036190571715145436,
      "loss": 2.4688,
      "theoretical_loss": 3.672551822382678,
      "tokens_seen": 935526400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036189568706118354,
      "loss": 2.8175,
      "theoretical_loss": 3.672527037238273,
      "tokens_seen": 935591936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618856569709127,
      "loss": 2.8405,
      "theoretical_loss": 3.6725022543160284,
      "tokens_seen": 935657472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618756268806419,
      "loss": 2.688,
      "theoretical_loss": 3.672477473615589,
      "tokens_seen": 935723008
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036186559679037114,
      "loss": 2.5455,
      "theoretical_loss": 3.6724526951366006,
      "tokens_seen": 935788544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618555667001003,
      "loss": 2.9469,
      "theoretical_loss": 3.672427918878708,
      "tokens_seen": 935854080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618455366098295,
      "loss": 2.7401,
      "theoretical_loss": 3.6724031448415575,
      "tokens_seen": 935919616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618355065195587,
      "loss": 2.937,
      "theoretical_loss": 3.672378373024794,
      "tokens_seen": 935985152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036182547642928787,
      "loss": 2.4816,
      "theoretical_loss": 3.672353603428063,
      "tokens_seen": 936050688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618154463390171,
      "loss": 2.674,
      "theoretical_loss": 3.6723288360510096,
      "tokens_seen": 936116224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036180541624874623,
      "loss": 2.4672,
      "theoretical_loss": 3.6723040708932806,
      "tokens_seen": 936181760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036179538615847546,
      "loss": 2.6237,
      "theoretical_loss": 3.672279307954521,
      "tokens_seen": 936247296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036178535606820464,
      "loss": 2.7813,
      "theoretical_loss": 3.6722545472343775,
      "tokens_seen": 936312832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617753259779338,
      "loss": 2.5787,
      "theoretical_loss": 3.672229788732495,
      "tokens_seen": 936378368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361765295887663,
      "loss": 2.676,
      "theoretical_loss": 3.67220503244852,
      "tokens_seen": 936443904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617552657973922,
      "loss": 2.5799,
      "theoretical_loss": 3.6721802783820987,
      "tokens_seen": 936509440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036174523570712137,
      "loss": 2.7577,
      "theoretical_loss": 3.6721555265328774,
      "tokens_seen": 936574976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617352056168506,
      "loss": 2.7343,
      "theoretical_loss": 3.6721307769005023,
      "tokens_seen": 936640512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036172517552657973,
      "loss": 2.6231,
      "theoretical_loss": 3.6721060294846195,
      "tokens_seen": 936706048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036171514543630897,
      "loss": 2.4971,
      "theoretical_loss": 3.6720812842848756,
      "tokens_seen": 936771584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617051153460381,
      "loss": 2.7271,
      "theoretical_loss": 3.6720565413009174,
      "tokens_seen": 936837120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036169508525576733,
      "loss": 2.8992,
      "theoretical_loss": 3.672031800532391,
      "tokens_seen": 936902656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616850551654965,
      "loss": 2.5012,
      "theoretical_loss": 3.6720070619789436,
      "tokens_seen": 936968192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616750250752257,
      "loss": 2.7448,
      "theoretical_loss": 3.671982325640222,
      "tokens_seen": 937033728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036166499498495487,
      "loss": 2.5139,
      "theoretical_loss": 3.6719575915158726,
      "tokens_seen": 937099264
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 524349,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2419161796569824,
      "objective/train/theoretical_loss": 3.6719328596055423,
      "objective/train/tokens_used": 957624800,
      "theoretical_loss": 3.6719328596055423,
      "tokens_seen": 937164800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036165496489468405,
      "loss": 2.5676,
      "theoretical_loss": 3.6719328596055423,
      "tokens_seen": 937164800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036164493480441323,
      "loss": 2.8033,
      "theoretical_loss": 3.6719081299088785,
      "tokens_seen": 937230336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036163490471414247,
      "loss": 2.5401,
      "theoretical_loss": 3.671883402425528,
      "tokens_seen": 937295872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616248746238716,
      "loss": 2.4762,
      "theoretical_loss": 3.671858677155138,
      "tokens_seen": 937361408
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036161484453360083,
      "loss": 2.4731,
      "theoretical_loss": 3.671833954097356,
      "tokens_seen": 937426944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036160481444333,
      "loss": 2.777,
      "theoretical_loss": 3.6718092332518295,
      "tokens_seen": 937492480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036160481444333,
      "loss": 2.7376,
      "theoretical_loss": 3.671784514618205,
      "tokens_seen": 937558016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615947843530592,
      "loss": 2.9631,
      "theoretical_loss": 3.671759798196131,
      "tokens_seen": 937623552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615847542627884,
      "loss": 2.8124,
      "theoretical_loss": 3.6717350839852547,
      "tokens_seen": 937689088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036157472417251756,
      "loss": 2.8745,
      "theoretical_loss": 3.671710371985223,
      "tokens_seen": 937754624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036156469408224674,
      "loss": 2.65,
      "theoretical_loss": 3.6716856621956855,
      "tokens_seen": 937820160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361554663991976,
      "loss": 2.861,
      "theoretical_loss": 3.6716609546162884,
      "tokens_seen": 937885696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615446339017051,
      "loss": 2.6636,
      "theoretical_loss": 3.6716362492466796,
      "tokens_seen": 937951232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036153460381143434,
      "loss": 2.8794,
      "theoretical_loss": 3.6716115460865084,
      "tokens_seen": 938016768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036152457372116346,
      "loss": 2.7265,
      "theoretical_loss": 3.6715868451354217,
      "tokens_seen": 938082304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615145436308927,
      "loss": 2.5481,
      "theoretical_loss": 3.671562146393068,
      "tokens_seen": 938147840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615045135406219,
      "loss": 2.7664,
      "theoretical_loss": 3.671537449859095,
      "tokens_seen": 938213376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036149448345035106,
      "loss": 2.7306,
      "theoretical_loss": 3.6715127555331524,
      "tokens_seen": 938278912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036148445336008024,
      "loss": 2.5884,
      "theoretical_loss": 3.671488063414887,
      "tokens_seen": 938344448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614744232698094,
      "loss": 2.6665,
      "theoretical_loss": 3.671463373503948,
      "tokens_seen": 938409984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614643931795386,
      "loss": 2.906,
      "theoretical_loss": 3.671438685799984,
      "tokens_seen": 938475520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036145436308926784,
      "loss": 2.8406,
      "theoretical_loss": 3.671414000302644,
      "tokens_seen": 938541056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036144433299899697,
      "loss": 2.7746,
      "theoretical_loss": 3.6713893170115757,
      "tokens_seen": 938606592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614343029087262,
      "loss": 2.8227,
      "theoretical_loss": 3.6713646359264285,
      "tokens_seen": 938672128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614242728184554,
      "loss": 2.6755,
      "theoretical_loss": 3.6713399570468512,
      "tokens_seen": 938737664
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 525468,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.555947780609131,
      "objective/train/theoretical_loss": 3.671315280372493,
      "objective/train/tokens_used": 959263200,
      "theoretical_loss": 3.671315280372493,
      "tokens_seen": 938803200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036141424272818456,
      "loss": 2.602,
      "theoretical_loss": 3.671315280372493,
      "tokens_seen": 938803200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036140421263791374,
      "loss": 2.8354,
      "theoretical_loss": 3.6712906059030024,
      "tokens_seen": 938868736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613941825476429,
      "loss": 2.774,
      "theoretical_loss": 3.6712659336380287,
      "tokens_seen": 938934272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613841524573721,
      "loss": 2.714,
      "theoretical_loss": 3.671241263577221,
      "tokens_seen": 938999808
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036137412236710134,
      "loss": 2.7993,
      "theoretical_loss": 3.671216595720229,
      "tokens_seen": 939065344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036136409227683047,
      "loss": 2.8738,
      "theoretical_loss": 3.671191930066702,
      "tokens_seen": 939130880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613540621865597,
      "loss": 2.5938,
      "theoretical_loss": 3.6711672666162887,
      "tokens_seen": 939196416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036134403209628883,
      "loss": 2.7551,
      "theoretical_loss": 3.6711426053686393,
      "tokens_seen": 939261952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036133400200601807,
      "loss": 2.9425,
      "theoretical_loss": 3.6711179463234034,
      "tokens_seen": 939327488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036132397191574725,
      "loss": 2.5775,
      "theoretical_loss": 3.6710932894802304,
      "tokens_seen": 939393024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036131394182547643,
      "loss": 2.737,
      "theoretical_loss": 3.67106863483877,
      "tokens_seen": 939458560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613039117352056,
      "loss": 2.7386,
      "theoretical_loss": 3.6710439823986722,
      "tokens_seen": 939524096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036129388164493484,
      "loss": 2.7826,
      "theoretical_loss": 3.6710193321595868,
      "tokens_seen": 939589632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036128385155466397,
      "loss": 2.5974,
      "theoretical_loss": 3.670994684121164,
      "tokens_seen": 939655168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612738214643932,
      "loss": 2.8526,
      "theoretical_loss": 3.6709700382830537,
      "tokens_seen": 939720704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036126379137412233,
      "loss": 2.3337,
      "theoretical_loss": 3.6709453946449058,
      "tokens_seen": 939786240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036125376128385157,
      "loss": 2.8718,
      "theoretical_loss": 3.6709207532063712,
      "tokens_seen": 939851776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036124373119358075,
      "loss": 2.7124,
      "theoretical_loss": 3.6708961139670997,
      "tokens_seen": 939917312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036123370110330993,
      "loss": 2.8604,
      "theoretical_loss": 3.6708714769267417,
      "tokens_seen": 939982848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612236710130391,
      "loss": 2.9014,
      "theoretical_loss": 3.6708468420849476,
      "tokens_seen": 940048384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612136409227683,
      "loss": 2.5795,
      "theoretical_loss": 3.6708222094413685,
      "tokens_seen": 940113920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612036108324975,
      "loss": 2.7011,
      "theoretical_loss": 3.670797578995655,
      "tokens_seen": 940179456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611935807422267,
      "loss": 2.8098,
      "theoretical_loss": 3.6707729507474567,
      "tokens_seen": 940244992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036118355065195584,
      "loss": 2.5256,
      "theoretical_loss": 3.670748324696426,
      "tokens_seen": 940310528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036117352056168507,
      "loss": 2.6763,
      "theoretical_loss": 3.6707237008422116,
      "tokens_seen": 940376064
    },
    {
      "epoch": 0.28,
      "objective/train/docs_used": 526013,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.817310333251953,
      "objective/train/theoretical_loss": 3.670699079184467,
      "objective/train/tokens_used": 960901600,
      "theoretical_loss": 3.670699079184467,
      "tokens_seen": 940441600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611634904714142,
      "loss": 2.6014,
      "theoretical_loss": 3.670699079184467,
      "tokens_seen": 940441600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036115346038114343,
      "loss": 2.648,
      "theoretical_loss": 3.6706744597228416,
      "tokens_seen": 940507136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611434302908726,
      "loss": 2.6984,
      "theoretical_loss": 3.670649842456987,
      "tokens_seen": 940572672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611334002006018,
      "loss": 2.5761,
      "theoretical_loss": 3.6706252273865543,
      "tokens_seen": 940638208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000361123370110331,
      "loss": 2.6609,
      "theoretical_loss": 3.6706006145111942,
      "tokens_seen": 940703744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611133400200602,
      "loss": 2.7398,
      "theoretical_loss": 3.6705760038305595,
      "tokens_seen": 940769280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611033099297894,
      "loss": 2.6214,
      "theoretical_loss": 3.6705513953443,
      "tokens_seen": 940834816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610932798395186,
      "loss": 2.7006,
      "theoretical_loss": 3.6705267890520688,
      "tokens_seen": 940900352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036108324974924776,
      "loss": 2.687,
      "theoretical_loss": 3.670502184953516,
      "tokens_seen": 940965888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036107321965897694,
      "loss": 2.4898,
      "theoretical_loss": 3.670477583048294,
      "tokens_seen": 941031424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610631895687062,
      "loss": 2.7117,
      "theoretical_loss": 3.670452983336055,
      "tokens_seen": 941096960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610531594784353,
      "loss": 2.6429,
      "theoretical_loss": 3.6704283858164497,
      "tokens_seen": 941162496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036104312938816454,
      "loss": 2.6568,
      "theoretical_loss": 3.6704037904891305,
      "tokens_seen": 941228032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036103309929789366,
      "loss": 2.7248,
      "theoretical_loss": 3.67037919735375,
      "tokens_seen": 941293568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610230692076229,
      "loss": 2.8445,
      "theoretical_loss": 3.6703546064099593,
      "tokens_seen": 941359104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610130391173521,
      "loss": 2.8391,
      "theoretical_loss": 3.6703300176574114,
      "tokens_seen": 941424640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036100300902708126,
      "loss": 2.8193,
      "theoretical_loss": 3.6703054310957577,
      "tokens_seen": 941490176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036099297893681044,
      "loss": 2.5903,
      "theoretical_loss": 3.6702808467246513,
      "tokens_seen": 941555712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609829488465396,
      "loss": 2.6982,
      "theoretical_loss": 3.670256264543744,
      "tokens_seen": 941621248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609729187562688,
      "loss": 2.7354,
      "theoretical_loss": 3.6702316845526886,
      "tokens_seen": 941686784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036096288866599804,
      "loss": 2.869,
      "theoretical_loss": 3.6702071067511373,
      "tokens_seen": 941752320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036095285857572717,
      "loss": 2.8926,
      "theoretical_loss": 3.670182531138743,
      "tokens_seen": 941817856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609428284854564,
      "loss": 2.8551,
      "theoretical_loss": 3.6701579577151584,
      "tokens_seen": 941883392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609327983951856,
      "loss": 2.7154,
      "theoretical_loss": 3.670133386480036,
      "tokens_seen": 941948928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036092276830491476,
      "loss": 2.5979,
      "theoretical_loss": 3.6701088174330287,
      "tokens_seen": 942014464
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 527234,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.644130229949951,
      "objective/train/theoretical_loss": 3.6700842505737903,
      "objective/train/tokens_used": 962540000,
      "theoretical_loss": 3.6700842505737903,
      "tokens_seen": 942080000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036091273821464394,
      "loss": 2.6034,
      "theoretical_loss": 3.6700842505737903,
      "tokens_seen": 942080000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609027081243731,
      "loss": 2.6743,
      "theoretical_loss": 3.6700596859019723,
      "tokens_seen": 942145536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608926780341023,
      "loss": 2.5416,
      "theoretical_loss": 3.670035123417229,
      "tokens_seen": 942211072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036088264794383154,
      "loss": 2.5893,
      "theoretical_loss": 3.670010563119213,
      "tokens_seen": 942276608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036087261785356067,
      "loss": 2.5442,
      "theoretical_loss": 3.669986005007578,
      "tokens_seen": 942342144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608625877632899,
      "loss": 2.6865,
      "theoretical_loss": 3.669961449081977,
      "tokens_seen": 942407680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036085255767301903,
      "loss": 2.372,
      "theoretical_loss": 3.6699368953420635,
      "tokens_seen": 942473216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036084252758274827,
      "loss": 2.467,
      "theoretical_loss": 3.669912343787491,
      "tokens_seen": 942538752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036083249749247745,
      "loss": 2.874,
      "theoretical_loss": 3.669887794417913,
      "tokens_seen": 942604288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036082246740220663,
      "loss": 2.83,
      "theoretical_loss": 3.669863247232983,
      "tokens_seen": 942669824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608124373119358,
      "loss": 2.7182,
      "theoretical_loss": 3.669838702232355,
      "tokens_seen": 942735360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036080240722166505,
      "loss": 2.5017,
      "theoretical_loss": 3.669814159415683,
      "tokens_seen": 942800896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036079237713139417,
      "loss": 2.6341,
      "theoretical_loss": 3.66978961878262,
      "tokens_seen": 942866432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607823470411234,
      "loss": 2.7817,
      "theoretical_loss": 3.6697650803328212,
      "tokens_seen": 942931968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036077231695085253,
      "loss": 2.6504,
      "theoretical_loss": 3.66974054406594,
      "tokens_seen": 942997504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036076228686058177,
      "loss": 2.7501,
      "theoretical_loss": 3.66971600998163,
      "tokens_seen": 943063040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036075225677031095,
      "loss": 2.8004,
      "theoretical_loss": 3.669691478079546,
      "tokens_seen": 943128576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036074222668004013,
      "loss": 2.7057,
      "theoretical_loss": 3.6696669483593425,
      "tokens_seen": 943194112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607321965897693,
      "loss": 2.4509,
      "theoretical_loss": 3.669642420820674,
      "tokens_seen": 943259648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607221664994985,
      "loss": 2.7561,
      "theoretical_loss": 3.669617895463194,
      "tokens_seen": 943325184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607121364092277,
      "loss": 2.3886,
      "theoretical_loss": 3.6695933722865575,
      "tokens_seen": 943390720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607021063189569,
      "loss": 2.7888,
      "theoretical_loss": 3.6695688512904194,
      "tokens_seen": 943456256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036069207622868604,
      "loss": 2.7435,
      "theoretical_loss": 3.669544332474434,
      "tokens_seen": 943521792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036068204613841527,
      "loss": 2.5464,
      "theoretical_loss": 3.669519815838256,
      "tokens_seen": 943587328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606720160481444,
      "loss": 2.7557,
      "theoretical_loss": 3.6694953013815406,
      "tokens_seen": 943652864
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 528141,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.465517044067383,
      "objective/train/theoretical_loss": 3.669470789103942,
      "objective/train/tokens_used": 964178400,
      "theoretical_loss": 3.669470789103942,
      "tokens_seen": 943718400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036066198595787364,
      "loss": 2.7745,
      "theoretical_loss": 3.669470789103942,
      "tokens_seen": 943718400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606519558676028,
      "loss": 2.8102,
      "theoretical_loss": 3.669446279005116,
      "tokens_seen": 943783936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360641925777332,
      "loss": 2.6492,
      "theoretical_loss": 3.6694217710847177,
      "tokens_seen": 943849472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606318956870612,
      "loss": 2.8314,
      "theoretical_loss": 3.6693972653424014,
      "tokens_seen": 943915008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606218655967904,
      "loss": 2.6344,
      "theoretical_loss": 3.669372761777823,
      "tokens_seen": 943980544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036061183550651954,
      "loss": 2.7509,
      "theoretical_loss": 3.669348260390638,
      "tokens_seen": 944046080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606018054162488,
      "loss": 3.0374,
      "theoretical_loss": 3.669323761180501,
      "tokens_seen": 944111616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605917753259779,
      "loss": 2.965,
      "theoretical_loss": 3.6692992641470683,
      "tokens_seen": 944177152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036058174523570714,
      "loss": 2.8973,
      "theoretical_loss": 3.6692747692899945,
      "tokens_seen": 944242688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605717151454363,
      "loss": 2.8169,
      "theoretical_loss": 3.6692502766089357,
      "tokens_seen": 944308224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605616850551655,
      "loss": 2.7218,
      "theoretical_loss": 3.669225786103548,
      "tokens_seen": 944373760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605516549648947,
      "loss": 2.5733,
      "theoretical_loss": 3.6692012977734865,
      "tokens_seen": 944439296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036054162487462386,
      "loss": 2.6159,
      "theoretical_loss": 3.6691768116184074,
      "tokens_seen": 944504832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036053159478435304,
      "loss": 2.5814,
      "theoretical_loss": 3.669152327637967,
      "tokens_seen": 944570368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605215646940823,
      "loss": 2.8545,
      "theoretical_loss": 3.66912784583182,
      "tokens_seen": 944635904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605115346038114,
      "loss": 2.841,
      "theoretical_loss": 3.669103366199624,
      "tokens_seen": 944701440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036050150451354064,
      "loss": 2.5827,
      "theoretical_loss": 3.6690788887410344,
      "tokens_seen": 944766976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036049147442326977,
      "loss": 2.6336,
      "theoretical_loss": 3.6690544134557075,
      "tokens_seen": 944832512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360481444332999,
      "loss": 2.6132,
      "theoretical_loss": 3.6690299403432998,
      "tokens_seen": 944898048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604714142427282,
      "loss": 2.5371,
      "theoretical_loss": 3.6690054694034675,
      "tokens_seen": 944963584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036046138415245737,
      "loss": 2.5715,
      "theoretical_loss": 3.6689810006358674,
      "tokens_seen": 945029120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036045135406218655,
      "loss": 2.7622,
      "theoretical_loss": 3.6689565340401558,
      "tokens_seen": 945094656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604413239719158,
      "loss": 2.835,
      "theoretical_loss": 3.6689320696159893,
      "tokens_seen": 945160192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604312938816449,
      "loss": 2.6585,
      "theoretical_loss": 3.6689076073630242,
      "tokens_seen": 945225728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036042126379137414,
      "loss": 2.9013,
      "theoretical_loss": 3.6688831472809182,
      "tokens_seen": 945291264
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 528757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7434470653533936,
      "objective/train/theoretical_loss": 3.668858689369327,
      "objective/train/tokens_used": 965816800,
      "theoretical_loss": 3.668858689369327,
      "tokens_seen": 945356800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036041123370110327,
      "loss": 2.6086,
      "theoretical_loss": 3.668858689369327,
      "tokens_seen": 945356800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604012036108325,
      "loss": 2.698,
      "theoretical_loss": 3.6688342336279094,
      "tokens_seen": 945422336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603911735205617,
      "loss": 2.7111,
      "theoretical_loss": 3.6688097800563213,
      "tokens_seen": 945487872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036038114343029087,
      "loss": 2.7297,
      "theoretical_loss": 3.668785328654219,
      "tokens_seen": 945553408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036037111334002005,
      "loss": 2.8327,
      "theoretical_loss": 3.668760879421261,
      "tokens_seen": 945618944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036036108324974923,
      "loss": 2.6444,
      "theoretical_loss": 3.668736432357104,
      "tokens_seen": 945684480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036035105315947847,
      "loss": 2.7526,
      "theoretical_loss": 3.6687119874614047,
      "tokens_seen": 945750016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036034102306920765,
      "loss": 2.7224,
      "theoretical_loss": 3.668687544733822,
      "tokens_seen": 945815552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036033099297893683,
      "loss": 2.6261,
      "theoretical_loss": 3.668663104174012,
      "tokens_seen": 945881088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360320962888666,
      "loss": 2.662,
      "theoretical_loss": 3.668638665781633,
      "tokens_seen": 945946624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036031093279839525,
      "loss": 2.6121,
      "theoretical_loss": 3.6686142295563426,
      "tokens_seen": 946012160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036030090270812437,
      "loss": 2.791,
      "theoretical_loss": 3.668589795497799,
      "tokens_seen": 946077696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602908726178536,
      "loss": 2.6844,
      "theoretical_loss": 3.6685653636056585,
      "tokens_seen": 946143232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036028084252758273,
      "loss": 2.9634,
      "theoretical_loss": 3.66854093387958,
      "tokens_seen": 946208768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036027081243731197,
      "loss": 2.7207,
      "theoretical_loss": 3.668516506319222,
      "tokens_seen": 946274304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036026078234704115,
      "loss": 2.5772,
      "theoretical_loss": 3.668492080924241,
      "tokens_seen": 946339840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036025075225677033,
      "loss": 2.6456,
      "theoretical_loss": 3.668467657694297,
      "tokens_seen": 946405376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602407221664995,
      "loss": 2.7981,
      "theoretical_loss": 3.668443236629047,
      "tokens_seen": 946470912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602306920762287,
      "loss": 2.8633,
      "theoretical_loss": 3.668418817728149,
      "tokens_seen": 946536448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602206619859579,
      "loss": 2.505,
      "theoretical_loss": 3.6683944009912617,
      "tokens_seen": 946601984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602106318956871,
      "loss": 2.7682,
      "theoretical_loss": 3.6683699864180443,
      "tokens_seen": 946667520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036020060180541624,
      "loss": 2.6689,
      "theoretical_loss": 3.6683455740081543,
      "tokens_seen": 946733056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601905717151455,
      "loss": 2.7219,
      "theoretical_loss": 3.6683211637612505,
      "tokens_seen": 946798592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601805416248746,
      "loss": 2.8951,
      "theoretical_loss": 3.668296755676992,
      "tokens_seen": 946864128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036017051153460384,
      "loss": 2.8538,
      "theoretical_loss": 3.668272349755037,
      "tokens_seen": 946929664
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 529862,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7047557830810547,
      "objective/train/theoretical_loss": 3.6682479459950446,
      "objective/train/tokens_used": 967455200,
      "theoretical_loss": 3.6682479459950446,
      "tokens_seen": 946995200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360160481444333,
      "loss": 2.571,
      "theoretical_loss": 3.6682479459950446,
      "tokens_seen": 946995200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601504513540622,
      "loss": 2.64,
      "theoretical_loss": 3.6682235443966733,
      "tokens_seen": 947060736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601404212637914,
      "loss": 2.7239,
      "theoretical_loss": 3.6681991449595825,
      "tokens_seen": 947126272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601303911735206,
      "loss": 2.706,
      "theoretical_loss": 3.668174747683431,
      "tokens_seen": 947191808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036012036108324974,
      "loss": 2.7165,
      "theoretical_loss": 3.6681503525678787,
      "tokens_seen": 947257344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360110330992979,
      "loss": 2.756,
      "theoretical_loss": 3.668125959612584,
      "tokens_seen": 947322880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601003009027081,
      "loss": 2.5381,
      "theoretical_loss": 3.668101568817206,
      "tokens_seen": 947388416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036009027081243734,
      "loss": 2.6841,
      "theoretical_loss": 3.6680771801814043,
      "tokens_seen": 947453952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600802407221665,
      "loss": 2.7336,
      "theoretical_loss": 3.668052793704838,
      "tokens_seen": 947519488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600702106318957,
      "loss": 2.679,
      "theoretical_loss": 3.668028409387168,
      "tokens_seen": 947585024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600601805416249,
      "loss": 2.7866,
      "theoretical_loss": 3.668004027228052,
      "tokens_seen": 947650560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036005015045135406,
      "loss": 2.6878,
      "theoretical_loss": 3.667979647227151,
      "tokens_seen": 947716096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036004012036108324,
      "loss": 2.6239,
      "theoretical_loss": 3.6679552693841244,
      "tokens_seen": 947781632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600300902708125,
      "loss": 2.6908,
      "theoretical_loss": 3.6679308936986317,
      "tokens_seen": 947847168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600200601805416,
      "loss": 2.5497,
      "theoretical_loss": 3.667906520170333,
      "tokens_seen": 947912704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036001003009027084,
      "loss": 2.7513,
      "theoretical_loss": 3.6678821487988884,
      "tokens_seen": 947978240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035999999999999997,
      "loss": 2.3677,
      "theoretical_loss": 3.667857779583958,
      "tokens_seen": 948043776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599899699097292,
      "loss": 2.6004,
      "theoretical_loss": 3.6678334125252015,
      "tokens_seen": 948109312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599799398194584,
      "loss": 2.6877,
      "theoretical_loss": 3.667809047622279,
      "tokens_seen": 948174848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035996990972918757,
      "loss": 2.8092,
      "theoretical_loss": 3.667784684874851,
      "tokens_seen": 948240384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035995987963891675,
      "loss": 2.6486,
      "theoretical_loss": 3.667760324282579,
      "tokens_seen": 948305920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359949849548646,
      "loss": 2.5051,
      "theoretical_loss": 3.6677359658451216,
      "tokens_seen": 948371456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599398194583751,
      "loss": 2.7112,
      "theoretical_loss": 3.6677116095621405,
      "tokens_seen": 948436992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035992978936810434,
      "loss": 2.5369,
      "theoretical_loss": 3.6676872554332958,
      "tokens_seen": 948502528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035991975927783347,
      "loss": 2.6575,
      "theoretical_loss": 3.667662903458248,
      "tokens_seen": 948568064
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 530380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.964254140853882,
      "objective/train/theoretical_loss": 3.6676385536366585,
      "objective/train/tokens_used": 969093600,
      "theoretical_loss": 3.6676385536366585,
      "tokens_seen": 948633600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599097291875627,
      "loss": 2.7343,
      "theoretical_loss": 3.6676385536366585,
      "tokens_seen": 948633600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598996990972919,
      "loss": 2.7499,
      "theoretical_loss": 3.6676142059681878,
      "tokens_seen": 948699136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035988966900702107,
      "loss": 2.8896,
      "theoretical_loss": 3.6675898604524964,
      "tokens_seen": 948764672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035987963891675025,
      "loss": 2.9244,
      "theoretical_loss": 3.667565517089246,
      "tokens_seen": 948830208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035986960882647943,
      "loss": 2.5677,
      "theoretical_loss": 3.667541175878097,
      "tokens_seen": 948895744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598595787362086,
      "loss": 2.8888,
      "theoretical_loss": 3.6675168368187103,
      "tokens_seen": 948961280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035984954864593785,
      "loss": 2.5331,
      "theoretical_loss": 3.6674924999107485,
      "tokens_seen": 949026816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359839518555667,
      "loss": 2.8319,
      "theoretical_loss": 3.6674681651538714,
      "tokens_seen": 949092352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598294884653962,
      "loss": 2.6815,
      "theoretical_loss": 3.6674438325477405,
      "tokens_seen": 949157888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035981945837512534,
      "loss": 2.7486,
      "theoretical_loss": 3.6674195020920184,
      "tokens_seen": 949223424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035980942828485457,
      "loss": 2.7056,
      "theoretical_loss": 3.6673951737863657,
      "tokens_seen": 949288960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035979939819458375,
      "loss": 2.6764,
      "theoretical_loss": 3.667370847630444,
      "tokens_seen": 949354496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035978936810431293,
      "loss": 2.5544,
      "theoretical_loss": 3.667346523623915,
      "tokens_seen": 949420032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597793380140421,
      "loss": 2.6938,
      "theoretical_loss": 3.6673222017664404,
      "tokens_seen": 949485568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035976930792377135,
      "loss": 2.4797,
      "theoretical_loss": 3.6672978820576825,
      "tokens_seen": 949551104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597592778335005,
      "loss": 2.8528,
      "theoretical_loss": 3.667273564497303,
      "tokens_seen": 949616640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597492477432297,
      "loss": 2.6105,
      "theoretical_loss": 3.6672492490849633,
      "tokens_seen": 949682176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035973921765295884,
      "loss": 2.5964,
      "theoretical_loss": 3.667224935820326,
      "tokens_seen": 949747712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597291875626881,
      "loss": 2.7149,
      "theoretical_loss": 3.667200624703053,
      "tokens_seen": 949813248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035971915747241726,
      "loss": 2.5913,
      "theoretical_loss": 3.6671763157328066,
      "tokens_seen": 949878784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035970912738214644,
      "loss": 2.6368,
      "theoretical_loss": 3.667152008909249,
      "tokens_seen": 949944320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596990972918756,
      "loss": 2.7971,
      "theoretical_loss": 3.6671277042320423,
      "tokens_seen": 950009856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596890672016048,
      "loss": 2.7024,
      "theoretical_loss": 3.6671034017008495,
      "tokens_seen": 950075392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359679037111334,
      "loss": 2.5597,
      "theoretical_loss": 3.667079101315333,
      "tokens_seen": 950140928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596690070210632,
      "loss": 2.7964,
      "theoretical_loss": 3.667054803075155,
      "tokens_seen": 950206464
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 531473,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.32260799407959,
      "objective/train/theoretical_loss": 3.6670305069799785,
      "objective/train/tokens_used": 970732000,
      "theoretical_loss": 3.6670305069799785,
      "tokens_seen": 950272000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035965897693079234,
      "loss": 2.6143,
      "theoretical_loss": 3.6670305069799785,
      "tokens_seen": 950272000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596489468405216,
      "loss": 2.657,
      "theoretical_loss": 3.667006213029466,
      "tokens_seen": 950337536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035963891675025076,
      "loss": 2.7829,
      "theoretical_loss": 3.6669819212232806,
      "tokens_seen": 950403072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035962888665997994,
      "loss": 2.7387,
      "theoretical_loss": 3.6669576315610843,
      "tokens_seen": 950468608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596188565697091,
      "loss": 2.7885,
      "theoretical_loss": 3.6669333440425413,
      "tokens_seen": 950534144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596088264794383,
      "loss": 2.854,
      "theoretical_loss": 3.666909058667314,
      "tokens_seen": 950599680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035959879638916754,
      "loss": 2.5672,
      "theoretical_loss": 3.666884775435066,
      "tokens_seen": 950665216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595887662988967,
      "loss": 2.4754,
      "theoretical_loss": 3.6668604943454595,
      "tokens_seen": 950730752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595787362086259,
      "loss": 2.856,
      "theoretical_loss": 3.6668362153981584,
      "tokens_seen": 950796288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595687061183551,
      "loss": 2.6182,
      "theoretical_loss": 3.6668119385928257,
      "tokens_seen": 950861824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035955867602808426,
      "loss": 2.7393,
      "theoretical_loss": 3.6667876639291257,
      "tokens_seen": 950927360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035954864593781344,
      "loss": 2.5813,
      "theoretical_loss": 3.6667633914067213,
      "tokens_seen": 950992896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595386158475427,
      "loss": 2.7133,
      "theoretical_loss": 3.6667391210252758,
      "tokens_seen": 951058432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595285857572718,
      "loss": 2.5476,
      "theoretical_loss": 3.666714852784453,
      "tokens_seen": 951123968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035951855566700104,
      "loss": 2.8157,
      "theoretical_loss": 3.6666905866839175,
      "tokens_seen": 951189504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035950852557673017,
      "loss": 2.8043,
      "theoretical_loss": 3.6666663227233314,
      "tokens_seen": 951255040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594984954864594,
      "loss": 2.7221,
      "theoretical_loss": 3.66664206090236,
      "tokens_seen": 951320576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594884653961886,
      "loss": 2.758,
      "theoretical_loss": 3.6666178012206663,
      "tokens_seen": 951386112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035947843530591777,
      "loss": 2.7675,
      "theoretical_loss": 3.666593543677915,
      "tokens_seen": 951451648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035946840521564695,
      "loss": 2.677,
      "theoretical_loss": 3.66656928827377,
      "tokens_seen": 951517184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594583751253762,
      "loss": 2.8365,
      "theoretical_loss": 3.6665450350078954,
      "tokens_seen": 951582720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594483450351053,
      "loss": 2.7003,
      "theoretical_loss": 3.666520783879956,
      "tokens_seen": 951648256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035943831494483454,
      "loss": 2.7854,
      "theoretical_loss": 3.6664965348896157,
      "tokens_seen": 951713792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035942828485456367,
      "loss": 2.5429,
      "theoretical_loss": 3.6664722880365384,
      "tokens_seen": 951779328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594182547642929,
      "loss": 2.8141,
      "theoretical_loss": 3.666448043320389,
      "tokens_seen": 951844864
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 532856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.025040864944458,
      "objective/train/theoretical_loss": 3.666423800740832,
      "objective/train/tokens_used": 972370400,
      "theoretical_loss": 3.666423800740832,
      "tokens_seen": 951910400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594082246740221,
      "loss": 2.7602,
      "theoretical_loss": 3.666423800740832,
      "tokens_seen": 951910400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035939819458375127,
      "loss": 2.6914,
      "theoretical_loss": 3.666399560297532,
      "tokens_seen": 951975936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035938816449348045,
      "loss": 2.6439,
      "theoretical_loss": 3.6663753219901545,
      "tokens_seen": 952041472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035937813440320963,
      "loss": 2.8814,
      "theoretical_loss": 3.6663510858183628,
      "tokens_seen": 952107008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593681043129388,
      "loss": 2.5812,
      "theoretical_loss": 3.6663268517818235,
      "tokens_seen": 952172544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035935807422266805,
      "loss": 2.9106,
      "theoretical_loss": 3.6663026198802,
      "tokens_seen": 952238080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593480441323972,
      "loss": 2.9617,
      "theoretical_loss": 3.666278390113158,
      "tokens_seen": 952303616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593380140421264,
      "loss": 2.798,
      "theoretical_loss": 3.6662541624803624,
      "tokens_seen": 952369152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035932798395185554,
      "loss": 2.7121,
      "theoretical_loss": 3.6662299369814786,
      "tokens_seen": 952434688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035931795386158477,
      "loss": 2.6545,
      "theoretical_loss": 3.6662057136161716,
      "tokens_seen": 952500224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035930792377131395,
      "loss": 2.729,
      "theoretical_loss": 3.6661814923841067,
      "tokens_seen": 952565760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035929789368104313,
      "loss": 2.8024,
      "theoretical_loss": 3.6661572732849494,
      "tokens_seen": 952631296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592878635907723,
      "loss": 2.5495,
      "theoretical_loss": 3.6661330563183654,
      "tokens_seen": 952696832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035927783350050155,
      "loss": 2.4999,
      "theoretical_loss": 3.66610884148402,
      "tokens_seen": 952762368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592678034102307,
      "loss": 2.8948,
      "theoretical_loss": 3.6660846287815794,
      "tokens_seen": 952827904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592577733199599,
      "loss": 2.654,
      "theoretical_loss": 3.6660604182107077,
      "tokens_seen": 952893440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035924774322968904,
      "loss": 2.8302,
      "theoretical_loss": 3.666036209771072,
      "tokens_seen": 952958976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592377131394183,
      "loss": 2.8485,
      "theoretical_loss": 3.6660120034623382,
      "tokens_seen": 953024512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035922768304914746,
      "loss": 2.7843,
      "theoretical_loss": 3.665987799284171,
      "tokens_seen": 953090048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035921765295887664,
      "loss": 2.351,
      "theoretical_loss": 3.665963597236238,
      "tokens_seen": 953155584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592076228686058,
      "loss": 2.6998,
      "theoretical_loss": 3.6659393973182044,
      "tokens_seen": 953221120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359197592778335,
      "loss": 2.5819,
      "theoretical_loss": 3.665915199529736,
      "tokens_seen": 953286656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591875626880642,
      "loss": 2.6821,
      "theoretical_loss": 3.6658910038704997,
      "tokens_seen": 953352192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591775325977934,
      "loss": 2.7457,
      "theoretical_loss": 3.665866810340162,
      "tokens_seen": 953417728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035916750250752254,
      "loss": 2.6965,
      "theoretical_loss": 3.665842618938388,
      "tokens_seen": 953483264
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 533452,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.747199773788452,
      "objective/train/theoretical_loss": 3.6658184296648457,
      "objective/train/tokens_used": 974008800,
      "theoretical_loss": 3.6658184296648457,
      "tokens_seen": 953548800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591574724172518,
      "loss": 2.8721,
      "theoretical_loss": 3.6658184296648457,
      "tokens_seen": 953548800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035914744232698096,
      "loss": 2.6881,
      "theoretical_loss": 3.6657942425192003,
      "tokens_seen": 953614336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035913741223671014,
      "loss": 2.6932,
      "theoretical_loss": 3.665770057501119,
      "tokens_seen": 953679872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591273821464393,
      "loss": 2.7589,
      "theoretical_loss": 3.665745874610269,
      "tokens_seen": 953745408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591173520561685,
      "loss": 2.8227,
      "theoretical_loss": 3.665721693846316,
      "tokens_seen": 953810944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591073219658977,
      "loss": 2.6922,
      "theoretical_loss": 3.6656975152089277,
      "tokens_seen": 953876480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590972918756269,
      "loss": 2.6638,
      "theoretical_loss": 3.6656733386977702,
      "tokens_seen": 953942016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035908726178535605,
      "loss": 2.9033,
      "theoretical_loss": 3.665649164312511,
      "tokens_seen": 954007552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590772316950853,
      "loss": 2.6242,
      "theoretical_loss": 3.665624992052817,
      "tokens_seen": 954073088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590672016048144,
      "loss": 2.7591,
      "theoretical_loss": 3.665600821918355,
      "tokens_seen": 954138624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035905717151454364,
      "loss": 2.748,
      "theoretical_loss": 3.6655766539087935,
      "tokens_seen": 954204160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590471414242728,
      "loss": 2.5731,
      "theoretical_loss": 3.6655524880237977,
      "tokens_seen": 954269696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359037111334002,
      "loss": 2.6532,
      "theoretical_loss": 3.665528324263037,
      "tokens_seen": 954335232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590270812437312,
      "loss": 2.5441,
      "theoretical_loss": 3.665504162626177,
      "tokens_seen": 954400768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035901705115346037,
      "loss": 2.5651,
      "theoretical_loss": 3.6654800031128865,
      "tokens_seen": 954466304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035900702106318955,
      "loss": 2.7077,
      "theoretical_loss": 3.665455845722833,
      "tokens_seen": 954531840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589969909729188,
      "loss": 2.6124,
      "theoretical_loss": 3.6654316904556827,
      "tokens_seen": 954597376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589869608826479,
      "loss": 2.7877,
      "theoretical_loss": 3.665407537311105,
      "tokens_seen": 954662912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035897693079237715,
      "loss": 2.6148,
      "theoretical_loss": 3.6653833862887675,
      "tokens_seen": 954728448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035896690070210633,
      "loss": 2.7657,
      "theoretical_loss": 3.665359237388337,
      "tokens_seen": 954793984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589568706118355,
      "loss": 2.6572,
      "theoretical_loss": 3.6653350906094824,
      "tokens_seen": 954859520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589468405215647,
      "loss": 2.5009,
      "theoretical_loss": 3.6653109459518713,
      "tokens_seen": 954925056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035893681043129387,
      "loss": 2.5484,
      "theoretical_loss": 3.6652868034151718,
      "tokens_seen": 954990592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035892678034102305,
      "loss": 2.7428,
      "theoretical_loss": 3.6652626629990523,
      "tokens_seen": 955056128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589167502507523,
      "loss": 2.4657,
      "theoretical_loss": 3.665238524703181,
      "tokens_seen": 955121664
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 534564,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8826329708099365,
      "objective/train/theoretical_loss": 3.6652143885272257,
      "objective/train/tokens_used": 975647200,
      "theoretical_loss": 3.6652143885272257,
      "tokens_seen": 955187200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589067201604814,
      "loss": 2.5385,
      "theoretical_loss": 3.6652143885272257,
      "tokens_seen": 955187200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035889669007021065,
      "loss": 2.8421,
      "theoretical_loss": 3.6651902544708554,
      "tokens_seen": 955252736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588866599799398,
      "loss": 2.7383,
      "theoretical_loss": 3.665166122533739,
      "tokens_seen": 955318272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358876629889669,
      "loss": 2.716,
      "theoretical_loss": 3.6651419927155438,
      "tokens_seen": 955383808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588665997993982,
      "loss": 2.7816,
      "theoretical_loss": 3.665117865015939,
      "tokens_seen": 955449344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588565697091274,
      "loss": 2.6429,
      "theoretical_loss": 3.6650937394345933,
      "tokens_seen": 955514880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588465396188566,
      "loss": 2.7899,
      "theoretical_loss": 3.665069615971176,
      "tokens_seen": 955580416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035883650952858574,
      "loss": 2.6999,
      "theoretical_loss": 3.665045494625355,
      "tokens_seen": 955645952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035882647943831497,
      "loss": 2.5648,
      "theoretical_loss": 3.6650213753967997,
      "tokens_seen": 955711488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035881644934804415,
      "loss": 2.669,
      "theoretical_loss": 3.664997258285179,
      "tokens_seen": 955777024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035880641925777333,
      "loss": 2.6379,
      "theoretical_loss": 3.664973143290162,
      "tokens_seen": 955842560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587963891675025,
      "loss": 2.6819,
      "theoretical_loss": 3.6649490304114183,
      "tokens_seen": 955908096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035878635907723175,
      "loss": 2.7801,
      "theoretical_loss": 3.664924919648616,
      "tokens_seen": 955973632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587763289869609,
      "loss": 2.6382,
      "theoretical_loss": 3.6649008110014254,
      "tokens_seen": 956039168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587662988966901,
      "loss": 2.8571,
      "theoretical_loss": 3.664876704469515,
      "tokens_seen": 956104704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035875626880641924,
      "loss": 2.9089,
      "theoretical_loss": 3.6648526000525554,
      "tokens_seen": 956170240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587462387161485,
      "loss": 2.8262,
      "theoretical_loss": 3.6648284977502152,
      "tokens_seen": 956235776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035873620862587766,
      "loss": 2.8647,
      "theoretical_loss": 3.664804397562164,
      "tokens_seen": 956301312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035872617853560684,
      "loss": 2.8062,
      "theoretical_loss": 3.664780299488072,
      "tokens_seen": 956366848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358716148445336,
      "loss": 2.5454,
      "theoretical_loss": 3.6647562035276082,
      "tokens_seen": 956432384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587061183550652,
      "loss": 2.7496,
      "theoretical_loss": 3.6647321096804433,
      "tokens_seen": 956497920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586960882647944,
      "loss": 2.6876,
      "theoretical_loss": 3.6647080179462463,
      "tokens_seen": 956563456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586860581745236,
      "loss": 2.9231,
      "theoretical_loss": 3.6646839283246875,
      "tokens_seen": 956628992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035867602808425274,
      "loss": 2.6204,
      "theoretical_loss": 3.664659840815437,
      "tokens_seen": 956694528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358665997993982,
      "loss": 2.719,
      "theoretical_loss": 3.664635755418165,
      "tokens_seen": 956760064
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 535115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7175190448760986,
      "objective/train/theoretical_loss": 3.6646116721325415,
      "objective/train/tokens_used": 977285600,
      "theoretical_loss": 3.6646116721325415,
      "tokens_seen": 956825600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035865596790371116,
      "loss": 2.5719,
      "theoretical_loss": 3.6646116721325415,
      "tokens_seen": 956825600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035864593781344034,
      "loss": 2.4455,
      "theoretical_loss": 3.664587590958237,
      "tokens_seen": 956891136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586359077231695,
      "loss": 2.6658,
      "theoretical_loss": 3.6645635118949214,
      "tokens_seen": 956956672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586258776328987,
      "loss": 2.3844,
      "theoretical_loss": 3.664539434942265,
      "tokens_seen": 957022208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586158475426279,
      "loss": 2.6877,
      "theoretical_loss": 3.664515360099939,
      "tokens_seen": 957087744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586058174523571,
      "loss": 2.7858,
      "theoretical_loss": 3.664491287367613,
      "tokens_seen": 957153280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035859578736208625,
      "loss": 2.8758,
      "theoretical_loss": 3.664467216744959,
      "tokens_seen": 957218816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585857572718155,
      "loss": 2.7512,
      "theoretical_loss": 3.664443148231647,
      "tokens_seen": 957284352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585757271815446,
      "loss": 2.9092,
      "theoretical_loss": 3.6644190818273463,
      "tokens_seen": 957349888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035856569709127384,
      "loss": 2.679,
      "theoretical_loss": 3.66439501753173,
      "tokens_seen": 957415424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358555667001003,
      "loss": 2.9037,
      "theoretical_loss": 3.6643709553444683,
      "tokens_seen": 957480960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585456369107322,
      "loss": 2.5161,
      "theoretical_loss": 3.6643468952652314,
      "tokens_seen": 957546496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585356068204614,
      "loss": 2.5333,
      "theoretical_loss": 3.664322837293691,
      "tokens_seen": 957612032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035852557673019057,
      "loss": 2.8931,
      "theoretical_loss": 3.664298781429518,
      "tokens_seen": 957677568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035851554663991975,
      "loss": 2.9314,
      "theoretical_loss": 3.6642747276723844,
      "tokens_seen": 957743104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358505516549649,
      "loss": 2.8517,
      "theoretical_loss": 3.66425067602196,
      "tokens_seen": 957808640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584954864593781,
      "loss": 2.5505,
      "theoretical_loss": 3.6642266264779177,
      "tokens_seen": 957874176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035848545636910735,
      "loss": 2.3913,
      "theoretical_loss": 3.6642025790399284,
      "tokens_seen": 957939712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035847542627883653,
      "loss": 2.8246,
      "theoretical_loss": 3.6641785337076627,
      "tokens_seen": 958005248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584653961885657,
      "loss": 2.536,
      "theoretical_loss": 3.6641544904807937,
      "tokens_seen": 958070784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584553660982949,
      "loss": 2.7839,
      "theoretical_loss": 3.664130449358992,
      "tokens_seen": 958136320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035844533600802407,
      "loss": 2.7568,
      "theoretical_loss": 3.6641064103419296,
      "tokens_seen": 958201856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035843530591775325,
      "loss": 2.5894,
      "theoretical_loss": 3.6640823734292782,
      "tokens_seen": 958267392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584252758274825,
      "loss": 2.809,
      "theoretical_loss": 3.66405833862071,
      "tokens_seen": 958332928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584152457372116,
      "loss": 2.5561,
      "theoretical_loss": 3.6640343059158966,
      "tokens_seen": 958398464
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 536290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.003551483154297,
      "objective/train/theoretical_loss": 3.66401027531451,
      "objective/train/tokens_used": 978924000,
      "theoretical_loss": 3.66401027531451,
      "tokens_seen": 958464000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035840521564694085,
      "loss": 2.7357,
      "theoretical_loss": 3.66401027531451,
      "tokens_seen": 958464000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035839518555667,
      "loss": 2.6729,
      "theoretical_loss": 3.663986246816223,
      "tokens_seen": 958529536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583851554663992,
      "loss": 2.7778,
      "theoretical_loss": 3.663962220420707,
      "tokens_seen": 958595072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583751253761284,
      "loss": 2.7285,
      "theoretical_loss": 3.663938196127634,
      "tokens_seen": 958660608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583650952858576,
      "loss": 2.4119,
      "theoretical_loss": 3.6639141739366776,
      "tokens_seen": 958726144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035835506519558676,
      "loss": 2.7283,
      "theoretical_loss": 3.663890153847509,
      "tokens_seen": 958791680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035834503510531594,
      "loss": 2.4578,
      "theoretical_loss": 3.663866135859801,
      "tokens_seen": 958857216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583350050150451,
      "loss": 2.7346,
      "theoretical_loss": 3.663842119973226,
      "tokens_seen": 958922752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035832497492477435,
      "loss": 2.6133,
      "theoretical_loss": 3.663818106187457,
      "tokens_seen": 958988288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583149448345035,
      "loss": 2.6151,
      "theoretical_loss": 3.6637940945021663,
      "tokens_seen": 959053824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583049147442327,
      "loss": 2.6634,
      "theoretical_loss": 3.663770084917027,
      "tokens_seen": 959119360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582948846539619,
      "loss": 2.7547,
      "theoretical_loss": 3.663746077431712,
      "tokens_seen": 959184896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582848545636911,
      "loss": 2.8404,
      "theoretical_loss": 3.6637220720458936,
      "tokens_seen": 959250432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035827482447342026,
      "loss": 2.5344,
      "theoretical_loss": 3.6636980687592455,
      "tokens_seen": 959315968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035826479438314944,
      "loss": 2.7036,
      "theoretical_loss": 3.66367406757144,
      "tokens_seen": 959381504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582547642928786,
      "loss": 2.6886,
      "theoretical_loss": 3.6636500684821507,
      "tokens_seen": 959447040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035824473420260786,
      "loss": 2.7948,
      "theoretical_loss": 3.663626071491051,
      "tokens_seen": 959512576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358234704112337,
      "loss": 2.8091,
      "theoretical_loss": 3.6636020765978135,
      "tokens_seen": 959578112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582246740220662,
      "loss": 2.6147,
      "theoretical_loss": 3.663578083802112,
      "tokens_seen": 959643648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035821464393179535,
      "loss": 2.6451,
      "theoretical_loss": 3.6635540931036195,
      "tokens_seen": 959709184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582046138415246,
      "loss": 2.6816,
      "theoretical_loss": 3.6635301045020103,
      "tokens_seen": 959774720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035819458375125376,
      "loss": 2.7626,
      "theoretical_loss": 3.663506117996958,
      "tokens_seen": 959840256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035818455366098294,
      "loss": 2.5617,
      "theoretical_loss": 3.6634821335881345,
      "tokens_seen": 959905792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581745235707121,
      "loss": 2.6693,
      "theoretical_loss": 3.663458151275216,
      "tokens_seen": 959971328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035816449348044136,
      "loss": 2.64,
      "theoretical_loss": 3.663434171057874,
      "tokens_seen": 960036864
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 537086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.786989688873291,
      "objective/train/theoretical_loss": 3.6634101929357836,
      "objective/train/tokens_used": 980562400,
      "theoretical_loss": 3.6634101929357836,
      "tokens_seen": 960102400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581544633901705,
      "loss": 2.581,
      "theoretical_loss": 3.6634101929357836,
      "tokens_seen": 960102400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581444332998997,
      "loss": 2.7225,
      "theoretical_loss": 3.663386216908619,
      "tokens_seen": 960167936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035813440320962885,
      "loss": 2.5938,
      "theoretical_loss": 3.6633622429760533,
      "tokens_seen": 960233472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581243731193581,
      "loss": 2.7526,
      "theoretical_loss": 3.6633382711377607,
      "tokens_seen": 960299008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035811434302908727,
      "loss": 2.7211,
      "theoretical_loss": 3.663314301393416,
      "tokens_seen": 960364544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035810431293881645,
      "loss": 2.4448,
      "theoretical_loss": 3.663290333742693,
      "tokens_seen": 960430080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580942828485457,
      "loss": 2.5715,
      "theoretical_loss": 3.6632663681852655,
      "tokens_seen": 960495616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580842527582748,
      "loss": 2.6514,
      "theoretical_loss": 3.663242404720809,
      "tokens_seen": 960561152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035807422266800404,
      "loss": 2.6498,
      "theoretical_loss": 3.663218443348997,
      "tokens_seen": 960626688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580641925777332,
      "loss": 2.7838,
      "theoretical_loss": 3.663194484069505,
      "tokens_seen": 960692224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580541624874624,
      "loss": 2.4934,
      "theoretical_loss": 3.6631705268820065,
      "tokens_seen": 960757760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580441323971916,
      "loss": 2.6763,
      "theoretical_loss": 3.6631465717861764,
      "tokens_seen": 960823296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035803410230692077,
      "loss": 2.7096,
      "theoretical_loss": 3.6631226187816894,
      "tokens_seen": 960888832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035802407221664995,
      "loss": 2.7574,
      "theoretical_loss": 3.663098667868221,
      "tokens_seen": 960954368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580140421263792,
      "loss": 2.3962,
      "theoretical_loss": 3.6630747190454453,
      "tokens_seen": 961019904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580040120361083,
      "loss": 2.5524,
      "theoretical_loss": 3.6630507723130377,
      "tokens_seen": 961085440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035799398194583755,
      "loss": 2.7353,
      "theoretical_loss": 3.663026827670673,
      "tokens_seen": 961150976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035798395185556673,
      "loss": 2.6363,
      "theoretical_loss": 3.6630028851180265,
      "tokens_seen": 961216512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579739217652959,
      "loss": 2.5274,
      "theoretical_loss": 3.662978944654773,
      "tokens_seen": 961282048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579638916750251,
      "loss": 2.7494,
      "theoretical_loss": 3.662955006280588,
      "tokens_seen": 961347584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035795386158475427,
      "loss": 2.7442,
      "theoretical_loss": 3.662931069995147,
      "tokens_seen": 961413120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035794383149448345,
      "loss": 2.8181,
      "theoretical_loss": 3.6629071357981244,
      "tokens_seen": 961478656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579338014042127,
      "loss": 2.6398,
      "theoretical_loss": 3.662883203689197,
      "tokens_seen": 961544192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579237713139418,
      "loss": 2.8939,
      "theoretical_loss": 3.6628592736680394,
      "tokens_seen": 961609728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035791374122367105,
      "loss": 2.6428,
      "theoretical_loss": 3.6628353457343277,
      "tokens_seen": 961675264
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 537653,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.703756809234619,
      "objective/train/theoretical_loss": 3.6628114198877375,
      "objective/train/tokens_used": 982200800,
      "theoretical_loss": 3.6628114198877375,
      "tokens_seen": 961740800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579037111334002,
      "loss": 2.5893,
      "theoretical_loss": 3.6628114198877375,
      "tokens_seen": 961740800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578936810431294,
      "loss": 2.5064,
      "theoretical_loss": 3.6627874961279447,
      "tokens_seen": 961806336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578836509528586,
      "loss": 2.6169,
      "theoretical_loss": 3.6627635744546243,
      "tokens_seen": 961871872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578736208625878,
      "loss": 2.5647,
      "theoretical_loss": 3.662739654867453,
      "tokens_seen": 961937408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035786359077231696,
      "loss": 2.7298,
      "theoretical_loss": 3.662715737366107,
      "tokens_seen": 962002944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035785356068204614,
      "loss": 2.5342,
      "theoretical_loss": 3.6626918219502613,
      "tokens_seen": 962068480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578435305917753,
      "loss": 2.5674,
      "theoretical_loss": 3.6626679086195923,
      "tokens_seen": 962134016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035783350050150455,
      "loss": 2.5954,
      "theoretical_loss": 3.6626439973737765,
      "tokens_seen": 962199552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578234704112337,
      "loss": 2.8665,
      "theoretical_loss": 3.6626200882124906,
      "tokens_seen": 962265088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578134403209629,
      "loss": 2.7817,
      "theoretical_loss": 3.6625961811354104,
      "tokens_seen": 962330624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578034102306921,
      "loss": 2.4717,
      "theoretical_loss": 3.662572276142212,
      "tokens_seen": 962396160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577933801404213,
      "loss": 2.6832,
      "theoretical_loss": 3.662548373232572,
      "tokens_seen": 962461696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035778335005015046,
      "loss": 2.5416,
      "theoretical_loss": 3.662524472406168,
      "tokens_seen": 962527232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035777331995987964,
      "loss": 2.7669,
      "theoretical_loss": 3.662500573662675,
      "tokens_seen": 962592768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577632898696088,
      "loss": 2.6586,
      "theoretical_loss": 3.66247667700177,
      "tokens_seen": 962658304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035775325977933806,
      "loss": 2.4959,
      "theoretical_loss": 3.662452782423131,
      "tokens_seen": 962723840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577432296890672,
      "loss": 2.7372,
      "theoretical_loss": 3.662428889926434,
      "tokens_seen": 962789376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577331995987964,
      "loss": 2.9573,
      "theoretical_loss": 3.662404999511356,
      "tokens_seen": 962854912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035772316950852555,
      "loss": 2.5443,
      "theoretical_loss": 3.6623811111775737,
      "tokens_seen": 962920448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577131394182548,
      "loss": 2.7655,
      "theoretical_loss": 3.6623572249247642,
      "tokens_seen": 962985984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035770310932798396,
      "loss": 2.6151,
      "theoretical_loss": 3.662333340752604,
      "tokens_seen": 963051520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035769307923771314,
      "loss": 2.7865,
      "theoretical_loss": 3.662309458660772,
      "tokens_seen": 963117056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576830491474423,
      "loss": 2.7037,
      "theoretical_loss": 3.6622855786489446,
      "tokens_seen": 963182592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035767301905717156,
      "loss": 2.4568,
      "theoretical_loss": 3.6622617007167992,
      "tokens_seen": 963248128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576629889669007,
      "loss": 2.5051,
      "theoretical_loss": 3.6622378248640124,
      "tokens_seen": 963313664
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 538877,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.671361207962036,
      "objective/train/theoretical_loss": 3.6622139510902625,
      "objective/train/tokens_used": 983839200,
      "theoretical_loss": 3.6622139510902625,
      "tokens_seen": 963379200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576529588766299,
      "loss": 2.464,
      "theoretical_loss": 3.6622139510902625,
      "tokens_seen": 963379200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035764292878635905,
      "loss": 2.8269,
      "theoretical_loss": 3.662190079395227,
      "tokens_seen": 963444736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576328986960883,
      "loss": 2.7938,
      "theoretical_loss": 3.6621662097785834,
      "tokens_seen": 963510272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035762286860581747,
      "loss": 2.6486,
      "theoretical_loss": 3.66214234224001,
      "tokens_seen": 963575808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035761283851554665,
      "loss": 2.713,
      "theoretical_loss": 3.662118476779183,
      "tokens_seen": 963641344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035760280842527583,
      "loss": 2.7186,
      "theoretical_loss": 3.662094613395782,
      "tokens_seen": 963706880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357592778335005,
      "loss": 2.4699,
      "theoretical_loss": 3.6620707520894835,
      "tokens_seen": 963772416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575827482447342,
      "loss": 2.7424,
      "theoretical_loss": 3.662046892859966,
      "tokens_seen": 963837952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575727181544634,
      "loss": 2.6724,
      "theoretical_loss": 3.662023035706908,
      "tokens_seen": 963903488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035756268806419255,
      "loss": 2.6567,
      "theoretical_loss": 3.661999180629987,
      "tokens_seen": 963969024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575526579739218,
      "loss": 2.5726,
      "theoretical_loss": 3.6619753276288822,
      "tokens_seen": 964034560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575426278836509,
      "loss": 2.6506,
      "theoretical_loss": 3.661951476703271,
      "tokens_seen": 964100096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035753259779338015,
      "loss": 2.6112,
      "theoretical_loss": 3.6619276278528314,
      "tokens_seen": 964165632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035752256770310933,
      "loss": 2.6589,
      "theoretical_loss": 3.6619037810772426,
      "tokens_seen": 964231168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575125376128385,
      "loss": 2.7243,
      "theoretical_loss": 3.6618799363761827,
      "tokens_seen": 964296704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575025075225677,
      "loss": 2.7628,
      "theoretical_loss": 3.6618560937493303,
      "tokens_seen": 964362240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035749247743229693,
      "loss": 2.6175,
      "theoretical_loss": 3.661832253196364,
      "tokens_seen": 964427776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035748244734202606,
      "loss": 2.6583,
      "theoretical_loss": 3.6618084147169627,
      "tokens_seen": 964493312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574724172517553,
      "loss": 2.9021,
      "theoretical_loss": 3.6617845783108054,
      "tokens_seen": 964558848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574623871614844,
      "loss": 2.653,
      "theoretical_loss": 3.6617607439775703,
      "tokens_seen": 964624384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035745235707121365,
      "loss": 2.4431,
      "theoretical_loss": 3.661736911716936,
      "tokens_seen": 964689920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035744232698094283,
      "loss": 2.6706,
      "theoretical_loss": 3.6617130815285828,
      "tokens_seen": 964755456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357432296890672,
      "loss": 2.6288,
      "theoretical_loss": 3.661689253412189,
      "tokens_seen": 964820992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574222668004012,
      "loss": 2.7101,
      "theoretical_loss": 3.661665427367434,
      "tokens_seen": 964886528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574122367101304,
      "loss": 2.7293,
      "theoretical_loss": 3.6616416033939965,
      "tokens_seen": 964952064
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 539426,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4708588123321533,
      "objective/train/theoretical_loss": 3.6616177814915556,
      "objective/train/tokens_used": 985477600,
      "theoretical_loss": 3.6616177814915556,
      "tokens_seen": 965017600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035740220661985956,
      "loss": 2.766,
      "theoretical_loss": 3.6616177814915556,
      "tokens_seen": 965017600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573921765295888,
      "loss": 2.9734,
      "theoretical_loss": 3.6615939616597917,
      "tokens_seen": 965083136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573821464393179,
      "loss": 2.5102,
      "theoretical_loss": 3.661570143898383,
      "tokens_seen": 965148672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035737211634904716,
      "loss": 2.6679,
      "theoretical_loss": 3.6615463282070104,
      "tokens_seen": 965214208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573620862587763,
      "loss": 2.6167,
      "theoretical_loss": 3.6615225145853523,
      "tokens_seen": 965279744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573520561685055,
      "loss": 2.6641,
      "theoretical_loss": 3.6614987030330886,
      "tokens_seen": 965345280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035734202607823475,
      "loss": 2.8822,
      "theoretical_loss": 3.661474893549899,
      "tokens_seen": 965410816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573319959879639,
      "loss": 2.8572,
      "theoretical_loss": 3.6614510861354637,
      "tokens_seen": 965476352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573219658976931,
      "loss": 2.6992,
      "theoretical_loss": 3.661427280789462,
      "tokens_seen": 965541888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573119358074223,
      "loss": 2.5701,
      "theoretical_loss": 3.6614034775115742,
      "tokens_seen": 965607424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573019057171515,
      "loss": 2.788,
      "theoretical_loss": 3.66137967630148,
      "tokens_seen": 965672960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035729187562688066,
      "loss": 2.9269,
      "theoretical_loss": 3.6613558771588592,
      "tokens_seen": 965738496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035728184553660984,
      "loss": 2.7455,
      "theoretical_loss": 3.6613320800833926,
      "tokens_seen": 965804032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357271815446339,
      "loss": 2.7057,
      "theoretical_loss": 3.6613082850747602,
      "tokens_seen": 965869568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035726178535606826,
      "loss": 2.7845,
      "theoretical_loss": 3.6612844921326415,
      "tokens_seen": 965935104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572517552657974,
      "loss": 2.6283,
      "theoretical_loss": 3.661260701256718,
      "tokens_seen": 966000640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572417251755266,
      "loss": 2.5378,
      "theoretical_loss": 3.66123691244667,
      "tokens_seen": 966066176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035723169508525575,
      "loss": 2.626,
      "theoretical_loss": 3.6612131257021767,
      "tokens_seen": 966131712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357221664994985,
      "loss": 2.6145,
      "theoretical_loss": 3.6611893410229204,
      "tokens_seen": 966197248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035721163490471416,
      "loss": 2.5952,
      "theoretical_loss": 3.6611655584085803,
      "tokens_seen": 966262784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035720160481444334,
      "loss": 2.3264,
      "theoretical_loss": 3.6611417778588375,
      "tokens_seen": 966328320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571915747241725,
      "loss": 2.5936,
      "theoretical_loss": 3.661117999373374,
      "tokens_seen": 966393856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035718154463390176,
      "loss": 2.7259,
      "theoretical_loss": 3.661094222951869,
      "tokens_seen": 966459392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571715145436309,
      "loss": 2.6951,
      "theoretical_loss": 3.6610704485940033,
      "tokens_seen": 966524928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571614844533601,
      "loss": 2.7819,
      "theoretical_loss": 3.661046676299459,
      "tokens_seen": 966590464
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 540676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.884268283843994,
      "objective/train/theoretical_loss": 3.6610229060679167,
      "objective/train/tokens_used": 987116000,
      "theoretical_loss": 3.6610229060679167,
      "tokens_seen": 966656000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035715145436308925,
      "loss": 2.5583,
      "theoretical_loss": 3.6610229060679167,
      "tokens_seen": 966656000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571414242728185,
      "loss": 2.6991,
      "theoretical_loss": 3.660999137899058,
      "tokens_seen": 966721536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035713139418254767,
      "loss": 2.7206,
      "theoretical_loss": 3.660975371792563,
      "tokens_seen": 966787072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035712136409227685,
      "loss": 2.681,
      "theoretical_loss": 3.660951607748114,
      "tokens_seen": 966852608
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035711133400200603,
      "loss": 2.5881,
      "theoretical_loss": 3.660927845765392,
      "tokens_seen": 966918144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571013039117352,
      "loss": 2.7071,
      "theoretical_loss": 3.660904085844078,
      "tokens_seen": 966983680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570912738214644,
      "loss": 2.7228,
      "theoretical_loss": 3.660880327983854,
      "tokens_seen": 967049216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570812437311936,
      "loss": 2.6058,
      "theoretical_loss": 3.660856572184401,
      "tokens_seen": 967114752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035707121364092275,
      "loss": 2.5509,
      "theoretical_loss": 3.6608328184454013,
      "tokens_seen": 967180288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357061183550652,
      "loss": 2.6556,
      "theoretical_loss": 3.6608090667665363,
      "tokens_seen": 967245824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570511534603811,
      "loss": 2.796,
      "theoretical_loss": 3.6607853171474876,
      "tokens_seen": 967311360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035704112337011035,
      "loss": 2.6946,
      "theoretical_loss": 3.660761569587937,
      "tokens_seen": 967376896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035703109327983953,
      "loss": 2.5267,
      "theoretical_loss": 3.660737824087567,
      "tokens_seen": 967442432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570210631895687,
      "loss": 2.7813,
      "theoretical_loss": 3.660714080646059,
      "tokens_seen": 967507968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570110330992979,
      "loss": 2.6722,
      "theoretical_loss": 3.660690339263095,
      "tokens_seen": 967573504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035700100300902713,
      "loss": 2.8932,
      "theoretical_loss": 3.6606665999383576,
      "tokens_seen": 967639040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035699097291875626,
      "loss": 2.636,
      "theoretical_loss": 3.660642862671528,
      "tokens_seen": 967704576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569809428284855,
      "loss": 2.7414,
      "theoretical_loss": 3.66061912746229,
      "tokens_seen": 967770112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569709127382146,
      "loss": 2.823,
      "theoretical_loss": 3.6605953943103247,
      "tokens_seen": 967835648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035696088264794385,
      "loss": 2.8707,
      "theoretical_loss": 3.6605716632153147,
      "tokens_seen": 967901184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035695085255767303,
      "loss": 2.7507,
      "theoretical_loss": 3.660547934176943,
      "tokens_seen": 967966720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569408224674022,
      "loss": 2.9235,
      "theoretical_loss": 3.660524207194891,
      "tokens_seen": 968032256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569307923771314,
      "loss": 2.5484,
      "theoretical_loss": 3.660500482268843,
      "tokens_seen": 968097792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569207622868606,
      "loss": 2.7966,
      "theoretical_loss": 3.66047675939848,
      "tokens_seen": 968163328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035691073219658976,
      "loss": 2.7508,
      "theoretical_loss": 3.660453038583486,
      "tokens_seen": 968228864
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 541190,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9151451587677,
      "objective/train/theoretical_loss": 3.660429319823543,
      "objective/train/tokens_used": 988754400,
      "theoretical_loss": 3.660429319823543,
      "tokens_seen": 968294400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356900702106319,
      "loss": 2.8765,
      "theoretical_loss": 3.660429319823543,
      "tokens_seen": 968294400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568906720160481,
      "loss": 2.9427,
      "theoretical_loss": 3.6604056031183343,
      "tokens_seen": 968359936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035688064192577736,
      "loss": 2.8068,
      "theoretical_loss": 3.6603818884675423,
      "tokens_seen": 968425472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568706118355065,
      "loss": 2.6032,
      "theoretical_loss": 3.660358175870851,
      "tokens_seen": 968491008
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568605817452357,
      "loss": 2.5964,
      "theoretical_loss": 3.660334465327943,
      "tokens_seen": 968556544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568505516549649,
      "loss": 2.7074,
      "theoretical_loss": 3.6603107568385016,
      "tokens_seen": 968622080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568405215646941,
      "loss": 2.8676,
      "theoretical_loss": 3.66028705040221,
      "tokens_seen": 968687616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035683049147442326,
      "loss": 2.6544,
      "theoretical_loss": 3.6602633460187506,
      "tokens_seen": 968753152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568204613841525,
      "loss": 2.7562,
      "theoretical_loss": 3.6602396436878077,
      "tokens_seen": 968818688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568104312938816,
      "loss": 2.3719,
      "theoretical_loss": 3.6602159434090655,
      "tokens_seen": 968884224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035680040120361086,
      "loss": 2.7534,
      "theoretical_loss": 3.660192245182206,
      "tokens_seen": 968949760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035679037111334,
      "loss": 2.534,
      "theoretical_loss": 3.6601685490069134,
      "tokens_seen": 969015296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567803410230692,
      "loss": 2.62,
      "theoretical_loss": 3.660144854882872,
      "tokens_seen": 969080832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567703109327984,
      "loss": 2.8146,
      "theoretical_loss": 3.660121162809764,
      "tokens_seen": 969146368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567602808425276,
      "loss": 2.7872,
      "theoretical_loss": 3.6600974727872746,
      "tokens_seen": 969211904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035675025075225677,
      "loss": 2.7477,
      "theoretical_loss": 3.6600737848150873,
      "tokens_seen": 969277440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035674022066198595,
      "loss": 2.8665,
      "theoretical_loss": 3.660050098892886,
      "tokens_seen": 969342976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035673019057171513,
      "loss": 2.9426,
      "theoretical_loss": 3.660026415020355,
      "tokens_seen": 969408512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035672016048144436,
      "loss": 2.5801,
      "theoretical_loss": 3.6600027331971776,
      "tokens_seen": 969474048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567101303911735,
      "loss": 2.5656,
      "theoretical_loss": 3.6599790534230383,
      "tokens_seen": 969539584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567001003009027,
      "loss": 2.6174,
      "theoretical_loss": 3.659955375697622,
      "tokens_seen": 969605120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035669007021063185,
      "loss": 2.5564,
      "theoretical_loss": 3.659931700020612,
      "tokens_seen": 969670656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566800401203611,
      "loss": 2.4917,
      "theoretical_loss": 3.659908026391693,
      "tokens_seen": 969736192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035667001003009027,
      "loss": 2.6846,
      "theoretical_loss": 3.6598843548105497,
      "tokens_seen": 969801728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035665997993981945,
      "loss": 2.7662,
      "theoretical_loss": 3.6598606852768665,
      "tokens_seen": 969867264
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 542499,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.745290517807007,
      "objective/train/theoretical_loss": 3.659837017790328,
      "objective/train/tokens_used": 990392800,
      "theoretical_loss": 3.659837017790328,
      "tokens_seen": 969932800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035664994984954863,
      "loss": 2.8151,
      "theoretical_loss": 3.659837017790328,
      "tokens_seen": 969932800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035663991975927787,
      "loss": 2.6502,
      "theoretical_loss": 3.6598133523506187,
      "tokens_seen": 969998336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356629889669007,
      "loss": 2.6904,
      "theoretical_loss": 3.6597896889574235,
      "tokens_seen": 970063872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035661985957873623,
      "loss": 2.7728,
      "theoretical_loss": 3.6597660276104262,
      "tokens_seen": 970129408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035660982948846536,
      "loss": 2.4748,
      "theoretical_loss": 3.6597423683093133,
      "tokens_seen": 970194944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565997993981946,
      "loss": 2.4516,
      "theoretical_loss": 3.6597187110537686,
      "tokens_seen": 970260480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565897693079238,
      "loss": 3.0593,
      "theoretical_loss": 3.6596950558434775,
      "tokens_seen": 970326016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035657973921765295,
      "loss": 2.6859,
      "theoretical_loss": 3.659671402678125,
      "tokens_seen": 970391552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565697091273822,
      "loss": 2.7334,
      "theoretical_loss": 3.6596477515573964,
      "tokens_seen": 970457088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565596790371113,
      "loss": 2.5428,
      "theoretical_loss": 3.6596241024809766,
      "tokens_seen": 970522624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035654964894684055,
      "loss": 2.6518,
      "theoretical_loss": 3.65960045544855,
      "tokens_seen": 970588160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035653961885656973,
      "loss": 2.608,
      "theoretical_loss": 3.6595768104598045,
      "tokens_seen": 970653696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565295887662989,
      "loss": 2.776,
      "theoretical_loss": 3.659553167514423,
      "tokens_seen": 970719232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565195586760281,
      "loss": 2.8407,
      "theoretical_loss": 3.659529526612092,
      "tokens_seen": 970784768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035650952858575733,
      "loss": 2.5555,
      "theoretical_loss": 3.659505887752497,
      "tokens_seen": 970850304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035649949849548646,
      "loss": 2.8917,
      "theoretical_loss": 3.6594822509353238,
      "tokens_seen": 970915840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564894684052157,
      "loss": 2.8816,
      "theoretical_loss": 3.659458616160258,
      "tokens_seen": 970981376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564794383149448,
      "loss": 2.5581,
      "theoretical_loss": 3.6594349834269844,
      "tokens_seen": 971046912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035646940822467405,
      "loss": 2.3331,
      "theoretical_loss": 3.65941135273519,
      "tokens_seen": 971112448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035645937813440323,
      "loss": 2.658,
      "theoretical_loss": 3.6593877240845605,
      "tokens_seen": 971177984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564493480441324,
      "loss": 2.8178,
      "theoretical_loss": 3.659364097474782,
      "tokens_seen": 971243520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564393179538616,
      "loss": 2.5768,
      "theoretical_loss": 3.65934047290554,
      "tokens_seen": 971309056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564292878635908,
      "loss": 2.509,
      "theoretical_loss": 3.6593168503765203,
      "tokens_seen": 971374592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035641925777331996,
      "loss": 2.5761,
      "theoretical_loss": 3.6592932298874103,
      "tokens_seen": 971440128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564092276830492,
      "loss": 2.8889,
      "theoretical_loss": 3.659269611437895,
      "tokens_seen": 971505664
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 543787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6714510917663574,
      "objective/train/theoretical_loss": 3.6592459950276623,
      "objective/train/tokens_used": 992031200,
      "theoretical_loss": 3.6592459950276623,
      "tokens_seen": 971571200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563991975927783,
      "loss": 2.7689,
      "theoretical_loss": 3.6592459950276623,
      "tokens_seen": 971571200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035638916750250756,
      "loss": 2.3759,
      "theoretical_loss": 3.659222380656397,
      "tokens_seen": 971636736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563791374122367,
      "loss": 2.6477,
      "theoretical_loss": 3.659198768323786,
      "tokens_seen": 971702272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563691073219659,
      "loss": 2.633,
      "theoretical_loss": 3.659175158029516,
      "tokens_seen": 971767808
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563590772316951,
      "loss": 2.8631,
      "theoretical_loss": 3.6591515497732736,
      "tokens_seen": 971833344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563490471414243,
      "loss": 2.9037,
      "theoretical_loss": 3.6591279435547452,
      "tokens_seen": 971898880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035633901705115346,
      "loss": 2.6792,
      "theoretical_loss": 3.6591043393736173,
      "tokens_seen": 971964416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563289869608827,
      "loss": 2.6147,
      "theoretical_loss": 3.659080737229578,
      "tokens_seen": 972029952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563189568706118,
      "loss": 2.5756,
      "theoretical_loss": 3.6590571371223133,
      "tokens_seen": 972095488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035630892678034106,
      "loss": 2.5859,
      "theoretical_loss": 3.6590335390515096,
      "tokens_seen": 972161024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562988966900702,
      "loss": 2.7238,
      "theoretical_loss": 3.6590099430168546,
      "tokens_seen": 972226560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562888665997994,
      "loss": 2.7529,
      "theoretical_loss": 3.6589863490180354,
      "tokens_seen": 972292096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562788365095286,
      "loss": 2.6947,
      "theoretical_loss": 3.6589627570547387,
      "tokens_seen": 972357632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562688064192578,
      "loss": 2.8994,
      "theoretical_loss": 3.6589391671266522,
      "tokens_seen": 972423168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035625877632898697,
      "loss": 2.6734,
      "theoretical_loss": 3.6589155792334633,
      "tokens_seen": 972488704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035624874623871615,
      "loss": 2.7424,
      "theoretical_loss": 3.6588919933748585,
      "tokens_seen": 972554240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035623871614844533,
      "loss": 2.8223,
      "theoretical_loss": 3.658868409550526,
      "tokens_seen": 972619776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035622868605817456,
      "loss": 2.7093,
      "theoretical_loss": 3.6588448277601526,
      "tokens_seen": 972685312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562186559679037,
      "loss": 2.7391,
      "theoretical_loss": 3.658821248003427,
      "tokens_seen": 972750848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562086258776329,
      "loss": 2.7881,
      "theoretical_loss": 3.6587976702800353,
      "tokens_seen": 972816384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035619859578736205,
      "loss": 2.6558,
      "theoretical_loss": 3.658774094589667,
      "tokens_seen": 972881920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561885656970913,
      "loss": 2.6641,
      "theoretical_loss": 3.658750520932008,
      "tokens_seen": 972947456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035617853560682047,
      "loss": 2.6948,
      "theoretical_loss": 3.6587269493067476,
      "tokens_seen": 973012992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035616850551654965,
      "loss": 2.6006,
      "theoretical_loss": 3.6587033797135726,
      "tokens_seen": 973078528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035615847542627883,
      "loss": 2.5074,
      "theoretical_loss": 3.658679812152172,
      "tokens_seen": 973144064
    },
    {
      "epoch": 0.29,
      "objective/train/docs_used": 544242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9670348167419434,
      "objective/train/theoretical_loss": 3.658656246622233,
      "objective/train/tokens_used": 993669600,
      "theoretical_loss": 3.658656246622233,
      "tokens_seen": 973209600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035614844533600807,
      "loss": 2.8928,
      "theoretical_loss": 3.658656246622233,
      "tokens_seen": 973209600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561384152457372,
      "loss": 2.6655,
      "theoretical_loss": 3.658632683123444,
      "tokens_seen": 973275136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035612838515546643,
      "loss": 2.6035,
      "theoretical_loss": 3.658609121655494,
      "tokens_seen": 973340672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035611835506519556,
      "loss": 2.646,
      "theoretical_loss": 3.65858556221807,
      "tokens_seen": 973406208
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561083249749248,
      "loss": 2.6821,
      "theoretical_loss": 3.6585620048108605,
      "tokens_seen": 973471744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035609829488465397,
      "loss": 2.8966,
      "theoretical_loss": 3.6585384494335544,
      "tokens_seen": 973537280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035608826479438315,
      "loss": 2.6911,
      "theoretical_loss": 3.65851489608584,
      "tokens_seen": 973602816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035607823470411233,
      "loss": 2.7678,
      "theoretical_loss": 3.658491344767406,
      "tokens_seen": 973668352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560682046138415,
      "loss": 2.8951,
      "theoretical_loss": 3.6584677954779403,
      "tokens_seen": 973733888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560581745235707,
      "loss": 2.68,
      "theoretical_loss": 3.6584442482171324,
      "tokens_seen": 973799424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035604814443329993,
      "loss": 2.8754,
      "theoretical_loss": 3.658420702984671,
      "tokens_seen": 973864960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035603811434302906,
      "loss": 2.6871,
      "theoretical_loss": 3.658397159780244,
      "tokens_seen": 973930496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560280842527583,
      "loss": 2.4913,
      "theoretical_loss": 3.658373618603541,
      "tokens_seen": 973996032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560180541624874,
      "loss": 2.5447,
      "theoretical_loss": 3.6583500794542503,
      "tokens_seen": 974061568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035600802407221666,
      "loss": 2.6892,
      "theoretical_loss": 3.6583265423320617,
      "tokens_seen": 974127104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035599799398194584,
      "loss": 2.5806,
      "theoretical_loss": 3.6583030072366647,
      "tokens_seen": 974192640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355987963891675,
      "loss": 2.5213,
      "theoretical_loss": 3.6582794741677467,
      "tokens_seen": 974258176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559779338014042,
      "loss": 2.6246,
      "theoretical_loss": 3.658255943124998,
      "tokens_seen": 974323712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035596790371113344,
      "loss": 2.7462,
      "theoretical_loss": 3.6582324141081086,
      "tokens_seen": 974389248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035595787362086256,
      "loss": 2.6027,
      "theoretical_loss": 3.6582088871167664,
      "tokens_seen": 974454784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559478435305918,
      "loss": 2.6232,
      "theoretical_loss": 3.6581853621506615,
      "tokens_seen": 974520320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559378134403209,
      "loss": 2.7216,
      "theoretical_loss": 3.658161839209484,
      "tokens_seen": 974585856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035592778335005016,
      "loss": 2.8452,
      "theoretical_loss": 3.658138318292922,
      "tokens_seen": 974651392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035591775325977934,
      "loss": 2.6781,
      "theoretical_loss": 3.658114799400666,
      "tokens_seen": 974716928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559077231695085,
      "loss": 2.5703,
      "theoretical_loss": 3.6580912825324057,
      "tokens_seen": 974782464
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 545526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5406079292297363,
      "objective/train/theoretical_loss": 3.6580677676878306,
      "objective/train/tokens_used": 995308000,
      "theoretical_loss": 3.6580677676878306,
      "tokens_seen": 974848000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558976930792377,
      "loss": 2.6219,
      "theoretical_loss": 3.6580677676878306,
      "tokens_seen": 974848000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558876629889669,
      "loss": 2.7312,
      "theoretical_loss": 3.658044254866631,
      "tokens_seen": 974913536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035587763289869607,
      "loss": 2.6663,
      "theoretical_loss": 3.6580207440684966,
      "tokens_seen": 974979072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558676028084253,
      "loss": 2.7667,
      "theoretical_loss": 3.657997235293117,
      "tokens_seen": 975044608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035585757271815443,
      "loss": 2.7782,
      "theoretical_loss": 3.6579737285401825,
      "tokens_seen": 975110144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035584754262788366,
      "loss": 2.3444,
      "theoretical_loss": 3.6579502238093835,
      "tokens_seen": 975175680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558375125376129,
      "loss": 2.7807,
      "theoretical_loss": 3.6579267211004094,
      "tokens_seen": 975241216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355827482447342,
      "loss": 2.7576,
      "theoretical_loss": 3.657903220412951,
      "tokens_seen": 975306752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035581745235707126,
      "loss": 2.8424,
      "theoretical_loss": 3.657879721746699,
      "tokens_seen": 975372288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558074222668004,
      "loss": 2.7173,
      "theoretical_loss": 3.657856225101343,
      "tokens_seen": 975437824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557973921765296,
      "loss": 2.7882,
      "theoretical_loss": 3.6578327304765734,
      "tokens_seen": 975503360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557873620862588,
      "loss": 2.7136,
      "theoretical_loss": 3.6578092378720815,
      "tokens_seen": 975568896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355777331995988,
      "loss": 2.7584,
      "theoretical_loss": 3.6577857472875572,
      "tokens_seen": 975634432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035576730190571717,
      "loss": 2.6165,
      "theoretical_loss": 3.6577622587226917,
      "tokens_seen": 975699968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035575727181544635,
      "loss": 2.7483,
      "theoretical_loss": 3.657738772177175,
      "tokens_seen": 975765504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035574724172517553,
      "loss": 2.5081,
      "theoretical_loss": 3.6577152876506984,
      "tokens_seen": 975831040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035573721163490476,
      "loss": 3.0255,
      "theoretical_loss": 3.6576918051429526,
      "tokens_seen": 975896576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557271815446339,
      "loss": 2.4618,
      "theoretical_loss": 3.6576683246536286,
      "tokens_seen": 975962112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557171514543631,
      "loss": 2.5939,
      "theoretical_loss": 3.657644846182417,
      "tokens_seen": 976027648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035570712136409225,
      "loss": 3.0158,
      "theoretical_loss": 3.65762136972901,
      "tokens_seen": 976093184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556970912738215,
      "loss": 2.6443,
      "theoretical_loss": 3.657597895293097,
      "tokens_seen": 976158720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035568706118355067,
      "loss": 2.648,
      "theoretical_loss": 3.6575744228743705,
      "tokens_seen": 976224256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035567703109327985,
      "loss": 2.5602,
      "theoretical_loss": 3.6575509524725214,
      "tokens_seen": 976289792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035566700100300903,
      "loss": 2.8696,
      "theoretical_loss": 3.657527484087241,
      "tokens_seen": 976355328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035565697091273827,
      "loss": 2.7943,
      "theoretical_loss": 3.6575040177182205,
      "tokens_seen": 976420864
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 546148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.82014799118042,
      "objective/train/theoretical_loss": 3.6574805533651515,
      "objective/train/tokens_used": 996946400,
      "theoretical_loss": 3.6574805533651515,
      "tokens_seen": 976486400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556469408224674,
      "loss": 2.6284,
      "theoretical_loss": 3.6574805533651515,
      "tokens_seen": 976486400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035563691073219663,
      "loss": 2.7933,
      "theoretical_loss": 3.6574570910277258,
      "tokens_seen": 976551936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035562688064192576,
      "loss": 2.8423,
      "theoretical_loss": 3.6574336307056345,
      "tokens_seen": 976617472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355616850551655,
      "loss": 2.6661,
      "theoretical_loss": 3.65741017239857,
      "tokens_seen": 976683008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035560682046138417,
      "loss": 2.6578,
      "theoretical_loss": 3.6573867161062235,
      "tokens_seen": 976748544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035559679037111335,
      "loss": 2.6173,
      "theoretical_loss": 3.6573632618282867,
      "tokens_seen": 976814080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035558676028084253,
      "loss": 2.6873,
      "theoretical_loss": 3.6573398095644514,
      "tokens_seen": 976879616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555767301905717,
      "loss": 2.5844,
      "theoretical_loss": 3.6573163593144105,
      "tokens_seen": 976945152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555667001003009,
      "loss": 2.7901,
      "theoretical_loss": 3.657292911077855,
      "tokens_seen": 977010688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035555667001003013,
      "loss": 2.47,
      "theoretical_loss": 3.657269464854477,
      "tokens_seen": 977076224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035554663991975926,
      "loss": 2.576,
      "theoretical_loss": 3.657246020643969,
      "tokens_seen": 977141760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555366098294885,
      "loss": 2.7507,
      "theoretical_loss": 3.6572225784460235,
      "tokens_seen": 977207296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555265797392176,
      "loss": 2.6684,
      "theoretical_loss": 3.6571991382603324,
      "tokens_seen": 977272832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035551654964894686,
      "loss": 2.6614,
      "theoretical_loss": 3.6571757000865874,
      "tokens_seen": 977338368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035550651955867604,
      "loss": 2.8629,
      "theoretical_loss": 3.6571522639244822,
      "tokens_seen": 977403904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554964894684052,
      "loss": 2.6595,
      "theoretical_loss": 3.6571288297737086,
      "tokens_seen": 977469440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554864593781344,
      "loss": 2.6743,
      "theoretical_loss": 3.657105397633959,
      "tokens_seen": 977534976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035547642928786364,
      "loss": 2.7305,
      "theoretical_loss": 3.657081967504926,
      "tokens_seen": 977600512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035546639919759276,
      "loss": 2.7035,
      "theoretical_loss": 3.657058539386303,
      "tokens_seen": 977666048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355456369107322,
      "loss": 2.7986,
      "theoretical_loss": 3.6570351132777823,
      "tokens_seen": 977731584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554463390170511,
      "loss": 2.6578,
      "theoretical_loss": 3.657011689179056,
      "tokens_seen": 977797120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035543630892678036,
      "loss": 2.6389,
      "theoretical_loss": 3.656988267089818,
      "tokens_seen": 977862656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035542627883650954,
      "loss": 2.5991,
      "theoretical_loss": 3.6569648470097613,
      "tokens_seen": 977928192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554162487462387,
      "loss": 2.7215,
      "theoretical_loss": 3.6569414289385778,
      "tokens_seen": 977993728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554062186559679,
      "loss": 2.5016,
      "theoretical_loss": 3.6569180128759617,
      "tokens_seen": 978059264
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 547377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.918895721435547,
      "objective/train/theoretical_loss": 3.656894598821606,
      "objective/train/tokens_used": 998584800,
      "theoretical_loss": 3.656894598821606,
      "tokens_seen": 978124800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553961885656971,
      "loss": 2.7372,
      "theoretical_loss": 3.656894598821606,
      "tokens_seen": 978124800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035538615847542627,
      "loss": 2.6627,
      "theoretical_loss": 3.6568711867752035,
      "tokens_seen": 978190336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553761283851555,
      "loss": 2.7191,
      "theoretical_loss": 3.6568477767364476,
      "tokens_seen": 978255872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035536609829488463,
      "loss": 2.7288,
      "theoretical_loss": 3.656824368705032,
      "tokens_seen": 978321408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035535606820461386,
      "loss": 2.7022,
      "theoretical_loss": 3.6568009626806495,
      "tokens_seen": 978386944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035534603811434304,
      "loss": 2.6384,
      "theoretical_loss": 3.6567775586629945,
      "tokens_seen": 978452480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553360080240722,
      "loss": 2.8189,
      "theoretical_loss": 3.6567541566517594,
      "tokens_seen": 978518016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553259779338014,
      "loss": 2.7914,
      "theoretical_loss": 3.6567307566466387,
      "tokens_seen": 978583552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553159478435306,
      "loss": 2.8269,
      "theoretical_loss": 3.656707358647326,
      "tokens_seen": 978649088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035530591775325977,
      "loss": 2.4179,
      "theoretical_loss": 3.6566839626535153,
      "tokens_seen": 978714624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355295887662989,
      "loss": 2.6729,
      "theoretical_loss": 3.6566605686648996,
      "tokens_seen": 978780160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035528585757271813,
      "loss": 2.6041,
      "theoretical_loss": 3.6566371766811727,
      "tokens_seen": 978845696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035527582748244737,
      "loss": 2.5674,
      "theoretical_loss": 3.65661378670203,
      "tokens_seen": 978911232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552657973921765,
      "loss": 2.6614,
      "theoretical_loss": 3.6565903987271637,
      "tokens_seen": 978976768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035525576730190573,
      "loss": 2.8475,
      "theoretical_loss": 3.6565670127562697,
      "tokens_seen": 979042304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552457372116349,
      "loss": 2.7853,
      "theoretical_loss": 3.656543628789041,
      "tokens_seen": 979107840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552357071213641,
      "loss": 2.7585,
      "theoretical_loss": 3.6565202468251723,
      "tokens_seen": 979173376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035522567703109327,
      "loss": 2.7385,
      "theoretical_loss": 3.6564968668643574,
      "tokens_seen": 979238912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035521564694082245,
      "loss": 2.7629,
      "theoretical_loss": 3.656473488906291,
      "tokens_seen": 979304448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035520561685055163,
      "loss": 2.501,
      "theoretical_loss": 3.656450112950668,
      "tokens_seen": 979369984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035519558676028087,
      "loss": 2.6088,
      "theoretical_loss": 3.6564267389971823,
      "tokens_seen": 979435520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035518555667001,
      "loss": 2.4397,
      "theoretical_loss": 3.6564033670455283,
      "tokens_seen": 979501056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035517552657973923,
      "loss": 2.6795,
      "theoretical_loss": 3.656379997095401,
      "tokens_seen": 979566592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551654964894684,
      "loss": 2.5704,
      "theoretical_loss": 3.6563566291464955,
      "tokens_seen": 979632128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551554663991976,
      "loss": 2.7986,
      "theoretical_loss": 3.6563332631985053,
      "tokens_seen": 979697664
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 547866,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.091559648513794,
      "objective/train/theoretical_loss": 3.6563098992511267,
      "objective/train/tokens_used": 1000223200,
      "theoretical_loss": 3.6563098992511267,
      "tokens_seen": 979763200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551454363089268,
      "loss": 2.9244,
      "theoretical_loss": 3.6563098992511267,
      "tokens_seen": 979763200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035513540621865596,
      "loss": 2.7605,
      "theoretical_loss": 3.656286537304054,
      "tokens_seen": 979828736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035512537612838514,
      "loss": 3.0127,
      "theoretical_loss": 3.6562631773569816,
      "tokens_seen": 979894272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035511534603811437,
      "loss": 2.7358,
      "theoretical_loss": 3.6562398194096053,
      "tokens_seen": 979959808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551053159478435,
      "loss": 2.5052,
      "theoretical_loss": 3.65621646346162,
      "tokens_seen": 980025344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035509528585757273,
      "loss": 2.6234,
      "theoretical_loss": 3.6561931095127207,
      "tokens_seen": 980090880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550852557673019,
      "loss": 2.7937,
      "theoretical_loss": 3.656169757562603,
      "tokens_seen": 980156416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550752256770311,
      "loss": 2.7786,
      "theoretical_loss": 3.6561464076109615,
      "tokens_seen": 980221952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035506519558676033,
      "loss": 2.5644,
      "theoretical_loss": 3.6561230596574927,
      "tokens_seen": 980287488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035505516549648946,
      "loss": 2.791,
      "theoretical_loss": 3.656099713701891,
      "tokens_seen": 980353024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550451354062187,
      "loss": 2.7434,
      "theoretical_loss": 3.6560763697438525,
      "tokens_seen": 980418560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550351053159478,
      "loss": 2.6858,
      "theoretical_loss": 3.6560530277830723,
      "tokens_seen": 980484096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035502507522567706,
      "loss": 2.6007,
      "theoretical_loss": 3.656029687819246,
      "tokens_seen": 980549632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035501504513540624,
      "loss": 2.7632,
      "theoretical_loss": 3.6560063498520705,
      "tokens_seen": 980615168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550050150451354,
      "loss": 2.6044,
      "theoretical_loss": 3.65598301388124,
      "tokens_seen": 980680704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549949849548646,
      "loss": 2.7278,
      "theoretical_loss": 3.6559596799064513,
      "tokens_seen": 980746240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035498495486459384,
      "loss": 2.6622,
      "theoretical_loss": 3.6559363479274003,
      "tokens_seen": 980811776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035497492477432296,
      "loss": 2.6397,
      "theoretical_loss": 3.6559130179437824,
      "tokens_seen": 980877312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549648946840522,
      "loss": 2.4674,
      "theoretical_loss": 3.655889689955294,
      "tokens_seen": 980942848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549548645937813,
      "loss": 2.5841,
      "theoretical_loss": 3.6558663639616307,
      "tokens_seen": 981008384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035494483450351056,
      "loss": 2.803,
      "theoretical_loss": 3.6558430399624893,
      "tokens_seen": 981073920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035493480441323974,
      "loss": 2.8485,
      "theoretical_loss": 3.655819717957566,
      "tokens_seen": 981139456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549247743229689,
      "loss": 2.7757,
      "theoretical_loss": 3.655796397946557,
      "tokens_seen": 981204992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549147442326981,
      "loss": 2.6201,
      "theoretical_loss": 3.6557730799291583,
      "tokens_seen": 981270528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549047141424273,
      "loss": 2.6034,
      "theoretical_loss": 3.655749763905067,
      "tokens_seen": 981336064
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 549153,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9054205417633057,
      "objective/train/theoretical_loss": 3.655726449873979,
      "objective/train/tokens_used": 1001861600,
      "theoretical_loss": 3.655726449873979,
      "tokens_seen": 981401600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035489468405215647,
      "loss": 2.9853,
      "theoretical_loss": 3.655726449873979,
      "tokens_seen": 981401600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548846539618857,
      "loss": 2.9481,
      "theoretical_loss": 3.6557031378355913,
      "tokens_seen": 981467136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035487462387161483,
      "loss": 2.6993,
      "theoretical_loss": 3.6556798277896005,
      "tokens_seen": 981532672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035486459378134406,
      "loss": 2.8664,
      "theoretical_loss": 3.655656519735703,
      "tokens_seen": 981598208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035485456369107324,
      "loss": 2.7443,
      "theoretical_loss": 3.655633213673596,
      "tokens_seen": 981663744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548445336008024,
      "loss": 2.4849,
      "theoretical_loss": 3.6556099096029757,
      "tokens_seen": 981729280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548345035105316,
      "loss": 2.8908,
      "theoretical_loss": 3.6555866075235395,
      "tokens_seen": 981794816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548244734202608,
      "loss": 2.4488,
      "theoretical_loss": 3.655563307434984,
      "tokens_seen": 981860352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035481444332998997,
      "loss": 2.7768,
      "theoretical_loss": 3.655540009337007,
      "tokens_seen": 981925888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548044132397192,
      "loss": 2.7123,
      "theoretical_loss": 3.6555167132293045,
      "tokens_seen": 981991424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035479438314944833,
      "loss": 2.587,
      "theoretical_loss": 3.6554934191115747,
      "tokens_seen": 982056960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035478435305917757,
      "loss": 2.6422,
      "theoretical_loss": 3.6554701269835146,
      "tokens_seen": 982122496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547743229689067,
      "loss": 2.6564,
      "theoretical_loss": 3.6554468368448205,
      "tokens_seen": 982188032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035476429287863593,
      "loss": 2.6752,
      "theoretical_loss": 3.6554235486951914,
      "tokens_seen": 982253568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547542627883651,
      "loss": 2.6921,
      "theoretical_loss": 3.6554002625343234,
      "tokens_seen": 982319104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547442326980943,
      "loss": 2.7904,
      "theoretical_loss": 3.6553769783619146,
      "tokens_seen": 982384640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035473420260782347,
      "loss": 2.7994,
      "theoretical_loss": 3.6553536961776625,
      "tokens_seen": 982450176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035472417251755265,
      "loss": 2.6377,
      "theoretical_loss": 3.6553304159812647,
      "tokens_seen": 982515712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035471414242728183,
      "loss": 2.8887,
      "theoretical_loss": 3.655307137772419,
      "tokens_seen": 982581248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035470411233701107,
      "loss": 2.7482,
      "theoretical_loss": 3.655283861550823,
      "tokens_seen": 982646784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546940822467402,
      "loss": 2.5649,
      "theoretical_loss": 3.6552605873161745,
      "tokens_seen": 982712320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035468405215646943,
      "loss": 2.7879,
      "theoretical_loss": 3.655237315068171,
      "tokens_seen": 982777856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546740220661986,
      "loss": 2.8483,
      "theoretical_loss": 3.655214044806512,
      "tokens_seen": 982843392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546639919759278,
      "loss": 2.7856,
      "theoretical_loss": 3.6551907765308935,
      "tokens_seen": 982908928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354653961885657,
      "loss": 2.7548,
      "theoretical_loss": 3.655167510241015,
      "tokens_seen": 982974464
    },
    {
      "debugging/Self-BLEU-5": 0.5000497629393682,
      "debugging/distinct-1-grams": 0.7839180687368331,
      "debugging/distinct-2-grams": 0.9723027630693535,
      "debugging/entropy-1-grams": 5.462779182741081,
      "debugging/entropy-2-grams": 6.200783858933273,
      "debugging/length": 666.1666666666666,
      "debugging/num_segments": 6,
      "epoch": 0.3,
      "objective/train/docs_used": 549695,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.27803897857666,
      "objective/train/theoretical_loss": 3.655144245936574,
      "objective/train/tokens_used": 1003500000,
      "theoretical_loss": 3.655144245936574,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035464393179538616,
      "loss": 2.4059,
      "theoretical_loss": 3.655144245936574,
      "tokens_seen": 983040000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035463390170511534,
      "loss": 2.8016,
      "theoretical_loss": 3.6551209836172696,
      "tokens_seen": 983105536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035462387161484457,
      "loss": 2.5838,
      "theoretical_loss": 3.6550977232827986,
      "tokens_seen": 983171072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546138415245737,
      "loss": 2.5899,
      "theoretical_loss": 3.65507446493286,
      "tokens_seen": 983236608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035460381143430293,
      "loss": 2.74,
      "theoretical_loss": 3.655051208567153,
      "tokens_seen": 983302144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035459378134403206,
      "loss": 2.8234,
      "theoretical_loss": 3.655027954185376,
      "tokens_seen": 983367680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545837512537613,
      "loss": 2.7215,
      "theoretical_loss": 3.6550047017872265,
      "tokens_seen": 983433216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545737211634905,
      "loss": 2.6825,
      "theoretical_loss": 3.6549814513724037,
      "tokens_seen": 983498752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035456369107321966,
      "loss": 2.6115,
      "theoretical_loss": 3.6549582029406062,
      "tokens_seen": 983564288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035455366098294884,
      "loss": 2.7599,
      "theoretical_loss": 3.654934956491533,
      "tokens_seen": 983629824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354543630892678,
      "loss": 2.7921,
      "theoretical_loss": 3.654911712024883,
      "tokens_seen": 983695360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545336008024072,
      "loss": 2.7611,
      "theoretical_loss": 3.6548884695403543,
      "tokens_seen": 983760896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035452357071213644,
      "loss": 2.7598,
      "theoretical_loss": 3.6548652290376467,
      "tokens_seen": 983826432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035451354062186556,
      "loss": 2.7005,
      "theoretical_loss": 3.654841990516459,
      "tokens_seen": 983891968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545035105315948,
      "loss": 2.5206,
      "theoretical_loss": 3.65481875397649,
      "tokens_seen": 983957504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354493480441324,
      "loss": 2.5426,
      "theoretical_loss": 3.654795519417439,
      "tokens_seen": 984023040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035448345035105316,
      "loss": 2.8029,
      "theoretical_loss": 3.6547722868390053,
      "tokens_seen": 984088576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035447342026078234,
      "loss": 2.813,
      "theoretical_loss": 3.654749056240888,
      "tokens_seen": 984154112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544633901705115,
      "loss": 2.5548,
      "theoretical_loss": 3.6547258276227863,
      "tokens_seen": 984219648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544533600802407,
      "loss": 2.6819,
      "theoretical_loss": 3.6547026009844004,
      "tokens_seen": 984285184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035444332998996994,
      "loss": 2.5199,
      "theoretical_loss": 3.6546793763254293,
      "tokens_seen": 984350720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035443329989969907,
      "loss": 2.7544,
      "theoretical_loss": 3.6546561536455724,
      "tokens_seen": 984416256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544232698094283,
      "loss": 2.6461,
      "theoretical_loss": 3.654632932944529,
      "tokens_seen": 984481792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035441323971915743,
      "loss": 2.6753,
      "theoretical_loss": 3.654609714222,
      "tokens_seen": 984547328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035440320962888667,
      "loss": 2.6239,
      "theoretical_loss": 3.654586497477684,
      "tokens_seen": 984612864
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 550994,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.944199562072754,
      "objective/train/theoretical_loss": 3.6545632827112806,
      "objective/train/tokens_used": 1005138400,
      "theoretical_loss": 3.6545632827112806,
      "tokens_seen": 984678400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035439317953861585,
      "loss": 2.7087,
      "theoretical_loss": 3.6545632827112806,
      "tokens_seen": 984678400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035438314944834503,
      "loss": 2.7407,
      "theoretical_loss": 3.654540069922491,
      "tokens_seen": 984743936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543731193580742,
      "loss": 2.7961,
      "theoretical_loss": 3.6545168591110135,
      "tokens_seen": 984809472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035436308926780344,
      "loss": 2.5277,
      "theoretical_loss": 3.6544936502765495,
      "tokens_seen": 984875008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543530591775326,
      "loss": 2.5603,
      "theoretical_loss": 3.6544704434187985,
      "tokens_seen": 984940544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543430290872618,
      "loss": 2.4406,
      "theoretical_loss": 3.6544472385374602,
      "tokens_seen": 985006080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354332998996991,
      "loss": 2.9052,
      "theoretical_loss": 3.654424035632236,
      "tokens_seen": 985071616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035432296890672017,
      "loss": 2.5012,
      "theoretical_loss": 3.654400834702825,
      "tokens_seen": 985137152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543129388164494,
      "loss": 2.8039,
      "theoretical_loss": 3.6543776357489275,
      "tokens_seen": 985202688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035430290872617853,
      "loss": 2.7412,
      "theoretical_loss": 3.6543544387702447,
      "tokens_seen": 985268224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035429287863590777,
      "loss": 2.9997,
      "theoretical_loss": 3.654331243766477,
      "tokens_seen": 985333760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542828485456369,
      "loss": 2.8078,
      "theoretical_loss": 3.654308050737324,
      "tokens_seen": 985399296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035427281845536613,
      "loss": 2.6121,
      "theoretical_loss": 3.6542848596824875,
      "tokens_seen": 985464832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542627883650953,
      "loss": 2.5112,
      "theoretical_loss": 3.6542616706016675,
      "tokens_seen": 985530368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542527582748245,
      "loss": 2.4367,
      "theoretical_loss": 3.6542384834945643,
      "tokens_seen": 985595904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035424272818455367,
      "loss": 2.6558,
      "theoretical_loss": 3.654215298360879,
      "tokens_seen": 985661440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035423269809428285,
      "loss": 2.5777,
      "theoretical_loss": 3.6541921152003134,
      "tokens_seen": 985726976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035422266800401203,
      "loss": 2.7154,
      "theoretical_loss": 3.654168934012567,
      "tokens_seen": 985792512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035421263791374127,
      "loss": 2.471,
      "theoretical_loss": 3.6541457547973417,
      "tokens_seen": 985858048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542026078234704,
      "loss": 2.8969,
      "theoretical_loss": 3.654122577554338,
      "tokens_seen": 985923584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035419257773319963,
      "loss": 2.5885,
      "theoretical_loss": 3.6540994022832574,
      "tokens_seen": 985989120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541825476429288,
      "loss": 2.7422,
      "theoretical_loss": 3.6540762289838007,
      "tokens_seen": 986054656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354172517552658,
      "loss": 2.8467,
      "theoretical_loss": 3.65405305765567,
      "tokens_seen": 986120192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541624874623872,
      "loss": 2.7443,
      "theoretical_loss": 3.654029888298566,
      "tokens_seen": 986185728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035415245737211636,
      "loss": 2.5481,
      "theoretical_loss": 3.6540067209121894,
      "tokens_seen": 986251264
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 551567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.91987681388855,
      "objective/train/theoretical_loss": 3.653983555496242,
      "objective/train/tokens_used": 1006776800,
      "theoretical_loss": 3.653983555496242,
      "tokens_seen": 986316800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035414242728184554,
      "loss": 2.6924,
      "theoretical_loss": 3.653983555496242,
      "tokens_seen": 986316800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035413239719157477,
      "loss": 2.7576,
      "theoretical_loss": 3.6539603920504264,
      "tokens_seen": 986382336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541223671013039,
      "loss": 2.6437,
      "theoretical_loss": 3.653937230574443,
      "tokens_seen": 986447872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035411233701103314,
      "loss": 2.4609,
      "theoretical_loss": 3.653914071067994,
      "tokens_seen": 986513408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035410230692076226,
      "loss": 2.7274,
      "theoretical_loss": 3.6538909135307813,
      "tokens_seen": 986578944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540922768304915,
      "loss": 2.7185,
      "theoretical_loss": 3.653867757962505,
      "tokens_seen": 986644480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540822467402207,
      "loss": 2.4144,
      "theoretical_loss": 3.6538446043628694,
      "tokens_seen": 986710016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035407221664994986,
      "loss": 2.9986,
      "theoretical_loss": 3.6538214527315747,
      "tokens_seen": 986775552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035406218655967904,
      "loss": 2.7329,
      "theoretical_loss": 3.6537983030683234,
      "tokens_seen": 986841088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540521564694082,
      "loss": 2.5728,
      "theoretical_loss": 3.6537751553728173,
      "tokens_seen": 986906624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540421263791374,
      "loss": 2.6988,
      "theoretical_loss": 3.6537520096447587,
      "tokens_seen": 986972160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035403209628886664,
      "loss": 2.8755,
      "theoretical_loss": 3.65372886588385,
      "tokens_seen": 987037696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035402206619859577,
      "loss": 2.4468,
      "theoretical_loss": 3.653705724089793,
      "tokens_seen": 987103232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354012036108325,
      "loss": 2.6212,
      "theoretical_loss": 3.65368258426229,
      "tokens_seen": 987168768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540020060180542,
      "loss": 2.5444,
      "theoretical_loss": 3.6536594464010435,
      "tokens_seen": 987234304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035399197592778336,
      "loss": 2.7203,
      "theoretical_loss": 3.653636310505756,
      "tokens_seen": 987299840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035398194583751254,
      "loss": 2.7516,
      "theoretical_loss": 3.6536131765761297,
      "tokens_seen": 987365376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539719157472417,
      "loss": 2.6541,
      "theoretical_loss": 3.653590044611867,
      "tokens_seen": 987430912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539618856569709,
      "loss": 2.6099,
      "theoretical_loss": 3.653566914612671,
      "tokens_seen": 987496448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035395185556670014,
      "loss": 2.7547,
      "theoretical_loss": 3.653543786578244,
      "tokens_seen": 987561984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035394182547642927,
      "loss": 2.8886,
      "theoretical_loss": 3.6535206605082893,
      "tokens_seen": 987627520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539317953861585,
      "loss": 2.7337,
      "theoretical_loss": 3.653497536402509,
      "tokens_seen": 987693056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035392176529588763,
      "loss": 2.7313,
      "theoretical_loss": 3.6534744142606064,
      "tokens_seen": 987758592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035391173520561687,
      "loss": 2.7238,
      "theoretical_loss": 3.653451294082284,
      "tokens_seen": 987824128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035390170511534605,
      "loss": 2.7589,
      "theoretical_loss": 3.653428175867245,
      "tokens_seen": 987889664
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 552650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.231137752532959,
      "objective/train/theoretical_loss": 3.653405059615193,
      "objective/train/tokens_used": 1008415200,
      "theoretical_loss": 3.653405059615193,
      "tokens_seen": 987955200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035389167502507523,
      "loss": 2.8518,
      "theoretical_loss": 3.653405059615193,
      "tokens_seen": 987955200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538816449348044,
      "loss": 2.7992,
      "theoretical_loss": 3.65338194532583,
      "tokens_seen": 988020736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035387161484453364,
      "loss": 2.4641,
      "theoretical_loss": 3.6533588329988604,
      "tokens_seen": 988086272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035386158475426277,
      "loss": 2.7381,
      "theoretical_loss": 3.6533357226339866,
      "tokens_seen": 988151808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353851554663992,
      "loss": 2.7929,
      "theoretical_loss": 3.653312614230912,
      "tokens_seen": 988217344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035384152457372113,
      "loss": 2.8832,
      "theoretical_loss": 3.653289507789341,
      "tokens_seen": 988282880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035383149448345037,
      "loss": 2.6008,
      "theoretical_loss": 3.653266403308976,
      "tokens_seen": 988348416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035382146439317955,
      "loss": 2.7089,
      "theoretical_loss": 3.65324330078952,
      "tokens_seen": 988413952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035381143430290873,
      "loss": 2.5198,
      "theoretical_loss": 3.6532202002306784,
      "tokens_seen": 988479488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538014042126379,
      "loss": 2.7445,
      "theoretical_loss": 3.653197101632153,
      "tokens_seen": 988545024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537913741223671,
      "loss": 2.627,
      "theoretical_loss": 3.653174004993648,
      "tokens_seen": 988610560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537813440320963,
      "loss": 2.7365,
      "theoretical_loss": 3.653150910314868,
      "tokens_seen": 988676096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537713139418255,
      "loss": 2.6624,
      "theoretical_loss": 3.653127817595516,
      "tokens_seen": 988741632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035376128385155464,
      "loss": 2.8601,
      "theoretical_loss": 3.6531047268352967,
      "tokens_seen": 988807168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035375125376128387,
      "loss": 2.7081,
      "theoretical_loss": 3.653081638033913,
      "tokens_seen": 988872704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353741223671013,
      "loss": 2.7607,
      "theoretical_loss": 3.65305855119107,
      "tokens_seen": 988938240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035373119358074223,
      "loss": 2.6208,
      "theoretical_loss": 3.6530354663064704,
      "tokens_seen": 989003776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537211634904714,
      "loss": 2.6153,
      "theoretical_loss": 3.65301238337982,
      "tokens_seen": 989069312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537111334002006,
      "loss": 2.839,
      "theoretical_loss": 3.6529893024108215,
      "tokens_seen": 989134848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537011033099298,
      "loss": 2.5253,
      "theoretical_loss": 3.65296622339918,
      "tokens_seen": 989200384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353691073219659,
      "loss": 2.7016,
      "theoretical_loss": 3.6529431463446,
      "tokens_seen": 989265920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035368104312938814,
      "loss": 2.6507,
      "theoretical_loss": 3.652920071246786,
      "tokens_seen": 989331456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536710130391174,
      "loss": 2.4649,
      "theoretical_loss": 3.652896998105442,
      "tokens_seen": 989396992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536609829488465,
      "loss": 2.844,
      "theoretical_loss": 3.652873926920272,
      "tokens_seen": 989462528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035365095285857574,
      "loss": 2.8394,
      "theoretical_loss": 3.652850857690982,
      "tokens_seen": 989528064
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 553159,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4328830242156982,
      "objective/train/theoretical_loss": 3.6528277904172755,
      "objective/train/tokens_used": 1010053600,
      "theoretical_loss": 3.6528277904172755,
      "tokens_seen": 989593600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536409227683049,
      "loss": 2.5894,
      "theoretical_loss": 3.6528277904172755,
      "tokens_seen": 989593600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536308926780341,
      "loss": 2.3536,
      "theoretical_loss": 3.652804725098858,
      "tokens_seen": 989659136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536208625877633,
      "loss": 2.737,
      "theoretical_loss": 3.6527816617354336,
      "tokens_seen": 989724672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035361083249749246,
      "loss": 2.4307,
      "theoretical_loss": 3.652758600326708,
      "tokens_seen": 989790208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536008024072217,
      "loss": 2.6266,
      "theoretical_loss": 3.652735540872385,
      "tokens_seen": 989855744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535907723169509,
      "loss": 2.5696,
      "theoretical_loss": 3.652712483372171,
      "tokens_seen": 989921280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035358074222668006,
      "loss": 2.7762,
      "theoretical_loss": 3.65268942782577,
      "tokens_seen": 989986816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035357071213640924,
      "loss": 2.707,
      "theoretical_loss": 3.652666374232887,
      "tokens_seen": 990052352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535606820461384,
      "loss": 2.5712,
      "theoretical_loss": 3.652643322593228,
      "tokens_seen": 990117888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535506519558676,
      "loss": 2.6914,
      "theoretical_loss": 3.652620272906498,
      "tokens_seen": 990183424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035354062186559684,
      "loss": 2.5723,
      "theoretical_loss": 3.6525972251724017,
      "tokens_seen": 990248960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035353059177532597,
      "loss": 2.4504,
      "theoretical_loss": 3.652574179390645,
      "tokens_seen": 990314496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535205616850552,
      "loss": 2.5014,
      "theoretical_loss": 3.6525511355609335,
      "tokens_seen": 990380032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535105315947844,
      "loss": 2.7155,
      "theoretical_loss": 3.6525280936829723,
      "tokens_seen": 990445568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035350050150451356,
      "loss": 2.6203,
      "theoretical_loss": 3.6525050537564674,
      "tokens_seen": 990511104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035349047141424274,
      "loss": 2.4302,
      "theoretical_loss": 3.652482015781124,
      "tokens_seen": 990576640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534804413239719,
      "loss": 2.7063,
      "theoretical_loss": 3.6524589797566476,
      "tokens_seen": 990642176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534704112337011,
      "loss": 2.6519,
      "theoretical_loss": 3.6524359456827447,
      "tokens_seen": 990707712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035346038114343034,
      "loss": 2.5755,
      "theoretical_loss": 3.6524129135591203,
      "tokens_seen": 990773248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035345035105315947,
      "loss": 2.6261,
      "theoretical_loss": 3.652389883385481,
      "tokens_seen": 990838784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534403209628887,
      "loss": 2.5731,
      "theoretical_loss": 3.6523668551615325,
      "tokens_seen": 990904320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035343029087261783,
      "loss": 2.7021,
      "theoretical_loss": 3.6523438288869805,
      "tokens_seen": 990969856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035342026078234707,
      "loss": 2.7016,
      "theoretical_loss": 3.6523208045615316,
      "tokens_seen": 991035392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035341023069207625,
      "loss": 2.5318,
      "theoretical_loss": 3.652297782184892,
      "tokens_seen": 991100928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035340020060180543,
      "loss": 2.6415,
      "theoretical_loss": 3.6522747617567672,
      "tokens_seen": 991166464
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 554508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6566503047943115,
      "objective/train/theoretical_loss": 3.6522517432768637,
      "objective/train/tokens_used": 1011692000,
      "theoretical_loss": 3.6522517432768637,
      "tokens_seen": 991232000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533901705115346,
      "loss": 2.6658,
      "theoretical_loss": 3.6522517432768637,
      "tokens_seen": 991232000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035338014042126384,
      "loss": 2.6459,
      "theoretical_loss": 3.6522287267448883,
      "tokens_seen": 991297536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035337011033099297,
      "loss": 2.5836,
      "theoretical_loss": 3.652205712160547,
      "tokens_seen": 991363072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533600802407222,
      "loss": 2.8489,
      "theoretical_loss": 3.652182699523546,
      "tokens_seen": 991428608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035335005015045133,
      "loss": 2.7108,
      "theoretical_loss": 3.6521596888335925,
      "tokens_seen": 991494144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035334002006018057,
      "loss": 2.5673,
      "theoretical_loss": 3.6521366800903925,
      "tokens_seen": 991559680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035332998996990975,
      "loss": 2.727,
      "theoretical_loss": 3.6521136732936528,
      "tokens_seen": 991625216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035331995987963893,
      "loss": 2.7111,
      "theoretical_loss": 3.6520906684430807,
      "tokens_seen": 991690752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533099297893681,
      "loss": 2.5756,
      "theoretical_loss": 3.652067665538382,
      "tokens_seen": 991756288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532998996990973,
      "loss": 2.5974,
      "theoretical_loss": 3.652044664579264,
      "tokens_seen": 991821824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532898696088265,
      "loss": 2.425,
      "theoretical_loss": 3.652021665565434,
      "tokens_seen": 991887360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532798395185557,
      "loss": 2.6744,
      "theoretical_loss": 3.6519986684965984,
      "tokens_seen": 991952896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035326980942828484,
      "loss": 2.7236,
      "theoretical_loss": 3.6519756733724646,
      "tokens_seen": 992018432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035325977933801407,
      "loss": 2.6378,
      "theoretical_loss": 3.651952680192739,
      "tokens_seen": 992083968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532497492477432,
      "loss": 2.6454,
      "theoretical_loss": 3.6519296889571295,
      "tokens_seen": 992149504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035323971915747243,
      "loss": 2.6259,
      "theoretical_loss": 3.651906699665343,
      "tokens_seen": 992215040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532296890672016,
      "loss": 2.5189,
      "theoretical_loss": 3.651883712317087,
      "tokens_seen": 992280576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532196589769308,
      "loss": 2.8199,
      "theoretical_loss": 3.6518607269120684,
      "tokens_seen": 992346112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035320962888666,
      "loss": 2.6205,
      "theoretical_loss": 3.6518377434499953,
      "tokens_seen": 992411648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531995987963892,
      "loss": 2.8705,
      "theoretical_loss": 3.651814761930575,
      "tokens_seen": 992477184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035318956870611834,
      "loss": 2.7282,
      "theoretical_loss": 3.651791782353514,
      "tokens_seen": 992542720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531795386158476,
      "loss": 2.5389,
      "theoretical_loss": 3.6517688047185217,
      "tokens_seen": 992608256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531695085255767,
      "loss": 2.6315,
      "theoretical_loss": 3.6517458290253044,
      "tokens_seen": 992673792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035315947843530594,
      "loss": 2.7756,
      "theoretical_loss": 3.6517228552735697,
      "tokens_seen": 992739328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531494483450351,
      "loss": 2.6943,
      "theoretical_loss": 3.651699883463026,
      "tokens_seen": 992804864
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 555131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2285315990448,
      "objective/train/theoretical_loss": 3.6516769135933815,
      "objective/train/tokens_used": 1013330400,
      "theoretical_loss": 3.6516769135933815,
      "tokens_seen": 992870400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531394182547643,
      "loss": 2.5577,
      "theoretical_loss": 3.6516769135933815,
      "tokens_seen": 992870400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531293881644935,
      "loss": 2.6386,
      "theoretical_loss": 3.6516539456643438,
      "tokens_seen": 992935936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035311935807422266,
      "loss": 2.8148,
      "theoretical_loss": 3.6516309796756206,
      "tokens_seen": 993001472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035310932798395184,
      "loss": 2.7412,
      "theoretical_loss": 3.65160801562692,
      "tokens_seen": 993067008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530992978936811,
      "loss": 2.8597,
      "theoretical_loss": 3.6515850535179504,
      "tokens_seen": 993132544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530892678034102,
      "loss": 2.4166,
      "theoretical_loss": 3.65156209334842,
      "tokens_seen": 993198080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035307923771313944,
      "loss": 2.5507,
      "theoretical_loss": 3.651539135118036,
      "tokens_seen": 993263616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035306920762286857,
      "loss": 2.7169,
      "theoretical_loss": 3.6515161788265083,
      "tokens_seen": 993329152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530591775325978,
      "loss": 2.607,
      "theoretical_loss": 3.6514932244735436,
      "tokens_seen": 993394688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353049147442327,
      "loss": 2.7023,
      "theoretical_loss": 3.6514702720588517,
      "tokens_seen": 993460224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035303911735205617,
      "loss": 2.3688,
      "theoretical_loss": 3.6514473215821406,
      "tokens_seen": 993525760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035302908726178535,
      "loss": 2.7075,
      "theoretical_loss": 3.6514243730431186,
      "tokens_seen": 993591296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530190571715146,
      "loss": 2.4497,
      "theoretical_loss": 3.6514014264414953,
      "tokens_seen": 993656832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530090270812437,
      "loss": 2.8638,
      "theoretical_loss": 3.6513784817769777,
      "tokens_seen": 993722368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035299899699097294,
      "loss": 2.8829,
      "theoretical_loss": 3.6513555390492756,
      "tokens_seen": 993787904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035298896690070207,
      "loss": 2.6749,
      "theoretical_loss": 3.6513325982580978,
      "tokens_seen": 993853440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529789368104313,
      "loss": 2.4084,
      "theoretical_loss": 3.651309659403153,
      "tokens_seen": 993918976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529689067201605,
      "loss": 2.6559,
      "theoretical_loss": 3.6512867224841505,
      "tokens_seen": 993984512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035295887662988967,
      "loss": 2.6972,
      "theoretical_loss": 3.6512637875007985,
      "tokens_seen": 994050048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035294884653961885,
      "loss": 2.8845,
      "theoretical_loss": 3.6512408544528063,
      "tokens_seen": 994115584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035293881644934803,
      "loss": 2.7391,
      "theoretical_loss": 3.6512179233398836,
      "tokens_seen": 994181120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529287863590772,
      "loss": 2.8006,
      "theoretical_loss": 3.651194994161739,
      "tokens_seen": 994246656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035291875626880645,
      "loss": 2.6687,
      "theoretical_loss": 3.6511720669180816,
      "tokens_seen": 994312192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529087261785356,
      "loss": 2.5931,
      "theoretical_loss": 3.6511491416086215,
      "tokens_seen": 994377728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528986960882648,
      "loss": 2.7499,
      "theoretical_loss": 3.6511262182330673,
      "tokens_seen": 994443264
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 556491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.311375617980957,
      "objective/train/theoretical_loss": 3.6511032967911285,
      "objective/train/tokens_used": 1014968800,
      "theoretical_loss": 3.6511032967911285,
      "tokens_seen": 994508800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035288866599799394,
      "loss": 2.5432,
      "theoretical_loss": 3.6511032967911285,
      "tokens_seen": 994508800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035287863590772317,
      "loss": 2.6023,
      "theoretical_loss": 3.6510803772825153,
      "tokens_seen": 994574336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035286860581745235,
      "loss": 2.411,
      "theoretical_loss": 3.6510574597069367,
      "tokens_seen": 994639872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035285857572718153,
      "loss": 2.6745,
      "theoretical_loss": 3.6510345440641023,
      "tokens_seen": 994705408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035284854563691077,
      "loss": 2.8324,
      "theoretical_loss": 3.6510116303537217,
      "tokens_seen": 994770944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035283851554663995,
      "loss": 2.6422,
      "theoretical_loss": 3.6509887185755048,
      "tokens_seen": 994836480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035282848545636913,
      "loss": 2.4072,
      "theoretical_loss": 3.650965808729161,
      "tokens_seen": 994902016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528184553660983,
      "loss": 2.4776,
      "theoretical_loss": 3.6509429008144014,
      "tokens_seen": 994967552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528084252758275,
      "loss": 2.7632,
      "theoretical_loss": 3.650919994830935,
      "tokens_seen": 995033088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527983951855567,
      "loss": 2.588,
      "theoretical_loss": 3.6508970907784715,
      "tokens_seen": 995098624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527883650952859,
      "loss": 2.5729,
      "theoretical_loss": 3.6508741886567213,
      "tokens_seen": 995164160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035277833500501504,
      "loss": 2.7314,
      "theoretical_loss": 3.650851288465395,
      "tokens_seen": 995229696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035276830491474427,
      "loss": 2.7634,
      "theoretical_loss": 3.650828390204202,
      "tokens_seen": 995295232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527582748244734,
      "loss": 2.6234,
      "theoretical_loss": 3.650805493872853,
      "tokens_seen": 995360768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035274824473420263,
      "loss": 2.8951,
      "theoretical_loss": 3.650782599471058,
      "tokens_seen": 995426304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527382146439318,
      "loss": 2.6232,
      "theoretical_loss": 3.650759706998528,
      "tokens_seen": 995491840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352728184553661,
      "loss": 2.6352,
      "theoretical_loss": 3.6507368164549727,
      "tokens_seen": 995557376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527181544633902,
      "loss": 2.9256,
      "theoretical_loss": 3.650713927840103,
      "tokens_seen": 995622912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527081243731194,
      "loss": 2.4447,
      "theoretical_loss": 3.650691041153629,
      "tokens_seen": 995688448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035269809428284854,
      "loss": 2.6017,
      "theoretical_loss": 3.6506681563952625,
      "tokens_seen": 995753984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526880641925778,
      "loss": 2.6119,
      "theoretical_loss": 3.6506452735647126,
      "tokens_seen": 995819520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526780341023069,
      "loss": 2.8533,
      "theoretical_loss": 3.650622392661691,
      "tokens_seen": 995885056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035266800401203614,
      "loss": 2.7495,
      "theoretical_loss": 3.6505995136859086,
      "tokens_seen": 995950592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526579739217653,
      "loss": 2.5176,
      "theoretical_loss": 3.6505766366370755,
      "tokens_seen": 996016128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526479438314945,
      "loss": 2.5834,
      "theoretical_loss": 3.6505537615149035,
      "tokens_seen": 996081664
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 557069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.978800654411316,
      "objective/train/theoretical_loss": 3.650530888319103,
      "objective/train/tokens_used": 1016607200,
      "theoretical_loss": 3.650530888319103,
      "tokens_seen": 996147200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526379137412237,
      "loss": 2.3804,
      "theoretical_loss": 3.650530888319103,
      "tokens_seen": 996147200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035262788365095286,
      "loss": 2.7149,
      "theoretical_loss": 3.6505080170493853,
      "tokens_seen": 996212736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035261785356068204,
      "loss": 2.6825,
      "theoretical_loss": 3.6504851477054614,
      "tokens_seen": 996278272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526078234704113,
      "loss": 2.9308,
      "theoretical_loss": 3.6504622802870426,
      "tokens_seen": 996343808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525977933801404,
      "loss": 2.7761,
      "theoretical_loss": 3.6504394147938406,
      "tokens_seen": 996409344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035258776328986964,
      "loss": 2.7075,
      "theoretical_loss": 3.6504165512255655,
      "tokens_seen": 996474880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035257773319959877,
      "loss": 2.6206,
      "theoretical_loss": 3.6503936895819296,
      "tokens_seen": 996540416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352567703109328,
      "loss": 2.5507,
      "theoretical_loss": 3.650370829862644,
      "tokens_seen": 996605952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525576730190572,
      "loss": 2.5328,
      "theoretical_loss": 3.6503479720674203,
      "tokens_seen": 996671488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035254764292878637,
      "loss": 2.7378,
      "theoretical_loss": 3.6503251161959707,
      "tokens_seen": 996737024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035253761283851555,
      "loss": 2.6422,
      "theoretical_loss": 3.6503022622480055,
      "tokens_seen": 996802560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525275827482448,
      "loss": 2.5801,
      "theoretical_loss": 3.650279410223237,
      "tokens_seen": 996868096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525175526579739,
      "loss": 2.6606,
      "theoretical_loss": 3.6502565601213774,
      "tokens_seen": 996933632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035250752256770314,
      "loss": 2.4713,
      "theoretical_loss": 3.6502337119421377,
      "tokens_seen": 996999168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035249749247743227,
      "loss": 2.68,
      "theoretical_loss": 3.6502108656852306,
      "tokens_seen": 997064704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524874623871615,
      "loss": 2.5103,
      "theoretical_loss": 3.6501880213503672,
      "tokens_seen": 997130240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524774322968907,
      "loss": 2.6101,
      "theoretical_loss": 3.65016517893726,
      "tokens_seen": 997195776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035246740220661987,
      "loss": 2.6795,
      "theoretical_loss": 3.650142338445621,
      "tokens_seen": 997261312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035245737211634905,
      "loss": 2.6608,
      "theoretical_loss": 3.6501194998751623,
      "tokens_seen": 997326848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035244734202607823,
      "loss": 2.7978,
      "theoretical_loss": 3.650096663225596,
      "tokens_seen": 997392384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524373119358074,
      "loss": 2.7366,
      "theoretical_loss": 3.650073828496634,
      "tokens_seen": 997457920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035242728184553665,
      "loss": 2.9977,
      "theoretical_loss": 3.650050995687989,
      "tokens_seen": 997523456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524172517552658,
      "loss": 2.9039,
      "theoretical_loss": 3.650028164799374,
      "tokens_seen": 997588992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352407221664995,
      "loss": 2.6053,
      "theoretical_loss": 3.6500053358305005,
      "tokens_seen": 997654528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035239719157472414,
      "loss": 2.6113,
      "theoretical_loss": 3.649982508781081,
      "tokens_seen": 997720064
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 558090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5171451568603516,
      "objective/train/theoretical_loss": 3.6499596836508283,
      "objective/train/tokens_used": 1018245600,
      "theoretical_loss": 3.6499596836508283,
      "tokens_seen": 997785600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035238716148445337,
      "loss": 2.7921,
      "theoretical_loss": 3.6499596836508283,
      "tokens_seen": 997785600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035237713139418255,
      "loss": 2.5081,
      "theoretical_loss": 3.649936860439455,
      "tokens_seen": 997851136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035236710130391173,
      "loss": 2.8577,
      "theoretical_loss": 3.6499140391466733,
      "tokens_seen": 997916672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523570712136409,
      "loss": 2.6762,
      "theoretical_loss": 3.649891219772197,
      "tokens_seen": 997982208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035234704112337015,
      "loss": 2.6309,
      "theoretical_loss": 3.6498684023157386,
      "tokens_seen": 998047744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523370110330993,
      "loss": 2.9457,
      "theoretical_loss": 3.64984558677701,
      "tokens_seen": 998113280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523269809428285,
      "loss": 2.7832,
      "theoretical_loss": 3.6498227731557247,
      "tokens_seen": 998178816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035231695085255764,
      "loss": 2.6664,
      "theoretical_loss": 3.649799961451596,
      "tokens_seen": 998244352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523069207622869,
      "loss": 2.4468,
      "theoretical_loss": 3.6497771516643365,
      "tokens_seen": 998309888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035229689067201606,
      "loss": 2.6363,
      "theoretical_loss": 3.64975434379366,
      "tokens_seen": 998375424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035228686058174524,
      "loss": 2.6484,
      "theoretical_loss": 3.649731537839279,
      "tokens_seen": 998440960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522768304914744,
      "loss": 2.6946,
      "theoretical_loss": 3.6497087338009067,
      "tokens_seen": 998506496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522668004012036,
      "loss": 2.8492,
      "theoretical_loss": 3.649685931678256,
      "tokens_seen": 998572032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522567703109328,
      "loss": 2.3118,
      "theoretical_loss": 3.6496631314710415,
      "tokens_seen": 998637568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352246740220662,
      "loss": 2.8642,
      "theoretical_loss": 3.6496403331789757,
      "tokens_seen": 998703104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035223671013039114,
      "loss": 2.4645,
      "theoretical_loss": 3.649617536801773,
      "tokens_seen": 998768640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522266800401204,
      "loss": 2.5693,
      "theoretical_loss": 3.6495947423391453,
      "tokens_seen": 998834176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035221664994984956,
      "loss": 2.7414,
      "theoretical_loss": 3.6495719497908077,
      "tokens_seen": 998899712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035220661985957874,
      "loss": 2.7699,
      "theoretical_loss": 3.6495491591564733,
      "tokens_seen": 998965248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521965897693079,
      "loss": 2.698,
      "theoretical_loss": 3.6495263704358556,
      "tokens_seen": 999030784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521865596790371,
      "loss": 2.4147,
      "theoretical_loss": 3.6495035836286687,
      "tokens_seen": 999096320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521765295887663,
      "loss": 2.511,
      "theoretical_loss": 3.6494807987346265,
      "tokens_seen": 999161856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521664994984955,
      "loss": 2.9637,
      "theoretical_loss": 3.649458015753442,
      "tokens_seen": 999227392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035215646940822465,
      "loss": 2.5515,
      "theoretical_loss": 3.649435234684831,
      "tokens_seen": 999292928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521464393179539,
      "loss": 2.7807,
      "theoretical_loss": 3.6494124555285063,
      "tokens_seen": 999358464
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 559222,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.206585645675659,
      "objective/train/theoretical_loss": 3.649389678284182,
      "objective/train/tokens_used": 1019884000,
      "theoretical_loss": 3.649389678284182,
      "tokens_seen": 999424000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352136409227683,
      "loss": 2.775,
      "theoretical_loss": 3.649389678284182,
      "tokens_seen": 999424000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035212637913741224,
      "loss": 2.7452,
      "theoretical_loss": 3.649366902951572,
      "tokens_seen": 999489536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521163490471414,
      "loss": 2.6143,
      "theoretical_loss": 3.649344129530391,
      "tokens_seen": 999555072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521063189568706,
      "loss": 2.8037,
      "theoretical_loss": 3.6493213580203534,
      "tokens_seen": 999620608
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035209628886659984,
      "loss": 2.6248,
      "theoretical_loss": 3.6492985884211735,
      "tokens_seen": 999686144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035208625877632897,
      "loss": 3.0173,
      "theoretical_loss": 3.6492758207325657,
      "tokens_seen": 999751680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520762286860582,
      "loss": 2.7024,
      "theoretical_loss": 3.6492530549542437,
      "tokens_seen": 999817216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520661985957874,
      "loss": 2.943,
      "theoretical_loss": 3.6492302910859227,
      "tokens_seen": 999882752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035205616850551657,
      "loss": 3.0572,
      "theoretical_loss": 3.6492075291273176,
      "tokens_seen": 999948288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035204613841524575,
      "loss": 2.9415,
      "theoretical_loss": 3.6491847690781425,
      "tokens_seen": 1000013824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352036108324975,
      "loss": 2.6723,
      "theoretical_loss": 3.649162010938112,
      "tokens_seen": 1000079360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520260782347041,
      "loss": 2.6927,
      "theoretical_loss": 3.6491392547069412,
      "tokens_seen": 1000144896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035201604814443334,
      "loss": 2.7726,
      "theoretical_loss": 3.649116500384345,
      "tokens_seen": 1000210432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035200601805416247,
      "loss": 2.7641,
      "theoretical_loss": 3.6490937479700376,
      "tokens_seen": 1000275968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519959879638917,
      "loss": 2.7668,
      "theoretical_loss": 3.649070997463735,
      "tokens_seen": 1000341504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519859578736209,
      "loss": 2.6113,
      "theoretical_loss": 3.649048248865152,
      "tokens_seen": 1000407040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035197592778335007,
      "loss": 2.6915,
      "theoretical_loss": 3.649025502174003,
      "tokens_seen": 1000472576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035196589769307925,
      "loss": 2.7293,
      "theoretical_loss": 3.649002757390003,
      "tokens_seen": 1000538112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035195586760280843,
      "loss": 2.4951,
      "theoretical_loss": 3.648980014512868,
      "tokens_seen": 1000603648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519458375125376,
      "loss": 2.6436,
      "theoretical_loss": 3.648957273542313,
      "tokens_seen": 1000669184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035193580742226685,
      "loss": 2.6603,
      "theoretical_loss": 3.6489345344780535,
      "tokens_seen": 1000734720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351925777331996,
      "loss": 2.6807,
      "theoretical_loss": 3.648911797319804,
      "tokens_seen": 1000800256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519157472417252,
      "loss": 2.4684,
      "theoretical_loss": 3.648889062067281,
      "tokens_seen": 1000865792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035190571715145434,
      "loss": 2.5266,
      "theoretical_loss": 3.6488663287202,
      "tokens_seen": 1000931328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035189568706118357,
      "loss": 2.7564,
      "theoretical_loss": 3.6488435972782756,
      "tokens_seen": 1000996864
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 559970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.601503372192383,
      "objective/train/theoretical_loss": 3.6488208677412244,
      "objective/train/tokens_used": 1021522400,
      "theoretical_loss": 3.6488208677412244,
      "tokens_seen": 1001062400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035188565697091275,
      "loss": 2.6304,
      "theoretical_loss": 3.6488208677412244,
      "tokens_seen": 1001062400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035187562688064193,
      "loss": 2.9246,
      "theoretical_loss": 3.6487981401087612,
      "tokens_seen": 1001127936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518655967903711,
      "loss": 2.6429,
      "theoretical_loss": 3.6487754143806024,
      "tokens_seen": 1001193472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035185556670010035,
      "loss": 2.6234,
      "theoretical_loss": 3.6487526905564636,
      "tokens_seen": 1001259008
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518455366098295,
      "loss": 2.7218,
      "theoretical_loss": 3.64872996863606,
      "tokens_seen": 1001324544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518355065195587,
      "loss": 2.9734,
      "theoretical_loss": 3.648707248619109,
      "tokens_seen": 1001390080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035182547642928784,
      "loss": 2.7579,
      "theoretical_loss": 3.6486845305053253,
      "tokens_seen": 1001455616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518154463390171,
      "loss": 2.746,
      "theoretical_loss": 3.6486618142944263,
      "tokens_seen": 1001521152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035180541624874626,
      "loss": 2.7835,
      "theoretical_loss": 3.6486390999861262,
      "tokens_seen": 1001586688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035179538615847544,
      "loss": 2.7529,
      "theoretical_loss": 3.6486163875801427,
      "tokens_seen": 1001652224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517853560682046,
      "loss": 2.6583,
      "theoretical_loss": 3.6485936770761915,
      "tokens_seen": 1001717760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517753259779338,
      "loss": 2.7264,
      "theoretical_loss": 3.6485709684739884,
      "tokens_seen": 1001783296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351765295887663,
      "loss": 2.4917,
      "theoretical_loss": 3.648548261773251,
      "tokens_seen": 1001848832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517552657973922,
      "loss": 2.6531,
      "theoretical_loss": 3.648525556973695,
      "tokens_seen": 1001914368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035174523570712134,
      "loss": 2.809,
      "theoretical_loss": 3.648502854075037,
      "tokens_seen": 1001979904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517352056168506,
      "loss": 2.6379,
      "theoretical_loss": 3.648480153076993,
      "tokens_seen": 1002045440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035172517552657976,
      "loss": 2.6711,
      "theoretical_loss": 3.64845745397928,
      "tokens_seen": 1002110976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035171514543630894,
      "loss": 2.7953,
      "theoretical_loss": 3.6484347567816147,
      "tokens_seen": 1002176512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517051153460381,
      "loss": 2.6388,
      "theoretical_loss": 3.648412061483714,
      "tokens_seen": 1002242048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516950852557673,
      "loss": 2.6233,
      "theoretical_loss": 3.6483893680852946,
      "tokens_seen": 1002307584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516850551654965,
      "loss": 2.7684,
      "theoretical_loss": 3.6483666765860727,
      "tokens_seen": 1002373120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516750250752257,
      "loss": 2.4296,
      "theoretical_loss": 3.648343986985766,
      "tokens_seen": 1002438656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035166499498495485,
      "loss": 2.7591,
      "theoretical_loss": 3.6483212992840914,
      "tokens_seen": 1002504192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516549648946841,
      "loss": 2.5995,
      "theoretical_loss": 3.6482986134807653,
      "tokens_seen": 1002569728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516449348044132,
      "loss": 2.9894,
      "theoretical_loss": 3.6482759295755054,
      "tokens_seen": 1002635264
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 560865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5041370391845703,
      "objective/train/theoretical_loss": 3.6482532475680287,
      "objective/train/tokens_used": 1023160800,
      "theoretical_loss": 3.6482532475680287,
      "tokens_seen": 1002700800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035163490471414244,
      "loss": 2.7673,
      "theoretical_loss": 3.6482532475680287,
      "tokens_seen": 1002700800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516248746238716,
      "loss": 2.7331,
      "theoretical_loss": 3.6482305674580524,
      "tokens_seen": 1002766336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516148445336008,
      "loss": 2.5743,
      "theoretical_loss": 3.6482078892452936,
      "tokens_seen": 1002831872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035160481444333,
      "loss": 2.6882,
      "theoretical_loss": 3.64818521292947,
      "tokens_seen": 1002897408
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035159478435305917,
      "loss": 2.816,
      "theoretical_loss": 3.648162538510298,
      "tokens_seen": 1002962944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035158475426278835,
      "loss": 2.6259,
      "theoretical_loss": 3.6481398659874964,
      "tokens_seen": 1003028480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515747241725176,
      "loss": 2.8063,
      "theoretical_loss": 3.648117195360782,
      "tokens_seen": 1003094016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515646940822467,
      "loss": 2.7044,
      "theoretical_loss": 3.6480945266298725,
      "tokens_seen": 1003159552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035155466399197595,
      "loss": 2.5904,
      "theoretical_loss": 3.6480718597944852,
      "tokens_seen": 1003225088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035154463390170513,
      "loss": 2.7078,
      "theoretical_loss": 3.648049194854338,
      "tokens_seen": 1003290624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515346038114343,
      "loss": 2.7708,
      "theoretical_loss": 3.6480265318091485,
      "tokens_seen": 1003356160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515245737211635,
      "loss": 2.6009,
      "theoretical_loss": 3.6480038706586355,
      "tokens_seen": 1003421696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035151454363089267,
      "loss": 2.866,
      "theoretical_loss": 3.6479812114025156,
      "tokens_seen": 1003487232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035150451354062185,
      "loss": 2.6386,
      "theoretical_loss": 3.6479585540405073,
      "tokens_seen": 1003552768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514944834503511,
      "loss": 2.6617,
      "theoretical_loss": 3.6479358985723285,
      "tokens_seen": 1003618304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514844533600802,
      "loss": 2.7367,
      "theoretical_loss": 3.647913244997697,
      "tokens_seen": 1003683840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035147442326980945,
      "loss": 2.7202,
      "theoretical_loss": 3.6478905933163315,
      "tokens_seen": 1003749376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514643931795386,
      "loss": 2.9854,
      "theoretical_loss": 3.6478679435279497,
      "tokens_seen": 1003814912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514543630892678,
      "loss": 2.692,
      "theoretical_loss": 3.6478452956322704,
      "tokens_seen": 1003880448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351444332998997,
      "loss": 2.6885,
      "theoretical_loss": 3.647822649629011,
      "tokens_seen": 1003945984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514343029087262,
      "loss": 2.9112,
      "theoretical_loss": 3.6478000055178907,
      "tokens_seen": 1004011520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035142427281845536,
      "loss": 2.8827,
      "theoretical_loss": 3.6477773632986272,
      "tokens_seen": 1004077056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035141424272818454,
      "loss": 2.5035,
      "theoretical_loss": 3.6477547229709395,
      "tokens_seen": 1004142592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514042126379137,
      "loss": 2.6845,
      "theoretical_loss": 3.647732084534546,
      "tokens_seen": 1004208128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035139418254764295,
      "loss": 2.8052,
      "theoretical_loss": 3.647709447989165,
      "tokens_seen": 1004273664
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 561903,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.305278778076172,
      "objective/train/theoretical_loss": 3.6476868133345155,
      "objective/train/tokens_used": 1024799200,
      "theoretical_loss": 3.6476868133345155,
      "tokens_seen": 1004339200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513841524573721,
      "loss": 2.7461,
      "theoretical_loss": 3.6476868133345155,
      "tokens_seen": 1004339200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513741223671013,
      "loss": 2.505,
      "theoretical_loss": 3.6476641805703163,
      "tokens_seen": 1004404736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513640922768305,
      "loss": 2.6162,
      "theoretical_loss": 3.647641549696286,
      "tokens_seen": 1004470272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513540621865597,
      "loss": 2.715,
      "theoretical_loss": 3.6476189207121434,
      "tokens_seen": 1004535808
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513440320962889,
      "loss": 2.6644,
      "theoretical_loss": 3.647596293617607,
      "tokens_seen": 1004601344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035133400200601804,
      "loss": 2.3282,
      "theoretical_loss": 3.647573668412397,
      "tokens_seen": 1004666880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513239719157473,
      "loss": 2.8704,
      "theoretical_loss": 3.6475510450962307,
      "tokens_seen": 1004732416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035131394182547646,
      "loss": 2.9088,
      "theoretical_loss": 3.6475284236688283,
      "tokens_seen": 1004797952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035130391173520564,
      "loss": 2.5969,
      "theoretical_loss": 3.647505804129909,
      "tokens_seen": 1004863488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512938816449348,
      "loss": 2.7106,
      "theoretical_loss": 3.647483186479192,
      "tokens_seen": 1004929024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351283851554664,
      "loss": 2.6929,
      "theoretical_loss": 3.6474605707163956,
      "tokens_seen": 1004994560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512738214643932,
      "loss": 2.7415,
      "theoretical_loss": 3.64743795684124,
      "tokens_seen": 1005060096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512637913741224,
      "loss": 2.8185,
      "theoretical_loss": 3.6474153448534445,
      "tokens_seen": 1005125632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035125376128385154,
      "loss": 2.6448,
      "theoretical_loss": 3.6473927347527284,
      "tokens_seen": 1005191168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512437311935808,
      "loss": 2.8856,
      "theoretical_loss": 3.647370126538811,
      "tokens_seen": 1005256704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035123370110330996,
      "loss": 2.484,
      "theoretical_loss": 3.647347520211412,
      "tokens_seen": 1005322240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035122367101303914,
      "loss": 2.9744,
      "theoretical_loss": 3.647324915770251,
      "tokens_seen": 1005387776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512136409227683,
      "loss": 2.7492,
      "theoretical_loss": 3.6473023132150484,
      "tokens_seen": 1005453312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512036108324975,
      "loss": 2.5703,
      "theoretical_loss": 3.6472797125455227,
      "tokens_seen": 1005518848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511935807422267,
      "loss": 2.8978,
      "theoretical_loss": 3.6472571137613943,
      "tokens_seen": 1005584384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511835506519559,
      "loss": 2.9041,
      "theoretical_loss": 3.647234516862383,
      "tokens_seen": 1005649920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035117352056168505,
      "loss": 2.6811,
      "theoretical_loss": 3.647211921848209,
      "tokens_seen": 1005715456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511634904714143,
      "loss": 2.6016,
      "theoretical_loss": 3.647189328718592,
      "tokens_seen": 1005780992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511534603811434,
      "loss": 2.8905,
      "theoretical_loss": 3.6471667374732517,
      "tokens_seen": 1005846528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035114343029087264,
      "loss": 2.6996,
      "theoretical_loss": 3.6471441481119093,
      "tokens_seen": 1005912064
    },
    {
      "epoch": 0.3,
      "objective/train/docs_used": 562546,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.896876335144043,
      "objective/train/theoretical_loss": 3.6471215606342833,
      "objective/train/tokens_used": 1026437600,
      "theoretical_loss": 3.6471215606342833,
      "tokens_seen": 1005977600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511334002006018,
      "loss": 2.8747,
      "theoretical_loss": 3.6471215606342833,
      "tokens_seen": 1005977600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351123370110331,
      "loss": 2.5018,
      "theoretical_loss": 3.6470989750400955,
      "tokens_seen": 1006043136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511133400200602,
      "loss": 2.5671,
      "theoretical_loss": 3.6470763913290654,
      "tokens_seen": 1006108672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035110330992978937,
      "loss": 2.6394,
      "theoretical_loss": 3.647053809500913,
      "tokens_seen": 1006174208
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035109327983951855,
      "loss": 2.7279,
      "theoretical_loss": 3.6470312295553597,
      "tokens_seen": 1006239744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510832497492478,
      "loss": 2.6031,
      "theoretical_loss": 3.6470086514921247,
      "tokens_seen": 1006305280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510732196589769,
      "loss": 2.6531,
      "theoretical_loss": 3.6469860753109296,
      "tokens_seen": 1006370816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035106318956870615,
      "loss": 2.7165,
      "theoretical_loss": 3.6469635010114945,
      "tokens_seen": 1006436352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035105315947843533,
      "loss": 2.9426,
      "theoretical_loss": 3.64694092859354,
      "tokens_seen": 1006501888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003510431293881645,
      "loss": 2.9879,
      "theoretical_loss": 3.646918358056787,
      "tokens_seen": 1006567424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003510330992978937,
      "loss": 2.6251,
      "theoretical_loss": 3.6468957894009564,
      "tokens_seen": 1006632960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035102306920762287,
      "loss": 2.629,
      "theoretical_loss": 3.6468732226257687,
      "tokens_seen": 1006698496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035101303911735205,
      "loss": 2.7709,
      "theoretical_loss": 3.6468506577309445,
      "tokens_seen": 1006764032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003510030090270813,
      "loss": 2.7407,
      "theoretical_loss": 3.6468280947162053,
      "tokens_seen": 1006829568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509929789368104,
      "loss": 2.6753,
      "theoretical_loss": 3.6468055335812717,
      "tokens_seen": 1006895104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035098294884653965,
      "loss": 2.6529,
      "theoretical_loss": 3.6467829743258653,
      "tokens_seen": 1006960640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509729187562688,
      "loss": 2.5762,
      "theoretical_loss": 3.6467604169497063,
      "tokens_seen": 1007026176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350962888665998,
      "loss": 2.9005,
      "theoretical_loss": 3.6467378614525163,
      "tokens_seen": 1007091712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509528585757272,
      "loss": 2.7911,
      "theoretical_loss": 3.6467153078340173,
      "tokens_seen": 1007157248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509428284854564,
      "loss": 2.5908,
      "theoretical_loss": 3.6466927560939295,
      "tokens_seen": 1007222784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035093279839518556,
      "loss": 2.6977,
      "theoretical_loss": 3.6466702062319745,
      "tokens_seen": 1007288320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035092276830491474,
      "loss": 2.7994,
      "theoretical_loss": 3.646647658247874,
      "tokens_seen": 1007353856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509127382146439,
      "loss": 2.5597,
      "theoretical_loss": 3.6466251121413493,
      "tokens_seen": 1007419392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035090270812437315,
      "loss": 2.5898,
      "theoretical_loss": 3.6466025679121223,
      "tokens_seen": 1007484928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508926780341023,
      "loss": 2.6098,
      "theoretical_loss": 3.646580025559914,
      "tokens_seen": 1007550464
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 563033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2577614784240723,
      "objective/train/theoretical_loss": 3.646557485084446,
      "objective/train/tokens_used": 1028076000,
      "theoretical_loss": 3.646557485084446,
      "tokens_seen": 1007616000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508826479438315,
      "loss": 2.7075,
      "theoretical_loss": 3.646557485084446,
      "tokens_seen": 1007616000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508726178535607,
      "loss": 2.5445,
      "theoretical_loss": 3.6465349464854406,
      "tokens_seen": 1007681536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508625877632899,
      "loss": 2.6022,
      "theoretical_loss": 3.646512409762619,
      "tokens_seen": 1007747072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035085255767301906,
      "loss": 2.8948,
      "theoretical_loss": 3.6464898749157033,
      "tokens_seen": 1007812608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035084252758274824,
      "loss": 2.8789,
      "theoretical_loss": 3.646467341944416,
      "tokens_seen": 1007878144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508324974924774,
      "loss": 2.4769,
      "theoretical_loss": 3.6464448108484775,
      "tokens_seen": 1007943680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035082246740220666,
      "loss": 2.6822,
      "theoretical_loss": 3.6464222816276113,
      "tokens_seen": 1008009216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508124373119358,
      "loss": 2.6443,
      "theoretical_loss": 3.6463997542815383,
      "tokens_seen": 1008074752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350802407221665,
      "loss": 2.7836,
      "theoretical_loss": 3.646377228809982,
      "tokens_seen": 1008140288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035079237713139415,
      "loss": 2.4362,
      "theoretical_loss": 3.646354705212663,
      "tokens_seen": 1008205824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507823470411234,
      "loss": 2.6115,
      "theoretical_loss": 3.6463321834893048,
      "tokens_seen": 1008271360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035077231695085256,
      "loss": 2.6302,
      "theoretical_loss": 3.646309663639629,
      "tokens_seen": 1008336896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035076228686058174,
      "loss": 2.5671,
      "theoretical_loss": 3.646287145663358,
      "tokens_seen": 1008402432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507522567703109,
      "loss": 2.6446,
      "theoretical_loss": 3.6462646295602146,
      "tokens_seen": 1008467968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035074222668004016,
      "loss": 2.787,
      "theoretical_loss": 3.6462421153299207,
      "tokens_seen": 1008533504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507321965897693,
      "loss": 2.3819,
      "theoretical_loss": 3.6462196029721996,
      "tokens_seen": 1008599040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507221664994985,
      "loss": 2.6316,
      "theoretical_loss": 3.6461970924867737,
      "tokens_seen": 1008664576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035071213640922765,
      "loss": 2.8287,
      "theoretical_loss": 3.646174583873365,
      "tokens_seen": 1008730112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507021063189569,
      "loss": 2.8155,
      "theoretical_loss": 3.6461520771316964,
      "tokens_seen": 1008795648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035069207622868607,
      "loss": 2.7533,
      "theoretical_loss": 3.646129572261491,
      "tokens_seen": 1008861184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035068204613841525,
      "loss": 2.8465,
      "theoretical_loss": 3.646107069262472,
      "tokens_seen": 1008926720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035067201604814443,
      "loss": 2.5893,
      "theoretical_loss": 3.6460845681343614,
      "tokens_seen": 1008992256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506619859578736,
      "loss": 2.809,
      "theoretical_loss": 3.646062068876883,
      "tokens_seen": 1009057792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506519558676028,
      "loss": 2.8392,
      "theoretical_loss": 3.646039571489759,
      "tokens_seen": 1009123328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350641925777332,
      "loss": 2.8639,
      "theoretical_loss": 3.646017075972713,
      "tokens_seen": 1009188864
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 564080,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.776188373565674,
      "objective/train/theoretical_loss": 3.645994582325468,
      "objective/train/tokens_used": 1029714400,
      "theoretical_loss": 3.645994582325468,
      "tokens_seen": 1009254400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035063189568706115,
      "loss": 2.9152,
      "theoretical_loss": 3.645994582325468,
      "tokens_seen": 1009254400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506218655967904,
      "loss": 2.8531,
      "theoretical_loss": 3.6459720905477466,
      "tokens_seen": 1009319936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506118355065195,
      "loss": 2.7551,
      "theoretical_loss": 3.6459496006392733,
      "tokens_seen": 1009385472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035060180541624875,
      "loss": 2.5909,
      "theoretical_loss": 3.6459271125997708,
      "tokens_seen": 1009451008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350591775325978,
      "loss": 2.5893,
      "theoretical_loss": 3.645904626428962,
      "tokens_seen": 1009516544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505817452357071,
      "loss": 2.863,
      "theoretical_loss": 3.6458821421265704,
      "tokens_seen": 1009582080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035057171514543635,
      "loss": 2.9149,
      "theoretical_loss": 3.64585965969232,
      "tokens_seen": 1009647616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035056168505516553,
      "loss": 2.4815,
      "theoretical_loss": 3.6458371791259343,
      "tokens_seen": 1009713152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505516549648947,
      "loss": 2.8522,
      "theoretical_loss": 3.6458147004271364,
      "tokens_seen": 1009778688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505416248746239,
      "loss": 2.4721,
      "theoretical_loss": 3.6457922235956506,
      "tokens_seen": 1009844224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035053159478435307,
      "loss": 2.6782,
      "theoretical_loss": 3.6457697486312,
      "tokens_seen": 1009909760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035052156469408225,
      "loss": 2.978,
      "theoretical_loss": 3.645747275533509,
      "tokens_seen": 1009975296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505115346038115,
      "loss": 2.7994,
      "theoretical_loss": 3.6457248043023007,
      "tokens_seen": 1010040832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505015045135406,
      "loss": 2.6268,
      "theoretical_loss": 3.6457023349372992,
      "tokens_seen": 1010106368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035049147442326985,
      "loss": 2.7156,
      "theoretical_loss": 3.645679867438229,
      "tokens_seen": 1010171904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350481444332999,
      "loss": 2.7835,
      "theoretical_loss": 3.6456574018048133,
      "tokens_seen": 1010237440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504714142427282,
      "loss": 2.8286,
      "theoretical_loss": 3.6456349380367774,
      "tokens_seen": 1010302976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504613841524574,
      "loss": 2.8143,
      "theoretical_loss": 3.6456124761338433,
      "tokens_seen": 1010368512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504513540621866,
      "loss": 2.613,
      "theoretical_loss": 3.6455900160957375,
      "tokens_seen": 1010434048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035044132397191576,
      "loss": 2.484,
      "theoretical_loss": 3.645567557922183,
      "tokens_seen": 1010499584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035043129388164494,
      "loss": 2.7924,
      "theoretical_loss": 3.645545101612904,
      "tokens_seen": 1010565120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504212637913741,
      "loss": 2.7572,
      "theoretical_loss": 3.645522647167626,
      "tokens_seen": 1010630656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035041123370110335,
      "loss": 2.6412,
      "theoretical_loss": 3.645500194586072,
      "tokens_seen": 1010696192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504012036108325,
      "loss": 2.8669,
      "theoretical_loss": 3.645477743867967,
      "tokens_seen": 1010761728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503911735205617,
      "loss": 2.8419,
      "theoretical_loss": 3.645455295013036,
      "tokens_seen": 1010827264
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 564461,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.960507869720459,
      "objective/train/theoretical_loss": 3.6454328480210028,
      "objective/train/tokens_used": 1031352800,
      "theoretical_loss": 3.6454328480210028,
      "tokens_seen": 1010892800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503811434302909,
      "loss": 2.9234,
      "theoretical_loss": 3.6454328480210028,
      "tokens_seen": 1010892800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503711133400201,
      "loss": 2.9007,
      "theoretical_loss": 3.6454104028915926,
      "tokens_seen": 1010958336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035036108324974926,
      "loss": 2.8377,
      "theoretical_loss": 3.64538795962453,
      "tokens_seen": 1011023872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035035105315947844,
      "loss": 2.7468,
      "theoretical_loss": 3.64536551821954,
      "tokens_seen": 1011089408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503410230692076,
      "loss": 2.8693,
      "theoretical_loss": 3.6453430786763468,
      "tokens_seen": 1011154944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035033099297893686,
      "loss": 2.4695,
      "theoretical_loss": 3.645320640994676,
      "tokens_seen": 1011220480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350320962888666,
      "loss": 2.6836,
      "theoretical_loss": 3.645298205174252,
      "tokens_seen": 1011286016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503109327983952,
      "loss": 2.7281,
      "theoretical_loss": 3.6452757712148,
      "tokens_seen": 1011351552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035030090270812435,
      "loss": 2.7976,
      "theoretical_loss": 3.645253339116045,
      "tokens_seen": 1011417088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502908726178536,
      "loss": 2.6928,
      "theoretical_loss": 3.6452309088777124,
      "tokens_seen": 1011482624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035028084252758276,
      "loss": 2.4886,
      "theoretical_loss": 3.645208480499527,
      "tokens_seen": 1011548160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035027081243731194,
      "loss": 2.6801,
      "theoretical_loss": 3.6451860539812144,
      "tokens_seen": 1011613696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502607823470411,
      "loss": 2.7794,
      "theoretical_loss": 3.6451636293225,
      "tokens_seen": 1011679232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035025075225677036,
      "loss": 2.5784,
      "theoretical_loss": 3.6451412065231086,
      "tokens_seen": 1011744768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502407221664995,
      "loss": 2.7135,
      "theoretical_loss": 3.645118785582765,
      "tokens_seen": 1011810304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502306920762287,
      "loss": 2.6828,
      "theoretical_loss": 3.6450963665011966,
      "tokens_seen": 1011875840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035022066198595785,
      "loss": 2.8697,
      "theoretical_loss": 3.6450739492781277,
      "tokens_seen": 1011941376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502106318956871,
      "loss": 2.6464,
      "theoretical_loss": 3.645051533913284,
      "tokens_seen": 1012006912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035020060180541627,
      "loss": 2.7643,
      "theoretical_loss": 3.6450291204063907,
      "tokens_seen": 1012072448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035019057171514545,
      "loss": 2.6983,
      "theoretical_loss": 3.645006708757175,
      "tokens_seen": 1012137984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035018054162487463,
      "loss": 2.6446,
      "theoretical_loss": 3.644984298965361,
      "tokens_seen": 1012203520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501705115346038,
      "loss": 2.5605,
      "theoretical_loss": 3.6449618910306754,
      "tokens_seen": 1012269056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350160481444333,
      "loss": 2.5893,
      "theoretical_loss": 3.6449394849528436,
      "tokens_seen": 1012334592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501504513540622,
      "loss": 2.6389,
      "theoretical_loss": 3.644917080731592,
      "tokens_seen": 1012400128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035014042126379135,
      "loss": 2.7151,
      "theoretical_loss": 3.6448946783666463,
      "tokens_seen": 1012465664
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 565625,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7371182441711426,
      "objective/train/theoretical_loss": 3.6448722778577327,
      "objective/train/tokens_used": 1032991200,
      "theoretical_loss": 3.6448722778577327,
      "tokens_seen": 1012531200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501303911735206,
      "loss": 2.836,
      "theoretical_loss": 3.6448722778577327,
      "tokens_seen": 1012531200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501203610832497,
      "loss": 2.6583,
      "theoretical_loss": 3.6448498792045774,
      "tokens_seen": 1012596736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035011033099297895,
      "loss": 2.5067,
      "theoretical_loss": 3.6448274824069067,
      "tokens_seen": 1012662272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035010030090270813,
      "loss": 2.6818,
      "theoretical_loss": 3.644805087464446,
      "tokens_seen": 1012727808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500902708124373,
      "loss": 2.3187,
      "theoretical_loss": 3.644782694376923,
      "tokens_seen": 1012793344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500802407221665,
      "loss": 2.6766,
      "theoretical_loss": 3.6447603031440625,
      "tokens_seen": 1012858880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035007021063189573,
      "loss": 2.4468,
      "theoretical_loss": 3.6447379137655913,
      "tokens_seen": 1012924416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035006018054162486,
      "loss": 2.6092,
      "theoretical_loss": 3.644715526241237,
      "tokens_seen": 1012989952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500501504513541,
      "loss": 2.5698,
      "theoretical_loss": 3.6446931405707246,
      "tokens_seen": 1013055488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500401203610832,
      "loss": 2.7784,
      "theoretical_loss": 3.644670756753782,
      "tokens_seen": 1013121024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035003009027081245,
      "loss": 2.5981,
      "theoretical_loss": 3.6446483747901346,
      "tokens_seen": 1013186560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035002006018054163,
      "loss": 2.6603,
      "theoretical_loss": 3.64462599467951,
      "tokens_seen": 1013252096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500100300902708,
      "loss": 2.9651,
      "theoretical_loss": 3.644603616421634,
      "tokens_seen": 1013317632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035,
      "loss": 2.5475,
      "theoretical_loss": 3.644581240016235,
      "tokens_seen": 1013383168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499899699097292,
      "loss": 2.9199,
      "theoretical_loss": 3.644558865463038,
      "tokens_seen": 1013448704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034997993981945836,
      "loss": 2.6318,
      "theoretical_loss": 3.6445364927617714,
      "tokens_seen": 1013514240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499699097291876,
      "loss": 2.6628,
      "theoretical_loss": 3.6445141219121617,
      "tokens_seen": 1013579776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499598796389167,
      "loss": 2.595,
      "theoretical_loss": 3.6444917529139356,
      "tokens_seen": 1013645312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034994984954864596,
      "loss": 2.8204,
      "theoretical_loss": 3.6444693857668207,
      "tokens_seen": 1013710848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499398194583751,
      "loss": 2.8346,
      "theoretical_loss": 3.6444470204705435,
      "tokens_seen": 1013776384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499297893681043,
      "loss": 2.7842,
      "theoretical_loss": 3.6444246570248318,
      "tokens_seen": 1013841920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499197592778335,
      "loss": 2.6016,
      "theoretical_loss": 3.644402295429413,
      "tokens_seen": 1013907456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499097291875627,
      "loss": 2.818,
      "theoretical_loss": 3.6443799356840136,
      "tokens_seen": 1013972992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034989969909729186,
      "loss": 2.5587,
      "theoretical_loss": 3.6443575777883614,
      "tokens_seen": 1014038528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498896690070211,
      "loss": 2.6353,
      "theoretical_loss": 3.6443352217421845,
      "tokens_seen": 1014104064
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 566086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7045507431030273,
      "objective/train/theoretical_loss": 3.6443128675452092,
      "objective/train/tokens_used": 1034629600,
      "theoretical_loss": 3.6443128675452092,
      "tokens_seen": 1014169600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498796389167502,
      "loss": 2.6942,
      "theoretical_loss": 3.6443128675452092,
      "tokens_seen": 1014169600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034986960882647946,
      "loss": 2.8548,
      "theoretical_loss": 3.6442905151971647,
      "tokens_seen": 1014235136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498595787362086,
      "loss": 2.7197,
      "theoretical_loss": 3.6442681646977766,
      "tokens_seen": 1014300672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498495486459378,
      "loss": 2.6459,
      "theoretical_loss": 3.6442458160467743,
      "tokens_seen": 1014366208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034983951855566706,
      "loss": 2.5951,
      "theoretical_loss": 3.644223469243885,
      "tokens_seen": 1014431744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498294884653962,
      "loss": 2.7552,
      "theoretical_loss": 3.644201124288836,
      "tokens_seen": 1014497280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498194583751254,
      "loss": 2.7001,
      "theoretical_loss": 3.6441787811813553,
      "tokens_seen": 1014562816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034980942828485455,
      "loss": 2.7129,
      "theoretical_loss": 3.6441564399211708,
      "tokens_seen": 1014628352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497993981945838,
      "loss": 2.6051,
      "theoretical_loss": 3.644134100508011,
      "tokens_seen": 1014693888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034978936810431296,
      "loss": 2.6101,
      "theoretical_loss": 3.6441117629416038,
      "tokens_seen": 1014759424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034977933801404214,
      "loss": 2.6598,
      "theoretical_loss": 3.644089427221677,
      "tokens_seen": 1014824960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497693079237713,
      "loss": 2.4322,
      "theoretical_loss": 3.644067093347958,
      "tokens_seen": 1014890496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034975927783350056,
      "loss": 2.7851,
      "theoretical_loss": 3.6440447613201767,
      "tokens_seen": 1014956032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497492477432297,
      "loss": 2.7539,
      "theoretical_loss": 3.64402243113806,
      "tokens_seen": 1015021568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497392176529589,
      "loss": 2.5749,
      "theoretical_loss": 3.6440001028013365,
      "tokens_seen": 1015087104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034972918756268805,
      "loss": 2.8158,
      "theoretical_loss": 3.643977776309735,
      "tokens_seen": 1015152640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497191574724173,
      "loss": 2.753,
      "theoretical_loss": 3.6439554516629835,
      "tokens_seen": 1015218176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034970912738214647,
      "loss": 2.7789,
      "theoretical_loss": 3.6439331288608106,
      "tokens_seen": 1015283712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034969909729187565,
      "loss": 2.6908,
      "theoretical_loss": 3.6439108079029445,
      "tokens_seen": 1015349248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034968906720160483,
      "loss": 2.742,
      "theoretical_loss": 3.643888488789115,
      "tokens_seen": 1015414784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349679037111334,
      "loss": 2.7681,
      "theoretical_loss": 3.6438661715190492,
      "tokens_seen": 1015480320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496690070210632,
      "loss": 2.8472,
      "theoretical_loss": 3.643843856092477,
      "tokens_seen": 1015545856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496589769307924,
      "loss": 2.5085,
      "theoretical_loss": 3.6438215425091265,
      "tokens_seen": 1015611392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034964894684052155,
      "loss": 2.9066,
      "theoretical_loss": 3.6437992307687264,
      "tokens_seen": 1015676928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496389167502508,
      "loss": 2.673,
      "theoretical_loss": 3.6437769208710065,
      "tokens_seen": 1015742464
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 567051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.23244047164917,
      "objective/train/theoretical_loss": 3.6437546128156946,
      "objective/train/tokens_used": 1036268000,
      "theoretical_loss": 3.6437546128156946,
      "tokens_seen": 1015808000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496288866599799,
      "loss": 2.4863,
      "theoretical_loss": 3.6437546128156946,
      "tokens_seen": 1015808000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034961885656970915,
      "loss": 2.7115,
      "theoretical_loss": 3.6437323066025207,
      "tokens_seen": 1015873536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034960882647943833,
      "loss": 2.6778,
      "theoretical_loss": 3.643710002231213,
      "tokens_seen": 1015939072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495987963891675,
      "loss": 2.7031,
      "theoretical_loss": 3.643687699701501,
      "tokens_seen": 1016004608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495887662988967,
      "loss": 2.8153,
      "theoretical_loss": 3.6436653990131145,
      "tokens_seen": 1016070144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034957873620862593,
      "loss": 2.6424,
      "theoretical_loss": 3.6436431001657814,
      "tokens_seen": 1016135680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034956870611835506,
      "loss": 2.6089,
      "theoretical_loss": 3.643620803159232,
      "tokens_seen": 1016201216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495586760280843,
      "loss": 2.6988,
      "theoretical_loss": 3.643598507993195,
      "tokens_seen": 1016266752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495486459378134,
      "loss": 2.7106,
      "theoretical_loss": 3.6435762146674007,
      "tokens_seen": 1016332288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034953861584754265,
      "loss": 2.4762,
      "theoretical_loss": 3.6435539231815777,
      "tokens_seen": 1016397824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034952858575727183,
      "loss": 2.8292,
      "theoretical_loss": 3.6435316335354555,
      "tokens_seen": 1016463360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349518555667001,
      "loss": 2.6217,
      "theoretical_loss": 3.6435093457287646,
      "tokens_seen": 1016528896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495085255767302,
      "loss": 2.7067,
      "theoretical_loss": 3.6434870597612337,
      "tokens_seen": 1016594432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494984954864594,
      "loss": 2.7773,
      "theoretical_loss": 3.6434647756325926,
      "tokens_seen": 1016659968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034948846539618856,
      "loss": 2.6261,
      "theoretical_loss": 3.6434424933425715,
      "tokens_seen": 1016725504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494784353059178,
      "loss": 2.7,
      "theoretical_loss": 3.6434202128909003,
      "tokens_seen": 1016791040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494684052156469,
      "loss": 2.7586,
      "theoretical_loss": 3.6433979342773077,
      "tokens_seen": 1016856576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034945837512537616,
      "loss": 2.8934,
      "theoretical_loss": 3.643375657501525,
      "tokens_seen": 1016922112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494483450351053,
      "loss": 2.6247,
      "theoretical_loss": 3.6433533825632813,
      "tokens_seen": 1016987648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494383149448345,
      "loss": 2.8865,
      "theoretical_loss": 3.6433311094623066,
      "tokens_seen": 1017053184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494282848545637,
      "loss": 2.6264,
      "theoretical_loss": 3.643308838198332,
      "tokens_seen": 1017118720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494182547642929,
      "loss": 2.6972,
      "theoretical_loss": 3.6432865687710865,
      "tokens_seen": 1017184256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034940822467402206,
      "loss": 2.7361,
      "theoretical_loss": 3.643264301180301,
      "tokens_seen": 1017249792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493981945837513,
      "loss": 2.6566,
      "theoretical_loss": 3.643242035425705,
      "tokens_seen": 1017315328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493881644934804,
      "loss": 2.6861,
      "theoretical_loss": 3.64321977150703,
      "tokens_seen": 1017380864
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 568180,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6473095417022705,
      "objective/train/theoretical_loss": 3.643197509424005,
      "objective/train/tokens_used": 1037906400,
      "theoretical_loss": 3.643197509424005,
      "tokens_seen": 1017446400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034937813440320966,
      "loss": 2.7746,
      "theoretical_loss": 3.643197509424005,
      "tokens_seen": 1017446400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493681043129388,
      "loss": 2.8067,
      "theoretical_loss": 3.6431752491763616,
      "tokens_seen": 1017511936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349358074222668,
      "loss": 2.6145,
      "theoretical_loss": 3.64315299076383,
      "tokens_seen": 1017577472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493480441323972,
      "loss": 2.9663,
      "theoretical_loss": 3.6431307341861396,
      "tokens_seen": 1017643008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493380140421264,
      "loss": 2.7876,
      "theoretical_loss": 3.6431084794430224,
      "tokens_seen": 1017708544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034932798395185557,
      "loss": 2.7973,
      "theoretical_loss": 3.6430862265342085,
      "tokens_seen": 1017774080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034931795386158475,
      "loss": 2.5958,
      "theoretical_loss": 3.643063975459429,
      "tokens_seen": 1017839616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034930792377131393,
      "loss": 2.8849,
      "theoretical_loss": 3.643041726218414,
      "tokens_seen": 1017905152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034929789368104316,
      "loss": 2.7891,
      "theoretical_loss": 3.643019478810895,
      "tokens_seen": 1017970688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492878635907723,
      "loss": 2.7187,
      "theoretical_loss": 3.6429972332366027,
      "tokens_seen": 1018036224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492778335005015,
      "loss": 2.7684,
      "theoretical_loss": 3.6429749894952677,
      "tokens_seen": 1018101760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034926780341023065,
      "loss": 2.3443,
      "theoretical_loss": 3.6429527475866212,
      "tokens_seen": 1018167296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492577733199599,
      "loss": 2.8958,
      "theoretical_loss": 3.6429305075103944,
      "tokens_seen": 1018232832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034924774322968907,
      "loss": 2.7228,
      "theoretical_loss": 3.642908269266318,
      "tokens_seen": 1018298368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034923771313941825,
      "loss": 2.9386,
      "theoretical_loss": 3.6428860328541233,
      "tokens_seen": 1018363904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034922768304914743,
      "loss": 2.6324,
      "theoretical_loss": 3.6428637982735417,
      "tokens_seen": 1018429440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034921765295887667,
      "loss": 2.6251,
      "theoretical_loss": 3.642841565524305,
      "tokens_seen": 1018494976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492076228686058,
      "loss": 2.6117,
      "theoretical_loss": 3.6428193346061435,
      "tokens_seen": 1018560512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034919759277833503,
      "loss": 2.6193,
      "theoretical_loss": 3.642797105518789,
      "tokens_seen": 1018626048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034918756268806416,
      "loss": 2.611,
      "theoretical_loss": 3.6427748782619735,
      "tokens_seen": 1018691584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491775325977934,
      "loss": 2.5485,
      "theoretical_loss": 3.6427526528354277,
      "tokens_seen": 1018757120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034916750250752257,
      "loss": 2.8654,
      "theoretical_loss": 3.642730429238884,
      "tokens_seen": 1018822656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034915747241725175,
      "loss": 2.8412,
      "theoretical_loss": 3.6427082074720727,
      "tokens_seen": 1018888192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034914744232698093,
      "loss": 2.5741,
      "theoretical_loss": 3.6426859875347266,
      "tokens_seen": 1018953728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491374122367101,
      "loss": 2.6678,
      "theoretical_loss": 3.6426637694265773,
      "tokens_seen": 1019019264
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 568941,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.903310775756836,
      "objective/train/theoretical_loss": 3.6426415531473566,
      "objective/train/tokens_used": 1039544800,
      "theoretical_loss": 3.6426415531473566,
      "tokens_seen": 1019084800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491273821464393,
      "loss": 2.701,
      "theoretical_loss": 3.6426415531473566,
      "tokens_seen": 1019084800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034911735205616853,
      "loss": 2.8147,
      "theoretical_loss": 3.6426193386967958,
      "tokens_seen": 1019150336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034910732196589766,
      "loss": 2.5661,
      "theoretical_loss": 3.642597126074627,
      "tokens_seen": 1019215872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490972918756269,
      "loss": 2.8054,
      "theoretical_loss": 3.6425749152805826,
      "tokens_seen": 1019281408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034908726178535613,
      "loss": 2.7761,
      "theoretical_loss": 3.6425527063143948,
      "tokens_seen": 1019346944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034907723169508526,
      "loss": 2.6967,
      "theoretical_loss": 3.6425304991757943,
      "tokens_seen": 1019412480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490672016048145,
      "loss": 2.9256,
      "theoretical_loss": 3.6425082938645144,
      "tokens_seen": 1019478016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490571715145436,
      "loss": 2.9273,
      "theoretical_loss": 3.642486090380287,
      "tokens_seen": 1019543552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034904714142427285,
      "loss": 2.6534,
      "theoretical_loss": 3.642463888722845,
      "tokens_seen": 1019609088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034903711133400203,
      "loss": 2.8481,
      "theoretical_loss": 3.6424416888919193,
      "tokens_seen": 1019674624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490270812437312,
      "loss": 2.6814,
      "theoretical_loss": 3.6424194908872436,
      "tokens_seen": 1019740160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490170511534604,
      "loss": 2.7889,
      "theoretical_loss": 3.64239729470855,
      "tokens_seen": 1019805696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490070210631896,
      "loss": 2.8075,
      "theoretical_loss": 3.6423751003555696,
      "tokens_seen": 1019871232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034899699097291876,
      "loss": 2.7066,
      "theoretical_loss": 3.6423529078280366,
      "tokens_seen": 1019936768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348986960882648,
      "loss": 2.6282,
      "theoretical_loss": 3.642330717125683,
      "tokens_seen": 1020002304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489769307923771,
      "loss": 2.9868,
      "theoretical_loss": 3.6423085282482415,
      "tokens_seen": 1020067840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034896690070210636,
      "loss": 2.8634,
      "theoretical_loss": 3.6422863411954447,
      "tokens_seen": 1020133376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489568706118355,
      "loss": 2.7896,
      "theoretical_loss": 3.642264155967026,
      "tokens_seen": 1020198912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489468405215647,
      "loss": 2.7245,
      "theoretical_loss": 3.642241972562717,
      "tokens_seen": 1020264448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489368104312939,
      "loss": 2.7539,
      "theoretical_loss": 3.642219790982251,
      "tokens_seen": 1020329984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489267803410231,
      "loss": 2.3399,
      "theoretical_loss": 3.6421976112253613,
      "tokens_seen": 1020395520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034891675025075226,
      "loss": 2.5317,
      "theoretical_loss": 3.64217543329178,
      "tokens_seen": 1020461056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489067201604815,
      "loss": 2.8228,
      "theoretical_loss": 3.6421532571812416,
      "tokens_seen": 1020526592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488966900702106,
      "loss": 2.6505,
      "theoretical_loss": 3.6421310828934783,
      "tokens_seen": 1020592128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034888665997993986,
      "loss": 2.7931,
      "theoretical_loss": 3.642108910428223,
      "tokens_seen": 1020657664
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 569860,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.728569507598877,
      "objective/train/theoretical_loss": 3.6420867397852095,
      "objective/train/tokens_used": 1041183200,
      "theoretical_loss": 3.6420867397852095,
      "tokens_seen": 1020723200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348876629889669,
      "loss": 2.6178,
      "theoretical_loss": 3.6420867397852095,
      "tokens_seen": 1020723200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488665997993982,
      "loss": 2.7344,
      "theoretical_loss": 3.6420645709641706,
      "tokens_seen": 1020788736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488565697091274,
      "loss": 2.82,
      "theoretical_loss": 3.64204240396484,
      "tokens_seen": 1020854272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488465396188566,
      "loss": 2.7387,
      "theoretical_loss": 3.6420202387869507,
      "tokens_seen": 1020919808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034883650952858577,
      "loss": 2.7402,
      "theoretical_loss": 3.6419980754302363,
      "tokens_seen": 1020985344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034882647943831495,
      "loss": 2.8513,
      "theoretical_loss": 3.64197591389443,
      "tokens_seen": 1021050880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034881644934804413,
      "loss": 2.9306,
      "theoretical_loss": 3.6419537541792657,
      "tokens_seen": 1021116416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034880641925777336,
      "loss": 2.7919,
      "theoretical_loss": 3.6419315962844774,
      "tokens_seen": 1021181952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487963891675025,
      "loss": 2.6795,
      "theoretical_loss": 3.6419094402097976,
      "tokens_seen": 1021247488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487863590772317,
      "loss": 2.6768,
      "theoretical_loss": 3.6418872859549607,
      "tokens_seen": 1021313024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034877632898696085,
      "loss": 2.8067,
      "theoretical_loss": 3.6418651335197003,
      "tokens_seen": 1021378560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487662988966901,
      "loss": 2.7436,
      "theoretical_loss": 3.641842982903751,
      "tokens_seen": 1021444096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034875626880641927,
      "loss": 2.8722,
      "theoretical_loss": 3.641820834106845,
      "tokens_seen": 1021509632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034874623871614845,
      "loss": 2.8338,
      "theoretical_loss": 3.641798687128718,
      "tokens_seen": 1021575168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034873620862587763,
      "loss": 2.9625,
      "theoretical_loss": 3.641776541969103,
      "tokens_seen": 1021640704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034872617853560687,
      "loss": 2.5987,
      "theoretical_loss": 3.641754398627734,
      "tokens_seen": 1021706240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348716148445336,
      "loss": 2.882,
      "theoretical_loss": 3.6417322571043456,
      "tokens_seen": 1021771776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034870611835506523,
      "loss": 2.762,
      "theoretical_loss": 3.641710117398672,
      "tokens_seen": 1021837312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034869608826479436,
      "loss": 2.5842,
      "theoretical_loss": 3.641687979510446,
      "tokens_seen": 1021902848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486860581745236,
      "loss": 2.7492,
      "theoretical_loss": 3.6416658434394042,
      "tokens_seen": 1021968384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034867602808425277,
      "loss": 2.8195,
      "theoretical_loss": 3.6416437091852787,
      "tokens_seen": 1022033920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034866599799398195,
      "loss": 2.9394,
      "theoretical_loss": 3.6416215767478057,
      "tokens_seen": 1022099456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034865596790371113,
      "loss": 2.5965,
      "theoretical_loss": 3.641599446126718,
      "tokens_seen": 1022164992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486459378134403,
      "loss": 2.7538,
      "theoretical_loss": 3.6415773173217514,
      "tokens_seen": 1022230528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486359077231695,
      "loss": 2.5377,
      "theoretical_loss": 3.6415551903326397,
      "tokens_seen": 1022296064
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 570478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2970290184020996,
      "objective/train/theoretical_loss": 3.641533065159118,
      "objective/train/tokens_used": 1042821600,
      "theoretical_loss": 3.641533065159118,
      "tokens_seen": 1022361600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034862587763289873,
      "loss": 2.6148,
      "theoretical_loss": 3.641533065159118,
      "tokens_seen": 1022361600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034861584754262786,
      "loss": 2.8847,
      "theoretical_loss": 3.6415109418009206,
      "tokens_seen": 1022427136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486058174523571,
      "loss": 2.7738,
      "theoretical_loss": 3.6414888202577824,
      "tokens_seen": 1022492672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485957873620862,
      "loss": 2.9431,
      "theoretical_loss": 3.6414667005294374,
      "tokens_seen": 1022558208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034858575727181546,
      "loss": 2.7978,
      "theoretical_loss": 3.641444582615622,
      "tokens_seen": 1022623744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034857572718154464,
      "loss": 2.7741,
      "theoretical_loss": 3.6414224665160697,
      "tokens_seen": 1022689280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485656970912738,
      "loss": 2.6496,
      "theoretical_loss": 3.641400352230516,
      "tokens_seen": 1022754816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348555667001003,
      "loss": 2.6782,
      "theoretical_loss": 3.6413782397586956,
      "tokens_seen": 1022820352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034854563691073223,
      "loss": 2.6638,
      "theoretical_loss": 3.641356129100344,
      "tokens_seen": 1022885888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034853560682046136,
      "loss": 2.7403,
      "theoretical_loss": 3.6413340202551954,
      "tokens_seen": 1022951424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485255767301906,
      "loss": 2.8024,
      "theoretical_loss": 3.641311913222986,
      "tokens_seen": 1023016960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485155466399197,
      "loss": 2.6215,
      "theoretical_loss": 3.6412898080034504,
      "tokens_seen": 1023082496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034850551654964896,
      "loss": 2.8317,
      "theoretical_loss": 3.6412677045963244,
      "tokens_seen": 1023148032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034849548645937814,
      "loss": 2.7537,
      "theoretical_loss": 3.6412456030013427,
      "tokens_seen": 1023213568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484854563691073,
      "loss": 2.6645,
      "theoretical_loss": 3.6412235032182414,
      "tokens_seen": 1023279104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484754262788365,
      "loss": 2.4196,
      "theoretical_loss": 3.6412014052467554,
      "tokens_seen": 1023344640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484653961885657,
      "loss": 2.4108,
      "theoretical_loss": 3.64117930908662,
      "tokens_seen": 1023410176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034845536609829486,
      "loss": 2.8733,
      "theoretical_loss": 3.6411572147375715,
      "tokens_seen": 1023475712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484453360080241,
      "loss": 2.6105,
      "theoretical_loss": 3.6411351221993447,
      "tokens_seen": 1023541248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484353059177532,
      "loss": 2.6572,
      "theoretical_loss": 3.6411130314716758,
      "tokens_seen": 1023606784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034842527582748246,
      "loss": 2.5466,
      "theoretical_loss": 3.6410909425542997,
      "tokens_seen": 1023672320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034841524573721164,
      "loss": 2.493,
      "theoretical_loss": 3.641068855446953,
      "tokens_seen": 1023737856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484052156469408,
      "loss": 2.6693,
      "theoretical_loss": 3.641046770149372,
      "tokens_seen": 1023803392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034839518555667,
      "loss": 2.8194,
      "theoretical_loss": 3.641024686661291,
      "tokens_seen": 1023868928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483851554663992,
      "loss": 2.6758,
      "theoretical_loss": 3.6410026049824475,
      "tokens_seen": 1023934464
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 571810,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.663266897201538,
      "objective/train/theoretical_loss": 3.640980525112577,
      "objective/train/tokens_used": 1044460000,
      "theoretical_loss": 3.640980525112577,
      "tokens_seen": 1024000000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034837512537612837,
      "loss": 2.5753,
      "theoretical_loss": 3.640980525112577,
      "tokens_seen": 1024000000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483650952858576,
      "loss": 2.7885,
      "theoretical_loss": 3.640958447051415,
      "tokens_seen": 1024065536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034835506519558673,
      "loss": 2.5495,
      "theoretical_loss": 3.6409363707986975,
      "tokens_seen": 1024131072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034834503510531597,
      "loss": 2.8806,
      "theoretical_loss": 3.6409142963541616,
      "tokens_seen": 1024196608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034833500501504515,
      "loss": 2.8696,
      "theoretical_loss": 3.6408922237175427,
      "tokens_seen": 1024262144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034832497492477433,
      "loss": 2.6233,
      "theoretical_loss": 3.640870152888578,
      "tokens_seen": 1024327680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034831494483450356,
      "loss": 2.7639,
      "theoretical_loss": 3.6408480838670023,
      "tokens_seen": 1024393216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483049147442327,
      "loss": 2.8172,
      "theoretical_loss": 3.6408260166525537,
      "tokens_seen": 1024458752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482948846539619,
      "loss": 2.7901,
      "theoretical_loss": 3.6408039512449677,
      "tokens_seen": 1024524288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034828485456369105,
      "loss": 2.7927,
      "theoretical_loss": 3.6407818876439806,
      "tokens_seen": 1024589824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482748244734203,
      "loss": 2.8879,
      "theoretical_loss": 3.6407598258493294,
      "tokens_seen": 1024655360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034826479438314947,
      "loss": 2.7932,
      "theoretical_loss": 3.6407377658607505,
      "tokens_seen": 1024720896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034825476429287865,
      "loss": 2.6392,
      "theoretical_loss": 3.640715707677981,
      "tokens_seen": 1024786432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034824473420260783,
      "loss": 2.8363,
      "theoretical_loss": 3.6406936513007566,
      "tokens_seen": 1024851968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034823470411233707,
      "loss": 2.6225,
      "theoretical_loss": 3.640671596728815,
      "tokens_seen": 1024917504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482246740220662,
      "loss": 2.8723,
      "theoretical_loss": 3.6406495439618927,
      "tokens_seen": 1024983040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034821464393179543,
      "loss": 2.6802,
      "theoretical_loss": 3.6406274929997267,
      "tokens_seen": 1025048576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034820461384152456,
      "loss": 2.6781,
      "theoretical_loss": 3.6406054438420536,
      "tokens_seen": 1025114112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481945837512538,
      "loss": 2.7748,
      "theoretical_loss": 3.6405833964886103,
      "tokens_seen": 1025179648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034818455366098297,
      "loss": 2.6838,
      "theoretical_loss": 3.6405613509391346,
      "tokens_seen": 1025245184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034817452357071215,
      "loss": 2.782,
      "theoretical_loss": 3.640539307193363,
      "tokens_seen": 1025310720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034816449348044133,
      "loss": 2.7386,
      "theoretical_loss": 3.640517265251033,
      "tokens_seen": 1025376256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481544633901705,
      "loss": 2.9111,
      "theoretical_loss": 3.6404952251118816,
      "tokens_seen": 1025441792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481444332998997,
      "loss": 2.5696,
      "theoretical_loss": 3.6404731867756457,
      "tokens_seen": 1025507328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034813440320962893,
      "loss": 2.5349,
      "theoretical_loss": 3.640451150242063,
      "tokens_seen": 1025572864
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 572959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5865912437438965,
      "objective/train/theoretical_loss": 3.6404291155108712,
      "objective/train/tokens_used": 1046098400,
      "theoretical_loss": 3.6404291155108712,
      "tokens_seen": 1025638400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034812437311935806,
      "loss": 2.5791,
      "theoretical_loss": 3.6404291155108712,
      "tokens_seen": 1025638400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481143430290873,
      "loss": 2.7193,
      "theoretical_loss": 3.640407082581807,
      "tokens_seen": 1025703936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481043129388164,
      "loss": 2.7651,
      "theoretical_loss": 3.6403850514546083,
      "tokens_seen": 1025769472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034809428284854566,
      "loss": 2.6543,
      "theoretical_loss": 3.6403630221290126,
      "tokens_seen": 1025835008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034808425275827484,
      "loss": 2.7465,
      "theoretical_loss": 3.640340994604758,
      "tokens_seen": 1025900544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348074222668004,
      "loss": 2.5764,
      "theoretical_loss": 3.640318968881581,
      "tokens_seen": 1025966080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480641925777332,
      "loss": 2.74,
      "theoretical_loss": 3.6402969449592195,
      "tokens_seen": 1026031616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034805416248746243,
      "loss": 2.4407,
      "theoretical_loss": 3.6402749228374125,
      "tokens_seen": 1026097152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034804413239719156,
      "loss": 2.5572,
      "theoretical_loss": 3.6402529025158965,
      "tokens_seen": 1026162688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480341023069208,
      "loss": 2.9396,
      "theoretical_loss": 3.64023088399441,
      "tokens_seen": 1026228224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480240722166499,
      "loss": 2.6433,
      "theoretical_loss": 3.640208867272691,
      "tokens_seen": 1026293760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034801404212637916,
      "loss": 2.8696,
      "theoretical_loss": 3.6401868523504772,
      "tokens_seen": 1026359296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034800401203610834,
      "loss": 2.6728,
      "theoretical_loss": 3.640164839227506,
      "tokens_seen": 1026424832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479939819458375,
      "loss": 2.6425,
      "theoretical_loss": 3.640142827903517,
      "tokens_seen": 1026490368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479839518555667,
      "loss": 2.7012,
      "theoretical_loss": 3.6401208183782474,
      "tokens_seen": 1026555904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479739217652959,
      "loss": 2.7336,
      "theoretical_loss": 3.640098810651435,
      "tokens_seen": 1026621440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034796389167502506,
      "loss": 2.8522,
      "theoretical_loss": 3.6400768047228187,
      "tokens_seen": 1026686976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479538615847543,
      "loss": 2.7545,
      "theoretical_loss": 3.6400548005921367,
      "tokens_seen": 1026752512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034794383149448343,
      "loss": 2.6936,
      "theoretical_loss": 3.6400327982591274,
      "tokens_seen": 1026818048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034793380140421266,
      "loss": 2.7252,
      "theoretical_loss": 3.640010797723529,
      "tokens_seen": 1026883584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034792377131394184,
      "loss": 2.9514,
      "theoretical_loss": 3.6399887989850805,
      "tokens_seen": 1026949120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347913741223671,
      "loss": 2.6321,
      "theoretical_loss": 3.63996680204352,
      "tokens_seen": 1027014656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479037111334002,
      "loss": 2.6391,
      "theoretical_loss": 3.639944806898585,
      "tokens_seen": 1027080192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478936810431294,
      "loss": 2.6399,
      "theoretical_loss": 3.639922813550016,
      "tokens_seen": 1027145728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034788365095285857,
      "loss": 2.8079,
      "theoretical_loss": 3.6399008219975513,
      "tokens_seen": 1027211264
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 573554,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3572163581848145,
      "objective/train/theoretical_loss": 3.639878832240929,
      "objective/train/tokens_used": 1047736800,
      "theoretical_loss": 3.639878832240929,
      "tokens_seen": 1027276800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478736208625878,
      "loss": 2.6548,
      "theoretical_loss": 3.639878832240929,
      "tokens_seen": 1027276800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034786359077231693,
      "loss": 2.6283,
      "theoretical_loss": 3.6398568442798878,
      "tokens_seen": 1027342336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034785356068204617,
      "loss": 2.7042,
      "theoretical_loss": 3.6398348581141677,
      "tokens_seen": 1027407872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478435305917753,
      "loss": 2.6247,
      "theoretical_loss": 3.639812873743506,
      "tokens_seen": 1027473408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034783350050150453,
      "loss": 2.6911,
      "theoretical_loss": 3.6397908911676424,
      "tokens_seen": 1027538944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478234704112337,
      "loss": 2.8631,
      "theoretical_loss": 3.6397689103863167,
      "tokens_seen": 1027604480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478134403209629,
      "loss": 2.7386,
      "theoretical_loss": 3.6397469313992667,
      "tokens_seen": 1027670016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034780341023069207,
      "loss": 2.7186,
      "theoretical_loss": 3.6397249542062324,
      "tokens_seen": 1027735552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034779338014042125,
      "loss": 2.8693,
      "theoretical_loss": 3.6397029788069526,
      "tokens_seen": 1027801088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034778335005015043,
      "loss": 2.5519,
      "theoretical_loss": 3.639681005201167,
      "tokens_seen": 1027866624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034777331995987967,
      "loss": 2.6078,
      "theoretical_loss": 3.6396590333886136,
      "tokens_seen": 1027932160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477632898696088,
      "loss": 2.5822,
      "theoretical_loss": 3.639637063369033,
      "tokens_seen": 1027997696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034775325977933803,
      "loss": 2.8317,
      "theoretical_loss": 3.6396150951421644,
      "tokens_seen": 1028063232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477432296890672,
      "loss": 2.6957,
      "theoretical_loss": 3.639593128707747,
      "tokens_seen": 1028128768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477331995987964,
      "loss": 2.6182,
      "theoretical_loss": 3.6395711640655204,
      "tokens_seen": 1028194304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477231695085256,
      "loss": 2.5697,
      "theoretical_loss": 3.639549201215224,
      "tokens_seen": 1028259840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034771313941825476,
      "loss": 2.4462,
      "theoretical_loss": 3.639527240156598,
      "tokens_seen": 1028325376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034770310932798394,
      "loss": 2.7429,
      "theoretical_loss": 3.6395052808893813,
      "tokens_seen": 1028390912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034769307923771317,
      "loss": 2.8586,
      "theoretical_loss": 3.639483323413314,
      "tokens_seen": 1028456448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476830491474423,
      "loss": 2.8388,
      "theoretical_loss": 3.6394613677281358,
      "tokens_seen": 1028521984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034767301905717153,
      "loss": 2.7639,
      "theoretical_loss": 3.6394394138335864,
      "tokens_seen": 1028587520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034766298896690066,
      "loss": 2.7978,
      "theoretical_loss": 3.639417461729406,
      "tokens_seen": 1028653056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476529588766299,
      "loss": 2.6436,
      "theoretical_loss": 3.6393955114153345,
      "tokens_seen": 1028718592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476429287863591,
      "loss": 2.7939,
      "theoretical_loss": 3.6393735628911115,
      "tokens_seen": 1028784128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034763289869608826,
      "loss": 2.563,
      "theoretical_loss": 3.639351616156478,
      "tokens_seen": 1028849664
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 574441,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6161293983459473,
      "objective/train/theoretical_loss": 3.639329671211173,
      "objective/train/tokens_used": 1049375200,
      "theoretical_loss": 3.639329671211173,
      "tokens_seen": 1028915200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034762286860581744,
      "loss": 2.5903,
      "theoretical_loss": 3.639329671211173,
      "tokens_seen": 1028915200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476128385155466,
      "loss": 2.5422,
      "theoretical_loss": 3.639307728054937,
      "tokens_seen": 1028980736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476028084252758,
      "loss": 2.5475,
      "theoretical_loss": 3.639285786687511,
      "tokens_seen": 1029046272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034759277833500504,
      "loss": 3.0258,
      "theoretical_loss": 3.6392638471086336,
      "tokens_seen": 1029111808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475827482447342,
      "loss": 2.5871,
      "theoretical_loss": 3.639241909318047,
      "tokens_seen": 1029177344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475727181544634,
      "loss": 2.6932,
      "theoretical_loss": 3.6392199733154897,
      "tokens_seen": 1029242880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034756268806419263,
      "loss": 2.6994,
      "theoretical_loss": 3.6391980391007035,
      "tokens_seen": 1029308416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034755265797392176,
      "loss": 2.6649,
      "theoretical_loss": 3.639176106673429,
      "tokens_seen": 1029373952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347542627883651,
      "loss": 2.4648,
      "theoretical_loss": 3.639154176033406,
      "tokens_seen": 1029439488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475325977933801,
      "loss": 2.7083,
      "theoretical_loss": 3.639132247180375,
      "tokens_seen": 1029505024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034752256770310936,
      "loss": 2.6017,
      "theoretical_loss": 3.639110320114077,
      "tokens_seen": 1029570560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034751253761283854,
      "loss": 2.6896,
      "theoretical_loss": 3.6390883948342534,
      "tokens_seen": 1029636096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475025075225677,
      "loss": 2.811,
      "theoretical_loss": 3.6390664713406435,
      "tokens_seen": 1029701632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474924774322969,
      "loss": 2.5898,
      "theoretical_loss": 3.639044549632989,
      "tokens_seen": 1029767168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474824473420261,
      "loss": 2.7057,
      "theoretical_loss": 3.6390226297110306,
      "tokens_seen": 1029832704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034747241725175527,
      "loss": 2.2744,
      "theoretical_loss": 3.6390007115745098,
      "tokens_seen": 1029898240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474623871614845,
      "loss": 2.6949,
      "theoretical_loss": 3.6389787952231663,
      "tokens_seen": 1029963776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034745235707121363,
      "loss": 2.5461,
      "theoretical_loss": 3.6389568806567416,
      "tokens_seen": 1030029312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034744232698094286,
      "loss": 2.6111,
      "theoretical_loss": 3.638934967874978,
      "tokens_seen": 1030094848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034743229689067204,
      "loss": 2.8444,
      "theoretical_loss": 3.638913056877615,
      "tokens_seen": 1030160384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474222668004012,
      "loss": 2.7156,
      "theoretical_loss": 3.6388911476643946,
      "tokens_seen": 1030225920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474122367101304,
      "loss": 2.7605,
      "theoretical_loss": 3.6388692402350578,
      "tokens_seen": 1030291456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474022066198596,
      "loss": 2.578,
      "theoretical_loss": 3.6388473345893457,
      "tokens_seen": 1030356992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034739217652958877,
      "loss": 2.6943,
      "theoretical_loss": 3.638825430727,
      "tokens_seen": 1030422528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347382146439318,
      "loss": 3.0443,
      "theoretical_loss": 3.6388035286477622,
      "tokens_seen": 1030488064
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 574891,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2433831691741943,
      "objective/train/theoretical_loss": 3.6387816283513734,
      "objective/train/tokens_used": 1051013600,
      "theoretical_loss": 3.6387816283513734,
      "tokens_seen": 1030553600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034737211634904713,
      "loss": 2.8729,
      "theoretical_loss": 3.6387816283513734,
      "tokens_seen": 1030553600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034736208625877637,
      "loss": 2.4595,
      "theoretical_loss": 3.6387597298375756,
      "tokens_seen": 1030619136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473520561685055,
      "loss": 2.7197,
      "theoretical_loss": 3.63873783310611,
      "tokens_seen": 1030684672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034734202607823473,
      "loss": 2.8291,
      "theoretical_loss": 3.638715938156718,
      "tokens_seen": 1030750208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473319959879639,
      "loss": 2.9111,
      "theoretical_loss": 3.638694044989142,
      "tokens_seen": 1030815744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473219658976931,
      "loss": 2.5698,
      "theoretical_loss": 3.638672153603123,
      "tokens_seen": 1030881280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034731193580742227,
      "loss": 2.6734,
      "theoretical_loss": 3.638650263998403,
      "tokens_seen": 1030946816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034730190571715145,
      "loss": 2.8147,
      "theoretical_loss": 3.638628376174724,
      "tokens_seen": 1031012352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034729187562688063,
      "loss": 2.7605,
      "theoretical_loss": 3.638606490131828,
      "tokens_seen": 1031077888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034728184553660987,
      "loss": 2.5385,
      "theoretical_loss": 3.6385846058694558,
      "tokens_seen": 1031143424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347271815446339,
      "loss": 2.8986,
      "theoretical_loss": 3.638562723387351,
      "tokens_seen": 1031208960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034726178535606823,
      "loss": 2.7973,
      "theoretical_loss": 3.638540842685255,
      "tokens_seen": 1031274496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472517552657974,
      "loss": 2.6124,
      "theoretical_loss": 3.63851896376291,
      "tokens_seen": 1031340032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472417251755266,
      "loss": 2.5257,
      "theoretical_loss": 3.6384970866200574,
      "tokens_seen": 1031405568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472316950852558,
      "loss": 2.7745,
      "theoretical_loss": 3.6384752112564405,
      "tokens_seen": 1031471104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034722166499498496,
      "loss": 2.8269,
      "theoretical_loss": 3.638453337671801,
      "tokens_seen": 1031536640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034721163490471414,
      "loss": 2.6715,
      "theoretical_loss": 3.638431465865881,
      "tokens_seen": 1031602176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034720160481444337,
      "loss": 2.8328,
      "theoretical_loss": 3.6384095958384237,
      "tokens_seen": 1031667712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471915747241725,
      "loss": 2.7759,
      "theoretical_loss": 3.638387727589171,
      "tokens_seen": 1031733248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034718154463390173,
      "loss": 3.0587,
      "theoretical_loss": 3.6383658611178653,
      "tokens_seen": 1031798784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034717151454363086,
      "loss": 2.4092,
      "theoretical_loss": 3.6383439964242488,
      "tokens_seen": 1031864320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471614844533601,
      "loss": 2.5852,
      "theoretical_loss": 3.638322133508065,
      "tokens_seen": 1031929856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471514543630893,
      "loss": 2.6847,
      "theoretical_loss": 3.6383002723690563,
      "tokens_seen": 1031995392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034714142427281846,
      "loss": 2.7011,
      "theoretical_loss": 3.6382784130069643,
      "tokens_seen": 1032060928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034713139418254764,
      "loss": 2.8238,
      "theoretical_loss": 3.6382565554215334,
      "tokens_seen": 1032126464
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 576151,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7170889377593994,
      "objective/train/theoretical_loss": 3.6382346996125055,
      "objective/train/tokens_used": 1052652000,
      "theoretical_loss": 3.6382346996125055,
      "tokens_seen": 1032192000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471213640922768,
      "loss": 2.6599,
      "theoretical_loss": 3.6382346996125055,
      "tokens_seen": 1032192000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347111334002006,
      "loss": 2.6971,
      "theoretical_loss": 3.638212845579623,
      "tokens_seen": 1032257536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034710130391173524,
      "loss": 2.6737,
      "theoretical_loss": 3.63819099332263,
      "tokens_seen": 1032323072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034709127382146436,
      "loss": 2.8611,
      "theoretical_loss": 3.6381691428412686,
      "tokens_seen": 1032388608
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470812437311936,
      "loss": 2.5908,
      "theoretical_loss": 3.6381472941352815,
      "tokens_seen": 1032454144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470712136409228,
      "loss": 2.7545,
      "theoretical_loss": 3.638125447204413,
      "tokens_seen": 1032519680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034706118355065196,
      "loss": 2.8943,
      "theoretical_loss": 3.6381036020484054,
      "tokens_seen": 1032585216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034705115346038114,
      "loss": 2.7134,
      "theoretical_loss": 3.638081758667002,
      "tokens_seen": 1032650752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470411233701103,
      "loss": 2.6238,
      "theoretical_loss": 3.6380599170599464,
      "tokens_seen": 1032716288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470310932798395,
      "loss": 2.7772,
      "theoretical_loss": 3.6380380772269807,
      "tokens_seen": 1032781824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034702106318956874,
      "loss": 2.7171,
      "theoretical_loss": 3.6380162391678494,
      "tokens_seen": 1032847360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034701103309929787,
      "loss": 2.7668,
      "theoretical_loss": 3.637994402882296,
      "tokens_seen": 1032912896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470010030090271,
      "loss": 2.5211,
      "theoretical_loss": 3.637972568370063,
      "tokens_seen": 1032978432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034699097291875623,
      "loss": 2.7032,
      "theoretical_loss": 3.6379507356308944,
      "tokens_seen": 1033043968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034698094282848547,
      "loss": 2.818,
      "theoretical_loss": 3.637928904664534,
      "tokens_seen": 1033109504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034697091273821465,
      "loss": 2.4894,
      "theoretical_loss": 3.637907075470725,
      "tokens_seen": 1033175040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034696088264794383,
      "loss": 2.7589,
      "theoretical_loss": 3.637885248049211,
      "tokens_seen": 1033240576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346950852557673,
      "loss": 2.5823,
      "theoretical_loss": 3.6378634223997364,
      "tokens_seen": 1033306112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034694082246740224,
      "loss": 2.6596,
      "theoretical_loss": 3.6378415985220443,
      "tokens_seen": 1033371648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034693079237713137,
      "loss": 2.7211,
      "theoretical_loss": 3.6378197764158786,
      "tokens_seen": 1033437184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469207622868606,
      "loss": 2.4441,
      "theoretical_loss": 3.637797956080983,
      "tokens_seen": 1033502720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034691073219658973,
      "loss": 2.701,
      "theoretical_loss": 3.6377761375171023,
      "tokens_seen": 1033568256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034690070210631897,
      "loss": 2.5034,
      "theoretical_loss": 3.637754320723979,
      "tokens_seen": 1033633792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034689067201604815,
      "loss": 2.615,
      "theoretical_loss": 3.6377325057013588,
      "tokens_seen": 1033699328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034688064192577733,
      "loss": 2.6485,
      "theoretical_loss": 3.6377106924489846,
      "tokens_seen": 1033764864
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 577510,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.366300106048584,
      "objective/train/theoretical_loss": 3.637688880966601,
      "objective/train/tokens_used": 1054290400,
      "theoretical_loss": 3.637688880966601,
      "tokens_seen": 1033830400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468706118355065,
      "loss": 2.4252,
      "theoretical_loss": 3.637688880966601,
      "tokens_seen": 1033830400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468605817452357,
      "loss": 2.768,
      "theoretical_loss": 3.6376670712539516,
      "tokens_seen": 1033895936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468505516549649,
      "loss": 2.7405,
      "theoretical_loss": 3.6376452633107816,
      "tokens_seen": 1033961472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468405215646941,
      "loss": 2.7376,
      "theoretical_loss": 3.637623457136835,
      "tokens_seen": 1034027008
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468304914744233,
      "loss": 2.6507,
      "theoretical_loss": 3.6376016527318558,
      "tokens_seen": 1034092544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034682046138415247,
      "loss": 2.808,
      "theoretical_loss": 3.6375798500955887,
      "tokens_seen": 1034158080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034681043129388165,
      "loss": 2.5972,
      "theoretical_loss": 3.6375580492277777,
      "tokens_seen": 1034223616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034680040120361083,
      "loss": 2.8287,
      "theoretical_loss": 3.637536250128168,
      "tokens_seen": 1034289152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034679037111334007,
      "loss": 2.6991,
      "theoretical_loss": 3.6375144527965038,
      "tokens_seen": 1034354688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467803410230692,
      "loss": 2.8229,
      "theoretical_loss": 3.6374926572325297,
      "tokens_seen": 1034420224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034677031093279843,
      "loss": 2.7673,
      "theoretical_loss": 3.63747086343599,
      "tokens_seen": 1034485760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467602808425276,
      "loss": 2.7937,
      "theoretical_loss": 3.63744907140663,
      "tokens_seen": 1034551296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467502507522568,
      "loss": 2.7354,
      "theoretical_loss": 3.637427281144195,
      "tokens_seen": 1034616832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346740220661986,
      "loss": 2.7832,
      "theoretical_loss": 3.637405492648428,
      "tokens_seen": 1034682368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034673019057171516,
      "loss": 2.7563,
      "theoretical_loss": 3.637383705919076,
      "tokens_seen": 1034747904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034672016048144434,
      "loss": 2.7772,
      "theoretical_loss": 3.637361920955882,
      "tokens_seen": 1034813440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034671013039117357,
      "loss": 2.7117,
      "theoretical_loss": 3.637340137758592,
      "tokens_seen": 1034878976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467001003009027,
      "loss": 2.6319,
      "theoretical_loss": 3.6373183563269516,
      "tokens_seen": 1034944512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034669007021063193,
      "loss": 2.608,
      "theoretical_loss": 3.637296576660705,
      "tokens_seen": 1035010048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034668004012036106,
      "loss": 2.5727,
      "theoretical_loss": 3.6372747987595977,
      "tokens_seen": 1035075584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466700100300903,
      "loss": 2.7038,
      "theoretical_loss": 3.6372530226233746,
      "tokens_seen": 1035141120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466599799398195,
      "loss": 2.7065,
      "theoretical_loss": 3.637231248251781,
      "tokens_seen": 1035206656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034664994984954866,
      "loss": 2.9293,
      "theoretical_loss": 3.6372094756445623,
      "tokens_seen": 1035272192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034663991975927784,
      "loss": 2.7916,
      "theoretical_loss": 3.6371877048014642,
      "tokens_seen": 1035337728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346629889669007,
      "loss": 2.5348,
      "theoretical_loss": 3.6371659357222317,
      "tokens_seen": 1035403264
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 578118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6748743057250977,
      "objective/train/theoretical_loss": 3.6371441684066097,
      "objective/train/tokens_used": 1055928800,
      "theoretical_loss": 3.6371441684066097,
      "tokens_seen": 1035468800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466198595787362,
      "loss": 2.7133,
      "theoretical_loss": 3.6371441684066097,
      "tokens_seen": 1035468800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034660982948846544,
      "loss": 2.5092,
      "theoretical_loss": 3.637122402854345,
      "tokens_seen": 1035534336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034659979939819456,
      "loss": 2.6481,
      "theoretical_loss": 3.6371006390651828,
      "tokens_seen": 1035599872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465897693079238,
      "loss": 2.7212,
      "theoretical_loss": 3.637078877038868,
      "tokens_seen": 1035665408
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346579739217653,
      "loss": 2.4035,
      "theoretical_loss": 3.6370571167751473,
      "tokens_seen": 1035730944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034656970912738216,
      "loss": 2.4784,
      "theoretical_loss": 3.6370353582737653,
      "tokens_seen": 1035796480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034655967903711134,
      "loss": 2.5488,
      "theoretical_loss": 3.6370136015344685,
      "tokens_seen": 1035862016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465496489468405,
      "loss": 2.713,
      "theoretical_loss": 3.6369918465570024,
      "tokens_seen": 1035927552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465396188565697,
      "loss": 2.7944,
      "theoretical_loss": 3.6369700933411133,
      "tokens_seen": 1035993088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034652958876629894,
      "loss": 2.6355,
      "theoretical_loss": 3.636948341886547,
      "tokens_seen": 1036058624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034651955867602807,
      "loss": 2.6931,
      "theoretical_loss": 3.636926592193049,
      "tokens_seen": 1036124160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465095285857573,
      "loss": 2.761,
      "theoretical_loss": 3.636904844260366,
      "tokens_seen": 1036189696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034649949849548643,
      "loss": 2.7571,
      "theoretical_loss": 3.636883098088244,
      "tokens_seen": 1036255232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034648946840521567,
      "loss": 2.7743,
      "theoretical_loss": 3.6368613536764283,
      "tokens_seen": 1036320768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034647943831494485,
      "loss": 2.2489,
      "theoretical_loss": 3.6368396110246666,
      "tokens_seen": 1036386304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034646940822467403,
      "loss": 2.7951,
      "theoretical_loss": 3.6368178701327034,
      "tokens_seen": 1036451840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464593781344032,
      "loss": 2.5367,
      "theoretical_loss": 3.6367961310002865,
      "tokens_seen": 1036517376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034644934804413244,
      "loss": 2.635,
      "theoretical_loss": 3.636774393627162,
      "tokens_seen": 1036582912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034643931795386157,
      "loss": 2.563,
      "theoretical_loss": 3.6367526580130756,
      "tokens_seen": 1036648448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464292878635908,
      "loss": 2.4748,
      "theoretical_loss": 3.6367309241577743,
      "tokens_seen": 1036713984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034641925777331993,
      "loss": 2.7516,
      "theoretical_loss": 3.636709192061004,
      "tokens_seen": 1036779520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034640922768304917,
      "loss": 2.6484,
      "theoretical_loss": 3.636687461722512,
      "tokens_seen": 1036845056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034639919759277835,
      "loss": 2.8035,
      "theoretical_loss": 3.6366657331420447,
      "tokens_seen": 1036910592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034638916750250753,
      "loss": 2.6139,
      "theoretical_loss": 3.636644006319349,
      "tokens_seen": 1036976128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463791374122367,
      "loss": 2.7835,
      "theoretical_loss": 3.636622281254171,
      "tokens_seen": 1037041664
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 579340,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.744521379470825,
      "objective/train/theoretical_loss": 3.6366005579462577,
      "objective/train/tokens_used": 1057567200,
      "theoretical_loss": 3.6366005579462577,
      "tokens_seen": 1037107200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463691073219659,
      "loss": 2.6837,
      "theoretical_loss": 3.6366005579462577,
      "tokens_seen": 1037107200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463590772316951,
      "loss": 2.7886,
      "theoretical_loss": 3.6365788363953566,
      "tokens_seen": 1037172736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463490471414243,
      "loss": 2.566,
      "theoretical_loss": 3.636557116601214,
      "tokens_seen": 1037238272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034633901705115344,
      "loss": 2.6344,
      "theoretical_loss": 3.636535398563576,
      "tokens_seen": 1037303808
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034632898696088267,
      "loss": 2.6984,
      "theoretical_loss": 3.6365136822821915,
      "tokens_seen": 1037369344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463189568706118,
      "loss": 2.5895,
      "theoretical_loss": 3.636491967756806,
      "tokens_seen": 1037434880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034630892678034103,
      "loss": 2.8962,
      "theoretical_loss": 3.6364702549871675,
      "tokens_seen": 1037500416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462988966900702,
      "loss": 2.7219,
      "theoretical_loss": 3.6364485439730223,
      "tokens_seen": 1037565952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462888665997994,
      "loss": 2.6679,
      "theoretical_loss": 3.6364268347141184,
      "tokens_seen": 1037631488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462788365095286,
      "loss": 2.7464,
      "theoretical_loss": 3.6364051272102027,
      "tokens_seen": 1037697024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462688064192578,
      "loss": 2.825,
      "theoretical_loss": 3.636383421461022,
      "tokens_seen": 1037762560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034625877632898694,
      "loss": 2.603,
      "theoretical_loss": 3.6363617174663245,
      "tokens_seen": 1037828096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462487462387162,
      "loss": 2.7439,
      "theoretical_loss": 3.6363400152258576,
      "tokens_seen": 1037893632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462387161484453,
      "loss": 2.6634,
      "theoretical_loss": 3.636318314739368,
      "tokens_seen": 1037959168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034622868605817454,
      "loss": 2.6705,
      "theoretical_loss": 3.6362966160066037,
      "tokens_seen": 1038024704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462186559679037,
      "loss": 2.5983,
      "theoretical_loss": 3.6362749190273123,
      "tokens_seen": 1038090240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462086258776329,
      "loss": 2.7591,
      "theoretical_loss": 3.6362532238012415,
      "tokens_seen": 1038155776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461985957873621,
      "loss": 2.3577,
      "theoretical_loss": 3.6362315303281383,
      "tokens_seen": 1038221312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034618856569709126,
      "loss": 2.5005,
      "theoretical_loss": 3.6362098386077513,
      "tokens_seen": 1038286848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034617853560682044,
      "loss": 2.6127,
      "theoretical_loss": 3.636188148639828,
      "tokens_seen": 1038352384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461685055165497,
      "loss": 2.8448,
      "theoretical_loss": 3.636166460424116,
      "tokens_seen": 1038417920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461584754262788,
      "loss": 2.3289,
      "theoretical_loss": 3.636144773960363,
      "tokens_seen": 1038483456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034614844533600804,
      "loss": 2.5375,
      "theoretical_loss": 3.6361230892483176,
      "tokens_seen": 1038548992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034613841524573717,
      "loss": 2.7036,
      "theoretical_loss": 3.636101406287727,
      "tokens_seen": 1038614528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461283851554664,
      "loss": 2.4236,
      "theoretical_loss": 3.63607972507834,
      "tokens_seen": 1038680064
    },
    {
      "epoch": 0.31,
      "objective/train/docs_used": 579796,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.493173837661743,
      "objective/train/theoretical_loss": 3.6360580456199036,
      "objective/train/tokens_used": 1059205600,
      "theoretical_loss": 3.6360580456199036,
      "tokens_seen": 1038745600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461183550651956,
      "loss": 2.5022,
      "theoretical_loss": 3.6360580456199036,
      "tokens_seen": 1038745600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034610832497492476,
      "loss": 2.841,
      "theoretical_loss": 3.636036367912167,
      "tokens_seen": 1038811136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034609829488465395,
      "loss": 2.7636,
      "theoretical_loss": 3.6360146919548786,
      "tokens_seen": 1038876672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460882647943832,
      "loss": 2.8077,
      "theoretical_loss": 3.6359930177477855,
      "tokens_seen": 1038942208
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034607823470411236,
      "loss": 2.6653,
      "theoretical_loss": 3.635971345290637,
      "tokens_seen": 1039007744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034606820461384154,
      "loss": 2.5918,
      "theoretical_loss": 3.635949674583181,
      "tokens_seen": 1039073280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460581745235707,
      "loss": 2.6781,
      "theoretical_loss": 3.635928005625166,
      "tokens_seen": 1039138816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460481444332999,
      "loss": 2.6605,
      "theoretical_loss": 3.6359063384163397,
      "tokens_seen": 1039204352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034603811434302914,
      "loss": 2.6383,
      "theoretical_loss": 3.635884672956452,
      "tokens_seen": 1039269888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034602808425275827,
      "loss": 2.7481,
      "theoretical_loss": 3.6358630092452504,
      "tokens_seen": 1039335424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460180541624875,
      "loss": 2.5954,
      "theoretical_loss": 3.6358413472824838,
      "tokens_seen": 1039400960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034600802407221663,
      "loss": 2.5512,
      "theoretical_loss": 3.635819687067901,
      "tokens_seen": 1039466496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034599799398194587,
      "loss": 2.6184,
      "theoretical_loss": 3.6357980286012506,
      "tokens_seen": 1039532032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034598796389167505,
      "loss": 2.7418,
      "theoretical_loss": 3.6357763718822813,
      "tokens_seen": 1039597568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034597793380140423,
      "loss": 2.6507,
      "theoretical_loss": 3.635754716910742,
      "tokens_seen": 1039663104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003459679037111334,
      "loss": 2.3988,
      "theoretical_loss": 3.6357330636863816,
      "tokens_seen": 1039728640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034595787362086264,
      "loss": 2.6205,
      "theoretical_loss": 3.6357114122089493,
      "tokens_seen": 1039794176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034594784353059177,
      "loss": 2.858,
      "theoretical_loss": 3.6356897624781936,
      "tokens_seen": 1039859712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345937813440321,
      "loss": 2.6748,
      "theoretical_loss": 3.635668114493863,
      "tokens_seen": 1039925248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034592778335005013,
      "loss": 2.5437,
      "theoretical_loss": 3.635646468255708,
      "tokens_seen": 1039990784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034591775325977937,
      "loss": 2.4441,
      "theoretical_loss": 3.635624823763477,
      "tokens_seen": 1040056320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034590772316950855,
      "loss": 2.562,
      "theoretical_loss": 3.6356031810169185,
      "tokens_seen": 1040121856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034589769307923773,
      "loss": 2.7612,
      "theoretical_loss": 3.635581540015783,
      "tokens_seen": 1040187392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458876629889669,
      "loss": 2.6531,
      "theoretical_loss": 3.635559900759819,
      "tokens_seen": 1040252928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458776328986961,
      "loss": 2.8936,
      "theoretical_loss": 3.635538263248776,
      "tokens_seen": 1040318464
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 580938,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.539257049560547,
      "objective/train/theoretical_loss": 3.635516627482403,
      "objective/train/tokens_used": 1060844000,
      "theoretical_loss": 3.635516627482403,
      "tokens_seen": 1040384000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458676028084253,
      "loss": 2.5525,
      "theoretical_loss": 3.635516627482403,
      "tokens_seen": 1040384000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458575727181545,
      "loss": 2.6589,
      "theoretical_loss": 3.6354949934604504,
      "tokens_seen": 1040449536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034584754262788364,
      "loss": 2.7115,
      "theoretical_loss": 3.6354733611826666,
      "tokens_seen": 1040515072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034583751253761287,
      "loss": 2.6772,
      "theoretical_loss": 3.6354517306488017,
      "tokens_seen": 1040580608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345827482447342,
      "loss": 2.6409,
      "theoretical_loss": 3.6354301018586055,
      "tokens_seen": 1040646144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034581745235707123,
      "loss": 2.9847,
      "theoretical_loss": 3.635408474811827,
      "tokens_seen": 1040711680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458074222668004,
      "loss": 2.5639,
      "theoretical_loss": 3.635386849508217,
      "tokens_seen": 1040777216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457973921765296,
      "loss": 2.6658,
      "theoretical_loss": 3.635365225947524,
      "tokens_seen": 1040842752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457873620862588,
      "loss": 2.5532,
      "theoretical_loss": 3.635343604129498,
      "tokens_seen": 1040908288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345777331995988,
      "loss": 2.7893,
      "theoretical_loss": 3.63532198405389,
      "tokens_seen": 1040973824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034576730190571714,
      "loss": 2.6635,
      "theoretical_loss": 3.6353003657204486,
      "tokens_seen": 1041039360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457572718154464,
      "loss": 2.5693,
      "theoretical_loss": 3.635278749128924,
      "tokens_seen": 1041104896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457472417251755,
      "loss": 2.6342,
      "theoretical_loss": 3.635257134279067,
      "tokens_seen": 1041170432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034573721163490474,
      "loss": 2.6137,
      "theoretical_loss": 3.6352355211706264,
      "tokens_seen": 1041235968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457271815446339,
      "loss": 2.7036,
      "theoretical_loss": 3.635213909803354,
      "tokens_seen": 1041301504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457171514543631,
      "loss": 2.712,
      "theoretical_loss": 3.635192300176998,
      "tokens_seen": 1041367040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457071213640923,
      "loss": 2.7841,
      "theoretical_loss": 3.63517069229131,
      "tokens_seen": 1041432576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034569709127382146,
      "loss": 2.7111,
      "theoretical_loss": 3.63514908614604,
      "tokens_seen": 1041498112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034568706118355064,
      "loss": 2.717,
      "theoretical_loss": 3.6351274817409376,
      "tokens_seen": 1041563648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456770310932799,
      "loss": 2.4966,
      "theoretical_loss": 3.6351058790757538,
      "tokens_seen": 1041629184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345667001003009,
      "loss": 2.7355,
      "theoretical_loss": 3.6350842781502393,
      "tokens_seen": 1041694720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034565697091273824,
      "loss": 2.691,
      "theoretical_loss": 3.635062678964144,
      "tokens_seen": 1041760256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034564694082246737,
      "loss": 2.6776,
      "theoretical_loss": 3.635041081517219,
      "tokens_seen": 1041825792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456369107321966,
      "loss": 2.5323,
      "theoretical_loss": 3.635019485809214,
      "tokens_seen": 1041891328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456268806419258,
      "loss": 2.701,
      "theoretical_loss": 3.6349978918398804,
      "tokens_seen": 1041956864
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 581557,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8283419609069824,
      "objective/train/theoretical_loss": 3.6349762996089683,
      "objective/train/tokens_used": 1062482400,
      "theoretical_loss": 3.6349762996089683,
      "tokens_seen": 1042022400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034561685055165496,
      "loss": 2.6366,
      "theoretical_loss": 3.6349762996089683,
      "tokens_seen": 1042022400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034560682046138415,
      "loss": 2.694,
      "theoretical_loss": 3.634954709116229,
      "tokens_seen": 1042087936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455967903711134,
      "loss": 2.6484,
      "theoretical_loss": 3.6349331203614126,
      "tokens_seen": 1042153472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455867602808425,
      "loss": 2.6699,
      "theoretical_loss": 3.6349115333442708,
      "tokens_seen": 1042219008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034557673019057174,
      "loss": 2.8731,
      "theoretical_loss": 3.6348899480645542,
      "tokens_seen": 1042284544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034556670010030087,
      "loss": 2.5907,
      "theoretical_loss": 3.634868364522013,
      "tokens_seen": 1042350080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455566700100301,
      "loss": 2.7887,
      "theoretical_loss": 3.634846782716399,
      "tokens_seen": 1042415616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455466399197593,
      "loss": 2.58,
      "theoretical_loss": 3.634825202647463,
      "tokens_seen": 1042481152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034553660982948847,
      "loss": 2.5548,
      "theoretical_loss": 3.634803624314956,
      "tokens_seen": 1042546688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034552657973921765,
      "loss": 2.6047,
      "theoretical_loss": 3.634782047718629,
      "tokens_seen": 1042612224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034551654964894683,
      "loss": 2.6368,
      "theoretical_loss": 3.6347604728582334,
      "tokens_seen": 1042677760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345506519558676,
      "loss": 2.752,
      "theoretical_loss": 3.6347388997335206,
      "tokens_seen": 1042743296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034549648946840525,
      "loss": 2.7275,
      "theoretical_loss": 3.6347173283442418,
      "tokens_seen": 1042808832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454864593781344,
      "loss": 2.6721,
      "theoretical_loss": 3.634695758690148,
      "tokens_seen": 1042874368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454764292878636,
      "loss": 2.7311,
      "theoretical_loss": 3.6346741907709914,
      "tokens_seen": 1042939904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034546639919759274,
      "loss": 2.7933,
      "theoretical_loss": 3.634652624586522,
      "tokens_seen": 1043005440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034545636910732197,
      "loss": 2.5669,
      "theoretical_loss": 3.634631060136493,
      "tokens_seen": 1043070976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034544633901705115,
      "loss": 2.6247,
      "theoretical_loss": 3.6346094974206546,
      "tokens_seen": 1043136512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034543630892678033,
      "loss": 2.7361,
      "theoretical_loss": 3.634587936438759,
      "tokens_seen": 1043202048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454262788365095,
      "loss": 2.385,
      "theoretical_loss": 3.6345663771905583,
      "tokens_seen": 1043267584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034541624874623875,
      "loss": 2.6379,
      "theoretical_loss": 3.6345448196758032,
      "tokens_seen": 1043333120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454062186559679,
      "loss": 2.7274,
      "theoretical_loss": 3.634523263894246,
      "tokens_seen": 1043398656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453961885656971,
      "loss": 2.6841,
      "theoretical_loss": 3.6345017098456385,
      "tokens_seen": 1043464192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034538615847542624,
      "loss": 2.5927,
      "theoretical_loss": 3.6344801575297323,
      "tokens_seen": 1043529728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453761283851555,
      "loss": 2.718,
      "theoretical_loss": 3.63445860694628,
      "tokens_seen": 1043595264
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 582812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.661972761154175,
      "objective/train/theoretical_loss": 3.634437058095033,
      "objective/train/tokens_used": 1064120800,
      "theoretical_loss": 3.634437058095033,
      "tokens_seen": 1043660800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034536609829488466,
      "loss": 2.6098,
      "theoretical_loss": 3.634437058095033,
      "tokens_seen": 1043660800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034535606820461384,
      "loss": 2.8238,
      "theoretical_loss": 3.6344155109757432,
      "tokens_seen": 1043726336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345346038114343,
      "loss": 2.8692,
      "theoretical_loss": 3.6343939655881625,
      "tokens_seen": 1043791872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453360080240722,
      "loss": 2.5667,
      "theoretical_loss": 3.634372421932044,
      "tokens_seen": 1043857408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034532597793380143,
      "loss": 2.9155,
      "theoretical_loss": 3.6343508800071387,
      "tokens_seen": 1043922944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453159478435306,
      "loss": 2.7324,
      "theoretical_loss": 3.634329339813199,
      "tokens_seen": 1043988480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453059177532598,
      "loss": 2.6672,
      "theoretical_loss": 3.634307801349978,
      "tokens_seen": 1044054016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345295887662989,
      "loss": 2.6618,
      "theoretical_loss": 3.6342862646172276,
      "tokens_seen": 1044119552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452858575727182,
      "loss": 2.7461,
      "theoretical_loss": 3.6342647296147,
      "tokens_seen": 1044185088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034527582748244734,
      "loss": 2.7409,
      "theoretical_loss": 3.6342431963421475,
      "tokens_seen": 1044250624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452657973921766,
      "loss": 2.7384,
      "theoretical_loss": 3.634221664799323,
      "tokens_seen": 1044316160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452557673019057,
      "loss": 2.7435,
      "theoretical_loss": 3.6342001349859783,
      "tokens_seen": 1044381696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034524573721163494,
      "loss": 2.7216,
      "theoretical_loss": 3.634178606901867,
      "tokens_seen": 1044447232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452357071213641,
      "loss": 2.6133,
      "theoretical_loss": 3.6341570805467405,
      "tokens_seen": 1044512768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452256770310933,
      "loss": 2.6797,
      "theoretical_loss": 3.6341355559203525,
      "tokens_seen": 1044578304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452156469408225,
      "loss": 2.7254,
      "theoretical_loss": 3.6341140330224553,
      "tokens_seen": 1044643840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034520561685055166,
      "loss": 2.627,
      "theoretical_loss": 3.634092511852802,
      "tokens_seen": 1044709376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034519558676028084,
      "loss": 2.5409,
      "theoretical_loss": 3.6340709924111447,
      "tokens_seen": 1044774912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451855566700101,
      "loss": 2.5336,
      "theoretical_loss": 3.634049474697237,
      "tokens_seen": 1044840448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451755265797392,
      "loss": 2.7567,
      "theoretical_loss": 3.6340279587108313,
      "tokens_seen": 1044905984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034516549648946844,
      "loss": 2.6346,
      "theoretical_loss": 3.634006444451681,
      "tokens_seen": 1044971520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034515546639919757,
      "loss": 2.5416,
      "theoretical_loss": 3.6339849319195388,
      "tokens_seen": 1045037056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451454363089268,
      "loss": 2.5917,
      "theoretical_loss": 3.6339634211141583,
      "tokens_seen": 1045102592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345135406218656,
      "loss": 2.7407,
      "theoretical_loss": 3.633941912035292,
      "tokens_seen": 1045168128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034512537612838517,
      "loss": 2.4952,
      "theoretical_loss": 3.633920404682693,
      "tokens_seen": 1045233664
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 583395,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6361796855926514,
      "objective/train/theoretical_loss": 3.633898899056115,
      "objective/train/tokens_used": 1065759200,
      "theoretical_loss": 3.633898899056115,
      "tokens_seen": 1045299200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034511534603811435,
      "loss": 2.7576,
      "theoretical_loss": 3.633898899056115,
      "tokens_seen": 1045299200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451053159478436,
      "loss": 2.5862,
      "theoretical_loss": 3.633877395155311,
      "tokens_seen": 1045364736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450952858575727,
      "loss": 2.631,
      "theoretical_loss": 3.6338558929800353,
      "tokens_seen": 1045430272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034508525576730194,
      "loss": 2.8207,
      "theoretical_loss": 3.63383439253004,
      "tokens_seen": 1045495808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034507522567703107,
      "loss": 2.5697,
      "theoretical_loss": 3.6338128938050787,
      "tokens_seen": 1045561344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450651955867603,
      "loss": 2.6425,
      "theoretical_loss": 3.633791396804905,
      "tokens_seen": 1045626880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450551654964895,
      "loss": 2.7739,
      "theoretical_loss": 3.6337699015292735,
      "tokens_seen": 1045692416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034504513540621867,
      "loss": 2.3681,
      "theoretical_loss": 3.633748407977936,
      "tokens_seen": 1045757952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034503510531594785,
      "loss": 2.8102,
      "theoretical_loss": 3.6337269161506476,
      "tokens_seen": 1045823488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034502507522567703,
      "loss": 2.4205,
      "theoretical_loss": 3.633705426047161,
      "tokens_seen": 1045889024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450150451354062,
      "loss": 2.5667,
      "theoretical_loss": 3.6336839376672305,
      "tokens_seen": 1045954560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034500501504513545,
      "loss": 2.4626,
      "theoretical_loss": 3.63366245101061,
      "tokens_seen": 1046020096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449949849548646,
      "loss": 2.503,
      "theoretical_loss": 3.633640966077053,
      "tokens_seen": 1046085632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449849548645938,
      "loss": 2.7242,
      "theoretical_loss": 3.633619482866313,
      "tokens_seen": 1046151168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034497492477432294,
      "loss": 2.7701,
      "theoretical_loss": 3.633598001378145,
      "tokens_seen": 1046216704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034496489468405217,
      "loss": 2.6704,
      "theoretical_loss": 3.633576521612302,
      "tokens_seen": 1046282240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034495486459378135,
      "loss": 2.5448,
      "theoretical_loss": 3.6335550435685384,
      "tokens_seen": 1046347776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034494483450351053,
      "loss": 2.559,
      "theoretical_loss": 3.6335335672466087,
      "tokens_seen": 1046413312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449348044132397,
      "loss": 2.853,
      "theoretical_loss": 3.6335120926462663,
      "tokens_seen": 1046478848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034492477432296895,
      "loss": 2.4466,
      "theoretical_loss": 3.633490619767266,
      "tokens_seen": 1046544384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449147442326981,
      "loss": 2.609,
      "theoretical_loss": 3.6334691486093615,
      "tokens_seen": 1046609920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449047141424273,
      "loss": 2.5306,
      "theoretical_loss": 3.633447679172308,
      "tokens_seen": 1046675456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034489468405215644,
      "loss": 2.4962,
      "theoretical_loss": 3.6334262114558586,
      "tokens_seen": 1046740992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448846539618857,
      "loss": 2.8402,
      "theoretical_loss": 3.6334047454597687,
      "tokens_seen": 1046806528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034487462387161486,
      "loss": 2.4984,
      "theoretical_loss": 3.633383281183792,
      "tokens_seen": 1046872064
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 584517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.44215989112854,
      "objective/train/theoretical_loss": 3.6333618186276837,
      "objective/train/tokens_used": 1067397600,
      "theoretical_loss": 3.6333618186276837,
      "tokens_seen": 1046937600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034486459378134404,
      "loss": 2.5829,
      "theoretical_loss": 3.6333618186276837,
      "tokens_seen": 1046937600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448545636910732,
      "loss": 2.6031,
      "theoretical_loss": 3.633340357791198,
      "tokens_seen": 1047003136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448445336008024,
      "loss": 2.5413,
      "theoretical_loss": 3.6333188986740894,
      "tokens_seen": 1047068672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448345035105316,
      "loss": 2.5823,
      "theoretical_loss": 3.6332974412761128,
      "tokens_seen": 1047134208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448244734202608,
      "loss": 2.5924,
      "theoretical_loss": 3.6332759855970225,
      "tokens_seen": 1047199744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034481444332998994,
      "loss": 2.7455,
      "theoretical_loss": 3.633254531636574,
      "tokens_seen": 1047265280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448044132397192,
      "loss": 2.4347,
      "theoretical_loss": 3.633233079394521,
      "tokens_seen": 1047330816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034479438314944836,
      "loss": 2.6795,
      "theoretical_loss": 3.633211628870619,
      "tokens_seen": 1047396352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034478435305917754,
      "loss": 2.5025,
      "theoretical_loss": 3.6331901800646236,
      "tokens_seen": 1047461888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447743229689067,
      "loss": 2.679,
      "theoretical_loss": 3.633168732976288,
      "tokens_seen": 1047527424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447642928786359,
      "loss": 2.6558,
      "theoretical_loss": 3.633147287605369,
      "tokens_seen": 1047592960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447542627883651,
      "loss": 2.6364,
      "theoretical_loss": 3.6331258439516203,
      "tokens_seen": 1047658496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447442326980943,
      "loss": 2.4161,
      "theoretical_loss": 3.633104402014798,
      "tokens_seen": 1047724032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034473420260782345,
      "loss": 2.5527,
      "theoretical_loss": 3.633082961794656,
      "tokens_seen": 1047789568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447241725175527,
      "loss": 2.5928,
      "theoretical_loss": 3.6330615232909507,
      "tokens_seen": 1047855104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447141424272818,
      "loss": 2.625,
      "theoretical_loss": 3.6330400865034367,
      "tokens_seen": 1047920640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034470411233701104,
      "loss": 2.6341,
      "theoretical_loss": 3.63301865143187,
      "tokens_seen": 1047986176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446940822467402,
      "loss": 2.5154,
      "theoretical_loss": 3.6329972180760057,
      "tokens_seen": 1048051712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446840521564694,
      "loss": 2.5054,
      "theoretical_loss": 3.632975786435598,
      "tokens_seen": 1048117248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446740220661986,
      "loss": 2.7753,
      "theoretical_loss": 3.632954356510404,
      "tokens_seen": 1048182784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034466399197592777,
      "loss": 2.8493,
      "theoretical_loss": 3.6329329283001783,
      "tokens_seen": 1048248320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034465396188565695,
      "loss": 2.6918,
      "theoretical_loss": 3.632911501804677,
      "tokens_seen": 1048313856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446439317953862,
      "loss": 2.6121,
      "theoretical_loss": 3.632890077023655,
      "tokens_seen": 1048379392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446339017051153,
      "loss": 2.5529,
      "theoretical_loss": 3.6328686539568684,
      "tokens_seen": 1048444928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034462387161484455,
      "loss": 2.4928,
      "theoretical_loss": 3.6328472326040737,
      "tokens_seen": 1048510464
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 585002,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9021494388580322,
      "objective/train/theoretical_loss": 3.6328258129650246,
      "objective/train/tokens_used": 1069036000,
      "theoretical_loss": 3.6328258129650246,
      "tokens_seen": 1048576000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034461384152457373,
      "loss": 2.5486,
      "theoretical_loss": 3.6328258129650246,
      "tokens_seen": 1048576000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446038114343029,
      "loss": 2.7026,
      "theoretical_loss": 3.6328043950394786,
      "tokens_seen": 1048641536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034459378134403214,
      "loss": 2.4971,
      "theoretical_loss": 3.632782978827191,
      "tokens_seen": 1048707072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034458375125376127,
      "loss": 2.5622,
      "theoretical_loss": 3.6327615643279176,
      "tokens_seen": 1048772608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445737211634905,
      "loss": 2.7683,
      "theoretical_loss": 3.6327401515414146,
      "tokens_seen": 1048838144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445636910732197,
      "loss": 2.612,
      "theoretical_loss": 3.632718740467438,
      "tokens_seen": 1048903680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034455366098294887,
      "loss": 2.8323,
      "theoretical_loss": 3.6326973311057436,
      "tokens_seen": 1048969216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034454363089267805,
      "loss": 2.6183,
      "theoretical_loss": 3.6326759234560875,
      "tokens_seen": 1049034752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034453360080240723,
      "loss": 2.6601,
      "theoretical_loss": 3.632654517518226,
      "tokens_seen": 1049100288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445235707121364,
      "loss": 2.8211,
      "theoretical_loss": 3.6326331132919156,
      "tokens_seen": 1049165824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034451354062186565,
      "loss": 2.6244,
      "theoretical_loss": 3.632611710776912,
      "tokens_seen": 1049231360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445035105315948,
      "loss": 2.4925,
      "theoretical_loss": 3.632590309972972,
      "tokens_seen": 1049296896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344493480441324,
      "loss": 2.9594,
      "theoretical_loss": 3.6325689108798516,
      "tokens_seen": 1049362432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034448345035105314,
      "loss": 2.7034,
      "theoretical_loss": 3.6325475134973066,
      "tokens_seen": 1049427968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034447342026078237,
      "loss": 2.5489,
      "theoretical_loss": 3.632526117825095,
      "tokens_seen": 1049493504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034446339017051155,
      "loss": 2.6613,
      "theoretical_loss": 3.632504723862972,
      "tokens_seen": 1049559040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034445336008024073,
      "loss": 2.3941,
      "theoretical_loss": 3.632483331610695,
      "tokens_seen": 1049624576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444433299899699,
      "loss": 2.6026,
      "theoretical_loss": 3.63246194106802,
      "tokens_seen": 1049690112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034443329989969915,
      "loss": 2.8111,
      "theoretical_loss": 3.632440552234704,
      "tokens_seen": 1049755648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444232698094283,
      "loss": 2.7581,
      "theoretical_loss": 3.6324191651105036,
      "tokens_seen": 1049821184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444132397191575,
      "loss": 2.5065,
      "theoretical_loss": 3.632397779695175,
      "tokens_seen": 1049886720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034440320962888664,
      "loss": 2.4674,
      "theoretical_loss": 3.632376395988476,
      "tokens_seen": 1049952256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443931795386159,
      "loss": 2.3677,
      "theoretical_loss": 3.632355013990163,
      "tokens_seen": 1050017792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034438314944834506,
      "loss": 2.6539,
      "theoretical_loss": 3.632333633699992,
      "tokens_seen": 1050083328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034437311935807424,
      "loss": 2.3852,
      "theoretical_loss": 3.6323122551177214,
      "tokens_seen": 1050148864
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 586227,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.932581663131714,
      "objective/train/theoretical_loss": 3.632290878243108,
      "objective/train/tokens_used": 1070674400,
      "theoretical_loss": 3.632290878243108,
      "tokens_seen": 1050214400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443630892678034,
      "loss": 2.6061,
      "theoretical_loss": 3.632290878243108,
      "tokens_seen": 1050214400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443530591775326,
      "loss": 2.6796,
      "theoretical_loss": 3.632269503075908,
      "tokens_seen": 1050279936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443430290872618,
      "loss": 2.7109,
      "theoretical_loss": 3.6322481296158786,
      "tokens_seen": 1050345472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344332998996991,
      "loss": 2.5937,
      "theoretical_loss": 3.632226757862778,
      "tokens_seen": 1050411008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034432296890672014,
      "loss": 2.7738,
      "theoretical_loss": 3.632205387816362,
      "tokens_seen": 1050476544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443129388164494,
      "loss": 2.7724,
      "theoretical_loss": 3.632184019476389,
      "tokens_seen": 1050542080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034430290872617856,
      "loss": 2.635,
      "theoretical_loss": 3.6321626528426156,
      "tokens_seen": 1050607616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034429287863590774,
      "loss": 2.5516,
      "theoretical_loss": 3.6321412879147994,
      "tokens_seen": 1050673152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442828485456369,
      "loss": 2.7096,
      "theoretical_loss": 3.632119924692698,
      "tokens_seen": 1050738688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442728184553661,
      "loss": 2.5264,
      "theoretical_loss": 3.632098563176068,
      "tokens_seen": 1050804224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442627883650953,
      "loss": 2.4598,
      "theoretical_loss": 3.632077203364668,
      "tokens_seen": 1050869760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442527582748245,
      "loss": 2.4605,
      "theoretical_loss": 3.632055845258255,
      "tokens_seen": 1050935296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034424272818455365,
      "loss": 2.6781,
      "theoretical_loss": 3.6320344888565863,
      "tokens_seen": 1051000832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442326980942829,
      "loss": 2.6028,
      "theoretical_loss": 3.6320131341594206,
      "tokens_seen": 1051066368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344222668004012,
      "loss": 2.5341,
      "theoretical_loss": 3.6319917811665143,
      "tokens_seen": 1051131904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034421263791374124,
      "loss": 2.5869,
      "theoretical_loss": 3.631970429877626,
      "tokens_seen": 1051197440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442026078234704,
      "loss": 2.7181,
      "theoretical_loss": 3.6319490802925136,
      "tokens_seen": 1051262976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441925777331996,
      "loss": 2.8044,
      "theoretical_loss": 3.631927732410934,
      "tokens_seen": 1051328512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441825476429288,
      "loss": 2.6143,
      "theoretical_loss": 3.6319063862326457,
      "tokens_seen": 1051394048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034417251755265797,
      "loss": 2.8513,
      "theoretical_loss": 3.631885041757407,
      "tokens_seen": 1051459584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034416248746238715,
      "loss": 2.5423,
      "theoretical_loss": 3.631863698984975,
      "tokens_seen": 1051525120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441524573721164,
      "loss": 2.5051,
      "theoretical_loss": 3.631842357915108,
      "tokens_seen": 1051590656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441424272818455,
      "loss": 2.4431,
      "theoretical_loss": 3.631821018547565,
      "tokens_seen": 1051656192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034413239719157475,
      "loss": 2.6653,
      "theoretical_loss": 3.6317996808821027,
      "tokens_seen": 1051721728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034412236710130393,
      "loss": 2.5794,
      "theoretical_loss": 3.631778344918481,
      "tokens_seen": 1051787264
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 587166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6821694374084473,
      "objective/train/theoretical_loss": 3.6317570106564565,
      "objective/train/tokens_used": 1072312800,
      "theoretical_loss": 3.6317570106564565,
      "tokens_seen": 1051852800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441123370110331,
      "loss": 2.6543,
      "theoretical_loss": 3.6317570106564565,
      "tokens_seen": 1051852800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441023069207623,
      "loss": 2.6049,
      "theoretical_loss": 3.6317356780957883,
      "tokens_seen": 1051918336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034409227683049147,
      "loss": 2.2439,
      "theoretical_loss": 3.6317143472362345,
      "tokens_seen": 1051983872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034408224674022065,
      "loss": 2.7155,
      "theoretical_loss": 3.631693018077554,
      "tokens_seen": 1052049408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440722166499499,
      "loss": 2.5276,
      "theoretical_loss": 3.6316716906195046,
      "tokens_seen": 1052114944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344062186559679,
      "loss": 2.512,
      "theoretical_loss": 3.6316503648618452,
      "tokens_seen": 1052180480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034405215646940825,
      "loss": 2.7207,
      "theoretical_loss": 3.6316290408043335,
      "tokens_seen": 1052246016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440421263791374,
      "loss": 2.6449,
      "theoretical_loss": 3.6316077184467295,
      "tokens_seen": 1052311552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440320962888666,
      "loss": 2.6526,
      "theoretical_loss": 3.6315863977887908,
      "tokens_seen": 1052377088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440220661985958,
      "loss": 2.6497,
      "theoretical_loss": 3.6315650788302762,
      "tokens_seen": 1052442624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344012036108325,
      "loss": 2.7508,
      "theoretical_loss": 3.6315437615709447,
      "tokens_seen": 1052508160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034400200601805416,
      "loss": 2.8709,
      "theoretical_loss": 3.631522446010555,
      "tokens_seen": 1052573696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034399197592778334,
      "loss": 2.565,
      "theoretical_loss": 3.6315011321488657,
      "tokens_seen": 1052639232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439819458375125,
      "loss": 2.6508,
      "theoretical_loss": 3.631479819985636,
      "tokens_seen": 1052704768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034397191574724175,
      "loss": 2.7176,
      "theoretical_loss": 3.6314585095206247,
      "tokens_seen": 1052770304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439618856569709,
      "loss": 2.7527,
      "theoretical_loss": 3.6314372007535907,
      "tokens_seen": 1052835840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439518555667001,
      "loss": 2.631,
      "theoretical_loss": 3.6314158936842933,
      "tokens_seen": 1052901376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439418254764293,
      "loss": 2.7249,
      "theoretical_loss": 3.6313945883124914,
      "tokens_seen": 1052966912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439317953861585,
      "loss": 2.6789,
      "theoretical_loss": 3.6313732846379434,
      "tokens_seen": 1053032448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034392176529588766,
      "loss": 2.8109,
      "theoretical_loss": 3.63135198266041,
      "tokens_seen": 1053097984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034391173520561684,
      "loss": 2.4002,
      "theoretical_loss": 3.631330682379649,
      "tokens_seen": 1053163520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343901705115346,
      "loss": 2.6808,
      "theoretical_loss": 3.631309383795421,
      "tokens_seen": 1053229056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034389167502507526,
      "loss": 2.5586,
      "theoretical_loss": 3.631288086907484,
      "tokens_seen": 1053294592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438816449348044,
      "loss": 2.6458,
      "theoretical_loss": 3.631266791715598,
      "tokens_seen": 1053360128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438716148445336,
      "loss": 2.9422,
      "theoretical_loss": 3.6312454982195224,
      "tokens_seen": 1053425664
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 587595,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.472946882247925,
      "objective/train/theoretical_loss": 3.631224206419017,
      "objective/train/tokens_used": 1073951200,
      "theoretical_loss": 3.631224206419017,
      "tokens_seen": 1053491200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034386158475426275,
      "loss": 2.5474,
      "theoretical_loss": 3.631224206419017,
      "tokens_seen": 1053491200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343851554663992,
      "loss": 2.5543,
      "theoretical_loss": 3.631202916313841,
      "tokens_seen": 1053556736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438415245737212,
      "loss": 2.6514,
      "theoretical_loss": 3.6311816279037537,
      "tokens_seen": 1053622272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034383149448345034,
      "loss": 2.71,
      "theoretical_loss": 3.6311603411885147,
      "tokens_seen": 1053687808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438214643931796,
      "loss": 2.6042,
      "theoretical_loss": 3.631139056167884,
      "tokens_seen": 1053753344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034381143430290876,
      "loss": 2.5687,
      "theoretical_loss": 3.6311177728416215,
      "tokens_seen": 1053818880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034380140421263794,
      "loss": 2.7691,
      "theoretical_loss": 3.6310964912094867,
      "tokens_seen": 1053884416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437913741223671,
      "loss": 2.7997,
      "theoretical_loss": 3.6310752112712392,
      "tokens_seen": 1053949952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437813440320963,
      "loss": 2.4598,
      "theoretical_loss": 3.631053933026639,
      "tokens_seen": 1054015488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437713139418255,
      "loss": 2.7018,
      "theoretical_loss": 3.6310326564754467,
      "tokens_seen": 1054081024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437612838515547,
      "loss": 2.596,
      "theoretical_loss": 3.631011381617421,
      "tokens_seen": 1054146560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034375125376128385,
      "loss": 2.9412,
      "theoretical_loss": 3.630990108452323,
      "tokens_seen": 1054212096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437412236710131,
      "loss": 2.6816,
      "theoretical_loss": 3.630968836979912,
      "tokens_seen": 1054277632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437311935807422,
      "loss": 2.6108,
      "theoretical_loss": 3.6309475671999483,
      "tokens_seen": 1054343168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034372116349047144,
      "loss": 2.6602,
      "theoretical_loss": 3.6309262991121924,
      "tokens_seen": 1054408704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437111334002006,
      "loss": 2.4562,
      "theoretical_loss": 3.630905032716404,
      "tokens_seen": 1054474240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437011033099298,
      "loss": 2.6896,
      "theoretical_loss": 3.630883768012344,
      "tokens_seen": 1054539776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343691073219659,
      "loss": 2.8709,
      "theoretical_loss": 3.6308625049997723,
      "tokens_seen": 1054605312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034368104312938817,
      "loss": 2.5781,
      "theoretical_loss": 3.630841243678449,
      "tokens_seen": 1054670848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034367101303911735,
      "loss": 2.7216,
      "theoretical_loss": 3.630819984048135,
      "tokens_seen": 1054736384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436609829488466,
      "loss": 2.6508,
      "theoretical_loss": 3.6307987261085906,
      "tokens_seen": 1054801920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436509528585757,
      "loss": 2.6724,
      "theoretical_loss": 3.630777469859576,
      "tokens_seen": 1054867456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034364092276830495,
      "loss": 2.6541,
      "theoretical_loss": 3.6307562153008517,
      "tokens_seen": 1054932992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034363089267803413,
      "loss": 2.901,
      "theoretical_loss": 3.6307349624321787,
      "tokens_seen": 1054998528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436208625877633,
      "loss": 2.5115,
      "theoretical_loss": 3.630713711253318,
      "tokens_seen": 1055064064
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 588937,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4390525817871094,
      "objective/train/theoretical_loss": 3.6306924617640295,
      "objective/train/tokens_used": 1075589600,
      "theoretical_loss": 3.6306924617640295,
      "tokens_seen": 1055129600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436108324974925,
      "loss": 2.5418,
      "theoretical_loss": 3.6306924617640295,
      "tokens_seen": 1055129600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034360080240722167,
      "loss": 2.7519,
      "theoretical_loss": 3.6306712139640744,
      "tokens_seen": 1055195136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034359077231695085,
      "loss": 2.6733,
      "theoretical_loss": 3.6306499678532127,
      "tokens_seen": 1055260672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435807422266801,
      "loss": 2.5918,
      "theoretical_loss": 3.6306287234312062,
      "tokens_seen": 1055326208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435707121364092,
      "loss": 2.5974,
      "theoretical_loss": 3.630607480697816,
      "tokens_seen": 1055391744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034356068204613845,
      "loss": 2.6814,
      "theoretical_loss": 3.6305862396528017,
      "tokens_seen": 1055457280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435506519558676,
      "loss": 2.6467,
      "theoretical_loss": 3.6305650002959258,
      "tokens_seen": 1055522816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435406218655968,
      "loss": 2.5512,
      "theoretical_loss": 3.6305437626269486,
      "tokens_seen": 1055588352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343530591775326,
      "loss": 2.6324,
      "theoretical_loss": 3.6305225266456302,
      "tokens_seen": 1055653888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435205616850552,
      "loss": 2.6879,
      "theoretical_loss": 3.6305012923517337,
      "tokens_seen": 1055719424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034351053159478436,
      "loss": 2.3758,
      "theoretical_loss": 3.630480059745019,
      "tokens_seen": 1055784960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034350050150451354,
      "loss": 2.7239,
      "theoretical_loss": 3.6304588288252475,
      "tokens_seen": 1055850496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434904714142427,
      "loss": 2.6559,
      "theoretical_loss": 3.630437599592181,
      "tokens_seen": 1055916032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034348044132397195,
      "loss": 2.4738,
      "theoretical_loss": 3.6304163720455795,
      "tokens_seen": 1055981568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434704112337011,
      "loss": 2.7831,
      "theoretical_loss": 3.630395146185206,
      "tokens_seen": 1056047104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434603811434303,
      "loss": 2.6243,
      "theoretical_loss": 3.630373922010821,
      "tokens_seen": 1056112640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434503510531595,
      "loss": 2.6946,
      "theoretical_loss": 3.6303526995221866,
      "tokens_seen": 1056178176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434403209628887,
      "loss": 2.6034,
      "theoretical_loss": 3.6303314787190635,
      "tokens_seen": 1056243712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034343029087261786,
      "loss": 2.5507,
      "theoretical_loss": 3.630310259601214,
      "tokens_seen": 1056309248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034342026078234704,
      "loss": 2.5951,
      "theoretical_loss": 3.630289042168399,
      "tokens_seen": 1056374784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434102306920762,
      "loss": 2.7226,
      "theoretical_loss": 3.630267826420381,
      "tokens_seen": 1056440320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034340020060180546,
      "loss": 2.7196,
      "theoretical_loss": 3.630246612356921,
      "tokens_seen": 1056505856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433901705115346,
      "loss": 2.6928,
      "theoretical_loss": 3.630225399977781,
      "tokens_seen": 1056571392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433801404212638,
      "loss": 2.5232,
      "theoretical_loss": 3.630204189282723,
      "tokens_seen": 1056636928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034337011033099295,
      "loss": 2.7453,
      "theoretical_loss": 3.6301829802715084,
      "tokens_seen": 1056702464
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 589659,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.753281354904175,
      "objective/train/theoretical_loss": 3.6301617729438993,
      "objective/train/tokens_used": 1077228000,
      "theoretical_loss": 3.6301617729438993,
      "tokens_seen": 1056768000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433600802407222,
      "loss": 2.8091,
      "theoretical_loss": 3.6301617729438993,
      "tokens_seen": 1056768000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034335005015045136,
      "loss": 2.6981,
      "theoretical_loss": 3.6301405672996583,
      "tokens_seen": 1056833536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034334002006018054,
      "loss": 2.6649,
      "theoretical_loss": 3.630119363338547,
      "tokens_seen": 1056899072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433299899699097,
      "loss": 2.6571,
      "theoretical_loss": 3.6300981610603267,
      "tokens_seen": 1056964608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433199598796389,
      "loss": 2.6702,
      "theoretical_loss": 3.6300769604647605,
      "tokens_seen": 1057030144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433099297893681,
      "loss": 2.8513,
      "theoretical_loss": 3.63005576155161,
      "tokens_seen": 1057095680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432998996990973,
      "loss": 2.744,
      "theoretical_loss": 3.6300345643206375,
      "tokens_seen": 1057161216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034328986960882645,
      "loss": 2.4055,
      "theoretical_loss": 3.6300133687716056,
      "tokens_seen": 1057226752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432798395185557,
      "loss": 2.6993,
      "theoretical_loss": 3.6299921749042765,
      "tokens_seen": 1057292288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034326980942828487,
      "loss": 2.3819,
      "theoretical_loss": 3.6299709827184117,
      "tokens_seen": 1057357824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034325977933801405,
      "loss": 2.8359,
      "theoretical_loss": 3.6299497922137745,
      "tokens_seen": 1057423360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034324974924774323,
      "loss": 2.6268,
      "theoretical_loss": 3.6299286033901272,
      "tokens_seen": 1057488896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432397191574724,
      "loss": 2.7105,
      "theoretical_loss": 3.6299074162472325,
      "tokens_seen": 1057554432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432296890672016,
      "loss": 2.4748,
      "theoretical_loss": 3.6298862307848516,
      "tokens_seen": 1057619968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432196589769308,
      "loss": 2.8688,
      "theoretical_loss": 3.629865047002749,
      "tokens_seen": 1057685504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034320962888665995,
      "loss": 2.6851,
      "theoretical_loss": 3.6298438649006863,
      "tokens_seen": 1057751040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431995987963892,
      "loss": 2.7079,
      "theoretical_loss": 3.629822684478426,
      "tokens_seen": 1057816576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431895687061183,
      "loss": 2.8602,
      "theoretical_loss": 3.629801505735731,
      "tokens_seen": 1057882112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034317953861584755,
      "loss": 2.5481,
      "theoretical_loss": 3.629780328672364,
      "tokens_seen": 1057947648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034316950852557673,
      "loss": 2.8164,
      "theoretical_loss": 3.629759153288089,
      "tokens_seen": 1058013184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431594784353059,
      "loss": 2.6221,
      "theoretical_loss": 3.6297379795826674,
      "tokens_seen": 1058078720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431494483450351,
      "loss": 2.563,
      "theoretical_loss": 3.629716807555863,
      "tokens_seen": 1058144256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034313941825476433,
      "loss": 2.6197,
      "theoretical_loss": 3.629695637207438,
      "tokens_seen": 1058209792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034312938816449345,
      "loss": 2.6229,
      "theoretical_loss": 3.6296744685371554,
      "tokens_seen": 1058275328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431193580742227,
      "loss": 2.4343,
      "theoretical_loss": 3.629653301544779,
      "tokens_seen": 1058340864
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 591033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.53261661529541,
      "objective/train/theoretical_loss": 3.6296321362300716,
      "objective/train/tokens_used": 1078866400,
      "theoretical_loss": 3.6296321362300716,
      "tokens_seen": 1058406400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431093279839518,
      "loss": 2.5809,
      "theoretical_loss": 3.6296321362300716,
      "tokens_seen": 1058406400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034309929789368105,
      "loss": 2.5506,
      "theoretical_loss": 3.629610972592797,
      "tokens_seen": 1058471936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430892678034103,
      "loss": 2.8012,
      "theoretical_loss": 3.6295898106327167,
      "tokens_seen": 1058537472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430792377131394,
      "loss": 2.7142,
      "theoretical_loss": 3.629568650349596,
      "tokens_seen": 1058603008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034306920762286865,
      "loss": 2.7891,
      "theoretical_loss": 3.6295474917431965,
      "tokens_seen": 1058668544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430591775325978,
      "loss": 2.7441,
      "theoretical_loss": 3.6295263348132822,
      "tokens_seen": 1058734080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343049147442327,
      "loss": 2.3854,
      "theoretical_loss": 3.629505179559617,
      "tokens_seen": 1058799616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430391173520562,
      "loss": 2.6415,
      "theoretical_loss": 3.6294840259819643,
      "tokens_seen": 1058865152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430290872617854,
      "loss": 2.8416,
      "theoretical_loss": 3.6294628740800867,
      "tokens_seen": 1058930688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034301905717151456,
      "loss": 2.6718,
      "theoretical_loss": 3.6294417238537484,
      "tokens_seen": 1058996224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034300902708124374,
      "loss": 2.5243,
      "theoretical_loss": 3.6294205753027127,
      "tokens_seen": 1059061760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429989969909729,
      "loss": 2.6848,
      "theoretical_loss": 3.6293994284267437,
      "tokens_seen": 1059127296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034298896690070215,
      "loss": 2.7442,
      "theoretical_loss": 3.6293782832256047,
      "tokens_seen": 1059192832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429789368104313,
      "loss": 2.6438,
      "theoretical_loss": 3.629357139699059,
      "tokens_seen": 1059258368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429689067201605,
      "loss": 2.7197,
      "theoretical_loss": 3.629335997846872,
      "tokens_seen": 1059323904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429588766298897,
      "loss": 2.4123,
      "theoretical_loss": 3.6293148576688057,
      "tokens_seen": 1059389440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429488465396189,
      "loss": 2.662,
      "theoretical_loss": 3.6292937191646253,
      "tokens_seen": 1059454976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034293881644934806,
      "loss": 2.433,
      "theoretical_loss": 3.6292725823340932,
      "tokens_seen": 1059520512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034292878635907724,
      "loss": 2.5597,
      "theoretical_loss": 3.6292514471769755,
      "tokens_seen": 1059586048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429187562688064,
      "loss": 2.9441,
      "theoretical_loss": 3.629230313693035,
      "tokens_seen": 1059651584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034290872617853566,
      "loss": 2.3666,
      "theoretical_loss": 3.629209181882035,
      "tokens_seen": 1059717120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428986960882648,
      "loss": 2.9156,
      "theoretical_loss": 3.629188051743741,
      "tokens_seen": 1059782656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342888665997994,
      "loss": 2.6805,
      "theoretical_loss": 3.6291669232779165,
      "tokens_seen": 1059848192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034287863590772315,
      "loss": 2.5692,
      "theoretical_loss": 3.629145796484326,
      "tokens_seen": 1059913728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428686058174524,
      "loss": 2.7361,
      "theoretical_loss": 3.6291246713627334,
      "tokens_seen": 1059979264
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 592062,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5544235706329346,
      "objective/train/theoretical_loss": 3.6291035479129032,
      "objective/train/tokens_used": 1080504800,
      "theoretical_loss": 3.6291035479129032,
      "tokens_seen": 1060044800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034285857572718156,
      "loss": 2.4871,
      "theoretical_loss": 3.6291035479129032,
      "tokens_seen": 1060044800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034284854563691074,
      "loss": 2.3404,
      "theoretical_loss": 3.6290824261346,
      "tokens_seen": 1060110336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428385155466399,
      "loss": 2.8714,
      "theoretical_loss": 3.6290613060275883,
      "tokens_seen": 1060175872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428284854563691,
      "loss": 2.7667,
      "theoretical_loss": 3.629040187591632,
      "tokens_seen": 1060241408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428184553660983,
      "loss": 2.7177,
      "theoretical_loss": 3.6290190708264958,
      "tokens_seen": 1060306944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428084252758275,
      "loss": 2.7029,
      "theoretical_loss": 3.6289979557319443,
      "tokens_seen": 1060372480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034279839518555665,
      "loss": 2.634,
      "theoretical_loss": 3.628976842307742,
      "tokens_seen": 1060438016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427883650952859,
      "loss": 2.5498,
      "theoretical_loss": 3.6289557305536544,
      "tokens_seen": 1060503552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034277833500501507,
      "loss": 2.6537,
      "theoretical_loss": 3.628934620469445,
      "tokens_seen": 1060569088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034276830491474425,
      "loss": 2.7612,
      "theoretical_loss": 3.6289135120548783,
      "tokens_seen": 1060634624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034275827482447343,
      "loss": 2.845,
      "theoretical_loss": 3.6288924053097205,
      "tokens_seen": 1060700160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427482447342026,
      "loss": 2.8157,
      "theoretical_loss": 3.628871300233736,
      "tokens_seen": 1060765696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427382146439318,
      "loss": 2.9871,
      "theoretical_loss": 3.628850196826689,
      "tokens_seen": 1060831232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342728184553661,
      "loss": 3.0334,
      "theoretical_loss": 3.628829095088345,
      "tokens_seen": 1060896768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034271815446339015,
      "loss": 2.5297,
      "theoretical_loss": 3.6288079950184686,
      "tokens_seen": 1060962304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427081243731194,
      "loss": 2.5689,
      "theoretical_loss": 3.628786896616825,
      "tokens_seen": 1061027840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426980942828485,
      "loss": 2.7144,
      "theoretical_loss": 3.6287657998831797,
      "tokens_seen": 1061093376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034268806419257775,
      "loss": 2.6815,
      "theoretical_loss": 3.628744704817297,
      "tokens_seen": 1061158912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034267803410230693,
      "loss": 2.7748,
      "theoretical_loss": 3.628723611418943,
      "tokens_seen": 1061224448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426680040120361,
      "loss": 2.6297,
      "theoretical_loss": 3.6287025196878826,
      "tokens_seen": 1061289984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426579739217653,
      "loss": 2.5459,
      "theoretical_loss": 3.6286814296238803,
      "tokens_seen": 1061355520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034264794383149453,
      "loss": 2.872,
      "theoretical_loss": 3.6286603412267024,
      "tokens_seen": 1061421056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034263791374122366,
      "loss": 2.6676,
      "theoretical_loss": 3.6286392544961137,
      "tokens_seen": 1061486592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426278836509529,
      "loss": 2.6477,
      "theoretical_loss": 3.62861816943188,
      "tokens_seen": 1061552128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342617853560682,
      "loss": 2.8712,
      "theoretical_loss": 3.6285970860337664,
      "tokens_seen": 1061617664
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 592448,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1389501094818115,
      "objective/train/theoretical_loss": 3.6285760043015385,
      "objective/train/tokens_used": 1082143200,
      "theoretical_loss": 3.6285760043015385,
      "tokens_seen": 1061683200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034260782347041125,
      "loss": 2.7504,
      "theoretical_loss": 3.6285760043015385,
      "tokens_seen": 1061683200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034259779338014043,
      "loss": 2.6796,
      "theoretical_loss": 3.6285549242349626,
      "tokens_seen": 1061748736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425877632898696,
      "loss": 2.6368,
      "theoretical_loss": 3.628533845833803,
      "tokens_seen": 1061814272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425777331995988,
      "loss": 2.8308,
      "theoretical_loss": 3.628512769097826,
      "tokens_seen": 1061879808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342567703109328,
      "loss": 2.613,
      "theoretical_loss": 3.6284916940267973,
      "tokens_seen": 1061945344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034255767301905716,
      "loss": 2.9996,
      "theoretical_loss": 3.628470620620482,
      "tokens_seen": 1062010880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425476429287864,
      "loss": 2.6594,
      "theoretical_loss": 3.6284495488786472,
      "tokens_seen": 1062076416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425376128385155,
      "loss": 2.523,
      "theoretical_loss": 3.628428478801058,
      "tokens_seen": 1062141952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034252758274824476,
      "loss": 2.87,
      "theoretical_loss": 3.62840741038748,
      "tokens_seen": 1062207488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425175526579739,
      "loss": 2.608,
      "theoretical_loss": 3.62838634363768,
      "tokens_seen": 1062273024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425075225677031,
      "loss": 2.6211,
      "theoretical_loss": 3.628365278551423,
      "tokens_seen": 1062338560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424974924774323,
      "loss": 2.8357,
      "theoretical_loss": 3.6283442151284753,
      "tokens_seen": 1062404096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424874623871615,
      "loss": 2.7023,
      "theoretical_loss": 3.628323153368603,
      "tokens_seen": 1062469632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034247743229689066,
      "loss": 2.2979,
      "theoretical_loss": 3.628302093271573,
      "tokens_seen": 1062535168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424674022066199,
      "loss": 2.5555,
      "theoretical_loss": 3.6282810348371504,
      "tokens_seen": 1062600704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342457372116349,
      "loss": 2.4632,
      "theoretical_loss": 3.6282599780651017,
      "tokens_seen": 1062666240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034244734202607826,
      "loss": 2.9378,
      "theoretical_loss": 3.6282389229551932,
      "tokens_seen": 1062731776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424373119358074,
      "loss": 2.6977,
      "theoretical_loss": 3.6282178695071914,
      "tokens_seen": 1062797312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424272818455366,
      "loss": 2.4779,
      "theoretical_loss": 3.6281968177208626,
      "tokens_seen": 1062862848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424172517552658,
      "loss": 2.7706,
      "theoretical_loss": 3.628175767595973,
      "tokens_seen": 1062928384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342407221664995,
      "loss": 2.7399,
      "theoretical_loss": 3.628154719132289,
      "tokens_seen": 1062993920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034239719157472416,
      "loss": 2.4667,
      "theoretical_loss": 3.628133672329578,
      "tokens_seen": 1063059456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034238716148445335,
      "loss": 2.7334,
      "theoretical_loss": 3.628112627187605,
      "tokens_seen": 1063124992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423771313941825,
      "loss": 2.8203,
      "theoretical_loss": 3.628091583706138,
      "tokens_seen": 1063190528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034236710130391176,
      "loss": 2.614,
      "theoretical_loss": 3.6280705418849433,
      "tokens_seen": 1063256064
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 593762,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.574810743331909,
      "objective/train/theoretical_loss": 3.628049501723787,
      "objective/train/tokens_used": 1083781600,
      "theoretical_loss": 3.628049501723787,
      "tokens_seen": 1063321600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423570712136409,
      "loss": 2.7115,
      "theoretical_loss": 3.628049501723787,
      "tokens_seen": 1063321600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423470411233701,
      "loss": 2.7239,
      "theoretical_loss": 3.628028463222436,
      "tokens_seen": 1063387136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423370110330993,
      "loss": 2.5938,
      "theoretical_loss": 3.628007426380657,
      "tokens_seen": 1063452672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423269809428285,
      "loss": 2.7494,
      "theoretical_loss": 3.6279863911982178,
      "tokens_seen": 1063518208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423169508525577,
      "loss": 2.5914,
      "theoretical_loss": 3.627965357674884,
      "tokens_seen": 1063583744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034230692076228685,
      "loss": 2.6498,
      "theoretical_loss": 3.6279443258104234,
      "tokens_seen": 1063649280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422968906720161,
      "loss": 2.9348,
      "theoretical_loss": 3.627923295604603,
      "tokens_seen": 1063714816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034228686058174527,
      "loss": 2.6254,
      "theoretical_loss": 3.6279022670571885,
      "tokens_seen": 1063780352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034227683049147445,
      "loss": 2.7505,
      "theoretical_loss": 3.627881240167949,
      "tokens_seen": 1063845888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034226680040120363,
      "loss": 2.6287,
      "theoretical_loss": 3.62786021493665,
      "tokens_seen": 1063911424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422567703109328,
      "loss": 2.6029,
      "theoretical_loss": 3.6278391913630594,
      "tokens_seen": 1063976960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342246740220662,
      "loss": 2.8587,
      "theoretical_loss": 3.627818169446944,
      "tokens_seen": 1064042496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422367101303912,
      "loss": 2.7094,
      "theoretical_loss": 3.6277971491880714,
      "tokens_seen": 1064108032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034222668004012035,
      "loss": 2.5434,
      "theoretical_loss": 3.627776130586209,
      "tokens_seen": 1064173568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422166499498496,
      "loss": 2.9045,
      "theoretical_loss": 3.6277551136411237,
      "tokens_seen": 1064239104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422066198595787,
      "loss": 2.8314,
      "theoretical_loss": 3.6277340983525828,
      "tokens_seen": 1064304640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034219658976930795,
      "loss": 2.473,
      "theoretical_loss": 3.6277130847203543,
      "tokens_seen": 1064370176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034218655967903713,
      "loss": 2.6314,
      "theoretical_loss": 3.627692072744206,
      "tokens_seen": 1064435712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421765295887663,
      "loss": 2.4801,
      "theoretical_loss": 3.6276710624239037,
      "tokens_seen": 1064501248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421664994984955,
      "loss": 2.6896,
      "theoretical_loss": 3.627650053759217,
      "tokens_seen": 1064566784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034215646940822473,
      "loss": 2.8167,
      "theoretical_loss": 3.6276290467499126,
      "tokens_seen": 1064632320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034214643931795386,
      "loss": 2.6443,
      "theoretical_loss": 3.6276080413957583,
      "tokens_seen": 1064697856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421364092276831,
      "loss": 2.8321,
      "theoretical_loss": 3.6275870376965216,
      "tokens_seen": 1064763392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421263791374122,
      "loss": 2.511,
      "theoretical_loss": 3.62756603565197,
      "tokens_seen": 1064828928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034211634904714145,
      "loss": 2.6694,
      "theoretical_loss": 3.6275450352618726,
      "tokens_seen": 1064894464
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 594388,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.325084686279297,
      "objective/train/theoretical_loss": 3.6275240365259958,
      "objective/train/tokens_used": 1085420000,
      "theoretical_loss": 3.6275240365259958,
      "tokens_seen": 1064960000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034210631895687063,
      "loss": 2.5833,
      "theoretical_loss": 3.6275240365259958,
      "tokens_seen": 1064960000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420962888665998,
      "loss": 2.7116,
      "theoretical_loss": 3.6275030394441083,
      "tokens_seen": 1065025536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342086258776329,
      "loss": 2.5262,
      "theoretical_loss": 3.627482044015978,
      "tokens_seen": 1065091072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420762286860582,
      "loss": 2.6584,
      "theoretical_loss": 3.627461050241373,
      "tokens_seen": 1065156608
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034206619859578736,
      "loss": 2.6601,
      "theoretical_loss": 3.6274400581200608,
      "tokens_seen": 1065222144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420561685055166,
      "loss": 2.6521,
      "theoretical_loss": 3.62741906765181,
      "tokens_seen": 1065287680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420461384152457,
      "loss": 2.5302,
      "theoretical_loss": 3.6273980788363884,
      "tokens_seen": 1065353216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034203610832497496,
      "loss": 2.4866,
      "theoretical_loss": 3.6273770916735644,
      "tokens_seen": 1065418752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420260782347041,
      "loss": 2.614,
      "theoretical_loss": 3.6273561061631066,
      "tokens_seen": 1065484288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420160481444333,
      "loss": 2.713,
      "theoretical_loss": 3.6273351223047823,
      "tokens_seen": 1065549824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420060180541625,
      "loss": 2.6868,
      "theoretical_loss": 3.6273141400983606,
      "tokens_seen": 1065615360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419959879638917,
      "loss": 2.4458,
      "theoretical_loss": 3.6272931595436098,
      "tokens_seen": 1065680896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034198595787362086,
      "loss": 2.6685,
      "theoretical_loss": 3.6272721806402983,
      "tokens_seen": 1065746432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419759277833501,
      "loss": 2.6327,
      "theoretical_loss": 3.6272512033881945,
      "tokens_seen": 1065811968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419658976930792,
      "loss": 2.6649,
      "theoretical_loss": 3.6272302277870665,
      "tokens_seen": 1065877504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034195586760280846,
      "loss": 2.6261,
      "theoretical_loss": 3.627209253836684,
      "tokens_seen": 1065943040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419458375125376,
      "loss": 2.7778,
      "theoretical_loss": 3.627188281536814,
      "tokens_seen": 1066008576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419358074222668,
      "loss": 2.6386,
      "theoretical_loss": 3.627167310887227,
      "tokens_seen": 1066074112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341925777331996,
      "loss": 2.6872,
      "theoretical_loss": 3.6271463418876904,
      "tokens_seen": 1066139648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419157472417252,
      "loss": 2.7311,
      "theoretical_loss": 3.6271253745379726,
      "tokens_seen": 1066205184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034190571715145436,
      "loss": 2.6386,
      "theoretical_loss": 3.6271044088378437,
      "tokens_seen": 1066270720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034189568706118355,
      "loss": 2.5569,
      "theoretical_loss": 3.627083444787072,
      "tokens_seen": 1066336256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418856569709127,
      "loss": 2.5788,
      "theoretical_loss": 3.6270624823854263,
      "tokens_seen": 1066401792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034187562688064196,
      "loss": 2.4098,
      "theoretical_loss": 3.6270415216326755,
      "tokens_seen": 1066467328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418655967903711,
      "loss": 2.4781,
      "theoretical_loss": 3.6270205625285885,
      "tokens_seen": 1066532864
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 595571,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6795873641967773,
      "objective/train/theoretical_loss": 3.6269996050729345,
      "objective/train/tokens_used": 1087058400,
      "theoretical_loss": 3.6269996050729345,
      "tokens_seen": 1066598400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418555667001003,
      "loss": 2.613,
      "theoretical_loss": 3.6269996050729345,
      "tokens_seen": 1066598400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034184553660982945,
      "loss": 2.5394,
      "theoretical_loss": 3.6269786492654825,
      "tokens_seen": 1066663936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418355065195587,
      "loss": 2.3815,
      "theoretical_loss": 3.626957695106002,
      "tokens_seen": 1066729472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034182547642928787,
      "loss": 2.4471,
      "theoretical_loss": 3.6269367425942614,
      "tokens_seen": 1066795008
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034181544633901705,
      "loss": 2.6606,
      "theoretical_loss": 3.626915791730031,
      "tokens_seen": 1066860544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034180541624874623,
      "loss": 2.637,
      "theoretical_loss": 3.626894842513079,
      "tokens_seen": 1066926080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034179538615847547,
      "loss": 2.8521,
      "theoretical_loss": 3.6268738949431754,
      "tokens_seen": 1066991616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417853560682046,
      "loss": 2.5181,
      "theoretical_loss": 3.6268529490200887,
      "tokens_seen": 1067057152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034177532597793383,
      "loss": 2.6079,
      "theoretical_loss": 3.626832004743589,
      "tokens_seen": 1067122688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034176529588766295,
      "loss": 2.6194,
      "theoretical_loss": 3.626811062113446,
      "tokens_seen": 1067188224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417552657973922,
      "loss": 2.6109,
      "theoretical_loss": 3.626790121129429,
      "tokens_seen": 1067253760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034174523570712137,
      "loss": 2.6758,
      "theoretical_loss": 3.626769181791307,
      "tokens_seen": 1067319296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034173520561685055,
      "loss": 2.6405,
      "theoretical_loss": 3.62674824409885,
      "tokens_seen": 1067384832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034172517552657973,
      "loss": 2.7384,
      "theoretical_loss": 3.626727308051828,
      "tokens_seen": 1067450368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417151454363089,
      "loss": 2.7122,
      "theoretical_loss": 3.6267063736500096,
      "tokens_seen": 1067515904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417051153460381,
      "loss": 2.741,
      "theoretical_loss": 3.6266854408931657,
      "tokens_seen": 1067581440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034169508525576733,
      "loss": 2.5662,
      "theoretical_loss": 3.6266645097810652,
      "tokens_seen": 1067646976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034168505516549646,
      "loss": 2.4996,
      "theoretical_loss": 3.6266435803134787,
      "tokens_seen": 1067712512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416750250752257,
      "loss": 2.5729,
      "theoretical_loss": 3.6266226524901755,
      "tokens_seen": 1067778048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416649949849548,
      "loss": 2.738,
      "theoretical_loss": 3.6266017263109256,
      "tokens_seen": 1067843584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034165496489468406,
      "loss": 2.7026,
      "theoretical_loss": 3.626580801775499,
      "tokens_seen": 1067909120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034164493480441324,
      "loss": 2.3374,
      "theoretical_loss": 3.6265598788836657,
      "tokens_seen": 1067974656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416349047141424,
      "loss": 2.6919,
      "theoretical_loss": 3.626538957635196,
      "tokens_seen": 1068040192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416248746238716,
      "loss": 2.8102,
      "theoretical_loss": 3.62651803802986,
      "tokens_seen": 1068105728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034161484453360083,
      "loss": 2.5649,
      "theoretical_loss": 3.626497120067427,
      "tokens_seen": 1068171264
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 595858,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9211819171905518,
      "objective/train/theoretical_loss": 3.6264762037476683,
      "objective/train/tokens_used": 1088696800,
      "theoretical_loss": 3.6264762037476683,
      "tokens_seen": 1068236800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034160481444332996,
      "loss": 2.5103,
      "theoretical_loss": 3.6264762037476683,
      "tokens_seen": 1068236800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415947843530592,
      "loss": 2.6812,
      "theoretical_loss": 3.6264552890703534,
      "tokens_seen": 1068302336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415847542627884,
      "loss": 2.6637,
      "theoretical_loss": 3.6264343760352533,
      "tokens_seen": 1068367872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034157472417251756,
      "loss": 2.7339,
      "theoretical_loss": 3.6264134646421375,
      "tokens_seen": 1068433408
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415646940822468,
      "loss": 2.6223,
      "theoretical_loss": 3.6263925548907765,
      "tokens_seen": 1068498944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415546639919759,
      "loss": 2.6919,
      "theoretical_loss": 3.6263716467809415,
      "tokens_seen": 1068564480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034154463390170516,
      "loss": 2.7286,
      "theoretical_loss": 3.626350740312402,
      "tokens_seen": 1068630016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415346038114343,
      "loss": 2.5326,
      "theoretical_loss": 3.6263298354849294,
      "tokens_seen": 1068695552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415245737211635,
      "loss": 2.5669,
      "theoretical_loss": 3.6263089322982935,
      "tokens_seen": 1068761088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415145436308927,
      "loss": 2.656,
      "theoretical_loss": 3.6262880307522654,
      "tokens_seen": 1068826624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415045135406219,
      "loss": 2.5146,
      "theoretical_loss": 3.626267130846615,
      "tokens_seen": 1068892160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034149448345035106,
      "loss": 2.5452,
      "theoretical_loss": 3.626246232581114,
      "tokens_seen": 1068957696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414844533600803,
      "loss": 2.5444,
      "theoretical_loss": 3.6262253359555325,
      "tokens_seen": 1069023232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414744232698094,
      "loss": 2.507,
      "theoretical_loss": 3.6262044409696417,
      "tokens_seen": 1069088768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034146439317953866,
      "loss": 2.5204,
      "theoretical_loss": 3.6261835476232123,
      "tokens_seen": 1069154304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414543630892678,
      "loss": 2.6365,
      "theoretical_loss": 3.6261626559160147,
      "tokens_seen": 1069219840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341444332998997,
      "loss": 2.5105,
      "theoretical_loss": 3.6261417658478208,
      "tokens_seen": 1069285376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414343029087262,
      "loss": 2.7935,
      "theoretical_loss": 3.6261208774184004,
      "tokens_seen": 1069350912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414242728184554,
      "loss": 2.575,
      "theoretical_loss": 3.6260999906275257,
      "tokens_seen": 1069416448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034141424272818456,
      "loss": 2.5472,
      "theoretical_loss": 3.6260791054749673,
      "tokens_seen": 1069481984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034140421263791375,
      "loss": 2.5069,
      "theoretical_loss": 3.6260582219604958,
      "tokens_seen": 1069547520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034139418254764293,
      "loss": 2.7579,
      "theoretical_loss": 3.626037340083883,
      "tokens_seen": 1069613056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034138415245737216,
      "loss": 2.6508,
      "theoretical_loss": 3.6260164598448994,
      "tokens_seen": 1069678592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413741223671013,
      "loss": 2.7109,
      "theoretical_loss": 3.625995581243317,
      "tokens_seen": 1069744128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413640922768305,
      "loss": 2.6212,
      "theoretical_loss": 3.6259747042789074,
      "tokens_seen": 1069809664
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 597140,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7327358722686768,
      "objective/train/theoretical_loss": 3.625953828951441,
      "objective/train/tokens_used": 1090335200,
      "theoretical_loss": 3.625953828951441,
      "tokens_seen": 1069875200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034135406218655965,
      "loss": 2.6413,
      "theoretical_loss": 3.625953828951441,
      "tokens_seen": 1069875200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413440320962889,
      "loss": 2.6578,
      "theoretical_loss": 3.625932955260689,
      "tokens_seen": 1069940736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034133400200601807,
      "loss": 2.6887,
      "theoretical_loss": 3.625912083206424,
      "tokens_seen": 1070006272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034132397191574725,
      "loss": 2.3918,
      "theoretical_loss": 3.625891212788417,
      "tokens_seen": 1070071808
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034131394182547643,
      "loss": 2.8379,
      "theoretical_loss": 3.625870344006439,
      "tokens_seen": 1070137344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034130391173520567,
      "loss": 2.8266,
      "theoretical_loss": 3.625849476860262,
      "tokens_seen": 1070202880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412938816449348,
      "loss": 2.6852,
      "theoretical_loss": 3.6258286113496574,
      "tokens_seen": 1070268416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034128385155466403,
      "loss": 2.5112,
      "theoretical_loss": 3.6258077474743975,
      "tokens_seen": 1070333952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034127382146439315,
      "loss": 2.6649,
      "theoretical_loss": 3.625786885234253,
      "tokens_seen": 1070399488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412637913741224,
      "loss": 2.5514,
      "theoretical_loss": 3.625766024628997,
      "tokens_seen": 1070465024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034125376128385157,
      "loss": 2.7262,
      "theoretical_loss": 3.6257451656584,
      "tokens_seen": 1070530560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034124373119358075,
      "loss": 2.6269,
      "theoretical_loss": 3.6257243083222344,
      "tokens_seen": 1070596096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034123370110330993,
      "loss": 2.5139,
      "theoretical_loss": 3.6257034526202725,
      "tokens_seen": 1070661632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412236710130391,
      "loss": 2.7592,
      "theoretical_loss": 3.6256825985522854,
      "tokens_seen": 1070727168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412136409227683,
      "loss": 2.6409,
      "theoretical_loss": 3.6256617461180456,
      "tokens_seen": 1070792704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034120361083249753,
      "loss": 2.6633,
      "theoretical_loss": 3.625640895317325,
      "tokens_seen": 1070858240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034119358074222666,
      "loss": 2.5714,
      "theoretical_loss": 3.625620046149895,
      "tokens_seen": 1070923776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411835506519559,
      "loss": 2.6975,
      "theoretical_loss": 3.6255991986155296,
      "tokens_seen": 1070989312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341173520561685,
      "loss": 2.8864,
      "theoretical_loss": 3.625578352713999,
      "tokens_seen": 1071054848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034116349047141426,
      "loss": 2.8835,
      "theoretical_loss": 3.625557508445076,
      "tokens_seen": 1071120384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034115346038114344,
      "loss": 2.434,
      "theoretical_loss": 3.625536665808534,
      "tokens_seen": 1071185920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411434302908726,
      "loss": 2.7717,
      "theoretical_loss": 3.6255158248041433,
      "tokens_seen": 1071251456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411334002006018,
      "loss": 2.6301,
      "theoretical_loss": 3.6254949854316783,
      "tokens_seen": 1071316992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034112337011033103,
      "loss": 2.6367,
      "theoretical_loss": 3.62547414769091,
      "tokens_seen": 1071382528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034111334002006016,
      "loss": 2.5804,
      "theoretical_loss": 3.6254533115816105,
      "tokens_seen": 1071448064
    },
    {
      "epoch": 0.32,
      "objective/train/docs_used": 598183,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3538804054260254,
      "objective/train/theoretical_loss": 3.625432477103554,
      "objective/train/tokens_used": 1091973600,
      "theoretical_loss": 3.625432477103554,
      "tokens_seen": 1071513600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411033099297894,
      "loss": 2.3987,
      "theoretical_loss": 3.625432477103554,
      "tokens_seen": 1071513600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410932798395185,
      "loss": 2.6763,
      "theoretical_loss": 3.6254116442565114,
      "tokens_seen": 1071579136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034108324974924776,
      "loss": 2.8075,
      "theoretical_loss": 3.625390813040257,
      "tokens_seen": 1071644672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034107321965897694,
      "loss": 2.6814,
      "theoretical_loss": 3.625369983454561,
      "tokens_seen": 1071710208
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410631895687061,
      "loss": 2.6329,
      "theoretical_loss": 3.6253491554991983,
      "tokens_seen": 1071775744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410531594784353,
      "loss": 2.3829,
      "theoretical_loss": 3.62532832917394,
      "tokens_seen": 1071841280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410431293881645,
      "loss": 2.6071,
      "theoretical_loss": 3.6253075044785605,
      "tokens_seen": 1071906816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034103309929789366,
      "loss": 2.7892,
      "theoretical_loss": 3.625286681412831,
      "tokens_seen": 1071972352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410230692076229,
      "loss": 2.7847,
      "theoretical_loss": 3.6252658599765257,
      "tokens_seen": 1072037888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341013039117352,
      "loss": 2.6912,
      "theoretical_loss": 3.6252450401694167,
      "tokens_seen": 1072103424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034100300902708126,
      "loss": 2.5225,
      "theoretical_loss": 3.6252242219912767,
      "tokens_seen": 1072168960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034099297893681044,
      "loss": 2.6905,
      "theoretical_loss": 3.6252034054418796,
      "tokens_seen": 1072234496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003409829488465396,
      "loss": 2.7699,
      "theoretical_loss": 3.6251825905209976,
      "tokens_seen": 1072300032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003409729187562688,
      "loss": 2.4818,
      "theoretical_loss": 3.625161777228404,
      "tokens_seen": 1072365568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340962888665998,
      "loss": 2.8261,
      "theoretical_loss": 3.6251409655638724,
      "tokens_seen": 1072431104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034095285857572717,
      "loss": 2.7078,
      "theoretical_loss": 3.6251201555271755,
      "tokens_seen": 1072496640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409428284854564,
      "loss": 2.5445,
      "theoretical_loss": 3.625099347118087,
      "tokens_seen": 1072562176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034093279839518553,
      "loss": 2.6212,
      "theoretical_loss": 3.625078540336379,
      "tokens_seen": 1072627712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034092276830491477,
      "loss": 2.5129,
      "theoretical_loss": 3.625057735181826,
      "tokens_seen": 1072693248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409127382146439,
      "loss": 2.6918,
      "theoretical_loss": 3.6250369316542006,
      "tokens_seen": 1072758784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034090270812437313,
      "loss": 2.4667,
      "theoretical_loss": 3.625016129753277,
      "tokens_seen": 1072824320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408926780341023,
      "loss": 2.4542,
      "theoretical_loss": 3.6249953294788284,
      "tokens_seen": 1072889856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408826479438315,
      "loss": 2.6622,
      "theoretical_loss": 3.6249745308306274,
      "tokens_seen": 1072955392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034087261785356067,
      "loss": 2.7896,
      "theoretical_loss": 3.6249537338084488,
      "tokens_seen": 1073020928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034086258776328985,
      "loss": 2.7108,
      "theoretical_loss": 3.624932938412065,
      "tokens_seen": 1073086464
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 598781,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9608399868011475,
      "objective/train/theoretical_loss": 3.6249121446412507,
      "objective/train/tokens_used": 1093612000,
      "theoretical_loss": 3.6249121446412507,
      "tokens_seen": 1073152000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034085255767301903,
      "loss": 2.7379,
      "theoretical_loss": 3.6249121446412507,
      "tokens_seen": 1073152000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034084252758274827,
      "loss": 2.5467,
      "theoretical_loss": 3.624891352495779,
      "tokens_seen": 1073217536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034083249749247745,
      "loss": 2.3039,
      "theoretical_loss": 3.624870561975423,
      "tokens_seen": 1073283072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034082246740220663,
      "loss": 2.6401,
      "theoretical_loss": 3.6248497730799585,
      "tokens_seen": 1073348608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034081243731193587,
      "loss": 2.4763,
      "theoretical_loss": 3.624828985809157,
      "tokens_seen": 1073414144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340802407221665,
      "loss": 2.5679,
      "theoretical_loss": 3.6248082001627937,
      "tokens_seen": 1073479680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034079237713139423,
      "loss": 2.7953,
      "theoretical_loss": 3.6247874161406415,
      "tokens_seen": 1073545216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034078234704112335,
      "loss": 2.6214,
      "theoretical_loss": 3.6247666337424755,
      "tokens_seen": 1073610752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407723169508526,
      "loss": 2.6529,
      "theoretical_loss": 3.6247458529680694,
      "tokens_seen": 1073676288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034076228686058177,
      "loss": 2.7148,
      "theoretical_loss": 3.6247250738171966,
      "tokens_seen": 1073741824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034075225677031095,
      "loss": 2.9034,
      "theoretical_loss": 3.6247042962896314,
      "tokens_seen": 1073807360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034074222668004013,
      "loss": 2.6231,
      "theoretical_loss": 3.6246835203851484,
      "tokens_seen": 1073872896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407321965897693,
      "loss": 2.6436,
      "theoretical_loss": 3.624662746103521,
      "tokens_seen": 1073938432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407221664994985,
      "loss": 2.5244,
      "theoretical_loss": 3.6246419734445245,
      "tokens_seen": 1074003968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034071213640922773,
      "loss": 2.6201,
      "theoretical_loss": 3.624621202407932,
      "tokens_seen": 1074069504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034070210631895686,
      "loss": 2.5998,
      "theoretical_loss": 3.6246004329935184,
      "tokens_seen": 1074135040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406920762286861,
      "loss": 2.4807,
      "theoretical_loss": 3.6245796652010585,
      "tokens_seen": 1074200576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406820461384152,
      "loss": 2.7288,
      "theoretical_loss": 3.6245588990303252,
      "tokens_seen": 1074266112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034067201604814446,
      "loss": 2.4171,
      "theoretical_loss": 3.624538134481095,
      "tokens_seen": 1074331648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034066198595787364,
      "loss": 2.6719,
      "theoretical_loss": 3.6245173715531402,
      "tokens_seen": 1074397184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406519558676028,
      "loss": 2.6694,
      "theoretical_loss": 3.6244966102462373,
      "tokens_seen": 1074462720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340641925777332,
      "loss": 2.8719,
      "theoretical_loss": 3.6244758505601595,
      "tokens_seen": 1074528256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034063189568706123,
      "loss": 2.687,
      "theoretical_loss": 3.624455092494682,
      "tokens_seen": 1074593792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034062186559679036,
      "loss": 2.7869,
      "theoretical_loss": 3.624434336049579,
      "tokens_seen": 1074659328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406118355065196,
      "loss": 2.5128,
      "theoretical_loss": 3.6244135812246263,
      "tokens_seen": 1074724864
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 599993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5279252529144287,
      "objective/train/theoretical_loss": 3.6243928280195976,
      "objective/train/tokens_used": 1095250400,
      "theoretical_loss": 3.6243928280195976,
      "tokens_seen": 1074790400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406018054162487,
      "loss": 2.6169,
      "theoretical_loss": 3.6243928280195976,
      "tokens_seen": 1074790400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034059177532597796,
      "loss": 2.6747,
      "theoretical_loss": 3.624372076434268,
      "tokens_seen": 1074855936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034058174523570714,
      "loss": 2.7016,
      "theoretical_loss": 3.6243513264684126,
      "tokens_seen": 1074921472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405717151454363,
      "loss": 2.7665,
      "theoretical_loss": 3.624330578121806,
      "tokens_seen": 1074987008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405616850551655,
      "loss": 2.859,
      "theoretical_loss": 3.624309831394223,
      "tokens_seen": 1075052544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405516549648947,
      "loss": 2.8091,
      "theoretical_loss": 3.624289086285439,
      "tokens_seen": 1075118080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034054162487462386,
      "loss": 2.787,
      "theoretical_loss": 3.6242683427952285,
      "tokens_seen": 1075183616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405315947843531,
      "loss": 2.6744,
      "theoretical_loss": 3.6242476009233675,
      "tokens_seen": 1075249152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405215646940822,
      "loss": 2.6618,
      "theoretical_loss": 3.62422686066963,
      "tokens_seen": 1075314688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034051153460381146,
      "loss": 2.4714,
      "theoretical_loss": 3.6242061220337916,
      "tokens_seen": 1075380224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034050150451354064,
      "loss": 2.6045,
      "theoretical_loss": 3.624185385015628,
      "tokens_seen": 1075445760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404914744232698,
      "loss": 2.5883,
      "theoretical_loss": 3.6241646496149134,
      "tokens_seen": 1075511296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340481444332999,
      "loss": 2.7271,
      "theoretical_loss": 3.624143915831424,
      "tokens_seen": 1075576832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404714142427282,
      "loss": 2.7838,
      "theoretical_loss": 3.624123183664935,
      "tokens_seen": 1075642368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034046138415245737,
      "loss": 2.7861,
      "theoretical_loss": 3.6241024531152215,
      "tokens_seen": 1075707904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404513540621866,
      "loss": 2.7807,
      "theoretical_loss": 3.6240817241820587,
      "tokens_seen": 1075773440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034044132397191573,
      "loss": 2.7066,
      "theoretical_loss": 3.624060996865223,
      "tokens_seen": 1075838976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034043129388164497,
      "loss": 2.6692,
      "theoretical_loss": 3.624040271164489,
      "tokens_seen": 1075904512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404212637913741,
      "loss": 2.6054,
      "theoretical_loss": 3.624019547079633,
      "tokens_seen": 1075970048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034041123370110333,
      "loss": 2.7437,
      "theoretical_loss": 3.6239988246104295,
      "tokens_seen": 1076035584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404012036108325,
      "loss": 2.651,
      "theoretical_loss": 3.6239781037566554,
      "tokens_seen": 1076101120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403911735205617,
      "loss": 2.6356,
      "theoretical_loss": 3.6239573845180857,
      "tokens_seen": 1076166656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034038114343029087,
      "loss": 2.4081,
      "theoretical_loss": 3.6239366668944966,
      "tokens_seen": 1076232192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034037111334002005,
      "loss": 2.6507,
      "theoretical_loss": 3.6239159508856638,
      "tokens_seen": 1076297728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034036108324974923,
      "loss": 2.6173,
      "theoretical_loss": 3.6238952364913626,
      "tokens_seen": 1076363264
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 601033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.560241460800171,
      "objective/train/theoretical_loss": 3.623874523711369,
      "objective/train/tokens_used": 1096888800,
      "theoretical_loss": 3.623874523711369,
      "tokens_seen": 1076428800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034035105315947847,
      "loss": 2.4641,
      "theoretical_loss": 3.623874523711369,
      "tokens_seen": 1076428800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403410230692076,
      "loss": 2.9075,
      "theoretical_loss": 3.6238538125454594,
      "tokens_seen": 1076494336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034033099297893683,
      "loss": 2.7106,
      "theoretical_loss": 3.623833102993409,
      "tokens_seen": 1076559872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340320962888666,
      "loss": 2.5626,
      "theoretical_loss": 3.623812395054995,
      "tokens_seen": 1076625408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403109327983952,
      "loss": 2.667,
      "theoretical_loss": 3.6237916887299924,
      "tokens_seen": 1076690944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403009027081244,
      "loss": 2.4823,
      "theoretical_loss": 3.623770984018178,
      "tokens_seen": 1076756480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034029087261785356,
      "loss": 2.514,
      "theoretical_loss": 3.623750280919327,
      "tokens_seen": 1076822016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034028084252758274,
      "loss": 2.8244,
      "theoretical_loss": 3.623729579433217,
      "tokens_seen": 1076887552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034027081243731197,
      "loss": 2.6843,
      "theoretical_loss": 3.623708879559623,
      "tokens_seen": 1076953088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402607823470411,
      "loss": 2.663,
      "theoretical_loss": 3.6236881812983217,
      "tokens_seen": 1077018624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034025075225677033,
      "loss": 2.7929,
      "theoretical_loss": 3.62366748464909,
      "tokens_seen": 1077084160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034024072216649946,
      "loss": 2.5873,
      "theoretical_loss": 3.623646789611703,
      "tokens_seen": 1077149696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402306920762287,
      "loss": 2.5329,
      "theoretical_loss": 3.623626096185938,
      "tokens_seen": 1077215232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402206619859579,
      "loss": 2.8096,
      "theoretical_loss": 3.6236054043715713,
      "tokens_seen": 1077280768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034021063189568706,
      "loss": 2.4211,
      "theoretical_loss": 3.6235847141683797,
      "tokens_seen": 1077346304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034020060180541624,
      "loss": 2.6624,
      "theoretical_loss": 3.623564025576139,
      "tokens_seen": 1077411840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401905717151454,
      "loss": 2.9951,
      "theoretical_loss": 3.6235433385946267,
      "tokens_seen": 1077477376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401805416248746,
      "loss": 2.6661,
      "theoretical_loss": 3.6235226532236187,
      "tokens_seen": 1077542912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034017051153460384,
      "loss": 2.5833,
      "theoretical_loss": 3.623501969462892,
      "tokens_seen": 1077608448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034016048144433296,
      "loss": 2.7677,
      "theoretical_loss": 3.623481287312223,
      "tokens_seen": 1077673984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401504513540622,
      "loss": 2.5585,
      "theoretical_loss": 3.6234606067713893,
      "tokens_seen": 1077739520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401404212637914,
      "loss": 2.8325,
      "theoretical_loss": 3.6234399278401668,
      "tokens_seen": 1077805056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034013039117352056,
      "loss": 2.5578,
      "theoretical_loss": 3.6234192505183325,
      "tokens_seen": 1077870592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034012036108324974,
      "loss": 2.8148,
      "theoretical_loss": 3.623398574805664,
      "tokens_seen": 1077936128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401103309929789,
      "loss": 2.8187,
      "theoretical_loss": 3.6233779007019375,
      "tokens_seen": 1078001664
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 601484,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.654538869857788,
      "objective/train/theoretical_loss": 3.62335722820693,
      "objective/train/tokens_used": 1098527200,
      "theoretical_loss": 3.62335722820693,
      "tokens_seen": 1078067200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401003009027081,
      "loss": 2.5519,
      "theoretical_loss": 3.62335722820693,
      "tokens_seen": 1078067200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034009027081243734,
      "loss": 2.8683,
      "theoretical_loss": 3.6233365573204193,
      "tokens_seen": 1078132736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400802407221665,
      "loss": 2.7965,
      "theoretical_loss": 3.623315888042182,
      "tokens_seen": 1078198272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400702106318957,
      "loss": 2.8259,
      "theoretical_loss": 3.623295220371995,
      "tokens_seen": 1078263808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400601805416249,
      "loss": 2.5405,
      "theoretical_loss": 3.623274554309636,
      "tokens_seen": 1078329344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034005015045135406,
      "loss": 2.4842,
      "theoretical_loss": 3.6232538898548814,
      "tokens_seen": 1078394880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400401203610833,
      "loss": 2.6957,
      "theoretical_loss": 3.623233227007509,
      "tokens_seen": 1078460416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400300902708124,
      "loss": 2.8692,
      "theoretical_loss": 3.6232125657672962,
      "tokens_seen": 1078525952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034002006018054166,
      "loss": 2.7149,
      "theoretical_loss": 3.6231919061340205,
      "tokens_seen": 1078591488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034001003009027084,
      "loss": 2.6374,
      "theoretical_loss": 3.6231712481074583,
      "tokens_seen": 1078657024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034,
      "loss": 2.5463,
      "theoretical_loss": 3.6231505916873887,
      "tokens_seen": 1078722560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399899699097292,
      "loss": 2.6222,
      "theoretical_loss": 3.623129936873587,
      "tokens_seen": 1078788096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399799398194584,
      "loss": 2.9549,
      "theoretical_loss": 3.623109283665833,
      "tokens_seen": 1078853632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033996990972918757,
      "loss": 2.5965,
      "theoretical_loss": 3.6230886320639026,
      "tokens_seen": 1078919168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399598796389168,
      "loss": 2.7922,
      "theoretical_loss": 3.6230679820675746,
      "tokens_seen": 1078984704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033994984954864593,
      "loss": 2.7641,
      "theoretical_loss": 3.6230473336766256,
      "tokens_seen": 1079050240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033993981945837517,
      "loss": 2.5767,
      "theoretical_loss": 3.6230266868908343,
      "tokens_seen": 1079115776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399297893681043,
      "loss": 2.7457,
      "theoretical_loss": 3.6230060417099774,
      "tokens_seen": 1079181312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033991975927783353,
      "loss": 2.7767,
      "theoretical_loss": 3.6229853981338334,
      "tokens_seen": 1079246848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399097291875627,
      "loss": 2.5381,
      "theoretical_loss": 3.62296475616218,
      "tokens_seen": 1079312384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398996990972919,
      "loss": 2.7448,
      "theoretical_loss": 3.622944115794795,
      "tokens_seen": 1079377920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033988966900702107,
      "loss": 2.8126,
      "theoretical_loss": 3.6229234770314562,
      "tokens_seen": 1079443456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033987963891675025,
      "loss": 2.7071,
      "theoretical_loss": 3.622902839871942,
      "tokens_seen": 1079508992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033986960882647943,
      "loss": 2.8773,
      "theoretical_loss": 3.6228822043160305,
      "tokens_seen": 1079574528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033985957873620867,
      "loss": 2.6932,
      "theoretical_loss": 3.622861570363499,
      "tokens_seen": 1079640064
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 602559,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.677448034286499,
      "objective/train/theoretical_loss": 3.622840938014126,
      "objective/train/tokens_used": 1100165600,
      "theoretical_loss": 3.622840938014126,
      "tokens_seen": 1079705600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398495486459378,
      "loss": 2.611,
      "theoretical_loss": 3.622840938014126,
      "tokens_seen": 1079705600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033983951855566703,
      "loss": 2.4328,
      "theoretical_loss": 3.6228203072676894,
      "tokens_seen": 1079771136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398294884653962,
      "loss": 2.6645,
      "theoretical_loss": 3.622799678123968,
      "tokens_seen": 1079836672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398194583751254,
      "loss": 2.8604,
      "theoretical_loss": 3.6227790505827393,
      "tokens_seen": 1079902208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398094282848546,
      "loss": 2.6952,
      "theoretical_loss": 3.6227584246437825,
      "tokens_seen": 1079967744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033979939819458376,
      "loss": 2.6203,
      "theoretical_loss": 3.622737800306875,
      "tokens_seen": 1080033280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033978936810431294,
      "loss": 2.6824,
      "theoretical_loss": 3.6227171775717952,
      "tokens_seen": 1080098816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033977933801404217,
      "loss": 2.4892,
      "theoretical_loss": 3.6226965564383224,
      "tokens_seen": 1080164352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397693079237713,
      "loss": 2.8857,
      "theoretical_loss": 3.6226759369062345,
      "tokens_seen": 1080229888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033975927783350053,
      "loss": 2.5533,
      "theoretical_loss": 3.62265531897531,
      "tokens_seen": 1080295424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033974924774322966,
      "loss": 2.7272,
      "theoretical_loss": 3.6226347026453274,
      "tokens_seen": 1080360960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397392176529589,
      "loss": 2.8356,
      "theoretical_loss": 3.622614087916065,
      "tokens_seen": 1080426496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397291875626881,
      "loss": 2.6642,
      "theoretical_loss": 3.622593474787302,
      "tokens_seen": 1080492032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033971915747241726,
      "loss": 2.7828,
      "theoretical_loss": 3.622572863258817,
      "tokens_seen": 1080557568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033970912738214644,
      "loss": 2.7391,
      "theoretical_loss": 3.6225522533303884,
      "tokens_seen": 1080623104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396990972918756,
      "loss": 2.7628,
      "theoretical_loss": 3.622531645001795,
      "tokens_seen": 1080688640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396890672016048,
      "loss": 2.6882,
      "theoretical_loss": 3.6225110382728163,
      "tokens_seen": 1080754176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033967903711133404,
      "loss": 2.7996,
      "theoretical_loss": 3.6224904331432297,
      "tokens_seen": 1080819712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033966900702106316,
      "loss": 2.74,
      "theoretical_loss": 3.6224698296128155,
      "tokens_seen": 1080885248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396589769307924,
      "loss": 2.9426,
      "theoretical_loss": 3.6224492276813516,
      "tokens_seen": 1080950784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396489468405216,
      "loss": 2.6184,
      "theoretical_loss": 3.622428627348618,
      "tokens_seen": 1081016320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033963891675025076,
      "loss": 2.6755,
      "theoretical_loss": 3.6224080286143936,
      "tokens_seen": 1081081856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033962888665997994,
      "loss": 2.8268,
      "theoretical_loss": 3.6223874314784563,
      "tokens_seen": 1081147392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396188565697091,
      "loss": 2.6844,
      "theoretical_loss": 3.622366835940586,
      "tokens_seen": 1081212928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396088264794383,
      "loss": 2.6764,
      "theoretical_loss": 3.622346242000562,
      "tokens_seen": 1081278464
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 603206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.54484224319458,
      "objective/train/theoretical_loss": 3.6223256496581637,
      "objective/train/tokens_used": 1101804000,
      "theoretical_loss": 3.6223256496581637,
      "tokens_seen": 1081344000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033959879638916754,
      "loss": 2.7759,
      "theoretical_loss": 3.6223256496581637,
      "tokens_seen": 1081344000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033958876629889667,
      "loss": 2.7449,
      "theoretical_loss": 3.62230505891317,
      "tokens_seen": 1081409536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395787362086259,
      "loss": 2.8069,
      "theoretical_loss": 3.6222844697653596,
      "tokens_seen": 1081475072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033956870611835503,
      "loss": 2.567,
      "theoretical_loss": 3.6222638822145132,
      "tokens_seen": 1081540608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033955867602808426,
      "loss": 2.7481,
      "theoretical_loss": 3.6222432962604088,
      "tokens_seen": 1081606144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033954864593781345,
      "loss": 2.5326,
      "theoretical_loss": 3.6222227119028263,
      "tokens_seen": 1081671680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395386158475426,
      "loss": 2.5933,
      "theoretical_loss": 3.622202129141546,
      "tokens_seen": 1081737216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395285857572718,
      "loss": 2.7438,
      "theoretical_loss": 3.6221815479763464,
      "tokens_seen": 1081802752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033951855566700104,
      "loss": 2.4687,
      "theoretical_loss": 3.622160968407007,
      "tokens_seen": 1081868288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033950852557673017,
      "loss": 2.7814,
      "theoretical_loss": 3.622140390433308,
      "tokens_seen": 1081933824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394984954864594,
      "loss": 2.5706,
      "theoretical_loss": 3.622119814055029,
      "tokens_seen": 1081999360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033948846539618853,
      "loss": 2.7052,
      "theoretical_loss": 3.6220992392719493,
      "tokens_seen": 1082064896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033947843530591777,
      "loss": 2.5586,
      "theoretical_loss": 3.6220786660838487,
      "tokens_seen": 1082130432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033946840521564695,
      "loss": 2.676,
      "theoretical_loss": 3.622058094490507,
      "tokens_seen": 1082195968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033945837512537613,
      "loss": 2.8831,
      "theoretical_loss": 3.622037524491704,
      "tokens_seen": 1082261504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394483450351053,
      "loss": 2.6192,
      "theoretical_loss": 3.6220169560872204,
      "tokens_seen": 1082327040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394383149448345,
      "loss": 2.6554,
      "theoretical_loss": 3.6219963892768345,
      "tokens_seen": 1082392576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394282848545637,
      "loss": 2.5606,
      "theoretical_loss": 3.6219758240603275,
      "tokens_seen": 1082458112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394182547642929,
      "loss": 2.7442,
      "theoretical_loss": 3.6219552604374785,
      "tokens_seen": 1082523648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033940822467402204,
      "loss": 2.5597,
      "theoretical_loss": 3.6219346984080687,
      "tokens_seen": 1082589184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033939819458375127,
      "loss": 2.7137,
      "theoretical_loss": 3.621914137971877,
      "tokens_seen": 1082654720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393881644934804,
      "loss": 2.8618,
      "theoretical_loss": 3.6218935791286846,
      "tokens_seen": 1082720256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033937813440320963,
      "loss": 2.5351,
      "theoretical_loss": 3.6218730218782706,
      "tokens_seen": 1082785792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393681043129388,
      "loss": 2.7446,
      "theoretical_loss": 3.6218524662204157,
      "tokens_seen": 1082851328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339358074222668,
      "loss": 2.6969,
      "theoretical_loss": 3.6218319121549003,
      "tokens_seen": 1082916864
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 603451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.421107530593872,
      "objective/train/theoretical_loss": 3.6218113596815042,
      "objective/train/tokens_used": 1103442400,
      "theoretical_loss": 3.6218113596815042,
      "tokens_seen": 1082982400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393480441323972,
      "loss": 2.5057,
      "theoretical_loss": 3.6218113596815042,
      "tokens_seen": 1082982400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393380140421264,
      "loss": 2.6102,
      "theoretical_loss": 3.6217908088000086,
      "tokens_seen": 1083047936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393279839518556,
      "loss": 2.8188,
      "theoretical_loss": 3.6217702595101935,
      "tokens_seen": 1083113472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393179538615848,
      "loss": 2.6628,
      "theoretical_loss": 3.6217497118118382,
      "tokens_seen": 1083179008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033930792377131396,
      "loss": 2.6231,
      "theoretical_loss": 3.621729165704725,
      "tokens_seen": 1083244544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033929789368104314,
      "loss": 2.3983,
      "theoretical_loss": 3.6217086211886333,
      "tokens_seen": 1083310080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033928786359077237,
      "loss": 2.6967,
      "theoretical_loss": 3.621688078263344,
      "tokens_seen": 1083375616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392778335005015,
      "loss": 2.74,
      "theoretical_loss": 3.6216675369286375,
      "tokens_seen": 1083441152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033926780341023073,
      "loss": 2.7629,
      "theoretical_loss": 3.621646997184295,
      "tokens_seen": 1083506688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033925777331995986,
      "loss": 2.6403,
      "theoretical_loss": 3.6216264590300966,
      "tokens_seen": 1083572224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392477432296891,
      "loss": 2.8355,
      "theoretical_loss": 3.6216059224658226,
      "tokens_seen": 1083637760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392377131394183,
      "loss": 2.7227,
      "theoretical_loss": 3.6215853874912547,
      "tokens_seen": 1083703296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033922768304914746,
      "loss": 2.7424,
      "theoretical_loss": 3.6215648541061736,
      "tokens_seen": 1083768832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033921765295887664,
      "loss": 2.6606,
      "theoretical_loss": 3.62154432231036,
      "tokens_seen": 1083834368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392076228686058,
      "loss": 2.7373,
      "theoretical_loss": 3.6215237921035945,
      "tokens_seen": 1083899904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339197592778335,
      "loss": 2.5256,
      "theoretical_loss": 3.6215032634856583,
      "tokens_seen": 1083965440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033918756268806424,
      "loss": 2.7521,
      "theoretical_loss": 3.6214827364563327,
      "tokens_seen": 1084030976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033917753259779336,
      "loss": 2.7552,
      "theoretical_loss": 3.6214622110153982,
      "tokens_seen": 1084096512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391675025075226,
      "loss": 2.5513,
      "theoretical_loss": 3.621441687162636,
      "tokens_seen": 1084162048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391574724172518,
      "loss": 2.7222,
      "theoretical_loss": 3.6214211648978276,
      "tokens_seen": 1084227584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033914744232698096,
      "loss": 2.6685,
      "theoretical_loss": 3.6214006442207536,
      "tokens_seen": 1084293120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033913741223671014,
      "loss": 2.5153,
      "theoretical_loss": 3.6213801251311954,
      "tokens_seen": 1084358656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391273821464393,
      "loss": 2.5689,
      "theoretical_loss": 3.6213596076289347,
      "tokens_seen": 1084424192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391173520561685,
      "loss": 2.7692,
      "theoretical_loss": 3.6213390917137525,
      "tokens_seen": 1084489728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033910732196589774,
      "loss": 2.5669,
      "theoretical_loss": 3.6213185773854297,
      "tokens_seen": 1084555264
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 603451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2022194862365723,
      "objective/train/theoretical_loss": 3.6212980646437485,
      "objective/train/tokens_used": 1105080800,
      "theoretical_loss": 3.6212980646437485,
      "tokens_seen": 1084620800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033909729187562687,
      "loss": 2.8285,
      "theoretical_loss": 3.6212980646437485,
      "tokens_seen": 1084620800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003390872617853561,
      "loss": 2.5917,
      "theoretical_loss": 3.62127755348849,
      "tokens_seen": 1084686336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033907723169508523,
      "loss": 2.5659,
      "theoretical_loss": 3.621257043919435,
      "tokens_seen": 1084751872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033906720160481446,
      "loss": 2.8947,
      "theoretical_loss": 3.621236535936366,
      "tokens_seen": 1084817408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033905717151454365,
      "loss": 2.7032,
      "theoretical_loss": 3.621216029539064,
      "tokens_seen": 1084882944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033904714142427283,
      "loss": 2.854,
      "theoretical_loss": 3.6211955247273107,
      "tokens_seen": 1084948480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339037111334002,
      "loss": 2.9125,
      "theoretical_loss": 3.6211750215008878,
      "tokens_seen": 1085014016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033902708124373124,
      "loss": 2.6352,
      "theoretical_loss": 3.6211545198595774,
      "tokens_seen": 1085079552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033901705115346037,
      "loss": 2.7928,
      "theoretical_loss": 3.62113401980316,
      "tokens_seen": 1085145088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003390070210631896,
      "loss": 2.8488,
      "theoretical_loss": 3.6211135213314183,
      "tokens_seen": 1085210624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033899699097291873,
      "loss": 2.7162,
      "theoretical_loss": 3.621093024444135,
      "tokens_seen": 1085276160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033898696088264797,
      "loss": 2.9749,
      "theoretical_loss": 3.62107252914109,
      "tokens_seen": 1085341696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033897693079237715,
      "loss": 2.8548,
      "theoretical_loss": 3.621052035422066,
      "tokens_seen": 1085407232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033896690070210633,
      "loss": 2.6953,
      "theoretical_loss": 3.621031543286845,
      "tokens_seen": 1085472768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389568706118355,
      "loss": 2.8561,
      "theoretical_loss": 3.62101105273521,
      "tokens_seen": 1085538304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389468405215647,
      "loss": 2.7889,
      "theoretical_loss": 3.6209905637669415,
      "tokens_seen": 1085603840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389368104312939,
      "loss": 2.7944,
      "theoretical_loss": 3.6209700763818224,
      "tokens_seen": 1085669376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389267803410231,
      "loss": 2.4984,
      "theoretical_loss": 3.6209495905796345,
      "tokens_seen": 1085734912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033891675025075224,
      "loss": 2.8853,
      "theoretical_loss": 3.62092910636016,
      "tokens_seen": 1085800448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033890672016048147,
      "loss": 2.8857,
      "theoretical_loss": 3.6209086237231816,
      "tokens_seen": 1085865984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388966900702106,
      "loss": 2.8294,
      "theoretical_loss": 3.6208881426684805,
      "tokens_seen": 1085931520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033888665997993983,
      "loss": 2.6671,
      "theoretical_loss": 3.6208676631958396,
      "tokens_seen": 1085997056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338876629889669,
      "loss": 2.7934,
      "theoretical_loss": 3.6208471853050415,
      "tokens_seen": 1086062592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388665997993982,
      "loss": 2.6889,
      "theoretical_loss": 3.6208267089958683,
      "tokens_seen": 1086128128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388565697091274,
      "loss": 2.6472,
      "theoretical_loss": 3.6208062342681027,
      "tokens_seen": 1086193664
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 604042,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7982475757598877,
      "objective/train/theoretical_loss": 3.6207857611215264,
      "objective/train/tokens_used": 1106719200,
      "theoretical_loss": 3.6207857611215264,
      "tokens_seen": 1086259200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388465396188566,
      "loss": 2.6263,
      "theoretical_loss": 3.6207857611215264,
      "tokens_seen": 1086259200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033883650952858574,
      "loss": 2.8603,
      "theoretical_loss": 3.6207652895559224,
      "tokens_seen": 1086324736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338826479438315,
      "loss": 2.5424,
      "theoretical_loss": 3.6207448195710734,
      "tokens_seen": 1086390272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388164493480441,
      "loss": 2.5165,
      "theoretical_loss": 3.620724351166762,
      "tokens_seen": 1086455808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033880641925777334,
      "loss": 2.7872,
      "theoretical_loss": 3.6207038843427704,
      "tokens_seen": 1086521344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387963891675025,
      "loss": 2.4773,
      "theoretical_loss": 3.6206834190988815,
      "tokens_seen": 1086586880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387863590772317,
      "loss": 2.7559,
      "theoretical_loss": 3.620662955434878,
      "tokens_seen": 1086652416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387763289869609,
      "loss": 2.975,
      "theoretical_loss": 3.6206424933505437,
      "tokens_seen": 1086717952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033876629889669006,
      "loss": 2.9142,
      "theoretical_loss": 3.6206220328456595,
      "tokens_seen": 1086783488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033875626880641924,
      "loss": 2.8429,
      "theoretical_loss": 3.620601573920009,
      "tokens_seen": 1086849024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387462387161485,
      "loss": 2.6618,
      "theoretical_loss": 3.620581116573376,
      "tokens_seen": 1086914560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387362086258776,
      "loss": 2.7312,
      "theoretical_loss": 3.620560660805542,
      "tokens_seen": 1086980096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033872617853560684,
      "loss": 2.5756,
      "theoretical_loss": 3.6205402066162913,
      "tokens_seen": 1087045632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033871614844533597,
      "loss": 2.7202,
      "theoretical_loss": 3.6205197540054064,
      "tokens_seen": 1087111168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387061183550652,
      "loss": 2.7871,
      "theoretical_loss": 3.6204993029726698,
      "tokens_seen": 1087176704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386960882647944,
      "loss": 2.8353,
      "theoretical_loss": 3.6204788535178656,
      "tokens_seen": 1087242240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033868605817452356,
      "loss": 2.6144,
      "theoretical_loss": 3.620458405640776,
      "tokens_seen": 1087307776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033867602808425275,
      "loss": 2.8865,
      "theoretical_loss": 3.6204379593411846,
      "tokens_seen": 1087373312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338665997993982,
      "loss": 2.8391,
      "theoretical_loss": 3.6204175146188744,
      "tokens_seen": 1087438848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386559679037111,
      "loss": 2.7494,
      "theoretical_loss": 3.6203970714736298,
      "tokens_seen": 1087504384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033864593781344034,
      "loss": 2.806,
      "theoretical_loss": 3.6203766299052327,
      "tokens_seen": 1087569920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033863590772316947,
      "loss": 2.6098,
      "theoretical_loss": 3.6203561899134673,
      "tokens_seen": 1087635456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386258776328987,
      "loss": 3.0417,
      "theoretical_loss": 3.620335751498116,
      "tokens_seen": 1087700992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386158475426279,
      "loss": 2.9476,
      "theoretical_loss": 3.6203153146589635,
      "tokens_seen": 1087766528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033860581745235707,
      "loss": 2.6936,
      "theoretical_loss": 3.620294879395793,
      "tokens_seen": 1087832064
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 604895,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.194870710372925,
      "objective/train/theoretical_loss": 3.6202744457083877,
      "objective/train/tokens_used": 1108357600,
      "theoretical_loss": 3.6202744457083877,
      "tokens_seen": 1087897600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033859578736208625,
      "loss": 2.5966,
      "theoretical_loss": 3.6202744457083877,
      "tokens_seen": 1087897600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033858575727181543,
      "loss": 2.8116,
      "theoretical_loss": 3.620254013596531,
      "tokens_seen": 1087963136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033857572718154467,
      "loss": 2.8302,
      "theoretical_loss": 3.620233583060007,
      "tokens_seen": 1088028672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033856569709127385,
      "loss": 2.77,
      "theoretical_loss": 3.6202131540985993,
      "tokens_seen": 1088094208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033855566700100303,
      "loss": 2.8116,
      "theoretical_loss": 3.6201927267120917,
      "tokens_seen": 1088159744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385456369107322,
      "loss": 2.9741,
      "theoretical_loss": 3.6201723009002675,
      "tokens_seen": 1088225280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033853560682046144,
      "loss": 2.8073,
      "theoretical_loss": 3.62015187666291,
      "tokens_seen": 1088290816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033852557673019057,
      "loss": 2.5589,
      "theoretical_loss": 3.6201314539998046,
      "tokens_seen": 1088356352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385155466399198,
      "loss": 2.812,
      "theoretical_loss": 3.620111032910734,
      "tokens_seen": 1088421888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033850551654964893,
      "loss": 2.8191,
      "theoretical_loss": 3.6200906133954827,
      "tokens_seen": 1088487424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033849548645937817,
      "loss": 2.9442,
      "theoretical_loss": 3.620070195453834,
      "tokens_seen": 1088552960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033848545636910735,
      "loss": 2.7046,
      "theoretical_loss": 3.620049779085573,
      "tokens_seen": 1088618496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033847542627883653,
      "loss": 2.8518,
      "theoretical_loss": 3.6200293642904824,
      "tokens_seen": 1088684032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384653961885657,
      "loss": 2.8441,
      "theoretical_loss": 3.620008951068347,
      "tokens_seen": 1088749568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384553660982949,
      "loss": 2.6895,
      "theoretical_loss": 3.6199885394189515,
      "tokens_seen": 1088815104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384453360080241,
      "loss": 2.9414,
      "theoretical_loss": 3.619968129342079,
      "tokens_seen": 1088880640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384353059177533,
      "loss": 2.9459,
      "theoretical_loss": 3.619947720837514,
      "tokens_seen": 1088946176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033842527582748244,
      "loss": 2.7135,
      "theoretical_loss": 3.619927313905041,
      "tokens_seen": 1089011712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033841524573721167,
      "loss": 2.7475,
      "theoretical_loss": 3.6199069085444444,
      "tokens_seen": 1089077248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384052156469408,
      "loss": 2.5889,
      "theoretical_loss": 3.6198865047555087,
      "tokens_seen": 1089142784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033839518555667003,
      "loss": 2.7581,
      "theoretical_loss": 3.6198661025380177,
      "tokens_seen": 1089208320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383851554663992,
      "loss": 2.7403,
      "theoretical_loss": 3.619845701891756,
      "tokens_seen": 1089273856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383751253761284,
      "loss": 2.6924,
      "theoretical_loss": 3.6198253028165084,
      "tokens_seen": 1089339392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383650952858576,
      "loss": 2.7452,
      "theoretical_loss": 3.6198049053120593,
      "tokens_seen": 1089404928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383550651955868,
      "loss": 2.8973,
      "theoretical_loss": 3.6197845093781926,
      "tokens_seen": 1089470464
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 606118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.570420980453491,
      "objective/train/theoretical_loss": 3.6197641150146938,
      "objective/train/tokens_used": 1109996000,
      "theoretical_loss": 3.6197641150146938,
      "tokens_seen": 1089536000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033834503510531594,
      "loss": 2.6581,
      "theoretical_loss": 3.6197641150146938,
      "tokens_seen": 1089536000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383350050150452,
      "loss": 2.684,
      "theoretical_loss": 3.6197437222213473,
      "tokens_seen": 1089601536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383249749247743,
      "loss": 2.8388,
      "theoretical_loss": 3.6197233309979375,
      "tokens_seen": 1089667072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033831494483450354,
      "loss": 2.7701,
      "theoretical_loss": 3.61970294134425,
      "tokens_seen": 1089732608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383049147442327,
      "loss": 2.7718,
      "theoretical_loss": 3.619682553260068,
      "tokens_seen": 1089798144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382948846539619,
      "loss": 2.6133,
      "theoretical_loss": 3.619662166745177,
      "tokens_seen": 1089863680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382848545636911,
      "loss": 2.8218,
      "theoretical_loss": 3.6196417817993627,
      "tokens_seen": 1089929216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033827482447342026,
      "loss": 2.9036,
      "theoretical_loss": 3.6196213984224093,
      "tokens_seen": 1089994752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033826479438314944,
      "loss": 2.937,
      "theoretical_loss": 3.619601016614102,
      "tokens_seen": 1090060288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382547642928787,
      "loss": 2.7739,
      "theoretical_loss": 3.6195806363742253,
      "tokens_seen": 1090125824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382447342026078,
      "loss": 2.7651,
      "theoretical_loss": 3.619560257702564,
      "tokens_seen": 1090191360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033823470411233704,
      "loss": 2.6242,
      "theoretical_loss": 3.6195398805989045,
      "tokens_seen": 1090256896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033822467402206617,
      "loss": 2.9022,
      "theoretical_loss": 3.619519505063031,
      "tokens_seen": 1090322432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382146439317954,
      "loss": 2.6241,
      "theoretical_loss": 3.619499131094728,
      "tokens_seen": 1090387968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382046138415246,
      "loss": 2.6972,
      "theoretical_loss": 3.6194787586937824,
      "tokens_seen": 1090453504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033819458375125376,
      "loss": 2.8839,
      "theoretical_loss": 3.6194583878599778,
      "tokens_seen": 1090519040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033818455366098295,
      "loss": 2.8666,
      "theoretical_loss": 3.6194380185931,
      "tokens_seen": 1090584576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381745235707122,
      "loss": 2.7848,
      "theoretical_loss": 3.6194176508929354,
      "tokens_seen": 1090650112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381644934804413,
      "loss": 2.8002,
      "theoretical_loss": 3.619397284759268,
      "tokens_seen": 1090715648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033815446339017054,
      "loss": 2.8289,
      "theoretical_loss": 3.6193769201918835,
      "tokens_seen": 1090781184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033814443329989967,
      "loss": 2.8468,
      "theoretical_loss": 3.6193565571905673,
      "tokens_seen": 1090846720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381344032096289,
      "loss": 2.9158,
      "theoretical_loss": 3.619336195755105,
      "tokens_seen": 1090912256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381243731193581,
      "loss": 2.8524,
      "theoretical_loss": 3.619315835885283,
      "tokens_seen": 1090977792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033811434302908727,
      "loss": 2.6606,
      "theoretical_loss": 3.619295477580885,
      "tokens_seen": 1091043328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033810431293881645,
      "loss": 3.0112,
      "theoretical_loss": 3.6192751208416984,
      "tokens_seen": 1091108864
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 606665,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6309762001037598,
      "objective/train/theoretical_loss": 3.6192547656675083,
      "objective/train/tokens_used": 1111634400,
      "theoretical_loss": 3.6192547656675083,
      "tokens_seen": 1091174400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033809428284854563,
      "loss": 2.6641,
      "theoretical_loss": 3.6192547656675083,
      "tokens_seen": 1091174400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380842527582748,
      "loss": 2.6976,
      "theoretical_loss": 3.6192344120581,
      "tokens_seen": 1091239936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033807422266800405,
      "loss": 2.7123,
      "theoretical_loss": 3.6192140600132596,
      "tokens_seen": 1091305472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380641925777332,
      "loss": 2.8793,
      "theoretical_loss": 3.619193709532773,
      "tokens_seen": 1091371008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380541624874624,
      "loss": 2.6522,
      "theoretical_loss": 3.6191733606164256,
      "tokens_seen": 1091436544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033804413239719154,
      "loss": 2.8821,
      "theoretical_loss": 3.619153013264003,
      "tokens_seen": 1091502080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033803410230692077,
      "loss": 2.7992,
      "theoretical_loss": 3.6191326674752924,
      "tokens_seen": 1091567616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033802407221664995,
      "loss": 2.979,
      "theoretical_loss": 3.6191123232500786,
      "tokens_seen": 1091633152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033801404212637913,
      "loss": 2.7271,
      "theoretical_loss": 3.6190919805881485,
      "tokens_seen": 1091698688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380040120361083,
      "loss": 2.8355,
      "theoretical_loss": 3.6190716394892872,
      "tokens_seen": 1091764224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033799398194583755,
      "loss": 2.7256,
      "theoretical_loss": 3.6190512999532807,
      "tokens_seen": 1091829760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379839518555667,
      "loss": 2.6524,
      "theoretical_loss": 3.6190309619799166,
      "tokens_seen": 1091895296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379739217652959,
      "loss": 2.8708,
      "theoretical_loss": 3.619010625568979,
      "tokens_seen": 1091960832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033796389167502504,
      "loss": 2.7305,
      "theoretical_loss": 3.618990290720256,
      "tokens_seen": 1092026368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379538615847543,
      "loss": 2.7191,
      "theoretical_loss": 3.618969957433533,
      "tokens_seen": 1092091904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033794383149448346,
      "loss": 2.8005,
      "theoretical_loss": 3.618949625708596,
      "tokens_seen": 1092157440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033793380140421264,
      "loss": 2.7105,
      "theoretical_loss": 3.6189292955452315,
      "tokens_seen": 1092222976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379237713139418,
      "loss": 2.79,
      "theoretical_loss": 3.6189089669432266,
      "tokens_seen": 1092288512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337913741223671,
      "loss": 3.0197,
      "theoretical_loss": 3.6188886399023668,
      "tokens_seen": 1092354048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379037111334002,
      "loss": 2.8093,
      "theoretical_loss": 3.618868314422439,
      "tokens_seen": 1092419584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378936810431294,
      "loss": 2.9259,
      "theoretical_loss": 3.61884799050323,
      "tokens_seen": 1092485120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033788365095285854,
      "loss": 2.807,
      "theoretical_loss": 3.6188276681445255,
      "tokens_seen": 1092550656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378736208625878,
      "loss": 2.4255,
      "theoretical_loss": 3.6188073473461126,
      "tokens_seen": 1092616192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378635907723169,
      "loss": 2.679,
      "theoretical_loss": 3.6187870281077785,
      "tokens_seen": 1092681728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033785356068204614,
      "loss": 2.7072,
      "theoretical_loss": 3.618766710429308,
      "tokens_seen": 1092747264
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 607551,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.578968048095703,
      "objective/train/theoretical_loss": 3.6187463943104903,
      "objective/train/tokens_used": 1113272800,
      "theoretical_loss": 3.6187463943104903,
      "tokens_seen": 1092812800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378435305917753,
      "loss": 2.717,
      "theoretical_loss": 3.6187463943104903,
      "tokens_seen": 1092812800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378335005015045,
      "loss": 2.8585,
      "theoretical_loss": 3.6187260797511103,
      "tokens_seen": 1092878336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033782347041123374,
      "loss": 2.6637,
      "theoretical_loss": 3.6187057667509555,
      "tokens_seen": 1092943872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378134403209629,
      "loss": 2.915,
      "theoretical_loss": 3.618685455309813,
      "tokens_seen": 1093009408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378034102306921,
      "loss": 2.9441,
      "theoretical_loss": 3.618665145427469,
      "tokens_seen": 1093074944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377933801404213,
      "loss": 2.9739,
      "theoretical_loss": 3.6186448371037105,
      "tokens_seen": 1093140480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033778335005015046,
      "loss": 2.8502,
      "theoretical_loss": 3.6186245303383253,
      "tokens_seen": 1093206016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033777331995987964,
      "loss": 2.7294,
      "theoretical_loss": 3.6186042251311,
      "tokens_seen": 1093271552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377632898696089,
      "loss": 2.8688,
      "theoretical_loss": 3.6185839214818207,
      "tokens_seen": 1093337088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337753259779338,
      "loss": 3.0609,
      "theoretical_loss": 3.618563619390276,
      "tokens_seen": 1093402624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033774322968906724,
      "loss": 2.6654,
      "theoretical_loss": 3.6185433188562515,
      "tokens_seen": 1093468160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033773319959879637,
      "loss": 2.6188,
      "theoretical_loss": 3.618523019879536,
      "tokens_seen": 1093533696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377231695085256,
      "loss": 2.814,
      "theoretical_loss": 3.6185027224599153,
      "tokens_seen": 1093599232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377131394182548,
      "loss": 2.7198,
      "theoretical_loss": 3.6184824265971773,
      "tokens_seen": 1093664768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033770310932798396,
      "loss": 2.6649,
      "theoretical_loss": 3.6184621322911097,
      "tokens_seen": 1093730304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033769307923771315,
      "loss": 2.8431,
      "theoretical_loss": 3.618441839541499,
      "tokens_seen": 1093795840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376830491474424,
      "loss": 2.9257,
      "theoretical_loss": 3.618421548348133,
      "tokens_seen": 1093861376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376730190571715,
      "loss": 2.6722,
      "theoretical_loss": 3.618401258710799,
      "tokens_seen": 1093926912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033766298896690074,
      "loss": 2.7113,
      "theoretical_loss": 3.6183809706292847,
      "tokens_seen": 1093992448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033765295887662987,
      "loss": 2.6744,
      "theoretical_loss": 3.6183606841033775,
      "tokens_seen": 1094057984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376429287863591,
      "loss": 2.9785,
      "theoretical_loss": 3.618340399132865,
      "tokens_seen": 1094123520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376328986960883,
      "loss": 2.7727,
      "theoretical_loss": 3.6183201157175344,
      "tokens_seen": 1094189056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033762286860581747,
      "loss": 2.9491,
      "theoretical_loss": 3.618299833857174,
      "tokens_seen": 1094254592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033761283851554665,
      "loss": 2.6171,
      "theoretical_loss": 3.6182795535515706,
      "tokens_seen": 1094320128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033760280842527583,
      "loss": 3.0635,
      "theoretical_loss": 3.618259274800513,
      "tokens_seen": 1094385664
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 607859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.727546453475952,
      "objective/train/theoretical_loss": 3.618238997603788,
      "objective/train/tokens_used": 1114911200,
      "theoretical_loss": 3.618238997603788,
      "tokens_seen": 1094451200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337592778335005,
      "loss": 2.8344,
      "theoretical_loss": 3.618238997603788,
      "tokens_seen": 1094451200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033758274824473425,
      "loss": 2.8925,
      "theoretical_loss": 3.618218721961184,
      "tokens_seen": 1094516736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375727181544634,
      "loss": 2.9069,
      "theoretical_loss": 3.618198447872489,
      "tokens_seen": 1094582272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375626880641926,
      "loss": 2.8822,
      "theoretical_loss": 3.6181781753374898,
      "tokens_seen": 1094647808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033755265797392174,
      "loss": 2.8686,
      "theoretical_loss": 3.618157904355975,
      "tokens_seen": 1094713344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033754262788365097,
      "loss": 2.5076,
      "theoretical_loss": 3.6181376349277334,
      "tokens_seen": 1094778880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033753259779338015,
      "loss": 2.857,
      "theoretical_loss": 3.6181173670525517,
      "tokens_seen": 1094844416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033752256770310933,
      "loss": 2.6923,
      "theoretical_loss": 3.6180971007302185,
      "tokens_seen": 1094909952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375125376128385,
      "loss": 2.4682,
      "theoretical_loss": 3.6180768359605215,
      "tokens_seen": 1094975488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033750250752256775,
      "loss": 2.9568,
      "theoretical_loss": 3.61805657274325,
      "tokens_seen": 1095041024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374924774322969,
      "loss": 2.7654,
      "theoretical_loss": 3.618036311078191,
      "tokens_seen": 1095106560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374824473420261,
      "loss": 2.8227,
      "theoretical_loss": 3.618016050965133,
      "tokens_seen": 1095172096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033747241725175524,
      "loss": 2.7536,
      "theoretical_loss": 3.617995792403864,
      "tokens_seen": 1095237632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374623871614845,
      "loss": 2.5663,
      "theoretical_loss": 3.617975535394173,
      "tokens_seen": 1095303168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033745235707121366,
      "loss": 2.8911,
      "theoretical_loss": 3.617955279935848,
      "tokens_seen": 1095368704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033744232698094284,
      "loss": 2.5522,
      "theoretical_loss": 3.617935026028677,
      "tokens_seen": 1095434240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337432296890672,
      "loss": 2.6732,
      "theoretical_loss": 3.6179147736724495,
      "tokens_seen": 1095499776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374222668004012,
      "loss": 2.9258,
      "theoretical_loss": 3.617894522866953,
      "tokens_seen": 1095565312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374122367101304,
      "loss": 2.6745,
      "theoretical_loss": 3.6178742736119758,
      "tokens_seen": 1095630848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374022066198596,
      "loss": 2.4884,
      "theoretical_loss": 3.617854025907307,
      "tokens_seen": 1095696384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033739217652958874,
      "loss": 2.5483,
      "theoretical_loss": 3.6178337797527353,
      "tokens_seen": 1095761920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337382146439318,
      "loss": 2.7345,
      "theoretical_loss": 3.6178135351480494,
      "tokens_seen": 1095827456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373721163490471,
      "loss": 2.825,
      "theoretical_loss": 3.617793292093037,
      "tokens_seen": 1095892992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033736208625877634,
      "loss": 2.8887,
      "theoretical_loss": 3.6177730505874885,
      "tokens_seen": 1095958528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373520561685055,
      "loss": 2.8044,
      "theoretical_loss": 3.617752810631191,
      "tokens_seen": 1096024064
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 608889,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.800088882446289,
      "objective/train/theoretical_loss": 3.6177325722239337,
      "objective/train/tokens_used": 1116549600,
      "theoretical_loss": 3.6177325722239337,
      "tokens_seen": 1096089600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373420260782347,
      "loss": 2.8135,
      "theoretical_loss": 3.6177325722239337,
      "tokens_seen": 1096089600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373319959879639,
      "loss": 2.7077,
      "theoretical_loss": 3.617712335365506,
      "tokens_seen": 1096155136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373219658976931,
      "loss": 2.7595,
      "theoretical_loss": 3.6176921000556965,
      "tokens_seen": 1096220672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033731193580742225,
      "loss": 2.9444,
      "theoretical_loss": 3.617671866294294,
      "tokens_seen": 1096286208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373019057171515,
      "loss": 2.6648,
      "theoretical_loss": 3.6176516340810876,
      "tokens_seen": 1096351744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372918756268806,
      "loss": 2.7699,
      "theoretical_loss": 3.6176314034158663,
      "tokens_seen": 1096417280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033728184553660984,
      "loss": 2.8946,
      "theoretical_loss": 3.6176111742984194,
      "tokens_seen": 1096482816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337271815446339,
      "loss": 2.7141,
      "theoretical_loss": 3.6175909467285354,
      "tokens_seen": 1096548352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372617853560682,
      "loss": 2.8175,
      "theoretical_loss": 3.617570720706004,
      "tokens_seen": 1096613888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372517552657974,
      "loss": 2.9539,
      "theoretical_loss": 3.6175504962306135,
      "tokens_seen": 1096679424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033724172517552657,
      "loss": 2.7993,
      "theoretical_loss": 3.6175302733021546,
      "tokens_seen": 1096744960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033723169508525575,
      "loss": 2.8731,
      "theoretical_loss": 3.617510051920415,
      "tokens_seen": 1096810496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337221664994985,
      "loss": 2.6325,
      "theoretical_loss": 3.6174898320851847,
      "tokens_seen": 1096876032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372116349047141,
      "loss": 2.826,
      "theoretical_loss": 3.617469613796253,
      "tokens_seen": 1096941568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033720160481444335,
      "loss": 2.8224,
      "theoretical_loss": 3.6174493970534094,
      "tokens_seen": 1097007104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371915747241725,
      "loss": 2.8515,
      "theoretical_loss": 3.6174291818564432,
      "tokens_seen": 1097072640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371815446339017,
      "loss": 2.7638,
      "theoretical_loss": 3.617408968205144,
      "tokens_seen": 1097138176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371715145436309,
      "loss": 2.8324,
      "theoretical_loss": 3.6173887560993006,
      "tokens_seen": 1097203712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033716148445336007,
      "loss": 2.891,
      "theoretical_loss": 3.6173685455387035,
      "tokens_seen": 1097269248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033715145436308925,
      "loss": 2.8952,
      "theoretical_loss": 3.6173483365231416,
      "tokens_seen": 1097334784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371414242728185,
      "loss": 2.7169,
      "theoretical_loss": 3.6173281290524053,
      "tokens_seen": 1097400320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371313941825476,
      "loss": 2.8957,
      "theoretical_loss": 3.6173079231262832,
      "tokens_seen": 1097465856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033712136409227685,
      "loss": 2.7076,
      "theoretical_loss": 3.6172877187445662,
      "tokens_seen": 1097531392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337111334002006,
      "loss": 2.7603,
      "theoretical_loss": 3.6172675159070424,
      "tokens_seen": 1097596928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371013039117352,
      "loss": 2.7749,
      "theoretical_loss": 3.617247314613503,
      "tokens_seen": 1097662464
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 610057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0604283809661865,
      "objective/train/theoretical_loss": 3.617227114863738,
      "objective/train/tokens_used": 1118188000,
      "theoretical_loss": 3.617227114863738,
      "tokens_seen": 1097728000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370912738214644,
      "loss": 3.0146,
      "theoretical_loss": 3.617227114863738,
      "tokens_seen": 1097728000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370812437311936,
      "loss": 2.8418,
      "theoretical_loss": 3.617206916657536,
      "tokens_seen": 1097793536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370712136409228,
      "loss": 2.7766,
      "theoretical_loss": 3.617186719994688,
      "tokens_seen": 1097859072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033706118355065194,
      "loss": 2.7925,
      "theoretical_loss": 3.6171665248749836,
      "tokens_seen": 1097924608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033705115346038117,
      "loss": 2.6988,
      "theoretical_loss": 3.6171463312982124,
      "tokens_seen": 1097990144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033704112337011035,
      "loss": 2.9721,
      "theoretical_loss": 3.617126139264165,
      "tokens_seen": 1098055680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033703109327983953,
      "loss": 2.8138,
      "theoretical_loss": 3.617105948772631,
      "tokens_seen": 1098121216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370210631895687,
      "loss": 2.5797,
      "theoretical_loss": 3.617085759823401,
      "tokens_seen": 1098186752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033701103309929795,
      "loss": 2.7429,
      "theoretical_loss": 3.6170655724162653,
      "tokens_seen": 1098252288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370010030090271,
      "loss": 2.6726,
      "theoretical_loss": 3.6170453865510135,
      "tokens_seen": 1098317824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369909729187563,
      "loss": 2.8859,
      "theoretical_loss": 3.6170252022274356,
      "tokens_seen": 1098383360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033698094282848544,
      "loss": 2.7337,
      "theoretical_loss": 3.6170050194453225,
      "tokens_seen": 1098448896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369709127382147,
      "loss": 2.7052,
      "theoretical_loss": 3.616984838204465,
      "tokens_seen": 1098514432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033696088264794386,
      "loss": 2.6465,
      "theoretical_loss": 3.616964658504652,
      "tokens_seen": 1098579968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033695085255767304,
      "loss": 2.5825,
      "theoretical_loss": 3.6169444803456754,
      "tokens_seen": 1098645504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369408224674022,
      "loss": 2.7459,
      "theoretical_loss": 3.616924303727325,
      "tokens_seen": 1098711040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369307923771314,
      "loss": 2.722,
      "theoretical_loss": 3.616904128649391,
      "tokens_seen": 1098776576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369207622868606,
      "loss": 2.6449,
      "theoretical_loss": 3.616883955111664,
      "tokens_seen": 1098842112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369107321965898,
      "loss": 2.7865,
      "theoretical_loss": 3.616863783113935,
      "tokens_seen": 1098907648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033690070210631894,
      "loss": 2.7845,
      "theoretical_loss": 3.6168436126559946,
      "tokens_seen": 1098973184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368906720160482,
      "loss": 2.5238,
      "theoretical_loss": 3.616823443737633,
      "tokens_seen": 1099038720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368806419257773,
      "loss": 2.7368,
      "theoretical_loss": 3.616803276358641,
      "tokens_seen": 1099104256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033687061183550654,
      "loss": 2.6779,
      "theoretical_loss": 3.61678311051881,
      "tokens_seen": 1099169792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368605817452357,
      "loss": 2.7011,
      "theoretical_loss": 3.61676294621793,
      "tokens_seen": 1099235328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368505516549649,
      "loss": 2.7203,
      "theoretical_loss": 3.6167427834557913,
      "tokens_seen": 1099300864
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 610651,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6474804878234863,
      "objective/train/theoretical_loss": 3.6167226222321864,
      "objective/train/tokens_used": 1119826400,
      "theoretical_loss": 3.6167226222321864,
      "tokens_seen": 1099366400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368405215646941,
      "loss": 2.5971,
      "theoretical_loss": 3.6167226222321864,
      "tokens_seen": 1099366400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368304914744233,
      "loss": 2.7681,
      "theoretical_loss": 3.616702462546905,
      "tokens_seen": 1099431936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033682046138415245,
      "loss": 2.737,
      "theoretical_loss": 3.6166823043997383,
      "tokens_seen": 1099497472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368104312938817,
      "loss": 2.7658,
      "theoretical_loss": 3.6166621477904775,
      "tokens_seen": 1099563008
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368004012036108,
      "loss": 2.4798,
      "theoretical_loss": 3.6166419927189133,
      "tokens_seen": 1099628544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033679037111334004,
      "loss": 2.4484,
      "theoretical_loss": 3.6166218391848366,
      "tokens_seen": 1099694080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367803410230692,
      "loss": 2.7835,
      "theoretical_loss": 3.616601687188039,
      "tokens_seen": 1099759616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367703109327984,
      "loss": 2.6162,
      "theoretical_loss": 3.6165815367283116,
      "tokens_seen": 1099825152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367602808425276,
      "loss": 2.8223,
      "theoretical_loss": 3.6165613878054454,
      "tokens_seen": 1099890688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033675025075225677,
      "loss": 2.8348,
      "theoretical_loss": 3.616541240419232,
      "tokens_seen": 1099956224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033674022066198595,
      "loss": 2.9311,
      "theoretical_loss": 3.6165210945694612,
      "tokens_seen": 1100021760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367301905717152,
      "loss": 2.9308,
      "theoretical_loss": 3.616500950255926,
      "tokens_seen": 1100087296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367201604814443,
      "loss": 2.5941,
      "theoretical_loss": 3.6164808074784176,
      "tokens_seen": 1100152832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033671013039117355,
      "loss": 2.7612,
      "theoretical_loss": 3.6164606662367262,
      "tokens_seen": 1100218368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033670010030090273,
      "loss": 2.7321,
      "theoretical_loss": 3.6164405265306447,
      "tokens_seen": 1100283904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366900702106319,
      "loss": 2.6555,
      "theoretical_loss": 3.6164203883599635,
      "tokens_seen": 1100349440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366800401203611,
      "loss": 2.7248,
      "theoretical_loss": 3.6164002517244738,
      "tokens_seen": 1100414976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033667001003009027,
      "loss": 2.9274,
      "theoretical_loss": 3.6163801166239686,
      "tokens_seen": 1100480512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033665997993981945,
      "loss": 2.6037,
      "theoretical_loss": 3.6163599830582385,
      "tokens_seen": 1100546048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366499498495487,
      "loss": 2.6833,
      "theoretical_loss": 3.6163398510270754,
      "tokens_seen": 1100611584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366399197592778,
      "loss": 2.9599,
      "theoretical_loss": 3.6163197205302704,
      "tokens_seen": 1100677120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033662988966900705,
      "loss": 2.6235,
      "theoretical_loss": 3.616299591567616,
      "tokens_seen": 1100742656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366198595787362,
      "loss": 2.6697,
      "theoretical_loss": 3.6162794641389038,
      "tokens_seen": 1100808192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366098294884654,
      "loss": 2.764,
      "theoretical_loss": 3.616259338243925,
      "tokens_seen": 1100873728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365997993981946,
      "loss": 2.7095,
      "theoretical_loss": 3.6162392138824724,
      "tokens_seen": 1100939264
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 611659,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3433709144592285,
      "objective/train/theoretical_loss": 3.6162190910543366,
      "objective/train/tokens_used": 1121464800,
      "theoretical_loss": 3.6162190910543366,
      "tokens_seen": 1101004800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365897693079238,
      "loss": 2.5755,
      "theoretical_loss": 3.6162190910543366,
      "tokens_seen": 1101004800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033657973921765295,
      "loss": 2.8004,
      "theoretical_loss": 3.616198969759311,
      "tokens_seen": 1101070336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033656970912738214,
      "loss": 2.7259,
      "theoretical_loss": 3.6161788499971865,
      "tokens_seen": 1101135872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365596790371113,
      "loss": 2.8029,
      "theoretical_loss": 3.6161587317677553,
      "tokens_seen": 1101201408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033654964894684055,
      "loss": 2.6944,
      "theoretical_loss": 3.6161386150708097,
      "tokens_seen": 1101266944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365396188565697,
      "loss": 2.7386,
      "theoretical_loss": 3.6161184999061415,
      "tokens_seen": 1101332480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365295887662989,
      "loss": 2.6169,
      "theoretical_loss": 3.616098386273543,
      "tokens_seen": 1101398016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365195586760281,
      "loss": 2.9178,
      "theoretical_loss": 3.6160782741728066,
      "tokens_seen": 1101463552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365095285857573,
      "loss": 2.8122,
      "theoretical_loss": 3.616058163603724,
      "tokens_seen": 1101529088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033649949849548646,
      "loss": 2.6384,
      "theoretical_loss": 3.6160380545660873,
      "tokens_seen": 1101594624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033648946840521564,
      "loss": 2.9066,
      "theoretical_loss": 3.6160179470596896,
      "tokens_seen": 1101660160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364794383149448,
      "loss": 2.8531,
      "theoretical_loss": 3.6159978410843223,
      "tokens_seen": 1101725696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033646940822467406,
      "loss": 2.5062,
      "theoretical_loss": 3.6159777366397785,
      "tokens_seen": 1101791232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364593781344032,
      "loss": 2.7402,
      "theoretical_loss": 3.6159576337258503,
      "tokens_seen": 1101856768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364493480441324,
      "loss": 2.8608,
      "theoretical_loss": 3.61593753234233,
      "tokens_seen": 1101922304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033643931795386154,
      "loss": 2.7708,
      "theoretical_loss": 3.6159174324890104,
      "tokens_seen": 1101987840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364292878635908,
      "loss": 3.0905,
      "theoretical_loss": 3.6158973341656835,
      "tokens_seen": 1102053376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033641925777331996,
      "loss": 2.9452,
      "theoretical_loss": 3.6158772373721426,
      "tokens_seen": 1102118912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033640922768304914,
      "loss": 2.8754,
      "theoretical_loss": 3.6158571421081795,
      "tokens_seen": 1102184448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363991975927783,
      "loss": 2.7918,
      "theoretical_loss": 3.6158370483735878,
      "tokens_seen": 1102249984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363891675025075,
      "loss": 3.0109,
      "theoretical_loss": 3.6158169561681595,
      "tokens_seen": 1102315520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363791374122367,
      "loss": 2.9653,
      "theoretical_loss": 3.6157968654916868,
      "tokens_seen": 1102381056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363691073219659,
      "loss": 2.9488,
      "theoretical_loss": 3.6157767763439637,
      "tokens_seen": 1102446592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033635907723169505,
      "loss": 2.751,
      "theoretical_loss": 3.6157566887247823,
      "tokens_seen": 1102512128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363490471414243,
      "loss": 2.9518,
      "theoretical_loss": 3.6157366026339357,
      "tokens_seen": 1102577664
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 612217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9327733516693115,
      "objective/train/theoretical_loss": 3.615716518071217,
      "objective/train/tokens_used": 1123103200,
      "theoretical_loss": 3.615716518071217,
      "tokens_seen": 1102643200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033633901705115346,
      "loss": 2.9102,
      "theoretical_loss": 3.615716518071217,
      "tokens_seen": 1102643200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033632898696088265,
      "loss": 2.7885,
      "theoretical_loss": 3.6156964350364182,
      "tokens_seen": 1102708736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363189568706119,
      "loss": 2.7765,
      "theoretical_loss": 3.615676353529333,
      "tokens_seen": 1102774272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336308926780341,
      "loss": 2.9011,
      "theoretical_loss": 3.6156562735497544,
      "tokens_seen": 1102839808
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033629889669007024,
      "loss": 2.8588,
      "theoretical_loss": 3.6156361950974754,
      "tokens_seen": 1102905344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362888665997994,
      "loss": 2.6921,
      "theoretical_loss": 3.615616118172289,
      "tokens_seen": 1102970880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362788365095286,
      "loss": 2.725,
      "theoretical_loss": 3.615596042773988,
      "tokens_seen": 1103036416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362688064192578,
      "loss": 2.8121,
      "theoretical_loss": 3.615575968902366,
      "tokens_seen": 1103101952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033625877632898697,
      "loss": 2.745,
      "theoretical_loss": 3.6155558965572165,
      "tokens_seen": 1103167488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033624874623871615,
      "loss": 2.7481,
      "theoretical_loss": 3.6155358257383323,
      "tokens_seen": 1103233024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362387161484454,
      "loss": 2.7961,
      "theoretical_loss": 3.6155157564455065,
      "tokens_seen": 1103298560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362286860581745,
      "loss": 2.536,
      "theoretical_loss": 3.615495688678533,
      "tokens_seen": 1103364096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033621865596790375,
      "loss": 2.5681,
      "theoretical_loss": 3.6154756224372044,
      "tokens_seen": 1103429632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033620862587763293,
      "loss": 2.7488,
      "theoretical_loss": 3.615455557721315,
      "tokens_seen": 1103495168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361985957873621,
      "loss": 2.7733,
      "theoretical_loss": 3.615435494530658,
      "tokens_seen": 1103560704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361885656970913,
      "loss": 2.7637,
      "theoretical_loss": 3.615415432865026,
      "tokens_seen": 1103626240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033617853560682047,
      "loss": 2.5859,
      "theoretical_loss": 3.6153953727242136,
      "tokens_seen": 1103691776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033616850551654965,
      "loss": 3.0173,
      "theoretical_loss": 3.6153753141080145,
      "tokens_seen": 1103757312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361584754262789,
      "loss": 2.6315,
      "theoretical_loss": 3.615355257016221,
      "tokens_seen": 1103822848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336148445336008,
      "loss": 2.6973,
      "theoretical_loss": 3.615335201448628,
      "tokens_seen": 1103888384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033613841524573725,
      "loss": 2.8397,
      "theoretical_loss": 3.6153151474050285,
      "tokens_seen": 1103953920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361283851554664,
      "loss": 3.0104,
      "theoretical_loss": 3.6152950948852167,
      "tokens_seen": 1104019456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361183550651956,
      "loss": 2.8473,
      "theoretical_loss": 3.6152750438889862,
      "tokens_seen": 1104084992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361083249749248,
      "loss": 2.7402,
      "theoretical_loss": 3.6152549944161314,
      "tokens_seen": 1104150528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336098294884654,
      "loss": 2.5521,
      "theoretical_loss": 3.615234946466445,
      "tokens_seen": 1104216064
    },
    {
      "epoch": 0.33,
      "objective/train/docs_used": 613498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7340283393859863,
      "objective/train/theoretical_loss": 3.615214900039721,
      "objective/train/tokens_used": 1124741600,
      "theoretical_loss": 3.615214900039721,
      "tokens_seen": 1104281600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033608826479438316,
      "loss": 2.6072,
      "theoretical_loss": 3.615214900039721,
      "tokens_seen": 1104281600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033607823470411234,
      "loss": 2.713,
      "theoretical_loss": 3.6151948551357544,
      "tokens_seen": 1104347136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360682046138415,
      "loss": 2.7713,
      "theoretical_loss": 3.6151748117543385,
      "tokens_seen": 1104412672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033605817452357075,
      "loss": 2.7701,
      "theoretical_loss": 3.615154769895267,
      "tokens_seen": 1104478208
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360481444332999,
      "loss": 2.7249,
      "theoretical_loss": 3.6151347295583345,
      "tokens_seen": 1104543744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360381143430291,
      "loss": 2.7956,
      "theoretical_loss": 3.615114690743335,
      "tokens_seen": 1104609280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360280842527583,
      "loss": 2.9324,
      "theoretical_loss": 3.6150946534500625,
      "tokens_seen": 1104674816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360180541624875,
      "loss": 2.7668,
      "theoretical_loss": 3.615074617678311,
      "tokens_seen": 1104740352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033600802407221666,
      "loss": 2.775,
      "theoretical_loss": 3.6150545834278756,
      "tokens_seen": 1104805888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033599799398194584,
      "loss": 2.7627,
      "theoretical_loss": 3.615034550698549,
      "tokens_seen": 1104871424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000335987963891675,
      "loss": 2.7189,
      "theoretical_loss": 3.615014519490127,
      "tokens_seen": 1104936960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033597793380140426,
      "loss": 2.8479,
      "theoretical_loss": 3.6149944898024033,
      "tokens_seen": 1105002496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359679037111334,
      "loss": 2.7308,
      "theoretical_loss": 3.614974461635172,
      "tokens_seen": 1105068032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359578736208626,
      "loss": 2.742,
      "theoretical_loss": 3.6149544349882277,
      "tokens_seen": 1105133568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033594784353059174,
      "loss": 2.8406,
      "theoretical_loss": 3.614934409861365,
      "tokens_seen": 1105199104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000335937813440321,
      "loss": 2.8199,
      "theoretical_loss": 3.6149143862543784,
      "tokens_seen": 1105264640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033592778335005016,
      "loss": 2.737,
      "theoretical_loss": 3.614894364167063,
      "tokens_seen": 1105330176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033591775325977934,
      "loss": 2.8659,
      "theoretical_loss": 3.614874343599212,
      "tokens_seen": 1105395712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359077231695085,
      "loss": 2.751,
      "theoretical_loss": 3.614854324550621,
      "tokens_seen": 1105461248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358976930792377,
      "loss": 2.978,
      "theoretical_loss": 3.614834307021084,
      "tokens_seen": 1105526784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358876629889669,
      "loss": 2.6978,
      "theoretical_loss": 3.614814291010397,
      "tokens_seen": 1105592320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358776328986961,
      "loss": 2.9026,
      "theoretical_loss": 3.6147942765183534,
      "tokens_seen": 1105657856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033586760280842525,
      "loss": 2.7393,
      "theoretical_loss": 3.6147742635447484,
      "tokens_seen": 1105723392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358575727181545,
      "loss": 2.8325,
      "theoretical_loss": 3.614754252089377,
      "tokens_seen": 1105788928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033584754262788366,
      "loss": 3.0232,
      "theoretical_loss": 3.614734242152034,
      "tokens_seen": 1105854464
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 614567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4206838607788086,
      "objective/train/theoretical_loss": 3.6147142337325135,
      "objective/train/tokens_used": 1126380000,
      "theoretical_loss": 3.6147142337325135,
      "tokens_seen": 1105920000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033583751253761285,
      "loss": 2.6926,
      "theoretical_loss": 3.6147142337325135,
      "tokens_seen": 1105920000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335827482447342,
      "loss": 2.8486,
      "theoretical_loss": 3.6146942268306117,
      "tokens_seen": 1105985536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358174523570712,
      "loss": 2.7974,
      "theoretical_loss": 3.6146742214461227,
      "tokens_seen": 1106051072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358074222668004,
      "loss": 2.7092,
      "theoretical_loss": 3.614654217578842,
      "tokens_seen": 1106116608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357973921765296,
      "loss": 2.5362,
      "theoretical_loss": 3.6146342152285644,
      "tokens_seen": 1106182144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033578736208625875,
      "loss": 2.8508,
      "theoretical_loss": 3.614614214395085,
      "tokens_seen": 1106247680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335777331995988,
      "loss": 2.4857,
      "theoretical_loss": 3.614594215078199,
      "tokens_seen": 1106313216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357673019057171,
      "loss": 2.8308,
      "theoretical_loss": 3.614574217277702,
      "tokens_seen": 1106378752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033575727181544635,
      "loss": 2.8693,
      "theoretical_loss": 3.614554220993388,
      "tokens_seen": 1106444288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033574724172517553,
      "loss": 2.7059,
      "theoretical_loss": 3.6145342262250533,
      "tokens_seen": 1106509824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357372116349047,
      "loss": 2.7467,
      "theoretical_loss": 3.6145142329724926,
      "tokens_seen": 1106575360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357271815446339,
      "loss": 2.6414,
      "theoretical_loss": 3.6144942412355023,
      "tokens_seen": 1106640896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033571715145436313,
      "loss": 2.8767,
      "theoretical_loss": 3.6144742510138768,
      "tokens_seen": 1106706432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033570712136409225,
      "loss": 2.8899,
      "theoretical_loss": 3.614454262307411,
      "tokens_seen": 1106771968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356970912738215,
      "loss": 2.8324,
      "theoretical_loss": 3.6144342751159018,
      "tokens_seen": 1106837504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356870611835506,
      "loss": 2.5821,
      "theoretical_loss": 3.614414289439144,
      "tokens_seen": 1106903040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033567703109327985,
      "loss": 2.6711,
      "theoretical_loss": 3.6143943052769325,
      "tokens_seen": 1106968576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033566700100300903,
      "loss": 2.7409,
      "theoretical_loss": 3.6143743226290637,
      "tokens_seen": 1107034112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356569709127382,
      "loss": 2.8463,
      "theoretical_loss": 3.614354341495333,
      "tokens_seen": 1107099648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356469408224674,
      "loss": 2.7817,
      "theoretical_loss": 3.614334361875536,
      "tokens_seen": 1107165184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356369107321966,
      "loss": 2.7057,
      "theoretical_loss": 3.614314383769469,
      "tokens_seen": 1107230720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033562688064192576,
      "loss": 2.9353,
      "theoretical_loss": 3.6142944071769265,
      "tokens_seen": 1107296256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335616850551655,
      "loss": 2.4972,
      "theoretical_loss": 3.614274432097705,
      "tokens_seen": 1107361792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356068204613841,
      "loss": 2.7738,
      "theoretical_loss": 3.6142544585316,
      "tokens_seen": 1107427328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033559679037111336,
      "loss": 2.7836,
      "theoretical_loss": 3.614234486478408,
      "tokens_seen": 1107492864
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 615343,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.606733798980713,
      "objective/train/theoretical_loss": 3.614214515937924,
      "objective/train/tokens_used": 1128018400,
      "theoretical_loss": 3.614214515937924,
      "tokens_seen": 1107558400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355867602808425,
      "loss": 2.9232,
      "theoretical_loss": 3.614214515937924,
      "tokens_seen": 1107558400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355767301905717,
      "loss": 2.8684,
      "theoretical_loss": 3.614194546909945,
      "tokens_seen": 1107623936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033556670010030095,
      "loss": 2.6234,
      "theoretical_loss": 3.6141745793942657,
      "tokens_seen": 1107689472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355566700100301,
      "loss": 2.6494,
      "theoretical_loss": 3.6141546133906832,
      "tokens_seen": 1107755008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355466399197593,
      "loss": 2.9153,
      "theoretical_loss": 3.6141346488989923,
      "tokens_seen": 1107820544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355366098294885,
      "loss": 2.7833,
      "theoretical_loss": 3.6141146859189908,
      "tokens_seen": 1107886080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355265797392177,
      "loss": 2.5932,
      "theoretical_loss": 3.6140947244504735,
      "tokens_seen": 1107951616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033551654964894686,
      "loss": 2.6819,
      "theoretical_loss": 3.614074764493237,
      "tokens_seen": 1108017152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033550651955867604,
      "loss": 2.9393,
      "theoretical_loss": 3.6140548060470774,
      "tokens_seen": 1108082688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354964894684052,
      "loss": 3.0045,
      "theoretical_loss": 3.614034849111791,
      "tokens_seen": 1108148224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033548645937813446,
      "loss": 2.7504,
      "theoretical_loss": 3.614014893687174,
      "tokens_seen": 1108213760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354764292878636,
      "loss": 2.8945,
      "theoretical_loss": 3.613994939773023,
      "tokens_seen": 1108279296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354663991975928,
      "loss": 2.7427,
      "theoretical_loss": 3.613974987369134,
      "tokens_seen": 1108344832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033545636910732195,
      "loss": 2.9324,
      "theoretical_loss": 3.613955036475304,
      "tokens_seen": 1108410368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354463390170512,
      "loss": 2.575,
      "theoretical_loss": 3.6139350870913285,
      "tokens_seen": 1108475904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033543630892678036,
      "loss": 2.6235,
      "theoretical_loss": 3.6139151392170046,
      "tokens_seen": 1108541440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033542627883650954,
      "loss": 2.7408,
      "theoretical_loss": 3.613895192852129,
      "tokens_seen": 1108606976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354162487462387,
      "loss": 2.7964,
      "theoretical_loss": 3.6138752479964973,
      "tokens_seen": 1108672512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354062186559679,
      "loss": 2.8134,
      "theoretical_loss": 3.6138553046499076,
      "tokens_seen": 1108738048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353961885656971,
      "loss": 2.683,
      "theoretical_loss": 3.6138353628121553,
      "tokens_seen": 1108803584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353861584754263,
      "loss": 2.4956,
      "theoretical_loss": 3.6138154224830377,
      "tokens_seen": 1108869120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033537612838515545,
      "loss": 2.766,
      "theoretical_loss": 3.6137954836623507,
      "tokens_seen": 1108934656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353660982948847,
      "loss": 2.6687,
      "theoretical_loss": 3.613775546349892,
      "tokens_seen": 1109000192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033535606820461386,
      "loss": 2.9337,
      "theoretical_loss": 3.613755610545458,
      "tokens_seen": 1109065728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033534603811434305,
      "loss": 2.6042,
      "theoretical_loss": 3.6137356762488464,
      "tokens_seen": 1109131264
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 616276,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.3585472106933594,
      "objective/train/theoretical_loss": 3.613715743459852,
      "objective/train/tokens_used": 1129656800,
      "theoretical_loss": 3.613715743459852,
      "tokens_seen": 1109196800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353360080240722,
      "loss": 3.0601,
      "theoretical_loss": 3.613715743459852,
      "tokens_seen": 1109196800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353259779338014,
      "loss": 2.673,
      "theoretical_loss": 3.613695812178274,
      "tokens_seen": 1109262336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353159478435306,
      "loss": 2.7332,
      "theoretical_loss": 3.613675882403908,
      "tokens_seen": 1109327872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353059177532598,
      "loss": 2.997,
      "theoretical_loss": 3.6136559541365507,
      "tokens_seen": 1109393408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033529588766298895,
      "loss": 2.732,
      "theoretical_loss": 3.6136360273760006,
      "tokens_seen": 1109458944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352858575727182,
      "loss": 2.7252,
      "theoretical_loss": 3.6136161021220534,
      "tokens_seen": 1109524480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352758274824473,
      "loss": 2.8383,
      "theoretical_loss": 3.6135961783745065,
      "tokens_seen": 1109590016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033526579739217655,
      "loss": 2.836,
      "theoretical_loss": 3.613576256133158,
      "tokens_seen": 1109655552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033525576730190573,
      "loss": 2.694,
      "theoretical_loss": 3.6135563353978037,
      "tokens_seen": 1109721088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352457372116349,
      "loss": 2.6563,
      "theoretical_loss": 3.613536416168242,
      "tokens_seen": 1109786624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352357071213641,
      "loss": 2.497,
      "theoretical_loss": 3.6135164984442696,
      "tokens_seen": 1109852160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033522567703109333,
      "loss": 2.6525,
      "theoretical_loss": 3.6134965822256837,
      "tokens_seen": 1109917696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033521564694082245,
      "loss": 2.7866,
      "theoretical_loss": 3.613476667512282,
      "tokens_seen": 1109983232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352056168505517,
      "loss": 2.7501,
      "theoretical_loss": 3.6134567543038614,
      "tokens_seen": 1110048768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351955867602808,
      "loss": 2.7559,
      "theoretical_loss": 3.61343684260022,
      "tokens_seen": 1110114304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033518555667001005,
      "loss": 2.8027,
      "theoretical_loss": 3.6134169324011545,
      "tokens_seen": 1110179840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033517552657973923,
      "loss": 2.7998,
      "theoretical_loss": 3.613397023706463,
      "tokens_seen": 1110245376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351654964894684,
      "loss": 2.9335,
      "theoretical_loss": 3.6133771165159425,
      "tokens_seen": 1110310912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351554663991976,
      "loss": 2.8296,
      "theoretical_loss": 3.6133572108293914,
      "tokens_seen": 1110376448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351454363089268,
      "loss": 2.7763,
      "theoretical_loss": 3.613337306646607,
      "tokens_seen": 1110441984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033513540621865596,
      "loss": 2.7825,
      "theoretical_loss": 3.613317403967386,
      "tokens_seen": 1110507520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351253761283852,
      "loss": 3.0032,
      "theoretical_loss": 3.6132975027915277,
      "tokens_seen": 1110573056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351153460381143,
      "loss": 2.4999,
      "theoretical_loss": 3.6132776031188287,
      "tokens_seen": 1110638592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033510531594784356,
      "loss": 2.7841,
      "theoretical_loss": 3.6132577049490866,
      "tokens_seen": 1110704128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350952858575727,
      "loss": 2.7022,
      "theoretical_loss": 3.6132378082821006,
      "tokens_seen": 1110769664
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 616833,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7672462463378906,
      "objective/train/theoretical_loss": 3.613217913117667,
      "objective/train/tokens_used": 1131295200,
      "theoretical_loss": 3.613217913117667,
      "tokens_seen": 1110835200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350852557673019,
      "loss": 2.8862,
      "theoretical_loss": 3.613217913117667,
      "tokens_seen": 1110835200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350752256770311,
      "loss": 3.0087,
      "theoretical_loss": 3.6131980194555844,
      "tokens_seen": 1110900736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350651955867603,
      "loss": 2.8562,
      "theoretical_loss": 3.6131781272956514,
      "tokens_seen": 1110966272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033505516549648946,
      "loss": 2.8075,
      "theoretical_loss": 3.6131582366376644,
      "tokens_seen": 1111031808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350451354062187,
      "loss": 2.7964,
      "theoretical_loss": 3.613138347481423,
      "tokens_seen": 1111097344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350351053159478,
      "loss": 2.6795,
      "theoretical_loss": 3.6131184598267243,
      "tokens_seen": 1111162880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033502507522567706,
      "loss": 2.5942,
      "theoretical_loss": 3.6130985736733665,
      "tokens_seen": 1111228416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350150451354062,
      "loss": 2.6254,
      "theoretical_loss": 3.613078689021148,
      "tokens_seen": 1111293952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350050150451354,
      "loss": 2.944,
      "theoretical_loss": 3.6130588058698665,
      "tokens_seen": 1111359488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349949849548646,
      "loss": 2.7659,
      "theoretical_loss": 3.613038924219321,
      "tokens_seen": 1111425024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349849548645938,
      "loss": 2.7844,
      "theoretical_loss": 3.6130190440693095,
      "tokens_seen": 1111490560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033497492477432296,
      "loss": 2.8088,
      "theoretical_loss": 3.6129991654196294,
      "tokens_seen": 1111556096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033496489468405215,
      "loss": 2.7411,
      "theoretical_loss": 3.61297928827008,
      "tokens_seen": 1111621632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349548645937813,
      "loss": 2.8393,
      "theoretical_loss": 3.612959412620459,
      "tokens_seen": 1111687168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033494483450351056,
      "loss": 2.7167,
      "theoretical_loss": 3.612939538470566,
      "tokens_seen": 1111752704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349348044132397,
      "loss": 2.8403,
      "theoretical_loss": 3.612919665820198,
      "tokens_seen": 1111818240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349247743229689,
      "loss": 2.7531,
      "theoretical_loss": 3.612899794669154,
      "tokens_seen": 1111883776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033491474423269805,
      "loss": 2.5888,
      "theoretical_loss": 3.6128799250172325,
      "tokens_seen": 1111949312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003349047141424273,
      "loss": 2.6246,
      "theoretical_loss": 3.6128600568642324,
      "tokens_seen": 1112014848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033489468405215647,
      "loss": 2.6785,
      "theoretical_loss": 3.612840190209952,
      "tokens_seen": 1112080384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033488465396188565,
      "loss": 2.512,
      "theoretical_loss": 3.61282032505419,
      "tokens_seen": 1112145920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033487462387161483,
      "loss": 2.6301,
      "theoretical_loss": 3.612800461396745,
      "tokens_seen": 1112211456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033486459378134406,
      "loss": 2.7051,
      "theoretical_loss": 3.612780599237416,
      "tokens_seen": 1112276992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348545636910732,
      "loss": 2.7752,
      "theoretical_loss": 3.612760738576001,
      "tokens_seen": 1112342528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348445336008024,
      "loss": 2.7236,
      "theoretical_loss": 3.6127408794123,
      "tokens_seen": 1112408064
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 617863,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2659218311309814,
      "objective/train/theoretical_loss": 3.6127210217461108,
      "objective/train/tokens_used": 1132933600,
      "theoretical_loss": 3.6127210217461108,
      "tokens_seen": 1112473600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348345035105316,
      "loss": 2.6578,
      "theoretical_loss": 3.6127210217461108,
      "tokens_seen": 1112473600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348244734202608,
      "loss": 3.0378,
      "theoretical_loss": 3.6127011655772323,
      "tokens_seen": 1112539136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033481444332999,
      "loss": 3.0138,
      "theoretical_loss": 3.6126813109054643,
      "tokens_seen": 1112604672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033480441323971915,
      "loss": 2.7494,
      "theoretical_loss": 3.6126614577306047,
      "tokens_seen": 1112670208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347943831494484,
      "loss": 2.7871,
      "theoretical_loss": 3.6126416060524535,
      "tokens_seen": 1112735744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347843530591775,
      "loss": 2.7748,
      "theoretical_loss": 3.6126217558708085,
      "tokens_seen": 1112801280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033477432296890675,
      "loss": 2.8669,
      "theoretical_loss": 3.6126019071854696,
      "tokens_seen": 1112866816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033476429287863593,
      "loss": 2.4874,
      "theoretical_loss": 3.612582059996236,
      "tokens_seen": 1112932352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347542627883651,
      "loss": 2.7502,
      "theoretical_loss": 3.6125622143029066,
      "tokens_seen": 1112997888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347442326980943,
      "loss": 2.6639,
      "theoretical_loss": 3.6125423701052806,
      "tokens_seen": 1113063424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033473420260782353,
      "loss": 2.7515,
      "theoretical_loss": 3.612522527403157,
      "tokens_seen": 1113128960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033472417251755265,
      "loss": 2.6853,
      "theoretical_loss": 3.6125026861963354,
      "tokens_seen": 1113194496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347141424272819,
      "loss": 2.7526,
      "theoretical_loss": 3.612482846484615,
      "tokens_seen": 1113260032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334704112337011,
      "loss": 2.6622,
      "theoretical_loss": 3.6124630082677944,
      "tokens_seen": 1113325568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033469408224674025,
      "loss": 2.6131,
      "theoretical_loss": 3.6124431715456744,
      "tokens_seen": 1113391104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033468405215646943,
      "loss": 2.6075,
      "theoretical_loss": 3.612423336318053,
      "tokens_seen": 1113456640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346740220661986,
      "loss": 2.7197,
      "theoretical_loss": 3.6124035025847308,
      "tokens_seen": 1113522176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346639919759278,
      "loss": 2.5603,
      "theoretical_loss": 3.6123836703455066,
      "tokens_seen": 1113587712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334653961885657,
      "loss": 3.03,
      "theoretical_loss": 3.6123638396001807,
      "tokens_seen": 1113653248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033464393179538616,
      "loss": 2.8079,
      "theoretical_loss": 3.6123440103485516,
      "tokens_seen": 1113718784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346339017051154,
      "loss": 2.8281,
      "theoretical_loss": 3.6123241825904193,
      "tokens_seen": 1113784320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346238716148445,
      "loss": 2.4955,
      "theoretical_loss": 3.6123043563255837,
      "tokens_seen": 1113849856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033461384152457376,
      "loss": 2.8836,
      "theoretical_loss": 3.6122845315538443,
      "tokens_seen": 1113915392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346038114343029,
      "loss": 2.9342,
      "theoretical_loss": 3.6122647082750006,
      "tokens_seen": 1113980928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345937813440321,
      "loss": 2.7755,
      "theoretical_loss": 3.6122448864888526,
      "tokens_seen": 1114046464
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 618528,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8394312858581543,
      "objective/train/theoretical_loss": 3.612225066195201,
      "objective/train/tokens_used": 1134572000,
      "theoretical_loss": 3.612225066195201,
      "tokens_seen": 1114112000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345837512537613,
      "loss": 2.8486,
      "theoretical_loss": 3.612225066195201,
      "tokens_seen": 1114112000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345737211634905,
      "loss": 2.487,
      "theoretical_loss": 3.6122052473938435,
      "tokens_seen": 1114177536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033456369107321966,
      "loss": 2.8196,
      "theoretical_loss": 3.612185430084582,
      "tokens_seen": 1114243072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345536609829489,
      "loss": 2.7459,
      "theoretical_loss": 3.6121656142672154,
      "tokens_seen": 1114308608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334543630892678,
      "loss": 2.6842,
      "theoretical_loss": 3.612145799941544,
      "tokens_seen": 1114374144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033453360080240726,
      "loss": 2.6307,
      "theoretical_loss": 3.612125987107367,
      "tokens_seen": 1114439680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345235707121364,
      "loss": 2.8632,
      "theoretical_loss": 3.6121061757644863,
      "tokens_seen": 1114505216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345135406218656,
      "loss": 2.64,
      "theoretical_loss": 3.6120863659127,
      "tokens_seen": 1114570752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345035105315948,
      "loss": 2.6673,
      "theoretical_loss": 3.612066557551809,
      "tokens_seen": 1114636288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334493480441324,
      "loss": 2.9368,
      "theoretical_loss": 3.6120467506816136,
      "tokens_seen": 1114701824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033448345035105316,
      "loss": 2.9501,
      "theoretical_loss": 3.612026945301914,
      "tokens_seen": 1114767360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033447342026078235,
      "loss": 2.9252,
      "theoretical_loss": 3.6120071414125103,
      "tokens_seen": 1114832896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344633901705115,
      "loss": 2.6258,
      "theoretical_loss": 3.611987339013202,
      "tokens_seen": 1114898432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033445336008024076,
      "loss": 2.807,
      "theoretical_loss": 3.611967538103791,
      "tokens_seen": 1114963968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344433299899699,
      "loss": 2.8856,
      "theoretical_loss": 3.611947738684076,
      "tokens_seen": 1115029504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344332998996991,
      "loss": 2.7013,
      "theoretical_loss": 3.6119279407538585,
      "tokens_seen": 1115095040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033442326980942825,
      "loss": 2.6865,
      "theoretical_loss": 3.611908144312938,
      "tokens_seen": 1115160576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344132397191575,
      "loss": 2.9106,
      "theoretical_loss": 3.6118883493611165,
      "tokens_seen": 1115226112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033440320962888667,
      "loss": 2.7511,
      "theoretical_loss": 3.6118685558981927,
      "tokens_seen": 1115291648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033439317953861585,
      "loss": 2.6814,
      "theoretical_loss": 3.611848763923968,
      "tokens_seen": 1115357184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033438314944834503,
      "loss": 2.902,
      "theoretical_loss": 3.611828973438243,
      "tokens_seen": 1115422720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033437311935807427,
      "loss": 2.7772,
      "theoretical_loss": 3.611809184440818,
      "tokens_seen": 1115488256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343630892678034,
      "loss": 2.8768,
      "theoretical_loss": 3.6117893969314947,
      "tokens_seen": 1115553792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033435305917753263,
      "loss": 2.3855,
      "theoretical_loss": 3.6117696109100716,
      "tokens_seen": 1115619328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033434302908726175,
      "loss": 2.8104,
      "theoretical_loss": 3.611749826376352,
      "tokens_seen": 1115684864
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 619820,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4851090908050537,
      "objective/train/theoretical_loss": 3.6117300433301347,
      "objective/train/tokens_used": 1136210400,
      "theoretical_loss": 3.6117300433301347,
      "tokens_seen": 1115750400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334332998996991,
      "loss": 2.7111,
      "theoretical_loss": 3.6117300433301347,
      "tokens_seen": 1115750400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033432296890672017,
      "loss": 2.8355,
      "theoretical_loss": 3.611710261771221,
      "tokens_seen": 1115815936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033431293881644935,
      "loss": 2.7384,
      "theoretical_loss": 3.611690481699412,
      "tokens_seen": 1115881472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033430290872617853,
      "loss": 2.6487,
      "theoretical_loss": 3.611670703114509,
      "tokens_seen": 1115947008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342928786359077,
      "loss": 2.8993,
      "theoretical_loss": 3.6116509260163125,
      "tokens_seen": 1116012544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342828485456369,
      "loss": 2.6758,
      "theoretical_loss": 3.611631150404623,
      "tokens_seen": 1116078080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033427281845536613,
      "loss": 2.5941,
      "theoretical_loss": 3.611611376279242,
      "tokens_seen": 1116143616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033426278836509526,
      "loss": 2.7803,
      "theoretical_loss": 3.6115916036399702,
      "tokens_seen": 1116209152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342527582748245,
      "loss": 2.5758,
      "theoretical_loss": 3.611571832486609,
      "tokens_seen": 1116274688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342427281845536,
      "loss": 2.6273,
      "theoretical_loss": 3.6115520628189595,
      "tokens_seen": 1116340224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033423269809428285,
      "loss": 2.9422,
      "theoretical_loss": 3.611532294636823,
      "tokens_seen": 1116405760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033422266800401204,
      "loss": 2.5773,
      "theoretical_loss": 3.6115125279399996,
      "tokens_seen": 1116471296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342126379137412,
      "loss": 2.7014,
      "theoretical_loss": 3.6114927627282922,
      "tokens_seen": 1116536832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342026078234704,
      "loss": 2.6084,
      "theoretical_loss": 3.611472999001501,
      "tokens_seen": 1116602368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033419257773319963,
      "loss": 2.8009,
      "theoretical_loss": 3.611453236759427,
      "tokens_seen": 1116667904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033418254764292876,
      "loss": 2.757,
      "theoretical_loss": 3.6114334760018725,
      "tokens_seen": 1116733440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334172517552658,
      "loss": 2.8419,
      "theoretical_loss": 3.6114137167286384,
      "tokens_seen": 1116798976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341624874623871,
      "loss": 2.444,
      "theoretical_loss": 3.6113939589395256,
      "tokens_seen": 1116864512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033415245737211636,
      "loss": 2.7549,
      "theoretical_loss": 3.6113742026343365,
      "tokens_seen": 1116930048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033414242728184554,
      "loss": 2.7793,
      "theoretical_loss": 3.6113544478128725,
      "tokens_seen": 1116995584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341323971915747,
      "loss": 2.4836,
      "theoretical_loss": 3.6113346944749343,
      "tokens_seen": 1117061120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341223671013039,
      "loss": 2.7046,
      "theoretical_loss": 3.611314942620324,
      "tokens_seen": 1117126656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341123370110331,
      "loss": 2.5733,
      "theoretical_loss": 3.611295192248843,
      "tokens_seen": 1117192192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033410230692076226,
      "loss": 2.7553,
      "theoretical_loss": 3.6112754433602934,
      "tokens_seen": 1117257728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340922768304915,
      "loss": 2.6006,
      "theoretical_loss": 3.6112556959544766,
      "tokens_seen": 1117323264
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 620252,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6974408626556396,
      "objective/train/theoretical_loss": 3.611235950031194,
      "objective/train/tokens_used": 1137848800,
      "theoretical_loss": 3.611235950031194,
      "tokens_seen": 1117388800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340822467402207,
      "loss": 2.6483,
      "theoretical_loss": 3.611235950031194,
      "tokens_seen": 1117388800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033407221664994986,
      "loss": 2.8296,
      "theoretical_loss": 3.6112162055902477,
      "tokens_seen": 1117454336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340621865596791,
      "loss": 2.6342,
      "theoretical_loss": 3.61119646263144,
      "tokens_seen": 1117519872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340521564694082,
      "loss": 2.7249,
      "theoretical_loss": 3.6111767211545716,
      "tokens_seen": 1117585408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033404212637913746,
      "loss": 2.6615,
      "theoretical_loss": 3.6111569811594455,
      "tokens_seen": 1117650944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340320962888666,
      "loss": 2.8549,
      "theoretical_loss": 3.6111372426458628,
      "tokens_seen": 1117716480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340220661985958,
      "loss": 2.7704,
      "theoretical_loss": 3.611117505613625,
      "tokens_seen": 1117782016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334012036108325,
      "loss": 2.79,
      "theoretical_loss": 3.6110977700625355,
      "tokens_seen": 1117847552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340020060180542,
      "loss": 2.5227,
      "theoretical_loss": 3.6110780359923957,
      "tokens_seen": 1117913088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033399197592778336,
      "loss": 3.0267,
      "theoretical_loss": 3.6110583034030075,
      "tokens_seen": 1117978624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033398194583751255,
      "loss": 2.4724,
      "theoretical_loss": 3.6110385722941727,
      "tokens_seen": 1118044160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339719157472417,
      "loss": 2.6887,
      "theoretical_loss": 3.611018842665694,
      "tokens_seen": 1118109696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033396188565697096,
      "loss": 2.7518,
      "theoretical_loss": 3.6109991145173734,
      "tokens_seen": 1118175232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339518555667001,
      "loss": 2.5372,
      "theoretical_loss": 3.6109793878490133,
      "tokens_seen": 1118240768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339418254764293,
      "loss": 2.7651,
      "theoretical_loss": 3.6109596626604157,
      "tokens_seen": 1118306304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033393179538615845,
      "loss": 2.7977,
      "theoretical_loss": 3.6109399389513825,
      "tokens_seen": 1118371840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339217652958877,
      "loss": 2.728,
      "theoretical_loss": 3.610920216721717,
      "tokens_seen": 1118437376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033391173520561687,
      "loss": 2.6523,
      "theoretical_loss": 3.610900495971221,
      "tokens_seen": 1118502912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033390170511534605,
      "loss": 2.7489,
      "theoretical_loss": 3.6108807766996964,
      "tokens_seen": 1118568448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033389167502507523,
      "loss": 2.8485,
      "theoretical_loss": 3.6108610589069468,
      "tokens_seen": 1118633984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033388164493480447,
      "loss": 2.7418,
      "theoretical_loss": 3.610841342592774,
      "tokens_seen": 1118699520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338716148445336,
      "loss": 2.6925,
      "theoretical_loss": 3.6108216277569802,
      "tokens_seen": 1118765056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033386158475426283,
      "loss": 2.8039,
      "theoretical_loss": 3.610801914399369,
      "tokens_seen": 1118830592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033385155466399195,
      "loss": 2.7368,
      "theoretical_loss": 3.610782202519742,
      "tokens_seen": 1118896128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338415245737212,
      "loss": 2.8413,
      "theoretical_loss": 3.6107624921179013,
      "tokens_seen": 1118961664
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 621194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.00057315826416,
      "objective/train/theoretical_loss": 3.6107427831936514,
      "objective/train/tokens_used": 1139487200,
      "theoretical_loss": 3.6107427831936514,
      "tokens_seen": 1119027200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033383149448345037,
      "loss": 2.9902,
      "theoretical_loss": 3.6107427831936514,
      "tokens_seen": 1119027200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033382146439317955,
      "loss": 2.743,
      "theoretical_loss": 3.610723075746794,
      "tokens_seen": 1119092736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033381143430290873,
      "loss": 2.6116,
      "theoretical_loss": 3.6107033697771316,
      "tokens_seen": 1119158272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338014042126379,
      "loss": 3.0118,
      "theoretical_loss": 3.6106836652844674,
      "tokens_seen": 1119223808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337913741223671,
      "loss": 2.5351,
      "theoretical_loss": 3.610663962268604,
      "tokens_seen": 1119289344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033378134403209633,
      "loss": 2.6608,
      "theoretical_loss": 3.6106442607293445,
      "tokens_seen": 1119354880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033377131394182546,
      "loss": 2.6882,
      "theoretical_loss": 3.6106245606664915,
      "tokens_seen": 1119420416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337612838515547,
      "loss": 2.6599,
      "theoretical_loss": 3.610604862079848,
      "tokens_seen": 1119485952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337512537612838,
      "loss": 2.7627,
      "theoretical_loss": 3.6105851649692173,
      "tokens_seen": 1119551488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033374122367101306,
      "loss": 2.6458,
      "theoretical_loss": 3.610565469334402,
      "tokens_seen": 1119617024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033373119358074224,
      "loss": 2.6959,
      "theoretical_loss": 3.610545775175205,
      "tokens_seen": 1119682560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337211634904714,
      "loss": 2.746,
      "theoretical_loss": 3.61052608249143,
      "tokens_seen": 1119748096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337111334002006,
      "loss": 2.6225,
      "theoretical_loss": 3.61050639128288,
      "tokens_seen": 1119813632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033370110330992983,
      "loss": 2.785,
      "theoretical_loss": 3.610486701549357,
      "tokens_seen": 1119879168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033369107321965896,
      "loss": 2.8137,
      "theoretical_loss": 3.610467013290666,
      "tokens_seen": 1119944704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336810431293882,
      "loss": 2.9026,
      "theoretical_loss": 3.61044732650661,
      "tokens_seen": 1120010240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336710130391173,
      "loss": 2.7828,
      "theoretical_loss": 3.6104276411969907,
      "tokens_seen": 1120075776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033366098294884656,
      "loss": 2.7948,
      "theoretical_loss": 3.6104079573616126,
      "tokens_seen": 1120141312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033365095285857574,
      "loss": 2.748,
      "theoretical_loss": 3.610388275000279,
      "tokens_seen": 1120206848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336409227683049,
      "loss": 2.8541,
      "theoretical_loss": 3.610368594112793,
      "tokens_seen": 1120272384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336308926780341,
      "loss": 2.7131,
      "theoretical_loss": 3.6103489146989585,
      "tokens_seen": 1120337920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336208625877633,
      "loss": 2.4622,
      "theoretical_loss": 3.610329236758578,
      "tokens_seen": 1120403456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033361083249749246,
      "loss": 2.7207,
      "theoretical_loss": 3.6103095602914563,
      "tokens_seen": 1120468992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336008024072217,
      "loss": 2.6971,
      "theoretical_loss": 3.610289885297396,
      "tokens_seen": 1120534528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335907723169508,
      "loss": 3.0744,
      "theoretical_loss": 3.610270211776201,
      "tokens_seen": 1120600064
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 621844,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9886622428894043,
      "objective/train/theoretical_loss": 3.6102505397276743,
      "objective/train/tokens_used": 1141125600,
      "theoretical_loss": 3.6102505397276743,
      "tokens_seen": 1120665600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033358074222668006,
      "loss": 2.812,
      "theoretical_loss": 3.6102505397276743,
      "tokens_seen": 1120665600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033357071213640924,
      "loss": 2.6918,
      "theoretical_loss": 3.6102308691516205,
      "tokens_seen": 1120731136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335606820461384,
      "loss": 2.6555,
      "theoretical_loss": 3.610211200047843,
      "tokens_seen": 1120796672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335506519558676,
      "loss": 2.9233,
      "theoretical_loss": 3.6101915324161453,
      "tokens_seen": 1120862208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335406218655968,
      "loss": 2.6204,
      "theoretical_loss": 3.6101718662563314,
      "tokens_seen": 1120927744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033353059177532597,
      "loss": 2.5356,
      "theoretical_loss": 3.610152201568205,
      "tokens_seen": 1120993280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335205616850552,
      "loss": 2.8194,
      "theoretical_loss": 3.6101325383515697,
      "tokens_seen": 1121058816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033351053159478433,
      "loss": 2.8566,
      "theoretical_loss": 3.6101128766062294,
      "tokens_seen": 1121124352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033350050150451356,
      "loss": 2.7643,
      "theoretical_loss": 3.6100932163319888,
      "tokens_seen": 1121189888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334904714142427,
      "loss": 2.7626,
      "theoretical_loss": 3.6100735575286507,
      "tokens_seen": 1121255424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334804413239719,
      "loss": 2.5128,
      "theoretical_loss": 3.6100539001960197,
      "tokens_seen": 1121320960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334704112337011,
      "loss": 2.7467,
      "theoretical_loss": 3.6100342443339004,
      "tokens_seen": 1121386496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334603811434303,
      "loss": 2.7479,
      "theoretical_loss": 3.6100145899420957,
      "tokens_seen": 1121452032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033345035105315947,
      "loss": 2.8238,
      "theoretical_loss": 3.6099949370204105,
      "tokens_seen": 1121517568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033344032096288865,
      "loss": 2.6709,
      "theoretical_loss": 3.6099752855686487,
      "tokens_seen": 1121583104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033343029087261783,
      "loss": 2.7122,
      "theoretical_loss": 3.6099556355866143,
      "tokens_seen": 1121648640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033342026078234707,
      "loss": 2.7043,
      "theoretical_loss": 3.6099359870741115,
      "tokens_seen": 1121714176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334102306920762,
      "loss": 2.6631,
      "theoretical_loss": 3.609916340030945,
      "tokens_seen": 1121779712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033340020060180543,
      "loss": 2.5839,
      "theoretical_loss": 3.609896694456919,
      "tokens_seen": 1121845248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333901705115346,
      "loss": 2.9894,
      "theoretical_loss": 3.6098770503518374,
      "tokens_seen": 1121910784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333801404212638,
      "loss": 2.6817,
      "theoretical_loss": 3.609857407715505,
      "tokens_seen": 1121976320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333370110330993,
      "loss": 2.2964,
      "theoretical_loss": 3.609837766547726,
      "tokens_seen": 1122041856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033336008024072215,
      "loss": 2.6476,
      "theoretical_loss": 3.609818126848305,
      "tokens_seen": 1122107392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033335005015045134,
      "loss": 2.7987,
      "theoretical_loss": 3.609798488617046,
      "tokens_seen": 1122172928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033334002006018057,
      "loss": 2.6787,
      "theoretical_loss": 3.609778851853754,
      "tokens_seen": 1122238464
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 622614,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.409770965576172,
      "objective/train/theoretical_loss": 3.609759216558233,
      "objective/train/tokens_used": 1142764000,
      "theoretical_loss": 3.609759216558233,
      "tokens_seen": 1122304000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033332998996990975,
      "loss": 2.6987,
      "theoretical_loss": 3.609759216558233,
      "tokens_seen": 1122304000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033331995987963893,
      "loss": 2.4774,
      "theoretical_loss": 3.6097395827302887,
      "tokens_seen": 1122369536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333099297893681,
      "loss": 2.7174,
      "theoretical_loss": 3.609719950369725,
      "tokens_seen": 1122435072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332998996990973,
      "loss": 2.8698,
      "theoretical_loss": 3.609700319476346,
      "tokens_seen": 1122500608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033328986960882653,
      "loss": 2.7169,
      "theoretical_loss": 3.6096806900499576,
      "tokens_seen": 1122566144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033327983951855566,
      "loss": 2.6388,
      "theoretical_loss": 3.6096610620903635,
      "tokens_seen": 1122631680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332698094282849,
      "loss": 2.608,
      "theoretical_loss": 3.609641435597369,
      "tokens_seen": 1122697216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333259779338014,
      "loss": 2.9848,
      "theoretical_loss": 3.609621810570779,
      "tokens_seen": 1122762752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033324974924774326,
      "loss": 2.5629,
      "theoretical_loss": 3.609602187010398,
      "tokens_seen": 1122828288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033323971915747244,
      "loss": 2.6,
      "theoretical_loss": 3.6095825649160314,
      "tokens_seen": 1122893824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332296890672016,
      "loss": 2.7688,
      "theoretical_loss": 3.6095629442874833,
      "tokens_seen": 1122959360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332196589769308,
      "loss": 2.555,
      "theoretical_loss": 3.6095433251245597,
      "tokens_seen": 1123024896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033320962888666003,
      "loss": 2.859,
      "theoretical_loss": 3.6095237074270647,
      "tokens_seen": 1123090432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033319959879638916,
      "loss": 2.8304,
      "theoretical_loss": 3.609504091194804,
      "tokens_seen": 1123155968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331895687061184,
      "loss": 2.5043,
      "theoretical_loss": 3.609484476427582,
      "tokens_seen": 1123221504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331795386158475,
      "loss": 2.7801,
      "theoretical_loss": 3.609464863125204,
      "tokens_seen": 1123287040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033316950852557676,
      "loss": 2.6406,
      "theoretical_loss": 3.609445251287476,
      "tokens_seen": 1123352576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033315947843530594,
      "loss": 2.6735,
      "theoretical_loss": 3.609425640914202,
      "tokens_seen": 1123418112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331494483450351,
      "loss": 2.6454,
      "theoretical_loss": 3.609406032005188,
      "tokens_seen": 1123483648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331394182547643,
      "loss": 2.6867,
      "theoretical_loss": 3.609386424560239,
      "tokens_seen": 1123549184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331293881644935,
      "loss": 2.5992,
      "theoretical_loss": 3.6093668185791605,
      "tokens_seen": 1123614720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033311935807422266,
      "loss": 2.6658,
      "theoretical_loss": 3.6093472140617573,
      "tokens_seen": 1123680256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331093279839519,
      "loss": 2.7022,
      "theoretical_loss": 3.609327611007835,
      "tokens_seen": 1123745792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333099297893681,
      "loss": 2.7329,
      "theoretical_loss": 3.6093080094172,
      "tokens_seen": 1123811328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033308926780341026,
      "loss": 2.5589,
      "theoretical_loss": 3.6092884092896558,
      "tokens_seen": 1123876864
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 624048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0764658451080322,
      "objective/train/theoretical_loss": 3.60926881062501,
      "objective/train/tokens_used": 1144402400,
      "theoretical_loss": 3.60926881062501,
      "tokens_seen": 1123942400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033307923771313944,
      "loss": 2.8697,
      "theoretical_loss": 3.60926881062501,
      "tokens_seen": 1123942400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330692076228686,
      "loss": 2.7104,
      "theoretical_loss": 3.609249213423066,
      "tokens_seen": 1124007936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330591775325978,
      "loss": 2.8209,
      "theoretical_loss": 3.6092296176836314,
      "tokens_seen": 1124073472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333049147442327,
      "loss": 2.7686,
      "theoretical_loss": 3.6092100234065105,
      "tokens_seen": 1124139008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033303911735205617,
      "loss": 2.6625,
      "theoretical_loss": 3.609190430591509,
      "tokens_seen": 1124204544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330290872617854,
      "loss": 2.7168,
      "theoretical_loss": 3.609170839238433,
      "tokens_seen": 1124270080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033301905717151453,
      "loss": 2.9399,
      "theoretical_loss": 3.6091512493470885,
      "tokens_seen": 1124335616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033300902708124376,
      "loss": 2.7011,
      "theoretical_loss": 3.60913166091728,
      "tokens_seen": 1124401152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329989969909729,
      "loss": 2.6836,
      "theoretical_loss": 3.6091120739488147,
      "tokens_seen": 1124466688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329889669007021,
      "loss": 2.8206,
      "theoretical_loss": 3.609092488441498,
      "tokens_seen": 1124532224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329789368104313,
      "loss": 2.6126,
      "theoretical_loss": 3.609072904395135,
      "tokens_seen": 1124597760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329689067201605,
      "loss": 2.6087,
      "theoretical_loss": 3.609053321809532,
      "tokens_seen": 1124663296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033295887662988967,
      "loss": 2.7225,
      "theoretical_loss": 3.609033740684496,
      "tokens_seen": 1124728832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033294884653961885,
      "loss": 2.9933,
      "theoretical_loss": 3.6090141610198314,
      "tokens_seen": 1124794368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033293881644934803,
      "loss": 2.831,
      "theoretical_loss": 3.6089945828153454,
      "tokens_seen": 1124859904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033292878635907727,
      "loss": 2.7406,
      "theoretical_loss": 3.608975006070843,
      "tokens_seen": 1124925440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329187562688064,
      "loss": 2.6774,
      "theoretical_loss": 3.608955430786131,
      "tokens_seen": 1124990976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033290872617853563,
      "loss": 2.7052,
      "theoretical_loss": 3.6089358569610153,
      "tokens_seen": 1125056512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328986960882648,
      "loss": 2.6838,
      "theoretical_loss": 3.608916284595302,
      "tokens_seen": 1125122048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332888665997994,
      "loss": 2.7067,
      "theoretical_loss": 3.6088967136887975,
      "tokens_seen": 1125187584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328786359077232,
      "loss": 2.7984,
      "theoretical_loss": 3.608877144241308,
      "tokens_seen": 1125253120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033286860581745235,
      "loss": 2.9188,
      "theoretical_loss": 3.6088575762526394,
      "tokens_seen": 1125318656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033285857572718154,
      "loss": 2.7095,
      "theoretical_loss": 3.6088380097225983,
      "tokens_seen": 1125384192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033284854563691077,
      "loss": 2.734,
      "theoretical_loss": 3.608818444650991,
      "tokens_seen": 1125449728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328385155466399,
      "loss": 2.5812,
      "theoretical_loss": 3.6087988810376235,
      "tokens_seen": 1125515264
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 624451,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6627368927001953,
      "objective/train/theoretical_loss": 3.608779318882303,
      "objective/train/tokens_used": 1146040800,
      "theoretical_loss": 3.608779318882303,
      "tokens_seen": 1125580800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033282848545636913,
      "loss": 2.6308,
      "theoretical_loss": 3.608779318882303,
      "tokens_seen": 1125580800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033281845536609826,
      "loss": 2.5499,
      "theoretical_loss": 3.608759758184836,
      "tokens_seen": 1125646336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328084252758275,
      "loss": 2.8517,
      "theoretical_loss": 3.6087401989450276,
      "tokens_seen": 1125711872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327983951855567,
      "loss": 2.6624,
      "theoretical_loss": 3.6087206411626855,
      "tokens_seen": 1125777408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033278836509528586,
      "loss": 2.751,
      "theoretical_loss": 3.6087010848376164,
      "tokens_seen": 1125842944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033277833500501504,
      "loss": 2.5501,
      "theoretical_loss": 3.608681529969626,
      "tokens_seen": 1125908480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327683049147442,
      "loss": 2.6393,
      "theoretical_loss": 3.6086619765585217,
      "tokens_seen": 1125974016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327582748244734,
      "loss": 2.6149,
      "theoretical_loss": 3.6086424246041098,
      "tokens_seen": 1126039552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033274824473420264,
      "loss": 2.805,
      "theoretical_loss": 3.6086228741061976,
      "tokens_seen": 1126105088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033273821464393176,
      "loss": 2.7527,
      "theoretical_loss": 3.608603325064591,
      "tokens_seen": 1126170624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332728184553661,
      "loss": 2.9454,
      "theoretical_loss": 3.608583777479097,
      "tokens_seen": 1126236160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327181544633902,
      "loss": 2.8227,
      "theoretical_loss": 3.6085642313495225,
      "tokens_seen": 1126301696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033270812437311936,
      "loss": 2.7314,
      "theoretical_loss": 3.6085446866756747,
      "tokens_seen": 1126367232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033269809428284854,
      "loss": 2.6964,
      "theoretical_loss": 3.60852514345736,
      "tokens_seen": 1126432768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326880641925777,
      "loss": 2.8856,
      "theoretical_loss": 3.6085056016943855,
      "tokens_seen": 1126498304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326780341023069,
      "loss": 2.9844,
      "theoretical_loss": 3.608486061386558,
      "tokens_seen": 1126563840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033266800401203614,
      "loss": 2.8392,
      "theoretical_loss": 3.6084665225336856,
      "tokens_seen": 1126629376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033265797392176527,
      "loss": 2.8072,
      "theoretical_loss": 3.608446985135574,
      "tokens_seen": 1126694912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326479438314945,
      "loss": 2.9037,
      "theoretical_loss": 3.6084274491920305,
      "tokens_seen": 1126760448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033263791374122363,
      "loss": 2.713,
      "theoretical_loss": 3.6084079147028625,
      "tokens_seen": 1126825984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033262788365095286,
      "loss": 2.622,
      "theoretical_loss": 3.608388381667877,
      "tokens_seen": 1126891520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033261785356068205,
      "loss": 2.7095,
      "theoretical_loss": 3.6083688500868814,
      "tokens_seen": 1126957056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326078234704112,
      "loss": 2.6724,
      "theoretical_loss": 3.6083493199596828,
      "tokens_seen": 1127022592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325977933801404,
      "loss": 2.6232,
      "theoretical_loss": 3.6083297912860886,
      "tokens_seen": 1127088128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033258776328986964,
      "loss": 2.7234,
      "theoretical_loss": 3.6083102640659055,
      "tokens_seen": 1127153664
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 625621,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.037780284881592,
      "objective/train/theoretical_loss": 3.608290738298942,
      "objective/train/tokens_used": 1147679200,
      "theoretical_loss": 3.608290738298942,
      "tokens_seen": 1127219200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325777331995988,
      "loss": 2.6553,
      "theoretical_loss": 3.608290738298942,
      "tokens_seen": 1127219200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332567703109328,
      "loss": 2.813,
      "theoretical_loss": 3.608271213985004,
      "tokens_seen": 1127284736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325576730190572,
      "loss": 2.5845,
      "theoretical_loss": 3.6082516911239004,
      "tokens_seen": 1127350272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033254764292878637,
      "loss": 2.8016,
      "theoretical_loss": 3.608232169715438,
      "tokens_seen": 1127415808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325376128385156,
      "loss": 2.8117,
      "theoretical_loss": 3.6082126497594236,
      "tokens_seen": 1127481344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033252758274824473,
      "loss": 2.7862,
      "theoretical_loss": 3.6081931312556654,
      "tokens_seen": 1127546880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033251755265797396,
      "loss": 2.8977,
      "theoretical_loss": 3.608173614203971,
      "tokens_seen": 1127612416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325075225677031,
      "loss": 2.6172,
      "theoretical_loss": 3.608154098604148,
      "tokens_seen": 1127677952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033249749247743233,
      "loss": 2.8162,
      "theoretical_loss": 3.6081345844560033,
      "tokens_seen": 1127743488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324874623871615,
      "loss": 2.5889,
      "theoretical_loss": 3.6081150717593458,
      "tokens_seen": 1127809024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324774322968907,
      "loss": 2.8404,
      "theoretical_loss": 3.6080955605139824,
      "tokens_seen": 1127874560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033246740220661987,
      "loss": 2.7708,
      "theoretical_loss": 3.6080760507197214,
      "tokens_seen": 1127940096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033245737211634905,
      "loss": 2.7785,
      "theoretical_loss": 3.6080565423763695,
      "tokens_seen": 1128005632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033244734202607823,
      "loss": 2.9172,
      "theoretical_loss": 3.6080370354837354,
      "tokens_seen": 1128071168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033243731193580747,
      "loss": 2.5575,
      "theoretical_loss": 3.608017530041627,
      "tokens_seen": 1128136704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324272818455366,
      "loss": 2.6802,
      "theoretical_loss": 3.6079980260498514,
      "tokens_seen": 1128202240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033241725175526583,
      "loss": 2.6072,
      "theoretical_loss": 3.6079785235082174,
      "tokens_seen": 1128267776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332407221664995,
      "loss": 2.6779,
      "theoretical_loss": 3.6079590224165328,
      "tokens_seen": 1128333312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323971915747242,
      "loss": 2.6611,
      "theoretical_loss": 3.607939522774605,
      "tokens_seen": 1128398848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323871614844534,
      "loss": 2.7856,
      "theoretical_loss": 3.607920024582242,
      "tokens_seen": 1128464384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033237713139418255,
      "loss": 2.6113,
      "theoretical_loss": 3.607900527839253,
      "tokens_seen": 1128529920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033236710130391174,
      "loss": 2.8294,
      "theoretical_loss": 3.6078810325454453,
      "tokens_seen": 1128595456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033235707121364097,
      "loss": 2.7486,
      "theoretical_loss": 3.607861538700627,
      "tokens_seen": 1128660992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323470411233701,
      "loss": 2.9031,
      "theoretical_loss": 3.607842046304606,
      "tokens_seen": 1128726528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033233701103309933,
      "loss": 2.8622,
      "theoretical_loss": 3.607822555357192,
      "tokens_seen": 1128792064
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 626861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9778614044189453,
      "objective/train/theoretical_loss": 3.6078030658581914,
      "objective/train/tokens_used": 1149317600,
      "theoretical_loss": 3.6078030658581914,
      "tokens_seen": 1128857600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033232698094282846,
      "loss": 2.7505,
      "theoretical_loss": 3.6078030658581914,
      "tokens_seen": 1128857600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323169508525577,
      "loss": 2.6772,
      "theoretical_loss": 3.607783577807413,
      "tokens_seen": 1128923136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323069207622869,
      "loss": 2.654,
      "theoretical_loss": 3.6077640912046665,
      "tokens_seen": 1128988672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033229689067201606,
      "loss": 2.6944,
      "theoretical_loss": 3.607744606049758,
      "tokens_seen": 1129054208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033228686058174524,
      "loss": 2.7522,
      "theoretical_loss": 3.6077251223424973,
      "tokens_seen": 1129119744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322768304914744,
      "loss": 2.7485,
      "theoretical_loss": 3.607705640082693,
      "tokens_seen": 1129185280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322668004012036,
      "loss": 2.78,
      "theoretical_loss": 3.607686159270153,
      "tokens_seen": 1129250816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033225677031093284,
      "loss": 2.8463,
      "theoretical_loss": 3.6076666799046855,
      "tokens_seen": 1129316352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033224674022066196,
      "loss": 2.7895,
      "theoretical_loss": 3.6076472019861,
      "tokens_seen": 1129381888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322367101303912,
      "loss": 2.5306,
      "theoretical_loss": 3.607627725514204,
      "tokens_seen": 1129447424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322266800401204,
      "loss": 2.5914,
      "theoretical_loss": 3.6076082504888074,
      "tokens_seen": 1129512960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033221664994984956,
      "loss": 2.6608,
      "theoretical_loss": 3.6075887769097177,
      "tokens_seen": 1129578496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033220661985957874,
      "loss": 2.8178,
      "theoretical_loss": 3.607569304776744,
      "tokens_seen": 1129644032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321965897693079,
      "loss": 2.6823,
      "theoretical_loss": 3.607549834089695,
      "tokens_seen": 1129709568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321865596790371,
      "loss": 2.7135,
      "theoretical_loss": 3.6075303648483796,
      "tokens_seen": 1129775104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033217652958876634,
      "loss": 2.7891,
      "theoretical_loss": 3.607510897052607,
      "tokens_seen": 1129840640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033216649949849547,
      "loss": 2.5299,
      "theoretical_loss": 3.6074914307021846,
      "tokens_seen": 1129906176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321564694082247,
      "loss": 2.7886,
      "theoretical_loss": 3.6074719657969228,
      "tokens_seen": 1129971712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033214643931795383,
      "loss": 2.8238,
      "theoretical_loss": 3.60745250233663,
      "tokens_seen": 1130037248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033213640922768306,
      "loss": 2.7838,
      "theoretical_loss": 3.607433040321115,
      "tokens_seen": 1130102784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033212637913741225,
      "loss": 2.7236,
      "theoretical_loss": 3.6074135797501867,
      "tokens_seen": 1130168320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321163490471414,
      "loss": 2.9937,
      "theoretical_loss": 3.6073941206236544,
      "tokens_seen": 1130233856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321063189568706,
      "loss": 2.8337,
      "theoretical_loss": 3.607374662941327,
      "tokens_seen": 1130299392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033209628886659984,
      "loss": 2.7751,
      "theoretical_loss": 3.607355206703013,
      "tokens_seen": 1130364928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033208625877632897,
      "loss": 2.7562,
      "theoretical_loss": 3.607335751908522,
      "tokens_seen": 1130430464
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 627282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7507452964782715,
      "objective/train/theoretical_loss": 3.6073162985576643,
      "objective/train/tokens_used": 1150956000,
      "theoretical_loss": 3.6073162985576643,
      "tokens_seen": 1130496000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320762286860582,
      "loss": 2.7378,
      "theoretical_loss": 3.6073162985576643,
      "tokens_seen": 1130496000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033206619859578733,
      "loss": 2.6933,
      "theoretical_loss": 3.6072968466502475,
      "tokens_seen": 1130561536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033205616850551657,
      "loss": 2.6401,
      "theoretical_loss": 3.6072773961860816,
      "tokens_seen": 1130627072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033204613841524575,
      "loss": 2.7799,
      "theoretical_loss": 3.6072579471649755,
      "tokens_seen": 1130692608
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033203610832497493,
      "loss": 2.6876,
      "theoretical_loss": 3.607238499586739,
      "tokens_seen": 1130758144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320260782347041,
      "loss": 2.7543,
      "theoretical_loss": 3.6072190534511805,
      "tokens_seen": 1130823680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320160481444333,
      "loss": 2.8909,
      "theoretical_loss": 3.60719960875811,
      "tokens_seen": 1130889216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320060180541625,
      "loss": 2.6733,
      "theoretical_loss": 3.6071801655073377,
      "tokens_seen": 1130954752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319959879638917,
      "loss": 2.5583,
      "theoretical_loss": 3.6071607236986716,
      "tokens_seen": 1131020288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033198595787362084,
      "loss": 2.9844,
      "theoretical_loss": 3.607141283331922,
      "tokens_seen": 1131085824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033197592778335007,
      "loss": 2.933,
      "theoretical_loss": 3.6071218444068984,
      "tokens_seen": 1131151360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319658976930792,
      "loss": 2.4985,
      "theoretical_loss": 3.60710240692341,
      "tokens_seen": 1131216896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033195586760280843,
      "loss": 2.7601,
      "theoretical_loss": 3.607082970881267,
      "tokens_seen": 1131282432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319458375125376,
      "loss": 2.9582,
      "theoretical_loss": 3.607063536280278,
      "tokens_seen": 1131347968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319358074222668,
      "loss": 2.6208,
      "theoretical_loss": 3.607044103120254,
      "tokens_seen": 1131413504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331925777331996,
      "loss": 2.9075,
      "theoretical_loss": 3.6070246714010037,
      "tokens_seen": 1131479040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319157472417252,
      "loss": 2.8108,
      "theoretical_loss": 3.6070052411223372,
      "tokens_seen": 1131544576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033190571715145434,
      "loss": 2.81,
      "theoretical_loss": 3.6069858122840643,
      "tokens_seen": 1131610112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318956870611836,
      "loss": 2.696,
      "theoretical_loss": 3.606966384885995,
      "tokens_seen": 1131675648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318856569709127,
      "loss": 2.6364,
      "theoretical_loss": 3.6069469589279386,
      "tokens_seen": 1131741184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033187562688064194,
      "loss": 2.7666,
      "theoretical_loss": 3.6069275344097047,
      "tokens_seen": 1131806720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318655967903711,
      "loss": 2.8148,
      "theoretical_loss": 3.606908111331105,
      "tokens_seen": 1131872256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318555667001003,
      "loss": 2.9294,
      "theoretical_loss": 3.6068886896919476,
      "tokens_seen": 1131937792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318455366098295,
      "loss": 2.6665,
      "theoretical_loss": 3.6068692694920426,
      "tokens_seen": 1132003328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033183550651955866,
      "loss": 2.5168,
      "theoretical_loss": 3.6068498507312015,
      "tokens_seen": 1132068864
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 628386,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4533891677856445,
      "objective/train/theoretical_loss": 3.6068304334092334,
      "objective/train/tokens_used": 1152594400,
      "theoretical_loss": 3.6068304334092334,
      "tokens_seen": 1132134400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318254764292879,
      "loss": 2.6008,
      "theoretical_loss": 3.6068304334092334,
      "tokens_seen": 1132134400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318154463390171,
      "loss": 2.8305,
      "theoretical_loss": 3.606811017525948,
      "tokens_seen": 1132199936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033180541624874626,
      "loss": 2.7182,
      "theoretical_loss": 3.606791603081156,
      "tokens_seen": 1132265472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033179538615847544,
      "loss": 2.8093,
      "theoretical_loss": 3.6067721900746674,
      "tokens_seen": 1132331008
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317853560682046,
      "loss": 2.5651,
      "theoretical_loss": 3.6067527785062925,
      "tokens_seen": 1132396544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317753259779338,
      "loss": 2.8494,
      "theoretical_loss": 3.606733368375842,
      "tokens_seen": 1132462080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033176529588766304,
      "loss": 2.7811,
      "theoretical_loss": 3.606713959683125,
      "tokens_seen": 1132527616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033175526579739216,
      "loss": 2.9312,
      "theoretical_loss": 3.6066945524279532,
      "tokens_seen": 1132593152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317452357071214,
      "loss": 2.7798,
      "theoretical_loss": 3.6066751466101357,
      "tokens_seen": 1132658688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317352056168506,
      "loss": 2.7387,
      "theoretical_loss": 3.606655742229483,
      "tokens_seen": 1132724224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033172517552657976,
      "loss": 2.6532,
      "theoretical_loss": 3.606636339285807,
      "tokens_seen": 1132789760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033171514543630894,
      "loss": 2.7307,
      "theoretical_loss": 3.6066169377789166,
      "tokens_seen": 1132855296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317051153460381,
      "loss": 2.7641,
      "theoretical_loss": 3.6065975377086232,
      "tokens_seen": 1132920832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316950852557673,
      "loss": 2.7869,
      "theoretical_loss": 3.6065781390747365,
      "tokens_seen": 1132986368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033168505516549654,
      "loss": 2.6226,
      "theoretical_loss": 3.606558741877068,
      "tokens_seen": 1133051904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033167502507522567,
      "loss": 2.7482,
      "theoretical_loss": 3.6065393461154276,
      "tokens_seen": 1133117440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316649949849549,
      "loss": 2.4595,
      "theoretical_loss": 3.6065199517896263,
      "tokens_seen": 1133182976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033165496489468403,
      "loss": 2.7327,
      "theoretical_loss": 3.6065005588994747,
      "tokens_seen": 1133248512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033164493480441326,
      "loss": 2.7135,
      "theoretical_loss": 3.606481167444783,
      "tokens_seen": 1133314048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033163490471414245,
      "loss": 2.6186,
      "theoretical_loss": 3.6064617774253627,
      "tokens_seen": 1133379584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316248746238716,
      "loss": 2.9134,
      "theoretical_loss": 3.606442388841024,
      "tokens_seen": 1133445120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316148445336008,
      "loss": 2.8247,
      "theoretical_loss": 3.606423001691579,
      "tokens_seen": 1133510656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033160481444333004,
      "loss": 2.6597,
      "theoretical_loss": 3.606403615976837,
      "tokens_seen": 1133576192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033159478435305917,
      "loss": 2.4916,
      "theoretical_loss": 3.606384231696609,
      "tokens_seen": 1133641728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315847542627884,
      "loss": 2.9193,
      "theoretical_loss": 3.6063648488507067,
      "tokens_seen": 1133707264
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 629006,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7128870487213135,
      "objective/train/theoretical_loss": 3.606345467438941,
      "objective/train/tokens_used": 1154232800,
      "theoretical_loss": 3.606345467438941,
      "tokens_seen": 1133772800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033157472417251753,
      "loss": 2.7728,
      "theoretical_loss": 3.606345467438941,
      "tokens_seen": 1133772800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033156469408224677,
      "loss": 2.7129,
      "theoretical_loss": 3.6063260874611225,
      "tokens_seen": 1133838336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033155466399197595,
      "loss": 2.7891,
      "theoretical_loss": 3.6063067089170624,
      "tokens_seen": 1133903872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033154463390170513,
      "loss": 2.6031,
      "theoretical_loss": 3.6062873318065716,
      "tokens_seen": 1133969408
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315346038114343,
      "loss": 2.3994,
      "theoretical_loss": 3.6062679561294617,
      "tokens_seen": 1134034944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315245737211635,
      "loss": 2.7607,
      "theoretical_loss": 3.606248581885543,
      "tokens_seen": 1134100480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315145436308927,
      "loss": 2.7066,
      "theoretical_loss": 3.6062292090746277,
      "tokens_seen": 1134166016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315045135406219,
      "loss": 2.6768,
      "theoretical_loss": 3.606209837696527,
      "tokens_seen": 1134231552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033149448345035104,
      "loss": 2.6928,
      "theoretical_loss": 3.6061904677510506,
      "tokens_seen": 1134297088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033148445336008027,
      "loss": 2.6556,
      "theoretical_loss": 3.6061710992380114,
      "tokens_seen": 1134362624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314744232698094,
      "loss": 2.9638,
      "theoretical_loss": 3.60615173215722,
      "tokens_seen": 1134428160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033146439317953863,
      "loss": 2.7812,
      "theoretical_loss": 3.606132366508488,
      "tokens_seen": 1134493696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314543630892678,
      "loss": 2.5463,
      "theoretical_loss": 3.6061130022916266,
      "tokens_seen": 1134559232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331444332998997,
      "loss": 2.8497,
      "theoretical_loss": 3.6060936395064473,
      "tokens_seen": 1134624768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314343029087262,
      "loss": 2.7433,
      "theoretical_loss": 3.606074278152762,
      "tokens_seen": 1134690304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314242728184554,
      "loss": 2.9172,
      "theoretical_loss": 3.606054918230382,
      "tokens_seen": 1134755840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033141424272818454,
      "loss": 2.6835,
      "theoretical_loss": 3.6060355597391176,
      "tokens_seen": 1134821376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314042126379138,
      "loss": 2.6681,
      "theoretical_loss": 3.6060162026787825,
      "tokens_seen": 1134886912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313941825476429,
      "loss": 2.6483,
      "theoretical_loss": 3.6059968470491865,
      "tokens_seen": 1134952448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033138415245737214,
      "loss": 2.723,
      "theoretical_loss": 3.6059774928501422,
      "tokens_seen": 1135017984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313741223671013,
      "loss": 2.6665,
      "theoretical_loss": 3.6059581400814613,
      "tokens_seen": 1135083520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313640922768305,
      "loss": 2.8749,
      "theoretical_loss": 3.6059387887429546,
      "tokens_seen": 1135149056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313540621865597,
      "loss": 2.6138,
      "theoretical_loss": 3.605919438834435,
      "tokens_seen": 1135214592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033134403209628886,
      "loss": 2.7763,
      "theoretical_loss": 3.605900090355714,
      "tokens_seen": 1135280128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033133400200601804,
      "loss": 2.7442,
      "theoretical_loss": 3.605880743306603,
      "tokens_seen": 1135345664
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 630145,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.701599597930908,
      "objective/train/theoretical_loss": 3.605861397686914,
      "objective/train/tokens_used": 1155871200,
      "theoretical_loss": 3.605861397686914,
      "tokens_seen": 1135411200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313239719157473,
      "loss": 2.7269,
      "theoretical_loss": 3.605861397686914,
      "tokens_seen": 1135411200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313139418254764,
      "loss": 2.8759,
      "theoretical_loss": 3.605842053496459,
      "tokens_seen": 1135476736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033130391173520564,
      "loss": 2.8524,
      "theoretical_loss": 3.60582271073505,
      "tokens_seen": 1135542272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033129388164493477,
      "loss": 2.7067,
      "theoretical_loss": 3.6058033694024987,
      "tokens_seen": 1135607808
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331283851554664,
      "loss": 2.7671,
      "theoretical_loss": 3.605784029498617,
      "tokens_seen": 1135673344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312738214643932,
      "loss": 2.5566,
      "theoretical_loss": 3.605764691023218,
      "tokens_seen": 1135738880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033126379137412236,
      "loss": 2.6674,
      "theoretical_loss": 3.605745353976112,
      "tokens_seen": 1135804416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033125376128385155,
      "loss": 3.0017,
      "theoretical_loss": 3.605726018357113,
      "tokens_seen": 1135869952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312437311935808,
      "loss": 2.8305,
      "theoretical_loss": 3.6057066841660315,
      "tokens_seen": 1135935488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312337011033099,
      "loss": 2.8144,
      "theoretical_loss": 3.605687351402681,
      "tokens_seen": 1136001024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033122367101303914,
      "loss": 2.9268,
      "theoretical_loss": 3.6056680200668723,
      "tokens_seen": 1136066560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033121364092276827,
      "loss": 2.7998,
      "theoretical_loss": 3.605648690158419,
      "tokens_seen": 1136132096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312036108324975,
      "loss": 2.8524,
      "theoretical_loss": 3.605629361677133,
      "tokens_seen": 1136197632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311935807422267,
      "loss": 2.7364,
      "theoretical_loss": 3.605610034622826,
      "tokens_seen": 1136263168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033118355065195587,
      "loss": 2.8107,
      "theoretical_loss": 3.6055907089953108,
      "tokens_seen": 1136328704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033117352056168505,
      "loss": 2.7178,
      "theoretical_loss": 3.6055713847944,
      "tokens_seen": 1136394240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033116349047141423,
      "loss": 2.8013,
      "theoretical_loss": 3.605552062019906,
      "tokens_seen": 1136459776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311534603811434,
      "loss": 2.8232,
      "theoretical_loss": 3.6055327406716406,
      "tokens_seen": 1136525312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033114343029087265,
      "loss": 2.6199,
      "theoretical_loss": 3.6055134207494177,
      "tokens_seen": 1136590848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033113340020060177,
      "loss": 2.763,
      "theoretical_loss": 3.605494102253048,
      "tokens_seen": 1136656384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331123370110331,
      "loss": 2.7053,
      "theoretical_loss": 3.6054747851823454,
      "tokens_seen": 1136721920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033111334002006013,
      "loss": 2.8259,
      "theoretical_loss": 3.605455469537122,
      "tokens_seen": 1136787456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033110330992978937,
      "loss": 2.8541,
      "theoretical_loss": 3.6054361553171903,
      "tokens_seen": 1136852992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033109327983951855,
      "loss": 2.7364,
      "theoretical_loss": 3.605416842522364,
      "tokens_seen": 1136918528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033108324974924773,
      "loss": 2.7473,
      "theoretical_loss": 3.605397531152454,
      "tokens_seen": 1136984064
    },
    {
      "epoch": 0.34,
      "objective/train/docs_used": 631187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.756531000137329,
      "objective/train/theoretical_loss": 3.6053782212072747,
      "objective/train/tokens_used": 1157509600,
      "theoretical_loss": 3.6053782212072747,
      "tokens_seen": 1137049600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033107321965897697,
      "loss": 2.6511,
      "theoretical_loss": 3.6053782212072747,
      "tokens_seen": 1137049600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033106318956870615,
      "loss": 2.4748,
      "theoretical_loss": 3.605358912686638,
      "tokens_seen": 1137115136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033105315947843533,
      "loss": 2.869,
      "theoretical_loss": 3.6053396055903564,
      "tokens_seen": 1137180672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310431293881645,
      "loss": 2.6319,
      "theoretical_loss": 3.605320299918244,
      "tokens_seen": 1137246208
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310330992978937,
      "loss": 2.7541,
      "theoretical_loss": 3.6053009956701128,
      "tokens_seen": 1137311744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310230692076229,
      "loss": 2.8495,
      "theoretical_loss": 3.605281692845776,
      "tokens_seen": 1137377280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310130391173521,
      "loss": 2.7188,
      "theoretical_loss": 3.605262391445047,
      "tokens_seen": 1137442816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033100300902708124,
      "loss": 2.6702,
      "theoretical_loss": 3.605243091467737,
      "tokens_seen": 1137508352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033099297893681047,
      "loss": 2.6905,
      "theoretical_loss": 3.605223792913661,
      "tokens_seen": 1137573888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309829488465396,
      "loss": 2.9471,
      "theoretical_loss": 3.6052044957826315,
      "tokens_seen": 1137639424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033097291875626883,
      "loss": 2.9353,
      "theoretical_loss": 3.6051852000744606,
      "tokens_seen": 1137704960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000330962888665998,
      "loss": 2.6811,
      "theoretical_loss": 3.6051659057889633,
      "tokens_seen": 1137770496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309528585757272,
      "loss": 2.8414,
      "theoretical_loss": 3.6051466129259513,
      "tokens_seen": 1137836032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309428284854564,
      "loss": 2.5861,
      "theoretical_loss": 3.6051273214852384,
      "tokens_seen": 1137901568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309327983951856,
      "loss": 2.8635,
      "theoretical_loss": 3.6051080314666377,
      "tokens_seen": 1137967104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033092276830491474,
      "loss": 2.6327,
      "theoretical_loss": 3.6050887428699623,
      "tokens_seen": 1138032640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000330912738214644,
      "loss": 2.5742,
      "theoretical_loss": 3.6050694556950256,
      "tokens_seen": 1138098176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309027081243731,
      "loss": 2.8405,
      "theoretical_loss": 3.6050501699416415,
      "tokens_seen": 1138163712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033089267803410234,
      "loss": 2.8262,
      "theoretical_loss": 3.6050308856096223,
      "tokens_seen": 1138229248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308826479438315,
      "loss": 2.6865,
      "theoretical_loss": 3.6050116026987826,
      "tokens_seen": 1138294784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308726178535607,
      "loss": 2.6635,
      "theoretical_loss": 3.6049923212089343,
      "tokens_seen": 1138360320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308625877632899,
      "loss": 2.8313,
      "theoretical_loss": 3.6049730411398926,
      "tokens_seen": 1138425856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033085255767301906,
      "loss": 2.6591,
      "theoretical_loss": 3.60495376249147,
      "tokens_seen": 1138491392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033084252758274824,
      "loss": 2.6289,
      "theoretical_loss": 3.604934485263481,
      "tokens_seen": 1138556928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003308324974924775,
      "loss": 2.7447,
      "theoretical_loss": 3.604915209455738,
      "tokens_seen": 1138622464
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 631834,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4626145362854004,
      "objective/train/theoretical_loss": 3.604895935068055,
      "objective/train/tokens_used": 1159148000,
      "theoretical_loss": 3.604895935068055,
      "tokens_seen": 1138688000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003308224674022066,
      "loss": 2.8697,
      "theoretical_loss": 3.604895935068055,
      "tokens_seen": 1138688000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033081243731193584,
      "loss": 2.8243,
      "theoretical_loss": 3.604876662100246,
      "tokens_seen": 1138753536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033080240722166497,
      "loss": 2.742,
      "theoretical_loss": 3.604857390552125,
      "tokens_seen": 1138819072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307923771313942,
      "loss": 2.7061,
      "theoretical_loss": 3.604838120423504,
      "tokens_seen": 1138884608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307823470411234,
      "loss": 2.7016,
      "theoretical_loss": 3.6048188517141995,
      "tokens_seen": 1138950144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033077231695085256,
      "loss": 2.7084,
      "theoretical_loss": 3.604799584424023,
      "tokens_seen": 1139015680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033076228686058175,
      "loss": 2.7386,
      "theoretical_loss": 3.604780318552789,
      "tokens_seen": 1139081216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000330752256770311,
      "loss": 2.7143,
      "theoretical_loss": 3.6047610541003117,
      "tokens_seen": 1139146752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307422266800401,
      "loss": 2.8785,
      "theoretical_loss": 3.604741791066405,
      "tokens_seen": 1139212288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033073219658976934,
      "loss": 2.7387,
      "theoretical_loss": 3.6047225294508825,
      "tokens_seen": 1139277824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033072216649949847,
      "loss": 2.6078,
      "theoretical_loss": 3.6047032692535588,
      "tokens_seen": 1139343360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307121364092277,
      "loss": 2.6801,
      "theoretical_loss": 3.6046840104742466,
      "tokens_seen": 1139408896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307021063189569,
      "loss": 2.6139,
      "theoretical_loss": 3.6046647531127616,
      "tokens_seen": 1139474432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033069207622868607,
      "loss": 2.7409,
      "theoretical_loss": 3.604645497168917,
      "tokens_seen": 1139539968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033068204613841525,
      "loss": 2.6384,
      "theoretical_loss": 3.6046262426425266,
      "tokens_seen": 1139605504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033067201604814443,
      "loss": 2.7318,
      "theoretical_loss": 3.604606989533405,
      "tokens_seen": 1139671040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306619859578736,
      "loss": 2.9143,
      "theoretical_loss": 3.6045877378413667,
      "tokens_seen": 1139736576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033065195586760285,
      "loss": 2.6698,
      "theoretical_loss": 3.604568487566225,
      "tokens_seen": 1139802112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033064192577733197,
      "loss": 2.7636,
      "theoretical_loss": 3.6045492387077958,
      "tokens_seen": 1139867648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306318956870612,
      "loss": 2.7425,
      "theoretical_loss": 3.604529991265892,
      "tokens_seen": 1139933184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033062186559679034,
      "loss": 2.602,
      "theoretical_loss": 3.6045107452403276,
      "tokens_seen": 1139998720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033061183550651957,
      "loss": 2.7949,
      "theoretical_loss": 3.6044915006309184,
      "tokens_seen": 1140064256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033060180541624875,
      "loss": 2.6797,
      "theoretical_loss": 3.6044722574374775,
      "tokens_seen": 1140129792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033059177532597793,
      "loss": 2.5068,
      "theoretical_loss": 3.60445301565982,
      "tokens_seen": 1140195328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305817452357071,
      "loss": 2.5564,
      "theoretical_loss": 3.6044337752977604,
      "tokens_seen": 1140260864
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 632825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7563414573669434,
      "objective/train/theoretical_loss": 3.604414536351113,
      "objective/train/tokens_used": 1160786400,
      "theoretical_loss": 3.604414536351113,
      "tokens_seen": 1140326400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033057171514543635,
      "loss": 2.8835,
      "theoretical_loss": 3.604414536351113,
      "tokens_seen": 1140326400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305616850551655,
      "loss": 2.8897,
      "theoretical_loss": 3.604395298819692,
      "tokens_seen": 1140391936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305516549648947,
      "loss": 2.8563,
      "theoretical_loss": 3.604376062703313,
      "tokens_seen": 1140457472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033054162487462384,
      "loss": 2.7353,
      "theoretical_loss": 3.60435682800179,
      "tokens_seen": 1140523008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305315947843531,
      "loss": 2.7761,
      "theoretical_loss": 3.604337594714937,
      "tokens_seen": 1140588544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033052156469408225,
      "loss": 2.6665,
      "theoretical_loss": 3.6043183628425695,
      "tokens_seen": 1140654080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033051153460381144,
      "loss": 2.8084,
      "theoretical_loss": 3.6042991323845017,
      "tokens_seen": 1140719616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305015045135406,
      "loss": 2.91,
      "theoretical_loss": 3.604279903340549,
      "tokens_seen": 1140785152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304914744232698,
      "loss": 2.5918,
      "theoretical_loss": 3.6042606757105258,
      "tokens_seen": 1140850688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000330481444332999,
      "loss": 2.7085,
      "theoretical_loss": 3.604241449494247,
      "tokens_seen": 1140916224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304714142427282,
      "loss": 2.6838,
      "theoretical_loss": 3.6042222246915276,
      "tokens_seen": 1140981760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033046138415245734,
      "loss": 2.591,
      "theoretical_loss": 3.604203001302182,
      "tokens_seen": 1141047296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304513540621866,
      "loss": 2.6663,
      "theoretical_loss": 3.604183779326026,
      "tokens_seen": 1141112832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304413239719157,
      "loss": 2.9393,
      "theoretical_loss": 3.604164558762873,
      "tokens_seen": 1141178368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033043129388164494,
      "loss": 2.6918,
      "theoretical_loss": 3.6041453396125394,
      "tokens_seen": 1141243904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304212637913741,
      "loss": 2.6888,
      "theoretical_loss": 3.60412612187484,
      "tokens_seen": 1141309440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304112337011033,
      "loss": 2.7761,
      "theoretical_loss": 3.604106905549589,
      "tokens_seen": 1141374976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304012036108325,
      "loss": 2.8362,
      "theoretical_loss": 3.604087690636603,
      "tokens_seen": 1141440512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303911735205617,
      "loss": 2.6173,
      "theoretical_loss": 3.6040684771356957,
      "tokens_seen": 1141506048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033038114343029084,
      "loss": 2.7558,
      "theoretical_loss": 3.604049265046683,
      "tokens_seen": 1141571584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303711133400201,
      "loss": 2.8015,
      "theoretical_loss": 3.60403005436938,
      "tokens_seen": 1141637120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303610832497492,
      "loss": 2.7054,
      "theoretical_loss": 3.604010845103602,
      "tokens_seen": 1141702656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033035105315947844,
      "loss": 2.8058,
      "theoretical_loss": 3.6039916372491643,
      "tokens_seen": 1141768192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303410230692076,
      "loss": 2.8196,
      "theoretical_loss": 3.6039724308058814,
      "tokens_seen": 1141833728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303309929789368,
      "loss": 2.581,
      "theoretical_loss": 3.60395322577357,
      "tokens_seen": 1141899264
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 633465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.615844488143921,
      "objective/train/theoretical_loss": 3.6039340221520444,
      "objective/train/tokens_used": 1162424800,
      "theoretical_loss": 3.6039340221520444,
      "tokens_seen": 1141964800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033032096288866604,
      "loss": 2.6319,
      "theoretical_loss": 3.6039340221520444,
      "tokens_seen": 1141964800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033031093279839517,
      "loss": 2.7764,
      "theoretical_loss": 3.6039148199411204,
      "tokens_seen": 1142030336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303009027081244,
      "loss": 2.768,
      "theoretical_loss": 3.6038956191406135,
      "tokens_seen": 1142095872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302908726178536,
      "loss": 2.7994,
      "theoretical_loss": 3.6038764197503395,
      "tokens_seen": 1142161408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033028084252758276,
      "loss": 2.818,
      "theoretical_loss": 3.603857221770113,
      "tokens_seen": 1142226944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033027081243731195,
      "loss": 2.7977,
      "theoretical_loss": 3.6038380251997504,
      "tokens_seen": 1142292480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302607823470412,
      "loss": 2.7323,
      "theoretical_loss": 3.603818830039067,
      "tokens_seen": 1142358016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302507522567703,
      "loss": 2.6949,
      "theoretical_loss": 3.603799636287878,
      "tokens_seen": 1142423552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033024072216649954,
      "loss": 2.7476,
      "theoretical_loss": 3.6037804439459995,
      "tokens_seen": 1142489088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033023069207622867,
      "loss": 2.8042,
      "theoretical_loss": 3.6037612530132477,
      "tokens_seen": 1142554624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302206619859579,
      "loss": 2.832,
      "theoretical_loss": 3.6037420634894373,
      "tokens_seen": 1142620160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302106318956871,
      "loss": 2.5317,
      "theoretical_loss": 3.6037228753743844,
      "tokens_seen": 1142685696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033020060180541627,
      "loss": 2.791,
      "theoretical_loss": 3.6037036886679052,
      "tokens_seen": 1142751232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033019057171514545,
      "loss": 2.543,
      "theoretical_loss": 3.603684503369815,
      "tokens_seen": 1142816768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033018054162487463,
      "loss": 2.9402,
      "theoretical_loss": 3.60366531947993,
      "tokens_seen": 1142882304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301705115346038,
      "loss": 2.9351,
      "theoretical_loss": 3.6036461369980657,
      "tokens_seen": 1142947840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033016048144433305,
      "loss": 2.765,
      "theoretical_loss": 3.603626955924039,
      "tokens_seen": 1143013376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301504513540622,
      "loss": 2.6821,
      "theoretical_loss": 3.6036077762576646,
      "tokens_seen": 1143078912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301404212637914,
      "loss": 2.6081,
      "theoretical_loss": 3.603588597998759,
      "tokens_seen": 1143144448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033013039117352054,
      "loss": 2.854,
      "theoretical_loss": 3.6035694211471387,
      "tokens_seen": 1143209984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033012036108324977,
      "loss": 2.6444,
      "theoretical_loss": 3.6035502457026194,
      "tokens_seen": 1143275520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033011033099297895,
      "loss": 2.5662,
      "theoretical_loss": 3.603531071665017,
      "tokens_seen": 1143341056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033010030090270813,
      "loss": 2.7041,
      "theoretical_loss": 3.6035118990341477,
      "tokens_seen": 1143406592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300902708124373,
      "loss": 2.7002,
      "theoretical_loss": 3.6034927278098277,
      "tokens_seen": 1143472128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033008024072216655,
      "loss": 2.7704,
      "theoretical_loss": 3.6034735579918737,
      "tokens_seen": 1143537664
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 634749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3306100368499756,
      "objective/train/theoretical_loss": 3.6034543895801017,
      "objective/train/tokens_used": 1164063200,
      "theoretical_loss": 3.6034543895801017,
      "tokens_seen": 1143603200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300702106318957,
      "loss": 2.654,
      "theoretical_loss": 3.6034543895801017,
      "tokens_seen": 1143603200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300601805416249,
      "loss": 2.7415,
      "theoretical_loss": 3.603435222574327,
      "tokens_seen": 1143668736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033005015045135404,
      "loss": 2.6088,
      "theoretical_loss": 3.6034160569743676,
      "tokens_seen": 1143734272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300401203610833,
      "loss": 2.5878,
      "theoretical_loss": 3.6033968927800384,
      "tokens_seen": 1143799808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033003009027081245,
      "loss": 2.7632,
      "theoretical_loss": 3.6033777299911565,
      "tokens_seen": 1143865344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033002006018054164,
      "loss": 2.9358,
      "theoretical_loss": 3.6033585686075384,
      "tokens_seen": 1143930880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300100300902708,
      "loss": 2.7065,
      "theoretical_loss": 3.603339408629,
      "tokens_seen": 1143996416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033,
      "loss": 2.7274,
      "theoretical_loss": 3.6033202500553587,
      "tokens_seen": 1144061952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299899699097292,
      "loss": 2.6696,
      "theoretical_loss": 3.6033010928864297,
      "tokens_seen": 1144127488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299799398194584,
      "loss": 2.6557,
      "theoretical_loss": 3.6032819371220306,
      "tokens_seen": 1144193024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032996990972918754,
      "loss": 2.838,
      "theoretical_loss": 3.603262782761978,
      "tokens_seen": 1144258560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299598796389168,
      "loss": 2.8272,
      "theoretical_loss": 3.6032436298060873,
      "tokens_seen": 1144324096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299498495486459,
      "loss": 2.7172,
      "theoretical_loss": 3.6032244782541767,
      "tokens_seen": 1144389632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032993981945837514,
      "loss": 2.7073,
      "theoretical_loss": 3.6032053281060623,
      "tokens_seen": 1144455168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299297893681043,
      "loss": 2.7657,
      "theoretical_loss": 3.6031861793615603,
      "tokens_seen": 1144520704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299197592778335,
      "loss": 2.8131,
      "theoretical_loss": 3.6031670320204885,
      "tokens_seen": 1144586240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299097291875627,
      "loss": 2.7006,
      "theoretical_loss": 3.6031478860826627,
      "tokens_seen": 1144651776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298996990972919,
      "loss": 2.7931,
      "theoretical_loss": 3.6031287415479003,
      "tokens_seen": 1144717312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032988966900702104,
      "loss": 2.7996,
      "theoretical_loss": 3.603109598416018,
      "tokens_seen": 1144782848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298796389167503,
      "loss": 2.4818,
      "theoretical_loss": 3.6030904566868323,
      "tokens_seen": 1144848384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298696088264794,
      "loss": 2.9172,
      "theoretical_loss": 3.603071316360161,
      "tokens_seen": 1144913920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032985957873620864,
      "loss": 2.7094,
      "theoretical_loss": 3.6030521774358206,
      "tokens_seen": 1144979456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298495486459378,
      "loss": 2.7017,
      "theoretical_loss": 3.603033039913628,
      "tokens_seen": 1145044992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329839518555667,
      "loss": 2.7211,
      "theoretical_loss": 3.6030139037934,
      "tokens_seen": 1145110528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298294884653962,
      "loss": 2.6337,
      "theoretical_loss": 3.602994769074954,
      "tokens_seen": 1145176064
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 635305,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9339799880981445,
      "objective/train/theoretical_loss": 3.6029756357581073,
      "objective/train/tokens_used": 1165701600,
      "theoretical_loss": 3.6029756357581073,
      "tokens_seen": 1145241600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032981945837512537,
      "loss": 2.6328,
      "theoretical_loss": 3.6029756357581073,
      "tokens_seen": 1145241600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032980942828485455,
      "loss": 2.7999,
      "theoretical_loss": 3.6029565038426767,
      "tokens_seen": 1145307136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297993981945838,
      "loss": 2.8183,
      "theoretical_loss": 3.60293737332848,
      "tokens_seen": 1145372672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297893681043129,
      "loss": 2.8277,
      "theoretical_loss": 3.602918244215333,
      "tokens_seen": 1145438208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032977933801404215,
      "loss": 2.6376,
      "theoretical_loss": 3.602899116503055,
      "tokens_seen": 1145503744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297693079237713,
      "loss": 2.7554,
      "theoretical_loss": 3.6028799901914614,
      "tokens_seen": 1145569280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297592778335005,
      "loss": 2.8153,
      "theoretical_loss": 3.6028608652803706,
      "tokens_seen": 1145634816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297492477432297,
      "loss": 2.4666,
      "theoretical_loss": 3.6028417417695993,
      "tokens_seen": 1145700352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032973921765295887,
      "loss": 2.726,
      "theoretical_loss": 3.6028226196589648,
      "tokens_seen": 1145765888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032972918756268805,
      "loss": 2.6854,
      "theoretical_loss": 3.6028034989482856,
      "tokens_seen": 1145831424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297191574724173,
      "loss": 2.5511,
      "theoretical_loss": 3.602784379637378,
      "tokens_seen": 1145896960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297091273821464,
      "loss": 2.547,
      "theoretical_loss": 3.60276526172606,
      "tokens_seen": 1145962496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032969909729187565,
      "loss": 2.7832,
      "theoretical_loss": 3.602746145214149,
      "tokens_seen": 1146028032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296890672016048,
      "loss": 2.6622,
      "theoretical_loss": 3.6027270301014624,
      "tokens_seen": 1146093568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329679037111334,
      "loss": 2.6686,
      "theoretical_loss": 3.602707916387818,
      "tokens_seen": 1146159104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296690070210632,
      "loss": 2.7567,
      "theoretical_loss": 3.6026888040730336,
      "tokens_seen": 1146224640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296589769307924,
      "loss": 2.6816,
      "theoretical_loss": 3.602669693156926,
      "tokens_seen": 1146290176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032964894684052155,
      "loss": 2.829,
      "theoretical_loss": 3.6026505836393143,
      "tokens_seen": 1146355712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032963891675025074,
      "loss": 2.7428,
      "theoretical_loss": 3.602631475520015,
      "tokens_seen": 1146421248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296288866599799,
      "loss": 2.6902,
      "theoretical_loss": 3.6026123687988463,
      "tokens_seen": 1146486784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032961885656970915,
      "loss": 2.7686,
      "theoretical_loss": 3.602593263475626,
      "tokens_seen": 1146552320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296088264794383,
      "loss": 2.7541,
      "theoretical_loss": 3.602574159550172,
      "tokens_seen": 1146617856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295987963891675,
      "loss": 2.4678,
      "theoretical_loss": 3.6025550570223013,
      "tokens_seen": 1146683392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295887662988967,
      "loss": 2.6954,
      "theoretical_loss": 3.6025359558918333,
      "tokens_seen": 1146748928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295787362086259,
      "loss": 2.6374,
      "theoretical_loss": 3.6025168561585845,
      "tokens_seen": 1146814464
    },
    {
      "debugging/Self-BLEU-5": 0.4786247835242382,
      "debugging/distinct-1-grams": 0.7609861614906261,
      "debugging/distinct-2-grams": 0.9220814316505062,
      "debugging/entropy-1-grams": 5.880984152522802,
      "debugging/entropy-2-grams": 6.5464310105420465,
      "debugging/length": 573.6,
      "debugging/num_segments": 10,
      "epoch": 0.35,
      "objective/train/docs_used": 636617,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.868852376937866,
      "objective/train/theoretical_loss": 3.6024977578223742,
      "objective/train/tokens_used": 1167340000,
      "theoretical_loss": 3.6024977578223742,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295687061183551,
      "loss": 2.8676,
      "theoretical_loss": 3.6024977578223742,
      "tokens_seen": 1146880000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032955867602808424,
      "loss": 2.6701,
      "theoretical_loss": 3.6024786608830195,
      "tokens_seen": 1146945536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295486459378135,
      "loss": 2.9224,
      "theoretical_loss": 3.6024595653403386,
      "tokens_seen": 1147011072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032953861584754265,
      "loss": 2.8055,
      "theoretical_loss": 3.602440471194149,
      "tokens_seen": 1147076608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032952858575727184,
      "loss": 2.686,
      "theoretical_loss": 3.6024213784442702,
      "tokens_seen": 1147142144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329518555667001,
      "loss": 2.8341,
      "theoretical_loss": 3.602402287090519,
      "tokens_seen": 1147207680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295085255767302,
      "loss": 2.8485,
      "theoretical_loss": 3.6023831971327143,
      "tokens_seen": 1147273216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294984954864594,
      "loss": 2.6909,
      "theoretical_loss": 3.602364108570674,
      "tokens_seen": 1147338752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294884653961886,
      "loss": 2.7537,
      "theoretical_loss": 3.6023450214042163,
      "tokens_seen": 1147404288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032947843530591774,
      "loss": 2.8729,
      "theoretical_loss": 3.60232593563316,
      "tokens_seen": 1147469824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329468405215647,
      "loss": 2.9974,
      "theoretical_loss": 3.602306851257322,
      "tokens_seen": 1147535360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294583751253761,
      "loss": 2.6533,
      "theoretical_loss": 3.602287768276523,
      "tokens_seen": 1147600896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032944834503510534,
      "loss": 2.793,
      "theoretical_loss": 3.6022686866905786,
      "tokens_seen": 1147666432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294383149448345,
      "loss": 2.6353,
      "theoretical_loss": 3.602249606499309,
      "tokens_seen": 1147731968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294282848545637,
      "loss": 2.8417,
      "theoretical_loss": 3.602230527702533,
      "tokens_seen": 1147797504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294182547642929,
      "loss": 2.7079,
      "theoretical_loss": 3.602211450300067,
      "tokens_seen": 1147863040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294082246740221,
      "loss": 2.8198,
      "theoretical_loss": 3.6021923742917314,
      "tokens_seen": 1147928576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032939819458375124,
      "loss": 2.6095,
      "theoretical_loss": 3.602173299677344,
      "tokens_seen": 1147994112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293881644934805,
      "loss": 2.7788,
      "theoretical_loss": 3.6021542264567237,
      "tokens_seen": 1148059648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293781344032096,
      "loss": 2.7369,
      "theoretical_loss": 3.6021351546296887,
      "tokens_seen": 1148125184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032936810431293884,
      "loss": 2.6818,
      "theoretical_loss": 3.6021160841960578,
      "tokens_seen": 1148190720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329358074222668,
      "loss": 2.8319,
      "theoretical_loss": 3.6020970151556497,
      "tokens_seen": 1148256256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293480441323972,
      "loss": 2.6298,
      "theoretical_loss": 3.602077947508283,
      "tokens_seen": 1148321792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293380140421264,
      "loss": 2.7902,
      "theoretical_loss": 3.602058881253777,
      "tokens_seen": 1148387328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032932798395185557,
      "loss": 2.7109,
      "theoretical_loss": 3.6020398163919496,
      "tokens_seen": 1148452864
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 637752,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6599230766296387,
      "objective/train/theoretical_loss": 3.6020207529226203,
      "objective/train/tokens_used": 1168978400,
      "theoretical_loss": 3.6020207529226203,
      "tokens_seen": 1148518400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032931795386158475,
      "loss": 2.7556,
      "theoretical_loss": 3.6020207529226203,
      "tokens_seen": 1148518400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329307923771314,
      "loss": 2.7028,
      "theoretical_loss": 3.6020016908456074,
      "tokens_seen": 1148583936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292978936810431,
      "loss": 2.6435,
      "theoretical_loss": 3.6019826301607303,
      "tokens_seen": 1148649472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032928786359077235,
      "loss": 2.7049,
      "theoretical_loss": 3.6019635708678077,
      "tokens_seen": 1148715008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292778335005015,
      "loss": 2.7472,
      "theoretical_loss": 3.601944512966658,
      "tokens_seen": 1148780544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292678034102307,
      "loss": 2.7465,
      "theoretical_loss": 3.601925456457101,
      "tokens_seen": 1148846080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292577733199599,
      "loss": 2.5858,
      "theoretical_loss": 3.6019064013389555,
      "tokens_seen": 1148911616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032924774322968907,
      "loss": 2.8702,
      "theoretical_loss": 3.60188734761204,
      "tokens_seen": 1148977152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032923771313941825,
      "loss": 2.6107,
      "theoretical_loss": 3.6018682952761747,
      "tokens_seen": 1149042688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292276830491475,
      "loss": 2.8954,
      "theoretical_loss": 3.6018492443311776,
      "tokens_seen": 1149108224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292176529588766,
      "loss": 2.5675,
      "theoretical_loss": 3.6018301947768685,
      "tokens_seen": 1149173760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032920762286860585,
      "loss": 2.7175,
      "theoretical_loss": 3.6018111466130662,
      "tokens_seen": 1149239296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329197592778335,
      "loss": 2.7522,
      "theoretical_loss": 3.6017920998395896,
      "tokens_seen": 1149304832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291875626880642,
      "loss": 2.6172,
      "theoretical_loss": 3.6017730544562596,
      "tokens_seen": 1149370368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291775325977934,
      "loss": 2.6186,
      "theoretical_loss": 3.6017540104628933,
      "tokens_seen": 1149435904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291675025075226,
      "loss": 2.8122,
      "theoretical_loss": 3.6017349678593114,
      "tokens_seen": 1149501440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032915747241725175,
      "loss": 2.5775,
      "theoretical_loss": 3.601715926645333,
      "tokens_seen": 1149566976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032914744232698094,
      "loss": 2.7746,
      "theoretical_loss": 3.6016968868207773,
      "tokens_seen": 1149632512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291374122367101,
      "loss": 2.5421,
      "theoretical_loss": 3.6016778483854637,
      "tokens_seen": 1149698048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032912738214643935,
      "loss": 2.4038,
      "theoretical_loss": 3.6016588113392114,
      "tokens_seen": 1149763584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291173520561685,
      "loss": 2.7402,
      "theoretical_loss": 3.601639775681841,
      "tokens_seen": 1149829120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291073219658977,
      "loss": 2.7332,
      "theoretical_loss": 3.6016207414131705,
      "tokens_seen": 1149894656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290972918756269,
      "loss": 2.7302,
      "theoretical_loss": 3.6016017085330203,
      "tokens_seen": 1149960192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290872617853561,
      "loss": 2.661,
      "theoretical_loss": 3.6015826770412103,
      "tokens_seen": 1150025728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032907723169508526,
      "loss": 2.7305,
      "theoretical_loss": 3.601563646937559,
      "tokens_seen": 1150091264
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 638347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3350415229797363,
      "objective/train/theoretical_loss": 3.6015446182218875,
      "objective/train/tokens_used": 1170616800,
      "theoretical_loss": 3.6015446182218875,
      "tokens_seen": 1150156800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032906720160481444,
      "loss": 2.6234,
      "theoretical_loss": 3.6015446182218875,
      "tokens_seen": 1150156800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290571715145436,
      "loss": 2.5445,
      "theoretical_loss": 3.6015255908940147,
      "tokens_seen": 1150222336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032904714142427286,
      "loss": 2.6305,
      "theoretical_loss": 3.60150656495376,
      "tokens_seen": 1150287872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329037111334002,
      "loss": 2.7704,
      "theoretical_loss": 3.6014875404009437,
      "tokens_seen": 1150353408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290270812437312,
      "loss": 2.8285,
      "theoretical_loss": 3.601468517235385,
      "tokens_seen": 1150418944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032901705115346034,
      "loss": 2.7081,
      "theoretical_loss": 3.6014494954569045,
      "tokens_seen": 1150484480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290070210631896,
      "loss": 2.7436,
      "theoretical_loss": 3.601430475065322,
      "tokens_seen": 1150550016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032899699097291876,
      "loss": 2.7384,
      "theoretical_loss": 3.6014114560604567,
      "tokens_seen": 1150615552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032898696088264794,
      "loss": 2.7074,
      "theoretical_loss": 3.6013924384421285,
      "tokens_seen": 1150681088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289769307923771,
      "loss": 2.671,
      "theoretical_loss": 3.6013734222101585,
      "tokens_seen": 1150746624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289669007021063,
      "loss": 2.6452,
      "theoretical_loss": 3.6013544073643655,
      "tokens_seen": 1150812160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289568706118355,
      "loss": 2.5962,
      "theoretical_loss": 3.60133539390457,
      "tokens_seen": 1150877696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289468405215647,
      "loss": 2.8025,
      "theoretical_loss": 3.601316381830592,
      "tokens_seen": 1150943232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032893681043129385,
      "loss": 2.6795,
      "theoretical_loss": 3.601297371142252,
      "tokens_seen": 1151008768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289267803410231,
      "loss": 2.7709,
      "theoretical_loss": 3.6012783618393693,
      "tokens_seen": 1151074304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032891675025075226,
      "loss": 2.7108,
      "theoretical_loss": 3.601259353921765,
      "tokens_seen": 1151139840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032890672016048145,
      "loss": 2.8169,
      "theoretical_loss": 3.6012403473892585,
      "tokens_seen": 1151205376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288966900702106,
      "loss": 2.7462,
      "theoretical_loss": 3.6012213422416703,
      "tokens_seen": 1151270912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288866599799398,
      "loss": 2.6478,
      "theoretical_loss": 3.6012023384788208,
      "tokens_seen": 1151336448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328876629889669,
      "loss": 2.7033,
      "theoretical_loss": 3.60118333610053,
      "tokens_seen": 1151401984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288665997993982,
      "loss": 2.7022,
      "theoretical_loss": 3.6011643351066183,
      "tokens_seen": 1151467520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032885656970912735,
      "loss": 2.8157,
      "theoretical_loss": 3.6011453354969065,
      "tokens_seen": 1151533056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288465396188566,
      "loss": 2.6348,
      "theoretical_loss": 3.6011263372712143,
      "tokens_seen": 1151598592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288365095285857,
      "loss": 2.7876,
      "theoretical_loss": 3.6011073404293628,
      "tokens_seen": 1151664128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032882647943831495,
      "loss": 2.5848,
      "theoretical_loss": 3.601088344971172,
      "tokens_seen": 1151729664
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 639523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5807628631591797,
      "objective/train/theoretical_loss": 3.6010693508964624,
      "objective/train/tokens_used": 1172255200,
      "theoretical_loss": 3.6010693508964624,
      "tokens_seen": 1151795200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288164493480442,
      "loss": 2.5903,
      "theoretical_loss": 3.6010693508964624,
      "tokens_seen": 1151795200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288064192577733,
      "loss": 2.5482,
      "theoretical_loss": 3.601050358205055,
      "tokens_seen": 1151860736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032879638916750255,
      "loss": 2.7664,
      "theoretical_loss": 3.6010313668967697,
      "tokens_seen": 1151926272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287863590772317,
      "loss": 2.7388,
      "theoretical_loss": 3.6010123769714273,
      "tokens_seen": 1151991808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287763289869609,
      "loss": 2.7822,
      "theoretical_loss": 3.600993388428849,
      "tokens_seen": 1152057344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287662988966901,
      "loss": 2.8795,
      "theoretical_loss": 3.600974401268855,
      "tokens_seen": 1152122880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032875626880641927,
      "loss": 2.5308,
      "theoretical_loss": 3.600955415491266,
      "tokens_seen": 1152188416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032874623871614845,
      "loss": 2.9124,
      "theoretical_loss": 3.6009364310959024,
      "tokens_seen": 1152253952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287362086258777,
      "loss": 2.7283,
      "theoretical_loss": 3.6009174480825856,
      "tokens_seen": 1152319488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287261785356068,
      "loss": 2.7157,
      "theoretical_loss": 3.6008984664511363,
      "tokens_seen": 1152385024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032871614844533605,
      "loss": 2.5741,
      "theoretical_loss": 3.6008794862013747,
      "tokens_seen": 1152450560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287061183550652,
      "loss": 2.9161,
      "theoretical_loss": 3.600860507333122,
      "tokens_seen": 1152516096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286960882647944,
      "loss": 2.6876,
      "theoretical_loss": 3.6008415298461993,
      "tokens_seen": 1152581632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286860581745236,
      "loss": 2.7743,
      "theoretical_loss": 3.6008225537404277,
      "tokens_seen": 1152647168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286760280842528,
      "loss": 2.8304,
      "theoretical_loss": 3.6008035790156274,
      "tokens_seen": 1152712704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032866599799398195,
      "loss": 2.9008,
      "theoretical_loss": 3.6007846056716204,
      "tokens_seen": 1152778240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032865596790371114,
      "loss": 2.7073,
      "theoretical_loss": 3.6007656337082268,
      "tokens_seen": 1152843776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286459378134403,
      "loss": 2.4973,
      "theoretical_loss": 3.6007466631252685,
      "tokens_seen": 1152909312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032863590772316955,
      "loss": 2.6311,
      "theoretical_loss": 3.6007276939225656,
      "tokens_seen": 1152974848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286258776328987,
      "loss": 2.5766,
      "theoretical_loss": 3.60070872609994,
      "tokens_seen": 1153040384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286158475426279,
      "loss": 2.5913,
      "theoretical_loss": 3.6006897596572127,
      "tokens_seen": 1153105920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286058174523571,
      "loss": 2.5173,
      "theoretical_loss": 3.6006707945942047,
      "tokens_seen": 1153171456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285957873620863,
      "loss": 2.6512,
      "theoretical_loss": 3.6006518309107376,
      "tokens_seen": 1153236992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032858575727181546,
      "loss": 2.8246,
      "theoretical_loss": 3.6006328686066325,
      "tokens_seen": 1153302528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032857572718154464,
      "loss": 2.8009,
      "theoretical_loss": 3.6006139076817103,
      "tokens_seen": 1153368064
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 639983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.798460006713867,
      "objective/train/theoretical_loss": 3.600594948135793,
      "objective/train/tokens_used": 1173893600,
      "theoretical_loss": 3.600594948135793,
      "tokens_seen": 1153433600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285656970912738,
      "loss": 2.6424,
      "theoretical_loss": 3.600594948135793,
      "tokens_seen": 1153433600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032855566700100306,
      "loss": 2.6186,
      "theoretical_loss": 3.6005759899687018,
      "tokens_seen": 1153499136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285456369107322,
      "loss": 2.8315,
      "theoretical_loss": 3.6005570331802574,
      "tokens_seen": 1153564672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285356068204614,
      "loss": 2.7475,
      "theoretical_loss": 3.600538077770282,
      "tokens_seen": 1153630208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032852557673019054,
      "loss": 2.7912,
      "theoretical_loss": 3.6005191237385965,
      "tokens_seen": 1153695744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285155466399198,
      "loss": 2.6463,
      "theoretical_loss": 3.6005001710850233,
      "tokens_seen": 1153761280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032850551654964896,
      "loss": 2.5891,
      "theoretical_loss": 3.6004812198093834,
      "tokens_seen": 1153826816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032849548645937814,
      "loss": 2.4599,
      "theoretical_loss": 3.600462269911498,
      "tokens_seen": 1153892352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284854563691073,
      "loss": 2.6544,
      "theoretical_loss": 3.6004433213911886,
      "tokens_seen": 1153957888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284754262788365,
      "loss": 2.5024,
      "theoretical_loss": 3.6004243742482775,
      "tokens_seen": 1154023424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284653961885657,
      "loss": 2.5078,
      "theoretical_loss": 3.600405428482586,
      "tokens_seen": 1154088960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284553660982949,
      "loss": 2.6429,
      "theoretical_loss": 3.600386484093936,
      "tokens_seen": 1154154496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032844533600802405,
      "loss": 2.6753,
      "theoretical_loss": 3.600367541082149,
      "tokens_seen": 1154220032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284353059177533,
      "loss": 2.6927,
      "theoretical_loss": 3.6003485994470474,
      "tokens_seen": 1154285568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032842527582748246,
      "loss": 2.5926,
      "theoretical_loss": 3.600329659188452,
      "tokens_seen": 1154351104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032841524573721165,
      "loss": 2.6645,
      "theoretical_loss": 3.6003107203061853,
      "tokens_seen": 1154416640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284052156469408,
      "loss": 2.8488,
      "theoretical_loss": 3.6002917828000687,
      "tokens_seen": 1154482176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032839518555667,
      "loss": 2.6092,
      "theoretical_loss": 3.6002728466699248,
      "tokens_seen": 1154547712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283851554663992,
      "loss": 2.67,
      "theoretical_loss": 3.6002539119155745,
      "tokens_seen": 1154613248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283751253761284,
      "loss": 2.57,
      "theoretical_loss": 3.6002349785368413,
      "tokens_seen": 1154678784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032836509528585755,
      "loss": 2.594,
      "theoretical_loss": 3.6002160465335455,
      "tokens_seen": 1154744320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283550651955868,
      "loss": 2.8139,
      "theoretical_loss": 3.6001971159055097,
      "tokens_seen": 1154809856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283450351053159,
      "loss": 2.6035,
      "theoretical_loss": 3.600178186652556,
      "tokens_seen": 1154875392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032833500501504515,
      "loss": 2.8044,
      "theoretical_loss": 3.6001592587745073,
      "tokens_seen": 1154940928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032832497492477433,
      "loss": 2.6125,
      "theoretical_loss": 3.6001403322711845,
      "tokens_seen": 1155006464
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 641200,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4352283477783203,
      "objective/train/theoretical_loss": 3.6001214071424106,
      "objective/train/tokens_used": 1175532000,
      "theoretical_loss": 3.6001214071424106,
      "tokens_seen": 1155072000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283149448345035,
      "loss": 2.718,
      "theoretical_loss": 3.6001214071424106,
      "tokens_seen": 1155072000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283049147442327,
      "loss": 2.7494,
      "theoretical_loss": 3.6001024833880075,
      "tokens_seen": 1155137536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282948846539619,
      "loss": 2.7,
      "theoretical_loss": 3.6000835610077973,
      "tokens_seen": 1155203072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032828485456369105,
      "loss": 2.6363,
      "theoretical_loss": 3.6000646400016025,
      "tokens_seen": 1155268608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282748244734203,
      "loss": 2.6263,
      "theoretical_loss": 3.6000457203692453,
      "tokens_seen": 1155334144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282647943831494,
      "loss": 2.5974,
      "theoretical_loss": 3.6000268021105475,
      "tokens_seen": 1155399680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032825476429287865,
      "loss": 2.9712,
      "theoretical_loss": 3.6000078852253328,
      "tokens_seen": 1155465216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032824473420260783,
      "loss": 2.5518,
      "theoretical_loss": 3.599988969713422,
      "tokens_seen": 1155530752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328234704112337,
      "loss": 2.6989,
      "theoretical_loss": 3.5999700555746386,
      "tokens_seen": 1155596288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282246740220662,
      "loss": 2.7771,
      "theoretical_loss": 3.5999511428088047,
      "tokens_seen": 1155661824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282146439317954,
      "loss": 2.6778,
      "theoretical_loss": 3.5999322314157425,
      "tokens_seen": 1155727360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032820461384152456,
      "loss": 2.5627,
      "theoretical_loss": 3.5999133213952756,
      "tokens_seen": 1155792896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281945837512538,
      "loss": 2.9634,
      "theoretical_loss": 3.599894412747225,
      "tokens_seen": 1155858432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281845536609829,
      "loss": 2.8353,
      "theoretical_loss": 3.599875505471415,
      "tokens_seen": 1155923968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032817452357071215,
      "loss": 2.9044,
      "theoretical_loss": 3.599856599567666,
      "tokens_seen": 1155989504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281644934804413,
      "loss": 2.6862,
      "theoretical_loss": 3.599837695035803,
      "tokens_seen": 1156055040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281544633901705,
      "loss": 2.8285,
      "theoretical_loss": 3.5998187918756472,
      "tokens_seen": 1156120576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281444332998997,
      "loss": 2.8731,
      "theoretical_loss": 3.599799890087022,
      "tokens_seen": 1156186112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281344032096289,
      "loss": 2.6179,
      "theoretical_loss": 3.5997809896697497,
      "tokens_seen": 1156251648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032812437311935806,
      "loss": 2.5605,
      "theoretical_loss": 3.5997620906236536,
      "tokens_seen": 1156317184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281143430290873,
      "loss": 2.6217,
      "theoretical_loss": 3.599743192948556,
      "tokens_seen": 1156382720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281043129388164,
      "loss": 2.7596,
      "theoretical_loss": 3.59972429664428,
      "tokens_seen": 1156448256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032809428284854566,
      "loss": 2.5259,
      "theoretical_loss": 3.599705401710648,
      "tokens_seen": 1156513792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280842527582748,
      "loss": 2.473,
      "theoretical_loss": 3.5996865081474843,
      "tokens_seen": 1156579328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328074222668004,
      "loss": 2.6603,
      "theoretical_loss": 3.59966761595461,
      "tokens_seen": 1156644864
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 641801,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.768510341644287,
      "objective/train/theoretical_loss": 3.59964872513185,
      "objective/train/tokens_used": 1177170400,
      "theoretical_loss": 3.59964872513185,
      "tokens_seen": 1156710400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032806419257773326,
      "loss": 2.7684,
      "theoretical_loss": 3.59964872513185,
      "tokens_seen": 1156710400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280541624874624,
      "loss": 2.7661,
      "theoretical_loss": 3.5996298356790257,
      "tokens_seen": 1156775936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280441323971916,
      "loss": 2.5801,
      "theoretical_loss": 3.599610947595961,
      "tokens_seen": 1156841472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032803410230692074,
      "loss": 2.8947,
      "theoretical_loss": 3.599592060882479,
      "tokens_seen": 1156907008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032802407221665,
      "loss": 2.8118,
      "theoretical_loss": 3.599573175538402,
      "tokens_seen": 1156972544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032801404212637916,
      "loss": 2.713,
      "theoretical_loss": 3.599554291563554,
      "tokens_seen": 1157038080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032800401203610834,
      "loss": 2.6654,
      "theoretical_loss": 3.5995354089577583,
      "tokens_seen": 1157103616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279939819458375,
      "loss": 2.4662,
      "theoretical_loss": 3.599516527720837,
      "tokens_seen": 1157169152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279839518555667,
      "loss": 2.7589,
      "theoretical_loss": 3.5994976478526146,
      "tokens_seen": 1157234688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279739217652959,
      "loss": 2.7968,
      "theoretical_loss": 3.599478769352914,
      "tokens_seen": 1157300224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279638916750251,
      "loss": 2.5783,
      "theoretical_loss": 3.5994598922215584,
      "tokens_seen": 1157365760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032795386158475425,
      "loss": 2.662,
      "theoretical_loss": 3.5994410164583703,
      "tokens_seen": 1157431296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279438314944835,
      "loss": 2.5359,
      "theoretical_loss": 3.5994221420631747,
      "tokens_seen": 1157496832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032793380140421266,
      "loss": 2.7496,
      "theoretical_loss": 3.5994032690357938,
      "tokens_seen": 1157562368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032792377131394185,
      "loss": 2.7392,
      "theoretical_loss": 3.5993843973760518,
      "tokens_seen": 1157627904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327913741223671,
      "loss": 2.6497,
      "theoretical_loss": 3.599365527083772,
      "tokens_seen": 1157693440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003279037111334002,
      "loss": 2.6444,
      "theoretical_loss": 3.599346658158777,
      "tokens_seen": 1157758976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278936810431294,
      "loss": 2.8481,
      "theoretical_loss": 3.599327790600891,
      "tokens_seen": 1157824512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278836509528586,
      "loss": 2.7322,
      "theoretical_loss": 3.5993089244099385,
      "tokens_seen": 1157890048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032787362086258775,
      "loss": 2.6314,
      "theoretical_loss": 3.5992900595857416,
      "tokens_seen": 1157955584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327863590772317,
      "loss": 2.5972,
      "theoretical_loss": 3.599271196128125,
      "tokens_seen": 1158021120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278535606820461,
      "loss": 2.6548,
      "theoretical_loss": 3.599252334036912,
      "tokens_seen": 1158086656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032784353059177535,
      "loss": 2.5857,
      "theoretical_loss": 3.5992334733119264,
      "tokens_seen": 1158152192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032783350050150453,
      "loss": 2.7087,
      "theoretical_loss": 3.5992146139529915,
      "tokens_seen": 1158217728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278234704112337,
      "loss": 2.5285,
      "theoretical_loss": 3.599195755959931,
      "tokens_seen": 1158283264
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 642800,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7663416862487793,
      "objective/train/theoretical_loss": 3.59917689933257,
      "objective/train/tokens_used": 1178808800,
      "theoretical_loss": 3.59917689933257,
      "tokens_seen": 1158348800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278134403209629,
      "loss": 2.7091,
      "theoretical_loss": 3.59917689933257,
      "tokens_seen": 1158348800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032780341023069213,
      "loss": 2.6744,
      "theoretical_loss": 3.599158044070731,
      "tokens_seen": 1158414336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032779338014042125,
      "loss": 2.5622,
      "theoretical_loss": 3.5991391901742382,
      "tokens_seen": 1158479872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277833500501505,
      "loss": 2.6717,
      "theoretical_loss": 3.599120337642916,
      "tokens_seen": 1158545408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277733199598796,
      "loss": 2.5785,
      "theoretical_loss": 3.599101486476587,
      "tokens_seen": 1158610944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032776328986960885,
      "loss": 2.5764,
      "theoretical_loss": 3.599082636675077,
      "tokens_seen": 1158676480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032775325977933803,
      "loss": 2.6596,
      "theoretical_loss": 3.599063788238209,
      "tokens_seen": 1158742016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277432296890672,
      "loss": 2.6874,
      "theoretical_loss": 3.599044941165807,
      "tokens_seen": 1158807552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277331995987964,
      "loss": 2.8803,
      "theoretical_loss": 3.5990260954576954,
      "tokens_seen": 1158873088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277231695085256,
      "loss": 2.793,
      "theoretical_loss": 3.599007251113698,
      "tokens_seen": 1158938624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032771313941825476,
      "loss": 2.7827,
      "theoretical_loss": 3.598988408133639,
      "tokens_seen": 1159004160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327703109327984,
      "loss": 2.7677,
      "theoretical_loss": 3.5989695665173427,
      "tokens_seen": 1159069696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276930792377131,
      "loss": 2.5865,
      "theoretical_loss": 3.598950726264633,
      "tokens_seen": 1159135232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032768304914744235,
      "loss": 2.6556,
      "theoretical_loss": 3.5989318873753344,
      "tokens_seen": 1159200768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276730190571715,
      "loss": 2.8157,
      "theoretical_loss": 3.5989130498492714,
      "tokens_seen": 1159266304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276629889669007,
      "loss": 2.6851,
      "theoretical_loss": 3.5988942136862674,
      "tokens_seen": 1159331840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276529588766299,
      "loss": 2.7396,
      "theoretical_loss": 3.598875378886148,
      "tokens_seen": 1159397376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276429287863591,
      "loss": 2.8136,
      "theoretical_loss": 3.5988565454487365,
      "tokens_seen": 1159462912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032763289869608826,
      "loss": 2.8422,
      "theoretical_loss": 3.598837713373858,
      "tokens_seen": 1159528448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276228686058175,
      "loss": 2.6567,
      "theoretical_loss": 3.598818882661336,
      "tokens_seen": 1159593984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276128385155466,
      "loss": 2.694,
      "theoretical_loss": 3.598800053310996,
      "tokens_seen": 1159659520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032760280842527586,
      "loss": 2.7281,
      "theoretical_loss": 3.5987812253226616,
      "tokens_seen": 1159725056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327592778335005,
      "loss": 2.7379,
      "theoretical_loss": 3.598762398696158,
      "tokens_seen": 1159790592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275827482447342,
      "loss": 2.8922,
      "theoretical_loss": 3.5987435734313094,
      "tokens_seen": 1159856128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275727181544634,
      "loss": 2.5817,
      "theoretical_loss": 3.598724749527941,
      "tokens_seen": 1159921664
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 643907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6200287342071533,
      "objective/train/theoretical_loss": 3.5987059269858763,
      "objective/train/tokens_used": 1180447200,
      "theoretical_loss": 3.5987059269858763,
      "tokens_seen": 1159987200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275626880641926,
      "loss": 2.7223,
      "theoretical_loss": 3.5987059269858763,
      "tokens_seen": 1159987200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032755265797392176,
      "loss": 2.7649,
      "theoretical_loss": 3.5986871058049408,
      "tokens_seen": 1160052736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032754262788365094,
      "loss": 2.5639,
      "theoretical_loss": 3.598668285984959,
      "tokens_seen": 1160118272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275325977933801,
      "loss": 2.7138,
      "theoretical_loss": 3.5986494675257554,
      "tokens_seen": 1160183808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032752256770310936,
      "loss": 2.6548,
      "theoretical_loss": 3.598630650427155,
      "tokens_seen": 1160249344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275125376128385,
      "loss": 2.6848,
      "theoretical_loss": 3.598611834688983,
      "tokens_seen": 1160314880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275025075225677,
      "loss": 2.7612,
      "theoretical_loss": 3.5985930203110628,
      "tokens_seen": 1160380416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032749247743229685,
      "loss": 2.9665,
      "theoretical_loss": 3.598574207293221,
      "tokens_seen": 1160445952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274824473420261,
      "loss": 2.8244,
      "theoretical_loss": 3.598555395635281,
      "tokens_seen": 1160511488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032747241725175527,
      "loss": 2.5306,
      "theoretical_loss": 3.5985365853370688,
      "tokens_seen": 1160577024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032746238716148445,
      "loss": 2.7923,
      "theoretical_loss": 3.5985177763984084,
      "tokens_seen": 1160642560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032745235707121363,
      "loss": 2.756,
      "theoretical_loss": 3.5984989688191256,
      "tokens_seen": 1160708096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032744232698094286,
      "loss": 2.6654,
      "theoretical_loss": 3.598480162599045,
      "tokens_seen": 1160773632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327432296890672,
      "loss": 2.8767,
      "theoretical_loss": 3.5984613577379916,
      "tokens_seen": 1160839168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274222668004012,
      "loss": 2.6347,
      "theoretical_loss": 3.598442554235791,
      "tokens_seen": 1160904704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032741223671013035,
      "loss": 2.6026,
      "theoretical_loss": 3.598423752092268,
      "tokens_seen": 1160970240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274022066198596,
      "loss": 2.5634,
      "theoretical_loss": 3.5984049513072467,
      "tokens_seen": 1161035776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032739217652958877,
      "loss": 2.6826,
      "theoretical_loss": 3.598386151880554,
      "tokens_seen": 1161101312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032738214643931795,
      "loss": 2.8304,
      "theoretical_loss": 3.598367353812014,
      "tokens_seen": 1161166848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032737211634904713,
      "loss": 2.6202,
      "theoretical_loss": 3.5983485571014526,
      "tokens_seen": 1161232384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273620862587763,
      "loss": 2.5217,
      "theoretical_loss": 3.5983297617486945,
      "tokens_seen": 1161297920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273520561685055,
      "loss": 2.6864,
      "theoretical_loss": 3.598310967753565,
      "tokens_seen": 1161363456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032734202607823473,
      "loss": 2.776,
      "theoretical_loss": 3.59829217511589,
      "tokens_seen": 1161428992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032733199598796386,
      "loss": 2.6287,
      "theoretical_loss": 3.5982733838354948,
      "tokens_seen": 1161494528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273219658976931,
      "loss": 2.5982,
      "theoretical_loss": 3.5982545939122037,
      "tokens_seen": 1161560064
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 644495,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7944414615631104,
      "objective/train/theoretical_loss": 3.598235805345843,
      "objective/train/tokens_used": 1182085600,
      "theoretical_loss": 3.598235805345843,
      "tokens_seen": 1161625600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032731193580742233,
      "loss": 2.7485,
      "theoretical_loss": 3.598235805345843,
      "tokens_seen": 1161625600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032730190571715145,
      "loss": 2.7042,
      "theoretical_loss": 3.598217018136239,
      "tokens_seen": 1161691136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272918756268807,
      "loss": 2.6849,
      "theoretical_loss": 3.598198232283216,
      "tokens_seen": 1161756672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272818455366098,
      "loss": 2.5112,
      "theoretical_loss": 3.598179447786599,
      "tokens_seen": 1161822208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032727181544633905,
      "loss": 2.5599,
      "theoretical_loss": 3.598160664646215,
      "tokens_seen": 1161887744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032726178535606823,
      "loss": 2.4779,
      "theoretical_loss": 3.598141882861889,
      "tokens_seen": 1161953280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272517552657974,
      "loss": 2.508,
      "theoretical_loss": 3.598123102433447,
      "tokens_seen": 1162018816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272417251755266,
      "loss": 2.7165,
      "theoretical_loss": 3.5981043233607135,
      "tokens_seen": 1162084352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272316950852558,
      "loss": 2.6167,
      "theoretical_loss": 3.5980855456435155,
      "tokens_seen": 1162149888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032722166499498496,
      "loss": 2.6884,
      "theoretical_loss": 3.5980667692816777,
      "tokens_seen": 1162215424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272116349047142,
      "loss": 2.5378,
      "theoretical_loss": 3.598047994275026,
      "tokens_seen": 1162280960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272016048144433,
      "loss": 2.6135,
      "theoretical_loss": 3.5980292206233875,
      "tokens_seen": 1162346496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032719157472417256,
      "loss": 2.6741,
      "theoretical_loss": 3.598010448326586,
      "tokens_seen": 1162412032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271815446339017,
      "loss": 2.7771,
      "theoretical_loss": 3.5979916773844494,
      "tokens_seen": 1162477568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271715145436309,
      "loss": 2.5643,
      "theoretical_loss": 3.5979729077968017,
      "tokens_seen": 1162543104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271614844533601,
      "loss": 2.6286,
      "theoretical_loss": 3.5979541395634698,
      "tokens_seen": 1162608640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271514543630893,
      "loss": 2.7594,
      "theoretical_loss": 3.59793537268428,
      "tokens_seen": 1162674176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032714142427281846,
      "loss": 2.5682,
      "theoretical_loss": 3.5979166071590574,
      "tokens_seen": 1162739712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271313941825477,
      "loss": 2.8591,
      "theoretical_loss": 3.5978978429876283,
      "tokens_seen": 1162805248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271213640922768,
      "loss": 2.5559,
      "theoretical_loss": 3.597879080169819,
      "tokens_seen": 1162870784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032711133400200606,
      "loss": 2.6989,
      "theoretical_loss": 3.5978603187054556,
      "tokens_seen": 1162936320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271013039117352,
      "loss": 2.4975,
      "theoretical_loss": 3.5978415585943635,
      "tokens_seen": 1163001856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270912738214644,
      "loss": 2.7157,
      "theoretical_loss": 3.5978227998363694,
      "tokens_seen": 1163067392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270812437311936,
      "loss": 2.5283,
      "theoretical_loss": 3.5978040424313003,
      "tokens_seen": 1163132928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270712136409228,
      "loss": 2.7726,
      "theoretical_loss": 3.5977852863789805,
      "tokens_seen": 1163198464
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 645795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8706414699554443,
      "objective/train/theoretical_loss": 3.5977665316792375,
      "objective/train/tokens_used": 1183724000,
      "theoretical_loss": 3.5977665316792375,
      "tokens_seen": 1163264000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032706118355065196,
      "loss": 2.708,
      "theoretical_loss": 3.5977665316792375,
      "tokens_seen": 1163264000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032705115346038114,
      "loss": 2.54,
      "theoretical_loss": 3.5977477783318976,
      "tokens_seen": 1163329536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270411233701103,
      "loss": 2.5949,
      "theoretical_loss": 3.597729026336787,
      "tokens_seen": 1163395072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032703109327983956,
      "loss": 2.7173,
      "theoretical_loss": 3.597710275693731,
      "tokens_seen": 1163460608
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270210631895687,
      "loss": 2.8881,
      "theoretical_loss": 3.5976915264025577,
      "tokens_seen": 1163526144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270110330992979,
      "loss": 2.7011,
      "theoretical_loss": 3.5976727784630924,
      "tokens_seen": 1163591680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032700100300902705,
      "loss": 2.7247,
      "theoretical_loss": 3.5976540318751615,
      "tokens_seen": 1163657216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269909729187563,
      "loss": 2.6947,
      "theoretical_loss": 3.597635286638592,
      "tokens_seen": 1163722752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032698094282848547,
      "loss": 2.7205,
      "theoretical_loss": 3.59761654275321,
      "tokens_seen": 1163788288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032697091273821465,
      "loss": 2.4699,
      "theoretical_loss": 3.597597800218842,
      "tokens_seen": 1163853824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032696088264794383,
      "loss": 2.5679,
      "theoretical_loss": 3.597579059035314,
      "tokens_seen": 1163919360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032695085255767306,
      "loss": 2.6229,
      "theoretical_loss": 3.5975603192024543,
      "tokens_seen": 1163984896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269408224674022,
      "loss": 2.8093,
      "theoretical_loss": 3.5975415807200886,
      "tokens_seen": 1164050432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269307923771314,
      "loss": 2.6794,
      "theoretical_loss": 3.5975228435880426,
      "tokens_seen": 1164115968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032692076228686055,
      "loss": 2.6717,
      "theoretical_loss": 3.5975041078061443,
      "tokens_seen": 1164181504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269107321965898,
      "loss": 2.6432,
      "theoretical_loss": 3.5974853733742194,
      "tokens_seen": 1164247040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032690070210631897,
      "loss": 2.4695,
      "theoretical_loss": 3.5974666402920956,
      "tokens_seen": 1164312576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032689067201604815,
      "loss": 2.6219,
      "theoretical_loss": 3.5974479085596,
      "tokens_seen": 1164378112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032688064192577733,
      "loss": 2.7599,
      "theoretical_loss": 3.5974291781765575,
      "tokens_seen": 1164443648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268706118355065,
      "loss": 2.849,
      "theoretical_loss": 3.5974104491427963,
      "tokens_seen": 1164509184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268605817452357,
      "loss": 2.7382,
      "theoretical_loss": 3.5973917214581435,
      "tokens_seen": 1164574720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032685055165496493,
      "loss": 2.6985,
      "theoretical_loss": 3.5973729951224254,
      "tokens_seen": 1164640256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032684052156469406,
      "loss": 2.619,
      "theoretical_loss": 3.597354270135469,
      "tokens_seen": 1164705792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268304914744233,
      "loss": 2.4724,
      "theoretical_loss": 3.5973355464971015,
      "tokens_seen": 1164771328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268204613841524,
      "loss": 2.5872,
      "theoretical_loss": 3.59731682420715,
      "tokens_seen": 1164836864
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 647021,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5365681648254395,
      "objective/train/theoretical_loss": 3.597298103265441,
      "objective/train/tokens_used": 1185362400,
      "theoretical_loss": 3.597298103265441,
      "tokens_seen": 1164902400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032681043129388165,
      "loss": 2.7167,
      "theoretical_loss": 3.597298103265441,
      "tokens_seen": 1164902400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032680040120361084,
      "loss": 2.7747,
      "theoretical_loss": 3.597279383671802,
      "tokens_seen": 1164967936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032679037111334,
      "loss": 2.7435,
      "theoretical_loss": 3.5972606654260604,
      "tokens_seen": 1165033472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267803410230692,
      "loss": 2.5863,
      "theoretical_loss": 3.5972419485280422,
      "tokens_seen": 1165099008
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032677031093279843,
      "loss": 2.6398,
      "theoretical_loss": 3.597223232977576,
      "tokens_seen": 1165164544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032676028084252756,
      "loss": 2.77,
      "theoretical_loss": 3.597204518774488,
      "tokens_seen": 1165230080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267502507522568,
      "loss": 2.7466,
      "theoretical_loss": 3.5971858059186057,
      "tokens_seen": 1165295616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267402206619859,
      "loss": 2.5286,
      "theoretical_loss": 3.5971670944097562,
      "tokens_seen": 1165361152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032673019057171516,
      "loss": 2.6559,
      "theoretical_loss": 3.5971483842477676,
      "tokens_seen": 1165426688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032672016048144434,
      "loss": 2.7506,
      "theoretical_loss": 3.597129675432466,
      "tokens_seen": 1165492224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267101303911735,
      "loss": 2.6339,
      "theoretical_loss": 3.5971109679636797,
      "tokens_seen": 1165557760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267001003009027,
      "loss": 2.6564,
      "theoretical_loss": 3.5970922618412353,
      "tokens_seen": 1165623296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266900702106319,
      "loss": 2.6649,
      "theoretical_loss": 3.5970735570649612,
      "tokens_seen": 1165688832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032668004012036106,
      "loss": 2.5365,
      "theoretical_loss": 3.597054853634684,
      "tokens_seen": 1165754368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266700100300903,
      "loss": 2.5501,
      "theoretical_loss": 3.597036151550231,
      "tokens_seen": 1165819904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266599799398194,
      "loss": 2.8168,
      "theoretical_loss": 3.597017450811431,
      "tokens_seen": 1165885440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032664994984954866,
      "loss": 2.511,
      "theoretical_loss": 3.5969987514181105,
      "tokens_seen": 1165950976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032663991975927784,
      "loss": 2.8327,
      "theoretical_loss": 3.596980053370097,
      "tokens_seen": 1166016512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326629889669007,
      "loss": 2.77,
      "theoretical_loss": 3.5969613566672187,
      "tokens_seen": 1166082048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266198595787362,
      "loss": 2.8883,
      "theoretical_loss": 3.596942661309303,
      "tokens_seen": 1166147584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266098294884654,
      "loss": 2.9158,
      "theoretical_loss": 3.5969239672961777,
      "tokens_seen": 1166213120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032659979939819457,
      "loss": 2.8605,
      "theoretical_loss": 3.5969052746276704,
      "tokens_seen": 1166278656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265897693079238,
      "loss": 2.8674,
      "theoretical_loss": 3.596886583303608,
      "tokens_seen": 1166344192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032657973921765293,
      "loss": 2.4217,
      "theoretical_loss": 3.59686789332382,
      "tokens_seen": 1166409728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032656970912738216,
      "loss": 2.7101,
      "theoretical_loss": 3.596849204688133,
      "tokens_seen": 1166475264
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 647459,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1023616790771484,
      "objective/train/theoretical_loss": 3.5968305173963744,
      "objective/train/tokens_used": 1187000800,
      "theoretical_loss": 3.5968305173963744,
      "tokens_seen": 1166540800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032655967903711135,
      "loss": 2.8593,
      "theoretical_loss": 3.5968305173963744,
      "tokens_seen": 1166540800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265496489468405,
      "loss": 2.5972,
      "theoretical_loss": 3.596811831448373,
      "tokens_seen": 1166606336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032653961885656976,
      "loss": 2.7343,
      "theoretical_loss": 3.596793146843957,
      "tokens_seen": 1166671872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265295887662989,
      "loss": 2.5819,
      "theoretical_loss": 3.5967744635829533,
      "tokens_seen": 1166737408
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265195586760281,
      "loss": 2.665,
      "theoretical_loss": 3.5967557816651903,
      "tokens_seen": 1166802944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032650952858575725,
      "loss": 2.6375,
      "theoretical_loss": 3.596737101090496,
      "tokens_seen": 1166868480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264994984954865,
      "loss": 2.5396,
      "theoretical_loss": 3.5967184218586983,
      "tokens_seen": 1166934016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032648946840521567,
      "loss": 2.6808,
      "theoretical_loss": 3.5966997439696256,
      "tokens_seen": 1166999552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032647943831494485,
      "loss": 2.6276,
      "theoretical_loss": 3.596681067423106,
      "tokens_seen": 1167065088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032646940822467403,
      "loss": 2.5045,
      "theoretical_loss": 3.596662392218967,
      "tokens_seen": 1167130624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032645937813440326,
      "loss": 2.4189,
      "theoretical_loss": 3.5966437183570372,
      "tokens_seen": 1167196160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264493480441324,
      "loss": 2.7956,
      "theoretical_loss": 3.596625045837145,
      "tokens_seen": 1167261696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264393179538616,
      "loss": 2.559,
      "theoretical_loss": 3.5966063746591184,
      "tokens_seen": 1167327232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032642928786359075,
      "loss": 2.7304,
      "theoretical_loss": 3.596587704822785,
      "tokens_seen": 1167392768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032641925777332,
      "loss": 2.6686,
      "theoretical_loss": 3.596569036327974,
      "tokens_seen": 1167458304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032640922768304917,
      "loss": 2.476,
      "theoretical_loss": 3.596550369174513,
      "tokens_seen": 1167523840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032639919759277835,
      "loss": 2.5304,
      "theoretical_loss": 3.596531703362231,
      "tokens_seen": 1167589376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032638916750250753,
      "loss": 2.7702,
      "theoretical_loss": 3.596513038890956,
      "tokens_seen": 1167654912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263791374122367,
      "loss": 2.6531,
      "theoretical_loss": 3.596494375760516,
      "tokens_seen": 1167720448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263691073219659,
      "loss": 2.7071,
      "theoretical_loss": 3.59647571397074,
      "tokens_seen": 1167785984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032635907723169513,
      "loss": 2.8008,
      "theoretical_loss": 3.596457053521457,
      "tokens_seen": 1167851520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032634904714142426,
      "loss": 2.6172,
      "theoretical_loss": 3.5964383944124947,
      "tokens_seen": 1167917056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263390170511535,
      "loss": 2.8027,
      "theoretical_loss": 3.596419736643681,
      "tokens_seen": 1167982592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263289869608826,
      "loss": 2.5849,
      "theoretical_loss": 3.5964010802148456,
      "tokens_seen": 1168048128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032631895687061185,
      "loss": 2.8076,
      "theoretical_loss": 3.596382425125817,
      "tokens_seen": 1168113664
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 648684,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9439427852630615,
      "objective/train/theoretical_loss": 3.5963637713764234,
      "objective/train/tokens_used": 1188639200,
      "theoretical_loss": 3.5963637713764234,
      "tokens_seen": 1168179200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032630892678034104,
      "loss": 2.6827,
      "theoretical_loss": 3.5963637713764234,
      "tokens_seen": 1168179200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262988966900702,
      "loss": 2.7814,
      "theoretical_loss": 3.5963451189664934,
      "tokens_seen": 1168244736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262888665997994,
      "loss": 2.6662,
      "theoretical_loss": 3.596326467895856,
      "tokens_seen": 1168310272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032627883650952863,
      "loss": 2.5774,
      "theoretical_loss": 3.5963078181643398,
      "tokens_seen": 1168375808
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032626880641925776,
      "loss": 2.7241,
      "theoretical_loss": 3.5962891697717736,
      "tokens_seen": 1168441344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326258776328987,
      "loss": 2.6469,
      "theoretical_loss": 3.596270522717986,
      "tokens_seen": 1168506880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262487462387161,
      "loss": 2.5015,
      "theoretical_loss": 3.596251877002806,
      "tokens_seen": 1168572416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032623871614844536,
      "loss": 2.8421,
      "theoretical_loss": 3.596233232626062,
      "tokens_seen": 1168637952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032622868605817454,
      "loss": 2.7734,
      "theoretical_loss": 3.596214589587584,
      "tokens_seen": 1168703488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262186559679037,
      "loss": 2.7135,
      "theoretical_loss": 3.5961959478871997,
      "tokens_seen": 1168769024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262086258776329,
      "loss": 2.5319,
      "theoretical_loss": 3.596177307524738,
      "tokens_seen": 1168834560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261985957873621,
      "loss": 2.6577,
      "theoretical_loss": 3.5961586685000295,
      "tokens_seen": 1168900096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032618856569709126,
      "loss": 2.4889,
      "theoretical_loss": 3.596140030812901,
      "tokens_seen": 1168965632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261785356068205,
      "loss": 2.5488,
      "theoretical_loss": 3.596121394463183,
      "tokens_seen": 1169031168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261685055165496,
      "loss": 2.6289,
      "theoretical_loss": 3.5961027594507042,
      "tokens_seen": 1169096704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032615847542627886,
      "loss": 2.5204,
      "theoretical_loss": 3.5960841257752936,
      "tokens_seen": 1169162240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032614844533600804,
      "loss": 2.8722,
      "theoretical_loss": 3.5960654934367806,
      "tokens_seen": 1169227776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261384152457372,
      "loss": 2.8192,
      "theoretical_loss": 3.5960468624349935,
      "tokens_seen": 1169293312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261283851554664,
      "loss": 2.7604,
      "theoretical_loss": 3.596028232769762,
      "tokens_seen": 1169358848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261183550651956,
      "loss": 2.5521,
      "theoretical_loss": 3.5960096044409164,
      "tokens_seen": 1169424384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032610832497492477,
      "loss": 2.8361,
      "theoretical_loss": 3.595990977448284,
      "tokens_seen": 1169489920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326098294884654,
      "loss": 2.7227,
      "theoretical_loss": 3.595972351791696,
      "tokens_seen": 1169555456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032608826479438313,
      "loss": 2.698,
      "theoretical_loss": 3.59595372747098,
      "tokens_seen": 1169620992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032607823470411236,
      "loss": 2.6011,
      "theoretical_loss": 3.5959351044859664,
      "tokens_seen": 1169686528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260682046138415,
      "loss": 2.7583,
      "theoretical_loss": 3.5959164828364845,
      "tokens_seen": 1169752064
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 649144,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.792980670928955,
      "objective/train/theoretical_loss": 3.5958978625223628,
      "objective/train/tokens_used": 1190277600,
      "theoretical_loss": 3.5958978625223628,
      "tokens_seen": 1169817600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260581745235707,
      "loss": 2.5967,
      "theoretical_loss": 3.5958978625223628,
      "tokens_seen": 1169817600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260481444332999,
      "loss": 2.7829,
      "theoretical_loss": 3.595879243543432,
      "tokens_seen": 1169883136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260381143430291,
      "loss": 2.6743,
      "theoretical_loss": 3.5958606258995207,
      "tokens_seen": 1169948672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032602808425275827,
      "loss": 2.5841,
      "theoretical_loss": 3.5958420095904593,
      "tokens_seen": 1170014208
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032601805416248745,
      "loss": 2.8274,
      "theoretical_loss": 3.5958233946160765,
      "tokens_seen": 1170079744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032600802407221663,
      "loss": 2.6443,
      "theoretical_loss": 3.595804780976202,
      "tokens_seen": 1170145280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032599799398194587,
      "loss": 2.3263,
      "theoretical_loss": 3.5957861686706654,
      "tokens_seen": 1170210816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000325987963891675,
      "loss": 2.4542,
      "theoretical_loss": 3.595767557699297,
      "tokens_seen": 1170276352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032597793380140423,
      "loss": 2.7791,
      "theoretical_loss": 3.595748948061925,
      "tokens_seen": 1170341888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003259679037111334,
      "loss": 2.6827,
      "theoretical_loss": 3.59573033975838,
      "tokens_seen": 1170407424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003259578736208626,
      "loss": 2.6128,
      "theoretical_loss": 3.5957117327884927,
      "tokens_seen": 1170472960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003259478435305918,
      "loss": 2.8103,
      "theoretical_loss": 3.595693127152091,
      "tokens_seen": 1170538496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032593781344032095,
      "loss": 2.7226,
      "theoretical_loss": 3.5956745228490057,
      "tokens_seen": 1170604032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032592778335005014,
      "loss": 2.8914,
      "theoretical_loss": 3.5956559198790665,
      "tokens_seen": 1170669568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032591775325977937,
      "loss": 2.6374,
      "theoretical_loss": 3.595637318242103,
      "tokens_seen": 1170735104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003259077231695085,
      "loss": 2.682,
      "theoretical_loss": 3.5956187179379455,
      "tokens_seen": 1170800640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032589769307923773,
      "loss": 2.6491,
      "theoretical_loss": 3.595600118966424,
      "tokens_seen": 1170866176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032588766298896686,
      "loss": 2.7347,
      "theoretical_loss": 3.5955815213273676,
      "tokens_seen": 1170931712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258776328986961,
      "loss": 2.5143,
      "theoretical_loss": 3.595562925020607,
      "tokens_seen": 1170997248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258676028084253,
      "loss": 2.7273,
      "theoretical_loss": 3.595544330045972,
      "tokens_seen": 1171062784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032585757271815446,
      "loss": 2.7069,
      "theoretical_loss": 3.5955257364032924,
      "tokens_seen": 1171128320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032584754262788364,
      "loss": 2.4612,
      "theoretical_loss": 3.595507144092399,
      "tokens_seen": 1171193856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258375125376128,
      "loss": 2.8743,
      "theoretical_loss": 3.595488553113121,
      "tokens_seen": 1171259392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000325827482447342,
      "loss": 2.6314,
      "theoretical_loss": 3.5954699634652894,
      "tokens_seen": 1171324928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032581745235707124,
      "loss": 2.4695,
      "theoretical_loss": 3.595451375148733,
      "tokens_seen": 1171390464
    },
    {
      "epoch": 0.35,
      "objective/train/docs_used": 649993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.923766851425171,
      "objective/train/theoretical_loss": 3.5954327881632837,
      "objective/train/tokens_used": 1191916000,
      "theoretical_loss": 3.5954327881632837,
      "tokens_seen": 1171456000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258074222668004,
      "loss": 2.8658,
      "theoretical_loss": 3.5954327881632837,
      "tokens_seen": 1171456000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257973921765296,
      "loss": 2.7553,
      "theoretical_loss": 3.5954142025087705,
      "tokens_seen": 1171521536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032578736208625883,
      "loss": 2.7047,
      "theoretical_loss": 3.595395618185024,
      "tokens_seen": 1171587072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032577733199598796,
      "loss": 3.073,
      "theoretical_loss": 3.5953770351918743,
      "tokens_seen": 1171652608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257673019057172,
      "loss": 2.5345,
      "theoretical_loss": 3.595358453529152,
      "tokens_seen": 1171718144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257572718154463,
      "loss": 2.5694,
      "theoretical_loss": 3.595339873196688,
      "tokens_seen": 1171783680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032574724172517556,
      "loss": 2.8471,
      "theoretical_loss": 3.5953212941943113,
      "tokens_seen": 1171849216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032573721163490474,
      "loss": 2.6878,
      "theoretical_loss": 3.5953027165218536,
      "tokens_seen": 1171914752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257271815446339,
      "loss": 2.6934,
      "theoretical_loss": 3.595284140179144,
      "tokens_seen": 1171980288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257171514543631,
      "loss": 2.6298,
      "theoretical_loss": 3.595265565166015,
      "tokens_seen": 1172045824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257071213640923,
      "loss": 2.5254,
      "theoretical_loss": 3.595246991482295,
      "tokens_seen": 1172111360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032569709127382146,
      "loss": 2.6913,
      "theoretical_loss": 3.5952284191278157,
      "tokens_seen": 1172176896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256870611835507,
      "loss": 2.5984,
      "theoretical_loss": 3.595209848102407,
      "tokens_seen": 1172242432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256770310932798,
      "loss": 2.5922,
      "theoretical_loss": 3.5951912784059,
      "tokens_seen": 1172307968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032566700100300906,
      "loss": 2.7977,
      "theoretical_loss": 3.5951727100381254,
      "tokens_seen": 1172373504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032565697091273824,
      "loss": 2.6315,
      "theoretical_loss": 3.5951541429989136,
      "tokens_seen": 1172439040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256469408224674,
      "loss": 3.0059,
      "theoretical_loss": 3.595135577288095,
      "tokens_seen": 1172504576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256369107321966,
      "loss": 2.5167,
      "theoretical_loss": 3.595117012905501,
      "tokens_seen": 1172570112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256268806419258,
      "loss": 2.4955,
      "theoretical_loss": 3.595098449850962,
      "tokens_seen": 1172635648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032561685055165497,
      "loss": 2.638,
      "theoretical_loss": 3.5950798881243085,
      "tokens_seen": 1172701184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256068204613842,
      "loss": 2.4716,
      "theoretical_loss": 3.5950613277253716,
      "tokens_seen": 1172766720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032559679037111333,
      "loss": 2.5471,
      "theoretical_loss": 3.595042768653982,
      "tokens_seen": 1172832256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032558676028084256,
      "loss": 2.829,
      "theoretical_loss": 3.595024210909971,
      "tokens_seen": 1172897792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255767301905717,
      "loss": 2.5036,
      "theoretical_loss": 3.5950056544931694,
      "tokens_seen": 1172963328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255667001003009,
      "loss": 2.784,
      "theoretical_loss": 3.594987099403407,
      "tokens_seen": 1173028864
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 651258,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.56889009475708,
      "objective/train/theoretical_loss": 3.5949685456405165,
      "objective/train/tokens_used": 1193554400,
      "theoretical_loss": 3.5949685456405165,
      "tokens_seen": 1173094400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255566700100301,
      "loss": 2.6219,
      "theoretical_loss": 3.5949685456405165,
      "tokens_seen": 1173094400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255466399197593,
      "loss": 2.7571,
      "theoretical_loss": 3.594949993204328,
      "tokens_seen": 1173159936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032553660982948847,
      "loss": 2.609,
      "theoretical_loss": 3.5949314420946727,
      "tokens_seen": 1173225472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032552657973921765,
      "loss": 2.7424,
      "theoretical_loss": 3.594912892311381,
      "tokens_seen": 1173291008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032551654964894683,
      "loss": 2.6436,
      "theoretical_loss": 3.594894343854285,
      "tokens_seen": 1173356544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032550651955867607,
      "loss": 2.7312,
      "theoretical_loss": 3.5948757967232154,
      "tokens_seen": 1173422080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254964894684052,
      "loss": 2.657,
      "theoretical_loss": 3.594857250918003,
      "tokens_seen": 1173487616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032548645937813443,
      "loss": 2.8031,
      "theoretical_loss": 3.5948387064384795,
      "tokens_seen": 1173553152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254764292878636,
      "loss": 2.6133,
      "theoretical_loss": 3.594820163284476,
      "tokens_seen": 1173618688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254663991975928,
      "loss": 2.9284,
      "theoretical_loss": 3.594801621455824,
      "tokens_seen": 1173684224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325456369107322,
      "loss": 2.6838,
      "theoretical_loss": 3.594783080952354,
      "tokens_seen": 1173749760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032544633901705115,
      "loss": 2.5229,
      "theoretical_loss": 3.5947645417738983,
      "tokens_seen": 1173815296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032543630892678034,
      "loss": 2.6614,
      "theoretical_loss": 3.594746003920287,
      "tokens_seen": 1173880832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032542627883650957,
      "loss": 2.7473,
      "theoretical_loss": 3.5947274673913525,
      "tokens_seen": 1173946368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254162487462387,
      "loss": 2.4982,
      "theoretical_loss": 3.594708932186926,
      "tokens_seen": 1174011904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032540621865596793,
      "loss": 2.7672,
      "theoretical_loss": 3.5946903983068386,
      "tokens_seen": 1174077440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032539618856569706,
      "loss": 2.6318,
      "theoretical_loss": 3.594671865750922,
      "tokens_seen": 1174142976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253861584754263,
      "loss": 2.7851,
      "theoretical_loss": 3.5946533345190073,
      "tokens_seen": 1174208512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253761283851555,
      "loss": 2.595,
      "theoretical_loss": 3.5946348046109273,
      "tokens_seen": 1174274048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032536609829488466,
      "loss": 2.7437,
      "theoretical_loss": 3.5946162760265117,
      "tokens_seen": 1174339584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032535606820461384,
      "loss": 2.9372,
      "theoretical_loss": 3.5945977487655933,
      "tokens_seen": 1174405120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325346038114343,
      "loss": 2.704,
      "theoretical_loss": 3.5945792228280036,
      "tokens_seen": 1174470656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253360080240722,
      "loss": 2.664,
      "theoretical_loss": 3.5945606982135736,
      "tokens_seen": 1174536192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032532597793380144,
      "loss": 2.7617,
      "theoretical_loss": 3.594542174922136,
      "tokens_seen": 1174601728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032531594784353056,
      "loss": 2.5863,
      "theoretical_loss": 3.5945236529535216,
      "tokens_seen": 1174667264
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 652492,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.737172842025757,
      "objective/train/theoretical_loss": 3.5945051323075625,
      "objective/train/tokens_used": 1195192800,
      "theoretical_loss": 3.5945051323075625,
      "tokens_seen": 1174732800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253059177532598,
      "loss": 2.6773,
      "theoretical_loss": 3.5945051323075625,
      "tokens_seen": 1174732800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325295887662989,
      "loss": 2.7729,
      "theoretical_loss": 3.594486612984091,
      "tokens_seen": 1174798336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032528585757271816,
      "loss": 2.7024,
      "theoretical_loss": 3.5944680949829375,
      "tokens_seen": 1174863872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032527582748244734,
      "loss": 2.5224,
      "theoretical_loss": 3.5944495783039354,
      "tokens_seen": 1174929408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252657973921765,
      "loss": 2.5866,
      "theoretical_loss": 3.5944310629469154,
      "tokens_seen": 1174994944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252557673019057,
      "loss": 2.7831,
      "theoretical_loss": 3.59441254891171,
      "tokens_seen": 1175060480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032524573721163494,
      "loss": 2.738,
      "theoretical_loss": 3.5943940361981506,
      "tokens_seen": 1175126016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032523570712136407,
      "loss": 2.5638,
      "theoretical_loss": 3.59437552480607,
      "tokens_seen": 1175191552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252256770310933,
      "loss": 2.6825,
      "theoretical_loss": 3.5943570147352997,
      "tokens_seen": 1175257088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032521564694082243,
      "loss": 2.6556,
      "theoretical_loss": 3.594338505985671,
      "tokens_seen": 1175322624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032520561685055166,
      "loss": 2.8164,
      "theoretical_loss": 3.594319998557017,
      "tokens_seen": 1175388160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032519558676028084,
      "loss": 2.7041,
      "theoretical_loss": 3.59430149244917,
      "tokens_seen": 1175453696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032518555667001,
      "loss": 2.5987,
      "theoretical_loss": 3.5942829876619613,
      "tokens_seen": 1175519232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251755265797392,
      "loss": 2.5567,
      "theoretical_loss": 3.594264484195223,
      "tokens_seen": 1175584768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032516549648946844,
      "loss": 2.7074,
      "theoretical_loss": 3.594245982048788,
      "tokens_seen": 1175650304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032515546639919757,
      "loss": 2.7701,
      "theoretical_loss": 3.5942274812224877,
      "tokens_seen": 1175715840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251454363089268,
      "loss": 2.6326,
      "theoretical_loss": 3.5942089817161547,
      "tokens_seen": 1175781376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032513540621865593,
      "loss": 2.608,
      "theoretical_loss": 3.594190483529621,
      "tokens_seen": 1175846912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032512537612838517,
      "loss": 2.5178,
      "theoretical_loss": 3.5941719866627193,
      "tokens_seen": 1175912448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032511534603811435,
      "loss": 2.737,
      "theoretical_loss": 3.594153491115282,
      "tokens_seen": 1175977984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032510531594784353,
      "loss": 2.5786,
      "theoretical_loss": 3.5941349968871408,
      "tokens_seen": 1176043520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250952858575727,
      "loss": 2.5329,
      "theoretical_loss": 3.5941165039781287,
      "tokens_seen": 1176109056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250852557673019,
      "loss": 2.4487,
      "theoretical_loss": 3.594098012388078,
      "tokens_seen": 1176174592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032507522567703107,
      "loss": 2.5077,
      "theoretical_loss": 3.5940795221168207,
      "tokens_seen": 1176240128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250651955867603,
      "loss": 2.5841,
      "theoretical_loss": 3.59406103316419,
      "tokens_seen": 1176305664
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 652965,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588792085647583,
      "objective/train/theoretical_loss": 3.5940425455300176,
      "objective/train/tokens_used": 1196831200,
      "theoretical_loss": 3.5940425455300176,
      "tokens_seen": 1176371200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250551654964895,
      "loss": 2.6309,
      "theoretical_loss": 3.5940425455300176,
      "tokens_seen": 1176371200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032504513540621867,
      "loss": 2.7376,
      "theoretical_loss": 3.594024059214137,
      "tokens_seen": 1176436736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032503510531594785,
      "loss": 2.9167,
      "theoretical_loss": 3.59400557421638,
      "tokens_seen": 1176502272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032502507522567703,
      "loss": 2.6344,
      "theoretical_loss": 3.5939870905365794,
      "tokens_seen": 1176567808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032501504513540627,
      "loss": 2.6758,
      "theoretical_loss": 3.593968608174568,
      "tokens_seen": 1176633344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250050150451354,
      "loss": 2.6595,
      "theoretical_loss": 3.593950127130178,
      "tokens_seen": 1176698880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032499498495486463,
      "loss": 2.6477,
      "theoretical_loss": 3.5939316474032426,
      "tokens_seen": 1176764416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003249849548645938,
      "loss": 2.6468,
      "theoretical_loss": 3.593913168993594,
      "tokens_seen": 1176829952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324974924774323,
      "loss": 2.7797,
      "theoretical_loss": 3.5938946919010655,
      "tokens_seen": 1176895488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003249648946840522,
      "loss": 2.6107,
      "theoretical_loss": 3.59387621612549,
      "tokens_seen": 1176961024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032495486459378135,
      "loss": 2.6228,
      "theoretical_loss": 3.593857741666699,
      "tokens_seen": 1177026560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032494483450351054,
      "loss": 2.5313,
      "theoretical_loss": 3.593839268524527,
      "tokens_seen": 1177092096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032493480441323977,
      "loss": 2.5962,
      "theoretical_loss": 3.593820796698806,
      "tokens_seen": 1177157632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003249247743229689,
      "loss": 2.7831,
      "theoretical_loss": 3.593802326189369,
      "tokens_seen": 1177223168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032491474423269813,
      "loss": 2.6414,
      "theoretical_loss": 3.5937838569960494,
      "tokens_seen": 1177288704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032490471414242726,
      "loss": 2.8084,
      "theoretical_loss": 3.5937653891186794,
      "tokens_seen": 1177354240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248946840521565,
      "loss": 2.7652,
      "theoretical_loss": 3.5937469225570924,
      "tokens_seen": 1177419776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248846539618857,
      "loss": 2.8432,
      "theoretical_loss": 3.593728457311122,
      "tokens_seen": 1177485312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032487462387161486,
      "loss": 2.6533,
      "theoretical_loss": 3.5937099933805996,
      "tokens_seen": 1177550848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032486459378134404,
      "loss": 2.6939,
      "theoretical_loss": 3.59369153076536,
      "tokens_seen": 1177616384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248545636910732,
      "loss": 2.4783,
      "theoretical_loss": 3.5936730694652352,
      "tokens_seen": 1177681920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248445336008024,
      "loss": 2.8857,
      "theoretical_loss": 3.593654609480059,
      "tokens_seen": 1177747456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032483450351053164,
      "loss": 2.8108,
      "theoretical_loss": 3.5936361508096644,
      "tokens_seen": 1177812992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032482447342026076,
      "loss": 2.7793,
      "theoretical_loss": 3.5936176934538846,
      "tokens_seen": 1177878528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032481444332999,
      "loss": 2.6711,
      "theoretical_loss": 3.593599237412553,
      "tokens_seen": 1177944064
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 652965,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3456740379333496,
      "objective/train/theoretical_loss": 3.5935807826855024,
      "objective/train/tokens_used": 1198469600,
      "theoretical_loss": 3.5935807826855024,
      "tokens_seen": 1178009600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248044132397192,
      "loss": 2.476,
      "theoretical_loss": 3.5935807826855024,
      "tokens_seen": 1178009600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032479438314944836,
      "loss": 2.7488,
      "theoretical_loss": 3.593562329272567,
      "tokens_seen": 1178075136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032478435305917754,
      "loss": 2.6266,
      "theoretical_loss": 3.5935438771735786,
      "tokens_seen": 1178140672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247743229689067,
      "loss": 2.5096,
      "theoretical_loss": 3.593525426388372,
      "tokens_seen": 1178206208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247642928786359,
      "loss": 2.8226,
      "theoretical_loss": 3.59350697691678,
      "tokens_seen": 1178271744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032475426278836514,
      "loss": 2.5758,
      "theoretical_loss": 3.593488528758636,
      "tokens_seen": 1178337280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032474423269809427,
      "loss": 2.7812,
      "theoretical_loss": 3.5934700819137735,
      "tokens_seen": 1178402816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247342026078235,
      "loss": 2.5558,
      "theoretical_loss": 3.593451636382026,
      "tokens_seen": 1178468352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032472417251755263,
      "loss": 2.8785,
      "theoretical_loss": 3.5934331921632268,
      "tokens_seen": 1178533888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032471414242728186,
      "loss": 2.5582,
      "theoretical_loss": 3.59341474925721,
      "tokens_seen": 1178599424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032470411233701104,
      "loss": 2.4771,
      "theoretical_loss": 3.593396307663809,
      "tokens_seen": 1178664960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246940822467402,
      "loss": 2.7665,
      "theoretical_loss": 3.593377867382857,
      "tokens_seen": 1178730496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246840521564694,
      "loss": 2.6074,
      "theoretical_loss": 3.593359428414188,
      "tokens_seen": 1178796032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032467402206619864,
      "loss": 2.727,
      "theoretical_loss": 3.5933409907576355,
      "tokens_seen": 1178861568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032466399197592777,
      "loss": 2.8406,
      "theoretical_loss": 3.593322554413033,
      "tokens_seen": 1178927104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324653961885657,
      "loss": 2.5648,
      "theoretical_loss": 3.593304119380215,
      "tokens_seen": 1178992640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032464393179538613,
      "loss": 2.9887,
      "theoretical_loss": 3.593285685659014,
      "tokens_seen": 1179058176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032463390170511537,
      "loss": 2.558,
      "theoretical_loss": 3.5932672532492647,
      "tokens_seen": 1179123712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032462387161484455,
      "loss": 2.4303,
      "theoretical_loss": 3.5932488221508008,
      "tokens_seen": 1179189248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032461384152457373,
      "loss": 2.68,
      "theoretical_loss": 3.593230392363456,
      "tokens_seen": 1179254784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246038114343029,
      "loss": 2.7311,
      "theoretical_loss": 3.5932119638870645,
      "tokens_seen": 1179320320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245937813440321,
      "loss": 2.4551,
      "theoretical_loss": 3.5931935367214596,
      "tokens_seen": 1179385856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032458375125376127,
      "loss": 2.7065,
      "theoretical_loss": 3.5931751108664756,
      "tokens_seen": 1179451392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245737211634905,
      "loss": 2.8443,
      "theoretical_loss": 3.5931566863219464,
      "tokens_seen": 1179516928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032456369107321963,
      "loss": 2.7742,
      "theoretical_loss": 3.5931382630877065,
      "tokens_seen": 1179582464
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 654223,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4708504676818848,
      "objective/train/theoretical_loss": 3.593119841163589,
      "objective/train/tokens_used": 1200108000,
      "theoretical_loss": 3.593119841163589,
      "tokens_seen": 1179648000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032455366098294887,
      "loss": 2.6758,
      "theoretical_loss": 3.593119841163589,
      "tokens_seen": 1179648000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324543630892678,
      "loss": 3.057,
      "theoretical_loss": 3.593101420549429,
      "tokens_seen": 1179713536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032453360080240723,
      "loss": 2.6789,
      "theoretical_loss": 3.593083001245059,
      "tokens_seen": 1179779072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245235707121364,
      "loss": 2.8285,
      "theoretical_loss": 3.5930645832503147,
      "tokens_seen": 1179844608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245135406218656,
      "loss": 2.6795,
      "theoretical_loss": 3.5930461665650295,
      "tokens_seen": 1179910144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245035105315948,
      "loss": 2.6497,
      "theoretical_loss": 3.5930277511890383,
      "tokens_seen": 1179975680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324493480441324,
      "loss": 2.676,
      "theoretical_loss": 3.593009337122174,
      "tokens_seen": 1180041216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032448345035105314,
      "loss": 2.731,
      "theoretical_loss": 3.5929909243642726,
      "tokens_seen": 1180106752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244734202607824,
      "loss": 2.6065,
      "theoretical_loss": 3.592972512915167,
      "tokens_seen": 1180172288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244633901705115,
      "loss": 2.7483,
      "theoretical_loss": 3.592954102774691,
      "tokens_seen": 1180237824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032445336008024074,
      "loss": 2.4523,
      "theoretical_loss": 3.5929356939426804,
      "tokens_seen": 1180303360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244433299899699,
      "loss": 2.4251,
      "theoretical_loss": 3.592917286418969,
      "tokens_seen": 1180368896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244332998996991,
      "loss": 2.6842,
      "theoretical_loss": 3.592898880203391,
      "tokens_seen": 1180434432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244232698094283,
      "loss": 2.6927,
      "theoretical_loss": 3.592880475295781,
      "tokens_seen": 1180499968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032441323971915746,
      "loss": 2.7892,
      "theoretical_loss": 3.5928620716959734,
      "tokens_seen": 1180565504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032440320962888664,
      "loss": 2.6003,
      "theoretical_loss": 3.592843669403803,
      "tokens_seen": 1180631040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243931795386159,
      "loss": 2.5782,
      "theoretical_loss": 3.5928252684191033,
      "tokens_seen": 1180696576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324383149448345,
      "loss": 2.6193,
      "theoretical_loss": 3.59280686874171,
      "tokens_seen": 1180762112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032437311935807424,
      "loss": 2.6897,
      "theoretical_loss": 3.592788470371457,
      "tokens_seen": 1180827648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032436308926780337,
      "loss": 2.7288,
      "theoretical_loss": 3.5927700733081798,
      "tokens_seen": 1180893184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243530591775326,
      "loss": 2.6043,
      "theoretical_loss": 3.5927516775517114,
      "tokens_seen": 1180958720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243430290872618,
      "loss": 2.7239,
      "theoretical_loss": 3.5927332831018877,
      "tokens_seen": 1181024256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032433299899699096,
      "loss": 2.5689,
      "theoretical_loss": 3.5927148899585433,
      "tokens_seen": 1181089792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243229689067202,
      "loss": 2.8611,
      "theoretical_loss": 3.5926964981215126,
      "tokens_seen": 1181155328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243129388164494,
      "loss": 2.531,
      "theoretical_loss": 3.5926781075906304,
      "tokens_seen": 1181220864
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 655414,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6210334300994873,
      "objective/train/theoretical_loss": 3.5926597183657316,
      "objective/train/tokens_used": 1201746400,
      "theoretical_loss": 3.5926597183657316,
      "tokens_seen": 1181286400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032430290872617856,
      "loss": 2.7083,
      "theoretical_loss": 3.5926597183657316,
      "tokens_seen": 1181286400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032429287863590774,
      "loss": 2.6754,
      "theoretical_loss": 3.592641330446651,
      "tokens_seen": 1181351936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242828485456369,
      "loss": 2.4938,
      "theoretical_loss": 3.592622943833223,
      "tokens_seen": 1181417472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242728184553661,
      "loss": 2.5009,
      "theoretical_loss": 3.5926045585252835,
      "tokens_seen": 1181483008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032426278836509534,
      "loss": 2.7818,
      "theoretical_loss": 3.5925861745226664,
      "tokens_seen": 1181548544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032425275827482447,
      "loss": 2.7641,
      "theoretical_loss": 3.592567791825207,
      "tokens_seen": 1181614080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242427281845537,
      "loss": 2.7128,
      "theoretical_loss": 3.5925494104327402,
      "tokens_seen": 1181679616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032423269809428283,
      "loss": 2.8377,
      "theoretical_loss": 3.5925310303451012,
      "tokens_seen": 1181745152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032422266800401206,
      "loss": 2.6549,
      "theoretical_loss": 3.592512651562125,
      "tokens_seen": 1181810688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032421263791374125,
      "loss": 2.502,
      "theoretical_loss": 3.592494274083646,
      "tokens_seen": 1181876224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242026078234704,
      "loss": 2.6759,
      "theoretical_loss": 3.5924758979095004,
      "tokens_seen": 1181941760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241925777331996,
      "loss": 2.6429,
      "theoretical_loss": 3.5924575230395224,
      "tokens_seen": 1182007296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032418254764292884,
      "loss": 2.3383,
      "theoretical_loss": 3.592439149473547,
      "tokens_seen": 1182072832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032417251755265797,
      "loss": 2.6148,
      "theoretical_loss": 3.5924207772114105,
      "tokens_seen": 1182138368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241624874623872,
      "loss": 2.6313,
      "theoretical_loss": 3.5924024062529476,
      "tokens_seen": 1182203904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032415245737211633,
      "loss": 2.7198,
      "theoretical_loss": 3.5923840365979927,
      "tokens_seen": 1182269440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032414242728184557,
      "loss": 2.6309,
      "theoretical_loss": 3.592365668246382,
      "tokens_seen": 1182334976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032413239719157475,
      "loss": 2.937,
      "theoretical_loss": 3.5923473011979508,
      "tokens_seen": 1182400512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032412236710130393,
      "loss": 2.6589,
      "theoretical_loss": 3.5923289354525334,
      "tokens_seen": 1182466048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241123370110331,
      "loss": 2.8447,
      "theoretical_loss": 3.5923105710099663,
      "tokens_seen": 1182531584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241023069207623,
      "loss": 2.7621,
      "theoretical_loss": 3.5922922078700847,
      "tokens_seen": 1182597120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032409227683049147,
      "loss": 2.7971,
      "theoretical_loss": 3.5922738460327235,
      "tokens_seen": 1182662656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240822467402207,
      "loss": 2.8789,
      "theoretical_loss": 3.5922554854977182,
      "tokens_seen": 1182728192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032407221664994984,
      "loss": 2.6273,
      "theoretical_loss": 3.592237126264905,
      "tokens_seen": 1182793728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032406218655967907,
      "loss": 2.8434,
      "theoretical_loss": 3.5922187683341185,
      "tokens_seen": 1182859264
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 655986,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3905091285705566,
      "objective/train/theoretical_loss": 3.5922004117051944,
      "objective/train/tokens_used": 1203384800,
      "theoretical_loss": 3.5922004117051944,
      "tokens_seen": 1182924800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240521564694082,
      "loss": 2.5661,
      "theoretical_loss": 3.5922004117051944,
      "tokens_seen": 1182924800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032404212637913743,
      "loss": 2.5477,
      "theoretical_loss": 3.592182056377969,
      "tokens_seen": 1182990336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240320962888666,
      "loss": 2.7005,
      "theoretical_loss": 3.592163702352277,
      "tokens_seen": 1183055872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240220661985958,
      "loss": 2.7133,
      "theoretical_loss": 3.5921453496279545,
      "tokens_seen": 1183121408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324012036108325,
      "loss": 2.7503,
      "theoretical_loss": 3.5921269982048374,
      "tokens_seen": 1183186944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240020060180542,
      "loss": 2.6778,
      "theoretical_loss": 3.5921086480827604,
      "tokens_seen": 1183252480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032399197592778334,
      "loss": 2.7109,
      "theoretical_loss": 3.59209029926156,
      "tokens_seen": 1183318016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239819458375126,
      "loss": 2.545,
      "theoretical_loss": 3.592071951741072,
      "tokens_seen": 1183383552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239719157472417,
      "loss": 2.7016,
      "theoretical_loss": 3.592053605521132,
      "tokens_seen": 1183449088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032396188565697094,
      "loss": 2.8656,
      "theoretical_loss": 3.592035260601575,
      "tokens_seen": 1183514624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239518555667001,
      "loss": 2.4547,
      "theoretical_loss": 3.5920169169822382,
      "tokens_seen": 1183580160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239418254764293,
      "loss": 2.7549,
      "theoretical_loss": 3.5919985746629566,
      "tokens_seen": 1183645696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239317953861585,
      "loss": 2.8236,
      "theoretical_loss": 3.5919802336435662,
      "tokens_seen": 1183711232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032392176529588766,
      "loss": 2.4944,
      "theoretical_loss": 3.5919618939239033,
      "tokens_seen": 1183776768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032391173520561684,
      "loss": 2.7984,
      "theoretical_loss": 3.5919435555038035,
      "tokens_seen": 1183842304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003239017051153461,
      "loss": 2.6072,
      "theoretical_loss": 3.5919252183831025,
      "tokens_seen": 1183907840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238916750250752,
      "loss": 2.6405,
      "theoretical_loss": 3.591906882561637,
      "tokens_seen": 1183973376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032388164493480444,
      "loss": 2.9206,
      "theoretical_loss": 3.5918885480392424,
      "tokens_seen": 1184038912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032387161484453357,
      "loss": 2.7311,
      "theoretical_loss": 3.5918702148157555,
      "tokens_seen": 1184104448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238615847542628,
      "loss": 2.7311,
      "theoretical_loss": 3.591851882891012,
      "tokens_seen": 1184169984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323851554663992,
      "loss": 2.7366,
      "theoretical_loss": 3.5918335522648475,
      "tokens_seen": 1184235520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032384152457372116,
      "loss": 2.3489,
      "theoretical_loss": 3.5918152229370985,
      "tokens_seen": 1184301056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032383149448345034,
      "loss": 2.6593,
      "theoretical_loss": 3.591796894907602,
      "tokens_seen": 1184366592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238214643931796,
      "loss": 2.8426,
      "theoretical_loss": 3.591778568176193,
      "tokens_seen": 1184432128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238114343029087,
      "loss": 2.7458,
      "theoretical_loss": 3.5917602427427084,
      "tokens_seen": 1184497664
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 656832,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.695746421813965,
      "objective/train/theoretical_loss": 3.591741918606984,
      "objective/train/tokens_used": 1205023200,
      "theoretical_loss": 3.591741918606984,
      "tokens_seen": 1184563200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032380140421263794,
      "loss": 2.874,
      "theoretical_loss": 3.591741918606984,
      "tokens_seen": 1184563200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032379137412236707,
      "loss": 2.6529,
      "theoretical_loss": 3.5917235957688574,
      "tokens_seen": 1184628736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237813440320963,
      "loss": 2.6853,
      "theoretical_loss": 3.591705274228163,
      "tokens_seen": 1184694272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237713139418255,
      "loss": 2.8111,
      "theoretical_loss": 3.591686953984739,
      "tokens_seen": 1184759808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032376128385155467,
      "loss": 2.6361,
      "theoretical_loss": 3.5916686350384204,
      "tokens_seen": 1184825344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032375125376128385,
      "loss": 2.5834,
      "theoretical_loss": 3.5916503173890444,
      "tokens_seen": 1184890880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032374122367101303,
      "loss": 2.905,
      "theoretical_loss": 3.591632001036447,
      "tokens_seen": 1184956416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237311935807422,
      "loss": 2.6953,
      "theoretical_loss": 3.591613685980465,
      "tokens_seen": 1185021952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032372116349047145,
      "loss": 2.8276,
      "theoretical_loss": 3.5915953722209353,
      "tokens_seen": 1185087488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032371113340020057,
      "loss": 2.5766,
      "theoretical_loss": 3.5915770597576935,
      "tokens_seen": 1185153024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237011033099298,
      "loss": 2.586,
      "theoretical_loss": 3.591558748590577,
      "tokens_seen": 1185218560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032369107321965893,
      "loss": 2.7271,
      "theoretical_loss": 3.591540438719422,
      "tokens_seen": 1185284096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032368104312938817,
      "loss": 2.8175,
      "theoretical_loss": 3.5915221301440647,
      "tokens_seen": 1185349632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032367101303911735,
      "loss": 2.823,
      "theoretical_loss": 3.5915038228643423,
      "tokens_seen": 1185415168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032366098294884653,
      "loss": 2.7177,
      "theoretical_loss": 3.5914855168800917,
      "tokens_seen": 1185480704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236509528585757,
      "loss": 2.648,
      "theoretical_loss": 3.5914672121911493,
      "tokens_seen": 1185546240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032364092276830495,
      "loss": 2.8443,
      "theoretical_loss": 3.591448908797352,
      "tokens_seen": 1185611776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236308926780341,
      "loss": 2.7258,
      "theoretical_loss": 3.591430606698536,
      "tokens_seen": 1185677312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236208625877633,
      "loss": 2.5077,
      "theoretical_loss": 3.591412305894539,
      "tokens_seen": 1185742848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032361083249749244,
      "loss": 2.7522,
      "theoretical_loss": 3.5913940063851975,
      "tokens_seen": 1185808384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236008024072217,
      "loss": 2.5417,
      "theoretical_loss": 3.591375708170348,
      "tokens_seen": 1185873920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032359077231695085,
      "loss": 2.8327,
      "theoretical_loss": 3.5913574112498274,
      "tokens_seen": 1185939456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032358074222668004,
      "loss": 2.6164,
      "theoretical_loss": 3.591339115623474,
      "tokens_seen": 1186004992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032357071213640927,
      "loss": 2.4979,
      "theoretical_loss": 3.5913208212911227,
      "tokens_seen": 1186070528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235606820461384,
      "loss": 2.7355,
      "theoretical_loss": 3.591302528252612,
      "tokens_seen": 1186136064
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 657301,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4484574794769287,
      "objective/train/theoretical_loss": 3.5912842365077777,
      "objective/train/tokens_used": 1206661600,
      "theoretical_loss": 3.5912842365077777,
      "tokens_seen": 1186201600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032355065195586763,
      "loss": 2.6418,
      "theoretical_loss": 3.5912842365077777,
      "tokens_seen": 1186201600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235406218655968,
      "loss": 2.7045,
      "theoretical_loss": 3.591265946056458,
      "tokens_seen": 1186267136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323530591775326,
      "loss": 2.6399,
      "theoretical_loss": 3.5912476568984895,
      "tokens_seen": 1186332672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235205616850552,
      "loss": 2.7264,
      "theoretical_loss": 3.5912293690337096,
      "tokens_seen": 1186398208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235105315947844,
      "loss": 2.6608,
      "theoretical_loss": 3.5912110824619545,
      "tokens_seen": 1186463744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032350050150451354,
      "loss": 2.6546,
      "theoretical_loss": 3.591192797183062,
      "tokens_seen": 1186529280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234904714142428,
      "loss": 2.6396,
      "theoretical_loss": 3.59117451319687,
      "tokens_seen": 1186594816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234804413239719,
      "loss": 2.711,
      "theoretical_loss": 3.5911562305032145,
      "tokens_seen": 1186660352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032347041123370114,
      "loss": 2.735,
      "theoretical_loss": 3.5911379491019333,
      "tokens_seen": 1186725888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234603811434303,
      "loss": 2.7617,
      "theoretical_loss": 3.5911196689928637,
      "tokens_seen": 1186791424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234503510531595,
      "loss": 2.6371,
      "theoretical_loss": 3.591101390175843,
      "tokens_seen": 1186856960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234403209628887,
      "loss": 2.6863,
      "theoretical_loss": 3.5910831126507086,
      "tokens_seen": 1186922496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032343029087261786,
      "loss": 2.7775,
      "theoretical_loss": 3.5910648364172975,
      "tokens_seen": 1186988032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032342026078234704,
      "loss": 2.9409,
      "theoretical_loss": 3.591046561475447,
      "tokens_seen": 1187053568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234102306920763,
      "loss": 2.9129,
      "theoretical_loss": 3.5910282878249955,
      "tokens_seen": 1187119104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234002006018054,
      "loss": 2.7658,
      "theoretical_loss": 3.5910100154657796,
      "tokens_seen": 1187184640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032339017051153464,
      "loss": 2.5633,
      "theoretical_loss": 3.590991744397637,
      "tokens_seen": 1187250176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032338014042126377,
      "loss": 2.6758,
      "theoretical_loss": 3.5909734746204056,
      "tokens_seen": 1187315712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323370110330993,
      "loss": 2.6059,
      "theoretical_loss": 3.5909552061339225,
      "tokens_seen": 1187381248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233600802407222,
      "loss": 2.6783,
      "theoretical_loss": 3.5909369389380252,
      "tokens_seen": 1187446784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032335005015045136,
      "loss": 2.5929,
      "theoretical_loss": 3.5909186730325517,
      "tokens_seen": 1187512320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032334002006018054,
      "loss": 2.6622,
      "theoretical_loss": 3.590900408417339,
      "tokens_seen": 1187577856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233299899699098,
      "loss": 2.4147,
      "theoretical_loss": 3.5908821450922255,
      "tokens_seen": 1187643392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233199598796389,
      "loss": 2.5664,
      "theoretical_loss": 3.5908638830570485,
      "tokens_seen": 1187708928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032330992978936814,
      "loss": 2.6783,
      "theoretical_loss": 3.5908456223116456,
      "tokens_seen": 1187774464
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 658636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8977458477020264,
      "objective/train/theoretical_loss": 3.590827362855855,
      "objective/train/tokens_used": 1208300000,
      "theoretical_loss": 3.590827362855855,
      "tokens_seen": 1187840000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032329989969909727,
      "loss": 2.8993,
      "theoretical_loss": 3.590827362855855,
      "tokens_seen": 1187840000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232898696088265,
      "loss": 2.895,
      "theoretical_loss": 3.590809104689514,
      "tokens_seen": 1187905536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232798395185557,
      "loss": 2.6605,
      "theoretical_loss": 3.5907908478124613,
      "tokens_seen": 1187971072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032326980942828487,
      "loss": 2.7268,
      "theoretical_loss": 3.5907725922245337,
      "tokens_seen": 1188036608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032325977933801405,
      "loss": 2.4671,
      "theoretical_loss": 3.5907543379255693,
      "tokens_seen": 1188102144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032324974924774323,
      "loss": 2.5579,
      "theoretical_loss": 3.590736084915406,
      "tokens_seen": 1188167680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232397191574724,
      "loss": 2.5451,
      "theoretical_loss": 3.5907178331938825,
      "tokens_seen": 1188233216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032322968906720165,
      "loss": 2.8114,
      "theoretical_loss": 3.5906995827608355,
      "tokens_seen": 1188298752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032321965897693077,
      "loss": 2.7601,
      "theoretical_loss": 3.5906813336161036,
      "tokens_seen": 1188364288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032320962888666,
      "loss": 2.8106,
      "theoretical_loss": 3.590663085759525,
      "tokens_seen": 1188429824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032319959879638913,
      "loss": 2.6096,
      "theoretical_loss": 3.5906448391909382,
      "tokens_seen": 1188495360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032318956870611837,
      "loss": 2.7287,
      "theoretical_loss": 3.59062659391018,
      "tokens_seen": 1188560896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032317953861584755,
      "loss": 2.734,
      "theoretical_loss": 3.5906083499170895,
      "tokens_seen": 1188626432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032316950852557673,
      "loss": 2.9007,
      "theoretical_loss": 3.590590107211504,
      "tokens_seen": 1188691968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231594784353059,
      "loss": 2.624,
      "theoretical_loss": 3.5905718657932626,
      "tokens_seen": 1188757504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032314944834503515,
      "loss": 2.5711,
      "theoretical_loss": 3.590553625662203,
      "tokens_seen": 1188823040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231394182547643,
      "loss": 2.6533,
      "theoretical_loss": 3.5905353868181633,
      "tokens_seen": 1188888576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231293881644935,
      "loss": 2.6393,
      "theoretical_loss": 3.5905171492609815,
      "tokens_seen": 1188954112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032311935807422264,
      "loss": 2.7133,
      "theoretical_loss": 3.5904989129904967,
      "tokens_seen": 1189019648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231093279839519,
      "loss": 2.5837,
      "theoretical_loss": 3.5904806780065472,
      "tokens_seen": 1189085184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032309929789368105,
      "loss": 2.8506,
      "theoretical_loss": 3.5904624443089705,
      "tokens_seen": 1189150720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032308926780341024,
      "loss": 2.9039,
      "theoretical_loss": 3.590444211897606,
      "tokens_seen": 1189216256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230792377131394,
      "loss": 2.7202,
      "theoretical_loss": 3.5904259807722907,
      "tokens_seen": 1189281792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230692076228686,
      "loss": 2.6139,
      "theoretical_loss": 3.5904077509328642,
      "tokens_seen": 1189347328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230591775325978,
      "loss": 2.5795,
      "theoretical_loss": 3.590389522379165,
      "tokens_seen": 1189412864
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 659771,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9299919605255127,
      "objective/train/theoretical_loss": 3.5903712951110305,
      "objective/train/tokens_used": 1209938400,
      "theoretical_loss": 3.5903712951110305,
      "tokens_seen": 1189478400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323049147442327,
      "loss": 2.7602,
      "theoretical_loss": 3.5903712951110305,
      "tokens_seen": 1189478400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032303911735205614,
      "loss": 2.4382,
      "theoretical_loss": 3.5903530691283,
      "tokens_seen": 1189543936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230290872617854,
      "loss": 2.8542,
      "theoretical_loss": 3.590334844430812,
      "tokens_seen": 1189609472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230190571715145,
      "loss": 2.6869,
      "theoretical_loss": 3.590316621018405,
      "tokens_seen": 1189675008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032300902708124374,
      "loss": 2.7318,
      "theoretical_loss": 3.5902983988909183,
      "tokens_seen": 1189740544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229989969909729,
      "loss": 2.6897,
      "theoretical_loss": 3.5902801780481886,
      "tokens_seen": 1189806080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229889669007021,
      "loss": 2.5943,
      "theoretical_loss": 3.590261958490057,
      "tokens_seen": 1189871616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229789368104313,
      "loss": 2.5884,
      "theoretical_loss": 3.5902437402163603,
      "tokens_seen": 1189937152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229689067201605,
      "loss": 2.7583,
      "theoretical_loss": 3.5902255232269376,
      "tokens_seen": 1190002688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032295887662988964,
      "loss": 2.6598,
      "theoretical_loss": 3.5902073075216285,
      "tokens_seen": 1190068224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229488465396189,
      "loss": 2.6565,
      "theoretical_loss": 3.590189093100271,
      "tokens_seen": 1190133760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322938816449348,
      "loss": 2.7323,
      "theoretical_loss": 3.590170879962704,
      "tokens_seen": 1190199296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032292878635907724,
      "loss": 2.5401,
      "theoretical_loss": 3.590152668108767,
      "tokens_seen": 1190264832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229187562688064,
      "loss": 2.7671,
      "theoretical_loss": 3.590134457538298,
      "tokens_seen": 1190330368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003229087261785356,
      "loss": 2.6562,
      "theoretical_loss": 3.590116248251136,
      "tokens_seen": 1190395904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228986960882648,
      "loss": 2.5468,
      "theoretical_loss": 3.59009804024712,
      "tokens_seen": 1190461440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032288866599799397,
      "loss": 2.8533,
      "theoretical_loss": 3.5900798335260897,
      "tokens_seen": 1190526976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032287863590772315,
      "loss": 2.8279,
      "theoretical_loss": 3.590061628087883,
      "tokens_seen": 1190592512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228686058174524,
      "loss": 2.6782,
      "theoretical_loss": 3.5900434239323396,
      "tokens_seen": 1190658048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228585757271815,
      "loss": 2.668,
      "theoretical_loss": 3.590025221059298,
      "tokens_seen": 1190723584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032284854563691074,
      "loss": 2.6663,
      "theoretical_loss": 3.5900070194685982,
      "tokens_seen": 1190789120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228385155466399,
      "loss": 2.9798,
      "theoretical_loss": 3.589988819160078,
      "tokens_seen": 1190854656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228284854563691,
      "loss": 2.5456,
      "theoretical_loss": 3.5899706201335775,
      "tokens_seen": 1190920192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032281845536609834,
      "loss": 2.7142,
      "theoretical_loss": 3.5899524223889356,
      "tokens_seen": 1190985728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032280842527582747,
      "loss": 2.6882,
      "theoretical_loss": 3.589934225925991,
      "tokens_seen": 1191051264
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 660497,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.826659917831421,
      "objective/train/theoretical_loss": 3.589916030744584,
      "objective/train/tokens_used": 1211576800,
      "theoretical_loss": 3.589916030744584,
      "tokens_seen": 1191116800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227983951855567,
      "loss": 2.8398,
      "theoretical_loss": 3.589916030744584,
      "tokens_seen": 1191116800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227883650952859,
      "loss": 2.536,
      "theoretical_loss": 3.589897836844553,
      "tokens_seen": 1191182336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032277833500501507,
      "loss": 2.5995,
      "theoretical_loss": 3.589879644225737,
      "tokens_seen": 1191247872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032276830491474425,
      "loss": 2.7609,
      "theoretical_loss": 3.589861452887976,
      "tokens_seen": 1191313408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032275827482447343,
      "loss": 2.5637,
      "theoretical_loss": 3.5898432628311094,
      "tokens_seen": 1191378944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227482447342026,
      "loss": 2.8383,
      "theoretical_loss": 3.5898250740549758,
      "tokens_seen": 1191444480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032273821464393185,
      "loss": 2.5828,
      "theoretical_loss": 3.5898068865594155,
      "tokens_seen": 1191510016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032272818455366097,
      "loss": 2.7367,
      "theoretical_loss": 3.589788700344267,
      "tokens_seen": 1191575552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227181544633902,
      "loss": 2.7134,
      "theoretical_loss": 3.5897705154093704,
      "tokens_seen": 1191641088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032270812437311933,
      "loss": 2.535,
      "theoretical_loss": 3.589752331754565,
      "tokens_seen": 1191706624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032269809428284857,
      "loss": 2.6662,
      "theoretical_loss": 3.58973414937969,
      "tokens_seen": 1191772160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032268806419257775,
      "loss": 2.5705,
      "theoretical_loss": 3.589715968284585,
      "tokens_seen": 1191837696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032267803410230693,
      "loss": 2.7252,
      "theoretical_loss": 3.58969778846909,
      "tokens_seen": 1191903232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226680040120361,
      "loss": 2.596,
      "theoretical_loss": 3.589679609933045,
      "tokens_seen": 1191968768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032265797392176535,
      "loss": 2.5493,
      "theoretical_loss": 3.589661432676288,
      "tokens_seen": 1192034304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226479438314945,
      "loss": 2.5605,
      "theoretical_loss": 3.58964325669866,
      "tokens_seen": 1192099840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226379137412237,
      "loss": 2.7074,
      "theoretical_loss": 3.589625082,
      "tokens_seen": 1192165376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032262788365095284,
      "loss": 2.7322,
      "theoretical_loss": 3.5896069085801483,
      "tokens_seen": 1192230912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226178535606821,
      "loss": 2.8322,
      "theoretical_loss": 3.5895887364389445,
      "tokens_seen": 1192296448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032260782347041125,
      "loss": 2.6919,
      "theoretical_loss": 3.589570565576228,
      "tokens_seen": 1192361984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032259779338014044,
      "loss": 2.708,
      "theoretical_loss": 3.5895523959918383,
      "tokens_seen": 1192427520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225877632898696,
      "loss": 2.7512,
      "theoretical_loss": 3.5895342276856157,
      "tokens_seen": 1192493056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225777331995988,
      "loss": 2.5763,
      "theoretical_loss": 3.5895160606574006,
      "tokens_seen": 1192558592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322567703109328,
      "loss": 2.7169,
      "theoretical_loss": 3.5894978949070317,
      "tokens_seen": 1192624128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225576730190572,
      "loss": 2.4533,
      "theoretical_loss": 3.58947973043435,
      "tokens_seen": 1192689664
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 661795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.610649347305298,
      "objective/train/theoretical_loss": 3.5894615672391947,
      "objective/train/tokens_used": 1213215200,
      "theoretical_loss": 3.5894615672391947,
      "tokens_seen": 1192755200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032254764292878634,
      "loss": 2.6943,
      "theoretical_loss": 3.5894615672391947,
      "tokens_seen": 1192755200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225376128385156,
      "loss": 2.7775,
      "theoretical_loss": 3.589443405321406,
      "tokens_seen": 1192820736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225275827482447,
      "loss": 2.541,
      "theoretical_loss": 3.5894252446808244,
      "tokens_seen": 1192886272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032251755265797394,
      "loss": 2.6342,
      "theoretical_loss": 3.589407085317289,
      "tokens_seen": 1192951808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225075225677031,
      "loss": 2.2844,
      "theoretical_loss": 3.5893889272306403,
      "tokens_seen": 1193017344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224974924774323,
      "loss": 2.7089,
      "theoretical_loss": 3.589370770420718,
      "tokens_seen": 1193082880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224874623871615,
      "loss": 2.4681,
      "theoretical_loss": 3.589352614887363,
      "tokens_seen": 1193148416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224774322968907,
      "loss": 2.833,
      "theoretical_loss": 3.5893344606304147,
      "tokens_seen": 1193213952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032246740220661984,
      "loss": 2.5555,
      "theoretical_loss": 3.589316307649714,
      "tokens_seen": 1193279488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224573721163491,
      "loss": 2.656,
      "theoretical_loss": 3.5892981559451003,
      "tokens_seen": 1193345024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224473420260782,
      "loss": 2.5065,
      "theoretical_loss": 3.5892800055164145,
      "tokens_seen": 1193410560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032243731193580744,
      "loss": 2.4987,
      "theoretical_loss": 3.589261856363496,
      "tokens_seen": 1193476096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224272818455366,
      "loss": 2.7401,
      "theoretical_loss": 3.589243708486186,
      "tokens_seen": 1193541632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224172517552658,
      "loss": 2.5925,
      "theoretical_loss": 3.5892255618843247,
      "tokens_seen": 1193607168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322407221664995,
      "loss": 2.8549,
      "theoretical_loss": 3.589207416557752,
      "tokens_seen": 1193672704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032239719157472417,
      "loss": 2.7572,
      "theoretical_loss": 3.589189272506308,
      "tokens_seen": 1193738240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032238716148445335,
      "loss": 2.676,
      "theoretical_loss": 3.5891711297298334,
      "tokens_seen": 1193803776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223771313941826,
      "loss": 2.5692,
      "theoretical_loss": 3.5891529882281694,
      "tokens_seen": 1193869312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223671013039117,
      "loss": 2.6654,
      "theoretical_loss": 3.5891348480011556,
      "tokens_seen": 1193934848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032235707121364095,
      "loss": 2.4853,
      "theoretical_loss": 3.589116709048633,
      "tokens_seen": 1194000384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223470411233701,
      "loss": 2.5927,
      "theoretical_loss": 3.589098571370441,
      "tokens_seen": 1194065920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223370110330993,
      "loss": 2.6482,
      "theoretical_loss": 3.589080434966421,
      "tokens_seen": 1194131456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223269809428285,
      "loss": 2.5284,
      "theoretical_loss": 3.5890622998364146,
      "tokens_seen": 1194196992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032231695085255767,
      "loss": 2.65,
      "theoretical_loss": 3.58904416598026,
      "tokens_seen": 1194262528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032230692076228685,
      "loss": 2.4034,
      "theoretical_loss": 3.5890260333977997,
      "tokens_seen": 1194328064
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 662071,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3905045986175537,
      "objective/train/theoretical_loss": 3.589007902088874,
      "objective/train/tokens_used": 1214853600,
      "theoretical_loss": 3.589007902088874,
      "tokens_seen": 1194393600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222968906720161,
      "loss": 2.4546,
      "theoretical_loss": 3.589007902088874,
      "tokens_seen": 1194393600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222868605817452,
      "loss": 2.8831,
      "theoretical_loss": 3.5889897720533233,
      "tokens_seen": 1194459136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032227683049147445,
      "loss": 2.6318,
      "theoretical_loss": 3.5889716432909875,
      "tokens_seen": 1194524672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222668004012036,
      "loss": 2.4259,
      "theoretical_loss": 3.588953515801709,
      "tokens_seen": 1194590208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222567703109328,
      "loss": 2.7315,
      "theoretical_loss": 3.5889353895853278,
      "tokens_seen": 1194655744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322246740220662,
      "loss": 2.7919,
      "theoretical_loss": 3.588917264641685,
      "tokens_seen": 1194721280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032223671013039117,
      "loss": 2.5828,
      "theoretical_loss": 3.5888991409706206,
      "tokens_seen": 1194786816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032222668004012035,
      "loss": 2.6261,
      "theoretical_loss": 3.5888810185719757,
      "tokens_seen": 1194852352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032221664994984953,
      "loss": 2.4561,
      "theoretical_loss": 3.5888628974455914,
      "tokens_seen": 1194917888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222066198595787,
      "loss": 2.5928,
      "theoretical_loss": 3.5888447775913095,
      "tokens_seen": 1194983424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032219658976930795,
      "loss": 2.5253,
      "theoretical_loss": 3.5888266590089692,
      "tokens_seen": 1195048960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221865596790371,
      "loss": 2.6944,
      "theoretical_loss": 3.588808541698413,
      "tokens_seen": 1195114496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221765295887663,
      "loss": 2.4546,
      "theoretical_loss": 3.588790425659481,
      "tokens_seen": 1195180032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221664994984955,
      "loss": 2.6356,
      "theoretical_loss": 3.588772310892015,
      "tokens_seen": 1195245568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221564694082247,
      "loss": 2.6211,
      "theoretical_loss": 3.5887541973958546,
      "tokens_seen": 1195311104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032214643931795386,
      "loss": 2.758,
      "theoretical_loss": 3.5887360851708423,
      "tokens_seen": 1195376640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032213640922768304,
      "loss": 2.7939,
      "theoretical_loss": 3.5887179742168187,
      "tokens_seen": 1195442176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221263791374122,
      "loss": 2.616,
      "theoretical_loss": 3.588699864533625,
      "tokens_seen": 1195507712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032211634904714145,
      "loss": 2.6652,
      "theoretical_loss": 3.5886817561211024,
      "tokens_seen": 1195573248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221063189568706,
      "loss": 2.7012,
      "theoretical_loss": 3.588663648979092,
      "tokens_seen": 1195638784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220962888665998,
      "loss": 2.7739,
      "theoretical_loss": 3.588645543107435,
      "tokens_seen": 1195704320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032208625877632894,
      "loss": 2.8342,
      "theoretical_loss": 3.5886274385059727,
      "tokens_seen": 1195769856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220762286860582,
      "loss": 2.6931,
      "theoretical_loss": 3.5886093351745467,
      "tokens_seen": 1195835392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220661985957874,
      "loss": 2.5972,
      "theoretical_loss": 3.5885912331129974,
      "tokens_seen": 1195900928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032205616850551654,
      "loss": 2.6297,
      "theoretical_loss": 3.5885731323211676,
      "tokens_seen": 1195966464
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 663277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6097090244293213,
      "objective/train/theoretical_loss": 3.5885550327988973,
      "objective/train/tokens_used": 1216492000,
      "theoretical_loss": 3.5885550327988973,
      "tokens_seen": 1196032000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220461384152458,
      "loss": 2.7212,
      "theoretical_loss": 3.5885550327988973,
      "tokens_seen": 1196032000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220361083249749,
      "loss": 2.7139,
      "theoretical_loss": 3.5885369345460285,
      "tokens_seen": 1196097536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032202607823470414,
      "loss": 2.5494,
      "theoretical_loss": 3.5885188375624026,
      "tokens_seen": 1196163072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220160481444333,
      "loss": 2.6198,
      "theoretical_loss": 3.588500741847861,
      "tokens_seen": 1196228608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220060180541625,
      "loss": 2.9059,
      "theoretical_loss": 3.5884826474022455,
      "tokens_seen": 1196294144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219959879638917,
      "loss": 2.7009,
      "theoretical_loss": 3.588464554225397,
      "tokens_seen": 1196359680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219859578736209,
      "loss": 2.7996,
      "theoretical_loss": 3.5884464623171572,
      "tokens_seen": 1196425216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032197592778335004,
      "loss": 2.7252,
      "theoretical_loss": 3.588428371677368,
      "tokens_seen": 1196490752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219658976930793,
      "loss": 2.623,
      "theoretical_loss": 3.588410282305871,
      "tokens_seen": 1196556288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219558676028084,
      "loss": 2.7392,
      "theoretical_loss": 3.588392194202507,
      "tokens_seen": 1196621824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032194583751253764,
      "loss": 2.8584,
      "theoretical_loss": 3.5883741073671187,
      "tokens_seen": 1196687360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219358074222668,
      "loss": 2.6099,
      "theoretical_loss": 3.5883560217995476,
      "tokens_seen": 1196752896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321925777331996,
      "loss": 2.8514,
      "theoretical_loss": 3.588337937499635,
      "tokens_seen": 1196818432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003219157472417252,
      "loss": 2.5845,
      "theoretical_loss": 3.5883198544672226,
      "tokens_seen": 1196883968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032190571715145437,
      "loss": 2.6211,
      "theoretical_loss": 3.588301772702153,
      "tokens_seen": 1196949504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032189568706118355,
      "loss": 2.783,
      "theoretical_loss": 3.588283692204267,
      "tokens_seen": 1197015040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218856569709128,
      "loss": 2.7052,
      "theoretical_loss": 3.5882656129734065,
      "tokens_seen": 1197080576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218756268806419,
      "loss": 2.7137,
      "theoretical_loss": 3.5882475350094136,
      "tokens_seen": 1197146112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032186559679037115,
      "loss": 2.5264,
      "theoretical_loss": 3.5882294583121306,
      "tokens_seen": 1197211648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218555667001003,
      "loss": 2.5417,
      "theoretical_loss": 3.5882113828813993,
      "tokens_seen": 1197277184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218455366098295,
      "loss": 2.6667,
      "theoretical_loss": 3.588193308717061,
      "tokens_seen": 1197342720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218355065195587,
      "loss": 2.6857,
      "theoretical_loss": 3.5881752358189583,
      "tokens_seen": 1197408256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032182547642928787,
      "loss": 2.7712,
      "theoretical_loss": 3.588157164186933,
      "tokens_seen": 1197473792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032181544633901705,
      "loss": 2.6639,
      "theoretical_loss": 3.5881390938208266,
      "tokens_seen": 1197539328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218054162487463,
      "loss": 2.7148,
      "theoretical_loss": 3.588121024720482,
      "tokens_seen": 1197604864
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 664536,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6411828994750977,
      "objective/train/theoretical_loss": 3.588102956885741,
      "objective/train/tokens_used": 1218130400,
      "theoretical_loss": 3.588102956885741,
      "tokens_seen": 1197670400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217953861584754,
      "loss": 2.83,
      "theoretical_loss": 3.588102956885741,
      "tokens_seen": 1197670400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032178535606820465,
      "loss": 2.5972,
      "theoretical_loss": 3.5880848903164457,
      "tokens_seen": 1197735936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217753259779338,
      "loss": 2.7223,
      "theoretical_loss": 3.5880668250124375,
      "tokens_seen": 1197801472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321765295887663,
      "loss": 2.7356,
      "theoretical_loss": 3.5880487609735594,
      "tokens_seen": 1197867008
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217552657973922,
      "loss": 2.7346,
      "theoretical_loss": 3.588030698199654,
      "tokens_seen": 1197932544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032174523570712137,
      "loss": 2.575,
      "theoretical_loss": 3.5880126366905625,
      "tokens_seen": 1197998080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032173520561685055,
      "loss": 2.4085,
      "theoretical_loss": 3.5879945764461274,
      "tokens_seen": 1198063616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032172517552657974,
      "loss": 2.6027,
      "theoretical_loss": 3.5879765174661915,
      "tokens_seen": 1198129152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217151454363089,
      "loss": 2.6716,
      "theoretical_loss": 3.5879584597505967,
      "tokens_seen": 1198194688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032170511534603815,
      "loss": 2.61,
      "theoretical_loss": 3.5879404032991853,
      "tokens_seen": 1198260224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216950852557673,
      "loss": 2.5906,
      "theoretical_loss": 3.5879223481117997,
      "tokens_seen": 1198325760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216850551654965,
      "loss": 2.5357,
      "theoretical_loss": 3.5879042941882826,
      "tokens_seen": 1198391296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216750250752257,
      "loss": 2.8647,
      "theoretical_loss": 3.5878862415284765,
      "tokens_seen": 1198456832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216649949849549,
      "loss": 2.5841,
      "theoretical_loss": 3.5878681901322227,
      "tokens_seen": 1198522368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032165496489468406,
      "loss": 2.6733,
      "theoretical_loss": 3.5878501399993645,
      "tokens_seen": 1198587904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032164493480441324,
      "loss": 2.5762,
      "theoretical_loss": 3.587832091129745,
      "tokens_seen": 1198653440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216349047141424,
      "loss": 2.6126,
      "theoretical_loss": 3.587814043523206,
      "tokens_seen": 1198718976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032162487462387165,
      "loss": 2.3205,
      "theoretical_loss": 3.5877959971795903,
      "tokens_seen": 1198784512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216148445336008,
      "loss": 2.6872,
      "theoretical_loss": 3.58777795209874,
      "tokens_seen": 1198850048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032160481444333,
      "loss": 2.7524,
      "theoretical_loss": 3.587759908280498,
      "tokens_seen": 1198915584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032159478435305914,
      "loss": 2.7265,
      "theoretical_loss": 3.5877418657247073,
      "tokens_seen": 1198981120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215847542627884,
      "loss": 2.7187,
      "theoretical_loss": 3.5877238244312104,
      "tokens_seen": 1199046656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032157472417251756,
      "loss": 2.8152,
      "theoretical_loss": 3.5877057843998497,
      "tokens_seen": 1199112192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032156469408224674,
      "loss": 2.6784,
      "theoretical_loss": 3.5876877456304683,
      "tokens_seen": 1199177728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215546639919759,
      "loss": 2.6942,
      "theoretical_loss": 3.5876697081229088,
      "tokens_seen": 1199243264
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 665055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5700125694274902,
      "objective/train/theoretical_loss": 3.587651671877014,
      "objective/train/tokens_used": 1219768800,
      "theoretical_loss": 3.587651671877014,
      "tokens_seen": 1199308800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215446339017051,
      "loss": 2.676,
      "theoretical_loss": 3.587651671877014,
      "tokens_seen": 1199308800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215346038114343,
      "loss": 2.5606,
      "theoretical_loss": 3.587633636892627,
      "tokens_seen": 1199374336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215245737211635,
      "loss": 2.5459,
      "theoretical_loss": 3.5876156031695903,
      "tokens_seen": 1199439872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032151454363089265,
      "loss": 2.517,
      "theoretical_loss": 3.587597570707746,
      "tokens_seen": 1199505408
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215045135406219,
      "loss": 2.6024,
      "theoretical_loss": 3.5875795395069385,
      "tokens_seen": 1199570944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032149448345035106,
      "loss": 2.8159,
      "theoretical_loss": 3.58756150956701,
      "tokens_seen": 1199636480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032148445336008024,
      "loss": 2.6025,
      "theoretical_loss": 3.5875434808878035,
      "tokens_seen": 1199702016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214744232698094,
      "loss": 2.4649,
      "theoretical_loss": 3.587525453469162,
      "tokens_seen": 1199767552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214643931795386,
      "loss": 2.8485,
      "theoretical_loss": 3.5875074273109284,
      "tokens_seen": 1199833088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214543630892678,
      "loss": 2.6042,
      "theoretical_loss": 3.587489402412946,
      "tokens_seen": 1199898624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321444332998997,
      "loss": 2.7668,
      "theoretical_loss": 3.5874713787750574,
      "tokens_seen": 1199964160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032143430290872615,
      "loss": 2.9209,
      "theoretical_loss": 3.587453356397106,
      "tokens_seen": 1200029696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214242728184554,
      "loss": 2.8785,
      "theoretical_loss": 3.587435335278935,
      "tokens_seen": 1200095232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214142427281845,
      "loss": 2.4521,
      "theoretical_loss": 3.587417315420388,
      "tokens_seen": 1200160768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032140421263791375,
      "loss": 2.8288,
      "theoretical_loss": 3.5873992968213066,
      "tokens_seen": 1200226304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032139418254764293,
      "loss": 2.6307,
      "theoretical_loss": 3.5873812794815363,
      "tokens_seen": 1200291840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213841524573721,
      "loss": 2.8476,
      "theoretical_loss": 3.587363263400918,
      "tokens_seen": 1200357376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213741223671013,
      "loss": 2.7485,
      "theoretical_loss": 3.5873452485792967,
      "tokens_seen": 1200422912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213640922768305,
      "loss": 2.7168,
      "theoretical_loss": 3.587327235016515,
      "tokens_seen": 1200488448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032135406218655965,
      "loss": 2.4365,
      "theoretical_loss": 3.587309222712416,
      "tokens_seen": 1200553984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213440320962889,
      "loss": 2.6422,
      "theoretical_loss": 3.5872912116668436,
      "tokens_seen": 1200619520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321334002006018,
      "loss": 2.6353,
      "theoretical_loss": 3.5872732018796407,
      "tokens_seen": 1200685056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032132397191574725,
      "loss": 2.6725,
      "theoretical_loss": 3.587255193350651,
      "tokens_seen": 1200750592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213139418254765,
      "loss": 2.6841,
      "theoretical_loss": 3.587237186079718,
      "tokens_seen": 1200816128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213039117352056,
      "loss": 2.5766,
      "theoretical_loss": 3.587219180066685,
      "tokens_seen": 1200881664
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 666208,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7520852088928223,
      "objective/train/theoretical_loss": 3.587201175311395,
      "objective/train/tokens_used": 1221407200,
      "theoretical_loss": 3.587201175311395,
      "tokens_seen": 1200947200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032129388164493485,
      "loss": 2.8138,
      "theoretical_loss": 3.587201175311395,
      "tokens_seen": 1200947200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321283851554664,
      "loss": 2.8434,
      "theoretical_loss": 3.5871831718136926,
      "tokens_seen": 1201012736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212738214643932,
      "loss": 2.8096,
      "theoretical_loss": 3.5871651695734204,
      "tokens_seen": 1201078272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212637913741224,
      "loss": 2.446,
      "theoretical_loss": 3.5871471685904224,
      "tokens_seen": 1201143808
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212537612838516,
      "loss": 2.709,
      "theoretical_loss": 3.5871291688645424,
      "tokens_seen": 1201209344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032124373119358075,
      "loss": 2.5374,
      "theoretical_loss": 3.5871111703956235,
      "tokens_seen": 1201274880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032123370110330994,
      "loss": 2.5882,
      "theoretical_loss": 3.5870931731835096,
      "tokens_seen": 1201340416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212236710130391,
      "loss": 2.496,
      "theoretical_loss": 3.5870751772280443,
      "tokens_seen": 1201405952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032121364092276835,
      "loss": 2.4976,
      "theoretical_loss": 3.5870571825290716,
      "tokens_seen": 1201471488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212036108324975,
      "loss": 2.6233,
      "theoretical_loss": 3.5870391890864353,
      "tokens_seen": 1201537024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211935807422267,
      "loss": 2.4834,
      "theoretical_loss": 3.587021196899978,
      "tokens_seen": 1201602560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211835506519559,
      "loss": 2.5302,
      "theoretical_loss": 3.5870032059695456,
      "tokens_seen": 1201668096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211735205616851,
      "loss": 2.6988,
      "theoretical_loss": 3.58698521629498,
      "tokens_seen": 1201733632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032116349047141426,
      "loss": 2.5982,
      "theoretical_loss": 3.586967227876126,
      "tokens_seen": 1201799168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032115346038114344,
      "loss": 2.7329,
      "theoretical_loss": 3.586949240712827,
      "tokens_seen": 1201864704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211434302908726,
      "loss": 2.6829,
      "theoretical_loss": 3.5869312548049277,
      "tokens_seen": 1201930240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032113340020060185,
      "loss": 2.883,
      "theoretical_loss": 3.5869132701522712,
      "tokens_seen": 1201995776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321123370110331,
      "loss": 2.6068,
      "theoretical_loss": 3.586895286754702,
      "tokens_seen": 1202061312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211133400200602,
      "loss": 2.6619,
      "theoretical_loss": 3.586877304612064,
      "tokens_seen": 1202126848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032110330992978934,
      "loss": 2.6131,
      "theoretical_loss": 3.5868593237242004,
      "tokens_seen": 1202192384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210932798395186,
      "loss": 2.737,
      "theoretical_loss": 3.5868413440909563,
      "tokens_seen": 1202257920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032108324974924776,
      "loss": 2.747,
      "theoretical_loss": 3.5868233657121755,
      "tokens_seen": 1202323456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032107321965897694,
      "loss": 2.7392,
      "theoretical_loss": 3.5868053885877016,
      "tokens_seen": 1202388992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210631895687061,
      "loss": 2.5638,
      "theoretical_loss": 3.5867874127173796,
      "tokens_seen": 1202454528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210531594784353,
      "loss": 2.6974,
      "theoretical_loss": 3.5867694381010535,
      "tokens_seen": 1202520064
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 666794,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.74890398979187,
      "objective/train/theoretical_loss": 3.5867514647385663,
      "objective/train/tokens_used": 1223045600,
      "theoretical_loss": 3.5867514647385663,
      "tokens_seen": 1202585600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210431293881645,
      "loss": 2.7497,
      "theoretical_loss": 3.5867514647385663,
      "tokens_seen": 1202585600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210330992978937,
      "loss": 2.5142,
      "theoretical_loss": 3.586733492629764,
      "tokens_seen": 1202651136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032102306920762285,
      "loss": 2.6491,
      "theoretical_loss": 3.5867155217744893,
      "tokens_seen": 1202716672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210130391173521,
      "loss": 2.8934,
      "theoretical_loss": 3.5866975521725877,
      "tokens_seen": 1202782208
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032100300902708126,
      "loss": 2.7662,
      "theoretical_loss": 3.5866795838239023,
      "tokens_seen": 1202847744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032099297893681044,
      "loss": 2.738,
      "theoretical_loss": 3.5866616167282785,
      "tokens_seen": 1202913280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209829488465396,
      "loss": 2.6258,
      "theoretical_loss": 3.58664365088556,
      "tokens_seen": 1202978816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209729187562688,
      "loss": 2.6219,
      "theoretical_loss": 3.5866256862955916,
      "tokens_seen": 1203044352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320962888665998,
      "loss": 2.5657,
      "theoretical_loss": 3.5866077229582176,
      "tokens_seen": 1203109888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209528585757272,
      "loss": 2.6135,
      "theoretical_loss": 3.5865897608732817,
      "tokens_seen": 1203175424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032094282848545635,
      "loss": 2.7004,
      "theoretical_loss": 3.5865718000406295,
      "tokens_seen": 1203240960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209327983951856,
      "loss": 2.7194,
      "theoretical_loss": 3.586553840460105,
      "tokens_seen": 1203306496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209227683049147,
      "loss": 2.6563,
      "theoretical_loss": 3.5865358821315527,
      "tokens_seen": 1203372032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032091273821464395,
      "loss": 2.797,
      "theoretical_loss": 3.5865179250548174,
      "tokens_seen": 1203437568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032090270812437313,
      "loss": 2.6798,
      "theoretical_loss": 3.586499969229743,
      "tokens_seen": 1203503104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208926780341023,
      "loss": 2.8114,
      "theoretical_loss": 3.5864820146561747,
      "tokens_seen": 1203568640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208826479438315,
      "loss": 2.5543,
      "theoretical_loss": 3.586464061333957,
      "tokens_seen": 1203634176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208726178535607,
      "loss": 2.6157,
      "theoretical_loss": 3.5864461092629343,
      "tokens_seen": 1203699712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032086258776328985,
      "loss": 2.8626,
      "theoretical_loss": 3.586428158442952,
      "tokens_seen": 1203765248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208525576730191,
      "loss": 2.3889,
      "theoretical_loss": 3.5864102088738536,
      "tokens_seen": 1203830784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208425275827482,
      "loss": 2.8346,
      "theoretical_loss": 3.586392260555485,
      "tokens_seen": 1203896320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032083249749247745,
      "loss": 2.9258,
      "theoretical_loss": 3.5863743134876906,
      "tokens_seen": 1203961856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032082246740220663,
      "loss": 2.5813,
      "theoretical_loss": 3.586356367670315,
      "tokens_seen": 1204027392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208124373119358,
      "loss": 2.7728,
      "theoretical_loss": 3.5863384231032036,
      "tokens_seen": 1204092928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320802407221665,
      "loss": 2.6154,
      "theoretical_loss": 3.5863204797862003,
      "tokens_seen": 1204158464
    },
    {
      "epoch": 0.36,
      "objective/train/docs_used": 667638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.698781728744507,
      "objective/train/theoretical_loss": 3.586302537719151,
      "objective/train/tokens_used": 1224684000,
      "theoretical_loss": 3.586302537719151,
      "tokens_seen": 1204224000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207923771313942,
      "loss": 2.659,
      "theoretical_loss": 3.586302537719151,
      "tokens_seen": 1204224000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032078234704112336,
      "loss": 2.6391,
      "theoretical_loss": 3.5862845969018995,
      "tokens_seen": 1204289536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207723169508526,
      "loss": 2.5357,
      "theoretical_loss": 3.586266657334292,
      "tokens_seen": 1204355072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207622868605817,
      "loss": 2.5632,
      "theoretical_loss": 3.5862487190161723,
      "tokens_seen": 1204420608
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032075225677031095,
      "loss": 2.7713,
      "theoretical_loss": 3.5862307819473864,
      "tokens_seen": 1204486144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207422266800401,
      "loss": 2.5229,
      "theoretical_loss": 3.5862128461277782,
      "tokens_seen": 1204551680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207321965897693,
      "loss": 2.493,
      "theoretical_loss": 3.586194911557194,
      "tokens_seen": 1204617216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207221664994985,
      "loss": 2.8192,
      "theoretical_loss": 3.586176978235478,
      "tokens_seen": 1204682752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207121364092277,
      "loss": 2.8019,
      "theoretical_loss": 3.586159046162476,
      "tokens_seen": 1204748288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032070210631895686,
      "loss": 2.4404,
      "theoretical_loss": 3.586141115338032,
      "tokens_seen": 1204813824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206920762286861,
      "loss": 2.7097,
      "theoretical_loss": 3.586123185761992,
      "tokens_seen": 1204879360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206820461384152,
      "loss": 2.8557,
      "theoretical_loss": 3.586105257434201,
      "tokens_seen": 1204944896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032067201604814446,
      "loss": 2.8357,
      "theoretical_loss": 3.5860873303545047,
      "tokens_seen": 1205010432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206619859578736,
      "loss": 2.8592,
      "theoretical_loss": 3.586069404522747,
      "tokens_seen": 1205075968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206519558676028,
      "loss": 2.8587,
      "theoretical_loss": 3.586051479938775,
      "tokens_seen": 1205141504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320641925777332,
      "loss": 2.7109,
      "theoretical_loss": 3.5860335566024326,
      "tokens_seen": 1205207040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206318956870612,
      "loss": 2.5293,
      "theoretical_loss": 3.586015634513566,
      "tokens_seen": 1205272576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032062186559679036,
      "loss": 2.7344,
      "theoretical_loss": 3.585997713672019,
      "tokens_seen": 1205338112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032061183550651954,
      "loss": 2.5098,
      "theoretical_loss": 3.5859797940776392,
      "tokens_seen": 1205403648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206018054162487,
      "loss": 2.7707,
      "theoretical_loss": 3.5859618757302707,
      "tokens_seen": 1205469184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032059177532597796,
      "loss": 2.7907,
      "theoretical_loss": 3.585943958629759,
      "tokens_seen": 1205534720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205817452357071,
      "loss": 2.6121,
      "theoretical_loss": 3.5859260427759496,
      "tokens_seen": 1205600256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205717151454363,
      "loss": 2.8092,
      "theoretical_loss": 3.5859081281686884,
      "tokens_seen": 1205665792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205616850551655,
      "loss": 2.4816,
      "theoretical_loss": 3.5858902148078204,
      "tokens_seen": 1205731328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205516549648947,
      "loss": 2.4486,
      "theoretical_loss": 3.5858723026931916,
      "tokens_seen": 1205796864
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 668120,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.765895366668701,
      "objective/train/theoretical_loss": 3.585854391824647,
      "objective/train/tokens_used": 1226322400,
      "theoretical_loss": 3.585854391824647,
      "tokens_seen": 1205862400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205416248746239,
      "loss": 3.0271,
      "theoretical_loss": 3.585854391824647,
      "tokens_seen": 1205862400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032053159478435305,
      "loss": 2.6968,
      "theoretical_loss": 3.5858364822020325,
      "tokens_seen": 1205927936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205215646940823,
      "loss": 2.7393,
      "theoretical_loss": 3.585818573825194,
      "tokens_seen": 1205993472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032051153460381146,
      "loss": 2.6489,
      "theoretical_loss": 3.585800666693977,
      "tokens_seen": 1206059008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032050150451354064,
      "loss": 2.6772,
      "theoretical_loss": 3.585782760808227,
      "tokens_seen": 1206124544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204914744232698,
      "loss": 2.5699,
      "theoretical_loss": 3.5857648561677893,
      "tokens_seen": 1206190080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320481444332999,
      "loss": 2.8226,
      "theoretical_loss": 3.585746952772511,
      "tokens_seen": 1206255616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204714142427282,
      "loss": 2.678,
      "theoretical_loss": 3.5857290506222363,
      "tokens_seen": 1206321152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204613841524574,
      "loss": 2.6826,
      "theoretical_loss": 3.585711149716812,
      "tokens_seen": 1206386688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032045135406218655,
      "loss": 2.725,
      "theoretical_loss": 3.5856932500560834,
      "tokens_seen": 1206452224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204413239719158,
      "loss": 2.7681,
      "theoretical_loss": 3.585675351639897,
      "tokens_seen": 1206517760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204312938816449,
      "loss": 2.7728,
      "theoretical_loss": 3.5856574544680977,
      "tokens_seen": 1206583296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032042126379137415,
      "loss": 2.6691,
      "theoretical_loss": 3.5856395585405325,
      "tokens_seen": 1206648832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032041123370110333,
      "loss": 2.6117,
      "theoretical_loss": 3.5856216638570464,
      "tokens_seen": 1206714368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204012036108325,
      "loss": 2.7022,
      "theoretical_loss": 3.5856037704174857,
      "tokens_seen": 1206779904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203911735205617,
      "loss": 2.757,
      "theoretical_loss": 3.5855858782216963,
      "tokens_seen": 1206845440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203811434302909,
      "loss": 2.7985,
      "theoretical_loss": 3.585567987269525,
      "tokens_seen": 1206910976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032037111334002005,
      "loss": 2.3973,
      "theoretical_loss": 3.5855500975608163,
      "tokens_seen": 1206976512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203610832497493,
      "loss": 2.6385,
      "theoretical_loss": 3.5855322090954176,
      "tokens_seen": 1207042048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203510531594784,
      "loss": 2.6184,
      "theoretical_loss": 3.585514321873174,
      "tokens_seen": 1207107584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032034102306920765,
      "loss": 2.8333,
      "theoretical_loss": 3.5854964358939325,
      "tokens_seen": 1207173120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032033099297893683,
      "loss": 2.6203,
      "theoretical_loss": 3.5854785511575393,
      "tokens_seen": 1207238656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032033099297893683,
      "loss": 2.6739,
      "theoretical_loss": 3.5854606676638396,
      "tokens_seen": 1207304192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320320962888666,
      "loss": 2.7132,
      "theoretical_loss": 3.5854427854126802,
      "tokens_seen": 1207369728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203109327983952,
      "loss": 2.6528,
      "theoretical_loss": 3.585424904403908,
      "tokens_seen": 1207435264
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 669193,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.440397024154663,
      "objective/train/theoretical_loss": 3.5854070246373677,
      "objective/train/tokens_used": 1227960800,
      "theoretical_loss": 3.5854070246373677,
      "tokens_seen": 1207500800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203009027081244,
      "loss": 2.4402,
      "theoretical_loss": 3.5854070246373677,
      "tokens_seen": 1207500800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032029087261785356,
      "loss": 2.8087,
      "theoretical_loss": 3.585389146112907,
      "tokens_seen": 1207566336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202808425275828,
      "loss": 2.79,
      "theoretical_loss": 3.5853712688303707,
      "tokens_seen": 1207631872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202708124373119,
      "loss": 2.7522,
      "theoretical_loss": 3.585353392789607,
      "tokens_seen": 1207697408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032026078234704115,
      "loss": 2.6718,
      "theoretical_loss": 3.585335517990461,
      "tokens_seen": 1207762944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202507522567703,
      "loss": 2.6665,
      "theoretical_loss": 3.58531764443278,
      "tokens_seen": 1207828480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202407221664995,
      "loss": 2.4913,
      "theoretical_loss": 3.585299772116409,
      "tokens_seen": 1207894016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202306920762287,
      "loss": 2.8258,
      "theoretical_loss": 3.5852819010411956,
      "tokens_seen": 1207959552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202206619859579,
      "loss": 2.7235,
      "theoretical_loss": 3.5852640312069863,
      "tokens_seen": 1208025088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032021063189568706,
      "loss": 2.5067,
      "theoretical_loss": 3.585246162613627,
      "tokens_seen": 1208090624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202006018054163,
      "loss": 2.7812,
      "theoretical_loss": 3.585228295260965,
      "tokens_seen": 1208156160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201905717151454,
      "loss": 2.7972,
      "theoretical_loss": 3.5852104291488462,
      "tokens_seen": 1208221696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032018054162487466,
      "loss": 2.5162,
      "theoretical_loss": 3.5851925642771167,
      "tokens_seen": 1208287232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201705115346038,
      "loss": 2.7096,
      "theoretical_loss": 3.5851747006456245,
      "tokens_seen": 1208352768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320160481444333,
      "loss": 2.7143,
      "theoretical_loss": 3.585156838254216,
      "tokens_seen": 1208418304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201504513540622,
      "loss": 2.818,
      "theoretical_loss": 3.585138977102736,
      "tokens_seen": 1208483840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201404212637914,
      "loss": 2.6218,
      "theoretical_loss": 3.5851211171910338,
      "tokens_seen": 1208549376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032013039117352056,
      "loss": 2.395,
      "theoretical_loss": 3.5851032585189544,
      "tokens_seen": 1208614912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032012036108324974,
      "loss": 2.3369,
      "theoretical_loss": 3.585085401086345,
      "tokens_seen": 1208680448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201103309929789,
      "loss": 2.5753,
      "theoretical_loss": 3.585067544893053,
      "tokens_seen": 1208745984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032010030090270816,
      "loss": 2.8804,
      "theoretical_loss": 3.5850496899389244,
      "tokens_seen": 1208811520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200902708124373,
      "loss": 2.6475,
      "theoretical_loss": 3.5850318362238056,
      "tokens_seen": 1208877056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200802407221665,
      "loss": 2.8044,
      "theoretical_loss": 3.5850139837475448,
      "tokens_seen": 1208942592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032007021063189565,
      "loss": 2.5214,
      "theoretical_loss": 3.584996132509988,
      "tokens_seen": 1209008128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200601805416249,
      "loss": 2.7903,
      "theoretical_loss": 3.5849782825109826,
      "tokens_seen": 1209073664
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 670409,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.449265480041504,
      "objective/train/theoretical_loss": 3.584960433750375,
      "objective/train/tokens_used": 1229599200,
      "theoretical_loss": 3.584960433750375,
      "tokens_seen": 1209139200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032005015045135407,
      "loss": 2.629,
      "theoretical_loss": 3.584960433750375,
      "tokens_seen": 1209139200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032004012036108325,
      "loss": 2.6543,
      "theoretical_loss": 3.5849425862280127,
      "tokens_seen": 1209204736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032003009027081243,
      "loss": 2.5549,
      "theoretical_loss": 3.5849247399437423,
      "tokens_seen": 1209270272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032002006018054166,
      "loss": 2.7935,
      "theoretical_loss": 3.5849068948974114,
      "tokens_seen": 1209335808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200100300902708,
      "loss": 2.693,
      "theoretical_loss": 3.584889051088866,
      "tokens_seen": 1209401344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032,
      "loss": 2.8575,
      "theoretical_loss": 3.5848712085179537,
      "tokens_seen": 1209466880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031998996990972915,
      "loss": 2.7422,
      "theoretical_loss": 3.584853367184522,
      "tokens_seen": 1209532416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199799398194584,
      "loss": 2.593,
      "theoretical_loss": 3.584835527088418,
      "tokens_seen": 1209597952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031996990972918757,
      "loss": 2.6048,
      "theoretical_loss": 3.584817688229488,
      "tokens_seen": 1209663488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031995987963891675,
      "loss": 2.7282,
      "theoretical_loss": 3.58479985060758,
      "tokens_seen": 1209729024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031994984954864593,
      "loss": 2.7549,
      "theoretical_loss": 3.584782014222541,
      "tokens_seen": 1209794560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199398194583751,
      "loss": 2.4356,
      "theoretical_loss": 3.5847641790742184,
      "tokens_seen": 1209860096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199297893681043,
      "loss": 2.677,
      "theoretical_loss": 3.584746345162459,
      "tokens_seen": 1209925632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031991975927783353,
      "loss": 2.7683,
      "theoretical_loss": 3.5847285124871107,
      "tokens_seen": 1209991168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031990972918756266,
      "loss": 2.6618,
      "theoretical_loss": 3.58471068104802,
      "tokens_seen": 1210056704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198996990972919,
      "loss": 2.6729,
      "theoretical_loss": 3.5846928508450353,
      "tokens_seen": 1210122240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319889669007021,
      "loss": 2.5604,
      "theoretical_loss": 3.5846750218780032,
      "tokens_seen": 1210187776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031987963891675025,
      "loss": 2.5945,
      "theoretical_loss": 3.584657194146771,
      "tokens_seen": 1210253312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031986960882647943,
      "loss": 2.5825,
      "theoretical_loss": 3.584639367651187,
      "tokens_seen": 1210318848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198595787362086,
      "loss": 2.6042,
      "theoretical_loss": 3.5846215423910976,
      "tokens_seen": 1210384384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198495486459378,
      "loss": 2.6627,
      "theoretical_loss": 3.5846037183663513,
      "tokens_seen": 1210449920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031983951855566703,
      "loss": 2.6505,
      "theoretical_loss": 3.5845858955767946,
      "tokens_seen": 1210515456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031982948846539616,
      "loss": 2.5319,
      "theoretical_loss": 3.5845680740222763,
      "tokens_seen": 1210580992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198194583751254,
      "loss": 2.6629,
      "theoretical_loss": 3.584550253702642,
      "tokens_seen": 1210646528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198094282848546,
      "loss": 2.6587,
      "theoretical_loss": 3.5845324346177416,
      "tokens_seen": 1210712064
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 670966,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8296167850494385,
      "objective/train/theoretical_loss": 3.584514616767421,
      "objective/train/tokens_used": 1231237600,
      "theoretical_loss": 3.584514616767421,
      "tokens_seen": 1210777600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031979939819458376,
      "loss": 2.6485,
      "theoretical_loss": 3.584514616767421,
      "tokens_seen": 1210777600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319789368104313,
      "loss": 2.6005,
      "theoretical_loss": 3.5844968001515287,
      "tokens_seen": 1210843136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197793380140421,
      "loss": 2.715,
      "theoretical_loss": 3.584478984769912,
      "tokens_seen": 1210908672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031976930792377135,
      "loss": 2.689,
      "theoretical_loss": 3.584461170622419,
      "tokens_seen": 1210974208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197592778335005,
      "loss": 2.8714,
      "theoretical_loss": 3.584443357708897,
      "tokens_seen": 1211039744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197492477432297,
      "loss": 2.2651,
      "theoretical_loss": 3.584425546029194,
      "tokens_seen": 1211105280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197392176529589,
      "loss": 2.6919,
      "theoretical_loss": 3.5844077355831576,
      "tokens_seen": 1211170816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197291875626881,
      "loss": 2.6406,
      "theoretical_loss": 3.584389926370635,
      "tokens_seen": 1211236352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031971915747241726,
      "loss": 2.7218,
      "theoretical_loss": 3.584372118391476,
      "tokens_seen": 1211301888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197091273821465,
      "loss": 2.6949,
      "theoretical_loss": 3.5843543116455265,
      "tokens_seen": 1211367424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196990972918756,
      "loss": 2.5578,
      "theoretical_loss": 3.584336506132635,
      "tokens_seen": 1211432960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031968906720160486,
      "loss": 2.6447,
      "theoretical_loss": 3.5843187018526494,
      "tokens_seen": 1211498496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319679037111334,
      "loss": 2.7112,
      "theoretical_loss": 3.5843008988054184,
      "tokens_seen": 1211564032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196690070210632,
      "loss": 2.6566,
      "theoretical_loss": 3.584283096990789,
      "tokens_seen": 1211629568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196589769307924,
      "loss": 2.6208,
      "theoretical_loss": 3.5842652964086095,
      "tokens_seen": 1211695104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196489468405216,
      "loss": 2.5844,
      "theoretical_loss": 3.584247497058728,
      "tokens_seen": 1211760640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031963891675025076,
      "loss": 2.6369,
      "theoretical_loss": 3.5842296989409923,
      "tokens_seen": 1211826176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031962888665997994,
      "loss": 2.7917,
      "theoretical_loss": 3.5842119020552508,
      "tokens_seen": 1211891712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196188565697091,
      "loss": 2.7858,
      "theoretical_loss": 3.584194106401351,
      "tokens_seen": 1211957248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031960882647943836,
      "loss": 2.6762,
      "theoretical_loss": 3.5841763119791423,
      "tokens_seen": 1212022784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195987963891675,
      "loss": 2.9897,
      "theoretical_loss": 3.5841585187884712,
      "tokens_seen": 1212088320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195887662988967,
      "loss": 2.5801,
      "theoretical_loss": 3.5841407268291876,
      "tokens_seen": 1212153856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031957873620862585,
      "loss": 2.7895,
      "theoretical_loss": 3.5841229361011386,
      "tokens_seen": 1212219392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195687061183551,
      "loss": 2.7946,
      "theoretical_loss": 3.5841051466041725,
      "tokens_seen": 1212284928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031955867602808427,
      "loss": 2.6184,
      "theoretical_loss": 3.5840873583381376,
      "tokens_seen": 1212350464
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 672128,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6712498664855957,
      "objective/train/theoretical_loss": 3.5840695713028827,
      "objective/train/tokens_used": 1232876000,
      "theoretical_loss": 3.5840695713028827,
      "tokens_seen": 1212416000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031954864593781345,
      "loss": 2.5049,
      "theoretical_loss": 3.5840695713028827,
      "tokens_seen": 1212416000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031953861584754263,
      "loss": 2.7097,
      "theoretical_loss": 3.5840517854982554,
      "tokens_seen": 1212481536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031952858575727186,
      "loss": 2.6616,
      "theoretical_loss": 3.584034000924105,
      "tokens_seen": 1212547072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319518555667001,
      "loss": 2.4007,
      "theoretical_loss": 3.584016217580279,
      "tokens_seen": 1212612608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195085255767302,
      "loss": 2.7693,
      "theoretical_loss": 3.583998435466626,
      "tokens_seen": 1212678144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031949849548645935,
      "loss": 2.7655,
      "theoretical_loss": 3.5839806545829944,
      "tokens_seen": 1212743680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194884653961886,
      "loss": 2.6329,
      "theoretical_loss": 3.583962874929233,
      "tokens_seen": 1212809216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031947843530591777,
      "loss": 2.5438,
      "theoretical_loss": 3.5839450965051904,
      "tokens_seen": 1212874752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031946840521564695,
      "loss": 2.6286,
      "theoretical_loss": 3.583927319310714,
      "tokens_seen": 1212940288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031945837512537613,
      "loss": 2.411,
      "theoretical_loss": 3.5839095433456536,
      "tokens_seen": 1213005824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194483450351053,
      "loss": 2.4915,
      "theoretical_loss": 3.583891768609857,
      "tokens_seen": 1213071360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194383149448345,
      "loss": 2.588,
      "theoretical_loss": 3.5838739951031737,
      "tokens_seen": 1213136896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031942828485456373,
      "loss": 2.5782,
      "theoretical_loss": 3.5838562228254514,
      "tokens_seen": 1213202432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031941825476429286,
      "loss": 2.6396,
      "theoretical_loss": 3.5838384517765385,
      "tokens_seen": 1213267968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194082246740221,
      "loss": 2.6197,
      "theoretical_loss": 3.5838206819562846,
      "tokens_seen": 1213333504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193981945837512,
      "loss": 2.762,
      "theoretical_loss": 3.5838029133645377,
      "tokens_seen": 1213399040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031938816449348045,
      "loss": 2.6373,
      "theoretical_loss": 3.583785146001147,
      "tokens_seen": 1213464576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031937813440320964,
      "loss": 2.6973,
      "theoretical_loss": 3.583767379865961,
      "tokens_seen": 1213530112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193681043129388,
      "loss": 2.5355,
      "theoretical_loss": 3.5837496149588284,
      "tokens_seen": 1213595648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319358074222668,
      "loss": 2.7368,
      "theoretical_loss": 3.583731851279598,
      "tokens_seen": 1213661184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031934804413239723,
      "loss": 2.8614,
      "theoretical_loss": 3.583714088828119,
      "tokens_seen": 1213726720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031933801404212636,
      "loss": 2.5393,
      "theoretical_loss": 3.5836963276042404,
      "tokens_seen": 1213792256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193279839518556,
      "loss": 2.7673,
      "theoretical_loss": 3.58367856760781,
      "tokens_seen": 1213857792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193179538615847,
      "loss": 2.7873,
      "theoretical_loss": 3.583660808838678,
      "tokens_seen": 1213923328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031930792377131396,
      "loss": 2.758,
      "theoretical_loss": 3.5836430512966926,
      "tokens_seen": 1213988864
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 672740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2003424167633057,
      "objective/train/theoretical_loss": 3.5836252949817027,
      "objective/train/tokens_used": 1234514400,
      "theoretical_loss": 3.5836252949817027,
      "tokens_seen": 1214054400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031929789368104314,
      "loss": 2.7173,
      "theoretical_loss": 3.5836252949817027,
      "tokens_seen": 1214054400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192878635907723,
      "loss": 2.7391,
      "theoretical_loss": 3.583607539893557,
      "tokens_seen": 1214119936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192778335005015,
      "loss": 2.6003,
      "theoretical_loss": 3.583589786032106,
      "tokens_seen": 1214185472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192678034102307,
      "loss": 2.8058,
      "theoretical_loss": 3.583572033397197,
      "tokens_seen": 1214251008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031925777331995986,
      "loss": 2.8194,
      "theoretical_loss": 3.5835542819886808,
      "tokens_seen": 1214316544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192477432296891,
      "loss": 2.5911,
      "theoretical_loss": 3.5835365318064047,
      "tokens_seen": 1214382080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192377131394182,
      "loss": 2.6463,
      "theoretical_loss": 3.5835187828502186,
      "tokens_seen": 1214447616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031922768304914746,
      "loss": 2.602,
      "theoretical_loss": 3.583501035119972,
      "tokens_seen": 1214513152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192176529588766,
      "loss": 2.4129,
      "theoretical_loss": 3.583483288615514,
      "tokens_seen": 1214578688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192076228686058,
      "loss": 2.7214,
      "theoretical_loss": 3.5834655433366933,
      "tokens_seen": 1214644224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319197592778335,
      "loss": 2.5718,
      "theoretical_loss": 3.583447799283359,
      "tokens_seen": 1214709760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191875626880642,
      "loss": 2.8878,
      "theoretical_loss": 3.583430056455361,
      "tokens_seen": 1214775296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031917753259779337,
      "loss": 2.3432,
      "theoretical_loss": 3.5834123148525485,
      "tokens_seen": 1214840832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191675025075226,
      "loss": 2.7502,
      "theoretical_loss": 3.583394574474771,
      "tokens_seen": 1214906368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031915747241725173,
      "loss": 2.6352,
      "theoretical_loss": 3.5833768353218773,
      "tokens_seen": 1214971904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031914744232698096,
      "loss": 2.4479,
      "theoretical_loss": 3.583359097393717,
      "tokens_seen": 1215037440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191374122367101,
      "loss": 2.6745,
      "theoretical_loss": 3.583341360690139,
      "tokens_seen": 1215102976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191273821464393,
      "loss": 2.7993,
      "theoretical_loss": 3.5833236252109937,
      "tokens_seen": 1215168512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191173520561685,
      "loss": 2.6827,
      "theoretical_loss": 3.5833058909561295,
      "tokens_seen": 1215234048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191073219658977,
      "loss": 2.8311,
      "theoretical_loss": 3.583288157925397,
      "tokens_seen": 1215299584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031909729187562687,
      "loss": 2.8424,
      "theoretical_loss": 3.583270426118645,
      "tokens_seen": 1215365120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031908726178535605,
      "loss": 2.927,
      "theoretical_loss": 3.583252695535723,
      "tokens_seen": 1215430656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031907723169508523,
      "loss": 2.7119,
      "theoretical_loss": 3.58323496617648,
      "tokens_seen": 1215496192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031906720160481447,
      "loss": 2.7552,
      "theoretical_loss": 3.583217238040767,
      "tokens_seen": 1215561728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031905717151454365,
      "loss": 2.6251,
      "theoretical_loss": 3.583199511128433,
      "tokens_seen": 1215627264
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 673834,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.846170425415039,
      "objective/train/theoretical_loss": 3.5831817854393266,
      "objective/train/tokens_used": 1236152800,
      "theoretical_loss": 3.5831817854393266,
      "tokens_seen": 1215692800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031904714142427283,
      "loss": 2.8292,
      "theoretical_loss": 3.5831817854393266,
      "tokens_seen": 1215692800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031903711133400206,
      "loss": 2.6614,
      "theoretical_loss": 3.5831640609732993,
      "tokens_seen": 1215758336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003190270812437312,
      "loss": 2.634,
      "theoretical_loss": 3.5831463377301995,
      "tokens_seen": 1215823872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003190170511534604,
      "loss": 2.6109,
      "theoretical_loss": 3.5831286157098767,
      "tokens_seen": 1215889408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031900702106318955,
      "loss": 2.6929,
      "theoretical_loss": 3.5831108949121817,
      "tokens_seen": 1215954944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189969909729188,
      "loss": 2.7014,
      "theoretical_loss": 3.5830931753369635,
      "tokens_seen": 1216020480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031898696088264797,
      "loss": 2.7554,
      "theoretical_loss": 3.5830754569840724,
      "tokens_seen": 1216086016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031897693079237715,
      "loss": 2.3739,
      "theoretical_loss": 3.5830577398533574,
      "tokens_seen": 1216151552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031896690070210633,
      "loss": 2.8485,
      "theoretical_loss": 3.5830400239446694,
      "tokens_seen": 1216217088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189568706118355,
      "loss": 2.9264,
      "theoretical_loss": 3.5830223092578573,
      "tokens_seen": 1216282624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189468405215647,
      "loss": 3.0185,
      "theoretical_loss": 3.583004595792772,
      "tokens_seen": 1216348160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031893681043129393,
      "loss": 2.384,
      "theoretical_loss": 3.582986883549262,
      "tokens_seen": 1216413696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031892678034102306,
      "loss": 2.6403,
      "theoretical_loss": 3.5829691725271786,
      "tokens_seen": 1216479232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189167502507523,
      "loss": 2.6133,
      "theoretical_loss": 3.5829514627263714,
      "tokens_seen": 1216544768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189067201604814,
      "loss": 2.5405,
      "theoretical_loss": 3.5829337541466906,
      "tokens_seen": 1216610304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031889669007021065,
      "loss": 2.6433,
      "theoretical_loss": 3.5829160467879855,
      "tokens_seen": 1216675840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031888665997993984,
      "loss": 2.6715,
      "theoretical_loss": 3.5828983406501065,
      "tokens_seen": 1216741376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318876629889669,
      "loss": 2.6068,
      "theoretical_loss": 3.582880635732904,
      "tokens_seen": 1216806912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188665997993982,
      "loss": 2.4911,
      "theoretical_loss": 3.5828629320362273,
      "tokens_seen": 1216872448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031885656970912743,
      "loss": 2.6389,
      "theoretical_loss": 3.5828452295599273,
      "tokens_seen": 1216937984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031884653961885656,
      "loss": 2.7881,
      "theoretical_loss": 3.5828275283038544,
      "tokens_seen": 1217003520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188365095285858,
      "loss": 2.6524,
      "theoretical_loss": 3.582809828267858,
      "tokens_seen": 1217069056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188264794383149,
      "loss": 2.6268,
      "theoretical_loss": 3.582792129451789,
      "tokens_seen": 1217134592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031881644934804416,
      "loss": 2.6762,
      "theoretical_loss": 3.582774431855497,
      "tokens_seen": 1217200128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031880641925777334,
      "loss": 2.2686,
      "theoretical_loss": 3.5827567354788323,
      "tokens_seen": 1217265664
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 674331,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5128471851348877,
      "objective/train/theoretical_loss": 3.5827390403216457,
      "objective/train/tokens_used": 1237791200,
      "theoretical_loss": 3.5827390403216457,
      "tokens_seen": 1217331200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187963891675025,
      "loss": 2.7099,
      "theoretical_loss": 3.5827390403216457,
      "tokens_seen": 1217331200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187863590772317,
      "loss": 2.8773,
      "theoretical_loss": 3.5827213463837877,
      "tokens_seen": 1217396736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187763289869609,
      "loss": 2.7363,
      "theoretical_loss": 3.5827036536651073,
      "tokens_seen": 1217462272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031876629889669006,
      "loss": 2.6986,
      "theoretical_loss": 3.5826859621654568,
      "tokens_seen": 1217527808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187562688064193,
      "loss": 2.4538,
      "theoretical_loss": 3.582668271884685,
      "tokens_seen": 1217593344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187462387161484,
      "loss": 2.6988,
      "theoretical_loss": 3.582650582822643,
      "tokens_seen": 1217658880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031873620862587766,
      "loss": 2.6411,
      "theoretical_loss": 3.5826328949791812,
      "tokens_seen": 1217724416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187261785356068,
      "loss": 2.4917,
      "theoretical_loss": 3.58261520835415,
      "tokens_seen": 1217789952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318716148445336,
      "loss": 2.7464,
      "theoretical_loss": 3.5825975229474,
      "tokens_seen": 1217855488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187061183550652,
      "loss": 2.669,
      "theoretical_loss": 3.582579838758782,
      "tokens_seen": 1217921024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186960882647944,
      "loss": 2.4795,
      "theoretical_loss": 3.5825621557881457,
      "tokens_seen": 1217986560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031868605817452357,
      "loss": 2.6991,
      "theoretical_loss": 3.5825444740353425,
      "tokens_seen": 1218052096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186760280842528,
      "loss": 2.6764,
      "theoretical_loss": 3.5825267935002225,
      "tokens_seen": 1218117632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031866599799398193,
      "loss": 2.5941,
      "theoretical_loss": 3.5825091141826366,
      "tokens_seen": 1218183168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031865596790371116,
      "loss": 2.6048,
      "theoretical_loss": 3.582491436082436,
      "tokens_seen": 1218248704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186459378134403,
      "loss": 2.5775,
      "theoretical_loss": 3.58247375919947,
      "tokens_seen": 1218314240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186359077231695,
      "loss": 2.7505,
      "theoretical_loss": 3.582456083533591,
      "tokens_seen": 1218379776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186258776328987,
      "loss": 2.2671,
      "theoretical_loss": 3.582438409084648,
      "tokens_seen": 1218445312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186158475426279,
      "loss": 2.9417,
      "theoretical_loss": 3.5824207358524935,
      "tokens_seen": 1218510848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031860581745235707,
      "loss": 2.6654,
      "theoretical_loss": 3.582403063836977,
      "tokens_seen": 1218576384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031859578736208625,
      "loss": 2.6196,
      "theoretical_loss": 3.5823853930379497,
      "tokens_seen": 1218641920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031858575727181543,
      "loss": 2.7017,
      "theoretical_loss": 3.5823677234552624,
      "tokens_seen": 1218707456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031857572718154467,
      "loss": 2.6233,
      "theoretical_loss": 3.5823500550887664,
      "tokens_seen": 1218772992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185656970912738,
      "loss": 2.3861,
      "theoretical_loss": 3.5823323879383118,
      "tokens_seen": 1218838528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031855566700100303,
      "loss": 2.766,
      "theoretical_loss": 3.5823147220037503,
      "tokens_seen": 1218904064
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 675627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.326138973236084,
      "objective/train/theoretical_loss": 3.582297057284933,
      "objective/train/tokens_used": 1239429600,
      "theoretical_loss": 3.582297057284933,
      "tokens_seen": 1218969600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185456369107322,
      "loss": 2.6397,
      "theoretical_loss": 3.582297057284933,
      "tokens_seen": 1218969600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185356068204614,
      "loss": 2.5176,
      "theoretical_loss": 3.58227939378171,
      "tokens_seen": 1219035136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031852557673019057,
      "loss": 2.5311,
      "theoretical_loss": 3.5822617314939325,
      "tokens_seen": 1219100672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031851554663991975,
      "loss": 2.6315,
      "theoretical_loss": 3.582244070421452,
      "tokens_seen": 1219166208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031850551654964893,
      "loss": 2.5092,
      "theoretical_loss": 3.5822264105641195,
      "tokens_seen": 1219231744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031849548645937817,
      "loss": 2.5286,
      "theoretical_loss": 3.582208751921786,
      "tokens_seen": 1219297280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184854563691073,
      "loss": 2.8103,
      "theoretical_loss": 3.582191094494302,
      "tokens_seen": 1219362816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031847542627883653,
      "loss": 2.4883,
      "theoretical_loss": 3.5821734382815196,
      "tokens_seen": 1219428352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031846539618856566,
      "loss": 2.7542,
      "theoretical_loss": 3.582155783283289,
      "tokens_seen": 1219493888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184553660982949,
      "loss": 2.481,
      "theoretical_loss": 3.5821381294994628,
      "tokens_seen": 1219559424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184453360080241,
      "loss": 2.7431,
      "theoretical_loss": 3.582120476929891,
      "tokens_seen": 1219624960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031843530591775326,
      "loss": 2.4548,
      "theoretical_loss": 3.5821028255744247,
      "tokens_seen": 1219690496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031842527582748244,
      "loss": 2.6645,
      "theoretical_loss": 3.582085175432916,
      "tokens_seen": 1219756032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184152457372116,
      "loss": 2.363,
      "theoretical_loss": 3.5820675265052158,
      "tokens_seen": 1219821568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184052156469408,
      "loss": 2.6981,
      "theoretical_loss": 3.5820498787911754,
      "tokens_seen": 1219887104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031839518555667004,
      "loss": 2.688,
      "theoretical_loss": 3.582032232290646,
      "tokens_seen": 1219952640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031838515546639916,
      "loss": 2.6851,
      "theoretical_loss": 3.5820145870034796,
      "tokens_seen": 1220018176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183751253761284,
      "loss": 2.6927,
      "theoretical_loss": 3.5819969429295266,
      "tokens_seen": 1220083712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183650952858576,
      "loss": 2.7519,
      "theoretical_loss": 3.5819793000686397,
      "tokens_seen": 1220149248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031835506519558676,
      "loss": 2.5282,
      "theoretical_loss": 3.581961658420669,
      "tokens_seen": 1220214784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031834503510531594,
      "loss": 2.6981,
      "theoretical_loss": 3.5819440179854674,
      "tokens_seen": 1220280320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183350050150451,
      "loss": 2.4955,
      "theoretical_loss": 3.5819263787628848,
      "tokens_seen": 1220345856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183249749247743,
      "loss": 2.6698,
      "theoretical_loss": 3.5819087407527737,
      "tokens_seen": 1220411392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031831494483450354,
      "loss": 2.8391,
      "theoretical_loss": 3.5818911039549857,
      "tokens_seen": 1220476928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183049147442327,
      "loss": 2.6485,
      "theoretical_loss": 3.581873468369372,
      "tokens_seen": 1220542464
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 676618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6377201080322266,
      "objective/train/theoretical_loss": 3.5818558339957844,
      "objective/train/tokens_used": 1241068000,
      "theoretical_loss": 3.5818558339957844,
      "tokens_seen": 1220608000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182948846539619,
      "loss": 2.5538,
      "theoretical_loss": 3.5818558339957844,
      "tokens_seen": 1220608000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182848545636911,
      "loss": 2.7519,
      "theoretical_loss": 3.581838200834075,
      "tokens_seen": 1220673536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031827482447342026,
      "loss": 2.6153,
      "theoretical_loss": 3.581820568884094,
      "tokens_seen": 1220739072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182647943831495,
      "loss": 2.7711,
      "theoretical_loss": 3.581802938145695,
      "tokens_seen": 1220804608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182547642928786,
      "loss": 2.7786,
      "theoretical_loss": 3.581785308618729,
      "tokens_seen": 1220870144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031824473420260786,
      "loss": 2.5908,
      "theoretical_loss": 3.5817676803030465,
      "tokens_seen": 1220935680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318234704112337,
      "loss": 2.6733,
      "theoretical_loss": 3.5817500531985007,
      "tokens_seen": 1221001216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182246740220662,
      "loss": 2.5456,
      "theoretical_loss": 3.581732427304943,
      "tokens_seen": 1221066752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182146439317954,
      "loss": 2.5502,
      "theoretical_loss": 3.5817148026222254,
      "tokens_seen": 1221132288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182046138415246,
      "loss": 2.8477,
      "theoretical_loss": 3.5816971791501993,
      "tokens_seen": 1221197824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031819458375125377,
      "loss": 2.6389,
      "theoretical_loss": 3.5816795568887168,
      "tokens_seen": 1221263360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318184553660983,
      "loss": 2.6474,
      "theoretical_loss": 3.58166193583763,
      "tokens_seen": 1221328896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031817452357071213,
      "loss": 2.6726,
      "theoretical_loss": 3.5816443159967903,
      "tokens_seen": 1221394432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031816449348044136,
      "loss": 2.7162,
      "theoretical_loss": 3.58162669736605,
      "tokens_seen": 1221459968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181544633901705,
      "loss": 2.629,
      "theoretical_loss": 3.5816090799452613,
      "tokens_seen": 1221525504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181444332998997,
      "loss": 2.5372,
      "theoretical_loss": 3.581591463734276,
      "tokens_seen": 1221591040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181344032096289,
      "loss": 2.5658,
      "theoretical_loss": 3.581573848732946,
      "tokens_seen": 1221656576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181243731193581,
      "loss": 2.5672,
      "theoretical_loss": 3.581556234941123,
      "tokens_seen": 1221722112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031811434302908727,
      "loss": 2.8258,
      "theoretical_loss": 3.58153862235866,
      "tokens_seen": 1221787648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031810431293881645,
      "loss": 2.8054,
      "theoretical_loss": 3.581521010985408,
      "tokens_seen": 1221853184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031809428284854563,
      "loss": 2.7113,
      "theoretical_loss": 3.5815034008212203,
      "tokens_seen": 1221918720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031808425275827487,
      "loss": 2.6263,
      "theoretical_loss": 3.5814857918659477,
      "tokens_seen": 1221984256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318074222668004,
      "loss": 2.9308,
      "theoretical_loss": 3.5814681841194442,
      "tokens_seen": 1222049792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031806419257773323,
      "loss": 2.6379,
      "theoretical_loss": 3.58145057758156,
      "tokens_seen": 1222115328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003180541624874624,
      "loss": 2.5758,
      "theoretical_loss": 3.581432972252149,
      "tokens_seen": 1222180864
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 677215,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3762056827545166,
      "objective/train/theoretical_loss": 3.5814153681310623,
      "objective/train/tokens_used": 1242706400,
      "theoretical_loss": 3.5814153681310623,
      "tokens_seen": 1222246400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003180441323971916,
      "loss": 2.7435,
      "theoretical_loss": 3.5814153681310623,
      "tokens_seen": 1222246400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031803410230692077,
      "loss": 2.716,
      "theoretical_loss": 3.581397765218153,
      "tokens_seen": 1222311936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031802407221664995,
      "loss": 2.6065,
      "theoretical_loss": 3.581380163513273,
      "tokens_seen": 1222377472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031801404212637913,
      "loss": 2.7569,
      "theoretical_loss": 3.5813625630162744,
      "tokens_seen": 1222443008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031800401203610837,
      "loss": 2.7203,
      "theoretical_loss": 3.58134496372701,
      "tokens_seen": 1222508544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179939819458375,
      "loss": 2.6444,
      "theoretical_loss": 3.581327365645332,
      "tokens_seen": 1222574080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031798395185556673,
      "loss": 2.5914,
      "theoretical_loss": 3.581309768771093,
      "tokens_seen": 1222639616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031797392176529586,
      "loss": 2.6799,
      "theoretical_loss": 3.581292173104145,
      "tokens_seen": 1222705152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179638916750251,
      "loss": 2.8491,
      "theoretical_loss": 3.581274578644341,
      "tokens_seen": 1222770688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179538615847543,
      "loss": 2.4735,
      "theoretical_loss": 3.5812569853915335,
      "tokens_seen": 1222836224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031794383149448346,
      "loss": 2.7763,
      "theoretical_loss": 3.5812393933455744,
      "tokens_seen": 1222901760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031793380140421264,
      "loss": 2.55,
      "theoretical_loss": 3.5812218025063167,
      "tokens_seen": 1222967296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003179237713139418,
      "loss": 2.4893,
      "theoretical_loss": 3.581204212873613,
      "tokens_seen": 1223032832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317913741223671,
      "loss": 2.7899,
      "theoretical_loss": 3.581186624447316,
      "tokens_seen": 1223098368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031790371113340024,
      "loss": 2.4937,
      "theoretical_loss": 3.5811690372272773,
      "tokens_seen": 1223163904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031789368104312936,
      "loss": 2.8664,
      "theoretical_loss": 3.581151451213351,
      "tokens_seen": 1223229440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178836509528586,
      "loss": 2.7589,
      "theoretical_loss": 3.5811338664053887,
      "tokens_seen": 1223294976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178736208625878,
      "loss": 2.4983,
      "theoretical_loss": 3.581116282803244,
      "tokens_seen": 1223360512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031786359077231696,
      "loss": 2.7072,
      "theoretical_loss": 3.581098700406769,
      "tokens_seen": 1223426048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031785356068204614,
      "loss": 2.6054,
      "theoretical_loss": 3.581081119215816,
      "tokens_seen": 1223491584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178435305917753,
      "loss": 2.4573,
      "theoretical_loss": 3.5810635392302395,
      "tokens_seen": 1223557120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178335005015045,
      "loss": 2.5697,
      "theoretical_loss": 3.5810459604498908,
      "tokens_seen": 1223622656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031782347041123374,
      "loss": 2.4891,
      "theoretical_loss": 3.5810283828746226,
      "tokens_seen": 1223688192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031781344032096287,
      "loss": 2.8321,
      "theoretical_loss": 3.5810108065042887,
      "tokens_seen": 1223753728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178034102306921,
      "loss": 2.7512,
      "theoretical_loss": 3.5809932313387414,
      "tokens_seen": 1223819264
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 678133,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7944464683532715,
      "objective/train/theoretical_loss": 3.5809756573778335,
      "objective/train/tokens_used": 1244344800,
      "theoretical_loss": 3.5809756573778335,
      "tokens_seen": 1223884800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031779338014042123,
      "loss": 2.6868,
      "theoretical_loss": 3.5809756573778335,
      "tokens_seen": 1223884800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031778335005015046,
      "loss": 2.837,
      "theoretical_loss": 3.5809580846214186,
      "tokens_seen": 1223950336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031777331995987964,
      "loss": 2.7034,
      "theoretical_loss": 3.580940513069349,
      "tokens_seen": 1224015872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177632898696088,
      "loss": 2.7941,
      "theoretical_loss": 3.580922942721478,
      "tokens_seen": 1224081408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317753259779338,
      "loss": 2.7049,
      "theoretical_loss": 3.580905373577659,
      "tokens_seen": 1224146944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177432296890672,
      "loss": 2.7759,
      "theoretical_loss": 3.580887805637744,
      "tokens_seen": 1224212480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031773319959879637,
      "loss": 2.6332,
      "theoretical_loss": 3.5808702389015865,
      "tokens_seen": 1224278016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177231695085256,
      "loss": 2.7317,
      "theoretical_loss": 3.5808526733690402,
      "tokens_seen": 1224343552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031771313941825473,
      "loss": 2.6885,
      "theoretical_loss": 3.580835109039958,
      "tokens_seen": 1224409088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031770310932798397,
      "loss": 2.6962,
      "theoretical_loss": 3.5808175459141918,
      "tokens_seen": 1224474624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031769307923771315,
      "loss": 2.6085,
      "theoretical_loss": 3.5807999839915965,
      "tokens_seen": 1224540160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031768304914744233,
      "loss": 2.5963,
      "theoretical_loss": 3.5807824232720247,
      "tokens_seen": 1224605696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176730190571715,
      "loss": 2.7513,
      "theoretical_loss": 3.580764863755329,
      "tokens_seen": 1224671232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176629889669007,
      "loss": 2.6441,
      "theoretical_loss": 3.580747305441363,
      "tokens_seen": 1224736768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031765295887662987,
      "loss": 2.3998,
      "theoretical_loss": 3.58072974832998,
      "tokens_seen": 1224802304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176429287863591,
      "loss": 2.7632,
      "theoretical_loss": 3.580712192421034,
      "tokens_seen": 1224867840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031763289869608823,
      "loss": 2.6429,
      "theoretical_loss": 3.580694637714377,
      "tokens_seen": 1224933376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031762286860581747,
      "loss": 2.5251,
      "theoretical_loss": 3.5806770842098636,
      "tokens_seen": 1224998912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176128385155466,
      "loss": 2.5689,
      "theoretical_loss": 3.580659531907346,
      "tokens_seen": 1225064448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031760280842527583,
      "loss": 2.8738,
      "theoretical_loss": 3.5806419808066785,
      "tokens_seen": 1225129984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317592778335005,
      "loss": 2.7404,
      "theoretical_loss": 3.580624430907714,
      "tokens_seen": 1225195520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175827482447342,
      "loss": 2.5556,
      "theoretical_loss": 3.5806068822103065,
      "tokens_seen": 1225261056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175727181544634,
      "loss": 2.8817,
      "theoretical_loss": 3.580589334714309,
      "tokens_seen": 1225326592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175626880641926,
      "loss": 2.719,
      "theoretical_loss": 3.580571788419575,
      "tokens_seen": 1225392128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175526579739218,
      "loss": 2.5082,
      "theoretical_loss": 3.5805542433259583,
      "tokens_seen": 1225457664
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 678803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4843032360076904,
      "objective/train/theoretical_loss": 3.5805366994333125,
      "objective/train/tokens_used": 1245983200,
      "theoretical_loss": 3.5805366994333125,
      "tokens_seen": 1225523200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031754262788365097,
      "loss": 2.6552,
      "theoretical_loss": 3.5805366994333125,
      "tokens_seen": 1225523200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031753259779338015,
      "loss": 2.7421,
      "theoretical_loss": 3.5805191567414907,
      "tokens_seen": 1225588736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031752256770310933,
      "loss": 2.7794,
      "theoretical_loss": 3.5805016152503466,
      "tokens_seen": 1225654272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031751253761283857,
      "loss": 2.7931,
      "theoretical_loss": 3.580484074959734,
      "tokens_seen": 1225719808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175025075225677,
      "loss": 2.6884,
      "theoretical_loss": 3.5804665358695074,
      "tokens_seen": 1225785344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031749247743229693,
      "loss": 2.5482,
      "theoretical_loss": 3.5804489979795187,
      "tokens_seen": 1225850880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031748244734202606,
      "loss": 2.7467,
      "theoretical_loss": 3.580431461289623,
      "tokens_seen": 1225916416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174724172517553,
      "loss": 2.7609,
      "theoretical_loss": 3.5804139257996734,
      "tokens_seen": 1225981952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174623871614845,
      "loss": 2.5784,
      "theoretical_loss": 3.580396391509524,
      "tokens_seen": 1226047488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031745235707121366,
      "loss": 2.6765,
      "theoretical_loss": 3.580378858419028,
      "tokens_seen": 1226113024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031744232698094284,
      "loss": 2.7612,
      "theoretical_loss": 3.58036132652804,
      "tokens_seen": 1226178560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317432296890672,
      "loss": 2.5738,
      "theoretical_loss": 3.5803437958364137,
      "tokens_seen": 1226244096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174222668004012,
      "loss": 2.6951,
      "theoretical_loss": 3.5803262663440023,
      "tokens_seen": 1226309632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031741223671013044,
      "loss": 2.6944,
      "theoretical_loss": 3.58030873805066,
      "tokens_seen": 1226375168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031740220661985956,
      "loss": 2.7802,
      "theoretical_loss": 3.580291210956241,
      "tokens_seen": 1226440704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173921765295888,
      "loss": 2.5398,
      "theoretical_loss": 3.5802736850605994,
      "tokens_seen": 1226506240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317382146439318,
      "loss": 2.6482,
      "theoretical_loss": 3.580256160363588,
      "tokens_seen": 1226571776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031737211634904716,
      "loss": 2.7186,
      "theoretical_loss": 3.5802386368650625,
      "tokens_seen": 1226637312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031736208625877634,
      "loss": 2.534,
      "theoretical_loss": 3.5802211145648757,
      "tokens_seen": 1226702848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173520561685055,
      "loss": 2.5662,
      "theoretical_loss": 3.5802035934628815,
      "tokens_seen": 1226768384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173420260782347,
      "loss": 2.7673,
      "theoretical_loss": 3.5801860735589344,
      "tokens_seen": 1226833920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031733199598796394,
      "loss": 2.5737,
      "theoretical_loss": 3.5801685548528894,
      "tokens_seen": 1226899456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031732196589769307,
      "loss": 2.7677,
      "theoretical_loss": 3.5801510373445993,
      "tokens_seen": 1226964992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173119358074223,
      "loss": 2.4772,
      "theoretical_loss": 3.5801335210339182,
      "tokens_seen": 1227030528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031730190571715143,
      "loss": 2.6551,
      "theoretical_loss": 3.5801160059207007,
      "tokens_seen": 1227096064
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 679787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1799423694610596,
      "objective/train/theoretical_loss": 3.580098492004802,
      "objective/train/tokens_used": 1247621600,
      "theoretical_loss": 3.580098492004802,
      "tokens_seen": 1227161600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031729187562688066,
      "loss": 2.7681,
      "theoretical_loss": 3.580098492004802,
      "tokens_seen": 1227161600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031728184553660984,
      "loss": 2.6133,
      "theoretical_loss": 3.580080979286074,
      "tokens_seen": 1227227136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317271815446339,
      "loss": 2.7949,
      "theoretical_loss": 3.580063467764373,
      "tokens_seen": 1227292672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172617853560682,
      "loss": 2.4662,
      "theoretical_loss": 3.5800459574395527,
      "tokens_seen": 1227358208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172517552657974,
      "loss": 2.7973,
      "theoretical_loss": 3.5800284483114666,
      "tokens_seen": 1227423744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031724172517552657,
      "loss": 2.4858,
      "theoretical_loss": 3.58001094037997,
      "tokens_seen": 1227489280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172316950852558,
      "loss": 2.7632,
      "theoretical_loss": 3.5799934336449173,
      "tokens_seen": 1227554816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031722166499498493,
      "loss": 2.4781,
      "theoretical_loss": 3.579975928106162,
      "tokens_seen": 1227620352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031721163490471417,
      "loss": 2.825,
      "theoretical_loss": 3.579958423763559,
      "tokens_seen": 1227685888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031720160481444335,
      "loss": 2.662,
      "theoretical_loss": 3.579940920616963,
      "tokens_seen": 1227751424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031719157472417253,
      "loss": 2.6949,
      "theoretical_loss": 3.579923418666228,
      "tokens_seen": 1227816960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171815446339017,
      "loss": 2.7449,
      "theoretical_loss": 3.5799059179112076,
      "tokens_seen": 1227882496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171715145436309,
      "loss": 2.7963,
      "theoretical_loss": 3.5798884183517585,
      "tokens_seen": 1227948032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031716148445336007,
      "loss": 2.7011,
      "theoretical_loss": 3.5798709199877337,
      "tokens_seen": 1228013568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171514543630893,
      "loss": 2.8564,
      "theoretical_loss": 3.5798534228189878,
      "tokens_seen": 1228079104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031714142427281843,
      "loss": 2.6509,
      "theoretical_loss": 3.5798359268453757,
      "tokens_seen": 1228144640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031713139418254767,
      "loss": 2.5021,
      "theoretical_loss": 3.5798184320667525,
      "tokens_seen": 1228210176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171213640922768,
      "loss": 2.6051,
      "theoretical_loss": 3.579800938482972,
      "tokens_seen": 1228275712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031711133400200603,
      "loss": 2.8244,
      "theoretical_loss": 3.579783446093889,
      "tokens_seen": 1228341248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171013039117352,
      "loss": 2.8286,
      "theoretical_loss": 3.579765954899358,
      "tokens_seen": 1228406784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170912738214644,
      "loss": 2.8978,
      "theoretical_loss": 3.579748464899234,
      "tokens_seen": 1228472320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170812437311936,
      "loss": 2.6003,
      "theoretical_loss": 3.579730976093372,
      "tokens_seen": 1228537856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170712136409228,
      "loss": 2.6917,
      "theoretical_loss": 3.5797134884816266,
      "tokens_seen": 1228603392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031706118355065194,
      "loss": 2.6362,
      "theoretical_loss": 3.5796960020638524,
      "tokens_seen": 1228668928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170511534603812,
      "loss": 2.604,
      "theoretical_loss": 3.5796785168399037,
      "tokens_seen": 1228734464
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 680312,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.374969482421875,
      "objective/train/theoretical_loss": 3.5796610328096365,
      "objective/train/tokens_used": 1249260000,
      "theoretical_loss": 3.5796610328096365,
      "tokens_seen": 1228800000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170411233701103,
      "loss": 2.6887,
      "theoretical_loss": 3.5796610328096365,
      "tokens_seen": 1228800000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031703109327983954,
      "loss": 2.7248,
      "theoretical_loss": 3.5796435499729045,
      "tokens_seen": 1228865536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170210631895687,
      "loss": 2.8177,
      "theoretical_loss": 3.5796260683295635,
      "tokens_seen": 1228931072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170110330992979,
      "loss": 2.6824,
      "theoretical_loss": 3.579608587879468,
      "tokens_seen": 1228996608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170010030090271,
      "loss": 2.9693,
      "theoretical_loss": 3.5795911086224725,
      "tokens_seen": 1229062144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031699097291875626,
      "loss": 2.7433,
      "theoretical_loss": 3.579573630558433,
      "tokens_seen": 1229127680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031698094282848544,
      "loss": 2.5788,
      "theoretical_loss": 3.5795561536872036,
      "tokens_seen": 1229193216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169709127382147,
      "loss": 2.7061,
      "theoretical_loss": 3.57953867800864,
      "tokens_seen": 1229258752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169608826479438,
      "loss": 2.6546,
      "theoretical_loss": 3.5795212035225963,
      "tokens_seen": 1229324288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031695085255767304,
      "loss": 2.3916,
      "theoretical_loss": 3.579503730228928,
      "tokens_seen": 1229389824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031694082246740217,
      "loss": 2.5509,
      "theoretical_loss": 3.57948625812749,
      "tokens_seen": 1229455360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169307923771314,
      "loss": 2.6121,
      "theoretical_loss": 3.5794687872181385,
      "tokens_seen": 1229520896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169207622868606,
      "loss": 2.3906,
      "theoretical_loss": 3.5794513175007276,
      "tokens_seen": 1229586432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031691073219658976,
      "loss": 2.6725,
      "theoretical_loss": 3.5794338489751123,
      "tokens_seen": 1229651968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031690070210631894,
      "loss": 2.7202,
      "theoretical_loss": 3.579416381641148,
      "tokens_seen": 1229717504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168906720160482,
      "loss": 2.8365,
      "theoretical_loss": 3.5793989154986905,
      "tokens_seen": 1229783040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168806419257773,
      "loss": 2.6397,
      "theoretical_loss": 3.5793814505475945,
      "tokens_seen": 1229848576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031687061183550654,
      "loss": 2.7826,
      "theoretical_loss": 3.5793639867877154,
      "tokens_seen": 1229914112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031686058174523567,
      "loss": 2.472,
      "theoretical_loss": 3.5793465242189075,
      "tokens_seen": 1229979648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168505516549649,
      "loss": 2.6565,
      "theoretical_loss": 3.579329062841028,
      "tokens_seen": 1230045184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168405215646941,
      "loss": 2.6165,
      "theoretical_loss": 3.579311602653931,
      "tokens_seen": 1230110720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031683049147442327,
      "loss": 2.807,
      "theoretical_loss": 3.5792941436574726,
      "tokens_seen": 1230176256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031682046138415245,
      "loss": 2.6677,
      "theoretical_loss": 3.579276685851507,
      "tokens_seen": 1230241792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031681043129388163,
      "loss": 2.5583,
      "theoretical_loss": 3.5792592292358902,
      "tokens_seen": 1230307328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031680040120361086,
      "loss": 2.8452,
      "theoretical_loss": 3.5792417738104785,
      "tokens_seen": 1230372864
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 681644,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6463749408721924,
      "objective/train/theoretical_loss": 3.579224319575126,
      "objective/train/tokens_used": 1250898400,
      "theoretical_loss": 3.579224319575126,
      "tokens_seen": 1230438400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031679037111334004,
      "loss": 2.5301,
      "theoretical_loss": 3.579224319575126,
      "tokens_seen": 1230438400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167803410230692,
      "loss": 2.6175,
      "theoretical_loss": 3.579206866529689,
      "tokens_seen": 1230503936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167703109327984,
      "loss": 2.7769,
      "theoretical_loss": 3.5791894146740226,
      "tokens_seen": 1230569472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167602808425276,
      "loss": 2.7163,
      "theoretical_loss": 3.579171964007983,
      "tokens_seen": 1230635008
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031675025075225677,
      "loss": 2.658,
      "theoretical_loss": 3.579154514531425,
      "tokens_seen": 1230700544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316740220661986,
      "loss": 2.7854,
      "theoretical_loss": 3.579137066244204,
      "tokens_seen": 1230766080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031673019057171513,
      "loss": 2.747,
      "theoretical_loss": 3.5791196191461774,
      "tokens_seen": 1230831616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031672016048144437,
      "loss": 2.5688,
      "theoretical_loss": 3.5791021732371986,
      "tokens_seen": 1230897152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031671013039117355,
      "loss": 2.8399,
      "theoretical_loss": 3.5790847285171243,
      "tokens_seen": 1230962688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031670010030090273,
      "loss": 2.5067,
      "theoretical_loss": 3.5790672849858103,
      "tokens_seen": 1231028224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166900702106319,
      "loss": 2.7153,
      "theoretical_loss": 3.579049842643112,
      "tokens_seen": 1231093760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166800401203611,
      "loss": 2.6044,
      "theoretical_loss": 3.5790324014888855,
      "tokens_seen": 1231159296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031667001003009027,
      "loss": 2.719,
      "theoretical_loss": 3.5790149615229856,
      "tokens_seen": 1231224832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166599799398195,
      "loss": 2.5149,
      "theoretical_loss": 3.5789975227452695,
      "tokens_seen": 1231290368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031664994984954863,
      "loss": 2.7169,
      "theoretical_loss": 3.5789800851555924,
      "tokens_seen": 1231355904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031663991975927787,
      "loss": 2.51,
      "theoretical_loss": 3.5789626487538095,
      "tokens_seen": 1231421440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316629889669007,
      "loss": 2.6168,
      "theoretical_loss": 3.5789452135397775,
      "tokens_seen": 1231486976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031661985957873623,
      "loss": 2.601,
      "theoretical_loss": 3.578927779513352,
      "tokens_seen": 1231552512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166098294884654,
      "loss": 2.7909,
      "theoretical_loss": 3.578910346674389,
      "tokens_seen": 1231618048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165997993981946,
      "loss": 2.6013,
      "theoretical_loss": 3.578892915022744,
      "tokens_seen": 1231683584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165897693079238,
      "loss": 2.7524,
      "theoretical_loss": 3.5788754845582735,
      "tokens_seen": 1231749120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316579739217653,
      "loss": 2.8369,
      "theoretical_loss": 3.5788580552808336,
      "tokens_seen": 1231814656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031656970912738214,
      "loss": 2.6386,
      "theoretical_loss": 3.5788406271902797,
      "tokens_seen": 1231880192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165596790371114,
      "loss": 2.6326,
      "theoretical_loss": 3.578823200286468,
      "tokens_seen": 1231945728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165496489468405,
      "loss": 2.5907,
      "theoretical_loss": 3.5788057745692554,
      "tokens_seen": 1232011264
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 682283,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9270589351654053,
      "objective/train/theoretical_loss": 3.578788350038497,
      "objective/train/tokens_used": 1252536800,
      "theoretical_loss": 3.578788350038497,
      "tokens_seen": 1232076800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031653961885656974,
      "loss": 2.7132,
      "theoretical_loss": 3.578788350038497,
      "tokens_seen": 1232076800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165295887662989,
      "loss": 2.8002,
      "theoretical_loss": 3.578770926694049,
      "tokens_seen": 1232142336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165195586760281,
      "loss": 2.4558,
      "theoretical_loss": 3.5787535045357686,
      "tokens_seen": 1232207872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165095285857573,
      "loss": 2.5559,
      "theoretical_loss": 3.57873608356351,
      "tokens_seen": 1232273408
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031649949849548646,
      "loss": 2.4907,
      "theoretical_loss": 3.5787186637771313,
      "tokens_seen": 1232338944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031648946840521564,
      "loss": 2.6496,
      "theoretical_loss": 3.5787012451764877,
      "tokens_seen": 1232404480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164794383149449,
      "loss": 2.7241,
      "theoretical_loss": 3.578683827761436,
      "tokens_seen": 1232470016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316469408224674,
      "loss": 2.7307,
      "theoretical_loss": 3.578666411531832,
      "tokens_seen": 1232535552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031645937813440324,
      "loss": 2.5383,
      "theoretical_loss": 3.578648996487532,
      "tokens_seen": 1232601088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031644934804413237,
      "loss": 2.729,
      "theoretical_loss": 3.5786315826283923,
      "tokens_seen": 1232666624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164393179538616,
      "loss": 2.7778,
      "theoretical_loss": 3.57861416995427,
      "tokens_seen": 1232732160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164292878635908,
      "loss": 2.5696,
      "theoretical_loss": 3.5785967584650207,
      "tokens_seen": 1232797696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031641925777331996,
      "loss": 2.4963,
      "theoretical_loss": 3.578579348160501,
      "tokens_seen": 1232863232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031640922768304914,
      "loss": 2.4366,
      "theoretical_loss": 3.5785619390405667,
      "tokens_seen": 1232928768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163991975927784,
      "loss": 2.8987,
      "theoretical_loss": 3.5785445311050754,
      "tokens_seen": 1232994304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163891675025075,
      "loss": 2.3145,
      "theoretical_loss": 3.578527124353883,
      "tokens_seen": 1233059840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031637913741223674,
      "loss": 2.6965,
      "theoretical_loss": 3.5785097187868455,
      "tokens_seen": 1233125376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031636910732196587,
      "loss": 2.6582,
      "theoretical_loss": 3.57849231440382,
      "tokens_seen": 1233190912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163590772316951,
      "loss": 2.5731,
      "theoretical_loss": 3.5784749112046637,
      "tokens_seen": 1233256448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163490471414243,
      "loss": 2.5886,
      "theoretical_loss": 3.578457509189232,
      "tokens_seen": 1233321984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031633901705115347,
      "loss": 2.6165,
      "theoretical_loss": 3.578440108357382,
      "tokens_seen": 1233387520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031632898696088265,
      "loss": 2.663,
      "theoretical_loss": 3.57842270870897,
      "tokens_seen": 1233453056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031631895687061183,
      "loss": 2.7879,
      "theoretical_loss": 3.578405310243853,
      "tokens_seen": 1233518592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316308926780341,
      "loss": 2.4878,
      "theoretical_loss": 3.578387912961887,
      "tokens_seen": 1233584128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031629889669007024,
      "loss": 2.8083,
      "theoretical_loss": 3.57837051686293,
      "tokens_seen": 1233649664
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 683327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9415130615234375,
      "objective/train/theoretical_loss": 3.5783531219468374,
      "objective/train/tokens_used": 1254175200,
      "theoretical_loss": 3.5783531219468374,
      "tokens_seen": 1233715200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031628886659979937,
      "loss": 2.7898,
      "theoretical_loss": 3.5783531219468374,
      "tokens_seen": 1233715200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162788365095286,
      "loss": 2.555,
      "theoretical_loss": 3.5783357282134665,
      "tokens_seen": 1233780736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031626880641925773,
      "loss": 2.7074,
      "theoretical_loss": 3.578318335662675,
      "tokens_seen": 1233846272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031625877632898697,
      "loss": 2.7915,
      "theoretical_loss": 3.5783009442943174,
      "tokens_seen": 1233911808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031624874623871615,
      "loss": 2.6769,
      "theoretical_loss": 3.5782835541082525,
      "tokens_seen": 1233977344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031623871614844533,
      "loss": 2.595,
      "theoretical_loss": 3.5782661651043366,
      "tokens_seen": 1234042880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162286860581745,
      "loss": 2.6416,
      "theoretical_loss": 3.5782487772824263,
      "tokens_seen": 1234108416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031621865596790375,
      "loss": 2.8331,
      "theoretical_loss": 3.5782313906423786,
      "tokens_seen": 1234173952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162086258776329,
      "loss": 2.4344,
      "theoretical_loss": 3.5782140051840505,
      "tokens_seen": 1234239488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161985957873621,
      "loss": 2.4601,
      "theoretical_loss": 3.578196620907299,
      "tokens_seen": 1234305024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031618856569709124,
      "loss": 2.7753,
      "theoretical_loss": 3.578179237811981,
      "tokens_seen": 1234370560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031617853560682047,
      "loss": 2.722,
      "theoretical_loss": 3.578161855897953,
      "tokens_seen": 1234436096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031616850551654965,
      "loss": 2.5857,
      "theoretical_loss": 3.5781444751650735,
      "tokens_seen": 1234501632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031615847542627883,
      "loss": 2.5077,
      "theoretical_loss": 3.5781270956131976,
      "tokens_seen": 1234567168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316148445336008,
      "loss": 2.5656,
      "theoretical_loss": 3.5781097172421834,
      "tokens_seen": 1234632704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161384152457372,
      "loss": 2.5073,
      "theoretical_loss": 3.5780923400518883,
      "tokens_seen": 1234698240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161283851554664,
      "loss": 2.6107,
      "theoretical_loss": 3.5780749640421687,
      "tokens_seen": 1234763776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161183550651956,
      "loss": 2.5824,
      "theoretical_loss": 3.5780575892128823,
      "tokens_seen": 1234829312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031610832497492474,
      "loss": 2.631,
      "theoretical_loss": 3.5780402155638855,
      "tokens_seen": 1234894848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316098294884654,
      "loss": 2.4193,
      "theoretical_loss": 3.5780228430950363,
      "tokens_seen": 1234960384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160882647943831,
      "loss": 2.7855,
      "theoretical_loss": 3.5780054718061916,
      "tokens_seen": 1235025920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031607823470411234,
      "loss": 2.6025,
      "theoretical_loss": 3.5779881016972093,
      "tokens_seen": 1235091456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160682046138415,
      "loss": 2.8302,
      "theoretical_loss": 3.577970732767945,
      "tokens_seen": 1235156992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160581745235707,
      "loss": 2.5224,
      "theoretical_loss": 3.5779533650182573,
      "tokens_seen": 1235222528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031604814443329994,
      "loss": 2.7319,
      "theoretical_loss": 3.5779359984480035,
      "tokens_seen": 1235288064
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 684175,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.004976272583008,
      "objective/train/theoretical_loss": 3.5779186330570405,
      "objective/train/tokens_used": 1255813600,
      "theoretical_loss": 3.5779186330570405,
      "tokens_seen": 1235353600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160381143430291,
      "loss": 2.6376,
      "theoretical_loss": 3.5779186330570405,
      "tokens_seen": 1235353600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160280842527583,
      "loss": 2.6485,
      "theoretical_loss": 3.577901268845226,
      "tokens_seen": 1235419136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160180541624875,
      "loss": 2.6592,
      "theoretical_loss": 3.5778839058124166,
      "tokens_seen": 1235484672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031600802407221666,
      "loss": 2.5898,
      "theoretical_loss": 3.577866543958471,
      "tokens_seen": 1235550208
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031599799398194584,
      "loss": 2.7968,
      "theoretical_loss": 3.577849183283245,
      "tokens_seen": 1235615744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159879638916751,
      "loss": 2.7619,
      "theoretical_loss": 3.577831823786598,
      "tokens_seen": 1235681280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159779338014042,
      "loss": 2.6657,
      "theoretical_loss": 3.577814465468386,
      "tokens_seen": 1235746816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031596790371113344,
      "loss": 2.5929,
      "theoretical_loss": 3.5777971083284674,
      "tokens_seen": 1235812352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031595787362086257,
      "loss": 2.6074,
      "theoretical_loss": 3.577779752366699,
      "tokens_seen": 1235877888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159478435305918,
      "loss": 2.5263,
      "theoretical_loss": 3.577762397582939,
      "tokens_seen": 1235943424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000315937813440321,
      "loss": 2.5976,
      "theoretical_loss": 3.5777450439770444,
      "tokens_seen": 1236008960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031592778335005016,
      "loss": 2.7342,
      "theoretical_loss": 3.5777276915488736,
      "tokens_seen": 1236074496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031591775325977934,
      "loss": 2.6336,
      "theoretical_loss": 3.5777103402982835,
      "tokens_seen": 1236140032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159077231695086,
      "loss": 2.8675,
      "theoretical_loss": 3.577692990225132,
      "tokens_seen": 1236205568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158976930792377,
      "loss": 2.791,
      "theoretical_loss": 3.577675641329277,
      "tokens_seen": 1236271104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031588766298896694,
      "loss": 2.7231,
      "theoretical_loss": 3.5776582936105754,
      "tokens_seen": 1236336640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031587763289869607,
      "loss": 2.575,
      "theoretical_loss": 3.577640947068886,
      "tokens_seen": 1236402176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158676028084253,
      "loss": 2.8744,
      "theoretical_loss": 3.577623601704066,
      "tokens_seen": 1236467712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158575727181545,
      "loss": 2.7638,
      "theoretical_loss": 3.5776062575159733,
      "tokens_seen": 1236533248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031584754262788367,
      "loss": 2.7146,
      "theoretical_loss": 3.577588914504466,
      "tokens_seen": 1236598784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031583751253761285,
      "loss": 2.8056,
      "theoretical_loss": 3.5775715726694015,
      "tokens_seen": 1236664320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031582748244734203,
      "loss": 2.5687,
      "theoretical_loss": 3.5775542320106375,
      "tokens_seen": 1236729856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158174523570712,
      "loss": 2.8633,
      "theoretical_loss": 3.577536892528032,
      "tokens_seen": 1236795392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031580742226680044,
      "loss": 2.4725,
      "theoretical_loss": 3.577519554221444,
      "tokens_seen": 1236860928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031579739217652957,
      "loss": 2.7498,
      "theoretical_loss": 3.5775022170907302,
      "tokens_seen": 1236926464
    },
    {
      "epoch": 0.37,
      "objective/train/docs_used": 684175,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.575035572052002,
      "objective/train/theoretical_loss": 3.5774848811357485,
      "objective/train/tokens_used": 1257452000,
      "theoretical_loss": 3.5774848811357485,
      "tokens_seen": 1236992000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157873620862588,
      "loss": 2.7459,
      "theoretical_loss": 3.5774848811357485,
      "tokens_seen": 1236992000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031577733199598793,
      "loss": 2.5454,
      "theoretical_loss": 3.5774675463563574,
      "tokens_seen": 1237057536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031576730190571717,
      "loss": 2.6331,
      "theoretical_loss": 3.5774502127524146,
      "tokens_seen": 1237123072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031575727181544635,
      "loss": 2.7631,
      "theoretical_loss": 3.5774328803237787,
      "tokens_seen": 1237188608
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031574724172517553,
      "loss": 2.841,
      "theoretical_loss": 3.577415549070307,
      "tokens_seen": 1237254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157372116349047,
      "loss": 2.6371,
      "theoretical_loss": 3.577398218991858,
      "tokens_seen": 1237319680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031572718154463395,
      "loss": 2.8772,
      "theoretical_loss": 3.57738089008829,
      "tokens_seen": 1237385216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157171514543631,
      "loss": 2.8408,
      "theoretical_loss": 3.577363562359461,
      "tokens_seen": 1237450752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003157071213640923,
      "loss": 2.6924,
      "theoretical_loss": 3.5773462358052286,
      "tokens_seen": 1237516288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031569709127382144,
      "loss": 2.6864,
      "theoretical_loss": 3.5773289104254515,
      "tokens_seen": 1237581824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031568706118355067,
      "loss": 2.7785,
      "theoretical_loss": 3.577311586219988,
      "tokens_seen": 1237647360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031567703109327985,
      "loss": 2.5432,
      "theoretical_loss": 3.577294263188696,
      "tokens_seen": 1237712896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031566700100300903,
      "loss": 2.5293,
      "theoretical_loss": 3.577276941331434,
      "tokens_seen": 1237778432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156569709127382,
      "loss": 2.6125,
      "theoretical_loss": 3.57725962064806,
      "tokens_seen": 1237843968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156469408224674,
      "loss": 2.5755,
      "theoretical_loss": 3.5772423011384324,
      "tokens_seen": 1237909504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156369107321966,
      "loss": 2.6375,
      "theoretical_loss": 3.5772249828024103,
      "tokens_seen": 1237975040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156268806419258,
      "loss": 2.4348,
      "theoretical_loss": 3.577207665639851,
      "tokens_seen": 1238040576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031561685055165494,
      "loss": 2.7653,
      "theoretical_loss": 3.5771903496506128,
      "tokens_seen": 1238106112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156068204613842,
      "loss": 2.8229,
      "theoretical_loss": 3.577173034834555,
      "tokens_seen": 1238171648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155967903711133,
      "loss": 2.6971,
      "theoretical_loss": 3.5771557211915352,
      "tokens_seen": 1238237184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031558676028084254,
      "loss": 2.7564,
      "theoretical_loss": 3.5771384087214124,
      "tokens_seen": 1238302720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155767301905717,
      "loss": 2.4933,
      "theoretical_loss": 3.577121097424045,
      "tokens_seen": 1238368256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155667001003009,
      "loss": 2.5807,
      "theoretical_loss": 3.5771037872992917,
      "tokens_seen": 1238433792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155566700100301,
      "loss": 2.5816,
      "theoretical_loss": 3.57708647834701,
      "tokens_seen": 1238499328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155466399197593,
      "loss": 2.9417,
      "theoretical_loss": 3.5770691705670594,
      "tokens_seen": 1238564864
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 685182,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7375147342681885,
      "objective/train/theoretical_loss": 3.5770518639592983,
      "objective/train/tokens_used": 1259090400,
      "theoretical_loss": 3.5770518639592983,
      "tokens_seen": 1238630400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031553660982948844,
      "loss": 2.6052,
      "theoretical_loss": 3.5770518639592983,
      "tokens_seen": 1238630400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155265797392177,
      "loss": 2.7781,
      "theoretical_loss": 3.577034558523585,
      "tokens_seen": 1238695936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155165496489468,
      "loss": 2.8369,
      "theoretical_loss": 3.577017254259779,
      "tokens_seen": 1238761472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031550651955867604,
      "loss": 2.7518,
      "theoretical_loss": 3.5769999511677377,
      "tokens_seen": 1238827008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154964894684052,
      "loss": 2.5688,
      "theoretical_loss": 3.576982649247321,
      "tokens_seen": 1238892544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154864593781344,
      "loss": 2.7044,
      "theoretical_loss": 3.576965348498386,
      "tokens_seen": 1238958080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154764292878636,
      "loss": 2.9206,
      "theoretical_loss": 3.5769480489207934,
      "tokens_seen": 1239023616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031546639919759277,
      "loss": 2.8187,
      "theoretical_loss": 3.5769307505144003,
      "tokens_seen": 1239089152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031545636910732195,
      "loss": 2.5674,
      "theoretical_loss": 3.576913453279066,
      "tokens_seen": 1239154688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154463390170512,
      "loss": 2.7482,
      "theoretical_loss": 3.57689615721465,
      "tokens_seen": 1239220224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154363089267803,
      "loss": 2.8608,
      "theoretical_loss": 3.57687886232101,
      "tokens_seen": 1239285760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031542627883650954,
      "loss": 2.6337,
      "theoretical_loss": 3.5768615685980056,
      "tokens_seen": 1239351296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154162487462387,
      "loss": 2.7525,
      "theoretical_loss": 3.5768442760454953,
      "tokens_seen": 1239416832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154062186559679,
      "loss": 2.6064,
      "theoretical_loss": 3.5768269846633385,
      "tokens_seen": 1239482368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153961885656971,
      "loss": 2.6995,
      "theoretical_loss": 3.576809694451393,
      "tokens_seen": 1239547904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031538615847542627,
      "loss": 2.5633,
      "theoretical_loss": 3.5767924054095195,
      "tokens_seen": 1239613440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031537612838515545,
      "loss": 2.7152,
      "theoretical_loss": 3.5767751175375753,
      "tokens_seen": 1239678976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153660982948847,
      "loss": 2.8287,
      "theoretical_loss": 3.5767578308354198,
      "tokens_seen": 1239744512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153560682046138,
      "loss": 2.8507,
      "theoretical_loss": 3.5767405453029126,
      "tokens_seen": 1239810048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031534603811434305,
      "loss": 2.799,
      "theoretical_loss": 3.5767232609399127,
      "tokens_seen": 1239875584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153360080240722,
      "loss": 2.725,
      "theoretical_loss": 3.576705977746278,
      "tokens_seen": 1239941120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153259779338014,
      "loss": 2.6619,
      "theoretical_loss": 3.5766886957218693,
      "tokens_seen": 1240006656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153159478435306,
      "loss": 2.8006,
      "theoretical_loss": 3.576671414866544,
      "tokens_seen": 1240072192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031530591775325977,
      "loss": 2.6383,
      "theoretical_loss": 3.5766541351801626,
      "tokens_seen": 1240137728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315295887662989,
      "loss": 2.5801,
      "theoretical_loss": 3.576636856662584,
      "tokens_seen": 1240203264
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 685844,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0006914138793945,
      "objective/train/theoretical_loss": 3.5766195793136664,
      "objective/train/tokens_used": 1260728800,
      "theoretical_loss": 3.5766195793136664,
      "tokens_seen": 1240268800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031528585757271813,
      "loss": 2.4128,
      "theoretical_loss": 3.5766195793136664,
      "tokens_seen": 1240268800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031527582748244737,
      "loss": 2.6779,
      "theoretical_loss": 3.5766023031332703,
      "tokens_seen": 1240334336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031526579739217655,
      "loss": 2.8779,
      "theoretical_loss": 3.5765850281212543,
      "tokens_seen": 1240399872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031525576730190573,
      "loss": 2.5602,
      "theoretical_loss": 3.5765677542774776,
      "tokens_seen": 1240465408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152457372116349,
      "loss": 2.7543,
      "theoretical_loss": 3.5765504816018,
      "tokens_seen": 1240530944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031523570712136415,
      "loss": 2.7578,
      "theoretical_loss": 3.57653321009408,
      "tokens_seen": 1240596480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152256770310933,
      "loss": 2.6434,
      "theoretical_loss": 3.5765159397541773,
      "tokens_seen": 1240662016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152156469408225,
      "loss": 2.6226,
      "theoretical_loss": 3.5764986705819517,
      "tokens_seen": 1240727552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031520561685055164,
      "loss": 2.9745,
      "theoretical_loss": 3.5764814025772615,
      "tokens_seen": 1240793088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031519558676028087,
      "loss": 2.8188,
      "theoretical_loss": 3.5764641357399674,
      "tokens_seen": 1240858624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031518555667001005,
      "loss": 2.7181,
      "theoretical_loss": 3.576446870069928,
      "tokens_seen": 1240924160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031517552657973924,
      "loss": 2.8341,
      "theoretical_loss": 3.576429605567003,
      "tokens_seen": 1240989696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151654964894684,
      "loss": 2.7724,
      "theoretical_loss": 3.576412342231052,
      "tokens_seen": 1241055232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151554663991976,
      "loss": 2.6043,
      "theoretical_loss": 3.576395080061934,
      "tokens_seen": 1241120768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151454363089268,
      "loss": 2.6358,
      "theoretical_loss": 3.576377819059509,
      "tokens_seen": 1241186304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315135406218656,
      "loss": 2.8804,
      "theoretical_loss": 3.5763605592236365,
      "tokens_seen": 1241251840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031512537612838514,
      "loss": 2.7021,
      "theoretical_loss": 3.5763433005541767,
      "tokens_seen": 1241317376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151153460381144,
      "loss": 2.3519,
      "theoretical_loss": 3.5763260430509876,
      "tokens_seen": 1241382912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151053159478435,
      "loss": 2.8287,
      "theoretical_loss": 3.57630878671393,
      "tokens_seen": 1241448448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031509528585757274,
      "loss": 2.3984,
      "theoretical_loss": 3.5762915315428634,
      "tokens_seen": 1241513984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150852557673019,
      "loss": 2.6735,
      "theoretical_loss": 3.576274277537647,
      "tokens_seen": 1241579520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150752256770311,
      "loss": 2.7345,
      "theoretical_loss": 3.5762570246981413,
      "tokens_seen": 1241645056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150651955867603,
      "loss": 2.6689,
      "theoretical_loss": 3.5762397730242053,
      "tokens_seen": 1241710592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150551654964895,
      "loss": 2.8214,
      "theoretical_loss": 3.5762225225156987,
      "tokens_seen": 1241776128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031504513540621864,
      "loss": 2.8196,
      "theoretical_loss": 3.576205273172482,
      "tokens_seen": 1241841664
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 687001,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8483855724334717,
      "objective/train/theoretical_loss": 3.5761880249944147,
      "objective/train/tokens_used": 1262367200,
      "theoretical_loss": 3.5761880249944147,
      "tokens_seen": 1241907200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150351053159479,
      "loss": 2.6797,
      "theoretical_loss": 3.5761880249944147,
      "tokens_seen": 1241907200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315025075225677,
      "loss": 2.8243,
      "theoretical_loss": 3.576170777981356,
      "tokens_seen": 1241972736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031501504513540624,
      "loss": 2.812,
      "theoretical_loss": 3.5761535321331666,
      "tokens_seen": 1242038272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150050150451354,
      "loss": 2.6161,
      "theoretical_loss": 3.5761362874497062,
      "tokens_seen": 1242103808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149949849548646,
      "loss": 2.5499,
      "theoretical_loss": 3.5761190439308344,
      "tokens_seen": 1242169344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149849548645938,
      "loss": 2.7484,
      "theoretical_loss": 3.576101801576411,
      "tokens_seen": 1242234880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031497492477432297,
      "loss": 2.7865,
      "theoretical_loss": 3.5760845603862963,
      "tokens_seen": 1242300416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031496489468405215,
      "loss": 2.7135,
      "theoretical_loss": 3.57606732036035,
      "tokens_seen": 1242365952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149548645937814,
      "loss": 2.8205,
      "theoretical_loss": 3.576050081498432,
      "tokens_seen": 1242431488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149448345035105,
      "loss": 2.6539,
      "theoretical_loss": 3.5760328438004025,
      "tokens_seen": 1242497024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031493480441323974,
      "loss": 2.5831,
      "theoretical_loss": 3.5760156072661218,
      "tokens_seen": 1242562560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149247743229689,
      "loss": 2.7412,
      "theoretical_loss": 3.57599837189545,
      "tokens_seen": 1242628096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149147442326981,
      "loss": 2.7729,
      "theoretical_loss": 3.575981137688246,
      "tokens_seen": 1242693632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149047141424273,
      "loss": 2.5824,
      "theoretical_loss": 3.5759639046443716,
      "tokens_seen": 1242759168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031489468405215647,
      "loss": 2.6411,
      "theoretical_loss": 3.575946672763686,
      "tokens_seen": 1242824704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031488465396188565,
      "loss": 2.7375,
      "theoretical_loss": 3.575929442046049,
      "tokens_seen": 1242890240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148746238716149,
      "loss": 2.7509,
      "theoretical_loss": 3.575912212491322,
      "tokens_seen": 1242955776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314864593781344,
      "loss": 2.7655,
      "theoretical_loss": 3.575894984099364,
      "tokens_seen": 1243021312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031485456369107325,
      "loss": 2.8217,
      "theoretical_loss": 3.575877756870036,
      "tokens_seen": 1243086848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148445336008024,
      "loss": 2.6609,
      "theoretical_loss": 3.575860530803198,
      "tokens_seen": 1243152384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148345035105316,
      "loss": 2.8145,
      "theoretical_loss": 3.57584330589871,
      "tokens_seen": 1243217920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148244734202608,
      "loss": 2.6594,
      "theoretical_loss": 3.575826082156432,
      "tokens_seen": 1243283456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031481444332998997,
      "loss": 2.5629,
      "theoretical_loss": 3.5758088595762256,
      "tokens_seen": 1243348992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031480441323971915,
      "loss": 2.5686,
      "theoretical_loss": 3.57579163815795,
      "tokens_seen": 1243414528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031479438314944833,
      "loss": 2.7092,
      "theoretical_loss": 3.5757744179014668,
      "tokens_seen": 1243480064
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 688085,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8447015285491943,
      "objective/train/theoretical_loss": 3.575757198806635,
      "objective/train/tokens_used": 1264005600,
      "theoretical_loss": 3.575757198806635,
      "tokens_seen": 1243545600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147843530591775,
      "loss": 2.9024,
      "theoretical_loss": 3.575757198806635,
      "tokens_seen": 1243545600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031477432296890675,
      "loss": 2.7652,
      "theoretical_loss": 3.5757399808733155,
      "tokens_seen": 1243611136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147642928786359,
      "loss": 2.8517,
      "theoretical_loss": 3.575722764101369,
      "tokens_seen": 1243676672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147542627883651,
      "loss": 2.7982,
      "theoretical_loss": 3.575705548490656,
      "tokens_seen": 1243742208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147442326980943,
      "loss": 2.4399,
      "theoretical_loss": 3.5756883340410366,
      "tokens_seen": 1243807744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147342026078235,
      "loss": 2.4896,
      "theoretical_loss": 3.5756711207523715,
      "tokens_seen": 1243873280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031472417251755266,
      "loss": 2.6552,
      "theoretical_loss": 3.5756539086245214,
      "tokens_seen": 1243938816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031471414242728184,
      "loss": 2.649,
      "theoretical_loss": 3.575636697657347,
      "tokens_seen": 1244004352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314704112337011,
      "loss": 2.7111,
      "theoretical_loss": 3.5756194878507084,
      "tokens_seen": 1244069888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031469408224674025,
      "loss": 2.6986,
      "theoretical_loss": 3.5756022792044666,
      "tokens_seen": 1244135424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146840521564694,
      "loss": 2.7551,
      "theoretical_loss": 3.5755850717184816,
      "tokens_seen": 1244200960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146740220661986,
      "loss": 2.7036,
      "theoretical_loss": 3.575567865392615,
      "tokens_seen": 1244266496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031466399197592774,
      "loss": 2.7044,
      "theoretical_loss": 3.5755506602267273,
      "tokens_seen": 1244332032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314653961885657,
      "loss": 2.715,
      "theoretical_loss": 3.575533456220678,
      "tokens_seen": 1244397568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031464393179538616,
      "loss": 2.6174,
      "theoretical_loss": 3.57551625337433,
      "tokens_seen": 1244463104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031463390170511534,
      "loss": 2.6853,
      "theoretical_loss": 3.575499051687542,
      "tokens_seen": 1244528640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146238716148445,
      "loss": 2.8587,
      "theoretical_loss": 3.575481851160176,
      "tokens_seen": 1244594176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146138415245737,
      "loss": 2.8025,
      "theoretical_loss": 3.5754646517920925,
      "tokens_seen": 1244659712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146038114343029,
      "loss": 2.8209,
      "theoretical_loss": 3.575447453583152,
      "tokens_seen": 1244725248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145937813440321,
      "loss": 2.5335,
      "theoretical_loss": 3.575430256533216,
      "tokens_seen": 1244790784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031458375125376125,
      "loss": 2.5279,
      "theoretical_loss": 3.575413060642145,
      "tokens_seen": 1244856320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145737211634905,
      "loss": 2.6258,
      "theoretical_loss": 3.5753958659097993,
      "tokens_seen": 1244921856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145636910732197,
      "loss": 2.6978,
      "theoretical_loss": 3.5753786723360412,
      "tokens_seen": 1244987392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031455366098294884,
      "loss": 2.6572,
      "theoretical_loss": 3.5753614799207307,
      "tokens_seen": 1245052928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145436308926781,
      "loss": 2.9123,
      "theoretical_loss": 3.5753442886637288,
      "tokens_seen": 1245118464
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 688769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5056591033935547,
      "objective/train/theoretical_loss": 3.5753270985648973,
      "objective/train/tokens_used": 1265644000,
      "theoretical_loss": 3.5753270985648973,
      "tokens_seen": 1245184000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145336008024072,
      "loss": 2.555,
      "theoretical_loss": 3.5753270985648973,
      "tokens_seen": 1245184000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031452357071213644,
      "loss": 2.5788,
      "theoretical_loss": 3.575309909624096,
      "tokens_seen": 1245249536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145135406218656,
      "loss": 2.7804,
      "theoretical_loss": 3.5752927218411865,
      "tokens_seen": 1245315072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145035105315948,
      "loss": 2.6175,
      "theoretical_loss": 3.57527553521603,
      "tokens_seen": 1245380608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314493480441324,
      "loss": 2.5547,
      "theoretical_loss": 3.5752583497484878,
      "tokens_seen": 1245446144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031448345035105317,
      "loss": 2.6196,
      "theoretical_loss": 3.5752411654384204,
      "tokens_seen": 1245511680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031447342026078235,
      "loss": 2.6699,
      "theoretical_loss": 3.5752239822856895,
      "tokens_seen": 1245577216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144633901705116,
      "loss": 2.7614,
      "theoretical_loss": 3.5752068002901565,
      "tokens_seen": 1245642752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144533600802407,
      "loss": 2.7317,
      "theoretical_loss": 3.5751896194516815,
      "tokens_seen": 1245708288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031444332998996994,
      "loss": 2.5388,
      "theoretical_loss": 3.5751724397701263,
      "tokens_seen": 1245773824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144332998996991,
      "loss": 2.6477,
      "theoretical_loss": 3.575155261245353,
      "tokens_seen": 1245839360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144232698094283,
      "loss": 2.7932,
      "theoretical_loss": 3.5751380838772215,
      "tokens_seen": 1245904896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144132397191575,
      "loss": 2.5666,
      "theoretical_loss": 3.5751209076655934,
      "tokens_seen": 1245970432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031440320962888667,
      "loss": 2.5904,
      "theoretical_loss": 3.575103732610331,
      "tokens_seen": 1246035968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031439317953861585,
      "loss": 2.7356,
      "theoretical_loss": 3.575086558711294,
      "tokens_seen": 1246101504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143831494483451,
      "loss": 2.6404,
      "theoretical_loss": 3.5750693859683453,
      "tokens_seen": 1246167040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143731193580742,
      "loss": 2.5866,
      "theoretical_loss": 3.5750522143813455,
      "tokens_seen": 1246232576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031436308926780345,
      "loss": 2.6652,
      "theoretical_loss": 3.575035043950156,
      "tokens_seen": 1246298112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143530591775326,
      "loss": 2.7431,
      "theoretical_loss": 3.5750178746746384,
      "tokens_seen": 1246363648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143430290872618,
      "loss": 2.7738,
      "theoretical_loss": 3.5750007065546545,
      "tokens_seen": 1246429184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314332998996991,
      "loss": 2.8112,
      "theoretical_loss": 3.574983539590065,
      "tokens_seen": 1246494720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031432296890672017,
      "loss": 2.6557,
      "theoretical_loss": 3.5749663737807325,
      "tokens_seen": 1246560256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031431293881644935,
      "loss": 2.7727,
      "theoretical_loss": 3.574949209126517,
      "tokens_seen": 1246625792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031430290872617853,
      "loss": 2.7943,
      "theoretical_loss": 3.574932045627281,
      "tokens_seen": 1246691328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142928786359077,
      "loss": 2.6418,
      "theoretical_loss": 3.5749148832828865,
      "tokens_seen": 1246756864
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 689881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9862053394317627,
      "objective/train/theoretical_loss": 3.5748977220931946,
      "objective/train/tokens_used": 1267282400,
      "theoretical_loss": 3.5748977220931946,
      "tokens_seen": 1246822400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031428284854563695,
      "loss": 2.6589,
      "theoretical_loss": 3.5748977220931946,
      "tokens_seen": 1246822400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142728184553661,
      "loss": 2.9614,
      "theoretical_loss": 3.5748805620580666,
      "tokens_seen": 1246887936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142627883650953,
      "loss": 2.4722,
      "theoretical_loss": 3.5748634031773645,
      "tokens_seen": 1246953472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142527582748245,
      "loss": 2.8278,
      "theoretical_loss": 3.5748462454509498,
      "tokens_seen": 1247019008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142427281845537,
      "loss": 2.5662,
      "theoretical_loss": 3.5748290888786842,
      "tokens_seen": 1247084544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031423269809428286,
      "loss": 2.7926,
      "theoretical_loss": 3.5748119334604302,
      "tokens_seen": 1247150080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031422266800401204,
      "loss": 2.7912,
      "theoretical_loss": 3.574794779196048,
      "tokens_seen": 1247215616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142126379137412,
      "loss": 2.6479,
      "theoretical_loss": 3.574777626085401,
      "tokens_seen": 1247281152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031420260782347045,
      "loss": 2.837,
      "theoretical_loss": 3.57476047412835,
      "tokens_seen": 1247346688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141925777331996,
      "loss": 2.6822,
      "theoretical_loss": 3.5747433233247565,
      "tokens_seen": 1247412224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141825476429288,
      "loss": 2.5263,
      "theoretical_loss": 3.5747261736744838,
      "tokens_seen": 1247477760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031417251755265794,
      "loss": 2.7709,
      "theoretical_loss": 3.574709025177392,
      "tokens_seen": 1247543296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141624874623872,
      "loss": 2.5264,
      "theoretical_loss": 3.5746918778333443,
      "tokens_seen": 1247608832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031415245737211636,
      "loss": 2.6754,
      "theoretical_loss": 3.574674731642202,
      "tokens_seen": 1247674368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031414242728184554,
      "loss": 2.8169,
      "theoretical_loss": 3.574657586603827,
      "tokens_seen": 1247739904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141323971915747,
      "loss": 2.8516,
      "theoretical_loss": 3.5746404427180813,
      "tokens_seen": 1247805440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141223671013039,
      "loss": 2.9212,
      "theoretical_loss": 3.574623299984827,
      "tokens_seen": 1247870976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141123370110331,
      "loss": 2.7596,
      "theoretical_loss": 3.5746061584039266,
      "tokens_seen": 1247936512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141023069207623,
      "loss": 2.8479,
      "theoretical_loss": 3.5745890179752413,
      "tokens_seen": 1248002048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031409227683049145,
      "loss": 2.6303,
      "theoretical_loss": 3.5745718786986336,
      "tokens_seen": 1248067584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140822467402207,
      "loss": 2.9746,
      "theoretical_loss": 3.5745547405739657,
      "tokens_seen": 1248133120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031407221664994986,
      "loss": 2.7913,
      "theoretical_loss": 3.5745376036010987,
      "tokens_seen": 1248198656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031406218655967904,
      "loss": 2.8253,
      "theoretical_loss": 3.5745204677798963,
      "tokens_seen": 1248264192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140521564694082,
      "loss": 2.7664,
      "theoretical_loss": 3.5745033331102194,
      "tokens_seen": 1248329728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140421263791374,
      "loss": 2.5198,
      "theoretical_loss": 3.5744861995919304,
      "tokens_seen": 1248395264
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 690418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.728623151779175,
      "objective/train/theoretical_loss": 3.574469067224892,
      "objective/train/tokens_used": 1268920800,
      "theoretical_loss": 3.574469067224892,
      "tokens_seen": 1248460800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140320962888666,
      "loss": 2.7498,
      "theoretical_loss": 3.574469067224892,
      "tokens_seen": 1248460800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140220661985958,
      "loss": 2.722,
      "theoretical_loss": 3.574451936008966,
      "tokens_seen": 1248526336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031401203610832495,
      "loss": 2.7592,
      "theoretical_loss": 3.5744348059440147,
      "tokens_seen": 1248591872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140020060180542,
      "loss": 2.8308,
      "theoretical_loss": 3.5744176770299,
      "tokens_seen": 1248657408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139919759277833,
      "loss": 2.6995,
      "theoretical_loss": 3.5744005492664854,
      "tokens_seen": 1248722944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031398194583751255,
      "loss": 2.8554,
      "theoretical_loss": 3.574383422653632,
      "tokens_seen": 1248788480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031397191574724173,
      "loss": 2.8966,
      "theoretical_loss": 3.5743662971912027,
      "tokens_seen": 1248854016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139618856569709,
      "loss": 2.6512,
      "theoretical_loss": 3.5743491728790593,
      "tokens_seen": 1248919552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139518555667001,
      "loss": 2.6799,
      "theoretical_loss": 3.574332049717065,
      "tokens_seen": 1248985088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139418254764293,
      "loss": 2.5296,
      "theoretical_loss": 3.574314927705082,
      "tokens_seen": 1249050624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031393179538615845,
      "loss": 2.6058,
      "theoretical_loss": 3.5742978068429716,
      "tokens_seen": 1249116160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139217652958877,
      "loss": 2.6187,
      "theoretical_loss": 3.574280687130598,
      "tokens_seen": 1249181696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139117352056168,
      "loss": 2.5322,
      "theoretical_loss": 3.5742635685678223,
      "tokens_seen": 1249247232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031390170511534605,
      "loss": 2.7232,
      "theoretical_loss": 3.574246451154508,
      "tokens_seen": 1249312768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031389167502507523,
      "loss": 2.7474,
      "theoretical_loss": 3.574229334890517,
      "tokens_seen": 1249378304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138816449348044,
      "loss": 2.6216,
      "theoretical_loss": 3.5742122197757125,
      "tokens_seen": 1249443840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138716148445336,
      "loss": 2.5586,
      "theoretical_loss": 3.574195105809956,
      "tokens_seen": 1249509376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138615847542628,
      "loss": 2.4887,
      "theoretical_loss": 3.5741779929931106,
      "tokens_seen": 1249574912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031385155466399196,
      "loss": 2.6734,
      "theoretical_loss": 3.5741608813250396,
      "tokens_seen": 1249640448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138415245737212,
      "loss": 2.8049,
      "theoretical_loss": 3.5741437708056045,
      "tokens_seen": 1249705984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138314944834503,
      "loss": 2.5464,
      "theoretical_loss": 3.5741266614346685,
      "tokens_seen": 1249771520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031382146439317955,
      "loss": 2.6368,
      "theoretical_loss": 3.574109553212095,
      "tokens_seen": 1249837056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031381143430290873,
      "loss": 2.5456,
      "theoretical_loss": 3.5740924461377457,
      "tokens_seen": 1249902592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138014042126379,
      "loss": 2.5049,
      "theoretical_loss": 3.5740753402114835,
      "tokens_seen": 1249968128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031379137412236715,
      "loss": 2.5967,
      "theoretical_loss": 3.5740582354331716,
      "tokens_seen": 1250033664
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 691448,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7786600589752197,
      "objective/train/theoretical_loss": 3.574041131802672,
      "objective/train/tokens_used": 1270559200,
      "theoretical_loss": 3.574041131802672,
      "tokens_seen": 1250099200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137813440320963,
      "loss": 2.703,
      "theoretical_loss": 3.574041131802672,
      "tokens_seen": 1250099200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137713139418255,
      "loss": 2.5466,
      "theoretical_loss": 3.5740240293198484,
      "tokens_seen": 1250164736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137612838515547,
      "loss": 2.8031,
      "theoretical_loss": 3.574006927984563,
      "tokens_seen": 1250230272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137512537612839,
      "loss": 2.6104,
      "theoretical_loss": 3.5739898277966793,
      "tokens_seen": 1250295808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031374122367101306,
      "loss": 2.7823,
      "theoretical_loss": 3.5739727287560594,
      "tokens_seen": 1250361344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031373119358074224,
      "loss": 2.7496,
      "theoretical_loss": 3.5739556308625673,
      "tokens_seen": 1250426880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137211634904714,
      "loss": 2.7246,
      "theoretical_loss": 3.5739385341160643,
      "tokens_seen": 1250492416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031371113340020065,
      "loss": 2.6489,
      "theoretical_loss": 3.5739214385164146,
      "tokens_seen": 1250557952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137011033099298,
      "loss": 2.5355,
      "theoretical_loss": 3.573904344063481,
      "tokens_seen": 1250623488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313691073219659,
      "loss": 2.7299,
      "theoretical_loss": 3.573887250757126,
      "tokens_seen": 1250689024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031368104312938814,
      "loss": 2.61,
      "theoretical_loss": 3.573870158597214,
      "tokens_seen": 1250754560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136710130391174,
      "loss": 2.885,
      "theoretical_loss": 3.573853067583606,
      "tokens_seen": 1250820096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031366098294884656,
      "loss": 2.8004,
      "theoretical_loss": 3.5738359777161666,
      "tokens_seen": 1250885632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031365095285857574,
      "loss": 2.8157,
      "theoretical_loss": 3.573818888994758,
      "tokens_seen": 1250951168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136409227683049,
      "loss": 2.6856,
      "theoretical_loss": 3.5738018014192434,
      "tokens_seen": 1251016704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136308926780341,
      "loss": 2.625,
      "theoretical_loss": 3.573784714989487,
      "tokens_seen": 1251082240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136208625877633,
      "loss": 2.7976,
      "theoretical_loss": 3.5737676297053502,
      "tokens_seen": 1251147776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136108324974925,
      "loss": 2.5148,
      "theoretical_loss": 3.573750545566698,
      "tokens_seen": 1251213312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031360080240722165,
      "loss": 2.4938,
      "theoretical_loss": 3.5737334625733927,
      "tokens_seen": 1251278848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135907723169509,
      "loss": 2.6541,
      "theoretical_loss": 3.5737163807252976,
      "tokens_seen": 1251344384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031358074222668006,
      "loss": 2.7878,
      "theoretical_loss": 3.5736993000222754,
      "tokens_seen": 1251409920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031357071213640924,
      "loss": 2.6138,
      "theoretical_loss": 3.5736822204641907,
      "tokens_seen": 1251475456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135606820461384,
      "loss": 2.6757,
      "theoretical_loss": 3.5736651420509054,
      "tokens_seen": 1251540992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135506519558676,
      "loss": 2.4925,
      "theoretical_loss": 3.573648064782284,
      "tokens_seen": 1251606528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135406218655968,
      "loss": 2.5749,
      "theoretical_loss": 3.5736309886581887,
      "tokens_seen": 1251672064
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 692076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7197704315185547,
      "objective/train/theoretical_loss": 3.573613913678484,
      "objective/train/tokens_used": 1272197600,
      "theoretical_loss": 3.573613913678484,
      "tokens_seen": 1251737600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313530591775326,
      "loss": 2.7437,
      "theoretical_loss": 3.573613913678484,
      "tokens_seen": 1251737600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031352056168505515,
      "loss": 2.6848,
      "theoretical_loss": 3.5735968398430327,
      "tokens_seen": 1251803136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135105315947844,
      "loss": 2.2709,
      "theoretical_loss": 3.5735797671516982,
      "tokens_seen": 1251868672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135005015045135,
      "loss": 2.3752,
      "theoretical_loss": 3.5735626956043443,
      "tokens_seen": 1251934208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031349047141424275,
      "loss": 2.6238,
      "theoretical_loss": 3.573545625200834,
      "tokens_seen": 1251999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031348044132397193,
      "loss": 2.3681,
      "theoretical_loss": 3.573528555941031,
      "tokens_seen": 1252065280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134704112337011,
      "loss": 2.7673,
      "theoretical_loss": 3.573511487824799,
      "tokens_seen": 1252130816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134603811434303,
      "loss": 2.812,
      "theoretical_loss": 3.5734944208520014,
      "tokens_seen": 1252196352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134503510531595,
      "loss": 2.654,
      "theoretical_loss": 3.5734773550225016,
      "tokens_seen": 1252261888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031344032096288865,
      "loss": 2.6109,
      "theoretical_loss": 3.573460290336163,
      "tokens_seen": 1252327424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134302908726179,
      "loss": 2.6974,
      "theoretical_loss": 3.57344322679285,
      "tokens_seen": 1252392960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313420260782347,
      "loss": 2.5873,
      "theoretical_loss": 3.5734261643924254,
      "tokens_seen": 1252458496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031341023069207625,
      "loss": 2.6487,
      "theoretical_loss": 3.5734091031347535,
      "tokens_seen": 1252524032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031340020060180543,
      "loss": 2.6721,
      "theoretical_loss": 3.5733920430196973,
      "tokens_seen": 1252589568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133901705115346,
      "loss": 2.7647,
      "theoretical_loss": 3.5733749840471214,
      "tokens_seen": 1252655104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133801404212638,
      "loss": 2.8827,
      "theoretical_loss": 3.5733579262168886,
      "tokens_seen": 1252720640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313370110330993,
      "loss": 2.7577,
      "theoretical_loss": 3.573340869528863,
      "tokens_seen": 1252786176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031336008024072216,
      "loss": 2.5466,
      "theoretical_loss": 3.573323813982909,
      "tokens_seen": 1252851712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133500501504514,
      "loss": 2.9619,
      "theoretical_loss": 3.573306759578889,
      "tokens_seen": 1252917248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133400200601805,
      "loss": 2.8034,
      "theoretical_loss": 3.573289706316668,
      "tokens_seen": 1252982784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031332998996990975,
      "loss": 2.592,
      "theoretical_loss": 3.57327265419611,
      "tokens_seen": 1253048320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133199598796389,
      "loss": 2.5597,
      "theoretical_loss": 3.5732556032170777,
      "tokens_seen": 1253113856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133099297893681,
      "loss": 2.5638,
      "theoretical_loss": 3.573238553379436,
      "tokens_seen": 1253179392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132998996990973,
      "loss": 2.7289,
      "theoretical_loss": 3.5732215046830484,
      "tokens_seen": 1253244928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132898696088265,
      "loss": 2.47,
      "theoretical_loss": 3.573204457127779,
      "tokens_seen": 1253310464
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 693502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.35072660446167,
      "objective/train/theoretical_loss": 3.5731874107134916,
      "objective/train/tokens_used": 1273836000,
      "theoretical_loss": 3.5731874107134916,
      "tokens_seen": 1253376000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031327983951855566,
      "loss": 2.595,
      "theoretical_loss": 3.5731874107134916,
      "tokens_seen": 1253376000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132698094282849,
      "loss": 2.8433,
      "theoretical_loss": 3.57317036544005,
      "tokens_seen": 1253441536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313259779338014,
      "loss": 2.8264,
      "theoretical_loss": 3.573153321307318,
      "tokens_seen": 1253507072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031324974924774326,
      "loss": 2.6044,
      "theoretical_loss": 3.5731362783151606,
      "tokens_seen": 1253572608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132397191574724,
      "loss": 2.6706,
      "theoretical_loss": 3.5731192364634414,
      "tokens_seen": 1253638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132296890672016,
      "loss": 2.5902,
      "theoretical_loss": 3.573102195752024,
      "tokens_seen": 1253703680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132196589769308,
      "loss": 2.6405,
      "theoretical_loss": 3.573085156180773,
      "tokens_seen": 1253769216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031320962888666,
      "loss": 2.6236,
      "theoretical_loss": 3.5730681177495525,
      "tokens_seen": 1253834752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031319959879638916,
      "loss": 2.7958,
      "theoretical_loss": 3.5730510804582267,
      "tokens_seen": 1253900288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031318956870611834,
      "loss": 2.6758,
      "theoretical_loss": 3.5730340443066595,
      "tokens_seen": 1253965824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131795386158475,
      "loss": 2.4292,
      "theoretical_loss": 3.5730170092947153,
      "tokens_seen": 1254031360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031316950852557676,
      "loss": 2.8543,
      "theoretical_loss": 3.572999975422258,
      "tokens_seen": 1254096896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131594784353059,
      "loss": 2.7389,
      "theoretical_loss": 3.572982942689152,
      "tokens_seen": 1254162432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131494483450351,
      "loss": 2.5684,
      "theoretical_loss": 3.572965911095262,
      "tokens_seen": 1254227968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031313941825476425,
      "loss": 2.6807,
      "theoretical_loss": 3.5729488806404515,
      "tokens_seen": 1254293504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131293881644935,
      "loss": 2.646,
      "theoretical_loss": 3.572931851324585,
      "tokens_seen": 1254359040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031311935807422267,
      "loss": 2.7752,
      "theoretical_loss": 3.572914823147528,
      "tokens_seen": 1254424576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031310932798395185,
      "loss": 2.8694,
      "theoretical_loss": 3.5728977961091433,
      "tokens_seen": 1254490112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031309929789368103,
      "loss": 2.6834,
      "theoretical_loss": 3.5728807702092955,
      "tokens_seen": 1254555648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031308926780341026,
      "loss": 2.7661,
      "theoretical_loss": 3.57286374544785,
      "tokens_seen": 1254621184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130792377131394,
      "loss": 2.5441,
      "theoretical_loss": 3.5728467218246704,
      "tokens_seen": 1254686720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130692076228686,
      "loss": 2.7302,
      "theoretical_loss": 3.5728296993396214,
      "tokens_seen": 1254752256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130591775325978,
      "loss": 2.5853,
      "theoretical_loss": 3.572812677992567,
      "tokens_seen": 1254817792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313049147442327,
      "loss": 2.6742,
      "theoretical_loss": 3.572795657783372,
      "tokens_seen": 1254883328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130391173520562,
      "loss": 2.7661,
      "theoretical_loss": 3.5727786387119016,
      "tokens_seen": 1254948864
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 693948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6206564903259277,
      "objective/train/theoretical_loss": 3.57276162077802,
      "objective/train/tokens_used": 1275474400,
      "theoretical_loss": 3.57276162077802,
      "tokens_seen": 1255014400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031302908726178535,
      "loss": 2.5951,
      "theoretical_loss": 3.57276162077802,
      "tokens_seen": 1255014400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130190571715146,
      "loss": 2.7347,
      "theoretical_loss": 3.572744603981591,
      "tokens_seen": 1255079936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130090270812437,
      "loss": 2.6406,
      "theoretical_loss": 3.57272758832248,
      "tokens_seen": 1255145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031299899699097295,
      "loss": 2.7352,
      "theoretical_loss": 3.572710573800551,
      "tokens_seen": 1255211008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031298896690070213,
      "loss": 2.5871,
      "theoretical_loss": 3.572693560415669,
      "tokens_seen": 1255276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129789368104313,
      "loss": 2.5636,
      "theoretical_loss": 3.5726765481676988,
      "tokens_seen": 1255342080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129689067201605,
      "loss": 2.6012,
      "theoretical_loss": 3.572659537056505,
      "tokens_seen": 1255407616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129588766298897,
      "loss": 2.8618,
      "theoretical_loss": 3.5726425270819515,
      "tokens_seen": 1255473152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031294884653961885,
      "loss": 2.5406,
      "theoretical_loss": 3.572625518243904,
      "tokens_seen": 1255538688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129388164493481,
      "loss": 2.8062,
      "theoretical_loss": 3.572608510542228,
      "tokens_seen": 1255604224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129287863590772,
      "loss": 2.7838,
      "theoretical_loss": 3.572591503976786,
      "tokens_seen": 1255669760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031291875626880645,
      "loss": 2.6065,
      "theoretical_loss": 3.5725744985474446,
      "tokens_seen": 1255735296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031290872617853563,
      "loss": 2.6006,
      "theoretical_loss": 3.572557494254068,
      "tokens_seen": 1255800832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128986960882648,
      "loss": 2.5443,
      "theoretical_loss": 3.572540491096521,
      "tokens_seen": 1255866368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312888665997994,
      "loss": 2.6986,
      "theoretical_loss": 3.572523489074668,
      "tokens_seen": 1255931904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128786359077232,
      "loss": 2.6317,
      "theoretical_loss": 3.5725064881883757,
      "tokens_seen": 1255997440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031286860581745236,
      "loss": 2.7196,
      "theoretical_loss": 3.572489488437507,
      "tokens_seen": 1256062976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128585757271816,
      "loss": 2.7257,
      "theoretical_loss": 3.5724724898219278,
      "tokens_seen": 1256128512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128485456369107,
      "loss": 2.6395,
      "theoretical_loss": 3.5724554923415024,
      "tokens_seen": 1256194048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031283851554663995,
      "loss": 2.7318,
      "theoretical_loss": 3.5724384959960966,
      "tokens_seen": 1256259584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128284854563691,
      "loss": 2.7062,
      "theoretical_loss": 3.572421500785575,
      "tokens_seen": 1256325120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128184553660983,
      "loss": 2.8028,
      "theoretical_loss": 3.5724045067098027,
      "tokens_seen": 1256390656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128084252758275,
      "loss": 2.8045,
      "theoretical_loss": 3.572387513768645,
      "tokens_seen": 1256456192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127983951855567,
      "loss": 2.6767,
      "theoretical_loss": 3.5723705219619664,
      "tokens_seen": 1256521728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031278836509528586,
      "loss": 2.5454,
      "theoretical_loss": 3.572353531289632,
      "tokens_seen": 1256587264
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 694968,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5616726875305176,
      "objective/train/theoretical_loss": 3.5723365417515076,
      "objective/train/tokens_used": 1277112800,
      "theoretical_loss": 3.5723365417515076,
      "tokens_seen": 1256652800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127783350050151,
      "loss": 2.7386,
      "theoretical_loss": 3.5723365417515076,
      "tokens_seen": 1256652800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127683049147442,
      "loss": 2.5781,
      "theoretical_loss": 3.5723195533474574,
      "tokens_seen": 1256718336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031275827482447346,
      "loss": 2.4681,
      "theoretical_loss": 3.5723025660773473,
      "tokens_seen": 1256783872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127482447342026,
      "loss": 2.7769,
      "theoretical_loss": 3.5722855799410427,
      "tokens_seen": 1256849408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127382146439318,
      "loss": 2.6774,
      "theoretical_loss": 3.572268594938408,
      "tokens_seen": 1256914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312728184553661,
      "loss": 2.7947,
      "theoretical_loss": 3.5722516110693086,
      "tokens_seen": 1256980480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127181544633902,
      "loss": 2.8695,
      "theoretical_loss": 3.572234628333611,
      "tokens_seen": 1257046016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031270812437311936,
      "loss": 2.6988,
      "theoretical_loss": 3.572217646731178,
      "tokens_seen": 1257111552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031269809428284854,
      "loss": 2.7323,
      "theoretical_loss": 3.5722006662618773,
      "tokens_seen": 1257177088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126880641925777,
      "loss": 2.5644,
      "theoretical_loss": 3.572183686925573,
      "tokens_seen": 1257242624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031267803410230696,
      "loss": 2.7084,
      "theoretical_loss": 3.572166708722131,
      "tokens_seen": 1257308160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126680040120361,
      "loss": 2.7757,
      "theoretical_loss": 3.5721497316514164,
      "tokens_seen": 1257373696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126579739217653,
      "loss": 2.8024,
      "theoretical_loss": 3.5721327557132945,
      "tokens_seen": 1257439232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031264794383149445,
      "loss": 2.4816,
      "theoretical_loss": 3.572115780907631,
      "tokens_seen": 1257504768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126379137412237,
      "loss": 2.614,
      "theoretical_loss": 3.572098807234291,
      "tokens_seen": 1257570304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031262788365095287,
      "loss": 2.636,
      "theoretical_loss": 3.5720818346931402,
      "tokens_seen": 1257635840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031261785356068205,
      "loss": 2.5206,
      "theoretical_loss": 3.572064863284044,
      "tokens_seen": 1257701376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031260782347041123,
      "loss": 3.0007,
      "theoretical_loss": 3.572047893006868,
      "tokens_seen": 1257766912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031259779338014046,
      "loss": 2.7366,
      "theoretical_loss": 3.5720309238614774,
      "tokens_seen": 1257832448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125877632898696,
      "loss": 2.6415,
      "theoretical_loss": 3.572013955847738,
      "tokens_seen": 1257897984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125777331995988,
      "loss": 2.5673,
      "theoretical_loss": 3.5719969889655157,
      "tokens_seen": 1257963520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031256770310932795,
      "loss": 2.9064,
      "theoretical_loss": 3.5719800232146754,
      "tokens_seen": 1258029056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125576730190572,
      "loss": 2.6507,
      "theoretical_loss": 3.5719630585950837,
      "tokens_seen": 1258094592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031254764292878637,
      "loss": 3.0106,
      "theoretical_loss": 3.5719460951066058,
      "tokens_seen": 1258160128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031253761283851555,
      "loss": 2.6872,
      "theoretical_loss": 3.5719291327491063,
      "tokens_seen": 1258225664
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 695499,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6464076042175293,
      "objective/train/theoretical_loss": 3.5719121715224524,
      "objective/train/tokens_used": 1278751200,
      "theoretical_loss": 3.5719121715224524,
      "tokens_seen": 1258291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031252758274824473,
      "loss": 2.5867,
      "theoretical_loss": 3.5719121715224524,
      "tokens_seen": 1258291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125175526579739,
      "loss": 2.7185,
      "theoretical_loss": 3.5718952114265092,
      "tokens_seen": 1258356736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125075225677031,
      "loss": 2.5542,
      "theoretical_loss": 3.5718782524611425,
      "tokens_seen": 1258422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031249749247743233,
      "loss": 2.7311,
      "theoretical_loss": 3.5718612946262183,
      "tokens_seen": 1258487808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031248746238716146,
      "loss": 2.6752,
      "theoretical_loss": 3.571844337921602,
      "tokens_seen": 1258553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124774322968907,
      "loss": 2.6093,
      "theoretical_loss": 3.571827382347159,
      "tokens_seen": 1258618880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124674022066198,
      "loss": 2.5055,
      "theoretical_loss": 3.5718104279027565,
      "tokens_seen": 1258684416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031245737211634905,
      "loss": 2.6449,
      "theoretical_loss": 3.571793474588259,
      "tokens_seen": 1258749952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031244734202607823,
      "loss": 2.606,
      "theoretical_loss": 3.571776522403533,
      "tokens_seen": 1258815488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124373119358074,
      "loss": 2.6274,
      "theoretical_loss": 3.5717595713484442,
      "tokens_seen": 1258881024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124272818455366,
      "loss": 2.8447,
      "theoretical_loss": 3.5717426214228594,
      "tokens_seen": 1258946560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031241725175526583,
      "loss": 2.7009,
      "theoretical_loss": 3.571725672626643,
      "tokens_seen": 1259012096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031240722166499496,
      "loss": 2.9124,
      "theoretical_loss": 3.5717087249596617,
      "tokens_seen": 1259077632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123971915747242,
      "loss": 2.7454,
      "theoretical_loss": 3.5716917784217816,
      "tokens_seen": 1259143168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123871614844533,
      "loss": 2.4614,
      "theoretical_loss": 3.571674833012869,
      "tokens_seen": 1259208704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031237713139418256,
      "loss": 2.8639,
      "theoretical_loss": 3.5716578887327897,
      "tokens_seen": 1259274240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031236710130391174,
      "loss": 2.7554,
      "theoretical_loss": 3.571640945581409,
      "tokens_seen": 1259339776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123570712136409,
      "loss": 2.7072,
      "theoretical_loss": 3.571624003558594,
      "tokens_seen": 1259405312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123470411233701,
      "loss": 2.4851,
      "theoretical_loss": 3.571607062664211,
      "tokens_seen": 1259470848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123370110330993,
      "loss": 2.5539,
      "theoretical_loss": 3.571590122898125,
      "tokens_seen": 1259536384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031232698094282846,
      "loss": 2.7576,
      "theoretical_loss": 3.5715731842602025,
      "tokens_seen": 1259601920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123169508525577,
      "loss": 2.5159,
      "theoretical_loss": 3.57155624675031,
      "tokens_seen": 1259667456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123069207622869,
      "loss": 2.6873,
      "theoretical_loss": 3.5715393103683137,
      "tokens_seen": 1259732992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031229689067201606,
      "loss": 2.6168,
      "theoretical_loss": 3.57152237511408,
      "tokens_seen": 1259798528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122868605817453,
      "loss": 2.6619,
      "theoretical_loss": 3.571505440987475,
      "tokens_seen": 1259864064
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 696245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5429320335388184,
      "objective/train/theoretical_loss": 3.5714885079883643,
      "objective/train/tokens_used": 1280389600,
      "theoretical_loss": 3.5714885079883643,
      "tokens_seen": 1259929600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122768304914744,
      "loss": 2.7015,
      "theoretical_loss": 3.5714885079883643,
      "tokens_seen": 1259929600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031226680040120366,
      "loss": 2.5002,
      "theoretical_loss": 3.5714715761166147,
      "tokens_seen": 1259995136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122567703109328,
      "loss": 2.6852,
      "theoretical_loss": 3.571454645372093,
      "tokens_seen": 1260060672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312246740220662,
      "loss": 2.5585,
      "theoretical_loss": 3.5714377157546644,
      "tokens_seen": 1260126208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122367101303912,
      "loss": 2.587,
      "theoretical_loss": 3.5714207872641963,
      "tokens_seen": 1260191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122266800401204,
      "loss": 2.7699,
      "theoretical_loss": 3.5714038599005544,
      "tokens_seen": 1260257280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031221664994984956,
      "loss": 2.8506,
      "theoretical_loss": 3.571386933663606,
      "tokens_seen": 1260322816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031220661985957874,
      "loss": 2.8147,
      "theoretical_loss": 3.5713700085532167,
      "tokens_seen": 1260388352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121965897693079,
      "loss": 2.7021,
      "theoretical_loss": 3.5713530845692527,
      "tokens_seen": 1260453888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031218655967903716,
      "loss": 2.5115,
      "theoretical_loss": 3.5713361617115815,
      "tokens_seen": 1260519424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121765295887663,
      "loss": 2.513,
      "theoretical_loss": 3.5713192399800686,
      "tokens_seen": 1260584960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121664994984955,
      "loss": 2.5299,
      "theoretical_loss": 3.571302319374581,
      "tokens_seen": 1260650496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031215646940822465,
      "loss": 2.706,
      "theoretical_loss": 3.571285399894985,
      "tokens_seen": 1260716032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121464393179539,
      "loss": 2.607,
      "theoretical_loss": 3.571268481541148,
      "tokens_seen": 1260781568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031213640922768307,
      "loss": 2.7006,
      "theoretical_loss": 3.571251564312936,
      "tokens_seen": 1260847104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031212637913741225,
      "loss": 2.7955,
      "theoretical_loss": 3.571234648210215,
      "tokens_seen": 1260912640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031211634904714143,
      "loss": 2.6337,
      "theoretical_loss": 3.5712177332328525,
      "tokens_seen": 1260978176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031210631895687066,
      "loss": 2.7536,
      "theoretical_loss": 3.5712008193807145,
      "tokens_seen": 1261043712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120962888665998,
      "loss": 2.7754,
      "theoretical_loss": 3.571183906653668,
      "tokens_seen": 1261109248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312086258776329,
      "loss": 2.5408,
      "theoretical_loss": 3.57116699505158,
      "tokens_seen": 1261174784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031207622868605815,
      "loss": 2.6892,
      "theoretical_loss": 3.5711500845743167,
      "tokens_seen": 1261240320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120661985957874,
      "loss": 2.6009,
      "theoretical_loss": 3.5711331752217452,
      "tokens_seen": 1261305856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031205616850551657,
      "loss": 2.7901,
      "theoretical_loss": 3.571116266993732,
      "tokens_seen": 1261371392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031204613841524575,
      "loss": 2.5838,
      "theoretical_loss": 3.5710993598901446,
      "tokens_seen": 1261436928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031203610832497493,
      "loss": 2.6067,
      "theoretical_loss": 3.5710824539108486,
      "tokens_seen": 1261502464
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 696550,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.979339122772217,
      "objective/train/theoretical_loss": 3.571065549055712,
      "objective/train/tokens_used": 1282028000,
      "theoretical_loss": 3.571065549055712,
      "tokens_seen": 1261568000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120260782347041,
      "loss": 2.7769,
      "theoretical_loss": 3.571065549055712,
      "tokens_seen": 1261568000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120160481444333,
      "loss": 2.5907,
      "theoretical_loss": 3.5710486453246,
      "tokens_seen": 1261633536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031200601805416253,
      "loss": 2.6867,
      "theoretical_loss": 3.5710317427173814,
      "tokens_seen": 1261699072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031199598796389166,
      "loss": 2.6573,
      "theoretical_loss": 3.5710148412339224,
      "tokens_seen": 1261764608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003119859578736209,
      "loss": 2.7326,
      "theoretical_loss": 3.57099794087409,
      "tokens_seen": 1261830144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031197592778335,
      "loss": 2.7819,
      "theoretical_loss": 3.5709810416377508,
      "tokens_seen": 1261895680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031196589769307925,
      "loss": 2.6791,
      "theoretical_loss": 3.5709641435247717,
      "tokens_seen": 1261961216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031195586760280843,
      "loss": 2.8725,
      "theoretical_loss": 3.5709472465350203,
      "tokens_seen": 1262026752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003119458375125376,
      "loss": 2.5824,
      "theoretical_loss": 3.5709303506683625,
      "tokens_seen": 1262092288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003119358074222668,
      "loss": 2.5955,
      "theoretical_loss": 3.570913455924667,
      "tokens_seen": 1262157824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031192577733199603,
      "loss": 2.4743,
      "theoretical_loss": 3.5708965623037994,
      "tokens_seen": 1262223360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031191574724172516,
      "loss": 2.6973,
      "theoretical_loss": 3.5708796698056275,
      "tokens_seen": 1262288896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003119057171514544,
      "loss": 2.6142,
      "theoretical_loss": 3.570862778430018,
      "tokens_seen": 1262354432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118956870611835,
      "loss": 2.7967,
      "theoretical_loss": 3.570845888176839,
      "tokens_seen": 1262419968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031188565697091276,
      "loss": 2.6092,
      "theoretical_loss": 3.5708289990459563,
      "tokens_seen": 1262485504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031187562688064194,
      "loss": 2.7344,
      "theoretical_loss": 3.570812111037238,
      "tokens_seen": 1262551040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118655967903711,
      "loss": 2.567,
      "theoretical_loss": 3.570795224150551,
      "tokens_seen": 1262616576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118555667001003,
      "loss": 2.5363,
      "theoretical_loss": 3.5707783383857623,
      "tokens_seen": 1262682112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118455366098295,
      "loss": 2.8668,
      "theoretical_loss": 3.5707614537427395,
      "tokens_seen": 1262747648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031183550651955866,
      "loss": 2.7614,
      "theoretical_loss": 3.5707445702213496,
      "tokens_seen": 1262813184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118254764292879,
      "loss": 2.6856,
      "theoretical_loss": 3.57072768782146,
      "tokens_seen": 1262878720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311815446339017,
      "loss": 2.4951,
      "theoretical_loss": 3.5707108065429383,
      "tokens_seen": 1262944256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031180541624874626,
      "loss": 2.8183,
      "theoretical_loss": 3.570693926385651,
      "tokens_seen": 1263009792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117953861584754,
      "loss": 2.6523,
      "theoretical_loss": 3.570677047349467,
      "tokens_seen": 1263075328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117853560682046,
      "loss": 2.5798,
      "theoretical_loss": 3.570660169434252,
      "tokens_seen": 1263140864
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 697604,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7571280002593994,
      "objective/train/theoretical_loss": 3.5706432926398746,
      "objective/train/tokens_used": 1283666400,
      "theoretical_loss": 3.5706432926398746,
      "tokens_seen": 1263206400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117753259779338,
      "loss": 2.8312,
      "theoretical_loss": 3.5706432926398746,
      "tokens_seen": 1263206400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311765295887663,
      "loss": 2.6696,
      "theoretical_loss": 3.570626416966201,
      "tokens_seen": 1263271936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031175526579739217,
      "loss": 2.6625,
      "theoretical_loss": 3.5706095424130995,
      "tokens_seen": 1263337472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117452357071214,
      "loss": 2.6203,
      "theoretical_loss": 3.5705926689804377,
      "tokens_seen": 1263403008
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031173520561685053,
      "loss": 2.7463,
      "theoretical_loss": 3.5705757966680833,
      "tokens_seen": 1263468544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031172517552657976,
      "loss": 2.7992,
      "theoretical_loss": 3.5705589254759023,
      "tokens_seen": 1263534080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117151454363089,
      "loss": 2.6173,
      "theoretical_loss": 3.570542055403764,
      "tokens_seen": 1263599616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117051153460381,
      "loss": 2.4746,
      "theoretical_loss": 3.570525186451535,
      "tokens_seen": 1263665152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116950852557673,
      "loss": 2.8257,
      "theoretical_loss": 3.5705083186190834,
      "tokens_seen": 1263730688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116850551654965,
      "loss": 2.5915,
      "theoretical_loss": 3.5704914519062756,
      "tokens_seen": 1263796224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031167502507522567,
      "loss": 2.6664,
      "theoretical_loss": 3.570474586312981,
      "tokens_seen": 1263861760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031166499498495485,
      "loss": 2.8993,
      "theoretical_loss": 3.570457721839066,
      "tokens_seen": 1263927296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031165496489468403,
      "loss": 2.4212,
      "theoretical_loss": 3.5704408584843987,
      "tokens_seen": 1263992832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031164493480441327,
      "loss": 2.7775,
      "theoretical_loss": 3.570423996248847,
      "tokens_seen": 1264058368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116349047141424,
      "loss": 2.809,
      "theoretical_loss": 3.570407135132278,
      "tokens_seen": 1264123904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031162487462387163,
      "loss": 2.7512,
      "theoretical_loss": 3.57039027513456,
      "tokens_seen": 1264189440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116148445336008,
      "loss": 2.5596,
      "theoretical_loss": 3.5703734162555607,
      "tokens_seen": 1264254976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031160481444333,
      "loss": 2.6572,
      "theoretical_loss": 3.570356558495148,
      "tokens_seen": 1264320512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031159478435305917,
      "loss": 2.6785,
      "theoretical_loss": 3.570339701853189,
      "tokens_seen": 1264386048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031158475426278835,
      "loss": 2.6975,
      "theoretical_loss": 3.570322846329552,
      "tokens_seen": 1264451584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031157472417251753,
      "loss": 2.7055,
      "theoretical_loss": 3.570305991924105,
      "tokens_seen": 1264517120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031156469408224677,
      "loss": 2.5921,
      "theoretical_loss": 3.5702891386367157,
      "tokens_seen": 1264582656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031155466399197595,
      "loss": 2.7323,
      "theoretical_loss": 3.570272286467252,
      "tokens_seen": 1264648192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031154463390170513,
      "loss": 2.6489,
      "theoretical_loss": 3.5702554354155813,
      "tokens_seen": 1264713728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115346038114343,
      "loss": 2.6341,
      "theoretical_loss": 3.5702385854815724,
      "tokens_seen": 1264779264
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 698805,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0502676963806152,
      "objective/train/theoretical_loss": 3.5702217366650935,
      "objective/train/tokens_used": 1285304800,
      "theoretical_loss": 3.5702217366650935,
      "tokens_seen": 1264844800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115245737211635,
      "loss": 2.8163,
      "theoretical_loss": 3.5702217366650935,
      "tokens_seen": 1264844800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031151454363089273,
      "loss": 2.5645,
      "theoretical_loss": 3.5702048889660114,
      "tokens_seen": 1264910336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031150451354062186,
      "loss": 2.6297,
      "theoretical_loss": 3.5701880423841947,
      "tokens_seen": 1264975872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114944834503511,
      "loss": 2.7847,
      "theoretical_loss": 3.5701711969195116,
      "tokens_seen": 1265041408
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114844533600802,
      "loss": 2.6326,
      "theoretical_loss": 3.57015435257183,
      "tokens_seen": 1265106944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031147442326980945,
      "loss": 2.6998,
      "theoretical_loss": 3.570137509341018,
      "tokens_seen": 1265172480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031146439317953863,
      "loss": 2.6843,
      "theoretical_loss": 3.570120667226943,
      "tokens_seen": 1265238016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114543630892678,
      "loss": 2.8271,
      "theoretical_loss": 3.5701038262294746,
      "tokens_seen": 1265303552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311444332998997,
      "loss": 2.7197,
      "theoretical_loss": 3.57008698634848,
      "tokens_seen": 1265369088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031143430290872623,
      "loss": 2.5454,
      "theoretical_loss": 3.5700701475838277,
      "tokens_seen": 1265434624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031142427281845536,
      "loss": 2.8299,
      "theoretical_loss": 3.5700533099353855,
      "tokens_seen": 1265500160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114142427281846,
      "loss": 2.5524,
      "theoretical_loss": 3.5700364734030217,
      "tokens_seen": 1265565696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114042126379137,
      "loss": 2.6633,
      "theoretical_loss": 3.5700196379866043,
      "tokens_seen": 1265631232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031139418254764296,
      "loss": 2.6066,
      "theoretical_loss": 3.5700028036860028,
      "tokens_seen": 1265696768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031138415245737214,
      "loss": 2.5845,
      "theoretical_loss": 3.569985970501084,
      "tokens_seen": 1265762304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113741223671013,
      "loss": 2.5936,
      "theoretical_loss": 3.5699691384317163,
      "tokens_seen": 1265827840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113640922768305,
      "loss": 2.4984,
      "theoretical_loss": 3.5699523074777693,
      "tokens_seen": 1265893376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113540621865597,
      "loss": 2.7509,
      "theoretical_loss": 3.5699354776391097,
      "tokens_seen": 1265958912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031134403209628886,
      "loss": 2.6516,
      "theoretical_loss": 3.569918648915607,
      "tokens_seen": 1266024448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113340020060181,
      "loss": 2.6516,
      "theoretical_loss": 3.5699018213071296,
      "tokens_seen": 1266089984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113239719157472,
      "loss": 2.6194,
      "theoretical_loss": 3.5698849948135454,
      "tokens_seen": 1266155520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031131394182547646,
      "loss": 2.6923,
      "theoretical_loss": 3.5698681694347227,
      "tokens_seen": 1266221056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113039117352056,
      "loss": 2.6186,
      "theoretical_loss": 3.5698513451705307,
      "tokens_seen": 1266286592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112938816449348,
      "loss": 2.6444,
      "theoretical_loss": 3.5698345220208374,
      "tokens_seen": 1266352128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311283851554664,
      "loss": 3.0244,
      "theoretical_loss": 3.5698176999855114,
      "tokens_seen": 1266417664
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 699330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7510995864868164,
      "objective/train/theoretical_loss": 3.569800879064421,
      "objective/train/tokens_used": 1286943200,
      "theoretical_loss": 3.569800879064421,
      "tokens_seen": 1266483200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112738214643932,
      "loss": 2.556,
      "theoretical_loss": 3.569800879064421,
      "tokens_seen": 1266483200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031126379137412237,
      "loss": 2.7012,
      "theoretical_loss": 3.569784059257435,
      "tokens_seen": 1266548736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112537612838516,
      "loss": 2.6631,
      "theoretical_loss": 3.569767240564422,
      "tokens_seen": 1266614272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031124373119358073,
      "loss": 2.6916,
      "theoretical_loss": 3.5697504229852504,
      "tokens_seen": 1266679808
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031123370110330996,
      "loss": 2.9995,
      "theoretical_loss": 3.5697336065197884,
      "tokens_seen": 1266745344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112236710130391,
      "loss": 2.5175,
      "theoretical_loss": 3.569716791167906,
      "tokens_seen": 1266810880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112136409227683,
      "loss": 2.6245,
      "theoretical_loss": 3.56969997692947,
      "tokens_seen": 1266876416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112036108324975,
      "loss": 2.8514,
      "theoretical_loss": 3.569683163804351,
      "tokens_seen": 1266941952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111935807422267,
      "loss": 2.6541,
      "theoretical_loss": 3.5696663517924163,
      "tokens_seen": 1267007488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031118355065195587,
      "loss": 2.6908,
      "theoretical_loss": 3.5696495408935354,
      "tokens_seen": 1267073024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031117352056168505,
      "loss": 2.7746,
      "theoretical_loss": 3.569632731107576,
      "tokens_seen": 1267138560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031116349047141423,
      "loss": 2.5903,
      "theoretical_loss": 3.569615922434408,
      "tokens_seen": 1267204096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031115346038114347,
      "loss": 2.5927,
      "theoretical_loss": 3.5695991148739,
      "tokens_seen": 1267269632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111434302908726,
      "loss": 2.7901,
      "theoretical_loss": 3.5695823084259204,
      "tokens_seen": 1267335168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031113340020060183,
      "loss": 2.7658,
      "theoretical_loss": 3.569565503090338,
      "tokens_seen": 1267400704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311123370110331,
      "loss": 2.7178,
      "theoretical_loss": 3.569548698867022,
      "tokens_seen": 1267466240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111133400200602,
      "loss": 2.8511,
      "theoretical_loss": 3.5695318957558415,
      "tokens_seen": 1267531776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031110330992978937,
      "loss": 2.7355,
      "theoretical_loss": 3.5695150937566646,
      "tokens_seen": 1267597312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031109327983951855,
      "loss": 2.7006,
      "theoretical_loss": 3.569498292869361,
      "tokens_seen": 1267662848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031108324974924773,
      "loss": 2.5922,
      "theoretical_loss": 3.5694814930937993,
      "tokens_seen": 1267728384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031107321965897697,
      "loss": 2.6327,
      "theoretical_loss": 3.5694646944298487,
      "tokens_seen": 1267793920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110631895687061,
      "loss": 2.594,
      "theoretical_loss": 3.569447896877378,
      "tokens_seen": 1267859456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031105315947843533,
      "loss": 2.5477,
      "theoretical_loss": 3.5694311004362556,
      "tokens_seen": 1267924992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031104312938816446,
      "loss": 2.523,
      "theoretical_loss": 3.569414305106352,
      "tokens_seen": 1267990528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110330992978937,
      "loss": 2.7821,
      "theoretical_loss": 3.5693975108875344,
      "tokens_seen": 1268056064
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 700491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.37016224861145,
      "objective/train/theoretical_loss": 3.5693807177796737,
      "objective/train/tokens_used": 1288581600,
      "theoretical_loss": 3.5693807177796737,
      "tokens_seen": 1268121600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110230692076229,
      "loss": 2.6417,
      "theoretical_loss": 3.5693807177796737,
      "tokens_seen": 1268121600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031101303911735206,
      "loss": 2.6757,
      "theoretical_loss": 3.5693639257826377,
      "tokens_seen": 1268187136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031100300902708124,
      "loss": 2.5802,
      "theoretical_loss": 3.5693471348962964,
      "tokens_seen": 1268252672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109929789368104,
      "loss": 2.6605,
      "theoretical_loss": 3.569330345120518,
      "tokens_seen": 1268318208
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109829488465396,
      "loss": 2.6493,
      "theoretical_loss": 3.569313556455173,
      "tokens_seen": 1268383744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031097291875626883,
      "loss": 2.7038,
      "theoretical_loss": 3.569296768900129,
      "tokens_seen": 1268449280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031096288866599796,
      "loss": 2.4932,
      "theoretical_loss": 3.5692799824552566,
      "tokens_seen": 1268514816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109528585757272,
      "loss": 2.8002,
      "theoretical_loss": 3.569263197120424,
      "tokens_seen": 1268580352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109428284854564,
      "loss": 2.7297,
      "theoretical_loss": 3.569246412895501,
      "tokens_seen": 1268645888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031093279839518556,
      "loss": 2.7536,
      "theoretical_loss": 3.5692296297803567,
      "tokens_seen": 1268711424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031092276830491474,
      "loss": 2.7096,
      "theoretical_loss": 3.569212847774861,
      "tokens_seen": 1268776960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109127382146439,
      "loss": 2.5073,
      "theoretical_loss": 3.569196066878882,
      "tokens_seen": 1268842496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109027081243731,
      "loss": 2.6679,
      "theoretical_loss": 3.5691792870922905,
      "tokens_seen": 1268908032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031089267803410234,
      "loss": 2.5551,
      "theoretical_loss": 3.5691625084149545,
      "tokens_seen": 1268973568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031088264794383147,
      "loss": 2.7831,
      "theoretical_loss": 3.5691457308467442,
      "tokens_seen": 1269039104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108726178535607,
      "loss": 2.7506,
      "theoretical_loss": 3.569128954387529,
      "tokens_seen": 1269104640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031086258776328983,
      "loss": 2.8059,
      "theoretical_loss": 3.5691121790371776,
      "tokens_seen": 1269170176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031085255767301906,
      "loss": 2.5739,
      "theoretical_loss": 3.5690954047955605,
      "tokens_seen": 1269235712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031084252758274824,
      "loss": 2.4954,
      "theoretical_loss": 3.5690786316625465,
      "tokens_seen": 1269301248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108324974924774,
      "loss": 2.4753,
      "theoretical_loss": 3.5690618596380057,
      "tokens_seen": 1269366784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108224674022066,
      "loss": 2.819,
      "theoretical_loss": 3.569045088721807,
      "tokens_seen": 1269432320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108124373119358,
      "loss": 2.7176,
      "theoretical_loss": 3.5690283189138197,
      "tokens_seen": 1269497856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000310802407221665,
      "loss": 2.6721,
      "theoretical_loss": 3.569011550213914,
      "tokens_seen": 1269563392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107923771313942,
      "loss": 2.6428,
      "theoretical_loss": 3.5689947826219592,
      "tokens_seen": 1269628928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107823470411234,
      "loss": 2.5535,
      "theoretical_loss": 3.5689780161378257,
      "tokens_seen": 1269694464
    },
    {
      "epoch": 0.38,
      "objective/train/docs_used": 701131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.939049243927002,
      "objective/train/theoretical_loss": 3.568961250761382,
      "objective/train/tokens_used": 1290220000,
      "theoretical_loss": 3.568961250761382,
      "tokens_seen": 1269760000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031077231695085257,
      "loss": 2.8504,
      "theoretical_loss": 3.568961250761382,
      "tokens_seen": 1269760000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107622868605818,
      "loss": 2.6175,
      "theoretical_loss": 3.568944486492498,
      "tokens_seen": 1269825536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031075225677031093,
      "loss": 2.6228,
      "theoretical_loss": 3.5689277233310435,
      "tokens_seen": 1269891072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031074222668004016,
      "loss": 2.799,
      "theoretical_loss": 3.568910961276889,
      "tokens_seen": 1269956608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107321965897693,
      "loss": 2.7415,
      "theoretical_loss": 3.5688942003299027,
      "tokens_seen": 1270022144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107221664994985,
      "loss": 2.737,
      "theoretical_loss": 3.5688774404899553,
      "tokens_seen": 1270087680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107121364092277,
      "loss": 2.627,
      "theoretical_loss": 3.5688606817569166,
      "tokens_seen": 1270153216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107021063189569,
      "loss": 2.5978,
      "theoretical_loss": 3.5688439241306567,
      "tokens_seen": 1270218752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031069207622868607,
      "loss": 2.9248,
      "theoretical_loss": 3.568827167611044,
      "tokens_seen": 1270284288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031068204613841525,
      "loss": 2.8998,
      "theoretical_loss": 3.5688104121979496,
      "tokens_seen": 1270349824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031067201604814443,
      "loss": 2.6636,
      "theoretical_loss": 3.568793657891243,
      "tokens_seen": 1270415360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031066198595787367,
      "loss": 2.8076,
      "theoretical_loss": 3.568776904690794,
      "tokens_seen": 1270480896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106519558676028,
      "loss": 2.8176,
      "theoretical_loss": 3.5687601525964725,
      "tokens_seen": 1270546432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031064192577733203,
      "loss": 2.8466,
      "theoretical_loss": 3.5687434016081485,
      "tokens_seen": 1270611968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106318956870612,
      "loss": 2.7659,
      "theoretical_loss": 3.568726651725692,
      "tokens_seen": 1270677504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106218655967904,
      "loss": 2.9059,
      "theoretical_loss": 3.5687099029489726,
      "tokens_seen": 1270743040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031061183550651957,
      "loss": 2.8367,
      "theoretical_loss": 3.5686931552778605,
      "tokens_seen": 1270808576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031060180541624875,
      "loss": 2.6191,
      "theoretical_loss": 3.5686764087122262,
      "tokens_seen": 1270874112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031059177532597793,
      "loss": 2.4859,
      "theoretical_loss": 3.568659663251939,
      "tokens_seen": 1270939648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031058174523570717,
      "loss": 2.7966,
      "theoretical_loss": 3.5686429188968694,
      "tokens_seen": 1271005184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105717151454363,
      "loss": 2.9504,
      "theoretical_loss": 3.5686261756468873,
      "tokens_seen": 1271070720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031056168505516553,
      "loss": 2.4731,
      "theoretical_loss": 3.5686094335018623,
      "tokens_seen": 1271136256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031055165496489466,
      "loss": 2.5467,
      "theoretical_loss": 3.5685926924616655,
      "tokens_seen": 1271201792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105416248746239,
      "loss": 2.6595,
      "theoretical_loss": 3.568575952526167,
      "tokens_seen": 1271267328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105315947843531,
      "loss": 2.8641,
      "theoretical_loss": 3.568559213695236,
      "tokens_seen": 1271332864
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 702434,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5973994731903076,
      "objective/train/theoretical_loss": 3.5685424759687434,
      "objective/train/tokens_used": 1291858400,
      "theoretical_loss": 3.5685424759687434,
      "tokens_seen": 1271398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031052156469408226,
      "loss": 2.7256,
      "theoretical_loss": 3.5685424759687434,
      "tokens_seen": 1271398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031051153460381144,
      "loss": 2.5591,
      "theoretical_loss": 3.568525739346559,
      "tokens_seen": 1271463936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105015045135406,
      "loss": 2.6854,
      "theoretical_loss": 3.568509003828553,
      "tokens_seen": 1271529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104914744232698,
      "loss": 2.6827,
      "theoretical_loss": 3.5684922694145964,
      "tokens_seen": 1271595008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031048144433299904,
      "loss": 2.6534,
      "theoretical_loss": 3.568475536104559,
      "tokens_seen": 1271660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031047141424272816,
      "loss": 2.7149,
      "theoretical_loss": 3.5684588038983107,
      "tokens_seen": 1271726080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104613841524574,
      "loss": 2.5994,
      "theoretical_loss": 3.5684420727957225,
      "tokens_seen": 1271791616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104513540621866,
      "loss": 2.7789,
      "theoretical_loss": 3.568425342796664,
      "tokens_seen": 1271857152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031044132397191576,
      "loss": 2.471,
      "theoretical_loss": 3.568408613901006,
      "tokens_seen": 1271922688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031043129388164494,
      "loss": 2.7984,
      "theoretical_loss": 3.5683918861086186,
      "tokens_seen": 1271988224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104212637913741,
      "loss": 2.7841,
      "theoretical_loss": 3.5683751594193733,
      "tokens_seen": 1272053760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104112337011033,
      "loss": 2.802,
      "theoretical_loss": 3.568358433833139,
      "tokens_seen": 1272119296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031040120361083254,
      "loss": 2.744,
      "theoretical_loss": 3.568341709349787,
      "tokens_seen": 1272184832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031039117352056167,
      "loss": 2.6926,
      "theoretical_loss": 3.5683249859691872,
      "tokens_seen": 1272250368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103811434302909,
      "loss": 2.7375,
      "theoretical_loss": 3.568308263691211,
      "tokens_seen": 1272315904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031037111334002003,
      "loss": 2.63,
      "theoretical_loss": 3.568291542515728,
      "tokens_seen": 1272381440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031036108324974926,
      "loss": 2.523,
      "theoretical_loss": 3.5682748224426093,
      "tokens_seen": 1272446976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031035105315947844,
      "loss": 2.729,
      "theoretical_loss": 3.5682581034717256,
      "tokens_seen": 1272512512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103410230692076,
      "loss": 2.5598,
      "theoretical_loss": 3.5682413856029465,
      "tokens_seen": 1272578048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103309929789368,
      "loss": 2.7564,
      "theoretical_loss": 3.5682246688361436,
      "tokens_seen": 1272643584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000310320962888666,
      "loss": 2.6907,
      "theoretical_loss": 3.5682079531711866,
      "tokens_seen": 1272709120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031031093279839517,
      "loss": 2.4328,
      "theoretical_loss": 3.5681912386079473,
      "tokens_seen": 1272774656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103009027081244,
      "loss": 2.7671,
      "theoretical_loss": 3.568174525146296,
      "tokens_seen": 1272840192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031029087261785353,
      "loss": 2.6715,
      "theoretical_loss": 3.5681578127861027,
      "tokens_seen": 1272905728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031028084252758277,
      "loss": 2.7638,
      "theoretical_loss": 3.5681411015272384,
      "tokens_seen": 1272971264
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 703530,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.578533887863159,
      "objective/train/theoretical_loss": 3.568124391369574,
      "objective/train/tokens_used": 1293496800,
      "theoretical_loss": 3.568124391369574,
      "tokens_seen": 1273036800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031027081243731195,
      "loss": 2.65,
      "theoretical_loss": 3.568124391369574,
      "tokens_seen": 1273036800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031026078234704113,
      "loss": 2.8089,
      "theoretical_loss": 3.5681076823129807,
      "tokens_seen": 1273102336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102507522567703,
      "loss": 2.5821,
      "theoretical_loss": 3.568090974357329,
      "tokens_seen": 1273167872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102407221664995,
      "loss": 2.7659,
      "theoretical_loss": 3.568074267502489,
      "tokens_seen": 1273233408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031023069207622867,
      "loss": 2.8061,
      "theoretical_loss": 3.568057561748332,
      "tokens_seen": 1273298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102206619859579,
      "loss": 2.8214,
      "theoretical_loss": 3.5680408570947284,
      "tokens_seen": 1273364480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031021063189568703,
      "loss": 2.8284,
      "theoretical_loss": 3.5680241535415504,
      "tokens_seen": 1273430016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031020060180541627,
      "loss": 2.9509,
      "theoretical_loss": 3.5680074510886675,
      "tokens_seen": 1273495552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101905717151454,
      "loss": 2.5731,
      "theoretical_loss": 3.5679907497359515,
      "tokens_seen": 1273561088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031018054162487463,
      "loss": 2.4667,
      "theoretical_loss": 3.5679740494832726,
      "tokens_seen": 1273626624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101705115346038,
      "loss": 2.606,
      "theoretical_loss": 3.567957350330502,
      "tokens_seen": 1273692160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000310160481444333,
      "loss": 2.6988,
      "theoretical_loss": 3.5679406522775112,
      "tokens_seen": 1273757696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101504513540622,
      "loss": 2.7648,
      "theoretical_loss": 3.5679239553241704,
      "tokens_seen": 1273823232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101404212637914,
      "loss": 2.8896,
      "theoretical_loss": 3.567907259470351,
      "tokens_seen": 1273888768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031013039117352054,
      "loss": 2.6081,
      "theoretical_loss": 3.567890564715924,
      "tokens_seen": 1273954304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031012036108324977,
      "loss": 2.6571,
      "theoretical_loss": 3.5678738710607605,
      "tokens_seen": 1274019840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101103309929789,
      "loss": 2.8599,
      "theoretical_loss": 3.5678571785047315,
      "tokens_seen": 1274085376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031010030090270813,
      "loss": 2.8499,
      "theoretical_loss": 3.5678404870477083,
      "tokens_seen": 1274150912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100902708124373,
      "loss": 2.7101,
      "theoretical_loss": 3.5678237966895616,
      "tokens_seen": 1274216448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100802407221665,
      "loss": 2.5664,
      "theoretical_loss": 3.5678071074301627,
      "tokens_seen": 1274281984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100702106318957,
      "loss": 2.9353,
      "theoretical_loss": 3.5677904192693832,
      "tokens_seen": 1274347520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031006018054162486,
      "loss": 2.5256,
      "theoretical_loss": 3.567773732207094,
      "tokens_seen": 1274413056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100501504513541,
      "loss": 2.8448,
      "theoretical_loss": 3.567757046243166,
      "tokens_seen": 1274478592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100401203610833,
      "loss": 2.582,
      "theoretical_loss": 3.5677403613774703,
      "tokens_seen": 1274544128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031003009027081246,
      "loss": 2.7362,
      "theoretical_loss": 3.5677236776098793,
      "tokens_seen": 1274609664
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 704259,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.703928232192993,
      "objective/train/theoretical_loss": 3.567706994940263,
      "objective/train/tokens_used": 1295135200,
      "theoretical_loss": 3.567706994940263,
      "tokens_seen": 1274675200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031002006018054164,
      "loss": 2.5891,
      "theoretical_loss": 3.567706994940263,
      "tokens_seen": 1274675200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100100300902708,
      "loss": 2.5855,
      "theoretical_loss": 3.567690313368493,
      "tokens_seen": 1274740736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031,
      "loss": 2.7842,
      "theoretical_loss": 3.5676736328944414,
      "tokens_seen": 1274806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030998996990972924,
      "loss": 2.7119,
      "theoretical_loss": 3.5676569535179787,
      "tokens_seen": 1274871808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030997993981945836,
      "loss": 2.541,
      "theoretical_loss": 3.5676402752389764,
      "tokens_seen": 1274937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099699097291876,
      "loss": 2.5326,
      "theoretical_loss": 3.567623598057306,
      "tokens_seen": 1275002880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099598796389168,
      "loss": 2.8544,
      "theoretical_loss": 3.5676069219728395,
      "tokens_seen": 1275068416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030994984954864596,
      "loss": 2.8376,
      "theoretical_loss": 3.5675902469854472,
      "tokens_seen": 1275133952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030993981945837514,
      "loss": 2.7259,
      "theoretical_loss": 3.567573573095001,
      "tokens_seen": 1275199488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099297893681043,
      "loss": 2.4658,
      "theoretical_loss": 3.5675569003013727,
      "tokens_seen": 1275265024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003099197592778335,
      "loss": 2.5857,
      "theoretical_loss": 3.5675402286044333,
      "tokens_seen": 1275330560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030990972918756274,
      "loss": 2.9346,
      "theoretical_loss": 3.567523558004055,
      "tokens_seen": 1275396096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030989969909729187,
      "loss": 2.5889,
      "theoretical_loss": 3.5675068885001084,
      "tokens_seen": 1275461632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098896690070211,
      "loss": 2.7123,
      "theoretical_loss": 3.567490220092466,
      "tokens_seen": 1275527168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030987963891675023,
      "loss": 2.7163,
      "theoretical_loss": 3.5674735527809984,
      "tokens_seen": 1275592704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030986960882647946,
      "loss": 2.6532,
      "theoretical_loss": 3.5674568865655782,
      "tokens_seen": 1275658240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030985957873620864,
      "loss": 2.7162,
      "theoretical_loss": 3.567440221446076,
      "tokens_seen": 1275723776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098495486459378,
      "loss": 2.5943,
      "theoretical_loss": 3.5674235574223645,
      "tokens_seen": 1275789312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309839518555667,
      "loss": 2.8053,
      "theoretical_loss": 3.5674068944943147,
      "tokens_seen": 1275854848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098294884653962,
      "loss": 2.7737,
      "theoretical_loss": 3.5673902326617983,
      "tokens_seen": 1275920384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030981945837512537,
      "loss": 2.6332,
      "theoretical_loss": 3.567373571924687,
      "tokens_seen": 1275985920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098094282848546,
      "loss": 2.9085,
      "theoretical_loss": 3.567356912282853,
      "tokens_seen": 1276051456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030979939819458373,
      "loss": 2.4999,
      "theoretical_loss": 3.567340253736168,
      "tokens_seen": 1276116992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030978936810431297,
      "loss": 2.9444,
      "theoretical_loss": 3.5673235962845027,
      "tokens_seen": 1276182528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030977933801404215,
      "loss": 2.6067,
      "theoretical_loss": 3.56730693992773,
      "tokens_seen": 1276248064
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 705368,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.666820764541626,
      "objective/train/theoretical_loss": 3.5672902846657215,
      "objective/train/tokens_used": 1296773600,
      "theoretical_loss": 3.5672902846657215,
      "tokens_seen": 1276313600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030976930792377133,
      "loss": 2.7037,
      "theoretical_loss": 3.5672902846657215,
      "tokens_seen": 1276313600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097592778335005,
      "loss": 2.7744,
      "theoretical_loss": 3.567273630498349,
      "tokens_seen": 1276379136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097492477432297,
      "loss": 2.7129,
      "theoretical_loss": 3.567256977425484,
      "tokens_seen": 1276444672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030973921765295887,
      "loss": 2.5538,
      "theoretical_loss": 3.567240325446999,
      "tokens_seen": 1276510208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097291875626881,
      "loss": 2.8296,
      "theoretical_loss": 3.5672236745627655,
      "tokens_seen": 1276575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030971915747241723,
      "loss": 2.4411,
      "theoretical_loss": 3.5672070247726553,
      "tokens_seen": 1276641280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030970912738214647,
      "loss": 2.5156,
      "theoretical_loss": 3.5671903760765407,
      "tokens_seen": 1276706816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096990972918756,
      "loss": 2.86,
      "theoretical_loss": 3.5671737284742937,
      "tokens_seen": 1276772352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030968906720160483,
      "loss": 2.6477,
      "theoretical_loss": 3.567157081965786,
      "tokens_seen": 1276837888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309679037111334,
      "loss": 2.4785,
      "theoretical_loss": 3.567140436550889,
      "tokens_seen": 1276903424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096690070210632,
      "loss": 2.4694,
      "theoretical_loss": 3.5671237922294763,
      "tokens_seen": 1276968960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096589769307924,
      "loss": 2.5146,
      "theoretical_loss": 3.567107149001419,
      "tokens_seen": 1277034496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096489468405216,
      "loss": 2.5783,
      "theoretical_loss": 3.567090506866589,
      "tokens_seen": 1277100032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030963891675025074,
      "loss": 2.5938,
      "theoretical_loss": 3.5670738658248586,
      "tokens_seen": 1277165568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030962888665997997,
      "loss": 2.6706,
      "theoretical_loss": 3.5670572258761,
      "tokens_seen": 1277231104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096188565697091,
      "loss": 2.6656,
      "theoretical_loss": 3.567040587020186,
      "tokens_seen": 1277296640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030960882647943833,
      "loss": 2.7574,
      "theoretical_loss": 3.5670239492569866,
      "tokens_seen": 1277362176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095987963891675,
      "loss": 2.5667,
      "theoretical_loss": 3.5670073125863766,
      "tokens_seen": 1277427712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095887662988967,
      "loss": 2.6286,
      "theoretical_loss": 3.566990677008227,
      "tokens_seen": 1277493248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095787362086259,
      "loss": 2.6909,
      "theoretical_loss": 3.5669740425224097,
      "tokens_seen": 1277558784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030956870611835506,
      "loss": 2.5731,
      "theoretical_loss": 3.566957409128798,
      "tokens_seen": 1277624320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030955867602808424,
      "loss": 2.5758,
      "theoretical_loss": 3.5669407768272627,
      "tokens_seen": 1277689856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095486459378135,
      "loss": 2.5629,
      "theoretical_loss": 3.566924145617677,
      "tokens_seen": 1277755392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095386158475426,
      "loss": 2.7665,
      "theoretical_loss": 3.5669075154999135,
      "tokens_seen": 1277820928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030952858575727184,
      "loss": 2.7295,
      "theoretical_loss": 3.5668908864738436,
      "tokens_seen": 1277886464
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 705943,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5287697315216064,
      "objective/train/theoretical_loss": 3.5668742585393405,
      "objective/train/tokens_used": 1298412000,
      "theoretical_loss": 3.5668742585393405,
      "tokens_seen": 1277952000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030951855566700096,
      "loss": 2.7983,
      "theoretical_loss": 3.5668742585393405,
      "tokens_seen": 1277952000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095085255767302,
      "loss": 2.4844,
      "theoretical_loss": 3.5668576316962763,
      "tokens_seen": 1278017536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094984954864594,
      "loss": 2.3433,
      "theoretical_loss": 3.5668410059445232,
      "tokens_seen": 1278083072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030948846539618856,
      "loss": 2.8347,
      "theoretical_loss": 3.566824381283954,
      "tokens_seen": 1278148608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030947843530591774,
      "loss": 2.5863,
      "theoretical_loss": 3.5668077577144404,
      "tokens_seen": 1278214144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309468405215647,
      "loss": 2.6227,
      "theoretical_loss": 3.5667911352358557,
      "tokens_seen": 1278279680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094583751253761,
      "loss": 2.7285,
      "theoretical_loss": 3.566774513848072,
      "tokens_seen": 1278345216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030944834503510534,
      "loss": 2.7424,
      "theoretical_loss": 3.566757893550962,
      "tokens_seen": 1278410752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030943831494483447,
      "loss": 2.6665,
      "theoretical_loss": 3.5667412743443982,
      "tokens_seen": 1278476288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094282848545637,
      "loss": 2.5997,
      "theoretical_loss": 3.566724656228253,
      "tokens_seen": 1278541824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094182547642929,
      "loss": 2.5413,
      "theoretical_loss": 3.566708039202399,
      "tokens_seen": 1278607360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030940822467402207,
      "loss": 2.8092,
      "theoretical_loss": 3.5666914232667084,
      "tokens_seen": 1278672896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030939819458375125,
      "loss": 2.7759,
      "theoretical_loss": 3.5666748084210544,
      "tokens_seen": 1278738432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030938816449348043,
      "loss": 2.6478,
      "theoretical_loss": 3.566658194665309,
      "tokens_seen": 1278803968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093781344032096,
      "loss": 2.6607,
      "theoretical_loss": 3.566641581999346,
      "tokens_seen": 1278869504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030936810431293884,
      "loss": 2.7618,
      "theoretical_loss": 3.566624970423037,
      "tokens_seen": 1278935040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030935807422266797,
      "loss": 2.8331,
      "theoretical_loss": 3.566608359936255,
      "tokens_seen": 1279000576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093480441323972,
      "loss": 2.7793,
      "theoretical_loss": 3.5665917505388727,
      "tokens_seen": 1279066112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030933801404212633,
      "loss": 2.6624,
      "theoretical_loss": 3.5665751422307634,
      "tokens_seen": 1279131648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030932798395185557,
      "loss": 2.6765,
      "theoretical_loss": 3.566558535011799,
      "tokens_seen": 1279197184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030931795386158475,
      "loss": 2.6573,
      "theoretical_loss": 3.5665419288818527,
      "tokens_seen": 1279262720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030930792377131393,
      "loss": 2.6487,
      "theoretical_loss": 3.5665253238407972,
      "tokens_seen": 1279328256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030929789368104317,
      "loss": 2.7496,
      "theoretical_loss": 3.566508719888505,
      "tokens_seen": 1279393792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030928786359077235,
      "loss": 2.552,
      "theoretical_loss": 3.56649211702485,
      "tokens_seen": 1279459328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030927783350050153,
      "loss": 2.4734,
      "theoretical_loss": 3.5664755152497043,
      "tokens_seen": 1279524864
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 707215,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.725428342819214,
      "objective/train/theoretical_loss": 3.566458914562941,
      "objective/train/tokens_used": 1300050400,
      "theoretical_loss": 3.566458914562941,
      "tokens_seen": 1279590400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092678034102307,
      "loss": 2.6628,
      "theoretical_loss": 3.566458914562941,
      "tokens_seen": 1279590400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092577733199599,
      "loss": 2.5535,
      "theoretical_loss": 3.566442314964432,
      "tokens_seen": 1279655936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030924774322968907,
      "loss": 2.4807,
      "theoretical_loss": 3.566425716454052,
      "tokens_seen": 1279721472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092377131394183,
      "loss": 2.7158,
      "theoretical_loss": 3.5664091190316727,
      "tokens_seen": 1279787008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030922768304914743,
      "loss": 2.5899,
      "theoretical_loss": 3.5663925226971678,
      "tokens_seen": 1279852544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030921765295887667,
      "loss": 2.6971,
      "theoretical_loss": 3.5663759274504097,
      "tokens_seen": 1279918080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092076228686058,
      "loss": 2.5819,
      "theoretical_loss": 3.5663593332912713,
      "tokens_seen": 1279983616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030919759277833503,
      "loss": 2.5817,
      "theoretical_loss": 3.5663427402196266,
      "tokens_seen": 1280049152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091875626880642,
      "loss": 2.8216,
      "theoretical_loss": 3.5663261482353477,
      "tokens_seen": 1280114688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091775325977934,
      "loss": 2.8327,
      "theoretical_loss": 3.5663095573383083,
      "tokens_seen": 1280180224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091675025075226,
      "loss": 2.63,
      "theoretical_loss": 3.5662929675283808,
      "tokens_seen": 1280245760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091574724172518,
      "loss": 2.6224,
      "theoretical_loss": 3.5662763788054392,
      "tokens_seen": 1280311296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030914744232698094,
      "loss": 2.7127,
      "theoretical_loss": 3.5662597911693563,
      "tokens_seen": 1280376832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030913741223671017,
      "loss": 2.7571,
      "theoretical_loss": 3.5662432046200045,
      "tokens_seen": 1280442368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091273821464393,
      "loss": 2.5277,
      "theoretical_loss": 3.5662266191572582,
      "tokens_seen": 1280507904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030911735205616853,
      "loss": 2.8114,
      "theoretical_loss": 3.56621003478099,
      "tokens_seen": 1280573440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091073219658977,
      "loss": 2.8727,
      "theoretical_loss": 3.566193451491073,
      "tokens_seen": 1280638976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090972918756269,
      "loss": 2.7229,
      "theoretical_loss": 3.566176869287381,
      "tokens_seen": 1280704512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090872617853561,
      "loss": 2.7898,
      "theoretical_loss": 3.5661602881697867,
      "tokens_seen": 1280770048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030907723169508526,
      "loss": 2.5655,
      "theoretical_loss": 3.5661437081381635,
      "tokens_seen": 1280835584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030906720160481444,
      "loss": 2.7232,
      "theoretical_loss": 3.5661271291923855,
      "tokens_seen": 1280901120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090571715145437,
      "loss": 2.7766,
      "theoretical_loss": 3.566110551332325,
      "tokens_seen": 1280966656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090471414242728,
      "loss": 2.7331,
      "theoretical_loss": 3.5660939745578553,
      "tokens_seen": 1281032192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030903711133400204,
      "loss": 2.7573,
      "theoretical_loss": 3.5660773988688503,
      "tokens_seen": 1281097728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030902708124373116,
      "loss": 2.6697,
      "theoretical_loss": 3.5660608242651834,
      "tokens_seen": 1281163264
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 708432,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4614956378936768,
      "objective/train/theoretical_loss": 3.566044250746728,
      "objective/train/tokens_used": 1301688800,
      "theoretical_loss": 3.566044250746728,
      "tokens_seen": 1281228800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090170511534604,
      "loss": 2.6463,
      "theoretical_loss": 3.566044250746728,
      "tokens_seen": 1281228800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090070210631896,
      "loss": 2.6089,
      "theoretical_loss": 3.5660276783133575,
      "tokens_seen": 1281294336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030899699097291876,
      "loss": 2.7834,
      "theoretical_loss": 3.566011106964946,
      "tokens_seen": 1281359872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030898696088264794,
      "loss": 2.4896,
      "theoretical_loss": 3.565994536701365,
      "tokens_seen": 1281425408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089769307923772,
      "loss": 2.5439,
      "theoretical_loss": 3.56597796752249,
      "tokens_seen": 1281490944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089669007021063,
      "loss": 2.5781,
      "theoretical_loss": 3.5659613994281942,
      "tokens_seen": 1281556480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030895687061183554,
      "loss": 2.6482,
      "theoretical_loss": 3.5659448324183503,
      "tokens_seen": 1281622016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030894684052156467,
      "loss": 2.5109,
      "theoretical_loss": 3.5659282664928322,
      "tokens_seen": 1281687552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089368104312939,
      "loss": 2.7859,
      "theoretical_loss": 3.565911701651514,
      "tokens_seen": 1281753088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089267803410231,
      "loss": 2.5165,
      "theoretical_loss": 3.565895137894269,
      "tokens_seen": 1281818624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030891675025075227,
      "loss": 2.6697,
      "theoretical_loss": 3.565878575220971,
      "tokens_seen": 1281884160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030890672016048145,
      "loss": 2.5868,
      "theoretical_loss": 3.565862013631493,
      "tokens_seen": 1281949696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030889669007021063,
      "loss": 2.6787,
      "theoretical_loss": 3.5658454531257093,
      "tokens_seen": 1282015232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088866599799398,
      "loss": 2.6519,
      "theoretical_loss": 3.565828893703493,
      "tokens_seen": 1282080768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030887662988966904,
      "loss": 2.785,
      "theoretical_loss": 3.5658123353647193,
      "tokens_seen": 1282146304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030886659979939817,
      "loss": 2.6485,
      "theoretical_loss": 3.56579577810926,
      "tokens_seen": 1282211840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088565697091274,
      "loss": 2.519,
      "theoretical_loss": 3.56577922193699,
      "tokens_seen": 1282277376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030884653961885653,
      "loss": 2.7086,
      "theoretical_loss": 3.5657626668477826,
      "tokens_seen": 1282342912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030883650952858577,
      "loss": 2.7554,
      "theoretical_loss": 3.5657461128415124,
      "tokens_seen": 1282408448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030882647943831495,
      "loss": 2.4995,
      "theoretical_loss": 3.5657295599180525,
      "tokens_seen": 1282473984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030881644934804413,
      "loss": 2.4584,
      "theoretical_loss": 3.5657130080772763,
      "tokens_seen": 1282539520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088064192577733,
      "loss": 2.6642,
      "theoretical_loss": 3.565696457319059,
      "tokens_seen": 1282605056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030879638916750255,
      "loss": 2.6031,
      "theoretical_loss": 3.5656799076432737,
      "tokens_seen": 1282670592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087863590772317,
      "loss": 2.5193,
      "theoretical_loss": 3.565663359049794,
      "tokens_seen": 1282736128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087763289869609,
      "loss": 2.488,
      "theoretical_loss": 3.5656468115384943,
      "tokens_seen": 1282801664
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 708993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5532875061035156,
      "objective/train/theoretical_loss": 3.565630265109249,
      "objective/train/tokens_used": 1303327200,
      "theoretical_loss": 3.565630265109249,
      "tokens_seen": 1282867200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030876629889669004,
      "loss": 2.6164,
      "theoretical_loss": 3.565630265109249,
      "tokens_seen": 1282867200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030875626880641927,
      "loss": 2.2776,
      "theoretical_loss": 3.565613719761931,
      "tokens_seen": 1282932736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030874623871614845,
      "loss": 2.7188,
      "theoretical_loss": 3.565597175496414,
      "tokens_seen": 1282998272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030873620862587763,
      "loss": 2.7206,
      "theoretical_loss": 3.565580632312574,
      "tokens_seen": 1283063808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087261785356068,
      "loss": 2.6497,
      "theoretical_loss": 3.5655640902102834,
      "tokens_seen": 1283129344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308716148445336,
      "loss": 2.8397,
      "theoretical_loss": 3.565547549189417,
      "tokens_seen": 1283194880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087061183550652,
      "loss": 2.5893,
      "theoretical_loss": 3.565531009249848,
      "tokens_seen": 1283260416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086960882647944,
      "loss": 2.5696,
      "theoretical_loss": 3.565514470391452,
      "tokens_seen": 1283325952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030868605817452354,
      "loss": 2.4794,
      "theoretical_loss": 3.5654979326141016,
      "tokens_seen": 1283391488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086760280842528,
      "loss": 2.5306,
      "theoretical_loss": 3.5654813959176717,
      "tokens_seen": 1283457024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086659979939819,
      "loss": 2.5971,
      "theoretical_loss": 3.5654648603020362,
      "tokens_seen": 1283522560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030865596790371114,
      "loss": 2.49,
      "theoretical_loss": 3.5654483257670693,
      "tokens_seen": 1283588096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086459378134403,
      "loss": 2.5352,
      "theoretical_loss": 3.5654317923126455,
      "tokens_seen": 1283653632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086359077231695,
      "loss": 2.4462,
      "theoretical_loss": 3.565415259938639,
      "tokens_seen": 1283719168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086258776328987,
      "loss": 2.6472,
      "theoretical_loss": 3.565398728644923,
      "tokens_seen": 1283784704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086158475426279,
      "loss": 2.6344,
      "theoretical_loss": 3.565382198431374,
      "tokens_seen": 1283850240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030860581745235704,
      "loss": 2.6027,
      "theoretical_loss": 3.565365669297864,
      "tokens_seen": 1283915776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085957873620863,
      "loss": 2.706,
      "theoretical_loss": 3.565349141244268,
      "tokens_seen": 1283981312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085857572718154,
      "loss": 2.759,
      "theoretical_loss": 3.5653326142704618,
      "tokens_seen": 1284046848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030857572718154464,
      "loss": 2.7042,
      "theoretical_loss": 3.565316088376317,
      "tokens_seen": 1284112384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085656970912738,
      "loss": 2.8362,
      "theoretical_loss": 3.5652995635617106,
      "tokens_seen": 1284177920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308555667001003,
      "loss": 2.7046,
      "theoretical_loss": 3.5652830398265154,
      "tokens_seen": 1284243456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030854563691073224,
      "loss": 2.7368,
      "theoretical_loss": 3.5652665171706066,
      "tokens_seen": 1284308992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030853560682046137,
      "loss": 2.4158,
      "theoretical_loss": 3.565249995593858,
      "tokens_seen": 1284374528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085255767301906,
      "loss": 2.6283,
      "theoretical_loss": 3.5652334750961447,
      "tokens_seen": 1284440064
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 710075,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5692903995513916,
      "objective/train/theoretical_loss": 3.5652169556773403,
      "objective/train/tokens_used": 1304965600,
      "theoretical_loss": 3.5652169556773403,
      "tokens_seen": 1284505600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085155466399198,
      "loss": 2.6864,
      "theoretical_loss": 3.5652169556773403,
      "tokens_seen": 1284505600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030850551654964896,
      "loss": 2.8175,
      "theoretical_loss": 3.56520043733732,
      "tokens_seen": 1284571136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030849548645937814,
      "loss": 2.7832,
      "theoretical_loss": 3.5651839200759587,
      "tokens_seen": 1284636672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084854563691074,
      "loss": 2.6762,
      "theoretical_loss": 3.5651674038931302,
      "tokens_seen": 1284702208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084754262788365,
      "loss": 2.5926,
      "theoretical_loss": 3.5651508887887093,
      "tokens_seen": 1284767744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030846539618856574,
      "loss": 2.8683,
      "theoretical_loss": 3.5651343747625703,
      "tokens_seen": 1284833280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030845536609829487,
      "loss": 2.6109,
      "theoretical_loss": 3.5651178618145885,
      "tokens_seen": 1284898816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084453360080241,
      "loss": 2.4738,
      "theoretical_loss": 3.565101349944638,
      "tokens_seen": 1284964352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084353059177533,
      "loss": 2.624,
      "theoretical_loss": 3.5650848391525933,
      "tokens_seen": 1285029888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030842527582748247,
      "loss": 2.631,
      "theoretical_loss": 3.565068329438329,
      "tokens_seen": 1285095424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030841524573721165,
      "loss": 2.7388,
      "theoretical_loss": 3.565051820801721,
      "tokens_seen": 1285160960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030840521564694083,
      "loss": 2.6277,
      "theoretical_loss": 3.5650353132426424,
      "tokens_seen": 1285226496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030839518555667,
      "loss": 2.6851,
      "theoretical_loss": 3.5650188067609685,
      "tokens_seen": 1285292032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030838515546639924,
      "loss": 2.6007,
      "theoretical_loss": 3.5650023013565746,
      "tokens_seen": 1285357568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030837512537612837,
      "loss": 2.5723,
      "theoretical_loss": 3.564985797029335,
      "tokens_seen": 1285423104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083650952858576,
      "loss": 2.4376,
      "theoretical_loss": 3.5649692937791246,
      "tokens_seen": 1285488640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030835506519558673,
      "loss": 2.6476,
      "theoretical_loss": 3.564952791605818,
      "tokens_seen": 1285554176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030834503510531597,
      "loss": 2.5162,
      "theoretical_loss": 3.56493629050929,
      "tokens_seen": 1285619712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030833500501504515,
      "loss": 2.7669,
      "theoretical_loss": 3.564919790489416,
      "tokens_seen": 1285685248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030832497492477433,
      "loss": 2.7172,
      "theoretical_loss": 3.56490329154607,
      "tokens_seen": 1285750784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083149448345035,
      "loss": 2.6172,
      "theoretical_loss": 3.564886793679128,
      "tokens_seen": 1285816320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030830491474423275,
      "loss": 2.6885,
      "theoretical_loss": 3.564870296888464,
      "tokens_seen": 1285881856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082948846539619,
      "loss": 2.6611,
      "theoretical_loss": 3.5648538011739537,
      "tokens_seen": 1285947392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082848545636911,
      "loss": 2.6482,
      "theoretical_loss": 3.5648373065354715,
      "tokens_seen": 1286012928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030827482447342024,
      "loss": 2.8201,
      "theoretical_loss": 3.564820812972892,
      "tokens_seen": 1286078464
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 710665,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8468892574310303,
      "objective/train/theoretical_loss": 3.5648043204860915,
      "objective/train/tokens_used": 1306604000,
      "theoretical_loss": 3.5648043204860915,
      "tokens_seen": 1286144000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030826479438314947,
      "loss": 2.7577,
      "theoretical_loss": 3.5648043204860915,
      "tokens_seen": 1286144000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030825476429287865,
      "loss": 2.6774,
      "theoretical_loss": 3.5647878290749433,
      "tokens_seen": 1286209536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030824473420260783,
      "loss": 2.4722,
      "theoretical_loss": 3.564771338739324,
      "tokens_seen": 1286275072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308234704112337,
      "loss": 2.7426,
      "theoretical_loss": 3.564754849479108,
      "tokens_seen": 1286340608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082246740220662,
      "loss": 2.6997,
      "theoretical_loss": 3.56473836129417,
      "tokens_seen": 1286406144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082146439317954,
      "loss": 2.5188,
      "theoretical_loss": 3.564721874184386,
      "tokens_seen": 1286471680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082046138415246,
      "loss": 2.6416,
      "theoretical_loss": 3.5647053881496307,
      "tokens_seen": 1286537216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030819458375125374,
      "loss": 2.6886,
      "theoretical_loss": 3.564688903189779,
      "tokens_seen": 1286602752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308184553660983,
      "loss": 2.8147,
      "theoretical_loss": 3.5646724193047064,
      "tokens_seen": 1286668288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081745235707121,
      "loss": 2.6382,
      "theoretical_loss": 3.5646559364942876,
      "tokens_seen": 1286733824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030816449348044134,
      "loss": 2.6757,
      "theoretical_loss": 3.5646394547583986,
      "tokens_seen": 1286799360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081544633901705,
      "loss": 2.8823,
      "theoretical_loss": 3.5646229740969138,
      "tokens_seen": 1286864896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081444332998997,
      "loss": 2.6111,
      "theoretical_loss": 3.564606494509709,
      "tokens_seen": 1286930432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081344032096289,
      "loss": 2.7325,
      "theoretical_loss": 3.5645900159966595,
      "tokens_seen": 1286995968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081243731193581,
      "loss": 2.8163,
      "theoretical_loss": 3.5645735385576405,
      "tokens_seen": 1287061504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030811434302908724,
      "loss": 2.6988,
      "theoretical_loss": 3.5645570621925264,
      "tokens_seen": 1287127040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081043129388165,
      "loss": 2.5156,
      "theoretical_loss": 3.5645405869011944,
      "tokens_seen": 1287192576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080942828485456,
      "loss": 2.4571,
      "theoretical_loss": 3.5645241126835185,
      "tokens_seen": 1287258112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030808425275827484,
      "loss": 2.6575,
      "theoretical_loss": 3.564507639539374,
      "tokens_seen": 1287323648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308074222668004,
      "loss": 2.7198,
      "theoretical_loss": 3.564491167468637,
      "tokens_seen": 1287389184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080641925777332,
      "loss": 2.6747,
      "theoretical_loss": 3.5644746964711826,
      "tokens_seen": 1287454720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080541624874624,
      "loss": 2.4686,
      "theoretical_loss": 3.5644582265468863,
      "tokens_seen": 1287520256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030804413239719157,
      "loss": 2.5834,
      "theoretical_loss": 3.5644417576956235,
      "tokens_seen": 1287585792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030803410230692075,
      "loss": 2.755,
      "theoretical_loss": 3.5644252899172697,
      "tokens_seen": 1287651328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030802407221665,
      "loss": 2.8153,
      "theoretical_loss": 3.564408823211701,
      "tokens_seen": 1287716864
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 711678,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.727043628692627,
      "objective/train/theoretical_loss": 3.5643923575787912,
      "objective/train/tokens_used": 1308242400,
      "theoretical_loss": 3.5643923575787912,
      "tokens_seen": 1287782400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080140421263791,
      "loss": 2.6927,
      "theoretical_loss": 3.5643923575787912,
      "tokens_seen": 1287782400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030800401203610834,
      "loss": 2.5504,
      "theoretical_loss": 3.5643758930184175,
      "tokens_seen": 1287847936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079939819458375,
      "loss": 2.5323,
      "theoretical_loss": 3.564359429530455,
      "tokens_seen": 1287913472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079839518555667,
      "loss": 2.6744,
      "theoretical_loss": 3.564342967114779,
      "tokens_seen": 1287979008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079739217652959,
      "loss": 2.4857,
      "theoretical_loss": 3.564326505771265,
      "tokens_seen": 1288044544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030796389167502507,
      "loss": 2.5586,
      "theoretical_loss": 3.5643100454997896,
      "tokens_seen": 1288110080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030795386158475425,
      "loss": 2.8358,
      "theoretical_loss": 3.564293586300227,
      "tokens_seen": 1288175616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079438314944835,
      "loss": 2.7387,
      "theoretical_loss": 3.5642771281724546,
      "tokens_seen": 1288241152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079338014042126,
      "loss": 2.7444,
      "theoretical_loss": 3.5642606711163465,
      "tokens_seen": 1288306688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030792377131394185,
      "loss": 2.7622,
      "theoretical_loss": 3.5642442151317795,
      "tokens_seen": 1288372224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000307913741223671,
      "loss": 2.9028,
      "theoretical_loss": 3.5642277602186283,
      "tokens_seen": 1288437760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079037111334002,
      "loss": 2.5815,
      "theoretical_loss": 3.5642113063767695,
      "tokens_seen": 1288503296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078936810431294,
      "loss": 2.6371,
      "theoretical_loss": 3.564194853606079,
      "tokens_seen": 1288568832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030788365095285857,
      "loss": 2.5217,
      "theoretical_loss": 3.5641784019064318,
      "tokens_seen": 1288634368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030787362086258775,
      "loss": 2.7574,
      "theoretical_loss": 3.564161951277704,
      "tokens_seen": 1288699904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030786359077231693,
      "loss": 2.6885,
      "theoretical_loss": 3.564145501719772,
      "tokens_seen": 1288765440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078535606820461,
      "loss": 2.4649,
      "theoretical_loss": 3.5641290532325107,
      "tokens_seen": 1288830976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030784353059177535,
      "loss": 2.7864,
      "theoretical_loss": 3.5641126058157964,
      "tokens_seen": 1288896512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078335005015045,
      "loss": 2.7527,
      "theoretical_loss": 3.5640961594695053,
      "tokens_seen": 1288962048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078234704112337,
      "loss": 2.3971,
      "theoretical_loss": 3.564079714193513,
      "tokens_seen": 1289027584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078134403209629,
      "loss": 2.8447,
      "theoretical_loss": 3.5640632699876957,
      "tokens_seen": 1289093120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078034102306921,
      "loss": 2.5957,
      "theoretical_loss": 3.564046826851929,
      "tokens_seen": 1289158656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077933801404213,
      "loss": 2.8842,
      "theoretical_loss": 3.564030384786089,
      "tokens_seen": 1289224192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030778335005015044,
      "loss": 2.5096,
      "theoretical_loss": 3.564013943790052,
      "tokens_seen": 1289289728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030777331995987967,
      "loss": 2.8288,
      "theoretical_loss": 3.5639975038636935,
      "tokens_seen": 1289355264
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 712324,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.687561273574829,
      "objective/train/theoretical_loss": 3.5639810650068897,
      "objective/train/tokens_used": 1309880800,
      "theoretical_loss": 3.5639810650068897,
      "tokens_seen": 1289420800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030776328986960885,
      "loss": 2.7162,
      "theoretical_loss": 3.5639810650068897,
      "tokens_seen": 1289420800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030775325977933803,
      "loss": 2.6648,
      "theoretical_loss": 3.563964627219517,
      "tokens_seen": 1289486336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077432296890672,
      "loss": 2.4395,
      "theoretical_loss": 3.5639481905014514,
      "tokens_seen": 1289551872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077331995987964,
      "loss": 2.6302,
      "theoretical_loss": 3.5639317548525686,
      "tokens_seen": 1289617408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077231695085256,
      "loss": 2.6502,
      "theoretical_loss": 3.563915320272745,
      "tokens_seen": 1289682944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077131394182548,
      "loss": 2.604,
      "theoretical_loss": 3.563898886761857,
      "tokens_seen": 1289748480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030770310932798394,
      "loss": 2.6581,
      "theoretical_loss": 3.56388245431978,
      "tokens_seen": 1289814016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076930792377132,
      "loss": 2.9133,
      "theoretical_loss": 3.563866022946391,
      "tokens_seen": 1289879552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076830491474423,
      "loss": 2.7273,
      "theoretical_loss": 3.5638495926415654,
      "tokens_seen": 1289945088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030767301905717154,
      "loss": 2.6406,
      "theoretical_loss": 3.5638331634051807,
      "tokens_seen": 1290010624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076629889669007,
      "loss": 2.7489,
      "theoretical_loss": 3.5638167352371117,
      "tokens_seen": 1290076160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076529588766299,
      "loss": 2.6471,
      "theoretical_loss": 3.5638003081372354,
      "tokens_seen": 1290141696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076429287863591,
      "loss": 2.8173,
      "theoretical_loss": 3.563783882105428,
      "tokens_seen": 1290207232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076328986960883,
      "loss": 2.789,
      "theoretical_loss": 3.563767457141566,
      "tokens_seen": 1290272768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030762286860581744,
      "loss": 2.732,
      "theoretical_loss": 3.5637510332455253,
      "tokens_seen": 1290338304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076128385155467,
      "loss": 2.7352,
      "theoretical_loss": 3.5637346104171828,
      "tokens_seen": 1290403840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076028084252758,
      "loss": 2.6191,
      "theoretical_loss": 3.5637181886564147,
      "tokens_seen": 1290469376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030759277833500504,
      "loss": 2.5478,
      "theoretical_loss": 3.5637017679630967,
      "tokens_seen": 1290534912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075827482447342,
      "loss": 2.5346,
      "theoretical_loss": 3.563685348337106,
      "tokens_seen": 1290600448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075727181544634,
      "loss": 2.6512,
      "theoretical_loss": 3.5636689297783186,
      "tokens_seen": 1290665984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075626880641926,
      "loss": 2.564,
      "theoretical_loss": 3.5636525122866116,
      "tokens_seen": 1290731520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030755265797392177,
      "loss": 2.4989,
      "theoretical_loss": 3.5636360958618605,
      "tokens_seen": 1290797056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030754262788365095,
      "loss": 2.745,
      "theoretical_loss": 3.563619680503942,
      "tokens_seen": 1290862592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075325977933802,
      "loss": 2.8103,
      "theoretical_loss": 3.563603266212734,
      "tokens_seen": 1290928128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075225677031093,
      "loss": 2.5077,
      "theoretical_loss": 3.563586852988111,
      "tokens_seen": 1290993664
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 713206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9165244102478027,
      "objective/train/theoretical_loss": 3.563570440829951,
      "objective/train/tokens_used": 1311519200,
      "theoretical_loss": 3.563570440829951,
      "tokens_seen": 1291059200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030751253761283854,
      "loss": 2.7767,
      "theoretical_loss": 3.563570440829951,
      "tokens_seen": 1291059200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075025075225677,
      "loss": 2.5896,
      "theoretical_loss": 3.56355402973813,
      "tokens_seen": 1291124736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074924774322969,
      "loss": 2.5465,
      "theoretical_loss": 3.5635376197125246,
      "tokens_seen": 1291190272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074824473420261,
      "loss": 2.5913,
      "theoretical_loss": 3.5635212107530116,
      "tokens_seen": 1291255808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030747241725175527,
      "loss": 2.6442,
      "theoretical_loss": 3.563504802859468,
      "tokens_seen": 1291321344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030746238716148445,
      "loss": 2.7839,
      "theoretical_loss": 3.5634883960317687,
      "tokens_seen": 1291386880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074523570712137,
      "loss": 2.7687,
      "theoretical_loss": 3.5634719902697927,
      "tokens_seen": 1291452416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074423269809428,
      "loss": 2.8589,
      "theoretical_loss": 3.563455585573416,
      "tokens_seen": 1291517952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030743229689067205,
      "loss": 2.8238,
      "theoretical_loss": 3.563439181942514,
      "tokens_seen": 1291583488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074222668004012,
      "loss": 2.6259,
      "theoretical_loss": 3.563422779376965,
      "tokens_seen": 1291649024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074122367101304,
      "loss": 2.5173,
      "theoretical_loss": 3.563406377876645,
      "tokens_seen": 1291714560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074022066198596,
      "loss": 2.7689,
      "theoretical_loss": 3.563389977441431,
      "tokens_seen": 1291780096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030739217652958877,
      "loss": 2.5174,
      "theoretical_loss": 3.5633735780711997,
      "tokens_seen": 1291845632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030738214643931795,
      "loss": 2.3589,
      "theoretical_loss": 3.563357179765828,
      "tokens_seen": 1291911168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030737211634904713,
      "loss": 2.7557,
      "theoretical_loss": 3.563340782525193,
      "tokens_seen": 1291976704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073620862587763,
      "loss": 2.6461,
      "theoretical_loss": 3.563324386349171,
      "tokens_seen": 1292042240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030735205616850555,
      "loss": 2.5942,
      "theoretical_loss": 3.5633079912376395,
      "tokens_seen": 1292107776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073420260782347,
      "loss": 2.5485,
      "theoretical_loss": 3.563291597190475,
      "tokens_seen": 1292173312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073319959879639,
      "loss": 2.8998,
      "theoretical_loss": 3.5632752042075544,
      "tokens_seen": 1292238848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073219658976931,
      "loss": 2.4768,
      "theoretical_loss": 3.5632588122887547,
      "tokens_seen": 1292304384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073119358074223,
      "loss": 2.6437,
      "theoretical_loss": 3.5632424214339533,
      "tokens_seen": 1292369920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030730190571715146,
      "loss": 2.5855,
      "theoretical_loss": 3.563226031643026,
      "tokens_seen": 1292435456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030729187562688064,
      "loss": 2.9889,
      "theoretical_loss": 3.5632096429158517,
      "tokens_seen": 1292500992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072818455366098,
      "loss": 2.6217,
      "theoretical_loss": 3.5631932552523056,
      "tokens_seen": 1292566528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030727181544633905,
      "loss": 2.6415,
      "theoretical_loss": 3.5631768686522656,
      "tokens_seen": 1292632064
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 713898,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8086705207824707,
      "objective/train/theoretical_loss": 3.5631604831156087,
      "objective/train/tokens_used": 1313157600,
      "theoretical_loss": 3.5631604831156087,
      "tokens_seen": 1292697600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072617853560682,
      "loss": 2.9427,
      "theoretical_loss": 3.5631604831156087,
      "tokens_seen": 1292697600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072517552657974,
      "loss": 2.5816,
      "theoretical_loss": 3.563144098642212,
      "tokens_seen": 1292763136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030724172517552654,
      "loss": 2.5896,
      "theoretical_loss": 3.5631277152319525,
      "tokens_seen": 1292828672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072316950852558,
      "loss": 2.4494,
      "theoretical_loss": 3.5631113328847075,
      "tokens_seen": 1292894208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030722166499498496,
      "loss": 2.7717,
      "theoretical_loss": 3.563094951600354,
      "tokens_seen": 1292959744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030721163490471414,
      "loss": 2.4096,
      "theoretical_loss": 3.563078571378769,
      "tokens_seen": 1293025280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072016048144433,
      "loss": 2.6803,
      "theoretical_loss": 3.5630621922198302,
      "tokens_seen": 1293090816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071915747241725,
      "loss": 2.7954,
      "theoretical_loss": 3.5630458141234147,
      "tokens_seen": 1293156352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071815446339017,
      "loss": 2.4529,
      "theoretical_loss": 3.563029437089399,
      "tokens_seen": 1293221888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071715145436309,
      "loss": 2.5035,
      "theoretical_loss": 3.5630130611176614,
      "tokens_seen": 1293287424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030716148445336005,
      "loss": 2.6606,
      "theoretical_loss": 3.562996686208078,
      "tokens_seen": 1293352960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071514543630893,
      "loss": 2.4751,
      "theoretical_loss": 3.5629803123605273,
      "tokens_seen": 1293418496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030714142427281846,
      "loss": 2.621,
      "theoretical_loss": 3.562963939574886,
      "tokens_seen": 1293484032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030713139418254764,
      "loss": 2.6346,
      "theoretical_loss": 3.5629475678510314,
      "tokens_seen": 1293549568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071213640922768,
      "loss": 2.5764,
      "theoretical_loss": 3.562931197188841,
      "tokens_seen": 1293615104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000307111334002006,
      "loss": 2.8367,
      "theoretical_loss": 3.562914827588192,
      "tokens_seen": 1293680640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071013039117352,
      "loss": 2.569,
      "theoretical_loss": 3.562898459048962,
      "tokens_seen": 1293746176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070912738214644,
      "loss": 2.8707,
      "theoretical_loss": 3.562882091571028,
      "tokens_seen": 1293811712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030708124373119355,
      "loss": 2.5853,
      "theoretical_loss": 3.5628657251542686,
      "tokens_seen": 1293877248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070712136409228,
      "loss": 2.5337,
      "theoretical_loss": 3.5628493597985598,
      "tokens_seen": 1293942784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070611835506519,
      "loss": 2.8478,
      "theoretical_loss": 3.56283299550378,
      "tokens_seen": 1294008320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030705115346038115,
      "loss": 2.5145,
      "theoretical_loss": 3.562816632269806,
      "tokens_seen": 1294073856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070411233701104,
      "loss": 2.727,
      "theoretical_loss": 3.562800270096516,
      "tokens_seen": 1294139392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070310932798395,
      "loss": 2.8267,
      "theoretical_loss": 3.5627839089837874,
      "tokens_seen": 1294204928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030702106318956874,
      "loss": 2.6774,
      "theoretical_loss": 3.5627675489314976,
      "tokens_seen": 1294270464
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 715089,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8689498901367188,
      "objective/train/theoretical_loss": 3.562751189939524,
      "objective/train/tokens_used": 1314796000,
      "theoretical_loss": 3.562751189939524,
      "tokens_seen": 1294336000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070110330992979,
      "loss": 2.6044,
      "theoretical_loss": 3.562751189939524,
      "tokens_seen": 1294336000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070010030090271,
      "loss": 2.5002,
      "theoretical_loss": 3.5627348320077443,
      "tokens_seen": 1294401536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069909729187563,
      "loss": 2.6302,
      "theoretical_loss": 3.5627184751360366,
      "tokens_seen": 1294467072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030698094282848547,
      "loss": 2.6883,
      "theoretical_loss": 3.5627021193242774,
      "tokens_seen": 1294532608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030697091273821465,
      "loss": 2.6914,
      "theoretical_loss": 3.562685764572346,
      "tokens_seen": 1294598144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069608826479439,
      "loss": 2.6019,
      "theoretical_loss": 3.5626694108801185,
      "tokens_seen": 1294663680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306950852557673,
      "loss": 2.8642,
      "theoretical_loss": 3.5626530582474736,
      "tokens_seen": 1294729216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030694082246740225,
      "loss": 2.8976,
      "theoretical_loss": 3.5626367066742883,
      "tokens_seen": 1294794752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069307923771314,
      "loss": 2.6636,
      "theoretical_loss": 3.5626203561604415,
      "tokens_seen": 1294860288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069207622868606,
      "loss": 2.6443,
      "theoretical_loss": 3.5626040067058096,
      "tokens_seen": 1294925824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069107321965898,
      "loss": 2.8291,
      "theoretical_loss": 3.562587658310271,
      "tokens_seen": 1294991360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030690070210631897,
      "loss": 2.6902,
      "theoretical_loss": 3.562571310973704,
      "tokens_seen": 1295056896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030689067201604815,
      "loss": 2.6235,
      "theoretical_loss": 3.562554964695985,
      "tokens_seen": 1295122432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030688064192577733,
      "loss": 2.5321,
      "theoretical_loss": 3.5625386194769932,
      "tokens_seen": 1295187968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068706118355065,
      "loss": 2.5641,
      "theoretical_loss": 3.5625222753166064,
      "tokens_seen": 1295253504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030686058174523575,
      "loss": 2.625,
      "theoretical_loss": 3.5625059322147017,
      "tokens_seen": 1295319040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068505516549649,
      "loss": 2.8001,
      "theoretical_loss": 3.5624895901711575,
      "tokens_seen": 1295384576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068405215646941,
      "loss": 2.7228,
      "theoretical_loss": 3.5624732491858513,
      "tokens_seen": 1295450112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068304914744233,
      "loss": 2.7979,
      "theoretical_loss": 3.5624569092586618,
      "tokens_seen": 1295515648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068204613841525,
      "loss": 2.6856,
      "theoretical_loss": 3.5624405703894664,
      "tokens_seen": 1295581184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030681043129388166,
      "loss": 2.581,
      "theoretical_loss": 3.562424232578143,
      "tokens_seen": 1295646720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030680040120361084,
      "loss": 2.7789,
      "theoretical_loss": 3.56240789582457,
      "tokens_seen": 1295712256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030679037111334,
      "loss": 2.7325,
      "theoretical_loss": 3.562391560128625,
      "tokens_seen": 1295777792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030678034102306925,
      "loss": 2.6629,
      "theoretical_loss": 3.562375225490187,
      "tokens_seen": 1295843328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067703109327984,
      "loss": 2.5312,
      "theoretical_loss": 3.562358891909132,
      "tokens_seen": 1295908864
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 716367,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.913911819458008,
      "objective/train/theoretical_loss": 3.5623425593853404,
      "objective/train/tokens_used": 1316434400,
      "theoretical_loss": 3.5623425593853404,
      "tokens_seen": 1295974400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067602808425276,
      "loss": 2.6319,
      "theoretical_loss": 3.5623425593853404,
      "tokens_seen": 1295974400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030675025075225674,
      "loss": 2.3928,
      "theoretical_loss": 3.5623262279186894,
      "tokens_seen": 1296039936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306740220661986,
      "loss": 2.5186,
      "theoretical_loss": 3.5623098975090564,
      "tokens_seen": 1296105472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030673019057171516,
      "loss": 2.7353,
      "theoretical_loss": 3.5622935681563206,
      "tokens_seen": 1296171008
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030672016048144434,
      "loss": 2.7635,
      "theoretical_loss": 3.56227723986036,
      "tokens_seen": 1296236544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067101303911735,
      "loss": 2.6512,
      "theoretical_loss": 3.562260912621052,
      "tokens_seen": 1296302080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067001003009027,
      "loss": 2.6434,
      "theoretical_loss": 3.5622445864382755,
      "tokens_seen": 1296367616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066900702106319,
      "loss": 2.5974,
      "theoretical_loss": 3.562228261311909,
      "tokens_seen": 1296433152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066800401203611,
      "loss": 2.635,
      "theoretical_loss": 3.56221193724183,
      "tokens_seen": 1296498688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030667001003009025,
      "loss": 2.8283,
      "theoretical_loss": 3.562195614227917,
      "tokens_seen": 1296564224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066599799398195,
      "loss": 2.5637,
      "theoretical_loss": 3.5621792922700486,
      "tokens_seen": 1296629760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030664994984954866,
      "loss": 2.7018,
      "theoretical_loss": 3.5621629713681027,
      "tokens_seen": 1296695296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030663991975927784,
      "loss": 2.8643,
      "theoretical_loss": 3.562146651521958,
      "tokens_seen": 1296760832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306629889669007,
      "loss": 2.7117,
      "theoretical_loss": 3.5621303327314924,
      "tokens_seen": 1296826368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066198595787362,
      "loss": 2.7569,
      "theoretical_loss": 3.5621140149965846,
      "tokens_seen": 1296891904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066098294884654,
      "loss": 2.7215,
      "theoretical_loss": 3.562097698317113,
      "tokens_seen": 1296957440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065997993981946,
      "loss": 2.75,
      "theoretical_loss": 3.5620813826929556,
      "tokens_seen": 1297022976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030658976930792375,
      "loss": 2.512,
      "theoretical_loss": 3.562065068123992,
      "tokens_seen": 1297088512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306579739217653,
      "loss": 2.7242,
      "theoretical_loss": 3.562048754610099,
      "tokens_seen": 1297154048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065697091273821,
      "loss": 2.687,
      "theoretical_loss": 3.562032442151156,
      "tokens_seen": 1297219584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030655967903711135,
      "loss": 2.6452,
      "theoretical_loss": 3.5620161307470415,
      "tokens_seen": 1297285120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030654964894684053,
      "loss": 2.5351,
      "theoretical_loss": 3.561999820397634,
      "tokens_seen": 1297350656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065396188565697,
      "loss": 2.7883,
      "theoretical_loss": 3.5619835111028113,
      "tokens_seen": 1297416192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065295887662989,
      "loss": 2.719,
      "theoretical_loss": 3.561967202862453,
      "tokens_seen": 1297481728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065195586760281,
      "loss": 2.6211,
      "theoretical_loss": 3.5619508956764374,
      "tokens_seen": 1297547264
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 716974,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6259679794311523,
      "objective/train/theoretical_loss": 3.5619345895446424,
      "objective/train/tokens_used": 1318072800,
      "theoretical_loss": 3.5619345895446424,
      "tokens_seen": 1297612800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030650952858575725,
      "loss": 2.6675,
      "theoretical_loss": 3.5619345895446424,
      "tokens_seen": 1297612800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064994984954865,
      "loss": 2.767,
      "theoretical_loss": 3.5619182844669472,
      "tokens_seen": 1297678336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064894684052156,
      "loss": 2.5396,
      "theoretical_loss": 3.561901980443231,
      "tokens_seen": 1297743872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030647943831494485,
      "loss": 2.743,
      "theoretical_loss": 3.561885677473371,
      "tokens_seen": 1297809408
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030646940822467403,
      "loss": 2.6216,
      "theoretical_loss": 3.561869375557247,
      "tokens_seen": 1297874944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064593781344032,
      "loss": 2.7258,
      "theoretical_loss": 3.5618530746947368,
      "tokens_seen": 1297940480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064493480441324,
      "loss": 2.7407,
      "theoretical_loss": 3.5618367748857205,
      "tokens_seen": 1298006016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064393179538616,
      "loss": 2.7233,
      "theoretical_loss": 3.561820476130075,
      "tokens_seen": 1298071552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030642928786359076,
      "loss": 2.863,
      "theoretical_loss": 3.561804178427681,
      "tokens_seen": 1298137088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030641925777332,
      "loss": 2.7729,
      "theoretical_loss": 3.5617878817784154,
      "tokens_seen": 1298202624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064092276830491,
      "loss": 2.4632,
      "theoretical_loss": 3.5617715861821586,
      "tokens_seen": 1298268160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030639919759277835,
      "loss": 2.4926,
      "theoretical_loss": 3.5617552916387885,
      "tokens_seen": 1298333696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063891675025075,
      "loss": 2.8324,
      "theoretical_loss": 3.561738998148184,
      "tokens_seen": 1298399232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063791374122367,
      "loss": 2.6675,
      "theoretical_loss": 3.561722705710224,
      "tokens_seen": 1298464768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063691073219659,
      "loss": 2.7435,
      "theoretical_loss": 3.561706414324788,
      "tokens_seen": 1298530304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063590772316951,
      "loss": 2.4899,
      "theoretical_loss": 3.5616901239917533,
      "tokens_seen": 1298595840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030634904714142426,
      "loss": 2.8086,
      "theoretical_loss": 3.5616738347110006,
      "tokens_seen": 1298661376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063390170511535,
      "loss": 2.5463,
      "theoretical_loss": 3.561657546482408,
      "tokens_seen": 1298726912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063289869608826,
      "loss": 2.5267,
      "theoretical_loss": 3.561641259305855,
      "tokens_seen": 1298792448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030631895687061186,
      "loss": 2.592,
      "theoretical_loss": 3.5616249731812193,
      "tokens_seen": 1298857984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306308926780341,
      "loss": 2.571,
      "theoretical_loss": 3.5616086881083806,
      "tokens_seen": 1298923520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062988966900702,
      "loss": 2.6245,
      "theoretical_loss": 3.5615924040872184,
      "tokens_seen": 1298989056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030628886659979945,
      "loss": 2.5617,
      "theoretical_loss": 3.5615761211176116,
      "tokens_seen": 1299054592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062788365095286,
      "loss": 2.6409,
      "theoretical_loss": 3.5615598391994383,
      "tokens_seen": 1299120128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062688064192578,
      "loss": 2.5904,
      "theoretical_loss": 3.5615435583325787,
      "tokens_seen": 1299185664
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 718007,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9354186058044434,
      "objective/train/theoretical_loss": 3.561527278516911,
      "objective/train/tokens_used": 1319711200,
      "theoretical_loss": 3.561527278516911,
      "tokens_seen": 1299251200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030625877632898694,
      "loss": 2.6152,
      "theoretical_loss": 3.561527278516911,
      "tokens_seen": 1299251200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062487462387162,
      "loss": 2.5821,
      "theoretical_loss": 3.561510999752315,
      "tokens_seen": 1299316736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030623871614844536,
      "loss": 2.758,
      "theoretical_loss": 3.5614947220386695,
      "tokens_seen": 1299382272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030622868605817454,
      "loss": 2.6641,
      "theoretical_loss": 3.561478445375854,
      "tokens_seen": 1299447808
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062186559679037,
      "loss": 2.8013,
      "theoretical_loss": 3.5614621697637476,
      "tokens_seen": 1299513344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062086258776329,
      "loss": 2.7587,
      "theoretical_loss": 3.5614458952022288,
      "tokens_seen": 1299578880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061985957873621,
      "loss": 2.5621,
      "theoretical_loss": 3.561429621691177,
      "tokens_seen": 1299644416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061885656970913,
      "loss": 2.94,
      "theoretical_loss": 3.5614133492304725,
      "tokens_seen": 1299709952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030617853560682045,
      "loss": 2.7403,
      "theoretical_loss": 3.5613970778199935,
      "tokens_seen": 1299775488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061685055165497,
      "loss": 2.6433,
      "theoretical_loss": 3.5613808074596194,
      "tokens_seen": 1299841024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030615847542627886,
      "loss": 2.4314,
      "theoretical_loss": 3.5613645381492294,
      "tokens_seen": 1299906560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030614844533600804,
      "loss": 2.788,
      "theoretical_loss": 3.5613482698887036,
      "tokens_seen": 1299972096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061384152457372,
      "loss": 2.676,
      "theoretical_loss": 3.5613320026779203,
      "tokens_seen": 1300037632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061283851554664,
      "loss": 2.8276,
      "theoretical_loss": 3.5613157365167596,
      "tokens_seen": 1300103168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061183550651956,
      "loss": 2.2945,
      "theoretical_loss": 3.5612994714051007,
      "tokens_seen": 1300168704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061083249749248,
      "loss": 2.3596,
      "theoretical_loss": 3.5612832073428224,
      "tokens_seen": 1300234240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030609829488465395,
      "loss": 2.5757,
      "theoretical_loss": 3.5612669443298053,
      "tokens_seen": 1300299776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060882647943832,
      "loss": 2.6133,
      "theoretical_loss": 3.5612506823659276,
      "tokens_seen": 1300365312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060782347041123,
      "loss": 2.7672,
      "theoretical_loss": 3.5612344214510694,
      "tokens_seen": 1300430848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030606820461384155,
      "loss": 2.8412,
      "theoretical_loss": 3.56121816158511,
      "tokens_seen": 1300496384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030605817452357073,
      "loss": 2.555,
      "theoretical_loss": 3.561201902767929,
      "tokens_seen": 1300561920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060481444332999,
      "loss": 2.5812,
      "theoretical_loss": 3.561185644999406,
      "tokens_seen": 1300627456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060381143430291,
      "loss": 2.6127,
      "theoretical_loss": 3.5611693882794198,
      "tokens_seen": 1300692992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060280842527583,
      "loss": 2.6288,
      "theoretical_loss": 3.561153132607851,
      "tokens_seen": 1300758528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030601805416248745,
      "loss": 2.9774,
      "theoretical_loss": 3.5611368779845787,
      "tokens_seen": 1300824064
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 718956,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0698211193084717,
      "objective/train/theoretical_loss": 3.561120624409482,
      "objective/train/tokens_used": 1321349600,
      "theoretical_loss": 3.561120624409482,
      "tokens_seen": 1300889600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060080240722167,
      "loss": 2.5853,
      "theoretical_loss": 3.561120624409482,
      "tokens_seen": 1300889600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059979939819458,
      "loss": 2.6357,
      "theoretical_loss": 3.5611043718824416,
      "tokens_seen": 1300955136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030598796389167505,
      "loss": 2.7555,
      "theoretical_loss": 3.561088120403336,
      "tokens_seen": 1301020672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030597793380140423,
      "loss": 2.7744,
      "theoretical_loss": 3.561071869972046,
      "tokens_seen": 1301086208
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059679037111334,
      "loss": 2.5788,
      "theoretical_loss": 3.5610556205884505,
      "tokens_seen": 1301151744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059578736208626,
      "loss": 2.7153,
      "theoretical_loss": 3.561039372252429,
      "tokens_seen": 1301217280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059478435305918,
      "loss": 2.5183,
      "theoretical_loss": 3.561023124963861,
      "tokens_seen": 1301282816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030593781344032096,
      "loss": 2.8893,
      "theoretical_loss": 3.561006878722628,
      "tokens_seen": 1301348352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059277833500502,
      "loss": 2.8397,
      "theoretical_loss": 3.5609906335286077,
      "tokens_seen": 1301413888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003059177532597793,
      "loss": 2.7949,
      "theoretical_loss": 3.5609743893816805,
      "tokens_seen": 1301479424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030590772316950855,
      "loss": 2.8534,
      "theoretical_loss": 3.560958146281727,
      "tokens_seen": 1301544960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058976930792377,
      "loss": 2.5166,
      "theoretical_loss": 3.560941904228626,
      "tokens_seen": 1301610496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058876629889669,
      "loss": 2.7456,
      "theoretical_loss": 3.5609256632222577,
      "tokens_seen": 1301676032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058776328986961,
      "loss": 2.9513,
      "theoretical_loss": 3.560909423262502,
      "tokens_seen": 1301741568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058676028084253,
      "loss": 2.42,
      "theoretical_loss": 3.560893184349239,
      "tokens_seen": 1301807104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030585757271815446,
      "loss": 2.6008,
      "theoretical_loss": 3.560876946482348,
      "tokens_seen": 1301872640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058475426278837,
      "loss": 2.5304,
      "theoretical_loss": 3.560860709661709,
      "tokens_seen": 1301938176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058375125376128,
      "loss": 2.8498,
      "theoretical_loss": 3.5608444738872027,
      "tokens_seen": 1302003712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030582748244734206,
      "loss": 2.7499,
      "theoretical_loss": 3.560828239158708,
      "tokens_seen": 1302069248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058174523570712,
      "loss": 2.9477,
      "theoretical_loss": 3.560812005476105,
      "tokens_seen": 1302134784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058074222668004,
      "loss": 2.5286,
      "theoretical_loss": 3.560795772839275,
      "tokens_seen": 1302200320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057973921765296,
      "loss": 2.7829,
      "theoretical_loss": 3.5607795412480963,
      "tokens_seen": 1302265856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057873620862588,
      "loss": 2.7022,
      "theoretical_loss": 3.56076331070245,
      "tokens_seen": 1302331392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030577733199598796,
      "loss": 2.7559,
      "theoretical_loss": 3.5607470812022157,
      "tokens_seen": 1302396928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030576730190571714,
      "loss": 2.6967,
      "theoretical_loss": 3.5607308527472736,
      "tokens_seen": 1302462464
    },
    {
      "epoch": 0.39,
      "objective/train/docs_used": 719575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2579822540283203,
      "objective/train/theoretical_loss": 3.560714625337504,
      "objective/train/tokens_used": 1322988000,
      "theoretical_loss": 3.560714625337504,
      "tokens_seen": 1302528000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057572718154463,
      "loss": 2.7297,
      "theoretical_loss": 3.560714625337504,
      "tokens_seen": 1302528000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030574724172517556,
      "loss": 2.74,
      "theoretical_loss": 3.5606983989727863,
      "tokens_seen": 1302593536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057372116349047,
      "loss": 2.5027,
      "theoretical_loss": 3.560682173653001,
      "tokens_seen": 1302659072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057271815446339,
      "loss": 2.5717,
      "theoretical_loss": 3.560665949378029,
      "tokens_seen": 1302724608
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030571715145436305,
      "loss": 2.5871,
      "theoretical_loss": 3.560649726147749,
      "tokens_seen": 1302790144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057071213640923,
      "loss": 2.6637,
      "theoretical_loss": 3.560633503962043,
      "tokens_seen": 1302855680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030569709127382147,
      "loss": 2.7712,
      "theoretical_loss": 3.560617282820789,
      "tokens_seen": 1302921216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030568706118355065,
      "loss": 2.5889,
      "theoretical_loss": 3.560601062723869,
      "tokens_seen": 1302986752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030567703109327983,
      "loss": 2.6207,
      "theoretical_loss": 3.5605848436711627,
      "tokens_seen": 1303052288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030566700100300906,
      "loss": 2.6828,
      "theoretical_loss": 3.56056862566255,
      "tokens_seen": 1303117824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056569709127382,
      "loss": 2.842,
      "theoretical_loss": 3.5605524086979123,
      "tokens_seen": 1303183360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056469408224674,
      "loss": 2.6398,
      "theoretical_loss": 3.5605361927771284,
      "tokens_seen": 1303248896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030563691073219655,
      "loss": 2.6295,
      "theoretical_loss": 3.5605199779000793,
      "tokens_seen": 1303314432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056268806419258,
      "loss": 2.1334,
      "theoretical_loss": 3.5605037640666457,
      "tokens_seen": 1303379968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030561685055165497,
      "loss": 2.3651,
      "theoretical_loss": 3.560487551276707,
      "tokens_seen": 1303445504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030560682046138415,
      "loss": 2.4331,
      "theoretical_loss": 3.560471339530145,
      "tokens_seen": 1303511040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030559679037111333,
      "loss": 2.7634,
      "theoretical_loss": 3.560455128826839,
      "tokens_seen": 1303576576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055867602808425,
      "loss": 2.8345,
      "theoretical_loss": 3.5604389191666694,
      "tokens_seen": 1303642112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055767301905717,
      "loss": 2.6223,
      "theoretical_loss": 3.5604227105495174,
      "tokens_seen": 1303707648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030556670010030093,
      "loss": 2.6005,
      "theoretical_loss": 3.5604065029752627,
      "tokens_seen": 1303773184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030555667001003006,
      "loss": 2.8058,
      "theoretical_loss": 3.560390296443787,
      "tokens_seen": 1303838720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055466399197593,
      "loss": 2.5682,
      "theoretical_loss": 3.5603740909549684,
      "tokens_seen": 1303904256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055366098294885,
      "loss": 2.6851,
      "theoretical_loss": 3.5603578865086902,
      "tokens_seen": 1303969792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030552657973921765,
      "loss": 2.783,
      "theoretical_loss": 3.560341683104831,
      "tokens_seen": 1304035328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055165496489469,
      "loss": 2.6548,
      "theoretical_loss": 3.560325480743272,
      "tokens_seen": 1304100864
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 720457,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.661524534225464,
      "objective/train/theoretical_loss": 3.560309279423894,
      "objective/train/tokens_used": 1324626400,
      "theoretical_loss": 3.560309279423894,
      "tokens_seen": 1304166400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305506519558676,
      "loss": 2.7846,
      "theoretical_loss": 3.560309279423894,
      "tokens_seen": 1304166400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030549648946840525,
      "loss": 2.6931,
      "theoretical_loss": 3.560293079146578,
      "tokens_seen": 1304231936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030548645937813443,
      "loss": 2.7123,
      "theoretical_loss": 3.560276879911203,
      "tokens_seen": 1304297472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054764292878636,
      "loss": 2.546,
      "theoretical_loss": 3.560260681717651,
      "tokens_seen": 1304363008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054663991975928,
      "loss": 2.5774,
      "theoretical_loss": 3.5602444845658026,
      "tokens_seen": 1304428544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305456369107322,
      "loss": 2.6763,
      "theoretical_loss": 3.560228288455538,
      "tokens_seen": 1304494080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030544633901705116,
      "loss": 2.6837,
      "theoretical_loss": 3.5602120933867383,
      "tokens_seen": 1304559616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054363089267804,
      "loss": 2.699,
      "theoretical_loss": 3.5601958993592833,
      "tokens_seen": 1304625152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054262788365095,
      "loss": 2.5079,
      "theoretical_loss": 3.560179706373055,
      "tokens_seen": 1304690688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030541624874623875,
      "loss": 2.7306,
      "theoretical_loss": 3.5601635144279333,
      "tokens_seen": 1304756224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054062186559679,
      "loss": 2.712,
      "theoretical_loss": 3.5601473235237995,
      "tokens_seen": 1304821760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053961885656971,
      "loss": 2.6428,
      "theoretical_loss": 3.5601311336605344,
      "tokens_seen": 1304887296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053861584754263,
      "loss": 2.6081,
      "theoretical_loss": 3.560114944838018,
      "tokens_seen": 1304952832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053761283851555,
      "loss": 2.6574,
      "theoretical_loss": 3.5600987570561324,
      "tokens_seen": 1305018368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030536609829488466,
      "loss": 2.8607,
      "theoretical_loss": 3.5600825703147576,
      "tokens_seen": 1305083904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053560682046139,
      "loss": 2.6453,
      "theoretical_loss": 3.560066384613774,
      "tokens_seen": 1305149440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305346038114343,
      "loss": 2.7926,
      "theoretical_loss": 3.560050199953064,
      "tokens_seen": 1305214976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030533600802407226,
      "loss": 2.7229,
      "theoretical_loss": 3.560034016332507,
      "tokens_seen": 1305280512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053259779338014,
      "loss": 2.6756,
      "theoretical_loss": 3.560017833751985,
      "tokens_seen": 1305346048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053159478435306,
      "loss": 2.7556,
      "theoretical_loss": 3.560001652211378,
      "tokens_seen": 1305411584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053059177532598,
      "loss": 2.5928,
      "theoretical_loss": 3.559985471710568,
      "tokens_seen": 1305477120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305295887662989,
      "loss": 2.6948,
      "theoretical_loss": 3.5599692922494355,
      "tokens_seen": 1305542656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030528585757271816,
      "loss": 2.6325,
      "theoretical_loss": 3.559953113827861,
      "tokens_seen": 1305608192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030527582748244734,
      "loss": 2.7057,
      "theoretical_loss": 3.5599369364457267,
      "tokens_seen": 1305673728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052657973921765,
      "loss": 2.8171,
      "theoretical_loss": 3.559920760102912,
      "tokens_seen": 1305739264
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 721070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7218596935272217,
      "objective/train/theoretical_loss": 3.5599045847993,
      "objective/train/tokens_used": 1326264800,
      "theoretical_loss": 3.5599045847993,
      "tokens_seen": 1305804800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030525576730190576,
      "loss": 2.6822,
      "theoretical_loss": 3.5599045847993,
      "tokens_seen": 1305804800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052457372116349,
      "loss": 2.6014,
      "theoretical_loss": 3.5598884105347706,
      "tokens_seen": 1305870336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052357071213641,
      "loss": 2.7997,
      "theoretical_loss": 3.5598722373092047,
      "tokens_seen": 1305935872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030522567703109325,
      "loss": 2.6595,
      "theoretical_loss": 3.5598560651224838,
      "tokens_seen": 1306001408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052156469408225,
      "loss": 2.5854,
      "theoretical_loss": 3.5598398939744893,
      "tokens_seen": 1306066944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030520561685055167,
      "loss": 2.6932,
      "theoretical_loss": 3.559823723865102,
      "tokens_seen": 1306132480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030519558676028085,
      "loss": 2.6856,
      "theoretical_loss": 3.5598075547942027,
      "tokens_seen": 1306198016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030518555667001003,
      "loss": 2.7419,
      "theoretical_loss": 3.5597913867616735,
      "tokens_seen": 1306263552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030517552657973926,
      "loss": 2.5488,
      "theoretical_loss": 3.559775219767395,
      "tokens_seen": 1306329088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051654964894684,
      "loss": 2.6137,
      "theoretical_loss": 3.559759053811249,
      "tokens_seen": 1306394624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051554663991976,
      "loss": 2.8409,
      "theoretical_loss": 3.5597428888931164,
      "tokens_seen": 1306460160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030514543630892675,
      "loss": 2.5936,
      "theoretical_loss": 3.559726725012878,
      "tokens_seen": 1306525696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305135406218656,
      "loss": 2.612,
      "theoretical_loss": 3.559710562170416,
      "tokens_seen": 1306591232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030512537612838517,
      "loss": 3.0133,
      "theoretical_loss": 3.5596944003656112,
      "tokens_seen": 1306656768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030511534603811435,
      "loss": 2.7824,
      "theoretical_loss": 3.5596782395983455,
      "tokens_seen": 1306722304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030510531594784353,
      "loss": 2.6491,
      "theoretical_loss": 3.5596620798684997,
      "tokens_seen": 1306787840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050952858575727,
      "loss": 2.6621,
      "theoretical_loss": 3.559645921175955,
      "tokens_seen": 1306853376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050852557673019,
      "loss": 2.5225,
      "theoretical_loss": 3.5596297635205927,
      "tokens_seen": 1306918912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030507522567703113,
      "loss": 2.5159,
      "theoretical_loss": 3.5596136069022952,
      "tokens_seen": 1306984448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030506519558676026,
      "loss": 2.6279,
      "theoretical_loss": 3.5595974513209434,
      "tokens_seen": 1307049984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050551654964895,
      "loss": 2.8104,
      "theoretical_loss": 3.559581296776419,
      "tokens_seen": 1307115520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050451354062186,
      "loss": 2.6946,
      "theoretical_loss": 3.5595651432686033,
      "tokens_seen": 1307181056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030503510531594785,
      "loss": 2.7261,
      "theoretical_loss": 3.559548990797377,
      "tokens_seen": 1307246592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030502507522567703,
      "loss": 2.6394,
      "theoretical_loss": 3.559532839362623,
      "tokens_seen": 1307312128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050150451354062,
      "loss": 2.6643,
      "theoretical_loss": 3.5595166889642216,
      "tokens_seen": 1307377664
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 722360,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.437825918197632,
      "objective/train/theoretical_loss": 3.5595005396020554,
      "objective/train/tokens_used": 1327903200,
      "theoretical_loss": 3.5595005396020554,
      "tokens_seen": 1307443200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050050150451354,
      "loss": 2.7504,
      "theoretical_loss": 3.5595005396020554,
      "tokens_seen": 1307443200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030499498495486463,
      "loss": 2.8347,
      "theoretical_loss": 3.5594843912760052,
      "tokens_seen": 1307508736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030498495486459376,
      "loss": 2.717,
      "theoretical_loss": 3.559468243985953,
      "tokens_seen": 1307574272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304974924774323,
      "loss": 2.6803,
      "theoretical_loss": 3.5594520977317803,
      "tokens_seen": 1307639808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049648946840521,
      "loss": 2.6731,
      "theoretical_loss": 3.559435952513369,
      "tokens_seen": 1307705344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030495486459378136,
      "loss": 2.6494,
      "theoretical_loss": 3.5594198083306003,
      "tokens_seen": 1307770880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030494483450351054,
      "loss": 2.656,
      "theoretical_loss": 3.559403665183356,
      "tokens_seen": 1307836416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049348044132397,
      "loss": 2.803,
      "theoretical_loss": 3.559387523071518,
      "tokens_seen": 1307901952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049247743229689,
      "loss": 2.7034,
      "theoretical_loss": 3.559371381994968,
      "tokens_seen": 1307967488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003049147442326981,
      "loss": 2.6816,
      "theoretical_loss": 3.559355241953588,
      "tokens_seen": 1308033024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030490471414242726,
      "loss": 2.7026,
      "theoretical_loss": 3.5593391029472587,
      "tokens_seen": 1308098560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048946840521565,
      "loss": 2.7462,
      "theoretical_loss": 3.5593229649758626,
      "tokens_seen": 1308164096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048846539618856,
      "loss": 2.7407,
      "theoretical_loss": 3.559306828039282,
      "tokens_seen": 1308229632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030487462387161486,
      "loss": 2.5191,
      "theoretical_loss": 3.559290692137398,
      "tokens_seen": 1308295168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304864593781344,
      "loss": 2.8017,
      "theoretical_loss": 3.5592745572700917,
      "tokens_seen": 1308360704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048545636910732,
      "loss": 2.6502,
      "theoretical_loss": 3.559258423437247,
      "tokens_seen": 1308426240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048445336008024,
      "loss": 2.6244,
      "theoretical_loss": 3.559242290638744,
      "tokens_seen": 1308491776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048345035105316,
      "loss": 2.5927,
      "theoretical_loss": 3.5592261588744654,
      "tokens_seen": 1308557312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030482447342026076,
      "loss": 2.4246,
      "theoretical_loss": 3.5592100281442924,
      "tokens_seen": 1308622848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030481444332999,
      "loss": 2.698,
      "theoretical_loss": 3.5591938984481084,
      "tokens_seen": 1308688384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048044132397192,
      "loss": 2.8172,
      "theoretical_loss": 3.5591777697857934,
      "tokens_seen": 1308753920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030479438314944836,
      "loss": 2.7269,
      "theoretical_loss": 3.559161642157231,
      "tokens_seen": 1308819456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030478435305917754,
      "loss": 2.7736,
      "theoretical_loss": 3.559145515562302,
      "tokens_seen": 1308884992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047743229689067,
      "loss": 2.312,
      "theoretical_loss": 3.5591293900008893,
      "tokens_seen": 1308950528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030476429287863596,
      "loss": 2.5476,
      "theoretical_loss": 3.5591132654728743,
      "tokens_seen": 1309016064
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 723452,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5018327236175537,
      "objective/train/theoretical_loss": 3.55909714197814,
      "objective/train/tokens_used": 1329541600,
      "theoretical_loss": 3.55909714197814,
      "tokens_seen": 1309081600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047542627883651,
      "loss": 2.7335,
      "theoretical_loss": 3.55909714197814,
      "tokens_seen": 1309081600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047442326980943,
      "loss": 2.3299,
      "theoretical_loss": 3.559081019516567,
      "tokens_seen": 1309147136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030473420260782345,
      "loss": 2.7316,
      "theoretical_loss": 3.559064898088038,
      "tokens_seen": 1309212672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047241725175527,
      "loss": 2.7617,
      "theoretical_loss": 3.5590487776924356,
      "tokens_seen": 1309278208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030471414242728187,
      "loss": 2.7093,
      "theoretical_loss": 3.5590326583296417,
      "tokens_seen": 1309343744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030470411233701105,
      "loss": 2.6649,
      "theoretical_loss": 3.559016539999538,
      "tokens_seen": 1309409280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030469408224674023,
      "loss": 2.6519,
      "theoretical_loss": 3.559000422702007,
      "tokens_seen": 1309474816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030468405215646946,
      "loss": 2.7074,
      "theoretical_loss": 3.558984306436931,
      "tokens_seen": 1309540352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046740220661986,
      "loss": 2.7133,
      "theoretical_loss": 3.558968191204192,
      "tokens_seen": 1309605888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046639919759278,
      "loss": 2.578,
      "theoretical_loss": 3.558952077003672,
      "tokens_seen": 1309671424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030465396188565695,
      "loss": 2.6657,
      "theoretical_loss": 3.558935963835254,
      "tokens_seen": 1309736960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046439317953862,
      "loss": 2.702,
      "theoretical_loss": 3.5589198516988194,
      "tokens_seen": 1309802496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030463390170511537,
      "loss": 2.6953,
      "theoretical_loss": 3.5589037405942507,
      "tokens_seen": 1309868032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030462387161484455,
      "loss": 2.5696,
      "theoretical_loss": 3.55888763052143,
      "tokens_seen": 1309933568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030461384152457373,
      "loss": 2.7087,
      "theoretical_loss": 3.5588715214802407,
      "tokens_seen": 1309999104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046038114343029,
      "loss": 2.6059,
      "theoretical_loss": 3.558855413470564,
      "tokens_seen": 1310064640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045937813440321,
      "loss": 2.6829,
      "theoretical_loss": 3.558839306492283,
      "tokens_seen": 1310130176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030458375125376133,
      "loss": 2.5815,
      "theoretical_loss": 3.5588232005452793,
      "tokens_seen": 1310195712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030457372116349046,
      "loss": 2.6662,
      "theoretical_loss": 3.5588070956294358,
      "tokens_seen": 1310261248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045636910732197,
      "loss": 2.6904,
      "theoretical_loss": 3.5587909917446345,
      "tokens_seen": 1310326784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045536609829488,
      "loss": 2.6112,
      "theoretical_loss": 3.558774888890758,
      "tokens_seen": 1310392320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030454363089267805,
      "loss": 2.629,
      "theoretical_loss": 3.5587587870676893,
      "tokens_seen": 1310457856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030453360080240723,
      "loss": 2.5755,
      "theoretical_loss": 3.5587426862753104,
      "tokens_seen": 1310523392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045235707121364,
      "loss": 2.7388,
      "theoretical_loss": 3.5587265865135036,
      "tokens_seen": 1310588928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045135406218656,
      "loss": 2.6498,
      "theoretical_loss": 3.558710487782152,
      "tokens_seen": 1310654464
    },
    {
      "debugging/Self-BLEU-5": 0.39316690427144263,
      "debugging/distinct-1-grams": 0.8245694632792082,
      "debugging/distinct-2-grams": 0.8227741660491869,
      "debugging/entropy-1-grams": 5.538189760337977,
      "debugging/entropy-2-grams": 6.130351591764565,
      "debugging/length": 468.5,
      "debugging/num_segments": 8,
      "epoch": 0.4,
      "objective/train/docs_used": 724043,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.692194938659668,
      "objective/train/theoretical_loss": 3.558694390081137,
      "objective/train/tokens_used": 1331180000,
      "theoretical_loss": 3.558694390081137,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030450351053159483,
      "loss": 2.7732,
      "theoretical_loss": 3.558694390081137,
      "tokens_seen": 1310720000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030449348044132396,
      "loss": 2.8132,
      "theoretical_loss": 3.5586782934103427,
      "tokens_seen": 1310785536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044834503510532,
      "loss": 2.7486,
      "theoretical_loss": 3.5586621977696504,
      "tokens_seen": 1310851072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044734202607823,
      "loss": 2.4709,
      "theoretical_loss": 3.5586461031589436,
      "tokens_seen": 1310916608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030446339017051156,
      "loss": 2.7782,
      "theoretical_loss": 3.5586300095781045,
      "tokens_seen": 1310982144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030445336008024074,
      "loss": 2.6988,
      "theoretical_loss": 3.558613917027015,
      "tokens_seen": 1311047680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044433299899699,
      "loss": 2.5095,
      "theoretical_loss": 3.558597825505559,
      "tokens_seen": 1311113216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044332998996991,
      "loss": 2.4097,
      "theoretical_loss": 3.558581735013618,
      "tokens_seen": 1311178752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044232698094283,
      "loss": 2.7494,
      "theoretical_loss": 3.5585656455510755,
      "tokens_seen": 1311244288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030441323971915746,
      "loss": 2.6686,
      "theoretical_loss": 3.5585495571178143,
      "tokens_seen": 1311309824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044032096288867,
      "loss": 2.5388,
      "theoretical_loss": 3.5585334697137165,
      "tokens_seen": 1311375360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043931795386158,
      "loss": 2.7446,
      "theoretical_loss": 3.558517383338665,
      "tokens_seen": 1311440896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030438314944834506,
      "loss": 2.7984,
      "theoretical_loss": 3.5585012979925428,
      "tokens_seen": 1311506432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043731193580742,
      "loss": 2.7733,
      "theoretical_loss": 3.5584852136752323,
      "tokens_seen": 1311571968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043630892678034,
      "loss": 2.7841,
      "theoretical_loss": 3.5584691303866167,
      "tokens_seen": 1311637504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043530591775326,
      "loss": 2.5939,
      "theoretical_loss": 3.5584530481265793,
      "tokens_seen": 1311703040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043430290872618,
      "loss": 2.6253,
      "theoretical_loss": 3.5584369668950018,
      "tokens_seen": 1311768576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030433299899699097,
      "loss": 2.6828,
      "theoretical_loss": 3.558420886691767,
      "tokens_seen": 1311834112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043229689067202,
      "loss": 2.6631,
      "theoretical_loss": 3.5584048075167587,
      "tokens_seen": 1311899648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030431293881644933,
      "loss": 2.6433,
      "theoretical_loss": 3.5583887293698595,
      "tokens_seen": 1311965184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030430290872617856,
      "loss": 2.5721,
      "theoretical_loss": 3.5583726522509522,
      "tokens_seen": 1312030720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042928786359077,
      "loss": 2.8131,
      "theoretical_loss": 3.5583565761599196,
      "tokens_seen": 1312096256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042828485456369,
      "loss": 2.6619,
      "theoretical_loss": 3.5583405010966453,
      "tokens_seen": 1312161792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042728184553661,
      "loss": 2.6893,
      "theoretical_loss": 3.5583244270610113,
      "tokens_seen": 1312227328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042627883650953,
      "loss": 2.9458,
      "theoretical_loss": 3.558308354052901,
      "tokens_seen": 1312292864
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 725321,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.910378932952881,
      "objective/train/theoretical_loss": 3.5582922820721974,
      "objective/train/tokens_used": 1332818400,
      "theoretical_loss": 3.5582922820721974,
      "tokens_seen": 1312358400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030425275827482447,
      "loss": 2.7906,
      "theoretical_loss": 3.5582922820721974,
      "tokens_seen": 1312358400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030424272818455365,
      "loss": 2.5895,
      "theoretical_loss": 3.558276211118784,
      "tokens_seen": 1312423936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030423269809428283,
      "loss": 2.5593,
      "theoretical_loss": 3.558260141192543,
      "tokens_seen": 1312489472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030422266800401207,
      "loss": 2.6732,
      "theoretical_loss": 3.558244072293358,
      "tokens_seen": 1312555008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042126379137412,
      "loss": 2.8627,
      "theoretical_loss": 3.558228004421112,
      "tokens_seen": 1312620544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030420260782347043,
      "loss": 2.6099,
      "theoretical_loss": 3.558211937575688,
      "tokens_seen": 1312686080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041925777331996,
      "loss": 2.781,
      "theoretical_loss": 3.5581958717569697,
      "tokens_seen": 1312751616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041825476429288,
      "loss": 2.6528,
      "theoretical_loss": 3.5581798069648394,
      "tokens_seen": 1312817152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030417251755265797,
      "loss": 2.6068,
      "theoretical_loss": 3.5581637431991804,
      "tokens_seen": 1312882688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030416248746238715,
      "loss": 2.5357,
      "theoretical_loss": 3.5581476804598764,
      "tokens_seen": 1312948224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030415245737211633,
      "loss": 2.7134,
      "theoretical_loss": 3.55813161874681,
      "tokens_seen": 1313013760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030414242728184557,
      "loss": 2.8565,
      "theoretical_loss": 3.558115558059865,
      "tokens_seen": 1313079296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041323971915747,
      "loss": 2.6083,
      "theoretical_loss": 3.558099498398924,
      "tokens_seen": 1313144832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030412236710130393,
      "loss": 2.8025,
      "theoretical_loss": 3.5580834397638705,
      "tokens_seen": 1313210368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030411233701103306,
      "loss": 2.8828,
      "theoretical_loss": 3.5580673821545883,
      "tokens_seen": 1313275904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041023069207623,
      "loss": 2.6021,
      "theoretical_loss": 3.55805132557096,
      "tokens_seen": 1313341440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040922768304915,
      "loss": 2.6631,
      "theoretical_loss": 3.558035270012869,
      "tokens_seen": 1313406976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030408224674022066,
      "loss": 2.5568,
      "theoretical_loss": 3.558019215480199,
      "tokens_seen": 1313472512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030407221664994984,
      "loss": 2.8165,
      "theoretical_loss": 3.5580031619728327,
      "tokens_seen": 1313538048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304062186559679,
      "loss": 2.6535,
      "theoretical_loss": 3.5579871094906546,
      "tokens_seen": 1313603584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030405215646940825,
      "loss": 2.8061,
      "theoretical_loss": 3.557971058033547,
      "tokens_seen": 1313669120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030404212637913743,
      "loss": 2.5655,
      "theoretical_loss": 3.557955007601394,
      "tokens_seen": 1313734656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040320962888666,
      "loss": 2.7296,
      "theoretical_loss": 3.5579389581940783,
      "tokens_seen": 1313800192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040220661985958,
      "loss": 2.7383,
      "theoretical_loss": 3.557922909811484,
      "tokens_seen": 1313865728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030401203610832503,
      "loss": 2.7464,
      "theoretical_loss": 3.5579068624534944,
      "tokens_seen": 1313931264
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 725665,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5255112648010254,
      "objective/train/theoretical_loss": 3.5578908161199934,
      "objective/train/tokens_used": 1334456800,
      "theoretical_loss": 3.5578908161199934,
      "tokens_seen": 1313996800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030400200601805416,
      "loss": 2.5646,
      "theoretical_loss": 3.5578908161199934,
      "tokens_seen": 1313996800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039919759277834,
      "loss": 2.719,
      "theoretical_loss": 3.557874770810863,
      "tokens_seen": 1314062336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039819458375125,
      "loss": 2.7278,
      "theoretical_loss": 3.5578587265259882,
      "tokens_seen": 1314127872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030397191574724176,
      "loss": 2.6767,
      "theoretical_loss": 3.5578426832652523,
      "tokens_seen": 1314193408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030396188565697094,
      "loss": 2.9329,
      "theoretical_loss": 3.5578266410285386,
      "tokens_seen": 1314258944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039518555667001,
      "loss": 2.7439,
      "theoretical_loss": 3.557810599815731,
      "tokens_seen": 1314324480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039418254764293,
      "loss": 2.7347,
      "theoretical_loss": 3.5577945596267124,
      "tokens_seen": 1314390016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039317953861585,
      "loss": 2.6394,
      "theoretical_loss": 3.557778520461367,
      "tokens_seen": 1314455552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030392176529588766,
      "loss": 2.4425,
      "theoretical_loss": 3.5577624823195784,
      "tokens_seen": 1314521088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003039117352056169,
      "loss": 2.7863,
      "theoretical_loss": 3.5577464452012295,
      "tokens_seen": 1314586624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303901705115346,
      "loss": 2.6364,
      "theoretical_loss": 3.557730409106205,
      "tokens_seen": 1314652160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030389167502507526,
      "loss": 2.7844,
      "theoretical_loss": 3.557714374034389,
      "tokens_seen": 1314717696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038816449348044,
      "loss": 2.7259,
      "theoretical_loss": 3.557698339985664,
      "tokens_seen": 1314783232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038716148445336,
      "loss": 2.6794,
      "theoretical_loss": 3.5576823069599133,
      "tokens_seen": 1314848768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038615847542628,
      "loss": 2.8248,
      "theoretical_loss": 3.5576662749570227,
      "tokens_seen": 1314914304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303851554663992,
      "loss": 2.6524,
      "theoretical_loss": 3.557650243976874,
      "tokens_seen": 1314979840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030384152457372117,
      "loss": 2.862,
      "theoretical_loss": 3.5576342140193518,
      "tokens_seen": 1315045376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038314944834504,
      "loss": 2.8933,
      "theoretical_loss": 3.5576181850843405,
      "tokens_seen": 1315110912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030382146439317953,
      "loss": 2.7301,
      "theoretical_loss": 3.557602157171723,
      "tokens_seen": 1315176448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030381143430290876,
      "loss": 2.6285,
      "theoretical_loss": 3.5575861302813827,
      "tokens_seen": 1315241984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038014042126379,
      "loss": 2.7635,
      "theoretical_loss": 3.5575701044132053,
      "tokens_seen": 1315307520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037913741223671,
      "loss": 2.5867,
      "theoretical_loss": 3.557554079567073,
      "tokens_seen": 1315373056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037813440320963,
      "loss": 2.6651,
      "theoretical_loss": 3.5575380557428704,
      "tokens_seen": 1315438592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037713139418255,
      "loss": 2.5734,
      "theoretical_loss": 3.5575220329404815,
      "tokens_seen": 1315504128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030376128385155467,
      "loss": 2.8709,
      "theoretical_loss": 3.55750601115979,
      "tokens_seen": 1315569664
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 726875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6938629150390625,
      "objective/train/theoretical_loss": 3.55748999040068,
      "objective/train/tokens_used": 1336095200,
      "theoretical_loss": 3.55748999040068,
      "tokens_seen": 1315635200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030375125376128385,
      "loss": 2.8485,
      "theoretical_loss": 3.55748999040068,
      "tokens_seen": 1315635200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030374122367101303,
      "loss": 2.6985,
      "theoretical_loss": 3.557473970663035,
      "tokens_seen": 1315700736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030373119358074227,
      "loss": 2.9056,
      "theoretical_loss": 3.5574579519467395,
      "tokens_seen": 1315766272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037211634904714,
      "loss": 2.5467,
      "theoretical_loss": 3.5574419342516777,
      "tokens_seen": 1315831808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030371113340020063,
      "loss": 2.5505,
      "theoretical_loss": 3.557425917577733,
      "tokens_seen": 1315897344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037011033099298,
      "loss": 2.829,
      "theoretical_loss": 3.5574099019247902,
      "tokens_seen": 1315962880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303691073219659,
      "loss": 2.763,
      "theoretical_loss": 3.557393887292733,
      "tokens_seen": 1316028416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030368104312938817,
      "loss": 2.6073,
      "theoretical_loss": 3.557377873681445,
      "tokens_seen": 1316093952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030367101303911735,
      "loss": 2.7531,
      "theoretical_loss": 3.5573618610908113,
      "tokens_seen": 1316159488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030366098294884653,
      "loss": 2.7179,
      "theoretical_loss": 3.5573458495207153,
      "tokens_seen": 1316225024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030365095285857577,
      "loss": 2.7346,
      "theoretical_loss": 3.5573298389710413,
      "tokens_seen": 1316290560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036409227683049,
      "loss": 2.6652,
      "theoretical_loss": 3.5573138294416733,
      "tokens_seen": 1316356096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030363089267803413,
      "loss": 2.7358,
      "theoretical_loss": 3.5572978209324964,
      "tokens_seen": 1316421632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030362086258776326,
      "loss": 2.6463,
      "theoretical_loss": 3.5572818134433937,
      "tokens_seen": 1316487168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036108324974925,
      "loss": 2.7587,
      "theoretical_loss": 3.5572658069742498,
      "tokens_seen": 1316552704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036008024072217,
      "loss": 2.6971,
      "theoretical_loss": 3.557249801524949,
      "tokens_seen": 1316618240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030359077231695086,
      "loss": 2.7197,
      "theoretical_loss": 3.5572337970953756,
      "tokens_seen": 1316683776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030358074222668004,
      "loss": 2.4087,
      "theoretical_loss": 3.557217793685414,
      "tokens_seen": 1316749312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035707121364092,
      "loss": 2.7399,
      "theoretical_loss": 3.557201791294948,
      "tokens_seen": 1316814848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035606820461384,
      "loss": 2.8805,
      "theoretical_loss": 3.5571857899238624,
      "tokens_seen": 1316880384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030355065195586763,
      "loss": 2.36,
      "theoretical_loss": 3.5571697895720416,
      "tokens_seen": 1316945920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030354062186559676,
      "loss": 2.7899,
      "theoretical_loss": 3.5571537902393695,
      "tokens_seen": 1317011456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303530591775326,
      "loss": 2.7434,
      "theoretical_loss": 3.5571377919257303,
      "tokens_seen": 1317076992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035205616850552,
      "loss": 2.9044,
      "theoretical_loss": 3.5571217946310094,
      "tokens_seen": 1317142528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030351053159478436,
      "loss": 2.6923,
      "theoretical_loss": 3.5571057983550904,
      "tokens_seen": 1317208064
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 727478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8409438133239746,
      "objective/train/theoretical_loss": 3.5570898030978584,
      "objective/train/tokens_used": 1337733600,
      "theoretical_loss": 3.5570898030978584,
      "tokens_seen": 1317273600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030350050150451354,
      "loss": 2.6263,
      "theoretical_loss": 3.5570898030978584,
      "tokens_seen": 1317273600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034904714142427,
      "loss": 2.693,
      "theoretical_loss": 3.5570738088591973,
      "tokens_seen": 1317339136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034804413239719,
      "loss": 2.6554,
      "theoretical_loss": 3.5570578156389914,
      "tokens_seen": 1317404672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030347041123370114,
      "loss": 2.5147,
      "theoretical_loss": 3.5570418234371255,
      "tokens_seen": 1317470208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030346038114343026,
      "loss": 2.7155,
      "theoretical_loss": 3.5570258322534842,
      "tokens_seen": 1317535744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034503510531595,
      "loss": 2.5846,
      "theoretical_loss": 3.5570098420879517,
      "tokens_seen": 1317601280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034403209628886,
      "loss": 2.7062,
      "theoretical_loss": 3.556993852940413,
      "tokens_seen": 1317666816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030343029087261786,
      "loss": 2.7364,
      "theoretical_loss": 3.556977864810752,
      "tokens_seen": 1317732352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030342026078234704,
      "loss": 2.7312,
      "theoretical_loss": 3.5569618776988543,
      "tokens_seen": 1317797888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034102306920762,
      "loss": 2.6489,
      "theoretical_loss": 3.5569458916046033,
      "tokens_seen": 1317863424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034002006018054,
      "loss": 2.5915,
      "theoretical_loss": 3.5569299065278845,
      "tokens_seen": 1317928960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033901705115346,
      "loss": 2.5329,
      "theoretical_loss": 3.556913922468582,
      "tokens_seen": 1317994496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030338014042126377,
      "loss": 2.6562,
      "theoretical_loss": 3.556897939426581,
      "tokens_seen": 1318060032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303370110330993,
      "loss": 2.7461,
      "theoretical_loss": 3.5568819574017665,
      "tokens_seen": 1318125568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030336008024072213,
      "loss": 2.7815,
      "theoretical_loss": 3.556865976394022,
      "tokens_seen": 1318191104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030335005015045137,
      "loss": 2.5943,
      "theoretical_loss": 3.5568499964032325,
      "tokens_seen": 1318256640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030334002006018055,
      "loss": 2.6025,
      "theoretical_loss": 3.5568340174292836,
      "tokens_seen": 1318322176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030332998996990973,
      "loss": 2.5328,
      "theoretical_loss": 3.556818039472059,
      "tokens_seen": 1318387712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033199598796389,
      "loss": 2.729,
      "theoretical_loss": 3.5568020625314443,
      "tokens_seen": 1318453248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033099297893681,
      "loss": 2.5643,
      "theoretical_loss": 3.5567860866073238,
      "tokens_seen": 1318518784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032998996990973,
      "loss": 2.741,
      "theoretical_loss": 3.5567701116995827,
      "tokens_seen": 1318584320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032898696088265,
      "loss": 2.7676,
      "theoretical_loss": 3.5567541378081056,
      "tokens_seen": 1318649856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032798395185557,
      "loss": 2.5141,
      "theoretical_loss": 3.556738164932777,
      "tokens_seen": 1318715392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030326980942828487,
      "loss": 2.6577,
      "theoretical_loss": 3.5567221930734823,
      "tokens_seen": 1318780928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030325977933801405,
      "loss": 2.7262,
      "theoretical_loss": 3.5567062222301065,
      "tokens_seen": 1318846464
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 728516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7909562587738037,
      "objective/train/theoretical_loss": 3.5566902524025332,
      "objective/train/tokens_used": 1339372000,
      "theoretical_loss": 3.5566902524025332,
      "tokens_seen": 1318912000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030324974924774323,
      "loss": 2.8554,
      "theoretical_loss": 3.5566902524025332,
      "tokens_seen": 1318912000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030323971915747247,
      "loss": 2.6728,
      "theoretical_loss": 3.556674283590649,
      "tokens_seen": 1318977536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032296890672016,
      "loss": 2.5645,
      "theoretical_loss": 3.5566583157943383,
      "tokens_seen": 1319043072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030321965897693083,
      "loss": 2.6148,
      "theoretical_loss": 3.5566423490134857,
      "tokens_seen": 1319108608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030320962888666,
      "loss": 2.7542,
      "theoretical_loss": 3.556626383247976,
      "tokens_seen": 1319174144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031995987963892,
      "loss": 2.7342,
      "theoretical_loss": 3.556610418497695,
      "tokens_seen": 1319239680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030318956870611837,
      "loss": 2.7786,
      "theoretical_loss": 3.5565944547625277,
      "tokens_seen": 1319305216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030317953861584755,
      "loss": 2.486,
      "theoretical_loss": 3.556578492042358,
      "tokens_seen": 1319370752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030316950852557673,
      "loss": 2.591,
      "theoretical_loss": 3.5565625303370725,
      "tokens_seen": 1319436288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030315947843530597,
      "loss": 2.5615,
      "theoretical_loss": 3.556546569646555,
      "tokens_seen": 1319501824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031494483450351,
      "loss": 2.4469,
      "theoretical_loss": 3.5565306099706904,
      "tokens_seen": 1319567360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030313941825476433,
      "loss": 2.6545,
      "theoretical_loss": 3.5565146513093655,
      "tokens_seen": 1319632896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030312938816449346,
      "loss": 2.6957,
      "theoretical_loss": 3.556498693662464,
      "tokens_seen": 1319698432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031193580742227,
      "loss": 2.5422,
      "theoretical_loss": 3.556482737029871,
      "tokens_seen": 1319763968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031093279839519,
      "loss": 2.7761,
      "theoretical_loss": 3.5564667814114728,
      "tokens_seen": 1319829504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030309929789368106,
      "loss": 2.7011,
      "theoretical_loss": 3.5564508268071533,
      "tokens_seen": 1319895040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030308926780341024,
      "loss": 2.5986,
      "theoretical_loss": 3.5564348732167987,
      "tokens_seen": 1319960576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030792377131394,
      "loss": 2.7712,
      "theoretical_loss": 3.5564189206402936,
      "tokens_seen": 1320026112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030692076228686,
      "loss": 2.5667,
      "theoretical_loss": 3.5564029690775234,
      "tokens_seen": 1320091648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030305917753259783,
      "loss": 2.5837,
      "theoretical_loss": 3.5563870185283735,
      "tokens_seen": 1320157184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030304914744232696,
      "loss": 2.8657,
      "theoretical_loss": 3.556371068992729,
      "tokens_seen": 1320222720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030391173520562,
      "loss": 2.734,
      "theoretical_loss": 3.5563551204704753,
      "tokens_seen": 1320288256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030290872617854,
      "loss": 2.4422,
      "theoretical_loss": 3.5563391729614975,
      "tokens_seen": 1320353792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030301905717151456,
      "loss": 2.8178,
      "theoretical_loss": 3.556323226465681,
      "tokens_seen": 1320419328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030300902708124374,
      "loss": 2.7261,
      "theoretical_loss": 3.556307280982912,
      "tokens_seen": 1320484864
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 729780,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1676318645477295,
      "objective/train/theoretical_loss": 3.556291336513074,
      "objective/train/tokens_used": 1341010400,
      "theoretical_loss": 3.556291336513074,
      "tokens_seen": 1320550400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029989969909729,
      "loss": 2.4284,
      "theoretical_loss": 3.556291336513074,
      "tokens_seen": 1320550400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029889669007021,
      "loss": 2.6546,
      "theoretical_loss": 3.5562753930560547,
      "tokens_seen": 1320615936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030297893681043134,
      "loss": 2.6635,
      "theoretical_loss": 3.556259450611738,
      "tokens_seen": 1320681472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030296890672016046,
      "loss": 2.6725,
      "theoretical_loss": 3.5562435091800095,
      "tokens_seen": 1320747008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029588766298897,
      "loss": 2.7965,
      "theoretical_loss": 3.5562275687607547,
      "tokens_seen": 1320812544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029488465396188,
      "loss": 2.6281,
      "theoretical_loss": 3.5562116293538595,
      "tokens_seen": 1320878080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030293881644934806,
      "loss": 2.5591,
      "theoretical_loss": 3.5561956909592087,
      "tokens_seen": 1320943616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030292878635907724,
      "loss": 2.8417,
      "theoretical_loss": 3.5561797535766884,
      "tokens_seen": 1321009152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029187562688064,
      "loss": 2.5182,
      "theoretical_loss": 3.5561638172061834,
      "tokens_seen": 1321074688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029087261785356,
      "loss": 2.7444,
      "theoretical_loss": 3.5561478818475805,
      "tokens_seen": 1321140224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028986960882648,
      "loss": 2.7431,
      "theoretical_loss": 3.556131947500764,
      "tokens_seen": 1321205760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030288866599799397,
      "loss": 2.7926,
      "theoretical_loss": 3.55611601416562,
      "tokens_seen": 1321271296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028786359077232,
      "loss": 2.8462,
      "theoretical_loss": 3.5561000818420343,
      "tokens_seen": 1321336832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030286860581745233,
      "loss": 2.6434,
      "theoretical_loss": 3.556084150529892,
      "tokens_seen": 1321402368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030285857572718157,
      "loss": 2.2269,
      "theoretical_loss": 3.5560682202290783,
      "tokens_seen": 1321467904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030284854563691075,
      "loss": 2.6968,
      "theoretical_loss": 3.5560522909394807,
      "tokens_seen": 1321533440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030283851554663993,
      "loss": 2.8215,
      "theoretical_loss": 3.5560363626609828,
      "tokens_seen": 1321598976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028284854563691,
      "loss": 2.6064,
      "theoretical_loss": 3.5560204353934717,
      "tokens_seen": 1321664512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028184553660983,
      "loss": 2.6011,
      "theoretical_loss": 3.5560045091368324,
      "tokens_seen": 1321730048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030280842527582747,
      "loss": 2.6131,
      "theoretical_loss": 3.555988583890951,
      "tokens_seen": 1321795584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027983951855567,
      "loss": 2.694,
      "theoretical_loss": 3.555972659655713,
      "tokens_seen": 1321861120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030278836509528583,
      "loss": 2.5547,
      "theoretical_loss": 3.5559567364310043,
      "tokens_seen": 1321926656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030277833500501507,
      "loss": 2.676,
      "theoretical_loss": 3.5559408142167106,
      "tokens_seen": 1321992192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027683049147442,
      "loss": 2.7288,
      "theoretical_loss": 3.555924893012717,
      "tokens_seen": 1322057728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030275827482447343,
      "loss": 2.7041,
      "theoretical_loss": 3.5559089728189113,
      "tokens_seen": 1322123264
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 730226,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.751304864883423,
      "objective/train/theoretical_loss": 3.555893053635177,
      "objective/train/tokens_used": 1342648800,
      "theoretical_loss": 3.555893053635177,
      "tokens_seen": 1322188800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027482447342026,
      "loss": 2.8815,
      "theoretical_loss": 3.555893053635177,
      "tokens_seen": 1322188800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027382146439318,
      "loss": 2.6246,
      "theoretical_loss": 3.555877135461401,
      "tokens_seen": 1322254336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302728184553661,
      "loss": 2.6019,
      "theoretical_loss": 3.55586121829747,
      "tokens_seen": 1322319872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027181544633902,
      "loss": 2.5323,
      "theoretical_loss": 3.5558453021432683,
      "tokens_seen": 1322385408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030270812437311934,
      "loss": 2.497,
      "theoretical_loss": 3.555829386998683,
      "tokens_seen": 1322450944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030269809428284857,
      "loss": 2.8508,
      "theoretical_loss": 3.5558134728635995,
      "tokens_seen": 1322516480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026880641925777,
      "loss": 2.629,
      "theoretical_loss": 3.5557975597379037,
      "tokens_seen": 1322582016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030267803410230693,
      "loss": 2.5966,
      "theoretical_loss": 3.555781647621482,
      "tokens_seen": 1322647552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026680040120361,
      "loss": 2.803,
      "theoretical_loss": 3.55576573651422,
      "tokens_seen": 1322713088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026579739217653,
      "loss": 2.7159,
      "theoretical_loss": 3.5557498264160037,
      "tokens_seen": 1322778624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026479438314945,
      "loss": 2.7146,
      "theoretical_loss": 3.5557339173267195,
      "tokens_seen": 1322844160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030263791374122366,
      "loss": 2.6778,
      "theoretical_loss": 3.555718009246253,
      "tokens_seen": 1322909696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030262788365095284,
      "loss": 2.6243,
      "theoretical_loss": 3.5557021021744903,
      "tokens_seen": 1322975232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026178535606821,
      "loss": 2.7398,
      "theoretical_loss": 3.5556861961113175,
      "tokens_seen": 1323040768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026078234704112,
      "loss": 2.6839,
      "theoretical_loss": 3.555670291056621,
      "tokens_seen": 1323106304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030259779338014044,
      "loss": 2.6276,
      "theoretical_loss": 3.5556543870102866,
      "tokens_seen": 1323171840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030258776328986956,
      "loss": 2.8223,
      "theoretical_loss": 3.5556384839722006,
      "tokens_seen": 1323237376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025777331995988,
      "loss": 2.6482,
      "theoretical_loss": 3.5556225819422496,
      "tokens_seen": 1323302912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302567703109328,
      "loss": 2.8187,
      "theoretical_loss": 3.5556066809203184,
      "tokens_seen": 1323368448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030255767301905716,
      "loss": 2.6949,
      "theoretical_loss": 3.5555907809062948,
      "tokens_seen": 1323433984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025476429287864,
      "loss": 2.4339,
      "theoretical_loss": 3.5555748819000637,
      "tokens_seen": 1323499520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025376128385156,
      "loss": 2.6973,
      "theoretical_loss": 3.555558983901512,
      "tokens_seen": 1323565056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030252758274824476,
      "loss": 2.6124,
      "theoretical_loss": 3.555543086910526,
      "tokens_seen": 1323630592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030251755265797394,
      "loss": 2.7644,
      "theoretical_loss": 3.5555271909269917,
      "tokens_seen": 1323696128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025075225677031,
      "loss": 2.8019,
      "theoretical_loss": 3.5555112959507955,
      "tokens_seen": 1323761664
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 731155,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7156481742858887,
      "objective/train/theoretical_loss": 3.5554954019818235,
      "objective/train/tokens_used": 1344287200,
      "theoretical_loss": 3.5554954019818235,
      "tokens_seen": 1323827200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024974924774323,
      "loss": 2.7315,
      "theoretical_loss": 3.5554954019818235,
      "tokens_seen": 1323827200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030248746238716154,
      "loss": 2.8048,
      "theoretical_loss": 3.5554795090199622,
      "tokens_seen": 1323892736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030247743229689066,
      "loss": 2.7554,
      "theoretical_loss": 3.555463617065098,
      "tokens_seen": 1323958272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024674022066199,
      "loss": 2.6749,
      "theoretical_loss": 3.5554477261171176,
      "tokens_seen": 1324023808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030245737211634903,
      "loss": 2.4962,
      "theoretical_loss": 3.5554318361759063,
      "tokens_seen": 1324089344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030244734202607826,
      "loss": 2.7971,
      "theoretical_loss": 3.555415947241351,
      "tokens_seen": 1324154880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030243731193580744,
      "loss": 2.6869,
      "theoretical_loss": 3.555400059313339,
      "tokens_seen": 1324220416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024272818455366,
      "loss": 2.8608,
      "theoretical_loss": 3.555384172391755,
      "tokens_seen": 1324285952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024172517552658,
      "loss": 2.7737,
      "theoretical_loss": 3.5553682864764875,
      "tokens_seen": 1324351488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302407221664995,
      "loss": 2.5956,
      "theoretical_loss": 3.555352401567421,
      "tokens_seen": 1324417024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030239719157472417,
      "loss": 2.5396,
      "theoretical_loss": 3.5553365176644434,
      "tokens_seen": 1324482560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023871614844534,
      "loss": 2.5868,
      "theoretical_loss": 3.5553206347674404,
      "tokens_seen": 1324548096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030237713139418253,
      "loss": 2.6205,
      "theoretical_loss": 3.555304752876299,
      "tokens_seen": 1324613632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030236710130391177,
      "loss": 2.673,
      "theoretical_loss": 3.5552888719909053,
      "tokens_seen": 1324679168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030235707121364095,
      "loss": 2.7001,
      "theoretical_loss": 3.5552729921111466,
      "tokens_seen": 1324744704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030234704112337013,
      "loss": 2.7891,
      "theoretical_loss": 3.555257113236908,
      "tokens_seen": 1324810240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023370110330993,
      "loss": 2.747,
      "theoretical_loss": 3.5552412353680776,
      "tokens_seen": 1324875776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023269809428285,
      "loss": 2.6373,
      "theoretical_loss": 3.5552253585045417,
      "tokens_seen": 1324941312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030231695085255767,
      "loss": 2.5577,
      "theoretical_loss": 3.5552094826461866,
      "tokens_seen": 1325006848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023069207622869,
      "loss": 2.6893,
      "theoretical_loss": 3.5551936077928987,
      "tokens_seen": 1325072384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030229689067201603,
      "loss": 2.5837,
      "theoretical_loss": 3.555177733944565,
      "tokens_seen": 1325137920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030228686058174527,
      "loss": 2.7776,
      "theoretical_loss": 3.5551618611010722,
      "tokens_seen": 1325203456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022768304914744,
      "loss": 2.6919,
      "theoretical_loss": 3.555145989262307,
      "tokens_seen": 1325268992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030226680040120363,
      "loss": 2.7789,
      "theoretical_loss": 3.5551301184281567,
      "tokens_seen": 1325334528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022567703109328,
      "loss": 2.6184,
      "theoretical_loss": 3.5551142485985068,
      "tokens_seen": 1325400064
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 731791,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.585141658782959,
      "objective/train/theoretical_loss": 3.555098379773245,
      "objective/train/tokens_used": 1345925600,
      "theoretical_loss": 3.555098379773245,
      "tokens_seen": 1325465600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302246740220662,
      "loss": 2.6396,
      "theoretical_loss": 3.555098379773245,
      "tokens_seen": 1325465600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022367101303912,
      "loss": 2.6297,
      "theoretical_loss": 3.555082511952258,
      "tokens_seen": 1325531136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022266800401204,
      "loss": 2.7135,
      "theoretical_loss": 3.555066645135432,
      "tokens_seen": 1325596672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030221664994984954,
      "loss": 2.6646,
      "theoretical_loss": 3.555050779322654,
      "tokens_seen": 1325662208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030220661985957877,
      "loss": 2.6371,
      "theoretical_loss": 3.5550349145138114,
      "tokens_seen": 1325727744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021965897693079,
      "loss": 2.6964,
      "theoretical_loss": 3.5550190507087907,
      "tokens_seen": 1325793280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030218655967903713,
      "loss": 2.8007,
      "theoretical_loss": 3.5550031879074786,
      "tokens_seen": 1325858816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021765295887663,
      "loss": 2.6598,
      "theoretical_loss": 3.554987326109762,
      "tokens_seen": 1325924352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021664994984955,
      "loss": 2.7457,
      "theoretical_loss": 3.5549714653155284,
      "tokens_seen": 1325989888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021564694082247,
      "loss": 2.7275,
      "theoretical_loss": 3.554955605524664,
      "tokens_seen": 1326055424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030214643931795386,
      "loss": 2.7028,
      "theoretical_loss": 3.5549397467370563,
      "tokens_seen": 1326120960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030213640922768304,
      "loss": 2.7375,
      "theoretical_loss": 3.554923888952591,
      "tokens_seen": 1326186496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021263791374123,
      "loss": 2.6846,
      "theoretical_loss": 3.5549080321711575,
      "tokens_seen": 1326252032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021163490471414,
      "loss": 2.7879,
      "theoretical_loss": 3.5548921763926407,
      "tokens_seen": 1326317568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030210631895687064,
      "loss": 2.6558,
      "theoretical_loss": 3.554876321616928,
      "tokens_seen": 1326383104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030209628886659976,
      "loss": 2.7538,
      "theoretical_loss": 3.5548604678439064,
      "tokens_seen": 1326448640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302086258776329,
      "loss": 2.5872,
      "theoretical_loss": 3.5548446150734643,
      "tokens_seen": 1326514176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020762286860582,
      "loss": 2.7702,
      "theoretical_loss": 3.554828763305487,
      "tokens_seen": 1326579712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030206619859578736,
      "loss": 2.672,
      "theoretical_loss": 3.5548129125398624,
      "tokens_seen": 1326645248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030205616850551654,
      "loss": 2.7365,
      "theoretical_loss": 3.5547970627764776,
      "tokens_seen": 1326710784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020461384152458,
      "loss": 2.6226,
      "theoretical_loss": 3.5547812140152195,
      "tokens_seen": 1326776320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020361083249749,
      "loss": 2.6412,
      "theoretical_loss": 3.5547653662559755,
      "tokens_seen": 1326841856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030202607823470414,
      "loss": 2.6594,
      "theoretical_loss": 3.5547495194986327,
      "tokens_seen": 1326907392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030201604814443327,
      "loss": 2.6627,
      "theoretical_loss": 3.554733673743078,
      "tokens_seen": 1326972928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020060180541625,
      "loss": 2.5631,
      "theoretical_loss": 3.554717828989199,
      "tokens_seen": 1327038464
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 732925,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.762134075164795,
      "objective/train/theoretical_loss": 3.554701985236883,
      "objective/train/tokens_used": 1347564000,
      "theoretical_loss": 3.554701985236883,
      "tokens_seen": 1327104000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003019959879638917,
      "loss": 2.5826,
      "theoretical_loss": 3.554701985236883,
      "tokens_seen": 1327104000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030198595787362087,
      "loss": 2.6966,
      "theoretical_loss": 3.5546861424860166,
      "tokens_seen": 1327169536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030197592778335005,
      "loss": 2.7382,
      "theoretical_loss": 3.554670300736487,
      "tokens_seen": 1327235072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030196589769307923,
      "loss": 2.68,
      "theoretical_loss": 3.554654459988183,
      "tokens_seen": 1327300608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003019558676028084,
      "loss": 2.7091,
      "theoretical_loss": 3.55463862024099,
      "tokens_seen": 1327366144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030194583751253764,
      "loss": 2.8205,
      "theoretical_loss": 3.554622781494796,
      "tokens_seen": 1327431680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030193580742226677,
      "loss": 3.0017,
      "theoretical_loss": 3.554606943749489,
      "tokens_seen": 1327497216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301925777331996,
      "loss": 2.7315,
      "theoretical_loss": 3.5545911070049554,
      "tokens_seen": 1327562752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030191574724172513,
      "loss": 2.74,
      "theoretical_loss": 3.5545752712610836,
      "tokens_seen": 1327628288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030190571715145437,
      "loss": 2.6388,
      "theoretical_loss": 3.5545594365177595,
      "tokens_seen": 1327693824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030189568706118355,
      "loss": 2.6028,
      "theoretical_loss": 3.5545436027748716,
      "tokens_seen": 1327759360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030188565697091273,
      "loss": 2.6851,
      "theoretical_loss": 3.554527770032307,
      "tokens_seen": 1327824896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018756268806419,
      "loss": 2.7161,
      "theoretical_loss": 3.554511938289953,
      "tokens_seen": 1327890432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030186559679037115,
      "loss": 2.6321,
      "theoretical_loss": 3.5544961075476977,
      "tokens_seen": 1327955968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018555667001003,
      "loss": 2.6999,
      "theoretical_loss": 3.5544802778054274,
      "tokens_seen": 1328021504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018455366098295,
      "loss": 2.5249,
      "theoretical_loss": 3.554464449063031,
      "tokens_seen": 1328087040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030183550651955864,
      "loss": 2.5905,
      "theoretical_loss": 3.554448621320395,
      "tokens_seen": 1328152576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030182547642928787,
      "loss": 2.7502,
      "theoretical_loss": 3.5544327945774072,
      "tokens_seen": 1328218112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030181544633901705,
      "loss": 2.7522,
      "theoretical_loss": 3.554416968833955,
      "tokens_seen": 1328283648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030180541624874623,
      "loss": 2.4335,
      "theoretical_loss": 3.554401144089926,
      "tokens_seen": 1328349184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030179538615847547,
      "loss": 2.753,
      "theoretical_loss": 3.554385320345209,
      "tokens_seen": 1328414720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017853560682046,
      "loss": 2.7503,
      "theoretical_loss": 3.5543694975996893,
      "tokens_seen": 1328480256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030177532597793383,
      "loss": 2.6822,
      "theoretical_loss": 3.554353675853256,
      "tokens_seen": 1328545792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301765295887663,
      "loss": 2.7244,
      "theoretical_loss": 3.554337855105797,
      "tokens_seen": 1328611328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017552657973922,
      "loss": 2.6174,
      "theoretical_loss": 3.554322035357199,
      "tokens_seen": 1328676864
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 733999,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7849137783050537,
      "objective/train/theoretical_loss": 3.5543062166073502,
      "objective/train/tokens_used": 1349202400,
      "theoretical_loss": 3.5543062166073502,
      "tokens_seen": 1328742400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017452357071214,
      "loss": 2.7786,
      "theoretical_loss": 3.5543062166073502,
      "tokens_seen": 1328742400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017352056168506,
      "loss": 2.9597,
      "theoretical_loss": 3.554290398856138,
      "tokens_seen": 1328807936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030172517552657974,
      "loss": 2.5949,
      "theoretical_loss": 3.5542745821034503,
      "tokens_seen": 1328873472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030171514543630897,
      "loss": 2.6613,
      "theoretical_loss": 3.554258766349175,
      "tokens_seen": 1328939008
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017051153460381,
      "loss": 2.5457,
      "theoretical_loss": 3.5542429515931993,
      "tokens_seen": 1329004544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030169508525576733,
      "loss": 2.7014,
      "theoretical_loss": 3.5542271378354116,
      "tokens_seen": 1329070080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016850551654965,
      "loss": 2.8066,
      "theoretical_loss": 3.5542113250756993,
      "tokens_seen": 1329135616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016750250752257,
      "loss": 2.824,
      "theoretical_loss": 3.5541955133139504,
      "tokens_seen": 1329201152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016649949849549,
      "loss": 2.6049,
      "theoretical_loss": 3.5541797025500523,
      "tokens_seen": 1329266688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030165496489468406,
      "loss": 2.7338,
      "theoretical_loss": 3.554163892783894,
      "tokens_seen": 1329332224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030164493480441324,
      "loss": 2.7905,
      "theoretical_loss": 3.5541480840153614,
      "tokens_seen": 1329397760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016349047141425,
      "loss": 2.6749,
      "theoretical_loss": 3.5541322762443444,
      "tokens_seen": 1329463296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016248746238716,
      "loss": 2.5686,
      "theoretical_loss": 3.5541164694707295,
      "tokens_seen": 1329528832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030161484453360084,
      "loss": 2.7111,
      "theoretical_loss": 3.5541006636944052,
      "tokens_seen": 1329594368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030160481444332996,
      "loss": 2.6742,
      "theoretical_loss": 3.5540848589152594,
      "tokens_seen": 1329659904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015947843530592,
      "loss": 2.6322,
      "theoretical_loss": 3.55406905513318,
      "tokens_seen": 1329725440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015847542627884,
      "loss": 2.7064,
      "theoretical_loss": 3.5540532523480546,
      "tokens_seen": 1329790976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030157472417251756,
      "loss": 2.7721,
      "theoretical_loss": 3.554037450559772,
      "tokens_seen": 1329856512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030156469408224674,
      "loss": 2.6562,
      "theoretical_loss": 3.554021649768219,
      "tokens_seen": 1329922048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301554663991976,
      "loss": 2.7299,
      "theoretical_loss": 3.5540058499732847,
      "tokens_seen": 1329987584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015446339017051,
      "loss": 2.7501,
      "theoretical_loss": 3.5539900511748566,
      "tokens_seen": 1330053120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030153460381143434,
      "loss": 2.6478,
      "theoretical_loss": 3.553974253372823,
      "tokens_seen": 1330118656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030152457372116347,
      "loss": 2.8084,
      "theoretical_loss": 3.553958456567072,
      "tokens_seen": 1330184192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015145436308927,
      "loss": 2.5444,
      "theoretical_loss": 3.5539426607574915,
      "tokens_seen": 1330249728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015045135406219,
      "loss": 2.2331,
      "theoretical_loss": 3.5539268659439696,
      "tokens_seen": 1330315264
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 734673,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9271119832992554,
      "objective/train/theoretical_loss": 3.553911072126394,
      "objective/train/tokens_used": 1350840800,
      "theoretical_loss": 3.553911072126394,
      "tokens_seen": 1330380800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030149448345035107,
      "loss": 2.3999,
      "theoretical_loss": 3.553911072126394,
      "tokens_seen": 1330380800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030148445336008025,
      "loss": 2.5087,
      "theoretical_loss": 3.553895279304654,
      "tokens_seen": 1330446336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030147442326980943,
      "loss": 2.7373,
      "theoretical_loss": 3.553879487478637,
      "tokens_seen": 1330511872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014643931795386,
      "loss": 2.8742,
      "theoretical_loss": 3.5538636966482313,
      "tokens_seen": 1330577408
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030145436308926784,
      "loss": 2.7182,
      "theoretical_loss": 3.553847906813325,
      "tokens_seen": 1330642944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030144433299899697,
      "loss": 2.5288,
      "theoretical_loss": 3.553832117973806,
      "tokens_seen": 1330708480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014343029087262,
      "loss": 2.5994,
      "theoretical_loss": 3.553816330129563,
      "tokens_seen": 1330774016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030142427281845533,
      "loss": 2.6242,
      "theoretical_loss": 3.553800543280485,
      "tokens_seen": 1330839552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030141424272818457,
      "loss": 2.4113,
      "theoretical_loss": 3.5537847574264587,
      "tokens_seen": 1330905088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030140421263791375,
      "loss": 2.6611,
      "theoretical_loss": 3.553768972567373,
      "tokens_seen": 1330970624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030139418254764293,
      "loss": 2.867,
      "theoretical_loss": 3.553753188703117,
      "tokens_seen": 1331036160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013841524573721,
      "loss": 2.8117,
      "theoretical_loss": 3.5537374058335782,
      "tokens_seen": 1331101696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030137412236710135,
      "loss": 2.6837,
      "theoretical_loss": 3.553721623958645,
      "tokens_seen": 1331167232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013640922768305,
      "loss": 2.5018,
      "theoretical_loss": 3.5537058430782054,
      "tokens_seen": 1331232768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013540621865597,
      "loss": 2.5259,
      "theoretical_loss": 3.553690063192149,
      "tokens_seen": 1331298304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030134403209628884,
      "loss": 2.9599,
      "theoretical_loss": 3.553674284300363,
      "tokens_seen": 1331363840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030133400200601807,
      "loss": 2.8089,
      "theoretical_loss": 3.553658506402736,
      "tokens_seen": 1331429376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030132397191574725,
      "loss": 2.7761,
      "theoretical_loss": 3.553642729499157,
      "tokens_seen": 1331494912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030131394182547643,
      "loss": 2.6749,
      "theoretical_loss": 3.5536269535895144,
      "tokens_seen": 1331560448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013039117352056,
      "loss": 2.6557,
      "theoretical_loss": 3.5536111786736964,
      "tokens_seen": 1331625984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012938816449348,
      "loss": 2.7507,
      "theoretical_loss": 3.553595404751591,
      "tokens_seen": 1331691520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301283851554664,
      "loss": 2.7159,
      "theoretical_loss": 3.5535796318230872,
      "tokens_seen": 1331757056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012738214643932,
      "loss": 2.6725,
      "theoretical_loss": 3.553563859888074,
      "tokens_seen": 1331822592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030126379137412234,
      "loss": 2.4583,
      "theoretical_loss": 3.5535480889464393,
      "tokens_seen": 1331888128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012537612838516,
      "loss": 2.5396,
      "theoretical_loss": 3.5535323189980716,
      "tokens_seen": 1331953664
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 735529,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.719951629638672,
      "objective/train/theoretical_loss": 3.55351655004286,
      "objective/train/tokens_used": 1352479200,
      "theoretical_loss": 3.55351655004286,
      "tokens_seen": 1332019200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012437311935807,
      "loss": 2.7935,
      "theoretical_loss": 3.55351655004286,
      "tokens_seen": 1332019200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030123370110330994,
      "loss": 2.939,
      "theoretical_loss": 3.5535007820806923,
      "tokens_seen": 1332084736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012236710130391,
      "loss": 2.7108,
      "theoretical_loss": 3.553485015111458,
      "tokens_seen": 1332150272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012136409227683,
      "loss": 2.6529,
      "theoretical_loss": 3.5534692491350452,
      "tokens_seen": 1332215808
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012036108324975,
      "loss": 2.8294,
      "theoretical_loss": 3.5534534841513428,
      "tokens_seen": 1332281344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011935807422267,
      "loss": 2.7404,
      "theoretical_loss": 3.553437720160239,
      "tokens_seen": 1332346880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030118355065195584,
      "loss": 2.7931,
      "theoretical_loss": 3.553421957161623,
      "tokens_seen": 1332412416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011735205616851,
      "loss": 2.7708,
      "theoretical_loss": 3.5534061951553837,
      "tokens_seen": 1332477952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011634904714142,
      "loss": 2.6398,
      "theoretical_loss": 3.5533904341414093,
      "tokens_seen": 1332543488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030115346038114344,
      "loss": 2.6252,
      "theoretical_loss": 3.5533746741195884,
      "tokens_seen": 1332609024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011434302908726,
      "loss": 2.6751,
      "theoretical_loss": 3.5533589150898104,
      "tokens_seen": 1332674560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011334002006018,
      "loss": 2.6118,
      "theoretical_loss": 3.5533431570519634,
      "tokens_seen": 1332740096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301123370110331,
      "loss": 2.7787,
      "theoretical_loss": 3.5533274000059363,
      "tokens_seen": 1332805632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030111334002006016,
      "loss": 2.7079,
      "theoretical_loss": 3.5533116439516186,
      "tokens_seen": 1332871168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030110330992978935,
      "loss": 2.7413,
      "theoretical_loss": 3.5532958888888984,
      "tokens_seen": 1332936704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010932798395186,
      "loss": 2.6334,
      "theoretical_loss": 3.5532801348176646,
      "tokens_seen": 1333002240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010832497492477,
      "loss": 2.684,
      "theoretical_loss": 3.5532643817378067,
      "tokens_seen": 1333067776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030107321965897694,
      "loss": 2.5295,
      "theoretical_loss": 3.5532486296492127,
      "tokens_seen": 1333133312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030106318956870607,
      "loss": 2.7846,
      "theoretical_loss": 3.553232878551772,
      "tokens_seen": 1333198848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010531594784353,
      "loss": 2.6775,
      "theoretical_loss": 3.553217128445374,
      "tokens_seen": 1333264384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030104312938816454,
      "loss": 2.6434,
      "theoretical_loss": 3.5532013793299067,
      "tokens_seen": 1333329920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030103309929789367,
      "loss": 2.5548,
      "theoretical_loss": 3.5531856312052597,
      "tokens_seen": 1333395456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010230692076229,
      "loss": 2.7114,
      "theoretical_loss": 3.553169884071321,
      "tokens_seen": 1333460992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010130391173521,
      "loss": 2.6786,
      "theoretical_loss": 3.553154137927981,
      "tokens_seen": 1333526528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030100300902708127,
      "loss": 2.4968,
      "theoretical_loss": 3.5531383927751277,
      "tokens_seen": 1333592064
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 735803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6573450565338135,
      "objective/train/theoretical_loss": 3.5531226486126504,
      "objective/train/tokens_used": 1354117600,
      "theoretical_loss": 3.5531226486126504,
      "tokens_seen": 1333657600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030099297893681045,
      "loss": 2.5315,
      "theoretical_loss": 3.5531226486126504,
      "tokens_seen": 1333657600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030098294884653963,
      "loss": 2.6829,
      "theoretical_loss": 3.5531069054404383,
      "tokens_seen": 1333723136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003009729187562688,
      "loss": 2.8954,
      "theoretical_loss": 3.5530911632583804,
      "tokens_seen": 1333788672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030096288866599804,
      "loss": 2.9845,
      "theoretical_loss": 3.5530754220663656,
      "tokens_seen": 1333854208
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030095285857572717,
      "loss": 2.7962,
      "theoretical_loss": 3.553059681864283,
      "tokens_seen": 1333919744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003009428284854564,
      "loss": 2.6451,
      "theoretical_loss": 3.553043942652022,
      "tokens_seen": 1333985280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030093279839518553,
      "loss": 2.913,
      "theoretical_loss": 3.553028204429471,
      "tokens_seen": 1334050816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030092276830491477,
      "loss": 2.4407,
      "theoretical_loss": 3.5530124671965204,
      "tokens_seen": 1334116352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030091273821464395,
      "loss": 2.6616,
      "theoretical_loss": 3.5529967309530583,
      "tokens_seen": 1334181888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030090270812437313,
      "loss": 2.4699,
      "theoretical_loss": 3.5529809956989746,
      "tokens_seen": 1334247424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008926780341023,
      "loss": 2.6506,
      "theoretical_loss": 3.5529652614341583,
      "tokens_seen": 1334312960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030088264794383155,
      "loss": 2.5312,
      "theoretical_loss": 3.5529495281584977,
      "tokens_seen": 1334378496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008726178535607,
      "loss": 2.53,
      "theoretical_loss": 3.5529337958718834,
      "tokens_seen": 1334444032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008625877632899,
      "loss": 2.5769,
      "theoretical_loss": 3.552918064574204,
      "tokens_seen": 1334509568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030085255767301904,
      "loss": 2.8111,
      "theoretical_loss": 3.5529023342653483,
      "tokens_seen": 1334575104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030084252758274827,
      "loss": 2.7617,
      "theoretical_loss": 3.552886604945207,
      "tokens_seen": 1334640640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030083249749247745,
      "loss": 2.5592,
      "theoretical_loss": 3.5528708766136683,
      "tokens_seen": 1334706176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030082246740220663,
      "loss": 2.5869,
      "theoretical_loss": 3.5528551492706217,
      "tokens_seen": 1334771712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008124373119358,
      "loss": 2.8061,
      "theoretical_loss": 3.552839422915956,
      "tokens_seen": 1334837248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300802407221665,
      "loss": 2.905,
      "theoretical_loss": 3.552823697549562,
      "tokens_seen": 1334902784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007923771313942,
      "loss": 2.7399,
      "theoretical_loss": 3.552807973171328,
      "tokens_seen": 1334968320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007823470411234,
      "loss": 2.8187,
      "theoretical_loss": 3.552792249781144,
      "tokens_seen": 1335033856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030077231695085254,
      "loss": 2.8911,
      "theoretical_loss": 3.5527765273788985,
      "tokens_seen": 1335099392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007622868605818,
      "loss": 2.7133,
      "theoretical_loss": 3.552760805964482,
      "tokens_seen": 1335164928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007522567703109,
      "loss": 2.7799,
      "theoretical_loss": 3.552745085537783,
      "tokens_seen": 1335230464
    },
    {
      "epoch": 0.4,
      "objective/train/docs_used": 736883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6608688831329346,
      "objective/train/theoretical_loss": 3.552729366098692,
      "objective/train/tokens_used": 1355756000,
      "theoretical_loss": 3.552729366098692,
      "tokens_seen": 1335296000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030074222668004014,
      "loss": 2.7968,
      "theoretical_loss": 3.552729366098692,
      "tokens_seen": 1335296000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007321965897693,
      "loss": 2.9199,
      "theoretical_loss": 3.5527136476470975,
      "tokens_seen": 1335361536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007221664994985,
      "loss": 2.7619,
      "theoretical_loss": 3.55269793018289,
      "tokens_seen": 1335427072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007121364092277,
      "loss": 2.5371,
      "theoretical_loss": 3.5526822137059577,
      "tokens_seen": 1335492608
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007021063189569,
      "loss": 2.8803,
      "theoretical_loss": 3.5526664982161913,
      "tokens_seen": 1335558144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030069207622868604,
      "loss": 2.7017,
      "theoretical_loss": 3.5526507837134798,
      "tokens_seen": 1335623680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006820461384153,
      "loss": 2.7533,
      "theoretical_loss": 3.5526350701977134,
      "tokens_seen": 1335689216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006720160481444,
      "loss": 2.6957,
      "theoretical_loss": 3.552619357668781,
      "tokens_seen": 1335754752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030066198595787364,
      "loss": 2.6309,
      "theoretical_loss": 3.552603646126572,
      "tokens_seen": 1335820288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006519558676028,
      "loss": 2.7543,
      "theoretical_loss": 3.552587935570977,
      "tokens_seen": 1335885824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300641925777332,
      "loss": 2.8341,
      "theoretical_loss": 3.5525722260018853,
      "tokens_seen": 1335951360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006318956870612,
      "loss": 2.8162,
      "theoretical_loss": 3.5525565174191858,
      "tokens_seen": 1336016896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030062186559679036,
      "loss": 2.8456,
      "theoretical_loss": 3.5525408098227693,
      "tokens_seen": 1336082432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030061183550651955,
      "loss": 2.6389,
      "theoretical_loss": 3.552525103212525,
      "tokens_seen": 1336147968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006018054162488,
      "loss": 2.6604,
      "theoretical_loss": 3.552509397588342,
      "tokens_seen": 1336213504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005917753259779,
      "loss": 2.9117,
      "theoretical_loss": 3.5524936929501107,
      "tokens_seen": 1336279040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030058174523570714,
      "loss": 2.623,
      "theoretical_loss": 3.5524779892977216,
      "tokens_seen": 1336344576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030057171514543627,
      "loss": 2.7888,
      "theoretical_loss": 3.552462286631063,
      "tokens_seen": 1336410112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005616850551655,
      "loss": 2.7198,
      "theoretical_loss": 3.5524465849500255,
      "tokens_seen": 1336475648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005516549648947,
      "loss": 2.7033,
      "theoretical_loss": 3.552430884254499,
      "tokens_seen": 1336541184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030054162487462387,
      "loss": 2.4967,
      "theoretical_loss": 3.5524151845443726,
      "tokens_seen": 1336606720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030053159478435305,
      "loss": 2.4972,
      "theoretical_loss": 3.552399485819537,
      "tokens_seen": 1336672256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005215646940823,
      "loss": 2.824,
      "theoretical_loss": 3.552383788079881,
      "tokens_seen": 1336737792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005115346038114,
      "loss": 2.9612,
      "theoretical_loss": 3.552368091325296,
      "tokens_seen": 1336803328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030050150451354065,
      "loss": 2.6908,
      "theoretical_loss": 3.552352395555671,
      "tokens_seen": 1336868864
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 737387,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.81619930267334,
      "objective/train/theoretical_loss": 3.552336700770896,
      "objective/train/tokens_used": 1357394400,
      "theoretical_loss": 3.552336700770896,
      "tokens_seen": 1336934400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004914744232698,
      "loss": 2.7363,
      "theoretical_loss": 3.552336700770896,
      "tokens_seen": 1336934400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300481444332999,
      "loss": 2.797,
      "theoretical_loss": 3.5523210069708604,
      "tokens_seen": 1336999936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004714142427282,
      "loss": 2.6272,
      "theoretical_loss": 3.552305314155455,
      "tokens_seen": 1337065472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030046138415245737,
      "loss": 2.8724,
      "theoretical_loss": 3.5522896223245697,
      "tokens_seen": 1337131008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030045135406218655,
      "loss": 2.7165,
      "theoretical_loss": 3.5522739314780942,
      "tokens_seen": 1337196544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030044132397191573,
      "loss": 2.9079,
      "theoretical_loss": 3.5522582416159176,
      "tokens_seen": 1337262080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004312938816449,
      "loss": 2.7132,
      "theoretical_loss": 3.552242552737932,
      "tokens_seen": 1337327616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030042126379137415,
      "loss": 2.9322,
      "theoretical_loss": 3.552226864844026,
      "tokens_seen": 1337393152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004112337011033,
      "loss": 2.8038,
      "theoretical_loss": 3.5522111779340895,
      "tokens_seen": 1337458688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004012036108325,
      "loss": 2.7096,
      "theoretical_loss": 3.552195492008013,
      "tokens_seen": 1337524224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003911735205617,
      "loss": 2.8169,
      "theoretical_loss": 3.552179807065687,
      "tokens_seen": 1337589760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003811434302909,
      "loss": 2.6221,
      "theoretical_loss": 3.5521641231070014,
      "tokens_seen": 1337655296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030037111334002006,
      "loss": 2.737,
      "theoretical_loss": 3.5521484401318455,
      "tokens_seen": 1337720832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030036108324974924,
      "loss": 2.7672,
      "theoretical_loss": 3.55213275814011,
      "tokens_seen": 1337786368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003510531594784,
      "loss": 2.9049,
      "theoretical_loss": 3.552117077131686,
      "tokens_seen": 1337851904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030034102306920765,
      "loss": 2.7507,
      "theoretical_loss": 3.552101397106462,
      "tokens_seen": 1337917440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003309929789368,
      "loss": 2.6721,
      "theoretical_loss": 3.5520857180643297,
      "tokens_seen": 1337982976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300320962888666,
      "loss": 2.6487,
      "theoretical_loss": 3.5520700400051783,
      "tokens_seen": 1338048512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030031093279839514,
      "loss": 2.6749,
      "theoretical_loss": 3.5520543629288976,
      "tokens_seen": 1338114048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003009027081244,
      "loss": 2.859,
      "theoretical_loss": 3.55203868683538,
      "tokens_seen": 1338179584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002908726178536,
      "loss": 2.8069,
      "theoretical_loss": 3.5520230117245135,
      "tokens_seen": 1338245120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030028084252758274,
      "loss": 2.6247,
      "theoretical_loss": 3.5520073375961894,
      "tokens_seen": 1338310656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300270812437312,
      "loss": 2.5632,
      "theoretical_loss": 3.551991664450298,
      "tokens_seen": 1338376192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002607823470411,
      "loss": 2.7317,
      "theoretical_loss": 3.5519759922867298,
      "tokens_seen": 1338441728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030025075225677034,
      "loss": 2.8898,
      "theoretical_loss": 3.5519603211053745,
      "tokens_seen": 1338507264
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 738438,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9319119453430176,
      "objective/train/theoretical_loss": 3.5519446509061225,
      "objective/train/tokens_used": 1359032800,
      "theoretical_loss": 3.5519446509061225,
      "tokens_seen": 1338572800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002407221664995,
      "loss": 2.7252,
      "theoretical_loss": 3.5519446509061225,
      "tokens_seen": 1338572800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002306920762287,
      "loss": 2.7483,
      "theoretical_loss": 3.5519289816888646,
      "tokens_seen": 1338638336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002206619859579,
      "loss": 2.7988,
      "theoretical_loss": 3.5519133134534915,
      "tokens_seen": 1338703872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002106318956871,
      "loss": 2.601,
      "theoretical_loss": 3.5518976461998926,
      "tokens_seen": 1338769408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030020060180541624,
      "loss": 2.6018,
      "theoretical_loss": 3.551881979927959,
      "tokens_seen": 1338834944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001905717151455,
      "loss": 2.8197,
      "theoretical_loss": 3.5518663146375813,
      "tokens_seen": 1338900480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001805416248746,
      "loss": 2.6897,
      "theoretical_loss": 3.5518506503286496,
      "tokens_seen": 1338966016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030017051153460384,
      "loss": 2.6883,
      "theoretical_loss": 3.551834987001054,
      "tokens_seen": 1339031552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300160481444333,
      "loss": 2.759,
      "theoretical_loss": 3.5518193246546863,
      "tokens_seen": 1339097088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001504513540622,
      "loss": 2.5986,
      "theoretical_loss": 3.5518036632894354,
      "tokens_seen": 1339162624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001404212637914,
      "loss": 2.682,
      "theoretical_loss": 3.551788002905193,
      "tokens_seen": 1339228160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030013039117352056,
      "loss": 2.6333,
      "theoretical_loss": 3.551772343501849,
      "tokens_seen": 1339293696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030012036108324975,
      "loss": 2.4821,
      "theoretical_loss": 3.5517566850792948,
      "tokens_seen": 1339359232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300110330992979,
      "loss": 2.5707,
      "theoretical_loss": 3.5517410276374197,
      "tokens_seen": 1339424768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001003009027081,
      "loss": 2.7059,
      "theoretical_loss": 3.5517253711761154,
      "tokens_seen": 1339490304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030009027081243734,
      "loss": 2.658,
      "theoretical_loss": 3.551709715695272,
      "tokens_seen": 1339555840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030008024072216647,
      "loss": 2.5603,
      "theoretical_loss": 3.5516940611947803,
      "tokens_seen": 1339621376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000702106318957,
      "loss": 2.7902,
      "theoretical_loss": 3.5516784076745314,
      "tokens_seen": 1339686912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000601805416249,
      "loss": 2.5795,
      "theoretical_loss": 3.551662755134415,
      "tokens_seen": 1339752448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030005015045135407,
      "loss": 2.6337,
      "theoretical_loss": 3.5516471035743225,
      "tokens_seen": 1339817984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030004012036108325,
      "loss": 2.6817,
      "theoretical_loss": 3.551631452994144,
      "tokens_seen": 1339883520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000300902708125,
      "loss": 2.6522,
      "theoretical_loss": 3.5516158033937706,
      "tokens_seen": 1339949056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000200601805416,
      "loss": 2.9809,
      "theoretical_loss": 3.5516001547730935,
      "tokens_seen": 1340014592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030001003009027085,
      "loss": 2.7736,
      "theoretical_loss": 3.5515845071320027,
      "tokens_seen": 1340080128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003,
      "loss": 2.7429,
      "theoretical_loss": 3.551568860470389,
      "tokens_seen": 1340145664
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 739370,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.132230043411255,
      "objective/train/theoretical_loss": 3.5515532147881443,
      "objective/train/tokens_used": 1360671200,
      "theoretical_loss": 3.5515532147881443,
      "tokens_seen": 1340211200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999899699097292,
      "loss": 2.9083,
      "theoretical_loss": 3.5515532147881443,
      "tokens_seen": 1340211200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999799398194584,
      "loss": 2.6438,
      "theoretical_loss": 3.5515375700851584,
      "tokens_seen": 1340276736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029996990972918757,
      "loss": 2.7103,
      "theoretical_loss": 3.551521926361322,
      "tokens_seen": 1340342272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029995987963891675,
      "loss": 2.5785,
      "theoretical_loss": 3.551506283616526,
      "tokens_seen": 1340407808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029994984954864593,
      "loss": 2.816,
      "theoretical_loss": 3.551490641850662,
      "tokens_seen": 1340473344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999398194583751,
      "loss": 2.5375,
      "theoretical_loss": 3.5514750010636202,
      "tokens_seen": 1340538880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029992978936810435,
      "loss": 2.6688,
      "theoretical_loss": 3.551459361255292,
      "tokens_seen": 1340604416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999197592778335,
      "loss": 2.9732,
      "theoretical_loss": 3.551443722425568,
      "tokens_seen": 1340669952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999097291875627,
      "loss": 2.5551,
      "theoretical_loss": 3.551428084574339,
      "tokens_seen": 1340735488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998996990972919,
      "loss": 2.9285,
      "theoretical_loss": 3.5514124477014963,
      "tokens_seen": 1340801024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998896690070211,
      "loss": 2.672,
      "theoretical_loss": 3.5513968118069306,
      "tokens_seen": 1340866560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029987963891675026,
      "loss": 2.6653,
      "theoretical_loss": 3.5513811768905335,
      "tokens_seen": 1340932096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029986960882647944,
      "loss": 2.644,
      "theoretical_loss": 3.551365542952195,
      "tokens_seen": 1340997632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998595787362086,
      "loss": 2.7729,
      "theoretical_loss": 3.5513499099918064,
      "tokens_seen": 1341063168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029984954864593785,
      "loss": 2.9094,
      "theoretical_loss": 3.551334278009259,
      "tokens_seen": 1341128704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299839518555667,
      "loss": 2.8276,
      "theoretical_loss": 3.551318647004444,
      "tokens_seen": 1341194240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998294884653962,
      "loss": 2.7759,
      "theoretical_loss": 3.5513030169772524,
      "tokens_seen": 1341259776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029981945837512534,
      "loss": 2.7611,
      "theoretical_loss": 3.551287387927575,
      "tokens_seen": 1341325312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998094282848546,
      "loss": 2.952,
      "theoretical_loss": 3.551271759855303,
      "tokens_seen": 1341390848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029979939819458376,
      "loss": 2.5767,
      "theoretical_loss": 3.5512561327603276,
      "tokens_seen": 1341456384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029978936810431294,
      "loss": 2.703,
      "theoretical_loss": 3.5512405066425403,
      "tokens_seen": 1341521920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997793380140421,
      "loss": 2.6325,
      "theoretical_loss": 3.551224881501832,
      "tokens_seen": 1341587456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997693079237713,
      "loss": 2.702,
      "theoretical_loss": 3.551209257338093,
      "tokens_seen": 1341652992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997592778335005,
      "loss": 2.6736,
      "theoretical_loss": 3.5511936341512156,
      "tokens_seen": 1341718528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997492477432297,
      "loss": 2.6862,
      "theoretical_loss": 3.551178011941091,
      "tokens_seen": 1341784064
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 740011,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.746211528778076,
      "objective/train/theoretical_loss": 3.55116239070761,
      "objective/train/tokens_used": 1362309600,
      "theoretical_loss": 3.55116239070761,
      "tokens_seen": 1341849600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029973921765295885,
      "loss": 2.5824,
      "theoretical_loss": 3.55116239070761,
      "tokens_seen": 1341849600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997291875626881,
      "loss": 2.8266,
      "theoretical_loss": 3.5511467704506643,
      "tokens_seen": 1341915136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029971915747241726,
      "loss": 2.6679,
      "theoretical_loss": 3.5511311511701447,
      "tokens_seen": 1341980672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029970912738214644,
      "loss": 2.541,
      "theoretical_loss": 3.5511155328659427,
      "tokens_seen": 1342046208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996990972918756,
      "loss": 2.7126,
      "theoretical_loss": 3.5510999155379492,
      "tokens_seen": 1342111744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996890672016048,
      "loss": 2.6137,
      "theoretical_loss": 3.551084299186056,
      "tokens_seen": 1342177280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299679037111334,
      "loss": 2.5994,
      "theoretical_loss": 3.551068683810154,
      "tokens_seen": 1342242816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996690070210632,
      "loss": 2.7638,
      "theoretical_loss": 3.5510530694101354,
      "tokens_seen": 1342308352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029965897693079235,
      "loss": 2.6661,
      "theoretical_loss": 3.551037455985891,
      "tokens_seen": 1342373888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996489468405216,
      "loss": 2.9079,
      "theoretical_loss": 3.5510218435373115,
      "tokens_seen": 1342439424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996389167502507,
      "loss": 2.7773,
      "theoretical_loss": 3.5510062320642897,
      "tokens_seen": 1342504960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029962888665997995,
      "loss": 2.8018,
      "theoretical_loss": 3.550990621566716,
      "tokens_seen": 1342570496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029961885656970913,
      "loss": 2.7557,
      "theoretical_loss": 3.550975012044482,
      "tokens_seen": 1342636032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996088264794383,
      "loss": 2.7851,
      "theoretical_loss": 3.55095940349748,
      "tokens_seen": 1342701568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995987963891675,
      "loss": 2.8343,
      "theoretical_loss": 3.5509437959256003,
      "tokens_seen": 1342767104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029958876629889667,
      "loss": 2.6933,
      "theoretical_loss": 3.550928189328735,
      "tokens_seen": 1342832640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029957873620862585,
      "loss": 2.8134,
      "theoretical_loss": 3.5509125837067756,
      "tokens_seen": 1342898176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995687061183551,
      "loss": 2.566,
      "theoretical_loss": 3.5508969790596137,
      "tokens_seen": 1342963712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995586760280842,
      "loss": 2.6247,
      "theoretical_loss": 3.5508813753871404,
      "tokens_seen": 1343029248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029954864593781345,
      "loss": 2.6155,
      "theoretical_loss": 3.5508657726892476,
      "tokens_seen": 1343094784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995386158475427,
      "loss": 2.5773,
      "theoretical_loss": 3.550850170965827,
      "tokens_seen": 1343160320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995285857572718,
      "loss": 2.9044,
      "theoretical_loss": 3.55083457021677,
      "tokens_seen": 1343225856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029951855566700105,
      "loss": 2.5965,
      "theoretical_loss": 3.550818970441968,
      "tokens_seen": 1343291392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995085255767302,
      "loss": 2.7735,
      "theoretical_loss": 3.5508033716413125,
      "tokens_seen": 1343356928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994984954864594,
      "loss": 2.6961,
      "theoretical_loss": 3.5507877738146965,
      "tokens_seen": 1343422464
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 740980,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.549888849258423,
      "objective/train/theoretical_loss": 3.5507721769620098,
      "objective/train/tokens_used": 1363948000,
      "theoretical_loss": 3.5507721769620098,
      "tokens_seen": 1343488000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994884653961886,
      "loss": 2.7331,
      "theoretical_loss": 3.5507721769620098,
      "tokens_seen": 1343488000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029947843530591777,
      "loss": 2.835,
      "theoretical_loss": 3.5507565810831454,
      "tokens_seen": 1343553536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029946840521564695,
      "loss": 2.7311,
      "theoretical_loss": 3.5507409861779946,
      "tokens_seen": 1343619072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029945837512537613,
      "loss": 2.8276,
      "theoretical_loss": 3.5507253922464486,
      "tokens_seen": 1343684608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994483450351053,
      "loss": 2.7563,
      "theoretical_loss": 3.5507097992884002,
      "tokens_seen": 1343750144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029943831494483455,
      "loss": 2.7471,
      "theoretical_loss": 3.5506942073037404,
      "tokens_seen": 1343815680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994282848545637,
      "loss": 2.637,
      "theoretical_loss": 3.550678616292361,
      "tokens_seen": 1343881216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994182547642929,
      "loss": 2.6875,
      "theoretical_loss": 3.550663026254154,
      "tokens_seen": 1343946752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994082246740221,
      "loss": 2.6548,
      "theoretical_loss": 3.5506474371890113,
      "tokens_seen": 1344012288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993981945837513,
      "loss": 2.7173,
      "theoretical_loss": 3.5506318490968245,
      "tokens_seen": 1344077824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029938816449348046,
      "loss": 2.759,
      "theoretical_loss": 3.5506162619774857,
      "tokens_seen": 1344143360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029937813440320964,
      "loss": 2.7381,
      "theoretical_loss": 3.5506006758308857,
      "tokens_seen": 1344208896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993681043129388,
      "loss": 2.7894,
      "theoretical_loss": 3.5505850906569183,
      "tokens_seen": 1344274432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029935807422266805,
      "loss": 2.6716,
      "theoretical_loss": 3.5505695064554734,
      "tokens_seen": 1344339968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993480441323972,
      "loss": 2.7826,
      "theoretical_loss": 3.5505539232264445,
      "tokens_seen": 1344405504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993380140421264,
      "loss": 2.8918,
      "theoretical_loss": 3.550538340969722,
      "tokens_seen": 1344471040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029932798395185554,
      "loss": 2.6461,
      "theoretical_loss": 3.5505227596851996,
      "tokens_seen": 1344536576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993179538615848,
      "loss": 2.8538,
      "theoretical_loss": 3.550507179372768,
      "tokens_seen": 1344602112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029930792377131396,
      "loss": 2.5681,
      "theoretical_loss": 3.5504916000323195,
      "tokens_seen": 1344667648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029929789368104314,
      "loss": 2.5183,
      "theoretical_loss": 3.550476021663746,
      "tokens_seen": 1344733184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992878635907723,
      "loss": 2.8899,
      "theoretical_loss": 3.55046044426694,
      "tokens_seen": 1344798720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992778335005015,
      "loss": 2.8328,
      "theoretical_loss": 3.550444867841793,
      "tokens_seen": 1344864256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992678034102307,
      "loss": 2.7164,
      "theoretical_loss": 3.550429292388197,
      "tokens_seen": 1344929792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992577733199599,
      "loss": 2.7924,
      "theoretical_loss": 3.5504137179060447,
      "tokens_seen": 1344995328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029924774322968905,
      "loss": 2.8093,
      "theoretical_loss": 3.5503981443952273,
      "tokens_seen": 1345060864
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 742066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.678600311279297,
      "objective/train/theoretical_loss": 3.5503825718556374,
      "objective/train/tokens_used": 1365586400,
      "theoretical_loss": 3.5503825718556374,
      "tokens_seen": 1345126400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992377131394183,
      "loss": 2.6478,
      "theoretical_loss": 3.5503825718556374,
      "tokens_seen": 1345126400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029922768304914746,
      "loss": 2.9484,
      "theoretical_loss": 3.5503670002871672,
      "tokens_seen": 1345191936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029921765295887664,
      "loss": 2.7338,
      "theoretical_loss": 3.5503514296897087,
      "tokens_seen": 1345257472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992076228686058,
      "loss": 2.8106,
      "theoretical_loss": 3.550335860063154,
      "tokens_seen": 1345323008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299197592778335,
      "loss": 2.5024,
      "theoretical_loss": 3.5503202914073952,
      "tokens_seen": 1345388544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991875626880642,
      "loss": 2.7086,
      "theoretical_loss": 3.5503047237223253,
      "tokens_seen": 1345454080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991775325977934,
      "loss": 2.5144,
      "theoretical_loss": 3.5502891570078354,
      "tokens_seen": 1345519616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029916750250752255,
      "loss": 2.7738,
      "theoretical_loss": 3.5502735912638177,
      "tokens_seen": 1345585152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991574724172518,
      "loss": 2.7315,
      "theoretical_loss": 3.550258026490165,
      "tokens_seen": 1345650688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991474423269809,
      "loss": 2.6275,
      "theoretical_loss": 3.5502424626867697,
      "tokens_seen": 1345716224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029913741223671015,
      "loss": 2.2516,
      "theoretical_loss": 3.5502268998535236,
      "tokens_seen": 1345781760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029912738214643933,
      "loss": 2.7417,
      "theoretical_loss": 3.5502113379903193,
      "tokens_seen": 1345847296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991173520561685,
      "loss": 2.6809,
      "theoretical_loss": 3.5501957770970494,
      "tokens_seen": 1345912832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991073219658977,
      "loss": 2.5831,
      "theoretical_loss": 3.550180217173605,
      "tokens_seen": 1345978368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029909729187562687,
      "loss": 2.5846,
      "theoretical_loss": 3.5501646582198796,
      "tokens_seen": 1346043904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029908726178535605,
      "loss": 2.4565,
      "theoretical_loss": 3.550149100235765,
      "tokens_seen": 1346109440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990772316950853,
      "loss": 2.7977,
      "theoretical_loss": 3.550133543221154,
      "tokens_seen": 1346174976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990672016048144,
      "loss": 2.746,
      "theoretical_loss": 3.5501179871759385,
      "tokens_seen": 1346240512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029905717151454365,
      "loss": 2.8822,
      "theoretical_loss": 3.5501024321000116,
      "tokens_seen": 1346306048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029904714142427283,
      "loss": 2.4634,
      "theoretical_loss": 3.550086877993265,
      "tokens_seen": 1346371584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299037111334002,
      "loss": 2.799,
      "theoretical_loss": 3.5500713248555913,
      "tokens_seen": 1346437120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990270812437312,
      "loss": 2.738,
      "theoretical_loss": 3.5500557726868838,
      "tokens_seen": 1346502656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990170511534604,
      "loss": 2.6585,
      "theoretical_loss": 3.5500402214870332,
      "tokens_seen": 1346568192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029900702106318956,
      "loss": 2.7074,
      "theoretical_loss": 3.550024671255934,
      "tokens_seen": 1346633728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989969909729188,
      "loss": 2.8023,
      "theoretical_loss": 3.550009121993477,
      "tokens_seen": 1346699264
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 742723,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1151106357574463,
      "objective/train/theoretical_loss": 3.549993573699556,
      "objective/train/tokens_used": 1367224800,
      "theoretical_loss": 3.549993573699556,
      "tokens_seen": 1346764800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989869608826479,
      "loss": 2.9029,
      "theoretical_loss": 3.549993573699556,
      "tokens_seen": 1346764800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029897693079237715,
      "loss": 2.801,
      "theoretical_loss": 3.5499780263740632,
      "tokens_seen": 1346830336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989669007021063,
      "loss": 2.5526,
      "theoretical_loss": 3.549962480016891,
      "tokens_seen": 1346895872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989568706118355,
      "loss": 2.8075,
      "theoretical_loss": 3.5499469346279318,
      "tokens_seen": 1346961408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989468405215647,
      "loss": 2.8386,
      "theoretical_loss": 3.549931390207078,
      "tokens_seen": 1347026944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989368104312939,
      "loss": 2.6807,
      "theoretical_loss": 3.5499158467542227,
      "tokens_seen": 1347092480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029892678034102306,
      "loss": 2.5482,
      "theoretical_loss": 3.5499003042692587,
      "tokens_seen": 1347158016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989167502507523,
      "loss": 2.6735,
      "theoretical_loss": 3.549884762752079,
      "tokens_seen": 1347223552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989067201604814,
      "loss": 2.716,
      "theoretical_loss": 3.5498692222025747,
      "tokens_seen": 1347289088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029889669007021066,
      "loss": 2.7231,
      "theoretical_loss": 3.5498536826206397,
      "tokens_seen": 1347354624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988866599799398,
      "loss": 2.6743,
      "theoretical_loss": 3.5498381440061664,
      "tokens_seen": 1347420160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298876629889669,
      "loss": 2.336,
      "theoretical_loss": 3.5498226063590477,
      "tokens_seen": 1347485696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988665997993982,
      "loss": 2.6778,
      "theoretical_loss": 3.549807069679176,
      "tokens_seen": 1347551232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988565697091274,
      "loss": 2.6835,
      "theoretical_loss": 3.5497915339664443,
      "tokens_seen": 1347616768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029884653961885656,
      "loss": 2.7947,
      "theoretical_loss": 3.5497759992207456,
      "tokens_seen": 1347682304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029883650952858574,
      "loss": 2.5446,
      "theoretical_loss": 3.549760465441972,
      "tokens_seen": 1347747840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988264794383149,
      "loss": 2.7514,
      "theoretical_loss": 3.549744932630017,
      "tokens_seen": 1347813376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029881644934804416,
      "loss": 2.7718,
      "theoretical_loss": 3.5497294007847726,
      "tokens_seen": 1347878912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988064192577733,
      "loss": 2.6986,
      "theoretical_loss": 3.5497138699061326,
      "tokens_seen": 1347944448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987963891675025,
      "loss": 2.8023,
      "theoretical_loss": 3.5496983399939888,
      "tokens_seen": 1348009984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987863590772317,
      "loss": 2.6328,
      "theoretical_loss": 3.5496828110482355,
      "tokens_seen": 1348075520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987763289869609,
      "loss": 2.6734,
      "theoretical_loss": 3.5496672830687643,
      "tokens_seen": 1348141056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987662988966901,
      "loss": 2.5762,
      "theoretical_loss": 3.549651756055469,
      "tokens_seen": 1348206592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029875626880641925,
      "loss": 2.7513,
      "theoretical_loss": 3.5496362300082414,
      "tokens_seen": 1348272128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987462387161485,
      "loss": 3.0143,
      "theoretical_loss": 3.5496207049269755,
      "tokens_seen": 1348337664
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 744066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.647536277770996,
      "objective/train/theoretical_loss": 3.5496051808115636,
      "objective/train/tokens_used": 1368863200,
      "theoretical_loss": 3.5496051808115636,
      "tokens_seen": 1348403200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029873620862587766,
      "loss": 2.4889,
      "theoretical_loss": 3.5496051808115636,
      "tokens_seen": 1348403200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029872617853560684,
      "loss": 2.4901,
      "theoretical_loss": 3.5495896576618993,
      "tokens_seen": 1348468736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298716148445336,
      "loss": 2.68,
      "theoretical_loss": 3.549574135477875,
      "tokens_seen": 1348534272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987061183550652,
      "loss": 2.5606,
      "theoretical_loss": 3.549558614259384,
      "tokens_seen": 1348599808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986960882647944,
      "loss": 2.7406,
      "theoretical_loss": 3.549543094006319,
      "tokens_seen": 1348665344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986860581745236,
      "loss": 2.659,
      "theoretical_loss": 3.549527574718574,
      "tokens_seen": 1348730880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029867602808425275,
      "loss": 2.9053,
      "theoretical_loss": 3.5495120563960407,
      "tokens_seen": 1348796416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298665997993982,
      "loss": 2.4369,
      "theoretical_loss": 3.5494965390386133,
      "tokens_seen": 1348861952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986559679037111,
      "loss": 2.6933,
      "theoretical_loss": 3.5494810226461846,
      "tokens_seen": 1348927488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029864593781344035,
      "loss": 2.5889,
      "theoretical_loss": 3.549465507218647,
      "tokens_seen": 1348993024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029863590772316953,
      "loss": 2.8407,
      "theoretical_loss": 3.5494499927558945,
      "tokens_seen": 1349058560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986258776328987,
      "loss": 2.5919,
      "theoretical_loss": 3.54943447925782,
      "tokens_seen": 1349124096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986158475426279,
      "loss": 2.6487,
      "theoretical_loss": 3.549418966724316,
      "tokens_seen": 1349189632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029860581745235707,
      "loss": 2.9139,
      "theoretical_loss": 3.549403455155277,
      "tokens_seen": 1349255168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029859578736208625,
      "loss": 2.617,
      "theoretical_loss": 3.549387944550595,
      "tokens_seen": 1349320704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985857572718155,
      "loss": 2.6767,
      "theoretical_loss": 3.549372434910164,
      "tokens_seen": 1349386240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985757271815446,
      "loss": 2.6744,
      "theoretical_loss": 3.549356926233876,
      "tokens_seen": 1349451776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029856569709127385,
      "loss": 2.6996,
      "theoretical_loss": 3.5493414185216263,
      "tokens_seen": 1349517312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029855566700100303,
      "loss": 2.6145,
      "theoretical_loss": 3.5493259117733063,
      "tokens_seen": 1349582848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985456369107322,
      "loss": 2.7506,
      "theoretical_loss": 3.54931040598881,
      "tokens_seen": 1349648384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985356068204614,
      "loss": 2.7514,
      "theoretical_loss": 3.549294901168031,
      "tokens_seen": 1349713920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985255767301906,
      "loss": 2.6993,
      "theoretical_loss": 3.549279397310862,
      "tokens_seen": 1349779456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029851554663991976,
      "loss": 2.4447,
      "theoretical_loss": 3.549263894417197,
      "tokens_seen": 1349844992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298505516549649,
      "loss": 2.6124,
      "theoretical_loss": 3.5492483924869287,
      "tokens_seen": 1349910528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984954864593781,
      "loss": 2.5441,
      "theoretical_loss": 3.5492328915199503,
      "tokens_seen": 1349976064
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 744638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0549464225769043,
      "objective/train/theoretical_loss": 3.5492173915161565,
      "objective/train/tokens_used": 1370501600,
      "theoretical_loss": 3.5492173915161565,
      "tokens_seen": 1350041600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029848545636910735,
      "loss": 3.036,
      "theoretical_loss": 3.5492173915161565,
      "tokens_seen": 1350041600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984754262788365,
      "loss": 2.7784,
      "theoretical_loss": 3.5492018924754394,
      "tokens_seen": 1350107136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984653961885657,
      "loss": 2.6132,
      "theoretical_loss": 3.549186394397693,
      "tokens_seen": 1350172672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984553660982949,
      "loss": 2.9592,
      "theoretical_loss": 3.54917089728281,
      "tokens_seen": 1350238208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984453360080241,
      "loss": 2.4644,
      "theoretical_loss": 3.549155401130685,
      "tokens_seen": 1350303744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029843530591775326,
      "loss": 2.7715,
      "theoretical_loss": 3.5491399059412103,
      "tokens_seen": 1350369280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984252758274825,
      "loss": 2.7839,
      "theoretical_loss": 3.5491244117142804,
      "tokens_seen": 1350434816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984152457372116,
      "loss": 2.6439,
      "theoretical_loss": 3.549108918449788,
      "tokens_seen": 1350500352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029840521564694086,
      "loss": 2.6755,
      "theoretical_loss": 3.5490934261476275,
      "tokens_seen": 1350565888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029839518555667,
      "loss": 2.6529,
      "theoretical_loss": 3.5490779348076917,
      "tokens_seen": 1350631424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983851554663992,
      "loss": 2.3959,
      "theoretical_loss": 3.5490624444298744,
      "tokens_seen": 1350696960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983751253761284,
      "loss": 2.5926,
      "theoretical_loss": 3.5490469550140693,
      "tokens_seen": 1350762496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983650952858576,
      "loss": 2.8891,
      "theoretical_loss": 3.5490314665601694,
      "tokens_seen": 1350828032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029835506519558676,
      "loss": 2.5595,
      "theoretical_loss": 3.5490159790680686,
      "tokens_seen": 1350893568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029834503510531594,
      "loss": 2.7735,
      "theoretical_loss": 3.549000492537661,
      "tokens_seen": 1350959104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983350050150451,
      "loss": 2.6252,
      "theoretical_loss": 3.54898500696884,
      "tokens_seen": 1351024640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029832497492477436,
      "loss": 2.6226,
      "theoretical_loss": 3.5489695223614985,
      "tokens_seen": 1351090176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983149448345035,
      "loss": 2.7547,
      "theoretical_loss": 3.5489540387155314,
      "tokens_seen": 1351155712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983049147442327,
      "loss": 2.6143,
      "theoretical_loss": 3.5489385560308317,
      "tokens_seen": 1351221248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029829488465396185,
      "loss": 2.5784,
      "theoretical_loss": 3.548923074307293,
      "tokens_seen": 1351286784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982848545636911,
      "loss": 2.8404,
      "theoretical_loss": 3.548907593544809,
      "tokens_seen": 1351352320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029827482447342026,
      "loss": 2.618,
      "theoretical_loss": 3.548892113743274,
      "tokens_seen": 1351417856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029826479438314945,
      "loss": 2.8315,
      "theoretical_loss": 3.548876634902581,
      "tokens_seen": 1351483392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982547642928786,
      "loss": 2.9895,
      "theoretical_loss": 3.5488611570226247,
      "tokens_seen": 1351548928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029824473420260786,
      "loss": 2.4631,
      "theoretical_loss": 3.548845680103298,
      "tokens_seen": 1351614464
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 745869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0698626041412354,
      "objective/train/theoretical_loss": 3.5488302041444952,
      "objective/train/tokens_used": 1372140000,
      "theoretical_loss": 3.5488302041444952,
      "tokens_seen": 1351680000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298234704112337,
      "loss": 2.8399,
      "theoretical_loss": 3.5488302041444952,
      "tokens_seen": 1351680000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982246740220662,
      "loss": 2.6906,
      "theoretical_loss": 3.54881472914611,
      "tokens_seen": 1351745536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029821464393179535,
      "loss": 2.7309,
      "theoretical_loss": 3.5487992551080363,
      "tokens_seen": 1351811072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982046138415246,
      "loss": 2.728,
      "theoretical_loss": 3.548783782030168,
      "tokens_seen": 1351876608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029819458375125377,
      "loss": 2.6026,
      "theoretical_loss": 3.5487683099123983,
      "tokens_seen": 1351942144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029818455366098295,
      "loss": 2.4077,
      "theoretical_loss": 3.5487528387546217,
      "tokens_seen": 1352007680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029817452357071213,
      "loss": 2.8385,
      "theoretical_loss": 3.5487373685567327,
      "tokens_seen": 1352073216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981644934804413,
      "loss": 2.7771,
      "theoretical_loss": 3.548721899318624,
      "tokens_seen": 1352138752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981544633901705,
      "loss": 2.7378,
      "theoretical_loss": 3.5487064310401903,
      "tokens_seen": 1352204288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029814443329989973,
      "loss": 2.6337,
      "theoretical_loss": 3.5486909637213255,
      "tokens_seen": 1352269824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029813440320962885,
      "loss": 2.7281,
      "theoretical_loss": 3.548675497361923,
      "tokens_seen": 1352335360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981243731193581,
      "loss": 2.8302,
      "theoretical_loss": 3.5486600319618775,
      "tokens_seen": 1352400896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981143430290872,
      "loss": 2.7914,
      "theoretical_loss": 3.548644567521083,
      "tokens_seen": 1352466432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029810431293881645,
      "loss": 2.9081,
      "theoretical_loss": 3.5486291040394327,
      "tokens_seen": 1352531968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029809428284854563,
      "loss": 2.7718,
      "theoretical_loss": 3.5486136415168215,
      "tokens_seen": 1352597504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980842527582748,
      "loss": 2.6835,
      "theoretical_loss": 3.548598179953143,
      "tokens_seen": 1352663040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298074222668004,
      "loss": 2.7298,
      "theoretical_loss": 3.5485827193482917,
      "tokens_seen": 1352728576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029806419257773323,
      "loss": 2.573,
      "theoretical_loss": 3.5485672597021614,
      "tokens_seen": 1352794112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029805416248746236,
      "loss": 2.6888,
      "theoretical_loss": 3.548551801014646,
      "tokens_seen": 1352859648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980441323971916,
      "loss": 2.6973,
      "theoretical_loss": 3.54853634328564,
      "tokens_seen": 1352925184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980341023069208,
      "loss": 2.5712,
      "theoretical_loss": 3.5485208865150373,
      "tokens_seen": 1352990720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029802407221664996,
      "loss": 2.6709,
      "theoretical_loss": 3.548505430702732,
      "tokens_seen": 1353056256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980140421263792,
      "loss": 2.4174,
      "theoretical_loss": 3.548489975848619,
      "tokens_seen": 1353121792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980040120361083,
      "loss": 2.7311,
      "theoretical_loss": 3.548474521952591,
      "tokens_seen": 1353187328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029799398194583755,
      "loss": 2.6042,
      "theoretical_loss": 3.548459069014544,
      "tokens_seen": 1353252864
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 746951,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.679593801498413,
      "objective/train/theoretical_loss": 3.548443617034371,
      "objective/train/tokens_used": 1373778400,
      "theoretical_loss": 3.548443617034371,
      "tokens_seen": 1353318400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979839518555667,
      "loss": 2.6849,
      "theoretical_loss": 3.548443617034371,
      "tokens_seen": 1353318400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979739217652959,
      "loss": 2.6242,
      "theoretical_loss": 3.5484281660119668,
      "tokens_seen": 1353383936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979638916750251,
      "loss": 2.6791,
      "theoretical_loss": 3.5484127159472254,
      "tokens_seen": 1353449472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979538615847543,
      "loss": 2.691,
      "theoretical_loss": 3.548397266840041,
      "tokens_seen": 1353515008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029794383149448346,
      "loss": 2.6914,
      "theoretical_loss": 3.5483818186903084,
      "tokens_seen": 1353580544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979338014042127,
      "loss": 2.5739,
      "theoretical_loss": 3.548366371497921,
      "tokens_seen": 1353646080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979237713139418,
      "loss": 2.7162,
      "theoretical_loss": 3.548350925262774,
      "tokens_seen": 1353711616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029791374122367106,
      "loss": 2.707,
      "theoretical_loss": 3.548335479984761,
      "tokens_seen": 1353777152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979037111334002,
      "loss": 2.5863,
      "theoretical_loss": 3.548320035663777,
      "tokens_seen": 1353842688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978936810431294,
      "loss": 2.7147,
      "theoretical_loss": 3.5483045922997167,
      "tokens_seen": 1353908224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978836509528586,
      "loss": 2.7314,
      "theoretical_loss": 3.5482891498924736,
      "tokens_seen": 1353973760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978736208625878,
      "loss": 2.7817,
      "theoretical_loss": 3.5482737084419425,
      "tokens_seen": 1354039296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029786359077231696,
      "loss": 2.7877,
      "theoretical_loss": 3.5482582679480172,
      "tokens_seen": 1354104832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029785356068204614,
      "loss": 2.6604,
      "theoretical_loss": 3.5482428284105936,
      "tokens_seen": 1354170368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978435305917753,
      "loss": 2.6569,
      "theoretical_loss": 3.5482273898295644,
      "tokens_seen": 1354235904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029783350050150456,
      "loss": 2.4882,
      "theoretical_loss": 3.5482119522048254,
      "tokens_seen": 1354301440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978234704112337,
      "loss": 2.8416,
      "theoretical_loss": 3.5481965155362705,
      "tokens_seen": 1354366976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978134403209629,
      "loss": 2.6932,
      "theoretical_loss": 3.548181079823795,
      "tokens_seen": 1354432512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029780341023069205,
      "loss": 2.7576,
      "theoretical_loss": 3.548165645067292,
      "tokens_seen": 1354498048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977933801404213,
      "loss": 2.603,
      "theoretical_loss": 3.548150211266657,
      "tokens_seen": 1354563584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029778335005015046,
      "loss": 2.6103,
      "theoretical_loss": 3.548134778421785,
      "tokens_seen": 1354629120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029777331995987965,
      "loss": 2.5713,
      "theoretical_loss": 3.5481193465325696,
      "tokens_seen": 1354694656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029776328986960883,
      "loss": 2.6096,
      "theoretical_loss": 3.5481039155989054,
      "tokens_seen": 1354760192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029775325977933806,
      "loss": 2.5663,
      "theoretical_loss": 3.548088485620688,
      "tokens_seen": 1354825728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977432296890672,
      "loss": 2.5378,
      "theoretical_loss": 3.5480730565978114,
      "tokens_seen": 1354891264
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 747636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4359793663024902,
      "objective/train/theoretical_loss": 3.54805762853017,
      "objective/train/tokens_used": 1375416800,
      "theoretical_loss": 3.54805762853017,
      "tokens_seen": 1354956800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977331995987964,
      "loss": 2.5447,
      "theoretical_loss": 3.54805762853017,
      "tokens_seen": 1354956800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029772316950852555,
      "loss": 2.6167,
      "theoretical_loss": 3.5480422014176582,
      "tokens_seen": 1355022336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977131394182548,
      "loss": 2.6217,
      "theoretical_loss": 3.548026775260172,
      "tokens_seen": 1355087872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029770310932798397,
      "loss": 2.7159,
      "theoretical_loss": 3.548011350057605,
      "tokens_seen": 1355153408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029769307923771315,
      "loss": 2.6077,
      "theoretical_loss": 3.547995925809852,
      "tokens_seen": 1355218944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029768304914744233,
      "loss": 2.473,
      "theoretical_loss": 3.5479805025168085,
      "tokens_seen": 1355284480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976730190571715,
      "loss": 2.5297,
      "theoretical_loss": 3.5479650801783684,
      "tokens_seen": 1355350016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976629889669007,
      "loss": 2.7728,
      "theoretical_loss": 3.547949658794427,
      "tokens_seen": 1355415552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029765295887662993,
      "loss": 2.7018,
      "theoretical_loss": 3.547934238364878,
      "tokens_seen": 1355481088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029764292878635905,
      "loss": 2.7848,
      "theoretical_loss": 3.5479188188896176,
      "tokens_seen": 1355546624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976328986960883,
      "loss": 2.732,
      "theoretical_loss": 3.54790340036854,
      "tokens_seen": 1355612160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976228686058174,
      "loss": 2.5588,
      "theoretical_loss": 3.54788798280154,
      "tokens_seen": 1355677696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029761283851554665,
      "loss": 2.66,
      "theoretical_loss": 3.547872566188513,
      "tokens_seen": 1355743232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029760280842527583,
      "loss": 2.5881,
      "theoretical_loss": 3.5478571505293526,
      "tokens_seen": 1355808768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297592778335005,
      "loss": 2.5998,
      "theoretical_loss": 3.5478417358239547,
      "tokens_seen": 1355874304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975827482447342,
      "loss": 2.6738,
      "theoretical_loss": 3.547826322072214,
      "tokens_seen": 1355939840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029757271815446343,
      "loss": 2.7197,
      "theoretical_loss": 3.5478109092740255,
      "tokens_seen": 1356005376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029756268806419256,
      "loss": 2.4993,
      "theoretical_loss": 3.547795497429284,
      "tokens_seen": 1356070912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975526579739218,
      "loss": 2.6149,
      "theoretical_loss": 3.547780086537884,
      "tokens_seen": 1356136448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975426278836509,
      "loss": 2.8272,
      "theoretical_loss": 3.547764676599721,
      "tokens_seen": 1356201984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029753259779338016,
      "loss": 2.5975,
      "theoretical_loss": 3.54774926761469,
      "tokens_seen": 1356267520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029752256770310934,
      "loss": 2.6422,
      "theoretical_loss": 3.547733859582686,
      "tokens_seen": 1356333056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975125376128385,
      "loss": 2.6204,
      "theoretical_loss": 3.547718452503604,
      "tokens_seen": 1356398592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975025075225677,
      "loss": 2.4732,
      "theoretical_loss": 3.5477030463773387,
      "tokens_seen": 1356464128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974924774322969,
      "loss": 2.6265,
      "theoretical_loss": 3.547687641203785,
      "tokens_seen": 1356529664
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 748231,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.774550676345825,
      "objective/train/theoretical_loss": 3.547672236982839,
      "objective/train/tokens_used": 1377055200,
      "theoretical_loss": 3.547672236982839,
      "tokens_seen": 1356595200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029748244734202606,
      "loss": 2.7848,
      "theoretical_loss": 3.547672236982839,
      "tokens_seen": 1356595200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974724172517553,
      "loss": 2.9087,
      "theoretical_loss": 3.5476568337143943,
      "tokens_seen": 1356660736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974623871614844,
      "loss": 2.6341,
      "theoretical_loss": 3.5476414313983473,
      "tokens_seen": 1356726272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029745235707121366,
      "loss": 2.6638,
      "theoretical_loss": 3.5476260300345928,
      "tokens_seen": 1356791808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974423269809428,
      "loss": 2.7465,
      "theoretical_loss": 3.547610629623025,
      "tokens_seen": 1356857344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297432296890672,
      "loss": 2.5829,
      "theoretical_loss": 3.54759523016354,
      "tokens_seen": 1356922880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974222668004012,
      "loss": 2.376,
      "theoretical_loss": 3.547579831656033,
      "tokens_seen": 1356988416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974122367101304,
      "loss": 2.7145,
      "theoretical_loss": 3.5475644341003987,
      "tokens_seen": 1357053952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029740220661985956,
      "loss": 2.6275,
      "theoretical_loss": 3.547549037496533,
      "tokens_seen": 1357119488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973921765295888,
      "loss": 2.4773,
      "theoretical_loss": 3.54753364184433,
      "tokens_seen": 1357185024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973821464393179,
      "loss": 2.6747,
      "theoretical_loss": 3.547518247143686,
      "tokens_seen": 1357250560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029737211634904716,
      "loss": 2.6817,
      "theoretical_loss": 3.5475028533944952,
      "tokens_seen": 1357316096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973620862587763,
      "loss": 2.4743,
      "theoretical_loss": 3.547487460596654,
      "tokens_seen": 1357381632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973520561685055,
      "loss": 2.5971,
      "theoretical_loss": 3.547472068750057,
      "tokens_seen": 1357447168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973420260782347,
      "loss": 2.6343,
      "theoretical_loss": 3.5474566778545995,
      "tokens_seen": 1357512704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973319959879639,
      "loss": 2.7745,
      "theoretical_loss": 3.5474412879101767,
      "tokens_seen": 1357578240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029732196589769307,
      "loss": 2.6456,
      "theoretical_loss": 3.5474258989166847,
      "tokens_seen": 1357643776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029731193580742225,
      "loss": 2.6217,
      "theoretical_loss": 3.547410510874018,
      "tokens_seen": 1357709312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029730190571715143,
      "loss": 2.9643,
      "theoretical_loss": 3.5473951237820724,
      "tokens_seen": 1357774848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029729187562688067,
      "loss": 2.7048,
      "theoretical_loss": 3.5473797376407425,
      "tokens_seen": 1357840384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029728184553660985,
      "loss": 2.6319,
      "theoretical_loss": 3.547364352449925,
      "tokens_seen": 1357905920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029727181544633903,
      "loss": 2.7396,
      "theoretical_loss": 3.5473489682095147,
      "tokens_seen": 1357971456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029726178535606826,
      "loss": 2.5027,
      "theoretical_loss": 3.5473335849194068,
      "tokens_seen": 1358036992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972517552657974,
      "loss": 2.7434,
      "theoretical_loss": 3.547318202579497,
      "tokens_seen": 1358102528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972417251755266,
      "loss": 2.4824,
      "theoretical_loss": 3.5473028211896804,
      "tokens_seen": 1358168064
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 749499,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2978737354278564,
      "objective/train/theoretical_loss": 3.547287440749853,
      "objective/train/tokens_used": 1378693600,
      "theoretical_loss": 3.547287440749853,
      "tokens_seen": 1358233600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029723169508525575,
      "loss": 2.4681,
      "theoretical_loss": 3.547287440749853,
      "tokens_seen": 1358233600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297221664994985,
      "loss": 2.7365,
      "theoretical_loss": 3.5472720612599096,
      "tokens_seen": 1358299136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029721163490471417,
      "loss": 2.7585,
      "theoretical_loss": 3.5472566827197465,
      "tokens_seen": 1358364672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029720160481444335,
      "loss": 2.6884,
      "theoretical_loss": 3.5472413051292593,
      "tokens_seen": 1358430208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029719157472417253,
      "loss": 2.7544,
      "theoretical_loss": 3.5472259284883423,
      "tokens_seen": 1358495744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971815446339017,
      "loss": 2.4551,
      "theoretical_loss": 3.5472105527968925,
      "tokens_seen": 1358561280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971715145436309,
      "loss": 2.6856,
      "theoretical_loss": 3.5471951780548046,
      "tokens_seen": 1358626816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029716148445336013,
      "loss": 2.8647,
      "theoretical_loss": 3.547179804261974,
      "tokens_seen": 1358692352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029715145436308926,
      "loss": 2.5976,
      "theoretical_loss": 3.5471644314182975,
      "tokens_seen": 1358757888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971414242728185,
      "loss": 2.6342,
      "theoretical_loss": 3.5471490595236697,
      "tokens_seen": 1358823424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971313941825476,
      "loss": 2.7117,
      "theoretical_loss": 3.5471336885779867,
      "tokens_seen": 1358888960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029712136409227685,
      "loss": 2.6165,
      "theoretical_loss": 3.547118318581144,
      "tokens_seen": 1358954496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029711133400200603,
      "loss": 2.5478,
      "theoretical_loss": 3.547102949533037,
      "tokens_seen": 1359020032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971013039117352,
      "loss": 2.5827,
      "theoretical_loss": 3.547087581433561,
      "tokens_seen": 1359085568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970912738214644,
      "loss": 2.7006,
      "theoretical_loss": 3.5470722142826134,
      "tokens_seen": 1359151104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029708124373119363,
      "loss": 2.7409,
      "theoretical_loss": 3.547056848080089,
      "tokens_seen": 1359216640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029707121364092276,
      "loss": 2.9835,
      "theoretical_loss": 3.5470414828258825,
      "tokens_seen": 1359282176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297061183550652,
      "loss": 2.738,
      "theoretical_loss": 3.5470261185198915,
      "tokens_seen": 1359347712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970511534603811,
      "loss": 2.6083,
      "theoretical_loss": 3.54701075516201,
      "tokens_seen": 1359413248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029704112337011036,
      "loss": 2.6068,
      "theoretical_loss": 3.546995392752135,
      "tokens_seen": 1359478784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029703109327983954,
      "loss": 2.6079,
      "theoretical_loss": 3.546980031290162,
      "tokens_seen": 1359544320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970210631895687,
      "loss": 2.963,
      "theoretical_loss": 3.5469646707759868,
      "tokens_seen": 1359609856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970110330992979,
      "loss": 2.5476,
      "theoretical_loss": 3.546949311209505,
      "tokens_seen": 1359675392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970010030090271,
      "loss": 2.636,
      "theoretical_loss": 3.546933952590613,
      "tokens_seen": 1359740928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029699097291875626,
      "loss": 2.6404,
      "theoretical_loss": 3.546918594919206,
      "tokens_seen": 1359806464
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 749946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6572976112365723,
      "objective/train/theoretical_loss": 3.5469032381951804,
      "objective/train/tokens_used": 1380332000,
      "theoretical_loss": 3.5469032381951804,
      "tokens_seen": 1359872000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969809428284855,
      "loss": 2.4872,
      "theoretical_loss": 3.5469032381951804,
      "tokens_seen": 1359872000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969709127382146,
      "loss": 2.815,
      "theoretical_loss": 3.5468878824184324,
      "tokens_seen": 1359937536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029696088264794386,
      "loss": 2.714,
      "theoretical_loss": 3.5468725275888566,
      "tokens_seen": 1360003072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296950852557673,
      "loss": 2.6146,
      "theoretical_loss": 3.5468571737063503,
      "tokens_seen": 1360068608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969408224674022,
      "loss": 2.4896,
      "theoretical_loss": 3.546841820770809,
      "tokens_seen": 1360134144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969307923771314,
      "loss": 2.7668,
      "theoretical_loss": 3.546826468782128,
      "tokens_seen": 1360199680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969207622868606,
      "loss": 2.5955,
      "theoretical_loss": 3.546811117740204,
      "tokens_seen": 1360265216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029691073219658976,
      "loss": 2.6055,
      "theoretical_loss": 3.5467957676449338,
      "tokens_seen": 1360330752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296900702106319,
      "loss": 2.5427,
      "theoretical_loss": 3.546780418496212,
      "tokens_seen": 1360396288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968906720160481,
      "loss": 2.3896,
      "theoretical_loss": 3.5467650702939353,
      "tokens_seen": 1360461824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029688064192577736,
      "loss": 2.7381,
      "theoretical_loss": 3.5467497230379994,
      "tokens_seen": 1360527360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968706118355065,
      "loss": 2.7378,
      "theoretical_loss": 3.546734376728301,
      "tokens_seen": 1360592896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968605817452357,
      "loss": 2.6143,
      "theoretical_loss": 3.546719031364735,
      "tokens_seen": 1360658432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968505516549649,
      "loss": 2.6222,
      "theoretical_loss": 3.5467036869471986,
      "tokens_seen": 1360723968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968405215646941,
      "loss": 2.8916,
      "theoretical_loss": 3.546688343475588,
      "tokens_seen": 1360789504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029683049147442327,
      "loss": 2.6858,
      "theoretical_loss": 3.5466730009497986,
      "tokens_seen": 1360855040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029682046138415245,
      "loss": 2.7056,
      "theoretical_loss": 3.5466576593697265,
      "tokens_seen": 1360920576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029681043129388163,
      "loss": 2.4382,
      "theoretical_loss": 3.546642318735269,
      "tokens_seen": 1360986112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029680040120361087,
      "loss": 2.6948,
      "theoretical_loss": 3.546626979046321,
      "tokens_seen": 1361051648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029679037111334,
      "loss": 2.7604,
      "theoretical_loss": 3.5466116403027796,
      "tokens_seen": 1361117184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029678034102306923,
      "loss": 2.631,
      "theoretical_loss": 3.5465963025045406,
      "tokens_seen": 1361182720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967703109327984,
      "loss": 2.432,
      "theoretical_loss": 3.5465809656515006,
      "tokens_seen": 1361248256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967602808425276,
      "loss": 2.4322,
      "theoretical_loss": 3.546565629743555,
      "tokens_seen": 1361313792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029675025075225677,
      "loss": 2.7459,
      "theoretical_loss": 3.5465502947806007,
      "tokens_seen": 1361379328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029674022066198595,
      "loss": 2.7613,
      "theoretical_loss": 3.546534960762534,
      "tokens_seen": 1361444864
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 750808,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7870075702667236,
      "objective/train/theoretical_loss": 3.5465196276892508,
      "objective/train/tokens_used": 1381970400,
      "theoretical_loss": 3.5465196276892508,
      "tokens_seen": 1361510400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029673019057171513,
      "loss": 2.8345,
      "theoretical_loss": 3.5465196276892508,
      "tokens_seen": 1361510400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029672016048144437,
      "loss": 2.4962,
      "theoretical_loss": 3.5465042955606485,
      "tokens_seen": 1361575936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967101303911735,
      "loss": 2.8696,
      "theoretical_loss": 3.5464889643766218,
      "tokens_seen": 1361641472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029670010030090273,
      "loss": 2.4112,
      "theoretical_loss": 3.546473634137068,
      "tokens_seen": 1361707008
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029669007021063186,
      "loss": 2.5725,
      "theoretical_loss": 3.5464583048418836,
      "tokens_seen": 1361772544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966800401203611,
      "loss": 2.7429,
      "theoretical_loss": 3.546442976490965,
      "tokens_seen": 1361838080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966700100300903,
      "loss": 2.6622,
      "theoretical_loss": 3.546427649084208,
      "tokens_seen": 1361903616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029665997993981946,
      "loss": 2.8366,
      "theoretical_loss": 3.546412322621509,
      "tokens_seen": 1361969152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029664994984954864,
      "loss": 2.6071,
      "theoretical_loss": 3.5463969971027653,
      "tokens_seen": 1362034688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966399197592778,
      "loss": 2.5876,
      "theoretical_loss": 3.5463816725278727,
      "tokens_seen": 1362100224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296629889669007,
      "loss": 2.6304,
      "theoretical_loss": 3.5463663488967274,
      "tokens_seen": 1362165760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029661985957873623,
      "loss": 2.8303,
      "theoretical_loss": 3.546351026209227,
      "tokens_seen": 1362231296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029660982948846536,
      "loss": 2.7284,
      "theoretical_loss": 3.5463357044652666,
      "tokens_seen": 1362296832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965997993981946,
      "loss": 2.8072,
      "theoretical_loss": 3.5463203836647437,
      "tokens_seen": 1362362368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965897693079238,
      "loss": 2.775,
      "theoretical_loss": 3.546305063807554,
      "tokens_seen": 1362427904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029657973921765296,
      "loss": 2.8564,
      "theoretical_loss": 3.5462897448935955,
      "tokens_seen": 1362493440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029656970912738214,
      "loss": 2.6761,
      "theoretical_loss": 3.5462744269227633,
      "tokens_seen": 1362558976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965596790371113,
      "loss": 2.5418,
      "theoretical_loss": 3.5462591098949545,
      "tokens_seen": 1362624512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965496489468405,
      "loss": 2.6304,
      "theoretical_loss": 3.5462437938100653,
      "tokens_seen": 1362690048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029653961885656974,
      "loss": 2.6772,
      "theoretical_loss": 3.5462284786679934,
      "tokens_seen": 1362755584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965295887662989,
      "loss": 2.753,
      "theoretical_loss": 3.546213164468634,
      "tokens_seen": 1362821120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965195586760281,
      "loss": 2.7359,
      "theoretical_loss": 3.546197851211885,
      "tokens_seen": 1362886656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965095285857573,
      "loss": 2.565,
      "theoretical_loss": 3.546182538897642,
      "tokens_seen": 1362952192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029649949849548646,
      "loss": 2.7624,
      "theoretical_loss": 3.5461672275258027,
      "tokens_seen": 1363017728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964894684052157,
      "loss": 2.5734,
      "theoretical_loss": 3.546151917096263,
      "tokens_seen": 1363083264
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 751921,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8145251274108887,
      "objective/train/theoretical_loss": 3.5461366076089202,
      "objective/train/tokens_used": 1383608800,
      "theoretical_loss": 3.5461366076089202,
      "tokens_seen": 1363148800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964794383149448,
      "loss": 2.8472,
      "theoretical_loss": 3.5461366076089202,
      "tokens_seen": 1363148800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029646940822467406,
      "loss": 2.6509,
      "theoretical_loss": 3.54612129906367,
      "tokens_seen": 1363214336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964593781344032,
      "loss": 2.6087,
      "theoretical_loss": 3.5461059914604105,
      "tokens_seen": 1363279872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964493480441324,
      "loss": 2.5793,
      "theoretical_loss": 3.5460906847990374,
      "tokens_seen": 1363345408
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964393179538616,
      "loss": 2.6705,
      "theoretical_loss": 3.5460753790794484,
      "tokens_seen": 1363410944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964292878635908,
      "loss": 2.7595,
      "theoretical_loss": 3.54606007430154,
      "tokens_seen": 1363476480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029641925777331996,
      "loss": 2.6003,
      "theoretical_loss": 3.5460447704652074,
      "tokens_seen": 1363542016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964092276830492,
      "loss": 2.6074,
      "theoretical_loss": 3.54602946757035,
      "tokens_seen": 1363607552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963991975927783,
      "loss": 2.6677,
      "theoretical_loss": 3.546014165616863,
      "tokens_seen": 1363673088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029638916750250756,
      "loss": 2.4795,
      "theoretical_loss": 3.545998864604644,
      "tokens_seen": 1363738624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963791374122367,
      "loss": 2.4361,
      "theoretical_loss": 3.5459835645335893,
      "tokens_seen": 1363804160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963691073219659,
      "loss": 2.8123,
      "theoretical_loss": 3.5459682654035958,
      "tokens_seen": 1363869696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963590772316951,
      "loss": 2.9173,
      "theoretical_loss": 3.5459529672145607,
      "tokens_seen": 1363935232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963490471414243,
      "loss": 2.4875,
      "theoretical_loss": 3.5459376699663814,
      "tokens_seen": 1364000768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029633901705115347,
      "loss": 2.7815,
      "theoretical_loss": 3.5459223736589536,
      "tokens_seen": 1364066304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029632898696088265,
      "loss": 2.6226,
      "theoretical_loss": 3.5459070782921756,
      "tokens_seen": 1364131840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029631895687061183,
      "loss": 2.4766,
      "theoretical_loss": 3.545891783865944,
      "tokens_seen": 1364197376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029630892678034107,
      "loss": 2.5906,
      "theoretical_loss": 3.5458764903801545,
      "tokens_seen": 1364262912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962988966900702,
      "loss": 2.5496,
      "theoretical_loss": 3.5458611978347054,
      "tokens_seen": 1364328448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029628886659979943,
      "loss": 2.4896,
      "theoretical_loss": 3.545845906229494,
      "tokens_seen": 1364393984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962788365095286,
      "loss": 2.6595,
      "theoretical_loss": 3.545830615564416,
      "tokens_seen": 1364459520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962688064192578,
      "loss": 2.523,
      "theoretical_loss": 3.54581532583937,
      "tokens_seen": 1364525056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029625877632898697,
      "loss": 2.6689,
      "theoretical_loss": 3.545800037054251,
      "tokens_seen": 1364590592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029624874623871615,
      "loss": 2.4986,
      "theoretical_loss": 3.5457847492089587,
      "tokens_seen": 1364656128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029623871614844533,
      "loss": 2.4405,
      "theoretical_loss": 3.5457694623033884,
      "tokens_seen": 1364721664
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 752367,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4154515266418457,
      "objective/train/theoretical_loss": 3.5457541763374376,
      "objective/train/tokens_used": 1385247200,
      "theoretical_loss": 3.5457541763374376,
      "tokens_seen": 1364787200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029622868605817457,
      "loss": 2.6149,
      "theoretical_loss": 3.5457541763374376,
      "tokens_seen": 1364787200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962186559679037,
      "loss": 2.8065,
      "theoretical_loss": 3.5457388913110033,
      "tokens_seen": 1364852736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029620862587763293,
      "loss": 2.678,
      "theoretical_loss": 3.545723607223983,
      "tokens_seen": 1364918272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029619859578736206,
      "loss": 2.706,
      "theoretical_loss": 3.5457083240762737,
      "tokens_seen": 1364983808
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961885656970913,
      "loss": 2.7492,
      "theoretical_loss": 3.545693041867773,
      "tokens_seen": 1365049344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961785356068205,
      "loss": 2.7696,
      "theoretical_loss": 3.545677760598377,
      "tokens_seen": 1365114880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029616850551654966,
      "loss": 2.4822,
      "theoretical_loss": 3.5456624802679837,
      "tokens_seen": 1365180416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029615847542627884,
      "loss": 2.8991,
      "theoretical_loss": 3.5456472008764903,
      "tokens_seen": 1365245952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296148445336008,
      "loss": 2.6352,
      "theoretical_loss": 3.5456319224237944,
      "tokens_seen": 1365311488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961384152457372,
      "loss": 2.4082,
      "theoretical_loss": 3.545616644909792,
      "tokens_seen": 1365377024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029612838515546643,
      "loss": 2.7084,
      "theoretical_loss": 3.545601368334382,
      "tokens_seen": 1365442560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029611835506519556,
      "loss": 2.687,
      "theoretical_loss": 3.5455860926974605,
      "tokens_seen": 1365508096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961083249749248,
      "loss": 2.5229,
      "theoretical_loss": 3.545570817998925,
      "tokens_seen": 1365573632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296098294884654,
      "loss": 2.9669,
      "theoretical_loss": 3.5455555442386735,
      "tokens_seen": 1365639168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029608826479438316,
      "loss": 2.5514,
      "theoretical_loss": 3.5455402714166024,
      "tokens_seen": 1365704704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029607823470411234,
      "loss": 2.6351,
      "theoretical_loss": 3.54552499953261,
      "tokens_seen": 1365770240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960682046138415,
      "loss": 2.584,
      "theoretical_loss": 3.545509728586593,
      "tokens_seen": 1365835776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960581745235707,
      "loss": 2.574,
      "theoretical_loss": 3.5454944585784487,
      "tokens_seen": 1365901312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029604814443329994,
      "loss": 2.6679,
      "theoretical_loss": 3.545479189508075,
      "tokens_seen": 1365966848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029603811434302906,
      "loss": 2.5933,
      "theoretical_loss": 3.5454639213753687,
      "tokens_seen": 1366032384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960280842527583,
      "loss": 2.781,
      "theoretical_loss": 3.545448654180228,
      "tokens_seen": 1366097920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002960180541624874,
      "loss": 2.5818,
      "theoretical_loss": 3.54543338792255,
      "tokens_seen": 1366163456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029600802407221666,
      "loss": 2.5303,
      "theoretical_loss": 3.545418122602232,
      "tokens_seen": 1366228992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029599799398194584,
      "loss": 2.5922,
      "theoretical_loss": 3.5454028582191723,
      "tokens_seen": 1366294528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295987963891675,
      "loss": 2.7615,
      "theoretical_loss": 3.545387594773267,
      "tokens_seen": 1366360064
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 752972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.119584321975708,
      "objective/train/theoretical_loss": 3.5453723322644146,
      "objective/train/tokens_used": 1386885600,
      "theoretical_loss": 3.5453723322644146,
      "tokens_seen": 1366425600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959779338014042,
      "loss": 2.7982,
      "theoretical_loss": 3.5453723322644146,
      "tokens_seen": 1366425600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959679037111334,
      "loss": 2.4731,
      "theoretical_loss": 3.5453570706925124,
      "tokens_seen": 1366491136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029595787362086257,
      "loss": 2.6281,
      "theoretical_loss": 3.545341810057458,
      "tokens_seen": 1366556672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959478435305918,
      "loss": 2.5008,
      "theoretical_loss": 3.5453265503591487,
      "tokens_seen": 1366622208
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029593781344032093,
      "loss": 2.5527,
      "theoretical_loss": 3.545311291597483,
      "tokens_seen": 1366687744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029592778335005016,
      "loss": 2.6509,
      "theoretical_loss": 3.545296033772357,
      "tokens_seen": 1366753280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029591775325977935,
      "loss": 2.6369,
      "theoretical_loss": 3.5452807768836694,
      "tokens_seen": 1366818816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029590772316950853,
      "loss": 2.7679,
      "theoretical_loss": 3.5452655209313173,
      "tokens_seen": 1366884352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958976930792377,
      "loss": 2.6728,
      "theoretical_loss": 3.545250265915199,
      "tokens_seen": 1366949888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958876629889669,
      "loss": 2.7048,
      "theoretical_loss": 3.5452350118352114,
      "tokens_seen": 1367015424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029587763289869607,
      "loss": 2.7995,
      "theoretical_loss": 3.5452197586912524,
      "tokens_seen": 1367080960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958676028084253,
      "loss": 2.5238,
      "theoretical_loss": 3.5452045064832207,
      "tokens_seen": 1367146496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029585757271815443,
      "loss": 2.6935,
      "theoretical_loss": 3.545189255211012,
      "tokens_seen": 1367212032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029584754262788367,
      "loss": 2.6453,
      "theoretical_loss": 3.545174004874526,
      "tokens_seen": 1367277568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958375125376128,
      "loss": 2.6736,
      "theoretical_loss": 3.5451587554736594,
      "tokens_seen": 1367343104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029582748244734203,
      "loss": 2.7711,
      "theoretical_loss": 3.5451435070083104,
      "tokens_seen": 1367408640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958174523570712,
      "loss": 2.6743,
      "theoretical_loss": 3.545128259478376,
      "tokens_seen": 1367474176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958074222668004,
      "loss": 2.7954,
      "theoretical_loss": 3.5451130128837547,
      "tokens_seen": 1367539712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957973921765296,
      "loss": 2.6541,
      "theoretical_loss": 3.5450977672243447,
      "tokens_seen": 1367605248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957873620862588,
      "loss": 2.7035,
      "theoretical_loss": 3.5450825225000426,
      "tokens_seen": 1367670784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295777331995988,
      "loss": 2.8651,
      "theoretical_loss": 3.545067278710747,
      "tokens_seen": 1367736320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029576730190571717,
      "loss": 2.5355,
      "theoretical_loss": 3.5450520358563558,
      "tokens_seen": 1367801856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029575727181544635,
      "loss": 2.5466,
      "theoretical_loss": 3.5450367939367666,
      "tokens_seen": 1367867392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029574724172517553,
      "loss": 2.6883,
      "theoretical_loss": 3.5450215529518774,
      "tokens_seen": 1367932928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029573721163490477,
      "loss": 2.4825,
      "theoretical_loss": 3.5450063129015863,
      "tokens_seen": 1367998464
    },
    {
      "epoch": 0.41,
      "objective/train/docs_used": 754199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.419557809829712,
      "objective/train/theoretical_loss": 3.5449910737857913,
      "objective/train/tokens_used": 1388524000,
      "theoretical_loss": 3.5449910737857913,
      "tokens_seen": 1368064000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957271815446339,
      "loss": 2.7239,
      "theoretical_loss": 3.5449910737857913,
      "tokens_seen": 1368064000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029571715145436313,
      "loss": 2.6982,
      "theoretical_loss": 3.5449758356043897,
      "tokens_seen": 1368129536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029570712136409226,
      "loss": 2.6743,
      "theoretical_loss": 3.54496059835728,
      "tokens_seen": 1368195072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956970912738215,
      "loss": 2.7385,
      "theoretical_loss": 3.54494536204436,
      "tokens_seen": 1368260608
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956870611835507,
      "loss": 2.6535,
      "theoretical_loss": 3.5449301266655273,
      "tokens_seen": 1368326144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029567703109327986,
      "loss": 2.6099,
      "theoretical_loss": 3.544914892220681,
      "tokens_seen": 1368391680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029566700100300904,
      "loss": 2.962,
      "theoretical_loss": 3.544899658709718,
      "tokens_seen": 1368457216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956569709127382,
      "loss": 2.5697,
      "theoretical_loss": 3.5448844261325365,
      "tokens_seen": 1368522752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956469408224674,
      "loss": 2.849,
      "theoretical_loss": 3.544869194489035,
      "tokens_seen": 1368588288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029563691073219663,
      "loss": 2.6937,
      "theoretical_loss": 3.5448539637791114,
      "tokens_seen": 1368653824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029562688064192576,
      "loss": 2.9566,
      "theoretical_loss": 3.5448387340026635,
      "tokens_seen": 1368719360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295616850551655,
      "loss": 2.5097,
      "theoretical_loss": 3.5448235051595898,
      "tokens_seen": 1368784896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956068204613842,
      "loss": 2.6633,
      "theoretical_loss": 3.5448082772497886,
      "tokens_seen": 1368850432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029559679037111336,
      "loss": 3.0324,
      "theoretical_loss": 3.5447930502731575,
      "tokens_seen": 1368915968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029558676028084254,
      "loss": 2.7005,
      "theoretical_loss": 3.544777824229594,
      "tokens_seen": 1368981504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955767301905717,
      "loss": 2.6599,
      "theoretical_loss": 3.544762599118998,
      "tokens_seen": 1369047040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955667001003009,
      "loss": 2.5976,
      "theoretical_loss": 3.5447473749412666,
      "tokens_seen": 1369112576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029555667001003014,
      "loss": 2.7322,
      "theoretical_loss": 3.544732151696298,
      "tokens_seen": 1369178112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029554663991975926,
      "loss": 2.6484,
      "theoretical_loss": 3.5447169293839904,
      "tokens_seen": 1369243648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955366098294885,
      "loss": 2.7548,
      "theoretical_loss": 3.5447017080042422,
      "tokens_seen": 1369309184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955265797392176,
      "loss": 2.7891,
      "theoretical_loss": 3.5446864875569517,
      "tokens_seen": 1369374720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029551654964894686,
      "loss": 2.6879,
      "theoretical_loss": 3.544671268042017,
      "tokens_seen": 1369440256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029550651955867604,
      "loss": 2.529,
      "theoretical_loss": 3.544656049459337,
      "tokens_seen": 1369505792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954964894684052,
      "loss": 2.8846,
      "theoretical_loss": 3.5446408318088087,
      "tokens_seen": 1369571328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954864593781344,
      "loss": 2.5813,
      "theoretical_loss": 3.5446256150903315,
      "tokens_seen": 1369636864
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 755490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9412763118743896,
      "objective/train/theoretical_loss": 3.544610399303803,
      "objective/train/tokens_used": 1390162400,
      "theoretical_loss": 3.544610399303803,
      "tokens_seen": 1369702400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954764292878636,
      "loss": 2.6904,
      "theoretical_loss": 3.544610399303803,
      "tokens_seen": 1369702400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029546639919759277,
      "loss": 2.6165,
      "theoretical_loss": 3.544595184449122,
      "tokens_seen": 1369767936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295456369107322,
      "loss": 2.7068,
      "theoretical_loss": 3.544579970526187,
      "tokens_seen": 1369833472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029544633901705113,
      "loss": 2.4895,
      "theoretical_loss": 3.544564757534896,
      "tokens_seen": 1369899008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029543630892678037,
      "loss": 2.6439,
      "theoretical_loss": 3.5445495454751477,
      "tokens_seen": 1369964544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029542627883650955,
      "loss": 2.5897,
      "theoretical_loss": 3.5445343343468396,
      "tokens_seen": 1370030080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029541624874623873,
      "loss": 2.5313,
      "theoretical_loss": 3.5445191241498715,
      "tokens_seen": 1370095616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954062186559679,
      "loss": 2.6542,
      "theoretical_loss": 3.5445039148841406,
      "tokens_seen": 1370161152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953961885656971,
      "loss": 2.6015,
      "theoretical_loss": 3.5444887065495463,
      "tokens_seen": 1370226688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029538615847542627,
      "loss": 2.9512,
      "theoretical_loss": 3.5444734991459868,
      "tokens_seen": 1370292224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953761283851555,
      "loss": 2.8088,
      "theoretical_loss": 3.54445829267336,
      "tokens_seen": 1370357760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029536609829488463,
      "loss": 2.7238,
      "theoretical_loss": 3.544443087131565,
      "tokens_seen": 1370423296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029535606820461387,
      "loss": 2.5758,
      "theoretical_loss": 3.5444278825205,
      "tokens_seen": 1370488832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295346038114343,
      "loss": 2.8293,
      "theoretical_loss": 3.5444126788400636,
      "tokens_seen": 1370554368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029533600802407223,
      "loss": 2.5181,
      "theoretical_loss": 3.544397476090155,
      "tokens_seen": 1370619904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953259779338014,
      "loss": 2.5581,
      "theoretical_loss": 3.5443822742706717,
      "tokens_seen": 1370685440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953159478435306,
      "loss": 2.5342,
      "theoretical_loss": 3.5443670733815127,
      "tokens_seen": 1370750976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953059177532598,
      "loss": 2.7378,
      "theoretical_loss": 3.5443518734225767,
      "tokens_seen": 1370816512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295295887662989,
      "loss": 2.709,
      "theoretical_loss": 3.544336674393762,
      "tokens_seen": 1370882048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029528585757271814,
      "loss": 2.4601,
      "theoretical_loss": 3.544321476294968,
      "tokens_seen": 1370947584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029527582748244737,
      "loss": 2.5891,
      "theoretical_loss": 3.544306279126092,
      "tokens_seen": 1371013120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952657973921765,
      "loss": 2.6917,
      "theoretical_loss": 3.5442910828870344,
      "tokens_seen": 1371078656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029525576730190573,
      "loss": 2.6312,
      "theoretical_loss": 3.544275887577692,
      "tokens_seen": 1371144192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952457372116349,
      "loss": 2.6885,
      "theoretical_loss": 3.544260693197965,
      "tokens_seen": 1371209728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952357071213641,
      "loss": 2.7528,
      "theoretical_loss": 3.5442454997477513,
      "tokens_seen": 1371275264
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 756189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.601017951965332,
      "objective/train/theoretical_loss": 3.5442303072269503,
      "objective/train/tokens_used": 1391800800,
      "theoretical_loss": 3.5442303072269503,
      "tokens_seen": 1371340800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952256770310933,
      "loss": 2.5892,
      "theoretical_loss": 3.5442303072269503,
      "tokens_seen": 1371340800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029521564694082246,
      "loss": 2.5543,
      "theoretical_loss": 3.5442151156354598,
      "tokens_seen": 1371406336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029520561685055164,
      "loss": 2.5929,
      "theoretical_loss": 3.544199924973179,
      "tokens_seen": 1371471872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951955867602809,
      "loss": 2.5026,
      "theoretical_loss": 3.5441847352400067,
      "tokens_seen": 1371537408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029518555667001,
      "loss": 2.7645,
      "theoretical_loss": 3.5441695464358416,
      "tokens_seen": 1371602944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029517552657973924,
      "loss": 2.8162,
      "theoretical_loss": 3.544154358560583,
      "tokens_seen": 1371668480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029516549648946836,
      "loss": 2.7797,
      "theoretical_loss": 3.5441391716141286,
      "tokens_seen": 1371734016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951554663991976,
      "loss": 2.6188,
      "theoretical_loss": 3.544123985596378,
      "tokens_seen": 1371799552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951454363089268,
      "loss": 2.4542,
      "theoretical_loss": 3.5441088005072303,
      "tokens_seen": 1371865088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029513540621865596,
      "loss": 2.518,
      "theoretical_loss": 3.544093616346584,
      "tokens_seen": 1371930624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029512537612838514,
      "loss": 2.4409,
      "theoretical_loss": 3.544078433114338,
      "tokens_seen": 1371996160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951153460381144,
      "loss": 2.7041,
      "theoretical_loss": 3.5440632508103906,
      "tokens_seen": 1372061696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951053159478435,
      "loss": 2.5683,
      "theoretical_loss": 3.544048069434642,
      "tokens_seen": 1372127232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029509528585757274,
      "loss": 2.5049,
      "theoretical_loss": 3.5440328889869903,
      "tokens_seen": 1372192768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029508525576730187,
      "loss": 2.9035,
      "theoretical_loss": 3.5440177094673344,
      "tokens_seen": 1372258304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950752256770311,
      "loss": 2.6459,
      "theoretical_loss": 3.544002530875573,
      "tokens_seen": 1372323840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950651955867603,
      "loss": 2.8768,
      "theoretical_loss": 3.543987353211606,
      "tokens_seen": 1372389376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029505516549648946,
      "loss": 2.8668,
      "theoretical_loss": 3.543972176475332,
      "tokens_seen": 1372454912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029504513540621865,
      "loss": 2.6032,
      "theoretical_loss": 3.543957000666649,
      "tokens_seen": 1372520448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950351053159478,
      "loss": 2.676,
      "theoretical_loss": 3.543941825785458,
      "tokens_seen": 1372585984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029502507522567706,
      "loss": 2.6996,
      "theoretical_loss": 3.543926651831656,
      "tokens_seen": 1372651520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029501504513540624,
      "loss": 2.8494,
      "theoretical_loss": 3.5439114788051436,
      "tokens_seen": 1372717056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950050150451354,
      "loss": 2.4998,
      "theoretical_loss": 3.543896306705819,
      "tokens_seen": 1372782592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949949849548646,
      "loss": 2.6772,
      "theoretical_loss": 3.543881135533581,
      "tokens_seen": 1372848128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949849548645938,
      "loss": 2.8906,
      "theoretical_loss": 3.5438659652883295,
      "tokens_seen": 1372913664
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 757380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.386281967163086,
      "objective/train/theoretical_loss": 3.5438507959699637,
      "objective/train/tokens_used": 1393439200,
      "theoretical_loss": 3.5438507959699637,
      "tokens_seen": 1372979200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029497492477432297,
      "loss": 2.5829,
      "theoretical_loss": 3.5438507959699637,
      "tokens_seen": 1372979200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949648946840522,
      "loss": 2.7449,
      "theoretical_loss": 3.5438356275783827,
      "tokens_seen": 1373044736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029495486459378133,
      "loss": 2.6824,
      "theoretical_loss": 3.5438204601134844,
      "tokens_seen": 1373110272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029494483450351057,
      "loss": 2.6819,
      "theoretical_loss": 3.5438052935751694,
      "tokens_seen": 1373175808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029493480441323975,
      "loss": 2.5504,
      "theoretical_loss": 3.5437901279633364,
      "tokens_seen": 1373241344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029492477432296893,
      "loss": 2.6439,
      "theoretical_loss": 3.543774963277884,
      "tokens_seen": 1373306880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949147442326981,
      "loss": 2.6087,
      "theoretical_loss": 3.5437597995187122,
      "tokens_seen": 1373372416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949047141424273,
      "loss": 2.3426,
      "theoretical_loss": 3.54374463668572,
      "tokens_seen": 1373437952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029489468405215647,
      "loss": 2.615,
      "theoretical_loss": 3.5437294747788064,
      "tokens_seen": 1373503488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948846539618857,
      "loss": 2.4632,
      "theoretical_loss": 3.5437143137978713,
      "tokens_seen": 1373569024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029487462387161483,
      "loss": 2.5999,
      "theoretical_loss": 3.5436991537428133,
      "tokens_seen": 1373634560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029486459378134407,
      "loss": 2.5646,
      "theoretical_loss": 3.5436839946135317,
      "tokens_seen": 1373700096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948545636910732,
      "loss": 2.4645,
      "theoretical_loss": 3.5436688364099265,
      "tokens_seen": 1373765632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029484453360080243,
      "loss": 2.7137,
      "theoretical_loss": 3.5436536791318964,
      "tokens_seen": 1373831168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948345035105316,
      "loss": 2.779,
      "theoretical_loss": 3.5436385227793408,
      "tokens_seen": 1373896704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948244734202608,
      "loss": 2.5723,
      "theoretical_loss": 3.543623367352159,
      "tokens_seen": 1373962240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029481444332999,
      "loss": 2.4863,
      "theoretical_loss": 3.5436082128502506,
      "tokens_seen": 1374027776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948044132397192,
      "loss": 2.4921,
      "theoretical_loss": 3.543593059273515,
      "tokens_seen": 1374093312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029479438314944834,
      "loss": 2.5952,
      "theoretical_loss": 3.5435779066218513,
      "tokens_seen": 1374158848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029478435305917757,
      "loss": 2.6188,
      "theoretical_loss": 3.5435627548951594,
      "tokens_seen": 1374224384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947743229689067,
      "loss": 2.6058,
      "theoretical_loss": 3.543547604093338,
      "tokens_seen": 1374289920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029476429287863593,
      "loss": 2.7356,
      "theoretical_loss": 3.543532454216287,
      "tokens_seen": 1374355456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947542627883651,
      "loss": 2.7198,
      "theoretical_loss": 3.543517305263906,
      "tokens_seen": 1374420992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947442326980943,
      "loss": 2.7524,
      "theoretical_loss": 3.5435021572360945,
      "tokens_seen": 1374486528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947342026078235,
      "loss": 2.8701,
      "theoretical_loss": 3.5434870101327514,
      "tokens_seen": 1374552064
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 758052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.612107992172241,
      "objective/train/theoretical_loss": 3.5434718639537763,
      "objective/train/tokens_used": 1395077600,
      "theoretical_loss": 3.5434718639537763,
      "tokens_seen": 1374617600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029472417251755266,
      "loss": 2.6773,
      "theoretical_loss": 3.5434718639537763,
      "tokens_seen": 1374617600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029471414242728184,
      "loss": 2.6392,
      "theoretical_loss": 3.54345671869907,
      "tokens_seen": 1374683136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947041123370111,
      "loss": 2.4857,
      "theoretical_loss": 3.54344157436853,
      "tokens_seen": 1374748672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946940822467402,
      "loss": 2.5248,
      "theoretical_loss": 3.5434264309620573,
      "tokens_seen": 1374814208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029468405215646944,
      "loss": 2.6708,
      "theoretical_loss": 3.5434112884795517,
      "tokens_seen": 1374879744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029467402206619856,
      "loss": 2.6134,
      "theoretical_loss": 3.5433961469209114,
      "tokens_seen": 1374945280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946639919759278,
      "loss": 2.6086,
      "theoretical_loss": 3.543381006286037,
      "tokens_seen": 1375010816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294653961885657,
      "loss": 2.6643,
      "theoretical_loss": 3.5433658665748284,
      "tokens_seen": 1375076352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029464393179538616,
      "loss": 2.8531,
      "theoretical_loss": 3.543350727787184,
      "tokens_seen": 1375141888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029463390170511534,
      "loss": 2.8589,
      "theoretical_loss": 3.5433355899230046,
      "tokens_seen": 1375207424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946238716148446,
      "loss": 2.6566,
      "theoretical_loss": 3.543320452982189,
      "tokens_seen": 1375272960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946138415245737,
      "loss": 2.6606,
      "theoretical_loss": 3.543305316964638,
      "tokens_seen": 1375338496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029460381143430294,
      "loss": 2.6512,
      "theoretical_loss": 3.5432901818702502,
      "tokens_seen": 1375404032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029459378134403207,
      "loss": 2.7956,
      "theoretical_loss": 3.5432750476989257,
      "tokens_seen": 1375469568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945837512537613,
      "loss": 2.6258,
      "theoretical_loss": 3.543259914450564,
      "tokens_seen": 1375535104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945737211634905,
      "loss": 2.6525,
      "theoretical_loss": 3.5432447821250657,
      "tokens_seen": 1375600640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029456369107321966,
      "loss": 2.5247,
      "theoretical_loss": 3.5432296507223295,
      "tokens_seen": 1375666176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029455366098294885,
      "loss": 2.5934,
      "theoretical_loss": 3.5432145202422562,
      "tokens_seen": 1375731712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294543630892678,
      "loss": 2.6191,
      "theoretical_loss": 3.5431993906847445,
      "tokens_seen": 1375797248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945336008024072,
      "loss": 2.7456,
      "theoretical_loss": 3.543184262049695,
      "tokens_seen": 1375862784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029452357071213644,
      "loss": 2.5578,
      "theoretical_loss": 3.543169134337007,
      "tokens_seen": 1375928320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029451354062186557,
      "loss": 2.9169,
      "theoretical_loss": 3.5431540075465806,
      "tokens_seen": 1375993856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945035105315948,
      "loss": 2.7079,
      "theoretical_loss": 3.543138881678316,
      "tokens_seen": 1376059392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029449348044132393,
      "loss": 2.4352,
      "theoretical_loss": 3.543123756732112,
      "tokens_seen": 1376124928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029448345035105317,
      "loss": 2.5168,
      "theoretical_loss": 3.54310863270787,
      "tokens_seen": 1376190464
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 759249,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.637690305709839,
      "objective/train/theoretical_loss": 3.5430935096054883,
      "objective/train/tokens_used": 1396716000,
      "theoretical_loss": 3.5430935096054883,
      "tokens_seen": 1376256000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029447342026078235,
      "loss": 2.5261,
      "theoretical_loss": 3.5430935096054883,
      "tokens_seen": 1376256000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029446339017051153,
      "loss": 2.5534,
      "theoretical_loss": 3.5430783874248686,
      "tokens_seen": 1376321536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944533600802407,
      "loss": 2.5707,
      "theoretical_loss": 3.543063266165909,
      "tokens_seen": 1376387072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029444332998996995,
      "loss": 2.5748,
      "theoretical_loss": 3.543048145828511,
      "tokens_seen": 1376452608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944332998996991,
      "loss": 2.6966,
      "theoretical_loss": 3.543033026412573,
      "tokens_seen": 1376518144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944232698094283,
      "loss": 2.5546,
      "theoretical_loss": 3.5430179079179958,
      "tokens_seen": 1376583680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029441323971915744,
      "loss": 2.5579,
      "theoretical_loss": 3.54300279034468,
      "tokens_seen": 1376649216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029440320962888667,
      "loss": 2.4806,
      "theoretical_loss": 3.542987673692525,
      "tokens_seen": 1376714752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029439317953861585,
      "loss": 2.6588,
      "theoretical_loss": 3.5429725579614306,
      "tokens_seen": 1376780288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029438314944834503,
      "loss": 2.7941,
      "theoretical_loss": 3.5429574431512973,
      "tokens_seen": 1376845824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943731193580742,
      "loss": 2.6557,
      "theoretical_loss": 3.542942329262025,
      "tokens_seen": 1376911360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943630892678034,
      "loss": 2.6905,
      "theoretical_loss": 3.5429272162935135,
      "tokens_seen": 1376976896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943530591775326,
      "loss": 2.5786,
      "theoretical_loss": 3.542912104245663,
      "tokens_seen": 1377042432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943430290872618,
      "loss": 2.6899,
      "theoretical_loss": 3.542896993118374,
      "tokens_seen": 1377107968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029433299899699094,
      "loss": 2.6485,
      "theoretical_loss": 3.542881882911546,
      "tokens_seen": 1377173504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943229689067202,
      "loss": 2.6091,
      "theoretical_loss": 3.5428667736250796,
      "tokens_seen": 1377239040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943129388164493,
      "loss": 2.5071,
      "theoretical_loss": 3.5428516652588753,
      "tokens_seen": 1377304576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029430290872617854,
      "loss": 2.6595,
      "theoretical_loss": 3.542836557812832,
      "tokens_seen": 1377370112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029429287863590777,
      "loss": 2.6319,
      "theoretical_loss": 3.542821451286851,
      "tokens_seen": 1377435648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942828485456369,
      "loss": 2.3688,
      "theoretical_loss": 3.542806345680832,
      "tokens_seen": 1377501184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029427281845536613,
      "loss": 2.643,
      "theoretical_loss": 3.5427912409946756,
      "tokens_seen": 1377566720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942627883650953,
      "loss": 2.614,
      "theoretical_loss": 3.542776137228282,
      "tokens_seen": 1377632256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942527582748245,
      "loss": 2.5967,
      "theoretical_loss": 3.5427610343815505,
      "tokens_seen": 1377697792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942427281845537,
      "loss": 2.6822,
      "theoretical_loss": 3.542745932454383,
      "tokens_seen": 1377763328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029423269809428286,
      "loss": 2.5199,
      "theoretical_loss": 3.5427308314466783,
      "tokens_seen": 1377828864
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 760358,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.197251319885254,
      "objective/train/theoretical_loss": 3.542715731358337,
      "objective/train/tokens_used": 1398354400,
      "theoretical_loss": 3.542715731358337,
      "tokens_seen": 1377894400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029422266800401204,
      "loss": 2.4138,
      "theoretical_loss": 3.542715731358337,
      "tokens_seen": 1377894400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942126379137413,
      "loss": 2.4419,
      "theoretical_loss": 3.54270063218926,
      "tokens_seen": 1377959936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942026078234704,
      "loss": 2.6702,
      "theoretical_loss": 3.5426855339393475,
      "tokens_seen": 1378025472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029419257773319964,
      "loss": 2.6982,
      "theoretical_loss": 3.5426704366084993,
      "tokens_seen": 1378091008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029418254764292876,
      "loss": 2.4193,
      "theoretical_loss": 3.542655340196616,
      "tokens_seen": 1378156544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294172517552658,
      "loss": 2.7807,
      "theoretical_loss": 3.5426402447035983,
      "tokens_seen": 1378222080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941624874623872,
      "loss": 2.7231,
      "theoretical_loss": 3.5426251501293464,
      "tokens_seen": 1378287616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029415245737211636,
      "loss": 2.6748,
      "theoretical_loss": 3.5426100564737606,
      "tokens_seen": 1378353152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029414242728184554,
      "loss": 2.716,
      "theoretical_loss": 3.542594963736741,
      "tokens_seen": 1378418688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941323971915748,
      "loss": 2.754,
      "theoretical_loss": 3.5425798719181882,
      "tokens_seen": 1378484224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941223671013039,
      "loss": 2.5394,
      "theoretical_loss": 3.5425647810180037,
      "tokens_seen": 1378549760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029411233701103314,
      "loss": 2.621,
      "theoretical_loss": 3.5425496910360863,
      "tokens_seen": 1378615296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029410230692076227,
      "loss": 2.585,
      "theoretical_loss": 3.5425346019723376,
      "tokens_seen": 1378680832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940922768304915,
      "loss": 2.5968,
      "theoretical_loss": 3.542519513826657,
      "tokens_seen": 1378746368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940822467402207,
      "loss": 2.478,
      "theoretical_loss": 3.5425044265989465,
      "tokens_seen": 1378811904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029407221664994986,
      "loss": 2.7627,
      "theoretical_loss": 3.542489340289106,
      "tokens_seen": 1378877440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029406218655967905,
      "loss": 2.4349,
      "theoretical_loss": 3.5424742548970354,
      "tokens_seen": 1378942976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940521564694082,
      "loss": 2.6054,
      "theoretical_loss": 3.542459170422636,
      "tokens_seen": 1379008512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940421263791374,
      "loss": 2.7946,
      "theoretical_loss": 3.542444086865808,
      "tokens_seen": 1379074048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029403209628886664,
      "loss": 2.7524,
      "theoretical_loss": 3.5424290042264523,
      "tokens_seen": 1379139584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029402206619859577,
      "loss": 2.8018,
      "theoretical_loss": 3.542413922504469,
      "tokens_seen": 1379205120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294012036108325,
      "loss": 2.7234,
      "theoretical_loss": 3.542398841699759,
      "tokens_seen": 1379270656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029400200601805413,
      "loss": 2.4455,
      "theoretical_loss": 3.5423837618122236,
      "tokens_seen": 1379336192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029399197592778337,
      "loss": 2.5743,
      "theoretical_loss": 3.5423686828417624,
      "tokens_seen": 1379401728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029398194583751255,
      "loss": 2.6166,
      "theoretical_loss": 3.542353604788276,
      "tokens_seen": 1379467264
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 760756,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8100745677948,
      "objective/train/theoretical_loss": 3.5423385276516663,
      "objective/train/tokens_used": 1399992800,
      "theoretical_loss": 3.5423385276516663,
      "tokens_seen": 1379532800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029397191574724173,
      "loss": 2.5965,
      "theoretical_loss": 3.5423385276516663,
      "tokens_seen": 1379532800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939618856569709,
      "loss": 2.5801,
      "theoretical_loss": 3.542323451431833,
      "tokens_seen": 1379598336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029395185556670015,
      "loss": 2.5508,
      "theoretical_loss": 3.542308376128677,
      "tokens_seen": 1379663872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939418254764293,
      "loss": 2.4807,
      "theoretical_loss": 3.5422933017420988,
      "tokens_seen": 1379729408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939317953861585,
      "loss": 2.7529,
      "theoretical_loss": 3.542278228271999,
      "tokens_seen": 1379794944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029392176529588764,
      "loss": 2.5369,
      "theoretical_loss": 3.5422631557182798,
      "tokens_seen": 1379860480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029391173520561687,
      "loss": 2.6885,
      "theoretical_loss": 3.542248084080841,
      "tokens_seen": 1379926016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029390170511534605,
      "loss": 2.6143,
      "theoretical_loss": 3.5422330133595823,
      "tokens_seen": 1379991552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029389167502507523,
      "loss": 2.5206,
      "theoretical_loss": 3.542217943554406,
      "tokens_seen": 1380057088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938816449348044,
      "loss": 2.5229,
      "theoretical_loss": 3.5422028746652123,
      "tokens_seen": 1380122624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938716148445336,
      "loss": 2.7405,
      "theoretical_loss": 3.542187806691903,
      "tokens_seen": 1380188160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938615847542628,
      "loss": 2.507,
      "theoretical_loss": 3.5421727396343767,
      "tokens_seen": 1380253696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293851554663992,
      "loss": 2.6167,
      "theoretical_loss": 3.542157673492537,
      "tokens_seen": 1380319232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029384152457372114,
      "loss": 2.7905,
      "theoretical_loss": 3.542142608266283,
      "tokens_seen": 1380384768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938314944834504,
      "loss": 2.7074,
      "theoretical_loss": 3.542127543955515,
      "tokens_seen": 1380450304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938214643931795,
      "loss": 2.5172,
      "theoretical_loss": 3.542112480560136,
      "tokens_seen": 1380515840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029381143430290874,
      "loss": 2.6529,
      "theoretical_loss": 3.542097418080046,
      "tokens_seen": 1380581376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938014042126379,
      "loss": 2.8394,
      "theoretical_loss": 3.5420823565151456,
      "tokens_seen": 1380646912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937913741223671,
      "loss": 2.6438,
      "theoretical_loss": 3.542067295865336,
      "tokens_seen": 1380712448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937813440320963,
      "loss": 2.588,
      "theoretical_loss": 3.542052236130518,
      "tokens_seen": 1380777984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937713139418255,
      "loss": 2.888,
      "theoretical_loss": 3.542037177310593,
      "tokens_seen": 1380843520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029376128385155464,
      "loss": 2.8485,
      "theoretical_loss": 3.5420221194054617,
      "tokens_seen": 1380909056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937512537612839,
      "loss": 2.633,
      "theoretical_loss": 3.542007062415025,
      "tokens_seen": 1380974592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293741223671013,
      "loss": 2.4821,
      "theoretical_loss": 3.541992006339184,
      "tokens_seen": 1381040128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029373119358074224,
      "loss": 2.6478,
      "theoretical_loss": 3.54197695117784,
      "tokens_seen": 1381105664
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 761793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.823936700820923,
      "objective/train/theoretical_loss": 3.5419618969308937,
      "objective/train/tokens_used": 1401631200,
      "theoretical_loss": 3.5419618969308937,
      "tokens_seen": 1381171200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937211634904714,
      "loss": 2.7591,
      "theoretical_loss": 3.5419618969308937,
      "tokens_seen": 1381171200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937111334002006,
      "loss": 2.7321,
      "theoretical_loss": 3.5419468435982466,
      "tokens_seen": 1381236736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937011033099298,
      "loss": 2.7409,
      "theoretical_loss": 3.5419317911798,
      "tokens_seen": 1381302272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029369107321965896,
      "loss": 2.6413,
      "theoretical_loss": 3.541916739675454,
      "tokens_seen": 1381367808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029368104312938815,
      "loss": 2.6433,
      "theoretical_loss": 3.54190168908511,
      "tokens_seen": 1381433344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936710130391174,
      "loss": 2.8848,
      "theoretical_loss": 3.54188663940867,
      "tokens_seen": 1381498880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936609829488465,
      "loss": 2.677,
      "theoretical_loss": 3.5418715906460343,
      "tokens_seen": 1381564416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029365095285857574,
      "loss": 2.5477,
      "theoretical_loss": 3.5418565427971043,
      "tokens_seen": 1381629952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029364092276830487,
      "loss": 2.5925,
      "theoretical_loss": 3.5418414958617817,
      "tokens_seen": 1381695488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936308926780341,
      "loss": 2.6368,
      "theoretical_loss": 3.541826449839967,
      "tokens_seen": 1381761024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936208625877633,
      "loss": 2.4387,
      "theoretical_loss": 3.5418114047315616,
      "tokens_seen": 1381826560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029361083249749247,
      "loss": 2.7707,
      "theoretical_loss": 3.541796360536467,
      "tokens_seen": 1381892096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029360080240722165,
      "loss": 2.6518,
      "theoretical_loss": 3.5417813172545842,
      "tokens_seen": 1381957632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935907723169509,
      "loss": 2.5689,
      "theoretical_loss": 3.5417662748858145,
      "tokens_seen": 1382023168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029358074222668,
      "loss": 2.5661,
      "theoretical_loss": 3.541751233430059,
      "tokens_seen": 1382088704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029357071213640925,
      "loss": 2.5498,
      "theoretical_loss": 3.5417361928872193,
      "tokens_seen": 1382154240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935606820461384,
      "loss": 2.6417,
      "theoretical_loss": 3.541721153257197,
      "tokens_seen": 1382219776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935506519558676,
      "loss": 2.6782,
      "theoretical_loss": 3.5417061145398927,
      "tokens_seen": 1382285312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029354062186559684,
      "loss": 2.8397,
      "theoretical_loss": 3.541691076735208,
      "tokens_seen": 1382350848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029353059177532597,
      "loss": 2.495,
      "theoretical_loss": 3.5416760398430442,
      "tokens_seen": 1382416384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935205616850552,
      "loss": 2.8524,
      "theoretical_loss": 3.5416610038633034,
      "tokens_seen": 1382481920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029351053159478433,
      "loss": 2.7181,
      "theoretical_loss": 3.541645968795886,
      "tokens_seen": 1382547456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029350050150451357,
      "loss": 2.7833,
      "theoretical_loss": 3.541630934640694,
      "tokens_seen": 1382612992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029349047141424275,
      "loss": 2.5108,
      "theoretical_loss": 3.541615901397628,
      "tokens_seen": 1382678528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029348044132397193,
      "loss": 2.6661,
      "theoretical_loss": 3.5416008690665906,
      "tokens_seen": 1382744064
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 762358,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7813034057617188,
      "objective/train/theoretical_loss": 3.5415858376474825,
      "objective/train/tokens_used": 1403269600,
      "theoretical_loss": 3.5415858376474825,
      "tokens_seen": 1382809600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934704112337011,
      "loss": 2.7684,
      "theoretical_loss": 3.5415858376474825,
      "tokens_seen": 1382809600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029346038114343035,
      "loss": 2.6129,
      "theoretical_loss": 3.5415708071402054,
      "tokens_seen": 1382875136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934503510531595,
      "loss": 2.6347,
      "theoretical_loss": 3.5415557775446604,
      "tokens_seen": 1382940672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934403209628887,
      "loss": 2.5538,
      "theoretical_loss": 3.5415407488607498,
      "tokens_seen": 1383006208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029343029087261784,
      "loss": 2.7402,
      "theoretical_loss": 3.5415257210883744,
      "tokens_seen": 1383071744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029342026078234707,
      "loss": 2.5256,
      "theoretical_loss": 3.541510694227436,
      "tokens_seen": 1383137280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029341023069207625,
      "loss": 2.6425,
      "theoretical_loss": 3.541495668277836,
      "tokens_seen": 1383202816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029340020060180543,
      "loss": 2.5158,
      "theoretical_loss": 3.541480643239476,
      "tokens_seen": 1383268352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933901705115346,
      "loss": 2.6083,
      "theoretical_loss": 3.541465619112258,
      "tokens_seen": 1383333888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933801404212638,
      "loss": 2.6756,
      "theoretical_loss": 3.541450595896083,
      "tokens_seen": 1383399424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293370110330993,
      "loss": 2.635,
      "theoretical_loss": 3.541435573590852,
      "tokens_seen": 1383464960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933600802407222,
      "loss": 2.581,
      "theoretical_loss": 3.5414205521964686,
      "tokens_seen": 1383530496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029335005015045134,
      "loss": 2.7649,
      "theoretical_loss": 3.5414055317128326,
      "tokens_seen": 1383596032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933400200601806,
      "loss": 2.5137,
      "theoretical_loss": 3.5413905121398463,
      "tokens_seen": 1383661568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933299899699097,
      "loss": 2.7317,
      "theoretical_loss": 3.5413754934774113,
      "tokens_seen": 1383727104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029331995987963894,
      "loss": 2.8341,
      "theoretical_loss": 3.54136047572543,
      "tokens_seen": 1383792640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933099297893681,
      "loss": 2.8915,
      "theoretical_loss": 3.5413454588838027,
      "tokens_seen": 1383858176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932998996990973,
      "loss": 2.8376,
      "theoretical_loss": 3.541330442952432,
      "tokens_seen": 1383923712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932898696088265,
      "loss": 2.5754,
      "theoretical_loss": 3.5413154279312193,
      "tokens_seen": 1383989248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932798395185557,
      "loss": 2.5504,
      "theoretical_loss": 3.5413004138200668,
      "tokens_seen": 1384054784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029326980942828484,
      "loss": 2.8385,
      "theoretical_loss": 3.541285400618876,
      "tokens_seen": 1384120320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932597793380141,
      "loss": 2.7744,
      "theoretical_loss": 3.5412703883275487,
      "tokens_seen": 1384185856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932497492477432,
      "loss": 2.6171,
      "theoretical_loss": 3.541255376945986,
      "tokens_seen": 1384251392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029323971915747244,
      "loss": 2.763,
      "theoretical_loss": 3.5412403664740904,
      "tokens_seen": 1384316928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932296890672016,
      "loss": 2.5659,
      "theoretical_loss": 3.541225356911764,
      "tokens_seen": 1384382464
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 763291,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.771702289581299,
      "objective/train/theoretical_loss": 3.5412103482589083,
      "objective/train/tokens_used": 1404908000,
      "theoretical_loss": 3.5412103482589083,
      "tokens_seen": 1384448000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932196589769308,
      "loss": 2.7747,
      "theoretical_loss": 3.5412103482589083,
      "tokens_seen": 1384448000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029320962888666,
      "loss": 2.5403,
      "theoretical_loss": 3.541195340515425,
      "tokens_seen": 1384513536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029319959879638916,
      "loss": 2.7576,
      "theoretical_loss": 3.5411803336812158,
      "tokens_seen": 1384579072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029318956870611835,
      "loss": 2.7168,
      "theoretical_loss": 3.5411653277561825,
      "tokens_seen": 1384644608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931795386158476,
      "loss": 2.7653,
      "theoretical_loss": 3.541150322740228,
      "tokens_seen": 1384710144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931695085255767,
      "loss": 2.7363,
      "theoretical_loss": 3.541135318633253,
      "tokens_seen": 1384775680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029315947843530594,
      "loss": 2.5727,
      "theoretical_loss": 3.5411203154351605,
      "tokens_seen": 1384841216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029314944834503507,
      "loss": 2.5424,
      "theoretical_loss": 3.5411053131458514,
      "tokens_seen": 1384906752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931394182547643,
      "loss": 2.6556,
      "theoretical_loss": 3.5410903117652284,
      "tokens_seen": 1384972288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931293881644935,
      "loss": 2.5225,
      "theoretical_loss": 3.5410753112931928,
      "tokens_seen": 1385037824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029311935807422267,
      "loss": 2.6571,
      "theoretical_loss": 3.5410603117296473,
      "tokens_seen": 1385103360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029310932798395185,
      "loss": 2.5987,
      "theoretical_loss": 3.5410453130744934,
      "tokens_seen": 1385168896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930992978936811,
      "loss": 2.544,
      "theoretical_loss": 3.5410303153276335,
      "tokens_seen": 1385234432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930892678034102,
      "loss": 2.7685,
      "theoretical_loss": 3.541015318488969,
      "tokens_seen": 1385299968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029307923771313945,
      "loss": 2.626,
      "theoretical_loss": 3.5410003225584026,
      "tokens_seen": 1385365504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930692076228686,
      "loss": 2.6125,
      "theoretical_loss": 3.5409853275358363,
      "tokens_seen": 1385431040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930591775325978,
      "loss": 2.4994,
      "theoretical_loss": 3.5409703334211717,
      "tokens_seen": 1385496576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293049147442327,
      "loss": 2.7026,
      "theoretical_loss": 3.5409553402143112,
      "tokens_seen": 1385562112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029303911735205617,
      "loss": 2.5906,
      "theoretical_loss": 3.540940347915157,
      "tokens_seen": 1385627648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029302908726178535,
      "loss": 2.7143,
      "theoretical_loss": 3.540925356523611,
      "tokens_seen": 1385693184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029301905717151453,
      "loss": 2.6731,
      "theoretical_loss": 3.5409103660395758,
      "tokens_seen": 1385758720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930090270812437,
      "loss": 2.6631,
      "theoretical_loss": 3.5408953764629527,
      "tokens_seen": 1385824256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029299899699097295,
      "loss": 2.8875,
      "theoretical_loss": 3.5408803877936443,
      "tokens_seen": 1385889792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929889669007021,
      "loss": 2.5833,
      "theoretical_loss": 3.540865400031553,
      "tokens_seen": 1385955328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929789368104313,
      "loss": 2.5372,
      "theoretical_loss": 3.540850413176581,
      "tokens_seen": 1386020864
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 764545,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.004178762435913,
      "objective/train/theoretical_loss": 3.5408354272286298,
      "objective/train/tokens_used": 1406546400,
      "theoretical_loss": 3.5408354272286298,
      "tokens_seen": 1386086400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929689067201605,
      "loss": 2.9131,
      "theoretical_loss": 3.5408354272286298,
      "tokens_seen": 1386086400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929588766298897,
      "loss": 2.6061,
      "theoretical_loss": 3.540820442187602,
      "tokens_seen": 1386151936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029294884653961885,
      "loss": 2.6359,
      "theoretical_loss": 3.540805458053401,
      "tokens_seen": 1386217472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029293881644934804,
      "loss": 2.5335,
      "theoretical_loss": 3.5407904748259273,
      "tokens_seen": 1386283008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929287863590772,
      "loss": 2.7497,
      "theoretical_loss": 3.5407754925050843,
      "tokens_seen": 1386348544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029291875626880645,
      "loss": 2.8104,
      "theoretical_loss": 3.540760511090774,
      "tokens_seen": 1386414080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929087261785356,
      "loss": 2.6975,
      "theoretical_loss": 3.540745530582898,
      "tokens_seen": 1386479616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928986960882648,
      "loss": 2.59,
      "theoretical_loss": 3.540730550981359,
      "tokens_seen": 1386545152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029288866599799394,
      "loss": 2.5904,
      "theoretical_loss": 3.54071557228606,
      "tokens_seen": 1386610688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928786359077232,
      "loss": 2.5001,
      "theoretical_loss": 3.540700594496903,
      "tokens_seen": 1386676224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029286860581745236,
      "loss": 2.5749,
      "theoretical_loss": 3.54068561761379,
      "tokens_seen": 1386741760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029285857572718154,
      "loss": 2.6409,
      "theoretical_loss": 3.5406706416366243,
      "tokens_seen": 1386807296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928485456369107,
      "loss": 2.6421,
      "theoretical_loss": 3.540655666565307,
      "tokens_seen": 1386872832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928385155466399,
      "loss": 2.4422,
      "theoretical_loss": 3.540640692399741,
      "tokens_seen": 1386938368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928284854563691,
      "loss": 2.6959,
      "theoretical_loss": 3.540625719139829,
      "tokens_seen": 1387003904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928184553660983,
      "loss": 2.5246,
      "theoretical_loss": 3.5406107467854735,
      "tokens_seen": 1387069440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029280842527582744,
      "loss": 2.4474,
      "theoretical_loss": 3.540595775336577,
      "tokens_seen": 1387134976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927983951855567,
      "loss": 2.6563,
      "theoretical_loss": 3.5405808047930405,
      "tokens_seen": 1387200512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927883650952859,
      "loss": 2.7903,
      "theoretical_loss": 3.540565835154769,
      "tokens_seen": 1387266048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029277833500501504,
      "loss": 2.6836,
      "theoretical_loss": 3.540550866421663,
      "tokens_seen": 1387331584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927683049147443,
      "loss": 2.59,
      "theoretical_loss": 3.5405358985936255,
      "tokens_seen": 1387397120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927582748244734,
      "loss": 2.516,
      "theoretical_loss": 3.5405209316705597,
      "tokens_seen": 1387462656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029274824473420264,
      "loss": 2.8805,
      "theoretical_loss": 3.5405059656523674,
      "tokens_seen": 1387528192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927382146439318,
      "loss": 2.623,
      "theoretical_loss": 3.5404910005389514,
      "tokens_seen": 1387593728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292728184553661,
      "loss": 2.5908,
      "theoretical_loss": 3.5404760363302143,
      "tokens_seen": 1387659264
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 765178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3576533794403076,
      "objective/train/theoretical_loss": 3.5404610730260595,
      "objective/train/tokens_used": 1408184800,
      "theoretical_loss": 3.5404610730260595,
      "tokens_seen": 1387724800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927181544633902,
      "loss": 2.3922,
      "theoretical_loss": 3.5404610730260595,
      "tokens_seen": 1387724800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029270812437311936,
      "loss": 2.5994,
      "theoretical_loss": 3.540446110626388,
      "tokens_seen": 1387790336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029269809428284855,
      "loss": 2.5658,
      "theoretical_loss": 3.540431149131103,
      "tokens_seen": 1387855872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926880641925778,
      "loss": 2.7686,
      "theoretical_loss": 3.5404161885401075,
      "tokens_seen": 1387921408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926780341023069,
      "loss": 2.4703,
      "theoretical_loss": 3.5404012288533044,
      "tokens_seen": 1387986944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029266800401203614,
      "loss": 2.8259,
      "theoretical_loss": 3.5403862700705955,
      "tokens_seen": 1388052480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029265797392176527,
      "loss": 2.5285,
      "theoretical_loss": 3.5403713121918843,
      "tokens_seen": 1388118016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926479438314945,
      "loss": 2.5711,
      "theoretical_loss": 3.5403563552170727,
      "tokens_seen": 1388183552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926379137412237,
      "loss": 2.849,
      "theoretical_loss": 3.540341399146064,
      "tokens_seen": 1388249088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029262788365095287,
      "loss": 2.4781,
      "theoretical_loss": 3.5403264439787607,
      "tokens_seen": 1388314624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029261785356068205,
      "loss": 2.6438,
      "theoretical_loss": 3.540311489715066,
      "tokens_seen": 1388380160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926078234704113,
      "loss": 2.6651,
      "theoretical_loss": 3.5402965363548815,
      "tokens_seen": 1388445696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925977933801404,
      "loss": 2.7053,
      "theoretical_loss": 3.5402815838981114,
      "tokens_seen": 1388511232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029258776328986965,
      "loss": 2.6005,
      "theoretical_loss": 3.5402666323446574,
      "tokens_seen": 1388576768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925777331995988,
      "loss": 2.5783,
      "theoretical_loss": 3.5402516816944227,
      "tokens_seen": 1388642304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292567703109328,
      "loss": 2.5394,
      "theoretical_loss": 3.54023673194731,
      "tokens_seen": 1388707840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925576730190572,
      "loss": 2.6339,
      "theoretical_loss": 3.5402217831032226,
      "tokens_seen": 1388773376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029254764292878637,
      "loss": 2.4864,
      "theoretical_loss": 3.5402068351620626,
      "tokens_seen": 1388838912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029253761283851555,
      "loss": 2.5444,
      "theoretical_loss": 3.5401918881237338,
      "tokens_seen": 1388904448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029252758274824473,
      "loss": 2.5277,
      "theoretical_loss": 3.540176941988138,
      "tokens_seen": 1388969984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925175526579739,
      "loss": 2.6024,
      "theoretical_loss": 3.5401619967551787,
      "tokens_seen": 1389035520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029250752256770315,
      "loss": 2.5055,
      "theoretical_loss": 3.540147052424759,
      "tokens_seen": 1389101056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924974924774323,
      "loss": 2.4434,
      "theoretical_loss": 3.540132108996781,
      "tokens_seen": 1389166592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924874623871615,
      "loss": 2.9739,
      "theoretical_loss": 3.5401171664711484,
      "tokens_seen": 1389232128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924774322968907,
      "loss": 2.6789,
      "theoretical_loss": 3.5401022248477636,
      "tokens_seen": 1389297664
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 766177,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7885921001434326,
      "objective/train/theoretical_loss": 3.540087284126531,
      "objective/train/tokens_used": 1409823200,
      "theoretical_loss": 3.540087284126531,
      "tokens_seen": 1389363200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924674022066199,
      "loss": 2.7667,
      "theoretical_loss": 3.540087284126531,
      "tokens_seen": 1389363200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029245737211634906,
      "loss": 2.6305,
      "theoretical_loss": 3.540072344307351,
      "tokens_seen": 1389428736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029244734202607824,
      "loss": 2.6864,
      "theoretical_loss": 3.540057405390129,
      "tokens_seen": 1389494272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924373119358074,
      "loss": 2.5925,
      "theoretical_loss": 3.5400424673747666,
      "tokens_seen": 1389559808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029242728184553665,
      "loss": 2.6318,
      "theoretical_loss": 3.5400275302611677,
      "tokens_seen": 1389625344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924172517552658,
      "loss": 2.6014,
      "theoretical_loss": 3.5400125940492346,
      "tokens_seen": 1389690880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292407221664995,
      "loss": 2.8096,
      "theoretical_loss": 3.5399976587388706,
      "tokens_seen": 1389756416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029239719157472414,
      "loss": 2.7303,
      "theoretical_loss": 3.5399827243299793,
      "tokens_seen": 1389821952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923871614844534,
      "loss": 2.6735,
      "theoretical_loss": 3.5399677908224625,
      "tokens_seen": 1389887488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029237713139418256,
      "loss": 2.5112,
      "theoretical_loss": 3.5399528582162247,
      "tokens_seen": 1389953024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029236710130391174,
      "loss": 2.6793,
      "theoretical_loss": 3.5399379265111683,
      "tokens_seen": 1390018560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923570712136409,
      "loss": 2.6415,
      "theoretical_loss": 3.539922995707197,
      "tokens_seen": 1390084096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923470411233701,
      "loss": 2.5822,
      "theoretical_loss": 3.5399080658042132,
      "tokens_seen": 1390149632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923370110330993,
      "loss": 2.6738,
      "theoretical_loss": 3.53989313680212,
      "tokens_seen": 1390215168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923269809428285,
      "loss": 2.5026,
      "theoretical_loss": 3.5398782087008214,
      "tokens_seen": 1390280704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029231695085255765,
      "loss": 2.5785,
      "theoretical_loss": 3.53986328150022,
      "tokens_seen": 1390346240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923069207622869,
      "loss": 2.6385,
      "theoretical_loss": 3.539848355200219,
      "tokens_seen": 1390411776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029229689067201606,
      "loss": 2.5758,
      "theoretical_loss": 3.539833429800722,
      "tokens_seen": 1390477312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029228686058174524,
      "loss": 2.6024,
      "theoretical_loss": 3.539818505301632,
      "tokens_seen": 1390542848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922768304914744,
      "loss": 2.8159,
      "theoretical_loss": 3.539803581702852,
      "tokens_seen": 1390608384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922668004012036,
      "loss": 2.589,
      "theoretical_loss": 3.5397886590042855,
      "tokens_seen": 1390673920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922567703109328,
      "loss": 2.5165,
      "theoretical_loss": 3.539773737205836,
      "tokens_seen": 1390739456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292246740220662,
      "loss": 2.3329,
      "theoretical_loss": 3.5397588163074065,
      "tokens_seen": 1390804992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029223671013039115,
      "loss": 2.5392,
      "theoretical_loss": 3.5397438963089005,
      "tokens_seen": 1390870528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922266800401204,
      "loss": 2.6437,
      "theoretical_loss": 3.5397289772102205,
      "tokens_seen": 1390936064
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 766696,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6347217559814453,
      "objective/train/theoretical_loss": 3.5397140590112715,
      "objective/train/tokens_used": 1411461600,
      "theoretical_loss": 3.5397140590112715,
      "tokens_seen": 1391001600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922166499498495,
      "loss": 2.618,
      "theoretical_loss": 3.5397140590112715,
      "tokens_seen": 1391001600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029220661985957875,
      "loss": 2.6174,
      "theoretical_loss": 3.5396991417119557,
      "tokens_seen": 1391067136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921965897693079,
      "loss": 2.5041,
      "theoretical_loss": 3.5396842253121763,
      "tokens_seen": 1391132672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921865596790371,
      "loss": 2.5927,
      "theoretical_loss": 3.539669309811837,
      "tokens_seen": 1391198208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921765295887663,
      "loss": 2.6222,
      "theoretical_loss": 3.5396543952108415,
      "tokens_seen": 1391263744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029216649949849547,
      "loss": 2.5897,
      "theoretical_loss": 3.5396394815090932,
      "tokens_seen": 1391329280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029215646940822465,
      "loss": 2.7557,
      "theoretical_loss": 3.5396245687064947,
      "tokens_seen": 1391394816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921464393179539,
      "loss": 2.5768,
      "theoretical_loss": 3.5396096568029503,
      "tokens_seen": 1391460352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292136409227683,
      "loss": 2.5327,
      "theoretical_loss": 3.5395947457983636,
      "tokens_seen": 1391525888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029212637913741225,
      "loss": 2.5041,
      "theoretical_loss": 3.539579835692637,
      "tokens_seen": 1391591424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029211634904714143,
      "loss": 2.5025,
      "theoretical_loss": 3.5395649264856752,
      "tokens_seen": 1391656960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921063189568706,
      "loss": 2.6626,
      "theoretical_loss": 3.5395500181773807,
      "tokens_seen": 1391722496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920962888665998,
      "loss": 2.5822,
      "theoretical_loss": 3.539535110767658,
      "tokens_seen": 1391788032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292086258776329,
      "loss": 2.6969,
      "theoretical_loss": 3.5395202042564096,
      "tokens_seen": 1391853568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029207622868605815,
      "loss": 2.4481,
      "theoretical_loss": 3.5395052986435394,
      "tokens_seen": 1391919104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920661985957874,
      "loss": 2.467,
      "theoretical_loss": 3.5394903939289515,
      "tokens_seen": 1391984640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920561685055165,
      "loss": 2.4963,
      "theoretical_loss": 3.539475490112549,
      "tokens_seen": 1392050176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029204613841524575,
      "loss": 2.6307,
      "theoretical_loss": 3.539460587194236,
      "tokens_seen": 1392115712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029203610832497493,
      "loss": 2.6562,
      "theoretical_loss": 3.539445685173915,
      "tokens_seen": 1392181248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920260782347041,
      "loss": 2.4054,
      "theoretical_loss": 3.5394307840514907,
      "tokens_seen": 1392246784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029201604814443335,
      "loss": 2.7607,
      "theoretical_loss": 3.5394158838268663,
      "tokens_seen": 1392312320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920060180541625,
      "loss": 2.5339,
      "theoretical_loss": 3.5394009844999457,
      "tokens_seen": 1392377856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919959879638917,
      "loss": 2.8024,
      "theoretical_loss": 3.539386086070632,
      "tokens_seen": 1392443392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919859578736209,
      "loss": 2.6481,
      "theoretical_loss": 3.5393711885388295,
      "tokens_seen": 1392508928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919759277833501,
      "loss": 2.7028,
      "theoretical_loss": 3.5393562919044417,
      "tokens_seen": 1392574464
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 767843,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7980923652648926,
      "objective/train/theoretical_loss": 3.539341396167372,
      "objective/train/tokens_used": 1413100000,
      "theoretical_loss": 3.539341396167372,
      "tokens_seen": 1392640000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029196589769307926,
      "loss": 2.6455,
      "theoretical_loss": 3.539341396167372,
      "tokens_seen": 1392640000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029195586760280844,
      "loss": 2.4258,
      "theoretical_loss": 3.5393265013275244,
      "tokens_seen": 1392705536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919458375125376,
      "loss": 2.6302,
      "theoretical_loss": 3.5393116073848025,
      "tokens_seen": 1392771072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029193580742226685,
      "loss": 2.6789,
      "theoretical_loss": 3.5392967143391107,
      "tokens_seen": 1392836608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291925777331996,
      "loss": 2.4883,
      "theoretical_loss": 3.539281822190352,
      "tokens_seen": 1392902144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002919157472417252,
      "loss": 2.4174,
      "theoretical_loss": 3.53926693093843,
      "tokens_seen": 1392967680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029190571715145434,
      "loss": 2.7658,
      "theoretical_loss": 3.5392520405832495,
      "tokens_seen": 1393033216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918956870611836,
      "loss": 2.7246,
      "theoretical_loss": 3.5392371511247136,
      "tokens_seen": 1393098752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029188565697091276,
      "loss": 2.5244,
      "theoretical_loss": 3.5392222625627263,
      "tokens_seen": 1393164288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029187562688064194,
      "loss": 2.5479,
      "theoretical_loss": 3.5392073748971913,
      "tokens_seen": 1393229824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918655967903711,
      "loss": 2.537,
      "theoretical_loss": 3.5391924881280126,
      "tokens_seen": 1393295360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918555667001003,
      "loss": 2.5608,
      "theoretical_loss": 3.5391776022550943,
      "tokens_seen": 1393360896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918455366098295,
      "loss": 2.7187,
      "theoretical_loss": 3.5391627172783395,
      "tokens_seen": 1393426432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918355065195587,
      "loss": 2.633,
      "theoretical_loss": 3.539147833197653,
      "tokens_seen": 1393491968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029182547642928785,
      "loss": 2.5427,
      "theoretical_loss": 3.539132950012938,
      "tokens_seen": 1393557504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918154463390171,
      "loss": 2.6454,
      "theoretical_loss": 3.5391180677240994,
      "tokens_seen": 1393623040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029180541624874626,
      "loss": 2.4717,
      "theoretical_loss": 3.5391031863310403,
      "tokens_seen": 1393688576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029179538615847544,
      "loss": 2.6568,
      "theoretical_loss": 3.5390883058336655,
      "tokens_seen": 1393754112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917853560682046,
      "loss": 2.5793,
      "theoretical_loss": 3.5390734262318775,
      "tokens_seen": 1393819648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917753259779338,
      "loss": 2.7018,
      "theoretical_loss": 3.539058547525581,
      "tokens_seen": 1393885184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291765295887663,
      "loss": 2.5705,
      "theoretical_loss": 3.5390436697146814,
      "tokens_seen": 1393950720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917552657973922,
      "loss": 2.7231,
      "theoretical_loss": 3.5390287927990807,
      "tokens_seen": 1394016256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029174523570712135,
      "loss": 2.607,
      "theoretical_loss": 3.539013916778684,
      "tokens_seen": 1394081792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917352056168506,
      "loss": 2.7236,
      "theoretical_loss": 3.538999041653395,
      "tokens_seen": 1394147328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917251755265797,
      "loss": 2.5751,
      "theoretical_loss": 3.5389841674231177,
      "tokens_seen": 1394212864
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 768328,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.26961350440979,
      "objective/train/theoretical_loss": 3.5389692940877566,
      "objective/train/tokens_used": 1414738400,
      "theoretical_loss": 3.5389692940877566,
      "tokens_seen": 1394278400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029171514543630895,
      "loss": 2.4375,
      "theoretical_loss": 3.5389692940877566,
      "tokens_seen": 1394278400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917051153460381,
      "loss": 2.6149,
      "theoretical_loss": 3.5389544216472153,
      "tokens_seen": 1394343936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916950852557673,
      "loss": 2.6194,
      "theoretical_loss": 3.5389395501013983,
      "tokens_seen": 1394409472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916850551654965,
      "loss": 2.7015,
      "theoretical_loss": 3.5389246794502096,
      "tokens_seen": 1394475008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029167502507522567,
      "loss": 2.9316,
      "theoretical_loss": 3.5389098096935534,
      "tokens_seen": 1394540544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029166499498495485,
      "loss": 2.5839,
      "theoretical_loss": 3.5388949408313337,
      "tokens_seen": 1394606080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916549648946841,
      "loss": 2.7859,
      "theoretical_loss": 3.5388800728634546,
      "tokens_seen": 1394671616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916449348044132,
      "loss": 2.5476,
      "theoretical_loss": 3.538865205789821,
      "tokens_seen": 1394737152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029163490471414245,
      "loss": 2.5964,
      "theoretical_loss": 3.538850339610336,
      "tokens_seen": 1394802688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029162487462387163,
      "loss": 2.5813,
      "theoretical_loss": 3.5388354743249044,
      "tokens_seen": 1394868224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916148445336008,
      "loss": 2.5654,
      "theoretical_loss": 3.5388206099334303,
      "tokens_seen": 1394933760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029160481444333,
      "loss": 2.2822,
      "theoretical_loss": 3.5388057464358185,
      "tokens_seen": 1394999296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915947843530592,
      "loss": 2.5525,
      "theoretical_loss": 3.5387908838319726,
      "tokens_seen": 1395064832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029158475426278835,
      "loss": 2.6056,
      "theoretical_loss": 3.5387760221217968,
      "tokens_seen": 1395130368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915747241725176,
      "loss": 2.648,
      "theoretical_loss": 3.538761161305196,
      "tokens_seen": 1395195904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915646940822467,
      "loss": 2.6446,
      "theoretical_loss": 3.5387463013820737,
      "tokens_seen": 1395261440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029155466399197595,
      "loss": 2.778,
      "theoretical_loss": 3.538731442352335,
      "tokens_seen": 1395326976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915446339017051,
      "loss": 2.6212,
      "theoretical_loss": 3.5387165842158836,
      "tokens_seen": 1395392512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915346038114343,
      "loss": 2.6653,
      "theoretical_loss": 3.538701726972625,
      "tokens_seen": 1395458048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915245737211635,
      "loss": 2.6928,
      "theoretical_loss": 3.5386868706224615,
      "tokens_seen": 1395523584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915145436308927,
      "loss": 2.7496,
      "theoretical_loss": 3.5386720151652997,
      "tokens_seen": 1395589120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029150451354062186,
      "loss": 2.4478,
      "theoretical_loss": 3.5386571606010424,
      "tokens_seen": 1395654656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914944834503511,
      "loss": 2.743,
      "theoretical_loss": 3.538642306929595,
      "tokens_seen": 1395720192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914844533600802,
      "loss": 2.8987,
      "theoretical_loss": 3.538627454150861,
      "tokens_seen": 1395785728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029147442326980946,
      "loss": 2.6115,
      "theoretical_loss": 3.5386126022647457,
      "tokens_seen": 1395851264
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 769315,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.528290033340454,
      "objective/train/theoretical_loss": 3.538597751271153,
      "objective/train/tokens_used": 1416376800,
      "theoretical_loss": 3.538597751271153,
      "tokens_seen": 1395916800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914643931795386,
      "loss": 2.5921,
      "theoretical_loss": 3.538597751271153,
      "tokens_seen": 1395916800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914543630892678,
      "loss": 2.6469,
      "theoretical_loss": 3.538582901169988,
      "tokens_seen": 1395982336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291444332998997,
      "loss": 2.7489,
      "theoretical_loss": 3.538568051961154,
      "tokens_seen": 1396047872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914343029087262,
      "loss": 2.7785,
      "theoretical_loss": 3.5385532036445566,
      "tokens_seen": 1396113408
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029142427281845536,
      "loss": 2.6316,
      "theoretical_loss": 3.5385383562201005,
      "tokens_seen": 1396178944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029141424272818454,
      "loss": 2.4348,
      "theoretical_loss": 3.538523509687689,
      "tokens_seen": 1396244480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914042126379137,
      "loss": 2.5896,
      "theoretical_loss": 3.538508664047227,
      "tokens_seen": 1396310016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029139418254764296,
      "loss": 2.6467,
      "theoretical_loss": 3.53849381929862,
      "tokens_seen": 1396375552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913841524573721,
      "loss": 2.577,
      "theoretical_loss": 3.5384789754417723,
      "tokens_seen": 1396441088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913741223671013,
      "loss": 2.5934,
      "theoretical_loss": 3.5384641324765873,
      "tokens_seen": 1396506624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029136409227683045,
      "loss": 2.8131,
      "theoretical_loss": 3.538449290402971,
      "tokens_seen": 1396572160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913540621865597,
      "loss": 2.5786,
      "theoretical_loss": 3.538434449220827,
      "tokens_seen": 1396637696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029134403209628886,
      "loss": 2.612,
      "theoretical_loss": 3.5384196089300604,
      "tokens_seen": 1396703232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029133400200601805,
      "loss": 2.2821,
      "theoretical_loss": 3.538404769530576,
      "tokens_seen": 1396768768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913239719157472,
      "loss": 2.9357,
      "theoretical_loss": 3.5383899310222775,
      "tokens_seen": 1396834304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029131394182547646,
      "loss": 2.7306,
      "theoretical_loss": 3.5383750934050715,
      "tokens_seen": 1396899840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913039117352056,
      "loss": 2.6315,
      "theoretical_loss": 3.538360256678861,
      "tokens_seen": 1396965376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912938816449348,
      "loss": 2.7702,
      "theoretical_loss": 3.5383454208435507,
      "tokens_seen": 1397030912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291283851554664,
      "loss": 2.5113,
      "theoretical_loss": 3.5383305858990464,
      "tokens_seen": 1397096448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912738214643932,
      "loss": 2.7669,
      "theoretical_loss": 3.538315751845252,
      "tokens_seen": 1397161984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912637913741224,
      "loss": 2.7036,
      "theoretical_loss": 3.5383009186820726,
      "tokens_seen": 1397227520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029125376128385155,
      "loss": 2.7275,
      "theoretical_loss": 3.538286086409413,
      "tokens_seen": 1397293056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912437311935808,
      "loss": 2.5954,
      "theoretical_loss": 3.5382712550271775,
      "tokens_seen": 1397358592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912337011033099,
      "loss": 2.4268,
      "theoretical_loss": 3.5382564245352714,
      "tokens_seen": 1397424128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029122367101303915,
      "loss": 2.494,
      "theoretical_loss": 3.5382415949335995,
      "tokens_seen": 1397489664
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 769972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.501089096069336,
      "objective/train/theoretical_loss": 3.538226766222066,
      "objective/train/tokens_used": 1418015200,
      "theoretical_loss": 3.538226766222066,
      "tokens_seen": 1397555200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029121364092276833,
      "loss": 2.532,
      "theoretical_loss": 3.538226766222066,
      "tokens_seen": 1397555200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912036108324975,
      "loss": 2.5163,
      "theoretical_loss": 3.5382119384005764,
      "tokens_seen": 1397620736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911935807422267,
      "loss": 2.6278,
      "theoretical_loss": 3.538197111469035,
      "tokens_seen": 1397686272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029118355065195587,
      "loss": 2.5544,
      "theoretical_loss": 3.5381822854273475,
      "tokens_seen": 1397751808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029117352056168505,
      "loss": 2.9922,
      "theoretical_loss": 3.538167460275418,
      "tokens_seen": 1397817344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911634904714143,
      "loss": 2.6333,
      "theoretical_loss": 3.5381526360131517,
      "tokens_seen": 1397882880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911534603811434,
      "loss": 2.4192,
      "theoretical_loss": 3.5381378126404535,
      "tokens_seen": 1397948416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029114343029087265,
      "loss": 2.4743,
      "theoretical_loss": 3.5381229901572278,
      "tokens_seen": 1398013952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029113340020060183,
      "loss": 2.5949,
      "theoretical_loss": 3.5381081685633804,
      "tokens_seen": 1398079488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291123370110331,
      "loss": 2.545,
      "theoretical_loss": 3.538093347858816,
      "tokens_seen": 1398145024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911133400200602,
      "loss": 2.8087,
      "theoretical_loss": 3.538078528043439,
      "tokens_seen": 1398210560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911033099297894,
      "loss": 2.519,
      "theoretical_loss": 3.538063709117155,
      "tokens_seen": 1398276096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029109327983951855,
      "loss": 2.6899,
      "theoretical_loss": 3.538048891079869,
      "tokens_seen": 1398341632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910832497492478,
      "loss": 2.4182,
      "theoretical_loss": 3.5380340739314855,
      "tokens_seen": 1398407168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910732196589769,
      "loss": 2.6658,
      "theoretical_loss": 3.53801925767191,
      "tokens_seen": 1398472704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029106318956870615,
      "loss": 2.8556,
      "theoretical_loss": 3.538004442301047,
      "tokens_seen": 1398538240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910531594784353,
      "loss": 2.6686,
      "theoretical_loss": 3.537989627818802,
      "tokens_seen": 1398603776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910431293881645,
      "loss": 2.7685,
      "theoretical_loss": 3.5379748142250804,
      "tokens_seen": 1398669312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910330992978937,
      "loss": 2.7405,
      "theoretical_loss": 3.537960001519786,
      "tokens_seen": 1398734848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910230692076229,
      "loss": 2.5712,
      "theoretical_loss": 3.5379451897028256,
      "tokens_seen": 1398800384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029101303911735206,
      "loss": 2.6448,
      "theoretical_loss": 3.537930378774103,
      "tokens_seen": 1398865920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910030090270813,
      "loss": 2.484,
      "theoretical_loss": 3.537915568733524,
      "tokens_seen": 1398931456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909929789368104,
      "loss": 2.8006,
      "theoretical_loss": 3.5379007595809933,
      "tokens_seen": 1398996992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029098294884653966,
      "loss": 2.3332,
      "theoretical_loss": 3.5378859513164165,
      "tokens_seen": 1399062528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909729187562688,
      "loss": 2.6569,
      "theoretical_loss": 3.5378711439396984,
      "tokens_seen": 1399128064
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 771272,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4476327896118164,
      "objective/train/theoretical_loss": 3.5378563374507443,
      "objective/train/tokens_used": 1419653600,
      "theoretical_loss": 3.5378563374507443,
      "tokens_seen": 1399193600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290962888665998,
      "loss": 2.5499,
      "theoretical_loss": 3.5378563374507443,
      "tokens_seen": 1399193600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909528585757272,
      "loss": 2.6638,
      "theoretical_loss": 3.53784153184946,
      "tokens_seen": 1399259136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909428284854564,
      "loss": 2.605,
      "theoretical_loss": 3.5378267271357493,
      "tokens_seen": 1399324672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029093279839518556,
      "loss": 2.5923,
      "theoretical_loss": 3.537811923309518,
      "tokens_seen": 1399390208
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029092276830491474,
      "loss": 2.56,
      "theoretical_loss": 3.537797120370673,
      "tokens_seen": 1399455744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909127382146439,
      "loss": 2.7205,
      "theoretical_loss": 3.537782318319117,
      "tokens_seen": 1399521280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029090270812437316,
      "loss": 2.4415,
      "theoretical_loss": 3.537767517154757,
      "tokens_seen": 1399586816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908926780341023,
      "loss": 2.7773,
      "theoretical_loss": 3.537752716877497,
      "tokens_seen": 1399652352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908826479438315,
      "loss": 2.6969,
      "theoretical_loss": 3.5377379174872434,
      "tokens_seen": 1399717888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029087261785356065,
      "loss": 2.6716,
      "theoretical_loss": 3.5377231189839016,
      "tokens_seen": 1399783424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908625877632899,
      "loss": 2.5032,
      "theoretical_loss": 3.5377083213673757,
      "tokens_seen": 1399848960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029085255767301906,
      "loss": 2.6298,
      "theoretical_loss": 3.5376935246375716,
      "tokens_seen": 1399914496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029084252758274825,
      "loss": 2.6113,
      "theoretical_loss": 3.537678728794396,
      "tokens_seen": 1399980032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908324974924774,
      "loss": 2.6666,
      "theoretical_loss": 3.5376639338377522,
      "tokens_seen": 1400045568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029082246740220666,
      "loss": 2.436,
      "theoretical_loss": 3.537649139767547,
      "tokens_seen": 1400111104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908124373119358,
      "loss": 2.594,
      "theoretical_loss": 3.5376343465836846,
      "tokens_seen": 1400176640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290802407221665,
      "loss": 2.4731,
      "theoretical_loss": 3.5376195542860716,
      "tokens_seen": 1400242176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029079237713139415,
      "loss": 2.7171,
      "theoretical_loss": 3.5376047628746132,
      "tokens_seen": 1400307712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907823470411234,
      "loss": 2.5599,
      "theoretical_loss": 3.537589972349214,
      "tokens_seen": 1400373248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029077231695085257,
      "loss": 2.6427,
      "theoretical_loss": 3.53757518270978,
      "tokens_seen": 1400438784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029076228686058175,
      "loss": 2.5244,
      "theoretical_loss": 3.5375603939562175,
      "tokens_seen": 1400504320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029075225677031093,
      "loss": 2.5821,
      "theoretical_loss": 3.5375456060884307,
      "tokens_seen": 1400569856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907422266800401,
      "loss": 2.3923,
      "theoretical_loss": 3.5375308191063257,
      "tokens_seen": 1400635392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907321965897693,
      "loss": 2.5818,
      "theoretical_loss": 3.537516033009808,
      "tokens_seen": 1400700928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029072216649949853,
      "loss": 2.6046,
      "theoretical_loss": 3.537501247798783,
      "tokens_seen": 1400766464
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 771955,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5957698822021484,
      "objective/train/theoretical_loss": 3.537486463473156,
      "objective/train/tokens_used": 1421292000,
      "theoretical_loss": 3.537486463473156,
      "tokens_seen": 1400832000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029071213640922765,
      "loss": 2.6511,
      "theoretical_loss": 3.537486463473156,
      "tokens_seen": 1400832000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907021063189569,
      "loss": 2.589,
      "theoretical_loss": 3.5374716800328336,
      "tokens_seen": 1400897536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290692076228686,
      "loss": 2.6191,
      "theoretical_loss": 3.5374568974777203,
      "tokens_seen": 1400963072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029068204613841525,
      "loss": 2.4094,
      "theoretical_loss": 3.537442115807722,
      "tokens_seen": 1401028608
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029067201604814443,
      "loss": 2.8195,
      "theoretical_loss": 3.537427335022744,
      "tokens_seen": 1401094144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906619859578736,
      "loss": 2.4319,
      "theoretical_loss": 3.5374125551226925,
      "tokens_seen": 1401159680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906519558676028,
      "loss": 2.224,
      "theoretical_loss": 3.537397776107473,
      "tokens_seen": 1401225216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029064192577733203,
      "loss": 2.5957,
      "theoretical_loss": 3.537382997976991,
      "tokens_seen": 1401290752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029063189568706116,
      "loss": 2.5608,
      "theoretical_loss": 3.5373682207311523,
      "tokens_seen": 1401356288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906218655967904,
      "loss": 2.8076,
      "theoretical_loss": 3.537353444369862,
      "tokens_seen": 1401421824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906118355065195,
      "loss": 2.4087,
      "theoretical_loss": 3.537338668893027,
      "tokens_seen": 1401487360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029060180541624876,
      "loss": 2.5909,
      "theoretical_loss": 3.5373238943005516,
      "tokens_seen": 1401552896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029059177532597794,
      "loss": 2.5853,
      "theoretical_loss": 3.5373091205923424,
      "tokens_seen": 1401618432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905817452357071,
      "loss": 2.6245,
      "theoretical_loss": 3.537294347768305,
      "tokens_seen": 1401683968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905717151454363,
      "loss": 2.5226,
      "theoretical_loss": 3.5372795758283453,
      "tokens_seen": 1401749504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905616850551655,
      "loss": 2.5014,
      "theoretical_loss": 3.5372648047723683,
      "tokens_seen": 1401815040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029055165496489466,
      "loss": 2.711,
      "theoretical_loss": 3.5372500346002806,
      "tokens_seen": 1401880576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905416248746239,
      "loss": 2.7036,
      "theoretical_loss": 3.5372352653119883,
      "tokens_seen": 1401946112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905315947843531,
      "loss": 2.4211,
      "theoretical_loss": 3.5372204969073957,
      "tokens_seen": 1402011648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029052156469408226,
      "loss": 2.6379,
      "theoretical_loss": 3.5372057293864096,
      "tokens_seen": 1402077184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905115346038115,
      "loss": 2.7198,
      "theoretical_loss": 3.5371909627489364,
      "tokens_seen": 1402142720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905015045135406,
      "loss": 2.7426,
      "theoretical_loss": 3.5371761969948805,
      "tokens_seen": 1402208256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029049147442326986,
      "loss": 2.5106,
      "theoretical_loss": 3.537161432124149,
      "tokens_seen": 1402273792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290481444332999,
      "loss": 2.9995,
      "theoretical_loss": 3.5371466681366477,
      "tokens_seen": 1402339328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904714142427282,
      "loss": 2.4158,
      "theoretical_loss": 3.5371319050322816,
      "tokens_seen": 1402404864
    },
    {
      "epoch": 0.42,
      "objective/train/docs_used": 773050,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4128201007843018,
      "objective/train/theoretical_loss": 3.5371171428109576,
      "objective/train/tokens_used": 1422930400,
      "theoretical_loss": 3.5371171428109576,
      "tokens_seen": 1402470400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904613841524574,
      "loss": 2.5512,
      "theoretical_loss": 3.5371171428109576,
      "tokens_seen": 1402470400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904513540621866,
      "loss": 2.5093,
      "theoretical_loss": 3.5371023814725806,
      "tokens_seen": 1402535936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029044132397191576,
      "loss": 2.6765,
      "theoretical_loss": 3.537087621017058,
      "tokens_seen": 1402601472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029043129388164494,
      "loss": 2.6637,
      "theoretical_loss": 3.537072861444294,
      "tokens_seen": 1402667008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904212637913741,
      "loss": 2.6707,
      "theoretical_loss": 3.5370581027541963,
      "tokens_seen": 1402732544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029041123370110336,
      "loss": 2.705,
      "theoretical_loss": 3.537043344946669,
      "tokens_seen": 1402798080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904012036108325,
      "loss": 2.3418,
      "theoretical_loss": 3.5370285880216197,
      "tokens_seen": 1402863616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903911735205617,
      "loss": 2.6118,
      "theoretical_loss": 3.5370138319789537,
      "tokens_seen": 1402929152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029038114343029085,
      "loss": 2.4865,
      "theoretical_loss": 3.536999076818577,
      "tokens_seen": 1402994688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903711133400201,
      "loss": 2.3003,
      "theoretical_loss": 3.536984322540396,
      "tokens_seen": 1403060224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029036108324974926,
      "loss": 2.605,
      "theoretical_loss": 3.5369695691443166,
      "tokens_seen": 1403125760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029035105315947845,
      "loss": 2.599,
      "theoretical_loss": 3.5369548166302445,
      "tokens_seen": 1403191296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903410230692076,
      "loss": 2.5252,
      "theoretical_loss": 3.5369400649980864,
      "tokens_seen": 1403256832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029033099297893686,
      "loss": 2.8072,
      "theoretical_loss": 3.536925314247748,
      "tokens_seen": 1403322368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000290320962888666,
      "loss": 2.5232,
      "theoretical_loss": 3.536910564379135,
      "tokens_seen": 1403387904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903109327983952,
      "loss": 2.4439,
      "theoretical_loss": 3.536895815392154,
      "tokens_seen": 1403453440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029030090270812435,
      "loss": 2.7899,
      "theoretical_loss": 3.536881067286712,
      "tokens_seen": 1403518976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902908726178536,
      "loss": 2.8269,
      "theoretical_loss": 3.5368663200627135,
      "tokens_seen": 1403584512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029028084252758277,
      "loss": 2.5834,
      "theoretical_loss": 3.5368515737200656,
      "tokens_seen": 1403650048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029027081243731195,
      "loss": 2.5682,
      "theoretical_loss": 3.5368368282586746,
      "tokens_seen": 1403715584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029026078234704113,
      "loss": 2.5021,
      "theoretical_loss": 3.5368220836784463,
      "tokens_seen": 1403781120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902507522567703,
      "loss": 2.5592,
      "theoretical_loss": 3.536807339979287,
      "tokens_seen": 1403846656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902407221664995,
      "loss": 2.4743,
      "theoretical_loss": 3.536792597161103,
      "tokens_seen": 1403912192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029023069207622873,
      "loss": 2.5822,
      "theoretical_loss": 3.5367778552238005,
      "tokens_seen": 1403977728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029022066198595785,
      "loss": 2.6146,
      "theoretical_loss": 3.536763114167286,
      "tokens_seen": 1404043264
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 774221,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6937434673309326,
      "objective/train/theoretical_loss": 3.5367483739914647,
      "objective/train/tokens_used": 1424568800,
      "theoretical_loss": 3.5367483739914647,
      "tokens_seen": 1404108800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902106318956871,
      "loss": 2.6184,
      "theoretical_loss": 3.5367483739914647,
      "tokens_seen": 1404108800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902006018054162,
      "loss": 2.6297,
      "theoretical_loss": 3.536733634696245,
      "tokens_seen": 1404174336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029019057171514545,
      "loss": 2.3651,
      "theoretical_loss": 3.536718896281531,
      "tokens_seen": 1404239872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029018054162487463,
      "loss": 2.6208,
      "theoretical_loss": 3.53670415874723,
      "tokens_seen": 1404305408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901705115346038,
      "loss": 2.5992,
      "theoretical_loss": 3.5366894220932483,
      "tokens_seen": 1404370944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000290160481444333,
      "loss": 2.4126,
      "theoretical_loss": 3.536674686319492,
      "tokens_seen": 1404436480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029015045135406223,
      "loss": 2.4655,
      "theoretical_loss": 3.5366599514258676,
      "tokens_seen": 1404502016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029014042126379136,
      "loss": 2.8439,
      "theoretical_loss": 3.536645217412282,
      "tokens_seen": 1404567552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901303911735206,
      "loss": 2.672,
      "theoretical_loss": 3.5366304842786405,
      "tokens_seen": 1404633088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901203610832497,
      "loss": 2.377,
      "theoretical_loss": 3.5366157520248507,
      "tokens_seen": 1404698624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029011033099297896,
      "loss": 2.5877,
      "theoretical_loss": 3.5366010206508176,
      "tokens_seen": 1404764160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029010030090270814,
      "loss": 2.4964,
      "theoretical_loss": 3.536586290156449,
      "tokens_seen": 1404829696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900902708124373,
      "loss": 2.643,
      "theoretical_loss": 3.5365715605416503,
      "tokens_seen": 1404895232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900802407221665,
      "loss": 2.7111,
      "theoretical_loss": 3.5365568318063287,
      "tokens_seen": 1404960768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900702106318957,
      "loss": 2.6959,
      "theoretical_loss": 3.5365421039503904,
      "tokens_seen": 1405026304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029006018054162486,
      "loss": 2.6788,
      "theoretical_loss": 3.5365273769737415,
      "tokens_seen": 1405091840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900501504513541,
      "loss": 2.7087,
      "theoretical_loss": 3.536512650876289,
      "tokens_seen": 1405157376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900401203610832,
      "loss": 2.5697,
      "theoretical_loss": 3.5364979256579394,
      "tokens_seen": 1405222912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029003009027081246,
      "loss": 2.5978,
      "theoretical_loss": 3.5364832013185996,
      "tokens_seen": 1405288448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900200601805416,
      "loss": 2.6004,
      "theoretical_loss": 3.5364684778581745,
      "tokens_seen": 1405353984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900100300902708,
      "loss": 2.4113,
      "theoretical_loss": 3.5364537552765727,
      "tokens_seen": 1405419520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029,
      "loss": 2.6976,
      "theoretical_loss": 3.536439033573699,
      "tokens_seen": 1405485056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899899699097292,
      "loss": 2.7273,
      "theoretical_loss": 3.5364243127494612,
      "tokens_seen": 1405550592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028997993981945836,
      "loss": 2.7515,
      "theoretical_loss": 3.536409592803766,
      "tokens_seen": 1405616128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899699097291876,
      "loss": 2.4516,
      "theoretical_loss": 3.536394873736519,
      "tokens_seen": 1405681664
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 774861,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.848618268966675,
      "objective/train/theoretical_loss": 3.5363801555476275,
      "objective/train/tokens_used": 1426207200,
      "theoretical_loss": 3.5363801555476275,
      "tokens_seen": 1405747200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899598796389167,
      "loss": 2.5522,
      "theoretical_loss": 3.5363801555476275,
      "tokens_seen": 1405747200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028994984954864596,
      "loss": 2.5532,
      "theoretical_loss": 3.536365438236998,
      "tokens_seen": 1405812736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899398194583751,
      "loss": 2.6049,
      "theoretical_loss": 3.536350721804537,
      "tokens_seen": 1405878272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899297893681043,
      "loss": 2.6394,
      "theoretical_loss": 3.5363360062501514,
      "tokens_seen": 1405943808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899197592778335,
      "loss": 2.5589,
      "theoretical_loss": 3.536321291573748,
      "tokens_seen": 1406009344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899097291875627,
      "loss": 2.7436,
      "theoretical_loss": 3.5363065777752327,
      "tokens_seen": 1406074880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028989969909729187,
      "loss": 2.683,
      "theoretical_loss": 3.5362918648545136,
      "tokens_seen": 1406140416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028988966900702105,
      "loss": 2.6996,
      "theoretical_loss": 3.536277152811496,
      "tokens_seen": 1406205952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028987963891675023,
      "loss": 2.5788,
      "theoretical_loss": 3.5362624416460875,
      "tokens_seen": 1406271488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028986960882647946,
      "loss": 2.622,
      "theoretical_loss": 3.536247731358195,
      "tokens_seen": 1406337024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898595787362086,
      "loss": 2.535,
      "theoretical_loss": 3.536233021947724,
      "tokens_seen": 1406402560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898495486459378,
      "loss": 2.4694,
      "theoretical_loss": 3.5362183134145826,
      "tokens_seen": 1406468096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289839518555667,
      "loss": 2.6784,
      "theoretical_loss": 3.5362036057586774,
      "tokens_seen": 1406533632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898294884653962,
      "loss": 2.5906,
      "theoretical_loss": 3.5361888989799146,
      "tokens_seen": 1406599168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028981945837512537,
      "loss": 2.5658,
      "theoretical_loss": 3.536174193078202,
      "tokens_seen": 1406664704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028980942828485455,
      "loss": 2.4758,
      "theoretical_loss": 3.5361594880534453,
      "tokens_seen": 1406730240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028979939819458373,
      "loss": 2.5771,
      "theoretical_loss": 3.536144783905552,
      "tokens_seen": 1406795776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028978936810431297,
      "loss": 2.5072,
      "theoretical_loss": 3.536130080634429,
      "tokens_seen": 1406861312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028977933801404215,
      "loss": 2.5125,
      "theoretical_loss": 3.536115378239983,
      "tokens_seen": 1406926848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028976930792377133,
      "loss": 2.4741,
      "theoretical_loss": 3.5361006767221204,
      "tokens_seen": 1406992384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897592778335005,
      "loss": 2.764,
      "theoretical_loss": 3.536085976080749,
      "tokens_seen": 1407057920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897492477432297,
      "loss": 2.4863,
      "theoretical_loss": 3.5360712763157753,
      "tokens_seen": 1407123456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028973921765295893,
      "loss": 2.6076,
      "theoretical_loss": 3.536056577427107,
      "tokens_seen": 1407188992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028972918756268805,
      "loss": 2.6422,
      "theoretical_loss": 3.5360418794146495,
      "tokens_seen": 1407254528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897191574724173,
      "loss": 2.6092,
      "theoretical_loss": 3.5360271822783114,
      "tokens_seen": 1407320064
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 776313,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.495975971221924,
      "objective/train/theoretical_loss": 3.536012486017998,
      "objective/train/tokens_used": 1427845600,
      "theoretical_loss": 3.536012486017998,
      "tokens_seen": 1407385600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897091273821464,
      "loss": 2.52,
      "theoretical_loss": 3.536012486017998,
      "tokens_seen": 1407385600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028969909729187565,
      "loss": 2.4312,
      "theoretical_loss": 3.5359977906336173,
      "tokens_seen": 1407451136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028968906720160483,
      "loss": 2.7821,
      "theoretical_loss": 3.5359830961250767,
      "tokens_seen": 1407516672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289679037111334,
      "loss": 2.4085,
      "theoretical_loss": 3.5359684024922826,
      "tokens_seen": 1407582208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896690070210632,
      "loss": 2.6334,
      "theoretical_loss": 3.535953709735143,
      "tokens_seen": 1407647744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028965897693079243,
      "loss": 2.561,
      "theoretical_loss": 3.535939017853563,
      "tokens_seen": 1407713280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028964894684052156,
      "loss": 2.4042,
      "theoretical_loss": 3.535924326847451,
      "tokens_seen": 1407778816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896389167502508,
      "loss": 2.6134,
      "theoretical_loss": 3.5359096367167138,
      "tokens_seen": 1407844352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896288866599799,
      "loss": 2.4784,
      "theoretical_loss": 3.5358949474612587,
      "tokens_seen": 1407909888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028961885656970916,
      "loss": 2.5981,
      "theoretical_loss": 3.535880259080993,
      "tokens_seen": 1407975424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028960882647943834,
      "loss": 2.6256,
      "theoretical_loss": 3.5358655715758225,
      "tokens_seen": 1408040960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895987963891675,
      "loss": 2.5376,
      "theoretical_loss": 3.5358508849456562,
      "tokens_seen": 1408106496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895887662988967,
      "loss": 2.4982,
      "theoretical_loss": 3.5358361991904,
      "tokens_seen": 1408172032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895787362086259,
      "loss": 2.5318,
      "theoretical_loss": 3.535821514309962,
      "tokens_seen": 1408237568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028956870611835506,
      "loss": 2.659,
      "theoretical_loss": 3.5358068303042485,
      "tokens_seen": 1408303104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895586760280843,
      "loss": 2.5674,
      "theoretical_loss": 3.535792147173167,
      "tokens_seen": 1408368640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895486459378134,
      "loss": 2.5131,
      "theoretical_loss": 3.5357774649166247,
      "tokens_seen": 1408434176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028953861584754266,
      "loss": 2.7549,
      "theoretical_loss": 3.535762783534529,
      "tokens_seen": 1408499712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895285857572718,
      "loss": 2.7811,
      "theoretical_loss": 3.5357481030267865,
      "tokens_seen": 1408565248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289518555667001,
      "loss": 2.6589,
      "theoretical_loss": 3.5357334233933058,
      "tokens_seen": 1408630784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895085255767302,
      "loss": 2.7636,
      "theoretical_loss": 3.535718744633993,
      "tokens_seen": 1408696320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894984954864594,
      "loss": 2.5477,
      "theoretical_loss": 3.535704066748755,
      "tokens_seen": 1408761856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028948846539618856,
      "loss": 2.5133,
      "theoretical_loss": 3.5356893897375006,
      "tokens_seen": 1408827392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894784353059178,
      "loss": 2.4739,
      "theoretical_loss": 3.5356747136001356,
      "tokens_seen": 1408892928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894684052156469,
      "loss": 2.5042,
      "theoretical_loss": 3.5356600383365686,
      "tokens_seen": 1408958464
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 776966,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5918385982513428,
      "objective/train/theoretical_loss": 3.5356453639467063,
      "objective/train/tokens_used": 1429484000,
      "theoretical_loss": 3.5356453639467063,
      "tokens_seen": 1409024000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028945837512537616,
      "loss": 2.4673,
      "theoretical_loss": 3.5356453639467063,
      "tokens_seen": 1409024000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894483450351053,
      "loss": 2.7497,
      "theoretical_loss": 3.5356306904304553,
      "tokens_seen": 1409089536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894383149448345,
      "loss": 2.5532,
      "theoretical_loss": 3.5356160177877247,
      "tokens_seen": 1409155072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894282848545637,
      "loss": 2.5717,
      "theoretical_loss": 3.5356013460184204,
      "tokens_seen": 1409220608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894182547642929,
      "loss": 2.551,
      "theoretical_loss": 3.5355866751224507,
      "tokens_seen": 1409286144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028940822467402207,
      "loss": 2.685,
      "theoretical_loss": 3.5355720050997226,
      "tokens_seen": 1409351680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028939819458375125,
      "loss": 2.6845,
      "theoretical_loss": 3.535557335950143,
      "tokens_seen": 1409417216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028938816449348043,
      "loss": 2.8001,
      "theoretical_loss": 3.5355426676736204,
      "tokens_seen": 1409482752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028937813440320966,
      "loss": 2.7927,
      "theoretical_loss": 3.5355280002700615,
      "tokens_seen": 1409548288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893681043129388,
      "loss": 2.6813,
      "theoretical_loss": 3.535513333739374,
      "tokens_seen": 1409613824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289358074222668,
      "loss": 2.5749,
      "theoretical_loss": 3.5354986680814657,
      "tokens_seen": 1409679360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893480441323972,
      "loss": 2.6353,
      "theoretical_loss": 3.5354840032962436,
      "tokens_seen": 1409744896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893380140421264,
      "loss": 2.6406,
      "theoretical_loss": 3.5354693393836154,
      "tokens_seen": 1409810432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028932798395185557,
      "loss": 2.6618,
      "theoretical_loss": 3.5354546763434884,
      "tokens_seen": 1409875968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028931795386158475,
      "loss": 2.3186,
      "theoretical_loss": 3.5354400141757703,
      "tokens_seen": 1409941504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028930792377131393,
      "loss": 2.7146,
      "theoretical_loss": 3.5354253528803685,
      "tokens_seen": 1410007040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028929789368104317,
      "loss": 2.638,
      "theoretical_loss": 3.535410692457191,
      "tokens_seen": 1410072576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892878635907723,
      "loss": 2.6464,
      "theoretical_loss": 3.535396032906145,
      "tokens_seen": 1410138112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028927783350050153,
      "loss": 2.6176,
      "theoretical_loss": 3.5353813742271383,
      "tokens_seen": 1410203648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028926780341023066,
      "loss": 2.5145,
      "theoretical_loss": 3.5353667164200786,
      "tokens_seen": 1410269184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892577733199599,
      "loss": 2.8026,
      "theoretical_loss": 3.5353520594848726,
      "tokens_seen": 1410334720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892477432296891,
      "loss": 2.3777,
      "theoretical_loss": 3.5353374034214293,
      "tokens_seen": 1410400256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028923771313941825,
      "loss": 2.6062,
      "theoretical_loss": 3.535322748229656,
      "tokens_seen": 1410465792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028922768304914744,
      "loss": 2.5894,
      "theoretical_loss": 3.5353080939094594,
      "tokens_seen": 1410531328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892176529588766,
      "loss": 2.8566,
      "theoretical_loss": 3.5352934404607477,
      "tokens_seen": 1410596864
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 777918,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2443416118621826,
      "objective/train/theoretical_loss": 3.535278787883429,
      "objective/train/tokens_used": 1431122400,
      "theoretical_loss": 3.535278787883429,
      "tokens_seen": 1410662400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892076228686058,
      "loss": 2.479,
      "theoretical_loss": 3.535278787883429,
      "tokens_seen": 1410662400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028919759277833503,
      "loss": 2.5928,
      "theoretical_loss": 3.5352641361774104,
      "tokens_seen": 1410727936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028918756268806416,
      "loss": 2.4782,
      "theoretical_loss": 3.5352494853426,
      "tokens_seen": 1410793472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891775325977934,
      "loss": 2.4867,
      "theoretical_loss": 3.5352348353789065,
      "tokens_seen": 1410859008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891675025075226,
      "loss": 2.7544,
      "theoretical_loss": 3.535220186286235,
      "tokens_seen": 1410924544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028915747241725176,
      "loss": 2.5548,
      "theoretical_loss": 3.535205538064496,
      "tokens_seen": 1410990080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028914744232698094,
      "loss": 2.8223,
      "theoretical_loss": 3.5351908907135954,
      "tokens_seen": 1411055616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891374122367101,
      "loss": 2.5629,
      "theoretical_loss": 3.5351762442334422,
      "tokens_seen": 1411121152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891273821464393,
      "loss": 2.6886,
      "theoretical_loss": 3.5351615986239437,
      "tokens_seen": 1411186688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028911735205616854,
      "loss": 2.5453,
      "theoretical_loss": 3.5351469538850075,
      "tokens_seen": 1411252224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028910732196589766,
      "loss": 2.746,
      "theoretical_loss": 3.535132310016542,
      "tokens_seen": 1411317760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890972918756269,
      "loss": 2.6494,
      "theoretical_loss": 3.5351176670184543,
      "tokens_seen": 1411383296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289087261785356,
      "loss": 2.577,
      "theoretical_loss": 3.535103024890653,
      "tokens_seen": 1411448832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028907723169508526,
      "loss": 2.6034,
      "theoretical_loss": 3.5350883836330453,
      "tokens_seen": 1411514368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028906720160481444,
      "loss": 2.4017,
      "theoretical_loss": 3.53507374324554,
      "tokens_seen": 1411579904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890571715145436,
      "loss": 2.392,
      "theoretical_loss": 3.5350591037280434,
      "tokens_seen": 1411645440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890471414242728,
      "loss": 2.6675,
      "theoretical_loss": 3.5350444650804658,
      "tokens_seen": 1411710976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289037111334002,
      "loss": 2.4364,
      "theoretical_loss": 3.5350298273027128,
      "tokens_seen": 1411776512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890270812437312,
      "loss": 2.5823,
      "theoretical_loss": 3.5350151903946934,
      "tokens_seen": 1411842048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890170511534604,
      "loss": 2.5937,
      "theoretical_loss": 3.5350005543563157,
      "tokens_seen": 1411907584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890070210631896,
      "loss": 2.5991,
      "theoretical_loss": 3.534985919187487,
      "tokens_seen": 1411973120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028899699097291876,
      "loss": 2.5725,
      "theoretical_loss": 3.5349712848881163,
      "tokens_seen": 1412038656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288986960882648,
      "loss": 2.5371,
      "theoretical_loss": 3.5349566514581103,
      "tokens_seen": 1412104192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889769307923771,
      "loss": 2.5297,
      "theoretical_loss": 3.534942018897378,
      "tokens_seen": 1412169728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028896690070210636,
      "loss": 2.6743,
      "theoretical_loss": 3.534927387205827,
      "tokens_seen": 1412235264
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 778542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5615313053131104,
      "objective/train/theoretical_loss": 3.534912756383366,
      "objective/train/tokens_used": 1432760800,
      "theoretical_loss": 3.534912756383366,
      "tokens_seen": 1412300800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889568706118355,
      "loss": 2.7307,
      "theoretical_loss": 3.534912756383366,
      "tokens_seen": 1412300800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889468405215647,
      "loss": 2.5797,
      "theoretical_loss": 3.534898126429902,
      "tokens_seen": 1412366336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889368104312939,
      "loss": 2.4541,
      "theoretical_loss": 3.534883497345344,
      "tokens_seen": 1412431872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889267803410231,
      "loss": 2.7214,
      "theoretical_loss": 3.534868869129599,
      "tokens_seen": 1412497408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028891675025075227,
      "loss": 2.8409,
      "theoretical_loss": 3.5348542417825763,
      "tokens_seen": 1412562944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028890672016048145,
      "loss": 2.7457,
      "theoretical_loss": 3.5348396153041834,
      "tokens_seen": 1412628480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028889669007021063,
      "loss": 2.613,
      "theoretical_loss": 3.5348249896943287,
      "tokens_seen": 1412694016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028888665997993986,
      "loss": 2.5372,
      "theoretical_loss": 3.5348103649529197,
      "tokens_seen": 1412759552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288876629889669,
      "loss": 2.5855,
      "theoretical_loss": 3.534795741079865,
      "tokens_seen": 1412825088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028886659979939823,
      "loss": 2.4765,
      "theoretical_loss": 3.534781118075073,
      "tokens_seen": 1412890624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888565697091274,
      "loss": 2.7239,
      "theoretical_loss": 3.5347664959384515,
      "tokens_seen": 1412956160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888465396188566,
      "loss": 2.7046,
      "theoretical_loss": 3.534751874669909,
      "tokens_seen": 1413021696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028883650952858577,
      "loss": 2.4911,
      "theoretical_loss": 3.534737254269353,
      "tokens_seen": 1413087232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028882647943831495,
      "loss": 2.6196,
      "theoretical_loss": 3.534722634736693,
      "tokens_seen": 1413152768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028881644934804413,
      "loss": 2.3404,
      "theoretical_loss": 3.5347080160718356,
      "tokens_seen": 1413218304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028880641925777337,
      "loss": 2.6794,
      "theoretical_loss": 3.5346933982746904,
      "tokens_seen": 1413283840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887963891675025,
      "loss": 2.6401,
      "theoretical_loss": 3.534678781345166,
      "tokens_seen": 1413349376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028878635907723173,
      "loss": 2.5482,
      "theoretical_loss": 3.5346641652831687,
      "tokens_seen": 1413414912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028877632898696086,
      "loss": 2.6347,
      "theoretical_loss": 3.5346495500886084,
      "tokens_seen": 1413480448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887662988966901,
      "loss": 2.843,
      "theoretical_loss": 3.5346349357613924,
      "tokens_seen": 1413545984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887562688064193,
      "loss": 2.6257,
      "theoretical_loss": 3.5346203223014303,
      "tokens_seen": 1413611520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028874623871614845,
      "loss": 2.9082,
      "theoretical_loss": 3.5346057097086296,
      "tokens_seen": 1413677056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028873620862587764,
      "loss": 2.5796,
      "theoretical_loss": 3.534591097982899,
      "tokens_seen": 1413742592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887261785356068,
      "loss": 2.6771,
      "theoretical_loss": 3.534576487124146,
      "tokens_seen": 1413808128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288716148445336,
      "loss": 2.6835,
      "theoretical_loss": 3.5345618771322798,
      "tokens_seen": 1413873664
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 779834,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2309916019439697,
      "objective/train/theoretical_loss": 3.5345472680072083,
      "objective/train/tokens_used": 1434399200,
      "theoretical_loss": 3.5345472680072083,
      "tokens_seen": 1413939200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028870611835506523,
      "loss": 2.6106,
      "theoretical_loss": 3.5345472680072083,
      "tokens_seen": 1413939200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028869608826479436,
      "loss": 2.6427,
      "theoretical_loss": 3.534532659748841,
      "tokens_seen": 1414004736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886860581745236,
      "loss": 2.5728,
      "theoretical_loss": 3.5345180523570847,
      "tokens_seen": 1414070272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886760280842528,
      "loss": 2.6264,
      "theoretical_loss": 3.5345034458318487,
      "tokens_seen": 1414135808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028866599799398196,
      "loss": 2.7622,
      "theoretical_loss": 3.534488840173042,
      "tokens_seen": 1414201344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028865596790371114,
      "loss": 2.7954,
      "theoretical_loss": 3.534474235380572,
      "tokens_seen": 1414266880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886459378134403,
      "loss": 2.5129,
      "theoretical_loss": 3.5344596314543475,
      "tokens_seen": 1414332416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886359077231695,
      "loss": 2.5018,
      "theoretical_loss": 3.534445028394277,
      "tokens_seen": 1414397952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028862587763289874,
      "loss": 2.5364,
      "theoretical_loss": 3.5344304262002693,
      "tokens_seen": 1414463488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028861584754262786,
      "loss": 2.824,
      "theoretical_loss": 3.534415824872233,
      "tokens_seen": 1414529024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886058174523571,
      "loss": 2.5825,
      "theoretical_loss": 3.5344012244100758,
      "tokens_seen": 1414594560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885957873620862,
      "loss": 2.5794,
      "theoretical_loss": 3.534386624813707,
      "tokens_seen": 1414660096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028858575727181546,
      "loss": 2.6594,
      "theoretical_loss": 3.5343720260830347,
      "tokens_seen": 1414725632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028857572718154464,
      "loss": 2.6169,
      "theoretical_loss": 3.5343574282179677,
      "tokens_seen": 1414791168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885656970912738,
      "loss": 2.6218,
      "theoretical_loss": 3.534342831218415,
      "tokens_seen": 1414856704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288555667001003,
      "loss": 2.7485,
      "theoretical_loss": 3.534328235084285,
      "tokens_seen": 1414922240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885456369107322,
      "loss": 2.6608,
      "theoretical_loss": 3.534313639815485,
      "tokens_seen": 1414987776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028853560682046137,
      "loss": 2.5013,
      "theoretical_loss": 3.534299045411925,
      "tokens_seen": 1415053312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885255767301906,
      "loss": 2.6526,
      "theoretical_loss": 3.534284451873514,
      "tokens_seen": 1415118848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028851554663991973,
      "loss": 2.4369,
      "theoretical_loss": 3.534269859200159,
      "tokens_seen": 1415184384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028850551654964896,
      "loss": 2.6079,
      "theoretical_loss": 3.534255267391771,
      "tokens_seen": 1415249920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028849548645937815,
      "loss": 2.5427,
      "theoretical_loss": 3.534240676448256,
      "tokens_seen": 1415315456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884854563691073,
      "loss": 2.7063,
      "theoretical_loss": 3.5342260863695247,
      "tokens_seen": 1415380992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884754262788365,
      "loss": 2.7161,
      "theoretical_loss": 3.534211497155485,
      "tokens_seen": 1415446528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884653961885657,
      "loss": 2.706,
      "theoretical_loss": 3.534196908806046,
      "tokens_seen": 1415512064
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 780488,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1402840614318848,
      "objective/train/theoretical_loss": 3.534182321321116,
      "objective/train/tokens_used": 1436037600,
      "theoretical_loss": 3.534182321321116,
      "tokens_seen": 1415577600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028845536609829487,
      "loss": 2.5412,
      "theoretical_loss": 3.534182321321116,
      "tokens_seen": 1415577600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884453360080241,
      "loss": 2.8081,
      "theoretical_loss": 3.5341677347006035,
      "tokens_seen": 1415643136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028843530591775323,
      "loss": 2.718,
      "theoretical_loss": 3.5341531489444185,
      "tokens_seen": 1415708672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028842527582748247,
      "loss": 2.5688,
      "theoretical_loss": 3.5341385640524683,
      "tokens_seen": 1415774208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884152457372116,
      "loss": 2.6132,
      "theoretical_loss": 3.534123980024663,
      "tokens_seen": 1415839744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028840521564694083,
      "loss": 2.5662,
      "theoretical_loss": 3.5341093968609103,
      "tokens_seen": 1415905280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028839518555667,
      "loss": 2.736,
      "theoretical_loss": 3.5340948145611195,
      "tokens_seen": 1415970816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883851554663992,
      "loss": 2.5976,
      "theoretical_loss": 3.5340802331251995,
      "tokens_seen": 1416036352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883751253761284,
      "loss": 2.6624,
      "theoretical_loss": 3.534065652553059,
      "tokens_seen": 1416101888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883650952858576,
      "loss": 2.7034,
      "theoretical_loss": 3.5340510728446075,
      "tokens_seen": 1416167424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028835506519558674,
      "loss": 2.7101,
      "theoretical_loss": 3.5340364939997526,
      "tokens_seen": 1416232960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028834503510531597,
      "loss": 2.7455,
      "theoretical_loss": 3.5340219160184043,
      "tokens_seen": 1416298496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883350050150451,
      "loss": 2.7348,
      "theoretical_loss": 3.534007338900471,
      "tokens_seen": 1416364032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028832497492477433,
      "loss": 2.814,
      "theoretical_loss": 3.5339927626458616,
      "tokens_seen": 1416429568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883149448345035,
      "loss": 2.5852,
      "theoretical_loss": 3.533978187254485,
      "tokens_seen": 1416495104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883049147442327,
      "loss": 2.7235,
      "theoretical_loss": 3.53396361272625,
      "tokens_seen": 1416560640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882948846539619,
      "loss": 2.6525,
      "theoretical_loss": 3.533949039061067,
      "tokens_seen": 1416626176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028828485456369106,
      "loss": 2.761,
      "theoretical_loss": 3.5339344662588426,
      "tokens_seen": 1416691712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882748244734203,
      "loss": 2.5754,
      "theoretical_loss": 3.533919894319488,
      "tokens_seen": 1416757248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882647943831495,
      "loss": 2.5589,
      "theoretical_loss": 3.5339053232429105,
      "tokens_seen": 1416822784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028825476429287866,
      "loss": 2.4349,
      "theoretical_loss": 3.53389075302902,
      "tokens_seen": 1416888320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028824473420260784,
      "loss": 2.5853,
      "theoretical_loss": 3.533876183677725,
      "tokens_seen": 1416953856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288234704112337,
      "loss": 2.5862,
      "theoretical_loss": 3.5338616151889353,
      "tokens_seen": 1417019392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882246740220662,
      "loss": 2.6607,
      "theoretical_loss": 3.5338470475625594,
      "tokens_seen": 1417084928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028821464393179543,
      "loss": 2.6889,
      "theoretical_loss": 3.533832480798506,
      "tokens_seen": 1417150464
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 781532,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.577233076095581,
      "objective/train/theoretical_loss": 3.5338179148966855,
      "objective/train/tokens_used": 1437676000,
      "theoretical_loss": 3.5338179148966855,
      "tokens_seen": 1417216000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028820461384152456,
      "loss": 2.587,
      "theoretical_loss": 3.5338179148966855,
      "tokens_seen": 1417216000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881945837512538,
      "loss": 2.6776,
      "theoretical_loss": 3.533803349857006,
      "tokens_seen": 1417281536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288184553660983,
      "loss": 2.6713,
      "theoretical_loss": 3.533788785679376,
      "tokens_seen": 1417347072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028817452357071216,
      "loss": 2.528,
      "theoretical_loss": 3.5337742223637063,
      "tokens_seen": 1417412608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028816449348044134,
      "loss": 2.6677,
      "theoretical_loss": 3.533759659909905,
      "tokens_seen": 1417478144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881544633901705,
      "loss": 2.7362,
      "theoretical_loss": 3.5337450983178806,
      "tokens_seen": 1417543680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881444332998997,
      "loss": 2.5738,
      "theoretical_loss": 3.5337305375875436,
      "tokens_seen": 1417609216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028813440320962894,
      "loss": 2.6745,
      "theoretical_loss": 3.5337159777188027,
      "tokens_seen": 1417674752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028812437311935806,
      "loss": 2.609,
      "theoretical_loss": 3.5337014187115674,
      "tokens_seen": 1417740288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881143430290873,
      "loss": 2.7586,
      "theoretical_loss": 3.533686860565746,
      "tokens_seen": 1417805824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881043129388164,
      "loss": 2.4825,
      "theoretical_loss": 3.533672303281248,
      "tokens_seen": 1417871360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028809428284854566,
      "loss": 2.7825,
      "theoretical_loss": 3.5336577468579833,
      "tokens_seen": 1417936896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028808425275827484,
      "loss": 2.6336,
      "theoretical_loss": 3.5336431912958606,
      "tokens_seen": 1418002432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288074222668004,
      "loss": 2.5171,
      "theoretical_loss": 3.5336286365947887,
      "tokens_seen": 1418067968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880641925777332,
      "loss": 2.836,
      "theoretical_loss": 3.533614082754678,
      "tokens_seen": 1418133504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880541624874624,
      "loss": 2.7926,
      "theoretical_loss": 3.5335995297754375,
      "tokens_seen": 1418199040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028804413239719157,
      "loss": 2.5955,
      "theoretical_loss": 3.5335849776569757,
      "tokens_seen": 1418264576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880341023069208,
      "loss": 2.6247,
      "theoretical_loss": 3.5335704263992023,
      "tokens_seen": 1418330112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028802407221664993,
      "loss": 2.4715,
      "theoretical_loss": 3.533555876002027,
      "tokens_seen": 1418395648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028801404212637916,
      "loss": 2.7533,
      "theoretical_loss": 3.5335413264653592,
      "tokens_seen": 1418461184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028800401203610835,
      "loss": 2.7117,
      "theoretical_loss": 3.533526777789108,
      "tokens_seen": 1418526720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879939819458375,
      "loss": 2.84,
      "theoretical_loss": 3.5335122299731827,
      "tokens_seen": 1418592256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879839518555667,
      "loss": 2.8294,
      "theoretical_loss": 3.533497683017492,
      "tokens_seen": 1418657792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879739217652959,
      "loss": 2.2382,
      "theoretical_loss": 3.533483136921947,
      "tokens_seen": 1418723328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028796389167502507,
      "loss": 2.6418,
      "theoretical_loss": 3.5334685916864554,
      "tokens_seen": 1418788864
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 782242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.07771635055542,
      "objective/train/theoretical_loss": 3.5334540473109275,
      "objective/train/tokens_used": 1439314400,
      "theoretical_loss": 3.5334540473109275,
      "tokens_seen": 1418854400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879538615847543,
      "loss": 2.4024,
      "theoretical_loss": 3.5334540473109275,
      "tokens_seen": 1418854400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028794383149448343,
      "loss": 2.8049,
      "theoretical_loss": 3.533439503795273,
      "tokens_seen": 1418919936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028793380140421267,
      "loss": 2.4388,
      "theoretical_loss": 3.533424961139401,
      "tokens_seen": 1418985472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879237713139418,
      "loss": 2.6352,
      "theoretical_loss": 3.5334104193432205,
      "tokens_seen": 1419051008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028791374122367103,
      "loss": 2.5944,
      "theoretical_loss": 3.5333958784066413,
      "tokens_seen": 1419116544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879037111334002,
      "loss": 2.7129,
      "theoretical_loss": 3.5333813383295736,
      "tokens_seen": 1419182080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878936810431294,
      "loss": 2.4101,
      "theoretical_loss": 3.533366799111926,
      "tokens_seen": 1419247616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878836509528586,
      "loss": 2.5793,
      "theoretical_loss": 3.533352260753608,
      "tokens_seen": 1419313152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878736208625878,
      "loss": 2.7171,
      "theoretical_loss": 3.5333377232545295,
      "tokens_seen": 1419378688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028786359077231694,
      "loss": 2.8174,
      "theoretical_loss": 3.5333231866146,
      "tokens_seen": 1419444224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028785356068204617,
      "loss": 2.4148,
      "theoretical_loss": 3.5333086508337295,
      "tokens_seen": 1419509760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878435305917753,
      "loss": 2.7083,
      "theoretical_loss": 3.5332941159118265,
      "tokens_seen": 1419575296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028783350050150453,
      "loss": 2.6961,
      "theoretical_loss": 3.533279581848802,
      "tokens_seen": 1419640832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878234704112337,
      "loss": 2.4613,
      "theoretical_loss": 3.533265048644564,
      "tokens_seen": 1419706368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878134403209629,
      "loss": 2.5528,
      "theoretical_loss": 3.5332505162990238,
      "tokens_seen": 1419771904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878034102306921,
      "loss": 2.8151,
      "theoretical_loss": 3.533235984812089,
      "tokens_seen": 1419837440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028779338014042126,
      "loss": 2.6704,
      "theoretical_loss": 3.5332214541836713,
      "tokens_seen": 1419902976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028778335005015044,
      "loss": 2.6604,
      "theoretical_loss": 3.5332069244136797,
      "tokens_seen": 1419968512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877733199598797,
      "loss": 2.5915,
      "theoretical_loss": 3.533192395502023,
      "tokens_seen": 1420034048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877632898696088,
      "loss": 2.5496,
      "theoretical_loss": 3.533177867448612,
      "tokens_seen": 1420099584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028775325977933804,
      "loss": 2.559,
      "theoretical_loss": 3.5331633402533553,
      "tokens_seen": 1420165120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028774322968906716,
      "loss": 2.7361,
      "theoretical_loss": 3.533148813916164,
      "tokens_seen": 1420230656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877331995987964,
      "loss": 2.4624,
      "theoretical_loss": 3.5331342884369468,
      "tokens_seen": 1420296192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877231695085256,
      "loss": 2.5075,
      "theoretical_loss": 3.5331197638156135,
      "tokens_seen": 1420361728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028771313941825476,
      "loss": 2.5413,
      "theoretical_loss": 3.533105240052074,
      "tokens_seen": 1420427264
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 783364,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6914069652557373,
      "objective/train/theoretical_loss": 3.533090717146238,
      "objective/train/tokens_used": 1440952800,
      "theoretical_loss": 3.533090717146238,
      "tokens_seen": 1420492800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028770310932798394,
      "loss": 2.7826,
      "theoretical_loss": 3.533090717146238,
      "tokens_seen": 1420492800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876930792377132,
      "loss": 2.7784,
      "theoretical_loss": 3.5330761950980163,
      "tokens_seen": 1420558336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876830491474423,
      "loss": 2.5332,
      "theoretical_loss": 3.533061673907317,
      "tokens_seen": 1420623872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028767301905717154,
      "loss": 2.5636,
      "theoretical_loss": 3.533047153574051,
      "tokens_seen": 1420689408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028766298896690067,
      "loss": 2.6052,
      "theoretical_loss": 3.5330326340981277,
      "tokens_seen": 1420754944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876529588766299,
      "loss": 2.4816,
      "theoretical_loss": 3.5330181154794573,
      "tokens_seen": 1420820480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876429287863591,
      "loss": 2.6522,
      "theoretical_loss": 3.533003597717949,
      "tokens_seen": 1420886016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028763289869608826,
      "loss": 2.7393,
      "theoretical_loss": 3.5329890808135134,
      "tokens_seen": 1420951552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028762286860581745,
      "loss": 2.4828,
      "theoretical_loss": 3.53297456476606,
      "tokens_seen": 1421017088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876128385155466,
      "loss": 2.6096,
      "theoretical_loss": 3.5329600495754985,
      "tokens_seen": 1421082624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876028084252758,
      "loss": 2.7357,
      "theoretical_loss": 3.5329455352417396,
      "tokens_seen": 1421148160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028759277833500504,
      "loss": 2.6544,
      "theoretical_loss": 3.532931021764692,
      "tokens_seen": 1421213696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028758274824473417,
      "loss": 2.7388,
      "theoretical_loss": 3.5329165091442665,
      "tokens_seen": 1421279232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875727181544634,
      "loss": 2.6667,
      "theoretical_loss": 3.532901997380373,
      "tokens_seen": 1421344768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028756268806419253,
      "loss": 2.8288,
      "theoretical_loss": 3.532887486472921,
      "tokens_seen": 1421410304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028755265797392177,
      "loss": 2.5947,
      "theoretical_loss": 3.5328729764218214,
      "tokens_seen": 1421475840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028754262788365095,
      "loss": 2.6598,
      "theoretical_loss": 3.5328584672269834,
      "tokens_seen": 1421541376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028753259779338013,
      "loss": 2.3975,
      "theoretical_loss": 3.532843958888317,
      "tokens_seen": 1421606912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028752256770310936,
      "loss": 2.5949,
      "theoretical_loss": 3.532829451405732,
      "tokens_seen": 1421672448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028751253761283855,
      "loss": 2.779,
      "theoretical_loss": 3.532814944779139,
      "tokens_seen": 1421737984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875025075225677,
      "loss": 2.577,
      "theoretical_loss": 3.532800439008448,
      "tokens_seen": 1421803520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874924774322969,
      "loss": 2.5758,
      "theoretical_loss": 3.532785934093569,
      "tokens_seen": 1421869056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874824473420261,
      "loss": 2.6191,
      "theoretical_loss": 3.5327714300344115,
      "tokens_seen": 1421934592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028747241725175527,
      "loss": 2.5578,
      "theoretical_loss": 3.532756926830886,
      "tokens_seen": 1422000128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874623871614845,
      "loss": 2.5465,
      "theoretical_loss": 3.532742424482903,
      "tokens_seen": 1422065664
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 783958,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4881510734558105,
      "objective/train/theoretical_loss": 3.532727922990372,
      "objective/train/tokens_used": 1442591200,
      "theoretical_loss": 3.532727922990372,
      "tokens_seen": 1422131200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028745235707121363,
      "loss": 2.5889,
      "theoretical_loss": 3.532727922990372,
      "tokens_seen": 1422131200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028744232698094287,
      "loss": 2.4254,
      "theoretical_loss": 3.5327134223532033,
      "tokens_seen": 1422196736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287432296890672,
      "loss": 2.7556,
      "theoretical_loss": 3.5326989225713077,
      "tokens_seen": 1422262272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028742226680040123,
      "loss": 2.8043,
      "theoretical_loss": 3.532684423644594,
      "tokens_seen": 1422327808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874122367101304,
      "loss": 2.5499,
      "theoretical_loss": 3.532669925572973,
      "tokens_seen": 1422393344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874022066198596,
      "loss": 2.7351,
      "theoretical_loss": 3.532655428356356,
      "tokens_seen": 1422458880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873921765295888,
      "loss": 2.5448,
      "theoretical_loss": 3.532640931994651,
      "tokens_seen": 1422524416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287382146439318,
      "loss": 2.6517,
      "theoretical_loss": 3.53262643648777,
      "tokens_seen": 1422589952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028737211634904714,
      "loss": 2.7519,
      "theoretical_loss": 3.5326119418356225,
      "tokens_seen": 1422655488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028736208625877637,
      "loss": 2.6739,
      "theoretical_loss": 3.5325974480381186,
      "tokens_seen": 1422721024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873520561685055,
      "loss": 2.8156,
      "theoretical_loss": 3.5325829550951693,
      "tokens_seen": 1422786560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028734202607823473,
      "loss": 2.4075,
      "theoretical_loss": 3.532568463006684,
      "tokens_seen": 1422852096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873319959879639,
      "loss": 2.6581,
      "theoretical_loss": 3.532553971772573,
      "tokens_seen": 1422917632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873219658976931,
      "loss": 2.754,
      "theoretical_loss": 3.532539481392747,
      "tokens_seen": 1422983168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873119358074223,
      "loss": 2.7735,
      "theoretical_loss": 3.532524991867116,
      "tokens_seen": 1423048704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028730190571715146,
      "loss": 2.849,
      "theoretical_loss": 3.5325105031955912,
      "tokens_seen": 1423114240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028729187562688064,
      "loss": 2.5006,
      "theoretical_loss": 3.532496015378081,
      "tokens_seen": 1423179776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872818455366099,
      "loss": 2.671,
      "theoretical_loss": 3.532481528414498,
      "tokens_seen": 1423245312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287271815446339,
      "loss": 2.4799,
      "theoretical_loss": 3.5324670423047513,
      "tokens_seen": 1423310848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028726178535606824,
      "loss": 2.6465,
      "theoretical_loss": 3.5324525570487513,
      "tokens_seen": 1423376384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028725175526579736,
      "loss": 2.4529,
      "theoretical_loss": 3.532438072646408,
      "tokens_seen": 1423441920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872417251755266,
      "loss": 2.6795,
      "theoretical_loss": 3.532423589097633,
      "tokens_seen": 1423507456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872316950852558,
      "loss": 2.7439,
      "theoretical_loss": 3.5324091064023357,
      "tokens_seen": 1423572992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028722166499498496,
      "loss": 2.4454,
      "theoretical_loss": 3.532394624560427,
      "tokens_seen": 1423638528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028721163490471414,
      "loss": 2.7103,
      "theoretical_loss": 3.5323801435718174,
      "tokens_seen": 1423704064
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 785018,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1949687004089355,
      "objective/train/theoretical_loss": 3.5323656634364164,
      "objective/train/tokens_used": 1444229600,
      "theoretical_loss": 3.5323656634364164,
      "tokens_seen": 1423769600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872016048144434,
      "loss": 2.8826,
      "theoretical_loss": 3.5323656634364164,
      "tokens_seen": 1423769600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871915747241725,
      "loss": 2.6976,
      "theoretical_loss": 3.5323511841541357,
      "tokens_seen": 1423835136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028718154463390174,
      "loss": 2.6035,
      "theoretical_loss": 3.5323367057248856,
      "tokens_seen": 1423900672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028717151454363087,
      "loss": 2.6493,
      "theoretical_loss": 3.5323222281485753,
      "tokens_seen": 1423966208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871614844533601,
      "loss": 2.6472,
      "theoretical_loss": 3.532307751425117,
      "tokens_seen": 1424031744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871514543630893,
      "loss": 2.7908,
      "theoretical_loss": 3.5322932755544203,
      "tokens_seen": 1424097280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028714142427281846,
      "loss": 2.5785,
      "theoretical_loss": 3.532278800536396,
      "tokens_seen": 1424162816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028713139418254765,
      "loss": 2.5572,
      "theoretical_loss": 3.5322643263709543,
      "tokens_seen": 1424228352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871213640922768,
      "loss": 2.5964,
      "theoretical_loss": 3.532249853058006,
      "tokens_seen": 1424293888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287111334002006,
      "loss": 2.5513,
      "theoretical_loss": 3.532235380597462,
      "tokens_seen": 1424359424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028710130391173524,
      "loss": 2.6485,
      "theoretical_loss": 3.532220908989232,
      "tokens_seen": 1424424960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028709127382146437,
      "loss": 2.9143,
      "theoretical_loss": 3.5322064382332274,
      "tokens_seen": 1424490496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870812437311936,
      "loss": 2.5105,
      "theoretical_loss": 3.5321919683293586,
      "tokens_seen": 1424556032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028707121364092273,
      "loss": 2.6386,
      "theoretical_loss": 3.532177499277536,
      "tokens_seen": 1424621568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028706118355065197,
      "loss": 2.6675,
      "theoretical_loss": 3.5321630310776704,
      "tokens_seen": 1424687104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028705115346038115,
      "loss": 2.587,
      "theoretical_loss": 3.5321485637296726,
      "tokens_seen": 1424752640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028704112337011033,
      "loss": 2.6668,
      "theoretical_loss": 3.532134097233453,
      "tokens_seen": 1424818176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870310932798395,
      "loss": 2.6154,
      "theoretical_loss": 3.5321196315889223,
      "tokens_seen": 1424883712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028702106318956875,
      "loss": 2.6835,
      "theoretical_loss": 3.5321051667959913,
      "tokens_seen": 1424949248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870110330992979,
      "loss": 2.6394,
      "theoretical_loss": 3.5320907028545707,
      "tokens_seen": 1425014784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870010030090271,
      "loss": 2.4343,
      "theoretical_loss": 3.532076239764571,
      "tokens_seen": 1425080320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028699097291875624,
      "loss": 2.7033,
      "theoretical_loss": 3.5320617775259033,
      "tokens_seen": 1425145856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028698094282848547,
      "loss": 2.6942,
      "theoretical_loss": 3.5320473161384784,
      "tokens_seen": 1425211392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028697091273821465,
      "loss": 2.7405,
      "theoretical_loss": 3.5320328556022065,
      "tokens_seen": 1425276928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028696088264794383,
      "loss": 2.5886,
      "theoretical_loss": 3.5320183959169986,
      "tokens_seen": 1425342464
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 786235,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.591965913772583,
      "objective/train/theoretical_loss": 3.532003937082766,
      "objective/train/tokens_used": 1445868000,
      "theoretical_loss": 3.532003937082766,
      "tokens_seen": 1425408000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286950852557673,
      "loss": 2.6586,
      "theoretical_loss": 3.532003937082766,
      "tokens_seen": 1425408000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869408224674022,
      "loss": 2.69,
      "theoretical_loss": 3.531989479099419,
      "tokens_seen": 1425473536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869307923771314,
      "loss": 2.6215,
      "theoretical_loss": 3.531975021966868,
      "tokens_seen": 1425539072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869207622868606,
      "loss": 2.7199,
      "theoretical_loss": 3.531960565685025,
      "tokens_seen": 1425604608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028691073219658974,
      "loss": 2.5658,
      "theoretical_loss": 3.5319461102537995,
      "tokens_seen": 1425670144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286900702106319,
      "loss": 2.5154,
      "theoretical_loss": 3.531931655673103,
      "tokens_seen": 1425735680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868906720160481,
      "loss": 2.6294,
      "theoretical_loss": 3.5319172019428464,
      "tokens_seen": 1425801216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028688064192577734,
      "loss": 2.7823,
      "theoretical_loss": 3.531902749062941,
      "tokens_seen": 1425866752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868706118355065,
      "loss": 2.538,
      "theoretical_loss": 3.531888297033297,
      "tokens_seen": 1425932288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868605817452357,
      "loss": 2.5781,
      "theoretical_loss": 3.5318738458538252,
      "tokens_seen": 1425997824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868505516549649,
      "loss": 2.664,
      "theoretical_loss": 3.531859395524437,
      "tokens_seen": 1426063360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868405215646941,
      "loss": 2.8786,
      "theoretical_loss": 3.5318449460450436,
      "tokens_seen": 1426128896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028683049147442324,
      "loss": 2.7226,
      "theoretical_loss": 3.531830497415555,
      "tokens_seen": 1426194432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868204613841525,
      "loss": 2.41,
      "theoretical_loss": 3.531816049635883,
      "tokens_seen": 1426259968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868104312938816,
      "loss": 2.4417,
      "theoretical_loss": 3.5318016027059382,
      "tokens_seen": 1426325504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028680040120361084,
      "loss": 2.62,
      "theoretical_loss": 3.531787156625631,
      "tokens_seen": 1426391040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028679037111334,
      "loss": 2.7438,
      "theoretical_loss": 3.531772711394874,
      "tokens_seen": 1426456576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867803410230692,
      "loss": 2.4956,
      "theoretical_loss": 3.531758267013577,
      "tokens_seen": 1426522112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028677031093279844,
      "loss": 2.4738,
      "theoretical_loss": 3.531743823481651,
      "tokens_seen": 1426587648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028676028084252756,
      "loss": 2.7312,
      "theoretical_loss": 3.5317293807990073,
      "tokens_seen": 1426653184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867502507522568,
      "loss": 2.4711,
      "theoretical_loss": 3.5317149389655573,
      "tokens_seen": 1426718720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286740220661986,
      "loss": 2.7506,
      "theoretical_loss": 3.5317004979812117,
      "tokens_seen": 1426784256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028673019057171516,
      "loss": 2.5439,
      "theoretical_loss": 3.531686057845881,
      "tokens_seen": 1426849792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028672016048144434,
      "loss": 2.5135,
      "theoretical_loss": 3.5316716185594776,
      "tokens_seen": 1426915328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867101303911736,
      "loss": 2.6132,
      "theoretical_loss": 3.531657180121911,
      "tokens_seen": 1426980864
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 786774,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.598083257675171,
      "objective/train/theoretical_loss": 3.5316427425330943,
      "objective/train/tokens_used": 1447506400,
      "theoretical_loss": 3.5316427425330943,
      "tokens_seen": 1427046400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867001003009027,
      "loss": 2.6186,
      "theoretical_loss": 3.5316427425330943,
      "tokens_seen": 1427046400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028669007021063194,
      "loss": 2.9038,
      "theoretical_loss": 3.531628305792937,
      "tokens_seen": 1427111936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028668004012036107,
      "loss": 2.6803,
      "theoretical_loss": 3.531613869901351,
      "tokens_seen": 1427177472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866700100300903,
      "loss": 2.6351,
      "theoretical_loss": 3.531599434858247,
      "tokens_seen": 1427243008
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866599799398195,
      "loss": 2.5262,
      "theoretical_loss": 3.5315850006635365,
      "tokens_seen": 1427308544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028664994984954866,
      "loss": 2.552,
      "theoretical_loss": 3.531570567317131,
      "tokens_seen": 1427374080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028663991975927785,
      "loss": 2.6389,
      "theoretical_loss": 3.531556134818941,
      "tokens_seen": 1427439616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286629889669007,
      "loss": 2.7612,
      "theoretical_loss": 3.531541703168878,
      "tokens_seen": 1427505152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866198595787362,
      "loss": 2.5592,
      "theoretical_loss": 3.531527272366853,
      "tokens_seen": 1427570688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028660982948846544,
      "loss": 2.4948,
      "theoretical_loss": 3.531512842412778,
      "tokens_seen": 1427636224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028659979939819457,
      "loss": 2.6488,
      "theoretical_loss": 3.531498413306564,
      "tokens_seen": 1427701760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865897693079238,
      "loss": 2.7581,
      "theoretical_loss": 3.5314839850481214,
      "tokens_seen": 1427767296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028657973921765293,
      "loss": 2.6461,
      "theoretical_loss": 3.5314695576373625,
      "tokens_seen": 1427832832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028656970912738217,
      "loss": 2.6874,
      "theoretical_loss": 3.531455131074198,
      "tokens_seen": 1427898368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028655967903711135,
      "loss": 2.625,
      "theoretical_loss": 3.531440705358539,
      "tokens_seen": 1427963904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028654964894684053,
      "loss": 2.5656,
      "theoretical_loss": 3.5314262804902974,
      "tokens_seen": 1428029440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865396188565697,
      "loss": 2.5828,
      "theoretical_loss": 3.5314118564693846,
      "tokens_seen": 1428094976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028652958876629895,
      "loss": 2.3775,
      "theoretical_loss": 3.531397433295712,
      "tokens_seen": 1428160512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865195586760281,
      "loss": 2.9139,
      "theoretical_loss": 3.53138301096919,
      "tokens_seen": 1428226048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865095285857573,
      "loss": 2.5729,
      "theoretical_loss": 3.5313685894897313,
      "tokens_seen": 1428291584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028649949849548644,
      "loss": 2.593,
      "theoretical_loss": 3.531354168857246,
      "tokens_seen": 1428357120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028648946840521567,
      "loss": 2.5411,
      "theoretical_loss": 3.5313397490716465,
      "tokens_seen": 1428422656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028647943831494485,
      "loss": 2.6507,
      "theoretical_loss": 3.5313253301328436,
      "tokens_seen": 1428488192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028646940822467403,
      "loss": 2.4573,
      "theoretical_loss": 3.531310912040749,
      "tokens_seen": 1428553728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864593781344032,
      "loss": 2.5802,
      "theoretical_loss": 3.5312964947952743,
      "tokens_seen": 1428619264
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 787815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.069486379623413,
      "objective/train/theoretical_loss": 3.5312820783963303,
      "objective/train/tokens_used": 1449144800,
      "theoretical_loss": 3.5312820783963303,
      "tokens_seen": 1428684800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864493480441324,
      "loss": 2.5377,
      "theoretical_loss": 3.5312820783963303,
      "tokens_seen": 1428684800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864393179538616,
      "loss": 2.7072,
      "theoretical_loss": 3.5312676628438293,
      "tokens_seen": 1428750336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864292878635908,
      "loss": 2.567,
      "theoretical_loss": 3.531253248137682,
      "tokens_seen": 1428815872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028641925777331994,
      "loss": 2.8516,
      "theoretical_loss": 3.5312388342778007,
      "tokens_seen": 1428881408
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864092276830492,
      "loss": 2.6305,
      "theoretical_loss": 3.5312244212640964,
      "tokens_seen": 1428946944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863991975927783,
      "loss": 2.917,
      "theoretical_loss": 3.5312100090964806,
      "tokens_seen": 1429012480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028638916750250754,
      "loss": 2.5045,
      "theoretical_loss": 3.531195597774865,
      "tokens_seen": 1429078016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863791374122367,
      "loss": 2.7884,
      "theoretical_loss": 3.531181187299161,
      "tokens_seen": 1429143552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863691073219659,
      "loss": 2.744,
      "theoretical_loss": 3.5311667776692803,
      "tokens_seen": 1429209088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863590772316951,
      "loss": 2.3637,
      "theoretical_loss": 3.5311523688851345,
      "tokens_seen": 1429274624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863490471414243,
      "loss": 2.8131,
      "theoretical_loss": 3.5311379609466353,
      "tokens_seen": 1429340160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028633901705115344,
      "loss": 2.5255,
      "theoretical_loss": 3.5311235538536936,
      "tokens_seen": 1429405696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863289869608827,
      "loss": 2.5705,
      "theoretical_loss": 3.5311091476062217,
      "tokens_seen": 1429471232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863189568706118,
      "loss": 2.6134,
      "theoretical_loss": 3.5310947422041314,
      "tokens_seen": 1429536768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028630892678034104,
      "loss": 2.4821,
      "theoretical_loss": 3.5310803376473334,
      "tokens_seen": 1429602304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862988966900702,
      "loss": 2.5878,
      "theoretical_loss": 3.53106593393574,
      "tokens_seen": 1429667840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862888665997994,
      "loss": 2.5424,
      "theoretical_loss": 3.531051531069263,
      "tokens_seen": 1429733376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862788365095286,
      "loss": 2.6448,
      "theoretical_loss": 3.5310371290478146,
      "tokens_seen": 1429798912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028626880641925776,
      "loss": 2.6663,
      "theoretical_loss": 3.5310227278713047,
      "tokens_seen": 1429864448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028625877632898694,
      "loss": 2.5672,
      "theoretical_loss": 3.5310083275396464,
      "tokens_seen": 1429929984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862487462387162,
      "loss": 2.7178,
      "theoretical_loss": 3.5309939280527516,
      "tokens_seen": 1429995520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862387161484453,
      "loss": 2.6479,
      "theoretical_loss": 3.530979529410531,
      "tokens_seen": 1430061056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028622868605817454,
      "loss": 2.7163,
      "theoretical_loss": 3.5309651316128967,
      "tokens_seen": 1430126592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028621865596790367,
      "loss": 2.6371,
      "theoretical_loss": 3.5309507346597613,
      "tokens_seen": 1430192128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862086258776329,
      "loss": 2.7673,
      "theoretical_loss": 3.5309363385510353,
      "tokens_seen": 1430257664
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 788429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.007455587387085,
      "objective/train/theoretical_loss": 3.5309219432866312,
      "objective/train/tokens_used": 1450783200,
      "theoretical_loss": 3.5309219432866312,
      "tokens_seen": 1430323200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861985957873621,
      "loss": 2.8257,
      "theoretical_loss": 3.5309219432866312,
      "tokens_seen": 1430323200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028618856569709127,
      "loss": 2.56,
      "theoretical_loss": 3.5309075488664607,
      "tokens_seen": 1430388736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028617853560682045,
      "loss": 2.7202,
      "theoretical_loss": 3.530893155290436,
      "tokens_seen": 1430454272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861685055165497,
      "loss": 2.4225,
      "theoretical_loss": 3.5308787625584683,
      "tokens_seen": 1430519808
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861584754262788,
      "loss": 2.4291,
      "theoretical_loss": 3.5308643706704697,
      "tokens_seen": 1430585344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028614844533600805,
      "loss": 2.6722,
      "theoretical_loss": 3.5308499796263515,
      "tokens_seen": 1430650880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028613841524573717,
      "loss": 2.7396,
      "theoretical_loss": 3.530835589426027,
      "tokens_seen": 1430716416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861283851554664,
      "loss": 2.7058,
      "theoretical_loss": 3.530821200069407,
      "tokens_seen": 1430781952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861183550651956,
      "loss": 2.6886,
      "theoretical_loss": 3.5308068115564026,
      "tokens_seen": 1430847488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028610832497492477,
      "loss": 2.5916,
      "theoretical_loss": 3.5307924238869273,
      "tokens_seen": 1430913024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028609829488465395,
      "loss": 2.5547,
      "theoretical_loss": 3.5307780370608923,
      "tokens_seen": 1430978560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028608826479438313,
      "loss": 2.4528,
      "theoretical_loss": 3.53076365107821,
      "tokens_seen": 1431044096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860782347041123,
      "loss": 2.6304,
      "theoretical_loss": 3.530749265938791,
      "tokens_seen": 1431109632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028606820461384155,
      "loss": 2.5961,
      "theoretical_loss": 3.530734881642549,
      "tokens_seen": 1431175168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860581745235707,
      "loss": 2.6846,
      "theoretical_loss": 3.530720498189395,
      "tokens_seen": 1431240704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860481444332999,
      "loss": 2.6643,
      "theoretical_loss": 3.5307061155792416,
      "tokens_seen": 1431306240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860381143430291,
      "loss": 2.6204,
      "theoretical_loss": 3.5306917338119996,
      "tokens_seen": 1431371776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860280842527583,
      "loss": 2.6394,
      "theoretical_loss": 3.530677352887582,
      "tokens_seen": 1431437312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860180541624875,
      "loss": 2.6282,
      "theoretical_loss": 3.530662972805901,
      "tokens_seen": 1431502848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028600802407221664,
      "loss": 2.8499,
      "theoretical_loss": 3.530648593566868,
      "tokens_seen": 1431568384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028599799398194587,
      "loss": 2.5693,
      "theoretical_loss": 3.5306342151703953,
      "tokens_seen": 1431633920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028598796389167505,
      "loss": 2.7649,
      "theoretical_loss": 3.5306198376163946,
      "tokens_seen": 1431699456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028597793380140423,
      "loss": 2.7873,
      "theoretical_loss": 3.5306054609047783,
      "tokens_seen": 1431764992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859679037111334,
      "loss": 2.676,
      "theoretical_loss": 3.530591085035459,
      "tokens_seen": 1431830528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859578736208626,
      "loss": 2.7262,
      "theoretical_loss": 3.530576710008348,
      "tokens_seen": 1431896064
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 789812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3145503997802734,
      "objective/train/theoretical_loss": 3.5305623358233578,
      "objective/train/tokens_used": 1452421600,
      "theoretical_loss": 3.5305623358233578,
      "tokens_seen": 1431961600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859478435305918,
      "loss": 2.6037,
      "theoretical_loss": 3.5305623358233578,
      "tokens_seen": 1431961600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285937813440321,
      "loss": 2.6274,
      "theoretical_loss": 3.5305479624804006,
      "tokens_seen": 1432027136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028592778335005014,
      "loss": 2.543,
      "theoretical_loss": 3.530533589979388,
      "tokens_seen": 1432092672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859177532597794,
      "loss": 2.5528,
      "theoretical_loss": 3.530519218320233,
      "tokens_seen": 1432158208
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859077231695085,
      "loss": 2.5332,
      "theoretical_loss": 3.530504847502847,
      "tokens_seen": 1432223744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028589769307923774,
      "loss": 2.7996,
      "theoretical_loss": 3.5304904775271426,
      "tokens_seen": 1432289280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858876629889669,
      "loss": 2.495,
      "theoretical_loss": 3.530476108393032,
      "tokens_seen": 1432354816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858776328986961,
      "loss": 2.5663,
      "theoretical_loss": 3.530461740100427,
      "tokens_seen": 1432420352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858676028084253,
      "loss": 2.899,
      "theoretical_loss": 3.5304473726492405,
      "tokens_seen": 1432485888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858575727181545,
      "loss": 2.7768,
      "theoretical_loss": 3.5304330060393845,
      "tokens_seen": 1432551424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028584754262788364,
      "loss": 2.6372,
      "theoretical_loss": 3.5304186402707707,
      "tokens_seen": 1432616960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858375125376129,
      "loss": 2.7577,
      "theoretical_loss": 3.5304042753433116,
      "tokens_seen": 1432682496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285827482447342,
      "loss": 2.6304,
      "theoretical_loss": 3.53038991125692,
      "tokens_seen": 1432748032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028581745235707124,
      "loss": 2.7153,
      "theoretical_loss": 3.530375548011508,
      "tokens_seen": 1432813568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858074222668004,
      "loss": 2.695,
      "theoretical_loss": 3.5303611856069876,
      "tokens_seen": 1432879104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857973921765296,
      "loss": 2.3783,
      "theoretical_loss": 3.5303468240432707,
      "tokens_seen": 1432944640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857873620862588,
      "loss": 2.8994,
      "theoretical_loss": 3.5303324633202706,
      "tokens_seen": 1433010176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028577733199598796,
      "loss": 2.5932,
      "theoretical_loss": 3.5303181034378994,
      "tokens_seen": 1433075712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028576730190571714,
      "loss": 2.5311,
      "theoretical_loss": 3.530303744396069,
      "tokens_seen": 1433141248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857572718154464,
      "loss": 2.5577,
      "theoretical_loss": 3.5302893861946925,
      "tokens_seen": 1433206784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857472417251755,
      "loss": 2.4627,
      "theoretical_loss": 3.5302750288336817,
      "tokens_seen": 1433272320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028573721163490474,
      "loss": 2.7041,
      "theoretical_loss": 3.5302606723129486,
      "tokens_seen": 1433337856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028572718154463387,
      "loss": 2.6183,
      "theoretical_loss": 3.5302463166324065,
      "tokens_seen": 1433403392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857171514543631,
      "loss": 2.8183,
      "theoretical_loss": 3.5302319617919675,
      "tokens_seen": 1433468928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857071213640923,
      "loss": 2.701,
      "theoretical_loss": 3.5302176077915437,
      "tokens_seen": 1433534464
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 791077,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5977752208709717,
      "objective/train/theoretical_loss": 3.530203254631048,
      "objective/train/tokens_used": 1454060000,
      "theoretical_loss": 3.530203254631048,
      "tokens_seen": 1433600000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028569709127382147,
      "loss": 2.5736,
      "theoretical_loss": 3.530203254631048,
      "tokens_seen": 1433600000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028568706118355065,
      "loss": 2.733,
      "theoretical_loss": 3.5301889023103925,
      "tokens_seen": 1433665536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856770310932799,
      "loss": 2.4584,
      "theoretical_loss": 3.5301745508294893,
      "tokens_seen": 1433731072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285667001003009,
      "loss": 2.4638,
      "theoretical_loss": 3.530160200188252,
      "tokens_seen": 1433796608
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028565697091273825,
      "loss": 2.8465,
      "theoretical_loss": 3.530145850386593,
      "tokens_seen": 1433862144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028564694082246737,
      "loss": 2.5399,
      "theoretical_loss": 3.5301315014244237,
      "tokens_seen": 1433927680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856369107321966,
      "loss": 2.4459,
      "theoretical_loss": 3.5301171533016573,
      "tokens_seen": 1433993216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856268806419258,
      "loss": 2.2933,
      "theoretical_loss": 3.5301028060182063,
      "tokens_seen": 1434058752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028561685055165497,
      "loss": 2.5866,
      "theoretical_loss": 3.5300884595739834,
      "tokens_seen": 1434124288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028560682046138415,
      "loss": 2.7661,
      "theoretical_loss": 3.530074113968901,
      "tokens_seen": 1434189824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028559679037111333,
      "loss": 2.8626,
      "theoretical_loss": 3.5300597692028717,
      "tokens_seen": 1434255360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855867602808425,
      "loss": 2.5412,
      "theoretical_loss": 3.530045425275808,
      "tokens_seen": 1434320896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028557673019057175,
      "loss": 2.579,
      "theoretical_loss": 3.530031082187622,
      "tokens_seen": 1434386432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855667001003009,
      "loss": 2.5718,
      "theoretical_loss": 3.5300167399382274,
      "tokens_seen": 1434451968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855566700100301,
      "loss": 2.7666,
      "theoretical_loss": 3.5300023985275364,
      "tokens_seen": 1434517504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855466399197593,
      "loss": 2.5841,
      "theoretical_loss": 3.5299880579554612,
      "tokens_seen": 1434583040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855366098294885,
      "loss": 2.5258,
      "theoretical_loss": 3.5299737182219153,
      "tokens_seen": 1434648576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028552657973921765,
      "loss": 2.6316,
      "theoretical_loss": 3.5299593793268107,
      "tokens_seen": 1434714112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028551654964894684,
      "loss": 2.5786,
      "theoretical_loss": 3.52994504127006,
      "tokens_seen": 1434779648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285506519558676,
      "loss": 2.968,
      "theoretical_loss": 3.5299307040515764,
      "tokens_seen": 1434845184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028549648946840525,
      "loss": 2.5524,
      "theoretical_loss": 3.529916367671272,
      "tokens_seen": 1434910720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854864593781344,
      "loss": 2.6716,
      "theoretical_loss": 3.5299020321290606,
      "tokens_seen": 1434976256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854764292878636,
      "loss": 2.7411,
      "theoretical_loss": 3.529887697424854,
      "tokens_seen": 1435041792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028546639919759274,
      "loss": 2.8467,
      "theoretical_loss": 3.529873363558565,
      "tokens_seen": 1435107328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285456369107322,
      "loss": 2.5434,
      "theoretical_loss": 3.529859030530106,
      "tokens_seen": 1435172864
    },
    {
      "epoch": 0.43,
      "objective/train/docs_used": 791531,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.479179620742798,
      "objective/train/theoretical_loss": 3.529844698339391,
      "objective/train/tokens_used": 1455698400,
      "theoretical_loss": 3.529844698339391,
      "tokens_seen": 1435238400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028544633901705116,
      "loss": 2.6274,
      "theoretical_loss": 3.529844698339391,
      "tokens_seen": 1435238400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028543630892678034,
      "loss": 2.9353,
      "theoretical_loss": 3.5298303669863316,
      "tokens_seen": 1435303936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854262788365095,
      "loss": 2.455,
      "theoretical_loss": 3.5298160364708413,
      "tokens_seen": 1435369472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854162487462387,
      "loss": 2.7144,
      "theoretical_loss": 3.5298017067928327,
      "tokens_seen": 1435435008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002854062186559679,
      "loss": 2.6812,
      "theoretical_loss": 3.5297873779522186,
      "tokens_seen": 1435500544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853961885656971,
      "loss": 2.6906,
      "theoretical_loss": 3.529773049948912,
      "tokens_seen": 1435566080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028538615847542624,
      "loss": 2.6389,
      "theoretical_loss": 3.5297587227828253,
      "tokens_seen": 1435631616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853761283851555,
      "loss": 2.5088,
      "theoretical_loss": 3.529744396453872,
      "tokens_seen": 1435697152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028536609829488466,
      "loss": 2.6032,
      "theoretical_loss": 3.529730070961964,
      "tokens_seen": 1435762688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028535606820461384,
      "loss": 2.5452,
      "theoretical_loss": 3.5297157463070157,
      "tokens_seen": 1435828224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285346038114343,
      "loss": 2.6159,
      "theoretical_loss": 3.5297014224889383,
      "tokens_seen": 1435893760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853360080240722,
      "loss": 2.5628,
      "theoretical_loss": 3.529687099507646,
      "tokens_seen": 1435959296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853259779338014,
      "loss": 2.4744,
      "theoretical_loss": 3.5296727773630514,
      "tokens_seen": 1436024832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853159478435306,
      "loss": 2.7097,
      "theoretical_loss": 3.529658456055067,
      "tokens_seen": 1436090368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028530591775325975,
      "loss": 2.7462,
      "theoretical_loss": 3.5296441355836063,
      "tokens_seen": 1436155904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285295887662989,
      "loss": 2.6551,
      "theoretical_loss": 3.529629815948582,
      "tokens_seen": 1436221440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852858575727181,
      "loss": 2.8108,
      "theoretical_loss": 3.5296154971499067,
      "tokens_seen": 1436286976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028527582748244735,
      "loss": 2.8623,
      "theoretical_loss": 3.5296011791874946,
      "tokens_seen": 1436352512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852657973921766,
      "loss": 2.7176,
      "theoretical_loss": 3.5295868620612576,
      "tokens_seen": 1436418048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852557673019057,
      "loss": 2.6384,
      "theoretical_loss": 3.5295725457711087,
      "tokens_seen": 1436483584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028524573721163494,
      "loss": 2.6418,
      "theoretical_loss": 3.5295582303169617,
      "tokens_seen": 1436549120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028523570712136407,
      "loss": 2.536,
      "theoretical_loss": 3.5295439156987287,
      "tokens_seen": 1436614656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852256770310933,
      "loss": 2.5104,
      "theoretical_loss": 3.529529601916324,
      "tokens_seen": 1436680192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852156469408225,
      "loss": 2.7834,
      "theoretical_loss": 3.5295152889696597,
      "tokens_seen": 1436745728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028520561685055167,
      "loss": 2.7306,
      "theoretical_loss": 3.5295009768586487,
      "tokens_seen": 1436811264
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 792701,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7126412391662598,
      "objective/train/theoretical_loss": 3.529486665583205,
      "objective/train/tokens_used": 1457336800,
      "theoretical_loss": 3.529486665583205,
      "tokens_seen": 1436876800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028519558676028085,
      "loss": 2.6347,
      "theoretical_loss": 3.529486665583205,
      "tokens_seen": 1436876800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851855566700101,
      "loss": 2.7781,
      "theoretical_loss": 3.529472355143241,
      "tokens_seen": 1436942336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851755265797392,
      "loss": 2.7436,
      "theoretical_loss": 3.52945804553867,
      "tokens_seen": 1437007872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028516549648946845,
      "loss": 2.6782,
      "theoretical_loss": 3.5294437367694056,
      "tokens_seen": 1437073408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028515546639919757,
      "loss": 2.6728,
      "theoretical_loss": 3.5294294288353605,
      "tokens_seen": 1437138944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851454363089268,
      "loss": 2.7392,
      "theoretical_loss": 3.529415121736448,
      "tokens_seen": 1437204480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285135406218656,
      "loss": 2.5219,
      "theoretical_loss": 3.529400815472581,
      "tokens_seen": 1437270016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028512537612838517,
      "loss": 2.7083,
      "theoretical_loss": 3.5293865100436728,
      "tokens_seen": 1437335552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028511534603811435,
      "loss": 2.7232,
      "theoretical_loss": 3.5293722054496364,
      "tokens_seen": 1437401088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028510531594784353,
      "loss": 2.8735,
      "theoretical_loss": 3.5293579016903855,
      "tokens_seen": 1437466624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850952858575727,
      "loss": 2.6501,
      "theoretical_loss": 3.5293435987658333,
      "tokens_seen": 1437532160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028508525576730195,
      "loss": 2.5175,
      "theoretical_loss": 3.529329296675893,
      "tokens_seen": 1437597696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850752256770311,
      "loss": 2.8487,
      "theoretical_loss": 3.5293149954204774,
      "tokens_seen": 1437663232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850651955867603,
      "loss": 2.4862,
      "theoretical_loss": 3.5293006949995007,
      "tokens_seen": 1437728768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850551654964895,
      "loss": 2.5771,
      "theoretical_loss": 3.529286395412875,
      "tokens_seen": 1437794304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850451354062187,
      "loss": 2.6454,
      "theoretical_loss": 3.5292720966605144,
      "tokens_seen": 1437859840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028503510531594785,
      "loss": 2.5789,
      "theoretical_loss": 3.5292577987423317,
      "tokens_seen": 1437925376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028502507522567704,
      "loss": 2.6764,
      "theoretical_loss": 3.529243501658241,
      "tokens_seen": 1437990912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850150451354062,
      "loss": 2.618,
      "theoretical_loss": 3.5292292054081544,
      "tokens_seen": 1438056448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028500501504513545,
      "loss": 2.5581,
      "theoretical_loss": 3.5292149099919867,
      "tokens_seen": 1438121984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849949849548646,
      "loss": 2.7268,
      "theoretical_loss": 3.52920061540965,
      "tokens_seen": 1438187520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849849548645938,
      "loss": 2.6949,
      "theoretical_loss": 3.529186321661059,
      "tokens_seen": 1438253056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028497492477432294,
      "loss": 2.6476,
      "theoretical_loss": 3.5291720287461255,
      "tokens_seen": 1438318592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849648946840522,
      "loss": 2.4817,
      "theoretical_loss": 3.5291577366647635,
      "tokens_seen": 1438384128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028495486459378136,
      "loss": 2.4231,
      "theoretical_loss": 3.529143445416887,
      "tokens_seen": 1438449664
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 793361,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5532398223876953,
      "objective/train/theoretical_loss": 3.5291291550024093,
      "objective/train/tokens_used": 1458975200,
      "theoretical_loss": 3.5291291550024093,
      "tokens_seen": 1438515200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028494483450351054,
      "loss": 2.614,
      "theoretical_loss": 3.5291291550024093,
      "tokens_seen": 1438515200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849348044132397,
      "loss": 2.7064,
      "theoretical_loss": 3.529114865421243,
      "tokens_seen": 1438580736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849247743229689,
      "loss": 2.5424,
      "theoretical_loss": 3.5291005766733026,
      "tokens_seen": 1438646272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849147442326981,
      "loss": 2.8834,
      "theoretical_loss": 3.5290862887585,
      "tokens_seen": 1438711808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849047141424273,
      "loss": 2.808,
      "theoretical_loss": 3.5290720016767514,
      "tokens_seen": 1438777344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028489468405215644,
      "loss": 2.4885,
      "theoretical_loss": 3.5290577154279674,
      "tokens_seen": 1438842880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848846539618857,
      "loss": 2.9221,
      "theoretical_loss": 3.529043430012063,
      "tokens_seen": 1438908416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028487462387161486,
      "loss": 2.5549,
      "theoretical_loss": 3.5290291454289515,
      "tokens_seen": 1438973952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028486459378134404,
      "loss": 2.589,
      "theoretical_loss": 3.5290148616785464,
      "tokens_seen": 1439039488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848545636910732,
      "loss": 2.6573,
      "theoretical_loss": 3.529000578760761,
      "tokens_seen": 1439105024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848445336008024,
      "loss": 2.6267,
      "theoretical_loss": 3.5289862966755092,
      "tokens_seen": 1439170560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848345035105316,
      "loss": 2.6466,
      "theoretical_loss": 3.5289720154227044,
      "tokens_seen": 1439236096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848244734202608,
      "loss": 2.5453,
      "theoretical_loss": 3.52895773500226,
      "tokens_seen": 1439301632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028481444332998995,
      "loss": 2.4943,
      "theoretical_loss": 3.52894345541409,
      "tokens_seen": 1439367168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848044132397192,
      "loss": 2.7117,
      "theoretical_loss": 3.528929176658108,
      "tokens_seen": 1439432704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847943831494483,
      "loss": 2.6462,
      "theoretical_loss": 3.528914898734227,
      "tokens_seen": 1439498240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028478435305917755,
      "loss": 2.4815,
      "theoretical_loss": 3.528900621642361,
      "tokens_seen": 1439563776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847743229689067,
      "loss": 2.6197,
      "theoretical_loss": 3.528886345382424,
      "tokens_seen": 1439629312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847642928786359,
      "loss": 2.8349,
      "theoretical_loss": 3.5288720699543292,
      "tokens_seen": 1439694848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847542627883651,
      "loss": 2.7815,
      "theoretical_loss": 3.52885779535799,
      "tokens_seen": 1439760384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028474423269809427,
      "loss": 2.5242,
      "theoretical_loss": 3.5288435215933207,
      "tokens_seen": 1439825920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028473420260782345,
      "loss": 2.5404,
      "theoretical_loss": 3.5288292486602355,
      "tokens_seen": 1439891456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847241725175527,
      "loss": 2.6967,
      "theoretical_loss": 3.5288149765586465,
      "tokens_seen": 1439956992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847141424272818,
      "loss": 2.6013,
      "theoretical_loss": 3.5288007052884685,
      "tokens_seen": 1440022528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028470411233701105,
      "loss": 2.5721,
      "theoretical_loss": 3.528786434849615,
      "tokens_seen": 1440088064
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 794550,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.535278558731079,
      "objective/train/theoretical_loss": 3.5287721652419997,
      "objective/train/tokens_used": 1460613600,
      "theoretical_loss": 3.5287721652419997,
      "tokens_seen": 1440153600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028469408224674023,
      "loss": 2.6271,
      "theoretical_loss": 3.5287721652419997,
      "tokens_seen": 1440153600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846840521564694,
      "loss": 2.4973,
      "theoretical_loss": 3.528757896465537,
      "tokens_seen": 1440219136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846740220661986,
      "loss": 2.6563,
      "theoretical_loss": 3.5287436285201395,
      "tokens_seen": 1440284672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846639919759278,
      "loss": 2.7874,
      "theoretical_loss": 3.5287293614057216,
      "tokens_seen": 1440350208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028465396188565695,
      "loss": 2.5078,
      "theoretical_loss": 3.5287150951221973,
      "tokens_seen": 1440415744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846439317953862,
      "loss": 2.3781,
      "theoretical_loss": 3.5287008296694804,
      "tokens_seen": 1440481280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846339017051153,
      "loss": 2.5335,
      "theoretical_loss": 3.528686565047484,
      "tokens_seen": 1440546816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028462387161484455,
      "loss": 2.5804,
      "theoretical_loss": 3.5286723012561225,
      "tokens_seen": 1440612352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846138415245737,
      "loss": 2.6424,
      "theoretical_loss": 3.52865803829531,
      "tokens_seen": 1440677888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846038114343029,
      "loss": 2.6213,
      "theoretical_loss": 3.52864377616496,
      "tokens_seen": 1440743424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845937813440321,
      "loss": 2.658,
      "theoretical_loss": 3.5286295148649867,
      "tokens_seen": 1440808960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845837512537613,
      "loss": 2.7613,
      "theoretical_loss": 3.5286152543953033,
      "tokens_seen": 1440874496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028457372116349046,
      "loss": 2.6722,
      "theoretical_loss": 3.528600994755824,
      "tokens_seen": 1440940032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845636910732197,
      "loss": 2.6129,
      "theoretical_loss": 3.528586735946463,
      "tokens_seen": 1441005568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845536609829488,
      "loss": 2.4905,
      "theoretical_loss": 3.528572477967134,
      "tokens_seen": 1441071104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028454363089267805,
      "loss": 2.8319,
      "theoretical_loss": 3.528558220817751,
      "tokens_seen": 1441136640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028453360080240724,
      "loss": 2.8369,
      "theoretical_loss": 3.528543964498228,
      "tokens_seen": 1441202176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845235707121364,
      "loss": 2.5964,
      "theoretical_loss": 3.528529709008479,
      "tokens_seen": 1441267712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028451354062186565,
      "loss": 2.6241,
      "theoretical_loss": 3.5285154543484176,
      "tokens_seen": 1441333248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845035105315948,
      "loss": 2.6402,
      "theoretical_loss": 3.528501200517958,
      "tokens_seen": 1441398784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284493480441324,
      "loss": 2.5355,
      "theoretical_loss": 3.528486947517015,
      "tokens_seen": 1441464320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028448345035105314,
      "loss": 2.5674,
      "theoretical_loss": 3.528472695345501,
      "tokens_seen": 1441529856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844734202607824,
      "loss": 2.7757,
      "theoretical_loss": 3.5284584440033315,
      "tokens_seen": 1441595392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028446339017051156,
      "loss": 2.6715,
      "theoretical_loss": 3.5284441934904196,
      "tokens_seen": 1441660928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028445336008024074,
      "loss": 2.6472,
      "theoretical_loss": 3.5284299438066804,
      "tokens_seen": 1441726464
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 795512,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.223959445953369,
      "objective/train/theoretical_loss": 3.5284156949520264,
      "objective/train/tokens_used": 1462252000,
      "theoretical_loss": 3.5284156949520264,
      "tokens_seen": 1441792000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844433299899699,
      "loss": 2.4238,
      "theoretical_loss": 3.5284156949520264,
      "tokens_seen": 1441792000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844332998996991,
      "loss": 2.8375,
      "theoretical_loss": 3.528401446926373,
      "tokens_seen": 1441857536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844232698094283,
      "loss": 2.7861,
      "theoretical_loss": 3.528387199729634,
      "tokens_seen": 1441923072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844132397191575,
      "loss": 2.5875,
      "theoretical_loss": 3.528372953361723,
      "tokens_seen": 1441988608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028440320962888664,
      "loss": 2.65,
      "theoretical_loss": 3.5283587078225542,
      "tokens_seen": 1442054144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843931795386159,
      "loss": 2.6652,
      "theoretical_loss": 3.5283444631120426,
      "tokens_seen": 1442119680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028438314944834506,
      "loss": 2.7144,
      "theoretical_loss": 3.528330219230102,
      "tokens_seen": 1442185216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028437311935807424,
      "loss": 2.7403,
      "theoretical_loss": 3.5283159761766454,
      "tokens_seen": 1442250752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843630892678034,
      "loss": 2.6111,
      "theoretical_loss": 3.5283017339515883,
      "tokens_seen": 1442316288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843530591775326,
      "loss": 2.6709,
      "theoretical_loss": 3.5282874925548446,
      "tokens_seen": 1442381824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843430290872618,
      "loss": 2.5262,
      "theoretical_loss": 3.5282732519863282,
      "tokens_seen": 1442447360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284332998996991,
      "loss": 2.6623,
      "theoretical_loss": 3.5282590122459534,
      "tokens_seen": 1442512896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028432296890672015,
      "loss": 2.6434,
      "theoretical_loss": 3.5282447733336344,
      "tokens_seen": 1442578432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843129388164494,
      "loss": 2.6882,
      "theoretical_loss": 3.5282305352492855,
      "tokens_seen": 1442643968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843029087261785,
      "loss": 2.6735,
      "theoretical_loss": 3.528216297992821,
      "tokens_seen": 1442709504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028429287863590775,
      "loss": 2.7104,
      "theoretical_loss": 3.5282020615641554,
      "tokens_seen": 1442775040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842828485456369,
      "loss": 2.7093,
      "theoretical_loss": 3.5281878259632027,
      "tokens_seen": 1442840576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842728184553661,
      "loss": 2.6069,
      "theoretical_loss": 3.5281735911898764,
      "tokens_seen": 1442906112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842627883650953,
      "loss": 2.7555,
      "theoretical_loss": 3.5281593572440926,
      "tokens_seen": 1442971648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028425275827482447,
      "loss": 2.5602,
      "theoretical_loss": 3.5281451241257638,
      "tokens_seen": 1443037184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028424272818455365,
      "loss": 2.8941,
      "theoretical_loss": 3.528130891834805,
      "tokens_seen": 1443102720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842326980942829,
      "loss": 2.6407,
      "theoretical_loss": 3.528116660371131,
      "tokens_seen": 1443168256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284222668004012,
      "loss": 2.8021,
      "theoretical_loss": 3.528102429734656,
      "tokens_seen": 1443233792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028421263791374125,
      "loss": 2.4425,
      "theoretical_loss": 3.5280881999252935,
      "tokens_seen": 1443299328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028420260782347043,
      "loss": 2.434,
      "theoretical_loss": 3.528073970942959,
      "tokens_seen": 1443364864
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 796104,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4091320037841797,
      "objective/train/theoretical_loss": 3.5280597427875664,
      "objective/train/tokens_used": 1463890400,
      "theoretical_loss": 3.5280597427875664,
      "tokens_seen": 1443430400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841925777331996,
      "loss": 2.4999,
      "theoretical_loss": 3.5280597427875664,
      "tokens_seen": 1443430400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841825476429288,
      "loss": 2.7,
      "theoretical_loss": 3.52804551545903,
      "tokens_seen": 1443495936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284172517552658,
      "loss": 2.5288,
      "theoretical_loss": 3.528031288957264,
      "tokens_seen": 1443561472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028416248746238715,
      "loss": 2.6487,
      "theoretical_loss": 3.5280170632821832,
      "tokens_seen": 1443627008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841524573721164,
      "loss": 2.8229,
      "theoretical_loss": 3.528002838433702,
      "tokens_seen": 1443692544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841424272818455,
      "loss": 2.8376,
      "theoretical_loss": 3.527988614411735,
      "tokens_seen": 1443758080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028413239719157475,
      "loss": 2.4499,
      "theoretical_loss": 3.5279743912161963,
      "tokens_seen": 1443823616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841223671013039,
      "loss": 2.6224,
      "theoretical_loss": 3.5279601688470006,
      "tokens_seen": 1443889152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841123370110331,
      "loss": 2.632,
      "theoretical_loss": 3.5279459473040626,
      "tokens_seen": 1443954688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841023069207623,
      "loss": 2.7038,
      "theoretical_loss": 3.527931726587296,
      "tokens_seen": 1444020224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840922768304915,
      "loss": 2.4187,
      "theoretical_loss": 3.527917506696616,
      "tokens_seen": 1444085760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028408224674022066,
      "loss": 2.6817,
      "theoretical_loss": 3.527903287631937,
      "tokens_seen": 1444151296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840722166499499,
      "loss": 2.8673,
      "theoretical_loss": 3.527889069393174,
      "tokens_seen": 1444216832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284062186559679,
      "loss": 2.4415,
      "theoretical_loss": 3.5278748519802408,
      "tokens_seen": 1444282368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028405215646940825,
      "loss": 2.5479,
      "theoretical_loss": 3.527860635393052,
      "tokens_seen": 1444347904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840421263791374,
      "loss": 2.7754,
      "theoretical_loss": 3.5278464196315227,
      "tokens_seen": 1444413440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840320962888666,
      "loss": 2.7913,
      "theoretical_loss": 3.5278322046955664,
      "tokens_seen": 1444478976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840220661985958,
      "loss": 2.6492,
      "theoretical_loss": 3.527817990585099,
      "tokens_seen": 1444544512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284012036108325,
      "loss": 2.4502,
      "theoretical_loss": 3.527803777300035,
      "tokens_seen": 1444610048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028400200601805416,
      "loss": 2.7361,
      "theoretical_loss": 3.5277895648402886,
      "tokens_seen": 1444675584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028399197592778334,
      "loss": 2.5574,
      "theoretical_loss": 3.527775353205774,
      "tokens_seen": 1444741120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839819458375125,
      "loss": 2.6315,
      "theoretical_loss": 3.5277611423964066,
      "tokens_seen": 1444806656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028397191574724176,
      "loss": 2.887,
      "theoretical_loss": 3.527746932412101,
      "tokens_seen": 1444872192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839618856569709,
      "loss": 2.6596,
      "theoretical_loss": 3.527732723252771,
      "tokens_seen": 1444937728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839518555667001,
      "loss": 2.6856,
      "theoretical_loss": 3.527718514918332,
      "tokens_seen": 1445003264
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 797347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.670140504837036,
      "objective/train/theoretical_loss": 3.5277043074086993,
      "objective/train/tokens_used": 1465528800,
      "theoretical_loss": 3.5277043074086993,
      "tokens_seen": 1445068800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028394182547642925,
      "loss": 2.6536,
      "theoretical_loss": 3.5277043074086993,
      "tokens_seen": 1445068800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839317953861585,
      "loss": 2.6025,
      "theoretical_loss": 3.527690100723787,
      "tokens_seen": 1445134336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028392176529588766,
      "loss": 2.7662,
      "theoretical_loss": 3.5276758948635094,
      "tokens_seen": 1445199872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028391173520561684,
      "loss": 2.6632,
      "theoretical_loss": 3.5276616898277817,
      "tokens_seen": 1445265408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283901705115346,
      "loss": 2.4511,
      "theoretical_loss": 3.5276474856165185,
      "tokens_seen": 1445330944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028389167502507526,
      "loss": 2.6842,
      "theoretical_loss": 3.527633282229635,
      "tokens_seen": 1445396480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838816449348044,
      "loss": 2.654,
      "theoretical_loss": 3.527619079667046,
      "tokens_seen": 1445462016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838716148445336,
      "loss": 2.7173,
      "theoretical_loss": 3.527604877928665,
      "tokens_seen": 1445527552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028386158475426275,
      "loss": 2.5172,
      "theoretical_loss": 3.5275906770144085,
      "tokens_seen": 1445593088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283851554663992,
      "loss": 2.631,
      "theoretical_loss": 3.5275764769241906,
      "tokens_seen": 1445658624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028384152457372117,
      "loss": 2.5484,
      "theoretical_loss": 3.5275622776579256,
      "tokens_seen": 1445724160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028383149448345035,
      "loss": 2.618,
      "theoretical_loss": 3.527548079215529,
      "tokens_seen": 1445789696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028382146439317953,
      "loss": 2.3792,
      "theoretical_loss": 3.5275338815969155,
      "tokens_seen": 1445855232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838114343029087,
      "loss": 2.7984,
      "theoretical_loss": 3.5275196848020003,
      "tokens_seen": 1445920768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838014042126379,
      "loss": 2.5403,
      "theoretical_loss": 3.5275054888306974,
      "tokens_seen": 1445986304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837913741223671,
      "loss": 2.7358,
      "theoretical_loss": 3.527491293682923,
      "tokens_seen": 1446051840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837813440320963,
      "loss": 2.8928,
      "theoretical_loss": 3.527477099358591,
      "tokens_seen": 1446117376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837713139418255,
      "loss": 2.7105,
      "theoretical_loss": 3.5274629058576163,
      "tokens_seen": 1446182912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028376128385155467,
      "loss": 2.6156,
      "theoretical_loss": 3.527448713179914,
      "tokens_seen": 1446248448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028375125376128385,
      "loss": 2.4957,
      "theoretical_loss": 3.5274345213253993,
      "tokens_seen": 1446313984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837412236710131,
      "loss": 2.9047,
      "theoretical_loss": 3.527420330293987,
      "tokens_seen": 1446379520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837311935807422,
      "loss": 2.7489,
      "theoretical_loss": 3.527406140085592,
      "tokens_seen": 1446445056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028372116349047145,
      "loss": 2.7722,
      "theoretical_loss": 3.52739195070013,
      "tokens_seen": 1446510592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028371113340020063,
      "loss": 2.6869,
      "theoretical_loss": 3.5273777621375144,
      "tokens_seen": 1446576128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837011033099298,
      "loss": 2.6216,
      "theoretical_loss": 3.5273635743976617,
      "tokens_seen": 1446641664
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 798075,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2298080921173096,
      "objective/train/theoretical_loss": 3.527349387480486,
      "objective/train/tokens_used": 1467167200,
      "theoretical_loss": 3.527349387480486,
      "tokens_seen": 1446707200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283691073219659,
      "loss": 2.4326,
      "theoretical_loss": 3.527349387480486,
      "tokens_seen": 1446707200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836810431293882,
      "loss": 2.5288,
      "theoretical_loss": 3.5273352013859034,
      "tokens_seen": 1446772736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028367101303911735,
      "loss": 2.5907,
      "theoretical_loss": 3.5273210161138278,
      "tokens_seen": 1446838272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836609829488466,
      "loss": 2.7755,
      "theoretical_loss": 3.5273068316641747,
      "tokens_seen": 1446903808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836509528585757,
      "loss": 2.8434,
      "theoretical_loss": 3.5272926480368594,
      "tokens_seen": 1446969344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028364092276830495,
      "loss": 2.7317,
      "theoretical_loss": 3.5272784652317966,
      "tokens_seen": 1447034880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836308926780341,
      "loss": 2.663,
      "theoretical_loss": 3.5272642832489014,
      "tokens_seen": 1447100416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836208625877633,
      "loss": 2.4391,
      "theoretical_loss": 3.5272501020880895,
      "tokens_seen": 1447165952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836108324974925,
      "loss": 2.5177,
      "theoretical_loss": 3.5272359217492752,
      "tokens_seen": 1447231488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836008024072217,
      "loss": 2.6772,
      "theoretical_loss": 3.527221742232374,
      "tokens_seen": 1447297024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028359077231695086,
      "loss": 2.5987,
      "theoretical_loss": 3.5272075635373015,
      "tokens_seen": 1447362560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835807422266801,
      "loss": 2.6612,
      "theoretical_loss": 3.5271933856639723,
      "tokens_seen": 1447428096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835707121364092,
      "loss": 2.5096,
      "theoretical_loss": 3.5271792086123015,
      "tokens_seen": 1447493632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028356068204613846,
      "loss": 2.4976,
      "theoretical_loss": 3.5271650323822046,
      "tokens_seen": 1447559168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835506519558676,
      "loss": 2.8078,
      "theoretical_loss": 3.5271508569735968,
      "tokens_seen": 1447624704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835406218655968,
      "loss": 2.5067,
      "theoretical_loss": 3.527136682386393,
      "tokens_seen": 1447690240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283530591775326,
      "loss": 2.6418,
      "theoretical_loss": 3.5271225086205087,
      "tokens_seen": 1447755776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835205616850552,
      "loss": 2.6431,
      "theoretical_loss": 3.527108335675859,
      "tokens_seen": 1447821312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028351053159478436,
      "loss": 2.8093,
      "theoretical_loss": 3.527094163552359,
      "tokens_seen": 1447886848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028350050150451354,
      "loss": 2.4425,
      "theoretical_loss": 3.5270799922499245,
      "tokens_seen": 1447952384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834904714142427,
      "loss": 2.5708,
      "theoretical_loss": 3.5270658217684705,
      "tokens_seen": 1448017920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028348044132397196,
      "loss": 2.6207,
      "theoretical_loss": 3.5270516521079123,
      "tokens_seen": 1448083456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834704112337011,
      "loss": 2.6184,
      "theoretical_loss": 3.527037483268165,
      "tokens_seen": 1448148992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834603811434303,
      "loss": 2.3861,
      "theoretical_loss": 3.527023315249144,
      "tokens_seen": 1448214528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028345035105315945,
      "loss": 2.7098,
      "theoretical_loss": 3.527009148050764,
      "tokens_seen": 1448280064
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 799347,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.110480546951294,
      "objective/train/theoretical_loss": 3.5269949816729422,
      "objective/train/tokens_used": 1468805600,
      "theoretical_loss": 3.5269949816729422,
      "tokens_seen": 1448345600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834403209628887,
      "loss": 2.6722,
      "theoretical_loss": 3.5269949816729422,
      "tokens_seen": 1448345600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028343029087261786,
      "loss": 2.7585,
      "theoretical_loss": 3.526980816115592,
      "tokens_seen": 1448411136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028342026078234705,
      "loss": 2.7324,
      "theoretical_loss": 3.526966651378629,
      "tokens_seen": 1448476672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834102306920762,
      "loss": 2.7324,
      "theoretical_loss": 3.52695248746197,
      "tokens_seen": 1448542208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028340020060180546,
      "loss": 2.7503,
      "theoretical_loss": 3.5269383243655295,
      "tokens_seen": 1448607744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833901705115346,
      "loss": 2.7735,
      "theoretical_loss": 3.526924162089222,
      "tokens_seen": 1448673280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833801404212638,
      "loss": 2.6382,
      "theoretical_loss": 3.5269100006329643,
      "tokens_seen": 1448738816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028337011033099295,
      "loss": 2.6276,
      "theoretical_loss": 3.526895839996671,
      "tokens_seen": 1448804352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833600802407222,
      "loss": 2.6102,
      "theoretical_loss": 3.5268816801802583,
      "tokens_seen": 1448869888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028335005015045137,
      "loss": 2.5135,
      "theoretical_loss": 3.5268675211836404,
      "tokens_seen": 1448935424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028334002006018055,
      "loss": 2.711,
      "theoretical_loss": 3.526853363006734,
      "tokens_seen": 1449000960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028332998996990973,
      "loss": 2.6132,
      "theoretical_loss": 3.526839205649454,
      "tokens_seen": 1449066496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833199598796389,
      "loss": 2.5544,
      "theoretical_loss": 3.526825049111716,
      "tokens_seen": 1449132032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833099297893681,
      "loss": 2.537,
      "theoretical_loss": 3.526810893393436,
      "tokens_seen": 1449197568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832998996990973,
      "loss": 2.6697,
      "theoretical_loss": 3.5267967384945287,
      "tokens_seen": 1449263104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028328986960882645,
      "loss": 2.58,
      "theoretical_loss": 3.526782584414909,
      "tokens_seen": 1449328640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832798395185557,
      "loss": 2.75,
      "theoretical_loss": 3.526768431154495,
      "tokens_seen": 1449394176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832698094282848,
      "loss": 2.7828,
      "theoretical_loss": 3.5267542787131996,
      "tokens_seen": 1449459712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028325977933801405,
      "loss": 2.4813,
      "theoretical_loss": 3.526740127090939,
      "tokens_seen": 1449525248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028324974924774323,
      "loss": 2.8291,
      "theoretical_loss": 3.52672597628763,
      "tokens_seen": 1449590784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832397191574724,
      "loss": 2.5883,
      "theoretical_loss": 3.5267118263031865,
      "tokens_seen": 1449656320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832296890672016,
      "loss": 2.8175,
      "theoretical_loss": 3.5266976771375256,
      "tokens_seen": 1449721856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028321965897693083,
      "loss": 2.5814,
      "theoretical_loss": 3.526683528790562,
      "tokens_seen": 1449787392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028320962888665996,
      "loss": 2.4963,
      "theoretical_loss": 3.5266693812622116,
      "tokens_seen": 1449852928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831995987963892,
      "loss": 2.5982,
      "theoretical_loss": 3.52665523455239,
      "tokens_seen": 1449918464
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 800715,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5400002002716064,
      "objective/train/theoretical_loss": 3.5266410886610124,
      "objective/train/tokens_used": 1470444000,
      "theoretical_loss": 3.5266410886610124,
      "tokens_seen": 1449984000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831895687061183,
      "loss": 2.5756,
      "theoretical_loss": 3.5266410886610124,
      "tokens_seen": 1449984000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028317953861584755,
      "loss": 2.6819,
      "theoretical_loss": 3.5266269435879956,
      "tokens_seen": 1450049536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028316950852557674,
      "loss": 2.5987,
      "theoretical_loss": 3.526612799333254,
      "tokens_seen": 1450115072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831594784353059,
      "loss": 2.593,
      "theoretical_loss": 3.526598655896704,
      "tokens_seen": 1450180608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831494483450351,
      "loss": 2.6084,
      "theoretical_loss": 3.5265845132782614,
      "tokens_seen": 1450246144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831394182547643,
      "loss": 2.6047,
      "theoretical_loss": 3.526570371477842,
      "tokens_seen": 1450311680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028312938816449346,
      "loss": 2.7709,
      "theoretical_loss": 3.5265562304953604,
      "tokens_seen": 1450377216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831193580742227,
      "loss": 2.6368,
      "theoretical_loss": 3.526542090330733,
      "tokens_seen": 1450442752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831093279839518,
      "loss": 2.6572,
      "theoretical_loss": 3.5265279509838763,
      "tokens_seen": 1450508288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028309929789368106,
      "loss": 2.4918,
      "theoretical_loss": 3.5265138124547057,
      "tokens_seen": 1450573824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830892678034102,
      "loss": 2.5653,
      "theoretical_loss": 3.5264996747431363,
      "tokens_seen": 1450639360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830792377131394,
      "loss": 2.5689,
      "theoretical_loss": 3.5264855378490845,
      "tokens_seen": 1450704896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830692076228686,
      "loss": 2.6756,
      "theoretical_loss": 3.526471401772465,
      "tokens_seen": 1450770432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830591775325978,
      "loss": 2.8919,
      "theoretical_loss": 3.5264572665131952,
      "tokens_seen": 1450835968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028304914744232696,
      "loss": 2.4432,
      "theoretical_loss": 3.5264431320711904,
      "tokens_seen": 1450901504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830391173520562,
      "loss": 2.7723,
      "theoretical_loss": 3.526428998446366,
      "tokens_seen": 1450967040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830290872617854,
      "loss": 2.7163,
      "theoretical_loss": 3.5264148656386385,
      "tokens_seen": 1451032576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028301905717151456,
      "loss": 2.6018,
      "theoretical_loss": 3.526400733647923,
      "tokens_seen": 1451098112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028300902708124374,
      "loss": 2.5447,
      "theoretical_loss": 3.526386602474136,
      "tokens_seen": 1451163648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829989969909729,
      "loss": 2.6576,
      "theoretical_loss": 3.5263724721171927,
      "tokens_seen": 1451229184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028298896690070216,
      "loss": 2.4897,
      "theoretical_loss": 3.5263583425770095,
      "tokens_seen": 1451294720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829789368104313,
      "loss": 2.4394,
      "theoretical_loss": 3.526344213853503,
      "tokens_seen": 1451360256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829689067201605,
      "loss": 2.7323,
      "theoretical_loss": 3.5263300859465874,
      "tokens_seen": 1451425792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028295887662988965,
      "loss": 2.6551,
      "theoretical_loss": 3.52631595885618,
      "tokens_seen": 1451491328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829488465396189,
      "loss": 2.5651,
      "theoretical_loss": 3.526301832582196,
      "tokens_seen": 1451556864
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 801482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5024585723876953,
      "objective/train/theoretical_loss": 3.5262877071245526,
      "objective/train/tokens_used": 1472082400,
      "theoretical_loss": 3.5262877071245526,
      "tokens_seen": 1451622400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028293881644934806,
      "loss": 2.6494,
      "theoretical_loss": 3.5262877071245526,
      "tokens_seen": 1451622400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028292878635907725,
      "loss": 2.6545,
      "theoretical_loss": 3.526273582483164,
      "tokens_seen": 1451687936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829187562688064,
      "loss": 2.3686,
      "theoretical_loss": 3.5262594586579477,
      "tokens_seen": 1451753472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028290872617853566,
      "loss": 2.4083,
      "theoretical_loss": 3.5262453356488184,
      "tokens_seen": 1451819008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828986960882648,
      "loss": 2.9027,
      "theoretical_loss": 3.5262312134556932,
      "tokens_seen": 1451884544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282888665997994,
      "loss": 2.4601,
      "theoretical_loss": 3.526217092078488,
      "tokens_seen": 1451950080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028287863590772315,
      "loss": 2.7495,
      "theoretical_loss": 3.526202971517118,
      "tokens_seen": 1452015616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828686058174524,
      "loss": 2.3887,
      "theoretical_loss": 3.5261888517715,
      "tokens_seen": 1452081152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028285857572718157,
      "loss": 2.4065,
      "theoretical_loss": 3.52617473284155,
      "tokens_seen": 1452146688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028284854563691075,
      "loss": 2.8447,
      "theoretical_loss": 3.526160614727184,
      "tokens_seen": 1452212224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028283851554663993,
      "loss": 2.7958,
      "theoretical_loss": 3.526146497428317,
      "tokens_seen": 1452277760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828284854563691,
      "loss": 2.6396,
      "theoretical_loss": 3.5261323809448673,
      "tokens_seen": 1452343296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828184553660983,
      "loss": 2.7114,
      "theoretical_loss": 3.5261182652767493,
      "tokens_seen": 1452408832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828084252758275,
      "loss": 2.6084,
      "theoretical_loss": 3.5261041504238797,
      "tokens_seen": 1452474368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028279839518555665,
      "loss": 2.4572,
      "theoretical_loss": 3.526090036386174,
      "tokens_seen": 1452539904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827883650952859,
      "loss": 2.681,
      "theoretical_loss": 3.5260759231635497,
      "tokens_seen": 1452605440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282778335005015,
      "loss": 2.6388,
      "theoretical_loss": 3.526061810755922,
      "tokens_seen": 1452670976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028276830491474425,
      "loss": 2.5954,
      "theoretical_loss": 3.526047699163207,
      "tokens_seen": 1452736512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028275827482447343,
      "loss": 2.5933,
      "theoretical_loss": 3.5260335883853213,
      "tokens_seen": 1452802048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827482447342026,
      "loss": 2.5419,
      "theoretical_loss": 3.5260194784221808,
      "tokens_seen": 1452867584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827382146439318,
      "loss": 2.3594,
      "theoretical_loss": 3.526005369273702,
      "tokens_seen": 1452933120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028272818455366103,
      "loss": 2.7973,
      "theoretical_loss": 3.5259912609398008,
      "tokens_seen": 1452998656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028271815446339016,
      "loss": 2.6136,
      "theoretical_loss": 3.525977153420394,
      "tokens_seen": 1453064192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827081243731194,
      "loss": 2.5128,
      "theoretical_loss": 3.525963046715397,
      "tokens_seen": 1453129728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826980942828485,
      "loss": 2.5333,
      "theoretical_loss": 3.525948940824726,
      "tokens_seen": 1453195264
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 802380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2598981857299805,
      "objective/train/theoretical_loss": 3.5259348357482985,
      "objective/train/tokens_used": 1473720800,
      "theoretical_loss": 3.5259348357482985,
      "tokens_seen": 1453260800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028268806419257775,
      "loss": 2.4851,
      "theoretical_loss": 3.5259348357482985,
      "tokens_seen": 1453260800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028267803410230694,
      "loss": 2.4297,
      "theoretical_loss": 3.52592073148603,
      "tokens_seen": 1453326336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826680040120361,
      "loss": 2.6273,
      "theoretical_loss": 3.525906628037837,
      "tokens_seen": 1453391872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826579739217653,
      "loss": 2.5379,
      "theoretical_loss": 3.525892525403635,
      "tokens_seen": 1453457408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826479438314945,
      "loss": 2.5902,
      "theoretical_loss": 3.5258784235833414,
      "tokens_seen": 1453522944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028263791374122366,
      "loss": 2.6003,
      "theoretical_loss": 3.525864322576872,
      "tokens_seen": 1453588480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826278836509529,
      "loss": 2.7424,
      "theoretical_loss": 3.5258502223841433,
      "tokens_seen": 1453654016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282617853560682,
      "loss": 2.9454,
      "theoretical_loss": 3.5258361230050714,
      "tokens_seen": 1453719552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028260782347041126,
      "loss": 2.6894,
      "theoretical_loss": 3.525822024439573,
      "tokens_seen": 1453785088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825977933801404,
      "loss": 2.6092,
      "theoretical_loss": 3.5258079266875644,
      "tokens_seen": 1453850624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825877632898696,
      "loss": 2.5426,
      "theoretical_loss": 3.525793829748962,
      "tokens_seen": 1453916160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825777331995988,
      "loss": 2.6994,
      "theoretical_loss": 3.5257797336236822,
      "tokens_seen": 1453981696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282567703109328,
      "loss": 2.4356,
      "theoretical_loss": 3.525765638311641,
      "tokens_seen": 1454047232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028255767301905716,
      "loss": 2.681,
      "theoretical_loss": 3.5257515438127554,
      "tokens_seen": 1454112768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825476429287864,
      "loss": 2.6066,
      "theoretical_loss": 3.5257374501269423,
      "tokens_seen": 1454178304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825376128385155,
      "loss": 2.7642,
      "theoretical_loss": 3.5257233572541167,
      "tokens_seen": 1454243840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028252758274824476,
      "loss": 2.6863,
      "theoretical_loss": 3.525709265194196,
      "tokens_seen": 1454309376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825175526579739,
      "loss": 2.657,
      "theoretical_loss": 3.5256951739470965,
      "tokens_seen": 1454374912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825075225677031,
      "loss": 2.8041,
      "theoretical_loss": 3.525681083512735,
      "tokens_seen": 1454440448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824974924774323,
      "loss": 2.5356,
      "theoretical_loss": 3.5256669938910274,
      "tokens_seen": 1454505984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824874623871615,
      "loss": 2.7467,
      "theoretical_loss": 3.525652905081891,
      "tokens_seen": 1454571520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028247743229689067,
      "loss": 2.8853,
      "theoretical_loss": 3.525638817085242,
      "tokens_seen": 1454637056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028246740220661985,
      "loss": 2.786,
      "theoretical_loss": 3.525624729900996,
      "tokens_seen": 1454702592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028245737211634903,
      "loss": 2.7916,
      "theoretical_loss": 3.5256106435290713,
      "tokens_seen": 1454768128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028244734202607826,
      "loss": 2.6602,
      "theoretical_loss": 3.5255965579693838,
      "tokens_seen": 1454833664
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 802899,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6358089447021484,
      "objective/train/theoretical_loss": 3.525582473221849,
      "objective/train/tokens_used": 1475359200,
      "theoretical_loss": 3.525582473221849,
      "tokens_seen": 1454899200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824373119358074,
      "loss": 2.8029,
      "theoretical_loss": 3.525582473221849,
      "tokens_seen": 1454899200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824272818455366,
      "loss": 2.6266,
      "theoretical_loss": 3.5255683892863843,
      "tokens_seen": 1454964736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824172517552658,
      "loss": 2.4783,
      "theoretical_loss": 3.525554306162907,
      "tokens_seen": 1455030272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282407221664995,
      "loss": 2.5727,
      "theoretical_loss": 3.5255402238513325,
      "tokens_seen": 1455095808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028239719157472417,
      "loss": 2.6995,
      "theoretical_loss": 3.5255261423515782,
      "tokens_seen": 1455161344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028238716148445335,
      "loss": 2.6435,
      "theoretical_loss": 3.525512061663561,
      "tokens_seen": 1455226880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028237713139418253,
      "loss": 2.7177,
      "theoretical_loss": 3.5254979817871965,
      "tokens_seen": 1455292416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028236710130391177,
      "loss": 2.7309,
      "theoretical_loss": 3.525483902722402,
      "tokens_seen": 1455357952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823570712136409,
      "loss": 2.7335,
      "theoretical_loss": 3.525469824469094,
      "tokens_seen": 1455423488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028234704112337013,
      "loss": 2.6619,
      "theoretical_loss": 3.52545574702719,
      "tokens_seen": 1455489024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028233701103309926,
      "loss": 2.5114,
      "theoretical_loss": 3.5254416703966056,
      "tokens_seen": 1455554560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823269809428285,
      "loss": 2.7022,
      "theoretical_loss": 3.5254275945772577,
      "tokens_seen": 1455620096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823169508525577,
      "loss": 2.8818,
      "theoretical_loss": 3.5254135195690637,
      "tokens_seen": 1455685632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028230692076228685,
      "loss": 2.5581,
      "theoretical_loss": 3.5253994453719395,
      "tokens_seen": 1455751168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028229689067201604,
      "loss": 2.8764,
      "theoretical_loss": 3.525385371985802,
      "tokens_seen": 1455816704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822868605817452,
      "loss": 2.7061,
      "theoretical_loss": 3.525371299410569,
      "tokens_seen": 1455882240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028227683049147445,
      "loss": 2.6916,
      "theoretical_loss": 3.525357227646156,
      "tokens_seen": 1455947776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028226680040120363,
      "loss": 2.6838,
      "theoretical_loss": 3.5253431566924807,
      "tokens_seen": 1456013312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822567703109328,
      "loss": 2.6026,
      "theoretical_loss": 3.5253290865494593,
      "tokens_seen": 1456078848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282246740220662,
      "loss": 2.5756,
      "theoretical_loss": 3.5253150172170087,
      "tokens_seen": 1456144384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028223671013039123,
      "loss": 2.5553,
      "theoretical_loss": 3.5253009486950457,
      "tokens_seen": 1456209920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028222668004012036,
      "loss": 2.589,
      "theoretical_loss": 3.525286880983488,
      "tokens_seen": 1456275456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822166499498496,
      "loss": 2.7555,
      "theoretical_loss": 3.5252728140822507,
      "tokens_seen": 1456340992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822066198595787,
      "loss": 2.5639,
      "theoretical_loss": 3.525258747991252,
      "tokens_seen": 1456406528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028219658976930795,
      "loss": 2.6847,
      "theoretical_loss": 3.525244682710409,
      "tokens_seen": 1456472064
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 804182,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6662416458129883,
      "objective/train/theoretical_loss": 3.525230618239638,
      "objective/train/tokens_used": 1476997600,
      "theoretical_loss": 3.525230618239638,
      "tokens_seen": 1456537600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028218655967903714,
      "loss": 2.5008,
      "theoretical_loss": 3.525230618239638,
      "tokens_seen": 1456537600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821765295887663,
      "loss": 2.5923,
      "theoretical_loss": 3.525216554578855,
      "tokens_seen": 1456603136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821664994984955,
      "loss": 2.7399,
      "theoretical_loss": 3.5252024917279785,
      "tokens_seen": 1456668672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821564694082247,
      "loss": 2.8464,
      "theoretical_loss": 3.525188429686925,
      "tokens_seen": 1456734208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028214643931795386,
      "loss": 2.6559,
      "theoretical_loss": 3.5251743684556107,
      "tokens_seen": 1456799744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821364092276831,
      "loss": 2.8568,
      "theoretical_loss": 3.5251603080339535,
      "tokens_seen": 1456865280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821263791374122,
      "loss": 2.468,
      "theoretical_loss": 3.525146248421869,
      "tokens_seen": 1456930816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028211634904714146,
      "loss": 2.6813,
      "theoretical_loss": 3.5251321896192764,
      "tokens_seen": 1456996352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821063189568706,
      "loss": 2.7486,
      "theoretical_loss": 3.5251181316260904,
      "tokens_seen": 1457061888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820962888665998,
      "loss": 2.3982,
      "theoretical_loss": 3.5251040744422295,
      "tokens_seen": 1457127424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282086258776329,
      "loss": 2.6629,
      "theoretical_loss": 3.52509001806761,
      "tokens_seen": 1457192960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820762286860582,
      "loss": 2.6004,
      "theoretical_loss": 3.5250759625021493,
      "tokens_seen": 1457258496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028206619859578736,
      "loss": 2.6276,
      "theoretical_loss": 3.525061907745764,
      "tokens_seen": 1457324032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820561685055166,
      "loss": 2.7723,
      "theoretical_loss": 3.5250478537983714,
      "tokens_seen": 1457389568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820461384152457,
      "loss": 2.5502,
      "theoretical_loss": 3.5250338006598887,
      "tokens_seen": 1457455104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028203610832497496,
      "loss": 2.6508,
      "theoretical_loss": 3.525019748330233,
      "tokens_seen": 1457520640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820260782347041,
      "loss": 2.5532,
      "theoretical_loss": 3.5250056968093206,
      "tokens_seen": 1457586176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820160481444333,
      "loss": 2.647,
      "theoretical_loss": 3.5249916460970696,
      "tokens_seen": 1457651712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820060180541625,
      "loss": 2.5241,
      "theoretical_loss": 3.5249775961933967,
      "tokens_seen": 1457717248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819959879638917,
      "loss": 2.5282,
      "theoretical_loss": 3.524963547098219,
      "tokens_seen": 1457782784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028198595787362087,
      "loss": 2.6907,
      "theoretical_loss": 3.524949498811454,
      "tokens_seen": 1457848320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028197592778335005,
      "loss": 2.7418,
      "theoretical_loss": 3.5249354513330178,
      "tokens_seen": 1457913856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028196589769307923,
      "loss": 2.4679,
      "theoretical_loss": 3.5249214046628286,
      "tokens_seen": 1457979392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028195586760280846,
      "loss": 2.6644,
      "theoretical_loss": 3.524907358800803,
      "tokens_seen": 1458044928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819458375125376,
      "loss": 2.8422,
      "theoretical_loss": 3.5248933137468588,
      "tokens_seen": 1458110464
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 804616,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4237711429595947,
      "objective/train/theoretical_loss": 3.5248792695009126,
      "objective/train/tokens_used": 1478636000,
      "theoretical_loss": 3.5248792695009126,
      "tokens_seen": 1458176000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819358074222668,
      "loss": 2.691,
      "theoretical_loss": 3.5248792695009126,
      "tokens_seen": 1458176000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281925777331996,
      "loss": 2.9381,
      "theoretical_loss": 3.5248652260628814,
      "tokens_seen": 1458241536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819157472417252,
      "loss": 2.7701,
      "theoretical_loss": 3.5248511834326837,
      "tokens_seen": 1458307072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028190571715145437,
      "loss": 2.7494,
      "theoretical_loss": 3.5248371416102353,
      "tokens_seen": 1458372608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028189568706118355,
      "loss": 2.8327,
      "theoretical_loss": 3.5248231005954542,
      "tokens_seen": 1458438144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028188565697091273,
      "loss": 2.5657,
      "theoretical_loss": 3.5248090603882574,
      "tokens_seen": 1458503680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028187562688064197,
      "loss": 2.7865,
      "theoretical_loss": 3.524795020988562,
      "tokens_seen": 1458569216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818655967903711,
      "loss": 2.5236,
      "theoretical_loss": 3.5247809823962855,
      "tokens_seen": 1458634752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028185556670010033,
      "loss": 2.5228,
      "theoretical_loss": 3.524766944611345,
      "tokens_seen": 1458700288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028184553660982946,
      "loss": 2.6959,
      "theoretical_loss": 3.524752907633659,
      "tokens_seen": 1458765824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818355065195587,
      "loss": 2.6638,
      "theoretical_loss": 3.5247388714631427,
      "tokens_seen": 1458831360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818254764292879,
      "loss": 2.5239,
      "theoretical_loss": 3.5247248360997148,
      "tokens_seen": 1458896896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028181544633901705,
      "loss": 2.58,
      "theoretical_loss": 3.524710801543293,
      "tokens_seen": 1458962432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028180541624874624,
      "loss": 2.3986,
      "theoretical_loss": 3.5246967677937935,
      "tokens_seen": 1459027968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817953861584754,
      "loss": 2.6534,
      "theoretical_loss": 3.524682734851134,
      "tokens_seen": 1459093504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817853560682046,
      "loss": 2.6836,
      "theoretical_loss": 3.524668702715232,
      "tokens_seen": 1459159040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028177532597793383,
      "loss": 2.7844,
      "theoretical_loss": 3.5246546713860054,
      "tokens_seen": 1459224576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028176529588766296,
      "loss": 2.4228,
      "theoretical_loss": 3.524640640863371,
      "tokens_seen": 1459290112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817552657973922,
      "loss": 2.5844,
      "theoretical_loss": 3.524626611147246,
      "tokens_seen": 1459355648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817452357071214,
      "loss": 2.443,
      "theoretical_loss": 3.524612582237549,
      "tokens_seen": 1459421184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028173520561685056,
      "loss": 2.8536,
      "theoretical_loss": 3.524598554134196,
      "tokens_seen": 1459486720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028172517552657974,
      "loss": 2.5898,
      "theoretical_loss": 3.5245845268371054,
      "tokens_seen": 1459552256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817151454363089,
      "loss": 2.8134,
      "theoretical_loss": 3.5245705003461936,
      "tokens_seen": 1459617792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817051153460381,
      "loss": 2.8013,
      "theoretical_loss": 3.5245564746613796,
      "tokens_seen": 1459683328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028169508525576734,
      "loss": 2.7783,
      "theoretical_loss": 3.52454244978258,
      "tokens_seen": 1459748864
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 805828,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6058542728424072,
      "objective/train/theoretical_loss": 3.524528425709712,
      "objective/train/tokens_used": 1480274400,
      "theoretical_loss": 3.524528425709712,
      "tokens_seen": 1459814400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028168505516549646,
      "loss": 2.7628,
      "theoretical_loss": 3.524528425709712,
      "tokens_seen": 1459814400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816750250752257,
      "loss": 2.6103,
      "theoretical_loss": 3.5245144024426938,
      "tokens_seen": 1459879936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816649949849548,
      "loss": 2.7047,
      "theoretical_loss": 3.5245003799814425,
      "tokens_seen": 1459945472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028165496489468406,
      "loss": 2.6731,
      "theoretical_loss": 3.5244863583258756,
      "tokens_seen": 1460011008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028164493480441324,
      "loss": 2.9568,
      "theoretical_loss": 3.524472337475911,
      "tokens_seen": 1460076544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816349047141424,
      "loss": 2.7297,
      "theoretical_loss": 3.524458317431466,
      "tokens_seen": 1460142080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816248746238716,
      "loss": 2.501,
      "theoretical_loss": 3.524444298192458,
      "tokens_seen": 1460207616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816148445336008,
      "loss": 2.7512,
      "theoretical_loss": 3.524430279758805,
      "tokens_seen": 1460273152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028160481444332997,
      "loss": 2.6057,
      "theoretical_loss": 3.5244162621304245,
      "tokens_seen": 1460338688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815947843530592,
      "loss": 2.6087,
      "theoretical_loss": 3.5244022453072343,
      "tokens_seen": 1460404224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028158475426278833,
      "loss": 2.7053,
      "theoretical_loss": 3.5243882292891513,
      "tokens_seen": 1460469760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028157472417251756,
      "loss": 2.7984,
      "theoretical_loss": 3.5243742140760936,
      "tokens_seen": 1460535296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028156469408224674,
      "loss": 2.5999,
      "theoretical_loss": 3.5243601996679788,
      "tokens_seen": 1460600832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815546639919759,
      "loss": 2.4843,
      "theoretical_loss": 3.524346186064725,
      "tokens_seen": 1460666368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815446339017051,
      "loss": 2.8821,
      "theoretical_loss": 3.5243321732662487,
      "tokens_seen": 1460731904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815346038114343,
      "loss": 2.413,
      "theoretical_loss": 3.5243181612724688,
      "tokens_seen": 1460797440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815245737211635,
      "loss": 2.6267,
      "theoretical_loss": 3.5243041500833026,
      "tokens_seen": 1460862976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815145436308927,
      "loss": 2.5767,
      "theoretical_loss": 3.5242901396986674,
      "tokens_seen": 1460928512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815045135406219,
      "loss": 2.6285,
      "theoretical_loss": 3.524276130118481,
      "tokens_seen": 1460994048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028149448345035107,
      "loss": 2.464,
      "theoretical_loss": 3.5242621213426615,
      "tokens_seen": 1461059584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028148445336008025,
      "loss": 2.6859,
      "theoretical_loss": 3.5242481133711268,
      "tokens_seen": 1461125120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028147442326980943,
      "loss": 2.6109,
      "theoretical_loss": 3.524234106203794,
      "tokens_seen": 1461190656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028146439317953866,
      "loss": 2.6543,
      "theoretical_loss": 3.5242200998405817,
      "tokens_seen": 1461256192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814543630892678,
      "loss": 2.8038,
      "theoretical_loss": 3.5242060942814066,
      "tokens_seen": 1461321728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281444332998997,
      "loss": 2.5912,
      "theoretical_loss": 3.5241920895261876,
      "tokens_seen": 1461387264
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 806380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.800405740737915,
      "objective/train/theoretical_loss": 3.5241780855748415,
      "objective/train/tokens_used": 1481912800,
      "theoretical_loss": 3.5241780855748415,
      "tokens_seen": 1461452800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814343029087262,
      "loss": 2.5523,
      "theoretical_loss": 3.5241780855748415,
      "tokens_seen": 1461452800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814242728184554,
      "loss": 2.8003,
      "theoretical_loss": 3.5241640824272866,
      "tokens_seen": 1461518336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028141424272818457,
      "loss": 2.5437,
      "theoretical_loss": 3.524150080083441,
      "tokens_seen": 1461583872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028140421263791375,
      "loss": 2.6448,
      "theoretical_loss": 3.524136078543222,
      "tokens_seen": 1461649408
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028139418254764293,
      "loss": 2.5327,
      "theoretical_loss": 3.5241220778065476,
      "tokens_seen": 1461714944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028138415245737217,
      "loss": 2.8325,
      "theoretical_loss": 3.5241080778733362,
      "tokens_seen": 1461780480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813741223671013,
      "loss": 2.6118,
      "theoretical_loss": 3.524094078743505,
      "tokens_seen": 1461846016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028136409227683053,
      "loss": 2.4199,
      "theoretical_loss": 3.524080080416972,
      "tokens_seen": 1461911552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028135406218655966,
      "loss": 2.6473,
      "theoretical_loss": 3.524066082893655,
      "tokens_seen": 1461977088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813440320962889,
      "loss": 2.4552,
      "theoretical_loss": 3.5240520861734725,
      "tokens_seen": 1462042624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813340020060181,
      "loss": 2.7012,
      "theoretical_loss": 3.524038090256342,
      "tokens_seen": 1462108160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028132397191574725,
      "loss": 2.5372,
      "theoretical_loss": 3.524024095142181,
      "tokens_seen": 1462173696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028131394182547644,
      "loss": 2.5762,
      "theoretical_loss": 3.5240101008309086,
      "tokens_seen": 1462239232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813039117352056,
      "loss": 2.8869,
      "theoretical_loss": 3.523996107322442,
      "tokens_seen": 1462304768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812938816449348,
      "loss": 2.7317,
      "theoretical_loss": 3.5239821146166985,
      "tokens_seen": 1462370304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028128385155466403,
      "loss": 2.5509,
      "theoretical_loss": 3.5239681227135975,
      "tokens_seen": 1462435840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028127382146439316,
      "loss": 2.6265,
      "theoretical_loss": 3.5239541316130563,
      "tokens_seen": 1462501376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812637913741224,
      "loss": 2.6827,
      "theoretical_loss": 3.5239401413149927,
      "tokens_seen": 1462566912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812537612838516,
      "loss": 2.7779,
      "theoretical_loss": 3.523926151819325,
      "tokens_seen": 1462632448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028124373119358076,
      "loss": 2.5768,
      "theoretical_loss": 3.523912163125971,
      "tokens_seen": 1462697984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028123370110330994,
      "loss": 2.6157,
      "theoretical_loss": 3.5238981752348493,
      "tokens_seen": 1462763520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812236710130391,
      "loss": 2.8139,
      "theoretical_loss": 3.5238841881458773,
      "tokens_seen": 1462829056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812136409227683,
      "loss": 2.5889,
      "theoretical_loss": 3.5238702018589736,
      "tokens_seen": 1462894592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028120361083249754,
      "loss": 2.5345,
      "theoretical_loss": 3.523856216374056,
      "tokens_seen": 1462960128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028119358074222666,
      "loss": 2.6998,
      "theoretical_loss": 3.5238422316910425,
      "tokens_seen": 1463025664
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 807855,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4178123474121094,
      "objective/train/theoretical_loss": 3.523828247809851,
      "objective/train/tokens_used": 1483551200,
      "theoretical_loss": 3.523828247809851,
      "tokens_seen": 1463091200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811835506519559,
      "loss": 2.4975,
      "theoretical_loss": 3.523828247809851,
      "tokens_seen": 1463091200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281173520561685,
      "loss": 2.6831,
      "theoretical_loss": 3.5238142647304,
      "tokens_seen": 1463156736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028116349047141426,
      "loss": 2.7051,
      "theoretical_loss": 3.5238002824526076,
      "tokens_seen": 1463222272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028115346038114344,
      "loss": 2.4668,
      "theoretical_loss": 3.5237863009763917,
      "tokens_seen": 1463287808
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811434302908726,
      "loss": 2.5234,
      "theoretical_loss": 3.523772320301671,
      "tokens_seen": 1463353344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811334002006018,
      "loss": 2.4375,
      "theoretical_loss": 3.5237583404283637,
      "tokens_seen": 1463418880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281123370110331,
      "loss": 2.5769,
      "theoretical_loss": 3.523744361356387,
      "tokens_seen": 1463484416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028111334002006017,
      "loss": 2.6807,
      "theoretical_loss": 3.5237303830856597,
      "tokens_seen": 1463549952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811033099297894,
      "loss": 2.6226,
      "theoretical_loss": 3.5237164056160997,
      "tokens_seen": 1463615488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028109327983951853,
      "loss": 2.5343,
      "theoretical_loss": 3.523702428947626,
      "tokens_seen": 1463681024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028108324974924776,
      "loss": 2.6531,
      "theoretical_loss": 3.5236884530801555,
      "tokens_seen": 1463746560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028107321965897695,
      "loss": 2.683,
      "theoretical_loss": 3.523674478013608,
      "tokens_seen": 1463812096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810631895687061,
      "loss": 2.8166,
      "theoretical_loss": 3.5236605037479007,
      "tokens_seen": 1463877632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810531594784353,
      "loss": 2.494,
      "theoretical_loss": 3.523646530282952,
      "tokens_seen": 1463943168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810431293881645,
      "loss": 2.5355,
      "theoretical_loss": 3.52363255761868,
      "tokens_seen": 1464008704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028103309929789367,
      "loss": 2.395,
      "theoretical_loss": 3.5236185857550044,
      "tokens_seen": 1464074240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810230692076229,
      "loss": 2.5215,
      "theoretical_loss": 3.523604614691841,
      "tokens_seen": 1464139776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028101303911735203,
      "loss": 2.5591,
      "theoretical_loss": 3.5235906444291105,
      "tokens_seen": 1464205312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028100300902708127,
      "loss": 2.5624,
      "theoretical_loss": 3.52357667496673,
      "tokens_seen": 1464270848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002809929789368104,
      "loss": 2.5935,
      "theoretical_loss": 3.5235627063046175,
      "tokens_seen": 1464336384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028098294884653963,
      "loss": 2.8998,
      "theoretical_loss": 3.5235487384426927,
      "tokens_seen": 1464401920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002809729187562688,
      "loss": 2.5753,
      "theoretical_loss": 3.5235347713808727,
      "tokens_seen": 1464467456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280962888665998,
      "loss": 2.5725,
      "theoretical_loss": 3.523520805119076,
      "tokens_seen": 1464532992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028095285857572717,
      "loss": 2.7663,
      "theoretical_loss": 3.5235068396572213,
      "tokens_seen": 1464598528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028094282848545635,
      "loss": 2.6161,
      "theoretical_loss": 3.5234928749952275,
      "tokens_seen": 1464664064
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 808908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.759751319885254,
      "objective/train/theoretical_loss": 3.523478911133012,
      "objective/train/tokens_used": 1485189600,
      "theoretical_loss": 3.523478911133012,
      "tokens_seen": 1464729600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028093279839518553,
      "loss": 2.6224,
      "theoretical_loss": 3.523478911133012,
      "tokens_seen": 1464729600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028092276830491477,
      "loss": 2.7913,
      "theoretical_loss": 3.523464948070494,
      "tokens_seen": 1464795136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002809127382146439,
      "loss": 2.7087,
      "theoretical_loss": 3.5234509858075915,
      "tokens_seen": 1464860672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028090270812437313,
      "loss": 2.6517,
      "theoretical_loss": 3.5234370243442226,
      "tokens_seen": 1464926208
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808926780341023,
      "loss": 2.6858,
      "theoretical_loss": 3.5234230636803066,
      "tokens_seen": 1464991744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808826479438315,
      "loss": 2.6047,
      "theoretical_loss": 3.5234091038157613,
      "tokens_seen": 1465057280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808726178535607,
      "loss": 2.6421,
      "theoretical_loss": 3.5233951447505056,
      "tokens_seen": 1465122816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028086258776328986,
      "loss": 2.443,
      "theoretical_loss": 3.5233811864844577,
      "tokens_seen": 1465188352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028085255767301904,
      "loss": 2.8726,
      "theoretical_loss": 3.523367229017537,
      "tokens_seen": 1465253888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808425275827483,
      "loss": 2.7869,
      "theoretical_loss": 3.5233532723496603,
      "tokens_seen": 1465319424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808324974924774,
      "loss": 2.5656,
      "theoretical_loss": 3.523339316480747,
      "tokens_seen": 1465384960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028082246740220664,
      "loss": 2.5999,
      "theoretical_loss": 3.5233253614107163,
      "tokens_seen": 1465450496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028081243731193576,
      "loss": 2.4881,
      "theoretical_loss": 3.5233114071394853,
      "tokens_seen": 1465516032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280802407221665,
      "loss": 2.5154,
      "theoretical_loss": 3.5232974536669737,
      "tokens_seen": 1465581568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807923771313942,
      "loss": 2.5077,
      "theoretical_loss": 3.5232835009931005,
      "tokens_seen": 1465647104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028078234704112336,
      "loss": 2.6592,
      "theoretical_loss": 3.523269549117783,
      "tokens_seen": 1465712640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807723169508526,
      "loss": 2.4471,
      "theoretical_loss": 3.52325559804094,
      "tokens_seen": 1465778176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807622868605818,
      "loss": 2.5608,
      "theoretical_loss": 3.5232416477624904,
      "tokens_seen": 1465843712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028075225677031096,
      "loss": 2.4759,
      "theoretical_loss": 3.523227698282353,
      "tokens_seen": 1465909248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028074222668004014,
      "loss": 2.4707,
      "theoretical_loss": 3.5232137496004468,
      "tokens_seen": 1465974784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807321965897693,
      "loss": 2.5881,
      "theoretical_loss": 3.5231998017166894,
      "tokens_seen": 1466040320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807221664994985,
      "loss": 2.593,
      "theoretical_loss": 3.523185854631,
      "tokens_seen": 1466105856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028071213640922774,
      "loss": 2.7124,
      "theoretical_loss": 3.523171908343297,
      "tokens_seen": 1466171392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028070210631895686,
      "loss": 2.4851,
      "theoretical_loss": 3.5231579628535004,
      "tokens_seen": 1466236928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806920762286861,
      "loss": 2.6762,
      "theoretical_loss": 3.523144018161527,
      "tokens_seen": 1466302464
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 809565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5152714252471924,
      "objective/train/theoretical_loss": 3.523130074267296,
      "objective/train/tokens_used": 1486828000,
      "theoretical_loss": 3.523130074267296,
      "tokens_seen": 1466368000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806820461384152,
      "loss": 2.5798,
      "theoretical_loss": 3.523130074267296,
      "tokens_seen": 1466368000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028067201604814446,
      "loss": 2.6714,
      "theoretical_loss": 3.5231161311707266,
      "tokens_seen": 1466433536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028066198595787364,
      "loss": 2.7615,
      "theoretical_loss": 3.5231021888717375,
      "tokens_seen": 1466499072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806519558676028,
      "loss": 2.6979,
      "theoretical_loss": 3.5230882473702474,
      "tokens_seen": 1466564608
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280641925777332,
      "loss": 2.7179,
      "theoretical_loss": 3.5230743066661745,
      "tokens_seen": 1466630144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806318956870612,
      "loss": 2.4576,
      "theoretical_loss": 3.523060366759438,
      "tokens_seen": 1466695680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028062186559679037,
      "loss": 2.6031,
      "theoretical_loss": 3.523046427649957,
      "tokens_seen": 1466761216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806118355065196,
      "loss": 2.5807,
      "theoretical_loss": 3.52303248933765,
      "tokens_seen": 1466826752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028060180541624873,
      "loss": 2.7375,
      "theoretical_loss": 3.5230185518224357,
      "tokens_seen": 1466892288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028059177532597796,
      "loss": 2.6703,
      "theoretical_loss": 3.5230046151042327,
      "tokens_seen": 1466957824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028058174523570715,
      "loss": 2.5978,
      "theoretical_loss": 3.5229906791829606,
      "tokens_seen": 1467023360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805717151454363,
      "loss": 2.6568,
      "theoretical_loss": 3.522976744058537,
      "tokens_seen": 1467088896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805616850551655,
      "loss": 2.5362,
      "theoretical_loss": 3.522962809730882,
      "tokens_seen": 1467154432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805516549648947,
      "loss": 2.6053,
      "theoretical_loss": 3.522948876199913,
      "tokens_seen": 1467219968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028054162487462387,
      "loss": 2.448,
      "theoretical_loss": 3.5229349434655504,
      "tokens_seen": 1467285504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805315947843531,
      "loss": 2.7329,
      "theoretical_loss": 3.5229210115277128,
      "tokens_seen": 1467351040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028052156469408223,
      "loss": 2.649,
      "theoretical_loss": 3.522907080386318,
      "tokens_seen": 1467416576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028051153460381147,
      "loss": 2.5478,
      "theoretical_loss": 3.5228931500412863,
      "tokens_seen": 1467482112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805015045135406,
      "loss": 2.5681,
      "theoretical_loss": 3.522879220492536,
      "tokens_seen": 1467547648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028049147442326983,
      "loss": 2.5845,
      "theoretical_loss": 3.5228652917399854,
      "tokens_seen": 1467613184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280481444332999,
      "loss": 2.613,
      "theoretical_loss": 3.5228513637835537,
      "tokens_seen": 1467678720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804714142427282,
      "loss": 2.6921,
      "theoretical_loss": 3.522837436623161,
      "tokens_seen": 1467744256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028046138415245737,
      "loss": 2.5427,
      "theoretical_loss": 3.5228235102587253,
      "tokens_seen": 1467809792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028045135406218655,
      "loss": 2.6837,
      "theoretical_loss": 3.5228095846901653,
      "tokens_seen": 1467875328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028044132397191574,
      "loss": 2.7453,
      "theoretical_loss": 3.522795659917401,
      "tokens_seen": 1467940864
    },
    {
      "epoch": 0.44,
      "objective/train/docs_used": 810763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.383155345916748,
      "objective/train/theoretical_loss": 3.5227817359403506,
      "objective/train/tokens_used": 1488466400,
      "theoretical_loss": 3.5227817359403506,
      "tokens_seen": 1468006400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028043129388164497,
      "loss": 2.4052,
      "theoretical_loss": 3.5227817359403506,
      "tokens_seen": 1468006400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804212637913741,
      "loss": 2.7047,
      "theoretical_loss": 3.522767812758933,
      "tokens_seen": 1468071936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028041123370110333,
      "loss": 2.646,
      "theoretical_loss": 3.5227538903730675,
      "tokens_seen": 1468137472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804012036108325,
      "loss": 2.5549,
      "theoretical_loss": 3.5227399687826733,
      "tokens_seen": 1468203008
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803911735205617,
      "loss": 2.5508,
      "theoretical_loss": 3.5227260479876694,
      "tokens_seen": 1468268544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803811434302909,
      "loss": 2.561,
      "theoretical_loss": 3.5227121279879743,
      "tokens_seen": 1468334080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028037111334002006,
      "loss": 2.5237,
      "theoretical_loss": 3.522698208783508,
      "tokens_seen": 1468399616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028036108324974924,
      "loss": 2.4907,
      "theoretical_loss": 3.522684290374189,
      "tokens_seen": 1468465152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803510531594785,
      "loss": 2.5964,
      "theoretical_loss": 3.5226703727599364,
      "tokens_seen": 1468530688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803410230692076,
      "loss": 2.6129,
      "theoretical_loss": 3.5226564559406697,
      "tokens_seen": 1468596224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028033099297893684,
      "loss": 2.3882,
      "theoretical_loss": 3.5226425399163075,
      "tokens_seen": 1468661760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028032096288866596,
      "loss": 2.753,
      "theoretical_loss": 3.522628624686769,
      "tokens_seen": 1468727296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803109327983952,
      "loss": 2.5923,
      "theoretical_loss": 3.5226147102519736,
      "tokens_seen": 1468792832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803009027081244,
      "loss": 2.5226,
      "theoretical_loss": 3.5226007966118407,
      "tokens_seen": 1468858368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028029087261785356,
      "loss": 2.919,
      "theoretical_loss": 3.522586883766289,
      "tokens_seen": 1468923904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028028084252758274,
      "loss": 2.6947,
      "theoretical_loss": 3.5225729717152374,
      "tokens_seen": 1468989440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000280270812437312,
      "loss": 2.6534,
      "theoretical_loss": 3.5225590604586055,
      "tokens_seen": 1469054976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802607823470411,
      "loss": 2.5873,
      "theoretical_loss": 3.5225451499963127,
      "tokens_seen": 1469120512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028025075225677034,
      "loss": 2.7467,
      "theoretical_loss": 3.522531240328278,
      "tokens_seen": 1469186048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028024072216649947,
      "loss": 2.4201,
      "theoretical_loss": 3.5225173314544205,
      "tokens_seen": 1469251584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802306920762287,
      "loss": 2.6165,
      "theoretical_loss": 3.5225034233746593,
      "tokens_seen": 1469317120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802206619859579,
      "loss": 2.7084,
      "theoretical_loss": 3.5224895160889145,
      "tokens_seen": 1469382656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028021063189568706,
      "loss": 2.6769,
      "theoretical_loss": 3.5224756095971044,
      "tokens_seen": 1469448192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028020060180541624,
      "loss": 2.6497,
      "theoretical_loss": 3.5224617038991486,
      "tokens_seen": 1469513728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801905717151454,
      "loss": 2.6049,
      "theoretical_loss": 3.5224477989949663,
      "tokens_seen": 1469579264
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 811321,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9198386669158936,
      "objective/train/theoretical_loss": 3.522433894884477,
      "objective/train/tokens_used": 1490104800,
      "theoretical_loss": 3.522433894884477,
      "tokens_seen": 1469644800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801805416248746,
      "loss": 2.7733,
      "theoretical_loss": 3.522433894884477,
      "tokens_seen": 1469644800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028017051153460384,
      "loss": 2.8255,
      "theoretical_loss": 3.5224199915675998,
      "tokens_seen": 1469710336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028016048144433297,
      "loss": 2.5859,
      "theoretical_loss": 3.5224060890442543,
      "tokens_seen": 1469775872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801504513540622,
      "loss": 2.6958,
      "theoretical_loss": 3.5223921873143595,
      "tokens_seen": 1469841408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028014042126379133,
      "loss": 2.5037,
      "theoretical_loss": 3.522378286377835,
      "tokens_seen": 1469906944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028013039117352057,
      "loss": 2.6737,
      "theoretical_loss": 3.5223643862346,
      "tokens_seen": 1469972480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028012036108324975,
      "loss": 2.543,
      "theoretical_loss": 3.522350486884574,
      "tokens_seen": 1470038016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028011033099297893,
      "loss": 2.5958,
      "theoretical_loss": 3.522336588327676,
      "tokens_seen": 1470103552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801003009027081,
      "loss": 2.667,
      "theoretical_loss": 3.5223226905638256,
      "tokens_seen": 1470169088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028009027081243735,
      "loss": 2.7794,
      "theoretical_loss": 3.522308793592943,
      "tokens_seen": 1470234624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028008024072216647,
      "loss": 2.7077,
      "theoretical_loss": 3.5222948974149464,
      "tokens_seen": 1470300160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002800702106318957,
      "loss": 2.7373,
      "theoretical_loss": 3.5222810020297555,
      "tokens_seen": 1470365696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028006018054162483,
      "loss": 2.7383,
      "theoretical_loss": 3.52226710743729,
      "tokens_seen": 1470431232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028005015045135407,
      "loss": 2.52,
      "theoretical_loss": 3.5222532136374696,
      "tokens_seen": 1470496768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028004012036108325,
      "loss": 2.6957,
      "theoretical_loss": 3.522239320630213,
      "tokens_seen": 1470562304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028003009027081243,
      "loss": 2.3979,
      "theoretical_loss": 3.5222254284154406,
      "tokens_seen": 1470627840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028002006018054167,
      "loss": 2.6116,
      "theoretical_loss": 3.522211536993071,
      "tokens_seen": 1470693376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002800100300902708,
      "loss": 2.6236,
      "theoretical_loss": 3.5221976463630242,
      "tokens_seen": 1470758912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028000000000000003,
      "loss": 2.6138,
      "theoretical_loss": 3.52218375652522,
      "tokens_seen": 1470824448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799899699097292,
      "loss": 2.705,
      "theoretical_loss": 3.522169867479577,
      "tokens_seen": 1470889984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799799398194584,
      "loss": 2.7807,
      "theoretical_loss": 3.5221559792260155,
      "tokens_seen": 1470955520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799699097291876,
      "loss": 2.619,
      "theoretical_loss": 3.5221420917644544,
      "tokens_seen": 1471021056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027995987963891675,
      "loss": 2.547,
      "theoretical_loss": 3.5221282050948144,
      "tokens_seen": 1471086592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027994984954864594,
      "loss": 2.7031,
      "theoretical_loss": 3.5221143192170135,
      "tokens_seen": 1471152128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027993981945837517,
      "loss": 2.6999,
      "theoretical_loss": 3.5221004341309725,
      "tokens_seen": 1471217664
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 812410,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.735546588897705,
      "objective/train/theoretical_loss": 3.5220865498366107,
      "objective/train/tokens_used": 1491743200,
      "theoretical_loss": 3.5220865498366107,
      "tokens_seen": 1471283200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799297893681043,
      "loss": 2.5636,
      "theoretical_loss": 3.5220865498366107,
      "tokens_seen": 1471283200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027991975927783353,
      "loss": 2.4444,
      "theoretical_loss": 3.5220726663338473,
      "tokens_seen": 1471348736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799097291875627,
      "loss": 2.7656,
      "theoretical_loss": 3.522058783622602,
      "tokens_seen": 1471414272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798996990972919,
      "loss": 2.4004,
      "theoretical_loss": 3.522044901702795,
      "tokens_seen": 1471479808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798896690070211,
      "loss": 2.4757,
      "theoretical_loss": 3.5220310205743455,
      "tokens_seen": 1471545344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027987963891675026,
      "loss": 2.6364,
      "theoretical_loss": 3.522017140237173,
      "tokens_seen": 1471610880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027986960882647944,
      "loss": 2.8141,
      "theoretical_loss": 3.5220032606911973,
      "tokens_seen": 1471676416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798595787362087,
      "loss": 2.8483,
      "theoretical_loss": 3.521989381936338,
      "tokens_seen": 1471741952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798495486459378,
      "loss": 2.6073,
      "theoretical_loss": 3.5219755039725156,
      "tokens_seen": 1471807488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027983951855566704,
      "loss": 2.4337,
      "theoretical_loss": 3.5219616267996487,
      "tokens_seen": 1471873024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027982948846539616,
      "loss": 2.6282,
      "theoretical_loss": 3.5219477504176573,
      "tokens_seen": 1471938560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798194583751254,
      "loss": 2.7561,
      "theoretical_loss": 3.5219338748264613,
      "tokens_seen": 1472004096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798094282848546,
      "loss": 2.663,
      "theoretical_loss": 3.5219200000259803,
      "tokens_seen": 1472069632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027979939819458376,
      "loss": 2.525,
      "theoretical_loss": 3.521906126016134,
      "tokens_seen": 1472135168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027978936810431294,
      "loss": 2.8497,
      "theoretical_loss": 3.5218922527968424,
      "tokens_seen": 1472200704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797793380140422,
      "loss": 2.3666,
      "theoretical_loss": 3.521878380368025,
      "tokens_seen": 1472266240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797693079237713,
      "loss": 2.7274,
      "theoretical_loss": 3.521864508729602,
      "tokens_seen": 1472331776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027975927783350054,
      "loss": 2.6469,
      "theoretical_loss": 3.521850637881492,
      "tokens_seen": 1472397312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027974924774322967,
      "loss": 2.5282,
      "theoretical_loss": 3.5218367678236167,
      "tokens_seen": 1472462848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797392176529589,
      "loss": 2.7885,
      "theoretical_loss": 3.521822898555895,
      "tokens_seen": 1472528384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797291875626881,
      "loss": 2.5835,
      "theoretical_loss": 3.521809030078246,
      "tokens_seen": 1472593920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027971915747241726,
      "loss": 2.6981,
      "theoretical_loss": 3.52179516239059,
      "tokens_seen": 1472659456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027970912738214644,
      "loss": 2.6311,
      "theoretical_loss": 3.521781295492847,
      "tokens_seen": 1472724992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796990972918756,
      "loss": 2.5466,
      "theoretical_loss": 3.521767429384937,
      "tokens_seen": 1472790528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796890672016048,
      "loss": 2.425,
      "theoretical_loss": 3.5217535640667794,
      "tokens_seen": 1472856064
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 813075,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4556291103363037,
      "objective/train/theoretical_loss": 3.521739699538295,
      "objective/train/tokens_used": 1493381600,
      "theoretical_loss": 3.521739699538295,
      "tokens_seen": 1472921600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027967903711133404,
      "loss": 2.5714,
      "theoretical_loss": 3.521739699538295,
      "tokens_seen": 1472921600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027966900702106317,
      "loss": 2.6501,
      "theoretical_loss": 3.521725835799403,
      "tokens_seen": 1472987136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796589769307924,
      "loss": 2.5784,
      "theoretical_loss": 3.521711972850023,
      "tokens_seen": 1473052672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027964894684052153,
      "loss": 2.7543,
      "theoretical_loss": 3.5216981106900755,
      "tokens_seen": 1473118208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027963891675025077,
      "loss": 2.3898,
      "theoretical_loss": 3.52168424931948,
      "tokens_seen": 1473183744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027962888665997995,
      "loss": 2.6985,
      "theoretical_loss": 3.5216703887381575,
      "tokens_seen": 1473249280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027961885656970913,
      "loss": 2.5342,
      "theoretical_loss": 3.5216565289460267,
      "tokens_seen": 1473314816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796088264794383,
      "loss": 2.4191,
      "theoretical_loss": 3.5216426699430077,
      "tokens_seen": 1473380352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027959879638916755,
      "loss": 2.4761,
      "theoretical_loss": 3.5216288117290206,
      "tokens_seen": 1473445888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027958876629889667,
      "loss": 2.6122,
      "theoretical_loss": 3.521614954303986,
      "tokens_seen": 1473511424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795787362086259,
      "loss": 2.7318,
      "theoretical_loss": 3.5216010976678236,
      "tokens_seen": 1473576960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027956870611835503,
      "loss": 2.7646,
      "theoretical_loss": 3.521587241820453,
      "tokens_seen": 1473642496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027955867602808427,
      "loss": 2.5787,
      "theoretical_loss": 3.521573386761794,
      "tokens_seen": 1473708032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027954864593781345,
      "loss": 2.6957,
      "theoretical_loss": 3.521559532491768,
      "tokens_seen": 1473773568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027953861584754263,
      "loss": 2.5986,
      "theoretical_loss": 3.5215456790102935,
      "tokens_seen": 1473839104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795285857572718,
      "loss": 2.7141,
      "theoretical_loss": 3.5215318263172914,
      "tokens_seen": 1473904640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279518555667001,
      "loss": 2.8128,
      "theoretical_loss": 3.5215179744126814,
      "tokens_seen": 1473970176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795085255767302,
      "loss": 2.8406,
      "theoretical_loss": 3.521504123296384,
      "tokens_seen": 1474035712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794984954864594,
      "loss": 2.5559,
      "theoretical_loss": 3.521490272968319,
      "tokens_seen": 1474101248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027948846539618854,
      "loss": 2.5268,
      "theoretical_loss": 3.521476423428407,
      "tokens_seen": 1474166784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794784353059178,
      "loss": 2.6032,
      "theoretical_loss": 3.5214625746765673,
      "tokens_seen": 1474232320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794684052156469,
      "loss": 2.6824,
      "theoretical_loss": 3.5214487267127206,
      "tokens_seen": 1474297856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027945837512537614,
      "loss": 2.5321,
      "theoretical_loss": 3.5214348795367862,
      "tokens_seen": 1474363392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794483450351053,
      "loss": 2.6556,
      "theoretical_loss": 3.5214210331486853,
      "tokens_seen": 1474428928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794383149448345,
      "loss": 2.6213,
      "theoretical_loss": 3.5214071875483377,
      "tokens_seen": 1474494464
    },
    {
      "debugging/Self-BLEU-5": 0.4457480258618696,
      "debugging/distinct-1-grams": 0.780869708352066,
      "debugging/distinct-2-grams": 0.9550381307455209,
      "debugging/entropy-1-grams": 5.691454067736992,
      "debugging/entropy-2-grams": 6.3380960139729146,
      "debugging/length": 545.0,
      "debugging/num_segments": 9,
      "epoch": 0.45,
      "objective/train/docs_used": 814377,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.860199451446533,
      "objective/train/theoretical_loss": 3.5213933427356636,
      "objective/train/tokens_used": 1495020000,
      "theoretical_loss": 3.5213933427356636,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794282848545637,
      "loss": 2.5364,
      "theoretical_loss": 3.5213933427356636,
      "tokens_seen": 1474560000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794182547642929,
      "loss": 2.5116,
      "theoretical_loss": 3.5213794987105826,
      "tokens_seen": 1474625536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027940822467402204,
      "loss": 2.4481,
      "theoretical_loss": 3.521365655473016,
      "tokens_seen": 1474691072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793981945837513,
      "loss": 2.3796,
      "theoretical_loss": 3.5213518130228834,
      "tokens_seen": 1474756608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793881644934804,
      "loss": 2.6261,
      "theoretical_loss": 3.5213379713601047,
      "tokens_seen": 1474822144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027937813440320964,
      "loss": 2.7711,
      "theoretical_loss": 3.521324130484601,
      "tokens_seen": 1474887680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793681043129388,
      "loss": 2.5423,
      "theoretical_loss": 3.521310290396291,
      "tokens_seen": 1474953216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279358074222668,
      "loss": 2.548,
      "theoretical_loss": 3.521296451095097,
      "tokens_seen": 1475018752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793480441323972,
      "loss": 2.4568,
      "theoretical_loss": 3.5212826125809373,
      "tokens_seen": 1475084288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027933801404212636,
      "loss": 2.6257,
      "theoretical_loss": 3.521268774853734,
      "tokens_seen": 1475149824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027932798395185554,
      "loss": 2.7034,
      "theoretical_loss": 3.521254937913406,
      "tokens_seen": 1475215360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793179538615848,
      "loss": 2.5689,
      "theoretical_loss": 3.521241101759874,
      "tokens_seen": 1475280896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793079237713139,
      "loss": 2.5344,
      "theoretical_loss": 3.521227266393059,
      "tokens_seen": 1475346432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027929789368104314,
      "loss": 2.875,
      "theoretical_loss": 3.52121343181288,
      "tokens_seen": 1475411968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027928786359077227,
      "loss": 2.5586,
      "theoretical_loss": 3.5211995980192583,
      "tokens_seen": 1475477504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792778335005015,
      "loss": 2.6094,
      "theoretical_loss": 3.521185765012114,
      "tokens_seen": 1475543040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027926780341023074,
      "loss": 2.6754,
      "theoretical_loss": 3.5211719327913675,
      "tokens_seen": 1475608576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027925777331995987,
      "loss": 2.6065,
      "theoretical_loss": 3.5211581013569395,
      "tokens_seen": 1475674112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792477432296891,
      "loss": 2.6373,
      "theoretical_loss": 3.5211442707087497,
      "tokens_seen": 1475739648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792377131394183,
      "loss": 2.6275,
      "theoretical_loss": 3.521130440846719,
      "tokens_seen": 1475805184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027922768304914746,
      "loss": 2.5502,
      "theoretical_loss": 3.5211166117707675,
      "tokens_seen": 1475870720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027921765295887664,
      "loss": 2.5877,
      "theoretical_loss": 3.5211027834808157,
      "tokens_seen": 1475936256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792076228686058,
      "loss": 2.6401,
      "theoretical_loss": 3.5210889559767837,
      "tokens_seen": 1476001792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279197592778335,
      "loss": 2.5757,
      "theoretical_loss": 3.521075129258593,
      "tokens_seen": 1476067328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027918756268806424,
      "loss": 2.6731,
      "theoretical_loss": 3.521061303326163,
      "tokens_seen": 1476132864
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 814873,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.263564109802246,
      "objective/train/theoretical_loss": 3.5210474781794145,
      "objective/train/tokens_used": 1496658400,
      "theoretical_loss": 3.5210474781794145,
      "tokens_seen": 1476198400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027917753259779337,
      "loss": 2.4547,
      "theoretical_loss": 3.5210474781794145,
      "tokens_seen": 1476198400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791675025075226,
      "loss": 2.5981,
      "theoretical_loss": 3.521033653818268,
      "tokens_seen": 1476263936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027915747241725173,
      "loss": 2.7949,
      "theoretical_loss": 3.5210198302426443,
      "tokens_seen": 1476329472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027914744232698097,
      "loss": 2.5497,
      "theoretical_loss": 3.5210060074524634,
      "tokens_seen": 1476395008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027913741223671015,
      "loss": 2.8282,
      "theoretical_loss": 3.5209921854476462,
      "tokens_seen": 1476460544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027912738214643933,
      "loss": 2.4958,
      "theoretical_loss": 3.5209783642281125,
      "tokens_seen": 1476526080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791173520561685,
      "loss": 2.7576,
      "theoretical_loss": 3.5209645437937835,
      "tokens_seen": 1476591616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791173520561685,
      "loss": 2.5902,
      "theoretical_loss": 3.52095072414458,
      "tokens_seen": 1476657152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027910732196589775,
      "loss": 2.5488,
      "theoretical_loss": 3.520936905280422,
      "tokens_seen": 1476722688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027909729187562687,
      "loss": 2.5285,
      "theoretical_loss": 3.52092308720123,
      "tokens_seen": 1476788224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790872617853561,
      "loss": 2.577,
      "theoretical_loss": 3.520909269906925,
      "tokens_seen": 1476853760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027907723169508523,
      "loss": 2.5842,
      "theoretical_loss": 3.520895453397427,
      "tokens_seen": 1476919296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027906720160481447,
      "loss": 2.6671,
      "theoretical_loss": 3.520881637672657,
      "tokens_seen": 1476984832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027905717151454365,
      "loss": 2.5175,
      "theoretical_loss": 3.5208678227325363,
      "tokens_seen": 1477050368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027904714142427283,
      "loss": 2.699,
      "theoretical_loss": 3.5208540085769844,
      "tokens_seen": 1477115904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279037111334002,
      "loss": 2.8946,
      "theoretical_loss": 3.5208401952059223,
      "tokens_seen": 1477181440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790270812437312,
      "loss": 2.6572,
      "theoretical_loss": 3.5208263826192705,
      "tokens_seen": 1477246976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790170511534604,
      "loss": 2.5727,
      "theoretical_loss": 3.5208125708169504,
      "tokens_seen": 1477312512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790070210631896,
      "loss": 2.7133,
      "theoretical_loss": 3.5207987597988817,
      "tokens_seen": 1477378048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027899699097291874,
      "loss": 2.6664,
      "theoretical_loss": 3.5207849495649857,
      "tokens_seen": 1477443584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278986960882648,
      "loss": 2.7478,
      "theoretical_loss": 3.520771140115183,
      "tokens_seen": 1477509120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789769307923771,
      "loss": 2.5194,
      "theoretical_loss": 3.520757331449394,
      "tokens_seen": 1477574656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027896690070210634,
      "loss": 2.5503,
      "theoretical_loss": 3.52074352356754,
      "tokens_seen": 1477640192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789568706118355,
      "loss": 2.3763,
      "theoretical_loss": 3.520729716469541,
      "tokens_seen": 1477705728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789468405215647,
      "loss": 2.7193,
      "theoretical_loss": 3.5207159101553183,
      "tokens_seen": 1477771264
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 816004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6182334423065186,
      "objective/train/theoretical_loss": 3.5207021046247924,
      "objective/train/tokens_used": 1498296800,
      "theoretical_loss": 3.5207021046247924,
      "tokens_seen": 1477836800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789368104312939,
      "loss": 2.7106,
      "theoretical_loss": 3.5207021046247924,
      "tokens_seen": 1477836800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789267803410231,
      "loss": 2.8983,
      "theoretical_loss": 3.520688299877884,
      "tokens_seen": 1477902336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027891675025075224,
      "loss": 2.7539,
      "theoretical_loss": 3.5206744959145144,
      "tokens_seen": 1477967872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789067201604815,
      "loss": 2.493,
      "theoretical_loss": 3.520660692734604,
      "tokens_seen": 1478033408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788966900702106,
      "loss": 2.607,
      "theoretical_loss": 3.520646890338073,
      "tokens_seen": 1478098944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027888665997993984,
      "loss": 2.7922,
      "theoretical_loss": 3.5206330887248436,
      "tokens_seen": 1478164480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278876629889669,
      "loss": 2.8261,
      "theoretical_loss": 3.520619287894835,
      "tokens_seen": 1478230016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788665997993982,
      "loss": 2.7453,
      "theoretical_loss": 3.520605487847969,
      "tokens_seen": 1478295552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788565697091274,
      "loss": 2.5752,
      "theoretical_loss": 3.520591688584167,
      "tokens_seen": 1478361088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027884653961885656,
      "loss": 2.5868,
      "theoretical_loss": 3.5205778901033487,
      "tokens_seen": 1478426624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027883650952858574,
      "loss": 2.3969,
      "theoretical_loss": 3.5205640924054356,
      "tokens_seen": 1478492160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278826479438315,
      "loss": 2.6064,
      "theoretical_loss": 3.520550295490348,
      "tokens_seen": 1478557696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788164493480441,
      "loss": 2.6417,
      "theoretical_loss": 3.5205364993580073,
      "tokens_seen": 1478623232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027880641925777334,
      "loss": 2.6554,
      "theoretical_loss": 3.520522704008334,
      "tokens_seen": 1478688768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027879638916750247,
      "loss": 2.6519,
      "theoretical_loss": 3.52050890944125,
      "tokens_seen": 1478754304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787863590772317,
      "loss": 2.6336,
      "theoretical_loss": 3.5204951156566753,
      "tokens_seen": 1478819840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787763289869609,
      "loss": 2.4454,
      "theoretical_loss": 3.520481322654531,
      "tokens_seen": 1478885376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027876629889669007,
      "loss": 2.6013,
      "theoretical_loss": 3.520467530434738,
      "tokens_seen": 1478950912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027875626880641925,
      "loss": 2.6672,
      "theoretical_loss": 3.5204537389972175,
      "tokens_seen": 1479016448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787462387161485,
      "loss": 2.3893,
      "theoretical_loss": 3.5204399483418904,
      "tokens_seen": 1479081984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787362086258776,
      "loss": 2.6592,
      "theoretical_loss": 3.520426158468678,
      "tokens_seen": 1479147520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027872617853560685,
      "loss": 2.6974,
      "theoretical_loss": 3.5204123693775,
      "tokens_seen": 1479213056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027871614844533597,
      "loss": 2.4663,
      "theoretical_loss": 3.520398581068279,
      "tokens_seen": 1479278592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787061183550652,
      "loss": 2.5473,
      "theoretical_loss": 3.520384793540935,
      "tokens_seen": 1479344128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786960882647944,
      "loss": 2.5937,
      "theoretical_loss": 3.5203710067953895,
      "tokens_seen": 1479409664
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 816537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.965928077697754,
      "objective/train/theoretical_loss": 3.5203572208315634,
      "objective/train/tokens_used": 1499935200,
      "theoretical_loss": 3.5203572208315634,
      "tokens_seen": 1479475200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027868605817452357,
      "loss": 2.6058,
      "theoretical_loss": 3.5203572208315634,
      "tokens_seen": 1479475200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027867602808425275,
      "loss": 2.5134,
      "theoretical_loss": 3.520343435649378,
      "tokens_seen": 1479540736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027866599799398193,
      "loss": 2.7719,
      "theoretical_loss": 3.520329651248754,
      "tokens_seen": 1479606272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786559679037111,
      "loss": 2.7405,
      "theoretical_loss": 3.5203158676296127,
      "tokens_seen": 1479671808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027864593781344035,
      "loss": 2.5804,
      "theoretical_loss": 3.520302084791875,
      "tokens_seen": 1479737344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786359077231695,
      "loss": 2.5212,
      "theoretical_loss": 3.5202883027354614,
      "tokens_seen": 1479802880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786258776328987,
      "loss": 2.7162,
      "theoretical_loss": 3.5202745214602946,
      "tokens_seen": 1479868416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786158475426279,
      "loss": 2.5461,
      "theoretical_loss": 3.5202607409662945,
      "tokens_seen": 1479933952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027860581745235707,
      "loss": 2.6608,
      "theoretical_loss": 3.520246961253383,
      "tokens_seen": 1479999488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027859578736208625,
      "loss": 2.5891,
      "theoretical_loss": 3.52023318232148,
      "tokens_seen": 1480065024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027858575727181544,
      "loss": 2.6021,
      "theoretical_loss": 3.5202194041705077,
      "tokens_seen": 1480130560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785757271815446,
      "loss": 2.6964,
      "theoretical_loss": 3.5202056268003874,
      "tokens_seen": 1480196096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027856569709127385,
      "loss": 2.9855,
      "theoretical_loss": 3.520191850211039,
      "tokens_seen": 1480261632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278555667001003,
      "loss": 2.6498,
      "theoretical_loss": 3.520178074402385,
      "tokens_seen": 1480327168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785456369107322,
      "loss": 2.5365,
      "theoretical_loss": 3.520164299374347,
      "tokens_seen": 1480392704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027853560682046134,
      "loss": 2.6188,
      "theoretical_loss": 3.5201505251268443,
      "tokens_seen": 1480458240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785255767301906,
      "loss": 2.638,
      "theoretical_loss": 3.5201367516597992,
      "tokens_seen": 1480523776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785155466399198,
      "loss": 2.7699,
      "theoretical_loss": 3.520122978973134,
      "tokens_seen": 1480589312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027850551654964894,
      "loss": 2.7147,
      "theoretical_loss": 3.520109207066768,
      "tokens_seen": 1480654848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784954864593782,
      "loss": 2.5845,
      "theoretical_loss": 3.5200954359406236,
      "tokens_seen": 1480720384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784854563691073,
      "loss": 2.6269,
      "theoretical_loss": 3.520081665594622,
      "tokens_seen": 1480785920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027847542627883654,
      "loss": 2.4266,
      "theoretical_loss": 3.5200678960286838,
      "tokens_seen": 1480851456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784653961885657,
      "loss": 2.4685,
      "theoretical_loss": 3.520054127242731,
      "tokens_seen": 1480916992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784553660982949,
      "loss": 2.7099,
      "theoretical_loss": 3.5200403592366847,
      "tokens_seen": 1480982528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784453360080241,
      "loss": 2.3486,
      "theoretical_loss": 3.520026592010466,
      "tokens_seen": 1481048064
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 817599,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7796854972839355,
      "objective/train/theoretical_loss": 3.520012825563997,
      "objective/train/tokens_used": 1501573600,
      "theoretical_loss": 3.520012825563997,
      "tokens_seen": 1481113600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784353059177533,
      "loss": 2.6445,
      "theoretical_loss": 3.520012825563997,
      "tokens_seen": 1481113600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027842527582748244,
      "loss": 2.7275,
      "theoretical_loss": 3.5199990598971977,
      "tokens_seen": 1481179136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784152457372117,
      "loss": 2.559,
      "theoretical_loss": 3.519985295009991,
      "tokens_seen": 1481244672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784052156469408,
      "loss": 2.6555,
      "theoretical_loss": 3.519971530902297,
      "tokens_seen": 1481310208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027839518555667004,
      "loss": 2.6921,
      "theoretical_loss": 3.519957767574038,
      "tokens_seen": 1481375744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783851554663992,
      "loss": 2.8146,
      "theoretical_loss": 3.5199440050251343,
      "tokens_seen": 1481441280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783751253761284,
      "loss": 2.5935,
      "theoretical_loss": 3.519930243255508,
      "tokens_seen": 1481506816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783650952858576,
      "loss": 2.5259,
      "theoretical_loss": 3.5199164822650806,
      "tokens_seen": 1481572352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027835506519558676,
      "loss": 2.454,
      "theoretical_loss": 3.5199027220537737,
      "tokens_seen": 1481637888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027834503510531594,
      "loss": 2.717,
      "theoretical_loss": 3.5198889626215077,
      "tokens_seen": 1481703424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783350050150452,
      "loss": 2.5306,
      "theoretical_loss": 3.5198752039682057,
      "tokens_seen": 1481768960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783249749247743,
      "loss": 2.5441,
      "theoretical_loss": 3.5198614460937874,
      "tokens_seen": 1481834496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027831494483450354,
      "loss": 2.5472,
      "theoretical_loss": 3.5198476889981754,
      "tokens_seen": 1481900032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027830491474423267,
      "loss": 2.6856,
      "theoretical_loss": 3.51983393268129,
      "tokens_seen": 1481965568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782948846539619,
      "loss": 2.4332,
      "theoretical_loss": 3.519820177143054,
      "tokens_seen": 1482031104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782848545636911,
      "loss": 2.7552,
      "theoretical_loss": 3.5198064223833887,
      "tokens_seen": 1482096640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027827482447342027,
      "loss": 2.5342,
      "theoretical_loss": 3.519792668402215,
      "tokens_seen": 1482162176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027826479438314945,
      "loss": 2.5159,
      "theoretical_loss": 3.5197789151994545,
      "tokens_seen": 1482227712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782547642928787,
      "loss": 2.5585,
      "theoretical_loss": 3.5197651627750295,
      "tokens_seen": 1482293248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782447342026078,
      "loss": 2.5253,
      "theoretical_loss": 3.5197514111288606,
      "tokens_seen": 1482358784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027823470411233705,
      "loss": 2.6442,
      "theoretical_loss": 3.5197376602608696,
      "tokens_seen": 1482424320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027822467402206617,
      "loss": 2.7405,
      "theoretical_loss": 3.5197239101709785,
      "tokens_seen": 1482489856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782146439317954,
      "loss": 2.7626,
      "theoretical_loss": 3.5197101608591086,
      "tokens_seen": 1482555392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782046138415246,
      "loss": 2.4868,
      "theoretical_loss": 3.5196964123251813,
      "tokens_seen": 1482620928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027819458375125377,
      "loss": 2.5604,
      "theoretical_loss": 3.5196826645691184,
      "tokens_seen": 1482686464
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 818835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.735487937927246,
      "objective/train/theoretical_loss": 3.5196689175908418,
      "objective/train/tokens_used": 1503212000,
      "theoretical_loss": 3.5196689175908418,
      "tokens_seen": 1482752000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027818455366098295,
      "loss": 2.6771,
      "theoretical_loss": 3.5196689175908418,
      "tokens_seen": 1482752000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027817452357071213,
      "loss": 2.4758,
      "theoretical_loss": 3.5196551713902724,
      "tokens_seen": 1482817536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781644934804413,
      "loss": 2.6044,
      "theoretical_loss": 3.5196414259673325,
      "tokens_seen": 1482883072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027815446339017055,
      "loss": 2.6185,
      "theoretical_loss": 3.519627681321943,
      "tokens_seen": 1482948608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781444332998997,
      "loss": 2.6867,
      "theoretical_loss": 3.519613937454026,
      "tokens_seen": 1483014144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781344032096289,
      "loss": 2.7563,
      "theoretical_loss": 3.5196001943635036,
      "tokens_seen": 1483079680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781243731193581,
      "loss": 2.7083,
      "theoretical_loss": 3.5195864520502975,
      "tokens_seen": 1483145216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781143430290873,
      "loss": 2.4394,
      "theoretical_loss": 3.5195727105143284,
      "tokens_seen": 1483210752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027810431293881645,
      "loss": 2.5391,
      "theoretical_loss": 3.5195589697555185,
      "tokens_seen": 1483276288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027809428284854564,
      "loss": 2.72,
      "theoretical_loss": 3.51954522977379,
      "tokens_seen": 1483341824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780842527582748,
      "loss": 2.7018,
      "theoretical_loss": 3.5195314905690642,
      "tokens_seen": 1483407360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027807422266800405,
      "loss": 2.66,
      "theoretical_loss": 3.519517752141262,
      "tokens_seen": 1483472896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780641925777332,
      "loss": 2.6427,
      "theoretical_loss": 3.519504014490307,
      "tokens_seen": 1483538432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780541624874624,
      "loss": 2.7975,
      "theoretical_loss": 3.5194902776161197,
      "tokens_seen": 1483603968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027804413239719154,
      "loss": 2.6736,
      "theoretical_loss": 3.519476541518622,
      "tokens_seen": 1483669504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780341023069208,
      "loss": 2.5769,
      "theoretical_loss": 3.5194628061977355,
      "tokens_seen": 1483735040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027802407221664996,
      "loss": 2.748,
      "theoretical_loss": 3.5194490716533826,
      "tokens_seen": 1483800576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027801404212637914,
      "loss": 2.4636,
      "theoretical_loss": 3.5194353378854846,
      "tokens_seen": 1483866112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780040120361083,
      "loss": 2.6612,
      "theoretical_loss": 3.5194216048939637,
      "tokens_seen": 1483931648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779939819458375,
      "loss": 2.4891,
      "theoretical_loss": 3.519407872678742,
      "tokens_seen": 1483997184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779839518555667,
      "loss": 2.8148,
      "theoretical_loss": 3.51939414123974,
      "tokens_seen": 1484062720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779739217652959,
      "loss": 2.8085,
      "theoretical_loss": 3.519380410576881,
      "tokens_seen": 1484128256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027796389167502504,
      "loss": 2.6915,
      "theoretical_loss": 3.5193666806900863,
      "tokens_seen": 1484193792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779538615847543,
      "loss": 2.613,
      "theoretical_loss": 3.519352951579277,
      "tokens_seen": 1484259328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027794383149448346,
      "loss": 2.4977,
      "theoretical_loss": 3.5193392232443763,
      "tokens_seen": 1484324864
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 819535,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2024571895599365,
      "objective/train/theoretical_loss": 3.519325495685306,
      "objective/train/tokens_used": 1504850400,
      "theoretical_loss": 3.519325495685306,
      "tokens_seen": 1484390400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027793380140421264,
      "loss": 2.6604,
      "theoretical_loss": 3.519325495685306,
      "tokens_seen": 1484390400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779237713139418,
      "loss": 2.6439,
      "theoretical_loss": 3.519311768901987,
      "tokens_seen": 1484455936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277913741223671,
      "loss": 2.5186,
      "theoretical_loss": 3.5192980428943414,
      "tokens_seen": 1484521472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779037111334002,
      "loss": 2.6832,
      "theoretical_loss": 3.5192843176622923,
      "tokens_seen": 1484587008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778936810431294,
      "loss": 2.6472,
      "theoretical_loss": 3.5192705932057606,
      "tokens_seen": 1484652544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027788365095285855,
      "loss": 2.66,
      "theoretical_loss": 3.5192568695246678,
      "tokens_seen": 1484718080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778736208625878,
      "loss": 2.6594,
      "theoretical_loss": 3.5192431466189373,
      "tokens_seen": 1484783616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778635907723169,
      "loss": 2.629,
      "theoretical_loss": 3.51922942448849,
      "tokens_seen": 1484849152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027785356068204614,
      "loss": 2.7417,
      "theoretical_loss": 3.5192157031332485,
      "tokens_seen": 1484914688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778435305917753,
      "loss": 2.7664,
      "theoretical_loss": 3.519201982553134,
      "tokens_seen": 1484980224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778335005015045,
      "loss": 2.6715,
      "theoretical_loss": 3.5191882627480693,
      "tokens_seen": 1485045760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778234704112337,
      "loss": 2.5785,
      "theoretical_loss": 3.519174543717976,
      "tokens_seen": 1485111296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027781344032096287,
      "loss": 2.6659,
      "theoretical_loss": 3.5191608254627766,
      "tokens_seen": 1485176832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027780341023069205,
      "loss": 2.5638,
      "theoretical_loss": 3.519147107982392,
      "tokens_seen": 1485242368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777933801404213,
      "loss": 2.8271,
      "theoretical_loss": 3.5191333912767457,
      "tokens_seen": 1485307904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777833500501504,
      "loss": 2.8528,
      "theoretical_loss": 3.5191196753457588,
      "tokens_seen": 1485373440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027777331995987965,
      "loss": 2.5073,
      "theoretical_loss": 3.519105960189354,
      "tokens_seen": 1485438976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777632898696089,
      "loss": 2.5813,
      "theoretical_loss": 3.5190922458074527,
      "tokens_seen": 1485504512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277753259779338,
      "loss": 2.5409,
      "theoretical_loss": 3.519078532199978,
      "tokens_seen": 1485570048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027774322968906725,
      "loss": 2.7484,
      "theoretical_loss": 3.51906481936685,
      "tokens_seen": 1485635584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027773319959879637,
      "loss": 2.4342,
      "theoretical_loss": 3.5190511073079938,
      "tokens_seen": 1485701120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777231695085256,
      "loss": 2.5802,
      "theoretical_loss": 3.5190373960233288,
      "tokens_seen": 1485766656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777131394182548,
      "loss": 2.5896,
      "theoretical_loss": 3.5190236855127788,
      "tokens_seen": 1485832192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027770310932798397,
      "loss": 2.7993,
      "theoretical_loss": 3.519009975776265,
      "tokens_seen": 1485897728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027769307923771315,
      "loss": 2.6962,
      "theoretical_loss": 3.5189962668137102,
      "tokens_seen": 1485963264
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 820839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.42441987991333,
      "objective/train/theoretical_loss": 3.5189825586250363,
      "objective/train/tokens_used": 1506488800,
      "theoretical_loss": 3.5189825586250363,
      "tokens_seen": 1486028800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027768304914744233,
      "loss": 2.7172,
      "theoretical_loss": 3.5189825586250363,
      "tokens_seen": 1486028800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776730190571715,
      "loss": 2.5732,
      "theoretical_loss": 3.5189688512101656,
      "tokens_seen": 1486094336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027766298896690075,
      "loss": 2.4703,
      "theoretical_loss": 3.518955144569021,
      "tokens_seen": 1486159872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776529588766299,
      "loss": 2.7413,
      "theoretical_loss": 3.518941438701523,
      "tokens_seen": 1486225408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776429287863591,
      "loss": 2.5478,
      "theoretical_loss": 3.518927733607595,
      "tokens_seen": 1486290944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776328986960883,
      "loss": 2.7123,
      "theoretical_loss": 3.518914029287159,
      "tokens_seen": 1486356480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776228686058175,
      "loss": 2.6896,
      "theoretical_loss": 3.518900325740137,
      "tokens_seen": 1486422016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027761283851554665,
      "loss": 2.7992,
      "theoretical_loss": 3.518886622966452,
      "tokens_seen": 1486487552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027760280842527584,
      "loss": 2.6367,
      "theoretical_loss": 3.518872920966025,
      "tokens_seen": 1486553088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277592778335005,
      "loss": 2.8035,
      "theoretical_loss": 3.51885921973878,
      "tokens_seen": 1486618624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027758274824473425,
      "loss": 2.6149,
      "theoretical_loss": 3.5188455192846377,
      "tokens_seen": 1486684160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775727181544634,
      "loss": 2.771,
      "theoretical_loss": 3.5188318196035215,
      "tokens_seen": 1486749696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775626880641926,
      "loss": 2.9162,
      "theoretical_loss": 3.518818120695353,
      "tokens_seen": 1486815232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027755265797392174,
      "loss": 2.8375,
      "theoretical_loss": 3.5188044225600548,
      "tokens_seen": 1486880768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277542627883651,
      "loss": 2.7337,
      "theoretical_loss": 3.518790725197549,
      "tokens_seen": 1486946304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027753259779338016,
      "loss": 2.836,
      "theoretical_loss": 3.5187770286077584,
      "tokens_seen": 1487011840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027752256770310934,
      "loss": 2.7107,
      "theoretical_loss": 3.5187633327906047,
      "tokens_seen": 1487077376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775125376128385,
      "loss": 2.5469,
      "theoretical_loss": 3.5187496377460112,
      "tokens_seen": 1487142912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775025075225677,
      "loss": 2.6512,
      "theoretical_loss": 3.5187359434738994,
      "tokens_seen": 1487208448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774924774322969,
      "loss": 2.7829,
      "theoretical_loss": 3.5187222499741924,
      "tokens_seen": 1487273984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774824473420261,
      "loss": 2.567,
      "theoretical_loss": 3.518708557246812,
      "tokens_seen": 1487339520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027747241725175524,
      "loss": 2.5233,
      "theoretical_loss": 3.518694865291681,
      "tokens_seen": 1487405056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774623871614845,
      "loss": 2.6127,
      "theoretical_loss": 3.518681174108721,
      "tokens_seen": 1487470592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027745235707121366,
      "loss": 2.5781,
      "theoretical_loss": 3.518667483697856,
      "tokens_seen": 1487536128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027744232698094284,
      "loss": 2.6465,
      "theoretical_loss": 3.5186537940590075,
      "tokens_seen": 1487601664
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 821509,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.194403648376465,
      "objective/train/theoretical_loss": 3.5186401051920977,
      "objective/train/tokens_used": 1508127200,
      "theoretical_loss": 3.5186401051920977,
      "tokens_seen": 1487667200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277432296890672,
      "loss": 2.4996,
      "theoretical_loss": 3.5186401051920977,
      "tokens_seen": 1487667200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774222668004012,
      "loss": 2.4979,
      "theoretical_loss": 3.5186264170970496,
      "tokens_seen": 1487732736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774122367101304,
      "loss": 2.7645,
      "theoretical_loss": 3.518612729773785,
      "tokens_seen": 1487798272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774022066198596,
      "loss": 2.5919,
      "theoretical_loss": 3.5185990432222267,
      "tokens_seen": 1487863808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027739217652958875,
      "loss": 2.7519,
      "theoretical_loss": 3.5185853574422983,
      "tokens_seen": 1487929344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277382146439318,
      "loss": 2.633,
      "theoretical_loss": 3.5185716724339207,
      "tokens_seen": 1487994880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773721163490471,
      "loss": 2.5317,
      "theoretical_loss": 3.518557988197017,
      "tokens_seen": 1488060416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027736208625877634,
      "loss": 2.6653,
      "theoretical_loss": 3.5185443047315106,
      "tokens_seen": 1488125952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773520561685055,
      "loss": 2.6096,
      "theoretical_loss": 3.5185306220373223,
      "tokens_seen": 1488191488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773420260782347,
      "loss": 2.7613,
      "theoretical_loss": 3.5185169401143765,
      "tokens_seen": 1488257024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773319959879639,
      "loss": 2.7551,
      "theoretical_loss": 3.5185032589625944,
      "tokens_seen": 1488322560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027732196589769307,
      "loss": 2.6489,
      "theoretical_loss": 3.518489578581899,
      "tokens_seen": 1488388096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027731193580742225,
      "loss": 2.6812,
      "theoretical_loss": 3.518475898972213,
      "tokens_seen": 1488453632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773019057171515,
      "loss": 2.6617,
      "theoretical_loss": 3.5184622201334594,
      "tokens_seen": 1488519168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772918756268806,
      "loss": 2.5584,
      "theoretical_loss": 3.51844854206556,
      "tokens_seen": 1488584704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027728184553660985,
      "loss": 2.6746,
      "theoretical_loss": 3.5184348647684383,
      "tokens_seen": 1488650240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027727181544633903,
      "loss": 2.698,
      "theoretical_loss": 3.5184211882420158,
      "tokens_seen": 1488715776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772617853560682,
      "loss": 2.8228,
      "theoretical_loss": 3.518407512486216,
      "tokens_seen": 1488781312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772517552657974,
      "loss": 2.5921,
      "theoretical_loss": 3.5183938375009616,
      "tokens_seen": 1488846848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027724172517552657,
      "loss": 2.7007,
      "theoretical_loss": 3.5183801632861744,
      "tokens_seen": 1488912384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027723169508525575,
      "loss": 2.8786,
      "theoretical_loss": 3.5183664898417786,
      "tokens_seen": 1488977920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277221664994985,
      "loss": 2.8498,
      "theoretical_loss": 3.5183528171676954,
      "tokens_seen": 1489043456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772116349047141,
      "loss": 2.6078,
      "theoretical_loss": 3.518339145263848,
      "tokens_seen": 1489108992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027720160481444335,
      "loss": 2.633,
      "theoretical_loss": 3.51832547413016,
      "tokens_seen": 1489174528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771915747241725,
      "loss": 2.8041,
      "theoretical_loss": 3.5183118037665526,
      "tokens_seen": 1489240064
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 822598,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.609518527984619,
      "objective/train/theoretical_loss": 3.5182981341729493,
      "objective/train/tokens_used": 1509765600,
      "theoretical_loss": 3.5182981341729493,
      "tokens_seen": 1489305600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771815446339017,
      "loss": 2.5348,
      "theoretical_loss": 3.5182981341729493,
      "tokens_seen": 1489305600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771715145436309,
      "loss": 2.7176,
      "theoretical_loss": 3.518284465349273,
      "tokens_seen": 1489371136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771614844533601,
      "loss": 2.5294,
      "theoretical_loss": 3.5182707972954463,
      "tokens_seen": 1489436672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027715145436308926,
      "loss": 2.8534,
      "theoretical_loss": 3.518257130011392,
      "tokens_seen": 1489502208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771414242728185,
      "loss": 2.445,
      "theoretical_loss": 3.5182434634970328,
      "tokens_seen": 1489567744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771313941825476,
      "loss": 2.6848,
      "theoretical_loss": 3.5182297977522916,
      "tokens_seen": 1489633280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027712136409227685,
      "loss": 2.7385,
      "theoretical_loss": 3.518216132777091,
      "tokens_seen": 1489698816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277111334002006,
      "loss": 2.5515,
      "theoretical_loss": 3.5182024685713538,
      "tokens_seen": 1489764352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771013039117352,
      "loss": 2.5135,
      "theoretical_loss": 3.518188805135003,
      "tokens_seen": 1489829888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770912738214644,
      "loss": 2.7728,
      "theoretical_loss": 3.5181751424679613,
      "tokens_seen": 1489895424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770812437311936,
      "loss": 2.7078,
      "theoretical_loss": 3.518161480570152,
      "tokens_seen": 1489960960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027707121364092276,
      "loss": 2.5694,
      "theoretical_loss": 3.5181478194414977,
      "tokens_seen": 1490026496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027706118355065194,
      "loss": 2.5484,
      "theoretical_loss": 3.518134159081921,
      "tokens_seen": 1490092032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770511534603811,
      "loss": 2.5302,
      "theoretical_loss": 3.518120499491345,
      "tokens_seen": 1490157568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027704112337011036,
      "loss": 2.7996,
      "theoretical_loss": 3.5181068406696925,
      "tokens_seen": 1490223104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770310932798395,
      "loss": 2.6767,
      "theoretical_loss": 3.5180931826168864,
      "tokens_seen": 1490288640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770210631895687,
      "loss": 2.6392,
      "theoretical_loss": 3.51807952533285,
      "tokens_seen": 1490354176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770110330992979,
      "loss": 2.8262,
      "theoretical_loss": 3.518065868817505,
      "tokens_seen": 1490419712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770010030090271,
      "loss": 2.6533,
      "theoretical_loss": 3.518052213070776,
      "tokens_seen": 1490485248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769909729187563,
      "loss": 2.5859,
      "theoretical_loss": 3.5180385580925853,
      "tokens_seen": 1490550784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027698094282848544,
      "loss": 2.3949,
      "theoretical_loss": 3.5180249038828553,
      "tokens_seen": 1490616320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769709127382147,
      "loss": 2.6479,
      "theoretical_loss": 3.5180112504415098,
      "tokens_seen": 1490681856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027696088264794386,
      "loss": 2.57,
      "theoretical_loss": 3.5179975977684714,
      "tokens_seen": 1490747392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027695085255767304,
      "loss": 2.6443,
      "theoretical_loss": 3.517983945863663,
      "tokens_seen": 1490812928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769408224674022,
      "loss": 2.4459,
      "theoretical_loss": 3.5179702947270073,
      "tokens_seen": 1490878464
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 823992,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9485185146331787,
      "objective/train/theoretical_loss": 3.517956644358428,
      "objective/train/tokens_used": 1511404000,
      "theoretical_loss": 3.517956644358428,
      "tokens_seen": 1490944000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769307923771314,
      "loss": 2.7354,
      "theoretical_loss": 3.517956644358428,
      "tokens_seen": 1490944000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769207622868606,
      "loss": 2.6115,
      "theoretical_loss": 3.517942994757848,
      "tokens_seen": 1491009536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769107321965898,
      "loss": 2.6366,
      "theoretical_loss": 3.51792934592519,
      "tokens_seen": 1491075072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027690070210631895,
      "loss": 2.5609,
      "theoretical_loss": 3.517915697860377,
      "tokens_seen": 1491140608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768906720160482,
      "loss": 2.812,
      "theoretical_loss": 3.5179020505633325,
      "tokens_seen": 1491206144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768806419257773,
      "loss": 2.6548,
      "theoretical_loss": 3.5178884040339797,
      "tokens_seen": 1491271680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027687061183550655,
      "loss": 2.7482,
      "theoretical_loss": 3.517874758272241,
      "tokens_seen": 1491337216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768605817452357,
      "loss": 2.458,
      "theoretical_loss": 3.5178611132780393,
      "tokens_seen": 1491402752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768505516549649,
      "loss": 2.5753,
      "theoretical_loss": 3.517847469051299,
      "tokens_seen": 1491468288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768405215646941,
      "loss": 2.5678,
      "theoretical_loss": 3.517833825591942,
      "tokens_seen": 1491533824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027683049147442327,
      "loss": 2.6824,
      "theoretical_loss": 3.517820182899892,
      "tokens_seen": 1491599360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027682046138415245,
      "loss": 2.5631,
      "theoretical_loss": 3.517806540975072,
      "tokens_seen": 1491664896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768104312938817,
      "loss": 2.7715,
      "theoretical_loss": 3.5177928998174055,
      "tokens_seen": 1491730432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768004012036108,
      "loss": 2.5204,
      "theoretical_loss": 3.517779259426815,
      "tokens_seen": 1491795968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027679037111334005,
      "loss": 2.8217,
      "theoretical_loss": 3.5177656198032237,
      "tokens_seen": 1491861504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027678034102306923,
      "loss": 2.7824,
      "theoretical_loss": 3.5177519809465556,
      "tokens_seen": 1491927040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767703109327984,
      "loss": 2.7031,
      "theoretical_loss": 3.517738342856733,
      "tokens_seen": 1491992576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767602808425276,
      "loss": 2.5397,
      "theoretical_loss": 3.5177247055336798,
      "tokens_seen": 1492058112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027675025075225677,
      "loss": 2.4636,
      "theoretical_loss": 3.5177110689773183,
      "tokens_seen": 1492123648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027674022066198595,
      "loss": 2.6908,
      "theoretical_loss": 3.5176974331875726,
      "tokens_seen": 1492189184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767301905717152,
      "loss": 2.4168,
      "theoretical_loss": 3.5176837981643656,
      "tokens_seen": 1492254720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767201604814443,
      "loss": 2.7574,
      "theoretical_loss": 3.5176701639076207,
      "tokens_seen": 1492320256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027671013039117355,
      "loss": 2.7032,
      "theoretical_loss": 3.517656530417261,
      "tokens_seen": 1492385792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767001003009027,
      "loss": 2.5744,
      "theoretical_loss": 3.5176428976932095,
      "tokens_seen": 1492451328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766900702106319,
      "loss": 2.6954,
      "theoretical_loss": 3.51762926573539,
      "tokens_seen": 1492516864
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 824636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.556931734085083,
      "objective/train/theoretical_loss": 3.5176156345437253,
      "objective/train/tokens_used": 1513042400,
      "theoretical_loss": 3.5176156345437253,
      "tokens_seen": 1492582400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766800401203611,
      "loss": 2.7401,
      "theoretical_loss": 3.5176156345437253,
      "tokens_seen": 1492582400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766700100300903,
      "loss": 2.4808,
      "theoretical_loss": 3.5176020041181397,
      "tokens_seen": 1492647936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027665997993981946,
      "loss": 2.4619,
      "theoretical_loss": 3.5175883744585548,
      "tokens_seen": 1492713472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766499498495487,
      "loss": 2.6713,
      "theoretical_loss": 3.517574745564896,
      "tokens_seen": 1492779008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766399197592778,
      "loss": 2.7158,
      "theoretical_loss": 3.5175611174370847,
      "tokens_seen": 1492844544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027662988966900705,
      "loss": 2.6191,
      "theoretical_loss": 3.5175474900750454,
      "tokens_seen": 1492910080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766198595787362,
      "loss": 2.637,
      "theoretical_loss": 3.5175338634787012,
      "tokens_seen": 1492975616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766098294884654,
      "loss": 2.5831,
      "theoretical_loss": 3.5175202376479753,
      "tokens_seen": 1493041152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765997993981946,
      "loss": 2.5333,
      "theoretical_loss": 3.517506612582791,
      "tokens_seen": 1493106688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765897693079238,
      "loss": 2.6254,
      "theoretical_loss": 3.5174929882830726,
      "tokens_seen": 1493172224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027657973921765296,
      "loss": 2.6903,
      "theoretical_loss": 3.517479364748742,
      "tokens_seen": 1493237760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027656970912738214,
      "loss": 2.5451,
      "theoretical_loss": 3.517465741979724,
      "tokens_seen": 1493303296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765596790371113,
      "loss": 2.3796,
      "theoretical_loss": 3.517452119975941,
      "tokens_seen": 1493368832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027654964894684056,
      "loss": 2.4264,
      "theoretical_loss": 3.5174384987373175,
      "tokens_seen": 1493434368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765396188565697,
      "loss": 2.4714,
      "theoretical_loss": 3.5174248782637756,
      "tokens_seen": 1493499904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765295887662989,
      "loss": 2.5993,
      "theoretical_loss": 3.51741125855524,
      "tokens_seen": 1493565440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027651955867602805,
      "loss": 2.6736,
      "theoretical_loss": 3.5173976396116333,
      "tokens_seen": 1493630976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765095285857573,
      "loss": 2.6955,
      "theoretical_loss": 3.5173840214328793,
      "tokens_seen": 1493696512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027649949849548646,
      "loss": 2.5953,
      "theoretical_loss": 3.5173704040189016,
      "tokens_seen": 1493762048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027648946840521564,
      "loss": 2.7007,
      "theoretical_loss": 3.517356787369624,
      "tokens_seen": 1493827584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764794383149448,
      "loss": 2.6716,
      "theoretical_loss": 3.517343171484969,
      "tokens_seen": 1493893120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027646940822467406,
      "loss": 2.5113,
      "theoretical_loss": 3.517329556364861,
      "tokens_seen": 1493958656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764593781344032,
      "loss": 2.4526,
      "theoretical_loss": 3.5173159420092235,
      "tokens_seen": 1494024192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764493480441324,
      "loss": 2.4199,
      "theoretical_loss": 3.51730232841798,
      "tokens_seen": 1494089728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027643931795386155,
      "loss": 2.7458,
      "theoretical_loss": 3.517288715591053,
      "tokens_seen": 1494155264
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 825478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.548858642578125,
      "objective/train/theoretical_loss": 3.517275103528368,
      "objective/train/tokens_used": 1514680800,
      "theoretical_loss": 3.517275103528368,
      "tokens_seen": 1494220800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764292878635908,
      "loss": 2.8416,
      "theoretical_loss": 3.517275103528368,
      "tokens_seen": 1494220800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027641925777331997,
      "loss": 2.7125,
      "theoretical_loss": 3.5172614922298466,
      "tokens_seen": 1494286336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027640922768304915,
      "loss": 2.6171,
      "theoretical_loss": 3.517247881695414,
      "tokens_seen": 1494351872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027639919759277833,
      "loss": 2.3508,
      "theoretical_loss": 3.5172342719249925,
      "tokens_seen": 1494417408
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763891675025075,
      "loss": 2.549,
      "theoretical_loss": 3.5172206629185068,
      "tokens_seen": 1494482944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763791374122367,
      "loss": 2.6508,
      "theoretical_loss": 3.51720705467588,
      "tokens_seen": 1494548480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763691073219659,
      "loss": 2.685,
      "theoretical_loss": 3.5171934471970356,
      "tokens_seen": 1494614016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027635907723169505,
      "loss": 2.7894,
      "theoretical_loss": 3.517179840481897,
      "tokens_seen": 1494679552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763490471414243,
      "loss": 2.5345,
      "theoretical_loss": 3.517166234530389,
      "tokens_seen": 1494745088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763390170511534,
      "loss": 2.8126,
      "theoretical_loss": 3.5171526293424344,
      "tokens_seen": 1494810624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027632898696088265,
      "loss": 2.6083,
      "theoretical_loss": 3.517139024917957,
      "tokens_seen": 1494876160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027631895687061183,
      "loss": 2.343,
      "theoretical_loss": 3.51712542125688,
      "tokens_seen": 1494941696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276308926780341,
      "loss": 2.7823,
      "theoretical_loss": 3.5171118183591283,
      "tokens_seen": 1495007232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762988966900702,
      "loss": 2.5965,
      "theoretical_loss": 3.5170982162246247,
      "tokens_seen": 1495072768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027628886659979943,
      "loss": 2.634,
      "theoretical_loss": 3.5170846148532933,
      "tokens_seen": 1495138304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027627883650952856,
      "loss": 2.7194,
      "theoretical_loss": 3.517071014245057,
      "tokens_seen": 1495203840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762688064192578,
      "loss": 2.3761,
      "theoretical_loss": 3.517057414399841,
      "tokens_seen": 1495269376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027625877632898697,
      "loss": 2.5364,
      "theoretical_loss": 3.5170438153175683,
      "tokens_seen": 1495334912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027624874623871615,
      "loss": 2.6406,
      "theoretical_loss": 3.5170302169981618,
      "tokens_seen": 1495400448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762387161484454,
      "loss": 2.6997,
      "theoretical_loss": 3.517016619441547,
      "tokens_seen": 1495465984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762286860581745,
      "loss": 2.6544,
      "theoretical_loss": 3.5170030226476463,
      "tokens_seen": 1495531520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027621865596790375,
      "loss": 2.7054,
      "theoretical_loss": 3.5169894266163837,
      "tokens_seen": 1495597056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762086258776329,
      "loss": 2.4439,
      "theoretical_loss": 3.516975831347684,
      "tokens_seen": 1495662592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761985957873621,
      "loss": 2.5708,
      "theoretical_loss": 3.51696223684147,
      "tokens_seen": 1495728128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761885656970913,
      "loss": 2.494,
      "theoretical_loss": 3.516948643097666,
      "tokens_seen": 1495793664
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 826035,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.259209156036377,
      "objective/train/theoretical_loss": 3.5169350501161953,
      "objective/train/tokens_used": 1516319200,
      "theoretical_loss": 3.5169350501161953,
      "tokens_seen": 1495859200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761785356068205,
      "loss": 2.5004,
      "theoretical_loss": 3.5169350501161953,
      "tokens_seen": 1495859200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027616850551654966,
      "loss": 2.532,
      "theoretical_loss": 3.5169214578969825,
      "tokens_seen": 1495924736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761584754262789,
      "loss": 2.5712,
      "theoretical_loss": 3.5169078664399507,
      "tokens_seen": 1495990272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276148445336008,
      "loss": 2.8897,
      "theoretical_loss": 3.516894275745025,
      "tokens_seen": 1496055808
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027613841524573725,
      "loss": 2.7066,
      "theoretical_loss": 3.5168806858121275,
      "tokens_seen": 1496121344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761283851554664,
      "loss": 2.7881,
      "theoretical_loss": 3.5168670966411835,
      "tokens_seen": 1496186880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761183550651956,
      "loss": 2.7021,
      "theoretical_loss": 3.5168535082321166,
      "tokens_seen": 1496252416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761083249749248,
      "loss": 2.7848,
      "theoretical_loss": 3.5168399205848506,
      "tokens_seen": 1496317952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276098294884654,
      "loss": 2.4479,
      "theoretical_loss": 3.516826333699309,
      "tokens_seen": 1496383488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027608826479438316,
      "loss": 2.4746,
      "theoretical_loss": 3.5168127475754165,
      "tokens_seen": 1496449024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027607823470411234,
      "loss": 2.8988,
      "theoretical_loss": 3.516799162213097,
      "tokens_seen": 1496514560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760682046138415,
      "loss": 2.6517,
      "theoretical_loss": 3.5167855776122736,
      "tokens_seen": 1496580096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027605817452357076,
      "loss": 2.6976,
      "theoretical_loss": 3.516771993772871,
      "tokens_seen": 1496645632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760481444332999,
      "loss": 2.6669,
      "theoretical_loss": 3.516758410694813,
      "tokens_seen": 1496711168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760381143430291,
      "loss": 2.7465,
      "theoretical_loss": 3.516744828378024,
      "tokens_seen": 1496776704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027602808425275825,
      "loss": 2.6385,
      "theoretical_loss": 3.5167312468224274,
      "tokens_seen": 1496842240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760180541624875,
      "loss": 2.703,
      "theoretical_loss": 3.5167176660279473,
      "tokens_seen": 1496907776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027600802407221666,
      "loss": 2.8551,
      "theoretical_loss": 3.516704085994508,
      "tokens_seen": 1496973312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027599799398194584,
      "loss": 2.8573,
      "theoretical_loss": 3.516690506722033,
      "tokens_seen": 1497038848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275987963891675,
      "loss": 2.6318,
      "theoretical_loss": 3.5166769282104475,
      "tokens_seen": 1497104384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027597793380140426,
      "loss": 2.7615,
      "theoretical_loss": 3.516663350459674,
      "tokens_seen": 1497169920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759679037111334,
      "loss": 2.6212,
      "theoretical_loss": 3.516649773469638,
      "tokens_seen": 1497235456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759578736208626,
      "loss": 2.8382,
      "theoretical_loss": 3.5166361972402624,
      "tokens_seen": 1497300992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027594784353059175,
      "loss": 2.5502,
      "theoretical_loss": 3.5166226217714724,
      "tokens_seen": 1497366528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275937813440321,
      "loss": 2.6544,
      "theoretical_loss": 3.5166090470631914,
      "tokens_seen": 1497432064
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 827284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4943959712982178,
      "objective/train/theoretical_loss": 3.516595473115344,
      "objective/train/tokens_used": 1517957600,
      "theoretical_loss": 3.516595473115344,
      "tokens_seen": 1497497600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027592778335005017,
      "loss": 2.6545,
      "theoretical_loss": 3.516595473115344,
      "tokens_seen": 1497497600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027591775325977935,
      "loss": 2.5794,
      "theoretical_loss": 3.5165818999278535,
      "tokens_seen": 1497563136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027590772316950853,
      "loss": 2.6723,
      "theoretical_loss": 3.5165683275006447,
      "tokens_seen": 1497628672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758976930792377,
      "loss": 2.8326,
      "theoretical_loss": 3.516554755833641,
      "tokens_seen": 1497694208
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758876629889669,
      "loss": 2.6049,
      "theoretical_loss": 3.5165411849267683,
      "tokens_seen": 1497759744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758776328986961,
      "loss": 2.7356,
      "theoretical_loss": 3.5165276147799487,
      "tokens_seen": 1497825280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027586760280842525,
      "loss": 2.7711,
      "theoretical_loss": 3.516514045393108,
      "tokens_seen": 1497890816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758575727181545,
      "loss": 2.611,
      "theoretical_loss": 3.516500476766169,
      "tokens_seen": 1497956352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758475426278836,
      "loss": 2.653,
      "theoretical_loss": 3.516486908899057,
      "tokens_seen": 1498021888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027583751253761285,
      "loss": 2.484,
      "theoretical_loss": 3.516473341791696,
      "tokens_seen": 1498087424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027582748244734203,
      "loss": 2.5086,
      "theoretical_loss": 3.5164597754440097,
      "tokens_seen": 1498152960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758174523570712,
      "loss": 2.4447,
      "theoretical_loss": 3.516446209855922,
      "tokens_seen": 1498218496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758074222668004,
      "loss": 2.4562,
      "theoretical_loss": 3.5164326450273586,
      "tokens_seen": 1498284032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027579739217652963,
      "loss": 2.6962,
      "theoretical_loss": 3.5164190809582427,
      "tokens_seen": 1498349568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027578736208625876,
      "loss": 2.6403,
      "theoretical_loss": 3.516405517648499,
      "tokens_seen": 1498415104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275777331995988,
      "loss": 2.6789,
      "theoretical_loss": 3.516391955098052,
      "tokens_seen": 1498480640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757673019057171,
      "loss": 2.5963,
      "theoretical_loss": 3.516378393306825,
      "tokens_seen": 1498546176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027575727181544635,
      "loss": 2.6649,
      "theoretical_loss": 3.5163648322747427,
      "tokens_seen": 1498611712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027574724172517554,
      "loss": 2.6449,
      "theoretical_loss": 3.51635127200173,
      "tokens_seen": 1498677248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757372116349047,
      "loss": 2.8132,
      "theoretical_loss": 3.5163377124877107,
      "tokens_seen": 1498742784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757271815446339,
      "loss": 2.698,
      "theoretical_loss": 3.516324153732609,
      "tokens_seen": 1498808320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757171514543631,
      "loss": 2.5638,
      "theoretical_loss": 3.51631059573635,
      "tokens_seen": 1498873856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027570712136409226,
      "loss": 2.6337,
      "theoretical_loss": 3.516297038498857,
      "tokens_seen": 1498939392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756970912738215,
      "loss": 2.6763,
      "theoretical_loss": 3.5162834820200555,
      "tokens_seen": 1499004928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756870611835506,
      "loss": 2.6062,
      "theoretical_loss": 3.516269926299869,
      "tokens_seen": 1499070464
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 827773,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0328586101531982,
      "objective/train/theoretical_loss": 3.5162563713382218,
      "objective/train/tokens_used": 1519596000,
      "theoretical_loss": 3.5162563713382218,
      "tokens_seen": 1499136000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027567703109327986,
      "loss": 2.8939,
      "theoretical_loss": 3.5162563713382218,
      "tokens_seen": 1499136000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275667001003009,
      "loss": 2.5476,
      "theoretical_loss": 3.516242817135039,
      "tokens_seen": 1499201536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756569709127382,
      "loss": 2.8686,
      "theoretical_loss": 3.516229263690245,
      "tokens_seen": 1499267072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756469408224674,
      "loss": 2.7172,
      "theoretical_loss": 3.5162157110037633,
      "tokens_seen": 1499332608
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756369107321966,
      "loss": 2.6608,
      "theoretical_loss": 3.516202159075519,
      "tokens_seen": 1499398144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027562688064192576,
      "loss": 2.5377,
      "theoretical_loss": 3.5161886079054367,
      "tokens_seen": 1499463680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275616850551655,
      "loss": 2.884,
      "theoretical_loss": 3.5161750574934407,
      "tokens_seen": 1499529216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756068204613841,
      "loss": 2.8149,
      "theoretical_loss": 3.5161615078394552,
      "tokens_seen": 1499594752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027559679037111336,
      "loss": 2.6367,
      "theoretical_loss": 3.516147958943405,
      "tokens_seen": 1499660288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755867602808425,
      "loss": 2.7369,
      "theoretical_loss": 3.516134410805214,
      "tokens_seen": 1499725824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755767301905717,
      "loss": 2.5466,
      "theoretical_loss": 3.5161208634248076,
      "tokens_seen": 1499791360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755667001003009,
      "loss": 2.6547,
      "theoretical_loss": 3.5161073168021098,
      "tokens_seen": 1499856896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755566700100301,
      "loss": 2.5219,
      "theoretical_loss": 3.5160937709370446,
      "tokens_seen": 1499922432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027554663991975927,
      "loss": 2.7234,
      "theoretical_loss": 3.516080225829538,
      "tokens_seen": 1499987968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027553660982948845,
      "loss": 2.635,
      "theoretical_loss": 3.5160666814795127,
      "tokens_seen": 1500053504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027552657973921763,
      "loss": 2.6421,
      "theoretical_loss": 3.5160531378868947,
      "tokens_seen": 1500119040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027551654964894686,
      "loss": 2.6204,
      "theoretical_loss": 3.516039595051608,
      "tokens_seen": 1500184576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027550651955867604,
      "loss": 2.4071,
      "theoretical_loss": 3.5160260529735767,
      "tokens_seen": 1500250112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754964894684052,
      "loss": 2.6619,
      "theoretical_loss": 3.5160125116527263,
      "tokens_seen": 1500315648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027548645937813446,
      "loss": 2.7885,
      "theoretical_loss": 3.515998971088981,
      "tokens_seen": 1500381184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754764292878636,
      "loss": 2.6149,
      "theoretical_loss": 3.515985431282265,
      "tokens_seen": 1500446720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754663991975928,
      "loss": 2.7472,
      "theoretical_loss": 3.515971892232504,
      "tokens_seen": 1500512256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027545636910732195,
      "loss": 2.4303,
      "theoretical_loss": 3.515958353939621,
      "tokens_seen": 1500577792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754463390170512,
      "loss": 2.7099,
      "theoretical_loss": 3.5159448164035423,
      "tokens_seen": 1500643328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027543630892678037,
      "loss": 2.8133,
      "theoretical_loss": 3.5159312796241915,
      "tokens_seen": 1500708864
    },
    {
      "epoch": 0.45,
      "objective/train/docs_used": 828910,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.892310380935669,
      "objective/train/theoretical_loss": 3.5159177436014932,
      "objective/train/tokens_used": 1521234400,
      "theoretical_loss": 3.5159177436014932,
      "tokens_seen": 1500774400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027542627883650955,
      "loss": 2.6749,
      "theoretical_loss": 3.5159177436014932,
      "tokens_seen": 1500774400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027541624874623873,
      "loss": 2.5888,
      "theoretical_loss": 3.515904208335373,
      "tokens_seen": 1500839936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754062186559679,
      "loss": 2.7679,
      "theoretical_loss": 3.515890673825754,
      "tokens_seen": 1500905472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753961885656971,
      "loss": 2.7315,
      "theoretical_loss": 3.515877140072563,
      "tokens_seen": 1500971008
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753861584754263,
      "loss": 2.6909,
      "theoretical_loss": 3.515863607075723,
      "tokens_seen": 1501036544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027537612838515545,
      "loss": 2.575,
      "theoretical_loss": 3.5158500748351593,
      "tokens_seen": 1501102080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753660982948847,
      "loss": 2.5857,
      "theoretical_loss": 3.5158365433507965,
      "tokens_seen": 1501167616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753560682046138,
      "loss": 2.5885,
      "theoretical_loss": 3.51582301262256,
      "tokens_seen": 1501233152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027534603811434305,
      "loss": 2.7469,
      "theoretical_loss": 3.5158094826503734,
      "tokens_seen": 1501298688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027533600802407223,
      "loss": 2.6433,
      "theoretical_loss": 3.5157959534341625,
      "tokens_seen": 1501364224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753259779338014,
      "loss": 2.7166,
      "theoretical_loss": 3.515782424973851,
      "tokens_seen": 1501429760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753159478435306,
      "loss": 2.5621,
      "theoretical_loss": 3.5157688972693646,
      "tokens_seen": 1501495296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027530591775325983,
      "loss": 2.6297,
      "theoretical_loss": 3.515755370320628,
      "tokens_seen": 1501560832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027529588766298896,
      "loss": 2.5006,
      "theoretical_loss": 3.5157418441275654,
      "tokens_seen": 1501626368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752858575727182,
      "loss": 2.7195,
      "theoretical_loss": 3.515728318690102,
      "tokens_seen": 1501691904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752758274824473,
      "loss": 2.7733,
      "theoretical_loss": 3.5157147940081623,
      "tokens_seen": 1501757440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027526579739217655,
      "loss": 2.8983,
      "theoretical_loss": 3.515701270081672,
      "tokens_seen": 1501822976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027525576730190574,
      "loss": 2.5811,
      "theoretical_loss": 3.515687746910555,
      "tokens_seen": 1501888512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752457372116349,
      "loss": 2.7298,
      "theoretical_loss": 3.5156742244947363,
      "tokens_seen": 1501954048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752357071213641,
      "loss": 2.8053,
      "theoretical_loss": 3.5156607028341416,
      "tokens_seen": 1502019584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752256770310933,
      "loss": 2.6385,
      "theoretical_loss": 3.5156471819286947,
      "tokens_seen": 1502085120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027521564694082246,
      "loss": 2.5712,
      "theoretical_loss": 3.5156336617783213,
      "tokens_seen": 1502150656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752056168505517,
      "loss": 2.7582,
      "theoretical_loss": 3.515620142382945,
      "tokens_seen": 1502216192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751955867602808,
      "loss": 2.658,
      "theoretical_loss": 3.5156066237424923,
      "tokens_seen": 1502281728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027518555667001006,
      "loss": 2.5665,
      "theoretical_loss": 3.5155931058568877,
      "tokens_seen": 1502347264
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 830192,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0300025939941406,
      "objective/train/theoretical_loss": 3.5155795887260552,
      "objective/train/tokens_used": 1522872800,
      "theoretical_loss": 3.5155795887260552,
      "tokens_seen": 1502412800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751755265797392,
      "loss": 2.5777,
      "theoretical_loss": 3.5155795887260552,
      "tokens_seen": 1502412800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751654964894684,
      "loss": 2.6391,
      "theoretical_loss": 3.5155660723499205,
      "tokens_seen": 1502478336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751554663991976,
      "loss": 2.6167,
      "theoretical_loss": 3.5155525567284087,
      "tokens_seen": 1502543872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751454363089268,
      "loss": 2.8656,
      "theoretical_loss": 3.5155390418614445,
      "tokens_seen": 1502609408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027513540621865596,
      "loss": 2.7072,
      "theoretical_loss": 3.5155255277489528,
      "tokens_seen": 1502674944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751253761283852,
      "loss": 2.7833,
      "theoretical_loss": 3.5155120143908585,
      "tokens_seen": 1502740480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751153460381143,
      "loss": 2.5559,
      "theoretical_loss": 3.5154985017870866,
      "tokens_seen": 1502806016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027510531594784356,
      "loss": 2.6557,
      "theoretical_loss": 3.515484989937563,
      "tokens_seen": 1502871552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750952858575727,
      "loss": 2.6358,
      "theoretical_loss": 3.515471478842211,
      "tokens_seen": 1502937088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750852557673019,
      "loss": 2.5667,
      "theoretical_loss": 3.5154579685009573,
      "tokens_seen": 1503002624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750752256770311,
      "loss": 2.6918,
      "theoretical_loss": 3.5154444589137253,
      "tokens_seen": 1503068160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750651955867603,
      "loss": 2.4683,
      "theoretical_loss": 3.5154309500804417,
      "tokens_seen": 1503133696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027505516549648947,
      "loss": 2.6315,
      "theoretical_loss": 3.5154174420010307,
      "tokens_seen": 1503199232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027504513540621865,
      "loss": 2.7197,
      "theoretical_loss": 3.515403934675417,
      "tokens_seen": 1503264768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027503510531594783,
      "loss": 2.8134,
      "theoretical_loss": 3.515390428103527,
      "tokens_seen": 1503330304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027502507522567706,
      "loss": 2.6432,
      "theoretical_loss": 3.5153769222852844,
      "tokens_seen": 1503395840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750150451354062,
      "loss": 2.8172,
      "theoretical_loss": 3.5153634172206143,
      "tokens_seen": 1503461376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750050150451354,
      "loss": 2.6194,
      "theoretical_loss": 3.5153499129094428,
      "tokens_seen": 1503526912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027499498495486455,
      "loss": 2.6861,
      "theoretical_loss": 3.515336409351695,
      "tokens_seen": 1503592448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749849548645938,
      "loss": 2.7088,
      "theoretical_loss": 3.5153229065472953,
      "tokens_seen": 1503657984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027497492477432297,
      "loss": 2.5598,
      "theoretical_loss": 3.515309404496169,
      "tokens_seen": 1503723520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027496489468405215,
      "loss": 2.8307,
      "theoretical_loss": 3.515295903198241,
      "tokens_seen": 1503789056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027495486459378133,
      "loss": 2.7083,
      "theoretical_loss": 3.5152824026534373,
      "tokens_seen": 1503854592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027494483450351057,
      "loss": 2.5902,
      "theoretical_loss": 3.515268902861682,
      "tokens_seen": 1503920128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749348044132397,
      "loss": 2.4284,
      "theoretical_loss": 3.515255403822902,
      "tokens_seen": 1503985664
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 830854,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.144463539123535,
      "objective/train/theoretical_loss": 3.5152419055370205,
      "objective/train/tokens_used": 1524511200,
      "theoretical_loss": 3.5152419055370205,
      "tokens_seen": 1504051200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027492477432296893,
      "loss": 2.6293,
      "theoretical_loss": 3.5152419055370205,
      "tokens_seen": 1504051200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027491474423269806,
      "loss": 2.597,
      "theoretical_loss": 3.515228408003964,
      "tokens_seen": 1504116736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749047141424273,
      "loss": 2.5466,
      "theoretical_loss": 3.515214911223657,
      "tokens_seen": 1504182272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027489468405215647,
      "loss": 2.5915,
      "theoretical_loss": 3.5152014151960254,
      "tokens_seen": 1504247808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027488465396188565,
      "loss": 2.7626,
      "theoretical_loss": 3.5151879199209937,
      "tokens_seen": 1504313344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027487462387161483,
      "loss": 2.6936,
      "theoretical_loss": 3.515174425398487,
      "tokens_seen": 1504378880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274864593781344,
      "loss": 2.6766,
      "theoretical_loss": 3.515160931628432,
      "tokens_seen": 1504444416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002748545636910732,
      "loss": 2.6261,
      "theoretical_loss": 3.515147438610753,
      "tokens_seen": 1504509952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027484453360080243,
      "loss": 2.4522,
      "theoretical_loss": 3.515133946345375,
      "tokens_seen": 1504575488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027483450351053156,
      "loss": 2.7393,
      "theoretical_loss": 3.515120454832223,
      "tokens_seen": 1504641024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002748244734202608,
      "loss": 2.5249,
      "theoretical_loss": 3.515106964071224,
      "tokens_seen": 1504706560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027481444332999,
      "loss": 2.7539,
      "theoretical_loss": 3.5150934740623017,
      "tokens_seen": 1504772096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027480441323971916,
      "loss": 2.6781,
      "theoretical_loss": 3.5150799848053818,
      "tokens_seen": 1504837632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027479438314944834,
      "loss": 2.7024,
      "theoretical_loss": 3.51506649630039,
      "tokens_seen": 1504903168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747843530591775,
      "loss": 2.6632,
      "theoretical_loss": 3.515053008547251,
      "tokens_seen": 1504968704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027477432296890675,
      "loss": 2.668,
      "theoretical_loss": 3.5150395215458907,
      "tokens_seen": 1505034240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027476429287863594,
      "loss": 2.5524,
      "theoretical_loss": 3.515026035296235,
      "tokens_seen": 1505099776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747542627883651,
      "loss": 2.5007,
      "theoretical_loss": 3.515012549798208,
      "tokens_seen": 1505165312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747442326980943,
      "loss": 2.5512,
      "theoretical_loss": 3.514999065051736,
      "tokens_seen": 1505230848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747342026078235,
      "loss": 2.7295,
      "theoretical_loss": 3.5149855810567434,
      "tokens_seen": 1505296384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027472417251755266,
      "loss": 2.7503,
      "theoretical_loss": 3.514972097813157,
      "tokens_seen": 1505361920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747141424272819,
      "loss": 2.6382,
      "theoretical_loss": 3.514958615320901,
      "tokens_seen": 1505427456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274704112337011,
      "loss": 2.7362,
      "theoretical_loss": 3.514945133579902,
      "tokens_seen": 1505492992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027469408224674026,
      "loss": 2.8275,
      "theoretical_loss": 3.514931652590084,
      "tokens_seen": 1505558528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746840521564694,
      "loss": 2.6193,
      "theoretical_loss": 3.514918172351374,
      "tokens_seen": 1505624064
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 831840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.831157684326172,
      "objective/train/theoretical_loss": 3.514904692863696,
      "objective/train/tokens_used": 1526149600,
      "theoretical_loss": 3.514904692863696,
      "tokens_seen": 1505689600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746740220661986,
      "loss": 2.6352,
      "theoretical_loss": 3.514904692863696,
      "tokens_seen": 1505689600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746639919759278,
      "loss": 2.729,
      "theoretical_loss": 3.514891214126976,
      "tokens_seen": 1505755136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274653961885657,
      "loss": 2.7844,
      "theoretical_loss": 3.5148777361411403,
      "tokens_seen": 1505820672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027464393179538616,
      "loss": 2.6508,
      "theoretical_loss": 3.5148642589061136,
      "tokens_seen": 1505886208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746339017051154,
      "loss": 2.7491,
      "theoretical_loss": 3.5148507824218216,
      "tokens_seen": 1505951744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746238716148445,
      "loss": 2.8469,
      "theoretical_loss": 3.5148373066881895,
      "tokens_seen": 1506017280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027461384152457376,
      "loss": 2.4959,
      "theoretical_loss": 3.5148238317051432,
      "tokens_seen": 1506082816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746038114343029,
      "loss": 2.4439,
      "theoretical_loss": 3.5148103574726077,
      "tokens_seen": 1506148352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745937813440321,
      "loss": 2.4271,
      "theoretical_loss": 3.5147968839905097,
      "tokens_seen": 1506213888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745837512537613,
      "loss": 2.6857,
      "theoretical_loss": 3.5147834112587737,
      "tokens_seen": 1506279424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745737211634905,
      "loss": 2.6605,
      "theoretical_loss": 3.5147699392773255,
      "tokens_seen": 1506344960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027456369107321967,
      "loss": 2.5731,
      "theoretical_loss": 3.5147564680460905,
      "tokens_seen": 1506410496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027455366098294885,
      "loss": 2.4873,
      "theoretical_loss": 3.514742997564995,
      "tokens_seen": 1506476032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027454363089267803,
      "loss": 2.5501,
      "theoretical_loss": 3.5147295278339636,
      "tokens_seen": 1506541568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027453360080240726,
      "loss": 2.7184,
      "theoretical_loss": 3.514716058852923,
      "tokens_seen": 1506607104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745235707121364,
      "loss": 2.649,
      "theoretical_loss": 3.514702590621798,
      "tokens_seen": 1506672640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745135406218656,
      "loss": 2.7671,
      "theoretical_loss": 3.5146891231405144,
      "tokens_seen": 1506738176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027450351053159475,
      "loss": 2.4505,
      "theoretical_loss": 3.514675656408998,
      "tokens_seen": 1506803712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274493480441324,
      "loss": 2.7606,
      "theoretical_loss": 3.5146621904271744,
      "tokens_seen": 1506869248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027448345035105317,
      "loss": 2.5961,
      "theoretical_loss": 3.514648725194969,
      "tokens_seen": 1506934784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027447342026078235,
      "loss": 2.4571,
      "theoretical_loss": 3.5146352607123084,
      "tokens_seen": 1507000320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027446339017051153,
      "loss": 2.6324,
      "theoretical_loss": 3.5146217969791174,
      "tokens_seen": 1507065856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027445336008024077,
      "loss": 2.7367,
      "theoretical_loss": 3.514608333995321,
      "tokens_seen": 1507131392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744433299899699,
      "loss": 2.7612,
      "theoretical_loss": 3.514594871760847,
      "tokens_seen": 1507196928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027443329989969913,
      "loss": 2.6376,
      "theoretical_loss": 3.5145814102756194,
      "tokens_seen": 1507262464
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 832546,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4045536518096924,
      "objective/train/theoretical_loss": 3.5145679495395643,
      "objective/train/tokens_used": 1527788000,
      "theoretical_loss": 3.5145679495395643,
      "tokens_seen": 1507328000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027442326980942826,
      "loss": 2.5427,
      "theoretical_loss": 3.5145679495395643,
      "tokens_seen": 1507328000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744132397191575,
      "loss": 2.8226,
      "theoretical_loss": 3.5145544895526077,
      "tokens_seen": 1507393536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027440320962888667,
      "loss": 2.6932,
      "theoretical_loss": 3.5145410303146756,
      "tokens_seen": 1507459072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027439317953861585,
      "loss": 2.7651,
      "theoretical_loss": 3.5145275718256928,
      "tokens_seen": 1507524608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027438314944834503,
      "loss": 2.4618,
      "theoretical_loss": 3.514514114085586,
      "tokens_seen": 1507590144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743731193580742,
      "loss": 2.6122,
      "theoretical_loss": 3.5145006570942803,
      "tokens_seen": 1507655680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743630892678034,
      "loss": 2.6113,
      "theoretical_loss": 3.5144872008517023,
      "tokens_seen": 1507721216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027435305917753263,
      "loss": 2.7369,
      "theoretical_loss": 3.514473745357777,
      "tokens_seen": 1507786752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027434302908726176,
      "loss": 2.6171,
      "theoretical_loss": 3.5144602906124307,
      "tokens_seen": 1507852288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274332998996991,
      "loss": 2.6005,
      "theoretical_loss": 3.5144468366155888,
      "tokens_seen": 1507917824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743229689067202,
      "loss": 2.724,
      "theoretical_loss": 3.5144333833671775,
      "tokens_seen": 1507983360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027431293881644936,
      "loss": 2.8599,
      "theoretical_loss": 3.5144199308671227,
      "tokens_seen": 1508048896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027430290872617854,
      "loss": 2.6357,
      "theoretical_loss": 3.5144064791153498,
      "tokens_seen": 1508114432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742928786359077,
      "loss": 2.642,
      "theoretical_loss": 3.514393028111785,
      "tokens_seen": 1508179968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742828485456369,
      "loss": 2.7981,
      "theoretical_loss": 3.5143795778563542,
      "tokens_seen": 1508245504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027427281845536614,
      "loss": 2.6059,
      "theoretical_loss": 3.5143661283489833,
      "tokens_seen": 1508311040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027426278836509526,
      "loss": 2.5728,
      "theoretical_loss": 3.514352679589598,
      "tokens_seen": 1508376576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742527582748245,
      "loss": 2.608,
      "theoretical_loss": 3.5143392315781243,
      "tokens_seen": 1508442112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742427281845536,
      "loss": 2.5706,
      "theoretical_loss": 3.5143257843144884,
      "tokens_seen": 1508507648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027423269809428286,
      "loss": 2.5239,
      "theoretical_loss": 3.514312337798615,
      "tokens_seen": 1508573184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027422266800401204,
      "loss": 2.4601,
      "theoretical_loss": 3.514298892030432,
      "tokens_seen": 1508638720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742126379137412,
      "loss": 2.5261,
      "theoretical_loss": 3.514285447009864,
      "tokens_seen": 1508704256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742026078234704,
      "loss": 2.6423,
      "theoretical_loss": 3.514272002736837,
      "tokens_seen": 1508769792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741925777331996,
      "loss": 2.6611,
      "theoretical_loss": 3.5142585592112776,
      "tokens_seen": 1508835328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027418254764292877,
      "loss": 2.825,
      "theoretical_loss": 3.514245116433111,
      "tokens_seen": 1508900864
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 833973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9619827270507812,
      "objective/train/theoretical_loss": 3.5142316744022644,
      "objective/train/tokens_used": 1529426400,
      "theoretical_loss": 3.5142316744022644,
      "tokens_seen": 1508966400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274172517552658,
      "loss": 2.6156,
      "theoretical_loss": 3.5142316744022644,
      "tokens_seen": 1508966400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027416248746238713,
      "loss": 2.5253,
      "theoretical_loss": 3.5142182331186627,
      "tokens_seen": 1509031936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027415245737211636,
      "loss": 2.5278,
      "theoretical_loss": 3.5142047925822317,
      "tokens_seen": 1509097472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027414242728184554,
      "loss": 2.8177,
      "theoretical_loss": 3.5141913527928983,
      "tokens_seen": 1509163008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741323971915747,
      "loss": 2.5251,
      "theoretical_loss": 3.514177913750588,
      "tokens_seen": 1509228544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741223671013039,
      "loss": 2.639,
      "theoretical_loss": 3.5141644754552277,
      "tokens_seen": 1509294080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741123370110331,
      "loss": 2.4835,
      "theoretical_loss": 3.514151037906742,
      "tokens_seen": 1509359616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027410230692076227,
      "loss": 2.6622,
      "theoretical_loss": 3.514137601105058,
      "tokens_seen": 1509425152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740922768304915,
      "loss": 2.5462,
      "theoretical_loss": 3.514124165050102,
      "tokens_seen": 1509490688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027408224674022063,
      "loss": 2.6131,
      "theoretical_loss": 3.514110729741799,
      "tokens_seen": 1509556224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027407221664994987,
      "loss": 2.6414,
      "theoretical_loss": 3.5140972951800764,
      "tokens_seen": 1509621760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274062186559679,
      "loss": 2.6016,
      "theoretical_loss": 3.514083861364859,
      "tokens_seen": 1509687296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027405215646940823,
      "loss": 2.5222,
      "theoretical_loss": 3.5140704282960735,
      "tokens_seen": 1509752832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740421263791374,
      "loss": 2.4032,
      "theoretical_loss": 3.514056995973646,
      "tokens_seen": 1509818368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740320962888666,
      "loss": 2.6077,
      "theoretical_loss": 3.514043564397503,
      "tokens_seen": 1509883904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740220661985958,
      "loss": 2.7701,
      "theoretical_loss": 3.5140301335675703,
      "tokens_seen": 1509949440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027401203610832495,
      "loss": 2.7075,
      "theoretical_loss": 3.514016703483774,
      "tokens_seen": 1510014976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740020060180542,
      "loss": 2.6965,
      "theoretical_loss": 3.5140032741460403,
      "tokens_seen": 1510080512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027399197592778337,
      "loss": 2.7415,
      "theoretical_loss": 3.5139898455542955,
      "tokens_seen": 1510146048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027398194583751255,
      "loss": 2.4527,
      "theoretical_loss": 3.5139764177084656,
      "tokens_seen": 1510211584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027397191574724173,
      "loss": 2.6699,
      "theoretical_loss": 3.5139629906084773,
      "tokens_seen": 1510277120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027396188565697097,
      "loss": 2.5838,
      "theoretical_loss": 3.513949564254256,
      "tokens_seen": 1510342656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002739518555667001,
      "loss": 2.7862,
      "theoretical_loss": 3.513936138645729,
      "tokens_seen": 1510408192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027394182547642933,
      "loss": 2.8045,
      "theoretical_loss": 3.5139227137828213,
      "tokens_seen": 1510473728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027393179538615846,
      "loss": 2.8602,
      "theoretical_loss": 3.5139092896654596,
      "tokens_seen": 1510539264
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 834645,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1922924518585205,
      "objective/train/theoretical_loss": 3.513895866293571,
      "objective/train/tokens_used": 1531064800,
      "theoretical_loss": 3.513895866293571,
      "tokens_seen": 1510604800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002739217652958877,
      "loss": 2.5753,
      "theoretical_loss": 3.513895866293571,
      "tokens_seen": 1510604800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027391173520561687,
      "loss": 2.3657,
      "theoretical_loss": 3.5138824436670806,
      "tokens_seen": 1510670336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027390170511534605,
      "loss": 2.5346,
      "theoretical_loss": 3.513869021785915,
      "tokens_seen": 1510735872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027389167502507524,
      "loss": 2.6478,
      "theoretical_loss": 3.513855600650001,
      "tokens_seen": 1510801408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738816449348044,
      "loss": 2.4694,
      "theoretical_loss": 3.513842180259264,
      "tokens_seen": 1510866944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738716148445336,
      "loss": 3.0212,
      "theoretical_loss": 3.513828760613631,
      "tokens_seen": 1510932480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027386158475426283,
      "loss": 2.6099,
      "theoretical_loss": 3.5138153417130287,
      "tokens_seen": 1510998016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027385155466399196,
      "loss": 2.6894,
      "theoretical_loss": 3.513801923557382,
      "tokens_seen": 1511063552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738415245737212,
      "loss": 2.5968,
      "theoretical_loss": 3.5137885061466188,
      "tokens_seen": 1511129088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738314944834504,
      "loss": 2.6717,
      "theoretical_loss": 3.5137750894806645,
      "tokens_seen": 1511194624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027382146439317956,
      "loss": 2.6688,
      "theoretical_loss": 3.5137616735594452,
      "tokens_seen": 1511260160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027381143430290874,
      "loss": 2.6462,
      "theoretical_loss": 3.513748258382888,
      "tokens_seen": 1511325696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738014042126379,
      "loss": 2.5841,
      "theoretical_loss": 3.5137348439509193,
      "tokens_seen": 1511391232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737913741223671,
      "loss": 2.433,
      "theoretical_loss": 3.513721430263465,
      "tokens_seen": 1511456768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027378134403209634,
      "loss": 2.7375,
      "theoretical_loss": 3.513708017320452,
      "tokens_seen": 1511522304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027377131394182546,
      "loss": 2.465,
      "theoretical_loss": 3.5136946051218065,
      "tokens_seen": 1511587840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737612838515547,
      "loss": 2.4808,
      "theoretical_loss": 3.5136811936674546,
      "tokens_seen": 1511653376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737512537612838,
      "loss": 2.5119,
      "theoretical_loss": 3.5136677829573233,
      "tokens_seen": 1511718912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027374122367101306,
      "loss": 2.5002,
      "theoretical_loss": 3.5136543729913385,
      "tokens_seen": 1511784448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027373119358074224,
      "loss": 2.5367,
      "theoretical_loss": 3.5136409637694266,
      "tokens_seen": 1511849984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737211634904714,
      "loss": 2.5183,
      "theoretical_loss": 3.5136275552915146,
      "tokens_seen": 1511915520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737111334002006,
      "loss": 2.6877,
      "theoretical_loss": 3.5136141475575293,
      "tokens_seen": 1511981056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737011033099298,
      "loss": 2.8269,
      "theoretical_loss": 3.513600740567396,
      "tokens_seen": 1512046592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027369107321965897,
      "loss": 2.6315,
      "theoretical_loss": 3.5135873343210418,
      "tokens_seen": 1512112128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736810431293882,
      "loss": 2.6766,
      "theoretical_loss": 3.5135739288183934,
      "tokens_seen": 1512177664
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 835710,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6051583290100098,
      "objective/train/theoretical_loss": 3.513560524059377,
      "objective/train/tokens_used": 1532703200,
      "theoretical_loss": 3.513560524059377,
      "tokens_seen": 1512243200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027367101303911733,
      "loss": 2.7525,
      "theoretical_loss": 3.513560524059377,
      "tokens_seen": 1512243200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027366098294884656,
      "loss": 2.6414,
      "theoretical_loss": 3.513547120043919,
      "tokens_seen": 1512308736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027365095285857574,
      "loss": 2.5715,
      "theoretical_loss": 3.5135337167719465,
      "tokens_seen": 1512374272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736409227683049,
      "loss": 2.872,
      "theoretical_loss": 3.5135203142433857,
      "tokens_seen": 1512439808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736308926780341,
      "loss": 2.5917,
      "theoretical_loss": 3.513506912458163,
      "tokens_seen": 1512505344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736208625877633,
      "loss": 2.5591,
      "theoretical_loss": 3.5134935114162054,
      "tokens_seen": 1512570880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027361083249749247,
      "loss": 2.6358,
      "theoretical_loss": 3.513480111117439,
      "tokens_seen": 1512636416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736008024072217,
      "loss": 2.3513,
      "theoretical_loss": 3.5134667115617906,
      "tokens_seen": 1512701952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027359077231695083,
      "loss": 2.6456,
      "theoretical_loss": 3.513453312749187,
      "tokens_seen": 1512767488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027358074222668007,
      "loss": 2.3847,
      "theoretical_loss": 3.5134399146795543,
      "tokens_seen": 1512833024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735707121364092,
      "loss": 2.5213,
      "theoretical_loss": 3.51342651735282,
      "tokens_seen": 1512898560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027356068204613843,
      "loss": 2.5991,
      "theoretical_loss": 3.51341312076891,
      "tokens_seen": 1512964096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735506519558676,
      "loss": 2.7351,
      "theoretical_loss": 3.5133997249277504,
      "tokens_seen": 1513029632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735406218655968,
      "loss": 2.5695,
      "theoretical_loss": 3.5133863298292694,
      "tokens_seen": 1513095168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027353059177532597,
      "loss": 2.6775,
      "theoretical_loss": 3.5133729354733925,
      "tokens_seen": 1513160704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027352056168505515,
      "loss": 2.5956,
      "theoretical_loss": 3.513359541860047,
      "tokens_seen": 1513226240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027351053159478433,
      "loss": 2.5487,
      "theoretical_loss": 3.5133461489891586,
      "tokens_seen": 1513291776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027350050150451357,
      "loss": 2.6104,
      "theoretical_loss": 3.513332756860655,
      "tokens_seen": 1513357312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734904714142427,
      "loss": 2.371,
      "theoretical_loss": 3.5133193654744623,
      "tokens_seen": 1513422848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027348044132397193,
      "loss": 2.7229,
      "theoretical_loss": 3.513305974830508,
      "tokens_seen": 1513488384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734704112337011,
      "loss": 2.7482,
      "theoretical_loss": 3.513292584928718,
      "tokens_seen": 1513553920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734603811434303,
      "loss": 2.587,
      "theoretical_loss": 3.5132791957690195,
      "tokens_seen": 1513619456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734503510531595,
      "loss": 2.487,
      "theoretical_loss": 3.513265807351339,
      "tokens_seen": 1513684992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027344032096288866,
      "loss": 2.5462,
      "theoretical_loss": 3.513252419675603,
      "tokens_seen": 1513750528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027343029087261784,
      "loss": 2.9182,
      "theoretical_loss": 3.5132390327417395,
      "tokens_seen": 1513816064
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 836217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6275956630706787,
      "objective/train/theoretical_loss": 3.5132256465496736,
      "objective/train/tokens_used": 1534341600,
      "theoretical_loss": 3.5132256465496736,
      "tokens_seen": 1513881600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734202607823471,
      "loss": 2.5214,
      "theoretical_loss": 3.5132256465496736,
      "tokens_seen": 1513881600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734102306920762,
      "loss": 2.5889,
      "theoretical_loss": 3.513212261099333,
      "tokens_seen": 1513947136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027340020060180544,
      "loss": 2.7688,
      "theoretical_loss": 3.5131988763906445,
      "tokens_seen": 1514012672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027339017051153456,
      "loss": 2.6498,
      "theoretical_loss": 3.513185492423535,
      "tokens_seen": 1514078208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733801404212638,
      "loss": 2.4727,
      "theoretical_loss": 3.5131721091979307,
      "tokens_seen": 1514143744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273370110330993,
      "loss": 2.6173,
      "theoretical_loss": 3.5131587267137587,
      "tokens_seen": 1514209280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027336008024072216,
      "loss": 2.5951,
      "theoretical_loss": 3.5131453449709458,
      "tokens_seen": 1514274816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027335005015045134,
      "loss": 2.379,
      "theoretical_loss": 3.5131319639694194,
      "tokens_seen": 1514340352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733400200601806,
      "loss": 2.6354,
      "theoretical_loss": 3.513118583709106,
      "tokens_seen": 1514405888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733299899699097,
      "loss": 2.7082,
      "theoretical_loss": 3.513105204189932,
      "tokens_seen": 1514471424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027331995987963894,
      "loss": 2.7024,
      "theoretical_loss": 3.5130918254118253,
      "tokens_seen": 1514536960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027330992978936807,
      "loss": 2.4955,
      "theoretical_loss": 3.5130784473747116,
      "tokens_seen": 1514602496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732998996990973,
      "loss": 2.5076,
      "theoretical_loss": 3.5130650700785186,
      "tokens_seen": 1514668032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732898696088265,
      "loss": 2.4884,
      "theoretical_loss": 3.513051693523173,
      "tokens_seen": 1514733568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027327983951855566,
      "loss": 2.4027,
      "theoretical_loss": 3.5130383177086015,
      "tokens_seen": 1514799104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732698094282849,
      "loss": 2.579,
      "theoretical_loss": 3.5130249426347318,
      "tokens_seen": 1514864640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273259779338014,
      "loss": 2.6875,
      "theoretical_loss": 3.51301156830149,
      "tokens_seen": 1514930176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027324974924774326,
      "loss": 2.6929,
      "theoretical_loss": 3.512998194708803,
      "tokens_seen": 1514995712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027323971915747244,
      "loss": 2.5525,
      "theoretical_loss": 3.512984821856598,
      "tokens_seen": 1515061248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732296890672016,
      "loss": 2.795,
      "theoretical_loss": 3.5129714497448026,
      "tokens_seen": 1515126784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732196589769308,
      "loss": 2.4369,
      "theoretical_loss": 3.5129580783733436,
      "tokens_seen": 1515192320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027320962888666,
      "loss": 2.6934,
      "theoretical_loss": 3.5129447077421467,
      "tokens_seen": 1515257856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027319959879638917,
      "loss": 2.6587,
      "theoretical_loss": 3.512931337851141,
      "tokens_seen": 1515323392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731895687061184,
      "loss": 2.7382,
      "theoretical_loss": 3.5129179687002514,
      "tokens_seen": 1515388928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027317953861584753,
      "loss": 2.6136,
      "theoretical_loss": 3.512904600289406,
      "tokens_seen": 1515454464
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 837242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.814265251159668,
      "objective/train/theoretical_loss": 3.512891232618532,
      "objective/train/tokens_used": 1535980000,
      "theoretical_loss": 3.512891232618532,
      "tokens_seen": 1515520000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027316950852557676,
      "loss": 2.709,
      "theoretical_loss": 3.512891232618532,
      "tokens_seen": 1515520000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027315947843530594,
      "loss": 2.5236,
      "theoretical_loss": 3.512877865687556,
      "tokens_seen": 1515585536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731494483450351,
      "loss": 2.8312,
      "theoretical_loss": 3.5128644994964056,
      "tokens_seen": 1515651072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731394182547643,
      "loss": 2.6342,
      "theoretical_loss": 3.512851134045007,
      "tokens_seen": 1515716608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731293881644935,
      "loss": 2.3248,
      "theoretical_loss": 3.512837769333288,
      "tokens_seen": 1515782144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027311935807422267,
      "loss": 2.5907,
      "theoretical_loss": 3.512824405361175,
      "tokens_seen": 1515847680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731093279839519,
      "loss": 2.486,
      "theoretical_loss": 3.5128110421285963,
      "tokens_seen": 1515913216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027309929789368103,
      "loss": 2.561,
      "theoretical_loss": 3.5127976796354776,
      "tokens_seen": 1515978752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027308926780341027,
      "loss": 2.8013,
      "theoretical_loss": 3.5127843178817475,
      "tokens_seen": 1516044288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730792377131394,
      "loss": 2.5745,
      "theoretical_loss": 3.5127709568673318,
      "tokens_seen": 1516109824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027306920762286863,
      "loss": 2.6471,
      "theoretical_loss": 3.5127575965921576,
      "tokens_seen": 1516175360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730591775325978,
      "loss": 2.5791,
      "theoretical_loss": 3.5127442370561535,
      "tokens_seen": 1516240896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273049147442327,
      "loss": 2.4793,
      "theoretical_loss": 3.5127308782592452,
      "tokens_seen": 1516306432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027303911735205617,
      "loss": 2.6887,
      "theoretical_loss": 3.5127175202013605,
      "tokens_seen": 1516371968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027302908726178535,
      "loss": 2.8017,
      "theoretical_loss": 3.512704162882427,
      "tokens_seen": 1516437504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027301905717151453,
      "loss": 2.6432,
      "theoretical_loss": 3.5126908063023707,
      "tokens_seen": 1516503040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027300902708124377,
      "loss": 2.6128,
      "theoretical_loss": 3.5126774504611196,
      "tokens_seen": 1516568576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729989969909729,
      "loss": 2.5469,
      "theoretical_loss": 3.512664095358601,
      "tokens_seen": 1516634112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027298896690070213,
      "loss": 2.6823,
      "theoretical_loss": 3.512650740994742,
      "tokens_seen": 1516699648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729789368104313,
      "loss": 2.6485,
      "theoretical_loss": 3.5126373873694696,
      "tokens_seen": 1516765184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729689067201605,
      "loss": 2.6336,
      "theoretical_loss": 3.512624034482711,
      "tokens_seen": 1516830720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729588766298897,
      "loss": 2.6261,
      "theoretical_loss": 3.512610682334394,
      "tokens_seen": 1516896256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027294884653961886,
      "loss": 2.7059,
      "theoretical_loss": 3.512597330924445,
      "tokens_seen": 1516961792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027293881644934804,
      "loss": 2.5078,
      "theoretical_loss": 3.512583980252792,
      "tokens_seen": 1517027328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729287863590773,
      "loss": 2.5923,
      "theoretical_loss": 3.5125706303193622,
      "tokens_seen": 1517092864
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 837775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.929088830947876,
      "objective/train/theoretical_loss": 3.5125572811240833,
      "objective/train/tokens_used": 1537618400,
      "theoretical_loss": 3.5125572811240833,
      "tokens_seen": 1517158400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729187562688064,
      "loss": 2.7973,
      "theoretical_loss": 3.5125572811240833,
      "tokens_seen": 1517158400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027290872617853564,
      "loss": 2.7104,
      "theoretical_loss": 3.512543932666881,
      "tokens_seen": 1517223936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027289869608826476,
      "loss": 2.5193,
      "theoretical_loss": 3.512530584947684,
      "tokens_seen": 1517289472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272888665997994,
      "loss": 2.7381,
      "theoretical_loss": 3.51251723796642,
      "tokens_seen": 1517355008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728786359077232,
      "loss": 2.6233,
      "theoretical_loss": 3.5125038917230147,
      "tokens_seen": 1517420544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027286860581745236,
      "loss": 2.5533,
      "theoretical_loss": 3.5124905462173963,
      "tokens_seen": 1517486080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027285857572718154,
      "loss": 2.5144,
      "theoretical_loss": 3.512477201449493,
      "tokens_seen": 1517551616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728485456369108,
      "loss": 2.6085,
      "theoretical_loss": 3.512463857419231,
      "tokens_seen": 1517617152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728385155466399,
      "loss": 2.4266,
      "theoretical_loss": 3.512450514126538,
      "tokens_seen": 1517682688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027282848545636914,
      "loss": 2.5963,
      "theoretical_loss": 3.5124371715713414,
      "tokens_seen": 1517748224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027281845536609827,
      "loss": 2.6626,
      "theoretical_loss": 3.5124238297535686,
      "tokens_seen": 1517813760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728084252758275,
      "loss": 2.674,
      "theoretical_loss": 3.5124104886731473,
      "tokens_seen": 1517879296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727983951855567,
      "loss": 2.6903,
      "theoretical_loss": 3.5123971483300043,
      "tokens_seen": 1517944832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027278836509528586,
      "loss": 2.3904,
      "theoretical_loss": 3.512383808724068,
      "tokens_seen": 1518010368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027277833500501504,
      "loss": 2.5493,
      "theoretical_loss": 3.5123704698552647,
      "tokens_seen": 1518075904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727683049147442,
      "loss": 2.5472,
      "theoretical_loss": 3.5123571317235225,
      "tokens_seen": 1518141440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727582748244734,
      "loss": 2.7771,
      "theoretical_loss": 3.5123437943287685,
      "tokens_seen": 1518206976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027274824473420264,
      "loss": 2.7707,
      "theoretical_loss": 3.512330457670931,
      "tokens_seen": 1518272512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027273821464393177,
      "loss": 2.6094,
      "theoretical_loss": 3.5123171217499363,
      "tokens_seen": 1518338048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272728184553661,
      "loss": 2.6912,
      "theoretical_loss": 3.5123037865657127,
      "tokens_seen": 1518403584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027271815446339013,
      "loss": 2.4024,
      "theoretical_loss": 3.5122904521181875,
      "tokens_seen": 1518469120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027270812437311937,
      "loss": 2.6141,
      "theoretical_loss": 3.512277118407288,
      "tokens_seen": 1518534656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027269809428284855,
      "loss": 2.7052,
      "theoretical_loss": 3.512263785432942,
      "tokens_seen": 1518600192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027268806419257773,
      "loss": 2.5501,
      "theoretical_loss": 3.512250453195077,
      "tokens_seen": 1518665728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726780341023069,
      "loss": 2.6798,
      "theoretical_loss": 3.5122371216936203,
      "tokens_seen": 1518731264
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 838761,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3899428844451904,
      "objective/train/theoretical_loss": 3.5122237909284997,
      "objective/train/tokens_used": 1539256800,
      "theoretical_loss": 3.5122237909284997,
      "tokens_seen": 1518796800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027266800401203614,
      "loss": 2.6283,
      "theoretical_loss": 3.5122237909284997,
      "tokens_seen": 1518796800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027265797392176527,
      "loss": 2.6057,
      "theoretical_loss": 3.5122104608996425,
      "tokens_seen": 1518862336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726479438314945,
      "loss": 2.4512,
      "theoretical_loss": 3.512197131606977,
      "tokens_seen": 1518927872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027263791374122363,
      "loss": 2.9519,
      "theoretical_loss": 3.5121838030504295,
      "tokens_seen": 1518993408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027262788365095287,
      "loss": 2.5698,
      "theoretical_loss": 3.5121704752299285,
      "tokens_seen": 1519058944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027261785356068205,
      "loss": 2.7011,
      "theoretical_loss": 3.512157148145402,
      "tokens_seen": 1519124480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027260782347041123,
      "loss": 2.572,
      "theoretical_loss": 3.5121438217967764,
      "tokens_seen": 1519190016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725977933801404,
      "loss": 2.7506,
      "theoretical_loss": 3.51213049618398,
      "tokens_seen": 1519255552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725877632898696,
      "loss": 2.653,
      "theoretical_loss": 3.5121171713069406,
      "tokens_seen": 1519321088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725777331995988,
      "loss": 2.878,
      "theoretical_loss": 3.5121038471655854,
      "tokens_seen": 1519386624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272567703109328,
      "loss": 2.5658,
      "theoretical_loss": 3.5120905237598423,
      "tokens_seen": 1519452160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027255767301905714,
      "loss": 2.4503,
      "theoretical_loss": 3.512077201089639,
      "tokens_seen": 1519517696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027254764292878637,
      "loss": 2.4628,
      "theoretical_loss": 3.512063879154903,
      "tokens_seen": 1519583232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725376128385155,
      "loss": 2.6997,
      "theoretical_loss": 3.512050557955562,
      "tokens_seen": 1519648768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027252758274824473,
      "loss": 2.7425,
      "theoretical_loss": 3.512037237491544,
      "tokens_seen": 1519714304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027251755265797397,
      "loss": 2.4202,
      "theoretical_loss": 3.512023917762776,
      "tokens_seen": 1519779840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725075225677031,
      "loss": 2.6174,
      "theoretical_loss": 3.512010598769187,
      "tokens_seen": 1519845376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027249749247743233,
      "loss": 2.4447,
      "theoretical_loss": 3.5119972805107036,
      "tokens_seen": 1519910912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724874623871615,
      "loss": 2.5532,
      "theoretical_loss": 3.5119839629872533,
      "tokens_seen": 1519976448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724774322968907,
      "loss": 2.6757,
      "theoretical_loss": 3.5119706461987654,
      "tokens_seen": 1520041984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724674022066199,
      "loss": 2.5968,
      "theoretical_loss": 3.511957330145166,
      "tokens_seen": 1520107520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027245737211634906,
      "loss": 2.6554,
      "theoretical_loss": 3.5119440148263834,
      "tokens_seen": 1520173056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027244734202607824,
      "loss": 2.7129,
      "theoretical_loss": 3.5119307002423454,
      "tokens_seen": 1520238592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724373119358075,
      "loss": 2.6401,
      "theoretical_loss": 3.5119173863929807,
      "tokens_seen": 1520304128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724272818455366,
      "loss": 2.6772,
      "theoretical_loss": 3.511904073278216,
      "tokens_seen": 1520369664
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 839881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8261830806732178,
      "objective/train/theoretical_loss": 3.5118907608979786,
      "objective/train/tokens_used": 1540895200,
      "theoretical_loss": 3.5118907608979786,
      "tokens_seen": 1520435200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027241725175526584,
      "loss": 2.734,
      "theoretical_loss": 3.5118907608979786,
      "tokens_seen": 1520435200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027240722166499496,
      "loss": 2.7,
      "theoretical_loss": 3.511877449252198,
      "tokens_seen": 1520500736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723971915747242,
      "loss": 2.5904,
      "theoretical_loss": 3.5118641383408007,
      "tokens_seen": 1520566272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723871614844534,
      "loss": 2.6964,
      "theoretical_loss": 3.511850828163715,
      "tokens_seen": 1520631808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027237713139418256,
      "loss": 2.6544,
      "theoretical_loss": 3.511837518720869,
      "tokens_seen": 1520697344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027236710130391174,
      "loss": 2.758,
      "theoretical_loss": 3.5118242100121897,
      "tokens_seen": 1520762880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272357071213641,
      "loss": 2.4927,
      "theoretical_loss": 3.5118109020376056,
      "tokens_seen": 1520828416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723470411233701,
      "loss": 2.6839,
      "theoretical_loss": 3.5117975947970446,
      "tokens_seen": 1520893952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027233701103309934,
      "loss": 2.6804,
      "theoretical_loss": 3.5117842882904347,
      "tokens_seen": 1520959488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027232698094282847,
      "loss": 2.6058,
      "theoretical_loss": 3.511770982517703,
      "tokens_seen": 1521025024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723169508525577,
      "loss": 2.4964,
      "theoretical_loss": 3.5117576774787786,
      "tokens_seen": 1521090560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723069207622869,
      "loss": 2.4434,
      "theoretical_loss": 3.5117443731735887,
      "tokens_seen": 1521156096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027229689067201606,
      "loss": 2.4178,
      "theoretical_loss": 3.511731069602061,
      "tokens_seen": 1521221632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027228686058174524,
      "loss": 2.7692,
      "theoretical_loss": 3.511717766764124,
      "tokens_seen": 1521287168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722768304914744,
      "loss": 2.7603,
      "theoretical_loss": 3.5117044646597058,
      "tokens_seen": 1521352704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722668004012036,
      "loss": 2.5612,
      "theoretical_loss": 3.5116911632887335,
      "tokens_seen": 1521418240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027225677031093284,
      "loss": 2.4733,
      "theoretical_loss": 3.511677862651136,
      "tokens_seen": 1521483776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027224674022066197,
      "loss": 2.7743,
      "theoretical_loss": 3.5116645627468404,
      "tokens_seen": 1521549312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722367101303912,
      "loss": 2.889,
      "theoretical_loss": 3.511651263575775,
      "tokens_seen": 1521614848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027222668004012033,
      "loss": 2.2772,
      "theoretical_loss": 3.511637965137868,
      "tokens_seen": 1521680384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027221664994984957,
      "loss": 2.4408,
      "theoretical_loss": 3.511624667433048,
      "tokens_seen": 1521745920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027220661985957875,
      "loss": 2.5401,
      "theoretical_loss": 3.5116113704612415,
      "tokens_seen": 1521811456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027219658976930793,
      "loss": 2.636,
      "theoretical_loss": 3.511598074222378,
      "tokens_seen": 1521876992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721865596790371,
      "loss": 2.4016,
      "theoretical_loss": 3.5115847787163843,
      "tokens_seen": 1521942528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027217652958876635,
      "loss": 2.5686,
      "theoretical_loss": 3.5115714839431895,
      "tokens_seen": 1522008064
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 840376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.44497013092041,
      "objective/train/theoretical_loss": 3.511558189902721,
      "objective/train/tokens_used": 1542533600,
      "theoretical_loss": 3.511558189902721,
      "tokens_seen": 1522073600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027216649949849547,
      "loss": 2.5371,
      "theoretical_loss": 3.511558189902721,
      "tokens_seen": 1522073600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721564694082247,
      "loss": 2.8807,
      "theoretical_loss": 3.511544896594907,
      "tokens_seen": 1522139136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027214643931795383,
      "loss": 2.5307,
      "theoretical_loss": 3.511531604019676,
      "tokens_seen": 1522204672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027213640922768307,
      "loss": 2.5698,
      "theoretical_loss": 3.5115183121769555,
      "tokens_seen": 1522270208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027212637913741225,
      "loss": 2.5847,
      "theoretical_loss": 3.511505021066674,
      "tokens_seen": 1522335744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027211634904714143,
      "loss": 2.515,
      "theoretical_loss": 3.5114917306887596,
      "tokens_seen": 1522401280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721063189568706,
      "loss": 2.7306,
      "theoretical_loss": 3.51147844104314,
      "tokens_seen": 1522466816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720962888665998,
      "loss": 2.7946,
      "theoretical_loss": 3.5114651521297437,
      "tokens_seen": 1522532352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272086258776329,
      "loss": 2.625,
      "theoretical_loss": 3.5114518639484986,
      "tokens_seen": 1522597888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720762286860582,
      "loss": 2.4889,
      "theoretical_loss": 3.5114385764993337,
      "tokens_seen": 1522663424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027206619859578734,
      "loss": 2.8904,
      "theoretical_loss": 3.511425289782176,
      "tokens_seen": 1522728960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027205616850551657,
      "loss": 2.716,
      "theoretical_loss": 3.511412003796954,
      "tokens_seen": 1522794496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720461384152457,
      "loss": 2.6501,
      "theoretical_loss": 3.511398718543596,
      "tokens_seen": 1522860032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027203610832497493,
      "loss": 2.6602,
      "theoretical_loss": 3.5113854340220305,
      "tokens_seen": 1522925568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720260782347041,
      "loss": 2.6252,
      "theoretical_loss": 3.5113721502321855,
      "tokens_seen": 1522991104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720160481444333,
      "loss": 2.6795,
      "theoretical_loss": 3.511358867173989,
      "tokens_seen": 1523056640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720060180541625,
      "loss": 2.5826,
      "theoretical_loss": 3.511345584847369,
      "tokens_seen": 1523122176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719959879638917,
      "loss": 2.4883,
      "theoretical_loss": 3.5113323032522548,
      "tokens_seen": 1523187712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027198595787362084,
      "loss": 2.5199,
      "theoretical_loss": 3.511319022388574,
      "tokens_seen": 1523253248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719759277833501,
      "loss": 2.7365,
      "theoretical_loss": 3.511305742256254,
      "tokens_seen": 1523318784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719658976930792,
      "loss": 2.5625,
      "theoretical_loss": 3.5112924628552244,
      "tokens_seen": 1523384320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027195586760280844,
      "loss": 2.7394,
      "theoretical_loss": 3.5112791841854127,
      "tokens_seen": 1523449856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719458375125376,
      "loss": 2.6528,
      "theoretical_loss": 3.5112659062467477,
      "tokens_seen": 1523515392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719358074222668,
      "loss": 2.6489,
      "theoretical_loss": 3.511252629039157,
      "tokens_seen": 1523580928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271925777331996,
      "loss": 2.7094,
      "theoretical_loss": 3.5112393525625696,
      "tokens_seen": 1523646464
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 841650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9947397708892822,
      "objective/train/theoretical_loss": 3.5112260768169135,
      "objective/train/tokens_used": 1544172000,
      "theoretical_loss": 3.5112260768169135,
      "tokens_seen": 1523712000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027191574724172516,
      "loss": 2.7517,
      "theoretical_loss": 3.5112260768169135,
      "tokens_seen": 1523712000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027190571715145434,
      "loss": 2.5517,
      "theoretical_loss": 3.5112128018021167,
      "tokens_seen": 1523777536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718956870611836,
      "loss": 2.5746,
      "theoretical_loss": 3.511199527518108,
      "tokens_seen": 1523843072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718856569709127,
      "loss": 2.4339,
      "theoretical_loss": 3.5111862539648158,
      "tokens_seen": 1523908608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027187562688064194,
      "loss": 2.5428,
      "theoretical_loss": 3.5111729811421686,
      "tokens_seen": 1523974144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027186559679037107,
      "loss": 2.8515,
      "theoretical_loss": 3.511159709050094,
      "tokens_seen": 1524039680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718555667001003,
      "loss": 2.6896,
      "theoretical_loss": 3.5111464376885206,
      "tokens_seen": 1524105216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718455366098295,
      "loss": 2.4945,
      "theoretical_loss": 3.5111331670573773,
      "tokens_seen": 1524170752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027183550651955867,
      "loss": 2.8215,
      "theoretical_loss": 3.5111198971565925,
      "tokens_seen": 1524236288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027182547642928785,
      "loss": 2.657,
      "theoretical_loss": 3.5111066279860936,
      "tokens_seen": 1524301824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718154463390171,
      "loss": 2.6308,
      "theoretical_loss": 3.51109335954581,
      "tokens_seen": 1524367360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718054162487462,
      "loss": 2.6359,
      "theoretical_loss": 3.5110800918356704,
      "tokens_seen": 1524432896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027179538615847544,
      "loss": 2.6419,
      "theoretical_loss": 3.511066824855602,
      "tokens_seen": 1524498432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027178535606820457,
      "loss": 2.5588,
      "theoretical_loss": 3.511053558605534,
      "tokens_seen": 1524563968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717753259779338,
      "loss": 2.6852,
      "theoretical_loss": 3.5110402930853946,
      "tokens_seen": 1524629504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027176529588766304,
      "loss": 2.6305,
      "theoretical_loss": 3.511027028295113,
      "tokens_seen": 1524695040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027175526579739217,
      "loss": 2.7099,
      "theoretical_loss": 3.5110137642346166,
      "tokens_seen": 1524760576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717452357071214,
      "loss": 2.5974,
      "theoretical_loss": 3.5110005009038345,
      "tokens_seen": 1524826112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027173520561685053,
      "loss": 2.5775,
      "theoretical_loss": 3.510987238302695,
      "tokens_seen": 1524891648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027172517552657977,
      "loss": 2.6237,
      "theoretical_loss": 3.510973976431127,
      "tokens_seen": 1524957184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027171514543630895,
      "loss": 2.5999,
      "theoretical_loss": 3.510960715289059,
      "tokens_seen": 1525022720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027170511534603813,
      "loss": 2.7274,
      "theoretical_loss": 3.5109474548764186,
      "tokens_seen": 1525088256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716950852557673,
      "loss": 2.625,
      "theoretical_loss": 3.510934195193135,
      "tokens_seen": 1525153792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027168505516549655,
      "loss": 2.5496,
      "theoretical_loss": 3.510920936239137,
      "tokens_seen": 1525219328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027167502507522567,
      "loss": 2.7176,
      "theoretical_loss": 3.5109076780143527,
      "tokens_seen": 1525284864
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 842112,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0185139179229736,
      "objective/train/theoretical_loss": 3.510894420518711,
      "objective/train/tokens_used": 1545810400,
      "theoretical_loss": 3.510894420518711,
      "tokens_seen": 1525350400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716649949849549,
      "loss": 2.7613,
      "theoretical_loss": 3.510894420518711,
      "tokens_seen": 1525350400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027165496489468403,
      "loss": 2.7588,
      "theoretical_loss": 3.51088116375214,
      "tokens_seen": 1525415936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027164493480441327,
      "loss": 2.7699,
      "theoretical_loss": 3.510867907714569,
      "tokens_seen": 1525481472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027163490471414245,
      "loss": 2.648,
      "theoretical_loss": 3.510854652405926,
      "tokens_seen": 1525547008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027162487462387163,
      "loss": 2.7531,
      "theoretical_loss": 3.51084139782614,
      "tokens_seen": 1525612544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716148445336008,
      "loss": 2.4952,
      "theoretical_loss": 3.5108281439751394,
      "tokens_seen": 1525678080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027160481444333,
      "loss": 2.8022,
      "theoretical_loss": 3.5108148908528527,
      "tokens_seen": 1525743616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715947843530592,
      "loss": 2.7302,
      "theoretical_loss": 3.5108016384592085,
      "tokens_seen": 1525809152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715847542627884,
      "loss": 2.514,
      "theoretical_loss": 3.5107883867941356,
      "tokens_seen": 1525874688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027157472417251754,
      "loss": 2.5953,
      "theoretical_loss": 3.5107751358575627,
      "tokens_seen": 1525940224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715646940822468,
      "loss": 2.6973,
      "theoretical_loss": 3.5107618856494187,
      "tokens_seen": 1526005760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715546639919759,
      "loss": 2.582,
      "theoretical_loss": 3.510748636169632,
      "tokens_seen": 1526071296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027154463390170514,
      "loss": 2.5885,
      "theoretical_loss": 3.5107353874181313,
      "tokens_seen": 1526136832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715346038114343,
      "loss": 2.5398,
      "theoretical_loss": 3.5107221393948453,
      "tokens_seen": 1526202368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715245737211635,
      "loss": 2.5864,
      "theoretical_loss": 3.5107088920997027,
      "tokens_seen": 1526267904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715145436308927,
      "loss": 2.5209,
      "theoretical_loss": 3.510695645532632,
      "tokens_seen": 1526333440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715045135406219,
      "loss": 2.4995,
      "theoretical_loss": 3.5106823996935623,
      "tokens_seen": 1526398976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027149448345035104,
      "loss": 2.6928,
      "theoretical_loss": 3.5106691545824225,
      "tokens_seen": 1526464512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714844533600803,
      "loss": 2.5935,
      "theoretical_loss": 3.510655910199141,
      "tokens_seen": 1526530048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714744232698094,
      "loss": 2.3438,
      "theoretical_loss": 3.510642666543646,
      "tokens_seen": 1526595584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027146439317953864,
      "loss": 2.5466,
      "theoretical_loss": 3.510629423615867,
      "tokens_seen": 1526661120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714543630892678,
      "loss": 2.675,
      "theoretical_loss": 3.510616181415733,
      "tokens_seen": 1526726656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271444332998997,
      "loss": 2.5762,
      "theoretical_loss": 3.510602939943172,
      "tokens_seen": 1526792192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714343029087262,
      "loss": 2.5912,
      "theoretical_loss": 3.510589699198114,
      "tokens_seen": 1526857728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027142427281845536,
      "loss": 2.6199,
      "theoretical_loss": 3.5105764591804864,
      "tokens_seen": 1526923264
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 843396,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7498199939727783,
      "objective/train/theoretical_loss": 3.510563219890219,
      "objective/train/tokens_used": 1547448800,
      "theoretical_loss": 3.510563219890219,
      "tokens_seen": 1526988800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027141424272818454,
      "loss": 2.7651,
      "theoretical_loss": 3.510563219890219,
      "tokens_seen": 1526988800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714042126379138,
      "loss": 2.5227,
      "theoretical_loss": 3.51054998132724,
      "tokens_seen": 1527054336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713941825476429,
      "loss": 2.4665,
      "theoretical_loss": 3.5105367434914783,
      "tokens_seen": 1527119872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027138415245737214,
      "loss": 2.7011,
      "theoretical_loss": 3.5105235063828633,
      "tokens_seen": 1527185408
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027137412236710127,
      "loss": 2.6465,
      "theoretical_loss": 3.5105102700013235,
      "tokens_seen": 1527250944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713640922768305,
      "loss": 2.3794,
      "theoretical_loss": 3.5104970343467876,
      "tokens_seen": 1527316480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713540621865597,
      "loss": 2.6201,
      "theoretical_loss": 3.510483799419185,
      "tokens_seen": 1527382016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027134403209628887,
      "loss": 2.5714,
      "theoretical_loss": 3.510470565218444,
      "tokens_seen": 1527447552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027133400200601805,
      "loss": 2.6446,
      "theoretical_loss": 3.510457331744494,
      "tokens_seen": 1527513088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713239719157473,
      "loss": 2.6866,
      "theoretical_loss": 3.5104440989972634,
      "tokens_seen": 1527578624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713139418254764,
      "loss": 2.691,
      "theoretical_loss": 3.510430866976681,
      "tokens_seen": 1527644160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027130391173520564,
      "loss": 2.6412,
      "theoretical_loss": 3.5104176356826766,
      "tokens_seen": 1527709696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027129388164493477,
      "loss": 2.5747,
      "theoretical_loss": 3.5104044051151786,
      "tokens_seen": 1527775232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271283851554664,
      "loss": 2.6853,
      "theoretical_loss": 3.5103911752741155,
      "tokens_seen": 1527840768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712738214643932,
      "loss": 2.6371,
      "theoretical_loss": 3.5103779461594176,
      "tokens_seen": 1527906304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027126379137412237,
      "loss": 2.7817,
      "theoretical_loss": 3.5103647177710124,
      "tokens_seen": 1527971840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027125376128385155,
      "loss": 2.7717,
      "theoretical_loss": 3.5103514901088295,
      "tokens_seen": 1528037376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027124373119358073,
      "loss": 2.4568,
      "theoretical_loss": 3.510338263172798,
      "tokens_seen": 1528102912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712337011033099,
      "loss": 2.8217,
      "theoretical_loss": 3.5103250369628465,
      "tokens_seen": 1528168448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027122367101303915,
      "loss": 2.6221,
      "theoretical_loss": 3.5103118114789047,
      "tokens_seen": 1528233984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712136409227683,
      "loss": 2.372,
      "theoretical_loss": 3.510298586720901,
      "tokens_seen": 1528299520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712036108324975,
      "loss": 2.5911,
      "theoretical_loss": 3.5102853626887645,
      "tokens_seen": 1528365056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711935807422267,
      "loss": 2.6072,
      "theoretical_loss": 3.5102721393824243,
      "tokens_seen": 1528430592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027118355065195587,
      "loss": 2.5907,
      "theoretical_loss": 3.51025891680181,
      "tokens_seen": 1528496128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027117352056168505,
      "loss": 2.8846,
      "theoretical_loss": 3.5102456949468497,
      "tokens_seen": 1528561664
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 844011,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.368853807449341,
      "objective/train/theoretical_loss": 3.5102324738174726,
      "objective/train/tokens_used": 1549087200,
      "theoretical_loss": 3.5102324738174726,
      "tokens_seen": 1528627200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027116349047141423,
      "loss": 2.7747,
      "theoretical_loss": 3.5102324738174726,
      "tokens_seen": 1528627200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711534603811434,
      "loss": 2.6397,
      "theoretical_loss": 3.5102192534136085,
      "tokens_seen": 1528692736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027114343029087265,
      "loss": 2.628,
      "theoretical_loss": 3.5102060337351864,
      "tokens_seen": 1528758272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711334002006018,
      "loss": 2.4658,
      "theoretical_loss": 3.5101928147821346,
      "tokens_seen": 1528823808
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271123370110331,
      "loss": 2.7907,
      "theoretical_loss": 3.5101795965543827,
      "tokens_seen": 1528889344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027111334002006014,
      "loss": 2.6682,
      "theoretical_loss": 3.51016637905186,
      "tokens_seen": 1528954880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711033099297894,
      "loss": 2.6562,
      "theoretical_loss": 3.5101531622744955,
      "tokens_seen": 1529020416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027109327983951856,
      "loss": 2.4884,
      "theoretical_loss": 3.510139946222218,
      "tokens_seen": 1529085952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027108324974924774,
      "loss": 2.6723,
      "theoretical_loss": 3.510126730894957,
      "tokens_seen": 1529151488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710732196589769,
      "loss": 2.5381,
      "theoretical_loss": 3.5101135162926416,
      "tokens_seen": 1529217024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710631895687061,
      "loss": 2.6156,
      "theoretical_loss": 3.510100302415201,
      "tokens_seen": 1529282560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710531594784353,
      "loss": 2.6253,
      "theoretical_loss": 3.510087089262564,
      "tokens_seen": 1529348096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710431293881645,
      "loss": 2.5449,
      "theoretical_loss": 3.5100738768346607,
      "tokens_seen": 1529413632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027103309929789364,
      "loss": 2.6174,
      "theoretical_loss": 3.5100606651314195,
      "tokens_seen": 1529479168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710230692076229,
      "loss": 2.845,
      "theoretical_loss": 3.5100474541527693,
      "tokens_seen": 1529544704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710130391173521,
      "loss": 2.6564,
      "theoretical_loss": 3.51003424389864,
      "tokens_seen": 1529610240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027100300902708124,
      "loss": 2.6652,
      "theoretical_loss": 3.510021034368961,
      "tokens_seen": 1529675776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709929789368105,
      "loss": 2.6993,
      "theoretical_loss": 3.510007825563661,
      "tokens_seen": 1529741312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709829488465396,
      "loss": 2.4723,
      "theoretical_loss": 3.5099946174826693,
      "tokens_seen": 1529806848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027097291875626884,
      "loss": 2.6199,
      "theoretical_loss": 3.5099814101259152,
      "tokens_seen": 1529872384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270962888665998,
      "loss": 2.33,
      "theoretical_loss": 3.5099682034933286,
      "tokens_seen": 1529937920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709528585757272,
      "loss": 2.6318,
      "theoretical_loss": 3.5099549975848374,
      "tokens_seen": 1530003456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709428284854564,
      "loss": 2.8543,
      "theoretical_loss": 3.5099417924003724,
      "tokens_seen": 1530068992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027093279839518556,
      "loss": 2.5853,
      "theoretical_loss": 3.509928587939862,
      "tokens_seen": 1530134528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027092276830491474,
      "loss": 2.764,
      "theoretical_loss": 3.509915384203236,
      "tokens_seen": 1530200064
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 845259,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.503072738647461,
      "objective/train/theoretical_loss": 3.509902181190423,
      "objective/train/tokens_used": 1550725600,
      "theoretical_loss": 3.509902181190423,
      "tokens_seen": 1530265600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270912738214644,
      "loss": 2.6888,
      "theoretical_loss": 3.509902181190423,
      "tokens_seen": 1530265600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709027081243731,
      "loss": 2.3594,
      "theoretical_loss": 3.5098889789013525,
      "tokens_seen": 1530331136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027089267803410234,
      "loss": 2.6613,
      "theoretical_loss": 3.509875777335955,
      "tokens_seen": 1530396672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027088264794383147,
      "loss": 2.6888,
      "theoretical_loss": 3.5098625764941582,
      "tokens_seen": 1530462208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708726178535607,
      "loss": 2.7254,
      "theoretical_loss": 3.5098493763758922,
      "tokens_seen": 1530527744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708625877632899,
      "loss": 2.7197,
      "theoretical_loss": 3.5098361769810866,
      "tokens_seen": 1530593280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027085255767301907,
      "loss": 2.5928,
      "theoretical_loss": 3.5098229783096704,
      "tokens_seen": 1530658816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027084252758274825,
      "loss": 2.653,
      "theoretical_loss": 3.509809780361573,
      "tokens_seen": 1530724352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708324974924775,
      "loss": 2.5659,
      "theoretical_loss": 3.509796583136724,
      "tokens_seen": 1530789888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708224674022066,
      "loss": 2.5242,
      "theoretical_loss": 3.5097833866350525,
      "tokens_seen": 1530855424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027081243731193584,
      "loss": 2.5583,
      "theoretical_loss": 3.5097701908564884,
      "tokens_seen": 1530920960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027080240722166497,
      "loss": 2.7689,
      "theoretical_loss": 3.509756995800961,
      "tokens_seen": 1530986496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707923771313942,
      "loss": 2.7224,
      "theoretical_loss": 3.5097438014683995,
      "tokens_seen": 1531052032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707823470411234,
      "loss": 2.9064,
      "theoretical_loss": 3.5097306078587334,
      "tokens_seen": 1531117568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027077231695085257,
      "loss": 2.5452,
      "theoretical_loss": 3.5097174149718917,
      "tokens_seen": 1531183104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027076228686058175,
      "loss": 2.727,
      "theoretical_loss": 3.5097042228078053,
      "tokens_seen": 1531248640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027075225677031093,
      "loss": 2.4024,
      "theoretical_loss": 3.509691031366402,
      "tokens_seen": 1531314176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707422266800401,
      "loss": 2.7561,
      "theoretical_loss": 3.509677840647612,
      "tokens_seen": 1531379712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027073219658976935,
      "loss": 2.4852,
      "theoretical_loss": 3.509664650651365,
      "tokens_seen": 1531445248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707221664994985,
      "loss": 2.8186,
      "theoretical_loss": 3.50965146137759,
      "tokens_seen": 1531510784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707121364092277,
      "loss": 2.7045,
      "theoretical_loss": 3.5096382728262174,
      "tokens_seen": 1531576320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707021063189569,
      "loss": 2.5804,
      "theoretical_loss": 3.509625084997176,
      "tokens_seen": 1531641856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027069207622868607,
      "loss": 2.8062,
      "theoretical_loss": 3.5096118978903954,
      "tokens_seen": 1531707392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027068204613841525,
      "loss": 2.5919,
      "theoretical_loss": 3.509598711505805,
      "tokens_seen": 1531772928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027067201604814443,
      "loss": 2.3991,
      "theoretical_loss": 3.5095855258433346,
      "tokens_seen": 1531838464
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 846517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4351906776428223,
      "objective/train/theoretical_loss": 3.5095723409029143,
      "objective/train/tokens_used": 1552364000,
      "theoretical_loss": 3.5095723409029143,
      "tokens_seen": 1531904000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706619859578736,
      "loss": 2.5955,
      "theoretical_loss": 3.5095723409029143,
      "tokens_seen": 1531904000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027065195586760285,
      "loss": 2.7506,
      "theoretical_loss": 3.5095591566844724,
      "tokens_seen": 1531969536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270641925777332,
      "loss": 2.6533,
      "theoretical_loss": 3.5095459731879397,
      "tokens_seen": 1532035072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706318956870612,
      "loss": 2.6163,
      "theoretical_loss": 3.5095327904132447,
      "tokens_seen": 1532100608
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027062186559679034,
      "loss": 2.7896,
      "theoretical_loss": 3.509519608360318,
      "tokens_seen": 1532166144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706118355065196,
      "loss": 2.5562,
      "theoretical_loss": 3.5095064270290885,
      "tokens_seen": 1532231680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027060180541624876,
      "loss": 2.6639,
      "theoretical_loss": 3.5094932464194866,
      "tokens_seen": 1532297216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027059177532597794,
      "loss": 2.336,
      "theoretical_loss": 3.5094800665314416,
      "tokens_seen": 1532362752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705817452357071,
      "loss": 2.5906,
      "theoretical_loss": 3.5094668873648827,
      "tokens_seen": 1532428288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705717151454363,
      "loss": 2.6533,
      "theoretical_loss": 3.5094537089197395,
      "tokens_seen": 1532493824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705616850551655,
      "loss": 2.5801,
      "theoretical_loss": 3.509440531195942,
      "tokens_seen": 1532559360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705516549648947,
      "loss": 2.6365,
      "theoretical_loss": 3.5094273541934204,
      "tokens_seen": 1532624896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027054162487462384,
      "loss": 2.5615,
      "theoretical_loss": 3.5094141779121037,
      "tokens_seen": 1532690432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705315947843531,
      "loss": 2.8698,
      "theoretical_loss": 3.5094010023519218,
      "tokens_seen": 1532755968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027052156469408226,
      "loss": 2.6892,
      "theoretical_loss": 3.5093878275128043,
      "tokens_seen": 1532821504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027051153460381144,
      "loss": 2.6264,
      "theoretical_loss": 3.509374653394681,
      "tokens_seen": 1532887040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705015045135406,
      "loss": 2.7703,
      "theoretical_loss": 3.509361479997481,
      "tokens_seen": 1532952576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704914744232698,
      "loss": 2.79,
      "theoretical_loss": 3.5093483073211353,
      "tokens_seen": 1533018112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270481444332999,
      "loss": 2.7154,
      "theoretical_loss": 3.509335135365573,
      "tokens_seen": 1533083648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704714142427282,
      "loss": 2.536,
      "theoretical_loss": 3.5093219641307236,
      "tokens_seen": 1533149184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027046138415245735,
      "loss": 2.4691,
      "theoretical_loss": 3.5093087936165173,
      "tokens_seen": 1533214720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704513540621866,
      "loss": 2.4994,
      "theoretical_loss": 3.509295623822883,
      "tokens_seen": 1533280256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704413239719157,
      "loss": 2.7187,
      "theoretical_loss": 3.5092824547497514,
      "tokens_seen": 1533345792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027043129388164494,
      "loss": 2.4823,
      "theoretical_loss": 3.5092692863970525,
      "tokens_seen": 1533411328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704212637913741,
      "loss": 2.5204,
      "theoretical_loss": 3.509256118764715,
      "tokens_seen": 1533476864
    },
    {
      "epoch": 0.46,
      "objective/train/docs_used": 847177,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.847522735595703,
      "objective/train/theoretical_loss": 3.50924295185267,
      "objective/train/tokens_used": 1554002400,
      "theoretical_loss": 3.50924295185267,
      "tokens_seen": 1533542400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704112337011033,
      "loss": 2.7513,
      "theoretical_loss": 3.50924295185267,
      "tokens_seen": 1533542400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704012036108325,
      "loss": 2.4923,
      "theoretical_loss": 3.509229785660846,
      "tokens_seen": 1533607936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027039117352056167,
      "loss": 2.5013,
      "theoretical_loss": 3.5092166201891737,
      "tokens_seen": 1533673472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027038114343029085,
      "loss": 2.7783,
      "theoretical_loss": 3.5092034554375826,
      "tokens_seen": 1533739008
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703711133400201,
      "loss": 2.7859,
      "theoretical_loss": 3.5091902914060027,
      "tokens_seen": 1533804544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703610832497492,
      "loss": 2.5113,
      "theoretical_loss": 3.509177128094364,
      "tokens_seen": 1533870080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027035105315947845,
      "loss": 2.6577,
      "theoretical_loss": 3.509163965502596,
      "tokens_seen": 1533935616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027034102306920763,
      "loss": 2.4836,
      "theoretical_loss": 3.509150803630629,
      "tokens_seen": 1534001152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703309929789368,
      "loss": 2.6504,
      "theoretical_loss": 3.5091376424783927,
      "tokens_seen": 1534066688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270320962888666,
      "loss": 2.6551,
      "theoretical_loss": 3.5091244820458165,
      "tokens_seen": 1534132224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027031093279839517,
      "loss": 2.627,
      "theoretical_loss": 3.5091113223328314,
      "tokens_seen": 1534197760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027030090270812435,
      "loss": 2.5064,
      "theoretical_loss": 3.5090981633393663,
      "tokens_seen": 1534263296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702908726178536,
      "loss": 2.4783,
      "theoretical_loss": 3.5090850050653515,
      "tokens_seen": 1534328832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702808425275827,
      "loss": 2.4495,
      "theoretical_loss": 3.509071847510717,
      "tokens_seen": 1534394368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027027081243731195,
      "loss": 2.6326,
      "theoretical_loss": 3.509058690675393,
      "tokens_seen": 1534459904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027026078234704113,
      "loss": 2.6848,
      "theoretical_loss": 3.5090455345593083,
      "tokens_seen": 1534525440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702507522567703,
      "loss": 2.6913,
      "theoretical_loss": 3.5090323791623943,
      "tokens_seen": 1534590976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027024072216649955,
      "loss": 2.6024,
      "theoretical_loss": 3.5090192244845806,
      "tokens_seen": 1534656512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702306920762287,
      "loss": 2.4758,
      "theoretical_loss": 3.5090060705257966,
      "tokens_seen": 1534722048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702206619859579,
      "loss": 2.6506,
      "theoretical_loss": 3.508992917285973,
      "tokens_seen": 1534787584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702106318956871,
      "loss": 2.5259,
      "theoretical_loss": 3.5089797647650394,
      "tokens_seen": 1534853120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027020060180541627,
      "loss": 2.5529,
      "theoretical_loss": 3.508966612962926,
      "tokens_seen": 1534918656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027019057171514545,
      "loss": 2.537,
      "theoretical_loss": 3.508953461879562,
      "tokens_seen": 1534984192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027018054162487463,
      "loss": 2.718,
      "theoretical_loss": 3.5089403115148787,
      "tokens_seen": 1535049728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701705115346038,
      "loss": 2.7283,
      "theoretical_loss": 3.5089271618688054,
      "tokens_seen": 1535115264
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 848526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7075657844543457,
      "objective/train/theoretical_loss": 3.508914012941273,
      "objective/train/tokens_used": 1555640800,
      "theoretical_loss": 3.508914012941273,
      "tokens_seen": 1535180800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027016048144433305,
      "loss": 2.7353,
      "theoretical_loss": 3.508914012941273,
      "tokens_seen": 1535180800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701504513540622,
      "loss": 2.609,
      "theoretical_loss": 3.5089008647322104,
      "tokens_seen": 1535246336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701404212637914,
      "loss": 2.5519,
      "theoretical_loss": 3.5088877172415485,
      "tokens_seen": 1535311872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027013039117352054,
      "loss": 2.6782,
      "theoretical_loss": 3.5088745704692164,
      "tokens_seen": 1535377408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701203610832498,
      "loss": 2.1937,
      "theoretical_loss": 3.5088614244151453,
      "tokens_seen": 1535442944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027011033099297896,
      "loss": 2.6903,
      "theoretical_loss": 3.508848279079265,
      "tokens_seen": 1535508480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027010030090270814,
      "loss": 2.657,
      "theoretical_loss": 3.5088351344615054,
      "tokens_seen": 1535574016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700902708124373,
      "loss": 2.9042,
      "theoretical_loss": 3.508821990561797,
      "tokens_seen": 1535639552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700802407221665,
      "loss": 2.7115,
      "theoretical_loss": 3.5088088473800694,
      "tokens_seen": 1535705088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700702106318957,
      "loss": 2.4773,
      "theoretical_loss": 3.508795704916253,
      "tokens_seen": 1535770624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700601805416249,
      "loss": 2.801,
      "theoretical_loss": 3.508782563170278,
      "tokens_seen": 1535836160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027005015045135404,
      "loss": 2.537,
      "theoretical_loss": 3.5087694221420747,
      "tokens_seen": 1535901696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700401203610833,
      "loss": 2.8042,
      "theoretical_loss": 3.5087562818315727,
      "tokens_seen": 1535967232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027003009027081246,
      "loss": 2.5515,
      "theoretical_loss": 3.5087431422387025,
      "tokens_seen": 1536032768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027002006018054164,
      "loss": 2.5543,
      "theoretical_loss": 3.508730003363395,
      "tokens_seen": 1536098304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700100300902708,
      "loss": 2.5299,
      "theoretical_loss": 3.5087168652055793,
      "tokens_seen": 1536163840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027,
      "loss": 2.5026,
      "theoretical_loss": 3.5087037277651865,
      "tokens_seen": 1536229376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699899699097292,
      "loss": 2.6706,
      "theoretical_loss": 3.5086905910421455,
      "tokens_seen": 1536294912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699799398194584,
      "loss": 2.5769,
      "theoretical_loss": 3.508677455036388,
      "tokens_seen": 1536360448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026996990972918755,
      "loss": 2.4922,
      "theoretical_loss": 3.5086643197478438,
      "tokens_seen": 1536425984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699598796389168,
      "loss": 2.4464,
      "theoretical_loss": 3.5086511851764426,
      "tokens_seen": 1536491520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699498495486459,
      "loss": 2.7297,
      "theoretical_loss": 3.508638051322116,
      "tokens_seen": 1536557056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026993981945837514,
      "loss": 2.6275,
      "theoretical_loss": 3.5086249181847924,
      "tokens_seen": 1536622592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699297893681043,
      "loss": 2.5365,
      "theoretical_loss": 3.508611785764403,
      "tokens_seen": 1536688128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699197592778335,
      "loss": 2.6263,
      "theoretical_loss": 3.5085986540608785,
      "tokens_seen": 1536753664
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 849205,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6184380054473877,
      "objective/train/theoretical_loss": 3.508585523074149,
      "objective/train/tokens_used": 1557279200,
      "theoretical_loss": 3.508585523074149,
      "tokens_seen": 1536819200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699097291875627,
      "loss": 2.833,
      "theoretical_loss": 3.508585523074149,
      "tokens_seen": 1536819200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026989969909729187,
      "loss": 2.6451,
      "theoretical_loss": 3.508572392804145,
      "tokens_seen": 1536884736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026988966900702105,
      "loss": 2.4547,
      "theoretical_loss": 3.5085592632507954,
      "tokens_seen": 1536950272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698796389167503,
      "loss": 2.5255,
      "theoretical_loss": 3.508546134414032,
      "tokens_seen": 1537015808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698696088264794,
      "loss": 2.4234,
      "theoretical_loss": 3.508533006293785,
      "tokens_seen": 1537081344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026985957873620865,
      "loss": 2.6398,
      "theoretical_loss": 3.5085198788899836,
      "tokens_seen": 1537146880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026984954864593783,
      "loss": 2.4771,
      "theoretical_loss": 3.50850675220256,
      "tokens_seen": 1537212416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269839518555667,
      "loss": 2.4766,
      "theoretical_loss": 3.5084936262314432,
      "tokens_seen": 1537277952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698294884653962,
      "loss": 2.6984,
      "theoretical_loss": 3.508480500976564,
      "tokens_seen": 1537343488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026981945837512537,
      "loss": 2.5505,
      "theoretical_loss": 3.508467376437853,
      "tokens_seen": 1537409024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026980942828485455,
      "loss": 2.6991,
      "theoretical_loss": 3.50845425261524,
      "tokens_seen": 1537474560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697993981945838,
      "loss": 2.3392,
      "theoretical_loss": 3.5084411295086557,
      "tokens_seen": 1537540096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697893681043129,
      "loss": 2.5386,
      "theoretical_loss": 3.508428007118031,
      "tokens_seen": 1537605632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026977933801404215,
      "loss": 2.5687,
      "theoretical_loss": 3.5084148854432957,
      "tokens_seen": 1537671168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697693079237713,
      "loss": 2.5538,
      "theoretical_loss": 3.50840176448438,
      "tokens_seen": 1537736704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697592778335005,
      "loss": 2.4781,
      "theoretical_loss": 3.508388644241215,
      "tokens_seen": 1537802240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697492477432297,
      "loss": 2.5457,
      "theoretical_loss": 3.5083755247137316,
      "tokens_seen": 1537867776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697392176529589,
      "loss": 2.509,
      "theoretical_loss": 3.5083624059018588,
      "tokens_seen": 1537933312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026972918756268806,
      "loss": 2.6265,
      "theoretical_loss": 3.5083492878055287,
      "tokens_seen": 1537998848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697191574724173,
      "loss": 2.4365,
      "theoretical_loss": 3.50833617042467,
      "tokens_seen": 1538064384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697091273821464,
      "loss": 2.4423,
      "theoretical_loss": 3.5083230537592147,
      "tokens_seen": 1538129920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026969909729187565,
      "loss": 2.5439,
      "theoretical_loss": 3.5083099378090927,
      "tokens_seen": 1538195456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696890672016048,
      "loss": 2.6968,
      "theoretical_loss": 3.5082968225742346,
      "tokens_seen": 1538260992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269679037111334,
      "loss": 2.6952,
      "theoretical_loss": 3.508283708054571,
      "tokens_seen": 1538326528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696690070210632,
      "loss": 2.488,
      "theoretical_loss": 3.508270594250032,
      "tokens_seen": 1538392064
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 849991,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9845397472381592,
      "objective/train/theoretical_loss": 3.5082574811605487,
      "objective/train/tokens_used": 1558917600,
      "theoretical_loss": 3.5082574811605487,
      "tokens_seen": 1538457600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696589769307924,
      "loss": 2.4194,
      "theoretical_loss": 3.5082574811605487,
      "tokens_seen": 1538457600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026964894684052156,
      "loss": 2.432,
      "theoretical_loss": 3.5082443687860514,
      "tokens_seen": 1538523136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026963891675025074,
      "loss": 2.3859,
      "theoretical_loss": 3.5082312571264707,
      "tokens_seen": 1538588672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696288866599799,
      "loss": 2.4426,
      "theoretical_loss": 3.508218146181737,
      "tokens_seen": 1538654208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026961885656970916,
      "loss": 2.6011,
      "theoretical_loss": 3.5082050359517813,
      "tokens_seen": 1538719744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696088264794383,
      "loss": 2.478,
      "theoretical_loss": 3.5081919264365338,
      "tokens_seen": 1538785280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695987963891675,
      "loss": 2.5434,
      "theoretical_loss": 3.508178817635925,
      "tokens_seen": 1538850816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026958876629889665,
      "loss": 2.7497,
      "theoretical_loss": 3.508165709549886,
      "tokens_seen": 1538916352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695787362086259,
      "loss": 2.5912,
      "theoretical_loss": 3.508152602178347,
      "tokens_seen": 1538981888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026956870611835506,
      "loss": 2.557,
      "theoretical_loss": 3.508139495521239,
      "tokens_seen": 1539047424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026955867602808424,
      "loss": 2.5539,
      "theoretical_loss": 3.5081263895784924,
      "tokens_seen": 1539112960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695486459378134,
      "loss": 2.2982,
      "theoretical_loss": 3.508113284350038,
      "tokens_seen": 1539178496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026953861584754266,
      "loss": 2.8349,
      "theoretical_loss": 3.5081001798358065,
      "tokens_seen": 1539244032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695285857572718,
      "loss": 2.7094,
      "theoretical_loss": 3.5080870760357277,
      "tokens_seen": 1539309568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269518555667001,
      "loss": 2.5591,
      "theoretical_loss": 3.5080739729497337,
      "tokens_seen": 1539375104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695085255767302,
      "loss": 2.4333,
      "theoretical_loss": 3.508060870577754,
      "tokens_seen": 1539440640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694984954864594,
      "loss": 2.5765,
      "theoretical_loss": 3.50804776891972,
      "tokens_seen": 1539506176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694884653961886,
      "loss": 2.84,
      "theoretical_loss": 3.5080346679755623,
      "tokens_seen": 1539571712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026947843530591775,
      "loss": 2.3706,
      "theoretical_loss": 3.508021567745211,
      "tokens_seen": 1539637248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269468405215647,
      "loss": 2.644,
      "theoretical_loss": 3.5080084682285975,
      "tokens_seen": 1539702784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694583751253761,
      "loss": 2.5932,
      "theoretical_loss": 3.5079953694256525,
      "tokens_seen": 1539768320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026944834503510534,
      "loss": 2.496,
      "theoretical_loss": 3.5079822713363065,
      "tokens_seen": 1539833856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694383149448345,
      "loss": 2.6042,
      "theoretical_loss": 3.5079691739604906,
      "tokens_seen": 1539899392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694282848545637,
      "loss": 2.6964,
      "theoretical_loss": 3.507956077298135,
      "tokens_seen": 1539964928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694182547642929,
      "loss": 2.7292,
      "theoretical_loss": 3.5079429813491716,
      "tokens_seen": 1540030464
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 850977,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2120277881622314,
      "objective/train/theoretical_loss": 3.5079298861135295,
      "objective/train/tokens_used": 1560556000,
      "theoretical_loss": 3.5079298861135295,
      "tokens_seen": 1540096000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026940822467402207,
      "loss": 2.6,
      "theoretical_loss": 3.5079298861135295,
      "tokens_seen": 1540096000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026939819458375125,
      "loss": 2.5619,
      "theoretical_loss": 3.507916791591141,
      "tokens_seen": 1540161536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693881644934805,
      "loss": 2.6881,
      "theoretical_loss": 3.507903697781936,
      "tokens_seen": 1540227072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693781344032096,
      "loss": 2.5687,
      "theoretical_loss": 3.507890604685845,
      "tokens_seen": 1540292608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026936810431293885,
      "loss": 2.6343,
      "theoretical_loss": 3.5078775123028003,
      "tokens_seen": 1540358144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026935807422266803,
      "loss": 2.7106,
      "theoretical_loss": 3.507864420632732,
      "tokens_seen": 1540423680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693480441323972,
      "loss": 2.4071,
      "theoretical_loss": 3.50785132967557,
      "tokens_seen": 1540489216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693380140421264,
      "loss": 2.624,
      "theoretical_loss": 3.5078382394312464,
      "tokens_seen": 1540554752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026932798395185557,
      "loss": 2.5093,
      "theoretical_loss": 3.5078251498996913,
      "tokens_seen": 1540620288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026931795386158475,
      "loss": 2.9694,
      "theoretical_loss": 3.5078120610808363,
      "tokens_seen": 1540685824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269307923771314,
      "loss": 2.6383,
      "theoretical_loss": 3.5077989729746117,
      "tokens_seen": 1540751360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692978936810431,
      "loss": 2.7317,
      "theoretical_loss": 3.5077858855809487,
      "tokens_seen": 1540816896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026928786359077235,
      "loss": 2.657,
      "theoretical_loss": 3.5077727988997776,
      "tokens_seen": 1540882432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692778335005015,
      "loss": 2.4937,
      "theoretical_loss": 3.5077597129310307,
      "tokens_seen": 1540947968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692678034102307,
      "loss": 2.5744,
      "theoretical_loss": 3.507746627674637,
      "tokens_seen": 1541013504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692577733199599,
      "loss": 2.6681,
      "theoretical_loss": 3.507733543130529,
      "tokens_seen": 1541079040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692477432296891,
      "loss": 2.5646,
      "theoretical_loss": 3.5077204592986364,
      "tokens_seen": 1541144576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026923771313941826,
      "loss": 2.4769,
      "theoretical_loss": 3.5077073761788915,
      "tokens_seen": 1541210112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692276830491475,
      "loss": 2.6539,
      "theoretical_loss": 3.507694293771224,
      "tokens_seen": 1541275648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692176529588766,
      "loss": 2.5781,
      "theoretical_loss": 3.5076812120755663,
      "tokens_seen": 1541341184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026920762286860585,
      "loss": 2.6934,
      "theoretical_loss": 3.507668131091848,
      "tokens_seen": 1541406720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269197592778335,
      "loss": 2.8051,
      "theoretical_loss": 3.5076550508200004,
      "tokens_seen": 1541472256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691875626880642,
      "loss": 2.6751,
      "theoretical_loss": 3.5076419712599547,
      "tokens_seen": 1541537792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691775325977934,
      "loss": 2.4307,
      "theoretical_loss": 3.5076288924116423,
      "tokens_seen": 1541603328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691675025075226,
      "loss": 2.5462,
      "theoretical_loss": 3.5076158142749936,
      "tokens_seen": 1541668864
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 851654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2880167961120605,
      "objective/train/theoretical_loss": 3.50760273684994,
      "objective/train/tokens_used": 1562194400,
      "theoretical_loss": 3.50760273684994,
      "tokens_seen": 1541734400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026915747241725176,
      "loss": 2.6604,
      "theoretical_loss": 3.50760273684994,
      "tokens_seen": 1541734400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026914744232698094,
      "loss": 2.5,
      "theoretical_loss": 3.5075896601364125,
      "tokens_seen": 1541799936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691374122367101,
      "loss": 2.7159,
      "theoretical_loss": 3.5075765841343416,
      "tokens_seen": 1541865472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026912738214643936,
      "loss": 2.4402,
      "theoretical_loss": 3.507563508843659,
      "tokens_seen": 1541931008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691173520561685,
      "loss": 2.7302,
      "theoretical_loss": 3.5075504342642954,
      "tokens_seen": 1541996544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691073219658977,
      "loss": 2.508,
      "theoretical_loss": 3.5075373603961824,
      "tokens_seen": 1542062080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026909729187562685,
      "loss": 2.6313,
      "theoretical_loss": 3.5075242872392502,
      "tokens_seen": 1542127616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690872617853561,
      "loss": 2.4254,
      "theoretical_loss": 3.5075112147934306,
      "tokens_seen": 1542193152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026907723169508526,
      "loss": 2.6111,
      "theoretical_loss": 3.5074981430586547,
      "tokens_seen": 1542258688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026906720160481444,
      "loss": 2.4878,
      "theoretical_loss": 3.5074850720348527,
      "tokens_seen": 1542324224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690571715145436,
      "loss": 2.5306,
      "theoretical_loss": 3.507472001721957,
      "tokens_seen": 1542389760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026904714142427286,
      "loss": 2.7115,
      "theoretical_loss": 3.5074589321198983,
      "tokens_seen": 1542455296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269037111334002,
      "loss": 2.6391,
      "theoretical_loss": 3.5074458632286074,
      "tokens_seen": 1542520832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690270812437312,
      "loss": 2.7007,
      "theoretical_loss": 3.5074327950480155,
      "tokens_seen": 1542586368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026901705115346035,
      "loss": 2.4498,
      "theoretical_loss": 3.5074197275780543,
      "tokens_seen": 1542651904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690070210631896,
      "loss": 2.6167,
      "theoretical_loss": 3.5074066608186545,
      "tokens_seen": 1542717440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026899699097291877,
      "loss": 2.4514,
      "theoretical_loss": 3.5073935947697468,
      "tokens_seen": 1542782976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026898696088264795,
      "loss": 2.5525,
      "theoretical_loss": 3.5073805294312637,
      "tokens_seen": 1542848512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026897693079237713,
      "loss": 2.4901,
      "theoretical_loss": 3.507367464803135,
      "tokens_seen": 1542914048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689669007021063,
      "loss": 2.6087,
      "theoretical_loss": 3.5073544008852933,
      "tokens_seen": 1542979584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689568706118355,
      "loss": 2.6454,
      "theoretical_loss": 3.5073413376776683,
      "tokens_seen": 1543045120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689468405215647,
      "loss": 2.5948,
      "theoretical_loss": 3.5073282751801926,
      "tokens_seen": 1543110656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026893681043129385,
      "loss": 2.6008,
      "theoretical_loss": 3.5073152133927965,
      "tokens_seen": 1543176192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689267803410231,
      "loss": 2.6828,
      "theoretical_loss": 3.507302152315412,
      "tokens_seen": 1543241728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689167502507522,
      "loss": 2.6454,
      "theoretical_loss": 3.5072890919479693,
      "tokens_seen": 1543307264
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 852770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.775528907775879,
      "objective/train/theoretical_loss": 3.5072760322904006,
      "objective/train/tokens_used": 1563832800,
      "theoretical_loss": 3.5072760322904006,
      "tokens_seen": 1543372800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026890672016048145,
      "loss": 2.7479,
      "theoretical_loss": 3.5072760322904006,
      "tokens_seen": 1543372800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026889669007021063,
      "loss": 2.7586,
      "theoretical_loss": 3.5072629733426375,
      "tokens_seen": 1543438336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688866599799398,
      "loss": 2.6097,
      "theoretical_loss": 3.5072499151046097,
      "tokens_seen": 1543503872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268876629889669,
      "loss": 2.868,
      "theoretical_loss": 3.50723685757625,
      "tokens_seen": 1543569408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026886659979939823,
      "loss": 2.4749,
      "theoretical_loss": 3.5072238007574894,
      "tokens_seen": 1543634944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026885656970912736,
      "loss": 2.523,
      "theoretical_loss": 3.5072107446482583,
      "tokens_seen": 1543700480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688465396188566,
      "loss": 2.4207,
      "theoretical_loss": 3.507197689248489,
      "tokens_seen": 1543766016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688365095285857,
      "loss": 2.456,
      "theoretical_loss": 3.5071846345581124,
      "tokens_seen": 1543831552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026882647943831495,
      "loss": 2.3869,
      "theoretical_loss": 3.5071715805770602,
      "tokens_seen": 1543897088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026881644934804413,
      "loss": 2.608,
      "theoretical_loss": 3.507158527305264,
      "tokens_seen": 1543962624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688064192577733,
      "loss": 2.6021,
      "theoretical_loss": 3.507145474742654,
      "tokens_seen": 1544028160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687963891675025,
      "loss": 2.5998,
      "theoretical_loss": 3.5071324228891623,
      "tokens_seen": 1544093696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687863590772317,
      "loss": 2.4491,
      "theoretical_loss": 3.5071193717447207,
      "tokens_seen": 1544159232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026877632898696086,
      "loss": 2.6654,
      "theoretical_loss": 3.5071063213092595,
      "tokens_seen": 1544224768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687662988966901,
      "loss": 2.4587,
      "theoretical_loss": 3.5070932715827112,
      "tokens_seen": 1544290304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687562688064193,
      "loss": 2.6008,
      "theoretical_loss": 3.5070802225650066,
      "tokens_seen": 1544355840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026874623871614846,
      "loss": 2.3871,
      "theoretical_loss": 3.5070671742560773,
      "tokens_seen": 1544421376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687362086258777,
      "loss": 2.6996,
      "theoretical_loss": 3.507054126655855,
      "tokens_seen": 1544486912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687261785356068,
      "loss": 2.3082,
      "theoretical_loss": 3.50704107976427,
      "tokens_seen": 1544552448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026871614844533605,
      "loss": 2.781,
      "theoretical_loss": 3.507028033581255,
      "tokens_seen": 1544617984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687061183550652,
      "loss": 2.6394,
      "theoretical_loss": 3.507014988106741,
      "tokens_seen": 1544683520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686960882647944,
      "loss": 2.8959,
      "theoretical_loss": 3.5070019433406596,
      "tokens_seen": 1544749056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686860581745236,
      "loss": 2.5596,
      "theoretical_loss": 3.506988899282942,
      "tokens_seen": 1544814592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686760280842528,
      "loss": 2.7267,
      "theoretical_loss": 3.5069758559335202,
      "tokens_seen": 1544880128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026866599799398196,
      "loss": 2.437,
      "theoretical_loss": 3.5069628132923256,
      "tokens_seen": 1544945664
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 853806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.327728509902954,
      "objective/train/theoretical_loss": 3.506949771359289,
      "objective/train/tokens_used": 1565471200,
      "theoretical_loss": 3.506949771359289,
      "tokens_seen": 1545011200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026865596790371114,
      "loss": 2.5891,
      "theoretical_loss": 3.506949771359289,
      "tokens_seen": 1545011200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686459378134403,
      "loss": 2.6977,
      "theoretical_loss": 3.5069367301343424,
      "tokens_seen": 1545076736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026863590772316956,
      "loss": 2.5861,
      "theoretical_loss": 3.506923689617417,
      "tokens_seen": 1545142272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686258776328987,
      "loss": 2.8947,
      "theoretical_loss": 3.5069106498084452,
      "tokens_seen": 1545207808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686158475426279,
      "loss": 2.6797,
      "theoretical_loss": 3.506897610707358,
      "tokens_seen": 1545273344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026860581745235705,
      "loss": 2.5427,
      "theoretical_loss": 3.5068845723140862,
      "tokens_seen": 1545338880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685957873620863,
      "loss": 2.569,
      "theoretical_loss": 3.5068715346285626,
      "tokens_seen": 1545404416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026858575727181546,
      "loss": 2.6444,
      "theoretical_loss": 3.506858497650718,
      "tokens_seen": 1545469952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026857572718154464,
      "loss": 2.5422,
      "theoretical_loss": 3.5068454613804843,
      "tokens_seen": 1545535488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685656970912738,
      "loss": 2.5319,
      "theoretical_loss": 3.5068324258177936,
      "tokens_seen": 1545601024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026855566700100306,
      "loss": 2.5023,
      "theoretical_loss": 3.506819390962576,
      "tokens_seen": 1545666560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685456369107322,
      "loss": 2.7061,
      "theoretical_loss": 3.506806356814765,
      "tokens_seen": 1545732096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685356068204614,
      "loss": 2.6222,
      "theoretical_loss": 3.5067933233742905,
      "tokens_seen": 1545797632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026852557673019055,
      "loss": 2.5943,
      "theoretical_loss": 3.5067802906410854,
      "tokens_seen": 1545863168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685155466399198,
      "loss": 2.6763,
      "theoretical_loss": 3.5067672586150804,
      "tokens_seen": 1545928704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026850551654964897,
      "loss": 2.4709,
      "theoretical_loss": 3.5067542272962076,
      "tokens_seen": 1545994240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026849548645937815,
      "loss": 2.4995,
      "theoretical_loss": 3.506741196684399,
      "tokens_seen": 1546059776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026848545636910733,
      "loss": 2.7153,
      "theoretical_loss": 3.5067281667795855,
      "tokens_seen": 1546125312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684754262788365,
      "loss": 2.5729,
      "theoretical_loss": 3.5067151375816996,
      "tokens_seen": 1546190848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684653961885657,
      "loss": 2.54,
      "theoretical_loss": 3.5067021090906723,
      "tokens_seen": 1546256384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684553660982949,
      "loss": 2.6817,
      "theoretical_loss": 3.506689081306435,
      "tokens_seen": 1546321920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026844533600802405,
      "loss": 2.5134,
      "theoretical_loss": 3.5066760542289206,
      "tokens_seen": 1546387456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684353059177533,
      "loss": 2.3954,
      "theoretical_loss": 3.5066630278580604,
      "tokens_seen": 1546452992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684252758274824,
      "loss": 2.5135,
      "theoretical_loss": 3.506650002193785,
      "tokens_seen": 1546518528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026841524573721165,
      "loss": 2.7969,
      "theoretical_loss": 3.506636977236028,
      "tokens_seen": 1546584064
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 854346,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0632946491241455,
      "objective/train/theoretical_loss": 3.50662395298472,
      "objective/train/tokens_used": 1567109600,
      "theoretical_loss": 3.50662395298472,
      "tokens_seen": 1546649600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026840521564694083,
      "loss": 2.4863,
      "theoretical_loss": 3.50662395298472,
      "tokens_seen": 1546649600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026839518555667,
      "loss": 2.6788,
      "theoretical_loss": 3.506610929439792,
      "tokens_seen": 1546715136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683851554663992,
      "loss": 2.7006,
      "theoretical_loss": 3.5065979066011774,
      "tokens_seen": 1546780672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026837512537612843,
      "loss": 2.6099,
      "theoretical_loss": 3.5065848844688077,
      "tokens_seen": 1546846208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026836509528585756,
      "loss": 2.6737,
      "theoretical_loss": 3.5065718630426135,
      "tokens_seen": 1546911744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683550651955868,
      "loss": 2.7256,
      "theoretical_loss": 3.5065588423225273,
      "tokens_seen": 1546977280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683450351053159,
      "loss": 2.5789,
      "theoretical_loss": 3.506545822308481,
      "tokens_seen": 1547042816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026833500501504515,
      "loss": 2.5767,
      "theoretical_loss": 3.5065328030004066,
      "tokens_seen": 1547108352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026832497492477433,
      "loss": 2.6697,
      "theoretical_loss": 3.506519784398235,
      "tokens_seen": 1547173888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683149448345035,
      "loss": 2.9253,
      "theoretical_loss": 3.5065067665018996,
      "tokens_seen": 1547239424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683049147442327,
      "loss": 2.7083,
      "theoretical_loss": 3.506493749311331,
      "tokens_seen": 1547304960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682948846539619,
      "loss": 2.4743,
      "theoretical_loss": 3.506480732826461,
      "tokens_seen": 1547370496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026828485456369106,
      "loss": 2.6113,
      "theoretical_loss": 3.5064677170472223,
      "tokens_seen": 1547436032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682748244734203,
      "loss": 2.4489,
      "theoretical_loss": 3.5064547019735457,
      "tokens_seen": 1547501568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682647943831494,
      "loss": 2.6039,
      "theoretical_loss": 3.5064416876053643,
      "tokens_seen": 1547567104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026825476429287866,
      "loss": 2.666,
      "theoretical_loss": 3.506428673942609,
      "tokens_seen": 1547632640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026825476429287866,
      "loss": 2.5713,
      "theoretical_loss": 3.506415660985212,
      "tokens_seen": 1547698176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682447342026078,
      "loss": 2.5002,
      "theoretical_loss": 3.506402648733105,
      "tokens_seen": 1547763712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268234704112337,
      "loss": 2.6364,
      "theoretical_loss": 3.5063896371862207,
      "tokens_seen": 1547829248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682246740220662,
      "loss": 2.5491,
      "theoretical_loss": 3.5063766263444895,
      "tokens_seen": 1547894784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682146439317954,
      "loss": 2.334,
      "theoretical_loss": 3.506363616207845,
      "tokens_seen": 1547960320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026820461384152456,
      "loss": 2.5686,
      "theoretical_loss": 3.506350606776219,
      "tokens_seen": 1548025856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681945837512538,
      "loss": 2.6391,
      "theoretical_loss": 3.5063375980495417,
      "tokens_seen": 1548091392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681845536609829,
      "loss": 2.6821,
      "theoretical_loss": 3.506324590027747,
      "tokens_seen": 1548156928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026817452357071216,
      "loss": 2.8725,
      "theoretical_loss": 3.506311582710766,
      "tokens_seen": 1548222464
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 855711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7896533012390137,
      "objective/train/theoretical_loss": 3.506298576098531,
      "objective/train/tokens_used": 1568748000,
      "theoretical_loss": 3.506298576098531,
      "tokens_seen": 1548288000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681644934804413,
      "loss": 2.6192,
      "theoretical_loss": 3.506298576098531,
      "tokens_seen": 1548288000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681544633901705,
      "loss": 2.4303,
      "theoretical_loss": 3.506285570190973,
      "tokens_seen": 1548353536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681444332998997,
      "loss": 2.5779,
      "theoretical_loss": 3.5062725649880253,
      "tokens_seen": 1548419072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681344032096289,
      "loss": 2.6168,
      "theoretical_loss": 3.5062595604896196,
      "tokens_seen": 1548484608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026812437311935807,
      "loss": 2.5535,
      "theoretical_loss": 3.5062465566956877,
      "tokens_seen": 1548550144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026811434302908725,
      "loss": 2.5247,
      "theoretical_loss": 3.506233553606161,
      "tokens_seen": 1548615680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026810431293881643,
      "loss": 2.5174,
      "theoretical_loss": 3.5062205512209728,
      "tokens_seen": 1548681216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026809428284854566,
      "loss": 2.5365,
      "theoretical_loss": 3.506207549540054,
      "tokens_seen": 1548746752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680842527582748,
      "loss": 2.6277,
      "theoretical_loss": 3.5061945485633377,
      "tokens_seen": 1548812288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268074222668004,
      "loss": 2.4795,
      "theoretical_loss": 3.5061815482907552,
      "tokens_seen": 1548877824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026806419257773315,
      "loss": 2.594,
      "theoretical_loss": 3.506168548722239,
      "tokens_seen": 1548943360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680541624874624,
      "loss": 2.5768,
      "theoretical_loss": 3.506155549857721,
      "tokens_seen": 1549008896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026804413239719157,
      "loss": 2.6516,
      "theoretical_loss": 3.506142551697133,
      "tokens_seen": 1549074432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026803410230692075,
      "loss": 2.5582,
      "theoretical_loss": 3.506129554240407,
      "tokens_seen": 1549139968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026802407221664993,
      "loss": 2.7333,
      "theoretical_loss": 3.506116557487476,
      "tokens_seen": 1549205504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026801404212637917,
      "loss": 2.6507,
      "theoretical_loss": 3.5061035614382714,
      "tokens_seen": 1549271040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026800401203610835,
      "loss": 2.5752,
      "theoretical_loss": 3.506090566092726,
      "tokens_seen": 1549336576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026799398194583753,
      "loss": 2.5158,
      "theoretical_loss": 3.5060775714507715,
      "tokens_seen": 1549402112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679839518555667,
      "loss": 2.6866,
      "theoretical_loss": 3.5060645775123396,
      "tokens_seen": 1549467648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679739217652959,
      "loss": 2.6036,
      "theoretical_loss": 3.5060515842773627,
      "tokens_seen": 1549533184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679638916750251,
      "loss": 2.4638,
      "theoretical_loss": 3.5060385917457735,
      "tokens_seen": 1549598720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026795386158475425,
      "loss": 2.7446,
      "theoretical_loss": 3.506025599917504,
      "tokens_seen": 1549664256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679438314944835,
      "loss": 2.5557,
      "theoretical_loss": 3.506012608792486,
      "tokens_seen": 1549729792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679338014042126,
      "loss": 2.6441,
      "theoretical_loss": 3.5059996183706517,
      "tokens_seen": 1549795328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026792377131394185,
      "loss": 2.5275,
      "theoretical_loss": 3.5059866286519337,
      "tokens_seen": 1549860864
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 856331,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5446994304656982,
      "objective/train/theoretical_loss": 3.505973639636264,
      "objective/train/tokens_used": 1570386400,
      "theoretical_loss": 3.505973639636264,
      "tokens_seen": 1549926400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026791374122367103,
      "loss": 2.5209,
      "theoretical_loss": 3.505973639636264,
      "tokens_seen": 1549926400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002679037111334002,
      "loss": 2.7458,
      "theoretical_loss": 3.505960651323575,
      "tokens_seen": 1549991936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678936810431294,
      "loss": 2.6458,
      "theoretical_loss": 3.5059476637137985,
      "tokens_seen": 1550057472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026788365095285863,
      "loss": 2.5634,
      "theoretical_loss": 3.5059346768068673,
      "tokens_seen": 1550123008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026787362086258776,
      "loss": 2.3974,
      "theoretical_loss": 3.5059216906027135,
      "tokens_seen": 1550188544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267863590772317,
      "loss": 2.2677,
      "theoretical_loss": 3.505908705101269,
      "tokens_seen": 1550254080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678535606820461,
      "loss": 2.5375,
      "theoretical_loss": 3.5058957203024663,
      "tokens_seen": 1550319616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026784353059177535,
      "loss": 2.4942,
      "theoretical_loss": 3.505882736206238,
      "tokens_seen": 1550385152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026783350050150453,
      "loss": 2.5636,
      "theoretical_loss": 3.505869752812516,
      "tokens_seen": 1550450688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678234704112337,
      "loss": 2.7162,
      "theoretical_loss": 3.505856770121232,
      "tokens_seen": 1550516224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678134403209629,
      "loss": 2.6569,
      "theoretical_loss": 3.50584378813232,
      "tokens_seen": 1550581760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678034102306921,
      "loss": 2.5652,
      "theoretical_loss": 3.5058308068457107,
      "tokens_seen": 1550647296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026779338014042126,
      "loss": 2.7396,
      "theoretical_loss": 3.505817826261337,
      "tokens_seen": 1550712832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677833500501505,
      "loss": 2.7289,
      "theoretical_loss": 3.5058048463791316,
      "tokens_seen": 1550778368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677733199598796,
      "loss": 2.6986,
      "theoretical_loss": 3.5057918671990267,
      "tokens_seen": 1550843904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026776328986960886,
      "loss": 2.6967,
      "theoretical_loss": 3.505778888720954,
      "tokens_seen": 1550909440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267753259779338,
      "loss": 2.6413,
      "theoretical_loss": 3.5057659109448465,
      "tokens_seen": 1550974976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677432296890672,
      "loss": 2.5232,
      "theoretical_loss": 3.5057529338706366,
      "tokens_seen": 1551040512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677331995987964,
      "loss": 2.5715,
      "theoretical_loss": 3.505739957498256,
      "tokens_seen": 1551106048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677231695085256,
      "loss": 2.5834,
      "theoretical_loss": 3.505726981827638,
      "tokens_seen": 1551171584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026771313941825476,
      "loss": 2.6938,
      "theoretical_loss": 3.5057140068587147,
      "tokens_seen": 1551237120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267703109327984,
      "loss": 2.5465,
      "theoretical_loss": 3.5057010325914177,
      "tokens_seen": 1551302656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676930792377131,
      "loss": 2.5455,
      "theoretical_loss": 3.505688059025681,
      "tokens_seen": 1551368192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026768304914744236,
      "loss": 2.7031,
      "theoretical_loss": 3.505675086161436,
      "tokens_seen": 1551433728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676730190571715,
      "loss": 2.6595,
      "theoretical_loss": 3.505662113998615,
      "tokens_seen": 1551499264
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 857659,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6628549098968506,
      "objective/train/theoretical_loss": 3.505649142537151,
      "objective/train/tokens_used": 1572024800,
      "theoretical_loss": 3.505649142537151,
      "tokens_seen": 1551564800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676629889669007,
      "loss": 2.6416,
      "theoretical_loss": 3.505649142537151,
      "tokens_seen": 1551564800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676529588766299,
      "loss": 2.7571,
      "theoretical_loss": 3.5056361717769757,
      "tokens_seen": 1551630336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676429287863591,
      "loss": 2.6626,
      "theoretical_loss": 3.5056232017180227,
      "tokens_seen": 1551695872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026763289869608827,
      "loss": 2.7074,
      "theoretical_loss": 3.5056102323602234,
      "tokens_seen": 1551761408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026762286860581745,
      "loss": 2.8588,
      "theoretical_loss": 3.5055972637035104,
      "tokens_seen": 1551826944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026761283851554663,
      "loss": 2.4444,
      "theoretical_loss": 3.505584295747817,
      "tokens_seen": 1551892480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026760280842527586,
      "loss": 2.5962,
      "theoretical_loss": 3.5055713284930756,
      "tokens_seen": 1551958016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267592778335005,
      "loss": 2.6036,
      "theoretical_loss": 3.505558361939218,
      "tokens_seen": 1552023552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675827482447342,
      "loss": 2.5306,
      "theoretical_loss": 3.505545396086177,
      "tokens_seen": 1552089088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026757271815446335,
      "loss": 2.6293,
      "theoretical_loss": 3.5055324309338856,
      "tokens_seen": 1552154624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675626880641926,
      "loss": 2.692,
      "theoretical_loss": 3.5055194664822755,
      "tokens_seen": 1552220160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026755265797392177,
      "loss": 2.6127,
      "theoretical_loss": 3.5055065027312797,
      "tokens_seen": 1552285696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026754262788365095,
      "loss": 2.6598,
      "theoretical_loss": 3.505493539680831,
      "tokens_seen": 1552351232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026753259779338013,
      "loss": 2.4989,
      "theoretical_loss": 3.5054805773308613,
      "tokens_seen": 1552416768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026752256770310937,
      "loss": 2.7468,
      "theoretical_loss": 3.5054676156813036,
      "tokens_seen": 1552482304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675125376128385,
      "loss": 2.6931,
      "theoretical_loss": 3.505454654732091,
      "tokens_seen": 1552547840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026750250752256773,
      "loss": 2.2726,
      "theoretical_loss": 3.5054416944831557,
      "tokens_seen": 1552613376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026749247743229686,
      "loss": 2.5311,
      "theoretical_loss": 3.50542873493443,
      "tokens_seen": 1552678912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674824473420261,
      "loss": 2.5356,
      "theoretical_loss": 3.5054157760858464,
      "tokens_seen": 1552744448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026747241725175527,
      "loss": 2.5747,
      "theoretical_loss": 3.505402817937338,
      "tokens_seen": 1552809984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026746238716148445,
      "loss": 2.6245,
      "theoretical_loss": 3.5053898604888376,
      "tokens_seen": 1552875520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026745235707121363,
      "loss": 2.5649,
      "theoretical_loss": 3.5053769037402773,
      "tokens_seen": 1552941056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674423269809428,
      "loss": 2.6746,
      "theoretical_loss": 3.50536394769159,
      "tokens_seen": 1553006592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267432296890672,
      "loss": 2.7146,
      "theoretical_loss": 3.5053509923427084,
      "tokens_seen": 1553072128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026742226680040123,
      "loss": 2.6613,
      "theoretical_loss": 3.505338037693565,
      "tokens_seen": 1553137664
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 858740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4517948627471924,
      "objective/train/theoretical_loss": 3.5053250837440926,
      "objective/train/tokens_used": 1573663200,
      "theoretical_loss": 3.5053250837440926,
      "tokens_seen": 1553203200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026741223671013036,
      "loss": 2.7005,
      "theoretical_loss": 3.5053250837440926,
      "tokens_seen": 1553203200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674022066198596,
      "loss": 2.5584,
      "theoretical_loss": 3.505312130494224,
      "tokens_seen": 1553268736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673921765295888,
      "loss": 2.6477,
      "theoretical_loss": 3.5052991779438916,
      "tokens_seen": 1553334272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026738214643931796,
      "loss": 2.5397,
      "theoretical_loss": 3.5052862260930286,
      "tokens_seen": 1553399808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026737211634904714,
      "loss": 2.7688,
      "theoretical_loss": 3.505273274941567,
      "tokens_seen": 1553465344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673620862587763,
      "loss": 2.6444,
      "theoretical_loss": 3.5052603244894405,
      "tokens_seen": 1553530880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673520561685055,
      "loss": 2.7496,
      "theoretical_loss": 3.505247374736581,
      "tokens_seen": 1553596416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026734202607823474,
      "loss": 2.5094,
      "theoretical_loss": 3.5052344256829215,
      "tokens_seen": 1553661952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026733199598796386,
      "loss": 2.682,
      "theoretical_loss": 3.505221477328395,
      "tokens_seen": 1553727488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673219658976931,
      "loss": 2.7016,
      "theoretical_loss": 3.5052085296729336,
      "tokens_seen": 1553793024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673119358074222,
      "loss": 2.7309,
      "theoretical_loss": 3.505195582716471,
      "tokens_seen": 1553858560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026730190571715146,
      "loss": 2.4301,
      "theoretical_loss": 3.5051826364589393,
      "tokens_seen": 1553924096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026729187562688064,
      "loss": 2.4826,
      "theoretical_loss": 3.5051696909002716,
      "tokens_seen": 1553989632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672818455366098,
      "loss": 2.5863,
      "theoretical_loss": 3.5051567460404005,
      "tokens_seen": 1554055168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267271815446339,
      "loss": 2.6842,
      "theoretical_loss": 3.5051438018792593,
      "tokens_seen": 1554120704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672617853560682,
      "loss": 2.4586,
      "theoretical_loss": 3.5051308584167806,
      "tokens_seen": 1554186240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672517552657974,
      "loss": 2.6992,
      "theoretical_loss": 3.5051179156528964,
      "tokens_seen": 1554251776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672417251755266,
      "loss": 2.6247,
      "theoretical_loss": 3.5051049735875406,
      "tokens_seen": 1554317312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672316950852558,
      "loss": 2.5235,
      "theoretical_loss": 3.5050920322206456,
      "tokens_seen": 1554382848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026722166499498496,
      "loss": 2.4592,
      "theoretical_loss": 3.5050790915521444,
      "tokens_seen": 1554448384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672116349047142,
      "loss": 2.7846,
      "theoretical_loss": 3.50506615158197,
      "tokens_seen": 1554513920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672016048144433,
      "loss": 2.4958,
      "theoretical_loss": 3.5050532123100546,
      "tokens_seen": 1554579456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026719157472417256,
      "loss": 2.8148,
      "theoretical_loss": 3.5050402737363315,
      "tokens_seen": 1554644992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671815446339017,
      "loss": 2.7222,
      "theoretical_loss": 3.5050273358607344,
      "tokens_seen": 1554710528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671715145436309,
      "loss": 2.5004,
      "theoretical_loss": 3.505014398683195,
      "tokens_seen": 1554776064
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 859255,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.550337314605713,
      "objective/train/theoretical_loss": 3.505001462203647,
      "objective/train/tokens_used": 1575301600,
      "theoretical_loss": 3.505001462203647,
      "tokens_seen": 1554841600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671614844533601,
      "loss": 2.6438,
      "theoretical_loss": 3.505001462203647,
      "tokens_seen": 1554841600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671514543630893,
      "loss": 2.6983,
      "theoretical_loss": 3.5049885264220224,
      "tokens_seen": 1554907136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026714142427281847,
      "loss": 2.7964,
      "theoretical_loss": 3.5049755913382548,
      "tokens_seen": 1554972672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026713139418254765,
      "loss": 2.6796,
      "theoretical_loss": 3.504962656952278,
      "tokens_seen": 1555038208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026712136409227683,
      "loss": 2.4732,
      "theoretical_loss": 3.504949723264023,
      "tokens_seen": 1555103744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026711133400200606,
      "loss": 2.6111,
      "theoretical_loss": 3.5049367902734243,
      "tokens_seen": 1555169280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671013039117352,
      "loss": 2.6091,
      "theoretical_loss": 3.5049238579804145,
      "tokens_seen": 1555234816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670912738214644,
      "loss": 2.747,
      "theoretical_loss": 3.504910926384926,
      "tokens_seen": 1555300352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026708124373119355,
      "loss": 2.6388,
      "theoretical_loss": 3.5048979954868926,
      "tokens_seen": 1555365888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670712136409228,
      "loss": 2.6924,
      "theoretical_loss": 3.504885065286247,
      "tokens_seen": 1555431424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026706118355065197,
      "loss": 2.5003,
      "theoretical_loss": 3.5048721357829216,
      "tokens_seen": 1555496960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026705115346038115,
      "loss": 2.7032,
      "theoretical_loss": 3.5048592069768505,
      "tokens_seen": 1555562496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026704112337011033,
      "loss": 2.7028,
      "theoretical_loss": 3.5048462788679657,
      "tokens_seen": 1555628032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026703109327983957,
      "loss": 2.4155,
      "theoretical_loss": 3.5048333514562016,
      "tokens_seen": 1555693568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670210631895687,
      "loss": 2.7196,
      "theoretical_loss": 3.5048204247414896,
      "tokens_seen": 1555759104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026701103309929793,
      "loss": 2.7427,
      "theoretical_loss": 3.5048074987237636,
      "tokens_seen": 1555824640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026700100300902706,
      "loss": 2.632,
      "theoretical_loss": 3.5047945734029566,
      "tokens_seen": 1555890176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669909729187563,
      "loss": 2.6611,
      "theoretical_loss": 3.504781648779002,
      "tokens_seen": 1555955712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026698094282848547,
      "loss": 2.6341,
      "theoretical_loss": 3.504768724851832,
      "tokens_seen": 1556021248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026697091273821465,
      "loss": 2.6446,
      "theoretical_loss": 3.5047558016213807,
      "tokens_seen": 1556086784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026696088264794383,
      "loss": 2.684,
      "theoretical_loss": 3.5047428790875808,
      "tokens_seen": 1556152320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266950852557673,
      "loss": 2.5949,
      "theoretical_loss": 3.504729957250365,
      "tokens_seen": 1556217856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669408224674022,
      "loss": 2.5733,
      "theoretical_loss": 3.5047170361096667,
      "tokens_seen": 1556283392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026693079237713143,
      "loss": 2.6359,
      "theoretical_loss": 3.5047041156654193,
      "tokens_seen": 1556348928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026692076228686056,
      "loss": 2.283,
      "theoretical_loss": 3.5046911959175553,
      "tokens_seen": 1556414464
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 860458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.509343385696411,
      "objective/train/theoretical_loss": 3.5046782768660085,
      "objective/train/tokens_used": 1576940000,
      "theoretical_loss": 3.5046782768660085,
      "tokens_seen": 1556480000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669107321965898,
      "loss": 2.5961,
      "theoretical_loss": 3.5046782768660085,
      "tokens_seen": 1556480000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266900702106319,
      "loss": 2.6762,
      "theoretical_loss": 3.5046653585107124,
      "tokens_seen": 1556545536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026689067201604816,
      "loss": 2.3836,
      "theoretical_loss": 3.504652440851599,
      "tokens_seen": 1556611072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026688064192577734,
      "loss": 2.4974,
      "theoretical_loss": 3.504639523888602,
      "tokens_seen": 1556676608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668706118355065,
      "loss": 2.412,
      "theoretical_loss": 3.5046266076216552,
      "tokens_seen": 1556742144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668605817452357,
      "loss": 2.7806,
      "theoretical_loss": 3.5046136920506905,
      "tokens_seen": 1556807680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026685055165496494,
      "loss": 2.4857,
      "theoretical_loss": 3.5046007771756424,
      "tokens_seen": 1556873216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026684052156469406,
      "loss": 2.5346,
      "theoretical_loss": 3.504587862996443,
      "tokens_seen": 1556938752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668304914744233,
      "loss": 2.6091,
      "theoretical_loss": 3.504574949513027,
      "tokens_seen": 1557004288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668204613841524,
      "loss": 2.6107,
      "theoretical_loss": 3.504562036725326,
      "tokens_seen": 1557069824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026681043129388166,
      "loss": 2.4473,
      "theoretical_loss": 3.504549124633274,
      "tokens_seen": 1557135360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026680040120361084,
      "loss": 2.5327,
      "theoretical_loss": 3.504536213236804,
      "tokens_seen": 1557200896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026679037111334,
      "loss": 2.6196,
      "theoretical_loss": 3.5045233025358495,
      "tokens_seen": 1557266432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667803410230692,
      "loss": 2.6453,
      "theoretical_loss": 3.504510392530344,
      "tokens_seen": 1557331968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667703109327984,
      "loss": 2.6411,
      "theoretical_loss": 3.5044974832202205,
      "tokens_seen": 1557397504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026676028084252757,
      "loss": 2.6207,
      "theoretical_loss": 3.504484574605412,
      "tokens_seen": 1557463040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667502507522568,
      "loss": 2.5174,
      "theoretical_loss": 3.5044716666858524,
      "tokens_seen": 1557528576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026674022066198593,
      "loss": 2.6171,
      "theoretical_loss": 3.5044587594614742,
      "tokens_seen": 1557594112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026673019057171516,
      "loss": 2.7117,
      "theoretical_loss": 3.5044458529322116,
      "tokens_seen": 1557659648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026672016048144434,
      "loss": 2.6978,
      "theoretical_loss": 3.5044329470979974,
      "tokens_seen": 1557725184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667101303911735,
      "loss": 2.5189,
      "theoretical_loss": 3.504420041958765,
      "tokens_seen": 1557790720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667001003009027,
      "loss": 2.3499,
      "theoretical_loss": 3.5044071375144474,
      "tokens_seen": 1557856256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666900702106319,
      "loss": 2.4713,
      "theoretical_loss": 3.5043942337649785,
      "tokens_seen": 1557921792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026668004012036107,
      "loss": 2.7262,
      "theoretical_loss": 3.5043813307102916,
      "tokens_seen": 1557987328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666700100300903,
      "loss": 2.7239,
      "theoretical_loss": 3.50436842835032,
      "tokens_seen": 1558052864
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 860986,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4487876892089844,
      "objective/train/theoretical_loss": 3.5043555266849973,
      "objective/train/tokens_used": 1578578400,
      "theoretical_loss": 3.5043555266849973,
      "tokens_seen": 1558118400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026665997993981943,
      "loss": 2.4259,
      "theoretical_loss": 3.5043555266849973,
      "tokens_seen": 1558118400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026664994984954867,
      "loss": 2.698,
      "theoretical_loss": 3.504342625714256,
      "tokens_seen": 1558183936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666399197592778,
      "loss": 2.7367,
      "theoretical_loss": 3.5043297254380303,
      "tokens_seen": 1558249472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026662988966900703,
      "loss": 2.613,
      "theoretical_loss": 3.5043168258562534,
      "tokens_seen": 1558315008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666198595787362,
      "loss": 2.4404,
      "theoretical_loss": 3.5043039269688587,
      "tokens_seen": 1558380544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666098294884654,
      "loss": 2.3962,
      "theoretical_loss": 3.50429102877578,
      "tokens_seen": 1558446080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026659979939819457,
      "loss": 2.7861,
      "theoretical_loss": 3.5042781312769495,
      "tokens_seen": 1558511616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026658976930792375,
      "loss": 2.5324,
      "theoretical_loss": 3.504265234472302,
      "tokens_seen": 1558577152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026657973921765293,
      "loss": 2.5691,
      "theoretical_loss": 3.5042523383617707,
      "tokens_seen": 1558642688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026656970912738217,
      "loss": 2.678,
      "theoretical_loss": 3.5042394429452886,
      "tokens_seen": 1558708224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665596790371113,
      "loss": 2.7404,
      "theoretical_loss": 3.50422654822279,
      "tokens_seen": 1558773760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026654964894684053,
      "loss": 2.411,
      "theoretical_loss": 3.504213654194207,
      "tokens_seen": 1558839296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665396188565697,
      "loss": 2.6935,
      "theoretical_loss": 3.504200760859474,
      "tokens_seen": 1558904832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665295887662989,
      "loss": 2.4234,
      "theoretical_loss": 3.504187868218524,
      "tokens_seen": 1558970368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665195586760281,
      "loss": 2.7451,
      "theoretical_loss": 3.5041749762712913,
      "tokens_seen": 1559035904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026650952858575726,
      "loss": 2.6332,
      "theoretical_loss": 3.5041620850177093,
      "tokens_seen": 1559101440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664994984954865,
      "loss": 2.564,
      "theoretical_loss": 3.5041491944577103,
      "tokens_seen": 1559166976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026648946840521567,
      "loss": 2.8275,
      "theoretical_loss": 3.5041363045912295,
      "tokens_seen": 1559232512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026647943831494485,
      "loss": 2.4974,
      "theoretical_loss": 3.5041234154181993,
      "tokens_seen": 1559298048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026646940822467403,
      "loss": 2.5962,
      "theoretical_loss": 3.504110526938554,
      "tokens_seen": 1559363584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664593781344032,
      "loss": 2.7282,
      "theoretical_loss": 3.504097639152226,
      "tokens_seen": 1559429120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664493480441324,
      "loss": 2.4502,
      "theoretical_loss": 3.5040847520591503,
      "tokens_seen": 1559494656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026643931795386163,
      "loss": 2.7018,
      "theoretical_loss": 3.5040718656592595,
      "tokens_seen": 1559560192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026642928786359076,
      "loss": 2.4882,
      "theoretical_loss": 3.5040589799524877,
      "tokens_seen": 1559625728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026641925777332,
      "loss": 2.5176,
      "theoretical_loss": 3.504046094938768,
      "tokens_seen": 1559691264
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 862298,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.92509126663208,
      "objective/train/theoretical_loss": 3.5040332106180347,
      "objective/train/tokens_used": 1580216800,
      "theoretical_loss": 3.5040332106180347,
      "tokens_seen": 1559756800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664092276830492,
      "loss": 2.678,
      "theoretical_loss": 3.5040332106180347,
      "tokens_seen": 1559756800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026639919759277836,
      "loss": 2.6686,
      "theoretical_loss": 3.5040203269902213,
      "tokens_seen": 1559822336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026638916750250754,
      "loss": 2.4537,
      "theoretical_loss": 3.5040074440552607,
      "tokens_seen": 1559887872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663791374122367,
      "loss": 2.575,
      "theoretical_loss": 3.5039945618130863,
      "tokens_seen": 1559953408
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663691073219659,
      "loss": 2.6144,
      "theoretical_loss": 3.5039816802636334,
      "tokens_seen": 1560018944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026635907723169514,
      "loss": 2.5371,
      "theoretical_loss": 3.5039687994068345,
      "tokens_seen": 1560084480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026634904714142426,
      "loss": 2.5255,
      "theoretical_loss": 3.5039559192426233,
      "tokens_seen": 1560150016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663390170511535,
      "loss": 2.7876,
      "theoretical_loss": 3.5039430397709337,
      "tokens_seen": 1560215552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663289869608826,
      "loss": 2.7222,
      "theoretical_loss": 3.503930160991699,
      "tokens_seen": 1560281088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026631895687061186,
      "loss": 2.4917,
      "theoretical_loss": 3.5039172829048537,
      "tokens_seen": 1560346624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026630892678034104,
      "loss": 2.6684,
      "theoretical_loss": 3.5039044055103306,
      "tokens_seen": 1560412160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662988966900702,
      "loss": 2.6276,
      "theoretical_loss": 3.503891528808064,
      "tokens_seen": 1560477696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662888665997994,
      "loss": 2.5074,
      "theoretical_loss": 3.5038786527979866,
      "tokens_seen": 1560543232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662788365095286,
      "loss": 2.555,
      "theoretical_loss": 3.503865777480034,
      "tokens_seen": 1560608768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026626880641925777,
      "loss": 2.551,
      "theoretical_loss": 3.5038529028541383,
      "tokens_seen": 1560674304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266258776328987,
      "loss": 2.6477,
      "theoretical_loss": 3.503840028920234,
      "tokens_seen": 1560739840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026624874623871613,
      "loss": 2.6122,
      "theoretical_loss": 3.5038271556782545,
      "tokens_seen": 1560805376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026623871614844536,
      "loss": 2.7329,
      "theoretical_loss": 3.5038142831281336,
      "tokens_seen": 1560870912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026622868605817454,
      "loss": 2.9113,
      "theoretical_loss": 3.5038014112698055,
      "tokens_seen": 1560936448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662186559679037,
      "loss": 2.4849,
      "theoretical_loss": 3.503788540103203,
      "tokens_seen": 1561001984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662086258776329,
      "loss": 2.5958,
      "theoretical_loss": 3.5037756696282614,
      "tokens_seen": 1561067520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661985957873621,
      "loss": 2.631,
      "theoretical_loss": 3.503762799844913,
      "tokens_seen": 1561133056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026618856569709127,
      "loss": 2.7846,
      "theoretical_loss": 3.503749930753093,
      "tokens_seen": 1561198592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661785356068205,
      "loss": 2.5079,
      "theoretical_loss": 3.5037370623527337,
      "tokens_seen": 1561264128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026616850551654963,
      "loss": 2.5375,
      "theoretical_loss": 3.50372419464377,
      "tokens_seen": 1561329664
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 862728,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6292545795440674,
      "objective/train/theoretical_loss": 3.503711327626135,
      "objective/train/tokens_used": 1581855200,
      "theoretical_loss": 3.503711327626135,
      "tokens_seen": 1561395200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026615847542627887,
      "loss": 2.7543,
      "theoretical_loss": 3.503711327626135,
      "tokens_seen": 1561395200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266148445336008,
      "loss": 2.6994,
      "theoretical_loss": 3.5036984612997637,
      "tokens_seen": 1561460736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026613841524573723,
      "loss": 2.5635,
      "theoretical_loss": 3.5036855956645887,
      "tokens_seen": 1561526272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661283851554664,
      "loss": 2.6385,
      "theoretical_loss": 3.5036727307205444,
      "tokens_seen": 1561591808
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661183550651956,
      "loss": 2.6418,
      "theoretical_loss": 3.503659866467565,
      "tokens_seen": 1561657344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026610832497492477,
      "loss": 2.6659,
      "theoretical_loss": 3.503647002905584,
      "tokens_seen": 1561722880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026609829488465395,
      "loss": 2.4739,
      "theoretical_loss": 3.503634140034535,
      "tokens_seen": 1561788416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026608826479438313,
      "loss": 2.8646,
      "theoretical_loss": 3.5036212778543527,
      "tokens_seen": 1561853952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026607823470411237,
      "loss": 2.7653,
      "theoretical_loss": 3.50360841636497,
      "tokens_seen": 1561919488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660682046138415,
      "loss": 2.6995,
      "theoretical_loss": 3.5035955555663216,
      "tokens_seen": 1561985024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026605817452357073,
      "loss": 2.6088,
      "theoretical_loss": 3.5035826954583413,
      "tokens_seen": 1562050560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660481444332999,
      "loss": 2.5755,
      "theoretical_loss": 3.5035698360409624,
      "tokens_seen": 1562116096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660381143430291,
      "loss": 2.6099,
      "theoretical_loss": 3.50355697731412,
      "tokens_seen": 1562181632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660280842527583,
      "loss": 2.6855,
      "theoretical_loss": 3.5035441192777466,
      "tokens_seen": 1562247168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026601805416248746,
      "loss": 2.8502,
      "theoretical_loss": 3.5035312619317773,
      "tokens_seen": 1562312704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026600802407221664,
      "loss": 2.4822,
      "theoretical_loss": 3.503518405276146,
      "tokens_seen": 1562378240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026599799398194587,
      "loss": 2.5384,
      "theoretical_loss": 3.503505549310786,
      "tokens_seen": 1562443776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265987963891675,
      "loss": 2.5567,
      "theoretical_loss": 3.503492694035632,
      "tokens_seen": 1562509312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026597793380140423,
      "loss": 2.6505,
      "theoretical_loss": 3.503479839450618,
      "tokens_seen": 1562574848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026596790371113336,
      "loss": 2.4879,
      "theoretical_loss": 3.503466985555677,
      "tokens_seen": 1562640384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659578736208626,
      "loss": 2.5259,
      "theoretical_loss": 3.503454132350744,
      "tokens_seen": 1562705920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659478435305918,
      "loss": 2.6594,
      "theoretical_loss": 3.503441279835753,
      "tokens_seen": 1562771456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026593781344032096,
      "loss": 2.561,
      "theoretical_loss": 3.503428428010637,
      "tokens_seen": 1562836992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026592778335005014,
      "loss": 2.6891,
      "theoretical_loss": 3.5034155768753314,
      "tokens_seen": 1562902528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659177532597794,
      "loss": 2.5436,
      "theoretical_loss": 3.5034027264297696,
      "tokens_seen": 1562968064
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 863704,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.2020814418792725,
      "objective/train/theoretical_loss": 3.503389876673886,
      "objective/train/tokens_used": 1583493600,
      "theoretical_loss": 3.503389876673886,
      "tokens_seen": 1563033600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002659077231695085,
      "loss": 2.7556,
      "theoretical_loss": 3.503389876673886,
      "tokens_seen": 1563033600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026589769307923774,
      "loss": 2.6175,
      "theoretical_loss": 3.503377027607614,
      "tokens_seen": 1563099136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026588766298896686,
      "loss": 2.6476,
      "theoretical_loss": 3.5033641792308883,
      "tokens_seen": 1563164672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658776328986961,
      "loss": 2.6468,
      "theoretical_loss": 3.5033513315436426,
      "tokens_seen": 1563230208
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658676028084253,
      "loss": 2.5511,
      "theoretical_loss": 3.5033384845458113,
      "tokens_seen": 1563295744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026585757271815446,
      "loss": 2.7945,
      "theoretical_loss": 3.503325638237328,
      "tokens_seen": 1563361280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026584754262788364,
      "loss": 2.6927,
      "theoretical_loss": 3.503312792618128,
      "tokens_seen": 1563426816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658375125376128,
      "loss": 2.559,
      "theoretical_loss": 3.5032999476881437,
      "tokens_seen": 1563492352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265827482447342,
      "loss": 2.557,
      "theoretical_loss": 3.5032871034473105,
      "tokens_seen": 1563557888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026581745235707124,
      "loss": 2.6794,
      "theoretical_loss": 3.503274259895562,
      "tokens_seen": 1563623424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026580742226680037,
      "loss": 2.4775,
      "theoretical_loss": 3.5032614170328333,
      "tokens_seen": 1563688960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657973921765296,
      "loss": 2.5448,
      "theoretical_loss": 3.5032485748590574,
      "tokens_seen": 1563754496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026578736208625873,
      "loss": 2.5522,
      "theoretical_loss": 3.5032357333741686,
      "tokens_seen": 1563820032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026577733199598797,
      "loss": 2.5624,
      "theoretical_loss": 3.5032228925781013,
      "tokens_seen": 1563885568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026576730190571715,
      "loss": 2.5355,
      "theoretical_loss": 3.50321005247079,
      "tokens_seen": 1563951104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026575727181544633,
      "loss": 2.6652,
      "theoretical_loss": 3.503197213052169,
      "tokens_seen": 1564016640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026574724172517556,
      "loss": 2.6279,
      "theoretical_loss": 3.5031843743221716,
      "tokens_seen": 1564082176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026573721163490474,
      "loss": 2.5509,
      "theoretical_loss": 3.5031715362807327,
      "tokens_seen": 1564147712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657271815446339,
      "loss": 2.4891,
      "theoretical_loss": 3.503158698927786,
      "tokens_seen": 1564213248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657171514543631,
      "loss": 2.5214,
      "theoretical_loss": 3.5031458622632665,
      "tokens_seen": 1564278784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657071213640923,
      "loss": 2.3911,
      "theoretical_loss": 3.503133026287108,
      "tokens_seen": 1564344320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026569709127382147,
      "loss": 2.5955,
      "theoretical_loss": 3.503120190999245,
      "tokens_seen": 1564409856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656870611835507,
      "loss": 2.5199,
      "theoretical_loss": 3.503107356399611,
      "tokens_seen": 1564475392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026567703109327983,
      "loss": 2.4505,
      "theoretical_loss": 3.5030945224881416,
      "tokens_seen": 1564540928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026566700100300907,
      "loss": 2.6076,
      "theoretical_loss": 3.5030816892647696,
      "tokens_seen": 1564606464
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 864198,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.547630548477173,
      "objective/train/theoretical_loss": 3.5030688567294304,
      "objective/train/tokens_used": 1585132000,
      "theoretical_loss": 3.5030688567294304,
      "tokens_seen": 1564672000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656569709127382,
      "loss": 2.6312,
      "theoretical_loss": 3.5030688567294304,
      "tokens_seen": 1564672000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026564694082246743,
      "loss": 2.3489,
      "theoretical_loss": 3.503056024882058,
      "tokens_seen": 1564737536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656369107321966,
      "loss": 2.605,
      "theoretical_loss": 3.5030431937225863,
      "tokens_seen": 1564803072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656268806419258,
      "loss": 2.5861,
      "theoretical_loss": 3.50303036325095,
      "tokens_seen": 1564868608
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026561685055165497,
      "loss": 2.6797,
      "theoretical_loss": 3.5030175334670837,
      "tokens_seen": 1564934144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026560682046138415,
      "loss": 2.5567,
      "theoretical_loss": 3.503004704370921,
      "tokens_seen": 1564999680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026559679037111333,
      "loss": 2.4033,
      "theoretical_loss": 3.5029918759623966,
      "tokens_seen": 1565065216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026558676028084257,
      "loss": 2.67,
      "theoretical_loss": 3.5029790482414453,
      "tokens_seen": 1565130752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655767301905717,
      "loss": 2.693,
      "theoretical_loss": 3.502966221208,
      "tokens_seen": 1565196288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026556670010030093,
      "loss": 2.4977,
      "theoretical_loss": 3.502953394861997,
      "tokens_seen": 1565261824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655566700100301,
      "loss": 2.749,
      "theoretical_loss": 3.50294056920337,
      "tokens_seen": 1565327360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655466399197593,
      "loss": 2.2335,
      "theoretical_loss": 3.5029277442320526,
      "tokens_seen": 1565392896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655366098294885,
      "loss": 2.4006,
      "theoretical_loss": 3.5029149199479797,
      "tokens_seen": 1565458432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026552657973921766,
      "loss": 2.5632,
      "theoretical_loss": 3.502902096351086,
      "tokens_seen": 1565523968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026551654964894684,
      "loss": 2.7731,
      "theoretical_loss": 3.5028892734413057,
      "tokens_seen": 1565589504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026550651955867607,
      "loss": 2.5731,
      "theoretical_loss": 3.5028764512185733,
      "tokens_seen": 1565655040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654964894684052,
      "loss": 2.5059,
      "theoretical_loss": 3.502863629682823,
      "tokens_seen": 1565720576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026548645937813443,
      "loss": 2.6426,
      "theoretical_loss": 3.5028508088339887,
      "tokens_seen": 1565786112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026547642928786356,
      "loss": 2.5619,
      "theoretical_loss": 3.5028379886720065,
      "tokens_seen": 1565851648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654663991975928,
      "loss": 2.3653,
      "theoretical_loss": 3.5028251691968095,
      "tokens_seen": 1565917184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265456369107322,
      "loss": 2.75,
      "theoretical_loss": 3.5028123504083326,
      "tokens_seen": 1565982720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026544633901705116,
      "loss": 2.5192,
      "theoretical_loss": 3.5027995323065104,
      "tokens_seen": 1566048256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026543630892678034,
      "loss": 2.3786,
      "theoretical_loss": 3.502786714891277,
      "tokens_seen": 1566113792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654262788365096,
      "loss": 2.7013,
      "theoretical_loss": 3.502773898162567,
      "tokens_seen": 1566179328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654162487462387,
      "loss": 2.5743,
      "theoretical_loss": 3.5027610821203146,
      "tokens_seen": 1566244864
    },
    {
      "epoch": 0.47,
      "objective/train/docs_used": 865502,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5836033821105957,
      "objective/train/theoretical_loss": 3.5027482667644554,
      "objective/train/tokens_used": 1586770400,
      "theoretical_loss": 3.5027482667644554,
      "tokens_seen": 1566310400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026540621865596794,
      "loss": 2.6275,
      "theoretical_loss": 3.5027482667644554,
      "tokens_seen": 1566310400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026539618856569707,
      "loss": 2.5517,
      "theoretical_loss": 3.5027354520949228,
      "tokens_seen": 1566375936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653861584754263,
      "loss": 2.6707,
      "theoretical_loss": 3.5027226381116523,
      "tokens_seen": 1566441472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653761283851555,
      "loss": 2.64,
      "theoretical_loss": 3.502709824814578,
      "tokens_seen": 1566507008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026536609829488466,
      "loss": 2.3741,
      "theoretical_loss": 3.5026970122036336,
      "tokens_seen": 1566572544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026535606820461384,
      "loss": 2.5363,
      "theoretical_loss": 3.502684200278755,
      "tokens_seen": 1566638080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265346038114343,
      "loss": 2.5694,
      "theoretical_loss": 3.5026713890398757,
      "tokens_seen": 1566703616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653360080240722,
      "loss": 2.6299,
      "theoretical_loss": 3.502658578486931,
      "tokens_seen": 1566769152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026532597793380144,
      "loss": 2.5698,
      "theoretical_loss": 3.5026457686198555,
      "tokens_seen": 1566834688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026531594784353057,
      "loss": 2.6487,
      "theoretical_loss": 3.5026329594385834,
      "tokens_seen": 1566900224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653059177532598,
      "loss": 2.5744,
      "theoretical_loss": 3.5026201509430495,
      "tokens_seen": 1566965760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026529588766298893,
      "loss": 2.6302,
      "theoretical_loss": 3.502607343133188,
      "tokens_seen": 1567031296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026528585757271817,
      "loss": 2.5629,
      "theoretical_loss": 3.5025945360089343,
      "tokens_seen": 1567096832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026527582748244735,
      "loss": 2.5357,
      "theoretical_loss": 3.502581729570222,
      "tokens_seen": 1567162368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026526579739217653,
      "loss": 2.5799,
      "theoretical_loss": 3.5025689238169866,
      "tokens_seen": 1567227904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652557673019057,
      "loss": 2.5381,
      "theoretical_loss": 3.5025561187491627,
      "tokens_seen": 1567293440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026524573721163494,
      "loss": 2.4347,
      "theoretical_loss": 3.502543314366685,
      "tokens_seen": 1567358976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026523570712136407,
      "loss": 2.5106,
      "theoretical_loss": 3.5025305106694873,
      "tokens_seen": 1567424512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652256770310933,
      "loss": 2.4497,
      "theoretical_loss": 3.502517707657505,
      "tokens_seen": 1567490048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026521564694082243,
      "loss": 2.4718,
      "theoretical_loss": 3.5025049053306727,
      "tokens_seen": 1567555584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026520561685055167,
      "loss": 2.6144,
      "theoretical_loss": 3.502492103688925,
      "tokens_seen": 1567621120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026519558676028085,
      "loss": 2.6231,
      "theoretical_loss": 3.502479302732197,
      "tokens_seen": 1567686656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026518555667001003,
      "loss": 2.3729,
      "theoretical_loss": 3.502466502460423,
      "tokens_seen": 1567752192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651755265797392,
      "loss": 2.6039,
      "theoretical_loss": 3.502453702873538,
      "tokens_seen": 1567817728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651654964894684,
      "loss": 2.5398,
      "theoretical_loss": 3.502440903971476,
      "tokens_seen": 1567883264
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 866057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.337247371673584,
      "objective/train/theoretical_loss": 3.5024281057541726,
      "objective/train/tokens_used": 1588408800,
      "theoretical_loss": 3.5024281057541726,
      "tokens_seen": 1567948800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651554663991976,
      "loss": 2.5498,
      "theoretical_loss": 3.5024281057541726,
      "tokens_seen": 1567948800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651454363089268,
      "loss": 2.7671,
      "theoretical_loss": 3.502415308221562,
      "tokens_seen": 1568014336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026513540621865594,
      "loss": 2.5185,
      "theoretical_loss": 3.502402511373579,
      "tokens_seen": 1568079872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026512537612838517,
      "loss": 2.528,
      "theoretical_loss": 3.5023897152101586,
      "tokens_seen": 1568145408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651153460381143,
      "loss": 2.4021,
      "theoretical_loss": 3.502376919731236,
      "tokens_seen": 1568210944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026510531594784353,
      "loss": 2.5057,
      "theoretical_loss": 3.5023641249367445,
      "tokens_seen": 1568276480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650952858575727,
      "loss": 2.653,
      "theoretical_loss": 3.5023513308266203,
      "tokens_seen": 1568342016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650852557673019,
      "loss": 2.7023,
      "theoretical_loss": 3.5023385374007985,
      "tokens_seen": 1568407552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650752256770311,
      "loss": 2.4164,
      "theoretical_loss": 3.5023257446592124,
      "tokens_seen": 1568473088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650651955867603,
      "loss": 2.5811,
      "theoretical_loss": 3.5023129526017973,
      "tokens_seen": 1568538624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026505516549648944,
      "loss": 2.6678,
      "theoretical_loss": 3.502300161228489,
      "tokens_seen": 1568604160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650451354062187,
      "loss": 2.3171,
      "theoretical_loss": 3.5022873705392215,
      "tokens_seen": 1568669696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650351053159478,
      "loss": 2.3869,
      "theoretical_loss": 3.5022745805339297,
      "tokens_seen": 1568735232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026502507522567704,
      "loss": 2.541,
      "theoretical_loss": 3.5022617912125487,
      "tokens_seen": 1568800768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650150451354062,
      "loss": 2.371,
      "theoretical_loss": 3.5022490025750126,
      "tokens_seen": 1568866304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650050150451354,
      "loss": 2.646,
      "theoretical_loss": 3.5022362146212576,
      "tokens_seen": 1568931840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026499498495486464,
      "loss": 2.6259,
      "theoretical_loss": 3.5022234273512174,
      "tokens_seen": 1568997376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026498495486459376,
      "loss": 2.4813,
      "theoretical_loss": 3.5022106407648277,
      "tokens_seen": 1569062912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264974924774323,
      "loss": 2.5337,
      "theoretical_loss": 3.5021978548620227,
      "tokens_seen": 1569128448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649648946840522,
      "loss": 2.6397,
      "theoretical_loss": 3.5021850696427377,
      "tokens_seen": 1569193984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026495486459378136,
      "loss": 2.7135,
      "theoretical_loss": 3.5021722851069077,
      "tokens_seen": 1569259520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026494483450351054,
      "loss": 2.5215,
      "theoretical_loss": 3.502159501254467,
      "tokens_seen": 1569325056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649348044132398,
      "loss": 2.7714,
      "theoretical_loss": 3.502146718085352,
      "tokens_seen": 1569390592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649247743229689,
      "loss": 2.5559,
      "theoretical_loss": 3.5021339355994963,
      "tokens_seen": 1569456128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026491474423269814,
      "loss": 2.5877,
      "theoretical_loss": 3.5021211537968346,
      "tokens_seen": 1569521664
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 867215,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3438003063201904,
      "objective/train/theoretical_loss": 3.502108372677303,
      "objective/train/tokens_used": 1590047200,
      "theoretical_loss": 3.502108372677303,
      "tokens_seen": 1569587200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026490471414242727,
      "loss": 2.5784,
      "theoretical_loss": 3.502108372677303,
      "tokens_seen": 1569587200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648946840521565,
      "loss": 2.536,
      "theoretical_loss": 3.5020955922408357,
      "tokens_seen": 1569652736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648846539618857,
      "loss": 2.5605,
      "theoretical_loss": 3.502082812487368,
      "tokens_seen": 1569718272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026487462387161486,
      "loss": 2.8492,
      "theoretical_loss": 3.5020700334168353,
      "tokens_seen": 1569783808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026486459378134404,
      "loss": 2.6621,
      "theoretical_loss": 3.5020572550291718,
      "tokens_seen": 1569849344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648545636910732,
      "loss": 2.4295,
      "theoretical_loss": 3.5020444773243122,
      "tokens_seen": 1569914880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648445336008024,
      "loss": 2.7394,
      "theoretical_loss": 3.502031700302193,
      "tokens_seen": 1569980416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026483450351053164,
      "loss": 2.6082,
      "theoretical_loss": 3.5020189239627477,
      "tokens_seen": 1570045952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026482447342026077,
      "loss": 2.4671,
      "theoretical_loss": 3.502006148305912,
      "tokens_seen": 1570111488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026481444332999,
      "loss": 2.412,
      "theoretical_loss": 3.5019933733316213,
      "tokens_seen": 1570177024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026480441323971913,
      "loss": 2.5336,
      "theoretical_loss": 3.5019805990398103,
      "tokens_seen": 1570242560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026479438314944837,
      "loss": 2.6074,
      "theoretical_loss": 3.5019678254304143,
      "tokens_seen": 1570308096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026478435305917755,
      "loss": 2.5739,
      "theoretical_loss": 3.5019550525033676,
      "tokens_seen": 1570373632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026477432296890673,
      "loss": 2.5441,
      "theoretical_loss": 3.5019422802586053,
      "tokens_seen": 1570439168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647642928786359,
      "loss": 2.3627,
      "theoretical_loss": 3.5019295086960636,
      "tokens_seen": 1570504704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026475426278836514,
      "loss": 2.5028,
      "theoretical_loss": 3.501916737815677,
      "tokens_seen": 1570570240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026474423269809427,
      "loss": 2.5644,
      "theoretical_loss": 3.5019039676173804,
      "tokens_seen": 1570635776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647342026078235,
      "loss": 2.6869,
      "theoretical_loss": 3.501891198101109,
      "tokens_seen": 1570701312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026472417251755263,
      "loss": 2.6599,
      "theoretical_loss": 3.5018784292667977,
      "tokens_seen": 1570766848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026471414242728187,
      "loss": 2.4728,
      "theoretical_loss": 3.5018656611143824,
      "tokens_seen": 1570832384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026470411233701105,
      "loss": 2.6884,
      "theoretical_loss": 3.5018528936437976,
      "tokens_seen": 1570897920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026469408224674023,
      "loss": 2.5154,
      "theoretical_loss": 3.5018401268549786,
      "tokens_seen": 1570963456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646840521564694,
      "loss": 2.728,
      "theoretical_loss": 3.5018273607478605,
      "tokens_seen": 1571028992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646740220661986,
      "loss": 2.5455,
      "theoretical_loss": 3.5018145953223785,
      "tokens_seen": 1571094528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646639919759278,
      "loss": 2.7501,
      "theoretical_loss": 3.5018018305784677,
      "tokens_seen": 1571160064
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 868323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.522857189178467,
      "objective/train/theoretical_loss": 3.5017890665160634,
      "objective/train/tokens_used": 1591685600,
      "theoretical_loss": 3.5017890665160634,
      "tokens_seen": 1571225600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264653961885657,
      "loss": 2.6023,
      "theoretical_loss": 3.5017890665160634,
      "tokens_seen": 1571225600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026464393179538614,
      "loss": 2.6551,
      "theoretical_loss": 3.5017763031351006,
      "tokens_seen": 1571291136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026463390170511537,
      "loss": 2.7338,
      "theoretical_loss": 3.5017635404355154,
      "tokens_seen": 1571356672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646238716148445,
      "loss": 2.4192,
      "theoretical_loss": 3.5017507784172412,
      "tokens_seen": 1571422208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026461384152457373,
      "loss": 2.6184,
      "theoretical_loss": 3.501738017080215,
      "tokens_seen": 1571487744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646038114343029,
      "loss": 2.5038,
      "theoretical_loss": 3.501725256424371,
      "tokens_seen": 1571553280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645937813440321,
      "loss": 2.4952,
      "theoretical_loss": 3.5017124964496453,
      "tokens_seen": 1571618816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645837512537613,
      "loss": 2.6781,
      "theoretical_loss": 3.501699737155972,
      "tokens_seen": 1571684352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645737211634905,
      "loss": 2.4646,
      "theoretical_loss": 3.501686978543287,
      "tokens_seen": 1571749888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026456369107321964,
      "loss": 2.6156,
      "theoretical_loss": 3.501674220611526,
      "tokens_seen": 1571815424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645536609829489,
      "loss": 2.455,
      "theoretical_loss": 3.501661463360623,
      "tokens_seen": 1571880960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264543630892678,
      "loss": 2.7022,
      "theoretical_loss": 3.5016487067905144,
      "tokens_seen": 1571946496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026453360080240724,
      "loss": 2.5932,
      "theoretical_loss": 3.501635950901135,
      "tokens_seen": 1572012032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645235707121364,
      "loss": 2.3791,
      "theoretical_loss": 3.501623195692421,
      "tokens_seen": 1572077568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645135406218656,
      "loss": 2.6157,
      "theoretical_loss": 3.5016104411643063,
      "tokens_seen": 1572143104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645035105315948,
      "loss": 2.5162,
      "theoretical_loss": 3.501597687316727,
      "tokens_seen": 1572208640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026449348044132396,
      "loss": 2.5154,
      "theoretical_loss": 3.5015849341496184,
      "tokens_seen": 1572274176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026448345035105314,
      "loss": 2.6475,
      "theoretical_loss": 3.5015721816629153,
      "tokens_seen": 1572339712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644734202607824,
      "loss": 2.5634,
      "theoretical_loss": 3.5015594298565538,
      "tokens_seen": 1572405248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644633901705115,
      "loss": 2.3788,
      "theoretical_loss": 3.5015466787304685,
      "tokens_seen": 1572470784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026445336008024074,
      "loss": 2.6087,
      "theoretical_loss": 3.5015339282845956,
      "tokens_seen": 1572536320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026444332998996987,
      "loss": 2.6118,
      "theoretical_loss": 3.5015211785188702,
      "tokens_seen": 1572601856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644332998996991,
      "loss": 2.7455,
      "theoretical_loss": 3.501508429433227,
      "tokens_seen": 1572667392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644232698094283,
      "loss": 2.5472,
      "theoretical_loss": 3.5014956810276026,
      "tokens_seen": 1572732928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026441323971915747,
      "loss": 2.5672,
      "theoretical_loss": 3.501482933301931,
      "tokens_seen": 1572798464
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 869061,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5812792778015137,
      "objective/train/theoretical_loss": 3.5014701862561486,
      "objective/train/tokens_used": 1593324000,
      "theoretical_loss": 3.5014701862561486,
      "tokens_seen": 1572864000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026440320962888665,
      "loss": 2.5332,
      "theoretical_loss": 3.5014701862561486,
      "tokens_seen": 1572864000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643931795386159,
      "loss": 2.4261,
      "theoretical_loss": 3.5014574398901903,
      "tokens_seen": 1572929536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264383149448345,
      "loss": 2.5914,
      "theoretical_loss": 3.5014446942039923,
      "tokens_seen": 1572995072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026437311935807424,
      "loss": 2.5054,
      "theoretical_loss": 3.501431949197489,
      "tokens_seen": 1573060608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026436308926780337,
      "loss": 2.2896,
      "theoretical_loss": 3.501419204870616,
      "tokens_seen": 1573126144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643530591775326,
      "loss": 2.5589,
      "theoretical_loss": 3.5014064612233096,
      "tokens_seen": 1573191680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643430290872618,
      "loss": 2.4529,
      "theoretical_loss": 3.5013937182555046,
      "tokens_seen": 1573257216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026433299899699097,
      "loss": 2.713,
      "theoretical_loss": 3.5013809759671366,
      "tokens_seen": 1573322752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026432296890672015,
      "loss": 2.6435,
      "theoretical_loss": 3.5013682343581407,
      "tokens_seen": 1573388288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026431293881644933,
      "loss": 2.4365,
      "theoretical_loss": 3.501355493428453,
      "tokens_seen": 1573453824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643029087261785,
      "loss": 2.5929,
      "theoretical_loss": 3.501342753178009,
      "tokens_seen": 1573519360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026429287863590775,
      "loss": 2.6807,
      "theoretical_loss": 3.5013300136067436,
      "tokens_seen": 1573584896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642828485456369,
      "loss": 2.5183,
      "theoretical_loss": 3.5013172747145926,
      "tokens_seen": 1573650432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642728184553661,
      "loss": 2.5825,
      "theoretical_loss": 3.5013045365014914,
      "tokens_seen": 1573715968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026426278836509534,
      "loss": 2.493,
      "theoretical_loss": 3.501291798967376,
      "tokens_seen": 1573781504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026425275827482447,
      "loss": 2.2586,
      "theoretical_loss": 3.5012790621121814,
      "tokens_seen": 1573847040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642427281845537,
      "loss": 2.4789,
      "theoretical_loss": 3.501266325935844,
      "tokens_seen": 1573912576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026423269809428283,
      "loss": 2.5391,
      "theoretical_loss": 3.5012535904382975,
      "tokens_seen": 1573978112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026422266800401207,
      "loss": 2.5963,
      "theoretical_loss": 3.5012408556194794,
      "tokens_seen": 1574043648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026421263791374125,
      "loss": 2.5903,
      "theoretical_loss": 3.5012281214793246,
      "tokens_seen": 1574109184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026420260782347043,
      "loss": 2.5422,
      "theoretical_loss": 3.5012153880177683,
      "tokens_seen": 1574174720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641925777331996,
      "loss": 2.4667,
      "theoretical_loss": 3.5012026552347466,
      "tokens_seen": 1574240256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641825476429288,
      "loss": 2.3859,
      "theoretical_loss": 3.501189923130195,
      "tokens_seen": 1574305792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264172517552658,
      "loss": 2.4377,
      "theoretical_loss": 3.5011771917040493,
      "tokens_seen": 1574371328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641624874623872,
      "loss": 2.8072,
      "theoretical_loss": 3.501164460956244,
      "tokens_seen": 1574436864
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 869692,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4734344482421875,
      "objective/train/theoretical_loss": 3.501151730886716,
      "objective/train/tokens_used": 1594962400,
      "theoretical_loss": 3.501151730886716,
      "tokens_seen": 1574502400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026415245737211634,
      "loss": 2.6231,
      "theoretical_loss": 3.501151730886716,
      "tokens_seen": 1574502400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026414242728184557,
      "loss": 2.5573,
      "theoretical_loss": 3.5011390014954005,
      "tokens_seen": 1574567936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641323971915747,
      "loss": 2.6861,
      "theoretical_loss": 3.5011262727822334,
      "tokens_seen": 1574633472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026412236710130393,
      "loss": 2.4734,
      "theoretical_loss": 3.5011135447471498,
      "tokens_seen": 1574699008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641123370110331,
      "loss": 2.515,
      "theoretical_loss": 3.5011008173900855,
      "tokens_seen": 1574764544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641023069207623,
      "loss": 2.5897,
      "theoretical_loss": 3.5010880907109767,
      "tokens_seen": 1574830080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640922768304915,
      "loss": 2.3304,
      "theoretical_loss": 3.501075364709758,
      "tokens_seen": 1574895616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640822467402207,
      "loss": 2.5674,
      "theoretical_loss": 3.501062639386366,
      "tokens_seen": 1574961152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026407221664994984,
      "loss": 2.8064,
      "theoretical_loss": 3.5010499147407366,
      "tokens_seen": 1575026688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640621865596791,
      "loss": 2.657,
      "theoretical_loss": 3.501037190772805,
      "tokens_seen": 1575092224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640521564694082,
      "loss": 2.8227,
      "theoretical_loss": 3.501024467482507,
      "tokens_seen": 1575157760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026404212637913744,
      "loss": 2.5477,
      "theoretical_loss": 3.5010117448697775,
      "tokens_seen": 1575223296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640320962888666,
      "loss": 2.5243,
      "theoretical_loss": 3.500999022934554,
      "tokens_seen": 1575288832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640220661985958,
      "loss": 2.6759,
      "theoretical_loss": 3.500986301676771,
      "tokens_seen": 1575354368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264012036108325,
      "loss": 2.6374,
      "theoretical_loss": 3.500973581096364,
      "tokens_seen": 1575419904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026400200601805416,
      "loss": 2.771,
      "theoretical_loss": 3.5009608611932697,
      "tokens_seen": 1575485440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026399197592778334,
      "loss": 2.5815,
      "theoretical_loss": 3.500948141967423,
      "tokens_seen": 1575550976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639819458375126,
      "loss": 2.4938,
      "theoretical_loss": 3.5009354234187606,
      "tokens_seen": 1575616512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639719157472417,
      "loss": 2.854,
      "theoretical_loss": 3.5009227055472176,
      "tokens_seen": 1575682048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026396188565697094,
      "loss": 2.3283,
      "theoretical_loss": 3.50090998835273,
      "tokens_seen": 1575747584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026395185556670007,
      "loss": 2.5448,
      "theoretical_loss": 3.5008972718352336,
      "tokens_seen": 1575813120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639418254764293,
      "loss": 2.8598,
      "theoretical_loss": 3.500884555994664,
      "tokens_seen": 1575878656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639317953861585,
      "loss": 2.4873,
      "theoretical_loss": 3.500871840830958,
      "tokens_seen": 1575944192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026392176529588767,
      "loss": 2.3254,
      "theoretical_loss": 3.500859126344049,
      "tokens_seen": 1576009728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026391173520561685,
      "loss": 2.6268,
      "theoretical_loss": 3.5008464125338756,
      "tokens_seen": 1576075264
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 870886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7491490840911865,
      "objective/train/theoretical_loss": 3.5008336994003724,
      "objective/train/tokens_used": 1596600800,
      "theoretical_loss": 3.5008336994003724,
      "tokens_seen": 1576140800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002639017051153461,
      "loss": 2.7081,
      "theoretical_loss": 3.5008336994003724,
      "tokens_seen": 1576140800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638916750250752,
      "loss": 2.4539,
      "theoretical_loss": 3.5008209869434754,
      "tokens_seen": 1576206336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026388164493480444,
      "loss": 2.5672,
      "theoretical_loss": 3.50080827516312,
      "tokens_seen": 1576271872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026387161484453357,
      "loss": 2.4068,
      "theoretical_loss": 3.500795564059243,
      "tokens_seen": 1576337408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638615847542628,
      "loss": 2.5358,
      "theoretical_loss": 3.5007828536317795,
      "tokens_seen": 1576402944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263851554663992,
      "loss": 2.4222,
      "theoretical_loss": 3.5007701438806658,
      "tokens_seen": 1576468480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026384152457372117,
      "loss": 2.513,
      "theoretical_loss": 3.5007574348058372,
      "tokens_seen": 1576534016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026383149448345035,
      "loss": 2.6312,
      "theoretical_loss": 3.50074472640723,
      "tokens_seen": 1576599552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026382146439317953,
      "loss": 2.492,
      "theoretical_loss": 3.5007320186847806,
      "tokens_seen": 1576665088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638114343029087,
      "loss": 2.3217,
      "theoretical_loss": 3.5007193116384245,
      "tokens_seen": 1576730624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026380140421263795,
      "loss": 2.574,
      "theoretical_loss": 3.500706605268097,
      "tokens_seen": 1576796160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637913741223671,
      "loss": 2.72,
      "theoretical_loss": 3.5006938995737356,
      "tokens_seen": 1576861696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637813440320963,
      "loss": 2.4526,
      "theoretical_loss": 3.5006811945552743,
      "tokens_seen": 1576927232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637713139418255,
      "loss": 2.38,
      "theoretical_loss": 3.500668490212651,
      "tokens_seen": 1576992768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026376128385155467,
      "loss": 2.3041,
      "theoretical_loss": 3.5006557865458,
      "tokens_seen": 1577058304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026375125376128385,
      "loss": 2.5886,
      "theoretical_loss": 3.5006430835546585,
      "tokens_seen": 1577123840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026374122367101303,
      "loss": 2.8158,
      "theoretical_loss": 3.5006303812391613,
      "tokens_seen": 1577189376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637311935807422,
      "loss": 2.4255,
      "theoretical_loss": 3.5006176795992454,
      "tokens_seen": 1577254912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026372116349047145,
      "loss": 2.5985,
      "theoretical_loss": 3.500604978634847,
      "tokens_seen": 1577320448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637111334002006,
      "loss": 2.5509,
      "theoretical_loss": 3.500592278345901,
      "tokens_seen": 1577385984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637011033099298,
      "loss": 2.5023,
      "theoretical_loss": 3.500579578732344,
      "tokens_seen": 1577451520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026369107321965894,
      "loss": 2.6507,
      "theoretical_loss": 3.5005668797941123,
      "tokens_seen": 1577517056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636810431293882,
      "loss": 2.5937,
      "theoretical_loss": 3.5005541815311414,
      "tokens_seen": 1577582592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026367101303911736,
      "loss": 2.7392,
      "theoretical_loss": 3.5005414839433673,
      "tokens_seen": 1577648128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026366098294884654,
      "loss": 2.6395,
      "theoretical_loss": 3.500528787030727,
      "tokens_seen": 1577713664
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 871486,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.030456781387329,
      "objective/train/theoretical_loss": 3.5005160907931554,
      "objective/train/tokens_used": 1598239200,
      "theoretical_loss": 3.5005160907931554,
      "tokens_seen": 1577779200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636509528585757,
      "loss": 3.0241,
      "theoretical_loss": 3.5005160907931554,
      "tokens_seen": 1577779200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636409227683049,
      "loss": 2.5359,
      "theoretical_loss": 3.5005033952305893,
      "tokens_seen": 1577844736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636308926780341,
      "loss": 2.6299,
      "theoretical_loss": 3.5004907003429646,
      "tokens_seen": 1577910272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636208625877633,
      "loss": 2.6755,
      "theoretical_loss": 3.5004780061302174,
      "tokens_seen": 1577975808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026361083249749244,
      "loss": 2.5483,
      "theoretical_loss": 3.5004653125922838,
      "tokens_seen": 1578041344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636008024072217,
      "loss": 2.6397,
      "theoretical_loss": 3.5004526197290993,
      "tokens_seen": 1578106880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026359077231695086,
      "loss": 2.5487,
      "theoretical_loss": 3.500439927540601,
      "tokens_seen": 1578172416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026358074222668004,
      "loss": 2.5478,
      "theoretical_loss": 3.5004272360267246,
      "tokens_seen": 1578237952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635707121364092,
      "loss": 2.3817,
      "theoretical_loss": 3.500414545187406,
      "tokens_seen": 1578303488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635606820461384,
      "loss": 2.562,
      "theoretical_loss": 3.5004018550225817,
      "tokens_seen": 1578369024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635506519558676,
      "loss": 2.3398,
      "theoretical_loss": 3.5003891655321877,
      "tokens_seen": 1578434560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635406218655968,
      "loss": 2.5235,
      "theoretical_loss": 3.50037647671616,
      "tokens_seen": 1578500096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026353059177532595,
      "loss": 2.6254,
      "theoretical_loss": 3.500363788574435,
      "tokens_seen": 1578565632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635205616850552,
      "loss": 2.5682,
      "theoretical_loss": 3.5003511011069492,
      "tokens_seen": 1578631168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026351053159478436,
      "loss": 2.624,
      "theoretical_loss": 3.500338414313638,
      "tokens_seen": 1578696704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026350050150451354,
      "loss": 2.5264,
      "theoretical_loss": 3.500325728194438,
      "tokens_seen": 1578762240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634904714142428,
      "loss": 2.5816,
      "theoretical_loss": 3.5003130427492852,
      "tokens_seen": 1578827776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634804413239719,
      "loss": 2.5874,
      "theoretical_loss": 3.5003003579781162,
      "tokens_seen": 1578893312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026347041123370114,
      "loss": 2.5642,
      "theoretical_loss": 3.500287673880867,
      "tokens_seen": 1578958848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026346038114343027,
      "loss": 2.4428,
      "theoretical_loss": 3.500274990457474,
      "tokens_seen": 1579024384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634503510531595,
      "loss": 2.6323,
      "theoretical_loss": 3.5002623077078727,
      "tokens_seen": 1579089920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634403209628887,
      "loss": 2.7493,
      "theoretical_loss": 3.5002496256320006,
      "tokens_seen": 1579155456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026343029087261787,
      "loss": 2.3776,
      "theoretical_loss": 3.5002369442297927,
      "tokens_seen": 1579220992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026342026078234705,
      "loss": 2.4109,
      "theoretical_loss": 3.5002242635011864,
      "tokens_seen": 1579286528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634102306920763,
      "loss": 2.5289,
      "theoretical_loss": 3.5002115834461165,
      "tokens_seen": 1579352064
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 872887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.328456401824951,
      "objective/train/theoretical_loss": 3.500198904064521,
      "objective/train/tokens_used": 1599877600,
      "theoretical_loss": 3.500198904064521,
      "tokens_seen": 1579417600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634002006018054,
      "loss": 2.5613,
      "theoretical_loss": 3.500198904064521,
      "tokens_seen": 1579417600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026339017051153464,
      "loss": 2.4966,
      "theoretical_loss": 3.5001862253563347,
      "tokens_seen": 1579483136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026338014042126377,
      "loss": 2.5745,
      "theoretical_loss": 3.500173547321495,
      "tokens_seen": 1579548672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263370110330993,
      "loss": 2.695,
      "theoretical_loss": 3.500160869959938,
      "tokens_seen": 1579614208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633600802407222,
      "loss": 2.5789,
      "theoretical_loss": 3.500148193271599,
      "tokens_seen": 1579679744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026335005015045137,
      "loss": 2.6317,
      "theoretical_loss": 3.5001355172564157,
      "tokens_seen": 1579745280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026334002006018055,
      "loss": 2.4367,
      "theoretical_loss": 3.5001228419143233,
      "tokens_seen": 1579810816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026332998996990973,
      "loss": 2.4399,
      "theoretical_loss": 3.5001101672452593,
      "tokens_seen": 1579876352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633199598796389,
      "loss": 2.3521,
      "theoretical_loss": 3.500097493249159,
      "tokens_seen": 1579941888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026330992978936815,
      "loss": 2.3943,
      "theoretical_loss": 3.5000848199259593,
      "tokens_seen": 1580007424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632998996990973,
      "loss": 2.5588,
      "theoretical_loss": 3.5000721472755965,
      "tokens_seen": 1580072960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632898696088265,
      "loss": 2.6983,
      "theoretical_loss": 3.5000594752980065,
      "tokens_seen": 1580138496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632798395185557,
      "loss": 2.5467,
      "theoretical_loss": 3.5000468039931265,
      "tokens_seen": 1580204032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026326980942828487,
      "loss": 2.529,
      "theoretical_loss": 3.5000341333608924,
      "tokens_seen": 1580269568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026325977933801405,
      "loss": 2.5671,
      "theoretical_loss": 3.5000214634012403,
      "tokens_seen": 1580335104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026324974924774323,
      "loss": 2.5333,
      "theoretical_loss": 3.500008794114107,
      "tokens_seen": 1580400640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632397191574724,
      "loss": 2.5233,
      "theoretical_loss": 3.49999612549943,
      "tokens_seen": 1580466176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026322968906720165,
      "loss": 2.6768,
      "theoretical_loss": 3.4999834575571436,
      "tokens_seen": 1580531712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632196589769308,
      "loss": 2.5567,
      "theoretical_loss": 3.4999707902871853,
      "tokens_seen": 1580597248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026320962888666,
      "loss": 2.6232,
      "theoretical_loss": 3.499958123689492,
      "tokens_seen": 1580662784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026319959879638914,
      "loss": 2.447,
      "theoretical_loss": 3.499945457763999,
      "tokens_seen": 1580728320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631895687061184,
      "loss": 2.5432,
      "theoretical_loss": 3.499932792510644,
      "tokens_seen": 1580793856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026317953861584756,
      "loss": 2.6017,
      "theoretical_loss": 3.4999201279293626,
      "tokens_seen": 1580859392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026316950852557674,
      "loss": 2.6208,
      "theoretical_loss": 3.4999074640200916,
      "tokens_seen": 1580924928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631594784353059,
      "loss": 2.8345,
      "theoretical_loss": 3.4998948007827675,
      "tokens_seen": 1580990464
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 874187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.422612428665161,
      "objective/train/theoretical_loss": 3.4998821382173264,
      "objective/train/tokens_used": 1601516000,
      "theoretical_loss": 3.4998821382173264,
      "tokens_seen": 1581056000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631494483450351,
      "loss": 2.5285,
      "theoretical_loss": 3.4998821382173264,
      "tokens_seen": 1581056000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631394182547643,
      "loss": 2.6874,
      "theoretical_loss": 3.4998694763237053,
      "tokens_seen": 1581121536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631293881644935,
      "loss": 2.4081,
      "theoretical_loss": 3.4998568151018405,
      "tokens_seen": 1581187072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026311935807422264,
      "loss": 2.5127,
      "theoretical_loss": 3.499844154551669,
      "tokens_seen": 1581252608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631093279839519,
      "loss": 2.4452,
      "theoretical_loss": 3.499831494673127,
      "tokens_seen": 1581318144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026309929789368106,
      "loss": 2.4203,
      "theoretical_loss": 3.4998188354661504,
      "tokens_seen": 1581383680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026308926780341024,
      "loss": 2.3916,
      "theoretical_loss": 3.4998061769306767,
      "tokens_seen": 1581449216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630792377131394,
      "loss": 2.4841,
      "theoretical_loss": 3.4997935190666416,
      "tokens_seen": 1581514752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630692076228686,
      "loss": 2.4694,
      "theoretical_loss": 3.4997808618739823,
      "tokens_seen": 1581580288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630591775325978,
      "loss": 2.4382,
      "theoretical_loss": 3.4997682053526358,
      "tokens_seen": 1581645824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263049147442327,
      "loss": 2.5936,
      "theoretical_loss": 3.499755549502537,
      "tokens_seen": 1581711360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026303911735205615,
      "loss": 2.7564,
      "theoretical_loss": 3.4997428943236244,
      "tokens_seen": 1581776896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630290872617854,
      "loss": 2.6199,
      "theoretical_loss": 3.4997302398158334,
      "tokens_seen": 1581842432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630190571715145,
      "loss": 2.8209,
      "theoretical_loss": 3.499717585979101,
      "tokens_seen": 1581907968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026300902708124374,
      "loss": 2.5797,
      "theoretical_loss": 3.499704932813364,
      "tokens_seen": 1581973504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629989969909729,
      "loss": 2.4459,
      "theoretical_loss": 3.499692280318558,
      "tokens_seen": 1582039040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629889669007021,
      "loss": 2.7433,
      "theoretical_loss": 3.499679628494621,
      "tokens_seen": 1582104576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629789368104313,
      "loss": 2.3748,
      "theoretical_loss": 3.4996669773414895,
      "tokens_seen": 1582170112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026296890672016047,
      "loss": 2.4761,
      "theoretical_loss": 3.4996543268590994,
      "tokens_seen": 1582235648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026295887662988965,
      "loss": 2.5291,
      "theoretical_loss": 3.4996416770473875,
      "tokens_seen": 1582301184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629488465396189,
      "loss": 2.2966,
      "theoretical_loss": 3.4996290279062907,
      "tokens_seen": 1582366720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262938816449348,
      "loss": 2.65,
      "theoretical_loss": 3.4996163794357456,
      "tokens_seen": 1582432256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026292878635907725,
      "loss": 2.5125,
      "theoretical_loss": 3.4996037316356894,
      "tokens_seen": 1582497792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026291875626880643,
      "loss": 2.7155,
      "theoretical_loss": 3.499591084506058,
      "tokens_seen": 1582563328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629087261785356,
      "loss": 2.4625,
      "theoretical_loss": 3.499578438046788,
      "tokens_seen": 1582628864
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 874795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7404651641845703,
      "objective/train/theoretical_loss": 3.499565792257817,
      "objective/train/tokens_used": 1603154400,
      "theoretical_loss": 3.499565792257817,
      "tokens_seen": 1582694400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628986960882648,
      "loss": 2.7593,
      "theoretical_loss": 3.499565792257817,
      "tokens_seen": 1582694400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026288866599799397,
      "loss": 2.5075,
      "theoretical_loss": 3.499553147139081,
      "tokens_seen": 1582759936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026287863590772315,
      "loss": 2.8189,
      "theoretical_loss": 3.499540502690517,
      "tokens_seen": 1582825472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628686058174524,
      "loss": 2.4139,
      "theoretical_loss": 3.4995278589120615,
      "tokens_seen": 1582891008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628585757271815,
      "loss": 2.5647,
      "theoretical_loss": 3.499515215803652,
      "tokens_seen": 1582956544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026284854563691075,
      "loss": 2.5506,
      "theoretical_loss": 3.4995025733652243,
      "tokens_seen": 1583022080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628385155466399,
      "loss": 2.525,
      "theoretical_loss": 3.4994899315967154,
      "tokens_seen": 1583087616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628284854563691,
      "loss": 2.4954,
      "theoretical_loss": 3.4994772904980627,
      "tokens_seen": 1583153152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628184553660983,
      "loss": 2.4888,
      "theoretical_loss": 3.4994646500692017,
      "tokens_seen": 1583218688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628084252758275,
      "loss": 2.4868,
      "theoretical_loss": 3.499452010310071,
      "tokens_seen": 1583284224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026279839518555666,
      "loss": 2.2767,
      "theoretical_loss": 3.499439371220606,
      "tokens_seen": 1583349760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627883650952859,
      "loss": 2.7694,
      "theoretical_loss": 3.4994267328007433,
      "tokens_seen": 1583415296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262778335005015,
      "loss": 2.5017,
      "theoretical_loss": 3.499414095050421,
      "tokens_seen": 1583480832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026276830491474425,
      "loss": 2.6436,
      "theoretical_loss": 3.499401457969575,
      "tokens_seen": 1583546368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026275827482447343,
      "loss": 2.5037,
      "theoretical_loss": 3.4993888215581426,
      "tokens_seen": 1583611904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627482447342026,
      "loss": 2.8348,
      "theoretical_loss": 3.49937618581606,
      "tokens_seen": 1583677440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026273821464393185,
      "loss": 2.6411,
      "theoretical_loss": 3.499363550743265,
      "tokens_seen": 1583742976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262728184553661,
      "loss": 2.5223,
      "theoretical_loss": 3.4993509163396936,
      "tokens_seen": 1583808512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627181544633902,
      "loss": 2.4816,
      "theoretical_loss": 3.499338282605283,
      "tokens_seen": 1583874048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026270812437311934,
      "loss": 2.5636,
      "theoretical_loss": 3.49932564953997,
      "tokens_seen": 1583939584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626980942828486,
      "loss": 2.6754,
      "theoretical_loss": 3.4993130171436917,
      "tokens_seen": 1584005120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026268806419257776,
      "loss": 2.8138,
      "theoretical_loss": 3.4993003854163844,
      "tokens_seen": 1584070656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026267803410230694,
      "loss": 2.6237,
      "theoretical_loss": 3.4992877543579857,
      "tokens_seen": 1584136192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626680040120361,
      "loss": 2.4506,
      "theoretical_loss": 3.4992751239684323,
      "tokens_seen": 1584201728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626579739217653,
      "loss": 2.6602,
      "theoretical_loss": 3.499262494247661,
      "tokens_seen": 1584267264
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 875946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.513859510421753,
      "objective/train/theoretical_loss": 3.4992498651956088,
      "objective/train/tokens_used": 1604792800,
      "theoretical_loss": 3.4992498651956088,
      "tokens_seen": 1584332800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626479438314945,
      "loss": 2.3576,
      "theoretical_loss": 3.4992498651956088,
      "tokens_seen": 1584332800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626379137412237,
      "loss": 2.5001,
      "theoretical_loss": 3.4992372368122124,
      "tokens_seen": 1584398336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026262788365095284,
      "loss": 2.5837,
      "theoretical_loss": 3.499224609097409,
      "tokens_seen": 1584463872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626178535606821,
      "loss": 2.7267,
      "theoretical_loss": 3.4992119820511354,
      "tokens_seen": 1584529408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026260782347041126,
      "loss": 2.3748,
      "theoretical_loss": 3.499199355673329,
      "tokens_seen": 1584594944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026259779338014044,
      "loss": 2.6236,
      "theoretical_loss": 3.4991867299639257,
      "tokens_seen": 1584660480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625877632898696,
      "loss": 2.7491,
      "theoretical_loss": 3.499174104922864,
      "tokens_seen": 1584726016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625777331995988,
      "loss": 2.3001,
      "theoretical_loss": 3.4991614805500797,
      "tokens_seen": 1584791552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262567703109328,
      "loss": 2.7681,
      "theoretical_loss": 3.49914885684551,
      "tokens_seen": 1584857088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625576730190572,
      "loss": 2.644,
      "theoretical_loss": 3.499136233809092,
      "tokens_seen": 1584922624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026254764292878635,
      "loss": 2.2981,
      "theoretical_loss": 3.4991236114407633,
      "tokens_seen": 1584988160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625376128385156,
      "loss": 2.6859,
      "theoretical_loss": 3.49911098974046,
      "tokens_seen": 1585053696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625275827482447,
      "loss": 2.6707,
      "theoretical_loss": 3.49909836870812,
      "tokens_seen": 1585119232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026251755265797394,
      "loss": 2.5614,
      "theoretical_loss": 3.4990857483436795,
      "tokens_seen": 1585184768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625075225677031,
      "loss": 2.3751,
      "theoretical_loss": 3.4990731286470758,
      "tokens_seen": 1585250304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624974924774323,
      "loss": 2.752,
      "theoretical_loss": 3.4990605096182463,
      "tokens_seen": 1585315840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624874623871615,
      "loss": 2.6411,
      "theoretical_loss": 3.4990478912571277,
      "tokens_seen": 1585381376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026247743229689067,
      "loss": 2.6529,
      "theoretical_loss": 3.499035273563657,
      "tokens_seen": 1585446912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026246740220661985,
      "loss": 2.6479,
      "theoretical_loss": 3.499022656537772,
      "tokens_seen": 1585512448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624573721163491,
      "loss": 2.6785,
      "theoretical_loss": 3.4990100401794084,
      "tokens_seen": 1585577984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624473420260782,
      "loss": 2.5063,
      "theoretical_loss": 3.498997424488505,
      "tokens_seen": 1585643520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026243731193580745,
      "loss": 2.7766,
      "theoretical_loss": 3.498984809464998,
      "tokens_seen": 1585709056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026242728184553663,
      "loss": 2.7247,
      "theoretical_loss": 3.498972195108824,
      "tokens_seen": 1585774592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624172517552658,
      "loss": 2.5592,
      "theoretical_loss": 3.4989595814199212,
      "tokens_seen": 1585840128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262407221664995,
      "loss": 2.7065,
      "theoretical_loss": 3.4989469683982257,
      "tokens_seen": 1585905664
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 876427,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.855236053466797,
      "objective/train/theoretical_loss": 3.4989343560436756,
      "objective/train/tokens_used": 1606431200,
      "theoretical_loss": 3.4989343560436756,
      "tokens_seen": 1585971200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026239719157472417,
      "loss": 2.5245,
      "theoretical_loss": 3.4989343560436756,
      "tokens_seen": 1585971200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026238716148445335,
      "loss": 2.444,
      "theoretical_loss": 3.4989217443562075,
      "tokens_seen": 1586036736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623771313941826,
      "loss": 2.449,
      "theoretical_loss": 3.4989091333357587,
      "tokens_seen": 1586102272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623671013039117,
      "loss": 2.5151,
      "theoretical_loss": 3.498896522982266,
      "tokens_seen": 1586167808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026235707121364095,
      "loss": 2.4415,
      "theoretical_loss": 3.4988839132956673,
      "tokens_seen": 1586233344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623470411233701,
      "loss": 2.6342,
      "theoretical_loss": 3.498871304275899,
      "tokens_seen": 1586298880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623370110330993,
      "loss": 2.492,
      "theoretical_loss": 3.498858695922899,
      "tokens_seen": 1586364416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623269809428285,
      "loss": 2.7537,
      "theoretical_loss": 3.498846088236604,
      "tokens_seen": 1586429952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623169508525577,
      "loss": 2.4092,
      "theoretical_loss": 3.4988334812169515,
      "tokens_seen": 1586495488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026230692076228686,
      "loss": 2.4392,
      "theoretical_loss": 3.4988208748638785,
      "tokens_seen": 1586561024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622968906720161,
      "loss": 2.4274,
      "theoretical_loss": 3.4988082691773226,
      "tokens_seen": 1586626560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622868605817452,
      "loss": 2.6292,
      "theoretical_loss": 3.49879566415722,
      "tokens_seen": 1586692096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026227683049147445,
      "loss": 2.6117,
      "theoretical_loss": 3.4987830598035092,
      "tokens_seen": 1586757632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622668004012036,
      "loss": 2.8439,
      "theoretical_loss": 3.498770456116127,
      "tokens_seen": 1586823168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622567703109328,
      "loss": 2.6063,
      "theoretical_loss": 3.49875785309501,
      "tokens_seen": 1586888704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262246740220662,
      "loss": 2.6568,
      "theoretical_loss": 3.4987452507400967,
      "tokens_seen": 1586954240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622367101303912,
      "loss": 2.6322,
      "theoretical_loss": 3.4987326490513238,
      "tokens_seen": 1587019776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026222668004012036,
      "loss": 2.299,
      "theoretical_loss": 3.498720048028628,
      "tokens_seen": 1587085312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026221664994984954,
      "loss": 2.609,
      "theoretical_loss": 3.498707447671947,
      "tokens_seen": 1587150848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622066198595787,
      "loss": 2.7251,
      "theoretical_loss": 3.498694847981219,
      "tokens_seen": 1587216384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026219658976930796,
      "loss": 2.4335,
      "theoretical_loss": 3.49868224895638,
      "tokens_seen": 1587281920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621865596790371,
      "loss": 2.5573,
      "theoretical_loss": 3.4986696505973676,
      "tokens_seen": 1587347456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621765295887663,
      "loss": 2.6674,
      "theoretical_loss": 3.49865705290412,
      "tokens_seen": 1587412992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026216649949849545,
      "loss": 2.6773,
      "theoretical_loss": 3.4986444558765735,
      "tokens_seen": 1587478528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621564694082247,
      "loss": 2.4559,
      "theoretical_loss": 3.4986318595146653,
      "tokens_seen": 1587544064
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 877583,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5934700965881348,
      "objective/train/theoretical_loss": 3.498619263818334,
      "objective/train/tokens_used": 1608069600,
      "theoretical_loss": 3.498619263818334,
      "tokens_seen": 1587609600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026214643931795386,
      "loss": 2.3889,
      "theoretical_loss": 3.498619263818334,
      "tokens_seen": 1587609600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026213640922768304,
      "loss": 2.5724,
      "theoretical_loss": 3.4986066687875157,
      "tokens_seen": 1587675136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621263791374122,
      "loss": 2.4605,
      "theoretical_loss": 3.498594074422149,
      "tokens_seen": 1587740672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026211634904714146,
      "loss": 2.5194,
      "theoretical_loss": 3.4985814807221702,
      "tokens_seen": 1587806208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621063189568706,
      "loss": 2.5424,
      "theoretical_loss": 3.498568887687517,
      "tokens_seen": 1587871744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620962888665998,
      "loss": 2.4043,
      "theoretical_loss": 3.498556295318127,
      "tokens_seen": 1587937280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026208625877632895,
      "loss": 2.6208,
      "theoretical_loss": 3.4985437036139375,
      "tokens_seen": 1588002816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620762286860582,
      "loss": 2.6661,
      "theoretical_loss": 3.4985311125748852,
      "tokens_seen": 1588068352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026206619859578737,
      "loss": 2.3886,
      "theoretical_loss": 3.498518522200909,
      "tokens_seen": 1588133888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026205616850551655,
      "loss": 2.6985,
      "theoretical_loss": 3.498505932491945,
      "tokens_seen": 1588199424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026204613841524573,
      "loss": 2.5593,
      "theoretical_loss": 3.4984933434479313,
      "tokens_seen": 1588264960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620361083249749,
      "loss": 2.819,
      "theoretical_loss": 3.4984807550688055,
      "tokens_seen": 1588330496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620260782347041,
      "loss": 2.5413,
      "theoretical_loss": 3.4984681673545044,
      "tokens_seen": 1588396032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620160481444333,
      "loss": 2.6242,
      "theoretical_loss": 3.4984555803049657,
      "tokens_seen": 1588461568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620060180541625,
      "loss": 2.2904,
      "theoretical_loss": 3.4984429939201274,
      "tokens_seen": 1588527104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619959879638917,
      "loss": 2.5513,
      "theoretical_loss": 3.4984304081999262,
      "tokens_seen": 1588592640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026198595787362087,
      "loss": 2.7006,
      "theoretical_loss": 3.4984178231443,
      "tokens_seen": 1588658176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026197592778335005,
      "loss": 2.8216,
      "theoretical_loss": 3.498405238753187,
      "tokens_seen": 1588723712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619658976930793,
      "loss": 2.2396,
      "theoretical_loss": 3.498392655026523,
      "tokens_seen": 1588789248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619558676028084,
      "loss": 2.4445,
      "theoretical_loss": 3.4983800719642466,
      "tokens_seen": 1588854784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026194583751253765,
      "loss": 2.5329,
      "theoretical_loss": 3.4983674895662955,
      "tokens_seen": 1588920320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026193580742226683,
      "loss": 2.5839,
      "theoretical_loss": 3.4983549078326064,
      "tokens_seen": 1588985856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261925777331996,
      "loss": 2.563,
      "theoretical_loss": 3.498342326763118,
      "tokens_seen": 1589051392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619157472417252,
      "loss": 2.5279,
      "theoretical_loss": 3.4983297463577667,
      "tokens_seen": 1589116928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026190571715145437,
      "loss": 2.613,
      "theoretical_loss": 3.4983171666164905,
      "tokens_seen": 1589182464
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 878291,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9362738132476807,
      "objective/train/theoretical_loss": 3.498304587539227,
      "objective/train/tokens_used": 1609708000,
      "theoretical_loss": 3.498304587539227,
      "tokens_seen": 1589248000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026189568706118355,
      "loss": 2.7078,
      "theoretical_loss": 3.498304587539227,
      "tokens_seen": 1589248000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618856569709128,
      "loss": 2.4795,
      "theoretical_loss": 3.498292009125914,
      "tokens_seen": 1589313536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618756268806419,
      "loss": 2.5528,
      "theoretical_loss": 3.498279431376489,
      "tokens_seen": 1589379072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026186559679037115,
      "loss": 2.3938,
      "theoretical_loss": 3.498266854290889,
      "tokens_seen": 1589444608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618555667001003,
      "loss": 2.7332,
      "theoretical_loss": 3.4982542778690524,
      "tokens_seen": 1589510144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618455366098295,
      "loss": 2.5084,
      "theoretical_loss": 3.4982417021109162,
      "tokens_seen": 1589575680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618355065195587,
      "loss": 2.5602,
      "theoretical_loss": 3.498229127016418,
      "tokens_seen": 1589641216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618254764292879,
      "loss": 2.6266,
      "theoretical_loss": 3.498216552585496,
      "tokens_seen": 1589706752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026181544633901706,
      "loss": 2.4609,
      "theoretical_loss": 3.498203978818087,
      "tokens_seen": 1589772288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618054162487463,
      "loss": 2.2296,
      "theoretical_loss": 3.49819140571413,
      "tokens_seen": 1589837824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617953861584754,
      "loss": 2.5298,
      "theoretical_loss": 3.4981788332735615,
      "tokens_seen": 1589903360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026178535606820465,
      "loss": 2.7224,
      "theoretical_loss": 3.498166261496319,
      "tokens_seen": 1589968896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617753259779338,
      "loss": 2.8246,
      "theoretical_loss": 3.498153690382341,
      "tokens_seen": 1590034432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261765295887663,
      "loss": 2.6064,
      "theoretical_loss": 3.4981411199315646,
      "tokens_seen": 1590099968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617552657973922,
      "loss": 2.6017,
      "theoretical_loss": 3.498128550143928,
      "tokens_seen": 1590165504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617452357071214,
      "loss": 2.4292,
      "theoretical_loss": 3.4981159810193683,
      "tokens_seen": 1590231040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026173520561685056,
      "loss": 2.4612,
      "theoretical_loss": 3.4981034125578234,
      "tokens_seen": 1590296576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026172517552657974,
      "loss": 2.5168,
      "theoretical_loss": 3.498090844759231,
      "tokens_seen": 1590362112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617151454363089,
      "loss": 2.6873,
      "theoretical_loss": 3.498078277623529,
      "tokens_seen": 1590427648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026170511534603816,
      "loss": 2.5425,
      "theoretical_loss": 3.4980657111506552,
      "tokens_seen": 1590493184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616950852557673,
      "loss": 2.6458,
      "theoretical_loss": 3.4980531453405463,
      "tokens_seen": 1590558720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616850551654965,
      "loss": 2.3692,
      "theoretical_loss": 3.4980405801931416,
      "tokens_seen": 1590624256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026167502507522565,
      "loss": 2.5229,
      "theoretical_loss": 3.498028015708378,
      "tokens_seen": 1590689792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616649949849549,
      "loss": 2.5577,
      "theoretical_loss": 3.498015451886193,
      "tokens_seen": 1590755328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026165496489468406,
      "loss": 2.6216,
      "theoretical_loss": 3.4980028887265253,
      "tokens_seen": 1590820864
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 879490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6622262001037598,
      "objective/train/theoretical_loss": 3.4979903262293117,
      "objective/train/tokens_used": 1611346400,
      "theoretical_loss": 3.4979903262293117,
      "tokens_seen": 1590886400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026164493480441324,
      "loss": 2.6067,
      "theoretical_loss": 3.4979903262293117,
      "tokens_seen": 1590886400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616349047141424,
      "loss": 2.6207,
      "theoretical_loss": 3.4979777643944905,
      "tokens_seen": 1590951936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026162487462387166,
      "loss": 2.5697,
      "theoretical_loss": 3.497965203221999,
      "tokens_seen": 1591017472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616148445336008,
      "loss": 2.7107,
      "theoretical_loss": 3.497952642711776,
      "tokens_seen": 1591083008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026160481444333,
      "loss": 2.6887,
      "theoretical_loss": 3.4979400828637583,
      "tokens_seen": 1591148544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026159478435305915,
      "loss": 2.4801,
      "theoretical_loss": 3.497927523677884,
      "tokens_seen": 1591214080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615847542627884,
      "loss": 2.3117,
      "theoretical_loss": 3.497914965154091,
      "tokens_seen": 1591279616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026157472417251757,
      "loss": 2.4725,
      "theoretical_loss": 3.4979024072923175,
      "tokens_seen": 1591345152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026156469408224675,
      "loss": 2.6624,
      "theoretical_loss": 3.4978898500925006,
      "tokens_seen": 1591410688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026155466399197593,
      "loss": 2.5702,
      "theoretical_loss": 3.4978772935545788,
      "tokens_seen": 1591476224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615446339017051,
      "loss": 2.6589,
      "theoretical_loss": 3.4978647376784897,
      "tokens_seen": 1591541760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615346038114343,
      "loss": 2.7033,
      "theoretical_loss": 3.497852182464171,
      "tokens_seen": 1591607296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615245737211635,
      "loss": 2.5444,
      "theoretical_loss": 3.4978396279115604,
      "tokens_seen": 1591672832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026151454363089265,
      "loss": 2.4963,
      "theoretical_loss": 3.497827074020597,
      "tokens_seen": 1591738368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615045135406219,
      "loss": 2.3806,
      "theoretical_loss": 3.4978145207912172,
      "tokens_seen": 1591803904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261494483450351,
      "loss": 2.6509,
      "theoretical_loss": 3.497801968223359,
      "tokens_seen": 1591869440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026148445336008025,
      "loss": 2.3885,
      "theoretical_loss": 3.4977894163169614,
      "tokens_seen": 1591934976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026147442326980943,
      "loss": 2.6258,
      "theoretical_loss": 3.4977768650719616,
      "tokens_seen": 1592000512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614643931795386,
      "loss": 2.8348,
      "theoretical_loss": 3.4977643144882977,
      "tokens_seen": 1592066048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614543630892678,
      "loss": 2.6699,
      "theoretical_loss": 3.4977517645659075,
      "tokens_seen": 1592131584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026144433299899703,
      "loss": 2.7163,
      "theoretical_loss": 3.4977392153047298,
      "tokens_seen": 1592197120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026143430290872616,
      "loss": 2.6766,
      "theoretical_loss": 3.497726666704701,
      "tokens_seen": 1592262656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614242728184554,
      "loss": 2.5892,
      "theoretical_loss": 3.49771411876576,
      "tokens_seen": 1592328192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614142427281845,
      "loss": 2.672,
      "theoretical_loss": 3.4977015714878443,
      "tokens_seen": 1592393728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026140421263791375,
      "loss": 2.6054,
      "theoretical_loss": 3.4976890248708923,
      "tokens_seen": 1592459264
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 880099,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.060887575149536,
      "objective/train/theoretical_loss": 3.4976764789148422,
      "objective/train/tokens_used": 1612984800,
      "theoretical_loss": 3.4976764789148422,
      "tokens_seen": 1592524800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026139418254764293,
      "loss": 2.8658,
      "theoretical_loss": 3.4976764789148422,
      "tokens_seen": 1592524800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613841524573721,
      "loss": 2.648,
      "theoretical_loss": 3.497663933619631,
      "tokens_seen": 1592590336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613741223671013,
      "loss": 2.7317,
      "theoretical_loss": 3.4976513889851977,
      "tokens_seen": 1592655872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613640922768305,
      "loss": 2.6132,
      "theoretical_loss": 3.4976388450114797,
      "tokens_seen": 1592721408
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026135406218655966,
      "loss": 2.5033,
      "theoretical_loss": 3.497626301698416,
      "tokens_seen": 1592786944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613440320962889,
      "loss": 2.4996,
      "theoretical_loss": 3.4976137590459433,
      "tokens_seen": 1592852480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261334002006018,
      "loss": 2.3833,
      "theoretical_loss": 3.497601217054,
      "tokens_seen": 1592918016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026132397191574726,
      "loss": 2.5479,
      "theoretical_loss": 3.497588675722525,
      "tokens_seen": 1592983552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613139418254764,
      "loss": 2.6199,
      "theoretical_loss": 3.4975761350514554,
      "tokens_seen": 1593049088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613039117352056,
      "loss": 2.7099,
      "theoretical_loss": 3.4975635950407296,
      "tokens_seen": 1593114624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612938816449348,
      "loss": 2.4258,
      "theoretical_loss": 3.497551055690286,
      "tokens_seen": 1593180160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261283851554664,
      "loss": 2.8454,
      "theoretical_loss": 3.4975385170000615,
      "tokens_seen": 1593245696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026127382146439316,
      "loss": 2.5293,
      "theoretical_loss": 3.4975259789699953,
      "tokens_seen": 1593311232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612637913741224,
      "loss": 2.6926,
      "theoretical_loss": 3.4975134416000255,
      "tokens_seen": 1593376768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612537612838516,
      "loss": 2.6239,
      "theoretical_loss": 3.4975009048900896,
      "tokens_seen": 1593442304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026124373119358076,
      "loss": 2.5472,
      "theoretical_loss": 3.497488368840126,
      "tokens_seen": 1593507840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026123370110330994,
      "loss": 2.6316,
      "theoretical_loss": 3.4974758334500726,
      "tokens_seen": 1593573376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612236710130391,
      "loss": 2.5666,
      "theoretical_loss": 3.497463298719868,
      "tokens_seen": 1593638912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026121364092276836,
      "loss": 2.8978,
      "theoretical_loss": 3.49745076464945,
      "tokens_seen": 1593704448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612036108324975,
      "loss": 2.622,
      "theoretical_loss": 3.4974382312387573,
      "tokens_seen": 1593769984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611935807422267,
      "loss": 2.5524,
      "theoretical_loss": 3.497425698487727,
      "tokens_seen": 1593835520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026118355065195585,
      "loss": 2.4529,
      "theoretical_loss": 3.497413166396298,
      "tokens_seen": 1593901056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611735205616851,
      "loss": 2.5538,
      "theoretical_loss": 3.497400634964408,
      "tokens_seen": 1593966592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026116349047141426,
      "loss": 2.4304,
      "theoretical_loss": 3.497388104191996,
      "tokens_seen": 1594032128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026115346038114344,
      "loss": 2.651,
      "theoretical_loss": 3.4973755740789993,
      "tokens_seen": 1594097664
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 881131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6384732723236084,
      "objective/train/theoretical_loss": 3.4973630446253563,
      "objective/train/tokens_used": 1614623200,
      "theoretical_loss": 3.4973630446253563,
      "tokens_seen": 1594163200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611434302908726,
      "loss": 2.5996,
      "theoretical_loss": 3.4973630446253563,
      "tokens_seen": 1594163200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026113340020060186,
      "loss": 2.7237,
      "theoretical_loss": 3.4973505158310054,
      "tokens_seen": 1594228736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261123370110331,
      "loss": 2.7224,
      "theoretical_loss": 3.497337987695885,
      "tokens_seen": 1594294272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611133400200602,
      "loss": 2.4421,
      "theoretical_loss": 3.4973254602199333,
      "tokens_seen": 1594359808
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026110330992978935,
      "loss": 2.4702,
      "theoretical_loss": 3.4973129334030877,
      "tokens_seen": 1594425344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610932798395186,
      "loss": 2.7693,
      "theoretical_loss": 3.4973004072452873,
      "tokens_seen": 1594490880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026108324974924777,
      "loss": 2.5035,
      "theoretical_loss": 3.4972878817464697,
      "tokens_seen": 1594556416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026107321965897695,
      "loss": 2.7898,
      "theoretical_loss": 3.4972753569065738,
      "tokens_seen": 1594621952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026106318956870613,
      "loss": 2.5371,
      "theoretical_loss": 3.4972628327255375,
      "tokens_seen": 1594687488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610531594784353,
      "loss": 2.617,
      "theoretical_loss": 3.4972503092032996,
      "tokens_seen": 1594753024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610431293881645,
      "loss": 2.4432,
      "theoretical_loss": 3.497237786339797,
      "tokens_seen": 1594818560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610330992978937,
      "loss": 2.5231,
      "theoretical_loss": 3.4972252641349693,
      "tokens_seen": 1594884096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026102306920762285,
      "loss": 2.6498,
      "theoretical_loss": 3.4972127425887547,
      "tokens_seen": 1594949632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610130391173521,
      "loss": 2.6556,
      "theoretical_loss": 3.4972002217010907,
      "tokens_seen": 1595015168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610030090270812,
      "loss": 2.5335,
      "theoretical_loss": 3.4971877014719164,
      "tokens_seen": 1595080704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026099297893681045,
      "loss": 2.7581,
      "theoretical_loss": 3.49717518190117,
      "tokens_seen": 1595146240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026098294884653963,
      "loss": 2.5752,
      "theoretical_loss": 3.4971626629887895,
      "tokens_seen": 1595211776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609729187562688,
      "loss": 2.8398,
      "theoretical_loss": 3.497150144734713,
      "tokens_seen": 1595277312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260962888665998,
      "loss": 2.5584,
      "theoretical_loss": 3.49713762713888,
      "tokens_seen": 1595342848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026095285857572723,
      "loss": 2.6581,
      "theoretical_loss": 3.4971251102012273,
      "tokens_seen": 1595408384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026094282848545636,
      "loss": 2.3576,
      "theoretical_loss": 3.4971125939216936,
      "tokens_seen": 1595473920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609327983951856,
      "loss": 2.801,
      "theoretical_loss": 3.4971000783002184,
      "tokens_seen": 1595539456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609227683049147,
      "loss": 2.613,
      "theoretical_loss": 3.4970875633367395,
      "tokens_seen": 1595604992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026091273821464395,
      "loss": 2.5997,
      "theoretical_loss": 3.497075049031195,
      "tokens_seen": 1595670528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026090270812437313,
      "loss": 2.4345,
      "theoretical_loss": 3.497062535383523,
      "tokens_seen": 1595736064
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 881793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.461489677429199,
      "objective/train/theoretical_loss": 3.4970500223936627,
      "objective/train/tokens_used": 1616261600,
      "theoretical_loss": 3.4970500223936627,
      "tokens_seen": 1595801600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608926780341023,
      "loss": 2.7317,
      "theoretical_loss": 3.4970500223936627,
      "tokens_seen": 1595801600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608826479438315,
      "loss": 2.2827,
      "theoretical_loss": 3.4970375100615523,
      "tokens_seen": 1595867136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608726178535607,
      "loss": 2.855,
      "theoretical_loss": 3.49702499838713,
      "tokens_seen": 1595932672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026086258776328986,
      "loss": 2.6554,
      "theoretical_loss": 3.497012487370334,
      "tokens_seen": 1595998208
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608525576730191,
      "loss": 2.4999,
      "theoretical_loss": 3.496999977011103,
      "tokens_seen": 1596063744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608425275827482,
      "loss": 2.4595,
      "theoretical_loss": 3.4969874673093755,
      "tokens_seen": 1596129280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026083249749247746,
      "loss": 2.7592,
      "theoretical_loss": 3.49697495826509,
      "tokens_seen": 1596194816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608224674022066,
      "loss": 2.752,
      "theoretical_loss": 3.496962449878185,
      "tokens_seen": 1596260352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608124373119358,
      "loss": 2.4624,
      "theoretical_loss": 3.4969499421485986,
      "tokens_seen": 1596325888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260802407221665,
      "loss": 2.5857,
      "theoretical_loss": 3.496937435076269,
      "tokens_seen": 1596391424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607923771313942,
      "loss": 2.7426,
      "theoretical_loss": 3.4969249286611364,
      "tokens_seen": 1596456960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026078234704112336,
      "loss": 2.6939,
      "theoretical_loss": 3.4969124229031374,
      "tokens_seen": 1596522496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607723169508526,
      "loss": 2.6724,
      "theoretical_loss": 3.496899917802211,
      "tokens_seen": 1596588032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607622868605817,
      "loss": 2.6715,
      "theoretical_loss": 3.496887413358296,
      "tokens_seen": 1596653568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026075225677031096,
      "loss": 2.6495,
      "theoretical_loss": 3.4968749095713307,
      "tokens_seen": 1596719104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607422266800401,
      "loss": 2.4519,
      "theoretical_loss": 3.496862406441254,
      "tokens_seen": 1596784640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607321965897693,
      "loss": 2.362,
      "theoretical_loss": 3.4968499039680037,
      "tokens_seen": 1596850176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607221664994985,
      "loss": 2.5348,
      "theoretical_loss": 3.4968374021515194,
      "tokens_seen": 1596915712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607121364092277,
      "loss": 2.9297,
      "theoretical_loss": 3.4968249009917383,
      "tokens_seen": 1596981248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026070210631895687,
      "loss": 2.5659,
      "theoretical_loss": 3.4968124004886003,
      "tokens_seen": 1597046784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026069207622868605,
      "loss": 2.5782,
      "theoretical_loss": 3.496799900642043,
      "tokens_seen": 1597112320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026068204613841523,
      "loss": 2.5968,
      "theoretical_loss": 3.4967874014520057,
      "tokens_seen": 1597177856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026067201604814446,
      "loss": 2.4417,
      "theoretical_loss": 3.4967749029184265,
      "tokens_seen": 1597243392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606619859578736,
      "loss": 2.5994,
      "theoretical_loss": 3.4967624050412436,
      "tokens_seen": 1597308928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606519558676028,
      "loss": 2.4952,
      "theoretical_loss": 3.496749907820397,
      "tokens_seen": 1597374464
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 882868,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.732029914855957,
      "objective/train/theoretical_loss": 3.4967374112558236,
      "objective/train/tokens_used": 1617900000,
      "theoretical_loss": 3.4967374112558236,
      "tokens_seen": 1597440000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026064192577733195,
      "loss": 2.4951,
      "theoretical_loss": 3.4967374112558236,
      "tokens_seen": 1597440000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606318956870612,
      "loss": 2.6173,
      "theoretical_loss": 3.496724915347463,
      "tokens_seen": 1597505536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026062186559679037,
      "loss": 2.4632,
      "theoretical_loss": 3.496712420095254,
      "tokens_seen": 1597571072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026061183550651955,
      "loss": 2.6571,
      "theoretical_loss": 3.4966999254991347,
      "tokens_seen": 1597636608
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026060180541624873,
      "loss": 2.6554,
      "theoretical_loss": 3.4966874315590437,
      "tokens_seen": 1597702144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026059177532597797,
      "loss": 2.6584,
      "theoretical_loss": 3.49667493827492,
      "tokens_seen": 1597767680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605817452357071,
      "loss": 2.771,
      "theoretical_loss": 3.4966624456467024,
      "tokens_seen": 1597833216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026057171514543633,
      "loss": 2.6361,
      "theoretical_loss": 3.496649953674329,
      "tokens_seen": 1597898752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026056168505516546,
      "loss": 2.6624,
      "theoretical_loss": 3.4966374623577385,
      "tokens_seen": 1597964288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605516549648947,
      "loss": 2.666,
      "theoretical_loss": 3.4966249716968703,
      "tokens_seen": 1598029824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026054162487462387,
      "loss": 2.6346,
      "theoretical_loss": 3.4966124816916624,
      "tokens_seen": 1598095360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026053159478435305,
      "loss": 2.5328,
      "theoretical_loss": 3.496599992342054,
      "tokens_seen": 1598160896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026052156469408223,
      "loss": 2.5264,
      "theoretical_loss": 3.496587503647983,
      "tokens_seen": 1598226432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605115346038114,
      "loss": 2.6111,
      "theoretical_loss": 3.4965750156093893,
      "tokens_seen": 1598291968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026050150451354065,
      "loss": 2.5371,
      "theoretical_loss": 3.4965625282262107,
      "tokens_seen": 1598357504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026049147442326983,
      "loss": 2.6252,
      "theoretical_loss": 3.496550041498386,
      "tokens_seen": 1598423040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260481444332999,
      "loss": 2.8174,
      "theoretical_loss": 3.496537555425854,
      "tokens_seen": 1598488576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604714142427282,
      "loss": 2.7547,
      "theoretical_loss": 3.496525070008554,
      "tokens_seen": 1598554112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026046138415245743,
      "loss": 2.6282,
      "theoretical_loss": 3.4965125852464243,
      "tokens_seen": 1598619648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026045135406218656,
      "loss": 2.4625,
      "theoretical_loss": 3.4965001011394037,
      "tokens_seen": 1598685184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604413239719158,
      "loss": 2.4502,
      "theoretical_loss": 3.496487617687431,
      "tokens_seen": 1598750720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604312938816449,
      "loss": 2.6997,
      "theoretical_loss": 3.496475134890445,
      "tokens_seen": 1598816256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026042126379137415,
      "loss": 2.6514,
      "theoretical_loss": 3.496462652748384,
      "tokens_seen": 1598881792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026041123370110333,
      "loss": 2.3242,
      "theoretical_loss": 3.4964501712611873,
      "tokens_seen": 1598947328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604012036108325,
      "loss": 2.7017,
      "theoretical_loss": 3.4964376904287935,
      "tokens_seen": 1599012864
    },
    {
      "epoch": 0.48,
      "objective/train/docs_used": 883378,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9928767681121826,
      "objective/train/theoretical_loss": 3.496425210251142,
      "objective/train/tokens_used": 1619538400,
      "theoretical_loss": 3.496425210251142,
      "tokens_seen": 1599078400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603911735205617,
      "loss": 2.5989,
      "theoretical_loss": 3.496425210251142,
      "tokens_seen": 1599078400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603811434302909,
      "loss": 2.6122,
      "theoretical_loss": 3.496412730728171,
      "tokens_seen": 1599143936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026037111334002006,
      "loss": 2.5824,
      "theoretical_loss": 3.4964002518598196,
      "tokens_seen": 1599209472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603610832497493,
      "loss": 2.4081,
      "theoretical_loss": 3.4963877736460267,
      "tokens_seen": 1599275008
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603510531594784,
      "loss": 2.474,
      "theoretical_loss": 3.4963752960867307,
      "tokens_seen": 1599340544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026034102306920766,
      "loss": 2.5778,
      "theoretical_loss": 3.496362819181871,
      "tokens_seen": 1599406080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603309929789368,
      "loss": 2.6896,
      "theoretical_loss": 3.496350342931386,
      "tokens_seen": 1599471616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260320962888666,
      "loss": 2.5757,
      "theoretical_loss": 3.496337867335215,
      "tokens_seen": 1599537152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603109327983952,
      "loss": 2.7654,
      "theoretical_loss": 3.496325392393296,
      "tokens_seen": 1599602688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603009027081244,
      "loss": 2.4686,
      "theoretical_loss": 3.496312918105569,
      "tokens_seen": 1599668224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026029087261785356,
      "loss": 2.6196,
      "theoretical_loss": 3.4963004444719723,
      "tokens_seen": 1599733760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602808425275828,
      "loss": 2.4787,
      "theoretical_loss": 3.4962879714924453,
      "tokens_seen": 1599799296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602708124373119,
      "loss": 2.6034,
      "theoretical_loss": 3.4962754991669263,
      "tokens_seen": 1599864832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026026078234704116,
      "loss": 2.7047,
      "theoretical_loss": 3.4962630274953543,
      "tokens_seen": 1599930368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602507522567703,
      "loss": 2.7864,
      "theoretical_loss": 3.4962505564776682,
      "tokens_seen": 1599995904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602407221664995,
      "loss": 2.6546,
      "theoretical_loss": 3.4962380861138076,
      "tokens_seen": 1600061440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602306920762287,
      "loss": 2.5206,
      "theoretical_loss": 3.496225616403711,
      "tokens_seen": 1600126976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602206619859579,
      "loss": 2.5829,
      "theoretical_loss": 3.496213147347317,
      "tokens_seen": 1600192512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026021063189568707,
      "loss": 2.5875,
      "theoretical_loss": 3.496200678944565,
      "tokens_seen": 1600258048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026020060180541625,
      "loss": 2.6611,
      "theoretical_loss": 3.4961882111953937,
      "tokens_seen": 1600323584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026019057171514543,
      "loss": 2.6666,
      "theoretical_loss": 3.4961757440997427,
      "tokens_seen": 1600389120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026018054162487466,
      "loss": 2.5267,
      "theoretical_loss": 3.49616327765755,
      "tokens_seen": 1600454656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002601705115346038,
      "loss": 2.4489,
      "theoretical_loss": 3.4961508118687554,
      "tokens_seen": 1600520192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000260160481444333,
      "loss": 2.5302,
      "theoretical_loss": 3.4961383467332974,
      "tokens_seen": 1600585728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026015045135406215,
      "loss": 2.6217,
      "theoretical_loss": 3.4961258822511154,
      "tokens_seen": 1600651264
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 884433,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5312654972076416,
      "objective/train/theoretical_loss": 3.4961134184221483,
      "objective/train/tokens_used": 1621176800,
      "theoretical_loss": 3.4961134184221483,
      "tokens_seen": 1600716800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002601404212637914,
      "loss": 2.5328,
      "theoretical_loss": 3.4961134184221483,
      "tokens_seen": 1600716800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026013039117352057,
      "loss": 2.5875,
      "theoretical_loss": 3.4961009552463347,
      "tokens_seen": 1600782336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026012036108324975,
      "loss": 2.8462,
      "theoretical_loss": 3.496088492723614,
      "tokens_seen": 1600847872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026011033099297893,
      "loss": 2.5826,
      "theoretical_loss": 3.496076030853926,
      "tokens_seen": 1600913408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026010030090270817,
      "loss": 2.5677,
      "theoretical_loss": 3.496063569637208,
      "tokens_seen": 1600978944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600902708124373,
      "loss": 2.6,
      "theoretical_loss": 3.4960511090734006,
      "tokens_seen": 1601044480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026008024072216653,
      "loss": 2.6757,
      "theoretical_loss": 3.4960386491624424,
      "tokens_seen": 1601110016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026007021063189566,
      "loss": 2.6101,
      "theoretical_loss": 3.4960261899042724,
      "tokens_seen": 1601175552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600601805416249,
      "loss": 2.674,
      "theoretical_loss": 3.496013731298829,
      "tokens_seen": 1601241088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026005015045135407,
      "loss": 2.7099,
      "theoretical_loss": 3.496001273346053,
      "tokens_seen": 1601306624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026004012036108325,
      "loss": 2.3682,
      "theoretical_loss": 3.4959888160458816,
      "tokens_seen": 1601372160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026003009027081243,
      "loss": 2.6105,
      "theoretical_loss": 3.495976359398255,
      "tokens_seen": 1601437696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600200601805416,
      "loss": 2.49,
      "theoretical_loss": 3.495963903403112,
      "tokens_seen": 1601503232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600100300902708,
      "loss": 2.3433,
      "theoretical_loss": 3.495951448060392,
      "tokens_seen": 1601568768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026000000000000003,
      "loss": 2.697,
      "theoretical_loss": 3.495938993370034,
      "tokens_seen": 1601634304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025998996990972916,
      "loss": 2.6611,
      "theoretical_loss": 3.495926539331977,
      "tokens_seen": 1601699840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599799398194584,
      "loss": 2.6991,
      "theoretical_loss": 3.4959140859461604,
      "tokens_seen": 1601765376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599699097291876,
      "loss": 2.5537,
      "theoretical_loss": 3.4959016332125232,
      "tokens_seen": 1601830912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025995987963891676,
      "loss": 2.6926,
      "theoretical_loss": 3.4958891811310044,
      "tokens_seen": 1601896448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025994984954864594,
      "loss": 2.6165,
      "theoretical_loss": 3.495876729701543,
      "tokens_seen": 1601961984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599398194583751,
      "loss": 2.6984,
      "theoretical_loss": 3.495864278924079,
      "tokens_seen": 1602027520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599297893681043,
      "loss": 2.4181,
      "theoretical_loss": 3.4958518287985507,
      "tokens_seen": 1602093056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025991975927783353,
      "loss": 2.6084,
      "theoretical_loss": 3.495839379324898,
      "tokens_seen": 1602158592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025990972918756266,
      "loss": 2.8353,
      "theoretical_loss": 3.4958269305030596,
      "tokens_seen": 1602224128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598996990972919,
      "loss": 2.5644,
      "theoretical_loss": 3.4958144823329755,
      "tokens_seen": 1602289664
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 884978,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.275560140609741,
      "objective/train/theoretical_loss": 3.4958020348145835,
      "objective/train/tokens_used": 1622815200,
      "theoretical_loss": 3.4958020348145835,
      "tokens_seen": 1602355200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259889669007021,
      "loss": 2.4181,
      "theoretical_loss": 3.4958020348145835,
      "tokens_seen": 1602355200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025987963891675026,
      "loss": 2.707,
      "theoretical_loss": 3.495789587947824,
      "tokens_seen": 1602420736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025986960882647944,
      "loss": 2.4409,
      "theoretical_loss": 3.4957771417326358,
      "tokens_seen": 1602486272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598595787362086,
      "loss": 2.5454,
      "theoretical_loss": 3.4957646961689584,
      "tokens_seen": 1602551808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598495486459378,
      "loss": 2.5064,
      "theoretical_loss": 3.4957522512567314,
      "tokens_seen": 1602617344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259839518555667,
      "loss": 2.8227,
      "theoretical_loss": 3.4957398069958927,
      "tokens_seen": 1602682880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025982948846539616,
      "loss": 2.6371,
      "theoretical_loss": 3.495727363386383,
      "tokens_seen": 1602748416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598194583751254,
      "loss": 2.4321,
      "theoretical_loss": 3.495714920428141,
      "tokens_seen": 1602813952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598094282848545,
      "loss": 2.4335,
      "theoretical_loss": 3.495702478121106,
      "tokens_seen": 1602879488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025979939819458376,
      "loss": 2.7261,
      "theoretical_loss": 3.4956900364652164,
      "tokens_seen": 1602945024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025978936810431294,
      "loss": 2.6027,
      "theoretical_loss": 3.4956775954604136,
      "tokens_seen": 1603010560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597793380140421,
      "loss": 2.4884,
      "theoretical_loss": 3.495665155106635,
      "tokens_seen": 1603076096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597693079237713,
      "loss": 2.7455,
      "theoretical_loss": 3.4956527154038213,
      "tokens_seen": 1603141632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597592778335005,
      "loss": 2.6733,
      "theoretical_loss": 3.4956402763519105,
      "tokens_seen": 1603207168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597492477432297,
      "loss": 2.5342,
      "theoretical_loss": 3.4956278379508428,
      "tokens_seen": 1603272704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597392176529589,
      "loss": 2.6573,
      "theoretical_loss": 3.4956154002005575,
      "tokens_seen": 1603338240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597291875626881,
      "loss": 2.5731,
      "theoretical_loss": 3.4956029631009935,
      "tokens_seen": 1603403776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025971915747241727,
      "loss": 2.6231,
      "theoretical_loss": 3.495590526652091,
      "tokens_seen": 1603469312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025970912738214645,
      "loss": 2.6964,
      "theoretical_loss": 3.4955780908537886,
      "tokens_seen": 1603534848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025969909729187563,
      "loss": 2.6514,
      "theoretical_loss": 3.4955656557060255,
      "tokens_seen": 1603600384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025968906720160486,
      "loss": 2.5183,
      "theoretical_loss": 3.4955532212087417,
      "tokens_seen": 1603665920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259679037111334,
      "loss": 2.7161,
      "theoretical_loss": 3.4955407873618762,
      "tokens_seen": 1603731456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596690070210632,
      "loss": 2.5747,
      "theoretical_loss": 3.4955283541653692,
      "tokens_seen": 1603796992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025965897693079235,
      "loss": 2.3093,
      "theoretical_loss": 3.495515921619159,
      "tokens_seen": 1603862528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596489468405216,
      "loss": 2.5467,
      "theoretical_loss": 3.495503489723186,
      "tokens_seen": 1603928064
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 886048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6932458877563477,
      "objective/train/theoretical_loss": 3.4954910584773886,
      "objective/train/tokens_used": 1624453600,
      "theoretical_loss": 3.4954910584773886,
      "tokens_seen": 1603993600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025963891675025077,
      "loss": 2.5246,
      "theoretical_loss": 3.4954910584773886,
      "tokens_seen": 1603993600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025962888665997995,
      "loss": 2.4522,
      "theoretical_loss": 3.495478627881707,
      "tokens_seen": 1604059136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025961885656970913,
      "loss": 2.7995,
      "theoretical_loss": 3.4954661979360804,
      "tokens_seen": 1604124672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025960882647943837,
      "loss": 2.6769,
      "theoretical_loss": 3.4954537686404485,
      "tokens_seen": 1604190208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595987963891675,
      "loss": 2.7504,
      "theoretical_loss": 3.49544133999475,
      "tokens_seen": 1604255744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025958876629889673,
      "loss": 2.5586,
      "theoretical_loss": 3.4954289119989252,
      "tokens_seen": 1604321280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025957873620862586,
      "loss": 2.6534,
      "theoretical_loss": 3.495416484652913,
      "tokens_seen": 1604386816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595687061183551,
      "loss": 2.5472,
      "theoretical_loss": 3.4954040579566534,
      "tokens_seen": 1604452352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025955867602808427,
      "loss": 2.5427,
      "theoretical_loss": 3.4953916319100857,
      "tokens_seen": 1604517888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025954864593781345,
      "loss": 2.5922,
      "theoretical_loss": 3.4953792065131495,
      "tokens_seen": 1604583424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025953861584754263,
      "loss": 2.7647,
      "theoretical_loss": 3.4953667817657834,
      "tokens_seen": 1604648960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595285857572718,
      "loss": 2.6222,
      "theoretical_loss": 3.495354357667928,
      "tokens_seen": 1604714496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259518555667001,
      "loss": 2.6812,
      "theoretical_loss": 3.495341934219523,
      "tokens_seen": 1604780032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025950852557673023,
      "loss": 2.5888,
      "theoretical_loss": 3.4953295114205067,
      "tokens_seen": 1604845568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025949849548645936,
      "loss": 2.6561,
      "theoretical_loss": 3.4953170892708196,
      "tokens_seen": 1604911104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594884653961886,
      "loss": 2.521,
      "theoretical_loss": 3.495304667770401,
      "tokens_seen": 1604976640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594784353059178,
      "loss": 2.6409,
      "theoretical_loss": 3.4952922469191905,
      "tokens_seen": 1605042176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025946840521564696,
      "loss": 2.6746,
      "theoretical_loss": 3.4952798267171277,
      "tokens_seen": 1605107712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025945837512537614,
      "loss": 2.6635,
      "theoretical_loss": 3.495267407164152,
      "tokens_seen": 1605173248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594483450351053,
      "loss": 2.5861,
      "theoretical_loss": 3.4952549882602026,
      "tokens_seen": 1605238784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594383149448345,
      "loss": 2.7291,
      "theoretical_loss": 3.49524257000522,
      "tokens_seen": 1605304320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025942828485456373,
      "loss": 2.6775,
      "theoretical_loss": 3.4952301523991434,
      "tokens_seen": 1605369856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025941825476429286,
      "loss": 2.6408,
      "theoretical_loss": 3.495217735441912,
      "tokens_seen": 1605435392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594082246740221,
      "loss": 2.4916,
      "theoretical_loss": 3.495205319133466,
      "tokens_seen": 1605500928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593981945837512,
      "loss": 2.5286,
      "theoretical_loss": 3.495192903473745,
      "tokens_seen": 1605566464
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 887036,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7378838062286377,
      "objective/train/theoretical_loss": 3.4951804884626885,
      "objective/train/tokens_used": 1626092000,
      "theoretical_loss": 3.4951804884626885,
      "tokens_seen": 1605632000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025938816449348046,
      "loss": 2.5401,
      "theoretical_loss": 3.4951804884626885,
      "tokens_seen": 1605632000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025937813440320964,
      "loss": 2.8037,
      "theoretical_loss": 3.4951680741002358,
      "tokens_seen": 1605697536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593681043129388,
      "loss": 2.4051,
      "theoretical_loss": 3.4951556603863265,
      "tokens_seen": 1605763072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259358074222668,
      "loss": 2.6145,
      "theoretical_loss": 3.4951432473209003,
      "tokens_seen": 1605828608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593480441323972,
      "loss": 2.5647,
      "theoretical_loss": 3.4951308349038976,
      "tokens_seen": 1605894144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025933801404212636,
      "loss": 2.636,
      "theoretical_loss": 3.4951184231352577,
      "tokens_seen": 1605959680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593279839518556,
      "loss": 2.3618,
      "theoretical_loss": 3.49510601201492,
      "tokens_seen": 1606025216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593179538615847,
      "loss": 2.5,
      "theoretical_loss": 3.4950936015428242,
      "tokens_seen": 1606090752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025930792377131396,
      "loss": 2.5412,
      "theoretical_loss": 3.49508119171891,
      "tokens_seen": 1606156288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025929789368104314,
      "loss": 2.4135,
      "theoretical_loss": 3.4950687825431177,
      "tokens_seen": 1606221824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592878635907723,
      "loss": 2.5794,
      "theoretical_loss": 3.4950563740153866,
      "tokens_seen": 1606287360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592778335005015,
      "loss": 2.6096,
      "theoretical_loss": 3.495043966135656,
      "tokens_seen": 1606352896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592678034102307,
      "loss": 2.6201,
      "theoretical_loss": 3.4950315589038654,
      "tokens_seen": 1606418432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025925777331995987,
      "loss": 2.5927,
      "theoretical_loss": 3.495019152319956,
      "tokens_seen": 1606483968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592477432296891,
      "loss": 2.7788,
      "theoretical_loss": 3.4950067463838663,
      "tokens_seen": 1606549504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025923771313941823,
      "loss": 2.6531,
      "theoretical_loss": 3.4949943410955364,
      "tokens_seen": 1606615040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025922768304914747,
      "loss": 2.4365,
      "theoretical_loss": 3.494981936454906,
      "tokens_seen": 1606680576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592176529588766,
      "loss": 2.5671,
      "theoretical_loss": 3.494969532461915,
      "tokens_seen": 1606746112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025920762286860583,
      "loss": 2.5943,
      "theoretical_loss": 3.494957129116503,
      "tokens_seen": 1606811648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259197592778335,
      "loss": 2.4938,
      "theoretical_loss": 3.49494472641861,
      "tokens_seen": 1606877184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591875626880642,
      "loss": 2.6406,
      "theoretical_loss": 3.494932324368176,
      "tokens_seen": 1606942720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025917753259779337,
      "loss": 2.7251,
      "theoretical_loss": 3.49491992296514,
      "tokens_seen": 1607008256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025916750250752255,
      "loss": 2.5622,
      "theoretical_loss": 3.4949075222094423,
      "tokens_seen": 1607073792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025915747241725173,
      "loss": 2.62,
      "theoretical_loss": 3.4948951221010227,
      "tokens_seen": 1607139328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025914744232698097,
      "loss": 2.6759,
      "theoretical_loss": 3.494882722639821,
      "tokens_seen": 1607204864
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 887423,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.676961898803711,
      "objective/train/theoretical_loss": 3.494870323825777,
      "objective/train/tokens_used": 1627730400,
      "theoretical_loss": 3.494870323825777,
      "tokens_seen": 1607270400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591374122367101,
      "loss": 2.7373,
      "theoretical_loss": 3.494870323825777,
      "tokens_seen": 1607270400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025912738214643933,
      "loss": 2.7514,
      "theoretical_loss": 3.494857925658831,
      "tokens_seen": 1607335936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591173520561685,
      "loss": 2.8276,
      "theoretical_loss": 3.4948455281389217,
      "tokens_seen": 1607401472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591073219658977,
      "loss": 2.5974,
      "theoretical_loss": 3.49483313126599,
      "tokens_seen": 1607467008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590972918756269,
      "loss": 2.5221,
      "theoretical_loss": 3.4948207350399754,
      "tokens_seen": 1607532544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025908726178535606,
      "loss": 2.4486,
      "theoretical_loss": 3.494808339460818,
      "tokens_seen": 1607598080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025907723169508524,
      "loss": 2.3917,
      "theoretical_loss": 3.494795944528457,
      "tokens_seen": 1607663616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025906720160481447,
      "loss": 2.606,
      "theoretical_loss": 3.4947835502428335,
      "tokens_seen": 1607729152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590571715145436,
      "loss": 2.6469,
      "theoretical_loss": 3.4947711566038864,
      "tokens_seen": 1607794688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025904714142427283,
      "loss": 2.762,
      "theoretical_loss": 3.4947587636115554,
      "tokens_seen": 1607860224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025903711133400196,
      "loss": 2.7188,
      "theoretical_loss": 3.494746371265781,
      "tokens_seen": 1607925760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590270812437312,
      "loss": 2.6411,
      "theoretical_loss": 3.494733979566503,
      "tokens_seen": 1607991296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590170511534604,
      "loss": 2.3122,
      "theoretical_loss": 3.494721588513662,
      "tokens_seen": 1608056832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025900702106318956,
      "loss": 2.571,
      "theoretical_loss": 3.4947091981071963,
      "tokens_seen": 1608122368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589969909729188,
      "loss": 2.7087,
      "theoretical_loss": 3.4946968083470473,
      "tokens_seen": 1608187904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258986960882648,
      "loss": 2.4094,
      "theoretical_loss": 3.4946844192331543,
      "tokens_seen": 1608253440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025897693079237716,
      "loss": 2.7878,
      "theoretical_loss": 3.4946720307654573,
      "tokens_seen": 1608318976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025896690070210634,
      "loss": 2.6973,
      "theoretical_loss": 3.4946596429438967,
      "tokens_seen": 1608384512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589568706118355,
      "loss": 2.8017,
      "theoretical_loss": 3.494647255768412,
      "tokens_seen": 1608450048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589468405215647,
      "loss": 2.8646,
      "theoretical_loss": 3.494634869238943,
      "tokens_seen": 1608515584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025893681043129393,
      "loss": 2.4821,
      "theoretical_loss": 3.4946224833554305,
      "tokens_seen": 1608581120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025892678034102306,
      "loss": 2.5314,
      "theoretical_loss": 3.4946100981178136,
      "tokens_seen": 1608646656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589167502507523,
      "loss": 2.5069,
      "theoretical_loss": 3.4945977135260327,
      "tokens_seen": 1608712192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589067201604814,
      "loss": 2.7645,
      "theoretical_loss": 3.494585329580028,
      "tokens_seen": 1608777728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025889669007021066,
      "loss": 2.6254,
      "theoretical_loss": 3.4945729462797397,
      "tokens_seen": 1608843264
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 888663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4536147117614746,
      "objective/train/theoretical_loss": 3.494560563625107,
      "objective/train/tokens_used": 1629368800,
      "theoretical_loss": 3.494560563625107,
      "tokens_seen": 1608908800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025888665997993984,
      "loss": 2.7273,
      "theoretical_loss": 3.494560563625107,
      "tokens_seen": 1608908800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258876629889669,
      "loss": 2.778,
      "theoretical_loss": 3.4945481816160706,
      "tokens_seen": 1608974336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588665997993982,
      "loss": 2.7873,
      "theoretical_loss": 3.4945358002525704,
      "tokens_seen": 1609039872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588565697091274,
      "loss": 2.6672,
      "theoretical_loss": 3.4945234195345463,
      "tokens_seen": 1609105408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025884653961885657,
      "loss": 2.6069,
      "theoretical_loss": 3.4945110394619388,
      "tokens_seen": 1609170944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588365095285858,
      "loss": 2.408,
      "theoretical_loss": 3.494498660034687,
      "tokens_seen": 1609236480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025882647943831493,
      "loss": 2.4127,
      "theoretical_loss": 3.4944862812527324,
      "tokens_seen": 1609302016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025881644934804416,
      "loss": 2.5735,
      "theoretical_loss": 3.4944739031160137,
      "tokens_seen": 1609367552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025880641925777334,
      "loss": 2.5958,
      "theoretical_loss": 3.4944615256244718,
      "tokens_seen": 1609433088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587963891675025,
      "loss": 2.7475,
      "theoretical_loss": 3.4944491487780467,
      "tokens_seen": 1609498624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587863590772317,
      "loss": 2.6384,
      "theoretical_loss": 3.4944367725766785,
      "tokens_seen": 1609564160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587763289869609,
      "loss": 2.5391,
      "theoretical_loss": 3.494424397020307,
      "tokens_seen": 1609629696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025876629889669007,
      "loss": 2.5014,
      "theoretical_loss": 3.4944120221088726,
      "tokens_seen": 1609695232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587562688064193,
      "loss": 2.5073,
      "theoretical_loss": 3.4943996478423154,
      "tokens_seen": 1609760768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025874623871614843,
      "loss": 2.7434,
      "theoretical_loss": 3.4943872742205757,
      "tokens_seen": 1609826304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025873620862587767,
      "loss": 2.4657,
      "theoretical_loss": 3.4943749012435936,
      "tokens_seen": 1609891840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587261785356068,
      "loss": 2.7081,
      "theoretical_loss": 3.494362528911309,
      "tokens_seen": 1609957376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025871614844533603,
      "loss": 2.7731,
      "theoretical_loss": 3.4943501572236624,
      "tokens_seen": 1610022912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587061183550652,
      "loss": 2.6584,
      "theoretical_loss": 3.4943377861805933,
      "tokens_seen": 1610088448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586960882647944,
      "loss": 2.63,
      "theoretical_loss": 3.494325415782043,
      "tokens_seen": 1610153984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025868605817452357,
      "loss": 2.4182,
      "theoretical_loss": 3.494313046027951,
      "tokens_seen": 1610219520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025867602808425275,
      "loss": 2.5206,
      "theoretical_loss": 3.4943006769182574,
      "tokens_seen": 1610285056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025866599799398193,
      "loss": 2.5546,
      "theoretical_loss": 3.4942883084529024,
      "tokens_seen": 1610350592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025865596790371117,
      "loss": 2.5437,
      "theoretical_loss": 3.4942759406318262,
      "tokens_seen": 1610416128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586459378134403,
      "loss": 2.6966,
      "theoretical_loss": 3.49426357345497,
      "tokens_seen": 1610481664
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 889165,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2135186195373535,
      "objective/train/theoretical_loss": 3.4942512069222724,
      "objective/train/tokens_used": 1631007200,
      "theoretical_loss": 3.4942512069222724,
      "tokens_seen": 1610547200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025863590772316953,
      "loss": 2.4609,
      "theoretical_loss": 3.4942512069222724,
      "tokens_seen": 1610547200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586258776328987,
      "loss": 2.51,
      "theoretical_loss": 3.494238841033675,
      "tokens_seen": 1610612736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586158475426279,
      "loss": 2.4031,
      "theoretical_loss": 3.4942264757891173,
      "tokens_seen": 1610678272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586058174523571,
      "loss": 2.9833,
      "theoretical_loss": 3.4942141111885396,
      "tokens_seen": 1610743808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025859578736208626,
      "loss": 2.598,
      "theoretical_loss": 3.494201747231883,
      "tokens_seen": 1610809344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025858575727181544,
      "loss": 2.836,
      "theoretical_loss": 3.4941893839190863,
      "tokens_seen": 1610874880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025857572718154467,
      "loss": 2.7011,
      "theoretical_loss": 3.494177021250091,
      "tokens_seen": 1610940416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585656970912738,
      "loss": 2.6018,
      "theoretical_loss": 3.494164659224837,
      "tokens_seen": 1611005952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025855566700100303,
      "loss": 2.5364,
      "theoretical_loss": 3.4941522978432644,
      "tokens_seen": 1611071488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025854563691073216,
      "loss": 2.8671,
      "theoretical_loss": 3.4941399371053143,
      "tokens_seen": 1611137024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585356068204614,
      "loss": 2.569,
      "theoretical_loss": 3.494127577010926,
      "tokens_seen": 1611202560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585255767301906,
      "loss": 2.7617,
      "theoretical_loss": 3.49411521756004,
      "tokens_seen": 1611268096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025851554663991976,
      "loss": 2.3829,
      "theoretical_loss": 3.494102858752597,
      "tokens_seen": 1611333632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025850551654964894,
      "loss": 2.6156,
      "theoretical_loss": 3.4940905005885368,
      "tokens_seen": 1611399168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584954864593782,
      "loss": 2.8408,
      "theoretical_loss": 3.4940781430678003,
      "tokens_seen": 1611464704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584854563691073,
      "loss": 2.6587,
      "theoretical_loss": 3.494065786190328,
      "tokens_seen": 1611530240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025847542627883654,
      "loss": 2.7119,
      "theoretical_loss": 3.4940534299560597,
      "tokens_seen": 1611595776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025846539618856566,
      "loss": 2.5316,
      "theoretical_loss": 3.4940410743649357,
      "tokens_seen": 1611661312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584553660982949,
      "loss": 2.4298,
      "theoretical_loss": 3.494028719416897,
      "tokens_seen": 1611726848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584453360080241,
      "loss": 2.6352,
      "theoretical_loss": 3.4940163651118836,
      "tokens_seen": 1611792384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025843530591775326,
      "loss": 2.6358,
      "theoretical_loss": 3.494004011449836,
      "tokens_seen": 1611857920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025842527582748244,
      "loss": 2.6609,
      "theoretical_loss": 3.4939916584306943,
      "tokens_seen": 1611923456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584152457372116,
      "loss": 2.6514,
      "theoretical_loss": 3.493979306054399,
      "tokens_seen": 1611988992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584052156469408,
      "loss": 2.7571,
      "theoretical_loss": 3.493966954320891,
      "tokens_seen": 1612054528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025839518555667004,
      "loss": 2.5126,
      "theoretical_loss": 3.49395460323011,
      "tokens_seen": 1612120064
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 890307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.018338203430176,
      "objective/train/theoretical_loss": 3.4939422527819968,
      "objective/train/tokens_used": 1632645600,
      "theoretical_loss": 3.4939422527819968,
      "tokens_seen": 1612185600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025838515546639917,
      "loss": 2.5821,
      "theoretical_loss": 3.4939422527819968,
      "tokens_seen": 1612185600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583751253761284,
      "loss": 2.5408,
      "theoretical_loss": 3.4939299029764923,
      "tokens_seen": 1612251136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025836509528585753,
      "loss": 2.4805,
      "theoretical_loss": 3.493917553813536,
      "tokens_seen": 1612316672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025835506519558677,
      "loss": 2.6354,
      "theoretical_loss": 3.493905205293069,
      "tokens_seen": 1612382208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025834503510531595,
      "loss": 2.7238,
      "theoretical_loss": 3.4938928574150316,
      "tokens_seen": 1612447744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025833500501504513,
      "loss": 2.7054,
      "theoretical_loss": 3.4938805101793644,
      "tokens_seen": 1612513280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583249749247743,
      "loss": 2.5066,
      "theoretical_loss": 3.4938681635860074,
      "tokens_seen": 1612578816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025831494483450354,
      "loss": 2.7947,
      "theoretical_loss": 3.4938558176349015,
      "tokens_seen": 1612644352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025830491474423267,
      "loss": 2.5333,
      "theoretical_loss": 3.4938434723259872,
      "tokens_seen": 1612709888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582948846539619,
      "loss": 2.6675,
      "theoretical_loss": 3.4938311276592047,
      "tokens_seen": 1612775424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025828485456369103,
      "loss": 2.6112,
      "theoretical_loss": 3.493818783634495,
      "tokens_seen": 1612840960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025827482447342027,
      "loss": 2.6136,
      "theoretical_loss": 3.493806440251798,
      "tokens_seen": 1612906496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025826479438314945,
      "loss": 2.6806,
      "theoretical_loss": 3.4937940975110546,
      "tokens_seen": 1612972032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025825476429287863,
      "loss": 2.7465,
      "theoretical_loss": 3.493781755412206,
      "tokens_seen": 1613037568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025824473420260787,
      "loss": 2.7219,
      "theoretical_loss": 3.4937694139551914,
      "tokens_seen": 1613103104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258234704112337,
      "loss": 2.5369,
      "theoretical_loss": 3.4937570731399523,
      "tokens_seen": 1613168640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025822467402206623,
      "loss": 2.7823,
      "theoretical_loss": 3.4937447329664284,
      "tokens_seen": 1613234176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582146439317954,
      "loss": 2.4617,
      "theoretical_loss": 3.493732393434561,
      "tokens_seen": 1613299712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582046138415246,
      "loss": 2.6741,
      "theoretical_loss": 3.4937200545442906,
      "tokens_seen": 1613365248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025819458375125377,
      "loss": 2.8025,
      "theoretical_loss": 3.4937077162955577,
      "tokens_seen": 1613430784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025818455366098295,
      "loss": 2.7791,
      "theoretical_loss": 3.4936953786883027,
      "tokens_seen": 1613496320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025817452357071213,
      "loss": 2.6624,
      "theoretical_loss": 3.4936830417224662,
      "tokens_seen": 1613561856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025816449348044137,
      "loss": 2.5387,
      "theoretical_loss": 3.4936707053979896,
      "tokens_seen": 1613627392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581544633901705,
      "loss": 2.5352,
      "theoretical_loss": 3.4936583697148125,
      "tokens_seen": 1613692928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025814443329989973,
      "loss": 2.6936,
      "theoretical_loss": 3.4936460346728753,
      "tokens_seen": 1613758464
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 891421,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5894956588745117,
      "objective/train/theoretical_loss": 3.49363370027212,
      "objective/train/tokens_used": 1634284000,
      "theoretical_loss": 3.49363370027212,
      "tokens_seen": 1613824000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581344032096289,
      "loss": 2.5598,
      "theoretical_loss": 3.49363370027212,
      "tokens_seen": 1613824000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581243731193581,
      "loss": 2.6864,
      "theoretical_loss": 3.493621366512486,
      "tokens_seen": 1613889536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581143430290873,
      "loss": 2.7065,
      "theoretical_loss": 3.4936090333939145,
      "tokens_seen": 1613955072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025810431293881646,
      "loss": 2.512,
      "theoretical_loss": 3.493596700916346,
      "tokens_seen": 1614020608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025809428284854564,
      "loss": 2.579,
      "theoretical_loss": 3.493584369079721,
      "tokens_seen": 1614086144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025808425275827487,
      "loss": 2.4163,
      "theoretical_loss": 3.4935720378839807,
      "tokens_seen": 1614151680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258074222668004,
      "loss": 2.4502,
      "theoretical_loss": 3.493559707329065,
      "tokens_seen": 1614217216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025806419257773323,
      "loss": 2.7627,
      "theoretical_loss": 3.4935473774149157,
      "tokens_seen": 1614282752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025805416248746236,
      "loss": 2.5223,
      "theoretical_loss": 3.4935350481414718,
      "tokens_seen": 1614348288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580441323971916,
      "loss": 2.5368,
      "theoretical_loss": 3.4935227195086758,
      "tokens_seen": 1614413824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580341023069208,
      "loss": 2.837,
      "theoretical_loss": 3.4935103915164674,
      "tokens_seen": 1614479360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025802407221664996,
      "loss": 2.733,
      "theoretical_loss": 3.4934980641647875,
      "tokens_seen": 1614544896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025801404212637914,
      "loss": 2.6878,
      "theoretical_loss": 3.493485737453577,
      "tokens_seen": 1614610432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580040120361084,
      "loss": 2.6848,
      "theoretical_loss": 3.493473411382776,
      "tokens_seen": 1614675968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579939819458375,
      "loss": 2.3273,
      "theoretical_loss": 3.493461085952326,
      "tokens_seen": 1614741504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025798395185556674,
      "loss": 2.4823,
      "theoretical_loss": 3.493448761162168,
      "tokens_seen": 1614807040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025797392176529586,
      "loss": 2.5212,
      "theoretical_loss": 3.4934364370122415,
      "tokens_seen": 1614872576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579638916750251,
      "loss": 2.8075,
      "theoretical_loss": 3.493424113502488,
      "tokens_seen": 1614938112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579538615847543,
      "loss": 2.8885,
      "theoretical_loss": 3.493411790632848,
      "tokens_seen": 1615003648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025794383149448346,
      "loss": 2.8604,
      "theoretical_loss": 3.493399468403263,
      "tokens_seen": 1615069184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025793380140421264,
      "loss": 2.6701,
      "theoretical_loss": 3.4933871468136735,
      "tokens_seen": 1615134720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579237713139418,
      "loss": 2.5409,
      "theoretical_loss": 3.49337482586402,
      "tokens_seen": 1615200256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257913741223671,
      "loss": 2.4081,
      "theoretical_loss": 3.493362505554243,
      "tokens_seen": 1615265792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025790371113340024,
      "loss": 2.7342,
      "theoretical_loss": 3.4933501858842844,
      "tokens_seen": 1615331328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025789368104312937,
      "loss": 2.7004,
      "theoretical_loss": 3.4933378668540835,
      "tokens_seen": 1615396864
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 892490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.780491828918457,
      "objective/train/theoretical_loss": 3.4933255484635826,
      "objective/train/tokens_used": 1635922400,
      "theoretical_loss": 3.4933255484635826,
      "tokens_seen": 1615462400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578836509528586,
      "loss": 2.6395,
      "theoretical_loss": 3.4933255484635826,
      "tokens_seen": 1615462400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025787362086258773,
      "loss": 2.589,
      "theoretical_loss": 3.4933132307127215,
      "tokens_seen": 1615527936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025786359077231697,
      "loss": 2.9653,
      "theoretical_loss": 3.4933009136014412,
      "tokens_seen": 1615593472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025785356068204615,
      "loss": 2.6433,
      "theoretical_loss": 3.493288597129683,
      "tokens_seen": 1615659008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025784353059177533,
      "loss": 2.7118,
      "theoretical_loss": 3.493276281297388,
      "tokens_seen": 1615724544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578335005015045,
      "loss": 2.8061,
      "theoretical_loss": 3.4932639661044957,
      "tokens_seen": 1615790080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025782347041123374,
      "loss": 2.6577,
      "theoretical_loss": 3.4932516515509486,
      "tokens_seen": 1615855616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025781344032096287,
      "loss": 2.7861,
      "theoretical_loss": 3.4932393376366866,
      "tokens_seen": 1615921152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578034102306921,
      "loss": 2.6685,
      "theoretical_loss": 3.4932270243616506,
      "tokens_seen": 1615986688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025779338014042123,
      "loss": 2.7509,
      "theoretical_loss": 3.4932147117257824,
      "tokens_seen": 1616052224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025778335005015047,
      "loss": 2.4067,
      "theoretical_loss": 3.4932023997290216,
      "tokens_seen": 1616117760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025777331995987965,
      "loss": 2.5045,
      "theoretical_loss": 3.4931900883713096,
      "tokens_seen": 1616183296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025776328986960883,
      "loss": 2.6771,
      "theoretical_loss": 3.493177777652588,
      "tokens_seen": 1616248832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257753259779338,
      "loss": 2.6147,
      "theoretical_loss": 3.493165467572797,
      "tokens_seen": 1616314368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577432296890672,
      "loss": 2.6992,
      "theoretical_loss": 3.4931531581318778,
      "tokens_seen": 1616379904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577331995987964,
      "loss": 2.7035,
      "theoretical_loss": 3.4931408493297713,
      "tokens_seen": 1616445440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577231695085256,
      "loss": 2.7856,
      "theoretical_loss": 3.493128541166418,
      "tokens_seen": 1616510976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025771313941825474,
      "loss": 2.6062,
      "theoretical_loss": 3.4931162336417594,
      "tokens_seen": 1616576512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025770310932798397,
      "loss": 2.6779,
      "theoretical_loss": 3.4931039267557367,
      "tokens_seen": 1616642048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576930792377131,
      "loss": 2.7855,
      "theoretical_loss": 3.4930916205082903,
      "tokens_seen": 1616707584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025768304914744233,
      "loss": 2.7018,
      "theoretical_loss": 3.493079314899361,
      "tokens_seen": 1616773120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576730190571715,
      "loss": 2.5119,
      "theoretical_loss": 3.4930670099288905,
      "tokens_seen": 1616838656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576629889669007,
      "loss": 2.6613,
      "theoretical_loss": 3.49305470559682,
      "tokens_seen": 1616904192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576529588766299,
      "loss": 2.8252,
      "theoretical_loss": 3.4930424019030895,
      "tokens_seen": 1616969728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576429287863591,
      "loss": 2.7648,
      "theoretical_loss": 3.4930300988476404,
      "tokens_seen": 1617035264
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 893217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8498308658599854,
      "objective/train/theoretical_loss": 3.493017796430414,
      "objective/train/tokens_used": 1637560800,
      "theoretical_loss": 3.493017796430414,
      "tokens_seen": 1617100800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025763289869608824,
      "loss": 2.6071,
      "theoretical_loss": 3.493017796430414,
      "tokens_seen": 1617100800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576228686058175,
      "loss": 2.7201,
      "theoretical_loss": 3.493005494651351,
      "tokens_seen": 1617166336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576128385155466,
      "loss": 2.7744,
      "theoretical_loss": 3.492993193510393,
      "tokens_seen": 1617231872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025760280842527584,
      "loss": 2.7385,
      "theoretical_loss": 3.49298089300748,
      "tokens_seen": 1617297408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257592778335005,
      "loss": 2.7485,
      "theoretical_loss": 3.492968593142554,
      "tokens_seen": 1617362944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575827482447342,
      "loss": 2.6081,
      "theoretical_loss": 3.492956293915556,
      "tokens_seen": 1617428480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575727181544634,
      "loss": 2.6366,
      "theoretical_loss": 3.4929439953264265,
      "tokens_seen": 1617494016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025756268806419256,
      "loss": 2.5578,
      "theoretical_loss": 3.4929316973751066,
      "tokens_seen": 1617559552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025755265797392174,
      "loss": 2.6992,
      "theoretical_loss": 3.4929194000615382,
      "tokens_seen": 1617625088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257542627883651,
      "loss": 2.7894,
      "theoretical_loss": 3.4929071033856616,
      "tokens_seen": 1617690624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575325977933801,
      "loss": 2.6224,
      "theoretical_loss": 3.4928948073474184,
      "tokens_seen": 1617756160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025752256770310934,
      "loss": 2.8395,
      "theoretical_loss": 3.4928825119467497,
      "tokens_seen": 1617821696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025751253761283847,
      "loss": 2.6206,
      "theoretical_loss": 3.492870217183596,
      "tokens_seen": 1617887232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575025075225677,
      "loss": 2.71,
      "theoretical_loss": 3.4928579230578984,
      "tokens_seen": 1617952768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025749247743229694,
      "loss": 2.5746,
      "theoretical_loss": 3.492845629569599,
      "tokens_seen": 1618018304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025748244734202606,
      "loss": 2.5882,
      "theoretical_loss": 3.4928333367186384,
      "tokens_seen": 1618083840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574724172517553,
      "loss": 2.5893,
      "theoretical_loss": 3.4928210445049577,
      "tokens_seen": 1618149376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574623871614845,
      "loss": 2.8691,
      "theoretical_loss": 3.492808752928498,
      "tokens_seen": 1618214912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025745235707121366,
      "loss": 2.6229,
      "theoretical_loss": 3.4927964619892005,
      "tokens_seen": 1618280448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025744232698094284,
      "loss": 2.6035,
      "theoretical_loss": 3.4927841716870063,
      "tokens_seen": 1618345984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257432296890672,
      "loss": 2.7588,
      "theoretical_loss": 3.4927718820218567,
      "tokens_seen": 1618411520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574222668004012,
      "loss": 2.522,
      "theoretical_loss": 3.4927595929936928,
      "tokens_seen": 1618477056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025741223671013044,
      "loss": 2.5641,
      "theoretical_loss": 3.4927473046024566,
      "tokens_seen": 1618542592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025740220661985957,
      "loss": 2.5197,
      "theoretical_loss": 3.492735016848088,
      "tokens_seen": 1618608128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573921765295888,
      "loss": 2.4852,
      "theoretical_loss": 3.4927227297305286,
      "tokens_seen": 1618673664
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 894105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0400853157043457,
      "objective/train/theoretical_loss": 3.4927104432497202,
      "objective/train/tokens_used": 1639199200,
      "theoretical_loss": 3.4927104432497202,
      "tokens_seen": 1618739200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025738214643931793,
      "loss": 2.7416,
      "theoretical_loss": 3.4927104432497202,
      "tokens_seen": 1618739200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025737211634904717,
      "loss": 2.8181,
      "theoretical_loss": 3.492698157405603,
      "tokens_seen": 1618804736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025736208625877635,
      "loss": 2.7422,
      "theoretical_loss": 3.4926858721981198,
      "tokens_seen": 1618870272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025735205616850553,
      "loss": 2.5819,
      "theoretical_loss": 3.49267358762721,
      "tokens_seen": 1618935808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573420260782347,
      "loss": 2.78,
      "theoretical_loss": 3.492661303692816,
      "tokens_seen": 1619001344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025733199598796394,
      "loss": 2.7969,
      "theoretical_loss": 3.492649020394879,
      "tokens_seen": 1619066880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025732196589769307,
      "loss": 2.5117,
      "theoretical_loss": 3.4926367377333403,
      "tokens_seen": 1619132416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573119358074223,
      "loss": 2.7497,
      "theoretical_loss": 3.4926244557081403,
      "tokens_seen": 1619197952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025730190571715143,
      "loss": 2.3726,
      "theoretical_loss": 3.492612174319221,
      "tokens_seen": 1619263488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025729187562688067,
      "loss": 2.4969,
      "theoretical_loss": 3.492599893566524,
      "tokens_seen": 1619329024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025728184553660985,
      "loss": 2.7764,
      "theoretical_loss": 3.49258761344999,
      "tokens_seen": 1619394560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025727181544633903,
      "loss": 2.7566,
      "theoretical_loss": 3.4925753339695604,
      "tokens_seen": 1619460096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572617853560682,
      "loss": 2.5046,
      "theoretical_loss": 3.492563055125177,
      "tokens_seen": 1619525632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572517552657974,
      "loss": 2.8057,
      "theoretical_loss": 3.4925507769167803,
      "tokens_seen": 1619591168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572417251755266,
      "loss": 2.6301,
      "theoretical_loss": 3.4925384993443123,
      "tokens_seen": 1619656704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572316950852558,
      "loss": 2.5833,
      "theoretical_loss": 3.492526222407714,
      "tokens_seen": 1619722240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025722166499498494,
      "loss": 2.6346,
      "theoretical_loss": 3.492513946106927,
      "tokens_seen": 1619787776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025721163490471417,
      "loss": 2.6943,
      "theoretical_loss": 3.4925016704418916,
      "tokens_seen": 1619853312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572016048144433,
      "loss": 2.5891,
      "theoretical_loss": 3.4924893954125507,
      "tokens_seen": 1619918848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025719157472417253,
      "loss": 2.5932,
      "theoretical_loss": 3.492477121018845,
      "tokens_seen": 1619984384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571815446339017,
      "loss": 2.7712,
      "theoretical_loss": 3.492464847260716,
      "tokens_seen": 1620049920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571715145436309,
      "loss": 2.9281,
      "theoretical_loss": 3.4924525741381047,
      "tokens_seen": 1620115456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571614844533601,
      "loss": 2.7198,
      "theoretical_loss": 3.492440301650953,
      "tokens_seen": 1620180992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571514543630893,
      "loss": 2.5301,
      "theoretical_loss": 3.4924280297992016,
      "tokens_seen": 1620246528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025714142427281844,
      "loss": 2.5522,
      "theoretical_loss": 3.4924157585827924,
      "tokens_seen": 1620312064
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 894566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7740554809570312,
      "objective/train/theoretical_loss": 3.492403488001667,
      "objective/train/tokens_used": 1640837600,
      "theoretical_loss": 3.492403488001667,
      "tokens_seen": 1620377600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571313941825477,
      "loss": 2.6027,
      "theoretical_loss": 3.492403488001667,
      "tokens_seen": 1620377600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571213640922768,
      "loss": 2.4531,
      "theoretical_loss": 3.492391218055767,
      "tokens_seen": 1620443136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025711133400200604,
      "loss": 2.5117,
      "theoretical_loss": 3.4923789487450327,
      "tokens_seen": 1620508672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571013039117352,
      "loss": 2.5483,
      "theoretical_loss": 3.492366680069406,
      "tokens_seen": 1620574208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570912738214644,
      "loss": 2.5429,
      "theoretical_loss": 3.4923544120288295,
      "tokens_seen": 1620639744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570812437311936,
      "loss": 2.9547,
      "theoretical_loss": 3.492342144623243,
      "tokens_seen": 1620705280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025707121364092276,
      "loss": 2.407,
      "theoretical_loss": 3.492329877852589,
      "tokens_seen": 1620770816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025706118355065194,
      "loss": 2.6359,
      "theoretical_loss": 3.4923176117168087,
      "tokens_seen": 1620836352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570511534603812,
      "loss": 2.7424,
      "theoretical_loss": 3.492305346215843,
      "tokens_seen": 1620901888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570411233701103,
      "loss": 2.5256,
      "theoretical_loss": 3.4922930813496347,
      "tokens_seen": 1620967424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025703109327983954,
      "loss": 2.8147,
      "theoretical_loss": 3.492280817118124,
      "tokens_seen": 1621032960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025702106318956867,
      "loss": 2.7526,
      "theoretical_loss": 3.492268553521253,
      "tokens_seen": 1621098496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570110330992979,
      "loss": 2.453,
      "theoretical_loss": 3.492256290558963,
      "tokens_seen": 1621164032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570010030090271,
      "loss": 2.7507,
      "theoretical_loss": 3.4922440282311955,
      "tokens_seen": 1621229568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025699097291875626,
      "loss": 2.6752,
      "theoretical_loss": 3.492231766537892,
      "tokens_seen": 1621295104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025698094282848545,
      "loss": 2.7155,
      "theoretical_loss": 3.4922195054789946,
      "tokens_seen": 1621360640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569709127382147,
      "loss": 2.6698,
      "theoretical_loss": 3.492207245054444,
      "tokens_seen": 1621426176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569608826479438,
      "loss": 2.4975,
      "theoretical_loss": 3.4921949852641827,
      "tokens_seen": 1621491712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025695085255767304,
      "loss": 2.4095,
      "theoretical_loss": 3.492182726108151,
      "tokens_seen": 1621557248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025694082246740217,
      "loss": 2.6184,
      "theoretical_loss": 3.4921704675862912,
      "tokens_seen": 1621622784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569307923771314,
      "loss": 2.6408,
      "theoretical_loss": 3.4921582096985455,
      "tokens_seen": 1621688320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569207622868606,
      "loss": 2.4466,
      "theoretical_loss": 3.4921459524448544,
      "tokens_seen": 1621753856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025691073219658977,
      "loss": 2.6971,
      "theoretical_loss": 3.4921336958251596,
      "tokens_seen": 1621819392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025690070210631895,
      "loss": 2.5353,
      "theoretical_loss": 3.492121439839403,
      "tokens_seen": 1621884928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025689067201604813,
      "loss": 2.5513,
      "theoretical_loss": 3.4921091844875267,
      "tokens_seen": 1621950464
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 895662,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5183286666870117,
      "objective/train/theoretical_loss": 3.4920969297694713,
      "objective/train/tokens_used": 1642476000,
      "theoretical_loss": 3.4920969297694713,
      "tokens_seen": 1622016000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568806419257773,
      "loss": 2.7824,
      "theoretical_loss": 3.4920969297694713,
      "tokens_seen": 1622016000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025687061183550655,
      "loss": 2.6162,
      "theoretical_loss": 3.4920846756851787,
      "tokens_seen": 1622081536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568605817452357,
      "loss": 2.7633,
      "theoretical_loss": 3.4920724222345907,
      "tokens_seen": 1622147072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568505516549649,
      "loss": 2.6291,
      "theoretical_loss": 3.492060169417649,
      "tokens_seen": 1622212608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025684052156469404,
      "loss": 2.5131,
      "theoretical_loss": 3.492047917234296,
      "tokens_seen": 1622278144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025683049147442327,
      "loss": 2.6453,
      "theoretical_loss": 3.4920356656844715,
      "tokens_seen": 1622343680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025682046138415245,
      "loss": 2.597,
      "theoretical_loss": 3.492023414768118,
      "tokens_seen": 1622409216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025681043129388163,
      "loss": 2.6343,
      "theoretical_loss": 3.492011164485178,
      "tokens_seen": 1622474752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568004012036108,
      "loss": 2.5363,
      "theoretical_loss": 3.4919989148355928,
      "tokens_seen": 1622540288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025679037111334005,
      "loss": 2.5657,
      "theoretical_loss": 3.4919866658193026,
      "tokens_seen": 1622605824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567803410230692,
      "loss": 2.6233,
      "theoretical_loss": 3.491974417436251,
      "tokens_seen": 1622671360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567703109327984,
      "loss": 2.6616,
      "theoretical_loss": 3.4919621696863787,
      "tokens_seen": 1622736896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025676028084252754,
      "loss": 2.3667,
      "theoretical_loss": 3.4919499225696278,
      "tokens_seen": 1622802432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567502507522568,
      "loss": 2.6855,
      "theoretical_loss": 3.4919376760859397,
      "tokens_seen": 1622867968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256740220661986,
      "loss": 2.9121,
      "theoretical_loss": 3.4919254302352565,
      "tokens_seen": 1622933504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025673019057171514,
      "loss": 2.7713,
      "theoretical_loss": 3.491913185017519,
      "tokens_seen": 1622999040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025672016048144437,
      "loss": 2.4044,
      "theoretical_loss": 3.49190094043267,
      "tokens_seen": 1623064576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567101303911735,
      "loss": 2.543,
      "theoretical_loss": 3.4918886964806513,
      "tokens_seen": 1623130112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025670010030090273,
      "loss": 2.8006,
      "theoretical_loss": 3.4918764531614035,
      "tokens_seen": 1623195648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566900702106319,
      "loss": 2.7173,
      "theoretical_loss": 3.491864210474869,
      "tokens_seen": 1623261184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566800401203611,
      "loss": 2.717,
      "theoretical_loss": 3.49185196842099,
      "tokens_seen": 1623326720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566700100300903,
      "loss": 2.4206,
      "theoretical_loss": 3.4918397269997077,
      "tokens_seen": 1623392256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566599799398195,
      "loss": 2.7348,
      "theoretical_loss": 3.491827486210964,
      "tokens_seen": 1623457792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025664994984954864,
      "loss": 2.5949,
      "theoretical_loss": 3.491815246054701,
      "tokens_seen": 1623523328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566399197592779,
      "loss": 2.8221,
      "theoretical_loss": 3.49180300653086,
      "tokens_seen": 1623588864
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 896620,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.650221347808838,
      "objective/train/theoretical_loss": 3.4917907676393827,
      "objective/train/tokens_used": 1644114400,
      "theoretical_loss": 3.4917907676393827,
      "tokens_seen": 1623654400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256629889669007,
      "loss": 2.4727,
      "theoretical_loss": 3.4917907676393827,
      "tokens_seen": 1623654400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025661985957873624,
      "loss": 2.7314,
      "theoretical_loss": 3.4917785293802117,
      "tokens_seen": 1623719936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566098294884654,
      "loss": 2.6896,
      "theoretical_loss": 3.4917662917532875,
      "tokens_seen": 1623785472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565997993981946,
      "loss": 2.7967,
      "theoretical_loss": 3.4917540547585535,
      "tokens_seen": 1623851008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565897693079238,
      "loss": 2.4523,
      "theoretical_loss": 3.4917418183959508,
      "tokens_seen": 1623916544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025657973921765296,
      "loss": 2.4015,
      "theoretical_loss": 3.491729582665421,
      "tokens_seen": 1623982080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025656970912738214,
      "loss": 2.8635,
      "theoretical_loss": 3.4917173475669063,
      "tokens_seen": 1624047616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565596790371114,
      "loss": 2.638,
      "theoretical_loss": 3.4917051131003483,
      "tokens_seen": 1624113152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565496489468405,
      "loss": 2.6574,
      "theoretical_loss": 3.4916928792656887,
      "tokens_seen": 1624178688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025653961885656974,
      "loss": 2.6169,
      "theoretical_loss": 3.49168064606287,
      "tokens_seen": 1624244224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025652958876629887,
      "loss": 2.6505,
      "theoretical_loss": 3.4916684134918334,
      "tokens_seen": 1624309760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565195586760281,
      "loss": 2.8218,
      "theoretical_loss": 3.4916561815525213,
      "tokens_seen": 1624375296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565095285857573,
      "loss": 2.7438,
      "theoretical_loss": 3.4916439502448755,
      "tokens_seen": 1624440832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025649949849548647,
      "loss": 2.5903,
      "theoretical_loss": 3.4916317195688373,
      "tokens_seen": 1624506368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025648946840521565,
      "loss": 2.6318,
      "theoretical_loss": 3.49161948952435,
      "tokens_seen": 1624571904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564794383149449,
      "loss": 2.4337,
      "theoretical_loss": 3.491607260111354,
      "tokens_seen": 1624637440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256469408224674,
      "loss": 2.681,
      "theoretical_loss": 3.4915950313297914,
      "tokens_seen": 1624702976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025645937813440324,
      "loss": 2.821,
      "theoretical_loss": 3.4915828031796052,
      "tokens_seen": 1624768512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025644934804413237,
      "loss": 2.5816,
      "theoretical_loss": 3.4915705756607363,
      "tokens_seen": 1624834048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564393179538616,
      "loss": 2.5739,
      "theoretical_loss": 3.491558348773127,
      "tokens_seen": 1624899584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564292878635908,
      "loss": 2.5323,
      "theoretical_loss": 3.49154612251672,
      "tokens_seen": 1624965120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025641925777331997,
      "loss": 2.7479,
      "theoretical_loss": 3.4915338968914558,
      "tokens_seen": 1625030656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025640922768304915,
      "loss": 2.6312,
      "theoretical_loss": 3.491521671897278,
      "tokens_seen": 1625096192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025639919759277833,
      "loss": 2.6383,
      "theoretical_loss": 3.491509447534127,
      "tokens_seen": 1625161728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563891675025075,
      "loss": 2.5722,
      "theoretical_loss": 3.4914972238019453,
      "tokens_seen": 1625227264
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 896956,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6397101879119873,
      "objective/train/theoretical_loss": 3.491485000700675,
      "objective/train/tokens_used": 1645752800,
      "theoretical_loss": 3.491485000700675,
      "tokens_seen": 1625292800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025637913741223675,
      "loss": 2.5791,
      "theoretical_loss": 3.491485000700675,
      "tokens_seen": 1625292800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563691073219659,
      "loss": 2.6691,
      "theoretical_loss": 3.491472778230259,
      "tokens_seen": 1625358336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563590772316951,
      "loss": 2.6711,
      "theoretical_loss": 3.4914605563906376,
      "tokens_seen": 1625423872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025634904714142424,
      "loss": 2.8036,
      "theoretical_loss": 3.4914483351817545,
      "tokens_seen": 1625489408
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025633901705115347,
      "loss": 2.6206,
      "theoretical_loss": 3.49143611460355,
      "tokens_seen": 1625554944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025632898696088265,
      "loss": 2.7595,
      "theoretical_loss": 3.4914238946559673,
      "tokens_seen": 1625620480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025631895687061183,
      "loss": 2.5415,
      "theoretical_loss": 3.4914116753389486,
      "tokens_seen": 1625686016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256308926780341,
      "loss": 2.663,
      "theoretical_loss": 3.4913994566524353,
      "tokens_seen": 1625751552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025629889669007025,
      "loss": 2.8635,
      "theoretical_loss": 3.49138723859637,
      "tokens_seen": 1625817088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562888665997994,
      "loss": 2.7911,
      "theoretical_loss": 3.4913750211706938,
      "tokens_seen": 1625882624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562788365095286,
      "loss": 2.7812,
      "theoretical_loss": 3.4913628043753495,
      "tokens_seen": 1625948160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025626880641925774,
      "loss": 2.7115,
      "theoretical_loss": 3.4913505882102793,
      "tokens_seen": 1626013696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256258776328987,
      "loss": 2.7131,
      "theoretical_loss": 3.491338372675425,
      "tokens_seen": 1626079232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025624874623871616,
      "loss": 2.4298,
      "theoretical_loss": 3.491326157770729,
      "tokens_seen": 1626144768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025623871614844534,
      "loss": 2.5261,
      "theoretical_loss": 3.4913139434961327,
      "tokens_seen": 1626210304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562286860581745,
      "loss": 2.7258,
      "theoretical_loss": 3.4913017298515783,
      "tokens_seen": 1626275840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562186559679037,
      "loss": 2.822,
      "theoretical_loss": 3.491289516837009,
      "tokens_seen": 1626341376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562086258776329,
      "loss": 2.6712,
      "theoretical_loss": 3.4912773044523657,
      "tokens_seen": 1626406912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561985957873621,
      "loss": 2.7299,
      "theoretical_loss": 3.491265092697591,
      "tokens_seen": 1626472448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025618856569709124,
      "loss": 2.5132,
      "theoretical_loss": 3.4912528815726276,
      "tokens_seen": 1626537984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561785356068205,
      "loss": 2.7295,
      "theoretical_loss": 3.491240671077417,
      "tokens_seen": 1626603520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025616850551654966,
      "loss": 2.7413,
      "theoretical_loss": 3.491228461211901,
      "tokens_seen": 1626669056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025615847542627884,
      "loss": 2.7521,
      "theoretical_loss": 3.4912162519760224,
      "tokens_seen": 1626734592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256148445336008,
      "loss": 2.6282,
      "theoretical_loss": 3.4912040433697236,
      "tokens_seen": 1626800128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561384152457372,
      "loss": 2.5982,
      "theoretical_loss": 3.4911918353929456,
      "tokens_seen": 1626865664
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 898227,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.830223321914673,
      "objective/train/theoretical_loss": 3.4911796280456313,
      "objective/train/tokens_used": 1647391200,
      "theoretical_loss": 3.4911796280456313,
      "tokens_seen": 1626931200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561283851554664,
      "loss": 2.7122,
      "theoretical_loss": 3.4911796280456313,
      "tokens_seen": 1626931200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561183550651956,
      "loss": 2.7158,
      "theoretical_loss": 3.4911674213277237,
      "tokens_seen": 1626996736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025610832497492475,
      "loss": 2.6925,
      "theoretical_loss": 3.4911552152391634,
      "tokens_seen": 1627062272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256098294884654,
      "loss": 2.5707,
      "theoretical_loss": 3.491143009779894,
      "tokens_seen": 1627127808
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560882647943831,
      "loss": 2.9297,
      "theoretical_loss": 3.491130804949857,
      "tokens_seen": 1627193344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025607823470411234,
      "loss": 2.7407,
      "theoretical_loss": 3.4911186007489947,
      "tokens_seen": 1627258880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560682046138415,
      "loss": 2.6176,
      "theoretical_loss": 3.491106397177249,
      "tokens_seen": 1627324416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560581745235707,
      "loss": 2.6364,
      "theoretical_loss": 3.491094194234563,
      "tokens_seen": 1627389952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560481444332999,
      "loss": 2.7198,
      "theoretical_loss": 3.491081991920878,
      "tokens_seen": 1627455488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025603811434302907,
      "loss": 2.7497,
      "theoretical_loss": 3.4910697902361374,
      "tokens_seen": 1627521024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025602808425275825,
      "loss": 2.499,
      "theoretical_loss": 3.491057589180282,
      "tokens_seen": 1627586560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560180541624875,
      "loss": 2.5999,
      "theoretical_loss": 3.491045388753255,
      "tokens_seen": 1627652096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560080240722166,
      "loss": 2.6363,
      "theoretical_loss": 3.491033188954999,
      "tokens_seen": 1627717632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025599799398194585,
      "loss": 2.731,
      "theoretical_loss": 3.491020989785455,
      "tokens_seen": 1627783168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559879638916751,
      "loss": 2.7878,
      "theoretical_loss": 3.4910087912445666,
      "tokens_seen": 1627848704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559779338014042,
      "loss": 2.7669,
      "theoretical_loss": 3.4909965933322757,
      "tokens_seen": 1627914240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025596790371113344,
      "loss": 2.5268,
      "theoretical_loss": 3.4909843960485243,
      "tokens_seen": 1627979776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025595787362086257,
      "loss": 2.6021,
      "theoretical_loss": 3.490972199393255,
      "tokens_seen": 1628045312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559478435305918,
      "loss": 2.6335,
      "theoretical_loss": 3.4909600033664097,
      "tokens_seen": 1628110848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000255937813440321,
      "loss": 2.4764,
      "theoretical_loss": 3.490947807967931,
      "tokens_seen": 1628176384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025592778335005017,
      "loss": 2.8449,
      "theoretical_loss": 3.4909356131977614,
      "tokens_seen": 1628241920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025591775325977935,
      "loss": 2.7262,
      "theoretical_loss": 3.4909234190558434,
      "tokens_seen": 1628307456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025590772316950853,
      "loss": 2.5077,
      "theoretical_loss": 3.490911225542119,
      "tokens_seen": 1628372992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558976930792377,
      "loss": 2.6785,
      "theoretical_loss": 3.4908990326565306,
      "tokens_seen": 1628438528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025588766298896695,
      "loss": 2.7593,
      "theoretical_loss": 3.4908868403990203,
      "tokens_seen": 1628504064
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 898909,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.812804937362671,
      "objective/train/theoretical_loss": 3.4908746487695304,
      "objective/train/tokens_used": 1649029600,
      "theoretical_loss": 3.4908746487695304,
      "tokens_seen": 1628569600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558776328986961,
      "loss": 2.7889,
      "theoretical_loss": 3.4908746487695304,
      "tokens_seen": 1628569600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558676028084253,
      "loss": 2.5556,
      "theoretical_loss": 3.4908624577680043,
      "tokens_seen": 1628635136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025585757271815444,
      "loss": 2.4617,
      "theoretical_loss": 3.4908502673943835,
      "tokens_seen": 1628700672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025584754262788367,
      "loss": 2.6531,
      "theoretical_loss": 3.4908380776486108,
      "tokens_seen": 1628766208
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025583751253761285,
      "loss": 2.7881,
      "theoretical_loss": 3.4908258885306283,
      "tokens_seen": 1628831744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025582748244734203,
      "loss": 2.8264,
      "theoretical_loss": 3.4908137000403783,
      "tokens_seen": 1628897280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558174523570712,
      "loss": 2.5923,
      "theoretical_loss": 3.4908015121778035,
      "tokens_seen": 1628962816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025580742226680045,
      "loss": 2.6108,
      "theoretical_loss": 3.4907893249428463,
      "tokens_seen": 1629028352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557973921765296,
      "loss": 2.5873,
      "theoretical_loss": 3.4907771383354493,
      "tokens_seen": 1629093888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557873620862588,
      "loss": 2.4197,
      "theoretical_loss": 3.490764952355555,
      "tokens_seen": 1629159424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025577733199598794,
      "loss": 2.9189,
      "theoretical_loss": 3.490752767003105,
      "tokens_seen": 1629224960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557673019057172,
      "loss": 2.6294,
      "theoretical_loss": 3.4907405822780424,
      "tokens_seen": 1629290496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025575727181544636,
      "loss": 2.7208,
      "theoretical_loss": 3.49072839818031,
      "tokens_seen": 1629356032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025574724172517554,
      "loss": 2.5332,
      "theoretical_loss": 3.49071621470985,
      "tokens_seen": 1629421568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557372116349047,
      "loss": 2.7087,
      "theoretical_loss": 3.4907040318666045,
      "tokens_seen": 1629487104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557271815446339,
      "loss": 2.4777,
      "theoretical_loss": 3.490691849650516,
      "tokens_seen": 1629552640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557171514543631,
      "loss": 2.6762,
      "theoretical_loss": 3.4906796680615275,
      "tokens_seen": 1629618176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557071213640923,
      "loss": 2.6705,
      "theoretical_loss": 3.4906674870995813,
      "tokens_seen": 1629683712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025569709127382144,
      "loss": 2.6812,
      "theoretical_loss": 3.4906553067646198,
      "tokens_seen": 1629749248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556870611835507,
      "loss": 2.6065,
      "theoretical_loss": 3.490643127056585,
      "tokens_seen": 1629814784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025567703109327986,
      "loss": 2.8383,
      "theoretical_loss": 3.490630947975421,
      "tokens_seen": 1629880320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025566700100300904,
      "loss": 2.4477,
      "theoretical_loss": 3.490618769521069,
      "tokens_seen": 1629945856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556569709127382,
      "loss": 2.6374,
      "theoretical_loss": 3.4906065916934716,
      "tokens_seen": 1630011392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556469408224674,
      "loss": 2.767,
      "theoretical_loss": 3.490594414492572,
      "tokens_seen": 1630076928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556369107321966,
      "loss": 2.6867,
      "theoretical_loss": 3.4905822379183125,
      "tokens_seen": 1630142464
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 899823,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1974503993988037,
      "objective/train/theoretical_loss": 3.490570061970635,
      "objective/train/tokens_used": 1650668000,
      "theoretical_loss": 3.490570061970635,
      "tokens_seen": 1630208000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556268806419258,
      "loss": 2.8756,
      "theoretical_loss": 3.490570061970635,
      "tokens_seen": 1630208000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025561685055165495,
      "loss": 2.7593,
      "theoretical_loss": 3.4905578866494826,
      "tokens_seen": 1630273536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556068204613842,
      "loss": 2.5659,
      "theoretical_loss": 3.4905457119547982,
      "tokens_seen": 1630339072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555967903711133,
      "loss": 2.6666,
      "theoretical_loss": 3.490533537886524,
      "tokens_seen": 1630404608
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025558676028084254,
      "loss": 2.5991,
      "theoretical_loss": 3.4905213644446027,
      "tokens_seen": 1630470144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555767301905717,
      "loss": 2.6245,
      "theoretical_loss": 3.490509191628977,
      "tokens_seen": 1630535680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555667001003009,
      "loss": 2.3732,
      "theoretical_loss": 3.490497019439589,
      "tokens_seen": 1630601216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555566700100301,
      "loss": 2.721,
      "theoretical_loss": 3.490484847876382,
      "tokens_seen": 1630666752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025554663991975927,
      "loss": 2.7833,
      "theoretical_loss": 3.490472676939298,
      "tokens_seen": 1630732288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025553660982948845,
      "loss": 2.5654,
      "theoretical_loss": 3.4904605066282803,
      "tokens_seen": 1630797824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555265797392177,
      "loss": 2.5818,
      "theoretical_loss": 3.4904483369432704,
      "tokens_seen": 1630863360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555165496489468,
      "loss": 2.7179,
      "theoretical_loss": 3.4904361678842126,
      "tokens_seen": 1630928896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025550651955867605,
      "loss": 2.4192,
      "theoretical_loss": 3.4904239994510484,
      "tokens_seen": 1630994432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025549648946840523,
      "loss": 2.4348,
      "theoretical_loss": 3.490411831643721,
      "tokens_seen": 1631059968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554864593781344,
      "loss": 2.7202,
      "theoretical_loss": 3.4903996644621724,
      "tokens_seen": 1631125504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554764292878636,
      "loss": 2.6147,
      "theoretical_loss": 3.4903874979063456,
      "tokens_seen": 1631191040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025546639919759277,
      "loss": 2.438,
      "theoretical_loss": 3.4903753319761837,
      "tokens_seen": 1631256576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025545636910732195,
      "loss": 2.8955,
      "theoretical_loss": 3.490363166671629,
      "tokens_seen": 1631322112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554463390170512,
      "loss": 2.5263,
      "theoretical_loss": 3.4903510019926243,
      "tokens_seen": 1631387648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554363089267803,
      "loss": 2.7364,
      "theoretical_loss": 3.4903388379391123,
      "tokens_seen": 1631453184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025542627883650955,
      "loss": 2.8961,
      "theoretical_loss": 3.490326674511035,
      "tokens_seen": 1631518720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554162487462387,
      "loss": 2.5371,
      "theoretical_loss": 3.4903145117083367,
      "tokens_seen": 1631584256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554062186559679,
      "loss": 2.4477,
      "theoretical_loss": 3.4903023495309586,
      "tokens_seen": 1631649792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553961885656971,
      "loss": 2.5275,
      "theoretical_loss": 3.490290187978845,
      "tokens_seen": 1631715328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553861584754263,
      "loss": 2.6323,
      "theoretical_loss": 3.4902780270519367,
      "tokens_seen": 1631780864
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 900688,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.873173475265503,
      "objective/train/theoretical_loss": 3.4902658667501774,
      "objective/train/tokens_used": 1652306400,
      "theoretical_loss": 3.4902658667501774,
      "tokens_seen": 1631846400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025537612838515546,
      "loss": 2.6775,
      "theoretical_loss": 3.4902658667501774,
      "tokens_seen": 1631846400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025536609829488464,
      "loss": 2.518,
      "theoretical_loss": 3.4902537070735105,
      "tokens_seen": 1631911936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553560682046138,
      "loss": 2.7277,
      "theoretical_loss": 3.490241548021878,
      "tokens_seen": 1631977472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025534603811434305,
      "loss": 2.6138,
      "theoretical_loss": 3.490229389595223,
      "tokens_seen": 1632043008
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553360080240722,
      "loss": 2.5422,
      "theoretical_loss": 3.490217231793488,
      "tokens_seen": 1632108544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553259779338014,
      "loss": 2.5416,
      "theoretical_loss": 3.4902050746166156,
      "tokens_seen": 1632174080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553159478435306,
      "loss": 2.4946,
      "theoretical_loss": 3.4901929180645492,
      "tokens_seen": 1632239616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553059177532598,
      "loss": 2.7374,
      "theoretical_loss": 3.490180762137231,
      "tokens_seen": 1632305152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025529588766298896,
      "loss": 2.3845,
      "theoretical_loss": 3.4901686068346045,
      "tokens_seen": 1632370688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025528585757271814,
      "loss": 2.7952,
      "theoretical_loss": 3.4901564521566124,
      "tokens_seen": 1632436224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552758274824473,
      "loss": 2.6978,
      "theoretical_loss": 3.490144298103197,
      "tokens_seen": 1632501760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025526579739217656,
      "loss": 2.5157,
      "theoretical_loss": 3.4901321446743014,
      "tokens_seen": 1632567296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552557673019057,
      "loss": 2.7612,
      "theoretical_loss": 3.490119991869869,
      "tokens_seen": 1632632832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552457372116349,
      "loss": 2.7876,
      "theoretical_loss": 3.4901078396898413,
      "tokens_seen": 1632698368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552357071213641,
      "loss": 2.6199,
      "theoretical_loss": 3.4900956881341623,
      "tokens_seen": 1632763904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552256770310933,
      "loss": 2.5505,
      "theoretical_loss": 3.490083537202774,
      "tokens_seen": 1632829440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552156469408225,
      "loss": 2.7916,
      "theoretical_loss": 3.490071386895621,
      "tokens_seen": 1632894976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025520561685055164,
      "loss": 2.4555,
      "theoretical_loss": 3.4900592372126438,
      "tokens_seen": 1632960512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551955867602809,
      "loss": 2.7613,
      "theoretical_loss": 3.490047088153787,
      "tokens_seen": 1633026048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025518555667001006,
      "loss": 2.2399,
      "theoretical_loss": 3.490034939718993,
      "tokens_seen": 1633091584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025517552657973924,
      "loss": 2.697,
      "theoretical_loss": 3.4900227919082045,
      "tokens_seen": 1633157120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551654964894684,
      "loss": 2.6479,
      "theoretical_loss": 3.490010644721365,
      "tokens_seen": 1633222656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551554663991976,
      "loss": 2.7111,
      "theoretical_loss": 3.4899984981584167,
      "tokens_seen": 1633288192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551454363089268,
      "loss": 2.6725,
      "theoretical_loss": 3.4899863522193026,
      "tokens_seen": 1633353728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000255135406218656,
      "loss": 2.6307,
      "theoretical_loss": 3.489974206903966,
      "tokens_seen": 1633419264
    },
    {
      "epoch": 0.49,
      "objective/train/docs_used": 901970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6137895584106445,
      "objective/train/theoretical_loss": 3.4899620622123497,
      "objective/train/tokens_used": 1653944800,
      "theoretical_loss": 3.4899620622123497,
      "tokens_seen": 1633484800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025512537612838515,
      "loss": 2.3214,
      "theoretical_loss": 3.4899620622123497,
      "tokens_seen": 1633484800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002551153460381144,
      "loss": 2.8042,
      "theoretical_loss": 3.4899499181443967,
      "tokens_seen": 1633550336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002551053159478435,
      "loss": 2.7133,
      "theoretical_loss": 3.4899377747000497,
      "tokens_seen": 1633615872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025509528585757274,
      "loss": 2.7481,
      "theoretical_loss": 3.489925631879252,
      "tokens_seen": 1633681408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550852557673019,
      "loss": 2.6438,
      "theoretical_loss": 3.489913489681946,
      "tokens_seen": 1633746944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550752256770311,
      "loss": 2.6686,
      "theoretical_loss": 3.4899013481080763,
      "tokens_seen": 1633812480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550651955867603,
      "loss": 2.8043,
      "theoretical_loss": 3.4898892071575833,
      "tokens_seen": 1633878016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025505516549648947,
      "loss": 2.4947,
      "theoretical_loss": 3.489877066830412,
      "tokens_seen": 1633943552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025504513540621865,
      "loss": 2.4075,
      "theoretical_loss": 3.489864927126505,
      "tokens_seen": 1634009088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550351053159479,
      "loss": 2.8582,
      "theoretical_loss": 3.4898527880458046,
      "tokens_seen": 1634074624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000255025075225677,
      "loss": 2.4703,
      "theoretical_loss": 3.4898406495882544,
      "tokens_seen": 1634140160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025501504513540625,
      "loss": 2.4688,
      "theoretical_loss": 3.4898285117537977,
      "tokens_seen": 1634205696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025500501504513543,
      "loss": 2.5317,
      "theoretical_loss": 3.489816374542377,
      "tokens_seen": 1634271232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549949849548646,
      "loss": 2.4912,
      "theoretical_loss": 3.489804237953935,
      "tokens_seen": 1634336768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549849548645938,
      "loss": 2.5076,
      "theoretical_loss": 3.489792101988416,
      "tokens_seen": 1634402304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025497492477432297,
      "loss": 2.7753,
      "theoretical_loss": 3.4897799666457616,
      "tokens_seen": 1634467840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025496489468405215,
      "loss": 2.548,
      "theoretical_loss": 3.489767831925916,
      "tokens_seen": 1634533376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549548645937814,
      "loss": 2.8576,
      "theoretical_loss": 3.4897556978288216,
      "tokens_seen": 1634598912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549448345035105,
      "loss": 2.5152,
      "theoretical_loss": 3.4897435643544217,
      "tokens_seen": 1634664448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025493480441323975,
      "loss": 2.5636,
      "theoretical_loss": 3.4897314315026593,
      "tokens_seen": 1634729984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549247743229689,
      "loss": 2.6746,
      "theoretical_loss": 3.4897192992734776,
      "tokens_seen": 1634795520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549147442326981,
      "loss": 2.6329,
      "theoretical_loss": 3.4897071676668197,
      "tokens_seen": 1634861056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549047141424273,
      "loss": 2.5526,
      "theoretical_loss": 3.4896950366826287,
      "tokens_seen": 1634926592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548946840521565,
      "loss": 2.4244,
      "theoretical_loss": 3.4896829063208474,
      "tokens_seen": 1634992128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025488465396188566,
      "loss": 2.5959,
      "theoretical_loss": 3.4896707765814194,
      "tokens_seen": 1635057664
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 902458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.477343797683716,
      "objective/train/theoretical_loss": 3.489658647464288,
      "objective/train/tokens_used": 1655583200,
      "theoretical_loss": 3.489658647464288,
      "tokens_seen": 1635123200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025487462387161484,
      "loss": 2.523,
      "theoretical_loss": 3.489658647464288,
      "tokens_seen": 1635123200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254864593781344,
      "loss": 2.637,
      "theoretical_loss": 3.489646518969395,
      "tokens_seen": 1635188736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025485456369107325,
      "loss": 2.6359,
      "theoretical_loss": 3.4896343910966854,
      "tokens_seen": 1635254272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548445336008024,
      "loss": 2.3162,
      "theoretical_loss": 3.4896222638461007,
      "tokens_seen": 1635319808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548345035105316,
      "loss": 2.6989,
      "theoretical_loss": 3.4896101372175856,
      "tokens_seen": 1635385344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548244734202608,
      "loss": 2.5139,
      "theoretical_loss": 3.489598011211082,
      "tokens_seen": 1635450880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025481444332999,
      "loss": 2.6453,
      "theoretical_loss": 3.489585885826533,
      "tokens_seen": 1635516416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025480441323971916,
      "loss": 2.7129,
      "theoretical_loss": 3.489573761063883,
      "tokens_seen": 1635581952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025479438314944834,
      "loss": 2.6297,
      "theoretical_loss": 3.4895616369230744,
      "tokens_seen": 1635647488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547843530591775,
      "loss": 2.6764,
      "theoretical_loss": 3.48954951340405,
      "tokens_seen": 1635713024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025477432296890676,
      "loss": 2.4747,
      "theoretical_loss": 3.4895373905067544,
      "tokens_seen": 1635778560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547642928786359,
      "loss": 2.6663,
      "theoretical_loss": 3.4895252682311293,
      "tokens_seen": 1635844096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547542627883651,
      "loss": 2.5264,
      "theoretical_loss": 3.489513146577119,
      "tokens_seen": 1635909632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025474423269809425,
      "loss": 2.5346,
      "theoretical_loss": 3.4895010255446657,
      "tokens_seen": 1635975168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547342026078235,
      "loss": 2.6398,
      "theoretical_loss": 3.4894889051337135,
      "tokens_seen": 1636040704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025472417251755266,
      "loss": 2.5351,
      "theoretical_loss": 3.489476785344205,
      "tokens_seen": 1636106240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025471414242728184,
      "loss": 2.7657,
      "theoretical_loss": 3.489464666176084,
      "tokens_seen": 1636171776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254704112337011,
      "loss": 2.864,
      "theoretical_loss": 3.4894525476292935,
      "tokens_seen": 1636237312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025469408224674026,
      "loss": 2.6022,
      "theoretical_loss": 3.4894404297037767,
      "tokens_seen": 1636302848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546840521564694,
      "loss": 2.7555,
      "theoretical_loss": 3.4894283123994767,
      "tokens_seen": 1636368384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546740220661986,
      "loss": 2.6694,
      "theoretical_loss": 3.489416195716337,
      "tokens_seen": 1636433920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025466399197592775,
      "loss": 2.425,
      "theoretical_loss": 3.4894040796543013,
      "tokens_seen": 1636499456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254653961885657,
      "loss": 2.4242,
      "theoretical_loss": 3.489391964213312,
      "tokens_seen": 1636564992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025464393179538616,
      "loss": 2.6197,
      "theoretical_loss": 3.489379849393314,
      "tokens_seen": 1636630528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025463390170511535,
      "loss": 2.6403,
      "theoretical_loss": 3.4893677351942483,
      "tokens_seen": 1636696064
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 903426,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7221128940582275,
      "objective/train/theoretical_loss": 3.48935562161606,
      "objective/train/tokens_used": 1657221600,
      "theoretical_loss": 3.48935562161606,
      "tokens_seen": 1636761600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025462387161484453,
      "loss": 2.6297,
      "theoretical_loss": 3.48935562161606,
      "tokens_seen": 1636761600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546138415245737,
      "loss": 2.5678,
      "theoretical_loss": 3.489343508658692,
      "tokens_seen": 1636827136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546038114343029,
      "loss": 2.5796,
      "theoretical_loss": 3.489331396322087,
      "tokens_seen": 1636892672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545937813440321,
      "loss": 2.7055,
      "theoretical_loss": 3.489319284606189,
      "tokens_seen": 1636958208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025458375125376125,
      "loss": 2.4974,
      "theoretical_loss": 3.489307173510941,
      "tokens_seen": 1637023744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545737211634905,
      "loss": 2.7415,
      "theoretical_loss": 3.4892950630362867,
      "tokens_seen": 1637089280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545636910732196,
      "loss": 2.647,
      "theoretical_loss": 3.489282953182169,
      "tokens_seen": 1637154816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025455366098294885,
      "loss": 2.6432,
      "theoretical_loss": 3.489270843948532,
      "tokens_seen": 1637220352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025454363089267803,
      "loss": 2.6731,
      "theoretical_loss": 3.489258735335318,
      "tokens_seen": 1637285888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545336008024072,
      "loss": 2.7159,
      "theoretical_loss": 3.489246627342471,
      "tokens_seen": 1637351424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545235707121364,
      "loss": 2.5777,
      "theoretical_loss": 3.4892345199699344,
      "tokens_seen": 1637416960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025451354062186563,
      "loss": 2.6603,
      "theoretical_loss": 3.489222413217652,
      "tokens_seen": 1637482496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545035105315948,
      "loss": 2.6612,
      "theoretical_loss": 3.4892103070855662,
      "tokens_seen": 1637548032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254493480441324,
      "loss": 2.7114,
      "theoretical_loss": 3.489198201573621,
      "tokens_seen": 1637613568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025448345035105317,
      "loss": 2.6173,
      "theoretical_loss": 3.48918609668176,
      "tokens_seen": 1637679104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025447342026078235,
      "loss": 2.5193,
      "theoretical_loss": 3.4891739924099268,
      "tokens_seen": 1637744640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544633901705116,
      "loss": 2.4744,
      "theoretical_loss": 3.4891618887580638,
      "tokens_seen": 1637810176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544533600802407,
      "loss": 2.6522,
      "theoretical_loss": 3.489149785726115,
      "tokens_seen": 1637875712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025444332998996995,
      "loss": 2.7144,
      "theoretical_loss": 3.4891376833140244,
      "tokens_seen": 1637941248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544332998996991,
      "loss": 2.6649,
      "theoretical_loss": 3.489125581521735,
      "tokens_seen": 1638006784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544232698094283,
      "loss": 2.5846,
      "theoretical_loss": 3.4891134803491894,
      "tokens_seen": 1638072320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544132397191575,
      "loss": 2.6829,
      "theoretical_loss": 3.489101379796333,
      "tokens_seen": 1638137856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544032096288867,
      "loss": 2.6106,
      "theoretical_loss": 3.4890892798631077,
      "tokens_seen": 1638203392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025439317953861586,
      "loss": 2.5666,
      "theoretical_loss": 3.489077180549457,
      "tokens_seen": 1638268928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025438314944834504,
      "loss": 2.8759,
      "theoretical_loss": 3.4890650818553257,
      "tokens_seen": 1638334464
    },
    {
      "debugging/Self-BLEU-5": 0.4993282995332528,
      "debugging/distinct-1-grams": 0.7782354038853119,
      "debugging/distinct-2-grams": 0.945600607571183,
      "debugging/entropy-1-grams": 5.791543305262103,
      "debugging/entropy-2-grams": 6.444393817535788,
      "debugging/length": 527.3,
      "debugging/num_segments": 10,
      "epoch": 0.5,
      "objective/train/docs_used": 904059,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.070502758026123,
      "objective/train/theoretical_loss": 3.489052983780656,
      "objective/train/tokens_used": 1658860000,
      "theoretical_loss": 3.489052983780656,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543731193580742,
      "loss": 2.8441,
      "theoretical_loss": 3.489052983780656,
      "tokens_seen": 1638400000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025436308926780345,
      "loss": 2.6732,
      "theoretical_loss": 3.489040886325392,
      "tokens_seen": 1638465536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543530591775326,
      "loss": 2.7544,
      "theoretical_loss": 3.489028789489477,
      "tokens_seen": 1638531072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543430290872618,
      "loss": 2.6605,
      "theoretical_loss": 3.4890166932728546,
      "tokens_seen": 1638596608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254332998996991,
      "loss": 2.5869,
      "theoretical_loss": 3.4890045976754687,
      "tokens_seen": 1638662144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543229689067202,
      "loss": 2.522,
      "theoretical_loss": 3.488992502697262,
      "tokens_seen": 1638727680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025431293881644936,
      "loss": 2.5652,
      "theoretical_loss": 3.4889804083381786,
      "tokens_seen": 1638793216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025430290872617854,
      "loss": 2.7658,
      "theoretical_loss": 3.488968314598162,
      "tokens_seen": 1638858752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542928786359077,
      "loss": 2.6907,
      "theoretical_loss": 3.4889562214771557,
      "tokens_seen": 1638924288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025428284854563696,
      "loss": 2.62,
      "theoretical_loss": 3.4889441289751035,
      "tokens_seen": 1638989824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542728184553661,
      "loss": 2.69,
      "theoretical_loss": 3.4889320370919483,
      "tokens_seen": 1639055360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542627883650953,
      "loss": 2.5552,
      "theoretical_loss": 3.4889199458276345,
      "tokens_seen": 1639120896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025425275827482445,
      "loss": 2.5333,
      "theoretical_loss": 3.4889078551821053,
      "tokens_seen": 1639186432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542427281845537,
      "loss": 2.6868,
      "theoretical_loss": 3.4888957651553043,
      "tokens_seen": 1639251968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025423269809428286,
      "loss": 2.416,
      "theoretical_loss": 3.488883675747175,
      "tokens_seen": 1639317504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025422266800401204,
      "loss": 2.6224,
      "theoretical_loss": 3.4888715869576616,
      "tokens_seen": 1639383040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542126379137412,
      "loss": 2.76,
      "theoretical_loss": 3.4888594987867068,
      "tokens_seen": 1639448576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025420260782347046,
      "loss": 2.6552,
      "theoretical_loss": 3.4888474112342545,
      "tokens_seen": 1639514112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541925777331996,
      "loss": 2.6712,
      "theoretical_loss": 3.488835324300249,
      "tokens_seen": 1639579648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541825476429288,
      "loss": 2.6095,
      "theoretical_loss": 3.488823237984633,
      "tokens_seen": 1639645184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025417251755265795,
      "loss": 2.5455,
      "theoretical_loss": 3.488811152287351,
      "tokens_seen": 1639710720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541624874623872,
      "loss": 2.6803,
      "theoretical_loss": 3.4887990672083458,
      "tokens_seen": 1639776256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025415245737211637,
      "loss": 2.7025,
      "theoretical_loss": 3.4887869827475617,
      "tokens_seen": 1639841792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025414242728184555,
      "loss": 2.6112,
      "theoretical_loss": 3.4887748989049423,
      "tokens_seen": 1639907328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025413239719157473,
      "loss": 2.4397,
      "theoretical_loss": 3.488762815680431,
      "tokens_seen": 1639972864
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 904549,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8121564388275146,
      "objective/train/theoretical_loss": 3.4887507330739718,
      "objective/train/tokens_used": 1660498400,
      "theoretical_loss": 3.4887507330739718,
      "tokens_seen": 1640038400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541223671013039,
      "loss": 2.8177,
      "theoretical_loss": 3.4887507330739718,
      "tokens_seen": 1640038400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541123370110331,
      "loss": 2.6086,
      "theoretical_loss": 3.4887386510855087,
      "tokens_seen": 1640103936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541023069207623,
      "loss": 2.5238,
      "theoretical_loss": 3.4887265697149843,
      "tokens_seen": 1640169472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025409227683049145,
      "loss": 2.571,
      "theoretical_loss": 3.4887144889623425,
      "tokens_seen": 1640235008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540822467402207,
      "loss": 2.6282,
      "theoretical_loss": 3.4887024088275282,
      "tokens_seen": 1640300544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540722166499498,
      "loss": 2.7475,
      "theoretical_loss": 3.4886903293104847,
      "tokens_seen": 1640366080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025406218655967905,
      "loss": 2.5826,
      "theoretical_loss": 3.4886782504111546,
      "tokens_seen": 1640431616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025405215646940823,
      "loss": 2.6086,
      "theoretical_loss": 3.4886661721294825,
      "tokens_seen": 1640497152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540421263791374,
      "loss": 2.7802,
      "theoretical_loss": 3.4886540944654127,
      "tokens_seen": 1640562688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540320962888666,
      "loss": 2.7161,
      "theoretical_loss": 3.488642017418888,
      "tokens_seen": 1640628224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025402206619859583,
      "loss": 2.6765,
      "theoretical_loss": 3.4886299409898522,
      "tokens_seen": 1640693760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025401203610832495,
      "loss": 2.7266,
      "theoretical_loss": 3.4886178651782496,
      "tokens_seen": 1640759296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540020060180542,
      "loss": 2.822,
      "theoretical_loss": 3.4886057899840237,
      "tokens_seen": 1640824832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539919759277833,
      "loss": 2.6459,
      "theoretical_loss": 3.4885937154071183,
      "tokens_seen": 1640890368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025398194583751255,
      "loss": 2.8957,
      "theoretical_loss": 3.4885816414474773,
      "tokens_seen": 1640955904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025397191574724173,
      "loss": 2.7312,
      "theoretical_loss": 3.4885695681050444,
      "tokens_seen": 1641021440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539618856569709,
      "loss": 2.5882,
      "theoretical_loss": 3.488557495379763,
      "tokens_seen": 1641086976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539518555667001,
      "loss": 2.6916,
      "theoretical_loss": 3.488545423271578,
      "tokens_seen": 1641152512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539418254764293,
      "loss": 2.6297,
      "theoretical_loss": 3.488533351780432,
      "tokens_seen": 1641218048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025393179538615846,
      "loss": 2.6452,
      "theoretical_loss": 3.4885212809062693,
      "tokens_seen": 1641283584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539217652958877,
      "loss": 2.5714,
      "theoretical_loss": 3.488509210649034,
      "tokens_seen": 1641349120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002539117352056168,
      "loss": 2.7709,
      "theoretical_loss": 3.4884971410086694,
      "tokens_seen": 1641414656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025390170511534606,
      "loss": 2.5685,
      "theoretical_loss": 3.48848507198512,
      "tokens_seen": 1641480192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538916750250752,
      "loss": 2.776,
      "theoretical_loss": 3.4884730035783296,
      "tokens_seen": 1641545728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538816449348044,
      "loss": 2.6852,
      "theoretical_loss": 3.4884609357882415,
      "tokens_seen": 1641611264
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 905573,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.890195608139038,
      "objective/train/theoretical_loss": 3.488448868614799,
      "objective/train/tokens_used": 1662136800,
      "theoretical_loss": 3.488448868614799,
      "tokens_seen": 1641676800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538716148445336,
      "loss": 2.5247,
      "theoretical_loss": 3.488448868614799,
      "tokens_seen": 1641676800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538615847542628,
      "loss": 2.5465,
      "theoretical_loss": 3.4884368020579477,
      "tokens_seen": 1641742336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025385155466399196,
      "loss": 2.8159,
      "theoretical_loss": 3.4884247361176306,
      "tokens_seen": 1641807872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538415245737212,
      "loss": 2.6373,
      "theoretical_loss": 3.4884126707937915,
      "tokens_seen": 1641873408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538314944834503,
      "loss": 2.5439,
      "theoretical_loss": 3.4884006060863744,
      "tokens_seen": 1641938944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025382146439317956,
      "loss": 2.6431,
      "theoretical_loss": 3.488388541995323,
      "tokens_seen": 1642004480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538114343029087,
      "loss": 2.5298,
      "theoretical_loss": 3.488376478520581,
      "tokens_seen": 1642070016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538014042126379,
      "loss": 2.5858,
      "theoretical_loss": 3.488364415662094,
      "tokens_seen": 1642135552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537913741223671,
      "loss": 2.7121,
      "theoretical_loss": 3.4883523534198035,
      "tokens_seen": 1642201088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537813440320963,
      "loss": 2.6309,
      "theoretical_loss": 3.488340291793655,
      "tokens_seen": 1642266624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025377131394182546,
      "loss": 2.4925,
      "theoretical_loss": 3.488328230783592,
      "tokens_seen": 1642332160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025376128385155465,
      "loss": 2.5332,
      "theoretical_loss": 3.488316170389558,
      "tokens_seen": 1642397696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537512537612839,
      "loss": 2.7789,
      "theoretical_loss": 3.488304110611498,
      "tokens_seen": 1642463232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025374122367101306,
      "loss": 2.8163,
      "theoretical_loss": 3.4882920514493554,
      "tokens_seen": 1642528768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025373119358074224,
      "loss": 2.8125,
      "theoretical_loss": 3.488279992903074,
      "tokens_seen": 1642594304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537211634904714,
      "loss": 2.5717,
      "theoretical_loss": 3.4882679349725985,
      "tokens_seen": 1642659840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025371113340020066,
      "loss": 2.6159,
      "theoretical_loss": 3.4882558776578714,
      "tokens_seen": 1642725376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537011033099298,
      "loss": 2.6103,
      "theoretical_loss": 3.4882438209588384,
      "tokens_seen": 1642790912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253691073219659,
      "loss": 2.8862,
      "theoretical_loss": 3.488231764875442,
      "tokens_seen": 1642856448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025368104312938815,
      "loss": 2.5776,
      "theoretical_loss": 3.4882197094076273,
      "tokens_seen": 1642921984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536710130391174,
      "loss": 2.6412,
      "theoretical_loss": 3.488207654555338,
      "tokens_seen": 1642987520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025366098294884657,
      "loss": 2.6165,
      "theoretical_loss": 3.4881956003185177,
      "tokens_seen": 1643053056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025365095285857575,
      "loss": 2.4645,
      "theoretical_loss": 3.4881835466971114,
      "tokens_seen": 1643118592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025364092276830493,
      "loss": 2.5726,
      "theoretical_loss": 3.488171493691062,
      "tokens_seen": 1643184128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536308926780341,
      "loss": 2.9511,
      "theoretical_loss": 3.4881594413003145,
      "tokens_seen": 1643249664
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 906761,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9660711288452148,
      "objective/train/theoretical_loss": 3.488147389524812,
      "objective/train/tokens_used": 1663775200,
      "theoretical_loss": 3.488147389524812,
      "tokens_seen": 1643315200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536208625877633,
      "loss": 2.4855,
      "theoretical_loss": 3.488147389524812,
      "tokens_seen": 1643315200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536108324974925,
      "loss": 2.4745,
      "theoretical_loss": 3.4881353383644997,
      "tokens_seen": 1643380736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025360080240722165,
      "loss": 2.5156,
      "theoretical_loss": 3.4881232878193207,
      "tokens_seen": 1643446272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535907723169509,
      "loss": 2.4046,
      "theoretical_loss": 3.4881112378892194,
      "tokens_seen": 1643511808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025358074222668,
      "loss": 2.4899,
      "theoretical_loss": 3.4880991885741404,
      "tokens_seen": 1643577344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025357071213640925,
      "loss": 2.5556,
      "theoretical_loss": 3.4880871398740267,
      "tokens_seen": 1643642880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025356068204613843,
      "loss": 2.7054,
      "theoretical_loss": 3.488075091788823,
      "tokens_seen": 1643708416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535506519558676,
      "loss": 2.8161,
      "theoretical_loss": 3.4880630443184737,
      "tokens_seen": 1643773952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535406218655968,
      "loss": 2.7645,
      "theoretical_loss": 3.4880509974629224,
      "tokens_seen": 1643839488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025353059177532603,
      "loss": 2.6721,
      "theoretical_loss": 3.488038951222114,
      "tokens_seen": 1643905024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025352056168505516,
      "loss": 2.6677,
      "theoretical_loss": 3.4880269055959916,
      "tokens_seen": 1643970560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535105315947844,
      "loss": 2.569,
      "theoretical_loss": 3.4880148605844994,
      "tokens_seen": 1644036096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535005015045135,
      "loss": 2.7944,
      "theoretical_loss": 3.4880028161875822,
      "tokens_seen": 1644101632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025349047141424275,
      "loss": 2.5979,
      "theoretical_loss": 3.487990772405184,
      "tokens_seen": 1644167168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025348044132397193,
      "loss": 2.5657,
      "theoretical_loss": 3.487978729237249,
      "tokens_seen": 1644232704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534704112337011,
      "loss": 2.668,
      "theoretical_loss": 3.4879666866837207,
      "tokens_seen": 1644298240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534603811434303,
      "loss": 2.5585,
      "theoretical_loss": 3.487954644744544,
      "tokens_seen": 1644363776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534503510531595,
      "loss": 2.5699,
      "theoretical_loss": 3.487942603419663,
      "tokens_seen": 1644429312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025344032096288866,
      "loss": 2.7541,
      "theoretical_loss": 3.487930562709021,
      "tokens_seen": 1644494848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534302908726179,
      "loss": 2.5378,
      "theoretical_loss": 3.4879185226125635,
      "tokens_seen": 1644560384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253420260782347,
      "loss": 2.7998,
      "theoretical_loss": 3.487906483130234,
      "tokens_seen": 1644625920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025341023069207626,
      "loss": 2.5876,
      "theoretical_loss": 3.4878944442619773,
      "tokens_seen": 1644691456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534002006018054,
      "loss": 2.5748,
      "theoretical_loss": 3.4878824060077367,
      "tokens_seen": 1644756992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533901705115346,
      "loss": 2.4012,
      "theoretical_loss": 3.4878703683674566,
      "tokens_seen": 1644822528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533801404212638,
      "loss": 2.6506,
      "theoretical_loss": 3.487858331341082,
      "tokens_seen": 1644888064
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 907335,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8264408111572266,
      "objective/train/theoretical_loss": 3.4878462949285556,
      "objective/train/tokens_used": 1665413600,
      "theoretical_loss": 3.4878462949285556,
      "tokens_seen": 1644953600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253370110330993,
      "loss": 2.9204,
      "theoretical_loss": 3.4878462949285556,
      "tokens_seen": 1644953600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025336008024072216,
      "loss": 2.4967,
      "theoretical_loss": 3.4878342591298237,
      "tokens_seen": 1645019136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533500501504514,
      "loss": 2.7797,
      "theoretical_loss": 3.487822223944829,
      "tokens_seen": 1645084672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533400200601805,
      "loss": 2.6993,
      "theoretical_loss": 3.4878101893735156,
      "tokens_seen": 1645150208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025332998996990976,
      "loss": 2.887,
      "theoretical_loss": 3.4877981554158293,
      "tokens_seen": 1645215744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533199598796389,
      "loss": 2.5216,
      "theoretical_loss": 3.4877861220717135,
      "tokens_seen": 1645281280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533099297893681,
      "loss": 2.7061,
      "theoretical_loss": 3.487774089341112,
      "tokens_seen": 1645346816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532998996990973,
      "loss": 2.3751,
      "theoretical_loss": 3.4877620572239696,
      "tokens_seen": 1645412352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532898696088265,
      "loss": 2.5392,
      "theoretical_loss": 3.487750025720231,
      "tokens_seen": 1645477888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025327983951855566,
      "loss": 2.5422,
      "theoretical_loss": 3.4877379948298395,
      "tokens_seen": 1645543424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025326980942828485,
      "loss": 2.6429,
      "theoretical_loss": 3.48772596455274,
      "tokens_seen": 1645608960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253259779338014,
      "loss": 2.4848,
      "theoretical_loss": 3.487713934888877,
      "tokens_seen": 1645674496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025324974924774326,
      "loss": 2.6671,
      "theoretical_loss": 3.487701905838194,
      "tokens_seen": 1645740032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532397191574724,
      "loss": 2.8147,
      "theoretical_loss": 3.487689877400637,
      "tokens_seen": 1645805568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532296890672016,
      "loss": 2.6472,
      "theoretical_loss": 3.4876778495761487,
      "tokens_seen": 1645871104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025321965897693075,
      "loss": 2.8622,
      "theoretical_loss": 3.487665822364674,
      "tokens_seen": 1645936640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025320962888666,
      "loss": 2.5986,
      "theoretical_loss": 3.487653795766157,
      "tokens_seen": 1646002176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025319959879638917,
      "loss": 2.5872,
      "theoretical_loss": 3.487641769780542,
      "tokens_seen": 1646067712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025318956870611835,
      "loss": 2.7704,
      "theoretical_loss": 3.4876297444077737,
      "tokens_seen": 1646133248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025317953861584753,
      "loss": 2.382,
      "theoretical_loss": 3.487617719647797,
      "tokens_seen": 1646198784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025316950852557677,
      "loss": 2.6681,
      "theoretical_loss": 3.487605695500555,
      "tokens_seen": 1646264320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531594784353059,
      "loss": 2.5585,
      "theoretical_loss": 3.4875936719659935,
      "tokens_seen": 1646329856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025314944834503513,
      "loss": 2.5622,
      "theoretical_loss": 3.4875816490440563,
      "tokens_seen": 1646395392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025313941825476425,
      "loss": 2.6301,
      "theoretical_loss": 3.487569626734687,
      "tokens_seen": 1646460928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531293881644935,
      "loss": 2.8262,
      "theoretical_loss": 3.487557605037831,
      "tokens_seen": 1646526464
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 908555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.054795503616333,
      "objective/train/theoretical_loss": 3.4875455839534326,
      "objective/train/tokens_used": 1667052000,
      "theoretical_loss": 3.4875455839534326,
      "tokens_seen": 1646592000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025311935807422267,
      "loss": 2.662,
      "theoretical_loss": 3.4875455839534326,
      "tokens_seen": 1646592000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025310932798395185,
      "loss": 2.4539,
      "theoretical_loss": 3.4875335634814357,
      "tokens_seen": 1646657536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025309929789368103,
      "loss": 2.5998,
      "theoretical_loss": 3.4875215436217855,
      "tokens_seen": 1646723072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530892678034102,
      "loss": 2.5366,
      "theoretical_loss": 3.487509524374426,
      "tokens_seen": 1646788608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530792377131394,
      "loss": 2.6798,
      "theoretical_loss": 3.487497505739301,
      "tokens_seen": 1646854144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025306920762286863,
      "loss": 2.667,
      "theoretical_loss": 3.4874854877163557,
      "tokens_seen": 1646919680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025305917753259776,
      "loss": 2.6236,
      "theoretical_loss": 3.487473470305535,
      "tokens_seen": 1646985216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253049147442327,
      "loss": 2.4557,
      "theoretical_loss": 3.487461453506783,
      "tokens_seen": 1647050752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530391173520562,
      "loss": 2.5589,
      "theoretical_loss": 3.487449437320043,
      "tokens_seen": 1647116288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025302908726178536,
      "loss": 2.5103,
      "theoretical_loss": 3.487437421745261,
      "tokens_seen": 1647181824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025301905717151454,
      "loss": 2.6911,
      "theoretical_loss": 3.487425406782381,
      "tokens_seen": 1647247360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530090270812437,
      "loss": 2.5283,
      "theoretical_loss": 3.487413392431348,
      "tokens_seen": 1647312896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025299899699097295,
      "loss": 2.5551,
      "theoretical_loss": 3.487401378692105,
      "tokens_seen": 1647378432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025298896690070213,
      "loss": 2.4347,
      "theoretical_loss": 3.487389365564598,
      "tokens_seen": 1647443968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529789368104313,
      "loss": 2.6411,
      "theoretical_loss": 3.487377353048771,
      "tokens_seen": 1647509504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529689067201605,
      "loss": 2.3506,
      "theoretical_loss": 3.487365341144568,
      "tokens_seen": 1647575040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529588766298897,
      "loss": 2.6173,
      "theoretical_loss": 3.487353329851935,
      "tokens_seen": 1647640576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025294884653961886,
      "loss": 2.7054,
      "theoretical_loss": 3.487341319170815,
      "tokens_seen": 1647706112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529388164493481,
      "loss": 2.6216,
      "theoretical_loss": 3.4873293091011535,
      "tokens_seen": 1647771648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529287863590772,
      "loss": 2.6152,
      "theoretical_loss": 3.487317299642894,
      "tokens_seen": 1647837184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025291875626880646,
      "loss": 2.6393,
      "theoretical_loss": 3.487305290795982,
      "tokens_seen": 1647902720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529087261785356,
      "loss": 2.5379,
      "theoretical_loss": 3.487293282560362,
      "tokens_seen": 1647968256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528986960882648,
      "loss": 2.6162,
      "theoretical_loss": 3.4872812749359787,
      "tokens_seen": 1648033792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252888665997994,
      "loss": 2.5202,
      "theoretical_loss": 3.487269267922776,
      "tokens_seen": 1648099328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528786359077232,
      "loss": 2.7426,
      "theoretical_loss": 3.487257261520699,
      "tokens_seen": 1648164864
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 909310,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.450800657272339,
      "objective/train/theoretical_loss": 3.4872452557296922,
      "objective/train/tokens_used": 1668690400,
      "theoretical_loss": 3.4872452557296922,
      "tokens_seen": 1648230400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025286860581745236,
      "loss": 2.5775,
      "theoretical_loss": 3.4872452557296922,
      "tokens_seen": 1648230400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528585757271816,
      "loss": 2.4144,
      "theoretical_loss": 3.4872332505497,
      "tokens_seen": 1648295936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528485456369107,
      "loss": 2.7607,
      "theoretical_loss": 3.487221245980667,
      "tokens_seen": 1648361472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025283851554663996,
      "loss": 2.5854,
      "theoretical_loss": 3.4872092420225385,
      "tokens_seen": 1648427008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528284854563691,
      "loss": 2.5705,
      "theoretical_loss": 3.4871972386752583,
      "tokens_seen": 1648492544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528184553660983,
      "loss": 2.6417,
      "theoretical_loss": 3.4871852359387714,
      "tokens_seen": 1648558080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528084252758275,
      "loss": 2.72,
      "theoretical_loss": 3.4871732338130226,
      "tokens_seen": 1648623616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527983951855567,
      "loss": 2.6058,
      "theoretical_loss": 3.487161232297956,
      "tokens_seen": 1648689152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025278836509528586,
      "loss": 2.5838,
      "theoretical_loss": 3.4871492313935164,
      "tokens_seen": 1648754688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025277833500501505,
      "loss": 2.486,
      "theoretical_loss": 3.4871372310996493,
      "tokens_seen": 1648820224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527683049147442,
      "loss": 2.5225,
      "theoretical_loss": 3.4871252314162984,
      "tokens_seen": 1648885760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025275827482447346,
      "loss": 2.5518,
      "theoretical_loss": 3.487113232343409,
      "tokens_seen": 1648951296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527482447342026,
      "loss": 2.8039,
      "theoretical_loss": 3.4871012338809253,
      "tokens_seen": 1649016832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527382146439318,
      "loss": 2.3799,
      "theoretical_loss": 3.487089236028792,
      "tokens_seen": 1649082368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025272818455366095,
      "loss": 2.5081,
      "theoretical_loss": 3.487077238786954,
      "tokens_seen": 1649147904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527181544633902,
      "loss": 2.3675,
      "theoretical_loss": 3.4870652421553565,
      "tokens_seen": 1649213440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025270812437311937,
      "loss": 2.6286,
      "theoretical_loss": 3.487053246133943,
      "tokens_seen": 1649278976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025269809428284855,
      "loss": 2.4927,
      "theoretical_loss": 3.487041250722659,
      "tokens_seen": 1649344512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025268806419257773,
      "loss": 2.657,
      "theoretical_loss": 3.4870292559214495,
      "tokens_seen": 1649410048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025267803410230697,
      "loss": 2.6827,
      "theoretical_loss": 3.4870172617302586,
      "tokens_seen": 1649475584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526680040120361,
      "loss": 2.5999,
      "theoretical_loss": 3.487005268149032,
      "tokens_seen": 1649541120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025265797392176533,
      "loss": 2.6097,
      "theoretical_loss": 3.4869932751777126,
      "tokens_seen": 1649606656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025264794383149445,
      "loss": 2.7098,
      "theoretical_loss": 3.486981282816247,
      "tokens_seen": 1649672192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526379137412237,
      "loss": 2.716,
      "theoretical_loss": 3.4869692910645798,
      "tokens_seen": 1649737728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025262788365095287,
      "loss": 2.6937,
      "theoretical_loss": 3.486957299922654,
      "tokens_seen": 1649803264
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 910512,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7155091762542725,
      "objective/train/theoretical_loss": 3.486945309390417,
      "objective/train/tokens_used": 1670328800,
      "theoretical_loss": 3.486945309390417,
      "tokens_seen": 1649868800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025261785356068205,
      "loss": 2.5642,
      "theoretical_loss": 3.486945309390417,
      "tokens_seen": 1649868800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025260782347041123,
      "loss": 2.7517,
      "theoretical_loss": 3.4869333194678114,
      "tokens_seen": 1649934336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525977933801404,
      "loss": 2.4548,
      "theoretical_loss": 3.486921330154783,
      "tokens_seen": 1649999872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525877632898696,
      "loss": 2.6268,
      "theoretical_loss": 3.486909341451276,
      "tokens_seen": 1650065408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025257773319959883,
      "loss": 2.6349,
      "theoretical_loss": 3.4868973533572363,
      "tokens_seen": 1650130944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025256770310932796,
      "loss": 2.4948,
      "theoretical_loss": 3.486885365872608,
      "tokens_seen": 1650196480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525576730190572,
      "loss": 2.5847,
      "theoretical_loss": 3.4868733789973354,
      "tokens_seen": 1650262016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525476429287864,
      "loss": 2.7459,
      "theoretical_loss": 3.486861392731364,
      "tokens_seen": 1650327552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025253761283851556,
      "loss": 2.6879,
      "theoretical_loss": 3.4868494070746388,
      "tokens_seen": 1650393088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025252758274824474,
      "loss": 2.695,
      "theoretical_loss": 3.4868374220271043,
      "tokens_seen": 1650458624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525175526579739,
      "loss": 2.7382,
      "theoretical_loss": 3.4868254375887053,
      "tokens_seen": 1650524160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525075225677031,
      "loss": 2.479,
      "theoretical_loss": 3.4868134537593867,
      "tokens_seen": 1650589696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025249749247743233,
      "loss": 2.7831,
      "theoretical_loss": 3.4868014705390937,
      "tokens_seen": 1650655232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025248746238716146,
      "loss": 2.5838,
      "theoretical_loss": 3.4867894879277705,
      "tokens_seen": 1650720768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524774322968907,
      "loss": 2.579,
      "theoretical_loss": 3.4867775059253625,
      "tokens_seen": 1650786304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524674022066198,
      "loss": 2.5511,
      "theoretical_loss": 3.4867655245318145,
      "tokens_seen": 1650851840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025245737211634906,
      "loss": 2.771,
      "theoretical_loss": 3.4867535437470716,
      "tokens_seen": 1650917376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025244734202607824,
      "loss": 2.6646,
      "theoretical_loss": 3.486741563571078,
      "tokens_seen": 1650982912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524373119358074,
      "loss": 2.6803,
      "theoretical_loss": 3.486729584003779,
      "tokens_seen": 1651048448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524272818455366,
      "loss": 2.6433,
      "theoretical_loss": 3.4867176050451203,
      "tokens_seen": 1651113984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524172517552658,
      "loss": 2.8157,
      "theoretical_loss": 3.4867056266950454,
      "tokens_seen": 1651179520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025240722166499496,
      "loss": 2.6017,
      "theoretical_loss": 3.4866936489535,
      "tokens_seen": 1651245056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523971915747242,
      "loss": 2.6364,
      "theoretical_loss": 3.4866816718204294,
      "tokens_seen": 1651310592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523871614844533,
      "loss": 2.73,
      "theoretical_loss": 3.4866696952957774,
      "tokens_seen": 1651376128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025237713139418256,
      "loss": 2.7331,
      "theoretical_loss": 3.4866577193794903,
      "tokens_seen": 1651441664
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 911219,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.751664876937866,
      "objective/train/theoretical_loss": 3.486645744071512,
      "objective/train/tokens_used": 1671967200,
      "theoretical_loss": 3.486645744071512,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025236710130391174,
      "loss": 2.7249,
      "theoretical_loss": 3.486645744071512,
      "tokens_seen": 1651507200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523570712136409,
      "loss": 2.566,
      "theoretical_loss": 3.486633769371788,
      "tokens_seen": 1651572736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523470411233701,
      "loss": 2.7991,
      "theoretical_loss": 3.486621795280263,
      "tokens_seen": 1651638272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523370110330993,
      "loss": 2.6551,
      "theoretical_loss": 3.4866098217968826,
      "tokens_seen": 1651703808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025232698094282847,
      "loss": 2.6979,
      "theoretical_loss": 3.4865978489215905,
      "tokens_seen": 1651769344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523169508525577,
      "loss": 2.5031,
      "theoretical_loss": 3.486585876654333,
      "tokens_seen": 1651834880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025230692076228683,
      "loss": 2.6585,
      "theoretical_loss": 3.486573904995055,
      "tokens_seen": 1651900416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025229689067201606,
      "loss": 2.4826,
      "theoretical_loss": 3.4865619339437,
      "tokens_seen": 1651965952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522868605817452,
      "loss": 2.8046,
      "theoretical_loss": 3.486549963500215,
      "tokens_seen": 1652031488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025227683049147443,
      "loss": 2.6915,
      "theoretical_loss": 3.4865379936645438,
      "tokens_seen": 1652097024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522668004012036,
      "loss": 2.5948,
      "theoretical_loss": 3.486526024436632,
      "tokens_seen": 1652162560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522567703109328,
      "loss": 2.4244,
      "theoretical_loss": 3.486514055816424,
      "tokens_seen": 1652228096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252246740220662,
      "loss": 2.6683,
      "theoretical_loss": 3.4865020878038653,
      "tokens_seen": 1652293632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025223671013039115,
      "loss": 2.6691,
      "theoretical_loss": 3.4864901203989014,
      "tokens_seen": 1652359168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522266800401204,
      "loss": 2.3956,
      "theoretical_loss": 3.4864781536014764,
      "tokens_seen": 1652424704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025221664994984957,
      "loss": 2.6905,
      "theoretical_loss": 3.4864661874115357,
      "tokens_seen": 1652490240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025220661985957875,
      "loss": 2.8024,
      "theoretical_loss": 3.4864542218290246,
      "tokens_seen": 1652555776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025219658976930793,
      "loss": 2.6582,
      "theoretical_loss": 3.486442256853888,
      "tokens_seen": 1652621312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025218655967903717,
      "loss": 2.531,
      "theoretical_loss": 3.4864302924860713,
      "tokens_seen": 1652686848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521765295887663,
      "loss": 2.4745,
      "theoretical_loss": 3.4864183287255193,
      "tokens_seen": 1652752384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025216649949849553,
      "loss": 2.5237,
      "theoretical_loss": 3.486406365572177,
      "tokens_seen": 1652817920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025215646940822465,
      "loss": 2.4202,
      "theoretical_loss": 3.4863944030259897,
      "tokens_seen": 1652883456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521464393179539,
      "loss": 2.6353,
      "theoretical_loss": 3.486382441086903,
      "tokens_seen": 1652948992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025213640922768307,
      "loss": 2.7251,
      "theoretical_loss": 3.4863704797548607,
      "tokens_seen": 1653014528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025212637913741225,
      "loss": 2.3471,
      "theoretical_loss": 3.4863585190298085,
      "tokens_seen": 1653080064
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 912444,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3914291858673096,
      "objective/train/theoretical_loss": 3.4863465589116927,
      "objective/train/tokens_used": 1673605600,
      "theoretical_loss": 3.4863465589116927,
      "tokens_seen": 1653145600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025211634904714143,
      "loss": 2.5272,
      "theoretical_loss": 3.4863465589116927,
      "tokens_seen": 1653145600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521063189568706,
      "loss": 2.5748,
      "theoretical_loss": 3.4863345994004566,
      "tokens_seen": 1653211136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520962888665998,
      "loss": 2.7575,
      "theoretical_loss": 3.486322640496047,
      "tokens_seen": 1653276672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025208625877632903,
      "loss": 2.7139,
      "theoretical_loss": 3.4863106821984076,
      "tokens_seen": 1653342208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025207622868605816,
      "loss": 2.5779,
      "theoretical_loss": 3.486298724507485,
      "tokens_seen": 1653407744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520661985957874,
      "loss": 2.4411,
      "theoretical_loss": 3.4862867674232225,
      "tokens_seen": 1653473280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520561685055166,
      "loss": 2.3808,
      "theoretical_loss": 3.4862748109455675,
      "tokens_seen": 1653538816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025204613841524576,
      "loss": 2.7139,
      "theoretical_loss": 3.4862628550744637,
      "tokens_seen": 1653604352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025203610832497494,
      "loss": 2.5578,
      "theoretical_loss": 3.486250899809857,
      "tokens_seen": 1653669888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520260782347041,
      "loss": 2.4489,
      "theoretical_loss": 3.4862389451516913,
      "tokens_seen": 1653735424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520160481444333,
      "loss": 2.6081,
      "theoretical_loss": 3.4862269910999135,
      "tokens_seen": 1653800960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025200601805416253,
      "loss": 2.4329,
      "theoretical_loss": 3.4862150376544685,
      "tokens_seen": 1653866496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025199598796389166,
      "loss": 2.6048,
      "theoretical_loss": 3.4862030848153003,
      "tokens_seen": 1653932032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002519859578736209,
      "loss": 2.6631,
      "theoretical_loss": 3.4861911325823556,
      "tokens_seen": 1653997568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025197592778335,
      "loss": 2.5916,
      "theoretical_loss": 3.4861791809555784,
      "tokens_seen": 1654063104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025196589769307926,
      "loss": 2.5941,
      "theoretical_loss": 3.486167229934915,
      "tokens_seen": 1654128640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025195586760280844,
      "loss": 2.7256,
      "theoretical_loss": 3.48615527952031,
      "tokens_seen": 1654194176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002519458375125376,
      "loss": 2.5466,
      "theoretical_loss": 3.486143329711709,
      "tokens_seen": 1654259712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002519358074222668,
      "loss": 2.3998,
      "theoretical_loss": 3.486131380509057,
      "tokens_seen": 1654325248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251925777331996,
      "loss": 2.6465,
      "theoretical_loss": 3.4861194319122992,
      "tokens_seen": 1654390784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025191574724172516,
      "loss": 2.4198,
      "theoretical_loss": 3.4861074839213813,
      "tokens_seen": 1654456320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002519057171514544,
      "loss": 2.3688,
      "theoretical_loss": 3.486095536536248,
      "tokens_seen": 1654521856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518956870611835,
      "loss": 2.7545,
      "theoretical_loss": 3.4860835897568454,
      "tokens_seen": 1654587392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025188565697091276,
      "loss": 2.7359,
      "theoretical_loss": 3.4860716435831183,
      "tokens_seen": 1654652928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025187562688064194,
      "loss": 2.6156,
      "theoretical_loss": 3.4860596980150116,
      "tokens_seen": 1654718464
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 913730,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4969518184661865,
      "objective/train/theoretical_loss": 3.486047753052471,
      "objective/train/tokens_used": 1675244000,
      "theoretical_loss": 3.486047753052471,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518655967903711,
      "loss": 2.45,
      "theoretical_loss": 3.486047753052471,
      "tokens_seen": 1654784000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518555667001003,
      "loss": 2.7326,
      "theoretical_loss": 3.4860358086954424,
      "tokens_seen": 1654849536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518455366098295,
      "loss": 2.4308,
      "theoretical_loss": 3.48602386494387,
      "tokens_seen": 1654915072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025183550651955867,
      "loss": 2.6557,
      "theoretical_loss": 3.4860119217977004,
      "tokens_seen": 1654980608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518254764292879,
      "loss": 2.3096,
      "theoretical_loss": 3.485999979256878,
      "tokens_seen": 1655046144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025181544633901703,
      "loss": 2.4677,
      "theoretical_loss": 3.485988037321348,
      "tokens_seen": 1655111680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025180541624874627,
      "loss": 2.5031,
      "theoretical_loss": 3.485976095991057,
      "tokens_seen": 1655177216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517953861584754,
      "loss": 2.6673,
      "theoretical_loss": 3.4859641552659486,
      "tokens_seen": 1655242752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025178535606820463,
      "loss": 2.5583,
      "theoretical_loss": 3.4859522151459696,
      "tokens_seen": 1655308288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517753259779338,
      "loss": 2.7385,
      "theoretical_loss": 3.485940275631065,
      "tokens_seen": 1655373824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251765295887663,
      "loss": 2.4868,
      "theoretical_loss": 3.4859283367211797,
      "tokens_seen": 1655439360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025175526579739217,
      "loss": 2.812,
      "theoretical_loss": 3.4859163984162596,
      "tokens_seen": 1655504896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025174523570712135,
      "loss": 2.7364,
      "theoretical_loss": 3.4859044607162506,
      "tokens_seen": 1655570432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025173520561685053,
      "loss": 2.4899,
      "theoretical_loss": 3.4858925236210965,
      "tokens_seen": 1655635968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025172517552657977,
      "loss": 2.649,
      "theoretical_loss": 3.485880587130744,
      "tokens_seen": 1655701504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517151454363089,
      "loss": 2.4772,
      "theoretical_loss": 3.4858686512451387,
      "tokens_seen": 1655767040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025170511534603813,
      "loss": 2.6075,
      "theoretical_loss": 3.485856715964225,
      "tokens_seen": 1655832576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516950852557673,
      "loss": 2.5502,
      "theoretical_loss": 3.4858447812879487,
      "tokens_seen": 1655898112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516850551654965,
      "loss": 2.3978,
      "theoretical_loss": 3.4858328472162556,
      "tokens_seen": 1655963648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516750250752257,
      "loss": 2.604,
      "theoretical_loss": 3.485820913749091,
      "tokens_seen": 1656029184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025166499498495486,
      "loss": 2.6016,
      "theoretical_loss": 3.4858089808864006,
      "tokens_seen": 1656094720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025165496489468404,
      "loss": 2.7385,
      "theoretical_loss": 3.4857970486281293,
      "tokens_seen": 1656160256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025164493480441327,
      "loss": 2.471,
      "theoretical_loss": 3.485785116974223,
      "tokens_seen": 1656225792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516349047141424,
      "loss": 2.6099,
      "theoretical_loss": 3.485773185924627,
      "tokens_seen": 1656291328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025162487462387163,
      "loss": 2.4141,
      "theoretical_loss": 3.4857612554792863,
      "tokens_seen": 1656356864
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 914357,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.1104023456573486,
      "objective/train/theoretical_loss": 3.4857493256381473,
      "objective/train/tokens_used": 1676882400,
      "theoretical_loss": 3.4857493256381473,
      "tokens_seen": 1656422400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025161484453360076,
      "loss": 2.6233,
      "theoretical_loss": 3.4857493256381473,
      "tokens_seen": 1656422400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025160481444333,
      "loss": 2.6848,
      "theoretical_loss": 3.485737396401155,
      "tokens_seen": 1656487936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515947843530592,
      "loss": 2.5044,
      "theoretical_loss": 3.4857254677682548,
      "tokens_seen": 1656553472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025158475426278836,
      "loss": 2.3072,
      "theoretical_loss": 3.485713539739393,
      "tokens_seen": 1656619008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025157472417251754,
      "loss": 2.5887,
      "theoretical_loss": 3.485701612314514,
      "tokens_seen": 1656684544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515646940822468,
      "loss": 2.6096,
      "theoretical_loss": 3.485689685493564,
      "tokens_seen": 1656750080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515546639919759,
      "loss": 2.6905,
      "theoretical_loss": 3.4856777592764883,
      "tokens_seen": 1656815616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025154463390170514,
      "loss": 2.589,
      "theoretical_loss": 3.4856658336632327,
      "tokens_seen": 1656881152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025153460381143426,
      "loss": 2.5405,
      "theoretical_loss": 3.4856539086537426,
      "tokens_seen": 1656946688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515245737211635,
      "loss": 2.4748,
      "theoretical_loss": 3.4856419842479633,
      "tokens_seen": 1657012224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515145436308927,
      "loss": 2.745,
      "theoretical_loss": 3.485630060445841,
      "tokens_seen": 1657077760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025150451354062186,
      "loss": 2.3802,
      "theoretical_loss": 3.48561813724732,
      "tokens_seen": 1657143296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514944834503511,
      "loss": 2.4292,
      "theoretical_loss": 3.485606214652347,
      "tokens_seen": 1657208832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514844533600802,
      "loss": 2.4401,
      "theoretical_loss": 3.485594292660868,
      "tokens_seen": 1657274368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025147442326980946,
      "loss": 2.6734,
      "theoretical_loss": 3.4855823712728276,
      "tokens_seen": 1657339904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025146439317953864,
      "loss": 2.4615,
      "theoretical_loss": 3.4855704504881713,
      "tokens_seen": 1657405440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514543630892678,
      "loss": 2.5108,
      "theoretical_loss": 3.4855585303068453,
      "tokens_seen": 1657470976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251444332998997,
      "loss": 2.5116,
      "theoretical_loss": 3.485546610728795,
      "tokens_seen": 1657536512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514343029087262,
      "loss": 2.6149,
      "theoretical_loss": 3.4855346917539665,
      "tokens_seen": 1657602048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025142427281845536,
      "loss": 2.5842,
      "theoretical_loss": 3.4855227733823044,
      "tokens_seen": 1657667584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514142427281846,
      "loss": 2.4293,
      "theoretical_loss": 3.4855108556137546,
      "tokens_seen": 1657733120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514042126379137,
      "loss": 2.621,
      "theoretical_loss": 3.4854989384482638,
      "tokens_seen": 1657798656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025139418254764296,
      "loss": 2.5053,
      "theoretical_loss": 3.4854870218857763,
      "tokens_seen": 1657864192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025138415245737214,
      "loss": 2.4417,
      "theoretical_loss": 3.4854751059262385,
      "tokens_seen": 1657929728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513741223671013,
      "loss": 2.8554,
      "theoretical_loss": 3.485463190569596,
      "tokens_seen": 1657995264
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 915563,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.297820806503296,
      "objective/train/theoretical_loss": 3.4854512758157945,
      "objective/train/tokens_used": 1678520800,
      "theoretical_loss": 3.4854512758157945,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513640922768305,
      "loss": 2.4281,
      "theoretical_loss": 3.4854512758157945,
      "tokens_seen": 1658060800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513540621865597,
      "loss": 2.4768,
      "theoretical_loss": 3.485439361664779,
      "tokens_seen": 1658126336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025134403209628887,
      "loss": 2.6372,
      "theoretical_loss": 3.4854274481164964,
      "tokens_seen": 1658191872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513340020060181,
      "loss": 2.4912,
      "theoretical_loss": 3.4854155351708913,
      "tokens_seen": 1658257408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025132397191574723,
      "loss": 2.5079,
      "theoretical_loss": 3.4854036228279095,
      "tokens_seen": 1658322944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025131394182547647,
      "loss": 2.7283,
      "theoretical_loss": 3.4853917110874972,
      "tokens_seen": 1658388480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513039117352056,
      "loss": 2.582,
      "theoretical_loss": 3.4853797999496,
      "tokens_seen": 1658454016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025129388164493483,
      "loss": 2.5734,
      "theoretical_loss": 3.4853678894141638,
      "tokens_seen": 1658519552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251283851554664,
      "loss": 2.6319,
      "theoretical_loss": 3.4853559794811337,
      "tokens_seen": 1658585088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512738214643932,
      "loss": 2.5871,
      "theoretical_loss": 3.4853440701504557,
      "tokens_seen": 1658650624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025126379137412237,
      "loss": 2.7078,
      "theoretical_loss": 3.4853321614220754,
      "tokens_seen": 1658716160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025125376128385155,
      "loss": 2.5763,
      "theoretical_loss": 3.4853202532959395,
      "tokens_seen": 1658781696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025124373119358073,
      "loss": 2.5561,
      "theoretical_loss": 3.4853083457719922,
      "tokens_seen": 1658847232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025123370110330997,
      "loss": 2.8845,
      "theoretical_loss": 3.4852964388501806,
      "tokens_seen": 1658912768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512236710130391,
      "loss": 2.8421,
      "theoretical_loss": 3.48528453253045,
      "tokens_seen": 1658978304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025121364092276833,
      "loss": 2.6076,
      "theoretical_loss": 3.485272626812746,
      "tokens_seen": 1659043840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512036108324975,
      "loss": 2.6308,
      "theoretical_loss": 3.4852607216970144,
      "tokens_seen": 1659109376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511935807422267,
      "loss": 2.5668,
      "theoretical_loss": 3.4852488171832006,
      "tokens_seen": 1659174912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511835506519559,
      "loss": 2.5898,
      "theoretical_loss": 3.4852369132712515,
      "tokens_seen": 1659240448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025117352056168506,
      "loss": 2.6859,
      "theoretical_loss": 3.4852250099611117,
      "tokens_seen": 1659305984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025116349047141424,
      "loss": 2.5701,
      "theoretical_loss": 3.4852131072527284,
      "tokens_seen": 1659371520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025115346038114347,
      "loss": 2.6191,
      "theoretical_loss": 3.485201205146046,
      "tokens_seen": 1659437056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511434302908726,
      "loss": 2.6428,
      "theoretical_loss": 3.485189303641011,
      "tokens_seen": 1659502592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025113340020060183,
      "loss": 2.2581,
      "theoretical_loss": 3.4851774027375693,
      "tokens_seen": 1659568128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025112337011033096,
      "loss": 2.6652,
      "theoretical_loss": 3.4851655024356667,
      "tokens_seen": 1659633664
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 916180,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4970815181732178,
      "objective/train/theoretical_loss": 3.4851536027352483,
      "objective/train/tokens_used": 1680159200,
      "theoretical_loss": 3.4851536027352483,
      "tokens_seen": 1659699200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511133400200602,
      "loss": 2.5663,
      "theoretical_loss": 3.4851536027352483,
      "tokens_seen": 1659699200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511033099297894,
      "loss": 2.5783,
      "theoretical_loss": 3.485141703636261,
      "tokens_seen": 1659764736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025109327983951856,
      "loss": 2.7343,
      "theoretical_loss": 3.48512980513865,
      "tokens_seen": 1659830272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025108324974924774,
      "loss": 2.5583,
      "theoretical_loss": 3.485117907242362,
      "tokens_seen": 1659895808
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251073219658977,
      "loss": 2.3295,
      "theoretical_loss": 3.4851060099473417,
      "tokens_seen": 1659961344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510631895687061,
      "loss": 2.6279,
      "theoretical_loss": 3.4850941132535356,
      "tokens_seen": 1660026880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025105315947843534,
      "loss": 2.5303,
      "theoretical_loss": 3.4850822171608895,
      "tokens_seen": 1660092416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025104312938816446,
      "loss": 2.6609,
      "theoretical_loss": 3.4850703216693493,
      "tokens_seen": 1660157952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510330992978937,
      "loss": 2.5563,
      "theoretical_loss": 3.4850584267788607,
      "tokens_seen": 1660223488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510230692076229,
      "loss": 2.4105,
      "theoretical_loss": 3.4850465324893705,
      "tokens_seen": 1660289024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025101303911735206,
      "loss": 2.578,
      "theoretical_loss": 3.4850346388008235,
      "tokens_seen": 1660354560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025100300902708124,
      "loss": 2.3747,
      "theoretical_loss": 3.4850227457131657,
      "tokens_seen": 1660420096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509929789368104,
      "loss": 2.7365,
      "theoretical_loss": 3.4850108532263437,
      "tokens_seen": 1660485632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509829488465396,
      "loss": 2.41,
      "theoretical_loss": 3.4849989613403034,
      "tokens_seen": 1660551168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025097291875626884,
      "loss": 2.3874,
      "theoretical_loss": 3.48498707005499,
      "tokens_seen": 1660616704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025096288866599797,
      "loss": 2.6576,
      "theoretical_loss": 3.48497517937035,
      "tokens_seen": 1660682240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509528585757272,
      "loss": 2.4309,
      "theoretical_loss": 3.4849632892863296,
      "tokens_seen": 1660747776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025094282848545633,
      "loss": 2.5207,
      "theoretical_loss": 3.4849513998028745,
      "tokens_seen": 1660813312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025093279839518556,
      "loss": 2.1752,
      "theoretical_loss": 3.48493951091993,
      "tokens_seen": 1660878848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025092276830491475,
      "loss": 2.641,
      "theoretical_loss": 3.484927622637443,
      "tokens_seen": 1660944384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509127382146439,
      "loss": 2.401,
      "theoretical_loss": 3.4849157349553592,
      "tokens_seen": 1661009920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509027081243731,
      "loss": 2.7633,
      "theoretical_loss": 3.4849038478736243,
      "tokens_seen": 1661075456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025089267803410234,
      "loss": 2.5639,
      "theoretical_loss": 3.4848919613921847,
      "tokens_seen": 1661140992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025088264794383147,
      "loss": 2.6735,
      "theoretical_loss": 3.4848800755109863,
      "tokens_seen": 1661206528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508726178535607,
      "loss": 2.709,
      "theoretical_loss": 3.484868190229975,
      "tokens_seen": 1661272064
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 917442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8525569438934326,
      "objective/train/theoretical_loss": 3.484856305549097,
      "objective/train/tokens_used": 1681797600,
      "theoretical_loss": 3.484856305549097,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025086258776328983,
      "loss": 2.5681,
      "theoretical_loss": 3.484856305549097,
      "tokens_seen": 1661337600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025085255767301907,
      "loss": 2.4369,
      "theoretical_loss": 3.484844421468298,
      "tokens_seen": 1661403136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025084252758274825,
      "loss": 2.6314,
      "theoretical_loss": 3.484832537987524,
      "tokens_seen": 1661468672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025083249749247743,
      "loss": 2.6084,
      "theoretical_loss": 3.4848206551067213,
      "tokens_seen": 1661534208
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508224674022066,
      "loss": 2.5684,
      "theoretical_loss": 3.4848087728258363,
      "tokens_seen": 1661599744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508124373119358,
      "loss": 2.5325,
      "theoretical_loss": 3.484796891144814,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250802407221665,
      "loss": 2.3666,
      "theoretical_loss": 3.484785010063602,
      "tokens_seen": 1661730816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507923771313942,
      "loss": 2.6535,
      "theoretical_loss": 3.4847731295821447,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025078234704112334,
      "loss": 2.5107,
      "theoretical_loss": 3.48476124970039,
      "tokens_seen": 1661861888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025077231695085257,
      "loss": 2.6435,
      "theoretical_loss": 3.4847493704182817,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507622868605817,
      "loss": 2.6296,
      "theoretical_loss": 3.4847374917357676,
      "tokens_seen": 1661992960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025075225677031093,
      "loss": 2.5506,
      "theoretical_loss": 3.4847256136527935,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025074222668004017,
      "loss": 2.5814,
      "theoretical_loss": 3.4847137361693052,
      "tokens_seen": 1662124032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507321965897693,
      "loss": 2.4772,
      "theoretical_loss": 3.4847018592852494,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025072216649949853,
      "loss": 2.4719,
      "theoretical_loss": 3.484689983000571,
      "tokens_seen": 1662255104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507121364092277,
      "loss": 2.5822,
      "theoretical_loss": 3.4846781073152173,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507021063189569,
      "loss": 2.755,
      "theoretical_loss": 3.484666232229134,
      "tokens_seen": 1662386176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506920762286861,
      "loss": 2.8523,
      "theoretical_loss": 3.484654357742267,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025068204613841526,
      "loss": 2.6019,
      "theoretical_loss": 3.4846424838545627,
      "tokens_seen": 1662517248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025067201604814444,
      "loss": 2.6645,
      "theoretical_loss": 3.4846306105659677,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025066198595787367,
      "loss": 2.6919,
      "theoretical_loss": 3.484618737876427,
      "tokens_seen": 1662648320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506519558676028,
      "loss": 2.5578,
      "theoretical_loss": 3.4846068657858877,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025064192577733203,
      "loss": 2.5369,
      "theoretical_loss": 3.484594994294296,
      "tokens_seen": 1662779392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025063189568706116,
      "loss": 2.6011,
      "theoretical_loss": 3.4845831234015976,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506218655967904,
      "loss": 2.3343,
      "theoretical_loss": 3.484571253107739,
      "tokens_seen": 1662910464
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 918485,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5903053283691406,
      "objective/train/theoretical_loss": 3.484559383412666,
      "objective/train/tokens_used": 1683436000,
      "theoretical_loss": 3.484559383412666,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506118355065196,
      "loss": 2.584,
      "theoretical_loss": 3.484559383412666,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025060180541624876,
      "loss": 2.5413,
      "theoretical_loss": 3.4845475143163256,
      "tokens_seen": 1663041536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025059177532597794,
      "loss": 2.3681,
      "theoretical_loss": 3.4845356458186627,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505817452357072,
      "loss": 2.6365,
      "theoretical_loss": 3.484523777919625,
      "tokens_seen": 1663172608
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505717151454363,
      "loss": 2.4343,
      "theoretical_loss": 3.4845119106191578,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025056168505516554,
      "loss": 2.4353,
      "theoretical_loss": 3.4845000439172074,
      "tokens_seen": 1663303680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025055165496489466,
      "loss": 2.4563,
      "theoretical_loss": 3.48448817781372,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505416248746239,
      "loss": 2.508,
      "theoretical_loss": 3.484476312308643,
      "tokens_seen": 1663434752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505315947843531,
      "loss": 2.6031,
      "theoretical_loss": 3.484464447401921,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025052156469408226,
      "loss": 2.5689,
      "theoretical_loss": 3.4844525830935007,
      "tokens_seen": 1663565824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025051153460381144,
      "loss": 2.7283,
      "theoretical_loss": 3.4844407193833282,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505015045135406,
      "loss": 2.5067,
      "theoretical_loss": 3.484428856271351,
      "tokens_seen": 1663696896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504914744232698,
      "loss": 2.6455,
      "theoretical_loss": 3.484416993757514,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025048144433299904,
      "loss": 2.624,
      "theoretical_loss": 3.484405131841765,
      "tokens_seen": 1663827968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025047141424272817,
      "loss": 2.4699,
      "theoretical_loss": 3.4843932705240483,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504613841524574,
      "loss": 2.4598,
      "theoretical_loss": 3.4843814098043113,
      "tokens_seen": 1663959040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025045135406218653,
      "loss": 2.6171,
      "theoretical_loss": 3.4843695496825005,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025044132397191576,
      "loss": 2.639,
      "theoretical_loss": 3.4843576901585616,
      "tokens_seen": 1664090112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025043129388164495,
      "loss": 2.5066,
      "theoretical_loss": 3.4843458312324413,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504212637913741,
      "loss": 2.7766,
      "theoretical_loss": 3.484333972904086,
      "tokens_seen": 1664221184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504112337011033,
      "loss": 2.481,
      "theoretical_loss": 3.484322115173442,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025040120361083254,
      "loss": 2.7617,
      "theoretical_loss": 3.4843102580404555,
      "tokens_seen": 1664352256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025039117352056167,
      "loss": 2.8213,
      "theoretical_loss": 3.4842984015050726,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503811434302909,
      "loss": 2.4564,
      "theoretical_loss": 3.4842865455672403,
      "tokens_seen": 1664483328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025037111334002003,
      "loss": 2.6656,
      "theoretical_loss": 3.484274690226904,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 919051,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.940255641937256,
      "objective/train/theoretical_loss": 3.484262835484011,
      "objective/train/tokens_used": 1685074400,
      "theoretical_loss": 3.484262835484011,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025036108324974927,
      "loss": 2.6709,
      "theoretical_loss": 3.484262835484011,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025035105315947845,
      "loss": 2.5675,
      "theoretical_loss": 3.484250981338507,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025034102306920763,
      "loss": 2.504,
      "theoretical_loss": 3.484239127790339,
      "tokens_seen": 1664745472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503309929789368,
      "loss": 2.6625,
      "theoretical_loss": 3.484227274839453,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250320962888666,
      "loss": 2.3716,
      "theoretical_loss": 3.4842154224857955,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503109327983952,
      "loss": 2.5582,
      "theoretical_loss": 3.484203570729313,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503009027081244,
      "loss": 2.5554,
      "theoretical_loss": 3.4841917195699517,
      "tokens_seen": 1665007616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025029087261785354,
      "loss": 2.4618,
      "theoretical_loss": 3.484179869007658,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025028084252758277,
      "loss": 2.4211,
      "theoretical_loss": 3.4841680190423787,
      "tokens_seen": 1665138688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502708124373119,
      "loss": 2.4952,
      "theoretical_loss": 3.4841561696740597,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025026078234704113,
      "loss": 2.7182,
      "theoretical_loss": 3.4841443209026477,
      "tokens_seen": 1665269760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502507522567703,
      "loss": 2.5909,
      "theoretical_loss": 3.4841324727280893,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502407221664995,
      "loss": 2.6523,
      "theoretical_loss": 3.4841206251503305,
      "tokens_seen": 1665400832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502306920762287,
      "loss": 2.7016,
      "theoretical_loss": 3.484108778169318,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502206619859579,
      "loss": 2.5623,
      "theoretical_loss": 3.4840969317849986,
      "tokens_seen": 1665531904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025021063189568704,
      "loss": 2.4777,
      "theoretical_loss": 3.484085085997318,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502006018054163,
      "loss": 2.6754,
      "theoretical_loss": 3.4840732408062234,
      "tokens_seen": 1665662976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501905717151454,
      "loss": 2.7799,
      "theoretical_loss": 3.484061396211661,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025018054162487464,
      "loss": 2.595,
      "theoretical_loss": 3.4840495522135777,
      "tokens_seen": 1665794048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501705115346038,
      "loss": 2.621,
      "theoretical_loss": 3.4840377088119188,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250160481444333,
      "loss": 2.7959,
      "theoretical_loss": 3.484025866006632,
      "tokens_seen": 1665925120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501504513540622,
      "loss": 2.7575,
      "theoretical_loss": 3.4840140237976636,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025014042126379136,
      "loss": 2.6141,
      "theoretical_loss": 3.48400218218496,
      "tokens_seen": 1666056192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025013039117352054,
      "loss": 2.7454,
      "theoretical_loss": 3.483990341168467,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501203610832498,
      "loss": 2.656,
      "theoretical_loss": 3.4839785007481323,
      "tokens_seen": 1666187264
    },
    {
      "epoch": 0.5,
      "objective/train/docs_used": 920018,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5550425052642822,
      "objective/train/theoretical_loss": 3.483966660923902,
      "objective/train/tokens_used": 1686712800,
      "theoretical_loss": 3.483966660923902,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501103309929789,
      "loss": 2.5749,
      "theoretical_loss": 3.483966660923902,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025010030090270814,
      "loss": 2.5773,
      "theoretical_loss": 3.483954821695723,
      "tokens_seen": 1666318336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025009027081243727,
      "loss": 2.8129,
      "theoretical_loss": 3.4839429830635407,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002500802407221665,
      "loss": 2.6336,
      "theoretical_loss": 3.4839311450273023,
      "tokens_seen": 1666449408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002500702106318957,
      "loss": 2.6115,
      "theoretical_loss": 3.4839193075869543,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025006018054162486,
      "loss": 2.6255,
      "theoretical_loss": 3.4839074707424444,
      "tokens_seen": 1666580480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025005015045135405,
      "loss": 2.6409,
      "theoretical_loss": 3.4838956344937175,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002500401203610833,
      "loss": 2.476,
      "theoretical_loss": 3.483883798840721,
      "tokens_seen": 1666711552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002500300902708124,
      "loss": 2.7829,
      "theoretical_loss": 3.4838719637834012,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025002006018054164,
      "loss": 2.5599,
      "theoretical_loss": 3.4838601293217053,
      "tokens_seen": 1666842624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025001003009027077,
      "loss": 2.6162,
      "theoretical_loss": 3.483848295455579,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025,
      "loss": 2.6758,
      "theoretical_loss": 3.4838364621849705,
      "tokens_seen": 1666973696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002499899699097292,
      "loss": 2.6641,
      "theoretical_loss": 3.483824629509824,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024997993981945837,
      "loss": 2.7141,
      "theoretical_loss": 3.483812797430088,
      "tokens_seen": 1667104768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024996990972918755,
      "loss": 2.6457,
      "theoretical_loss": 3.4838009659457088,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024995987963891673,
      "loss": 2.8476,
      "theoretical_loss": 3.4837891350566323,
      "tokens_seen": 1667235840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024994984954864597,
      "loss": 2.6696,
      "theoretical_loss": 3.4837773047628064,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024993981945837515,
      "loss": 2.5558,
      "theoretical_loss": 3.4837654750641764,
      "tokens_seen": 1667366912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024992978936810433,
      "loss": 2.3257,
      "theoretical_loss": 3.4837536459606904,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002499197592778335,
      "loss": 2.6005,
      "theoretical_loss": 3.4837418174522936,
      "tokens_seen": 1667497984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002499097291875627,
      "loss": 2.6535,
      "theoretical_loss": 3.4837299895389333,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024989969909729187,
      "loss": 2.634,
      "theoretical_loss": 3.4837181622205566,
      "tokens_seen": 1667629056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024988966900702105,
      "loss": 2.8475,
      "theoretical_loss": 3.48370633549711,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024987963891675023,
      "loss": 2.4116,
      "theoretical_loss": 3.483694509368539,
      "tokens_seen": 1667760128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498696088264794,
      "loss": 2.6517,
      "theoretical_loss": 3.4836826838347923,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 920691,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.782200574874878,
      "objective/train/theoretical_loss": 3.483670858895815,
      "objective/train/tokens_used": 1688351200,
      "theoretical_loss": 3.483670858895815,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024985957873620865,
      "loss": 2.7599,
      "theoretical_loss": 3.483670858895815,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024984954864593783,
      "loss": 2.4435,
      "theoretical_loss": 3.483659034551555,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249839518555667,
      "loss": 2.601,
      "theoretical_loss": 3.4836472108019585,
      "tokens_seen": 1668022272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498294884653962,
      "loss": 2.6008,
      "theoretical_loss": 3.483635387646972,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498194583751254,
      "loss": 2.5298,
      "theoretical_loss": 3.483623565086542,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024980942828485455,
      "loss": 2.7605,
      "theoretical_loss": 3.4836117431206164,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024979939819458374,
      "loss": 2.5033,
      "theoretical_loss": 3.483599921749141,
      "tokens_seen": 1668284416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497893681043129,
      "loss": 2.4498,
      "theoretical_loss": 3.483588100972063,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497793380140421,
      "loss": 2.4715,
      "theoretical_loss": 3.483576280789328,
      "tokens_seen": 1668415488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024976930792377133,
      "loss": 2.6093,
      "theoretical_loss": 3.483564461200885,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497592778335005,
      "loss": 2.6071,
      "theoretical_loss": 3.4835526422066785,
      "tokens_seen": 1668546560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497492477432297,
      "loss": 2.6817,
      "theoretical_loss": 3.4835408238066567,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497392176529589,
      "loss": 2.533,
      "theoretical_loss": 3.483529006000766,
      "tokens_seen": 1668677632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024972918756268806,
      "loss": 2.7993,
      "theoretical_loss": 3.4835171887889533,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024971915747241724,
      "loss": 2.7256,
      "theoretical_loss": 3.483505372171165,
      "tokens_seen": 1668808704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497091273821464,
      "loss": 2.8359,
      "theoretical_loss": 3.483493556147349,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496990972918756,
      "loss": 2.5652,
      "theoretical_loss": 3.4834817407174503,
      "tokens_seen": 1668939776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496890672016048,
      "loss": 2.6362,
      "theoretical_loss": 3.4834699258814172,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249679037111334,
      "loss": 2.4392,
      "theoretical_loss": 3.4834581116391963,
      "tokens_seen": 1669070848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496690070210632,
      "loss": 2.7295,
      "theoretical_loss": 3.483446297990734,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496589769307924,
      "loss": 2.5246,
      "theoretical_loss": 3.4834344849359775,
      "tokens_seen": 1669201920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024964894684052156,
      "loss": 2.535,
      "theoretical_loss": 3.4834226724748736,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024963891675025074,
      "loss": 2.6274,
      "theoretical_loss": 3.483410860607369,
      "tokens_seen": 1669332992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496288866599799,
      "loss": 2.5104,
      "theoretical_loss": 3.4833990493334106,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024961885656970916,
      "loss": 2.4023,
      "theoretical_loss": 3.4833872386529454,
      "tokens_seen": 1669464064
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 921959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.789898157119751,
      "objective/train/theoretical_loss": 3.48337542856592,
      "objective/train/tokens_used": 1689989600,
      "theoretical_loss": 3.48337542856592,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024960882647943834,
      "loss": 2.6686,
      "theoretical_loss": 3.48337542856592,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495987963891675,
      "loss": 2.4842,
      "theoretical_loss": 3.483363619072282,
      "tokens_seen": 1669595136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495887662988967,
      "loss": 2.4248,
      "theoretical_loss": 3.4833518101719774,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495787362086259,
      "loss": 2.5858,
      "theoretical_loss": 3.4833400018649536,
      "tokens_seen": 1669726208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024956870611835506,
      "loss": 2.5778,
      "theoretical_loss": 3.4833281941511576,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024955867602808425,
      "loss": 2.7988,
      "theoretical_loss": 3.483316387030536,
      "tokens_seen": 1669857280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495486459378135,
      "loss": 2.6397,
      "theoretical_loss": 3.4833045805030363,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024953861584754266,
      "loss": 2.7531,
      "theoretical_loss": 3.4832927745686044,
      "tokens_seen": 1669988352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024952858575727184,
      "loss": 2.4388,
      "theoretical_loss": 3.483280969227188,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249518555667001,
      "loss": 2.5885,
      "theoretical_loss": 3.4832691644787337,
      "tokens_seen": 1670119424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495085255767302,
      "loss": 2.7178,
      "theoretical_loss": 3.4832573603231887,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494984954864594,
      "loss": 2.5623,
      "theoretical_loss": 3.4832455567605,
      "tokens_seen": 1670250496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024948846539618857,
      "loss": 2.7961,
      "theoretical_loss": 3.4832337537906146,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024947843530591775,
      "loss": 2.6129,
      "theoretical_loss": 3.483221951413479,
      "tokens_seen": 1670381568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024946840521564693,
      "loss": 2.5338,
      "theoretical_loss": 3.4832101496290404,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024945837512537617,
      "loss": 2.7087,
      "theoretical_loss": 3.483198348437246,
      "tokens_seen": 1670512640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024944834503510535,
      "loss": 2.4963,
      "theoretical_loss": 3.4831865478380433,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024943831494483453,
      "loss": 2.5697,
      "theoretical_loss": 3.483174747831378,
      "tokens_seen": 1670643712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494282848545637,
      "loss": 2.4321,
      "theoretical_loss": 3.4831629484171978,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494182547642929,
      "loss": 2.5954,
      "theoretical_loss": 3.4831511495954497,
      "tokens_seen": 1670774784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024940822467402207,
      "loss": 2.6167,
      "theoretical_loss": 3.4831393513660807,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024939819458375125,
      "loss": 2.4674,
      "theoretical_loss": 3.483127553729038,
      "tokens_seen": 1670905856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024938816449348043,
      "loss": 2.5128,
      "theoretical_loss": 3.483115756684268,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493781344032096,
      "loss": 2.6061,
      "theoretical_loss": 3.483103960231719,
      "tokens_seen": 1671036928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024936810431293885,
      "loss": 2.448,
      "theoretical_loss": 3.4830921643713366,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 922543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3563973903656006,
      "objective/train/theoretical_loss": 3.4830803691030683,
      "objective/train/tokens_used": 1691628000,
      "theoretical_loss": 3.4830803691030683,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024935807422266803,
      "loss": 2.4339,
      "theoretical_loss": 3.4830803691030683,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493480441323972,
      "loss": 2.6399,
      "theoretical_loss": 3.483068574426862,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493380140421264,
      "loss": 2.5725,
      "theoretical_loss": 3.4830567803426637,
      "tokens_seen": 1671299072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493279839518556,
      "loss": 2.5871,
      "theoretical_loss": 3.483044986850421,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024931795386158476,
      "loss": 2.685,
      "theoretical_loss": 3.4830331939500807,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024930792377131394,
      "loss": 2.6343,
      "theoretical_loss": 3.48302140164159,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492978936810431,
      "loss": 2.6221,
      "theoretical_loss": 3.483009609924897,
      "tokens_seen": 1671561216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492878635907723,
      "loss": 2.5334,
      "theoretical_loss": 3.482997818799947,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024927783350050153,
      "loss": 2.65,
      "theoretical_loss": 3.482986028266688,
      "tokens_seen": 1671692288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492678034102307,
      "loss": 2.4353,
      "theoretical_loss": 3.4829742383250673,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492577733199599,
      "loss": 2.5522,
      "theoretical_loss": 3.4829624489750315,
      "tokens_seen": 1671823360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492477432296891,
      "loss": 2.3406,
      "theoretical_loss": 3.4829506602165283,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024923771313941826,
      "loss": 2.6512,
      "theoretical_loss": 3.4829388720495045,
      "tokens_seen": 1671954432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024922768304914744,
      "loss": 2.82,
      "theoretical_loss": 3.482927084473907,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492176529588766,
      "loss": 2.575,
      "theoretical_loss": 3.4829152974896838,
      "tokens_seen": 1672085504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492076228686058,
      "loss": 2.5564,
      "theoretical_loss": 3.482903511096781,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249197592778335,
      "loss": 2.5607,
      "theoretical_loss": 3.482891725295146,
      "tokens_seen": 1672216576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491875626880642,
      "loss": 2.6501,
      "theoretical_loss": 3.482879940084727,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491775325977934,
      "loss": 2.6407,
      "theoretical_loss": 3.48286815546547,
      "tokens_seen": 1672347648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491675025075226,
      "loss": 2.7688,
      "theoretical_loss": 3.4828563714373226,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024915747241725176,
      "loss": 2.6281,
      "theoretical_loss": 3.482844588000232,
      "tokens_seen": 1672478720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024914744232698094,
      "loss": 2.5557,
      "theoretical_loss": 3.4828328051541453,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491374122367101,
      "loss": 2.4853,
      "theoretical_loss": 3.48282102289901,
      "tokens_seen": 1672609792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491273821464393,
      "loss": 2.7953,
      "theoretical_loss": 3.482809241234773,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491173520561685,
      "loss": 3.0082,
      "theoretical_loss": 3.4827974601613816,
      "tokens_seen": 1672740864
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 923228,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.24631667137146,
      "objective/train/theoretical_loss": 3.482785679678783,
      "objective/train/tokens_used": 1693266400,
      "theoretical_loss": 3.482785679678783,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024910732196589767,
      "loss": 2.4982,
      "theoretical_loss": 3.482785679678783,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490972918756269,
      "loss": 2.6848,
      "theoretical_loss": 3.482773899786924,
      "tokens_seen": 1672871936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490872617853561,
      "loss": 2.6543,
      "theoretical_loss": 3.482762120485753,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024907723169508526,
      "loss": 2.5691,
      "theoretical_loss": 3.4827503417752164,
      "tokens_seen": 1673003008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024906720160481445,
      "loss": 2.7772,
      "theoretical_loss": 3.482738563655261,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490571715145436,
      "loss": 2.3038,
      "theoretical_loss": 3.482726786125835,
      "tokens_seen": 1673134080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490471414242728,
      "loss": 2.7516,
      "theoretical_loss": 3.4827150091868853,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249037111334002,
      "loss": 2.5809,
      "theoretical_loss": 3.482703232838359,
      "tokens_seen": 1673265152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024902708124373117,
      "loss": 2.5476,
      "theoretical_loss": 3.482691457080204,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024901705115346035,
      "loss": 2.4147,
      "theoretical_loss": 3.4826796819123667,
      "tokens_seen": 1673396224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490070210631896,
      "loss": 2.5065,
      "theoretical_loss": 3.482667907334795,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024899699097291877,
      "loss": 2.4006,
      "theoretical_loss": 3.4826561333474357,
      "tokens_seen": 1673527296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024898696088264795,
      "loss": 2.8972,
      "theoretical_loss": 3.482644359950237,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024897693079237713,
      "loss": 2.664,
      "theoretical_loss": 3.482632587143145,
      "tokens_seen": 1673658368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489669007021063,
      "loss": 2.3283,
      "theoretical_loss": 3.4826208149261078,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489568706118355,
      "loss": 2.5005,
      "theoretical_loss": 3.482609043299073,
      "tokens_seen": 1673789440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489468405215647,
      "loss": 2.5503,
      "theoretical_loss": 3.482597272261987,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024893681043129385,
      "loss": 2.6824,
      "theoretical_loss": 3.482585501814798,
      "tokens_seen": 1673920512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024892678034102304,
      "loss": 2.5329,
      "theoretical_loss": 3.4825737319574523,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024891675025075227,
      "loss": 2.4711,
      "theoretical_loss": 3.482561962689899,
      "tokens_seen": 1674051584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024890672016048145,
      "loss": 2.6805,
      "theoretical_loss": 3.4825501940120835,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024889669007021063,
      "loss": 2.6404,
      "theoretical_loss": 3.4825384259239547,
      "tokens_seen": 1674182656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488866599799398,
      "loss": 2.7344,
      "theoretical_loss": 3.4825266584254586,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248876629889669,
      "loss": 2.6601,
      "theoretical_loss": 3.482514891516544,
      "tokens_seen": 1674313728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024886659979939823,
      "loss": 2.522,
      "theoretical_loss": 3.482503125197157,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 924425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5553576946258545,
      "objective/train/theoretical_loss": 3.4824913594672458,
      "objective/train/tokens_used": 1694904800,
      "theoretical_loss": 3.4824913594672458,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488565697091274,
      "loss": 2.4835,
      "theoretical_loss": 3.4824913594672458,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488465396188566,
      "loss": 2.5052,
      "theoretical_loss": 3.4824795943267577,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488365095285858,
      "loss": 2.4857,
      "theoretical_loss": 3.48246782977564,
      "tokens_seen": 1674575872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024882647943831496,
      "loss": 2.4263,
      "theoretical_loss": 3.4824560658138397,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024881644934804414,
      "loss": 2.6956,
      "theoretical_loss": 3.482444302441305,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488064192577733,
      "loss": 2.5573,
      "theoretical_loss": 3.482432539657983,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487963891675025,
      "loss": 2.6223,
      "theoretical_loss": 3.4824207774638207,
      "tokens_seen": 1674838016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024878635907723173,
      "loss": 2.5977,
      "theoretical_loss": 3.4824090158587664,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487763289869609,
      "loss": 2.6278,
      "theoretical_loss": 3.482397254842767,
      "tokens_seen": 1674969088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487662988966901,
      "loss": 2.6206,
      "theoretical_loss": 3.4823854944157695,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487562688064193,
      "loss": 2.7115,
      "theoretical_loss": 3.482373734577722,
      "tokens_seen": 1675100160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024874623871614846,
      "loss": 2.5594,
      "theoretical_loss": 3.482361975328572,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024873620862587764,
      "loss": 2.6591,
      "theoretical_loss": 3.4823502166682667,
      "tokens_seen": 1675231232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487261785356068,
      "loss": 2.5634,
      "theoretical_loss": 3.482338458596754,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248716148445336,
      "loss": 2.5773,
      "theoretical_loss": 3.4823267011139807,
      "tokens_seen": 1675362304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487061183550652,
      "loss": 2.8888,
      "theoretical_loss": 3.482314944219895,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486960882647944,
      "loss": 2.7223,
      "theoretical_loss": 3.482303187914444,
      "tokens_seen": 1675493376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486860581745236,
      "loss": 2.8306,
      "theoretical_loss": 3.482291432197575,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486760280842528,
      "loss": 2.5997,
      "theoretical_loss": 3.4822796770692364,
      "tokens_seen": 1675624448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024866599799398196,
      "loss": 2.5176,
      "theoretical_loss": 3.482267922529374,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024865596790371114,
      "loss": 2.4726,
      "theoretical_loss": 3.4822561685779374,
      "tokens_seen": 1675755520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486459378134403,
      "loss": 2.7163,
      "theoretical_loss": 3.482244415214873,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486359077231695,
      "loss": 2.5854,
      "theoretical_loss": 3.482232662440128,
      "tokens_seen": 1675886592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486258776328987,
      "loss": 2.842,
      "theoretical_loss": 3.482220910253651,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024861584754262787,
      "loss": 2.459,
      "theoretical_loss": 3.4822091586553885,
      "tokens_seen": 1676017664
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 924948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4817583560943604,
      "objective/train/theoretical_loss": 3.482197407645289,
      "objective/train/tokens_used": 1696543200,
      "theoretical_loss": 3.482197407645289,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486058174523571,
      "loss": 2.7214,
      "theoretical_loss": 3.482197407645289,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485957873620863,
      "loss": 2.4936,
      "theoretical_loss": 3.482185657223299,
      "tokens_seen": 1676148736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024858575727181546,
      "loss": 2.5992,
      "theoretical_loss": 3.4821739073893676,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024857572718154465,
      "loss": 2.4888,
      "theoretical_loss": 3.482162158143441,
      "tokens_seen": 1676279808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485656970912738,
      "loss": 2.4368,
      "theoretical_loss": 3.482150409485467,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248555667001003,
      "loss": 2.5591,
      "theoretical_loss": 3.4821386614153944,
      "tokens_seen": 1676410880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485456369107322,
      "loss": 2.7247,
      "theoretical_loss": 3.482126913933169,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024853560682046137,
      "loss": 2.8482,
      "theoretical_loss": 3.4821151670387396,
      "tokens_seen": 1676541952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024852557673019055,
      "loss": 2.5457,
      "theoretical_loss": 3.482103420732053,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485155466399198,
      "loss": 2.715,
      "theoretical_loss": 3.482091675013058,
      "tokens_seen": 1676673024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024850551654964897,
      "loss": 2.6002,
      "theoretical_loss": 3.482079929881701,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024849548645937815,
      "loss": 2.6328,
      "theoretical_loss": 3.4820681853379307,
      "tokens_seen": 1676804096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024848545636910733,
      "loss": 2.4735,
      "theoretical_loss": 3.482056441381694,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484754262788365,
      "loss": 2.6296,
      "theoretical_loss": 3.4820446980129383,
      "tokens_seen": 1676935168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484653961885657,
      "loss": 2.8281,
      "theoretical_loss": 3.4820329552316123,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484553660982949,
      "loss": 2.5735,
      "theoretical_loss": 3.4820212130376627,
      "tokens_seen": 1677066240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024844533600802405,
      "loss": 2.5538,
      "theoretical_loss": 3.482009471431038,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024843530591775324,
      "loss": 2.5581,
      "theoretical_loss": 3.4819977304116847,
      "tokens_seen": 1677197312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024842527582748247,
      "loss": 2.6009,
      "theoretical_loss": 3.4819859899795516,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024841524573721165,
      "loss": 2.5509,
      "theoretical_loss": 3.481974250134586,
      "tokens_seen": 1677328384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024840521564694083,
      "loss": 2.427,
      "theoretical_loss": 3.481962510876736,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024839518555667,
      "loss": 2.6524,
      "theoretical_loss": 3.481950772205948,
      "tokens_seen": 1677459456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483851554663992,
      "loss": 2.6628,
      "theoretical_loss": 3.481939034122171,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483751253761284,
      "loss": 2.5665,
      "theoretical_loss": 3.4819272966253525,
      "tokens_seen": 1677590528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024836509528585756,
      "loss": 2.5822,
      "theoretical_loss": 3.4819155597154396,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 926151,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4123456478118896,
      "objective/train/theoretical_loss": 3.481903823392381,
      "objective/train/tokens_used": 1698181600,
      "theoretical_loss": 3.481903823392381,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024835506519558674,
      "loss": 2.367,
      "theoretical_loss": 3.481903823392381,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248345035105316,
      "loss": 2.679,
      "theoretical_loss": 3.4818920876561235,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024833500501504516,
      "loss": 2.6183,
      "theoretical_loss": 3.4818803525066153,
      "tokens_seen": 1677852672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024832497492477434,
      "loss": 2.5849,
      "theoretical_loss": 3.481868617943804,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483149448345035,
      "loss": 2.6098,
      "theoretical_loss": 3.4818568839676374,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483049147442327,
      "loss": 2.5749,
      "theoretical_loss": 3.481845150578063,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482948846539619,
      "loss": 2.3642,
      "theoretical_loss": 3.4818334177750296,
      "tokens_seen": 1678114816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024828485456369106,
      "loss": 2.7963,
      "theoretical_loss": 3.481821685558484,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024827482447342024,
      "loss": 2.6577,
      "theoretical_loss": 3.481809953928374,
      "tokens_seen": 1678245888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482647943831494,
      "loss": 2.7375,
      "theoretical_loss": 3.4817982228846476,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024825476429287866,
      "loss": 2.4198,
      "theoretical_loss": 3.4817864924272524,
      "tokens_seen": 1678376960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024824473420260784,
      "loss": 2.5,
      "theoretical_loss": 3.481774762556137,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248234704112337,
      "loss": 2.4954,
      "theoretical_loss": 3.4817630332712484,
      "tokens_seen": 1678508032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482246740220662,
      "loss": 2.683,
      "theoretical_loss": 3.4817513045725343,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482146439317954,
      "loss": 2.5708,
      "theoretical_loss": 3.481739576459943,
      "tokens_seen": 1678639104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024820461384152456,
      "loss": 2.5432,
      "theoretical_loss": 3.4817278489334225,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024819458375125375,
      "loss": 2.6931,
      "theoretical_loss": 3.48171612199292,
      "tokens_seen": 1678770176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481845536609829,
      "loss": 2.5562,
      "theoretical_loss": 3.481704395638383,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481745235707121,
      "loss": 2.6194,
      "theoretical_loss": 3.4816926698697612,
      "tokens_seen": 1678901248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024816449348044134,
      "loss": 2.702,
      "theoretical_loss": 3.4816809446870005,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481544633901705,
      "loss": 2.6251,
      "theoretical_loss": 3.48166922009005,
      "tokens_seen": 1679032320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481444332998997,
      "loss": 2.535,
      "theoretical_loss": 3.4816574960788564,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481344032096289,
      "loss": 2.6766,
      "theoretical_loss": 3.481645772653369,
      "tokens_seen": 1679163392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024812437311935807,
      "loss": 2.7735,
      "theoretical_loss": 3.4816340498135343,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481143430290873,
      "loss": 2.6769,
      "theoretical_loss": 3.4816223275593012,
      "tokens_seen": 1679294464
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 926923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.485356330871582,
      "objective/train/theoretical_loss": 3.4816106058906175,
      "objective/train/tokens_used": 1699820000,
      "theoretical_loss": 3.4816106058906175,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481043129388165,
      "loss": 2.5618,
      "theoretical_loss": 3.4816106058906175,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024809428284854566,
      "loss": 2.7246,
      "theoretical_loss": 3.48159888480743,
      "tokens_seen": 1679425536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024808425275827485,
      "loss": 2.5428,
      "theoretical_loss": 3.481587164309688,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024807422266800403,
      "loss": 2.7624,
      "theoretical_loss": 3.481575444397339,
      "tokens_seen": 1679556608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480641925777332,
      "loss": 2.7153,
      "theoretical_loss": 3.4815637250703304,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480541624874624,
      "loss": 2.6109,
      "theoretical_loss": 3.4815520063286103,
      "tokens_seen": 1679687680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024804413239719157,
      "loss": 2.4744,
      "theoretical_loss": 3.4815402881721274,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024803410230692075,
      "loss": 2.7151,
      "theoretical_loss": 3.4815285706008288,
      "tokens_seen": 1679818752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024802407221665,
      "loss": 2.7185,
      "theoretical_loss": 3.4815168536146626,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024801404212637917,
      "loss": 2.7745,
      "theoretical_loss": 3.481505137213577,
      "tokens_seen": 1679949824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024800401203610835,
      "loss": 2.5249,
      "theoretical_loss": 3.48149342139752,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024799398194583753,
      "loss": 2.6452,
      "theoretical_loss": 3.4814817061664396,
      "tokens_seen": 1680080896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479839518555667,
      "loss": 2.5947,
      "theoretical_loss": 3.481469991520283,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479739217652959,
      "loss": 2.6802,
      "theoretical_loss": 3.481458277458999,
      "tokens_seen": 1680211968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479638916750251,
      "loss": 2.8221,
      "theoretical_loss": 3.4814465639825354,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024795386158475425,
      "loss": 2.6334,
      "theoretical_loss": 3.4814348510908397,
      "tokens_seen": 1680343040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024794383149448344,
      "loss": 2.5991,
      "theoretical_loss": 3.481423138783861,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024793380140421267,
      "loss": 2.5965,
      "theoretical_loss": 3.4814114270615466,
      "tokens_seen": 1680474112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024792377131394185,
      "loss": 2.7437,
      "theoretical_loss": 3.4813997159238443,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024791374122367103,
      "loss": 2.726,
      "theoretical_loss": 3.481388005370702,
      "tokens_seen": 1680605184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002479037111334002,
      "loss": 2.7132,
      "theoretical_loss": 3.481376295402069,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478936810431294,
      "loss": 2.6958,
      "theoretical_loss": 3.481364586017892,
      "tokens_seen": 1680736256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478836509528586,
      "loss": 2.6933,
      "theoretical_loss": 3.4813528772181193,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024787362086258776,
      "loss": 2.5366,
      "theoretical_loss": 3.4813411690026994,
      "tokens_seen": 1680867328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024786359077231694,
      "loss": 2.6517,
      "theoretical_loss": 3.4813294613715797,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 927968,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.374573230743408,
      "objective/train/theoretical_loss": 3.4813177543247096,
      "objective/train/tokens_used": 1701458400,
      "theoretical_loss": 3.4813177543247096,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478535606820462,
      "loss": 2.6067,
      "theoretical_loss": 3.4813177543247096,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024784353059177536,
      "loss": 2.713,
      "theoretical_loss": 3.4813060478620352,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024783350050150454,
      "loss": 2.6515,
      "theoretical_loss": 3.4812943419835065,
      "tokens_seen": 1681129472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478234704112337,
      "loss": 2.4904,
      "theoretical_loss": 3.48128263668907,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478134403209629,
      "loss": 2.4253,
      "theoretical_loss": 3.481270931978675,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478034102306921,
      "loss": 2.8433,
      "theoretical_loss": 3.4812592278522687,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024779338014042126,
      "loss": 2.5101,
      "theoretical_loss": 3.4812475243098,
      "tokens_seen": 1681391616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024778335005015044,
      "loss": 2.7958,
      "theoretical_loss": 3.4812358213512162,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477733199598796,
      "loss": 2.5191,
      "theoretical_loss": 3.4812241189764657,
      "tokens_seen": 1681522688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024776328986960886,
      "loss": 2.4098,
      "theoretical_loss": 3.4812124171854966,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024775325977933804,
      "loss": 2.51,
      "theoretical_loss": 3.4812007159782574,
      "tokens_seen": 1681653760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477432296890672,
      "loss": 2.6385,
      "theoretical_loss": 3.481189015354696,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477331995987964,
      "loss": 2.626,
      "theoretical_loss": 3.4811773153147607,
      "tokens_seen": 1681784832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477231695085256,
      "loss": 2.7364,
      "theoretical_loss": 3.481165615858399,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024771313941825476,
      "loss": 2.6888,
      "theoretical_loss": 3.48115391698556,
      "tokens_seen": 1681915904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024770310932798395,
      "loss": 2.6936,
      "theoretical_loss": 3.481142218696191,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476930792377131,
      "loss": 2.6622,
      "theoretical_loss": 3.4811305209902406,
      "tokens_seen": 1682046976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476830491474423,
      "loss": 2.3932,
      "theoretical_loss": 3.481118823867657,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024767301905717154,
      "loss": 2.7226,
      "theoretical_loss": 3.4811071273283885,
      "tokens_seen": 1682178048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476629889669007,
      "loss": 2.7015,
      "theoretical_loss": 3.481095431372383,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476529588766299,
      "loss": 2.779,
      "theoretical_loss": 3.4810837359995883,
      "tokens_seen": 1682309120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476429287863591,
      "loss": 2.5293,
      "theoretical_loss": 3.481072041209954,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024763289869608827,
      "loss": 2.5135,
      "theoretical_loss": 3.4810603470034263,
      "tokens_seen": 1682440192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024762286860581745,
      "loss": 2.6093,
      "theoretical_loss": 3.481048653379955,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024761283851554663,
      "loss": 2.7552,
      "theoretical_loss": 3.4810369603394884,
      "tokens_seen": 1682571264
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 928503,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.304736614227295,
      "objective/train/theoretical_loss": 3.4810252678819733,
      "objective/train/tokens_used": 1703096800,
      "theoretical_loss": 3.4810252678819733,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476028084252758,
      "loss": 2.6536,
      "theoretical_loss": 3.4810252678819733,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247592778335005,
      "loss": 2.6542,
      "theoretical_loss": 3.481013576007359,
      "tokens_seen": 1682702336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024758274824473423,
      "loss": 2.6794,
      "theoretical_loss": 3.481001884715594,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475727181544634,
      "loss": 2.6845,
      "theoretical_loss": 3.4809901940066257,
      "tokens_seen": 1682833408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475626880641926,
      "loss": 2.7104,
      "theoretical_loss": 3.4809785038804026,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024755265797392177,
      "loss": 2.5928,
      "theoretical_loss": 3.480966814336873,
      "tokens_seen": 1682964480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024754262788365095,
      "loss": 2.6585,
      "theoretical_loss": 3.480955125375986,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024753259779338013,
      "loss": 2.6729,
      "theoretical_loss": 3.4809434369976886,
      "tokens_seen": 1683095552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475225677031093,
      "loss": 2.4823,
      "theoretical_loss": 3.48093174920193,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475125376128385,
      "loss": 2.9592,
      "theoretical_loss": 3.480920061988657,
      "tokens_seen": 1683226624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475025075225677,
      "loss": 2.5714,
      "theoretical_loss": 3.48090837535782,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474924774322969,
      "loss": 2.626,
      "theoretical_loss": 3.480896689309366,
      "tokens_seen": 1683357696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474824473420261,
      "loss": 2.4919,
      "theoretical_loss": 3.4808850038432437,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474724172517553,
      "loss": 2.4615,
      "theoretical_loss": 3.480873318959401,
      "tokens_seen": 1683488768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024746238716148445,
      "loss": 2.6045,
      "theoretical_loss": 3.4808616346577868,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024745235707121364,
      "loss": 2.7379,
      "theoretical_loss": 3.480849950938349,
      "tokens_seen": 1683619840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474423269809428,
      "loss": 2.5749,
      "theoretical_loss": 3.4808382678010363,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247432296890672,
      "loss": 2.4415,
      "theoretical_loss": 3.480826585245797,
      "tokens_seen": 1683750912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474222668004012,
      "loss": 2.6679,
      "theoretical_loss": 3.480814903272579,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024741223671013036,
      "loss": 2.3961,
      "theoretical_loss": 3.4808032218813305,
      "tokens_seen": 1683881984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474022066198596,
      "loss": 2.729,
      "theoretical_loss": 3.4807915410720005,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473921765295888,
      "loss": 2.7828,
      "theoretical_loss": 3.480779860844538,
      "tokens_seen": 1684013056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024738214643931796,
      "loss": 2.4455,
      "theoretical_loss": 3.4807681811988895,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024737211634904714,
      "loss": 2.6618,
      "theoretical_loss": 3.480756502135005,
      "tokens_seen": 1684144128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473620862587764,
      "loss": 2.4953,
      "theoretical_loss": 3.4807448236528318,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 929812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8366990089416504,
      "objective/train/theoretical_loss": 3.480733145752319,
      "objective/train/tokens_used": 1704735200,
      "theoretical_loss": 3.480733145752319,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024735205616850556,
      "loss": 2.5936,
      "theoretical_loss": 3.480733145752319,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024734202607823474,
      "loss": 2.5337,
      "theoretical_loss": 3.4807214684334147,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473319959879639,
      "loss": 2.7501,
      "theoretical_loss": 3.480709791696068,
      "tokens_seen": 1684406272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473219658976931,
      "loss": 2.4478,
      "theoretical_loss": 3.4806981155402257,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473119358074223,
      "loss": 2.6837,
      "theoretical_loss": 3.4806864399658375,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024730190571715146,
      "loss": 2.6195,
      "theoretical_loss": 3.4806747649728518,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024729187562688064,
      "loss": 2.8179,
      "theoretical_loss": 3.480663090561217,
      "tokens_seen": 1684668416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472818455366098,
      "loss": 2.4978,
      "theoretical_loss": 3.480651416730881,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024727181544633906,
      "loss": 2.5977,
      "theoretical_loss": 3.4806397434817926,
      "tokens_seen": 1684799488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024726178535606824,
      "loss": 2.5959,
      "theoretical_loss": 3.4806280708139,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472517552657974,
      "loss": 2.5551,
      "theoretical_loss": 3.480616398727152,
      "tokens_seen": 1684930560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472417251755266,
      "loss": 2.8213,
      "theoretical_loss": 3.480604727221497,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472316950852558,
      "loss": 2.8073,
      "theoretical_loss": 3.4805930562968834,
      "tokens_seen": 1685061632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024722166499498496,
      "loss": 2.466,
      "theoretical_loss": 3.4805813859532595,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024721163490471415,
      "loss": 2.5407,
      "theoretical_loss": 3.4805697161905744,
      "tokens_seen": 1685192704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472016048144433,
      "loss": 2.6209,
      "theoretical_loss": 3.480558047008776,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471915747241725,
      "loss": 2.6271,
      "theoretical_loss": 3.480546378407812,
      "tokens_seen": 1685323776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024718154463390174,
      "loss": 2.6273,
      "theoretical_loss": 3.4805347103876327,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471715145436309,
      "loss": 2.7374,
      "theoretical_loss": 3.480523042948186,
      "tokens_seen": 1685454848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471614844533601,
      "loss": 2.5767,
      "theoretical_loss": 3.48051137608942,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471514543630893,
      "loss": 2.8647,
      "theoretical_loss": 3.480499709811283,
      "tokens_seen": 1685585920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024714142427281847,
      "loss": 2.6571,
      "theoretical_loss": 3.4804880441137245,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024713139418254765,
      "loss": 2.4605,
      "theoretical_loss": 3.480476378996692,
      "tokens_seen": 1685716992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024712136409227683,
      "loss": 2.6362,
      "theoretical_loss": 3.4804647144601346,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247111334002006,
      "loss": 2.5492,
      "theoretical_loss": 3.480453050504001,
      "tokens_seen": 1685848064
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 931049,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3117072582244873,
      "objective/train/theoretical_loss": 3.4804413871282396,
      "objective/train/tokens_used": 1706373600,
      "theoretical_loss": 3.4804413871282396,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471013039117352,
      "loss": 2.6159,
      "theoretical_loss": 3.4804413871282396,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024709127382146443,
      "loss": 2.3377,
      "theoretical_loss": 3.480429724332798,
      "tokens_seen": 1685979136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470812437311936,
      "loss": 2.4654,
      "theoretical_loss": 3.480418062117627,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470712136409228,
      "loss": 2.593,
      "theoretical_loss": 3.480406400482673,
      "tokens_seen": 1686110208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024706118355065197,
      "loss": 2.6179,
      "theoretical_loss": 3.4803947394278856,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024705115346038115,
      "loss": 2.5479,
      "theoretical_loss": 3.4803830789532126,
      "tokens_seen": 1686241280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024704112337011033,
      "loss": 2.686,
      "theoretical_loss": 3.4803714190586037,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470310932798395,
      "loss": 2.7621,
      "theoretical_loss": 3.4803597597440072,
      "tokens_seen": 1686372352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470210631895687,
      "loss": 2.6463,
      "theoretical_loss": 3.4803481010093718,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470110330992979,
      "loss": 2.427,
      "theoretical_loss": 3.4803364428546453,
      "tokens_seen": 1686503424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470010030090271,
      "loss": 2.6326,
      "theoretical_loss": 3.4803247852797767,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469909729187563,
      "loss": 2.6612,
      "theoretical_loss": 3.480313128284715,
      "tokens_seen": 1686634496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469809428284855,
      "loss": 2.6253,
      "theoretical_loss": 3.4803014718694087,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024697091273821466,
      "loss": 2.6363,
      "theoretical_loss": 3.4802898160338063,
      "tokens_seen": 1686765568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024696088264794384,
      "loss": 2.699,
      "theoretical_loss": 3.480278160777856,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246950852557673,
      "loss": 2.5976,
      "theoretical_loss": 3.4802665061015077,
      "tokens_seen": 1686896640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469408224674022,
      "loss": 2.6327,
      "theoretical_loss": 3.480254852004709,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469307923771314,
      "loss": 2.5549,
      "theoretical_loss": 3.4802431984874094,
      "tokens_seen": 1687027712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024692076228686056,
      "loss": 2.5634,
      "theoretical_loss": 3.4802315455495565,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002469107321965898,
      "loss": 2.6633,
      "theoretical_loss": 3.4802198931910993,
      "tokens_seen": 1687158784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246900702106319,
      "loss": 2.4729,
      "theoretical_loss": 3.480208241411987,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024689067201604816,
      "loss": 2.4689,
      "theoretical_loss": 3.4801965902121683,
      "tokens_seen": 1687289856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024688064192577734,
      "loss": 2.6025,
      "theoretical_loss": 3.480184939591591,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468706118355065,
      "loss": 2.5987,
      "theoretical_loss": 3.480173289550205,
      "tokens_seen": 1687420928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468605817452357,
      "loss": 2.4535,
      "theoretical_loss": 3.4801616400879585,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 931336,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4152073860168457,
      "objective/train/theoretical_loss": 3.4801499912048,
      "objective/train/tokens_used": 1708012000,
      "theoretical_loss": 3.4801499912048,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468505516549649,
      "loss": 2.6822,
      "theoretical_loss": 3.4801499912048,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024684052156469406,
      "loss": 2.5953,
      "theoretical_loss": 3.480138342900678,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024683049147442325,
      "loss": 2.488,
      "theoretical_loss": 3.480126695175542,
      "tokens_seen": 1687683072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468204613841525,
      "loss": 2.6257,
      "theoretical_loss": 3.48011504802934,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024681043129388166,
      "loss": 2.5836,
      "theoretical_loss": 3.4801034014620216,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024680040120361084,
      "loss": 2.5873,
      "theoretical_loss": 3.4800917554735347,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024679037111334,
      "loss": 2.6314,
      "theoretical_loss": 3.4800801100638283,
      "tokens_seen": 1687945216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467803410230692,
      "loss": 2.4858,
      "theoretical_loss": 3.480068465232852,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467703109327984,
      "loss": 2.6978,
      "theoretical_loss": 3.480056820980553,
      "tokens_seen": 1688076288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024676028084252757,
      "loss": 2.5979,
      "theoretical_loss": 3.4800451773068817,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024675025075225675,
      "loss": 2.6652,
      "theoretical_loss": 3.4800335342117856,
      "tokens_seen": 1688207360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024674022066198593,
      "loss": 2.6697,
      "theoretical_loss": 3.4800218916952144,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024673019057171516,
      "loss": 2.7536,
      "theoretical_loss": 3.4800102497571164,
      "tokens_seen": 1688338432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024672016048144435,
      "loss": 2.4352,
      "theoretical_loss": 3.47999860839744,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467101303911735,
      "loss": 2.6223,
      "theoretical_loss": 3.479986967616135,
      "tokens_seen": 1688469504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467001003009027,
      "loss": 2.5931,
      "theoretical_loss": 3.4799753274131495,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466900702106319,
      "loss": 2.4609,
      "theoretical_loss": 3.479963687788433,
      "tokens_seen": 1688600576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024668004012036107,
      "loss": 2.3484,
      "theoretical_loss": 3.4799520487419335,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024667001003009025,
      "loss": 2.6675,
      "theoretical_loss": 3.4799404102736005,
      "tokens_seen": 1688731648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024665997993981943,
      "loss": 2.5314,
      "theoretical_loss": 3.4799287723833823,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466499498495486,
      "loss": 2.7952,
      "theoretical_loss": 3.4799171350712284,
      "tokens_seen": 1688862720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024663991975927785,
      "loss": 2.7406,
      "theoretical_loss": 3.4799054983370867,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024662988966900703,
      "loss": 2.5602,
      "theoretical_loss": 3.4798938621809077,
      "tokens_seen": 1688993792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466198595787362,
      "loss": 2.4935,
      "theoretical_loss": 3.4798822266026384,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466098294884654,
      "loss": 2.5577,
      "theoretical_loss": 3.4798705916022286,
      "tokens_seen": 1689124864
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 932508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9064462184906006,
      "objective/train/theoretical_loss": 3.4798589571796272,
      "objective/train/tokens_used": 1709650400,
      "theoretical_loss": 3.4798589571796272,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024659979939819463,
      "loss": 2.7119,
      "theoretical_loss": 3.4798589571796272,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465897693079238,
      "loss": 2.6507,
      "theoretical_loss": 3.4798473233347833,
      "tokens_seen": 1689255936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246579739217653,
      "loss": 2.699,
      "theoretical_loss": 3.4798356900676444,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024656970912738217,
      "loss": 2.5551,
      "theoretical_loss": 3.4798240573781616,
      "tokens_seen": 1689387008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024655967903711135,
      "loss": 2.6385,
      "theoretical_loss": 3.479812425266282,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024654964894684053,
      "loss": 2.6437,
      "theoretical_loss": 3.4798007937319557,
      "tokens_seen": 1689518080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465396188565697,
      "loss": 2.6489,
      "theoretical_loss": 3.479789162775131,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465295887662989,
      "loss": 2.5205,
      "theoretical_loss": 3.4797775323957567,
      "tokens_seen": 1689649152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465195586760281,
      "loss": 2.4516,
      "theoretical_loss": 3.479765902593782,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465095285857573,
      "loss": 2.8956,
      "theoretical_loss": 3.479754273369156,
      "tokens_seen": 1689780224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464994984954865,
      "loss": 2.6378,
      "theoretical_loss": 3.4797426447218274,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464894684052157,
      "loss": 2.594,
      "theoretical_loss": 3.479731016651745,
      "tokens_seen": 1689911296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024647943831494486,
      "loss": 2.5246,
      "theoretical_loss": 3.4797193891588583,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024646940822467404,
      "loss": 2.3676,
      "theoretical_loss": 3.479707762243116,
      "tokens_seen": 1690042368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464593781344032,
      "loss": 2.6035,
      "theoretical_loss": 3.479696135904467,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464493480441324,
      "loss": 2.4016,
      "theoretical_loss": 3.47968451014286,
      "tokens_seen": 1690173440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464393179538616,
      "loss": 2.5843,
      "theoretical_loss": 3.479672884958245,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024642928786359076,
      "loss": 2.8138,
      "theoretical_loss": 3.4796612603505697,
      "tokens_seen": 1690304512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024641925777332,
      "loss": 2.5247,
      "theoretical_loss": 3.4796496363197837,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464092276830492,
      "loss": 2.4319,
      "theoretical_loss": 3.479638012865836,
      "tokens_seen": 1690435584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024639919759277836,
      "loss": 2.577,
      "theoretical_loss": 3.4796263899886757,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024638916750250754,
      "loss": 2.7595,
      "theoretical_loss": 3.479614767688252,
      "tokens_seen": 1690566656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463791374122367,
      "loss": 2.3949,
      "theoretical_loss": 3.4796031459645134,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463691073219659,
      "loss": 2.5013,
      "theoretical_loss": 3.479591524817409,
      "tokens_seen": 1690697728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463590772316951,
      "loss": 2.3728,
      "theoretical_loss": 3.4795799042468882,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 933011,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7067461013793945,
      "objective/train/theoretical_loss": 3.4795682842529,
      "objective/train/tokens_used": 1711288800,
      "theoretical_loss": 3.4795682842529,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024634904714142426,
      "loss": 2.5548,
      "theoretical_loss": 3.4795682842529,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024633901705115345,
      "loss": 2.7218,
      "theoretical_loss": 3.4795566648353935,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463289869608827,
      "loss": 2.5683,
      "theoretical_loss": 3.4795450459943167,
      "tokens_seen": 1690959872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024631895687061186,
      "loss": 2.4241,
      "theoretical_loss": 3.4795334277296206,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024630892678034104,
      "loss": 2.4936,
      "theoretical_loss": 3.479521810041253,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462988966900702,
      "loss": 2.6301,
      "theoretical_loss": 3.4795101929291627,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462888665997994,
      "loss": 2.5857,
      "theoretical_loss": 3.4794985763932997,
      "tokens_seen": 1691222016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462788365095286,
      "loss": 2.5328,
      "theoretical_loss": 3.4794869604336123,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024626880641925777,
      "loss": 2.7977,
      "theoretical_loss": 3.47947534505005,
      "tokens_seen": 1691353088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024625877632898695,
      "loss": 2.4109,
      "theoretical_loss": 3.4794637302425624,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024624874623871613,
      "loss": 2.6785,
      "theoretical_loss": 3.479452116011098,
      "tokens_seen": 1691484160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024623871614844536,
      "loss": 2.534,
      "theoretical_loss": 3.4794405023556054,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024622868605817455,
      "loss": 2.6258,
      "theoretical_loss": 3.4794288892760346,
      "tokens_seen": 1691615232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462186559679037,
      "loss": 2.3829,
      "theoretical_loss": 3.4794172767723346,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462086258776329,
      "loss": 2.8251,
      "theoretical_loss": 3.479405664844454,
      "tokens_seen": 1691746304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461985957873621,
      "loss": 2.4629,
      "theoretical_loss": 3.479394053492343,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024618856569709127,
      "loss": 2.6932,
      "theoretical_loss": 3.4793824427159494,
      "tokens_seen": 1691877376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024617853560682045,
      "loss": 2.5929,
      "theoretical_loss": 3.4793708325152237,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024616850551654963,
      "loss": 2.5565,
      "theoretical_loss": 3.479359222890114,
      "tokens_seen": 1692008448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461584754262788,
      "loss": 2.5373,
      "theoretical_loss": 3.47934761384057,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024614844533600805,
      "loss": 2.7793,
      "theoretical_loss": 3.479336005366541,
      "tokens_seen": 1692139520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024613841524573723,
      "loss": 2.5573,
      "theoretical_loss": 3.4793243974679755,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461283851554664,
      "loss": 2.7503,
      "theoretical_loss": 3.479312790144823,
      "tokens_seen": 1692270592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461183550651956,
      "loss": 2.4464,
      "theoretical_loss": 3.479301183397033,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461083249749248,
      "loss": 2.3415,
      "theoretical_loss": 3.479289577224555,
      "tokens_seen": 1692401664
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 934010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.963383436203003,
      "objective/train/theoretical_loss": 3.4792779716273365,
      "objective/train/tokens_used": 1712927200,
      "theoretical_loss": 3.4792779716273365,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024609829488465395,
      "loss": 2.5812,
      "theoretical_loss": 3.4792779716273365,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024608826479438314,
      "loss": 2.4986,
      "theoretical_loss": 3.479266366605329,
      "tokens_seen": 1692532736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460782347041123,
      "loss": 2.3378,
      "theoretical_loss": 3.4792547621584804,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460682046138415,
      "loss": 2.6923,
      "theoretical_loss": 3.47924315828674,
      "tokens_seen": 1692663808
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024605817452357073,
      "loss": 2.7482,
      "theoretical_loss": 3.479231554990057,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460481444332999,
      "loss": 2.5195,
      "theoretical_loss": 3.479219952268381,
      "tokens_seen": 1692794880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460381143430291,
      "loss": 2.5607,
      "theoretical_loss": 3.4792083501216613,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460280842527583,
      "loss": 2.5865,
      "theoretical_loss": 3.479196748549847,
      "tokens_seen": 1692925952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024601805416248746,
      "loss": 2.7438,
      "theoretical_loss": 3.4791851475528874,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024600802407221664,
      "loss": 2.6863,
      "theoretical_loss": 3.4791735471307312,
      "tokens_seen": 1693057024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459979939819458,
      "loss": 2.5338,
      "theoretical_loss": 3.4791619472833286,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245987963891675,
      "loss": 2.3709,
      "theoretical_loss": 3.479150348010628,
      "tokens_seen": 1693188096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459779338014042,
      "loss": 2.6146,
      "theoretical_loss": 3.479138749312579,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459679037111334,
      "loss": 2.5514,
      "theoretical_loss": 3.4791271511891315,
      "tokens_seen": 1693319168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459578736208626,
      "loss": 2.5177,
      "theoretical_loss": 3.4791155536402343,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459478435305918,
      "loss": 2.5815,
      "theoretical_loss": 3.479103956665836,
      "tokens_seen": 1693450240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024593781344032096,
      "loss": 2.7414,
      "theoretical_loss": 3.479092360265887,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024592778335005014,
      "loss": 2.7175,
      "theoretical_loss": 3.4790807644403365,
      "tokens_seen": 1693581312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459177532597793,
      "loss": 2.5297,
      "theoretical_loss": 3.479069169189133,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002459077231695085,
      "loss": 2.8493,
      "theoretical_loss": 3.479057574512227,
      "tokens_seen": 1693712384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458976930792377,
      "loss": 2.5372,
      "theoretical_loss": 3.4790459804095666,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024588766298896687,
      "loss": 2.6134,
      "theoretical_loss": 3.479034386881102,
      "tokens_seen": 1693843456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458776328986961,
      "loss": 2.5648,
      "theoretical_loss": 3.4790227939267826,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458676028084253,
      "loss": 2.6549,
      "theoretical_loss": 3.479011201546557,
      "tokens_seen": 1693974528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024585757271815446,
      "loss": 2.5179,
      "theoretical_loss": 3.478999609740375,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 934404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.564699172973633,
      "objective/train/theoretical_loss": 3.4789880185081863,
      "objective/train/tokens_used": 1714565600,
      "theoretical_loss": 3.4789880185081863,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024584754262788365,
      "loss": 2.6838,
      "theoretical_loss": 3.4789880185081863,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458375125376129,
      "loss": 2.4893,
      "theoretical_loss": 3.4789764278499398,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024582748244734206,
      "loss": 2.5943,
      "theoretical_loss": 3.478964837765585,
      "tokens_seen": 1694236672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024581745235707124,
      "loss": 2.9362,
      "theoretical_loss": 3.4789532482550714,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458074222668004,
      "loss": 2.7273,
      "theoretical_loss": 3.478941659318348,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457973921765296,
      "loss": 2.6358,
      "theoretical_loss": 3.478930070955365,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457873620862588,
      "loss": 2.6806,
      "theoretical_loss": 3.478918483166071,
      "tokens_seen": 1694498816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024577733199598797,
      "loss": 2.4375,
      "theoretical_loss": 3.478906895950416,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024576730190571715,
      "loss": 2.4128,
      "theoretical_loss": 3.4788953093083492,
      "tokens_seen": 1694629888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024575727181544633,
      "loss": 2.6627,
      "theoretical_loss": 3.4788837232398198,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024574724172517556,
      "loss": 2.4786,
      "theoretical_loss": 3.4788721377447773,
      "tokens_seen": 1694760960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024573721163490475,
      "loss": 2.6385,
      "theoretical_loss": 3.4788605528231713,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024572718154463393,
      "loss": 2.7068,
      "theoretical_loss": 3.478848968474951,
      "tokens_seen": 1694892032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457171514543631,
      "loss": 2.4245,
      "theoretical_loss": 3.4788373847000664,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457071213640923,
      "loss": 2.5818,
      "theoretical_loss": 3.4788258014984663,
      "tokens_seen": 1695023104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024569709127382147,
      "loss": 2.556,
      "theoretical_loss": 3.478814218870101,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024568706118355065,
      "loss": 2.7009,
      "theoretical_loss": 3.478802636814919,
      "tokens_seen": 1695154176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024567703109327983,
      "loss": 2.5312,
      "theoretical_loss": 3.4787910553328705,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245667001003009,
      "loss": 2.5586,
      "theoretical_loss": 3.4787794744239044,
      "tokens_seen": 1695285248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024565697091273825,
      "loss": 2.3999,
      "theoretical_loss": 3.4787678940879707,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024564694082246743,
      "loss": 2.6618,
      "theoretical_loss": 3.4787563143250186,
      "tokens_seen": 1695416320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456369107321966,
      "loss": 2.5304,
      "theoretical_loss": 3.478744735134998,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456268806419258,
      "loss": 2.5885,
      "theoretical_loss": 3.4787331565178574,
      "tokens_seen": 1695547392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245616850551655,
      "loss": 2.5677,
      "theoretical_loss": 3.4787215784735475,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024560682046138415,
      "loss": 2.7126,
      "theoretical_loss": 3.4787100010020175,
      "tokens_seen": 1695678464
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 935560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5417470932006836,
      "objective/train/theoretical_loss": 3.478698424103216,
      "objective/train/tokens_used": 1716204000,
      "theoretical_loss": 3.478698424103216,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024559679037111334,
      "loss": 2.7574,
      "theoretical_loss": 3.478698424103216,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455867602808425,
      "loss": 2.5035,
      "theoretical_loss": 3.4786868477770945,
      "tokens_seen": 1695809536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455767301905717,
      "loss": 2.624,
      "theoretical_loss": 3.4786752720236005,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024556670010030093,
      "loss": 2.4596,
      "theoretical_loss": 3.4786636968426845,
      "tokens_seen": 1695940608
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455566700100301,
      "loss": 2.6585,
      "theoretical_loss": 3.478652122234296,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455466399197593,
      "loss": 2.5577,
      "theoretical_loss": 3.4786405481983844,
      "tokens_seen": 1696071680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455366098294885,
      "loss": 2.9418,
      "theoretical_loss": 3.4786289747348995,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024552657973921766,
      "loss": 2.626,
      "theoretical_loss": 3.478617401843791,
      "tokens_seen": 1696202752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024551654964894684,
      "loss": 2.7091,
      "theoretical_loss": 3.4786058295250077,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245506519558676,
      "loss": 2.6192,
      "theoretical_loss": 3.4785942577785,
      "tokens_seen": 1696333824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454964894684052,
      "loss": 2.6024,
      "theoretical_loss": 3.4785826866042173,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454864593781344,
      "loss": 2.6667,
      "theoretical_loss": 3.478571116002109,
      "tokens_seen": 1696464896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454764292878636,
      "loss": 2.3264,
      "theoretical_loss": 3.4785595459721246,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454663991975928,
      "loss": 2.2744,
      "theoretical_loss": 3.478547976514214,
      "tokens_seen": 1696595968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245456369107322,
      "loss": 2.5428,
      "theoretical_loss": 3.4785364076283267,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024544633901705116,
      "loss": 2.7503,
      "theoretical_loss": 3.478524839314413,
      "tokens_seen": 1696727040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024543630892678034,
      "loss": 2.7322,
      "theoretical_loss": 3.478513271572421,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454262788365095,
      "loss": 2.6492,
      "theoretical_loss": 3.4785017044023014,
      "tokens_seen": 1696858112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454162487462387,
      "loss": 2.6131,
      "theoretical_loss": 3.4784901378040036,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454062186559679,
      "loss": 2.5254,
      "theoretical_loss": 3.4784785717774778,
      "tokens_seen": 1696989184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024539618856569707,
      "loss": 2.8529,
      "theoretical_loss": 3.4784670063226732,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453861584754263,
      "loss": 2.4435,
      "theoretical_loss": 3.478455441439539,
      "tokens_seen": 1697120256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453761283851555,
      "loss": 2.5408,
      "theoretical_loss": 3.4784438771280257,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024536609829488466,
      "loss": 2.5826,
      "theoretical_loss": 3.4784323133880823,
      "tokens_seen": 1697251328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024535606820461385,
      "loss": 2.6413,
      "theoretical_loss": 3.4784207502196587,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 936050,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.872040271759033,
      "objective/train/theoretical_loss": 3.4784091876227046,
      "objective/train/tokens_used": 1717842400,
      "theoretical_loss": 3.4784091876227046,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245346038114343,
      "loss": 2.6255,
      "theoretical_loss": 3.4784091876227046,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453360080240722,
      "loss": 2.57,
      "theoretical_loss": 3.47839762559717,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453259779338014,
      "loss": 2.569,
      "theoretical_loss": 3.478386064143004,
      "tokens_seen": 1697513472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024531594784353057,
      "loss": 2.4512,
      "theoretical_loss": 3.4783745032601567,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024530591775325975,
      "loss": 2.5866,
      "theoretical_loss": 3.4783629429485785,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245295887662989,
      "loss": 2.6212,
      "theoretical_loss": 3.4783513832082176,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024528585757271817,
      "loss": 2.6184,
      "theoretical_loss": 3.4783398240390246,
      "tokens_seen": 1697775616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024527582748244735,
      "loss": 2.51,
      "theoretical_loss": 3.4783282654409495,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024526579739217653,
      "loss": 2.9548,
      "theoretical_loss": 3.478316707413941,
      "tokens_seen": 1697906688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452557673019057,
      "loss": 2.49,
      "theoretical_loss": 3.4783051499579507,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452457372116349,
      "loss": 2.4532,
      "theoretical_loss": 3.478293593072926,
      "tokens_seen": 1698037760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452357071213641,
      "loss": 2.6569,
      "theoretical_loss": 3.4782820367588183,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024522567703109325,
      "loss": 2.5143,
      "theoretical_loss": 3.478270481015577,
      "tokens_seen": 1698168832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024521564694082244,
      "loss": 2.6679,
      "theoretical_loss": 3.4782589258431518,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024520561685055167,
      "loss": 2.5766,
      "theoretical_loss": 3.4782473712414923,
      "tokens_seen": 1698299904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024519558676028085,
      "loss": 2.5808,
      "theoretical_loss": 3.4782358172105483,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024518555667001003,
      "loss": 2.6037,
      "theoretical_loss": 3.47822426375027,
      "tokens_seen": 1698430976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451755265797392,
      "loss": 2.5633,
      "theoretical_loss": 3.478212710860607,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451654964894684,
      "loss": 2.5854,
      "theoretical_loss": 3.4782011585415087,
      "tokens_seen": 1698562048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451554663991976,
      "loss": 2.4266,
      "theoretical_loss": 3.4781896067929257,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024514543630892676,
      "loss": 2.3192,
      "theoretical_loss": 3.4781780556148068,
      "tokens_seen": 1698693120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024513540621865594,
      "loss": 2.7456,
      "theoretical_loss": 3.4781665050071027,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451253761283852,
      "loss": 2.7737,
      "theoretical_loss": 3.478154954969763,
      "tokens_seen": 1698824192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024511534603811435,
      "loss": 2.5488,
      "theoretical_loss": 3.4781434055027365,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024510531594784354,
      "loss": 2.5229,
      "theoretical_loss": 3.478131856605975,
      "tokens_seen": 1698955264
    },
    {
      "epoch": 0.51,
      "objective/train/docs_used": 937278,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.28928804397583,
      "objective/train/theoretical_loss": 3.4781203082794265,
      "objective/train/tokens_used": 1719480800,
      "theoretical_loss": 3.4781203082794265,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450952858575727,
      "loss": 2.5324,
      "theoretical_loss": 3.4781203082794265,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450852557673019,
      "loss": 2.6336,
      "theoretical_loss": 3.4781087605230425,
      "tokens_seen": 1699086336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024507522567703113,
      "loss": 2.7146,
      "theoretical_loss": 3.4780972133367714,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450651955867603,
      "loss": 2.5599,
      "theoretical_loss": 3.478085666720564,
      "tokens_seen": 1699217408
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450551654964895,
      "loss": 2.5312,
      "theoretical_loss": 3.47807412067437,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450451354062187,
      "loss": 2.6779,
      "theoretical_loss": 3.4780625751981384,
      "tokens_seen": 1699348480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024503510531594786,
      "loss": 2.632,
      "theoretical_loss": 3.478051030291821,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024502507522567704,
      "loss": 2.5656,
      "theoretical_loss": 3.4780394859553656,
      "tokens_seen": 1699479552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002450150451354062,
      "loss": 2.5646,
      "theoretical_loss": 3.4780279421887235,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002450050150451354,
      "loss": 2.4481,
      "theoretical_loss": 3.478016398991844,
      "tokens_seen": 1699610624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449949849548646,
      "loss": 2.7247,
      "theoretical_loss": 3.478004856364677,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449849548645938,
      "loss": 2.7444,
      "theoretical_loss": 3.477993314307173,
      "tokens_seen": 1699741696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244974924774323,
      "loss": 2.429,
      "theoretical_loss": 3.4779817728192803,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449648946840522,
      "loss": 2.5336,
      "theoretical_loss": 3.477970231900951,
      "tokens_seen": 1699872768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024495486459378136,
      "loss": 2.5954,
      "theoretical_loss": 3.477958691552134,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024494483450351054,
      "loss": 2.7614,
      "theoretical_loss": 3.4779471517727787,
      "tokens_seen": 1700003840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449348044132397,
      "loss": 2.6846,
      "theoretical_loss": 3.4779356125628365,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449247743229689,
      "loss": 2.493,
      "theoretical_loss": 3.4779240739222557,
      "tokens_seen": 1700134912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449147442326981,
      "loss": 2.3463,
      "theoretical_loss": 3.4779125358509875,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024490471414242727,
      "loss": 2.668,
      "theoretical_loss": 3.4779009983489813,
      "tokens_seen": 1700265984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448946840521565,
      "loss": 2.508,
      "theoretical_loss": 3.4778894614161873,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448846539618857,
      "loss": 2.6657,
      "theoretical_loss": 3.477877925052555,
      "tokens_seen": 1700397056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024487462387161486,
      "loss": 2.7525,
      "theoretical_loss": 3.4778663892580353,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024486459378134405,
      "loss": 2.3541,
      "theoretical_loss": 3.477854854032577,
      "tokens_seen": 1700528128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448545636910732,
      "loss": 2.665,
      "theoretical_loss": 3.477843319376131,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 938365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5744383335113525,
      "objective/train/theoretical_loss": 3.4778317852886476,
      "objective/train/tokens_used": 1721119200,
      "theoretical_loss": 3.4778317852886476,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448445336008024,
      "loss": 2.7538,
      "theoretical_loss": 3.4778317852886476,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448345035105316,
      "loss": 2.5944,
      "theoretical_loss": 3.4778202517700754,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024482447342026077,
      "loss": 2.6273,
      "theoretical_loss": 3.4778087188203655,
      "tokens_seen": 1700790272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024481444332998995,
      "loss": 2.4997,
      "theoretical_loss": 3.477797186439468,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448044132397192,
      "loss": 2.3106,
      "theoretical_loss": 3.4777856546273327,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024479438314944837,
      "loss": 2.4189,
      "theoretical_loss": 3.477774123383909,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024478435305917755,
      "loss": 2.2807,
      "theoretical_loss": 3.477762592709148,
      "tokens_seen": 1701052416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024477432296890673,
      "loss": 2.6812,
      "theoretical_loss": 3.477751062602999,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447642928786359,
      "loss": 2.6293,
      "theoretical_loss": 3.477739533065412,
      "tokens_seen": 1701183488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447542627883651,
      "loss": 2.6103,
      "theoretical_loss": 3.4777280040963383,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447442326980943,
      "loss": 2.6884,
      "theoretical_loss": 3.4777164756957264,
      "tokens_seen": 1701314560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024473420260782345,
      "loss": 2.6038,
      "theoretical_loss": 3.477704947863527,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024472417251755264,
      "loss": 2.6905,
      "theoretical_loss": 3.4776934205996906,
      "tokens_seen": 1701445632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024471414242728187,
      "loss": 2.6287,
      "theoretical_loss": 3.4776818939041663,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024470411233701105,
      "loss": 2.6191,
      "theoretical_loss": 3.477670367776905,
      "tokens_seen": 1701576704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024469408224674023,
      "loss": 2.6578,
      "theoretical_loss": 3.477658842217857,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446840521564694,
      "loss": 2.6265,
      "theoretical_loss": 3.4776473172269715,
      "tokens_seen": 1701707776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446740220661986,
      "loss": 2.5431,
      "theoretical_loss": 3.4776357928041994,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446639919759278,
      "loss": 2.7685,
      "theoretical_loss": 3.4776242689494903,
      "tokens_seen": 1701838848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024465396188565696,
      "loss": 2.5846,
      "theoretical_loss": 3.4776127456627948,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024464393179538614,
      "loss": 2.674,
      "theoretical_loss": 3.4776012229440623,
      "tokens_seen": 1701969920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446339017051154,
      "loss": 2.6775,
      "theoretical_loss": 3.477589700793244,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024462387161484456,
      "loss": 2.4812,
      "theoretical_loss": 3.4775781792102887,
      "tokens_seen": 1702100992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024461384152457374,
      "loss": 2.9415,
      "theoretical_loss": 3.477566658195148,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446038114343029,
      "loss": 2.8512,
      "theoretical_loss": 3.477555137747771,
      "tokens_seen": 1702232064
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 938945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.693019390106201,
      "objective/train/theoretical_loss": 3.4775436178681085,
      "objective/train/tokens_used": 1722757600,
      "theoretical_loss": 3.4775436178681085,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445937813440321,
      "loss": 2.7421,
      "theoretical_loss": 3.4775436178681085,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445837512537613,
      "loss": 2.4748,
      "theoretical_loss": 3.4775320985561105,
      "tokens_seen": 1702363136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024457372116349046,
      "loss": 2.6635,
      "theoretical_loss": 3.477520579811727,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024456369107321964,
      "loss": 2.6122,
      "theoretical_loss": 3.477509061634908,
      "tokens_seen": 1702494208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445536609829488,
      "loss": 2.8061,
      "theoretical_loss": 3.4774975440256037,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024454363089267806,
      "loss": 2.5901,
      "theoretical_loss": 3.4774860269837653,
      "tokens_seen": 1702625280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024453360080240724,
      "loss": 2.8434,
      "theoretical_loss": 3.4774745105093414,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445235707121364,
      "loss": 2.6836,
      "theoretical_loss": 3.4774629946022837,
      "tokens_seen": 1702756352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445135406218656,
      "loss": 2.7571,
      "theoretical_loss": 3.477451479262541,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445035105315948,
      "loss": 2.4806,
      "theoretical_loss": 3.477439964490065,
      "tokens_seen": 1702887424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024449348044132396,
      "loss": 2.6898,
      "theoretical_loss": 3.477428450284805,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024448345035105315,
      "loss": 2.1839,
      "theoretical_loss": 3.4774169366467116,
      "tokens_seen": 1703018496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444734202607823,
      "loss": 2.4535,
      "theoretical_loss": 3.4774054235757346,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444633901705115,
      "loss": 2.5987,
      "theoretical_loss": 3.4773939110718244,
      "tokens_seen": 1703149568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024445336008024074,
      "loss": 2.5461,
      "theoretical_loss": 3.4773823991349317,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444433299899699,
      "loss": 2.6098,
      "theoretical_loss": 3.4773708877650065,
      "tokens_seen": 1703280640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444332998996991,
      "loss": 2.7719,
      "theoretical_loss": 3.477359376961999,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444232698094283,
      "loss": 2.5652,
      "theoretical_loss": 3.477347866725859,
      "tokens_seen": 1703411712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024441323971915747,
      "loss": 2.5395,
      "theoretical_loss": 3.4773363570565374,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024440320962888665,
      "loss": 2.5395,
      "theoretical_loss": 3.4773248479539847,
      "tokens_seen": 1703542784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024439317953861583,
      "loss": 2.5668,
      "theoretical_loss": 3.47731333941815,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244383149448345,
      "loss": 2.5848,
      "theoretical_loss": 3.477301831448985,
      "tokens_seen": 1703673856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443731193580742,
      "loss": 2.4327,
      "theoretical_loss": 3.4772903240464395,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443630892678034,
      "loss": 2.5092,
      "theoretical_loss": 3.477278817210463,
      "tokens_seen": 1703804928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443530591775326,
      "loss": 2.6484,
      "theoretical_loss": 3.4772673109410075,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 940087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.732879877090454,
      "objective/train/theoretical_loss": 3.4772558052380216,
      "objective/train/tokens_used": 1724396000,
      "theoretical_loss": 3.4772558052380216,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443430290872618,
      "loss": 2.7232,
      "theoretical_loss": 3.4772558052380216,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024433299899699097,
      "loss": 2.7388,
      "theoretical_loss": 3.4772443001014564,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024432296890672015,
      "loss": 2.4559,
      "theoretical_loss": 3.4772327955312625,
      "tokens_seen": 1704067072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443129388164494,
      "loss": 2.6903,
      "theoretical_loss": 3.47722129152739,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024430290872617857,
      "loss": 2.6098,
      "theoretical_loss": 3.477209788089789,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024429287863590775,
      "loss": 2.6533,
      "theoretical_loss": 3.4771982852184102,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024428284854563693,
      "loss": 2.6222,
      "theoretical_loss": 3.4771867829132033,
      "tokens_seen": 1704329216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442728184553661,
      "loss": 2.4471,
      "theoretical_loss": 3.4771752811741194,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442627883650953,
      "loss": 2.5374,
      "theoretical_loss": 3.477163780001109,
      "tokens_seen": 1704460288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442527582748245,
      "loss": 2.557,
      "theoretical_loss": 3.477152279394122,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024424272818455365,
      "loss": 2.5493,
      "theoretical_loss": 3.4771407793531086,
      "tokens_seen": 1704591360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024423269809428284,
      "loss": 2.5952,
      "theoretical_loss": 3.4771292798780196,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024422266800401207,
      "loss": 2.7427,
      "theoretical_loss": 3.477117780968805,
      "tokens_seen": 1704722432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024421263791374125,
      "loss": 2.3476,
      "theoretical_loss": 3.4771062826254155,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024420260782347043,
      "loss": 2.7885,
      "theoretical_loss": 3.477094784847802,
      "tokens_seen": 1704853504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441925777331996,
      "loss": 2.7579,
      "theoretical_loss": 3.477083287635914,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441825476429288,
      "loss": 2.8088,
      "theoretical_loss": 3.4770717909897026,
      "tokens_seen": 1704984576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244172517552658,
      "loss": 2.5156,
      "theoretical_loss": 3.4770602949091174,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024416248746238716,
      "loss": 2.7632,
      "theoretical_loss": 3.47704879939411,
      "tokens_seen": 1705115648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024415245737211634,
      "loss": 2.6471,
      "theoretical_loss": 3.4770373044446297,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441424272818456,
      "loss": 2.4826,
      "theoretical_loss": 3.4770258100606277,
      "tokens_seen": 1705246720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024413239719157473,
      "loss": 2.4252,
      "theoretical_loss": 3.4770143162420544,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441223671013039,
      "loss": 2.9122,
      "theoretical_loss": 3.47700282298886,
      "tokens_seen": 1705377792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024411233701103312,
      "loss": 2.7656,
      "theoretical_loss": 3.4769913303009945,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441023069207623,
      "loss": 2.4158,
      "theoretical_loss": 3.4769798381784094,
      "tokens_seen": 1705508864
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 940725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8139090538024902,
      "objective/train/theoretical_loss": 3.4769683466210544,
      "objective/train/tokens_used": 1726034400,
      "theoretical_loss": 3.4769683466210544,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024409227683049148,
      "loss": 2.4095,
      "theoretical_loss": 3.4769683466210544,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024408224674022066,
      "loss": 2.7397,
      "theoretical_loss": 3.47695685562888,
      "tokens_seen": 1705639936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024407221664994987,
      "loss": 2.7071,
      "theoretical_loss": 3.4769453652018374,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024406218655967905,
      "loss": 2.6431,
      "theoretical_loss": 3.476933875339877,
      "tokens_seen": 1705771008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024405215646940823,
      "loss": 2.8041,
      "theoretical_loss": 3.4769223860429483,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002440421263791374,
      "loss": 2.765,
      "theoretical_loss": 3.4769108973110026,
      "tokens_seen": 1705902080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002440320962888666,
      "loss": 2.5028,
      "theoretical_loss": 3.47689940914399,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002440220661985958,
      "loss": 2.4679,
      "theoretical_loss": 3.476887921541862,
      "tokens_seen": 1706033152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024401203610832498,
      "loss": 2.3727,
      "theoretical_loss": 3.4768764345045677,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024400200601805416,
      "loss": 2.623,
      "theoretical_loss": 3.476864948032059,
      "tokens_seen": 1706164224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024399197592778335,
      "loss": 2.6275,
      "theoretical_loss": 3.4768534621242853,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024398194583751255,
      "loss": 2.4081,
      "theoretical_loss": 3.4768419767811976,
      "tokens_seen": 1706295296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024397191574724173,
      "loss": 2.4711,
      "theoretical_loss": 3.4768304920027466,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024396188565697092,
      "loss": 2.6891,
      "theoretical_loss": 3.4768190077888823,
      "tokens_seen": 1706426368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002439518555667001,
      "loss": 2.6241,
      "theoretical_loss": 3.4768075241395566,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024394182547642928,
      "loss": 2.7675,
      "theoretical_loss": 3.4767960410547185,
      "tokens_seen": 1706557440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024393179538615849,
      "loss": 2.5211,
      "theoretical_loss": 3.4767845585343196,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024392176529588767,
      "loss": 2.6717,
      "theoretical_loss": 3.4767730765783096,
      "tokens_seen": 1706688512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024391173520561685,
      "loss": 2.5578,
      "theoretical_loss": 3.47676159518664,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024390170511534603,
      "loss": 2.4561,
      "theoretical_loss": 3.4767501143592607,
      "tokens_seen": 1706819584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024389167502507524,
      "loss": 2.6173,
      "theoretical_loss": 3.4767386340961233,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024388164493480442,
      "loss": 2.5474,
      "theoretical_loss": 3.476727154397177,
      "tokens_seen": 1706950656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438716148445336,
      "loss": 2.7327,
      "theoretical_loss": 3.4767156752623736,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024386158475426278,
      "loss": 2.742,
      "theoretical_loss": 3.4767041966916628,
      "tokens_seen": 1707081728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024385155466399196,
      "loss": 2.5371,
      "theoretical_loss": 3.476692718684996,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 941668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7325565814971924,
      "objective/train/theoretical_loss": 3.476681241242323,
      "objective/train/tokens_used": 1727672800,
      "theoretical_loss": 3.476681241242323,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024384152457372117,
      "loss": 2.609,
      "theoretical_loss": 3.476681241242323,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024383149448345035,
      "loss": 2.6995,
      "theoretical_loss": 3.4766697643635958,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024382146439317953,
      "loss": 2.7118,
      "theoretical_loss": 3.4766582880487635,
      "tokens_seen": 1707343872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024381143430290871,
      "loss": 2.5457,
      "theoretical_loss": 3.4766468122977776,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024380140421263792,
      "loss": 2.5599,
      "theoretical_loss": 3.4766353371105887,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437913741223671,
      "loss": 2.6873,
      "theoretical_loss": 3.4766238624871475,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024378134403209628,
      "loss": 2.7123,
      "theoretical_loss": 3.4766123884274043,
      "tokens_seen": 1707606016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024377131394182547,
      "loss": 2.6008,
      "theoretical_loss": 3.4766009149313097,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024376128385155467,
      "loss": 2.5601,
      "theoretical_loss": 3.476589441998815,
      "tokens_seen": 1707737088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024375125376128385,
      "loss": 2.6408,
      "theoretical_loss": 3.4765779696298704,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024374122367101304,
      "loss": 2.54,
      "theoretical_loss": 3.476566497824427,
      "tokens_seen": 1707868160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024373119358074222,
      "loss": 2.6129,
      "theoretical_loss": 3.476555026582435,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437211634904714,
      "loss": 2.8281,
      "theoretical_loss": 3.476543555903846,
      "tokens_seen": 1707999232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437111334002006,
      "loss": 2.6857,
      "theoretical_loss": 3.4765320857886097,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437011033099298,
      "loss": 2.7261,
      "theoretical_loss": 3.476520616236677,
      "tokens_seen": 1708130304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024369107321965897,
      "loss": 2.5562,
      "theoretical_loss": 3.476509147247999,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024368104312938815,
      "loss": 2.3866,
      "theoretical_loss": 3.4764976788225264,
      "tokens_seen": 1708261376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024367101303911736,
      "loss": 2.4865,
      "theoretical_loss": 3.4764862109602097,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024366098294884654,
      "loss": 2.4337,
      "theoretical_loss": 3.4764747436609995,
      "tokens_seen": 1708392448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024365095285857572,
      "loss": 2.4247,
      "theoretical_loss": 3.476463276924847,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436409227683049,
      "loss": 2.3957,
      "theoretical_loss": 3.476451810751703,
      "tokens_seen": 1708523520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024363089267803408,
      "loss": 2.6425,
      "theoretical_loss": 3.4764403451415173,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436208625877633,
      "loss": 2.6872,
      "theoretical_loss": 3.476428880094242,
      "tokens_seen": 1708654592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436108324974925,
      "loss": 2.702,
      "theoretical_loss": 3.476417415609827,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024360080240722168,
      "loss": 2.6886,
      "theoretical_loss": 3.476405951688223,
      "tokens_seen": 1708785664
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 943105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6183764934539795,
      "objective/train/theoretical_loss": 3.476394488329382,
      "objective/train/tokens_used": 1729311200,
      "theoretical_loss": 3.476394488329382,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024359077231695086,
      "loss": 2.6029,
      "theoretical_loss": 3.476394488329382,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024358074222668007,
      "loss": 2.5277,
      "theoretical_loss": 3.476383025533253,
      "tokens_seen": 1708916736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024357071213640925,
      "loss": 2.4737,
      "theoretical_loss": 3.4763715632997885,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024356068204613843,
      "loss": 2.5663,
      "theoretical_loss": 3.476360101628938,
      "tokens_seen": 1709047808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002435506519558676,
      "loss": 2.6589,
      "theoretical_loss": 3.476348640520653,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002435406218655968,
      "loss": 2.3847,
      "theoretical_loss": 3.476337179974884,
      "tokens_seen": 1709178880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000243530591775326,
      "loss": 2.6175,
      "theoretical_loss": 3.476325719991582,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024352056168505518,
      "loss": 2.6375,
      "theoretical_loss": 3.476314260570698,
      "tokens_seen": 1709309952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024351053159478436,
      "loss": 2.5795,
      "theoretical_loss": 3.4763028017121824,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024350050150451355,
      "loss": 2.5274,
      "theoretical_loss": 3.4762913434159866,
      "tokens_seen": 1709441024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024349047141424275,
      "loss": 2.7297,
      "theoretical_loss": 3.476279885682061,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024348044132397193,
      "loss": 2.6506,
      "theoretical_loss": 3.4762684285103562,
      "tokens_seen": 1709572096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024347041123370112,
      "loss": 2.7007,
      "theoretical_loss": 3.476256971900824,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002434603811434303,
      "loss": 2.6157,
      "theoretical_loss": 3.4762455158534147,
      "tokens_seen": 1709703168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024345035105315948,
      "loss": 2.3517,
      "theoretical_loss": 3.476234060368079,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024344032096288869,
      "loss": 2.69,
      "theoretical_loss": 3.4762226054447676,
      "tokens_seen": 1709834240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024343029087261787,
      "loss": 2.7542,
      "theoretical_loss": 3.476211151083432,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024342026078234705,
      "loss": 2.6273,
      "theoretical_loss": 3.4761996972840232,
      "tokens_seen": 1709965312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024341023069207623,
      "loss": 2.731,
      "theoretical_loss": 3.476188244046491,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024340020060180544,
      "loss": 2.5136,
      "theoretical_loss": 3.4761767913707877,
      "tokens_seen": 1710096384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024339017051153462,
      "loss": 2.8703,
      "theoretical_loss": 3.4761653392568634,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002433801404212638,
      "loss": 2.4538,
      "theoretical_loss": 3.4761538877046694,
      "tokens_seen": 1710227456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024337011033099298,
      "loss": 2.4375,
      "theoretical_loss": 3.476142436714156,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024336008024072216,
      "loss": 2.5465,
      "theoretical_loss": 3.4761309862852747,
      "tokens_seen": 1710358528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024335005015045137,
      "loss": 2.9604,
      "theoretical_loss": 3.4761195364179764,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 943708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.687971591949463,
      "objective/train/theoretical_loss": 3.4761080871122116,
      "objective/train/tokens_used": 1730949600,
      "theoretical_loss": 3.4761080871122116,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024334002006018055,
      "loss": 2.4426,
      "theoretical_loss": 3.4761080871122116,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024332998996990973,
      "loss": 2.5618,
      "theoretical_loss": 3.476096638367932,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024331995987963891,
      "loss": 2.7271,
      "theoretical_loss": 3.4760851901850875,
      "tokens_seen": 1710620672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024330992978936812,
      "loss": 2.4714,
      "theoretical_loss": 3.4760737425636297,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432998996990973,
      "loss": 2.6132,
      "theoretical_loss": 3.4760622955035103,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024328986960882648,
      "loss": 2.6443,
      "theoretical_loss": 3.476050849004679,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024327983951855567,
      "loss": 2.4785,
      "theoretical_loss": 3.4760394030670874,
      "tokens_seen": 1710882816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024326980942828487,
      "loss": 2.6931,
      "theoretical_loss": 3.476027957690686,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024325977933801405,
      "loss": 2.509,
      "theoretical_loss": 3.476016512875426,
      "tokens_seen": 1711013888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024324974924774324,
      "loss": 2.378,
      "theoretical_loss": 3.4760050686212596,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024323971915747242,
      "loss": 2.4291,
      "theoretical_loss": 3.475993624928136,
      "tokens_seen": 1711144960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432296890672016,
      "loss": 2.4729,
      "theoretical_loss": 3.475982181796007,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432196589769308,
      "loss": 2.581,
      "theoretical_loss": 3.475970739224824,
      "tokens_seen": 1711276032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024320962888666,
      "loss": 2.5046,
      "theoretical_loss": 3.475959297214537,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024319959879638917,
      "loss": 2.6216,
      "theoretical_loss": 3.475947855765098,
      "tokens_seen": 1711407104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024318956870611835,
      "loss": 2.3524,
      "theoretical_loss": 3.4759364148764575,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024317953861584756,
      "loss": 2.6154,
      "theoretical_loss": 3.4759249745485667,
      "tokens_seen": 1711538176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024316950852557674,
      "loss": 2.6872,
      "theoretical_loss": 3.4759135347813768,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024315947843530592,
      "loss": 2.4596,
      "theoretical_loss": 3.4759020955748388,
      "tokens_seen": 1711669248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431494483450351,
      "loss": 2.5897,
      "theoretical_loss": 3.4758906569289034,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024313941825476428,
      "loss": 2.5229,
      "theoretical_loss": 3.475879218843522,
      "tokens_seen": 1711800320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431293881644935,
      "loss": 2.7495,
      "theoretical_loss": 3.475867781318646,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024311935807422267,
      "loss": 2.5802,
      "theoretical_loss": 3.4758563443542254,
      "tokens_seen": 1711931392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024310932798395185,
      "loss": 2.7255,
      "theoretical_loss": 3.4758449079502123,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024309929789368103,
      "loss": 2.6508,
      "theoretical_loss": 3.475833472106557,
      "tokens_seen": 1712062464
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 944906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.916214942932129,
      "objective/train/theoretical_loss": 3.475822036823212,
      "objective/train/tokens_used": 1732588000,
      "theoretical_loss": 3.475822036823212,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024308926780341024,
      "loss": 2.8881,
      "theoretical_loss": 3.475822036823212,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024307923771313942,
      "loss": 2.6715,
      "theoretical_loss": 3.4758106021001267,
      "tokens_seen": 1712193536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002430692076228686,
      "loss": 2.5697,
      "theoretical_loss": 3.4757991679372533,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024305917753259779,
      "loss": 2.6192,
      "theoretical_loss": 3.4757877343345425,
      "tokens_seen": 1712324608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024304914744232697,
      "loss": 2.5016,
      "theoretical_loss": 3.4757763012919454,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024303911735205618,
      "loss": 3.0842,
      "theoretical_loss": 3.475764868809413,
      "tokens_seen": 1712455680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024302908726178536,
      "loss": 2.5743,
      "theoretical_loss": 3.4757534368868974,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024301905717151454,
      "loss": 2.6028,
      "theoretical_loss": 3.4757420055243484,
      "tokens_seen": 1712586752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024300902708124372,
      "loss": 2.6452,
      "theoretical_loss": 3.4757305747217178,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024299899699097293,
      "loss": 2.7392,
      "theoretical_loss": 3.4757191444789566,
      "tokens_seen": 1712717824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429889669007021,
      "loss": 2.6204,
      "theoretical_loss": 3.475707714796016,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429789368104313,
      "loss": 2.7439,
      "theoretical_loss": 3.4756962856728473,
      "tokens_seen": 1712848896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024296890672016047,
      "loss": 2.7054,
      "theoretical_loss": 3.4756848571094014,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024295887662988965,
      "loss": 2.6256,
      "theoretical_loss": 3.4756734291056297,
      "tokens_seen": 1712979968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024294884653961886,
      "loss": 2.7622,
      "theoretical_loss": 3.475662001661483,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024293881644934804,
      "loss": 2.4348,
      "theoretical_loss": 3.475650574776914,
      "tokens_seen": 1713111040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024292878635907722,
      "loss": 2.7899,
      "theoretical_loss": 3.4756391484518714,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429187562688064,
      "loss": 2.6638,
      "theoretical_loss": 3.4756277226863084,
      "tokens_seen": 1713242112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429087261785356,
      "loss": 2.6766,
      "theoretical_loss": 3.475616297480175,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002428986960882648,
      "loss": 2.4978,
      "theoretical_loss": 3.475604872833424,
      "tokens_seen": 1713373184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024288866599799397,
      "loss": 2.8022,
      "theoretical_loss": 3.4755934487460047,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024287863590772315,
      "loss": 2.7067,
      "theoretical_loss": 3.475582025217869,
      "tokens_seen": 1713504256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024286860581745234,
      "loss": 2.6932,
      "theoretical_loss": 3.4755706022489687,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024285857572718157,
      "loss": 2.5511,
      "theoretical_loss": 3.475559179839254,
      "tokens_seen": 1713635328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024284854563691075,
      "loss": 2.6009,
      "theoretical_loss": 3.4755477579886778,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 945396,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3279669284820557,
      "objective/train/theoretical_loss": 3.4755363366971896,
      "objective/train/tokens_used": 1734226400,
      "theoretical_loss": 3.4755363366971896,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024283851554663993,
      "loss": 2.3165,
      "theoretical_loss": 3.4755363366971896,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024282848545636911,
      "loss": 2.7255,
      "theoretical_loss": 3.4755249159647414,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024281845536609832,
      "loss": 2.445,
      "theoretical_loss": 3.4755134957912848,
      "tokens_seen": 1713897472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002428084252758275,
      "loss": 2.7067,
      "theoretical_loss": 3.4755020761767703,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024279839518555668,
      "loss": 2.4756,
      "theoretical_loss": 3.47549065712115,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024278836509528587,
      "loss": 2.8024,
      "theoretical_loss": 3.4754792386243745,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024277833500501507,
      "loss": 2.606,
      "theoretical_loss": 3.4754678206863954,
      "tokens_seen": 1714159616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024276830491474426,
      "loss": 2.5686,
      "theoretical_loss": 3.475456403307164,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024275827482447344,
      "loss": 2.5749,
      "theoretical_loss": 3.4754449864866315,
      "tokens_seen": 1714290688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024274824473420262,
      "loss": 2.4543,
      "theoretical_loss": 3.475433570224749,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427382146439318,
      "loss": 2.7523,
      "theoretical_loss": 3.475422154521468,
      "tokens_seen": 1714421760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242728184553661,
      "loss": 2.5017,
      "theoretical_loss": 3.47541073937674,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427181544633902,
      "loss": 2.6512,
      "theoretical_loss": 3.475399324790516,
      "tokens_seen": 1714552832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024270812437311937,
      "loss": 2.686,
      "theoretical_loss": 3.4753879107627483,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024269809428284855,
      "loss": 2.4969,
      "theoretical_loss": 3.4753764972933867,
      "tokens_seen": 1714683904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024268806419257776,
      "loss": 2.7049,
      "theoretical_loss": 3.4753650843823833,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024267803410230694,
      "loss": 2.4908,
      "theoretical_loss": 3.47535367202969,
      "tokens_seen": 1714814976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024266800401203612,
      "loss": 3.0286,
      "theoretical_loss": 3.475342260235257,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002426579739217653,
      "loss": 2.6383,
      "theoretical_loss": 3.475330848999036,
      "tokens_seen": 1714946048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024264794383149448,
      "loss": 2.6461,
      "theoretical_loss": 3.4753194383209793,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002426379137412237,
      "loss": 2.3834,
      "theoretical_loss": 3.475308028201037,
      "tokens_seen": 1715077120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024262788365095287,
      "loss": 2.6578,
      "theoretical_loss": 3.4752966186391614,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024261785356068205,
      "loss": 2.5254,
      "theoretical_loss": 3.4752852096353033,
      "tokens_seen": 1715208192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024260782347041123,
      "loss": 2.6649,
      "theoretical_loss": 3.4752738011894144,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024259779338014044,
      "loss": 2.606,
      "theoretical_loss": 3.4752623933014464,
      "tokens_seen": 1715339264
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 946450,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.933450698852539,
      "objective/train/theoretical_loss": 3.47525098597135,
      "objective/train/tokens_used": 1735864800,
      "theoretical_loss": 3.47525098597135,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024258776328986962,
      "loss": 2.544,
      "theoretical_loss": 3.47525098597135,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002425777331995988,
      "loss": 2.6517,
      "theoretical_loss": 3.475239579199077,
      "tokens_seen": 1715470336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024256770310932799,
      "loss": 2.5794,
      "theoretical_loss": 3.4752281729845786,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024255767301905717,
      "loss": 2.5423,
      "theoretical_loss": 3.4752167673278063,
      "tokens_seen": 1715601408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024254764292878638,
      "loss": 2.6021,
      "theoretical_loss": 3.475205362228712,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024253761283851556,
      "loss": 2.7333,
      "theoretical_loss": 3.475193957687246,
      "tokens_seen": 1715732480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024252758274824474,
      "loss": 2.5979,
      "theoretical_loss": 3.4751825537033607,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024251755265797392,
      "loss": 2.861,
      "theoretical_loss": 3.4751711502770077,
      "tokens_seen": 1715863552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024250752256770313,
      "loss": 2.3576,
      "theoretical_loss": 3.4751597474081377,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424974924774323,
      "loss": 2.5181,
      "theoretical_loss": 3.4751483450967027,
      "tokens_seen": 1715994624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424874623871615,
      "loss": 2.6962,
      "theoretical_loss": 3.4751369433426538,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024247743229689067,
      "loss": 2.4864,
      "theoretical_loss": 3.4751255421459426,
      "tokens_seen": 1716125696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024246740220661985,
      "loss": 2.6699,
      "theoretical_loss": 3.4751141415065208,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024245737211634906,
      "loss": 2.6684,
      "theoretical_loss": 3.4751027414243394,
      "tokens_seen": 1716256768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024244734202607824,
      "loss": 2.5157,
      "theoretical_loss": 3.4750913418993505,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024243731193580742,
      "loss": 2.6409,
      "theoretical_loss": 3.4750799429315054,
      "tokens_seen": 1716387840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424272818455366,
      "loss": 2.5722,
      "theoretical_loss": 3.475068544520755,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424172517552658,
      "loss": 2.6126,
      "theoretical_loss": 3.4750571466670515,
      "tokens_seen": 1716518912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242407221664995,
      "loss": 2.5443,
      "theoretical_loss": 3.475045749370346,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024239719157472417,
      "loss": 2.5288,
      "theoretical_loss": 3.4750343526305905,
      "tokens_seen": 1716649984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024238716148445335,
      "loss": 2.6954,
      "theoretical_loss": 3.475022956447736,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024237713139418254,
      "loss": 2.6162,
      "theoretical_loss": 3.4750115608217342,
      "tokens_seen": 1716781056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024236710130391174,
      "loss": 2.5241,
      "theoretical_loss": 3.4750001657525367,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024235707121364092,
      "loss": 2.6473,
      "theoretical_loss": 3.4749887712400955,
      "tokens_seen": 1716912128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423470411233701,
      "loss": 2.7635,
      "theoretical_loss": 3.474977377284361,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 947070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.163306951522827,
      "objective/train/theoretical_loss": 3.4749659838852858,
      "objective/train/tokens_used": 1737503200,
      "theoretical_loss": 3.4749659838852858,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423370110330993,
      "loss": 2.537,
      "theoretical_loss": 3.4749659838852858,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423269809428285,
      "loss": 2.6883,
      "theoretical_loss": 3.4749545910428212,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024231695085255768,
      "loss": 2.768,
      "theoretical_loss": 3.474943198756918,
      "tokens_seen": 1717174272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024230692076228686,
      "loss": 2.426,
      "theoretical_loss": 3.474931807027529,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024229689067201604,
      "loss": 2.5521,
      "theoretical_loss": 3.474920415854605,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024228686058174522,
      "loss": 2.5798,
      "theoretical_loss": 3.474909025238098,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024227683049147443,
      "loss": 2.6848,
      "theoretical_loss": 3.474897635177959,
      "tokens_seen": 1717436416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422668004012036,
      "loss": 2.3961,
      "theoretical_loss": 3.4748862456741403,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422567703109328,
      "loss": 2.5408,
      "theoretical_loss": 3.4748748567265926,
      "tokens_seen": 1717567488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024224674022066197,
      "loss": 2.4365,
      "theoretical_loss": 3.474863468335269,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024223671013039118,
      "loss": 2.4768,
      "theoretical_loss": 3.4748520805001193,
      "tokens_seen": 1717698560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024222668004012036,
      "loss": 2.6012,
      "theoretical_loss": 3.474840693221096,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024221664994984954,
      "loss": 2.4457,
      "theoretical_loss": 3.4748293064981515,
      "tokens_seen": 1717829632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024220661985957872,
      "loss": 2.5483,
      "theoretical_loss": 3.474817920331236,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024219658976930793,
      "loss": 2.6537,
      "theoretical_loss": 3.4748065347203023,
      "tokens_seen": 1717960704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421865596790371,
      "loss": 2.7603,
      "theoretical_loss": 3.474795149665301,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421765295887663,
      "loss": 2.7552,
      "theoretical_loss": 3.474783765166185,
      "tokens_seen": 1718091776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024216649949849547,
      "loss": 2.7307,
      "theoretical_loss": 3.4747723812229045,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024215646940822466,
      "loss": 2.3599,
      "theoretical_loss": 3.4747609978354124,
      "tokens_seen": 1718222848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024214643931795386,
      "loss": 2.7262,
      "theoretical_loss": 3.4747496150036596,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024213640922768305,
      "loss": 2.4759,
      "theoretical_loss": 3.4747382327275984,
      "tokens_seen": 1718353920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024212637913741223,
      "loss": 2.343,
      "theoretical_loss": 3.47472685100718,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024211634904714143,
      "loss": 2.3133,
      "theoretical_loss": 3.474715469842356,
      "tokens_seen": 1718484992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024210631895687064,
      "loss": 2.5425,
      "theoretical_loss": 3.4747040892330787,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024209628886659982,
      "loss": 2.6143,
      "theoretical_loss": 3.474692709179299,
      "tokens_seen": 1718616064
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 948336,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2460803985595703,
      "objective/train/theoretical_loss": 3.474681329680969,
      "objective/train/tokens_used": 1739141600,
      "theoretical_loss": 3.474681329680969,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242086258776329,
      "loss": 2.6059,
      "theoretical_loss": 3.474681329680969,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024207622868605819,
      "loss": 2.5097,
      "theoretical_loss": 3.4746699507380407,
      "tokens_seen": 1718747136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024206619859578737,
      "loss": 2.5548,
      "theoretical_loss": 3.4746585723504655,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024205616850551658,
      "loss": 2.6538,
      "theoretical_loss": 3.4746471945181954,
      "tokens_seen": 1718878208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024204613841524576,
      "loss": 2.2675,
      "theoretical_loss": 3.4746358172411815,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024203610832497494,
      "loss": 2.3277,
      "theoretical_loss": 3.474624440519376,
      "tokens_seen": 1719009280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024202607823470412,
      "loss": 2.6448,
      "theoretical_loss": 3.4746130643527304,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024201604814443333,
      "loss": 2.515,
      "theoretical_loss": 3.474601688741197,
      "tokens_seen": 1719140352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002420060180541625,
      "loss": 2.5759,
      "theoretical_loss": 3.474590313684727,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002419959879638917,
      "loss": 2.4871,
      "theoretical_loss": 3.474578939183272,
      "tokens_seen": 1719271424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024198595787362087,
      "loss": 2.515,
      "theoretical_loss": 3.474567565236785,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024197592778335005,
      "loss": 2.4587,
      "theoretical_loss": 3.474556191845216,
      "tokens_seen": 1719402496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024196589769307926,
      "loss": 2.5038,
      "theoretical_loss": 3.4745448190085177,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024195586760280844,
      "loss": 2.3805,
      "theoretical_loss": 3.4745334467266424,
      "tokens_seen": 1719533568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024194583751253762,
      "loss": 2.6151,
      "theoretical_loss": 3.4745220749995407,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002419358074222668,
      "loss": 2.5913,
      "theoretical_loss": 3.474510703827165,
      "tokens_seen": 1719664640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241925777331996,
      "loss": 2.4791,
      "theoretical_loss": 3.474499333209468,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002419157472417252,
      "loss": 2.5494,
      "theoretical_loss": 3.4744879631464,
      "tokens_seen": 1719795712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024190571715145437,
      "loss": 2.5839,
      "theoretical_loss": 3.474476593637913,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024189568706118355,
      "loss": 2.8499,
      "theoretical_loss": 3.47446522468396,
      "tokens_seen": 1719926784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024188565697091274,
      "loss": 2.3887,
      "theoretical_loss": 3.4744538562844913,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024187562688064194,
      "loss": 2.3999,
      "theoretical_loss": 3.4744424884394602,
      "tokens_seen": 1720057856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024186559679037113,
      "loss": 2.6067,
      "theoretical_loss": 3.4744311211488172,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418555667001003,
      "loss": 2.7142,
      "theoretical_loss": 3.4744197544125153,
      "tokens_seen": 1720188928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418455366098295,
      "loss": 2.438,
      "theoretical_loss": 3.4744083882305055,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 948940,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4110143184661865,
      "objective/train/theoretical_loss": 3.4743970226027403,
      "objective/train/tokens_used": 1740780000,
      "theoretical_loss": 3.4743970226027403,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418355065195587,
      "loss": 2.3539,
      "theoretical_loss": 3.4743970226027403,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024182547642928788,
      "loss": 2.7353,
      "theoretical_loss": 3.474385657529171,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024181544633901706,
      "loss": 2.4576,
      "theoretical_loss": 3.47437429300975,
      "tokens_seen": 1720451072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024180541624874624,
      "loss": 2.383,
      "theoretical_loss": 3.4743629290444282,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024179538615847542,
      "loss": 2.6433,
      "theoretical_loss": 3.4743515656331594,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024178535606820463,
      "loss": 2.5357,
      "theoretical_loss": 3.474340202775893,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002417753259779338,
      "loss": 2.5863,
      "theoretical_loss": 3.474328840472583,
      "tokens_seen": 1720713216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241765295887663,
      "loss": 2.4882,
      "theoretical_loss": 3.47431747872318,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024175526579739217,
      "loss": 2.7621,
      "theoretical_loss": 3.4743061175276364,
      "tokens_seen": 1720844288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024174523570712138,
      "loss": 2.6733,
      "theoretical_loss": 3.474294756885904,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024173520561685056,
      "loss": 2.4585,
      "theoretical_loss": 3.4742833967979347,
      "tokens_seen": 1720975360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024172517552657974,
      "loss": 2.7095,
      "theoretical_loss": 3.474272037263681,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024171514543630892,
      "loss": 2.7055,
      "theoretical_loss": 3.474260678283094,
      "tokens_seen": 1721106432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024170511534603813,
      "loss": 2.6254,
      "theoretical_loss": 3.474249319856126,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416950852557673,
      "loss": 2.6802,
      "theoretical_loss": 3.4742379619827286,
      "tokens_seen": 1721237504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416850551654965,
      "loss": 2.7691,
      "theoretical_loss": 3.4742266046628543,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024167502507522567,
      "loss": 2.6447,
      "theoretical_loss": 3.4742152478964545,
      "tokens_seen": 1721368576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024166499498495486,
      "loss": 2.5993,
      "theoretical_loss": 3.474203891683482,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024165496489468406,
      "loss": 2.6238,
      "theoretical_loss": 3.474192536023888,
      "tokens_seen": 1721499648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024164493480441325,
      "loss": 2.6828,
      "theoretical_loss": 3.474181180917624,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024163490471414243,
      "loss": 2.5706,
      "theoretical_loss": 3.474169826364643,
      "tokens_seen": 1721630720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416248746238716,
      "loss": 2.7476,
      "theoretical_loss": 3.4741584723648975,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024161484453360082,
      "loss": 2.5268,
      "theoretical_loss": 3.4741471189183377,
      "tokens_seen": 1721761792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024160481444333,
      "loss": 2.7344,
      "theoretical_loss": 3.474135766024917,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024159478435305918,
      "loss": 2.4882,
      "theoretical_loss": 3.4741244136845864,
      "tokens_seen": 1721892864
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 949989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3777735233306885,
      "objective/train/theoretical_loss": 3.474113061897299,
      "objective/train/tokens_used": 1742418400,
      "theoretical_loss": 3.474113061897299,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024158475426278836,
      "loss": 2.4078,
      "theoretical_loss": 3.474113061897299,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024157472417251754,
      "loss": 2.7387,
      "theoretical_loss": 3.474101710663006,
      "tokens_seen": 1722023936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024156469408224675,
      "loss": 2.7039,
      "theoretical_loss": 3.474090359981659,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024155466399197593,
      "loss": 2.6227,
      "theoretical_loss": 3.4740790098532113,
      "tokens_seen": 1722155008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415446339017051,
      "loss": 2.6943,
      "theoretical_loss": 3.4740676602776146,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415346038114343,
      "loss": 2.5623,
      "theoretical_loss": 3.47405631125482,
      "tokens_seen": 1722286080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415245737211635,
      "loss": 2.6723,
      "theoretical_loss": 3.4740449627847805,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024151454363089268,
      "loss": 2.4402,
      "theoretical_loss": 3.4740336148674476,
      "tokens_seen": 1722417152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024150451354062186,
      "loss": 2.5038,
      "theoretical_loss": 3.4740222675027734,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024149448345035104,
      "loss": 2.4965,
      "theoretical_loss": 3.474010920690711,
      "tokens_seen": 1722548224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024148445336008022,
      "loss": 2.5898,
      "theoretical_loss": 3.4739995744312107,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024147442326980943,
      "loss": 2.6954,
      "theoretical_loss": 3.4739882287242256,
      "tokens_seen": 1722679296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024146439317953861,
      "loss": 2.5603,
      "theoretical_loss": 3.473976883569708,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002414543630892678,
      "loss": 2.5534,
      "theoretical_loss": 3.47396553896761,
      "tokens_seen": 1722810368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024144433299899698,
      "loss": 2.503,
      "theoretical_loss": 3.4739541949178827,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024143430290872618,
      "loss": 2.4113,
      "theoretical_loss": 3.4739428514204787,
      "tokens_seen": 1722941440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024142427281845537,
      "loss": 2.5192,
      "theoretical_loss": 3.47393150847535,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024141424272818455,
      "loss": 2.5449,
      "theoretical_loss": 3.4739201660824497,
      "tokens_seen": 1723072512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024140421263791373,
      "loss": 2.7661,
      "theoretical_loss": 3.473908824241729,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413941825476429,
      "loss": 2.4586,
      "theoretical_loss": 3.4738974829531397,
      "tokens_seen": 1723203584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024138415245737212,
      "loss": 2.574,
      "theoretical_loss": 3.4738861422166343,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413741223671013,
      "loss": 2.5505,
      "theoretical_loss": 3.4738748020321655,
      "tokens_seen": 1723334656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413640922768305,
      "loss": 2.3738,
      "theoretical_loss": 3.4738634623996845,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413540621865597,
      "loss": 2.6772,
      "theoretical_loss": 3.473852123319144,
      "tokens_seen": 1723465728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413440320962889,
      "loss": 2.5526,
      "theoretical_loss": 3.4738407847904966,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 951217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6206398010253906,
      "objective/train/theoretical_loss": 3.4738294468136934,
      "objective/train/tokens_used": 1744056800,
      "theoretical_loss": 3.4738294468136934,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024133400200601808,
      "loss": 2.713,
      "theoretical_loss": 3.4738294468136934,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024132397191574726,
      "loss": 2.6853,
      "theoretical_loss": 3.473818109388687,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024131394182547644,
      "loss": 2.5389,
      "theoretical_loss": 3.47380677251543,
      "tokens_seen": 1723727872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024130391173520562,
      "loss": 2.8079,
      "theoretical_loss": 3.473795436193874,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024129388164493483,
      "loss": 2.4931,
      "theoretical_loss": 3.473784100423971,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000241283851554664,
      "loss": 2.5109,
      "theoretical_loss": 3.4737727652056742,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412738214643932,
      "loss": 2.5765,
      "theoretical_loss": 3.4737614305389344,
      "tokens_seen": 1723990016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024126379137412237,
      "loss": 2.6163,
      "theoretical_loss": 3.473750096423705,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024125376128385158,
      "loss": 2.5448,
      "theoretical_loss": 3.473738762859938,
      "tokens_seen": 1724121088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024124373119358076,
      "loss": 2.6093,
      "theoretical_loss": 3.4737274298475844,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024123370110330994,
      "loss": 2.4478,
      "theoretical_loss": 3.4737160973865984,
      "tokens_seen": 1724252160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024122367101303912,
      "loss": 2.6115,
      "theoretical_loss": 3.4737047654769304,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024121364092276833,
      "loss": 2.5017,
      "theoretical_loss": 3.473693434118534,
      "tokens_seen": 1724383232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412036108324975,
      "loss": 2.4799,
      "theoretical_loss": 3.4736821033113605,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411935807422267,
      "loss": 2.494,
      "theoretical_loss": 3.4736707730553626,
      "tokens_seen": 1724514304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024118355065195587,
      "loss": 2.4488,
      "theoretical_loss": 3.473659443350493,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024117352056168506,
      "loss": 2.5301,
      "theoretical_loss": 3.4736481141967026,
      "tokens_seen": 1724645376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024116349047141426,
      "loss": 2.6448,
      "theoretical_loss": 3.4736367855939445,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024115346038114345,
      "loss": 2.5604,
      "theoretical_loss": 3.4736254575421706,
      "tokens_seen": 1724776448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024114343029087263,
      "loss": 2.4759,
      "theoretical_loss": 3.4736141300413337,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411334002006018,
      "loss": 2.7043,
      "theoretical_loss": 3.4736028030913864,
      "tokens_seen": 1724907520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024112337011033102,
      "loss": 2.6217,
      "theoretical_loss": 3.47359147669228,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411133400200602,
      "loss": 2.6303,
      "theoretical_loss": 3.473580150843967,
      "tokens_seen": 1725038592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024110330992978938,
      "loss": 2.7545,
      "theoretical_loss": 3.4735688255464003,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024109327983951856,
      "loss": 2.5203,
      "theoretical_loss": 3.4735575007995316,
      "tokens_seen": 1725169664
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 951846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6185786724090576,
      "objective/train/theoretical_loss": 3.4735461766033136,
      "objective/train/tokens_used": 1745695200,
      "theoretical_loss": 3.4735461766033136,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024108324974924774,
      "loss": 2.7447,
      "theoretical_loss": 3.4735461766033136,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024107321965897695,
      "loss": 2.5946,
      "theoretical_loss": 3.473534852957698,
      "tokens_seen": 1725300736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024106318956870613,
      "loss": 2.6764,
      "theoretical_loss": 3.4735235298626375,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410531594784353,
      "loss": 2.5529,
      "theoretical_loss": 3.473512207318085,
      "tokens_seen": 1725431808
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410431293881645,
      "loss": 2.5411,
      "theoretical_loss": 3.473500885323992,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410330992978937,
      "loss": 2.6779,
      "theoretical_loss": 3.473489563880311,
      "tokens_seen": 1725562880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024102306920762288,
      "loss": 2.7473,
      "theoretical_loss": 3.473478242986994,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024101303911735206,
      "loss": 2.4551,
      "theoretical_loss": 3.4734669226439943,
      "tokens_seen": 1725693952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024100300902708124,
      "loss": 2.7682,
      "theoretical_loss": 3.473455602851264,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024099297893681042,
      "loss": 2.6126,
      "theoretical_loss": 3.473444283608755,
      "tokens_seen": 1725825024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024098294884653963,
      "loss": 2.5969,
      "theoretical_loss": 3.4734329649164195,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024097291875626881,
      "loss": 2.4276,
      "theoretical_loss": 3.4734216467742103,
      "tokens_seen": 1725956096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240962888665998,
      "loss": 2.576,
      "theoretical_loss": 3.47341032918208,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024095285857572718,
      "loss": 2.6256,
      "theoretical_loss": 3.4733990121399807,
      "tokens_seen": 1726087168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024094282848545638,
      "loss": 2.6355,
      "theoretical_loss": 3.4733876956478644,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024093279839518557,
      "loss": 2.5382,
      "theoretical_loss": 3.4733763797056842,
      "tokens_seen": 1726218240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024092276830491475,
      "loss": 2.7617,
      "theoretical_loss": 3.473365064313392,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024091273821464393,
      "loss": 2.6143,
      "theoretical_loss": 3.4733537494709408,
      "tokens_seen": 1726349312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002409027081243731,
      "loss": 2.4009,
      "theoretical_loss": 3.4733424351782816,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024089267803410232,
      "loss": 2.7579,
      "theoretical_loss": 3.4733311214353684,
      "tokens_seen": 1726480384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408826479438315,
      "loss": 2.509,
      "theoretical_loss": 3.4733198082421533,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024087261785356068,
      "loss": 2.5887,
      "theoretical_loss": 3.473308495598588,
      "tokens_seen": 1726611456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024086258776328986,
      "loss": 2.7457,
      "theoretical_loss": 3.4732971835046254,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024085255767301907,
      "loss": 2.6556,
      "theoretical_loss": 3.4732858719602175,
      "tokens_seen": 1726742528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024084252758274825,
      "loss": 2.4722,
      "theoretical_loss": 3.4732745609653177,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 953107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.782886505126953,
      "objective/train/theoretical_loss": 3.473263250519878,
      "objective/train/tokens_used": 1747333600,
      "theoretical_loss": 3.473263250519878,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024083249749247743,
      "loss": 2.6671,
      "theoretical_loss": 3.473263250519878,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408224674022066,
      "loss": 2.5493,
      "theoretical_loss": 3.4732519406238507,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408124373119358,
      "loss": 2.7184,
      "theoretical_loss": 3.4732406312771884,
      "tokens_seen": 1727004672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240802407221665,
      "loss": 2.4421,
      "theoretical_loss": 3.473229322479843,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024079237713139418,
      "loss": 2.5123,
      "theoretical_loss": 3.473218014231768,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024078234704112336,
      "loss": 2.632,
      "theoretical_loss": 3.473206706532915,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024077231695085254,
      "loss": 2.6943,
      "theoretical_loss": 3.473195399383237,
      "tokens_seen": 1727266816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024076228686058175,
      "loss": 2.6104,
      "theoretical_loss": 3.4731840927826867,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024075225677031093,
      "loss": 2.6378,
      "theoretical_loss": 3.473172786731216,
      "tokens_seen": 1727397888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024074222668004012,
      "loss": 2.8157,
      "theoretical_loss": 3.4731614812287774,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002407321965897693,
      "loss": 2.6568,
      "theoretical_loss": 3.4731501762753236,
      "tokens_seen": 1727528960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024072216649949848,
      "loss": 2.6077,
      "theoretical_loss": 3.4731388718708076,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024071213640922769,
      "loss": 2.5182,
      "theoretical_loss": 3.4731275680151814,
      "tokens_seen": 1727660032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024070210631895687,
      "loss": 2.7086,
      "theoretical_loss": 3.4731162647083975,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024069207622868605,
      "loss": 2.6064,
      "theoretical_loss": 3.4731049619504084,
      "tokens_seen": 1727791104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024068204613841523,
      "loss": 2.4214,
      "theoretical_loss": 3.473093659741167,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024067201604814444,
      "loss": 2.7116,
      "theoretical_loss": 3.4730823580806254,
      "tokens_seen": 1727922176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024066198595787362,
      "loss": 2.4066,
      "theoretical_loss": 3.473071056968737,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002406519558676028,
      "loss": 2.5823,
      "theoretical_loss": 3.473059756405453,
      "tokens_seen": 1728053248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024064192577733198,
      "loss": 2.5607,
      "theoretical_loss": 3.4730484563907273,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024063189568706116,
      "loss": 2.8311,
      "theoretical_loss": 3.4730371569245113,
      "tokens_seen": 1728184320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024062186559679037,
      "loss": 2.5769,
      "theoretical_loss": 3.473025858006759,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024061183550651958,
      "loss": 2.288,
      "theoretical_loss": 3.4730145596374213,
      "tokens_seen": 1728315392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024060180541624876,
      "loss": 2.7486,
      "theoretical_loss": 3.4730032618164524,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024059177532597794,
      "loss": 2.7213,
      "theoretical_loss": 3.4729919645438034,
      "tokens_seen": 1728446464
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 953628,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.81854510307312,
      "objective/train/theoretical_loss": 3.472980667819428,
      "objective/train/tokens_used": 1748972000,
      "theoretical_loss": 3.472980667819428,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024058174523570715,
      "loss": 2.6384,
      "theoretical_loss": 3.472980667819428,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024057171514543633,
      "loss": 2.4379,
      "theoretical_loss": 3.4729693716432783,
      "tokens_seen": 1728577536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405616850551655,
      "loss": 2.6984,
      "theoretical_loss": 3.472958076015307,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405516549648947,
      "loss": 2.5979,
      "theoretical_loss": 3.472946780935467,
      "tokens_seen": 1728708608
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405416248746239,
      "loss": 2.5205,
      "theoretical_loss": 3.4729354864037107,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024053159478435308,
      "loss": 2.4842,
      "theoretical_loss": 3.4729241924199905,
      "tokens_seen": 1728839680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024052156469408226,
      "loss": 2.5631,
      "theoretical_loss": 3.4729128989842595,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024051153460381144,
      "loss": 2.7279,
      "theoretical_loss": 3.4729016060964697,
      "tokens_seen": 1728970752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024050150451354062,
      "loss": 2.6394,
      "theoretical_loss": 3.4728903137565745,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024049147442326983,
      "loss": 2.4192,
      "theoretical_loss": 3.472879021964526,
      "tokens_seen": 1729101824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024048144433299901,
      "loss": 2.6708,
      "theoretical_loss": 3.472867730720277,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404714142427282,
      "loss": 2.4806,
      "theoretical_loss": 3.4728564400237802,
      "tokens_seen": 1729232896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024046138415245738,
      "loss": 2.6215,
      "theoretical_loss": 3.4728451498749884,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024045135406218658,
      "loss": 2.6727,
      "theoretical_loss": 3.4728338602738544,
      "tokens_seen": 1729363968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024044132397191577,
      "loss": 2.5882,
      "theoretical_loss": 3.4728225712203304,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024043129388164495,
      "loss": 2.5031,
      "theoretical_loss": 3.4728112827143693,
      "tokens_seen": 1729495040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024042126379137413,
      "loss": 2.6659,
      "theoretical_loss": 3.472799994755924,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404112337011033,
      "loss": 2.436,
      "theoretical_loss": 3.4727887073449466,
      "tokens_seen": 1729626112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024040120361083252,
      "loss": 2.6131,
      "theoretical_loss": 3.4727774204813904,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403911735205617,
      "loss": 2.6324,
      "theoretical_loss": 3.472766134165208,
      "tokens_seen": 1729757184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024038114343029088,
      "loss": 2.6855,
      "theoretical_loss": 3.472754848396352,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024037111334002006,
      "loss": 2.6523,
      "theoretical_loss": 3.4727435631747756,
      "tokens_seen": 1729888256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024036108324974927,
      "loss": 2.661,
      "theoretical_loss": 3.4727322785004304,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024035105315947845,
      "loss": 2.512,
      "theoretical_loss": 3.4727209943732706,
      "tokens_seen": 1730019328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024034102306920763,
      "loss": 2.8287,
      "theoretical_loss": 3.4727097107932474,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 954813,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0557727813720703,
      "objective/train/theoretical_loss": 3.4726984277603146,
      "objective/train/tokens_used": 1750610400,
      "theoretical_loss": 3.4726984277603146,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403309929789368,
      "loss": 2.8815,
      "theoretical_loss": 3.4726984277603146,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240320962888666,
      "loss": 2.5249,
      "theoretical_loss": 3.472687145274425,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403109327983952,
      "loss": 2.5654,
      "theoretical_loss": 3.472675863335531,
      "tokens_seen": 1730281472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024030090270812438,
      "loss": 2.5693,
      "theoretical_loss": 3.472664581943585,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024029087261785356,
      "loss": 2.6933,
      "theoretical_loss": 3.47265330109854,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024028084252758274,
      "loss": 2.8645,
      "theoretical_loss": 3.4726420208003494,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024027081243731195,
      "loss": 2.3471,
      "theoretical_loss": 3.472630741048966,
      "tokens_seen": 1730543616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024026078234704113,
      "loss": 2.4877,
      "theoretical_loss": 3.4726194618443413,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024025075225677032,
      "loss": 2.5954,
      "theoretical_loss": 3.4726081831864293,
      "tokens_seen": 1730674688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002402407221664995,
      "loss": 2.4906,
      "theoretical_loss": 3.472596905075182,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024023069207622868,
      "loss": 2.9431,
      "theoretical_loss": 3.4725856275105533,
      "tokens_seen": 1730805760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024022066198595789,
      "loss": 2.5397,
      "theoretical_loss": 3.472574350492495,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024021063189568707,
      "loss": 2.4545,
      "theoretical_loss": 3.47256307402096,
      "tokens_seen": 1730936832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024020060180541625,
      "loss": 2.5122,
      "theoretical_loss": 3.4725517980959015,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024019057171514543,
      "loss": 2.5027,
      "theoretical_loss": 3.4725405227172725,
      "tokens_seen": 1731067904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024018054162487464,
      "loss": 2.4191,
      "theoretical_loss": 3.472529247885025,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024017051153460382,
      "loss": 2.7288,
      "theoretical_loss": 3.472517973599113,
      "tokens_seen": 1731198976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000240160481444333,
      "loss": 2.4399,
      "theoretical_loss": 3.4725066998594882,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024015045135406218,
      "loss": 2.5115,
      "theoretical_loss": 3.4724954266661046,
      "tokens_seen": 1731330048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024014042126379136,
      "loss": 2.5681,
      "theoretical_loss": 3.4724841540189137,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024013039117352057,
      "loss": 2.6397,
      "theoretical_loss": 3.472472881917869,
      "tokens_seen": 1731461120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024012036108324975,
      "loss": 2.5213,
      "theoretical_loss": 3.4724616103629242,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024011033099297893,
      "loss": 2.5727,
      "theoretical_loss": 3.4724503393540314,
      "tokens_seen": 1731592192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024010030090270811,
      "loss": 2.3874,
      "theoretical_loss": 3.472439068891143,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024009027081243732,
      "loss": 2.7607,
      "theoretical_loss": 3.4724277989742127,
      "tokens_seen": 1731723264
    },
    {
      "epoch": 0.52,
      "objective/train/docs_used": 956095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2430784702301025,
      "objective/train/theoretical_loss": 3.4724165296031924,
      "objective/train/tokens_used": 1752248800,
      "theoretical_loss": 3.4724165296031924,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002400802407221665,
      "loss": 2.2866,
      "theoretical_loss": 3.4724165296031924,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024007021063189568,
      "loss": 2.7156,
      "theoretical_loss": 3.4724052607780367,
      "tokens_seen": 1731854336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024006018054162487,
      "loss": 2.6898,
      "theoretical_loss": 3.472393992498697,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024005015045135407,
      "loss": 2.6742,
      "theoretical_loss": 3.4723827247651267,
      "tokens_seen": 1731985408
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024004012036108325,
      "loss": 2.4651,
      "theoretical_loss": 3.472371457577279,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024003009027081244,
      "loss": 2.5455,
      "theoretical_loss": 3.472360190935106,
      "tokens_seen": 1732116480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024002006018054162,
      "loss": 2.1577,
      "theoretical_loss": 3.472348924838562,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002400100300902708,
      "loss": 2.6005,
      "theoretical_loss": 3.4723376592875983,
      "tokens_seen": 1732247552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024,
      "loss": 2.559,
      "theoretical_loss": 3.4723263942821694,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002399899699097292,
      "loss": 2.4253,
      "theoretical_loss": 3.472315129822227,
      "tokens_seen": 1732378624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00023997993981945837,
      "loss": 2.8419,
      "theoretical_loss": 3.4723038659077248,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023996990972918755,
      "loss": 2.405,
      "theoretical_loss": 3.4722926025386154,
      "tokens_seen": 1732509696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023995987963891676,
      "loss": 2.6941,
      "theoretical_loss": 3.4722813397148515,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023994984954864594,
      "loss": 2.6867,
      "theoretical_loss": 3.472270077436387,
      "tokens_seen": 1732640768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023993981945837512,
      "loss": 2.6475,
      "theoretical_loss": 3.4722588157031744,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002399297893681043,
      "loss": 2.5124,
      "theoretical_loss": 3.472247554515166,
      "tokens_seen": 1732771840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023991975927783348,
      "loss": 2.4132,
      "theoretical_loss": 3.4722362938723155,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002399097291875627,
      "loss": 2.4379,
      "theoretical_loss": 3.4722250337745764,
      "tokens_seen": 1732902912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023989969909729187,
      "loss": 2.701,
      "theoretical_loss": 3.472213774221901,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023988966900702105,
      "loss": 2.4589,
      "theoretical_loss": 3.4722025152142413,
      "tokens_seen": 1733033984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023987963891675023,
      "loss": 2.7256,
      "theoretical_loss": 3.472191256751552,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023986960882647944,
      "loss": 2.4173,
      "theoretical_loss": 3.472179998833786,
      "tokens_seen": 1733165056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023985957873620865,
      "loss": 2.6302,
      "theoretical_loss": 3.4721687414608953,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023984954864593783,
      "loss": 2.7278,
      "theoretical_loss": 3.4721574846328336,
      "tokens_seen": 1733296128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239839518555667,
      "loss": 2.4839,
      "theoretical_loss": 3.472146228349554,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 956643,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 1.9616906642913818,
      "objective/train/theoretical_loss": 3.472134972611009,
      "objective/train/tokens_used": 1753887200,
      "theoretical_loss": 3.472134972611009,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002398294884653962,
      "loss": 2.4031,
      "theoretical_loss": 3.472134972611009,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002398194583751254,
      "loss": 2.603,
      "theoretical_loss": 3.472123717417152,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023980942828485458,
      "loss": 2.4512,
      "theoretical_loss": 3.472112462767936,
      "tokens_seen": 1733558272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023979939819458376,
      "loss": 2.3502,
      "theoretical_loss": 3.4721012086633145,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023978936810431295,
      "loss": 2.655,
      "theoretical_loss": 3.4720899551032396,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023977933801404215,
      "loss": 2.7359,
      "theoretical_loss": 3.472078702087665,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023976930792377133,
      "loss": 2.521,
      "theoretical_loss": 3.472067449616544,
      "tokens_seen": 1733820416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023975927783350052,
      "loss": 2.5113,
      "theoretical_loss": 3.4720561976898288,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002397492477432297,
      "loss": 2.5476,
      "theoretical_loss": 3.4720449463074736,
      "tokens_seen": 1733951488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023973921765295888,
      "loss": 2.6874,
      "theoretical_loss": 3.47203369546943,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023972918756268809,
      "loss": 2.649,
      "theoretical_loss": 3.472022445175653,
      "tokens_seen": 1734082560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023971915747241727,
      "loss": 2.6367,
      "theoretical_loss": 3.472011195426095,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023970912738214645,
      "loss": 2.5268,
      "theoretical_loss": 3.4719999462207083,
      "tokens_seen": 1734213632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023969909729187563,
      "loss": 2.475,
      "theoretical_loss": 3.4719886975594463,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023968906720160484,
      "loss": 2.6148,
      "theoretical_loss": 3.471977449442263,
      "tokens_seen": 1734344704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023967903711133402,
      "loss": 2.5215,
      "theoretical_loss": 3.47196620186911,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002396690070210632,
      "loss": 2.5159,
      "theoretical_loss": 3.471954954839942,
      "tokens_seen": 1734475776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023965897693079238,
      "loss": 2.4031,
      "theoretical_loss": 3.4719437083547113,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023964894684052156,
      "loss": 2.682,
      "theoretical_loss": 3.4719324624133714,
      "tokens_seen": 1734606848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023963891675025077,
      "loss": 2.493,
      "theoretical_loss": 3.4719212170158755,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023962888665997995,
      "loss": 2.4958,
      "theoretical_loss": 3.471909972162176,
      "tokens_seen": 1734737920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023961885656970913,
      "loss": 2.7389,
      "theoretical_loss": 3.471898727852227,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023960882647943831,
      "loss": 2.5442,
      "theoretical_loss": 3.471887484085981,
      "tokens_seen": 1734868992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023959879638916752,
      "loss": 2.5183,
      "theoretical_loss": 3.4718762408633914,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395887662988967,
      "loss": 2.5784,
      "theoretical_loss": 3.471864998184411,
      "tokens_seen": 1735000064
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 957795,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.660010576248169,
      "objective/train/theoretical_loss": 3.471853756048994,
      "objective/train/tokens_used": 1755525600,
      "theoretical_loss": 3.471853756048994,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023957873620862588,
      "loss": 2.6239,
      "theoretical_loss": 3.471853756048994,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023956870611835507,
      "loss": 2.4088,
      "theoretical_loss": 3.4718425144570926,
      "tokens_seen": 1735131136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023955867602808427,
      "loss": 2.3937,
      "theoretical_loss": 3.471831273408661,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023954864593781345,
      "loss": 2.5568,
      "theoretical_loss": 3.471820032903651,
      "tokens_seen": 1735262208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023953861584754264,
      "loss": 2.6499,
      "theoretical_loss": 3.471808792942017,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023952858575727182,
      "loss": 2.5586,
      "theoretical_loss": 3.4717975535237118,
      "tokens_seen": 1735393280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239518555667001,
      "loss": 2.6822,
      "theoretical_loss": 3.4717863146486887,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395085255767302,
      "loss": 2.5667,
      "theoretical_loss": 3.4717750763169004,
      "tokens_seen": 1735524352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394984954864594,
      "loss": 2.4325,
      "theoretical_loss": 3.471763838528301,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023948846539618857,
      "loss": 2.69,
      "theoretical_loss": 3.471752601282843,
      "tokens_seen": 1735655424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023947843530591775,
      "loss": 2.585,
      "theoretical_loss": 3.47174136458048,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023946840521564696,
      "loss": 2.4108,
      "theoretical_loss": 3.4717301284211657,
      "tokens_seen": 1735786496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023945837512537614,
      "loss": 2.5635,
      "theoretical_loss": 3.4717188928048524,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023944834503510532,
      "loss": 2.6549,
      "theoretical_loss": 3.4717076577314936,
      "tokens_seen": 1735917568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394383149448345,
      "loss": 2.541,
      "theoretical_loss": 3.471696423201043,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023942828485456368,
      "loss": 2.4685,
      "theoretical_loss": 3.4716851892134537,
      "tokens_seen": 1736048640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394182547642929,
      "loss": 2.6719,
      "theoretical_loss": 3.4716739557686793,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023940822467402207,
      "loss": 2.6612,
      "theoretical_loss": 3.471662722866672,
      "tokens_seen": 1736179712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023939819458375125,
      "loss": 2.7224,
      "theoretical_loss": 3.4716514905073863,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023938816449348043,
      "loss": 2.6527,
      "theoretical_loss": 3.471640258690775,
      "tokens_seen": 1736310784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023937813440320964,
      "loss": 2.5055,
      "theoretical_loss": 3.4716290274167907,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023936810431293882,
      "loss": 2.6461,
      "theoretical_loss": 3.471617796685388,
      "tokens_seen": 1736441856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239358074222668,
      "loss": 2.5646,
      "theoretical_loss": 3.4716065664965194,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023934804413239719,
      "loss": 2.7423,
      "theoretical_loss": 3.4715953368501387,
      "tokens_seen": 1736572928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023933801404212637,
      "loss": 2.6278,
      "theoretical_loss": 3.4715841077461986,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 958202,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4927234649658203,
      "objective/train/theoretical_loss": 3.471572879184653,
      "objective/train/tokens_used": 1757164000,
      "theoretical_loss": 3.471572879184653,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023932798395185557,
      "loss": 2.4453,
      "theoretical_loss": 3.471572879184653,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023931795386158476,
      "loss": 2.4793,
      "theoretical_loss": 3.4715616511654552,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023930792377131394,
      "loss": 2.6568,
      "theoretical_loss": 3.4715504236885577,
      "tokens_seen": 1736835072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023929789368104312,
      "loss": 2.4734,
      "theoretical_loss": 3.4715391967539153,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023928786359077233,
      "loss": 2.7245,
      "theoretical_loss": 3.47152797036148,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392778335005015,
      "loss": 2.6338,
      "theoretical_loss": 3.471516744511206,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392678034102307,
      "loss": 2.6537,
      "theoretical_loss": 3.471505519203046,
      "tokens_seen": 1737097216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023925777331995987,
      "loss": 2.7103,
      "theoretical_loss": 3.471494294436954,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023924774322968905,
      "loss": 2.6439,
      "theoretical_loss": 3.471483070212883,
      "tokens_seen": 1737228288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023923771313941826,
      "loss": 2.6977,
      "theoretical_loss": 3.471471846530787,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023922768304914744,
      "loss": 2.5287,
      "theoretical_loss": 3.471460623390618,
      "tokens_seen": 1737359360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023921765295887662,
      "loss": 2.688,
      "theoretical_loss": 3.4714494007923307,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392076228686058,
      "loss": 2.5712,
      "theoretical_loss": 3.471438178735878,
      "tokens_seen": 1737490432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239197592778335,
      "loss": 2.5644,
      "theoretical_loss": 3.4714269572212135,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391875626880642,
      "loss": 2.4586,
      "theoretical_loss": 3.4714157362482907,
      "tokens_seen": 1737621504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023917753259779337,
      "loss": 2.5314,
      "theoretical_loss": 3.4714045158170626,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023916750250752255,
      "loss": 2.4782,
      "theoretical_loss": 3.4713932959274825,
      "tokens_seen": 1737752576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023915747241725174,
      "loss": 2.518,
      "theoretical_loss": 3.4713820765795047,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023914744232698094,
      "loss": 2.497,
      "theoretical_loss": 3.4713708577730817,
      "tokens_seen": 1737883648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023913741223671012,
      "loss": 2.3968,
      "theoretical_loss": 3.471359639508168,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391273821464393,
      "loss": 2.6052,
      "theoretical_loss": 3.4713484217847155,
      "tokens_seen": 1738014720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391173520561685,
      "loss": 2.3803,
      "theoretical_loss": 3.471337204602679,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023910732196589772,
      "loss": 2.566,
      "theoretical_loss": 3.471325987962011,
      "tokens_seen": 1738145792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390972918756269,
      "loss": 2.7858,
      "theoretical_loss": 3.471314771862666,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023908726178535608,
      "loss": 2.7492,
      "theoretical_loss": 3.4713035563045964,
      "tokens_seen": 1738276864
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 959288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.888723134994507,
      "objective/train/theoretical_loss": 3.4712923412877563,
      "objective/train/tokens_used": 1758802400,
      "theoretical_loss": 3.4712923412877563,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023907723169508527,
      "loss": 2.6595,
      "theoretical_loss": 3.4712923412877563,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023906720160481447,
      "loss": 2.7036,
      "theoretical_loss": 3.4712811268120993,
      "tokens_seen": 1738407936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023905717151454365,
      "loss": 2.6806,
      "theoretical_loss": 3.4712699128775784,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023904714142427284,
      "loss": 2.4649,
      "theoretical_loss": 3.4712586994841477,
      "tokens_seen": 1738539008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023903711133400202,
      "loss": 2.594,
      "theoretical_loss": 3.4712474866317597,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390270812437312,
      "loss": 2.5804,
      "theoretical_loss": 3.4712362743203693,
      "tokens_seen": 1738670080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390170511534604,
      "loss": 2.7387,
      "theoretical_loss": 3.4712250625499284,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002390070210631896,
      "loss": 2.5868,
      "theoretical_loss": 3.471213851320392,
      "tokens_seen": 1738801152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023899699097291877,
      "loss": 2.4995,
      "theoretical_loss": 3.471202640631713,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023898696088264795,
      "loss": 2.488,
      "theoretical_loss": 3.4711914304838443,
      "tokens_seen": 1738932224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023897693079237716,
      "loss": 2.4439,
      "theoretical_loss": 3.4711802208767404,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023896690070210634,
      "loss": 2.4648,
      "theoretical_loss": 3.471169011810354,
      "tokens_seen": 1739063296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023895687061183552,
      "loss": 2.7214,
      "theoretical_loss": 3.4711578032846404,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002389468405215647,
      "loss": 2.5167,
      "theoretical_loss": 3.4711465952995506,
      "tokens_seen": 1739194368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023893681043129388,
      "loss": 2.5828,
      "theoretical_loss": 3.47113538785504,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002389267803410231,
      "loss": 2.5949,
      "theoretical_loss": 3.471124180951061,
      "tokens_seen": 1739325440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023891675025075227,
      "loss": 2.56,
      "theoretical_loss": 3.471112974587568,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023890672016048145,
      "loss": 2.7107,
      "theoretical_loss": 3.4711017687645147,
      "tokens_seen": 1739456512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023889669007021063,
      "loss": 2.5704,
      "theoretical_loss": 3.471090563481854,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023888665997993984,
      "loss": 2.4975,
      "theoretical_loss": 3.4710793587395394,
      "tokens_seen": 1739587584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023887662988966902,
      "loss": 2.5815,
      "theoretical_loss": 3.4710681545375253,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002388665997993982,
      "loss": 2.4019,
      "theoretical_loss": 3.4710569508757647,
      "tokens_seen": 1739718656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023885656970912739,
      "loss": 2.4155,
      "theoretical_loss": 3.471045747754211,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023884653961885657,
      "loss": 2.4413,
      "theoretical_loss": 3.4710345451728184,
      "tokens_seen": 1739849728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023883650952858577,
      "loss": 2.7468,
      "theoretical_loss": 3.4710233431315407,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 960545,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8378190994262695,
      "objective/train/theoretical_loss": 3.4710121416303306,
      "objective/train/tokens_used": 1760440800,
      "theoretical_loss": 3.4710121416303306,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023882647943831496,
      "loss": 2.6464,
      "theoretical_loss": 3.4710121416303306,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023881644934804414,
      "loss": 2.7746,
      "theoretical_loss": 3.471000940669142,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023880641925777332,
      "loss": 2.5691,
      "theoretical_loss": 3.470989740247929,
      "tokens_seen": 1740111872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023879638916750253,
      "loss": 2.7217,
      "theoretical_loss": 3.4709785403666453,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387863590772317,
      "loss": 2.4845,
      "theoretical_loss": 3.4709673410252435,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387763289869609,
      "loss": 2.612,
      "theoretical_loss": 3.4709561422236783,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023876629889669007,
      "loss": 2.3012,
      "theoretical_loss": 3.470944943961903,
      "tokens_seen": 1740374016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023875626880641925,
      "loss": 2.5959,
      "theoretical_loss": 3.4709337462398713,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023874623871614846,
      "loss": 2.6713,
      "theoretical_loss": 3.470922549057537,
      "tokens_seen": 1740505088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023873620862587764,
      "loss": 2.6496,
      "theoretical_loss": 3.470911352414853,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023872617853560682,
      "loss": 2.6419,
      "theoretical_loss": 3.4709001563117736,
      "tokens_seen": 1740636160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000238716148445336,
      "loss": 2.6642,
      "theoretical_loss": 3.470888960748253,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387061183550652,
      "loss": 2.5172,
      "theoretical_loss": 3.470877765724244,
      "tokens_seen": 1740767232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386960882647944,
      "loss": 2.4762,
      "theoretical_loss": 3.4708665712397004,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023868605817452357,
      "loss": 2.4927,
      "theoretical_loss": 3.4708553772945763,
      "tokens_seen": 1740898304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023867602808425275,
      "loss": 2.5429,
      "theoretical_loss": 3.4708441838888255,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023866599799398194,
      "loss": 2.5757,
      "theoretical_loss": 3.4708329910224007,
      "tokens_seen": 1741029376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023865596790371114,
      "loss": 2.3667,
      "theoretical_loss": 3.4708217986952574,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023864593781344032,
      "loss": 2.4257,
      "theoretical_loss": 3.470810606907347,
      "tokens_seen": 1741160448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386359077231695,
      "loss": 2.4215,
      "theoretical_loss": 3.4707994156586253,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386258776328987,
      "loss": 2.6773,
      "theoretical_loss": 3.470788224949045,
      "tokens_seen": 1741291520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386158475426279,
      "loss": 2.5886,
      "theoretical_loss": 3.47077703477856,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023860581745235708,
      "loss": 2.4887,
      "theoretical_loss": 3.470765845147124,
      "tokens_seen": 1741422592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023859578736208626,
      "loss": 2.4459,
      "theoretical_loss": 3.470754656054691,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023858575727181544,
      "loss": 2.5437,
      "theoretical_loss": 3.4707434675012148,
      "tokens_seen": 1741553664
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 961110,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9698407649993896,
      "objective/train/theoretical_loss": 3.470732279486649,
      "objective/train/tokens_used": 1762079200,
      "theoretical_loss": 3.470732279486649,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023857572718154462,
      "loss": 2.692,
      "theoretical_loss": 3.470732279486649,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023856569709127383,
      "loss": 2.548,
      "theoretical_loss": 3.4707210920109466,
      "tokens_seen": 1741684736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000238555667001003,
      "loss": 2.6071,
      "theoretical_loss": 3.470709905074062,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002385456369107322,
      "loss": 2.6604,
      "theoretical_loss": 3.47069871867595,
      "tokens_seen": 1741815808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023853560682046137,
      "loss": 2.675,
      "theoretical_loss": 3.470687532816563,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023852557673019058,
      "loss": 2.4003,
      "theoretical_loss": 3.4706763474958553,
      "tokens_seen": 1741946880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023851554663991976,
      "loss": 2.5626,
      "theoretical_loss": 3.4706651627137806,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023850551654964894,
      "loss": 2.4609,
      "theoretical_loss": 3.4706539784702928,
      "tokens_seen": 1742077952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023849548645937812,
      "loss": 2.6941,
      "theoretical_loss": 3.470642794765345,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023848545636910733,
      "loss": 2.472,
      "theoretical_loss": 3.4706316115988924,
      "tokens_seen": 1742209024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384754262788365,
      "loss": 2.6402,
      "theoretical_loss": 3.4706204289708875,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384653961885657,
      "loss": 2.6983,
      "theoretical_loss": 3.4706092468812852,
      "tokens_seen": 1742340096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023845536609829487,
      "loss": 2.5154,
      "theoretical_loss": 3.4705980653300386,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023844533600802406,
      "loss": 2.444,
      "theoretical_loss": 3.4705868843171017,
      "tokens_seen": 1742471168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023843530591775326,
      "loss": 2.3977,
      "theoretical_loss": 3.4705757038424285,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023842527582748244,
      "loss": 2.5143,
      "theoretical_loss": 3.4705645239059724,
      "tokens_seen": 1742602240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023841524573721163,
      "loss": 2.6591,
      "theoretical_loss": 3.470553344507688,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384052156469408,
      "loss": 2.8745,
      "theoretical_loss": 3.4705421656475286,
      "tokens_seen": 1742733312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023839518555667002,
      "loss": 2.5744,
      "theoretical_loss": 3.470530987325448,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002383851554663992,
      "loss": 2.6227,
      "theoretical_loss": 3.4705198095414005,
      "tokens_seen": 1742864384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023837512537612838,
      "loss": 2.4502,
      "theoretical_loss": 3.4705086322953393,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023836509528585756,
      "loss": 2.6179,
      "theoretical_loss": 3.4704974555872194,
      "tokens_seen": 1742995456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023835506519558677,
      "loss": 2.4324,
      "theoretical_loss": 3.4704862794169937,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023834503510531598,
      "loss": 2.6691,
      "theoretical_loss": 3.470475103784616,
      "tokens_seen": 1743126528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023833500501504516,
      "loss": 2.7274,
      "theoretical_loss": 3.470463928690041,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 962337,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.721731185913086,
      "objective/train/theoretical_loss": 3.4704527541332224,
      "objective/train/tokens_used": 1763717600,
      "theoretical_loss": 3.4704527541332224,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023832497492477434,
      "loss": 2.6965,
      "theoretical_loss": 3.4704527541332224,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023831494483450352,
      "loss": 2.6668,
      "theoretical_loss": 3.4704415801141133,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023830491474423273,
      "loss": 2.6952,
      "theoretical_loss": 3.4704304066326683,
      "tokens_seen": 1743388672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382948846539619,
      "loss": 2.6102,
      "theoretical_loss": 3.4704192336888413,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382848545636911,
      "loss": 2.6455,
      "theoretical_loss": 3.470408061282587,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023827482447342027,
      "loss": 2.6401,
      "theoretical_loss": 3.4703968894138573,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023826479438314945,
      "loss": 2.6242,
      "theoretical_loss": 3.470385718082608,
      "tokens_seen": 1743650816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023825476429287866,
      "loss": 2.4711,
      "theoretical_loss": 3.470374547288792,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023824473420260784,
      "loss": 2.457,
      "theoretical_loss": 3.470363377032364,
      "tokens_seen": 1743781888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023823470411233702,
      "loss": 2.4616,
      "theoretical_loss": 3.4703522073132773,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382246740220662,
      "loss": 2.4347,
      "theoretical_loss": 3.470341038131486,
      "tokens_seen": 1743912960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382146439317954,
      "loss": 2.7024,
      "theoretical_loss": 3.470329869486944,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382046138415246,
      "loss": 2.4734,
      "theoretical_loss": 3.470318701379606,
      "tokens_seen": 1744044032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023819458375125377,
      "loss": 2.4649,
      "theoretical_loss": 3.4703075338094256,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023818455366098295,
      "loss": 2.4543,
      "theoretical_loss": 3.4702963667763562,
      "tokens_seen": 1744175104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023817452357071214,
      "loss": 2.4734,
      "theoretical_loss": 3.470285200280352,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023816449348044134,
      "loss": 2.7676,
      "theoretical_loss": 3.470274034321368,
      "tokens_seen": 1744306176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023815446339017052,
      "loss": 2.6812,
      "theoretical_loss": 3.470262868899357,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381444332998997,
      "loss": 2.6326,
      "theoretical_loss": 3.470251704014273,
      "tokens_seen": 1744437248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381344032096289,
      "loss": 2.6317,
      "theoretical_loss": 3.4702405396660705,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381243731193581,
      "loss": 2.5154,
      "theoretical_loss": 3.4702293758547036,
      "tokens_seen": 1744568320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023811434302908728,
      "loss": 2.5716,
      "theoretical_loss": 3.470218212580126,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023810431293881646,
      "loss": 2.7223,
      "theoretical_loss": 3.4702070498422923,
      "tokens_seen": 1744699392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023809428284854564,
      "loss": 2.6792,
      "theoretical_loss": 3.4701958876411556,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023808425275827482,
      "loss": 2.7996,
      "theoretical_loss": 3.4701847259766705,
      "tokens_seen": 1744830464
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 963376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.288900136947632,
      "objective/train/theoretical_loss": 3.4701735648487912,
      "objective/train/tokens_used": 1765356000,
      "theoretical_loss": 3.4701735648487912,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023807422266800403,
      "loss": 2.5,
      "theoretical_loss": 3.4701735648487912,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002380641925777332,
      "loss": 2.4213,
      "theoretical_loss": 3.4701624042574717,
      "tokens_seen": 1744961536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002380541624874624,
      "loss": 2.5726,
      "theoretical_loss": 3.4701512442026656,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023804413239719157,
      "loss": 2.5282,
      "theoretical_loss": 3.470140084684327,
      "tokens_seen": 1745092608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023803410230692078,
      "loss": 2.3577,
      "theoretical_loss": 3.4701289257024106,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023802407221664996,
      "loss": 2.5778,
      "theoretical_loss": 3.47011776725687,
      "tokens_seen": 1745223680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023801404212637914,
      "loss": 2.8681,
      "theoretical_loss": 3.47010660934766,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023800401203610832,
      "loss": 2.5045,
      "theoretical_loss": 3.470095451974733,
      "tokens_seen": 1745354752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023799398194583753,
      "loss": 2.4261,
      "theoretical_loss": 3.4700842951380446,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002379839518555667,
      "loss": 2.5177,
      "theoretical_loss": 3.4700731388375483,
      "tokens_seen": 1745485824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002379739217652959,
      "loss": 2.6188,
      "theoretical_loss": 3.4700619830731982,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023796389167502507,
      "loss": 2.6394,
      "theoretical_loss": 3.470050827844949,
      "tokens_seen": 1745616896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023795386158475426,
      "loss": 2.5905,
      "theoretical_loss": 3.4700396731527543,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023794383149448346,
      "loss": 2.6539,
      "theoretical_loss": 3.470028518996568,
      "tokens_seen": 1745747968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023793380140421265,
      "loss": 2.6534,
      "theoretical_loss": 3.4700173653763446,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023792377131394183,
      "loss": 2.6156,
      "theoretical_loss": 3.470006212292038,
      "tokens_seen": 1745879040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237913741223671,
      "loss": 2.7123,
      "theoretical_loss": 3.469995059743603,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023790371113340022,
      "loss": 2.7134,
      "theoretical_loss": 3.4699839077309926,
      "tokens_seen": 1746010112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378936810431294,
      "loss": 2.3377,
      "theoretical_loss": 3.469972756254162,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023788365095285858,
      "loss": 2.8397,
      "theoretical_loss": 3.469961605313064,
      "tokens_seen": 1746141184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023787362086258776,
      "loss": 2.5112,
      "theoretical_loss": 3.4699504549076545,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023786359077231694,
      "loss": 2.6862,
      "theoretical_loss": 3.4699393050378866,
      "tokens_seen": 1746272256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023785356068204615,
      "loss": 2.6016,
      "theoretical_loss": 3.4699281557037147,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023784353059177533,
      "loss": 2.5413,
      "theoretical_loss": 3.469917006905093,
      "tokens_seen": 1746403328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378335005015045,
      "loss": 2.4943,
      "theoretical_loss": 3.4699058586419755,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 963934,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6536614894866943,
      "objective/train/theoretical_loss": 3.469894710914317,
      "objective/train/tokens_used": 1766994400,
      "theoretical_loss": 3.469894710914317,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378234704112337,
      "loss": 2.4004,
      "theoretical_loss": 3.469894710914317,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378134403209629,
      "loss": 2.378,
      "theoretical_loss": 3.46988356372207,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023780341023069208,
      "loss": 2.4076,
      "theoretical_loss": 3.469872417065191,
      "tokens_seen": 1746665472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023779338014042126,
      "loss": 2.711,
      "theoretical_loss": 3.4698612709436327,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023778335005015044,
      "loss": 2.6406,
      "theoretical_loss": 3.46985012535735,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023777331995987962,
      "loss": 2.6731,
      "theoretical_loss": 3.469838980306297,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023776328986960883,
      "loss": 2.6465,
      "theoretical_loss": 3.469827835790427,
      "tokens_seen": 1746927616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023775325977933801,
      "loss": 2.542,
      "theoretical_loss": 3.4698166918096955,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002377432296890672,
      "loss": 2.4247,
      "theoretical_loss": 3.469805548364056,
      "tokens_seen": 1747058688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023773319959879638,
      "loss": 2.7093,
      "theoretical_loss": 3.469794405453463,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023772316950852558,
      "loss": 2.5949,
      "theoretical_loss": 3.4697832630778707,
      "tokens_seen": 1747189760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023771313941825477,
      "loss": 2.6132,
      "theoretical_loss": 3.4697721212372334,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023770310932798395,
      "loss": 2.6592,
      "theoretical_loss": 3.4697609799315052,
      "tokens_seen": 1747320832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023769307923771313,
      "loss": 2.4826,
      "theoretical_loss": 3.4697498391606403,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002376830491474423,
      "loss": 2.7855,
      "theoretical_loss": 3.4697386989245933,
      "tokens_seen": 1747451904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023767301905717152,
      "loss": 2.4906,
      "theoretical_loss": 3.469727559223318,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002376629889669007,
      "loss": 2.5405,
      "theoretical_loss": 3.469716420056769,
      "tokens_seen": 1747582976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023765295887662988,
      "loss": 2.6671,
      "theoretical_loss": 3.469705281424901,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023764292878635906,
      "loss": 2.6468,
      "theoretical_loss": 3.4696941433276676,
      "tokens_seen": 1747714048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023763289869608827,
      "loss": 2.7697,
      "theoretical_loss": 3.469683005765023,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023762286860581745,
      "loss": 2.5424,
      "theoretical_loss": 3.4696718687369223,
      "tokens_seen": 1747845120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023761283851554663,
      "loss": 2.6695,
      "theoretical_loss": 3.469660732243319,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023760280842527584,
      "loss": 2.6265,
      "theoretical_loss": 3.4696495962841674,
      "tokens_seen": 1747976192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023759277833500502,
      "loss": 2.6374,
      "theoretical_loss": 3.469638460859423,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023758274824473423,
      "loss": 2.5867,
      "theoretical_loss": 3.4696273259690384,
      "tokens_seen": 1748107264
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 965147,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.321573257446289,
      "objective/train/theoretical_loss": 3.4696161916129693,
      "objective/train/tokens_used": 1768632800,
      "theoretical_loss": 3.4696161916129693,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002375727181544634,
      "loss": 2.556,
      "theoretical_loss": 3.4696161916129693,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002375626880641926,
      "loss": 2.7989,
      "theoretical_loss": 3.4696050577911692,
      "tokens_seen": 1748238336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023755265797392177,
      "loss": 2.5279,
      "theoretical_loss": 3.469593924503593,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023754262788365098,
      "loss": 2.7597,
      "theoretical_loss": 3.4695827917501947,
      "tokens_seen": 1748369408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023753259779338016,
      "loss": 2.5928,
      "theoretical_loss": 3.4695716595309287,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023752256770310934,
      "loss": 2.4489,
      "theoretical_loss": 3.4695605278457493,
      "tokens_seen": 1748500480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023751253761283852,
      "loss": 2.7546,
      "theoretical_loss": 3.469549396694611,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023750250752256773,
      "loss": 2.5856,
      "theoretical_loss": 3.4695382660774685,
      "tokens_seen": 1748631552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374924774322969,
      "loss": 2.7554,
      "theoretical_loss": 3.469527135994275,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374824473420261,
      "loss": 2.7072,
      "theoretical_loss": 3.4695160064449864,
      "tokens_seen": 1748762624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023747241725175527,
      "loss": 2.4636,
      "theoretical_loss": 3.469504877429556,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023746238716148446,
      "loss": 2.5358,
      "theoretical_loss": 3.4694937489479383,
      "tokens_seen": 1748893696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023745235707121366,
      "loss": 2.4916,
      "theoretical_loss": 3.4694826210000884,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023744232698094285,
      "loss": 2.7236,
      "theoretical_loss": 3.46947149358596,
      "tokens_seen": 1749024768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023743229689067203,
      "loss": 2.5232,
      "theoretical_loss": 3.4694603667055075,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374222668004012,
      "loss": 2.4911,
      "theoretical_loss": 3.469449240358686,
      "tokens_seen": 1749155840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023741223671013042,
      "loss": 2.5438,
      "theoretical_loss": 3.4694381145454494,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374022066198596,
      "loss": 2.5329,
      "theoretical_loss": 3.469426989265752,
      "tokens_seen": 1749286912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023739217652958878,
      "loss": 2.579,
      "theoretical_loss": 3.469415864519548,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023738214643931796,
      "loss": 2.6647,
      "theoretical_loss": 3.469404740306793,
      "tokens_seen": 1749417984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023737211634904714,
      "loss": 2.6402,
      "theoretical_loss": 3.469393616627441,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023736208625877635,
      "loss": 2.6538,
      "theoretical_loss": 3.469382493481445,
      "tokens_seen": 1749549056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023735205616850553,
      "loss": 2.469,
      "theoretical_loss": 3.4693713708687612,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373420260782347,
      "loss": 2.7122,
      "theoretical_loss": 3.469360248789343,
      "tokens_seen": 1749680128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373319959879639,
      "loss": 2.6052,
      "theoretical_loss": 3.4693491272431456,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 965776,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.915630578994751,
      "objective/train/theoretical_loss": 3.469338006230123,
      "objective/train/tokens_used": 1770271200,
      "theoretical_loss": 3.469338006230123,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373219658976931,
      "loss": 2.691,
      "theoretical_loss": 3.469338006230123,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023731193580742228,
      "loss": 2.5569,
      "theoretical_loss": 3.46932688575023,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023730190571715146,
      "loss": 2.6282,
      "theoretical_loss": 3.4693157658034206,
      "tokens_seen": 1749942272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023729187562688064,
      "loss": 2.7118,
      "theoretical_loss": 3.46930464638965,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023728184553660982,
      "loss": 2.7004,
      "theoretical_loss": 3.469293527508872,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023727181544633903,
      "loss": 2.6706,
      "theoretical_loss": 3.4692824091610412,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023726178535606821,
      "loss": 2.5025,
      "theoretical_loss": 3.4692712913461126,
      "tokens_seen": 1750204416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002372517552657974,
      "loss": 2.491,
      "theoretical_loss": 3.4692601740640403,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023724172517552658,
      "loss": 2.4942,
      "theoretical_loss": 3.4692490573147787,
      "tokens_seen": 1750335488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023723169508525578,
      "loss": 2.7064,
      "theoretical_loss": 3.4692379410982825,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023722166499498497,
      "loss": 2.5348,
      "theoretical_loss": 3.4692268254145064,
      "tokens_seen": 1750466560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023721163490471415,
      "loss": 2.5545,
      "theoretical_loss": 3.4692157102634047,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023720160481444333,
      "loss": 2.3571,
      "theoretical_loss": 3.469204595644932,
      "tokens_seen": 1750597632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371915747241725,
      "loss": 2.5971,
      "theoretical_loss": 3.4691934815590426,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023718154463390172,
      "loss": 2.5332,
      "theoretical_loss": 3.4691823680056917,
      "tokens_seen": 1750728704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371715145436309,
      "loss": 2.4639,
      "theoretical_loss": 3.4691712549848326,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023716148445336008,
      "loss": 2.7934,
      "theoretical_loss": 3.4691601424964214,
      "tokens_seen": 1750859776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023715145436308926,
      "loss": 2.4834,
      "theoretical_loss": 3.469149030540412,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023714142427281847,
      "loss": 2.7801,
      "theoretical_loss": 3.469137919116758,
      "tokens_seen": 1750990848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023713139418254765,
      "loss": 2.5962,
      "theoretical_loss": 3.4691268082254156,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023712136409227683,
      "loss": 2.7476,
      "theoretical_loss": 3.4691156978663384,
      "tokens_seen": 1751121920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237111334002006,
      "loss": 2.8098,
      "theoretical_loss": 3.4691045880394817,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371013039117352,
      "loss": 2.7205,
      "theoretical_loss": 3.469093478744799,
      "tokens_seen": 1751252992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002370912738214644,
      "loss": 2.637,
      "theoretical_loss": 3.4690823699822455,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023708124373119358,
      "loss": 2.6251,
      "theoretical_loss": 3.4690712617517763,
      "tokens_seen": 1751384064
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 966555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3122029304504395,
      "objective/train/theoretical_loss": 3.469060154053345,
      "objective/train/tokens_used": 1771909600,
      "theoretical_loss": 3.469060154053345,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023707121364092276,
      "loss": 2.5506,
      "theoretical_loss": 3.469060154053345,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023706118355065194,
      "loss": 2.404,
      "theoretical_loss": 3.469049046886907,
      "tokens_seen": 1751515136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023705115346038115,
      "loss": 2.8042,
      "theoretical_loss": 3.4690379402524165,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023704112337011033,
      "loss": 2.4619,
      "theoretical_loss": 3.4690268341498287,
      "tokens_seen": 1751646208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023703109327983952,
      "loss": 2.4878,
      "theoretical_loss": 3.4690157285790972,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002370210631895687,
      "loss": 2.6142,
      "theoretical_loss": 3.4690046235401777,
      "tokens_seen": 1751777280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023701103309929788,
      "loss": 2.724,
      "theoretical_loss": 3.468993519033024,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023700100300902709,
      "loss": 2.5327,
      "theoretical_loss": 3.4689824150575914,
      "tokens_seen": 1751908352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023699097291875627,
      "loss": 2.7701,
      "theoretical_loss": 3.468971311613834,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023698094282848545,
      "loss": 2.6324,
      "theoretical_loss": 3.4689602087017066,
      "tokens_seen": 1752039424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023697091273821463,
      "loss": 2.5166,
      "theoretical_loss": 3.468949106321164,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023696088264794384,
      "loss": 2.5226,
      "theoretical_loss": 3.468938004472161,
      "tokens_seen": 1752170496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023695085255767302,
      "loss": 2.6971,
      "theoretical_loss": 3.468926903154652,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002369408224674022,
      "loss": 2.4385,
      "theoretical_loss": 3.468915802368592,
      "tokens_seen": 1752301568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023693079237713138,
      "loss": 2.4263,
      "theoretical_loss": 3.4689047021139356,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023692076228686056,
      "loss": 2.3891,
      "theoretical_loss": 3.4688936023906374,
      "tokens_seen": 1752432640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023691073219658977,
      "loss": 2.6672,
      "theoretical_loss": 3.4688825031986514,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023690070210631895,
      "loss": 2.6065,
      "theoretical_loss": 3.468871404537934,
      "tokens_seen": 1752563712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023689067201604813,
      "loss": 2.5117,
      "theoretical_loss": 3.4688603064084376,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368806419257773,
      "loss": 2.7411,
      "theoretical_loss": 3.468849208810119,
      "tokens_seen": 1752694784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023687061183550652,
      "loss": 2.3069,
      "theoretical_loss": 3.4688381117429317,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023686058174523573,
      "loss": 2.4558,
      "theoretical_loss": 3.468827015206831,
      "tokens_seen": 1752825856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368505516549649,
      "loss": 2.8347,
      "theoretical_loss": 3.4688159192017713,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368405215646941,
      "loss": 2.7041,
      "theoretical_loss": 3.468804823727708,
      "tokens_seen": 1752956928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368304914744233,
      "loss": 2.5444,
      "theoretical_loss": 3.4687937287845947,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 967544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.570122241973877,
      "objective/train/theoretical_loss": 3.468782634372387,
      "objective/train/tokens_used": 1773548000,
      "theoretical_loss": 3.468782634372387,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023682046138415248,
      "loss": 2.5025,
      "theoretical_loss": 3.468782634372387,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023681043129388166,
      "loss": 2.8496,
      "theoretical_loss": 3.4687715404910398,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023680040120361084,
      "loss": 2.4839,
      "theoretical_loss": 3.4687604471405065,
      "tokens_seen": 1753219072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023679037111334002,
      "loss": 2.3205,
      "theoretical_loss": 3.4687493543207433,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023678034102306923,
      "loss": 2.5001,
      "theoretical_loss": 3.468738262031705,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023677031093279841,
      "loss": 2.5622,
      "theoretical_loss": 3.468727170273345,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002367602808425276,
      "loss": 2.5747,
      "theoretical_loss": 3.4687160790456195,
      "tokens_seen": 1753481216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023675025075225678,
      "loss": 2.8216,
      "theoretical_loss": 3.468704988348483,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023674022066198598,
      "loss": 2.2564,
      "theoretical_loss": 3.4686938981818893,
      "tokens_seen": 1753612288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023673019057171517,
      "loss": 2.3816,
      "theoretical_loss": 3.468682808545794,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023672016048144435,
      "loss": 2.4803,
      "theoretical_loss": 3.4686717194401524,
      "tokens_seen": 1753743360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023671013039117353,
      "loss": 2.4115,
      "theoretical_loss": 3.4686606308649184,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002367001003009027,
      "loss": 2.5712,
      "theoretical_loss": 3.468649542820047,
      "tokens_seen": 1753874432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023669007021063192,
      "loss": 2.7387,
      "theoretical_loss": 3.468638455305493,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366800401203611,
      "loss": 2.5265,
      "theoretical_loss": 3.4686273683212114,
      "tokens_seen": 1754005504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023667001003009028,
      "loss": 2.6911,
      "theoretical_loss": 3.468616281867157,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023665997993981946,
      "loss": 2.5217,
      "theoretical_loss": 3.468605195943285,
      "tokens_seen": 1754136576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023664994984954867,
      "loss": 2.6751,
      "theoretical_loss": 3.46859411054955,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023663991975927785,
      "loss": 2.5037,
      "theoretical_loss": 3.468583025685906,
      "tokens_seen": 1754267648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023662988966900703,
      "loss": 2.663,
      "theoretical_loss": 3.468571941352309,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366198595787362,
      "loss": 2.5354,
      "theoretical_loss": 3.4685608575487135,
      "tokens_seen": 1754398720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366098294884654,
      "loss": 2.621,
      "theoretical_loss": 3.4685497742750737,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365997993981946,
      "loss": 2.5943,
      "theoretical_loss": 3.4685386915313456,
      "tokens_seen": 1754529792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023658976930792378,
      "loss": 2.9202,
      "theoretical_loss": 3.4685276093174835,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023657973921765296,
      "loss": 2.6104,
      "theoretical_loss": 3.468516527633442,
      "tokens_seen": 1754660864
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 968097,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0997824668884277,
      "objective/train/theoretical_loss": 3.4685054464791767,
      "objective/train/tokens_used": 1775186400,
      "theoretical_loss": 3.4685054464791767,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023656970912738214,
      "loss": 2.424,
      "theoretical_loss": 3.4685054464791767,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023655967903711135,
      "loss": 2.6244,
      "theoretical_loss": 3.4684943658546414,
      "tokens_seen": 1754791936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023654964894684053,
      "loss": 2.754,
      "theoretical_loss": 3.4684832857597923,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023653961885656972,
      "loss": 2.5408,
      "theoretical_loss": 3.4684722061945834,
      "tokens_seen": 1754923008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365295887662989,
      "loss": 2.6838,
      "theoretical_loss": 3.46846112715897,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023651955867602808,
      "loss": 2.5606,
      "theoretical_loss": 3.4684500486529064,
      "tokens_seen": 1755054080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023650952858575729,
      "loss": 2.5871,
      "theoretical_loss": 3.468438970676348,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023649949849548647,
      "loss": 2.6585,
      "theoretical_loss": 3.4684278932292503,
      "tokens_seen": 1755185152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023648946840521565,
      "loss": 2.6231,
      "theoretical_loss": 3.468416816311567,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023647943831494483,
      "loss": 2.6954,
      "theoretical_loss": 3.4684057399232544,
      "tokens_seen": 1755316224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023646940822467404,
      "loss": 2.503,
      "theoretical_loss": 3.468394664064266,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023645937813440322,
      "loss": 2.8229,
      "theoretical_loss": 3.468383588734558,
      "tokens_seen": 1755447296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364493480441324,
      "loss": 2.616,
      "theoretical_loss": 3.468372513934084,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023643931795386158,
      "loss": 2.6403,
      "theoretical_loss": 3.468361439662801,
      "tokens_seen": 1755578368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023642928786359076,
      "loss": 2.5552,
      "theoretical_loss": 3.468350365920662,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023641925777331997,
      "loss": 2.6282,
      "theoretical_loss": 3.468339292707623,
      "tokens_seen": 1755709440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023640922768304915,
      "loss": 2.7027,
      "theoretical_loss": 3.468328220023638,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023639919759277833,
      "loss": 2.3839,
      "theoretical_loss": 3.468317147868663,
      "tokens_seen": 1755840512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363891675025075,
      "loss": 2.4802,
      "theoretical_loss": 3.4683060762426523,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023637913741223672,
      "loss": 2.5935,
      "theoretical_loss": 3.4682950051455617,
      "tokens_seen": 1755971584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363691073219659,
      "loss": 2.5608,
      "theoretical_loss": 3.4682839345773457,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023635907723169508,
      "loss": 2.8494,
      "theoretical_loss": 3.468272864537959,
      "tokens_seen": 1756102656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023634904714142426,
      "loss": 2.5348,
      "theoretical_loss": 3.468261795027357,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023633901705115347,
      "loss": 2.7235,
      "theoretical_loss": 3.4682507260454942,
      "tokens_seen": 1756233728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023632898696088265,
      "loss": 2.5387,
      "theoretical_loss": 3.4682396575923264,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 969105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.096782922744751,
      "objective/train/theoretical_loss": 3.4682285896678087,
      "objective/train/tokens_used": 1776824800,
      "theoretical_loss": 3.4682285896678087,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023631895687061184,
      "loss": 2.3963,
      "theoretical_loss": 3.4682285896678087,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023630892678034102,
      "loss": 2.6202,
      "theoretical_loss": 3.468217522271895,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362988966900702,
      "loss": 2.7269,
      "theoretical_loss": 3.468206455404541,
      "tokens_seen": 1756495872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362888665997994,
      "loss": 2.5339,
      "theoretical_loss": 3.468195389065702,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362788365095286,
      "loss": 2.4248,
      "theoretical_loss": 3.4681843232553327,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023626880641925777,
      "loss": 2.4568,
      "theoretical_loss": 3.4681732579733886,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023625877632898695,
      "loss": 2.8597,
      "theoretical_loss": 3.4681621932198237,
      "tokens_seen": 1756758016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023624874623871616,
      "loss": 2.5498,
      "theoretical_loss": 3.4681511289945943,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023623871614844534,
      "loss": 2.7137,
      "theoretical_loss": 3.4681400652976544,
      "tokens_seen": 1756889088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023622868605817452,
      "loss": 2.6745,
      "theoretical_loss": 3.46812900212896,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362186559679037,
      "loss": 2.6714,
      "theoretical_loss": 3.4681179394884656,
      "tokens_seen": 1757020160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023620862587763288,
      "loss": 2.5536,
      "theoretical_loss": 3.468106877376126,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361985957873621,
      "loss": 2.6876,
      "theoretical_loss": 3.4680958157918975,
      "tokens_seen": 1757151232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023618856569709127,
      "loss": 2.6691,
      "theoretical_loss": 3.468084754735734,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023617853560682045,
      "loss": 2.5732,
      "theoretical_loss": 3.4680736942075914,
      "tokens_seen": 1757282304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023616850551654963,
      "loss": 2.5585,
      "theoretical_loss": 3.468062634207424,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023615847542627884,
      "loss": 2.7523,
      "theoretical_loss": 3.4680515747351874,
      "tokens_seen": 1757413376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023614844533600802,
      "loss": 2.714,
      "theoretical_loss": 3.4680405157908365,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361384152457372,
      "loss": 2.7009,
      "theoretical_loss": 3.4680294573743264,
      "tokens_seen": 1757544448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023612838515546639,
      "loss": 2.7764,
      "theoretical_loss": 3.4680183994856133,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023611835506519557,
      "loss": 2.5591,
      "theoretical_loss": 3.4680073421246504,
      "tokens_seen": 1757675520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361083249749248,
      "loss": 2.479,
      "theoretical_loss": 3.4679962852913944,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023609829488465398,
      "loss": 2.7692,
      "theoretical_loss": 3.4679852289857998,
      "tokens_seen": 1757806592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023608826479438316,
      "loss": 2.5419,
      "theoretical_loss": 3.4679741732078213,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023607823470411234,
      "loss": 2.636,
      "theoretical_loss": 3.4679631179574155,
      "tokens_seen": 1757937664
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 969814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.568777322769165,
      "objective/train/theoretical_loss": 3.467952063234536,
      "objective/train/tokens_used": 1778463200,
      "theoretical_loss": 3.467952063234536,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023606820461384155,
      "loss": 2.6146,
      "theoretical_loss": 3.467952063234536,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023605817452357073,
      "loss": 2.7029,
      "theoretical_loss": 3.467941009039139,
      "tokens_seen": 1758068736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023604814443329992,
      "loss": 2.5418,
      "theoretical_loss": 3.4679299553711793,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002360381143430291,
      "loss": 2.5246,
      "theoretical_loss": 3.4679189022306116,
      "tokens_seen": 1758199808
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023602808425275828,
      "loss": 2.6141,
      "theoretical_loss": 3.467907849617392,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023601805416248749,
      "loss": 2.5974,
      "theoretical_loss": 3.467896797531475,
      "tokens_seen": 1758330880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023600802407221667,
      "loss": 2.4464,
      "theoretical_loss": 3.467885745972816,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023599799398194585,
      "loss": 2.4103,
      "theoretical_loss": 3.4678746949413704,
      "tokens_seen": 1758461952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023598796389167503,
      "loss": 2.3585,
      "theoretical_loss": 3.467863644437093,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023597793380140424,
      "loss": 2.6201,
      "theoretical_loss": 3.467852594459939,
      "tokens_seen": 1758593024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023596790371113342,
      "loss": 2.6758,
      "theoretical_loss": 3.4678415450098643,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002359578736208626,
      "loss": 2.7116,
      "theoretical_loss": 3.4678304960868234,
      "tokens_seen": 1758724096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023594784353059178,
      "loss": 2.55,
      "theoretical_loss": 3.467819447690772,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023593781344032096,
      "loss": 2.6693,
      "theoretical_loss": 3.4678083998216653,
      "tokens_seen": 1758855168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023592778335005017,
      "loss": 2.6492,
      "theoretical_loss": 3.467797352479458,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023591775325977935,
      "loss": 2.6536,
      "theoretical_loss": 3.4677863056641054,
      "tokens_seen": 1758986240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023590772316950853,
      "loss": 2.5829,
      "theoretical_loss": 3.4677752593755633,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358976930792377,
      "loss": 2.6594,
      "theoretical_loss": 3.467764213613787,
      "tokens_seen": 1759117312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023588766298896692,
      "loss": 2.5569,
      "theoretical_loss": 3.4677531683787306,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358776328986961,
      "loss": 2.6547,
      "theoretical_loss": 3.4677421236703507,
      "tokens_seen": 1759248384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023586760280842528,
      "loss": 2.4016,
      "theoretical_loss": 3.4677310794886025,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023585757271815447,
      "loss": 2.7557,
      "theoretical_loss": 3.46772003583344,
      "tokens_seen": 1759379456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023584754262788367,
      "loss": 2.5023,
      "theoretical_loss": 3.46770899270482,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023583751253761285,
      "loss": 2.5663,
      "theoretical_loss": 3.467697950102697,
      "tokens_seen": 1759510528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023582748244734204,
      "loss": 2.7471,
      "theoretical_loss": 3.467686908027026,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 971034,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7307450771331787,
      "objective/train/theoretical_loss": 3.467675866477763,
      "objective/train/tokens_used": 1780101600,
      "theoretical_loss": 3.467675866477763,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023581745235707122,
      "loss": 2.7636,
      "theoretical_loss": 3.467675866477763,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358074222668004,
      "loss": 2.5667,
      "theoretical_loss": 3.4676648254548628,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357973921765296,
      "loss": 2.593,
      "theoretical_loss": 3.467653784958281,
      "tokens_seen": 1759772672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357873620862588,
      "loss": 2.6615,
      "theoretical_loss": 3.467642744987973,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023577733199598797,
      "loss": 2.5576,
      "theoretical_loss": 3.467631705543894,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023576730190571715,
      "loss": 2.6404,
      "theoretical_loss": 3.467620666625999,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023575727181544636,
      "loss": 2.8915,
      "theoretical_loss": 3.4676096282342432,
      "tokens_seen": 1760034816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023574724172517554,
      "loss": 2.7583,
      "theoretical_loss": 3.4675985903685826,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023573721163490472,
      "loss": 2.5498,
      "theoretical_loss": 3.4675875530289724,
      "tokens_seen": 1760165888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357271815446339,
      "loss": 2.6749,
      "theoretical_loss": 3.4675765162153676,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023571715145436308,
      "loss": 2.4578,
      "theoretical_loss": 3.467565479927724,
      "tokens_seen": 1760296960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357071213640923,
      "loss": 2.9108,
      "theoretical_loss": 3.4675544441659962,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023569709127382147,
      "loss": 2.8044,
      "theoretical_loss": 3.4675434089301405,
      "tokens_seen": 1760428032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023568706118355065,
      "loss": 2.558,
      "theoretical_loss": 3.467532374220112,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023567703109327983,
      "loss": 2.5134,
      "theoretical_loss": 3.4675213400358653,
      "tokens_seen": 1760559104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023566700100300904,
      "loss": 2.4334,
      "theoretical_loss": 3.4675103063773567,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023565697091273822,
      "loss": 2.4155,
      "theoretical_loss": 3.4674992732445413,
      "tokens_seen": 1760690176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002356469408224674,
      "loss": 2.4986,
      "theoretical_loss": 3.4674882406373744,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023563691073219659,
      "loss": 2.645,
      "theoretical_loss": 3.4674772085558114,
      "tokens_seen": 1760821248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023562688064192577,
      "loss": 2.3818,
      "theoretical_loss": 3.4674661769998076,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023561685055165497,
      "loss": 2.6899,
      "theoretical_loss": 3.4674551459693186,
      "tokens_seen": 1760952320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023560682046138416,
      "loss": 2.532,
      "theoretical_loss": 3.4674441154643,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023559679037111334,
      "loss": 2.6592,
      "theoretical_loss": 3.467433085484706,
      "tokens_seen": 1761083392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023558676028084252,
      "loss": 2.7259,
      "theoretical_loss": 3.467422056030494,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023557673019057173,
      "loss": 2.4571,
      "theoretical_loss": 3.467411027101618,
      "tokens_seen": 1761214464
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 971632,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7771389484405518,
      "objective/train/theoretical_loss": 3.467399998698034,
      "objective/train/tokens_used": 1781740000,
      "theoretical_loss": 3.467399998698034,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002355667001003009,
      "loss": 2.6972,
      "theoretical_loss": 3.467399998698034,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002355566700100301,
      "loss": 2.6388,
      "theoretical_loss": 3.467388970819697,
      "tokens_seen": 1761345536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023554663991975927,
      "loss": 2.6961,
      "theoretical_loss": 3.4673779434665635,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023553660982948845,
      "loss": 2.3988,
      "theoretical_loss": 3.4673669166385874,
      "tokens_seen": 1761476608
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023552657973921766,
      "loss": 2.3356,
      "theoretical_loss": 3.4673558903357247,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023551654964894684,
      "loss": 2.5909,
      "theoretical_loss": 3.4673448645579317,
      "tokens_seen": 1761607680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023550651955867602,
      "loss": 2.444,
      "theoretical_loss": 3.467333839305163,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354964894684052,
      "loss": 2.5427,
      "theoretical_loss": 3.4673228145773747,
      "tokens_seen": 1761738752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354864593781344,
      "loss": 2.4788,
      "theoretical_loss": 3.467311790374521,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354764292878636,
      "loss": 2.7955,
      "theoretical_loss": 3.467300766696559,
      "tokens_seen": 1761869824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023546639919759277,
      "loss": 2.5186,
      "theoretical_loss": 3.467289743543443,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023545636910732195,
      "loss": 2.5692,
      "theoretical_loss": 3.4672787209151292,
      "tokens_seen": 1762000896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023544633901705113,
      "loss": 2.7837,
      "theoretical_loss": 3.4672676988115727,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023543630892678034,
      "loss": 2.3005,
      "theoretical_loss": 3.4672566772327293,
      "tokens_seen": 1762131968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023542627883650952,
      "loss": 2.4069,
      "theoretical_loss": 3.4672456561785543,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354162487462387,
      "loss": 2.722,
      "theoretical_loss": 3.467234635649003,
      "tokens_seen": 1762263040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354062186559679,
      "loss": 2.5592,
      "theoretical_loss": 3.4672236156440315,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353961885656971,
      "loss": 2.8511,
      "theoretical_loss": 3.467212596163595,
      "tokens_seen": 1762394112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023538615847542628,
      "loss": 2.7208,
      "theoretical_loss": 3.467201577207649,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023537612838515546,
      "loss": 2.477,
      "theoretical_loss": 3.467190558776149,
      "tokens_seen": 1762525184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023536609829488464,
      "loss": 2.7212,
      "theoretical_loss": 3.4671795408690507,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023535606820461387,
      "loss": 2.6394,
      "theoretical_loss": 3.4671685234863094,
      "tokens_seen": 1762656256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023534603811434305,
      "loss": 2.6173,
      "theoretical_loss": 3.467157506627881,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023533600802407224,
      "loss": 2.2606,
      "theoretical_loss": 3.4671464902937204,
      "tokens_seen": 1762787328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023532597793380142,
      "loss": 2.6508,
      "theoretical_loss": 3.4671354744837837,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 972718,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7899014949798584,
      "objective/train/theoretical_loss": 3.4671244591980264,
      "objective/train/tokens_used": 1783378400,
      "theoretical_loss": 3.4671244591980264,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353159478435306,
      "loss": 2.4742,
      "theoretical_loss": 3.4671244591980264,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353059177532598,
      "loss": 2.6062,
      "theoretical_loss": 3.4671134444364045,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000235295887662989,
      "loss": 2.7169,
      "theoretical_loss": 3.467102430198873,
      "tokens_seen": 1763049472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023528585757271817,
      "loss": 2.5204,
      "theoretical_loss": 3.4670914164853874,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023527582748244735,
      "loss": 2.5964,
      "theoretical_loss": 3.4670804032959035,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023526579739217656,
      "loss": 2.5988,
      "theoretical_loss": 3.4670693906303773,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023525576730190574,
      "loss": 2.5434,
      "theoretical_loss": 3.4670583784887636,
      "tokens_seen": 1763311616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023524573721163492,
      "loss": 2.6381,
      "theoretical_loss": 3.4670473668710184,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002352357071213641,
      "loss": 2.668,
      "theoretical_loss": 3.467036355777098,
      "tokens_seen": 1763442688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023522567703109328,
      "loss": 2.4475,
      "theoretical_loss": 3.467025345206956,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002352156469408225,
      "loss": 2.5262,
      "theoretical_loss": 3.4670143351605502,
      "tokens_seen": 1763573760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023520561685055167,
      "loss": 2.3041,
      "theoretical_loss": 3.4670033256378354,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023519558676028085,
      "loss": 2.5624,
      "theoretical_loss": 3.466992316638767,
      "tokens_seen": 1763704832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023518555667001003,
      "loss": 2.4205,
      "theoretical_loss": 3.4669813081633007,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023517552657973924,
      "loss": 2.6358,
      "theoretical_loss": 3.466970300211393,
      "tokens_seen": 1763835904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023516549648946842,
      "loss": 2.7635,
      "theoretical_loss": 3.466959292782998,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002351554663991976,
      "loss": 2.7832,
      "theoretical_loss": 3.4669482858780727,
      "tokens_seen": 1763966976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023514543630892679,
      "loss": 2.6588,
      "theoretical_loss": 3.4669372794965723,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023513540621865597,
      "loss": 2.6708,
      "theoretical_loss": 3.4669262736384523,
      "tokens_seen": 1764098048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023512537612838517,
      "loss": 2.4572,
      "theoretical_loss": 3.4669152683036684,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023511534603811436,
      "loss": 2.5199,
      "theoretical_loss": 3.466904263492176,
      "tokens_seen": 1764229120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023510531594784354,
      "loss": 2.4272,
      "theoretical_loss": 3.466893259203932,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023509528585757272,
      "loss": 2.5436,
      "theoretical_loss": 3.4668822554388905,
      "tokens_seen": 1764360192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023508525576730193,
      "loss": 2.5252,
      "theoretical_loss": 3.4668712521970084,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350752256770311,
      "loss": 2.65,
      "theoretical_loss": 3.4668602494782403,
      "tokens_seen": 1764491264
    },
    {
      "epoch": 0.53,
      "objective/train/docs_used": 973845,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7647204399108887,
      "objective/train/theoretical_loss": 3.466849247282543,
      "objective/train/tokens_used": 1785016800,
      "theoretical_loss": 3.466849247282543,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350651955867603,
      "loss": 2.5482,
      "theoretical_loss": 3.466849247282543,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023505516549648947,
      "loss": 2.466,
      "theoretical_loss": 3.466838245609872,
      "tokens_seen": 1764622336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023504513540621865,
      "loss": 2.2684,
      "theoretical_loss": 3.466827244460182,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023503510531594786,
      "loss": 2.482,
      "theoretical_loss": 3.46681624383343,
      "tokens_seen": 1764753408
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023502507522567704,
      "loss": 2.5588,
      "theoretical_loss": 3.466805243729571,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023501504513540622,
      "loss": 2.471,
      "theoretical_loss": 3.466794244148561,
      "tokens_seen": 1764884480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350050150451354,
      "loss": 2.4617,
      "theoretical_loss": 3.4667832450903555,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349949849548646,
      "loss": 2.6585,
      "theoretical_loss": 3.4667722465549105,
      "tokens_seen": 1765015552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349849548645938,
      "loss": 2.6352,
      "theoretical_loss": 3.4667612485421815,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023497492477432297,
      "loss": 2.4554,
      "theoretical_loss": 3.4667502510521246,
      "tokens_seen": 1765146624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023496489468405215,
      "loss": 2.4898,
      "theoretical_loss": 3.466739254084695,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023495486459378134,
      "loss": 2.6801,
      "theoretical_loss": 3.466728257639849,
      "tokens_seen": 1765277696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023494483450351054,
      "loss": 2.6776,
      "theoretical_loss": 3.4667172617175424,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023493480441323972,
      "loss": 2.4368,
      "theoretical_loss": 3.4667062663177304,
      "tokens_seen": 1765408768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002349247743229689,
      "loss": 2.4614,
      "theoretical_loss": 3.4666952714403694,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002349147442326981,
      "loss": 2.5509,
      "theoretical_loss": 3.4666842770854145,
      "tokens_seen": 1765539840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002349047141424273,
      "loss": 2.6464,
      "theoretical_loss": 3.466673283252822,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023489468405215648,
      "loss": 2.4267,
      "theoretical_loss": 3.4666622899425477,
      "tokens_seen": 1765670912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023488465396188566,
      "loss": 2.7684,
      "theoretical_loss": 3.4666512971545473,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023487462387161484,
      "loss": 2.5994,
      "theoretical_loss": 3.4666403048887764,
      "tokens_seen": 1765801984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023486459378134402,
      "loss": 2.5682,
      "theoretical_loss": 3.4666293131451917,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023485456369107323,
      "loss": 2.4093,
      "theoretical_loss": 3.4666183219237476,
      "tokens_seen": 1765933056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348445336008024,
      "loss": 2.3661,
      "theoretical_loss": 3.4666073312244006,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348345035105316,
      "loss": 2.5007,
      "theoretical_loss": 3.466596341047107,
      "tokens_seen": 1766064128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023482447342026077,
      "loss": 2.6515,
      "theoretical_loss": 3.4665853513918217,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 974265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4899380207061768,
      "objective/train/theoretical_loss": 3.466574362258501,
      "objective/train/tokens_used": 1786655200,
      "theoretical_loss": 3.466574362258501,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023481444332998998,
      "loss": 2.4445,
      "theoretical_loss": 3.466574362258501,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023480441323971916,
      "loss": 2.4758,
      "theoretical_loss": 3.4665633736471015,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023479438314944834,
      "loss": 2.6716,
      "theoretical_loss": 3.466552385557578,
      "tokens_seen": 1766326272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023478435305917752,
      "loss": 2.5998,
      "theoretical_loss": 3.4665413979898863,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023477432296890673,
      "loss": 2.4649,
      "theoretical_loss": 3.466530410943983,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347642928786359,
      "loss": 2.571,
      "theoretical_loss": 3.466519424419823,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347542627883651,
      "loss": 2.3291,
      "theoretical_loss": 3.4665084384173634,
      "tokens_seen": 1766588416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023474423269809427,
      "loss": 2.5003,
      "theoretical_loss": 3.4664974529365593,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023473420260782346,
      "loss": 2.7273,
      "theoretical_loss": 3.466486467977367,
      "tokens_seen": 1766719488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023472417251755266,
      "loss": 2.4653,
      "theoretical_loss": 3.4664754835397416,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023471414242728184,
      "loss": 2.5274,
      "theoretical_loss": 3.4664644996236396,
      "tokens_seen": 1766850560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023470411233701103,
      "loss": 2.5581,
      "theoretical_loss": 3.4664535162290164,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346940822467402,
      "loss": 2.1795,
      "theoretical_loss": 3.466442533355829,
      "tokens_seen": 1766981632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023468405215646942,
      "loss": 2.595,
      "theoretical_loss": 3.466431551004032,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346740220661986,
      "loss": 2.5448,
      "theoretical_loss": 3.4664205691735823,
      "tokens_seen": 1767112704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023466399197592778,
      "loss": 2.5241,
      "theoretical_loss": 3.466409587864436,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023465396188565696,
      "loss": 2.4458,
      "theoretical_loss": 3.466398607076547,
      "tokens_seen": 1767243776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023464393179538614,
      "loss": 2.5523,
      "theoretical_loss": 3.4663876268098734,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023463390170511535,
      "loss": 2.385,
      "theoretical_loss": 3.466376647064371,
      "tokens_seen": 1767374848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023462387161484453,
      "loss": 2.616,
      "theoretical_loss": 3.4663656678399946,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346138415245737,
      "loss": 2.5692,
      "theoretical_loss": 3.4663546891367005,
      "tokens_seen": 1767505920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023460381143430292,
      "loss": 2.582,
      "theoretical_loss": 3.4663437109544453,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023459378134403213,
      "loss": 2.5744,
      "theoretical_loss": 3.466332733293184,
      "tokens_seen": 1767636992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345837512537613,
      "loss": 2.5657,
      "theoretical_loss": 3.4663217561528734,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345737211634905,
      "loss": 2.5531,
      "theoretical_loss": 3.466310779533469,
      "tokens_seen": 1767768064
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 975251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6392266750335693,
      "objective/train/theoretical_loss": 3.466299803434927,
      "objective/train/tokens_used": 1788293600,
      "theoretical_loss": 3.466299803434927,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023456369107321967,
      "loss": 2.5895,
      "theoretical_loss": 3.466299803434927,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023455366098294885,
      "loss": 2.5655,
      "theoretical_loss": 3.4662888278572033,
      "tokens_seen": 1767899136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023454363089267806,
      "loss": 2.6469,
      "theoretical_loss": 3.4662778528002534,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023453360080240724,
      "loss": 2.711,
      "theoretical_loss": 3.466266878264034,
      "tokens_seen": 1768030208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023452357071213642,
      "loss": 2.4469,
      "theoretical_loss": 3.4662559042485013,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345135406218656,
      "loss": 2.609,
      "theoretical_loss": 3.4662449307536103,
      "tokens_seen": 1768161280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345035105315948,
      "loss": 2.7191,
      "theoretical_loss": 3.466233957779318,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000234493480441324,
      "loss": 2.6533,
      "theoretical_loss": 3.4662229853255795,
      "tokens_seen": 1768292352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023448345035105317,
      "loss": 2.5558,
      "theoretical_loss": 3.466212013392351,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023447342026078235,
      "loss": 2.4606,
      "theoretical_loss": 3.4662010419795894,
      "tokens_seen": 1768423424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023446339017051154,
      "loss": 2.6137,
      "theoretical_loss": 3.46619007108725,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023445336008024074,
      "loss": 2.3096,
      "theoretical_loss": 3.4661791007152885,
      "tokens_seen": 1768554496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023444332998996992,
      "loss": 2.6895,
      "theoretical_loss": 3.466168130863662,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344332998996991,
      "loss": 2.3721,
      "theoretical_loss": 3.4661571615323252,
      "tokens_seen": 1768685568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344232698094283,
      "loss": 2.4794,
      "theoretical_loss": 3.466146192721235,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344132397191575,
      "loss": 2.4569,
      "theoretical_loss": 3.4661352244303476,
      "tokens_seen": 1768816640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023440320962888668,
      "loss": 2.7832,
      "theoretical_loss": 3.4661242566596187,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023439317953861586,
      "loss": 2.4007,
      "theoretical_loss": 3.4661132894090043,
      "tokens_seen": 1768947712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023438314944834504,
      "loss": 2.6413,
      "theoretical_loss": 3.46610232267846,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023437311935807422,
      "loss": 2.4071,
      "theoretical_loss": 3.4660913564679436,
      "tokens_seen": 1769078784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023436308926780343,
      "loss": 2.7769,
      "theoretical_loss": 3.4660803907774094,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002343530591775326,
      "loss": 2.4714,
      "theoretical_loss": 3.4660694256068143,
      "tokens_seen": 1769209856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002343430290872618,
      "loss": 2.513,
      "theoretical_loss": 3.4660584609561145,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023433299899699097,
      "loss": 2.4569,
      "theoretical_loss": 3.466047496825265,
      "tokens_seen": 1769340928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023432296890672018,
      "loss": 2.6069,
      "theoretical_loss": 3.4660365332142233,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 975838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.774744749069214,
      "objective/train/theoretical_loss": 3.4660255701229445,
      "objective/train/tokens_used": 1789932000,
      "theoretical_loss": 3.4660255701229445,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023431293881644936,
      "loss": 2.6818,
      "theoretical_loss": 3.4660255701229445,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023430290872617854,
      "loss": 2.6546,
      "theoretical_loss": 3.4660146075513856,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023429287863590772,
      "loss": 2.6377,
      "theoretical_loss": 3.466003645499502,
      "tokens_seen": 1769603072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342828485456369,
      "loss": 2.707,
      "theoretical_loss": 3.46599268396725,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342728184553661,
      "loss": 2.602,
      "theoretical_loss": 3.465981722954586,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342627883650953,
      "loss": 2.4092,
      "theoretical_loss": 3.465970762461466,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023425275827482447,
      "loss": 2.4361,
      "theoretical_loss": 3.4659598024878457,
      "tokens_seen": 1769865216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023424272818455366,
      "loss": 2.5396,
      "theoretical_loss": 3.4659488430336816,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023423269809428286,
      "loss": 2.3076,
      "theoretical_loss": 3.46593788409893,
      "tokens_seen": 1769996288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023422266800401204,
      "loss": 2.5533,
      "theoretical_loss": 3.4659269256835468,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023421263791374123,
      "loss": 2.8028,
      "theoretical_loss": 3.4659159677874882,
      "tokens_seen": 1770127360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342026078234704,
      "loss": 2.5231,
      "theoretical_loss": 3.4659050104107103,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023419257773319962,
      "loss": 2.5688,
      "theoretical_loss": 3.4658940535531695,
      "tokens_seen": 1770258432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341825476429288,
      "loss": 2.6782,
      "theoretical_loss": 3.465883097214822,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023417251755265798,
      "loss": 2.5465,
      "theoretical_loss": 3.4658721413956237,
      "tokens_seen": 1770389504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023416248746238716,
      "loss": 2.4251,
      "theoretical_loss": 3.465861186095531,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023415245737211634,
      "loss": 2.4895,
      "theoretical_loss": 3.4658502313144997,
      "tokens_seen": 1770520576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023414242728184555,
      "loss": 2.376,
      "theoretical_loss": 3.4658392770524866,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023413239719157473,
      "loss": 2.5522,
      "theoretical_loss": 3.4658283233094473,
      "tokens_seen": 1770651648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341223671013039,
      "loss": 2.6664,
      "theoretical_loss": 3.4658173700853387,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341123370110331,
      "loss": 2.4874,
      "theoretical_loss": 3.4658064173801164,
      "tokens_seen": 1770782720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341023069207623,
      "loss": 2.4227,
      "theoretical_loss": 3.465795465193737,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023409227683049148,
      "loss": 2.6215,
      "theoretical_loss": 3.465784513526156,
      "tokens_seen": 1770913792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023408224674022066,
      "loss": 2.4806,
      "theoretical_loss": 3.465773562377331,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023407221664994984,
      "loss": 2.5349,
      "theoretical_loss": 3.465762611747217,
      "tokens_seen": 1771044864
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 976909,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4867255687713623,
      "objective/train/theoretical_loss": 3.4657516616357706,
      "objective/train/tokens_used": 1791570400,
      "theoretical_loss": 3.4657516616357706,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023406218655967902,
      "loss": 2.434,
      "theoretical_loss": 3.4657516616357706,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023405215646940823,
      "loss": 2.7789,
      "theoretical_loss": 3.4657407120429484,
      "tokens_seen": 1771175936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002340421263791374,
      "loss": 2.5983,
      "theoretical_loss": 3.4657297629687056,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002340320962888666,
      "loss": 2.3371,
      "theoretical_loss": 3.465718814413,
      "tokens_seen": 1771307008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023402206619859578,
      "loss": 2.5422,
      "theoretical_loss": 3.4657078663757868,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023401203610832498,
      "loss": 2.8877,
      "theoretical_loss": 3.4656969188570224,
      "tokens_seen": 1771438080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023400200601805416,
      "loss": 2.5032,
      "theoretical_loss": 3.4656859718566633,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023399197592778335,
      "loss": 2.6246,
      "theoretical_loss": 3.465675025374666,
      "tokens_seen": 1771569152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023398194583751253,
      "loss": 2.6825,
      "theoretical_loss": 3.4656640794109856,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002339719157472417,
      "loss": 2.565,
      "theoretical_loss": 3.4656531339655796,
      "tokens_seen": 1771700224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023396188565697092,
      "loss": 2.5561,
      "theoretical_loss": 3.4656421890384044,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002339518555667001,
      "loss": 2.4401,
      "theoretical_loss": 3.4656312446294155,
      "tokens_seen": 1771831296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023394182547642928,
      "loss": 2.6505,
      "theoretical_loss": 3.465620300738569,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023393179538615846,
      "loss": 2.5684,
      "theoretical_loss": 3.4656093573658224,
      "tokens_seen": 1771962368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023392176529588767,
      "loss": 2.375,
      "theoretical_loss": 3.465598414511131,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023391173520561685,
      "loss": 2.4267,
      "theoretical_loss": 3.4655874721744517,
      "tokens_seen": 1772093440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023391173520561685,
      "loss": 2.5981,
      "theoretical_loss": 3.46557653035574,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023390170511534603,
      "loss": 2.5032,
      "theoretical_loss": 3.465565589054954,
      "tokens_seen": 1772224512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002338916750250752,
      "loss": 2.5082,
      "theoretical_loss": 3.4655546482720476,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002338816449348044,
      "loss": 2.4401,
      "theoretical_loss": 3.4655437080069786,
      "tokens_seen": 1772355584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002338716148445336,
      "loss": 2.5238,
      "theoretical_loss": 3.4655327682597035,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023386158475426278,
      "loss": 2.7075,
      "theoretical_loss": 3.4655218290301786,
      "tokens_seen": 1772486656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000233851554663992,
      "loss": 2.5778,
      "theoretical_loss": 3.465510890318359,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023384152457372117,
      "loss": 2.6095,
      "theoretical_loss": 3.465499952124203,
      "tokens_seen": 1772617728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023383149448345038,
      "loss": 2.3762,
      "theoretical_loss": 3.465489014447665,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 977811,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.209491491317749,
      "objective/train/theoretical_loss": 3.4654780772887026,
      "objective/train/tokens_used": 1793208800,
      "theoretical_loss": 3.4654780772887026,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023382146439317956,
      "loss": 2.6243,
      "theoretical_loss": 3.4654780772887026,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023381143430290874,
      "loss": 2.2984,
      "theoretical_loss": 3.465467140647272,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023380140421263792,
      "loss": 2.6788,
      "theoretical_loss": 3.4654562045233295,
      "tokens_seen": 1772879872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337913741223671,
      "loss": 2.5767,
      "theoretical_loss": 3.465445268916832,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337813440320963,
      "loss": 2.7247,
      "theoretical_loss": 3.4654343338277345,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337713139418255,
      "loss": 2.6301,
      "theoretical_loss": 3.4654233992559944,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023376128385155467,
      "loss": 2.5464,
      "theoretical_loss": 3.465412465201568,
      "tokens_seen": 1773142016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023375125376128386,
      "loss": 2.6401,
      "theoretical_loss": 3.465401531664412,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023374122367101306,
      "loss": 2.6863,
      "theoretical_loss": 3.4653905986444817,
      "tokens_seen": 1773273088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023373119358074224,
      "loss": 2.6016,
      "theoretical_loss": 3.465379666141735,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023372116349047143,
      "loss": 2.4497,
      "theoretical_loss": 3.4653687341561277,
      "tokens_seen": 1773404160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337111334002006,
      "loss": 2.5513,
      "theoretical_loss": 3.4653578026876155,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023370110330992982,
      "loss": 2.6074,
      "theoretical_loss": 3.4653468717361555,
      "tokens_seen": 1773535232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000233691073219659,
      "loss": 2.6423,
      "theoretical_loss": 3.4653359413017046,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023368104312938818,
      "loss": 2.6018,
      "theoretical_loss": 3.465325011384219,
      "tokens_seen": 1773666304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023367101303911736,
      "loss": 2.4703,
      "theoretical_loss": 3.465314081983654,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023366098294884654,
      "loss": 2.5259,
      "theoretical_loss": 3.4653031530999674,
      "tokens_seen": 1773797376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023365095285857575,
      "loss": 2.6595,
      "theoretical_loss": 3.4652922247331155,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023364092276830493,
      "loss": 2.4747,
      "theoretical_loss": 3.465281296883054,
      "tokens_seen": 1773928448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336308926780341,
      "loss": 2.6439,
      "theoretical_loss": 3.4652703695497404,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336208625877633,
      "loss": 2.5947,
      "theoretical_loss": 3.46525944273313,
      "tokens_seen": 1774059520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336108324974925,
      "loss": 2.3789,
      "theoretical_loss": 3.4652485164331805,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023360080240722168,
      "loss": 2.6416,
      "theoretical_loss": 3.4652375906498474,
      "tokens_seen": 1774190592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023359077231695086,
      "loss": 2.6127,
      "theoretical_loss": 3.4652266653830877,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023358074222668004,
      "loss": 2.7866,
      "theoretical_loss": 3.465215740632858,
      "tokens_seen": 1774321664
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 978517,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6737301349639893,
      "objective/train/theoretical_loss": 3.465204816399114,
      "objective/train/tokens_used": 1794847200,
      "theoretical_loss": 3.465204816399114,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023357071213640922,
      "loss": 2.5206,
      "theoretical_loss": 3.465204816399114,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023356068204613843,
      "loss": 2.5331,
      "theoretical_loss": 3.465193892681813,
      "tokens_seen": 1774452736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002335506519558676,
      "loss": 2.337,
      "theoretical_loss": 3.4651829694809115,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002335406218655968,
      "loss": 2.412,
      "theoretical_loss": 3.465172046796366,
      "tokens_seen": 1774583808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023353059177532598,
      "loss": 2.4402,
      "theoretical_loss": 3.4651611246281324,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023352056168505518,
      "loss": 2.3491,
      "theoretical_loss": 3.465150202976168,
      "tokens_seen": 1774714880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023351053159478437,
      "loss": 2.7487,
      "theoretical_loss": 3.465139281840429,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023350050150451355,
      "loss": 2.8213,
      "theoretical_loss": 3.465128361220872,
      "tokens_seen": 1774845952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023349047141424273,
      "loss": 2.6398,
      "theoretical_loss": 3.465117441117453,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334804413239719,
      "loss": 2.466,
      "theoretical_loss": 3.4651065215301298,
      "tokens_seen": 1774977024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023347041123370112,
      "loss": 2.6394,
      "theoretical_loss": 3.4650956024588577,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334603811434303,
      "loss": 2.6232,
      "theoretical_loss": 3.4650846839035943,
      "tokens_seen": 1775108096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023345035105315948,
      "loss": 2.1572,
      "theoretical_loss": 3.465073765864295,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023344032096288866,
      "loss": 2.395,
      "theoretical_loss": 3.465062848340917,
      "tokens_seen": 1775239168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023343029087261787,
      "loss": 2.5863,
      "theoretical_loss": 3.4650519313334174,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023342026078234705,
      "loss": 2.5454,
      "theoretical_loss": 3.465041014841751,
      "tokens_seen": 1775370240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023341023069207623,
      "loss": 2.7774,
      "theoretical_loss": 3.465030098865877,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334002006018054,
      "loss": 2.4523,
      "theoretical_loss": 3.4650191834057504,
      "tokens_seen": 1775501312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333901705115346,
      "loss": 2.4759,
      "theoretical_loss": 3.465008268461328,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333801404212638,
      "loss": 2.4438,
      "theoretical_loss": 3.4649973540325663,
      "tokens_seen": 1775632384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023337011033099298,
      "loss": 2.6314,
      "theoretical_loss": 3.4649864401194215,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023336008024072216,
      "loss": 2.6174,
      "theoretical_loss": 3.4649755267218514,
      "tokens_seen": 1775763456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023335005015045134,
      "loss": 2.6016,
      "theoretical_loss": 3.464964613839812,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023334002006018055,
      "loss": 2.4666,
      "theoretical_loss": 3.4649537014732594,
      "tokens_seen": 1775894528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023332998996990973,
      "loss": 2.6571,
      "theoretical_loss": 3.464942789622151,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 979554,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.405682325363159,
      "objective/train/theoretical_loss": 3.464931878286443,
      "objective/train/tokens_used": 1796485600,
      "theoretical_loss": 3.464931878286443,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023331995987963891,
      "loss": 2.5476,
      "theoretical_loss": 3.464931878286443,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333099297893681,
      "loss": 2.612,
      "theoretical_loss": 3.4649209674660923,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023329989969909728,
      "loss": 2.6099,
      "theoretical_loss": 3.464910057161056,
      "tokens_seen": 1776156672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023328986960882649,
      "loss": 2.4333,
      "theoretical_loss": 3.4648991473712893,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023327983951855567,
      "loss": 2.4825,
      "theoretical_loss": 3.4648882380967505,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023326980942828485,
      "loss": 2.4924,
      "theoretical_loss": 3.464877329337395,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023325977933801403,
      "loss": 2.2428,
      "theoretical_loss": 3.4648664210931797,
      "tokens_seen": 1776418816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023324974924774324,
      "loss": 2.5475,
      "theoretical_loss": 3.4648555133640624,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023323971915747242,
      "loss": 2.7417,
      "theoretical_loss": 3.464844606149998,
      "tokens_seen": 1776549888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002332296890672016,
      "loss": 2.5191,
      "theoretical_loss": 3.4648336994509448,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023321965897693078,
      "loss": 2.6846,
      "theoretical_loss": 3.4648227932668583,
      "tokens_seen": 1776680960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023320962888665996,
      "loss": 2.4009,
      "theoretical_loss": 3.464811887597696,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023319959879638917,
      "loss": 2.8033,
      "theoretical_loss": 3.464800982443414,
      "tokens_seen": 1776812032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023318956870611835,
      "loss": 2.4257,
      "theoretical_loss": 3.4647900778039693,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023317953861584753,
      "loss": 2.4967,
      "theoretical_loss": 3.464779173679319,
      "tokens_seen": 1776943104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331695085255767,
      "loss": 2.6116,
      "theoretical_loss": 3.4647682700694187,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023315947843530592,
      "loss": 2.5849,
      "theoretical_loss": 3.464757366974226,
      "tokens_seen": 1777074176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331494483450351,
      "loss": 2.6434,
      "theoretical_loss": 3.4647464643936976,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023313941825476428,
      "loss": 2.8021,
      "theoretical_loss": 3.4647355623277902,
      "tokens_seen": 1777205248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023312938816449346,
      "loss": 2.6791,
      "theoretical_loss": 3.4647246607764606,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023311935807422267,
      "loss": 2.4866,
      "theoretical_loss": 3.4647137597396647,
      "tokens_seen": 1777336320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023310932798395185,
      "loss": 2.6777,
      "theoretical_loss": 3.4647028592173603,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023309929789368106,
      "loss": 2.5901,
      "theoretical_loss": 3.464691959209503,
      "tokens_seen": 1777467392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023308926780341024,
      "loss": 2.5465,
      "theoretical_loss": 3.4646810597160513,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023307923771313942,
      "loss": 2.4744,
      "theoretical_loss": 3.46467016073696,
      "tokens_seen": 1777598464
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 979848,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8063058853149414,
      "objective/train/theoretical_loss": 3.4646592622721872,
      "objective/train/tokens_used": 1798124000,
      "theoretical_loss": 3.4646592622721872,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023306920762286863,
      "loss": 2.6089,
      "theoretical_loss": 3.4646592622721872,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023305917753259781,
      "loss": 2.7058,
      "theoretical_loss": 3.4646483643216897,
      "tokens_seen": 1777729536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000233049147442327,
      "loss": 2.5138,
      "theoretical_loss": 3.4646374668854234,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023303911735205618,
      "loss": 2.6242,
      "theoretical_loss": 3.4646265699633454,
      "tokens_seen": 1777860608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023302908726178538,
      "loss": 2.5853,
      "theoretical_loss": 3.464615673555413,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023301905717151457,
      "loss": 2.6789,
      "theoretical_loss": 3.4646047776615823,
      "tokens_seen": 1777991680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023300902708124375,
      "loss": 2.5806,
      "theoretical_loss": 3.4645938822818105,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023299899699097293,
      "loss": 2.5465,
      "theoretical_loss": 3.4645829874160543,
      "tokens_seen": 1778122752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002329889669007021,
      "loss": 2.6129,
      "theoretical_loss": 3.4645720930642705,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023297893681043132,
      "loss": 2.72,
      "theoretical_loss": 3.4645611992264156,
      "tokens_seen": 1778253824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002329689067201605,
      "loss": 2.5397,
      "theoretical_loss": 3.4645503059024474,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023295887662988968,
      "loss": 2.6537,
      "theoretical_loss": 3.4645394130923215,
      "tokens_seen": 1778384896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023294884653961886,
      "loss": 2.5832,
      "theoretical_loss": 3.4645285207959953,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023293881644934807,
      "loss": 2.5742,
      "theoretical_loss": 3.464517629013426,
      "tokens_seen": 1778515968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023292878635907725,
      "loss": 2.5002,
      "theoretical_loss": 3.4645067377445695,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023291875626880643,
      "loss": 2.7155,
      "theoretical_loss": 3.464495846989384,
      "tokens_seen": 1778647040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002329087261785356,
      "loss": 2.764,
      "theoretical_loss": 3.4644849567478246,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328986960882648,
      "loss": 2.6353,
      "theoretical_loss": 3.4644740670198497,
      "tokens_seen": 1778778112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232888665997994,
      "loss": 2.5242,
      "theoretical_loss": 3.464463177805415,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023287863590772318,
      "loss": 2.5894,
      "theoretical_loss": 3.4644522891044787,
      "tokens_seen": 1778909184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023286860581745236,
      "loss": 2.633,
      "theoretical_loss": 3.4644414009169964,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023285857572718154,
      "loss": 2.6758,
      "theoretical_loss": 3.4644305132429256,
      "tokens_seen": 1779040256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023284854563691075,
      "loss": 2.7131,
      "theoretical_loss": 3.464419626082223,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023283851554663993,
      "loss": 2.7512,
      "theoretical_loss": 3.4644087394348455,
      "tokens_seen": 1779171328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023282848545636911,
      "loss": 2.5803,
      "theoretical_loss": 3.46439785330075,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 981168,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9261651039123535,
      "objective/train/theoretical_loss": 3.4643869676798937,
      "objective/train/tokens_used": 1799762400,
      "theoretical_loss": 3.4643869676798937,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328184553660983,
      "loss": 2.547,
      "theoretical_loss": 3.4643869676798937,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023280842527582748,
      "loss": 2.5714,
      "theoretical_loss": 3.464376082572233,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023279839518555669,
      "loss": 2.5699,
      "theoretical_loss": 3.464365197977725,
      "tokens_seen": 1779433472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023278836509528587,
      "loss": 2.4846,
      "theoretical_loss": 3.464354313896327,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023277833500501505,
      "loss": 2.4584,
      "theoretical_loss": 3.464343430327995,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023276830491474423,
      "loss": 2.5428,
      "theoretical_loss": 3.4643325472726865,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023275827482447344,
      "loss": 2.6426,
      "theoretical_loss": 3.4643216647303583,
      "tokens_seen": 1779695616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023274824473420262,
      "loss": 2.6842,
      "theoretical_loss": 3.4643107827009674,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327382146439318,
      "loss": 2.4945,
      "theoretical_loss": 3.4642999011844715,
      "tokens_seen": 1779826688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023272818455366098,
      "loss": 2.4912,
      "theoretical_loss": 3.464289020180826,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023271815446339016,
      "loss": 2.4434,
      "theoretical_loss": 3.464278139689989,
      "tokens_seen": 1779957760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023270812437311937,
      "loss": 2.495,
      "theoretical_loss": 3.4642672597119173,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023269809428284855,
      "loss": 2.4342,
      "theoretical_loss": 3.4642563802465673,
      "tokens_seen": 1780088832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023268806419257773,
      "loss": 2.6704,
      "theoretical_loss": 3.464245501293896,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326780341023069,
      "loss": 2.5062,
      "theoretical_loss": 3.464234622853861,
      "tokens_seen": 1780219904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023266800401203612,
      "loss": 2.6159,
      "theoretical_loss": 3.4642237449264193,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326579739217653,
      "loss": 2.7678,
      "theoretical_loss": 3.4642128675115273,
      "tokens_seen": 1780350976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023264794383149448,
      "loss": 2.7737,
      "theoretical_loss": 3.4642019906091424,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023263791374122366,
      "loss": 2.5223,
      "theoretical_loss": 3.464191114219221,
      "tokens_seen": 1780482048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023262788365095287,
      "loss": 2.6733,
      "theoretical_loss": 3.464180238341721,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023261785356068205,
      "loss": 2.5102,
      "theoretical_loss": 3.4641693629765986,
      "tokens_seen": 1780613120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023260782347041124,
      "loss": 2.5524,
      "theoretical_loss": 3.464158488123811,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023259779338014042,
      "loss": 2.6851,
      "theoretical_loss": 3.4641476137833154,
      "tokens_seen": 1780744192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325877632898696,
      "loss": 2.5742,
      "theoretical_loss": 3.464136739955069,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325777331995988,
      "loss": 2.5445,
      "theoretical_loss": 3.464125866639028,
      "tokens_seen": 1780875264
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 981867,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0243661403656006,
      "objective/train/theoretical_loss": 3.4641149938351505,
      "objective/train/tokens_used": 1801400800,
      "theoretical_loss": 3.4641149938351505,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232567703109328,
      "loss": 2.5917,
      "theoretical_loss": 3.4641149938351505,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023255767301905717,
      "loss": 2.3419,
      "theoretical_loss": 3.464104121543393,
      "tokens_seen": 1781006336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023254764292878635,
      "loss": 2.4154,
      "theoretical_loss": 3.4640932497637125,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023253761283851556,
      "loss": 2.5471,
      "theoretical_loss": 3.464082378496066,
      "tokens_seen": 1781137408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023252758274824474,
      "loss": 2.6907,
      "theoretical_loss": 3.4640715077404103,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023251755265797392,
      "loss": 2.5956,
      "theoretical_loss": 3.4640606374967033,
      "tokens_seen": 1781268480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325075225677031,
      "loss": 2.6235,
      "theoretical_loss": 3.464049767764901,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023249749247743228,
      "loss": 2.7565,
      "theoretical_loss": 3.4640388985449615,
      "tokens_seen": 1781399552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324874623871615,
      "loss": 2.4995,
      "theoretical_loss": 3.4640280298368413,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023247743229689067,
      "loss": 2.7173,
      "theoretical_loss": 3.4640171616404976,
      "tokens_seen": 1781530624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023246740220661985,
      "loss": 2.5211,
      "theoretical_loss": 3.4640062939558867,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023245737211634903,
      "loss": 2.5921,
      "theoretical_loss": 3.4639954267829673,
      "tokens_seen": 1781661696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023244734202607824,
      "loss": 2.6692,
      "theoretical_loss": 3.4639845601216948,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023243731193580742,
      "loss": 2.6406,
      "theoretical_loss": 3.4639736939720276,
      "tokens_seen": 1781792768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324272818455366,
      "loss": 2.571,
      "theoretical_loss": 3.4639628283339223,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023241725175526578,
      "loss": 2.3942,
      "theoretical_loss": 3.463951963207336,
      "tokens_seen": 1781923840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023240722166499497,
      "loss": 2.5153,
      "theoretical_loss": 3.463941098592225,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023239719157472417,
      "loss": 2.7882,
      "theoretical_loss": 3.463930234488548,
      "tokens_seen": 1782054912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023238716148445336,
      "loss": 2.7587,
      "theoretical_loss": 3.4639193708962615,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023237713139418254,
      "loss": 2.4593,
      "theoretical_loss": 3.4639085078153222,
      "tokens_seen": 1782185984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023236710130391172,
      "loss": 2.5567,
      "theoretical_loss": 3.4638976452456873,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023235707121364095,
      "loss": 2.6562,
      "theoretical_loss": 3.4638867831873137,
      "tokens_seen": 1782317056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023234704112337013,
      "loss": 2.6656,
      "theoretical_loss": 3.4638759216401596,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023233701103309932,
      "loss": 2.5727,
      "theoretical_loss": 3.463865060604181,
      "tokens_seen": 1782448128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002323269809428285,
      "loss": 2.6451,
      "theoretical_loss": 3.4638542000793358,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 982914,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7162647247314453,
      "objective/train/theoretical_loss": 3.463843340065581,
      "objective/train/tokens_used": 1803039200,
      "theoretical_loss": 3.463843340065581,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023231695085255768,
      "loss": 2.5877,
      "theoretical_loss": 3.463843340065581,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023230692076228689,
      "loss": 2.7033,
      "theoretical_loss": 3.463832480562873,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023229689067201607,
      "loss": 2.7852,
      "theoretical_loss": 3.463821621571171,
      "tokens_seen": 1782710272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023228686058174525,
      "loss": 2.436,
      "theoretical_loss": 3.4638107630904296,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023227683049147443,
      "loss": 2.6596,
      "theoretical_loss": 3.4637999051206076,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023226680040120364,
      "loss": 2.5633,
      "theoretical_loss": 3.4637890476616615,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023225677031093282,
      "loss": 2.6062,
      "theoretical_loss": 3.4637781907135485,
      "tokens_seen": 1782972416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232246740220662,
      "loss": 2.6656,
      "theoretical_loss": 3.4637673342762265,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023223671013039118,
      "loss": 2.5826,
      "theoretical_loss": 3.463756478349652,
      "tokens_seen": 1783103488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023222668004012036,
      "loss": 2.5336,
      "theoretical_loss": 3.463745622933782,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023221664994984957,
      "loss": 2.6512,
      "theoretical_loss": 3.4637347680285746,
      "tokens_seen": 1783234560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023220661985957875,
      "loss": 2.6612,
      "theoretical_loss": 3.463723913633987,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023219658976930793,
      "loss": 2.6005,
      "theoretical_loss": 3.463713059749975,
      "tokens_seen": 1783365632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321865596790371,
      "loss": 2.8162,
      "theoretical_loss": 3.4637022063764977,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023217652958876632,
      "loss": 2.4618,
      "theoretical_loss": 3.4636913535135108,
      "tokens_seen": 1783496704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321664994984955,
      "loss": 2.4945,
      "theoretical_loss": 3.4636805011609724,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023215646940822468,
      "loss": 2.6025,
      "theoretical_loss": 3.463669649318839,
      "tokens_seen": 1783627776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023214643931795386,
      "loss": 2.6947,
      "theoretical_loss": 3.4636587979870685,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023213640922768307,
      "loss": 2.6345,
      "theoretical_loss": 3.463647947165618,
      "tokens_seen": 1783758848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023212637913741225,
      "loss": 2.5436,
      "theoretical_loss": 3.4636370968544448,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023211634904714144,
      "loss": 2.6925,
      "theoretical_loss": 3.463626247053506,
      "tokens_seen": 1783889920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023210631895687062,
      "loss": 2.6989,
      "theoretical_loss": 3.4636153977627586,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320962888665998,
      "loss": 2.6169,
      "theoretical_loss": 3.4636045489821607,
      "tokens_seen": 1784020992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232086258776329,
      "loss": 2.4932,
      "theoretical_loss": 3.4635937007116686,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320762286860582,
      "loss": 2.5965,
      "theoretical_loss": 3.4635828529512405,
      "tokens_seen": 1784152064
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 984019,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.507772207260132,
      "objective/train/theoretical_loss": 3.4635720057008332,
      "objective/train/tokens_used": 1804677600,
      "theoretical_loss": 3.4635720057008332,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023206619859578737,
      "loss": 2.4134,
      "theoretical_loss": 3.4635720057008332,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023205616850551655,
      "loss": 2.4492,
      "theoretical_loss": 3.463561158960404,
      "tokens_seen": 1784283136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023204613841524576,
      "loss": 2.4714,
      "theoretical_loss": 3.4635503127299097,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023203610832497494,
      "loss": 2.618,
      "theoretical_loss": 3.4635394670093085,
      "tokens_seen": 1784414208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023202607823470412,
      "loss": 2.719,
      "theoretical_loss": 3.4635286217985572,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320160481444333,
      "loss": 2.4438,
      "theoretical_loss": 3.4635177770976133,
      "tokens_seen": 1784545280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023200601805416248,
      "loss": 2.5049,
      "theoretical_loss": 3.463506932906434,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319959879638917,
      "loss": 2.4785,
      "theoretical_loss": 3.463496089224977,
      "tokens_seen": 1784676352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023198595787362087,
      "loss": 2.4071,
      "theoretical_loss": 3.4634852460531986,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023197592778335005,
      "loss": 2.5804,
      "theoretical_loss": 3.463474403391057,
      "tokens_seen": 1784807424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023196589769307923,
      "loss": 2.5506,
      "theoretical_loss": 3.46346356123851,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023195586760280844,
      "loss": 2.5477,
      "theoretical_loss": 3.4634527195955136,
      "tokens_seen": 1784938496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023194583751253762,
      "loss": 2.6944,
      "theoretical_loss": 3.463441878462026,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319358074222668,
      "loss": 2.8006,
      "theoretical_loss": 3.463431037838004,
      "tokens_seen": 1785069568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023192577733199599,
      "loss": 2.5052,
      "theoretical_loss": 3.463420197723406,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023191574724172517,
      "loss": 2.692,
      "theoretical_loss": 3.463409358118188,
      "tokens_seen": 1785200640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023190571715145437,
      "loss": 2.4587,
      "theoretical_loss": 3.4633985190223084,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023189568706118356,
      "loss": 2.7668,
      "theoretical_loss": 3.4633876804357246,
      "tokens_seen": 1785331712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023188565697091274,
      "loss": 2.5144,
      "theoretical_loss": 3.463376842358393,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023187562688064192,
      "loss": 2.336,
      "theoretical_loss": 3.463366004790272,
      "tokens_seen": 1785462784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023186559679037113,
      "loss": 2.7589,
      "theoretical_loss": 3.4633551677313186,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002318555667001003,
      "loss": 2.4846,
      "theoretical_loss": 3.46334433118149,
      "tokens_seen": 1785593856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002318455366098295,
      "loss": 2.598,
      "theoretical_loss": 3.463333495140744,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023183550651955867,
      "loss": 2.6523,
      "theoretical_loss": 3.4633226596090374,
      "tokens_seen": 1785724928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023182547642928785,
      "loss": 2.5686,
      "theoretical_loss": 3.463311824586328,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 984652,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3390166759490967,
      "objective/train/theoretical_loss": 3.4633009900725735,
      "objective/train/tokens_used": 1806316000,
      "theoretical_loss": 3.4633009900725735,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023181544633901706,
      "loss": 2.5698,
      "theoretical_loss": 3.4633009900725735,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023180541624874624,
      "loss": 2.4275,
      "theoretical_loss": 3.463290156067731,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023179538615847542,
      "loss": 2.6638,
      "theoretical_loss": 3.4632793225717573,
      "tokens_seen": 1785987072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002317853560682046,
      "loss": 2.7177,
      "theoretical_loss": 3.463268489584611,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002317753259779338,
      "loss": 2.5091,
      "theoretical_loss": 3.463257657106249,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231765295887663,
      "loss": 2.5267,
      "theoretical_loss": 3.463246825136628,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023175526579739217,
      "loss": 2.5578,
      "theoretical_loss": 3.463235993675707,
      "tokens_seen": 1786249216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023174523570712135,
      "loss": 2.3482,
      "theoretical_loss": 3.4632251627234423,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023173520561685053,
      "loss": 2.6927,
      "theoretical_loss": 3.4632143322797915,
      "tokens_seen": 1786380288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023172517552657974,
      "loss": 2.631,
      "theoretical_loss": 3.4632035023447125,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023171514543630892,
      "loss": 2.5262,
      "theoretical_loss": 3.4631926729181624,
      "tokens_seen": 1786511360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002317051153460381,
      "loss": 2.6821,
      "theoretical_loss": 3.4631818440000988,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023169508525576729,
      "loss": 2.7247,
      "theoretical_loss": 3.4631710155904787,
      "tokens_seen": 1786642432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316850551654965,
      "loss": 2.7414,
      "theoretical_loss": 3.4631601876892604,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023167502507522568,
      "loss": 2.5233,
      "theoretical_loss": 3.4631493602964007,
      "tokens_seen": 1786773504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023166499498495486,
      "loss": 2.6378,
      "theoretical_loss": 3.4631385334118576,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023165496489468404,
      "loss": 2.7249,
      "theoretical_loss": 3.4631277070355884,
      "tokens_seen": 1786904576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023164493480441322,
      "loss": 2.5564,
      "theoretical_loss": 3.463116881167551,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023163490471414243,
      "loss": 2.4779,
      "theoretical_loss": 3.4631060558077014,
      "tokens_seen": 1787035648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316248746238716,
      "loss": 2.6421,
      "theoretical_loss": 3.463095230955999,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316148445336008,
      "loss": 2.5259,
      "theoretical_loss": 3.4630844066124,
      "tokens_seen": 1787166720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023160481444333,
      "loss": 2.5863,
      "theoretical_loss": 3.463073582776863,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315947843530592,
      "loss": 2.5898,
      "theoretical_loss": 3.463062759449344,
      "tokens_seen": 1787297792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315847542627884,
      "loss": 2.6064,
      "theoretical_loss": 3.4630519366298023,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023157472417251757,
      "loss": 2.3533,
      "theoretical_loss": 3.4630411143181945,
      "tokens_seen": 1787428864
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 985931,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8117852210998535,
      "objective/train/theoretical_loss": 3.4630302925144782,
      "objective/train/tokens_used": 1807954400,
      "theoretical_loss": 3.4630302925144782,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023156469408224675,
      "loss": 2.6035,
      "theoretical_loss": 3.4630302925144782,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023155466399197596,
      "loss": 2.7356,
      "theoretical_loss": 3.463019471218611,
      "tokens_seen": 1787559936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023154463390170514,
      "loss": 2.6366,
      "theoretical_loss": 3.4630086504305497,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023153460381143432,
      "loss": 2.647,
      "theoretical_loss": 3.462997830150253,
      "tokens_seen": 1787691008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315245737211635,
      "loss": 2.6878,
      "theoretical_loss": 3.4629870103776783,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023151454363089268,
      "loss": 2.4871,
      "theoretical_loss": 3.462976191112783,
      "tokens_seen": 1787822080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315045135406219,
      "loss": 2.4998,
      "theoretical_loss": 3.4629653723555247,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023149448345035107,
      "loss": 2.4069,
      "theoretical_loss": 3.4629545541058606,
      "tokens_seen": 1787953152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023148445336008025,
      "loss": 2.4436,
      "theoretical_loss": 3.4629437363637483,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023147442326980943,
      "loss": 2.5846,
      "theoretical_loss": 3.462932919129146,
      "tokens_seen": 1788084224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023146439317953864,
      "loss": 2.6915,
      "theoretical_loss": 3.4629221024020107,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023145436308926782,
      "loss": 2.4596,
      "theoretical_loss": 3.4629112861823,
      "tokens_seen": 1788215296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231444332998997,
      "loss": 2.6897,
      "theoretical_loss": 3.4629004704699726,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023143430290872619,
      "loss": 2.4341,
      "theoretical_loss": 3.462889655264984,
      "tokens_seen": 1788346368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023142427281845537,
      "loss": 2.4791,
      "theoretical_loss": 3.462878840567294,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023141424272818457,
      "loss": 2.6216,
      "theoretical_loss": 3.4628680263768588,
      "tokens_seen": 1788477440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023140421263791376,
      "loss": 2.5357,
      "theoretical_loss": 3.4628572126936366,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023139418254764294,
      "loss": 2.5171,
      "theoretical_loss": 3.4628463995175847,
      "tokens_seen": 1788608512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023138415245737212,
      "loss": 2.8277,
      "theoretical_loss": 3.4628355868486613,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023137412236710133,
      "loss": 2.3535,
      "theoretical_loss": 3.4628247746868235,
      "tokens_seen": 1788739584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002313640922768305,
      "loss": 2.6147,
      "theoretical_loss": 3.462813963032029,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002313540621865597,
      "loss": 2.6034,
      "theoretical_loss": 3.4628031518842355,
      "tokens_seen": 1788870656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023134403209628887,
      "loss": 2.3671,
      "theoretical_loss": 3.462792341243401,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023133400200601805,
      "loss": 2.7327,
      "theoretical_loss": 3.4627815311094823,
      "tokens_seen": 1789001728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023132397191574726,
      "loss": 2.5874,
      "theoretical_loss": 3.462770721482438,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 986490,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7999305725097656,
      "objective/train/theoretical_loss": 3.4627599123622255,
      "objective/train/tokens_used": 1809592800,
      "theoretical_loss": 3.4627599123622255,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023131394182547644,
      "loss": 2.7609,
      "theoretical_loss": 3.4627599123622255,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023130391173520562,
      "loss": 2.5138,
      "theoretical_loss": 3.462749103748802,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312938816449348,
      "loss": 2.6086,
      "theoretical_loss": 3.4627382956421258,
      "tokens_seen": 1789263872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231283851554664,
      "loss": 2.5807,
      "theoretical_loss": 3.4627274880421544,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312738214643932,
      "loss": 2.7057,
      "theoretical_loss": 3.462716680948845,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023126379137412237,
      "loss": 2.5584,
      "theoretical_loss": 3.462705874362156,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023125376128385155,
      "loss": 2.4434,
      "theoretical_loss": 3.462695068282045,
      "tokens_seen": 1789526016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023124373119358073,
      "loss": 2.7263,
      "theoretical_loss": 3.4626842627084695,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023123370110330994,
      "loss": 2.5172,
      "theoretical_loss": 3.4626734576413867,
      "tokens_seen": 1789657088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023122367101303912,
      "loss": 2.342,
      "theoretical_loss": 3.4626626530807547,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312136409227683,
      "loss": 2.473,
      "theoretical_loss": 3.462651849026532,
      "tokens_seen": 1789788160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312036108324975,
      "loss": 2.1979,
      "theoretical_loss": 3.4626410454786756,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311935807422267,
      "loss": 2.4966,
      "theoretical_loss": 3.462630242437143,
      "tokens_seen": 1789919232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023118355065195588,
      "loss": 2.4444,
      "theoretical_loss": 3.462619439901892,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023117352056168506,
      "loss": 2.7513,
      "theoretical_loss": 3.462608637872881,
      "tokens_seen": 1790050304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023116349047141424,
      "loss": 2.4551,
      "theoretical_loss": 3.462597836350067,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023115346038114342,
      "loss": 2.5696,
      "theoretical_loss": 3.4625870353334083,
      "tokens_seen": 1790181376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023114343029087263,
      "loss": 2.5565,
      "theoretical_loss": 3.462576234822863,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311334002006018,
      "loss": 2.589,
      "theoretical_loss": 3.462565434818387,
      "tokens_seen": 1790312448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231123370110331,
      "loss": 2.5951,
      "theoretical_loss": 3.46255463531994,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023111334002006017,
      "loss": 2.1212,
      "theoretical_loss": 3.4625438363274794,
      "tokens_seen": 1790443520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023110330992978938,
      "loss": 2.538,
      "theoretical_loss": 3.462533037840962,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023109327983951856,
      "loss": 2.5986,
      "theoretical_loss": 3.4625222398603466,
      "tokens_seen": 1790574592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023108324974924774,
      "loss": 2.2605,
      "theoretical_loss": 3.4625114423855905,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023107321965897692,
      "loss": 2.5938,
      "theoretical_loss": 3.4625006454166516,
      "tokens_seen": 1790705664
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 987679,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3735997676849365,
      "objective/train/theoretical_loss": 3.4624898489534877,
      "objective/train/tokens_used": 1811231200,
      "theoretical_loss": 3.4624898489534877,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310631895687061,
      "loss": 2.5018,
      "theoretical_loss": 3.4624898489534877,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310531594784353,
      "loss": 2.6029,
      "theoretical_loss": 3.4624790529960574,
      "tokens_seen": 1790836736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310431293881645,
      "loss": 2.3636,
      "theoretical_loss": 3.462468257544317,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023103309929789367,
      "loss": 2.513,
      "theoretical_loss": 3.4624574625982247,
      "tokens_seen": 1790967808
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023102306920762286,
      "loss": 2.7336,
      "theoretical_loss": 3.462446668157739,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023101303911735206,
      "loss": 2.692,
      "theoretical_loss": 3.4624358742228174,
      "tokens_seen": 1791098880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023100300902708124,
      "loss": 2.6167,
      "theoretical_loss": 3.462425080793418,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023099297893681043,
      "loss": 2.6028,
      "theoretical_loss": 3.462414287869498,
      "tokens_seen": 1791229952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002309829488465396,
      "loss": 2.608,
      "theoretical_loss": 3.4624034954510154,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023097291875626881,
      "loss": 2.4834,
      "theoretical_loss": 3.4623927035379283,
      "tokens_seen": 1791361024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230962888665998,
      "loss": 2.3378,
      "theoretical_loss": 3.462381912130194,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023095285857572718,
      "loss": 2.6399,
      "theoretical_loss": 3.462371121227772,
      "tokens_seen": 1791492096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023094282848545636,
      "loss": 2.5145,
      "theoretical_loss": 3.462360330830618,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023093279839518554,
      "loss": 2.4724,
      "theoretical_loss": 3.462349540938691,
      "tokens_seen": 1791623168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023092276830491475,
      "loss": 2.4102,
      "theoretical_loss": 3.4623387515519486,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023091273821464393,
      "loss": 2.5437,
      "theoretical_loss": 3.462327962670349,
      "tokens_seen": 1791754240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002309027081243731,
      "loss": 2.6691,
      "theoretical_loss": 3.462317174293849,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308926780341023,
      "loss": 2.5777,
      "theoretical_loss": 3.462306386422408,
      "tokens_seen": 1791885312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308826479438315,
      "loss": 2.3827,
      "theoretical_loss": 3.462295599055983,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023087261785356068,
      "loss": 2.3883,
      "theoretical_loss": 3.4622848121945324,
      "tokens_seen": 1792016384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023086258776328986,
      "loss": 2.675,
      "theoretical_loss": 3.4622740258380134,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023085255767301907,
      "loss": 2.4986,
      "theoretical_loss": 3.4622632399863846,
      "tokens_seen": 1792147456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023084252758274825,
      "loss": 2.4722,
      "theoretical_loss": 3.4622524546396027,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023083249749247746,
      "loss": 2.5978,
      "theoretical_loss": 3.4622416697976277,
      "tokens_seen": 1792278528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023082246740220664,
      "loss": 2.5207,
      "theoretical_loss": 3.4622308854604156,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 988330,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3722500801086426,
      "objective/train/theoretical_loss": 3.462220101627925,
      "objective/train/tokens_used": 1812869600,
      "theoretical_loss": 3.462220101627925,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023081243731193582,
      "loss": 2.4544,
      "theoretical_loss": 3.462220101627925,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230802407221665,
      "loss": 2.4965,
      "theoretical_loss": 3.462209318300114,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307923771313942,
      "loss": 2.6264,
      "theoretical_loss": 3.46219853547694,
      "tokens_seen": 1792540672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307823470411234,
      "loss": 2.9242,
      "theoretical_loss": 3.4621877531583616,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023077231695085257,
      "loss": 2.4452,
      "theoretical_loss": 3.462176971344337,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023076228686058175,
      "loss": 2.6649,
      "theoretical_loss": 3.4621661900348224,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023075225677031094,
      "loss": 2.7592,
      "theoretical_loss": 3.4621554092297777,
      "tokens_seen": 1792802816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023074222668004014,
      "loss": 2.6149,
      "theoretical_loss": 3.46214462892916,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023073219658976932,
      "loss": 2.4829,
      "theoretical_loss": 3.462133849132927,
      "tokens_seen": 1792933888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307221664994985,
      "loss": 2.5123,
      "theoretical_loss": 3.462123069841038,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307121364092277,
      "loss": 2.5543,
      "theoretical_loss": 3.462112291053449,
      "tokens_seen": 1793064960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307021063189569,
      "loss": 2.5719,
      "theoretical_loss": 3.4621015127701193,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023069207622868608,
      "loss": 2.3829,
      "theoretical_loss": 3.4620907349910066,
      "tokens_seen": 1793196032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023068204613841526,
      "loss": 2.7034,
      "theoretical_loss": 3.462079957716069,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023067201604814444,
      "loss": 2.363,
      "theoretical_loss": 3.4620691809452637,
      "tokens_seen": 1793327104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023066198595787362,
      "loss": 2.6649,
      "theoretical_loss": 3.46205840467855,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023065195586760283,
      "loss": 2.5863,
      "theoretical_loss": 3.462047628915885,
      "tokens_seen": 1793458176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230641925777332,
      "loss": 2.5387,
      "theoretical_loss": 3.462036853657227,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306318956870612,
      "loss": 2.7292,
      "theoretical_loss": 3.4620260789025337,
      "tokens_seen": 1793589248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023062186559679037,
      "loss": 2.4789,
      "theoretical_loss": 3.4620153046517634,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023061183550651958,
      "loss": 2.4573,
      "theoretical_loss": 3.462004530904874,
      "tokens_seen": 1793720320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023060180541624876,
      "loss": 2.7241,
      "theoretical_loss": 3.461993757661824,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023059177532597794,
      "loss": 2.5425,
      "theoretical_loss": 3.461982984922571,
      "tokens_seen": 1793851392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023058174523570712,
      "loss": 2.8107,
      "theoretical_loss": 3.4619722126870727,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002305717151454363,
      "loss": 2.4525,
      "theoretical_loss": 3.461961440955288,
      "tokens_seen": 1793982464
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 989671,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0594289302825928,
      "objective/train/theoretical_loss": 3.461950669727174,
      "objective/train/tokens_used": 1814508000,
      "theoretical_loss": 3.461950669727174,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002305616850551655,
      "loss": 2.6317,
      "theoretical_loss": 3.461950669727174,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002305516549648947,
      "loss": 2.7489,
      "theoretical_loss": 3.4619398990026893,
      "tokens_seen": 1794113536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023054162487462387,
      "loss": 2.6713,
      "theoretical_loss": 3.461929128781792,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023053159478435306,
      "loss": 2.5276,
      "theoretical_loss": 3.4619183590644402,
      "tokens_seen": 1794244608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023052156469408226,
      "loss": 2.5841,
      "theoretical_loss": 3.4619075898505915,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023051153460381144,
      "loss": 2.441,
      "theoretical_loss": 3.461896821140204,
      "tokens_seen": 1794375680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023050150451354063,
      "loss": 2.4449,
      "theoretical_loss": 3.4618860529332363,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304914744232698,
      "loss": 2.5007,
      "theoretical_loss": 3.461875285229646,
      "tokens_seen": 1794506752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023048144433299902,
      "loss": 2.626,
      "theoretical_loss": 3.461864518029392,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304714142427282,
      "loss": 2.5948,
      "theoretical_loss": 3.4618537513324315,
      "tokens_seen": 1794637824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023046138415245738,
      "loss": 2.6399,
      "theoretical_loss": 3.4618429851387233,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023045135406218656,
      "loss": 2.4059,
      "theoretical_loss": 3.4618322194482243,
      "tokens_seen": 1794768896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023044132397191574,
      "loss": 2.6776,
      "theoretical_loss": 3.4618214542608943,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023043129388164495,
      "loss": 2.7367,
      "theoretical_loss": 3.46181068957669,
      "tokens_seen": 1794899968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023042126379137413,
      "loss": 2.1751,
      "theoretical_loss": 3.46179992539557,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304112337011033,
      "loss": 2.5735,
      "theoretical_loss": 3.4617891617174923,
      "tokens_seen": 1795031040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304012036108325,
      "loss": 2.7229,
      "theoretical_loss": 3.4617783985424158,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303911735205617,
      "loss": 2.4058,
      "theoretical_loss": 3.4617676358702973,
      "tokens_seen": 1795162112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023038114343029088,
      "loss": 2.424,
      "theoretical_loss": 3.461756873701096,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023037111334002006,
      "loss": 2.5511,
      "theoretical_loss": 3.4617461120347697,
      "tokens_seen": 1795293184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023036108324974924,
      "loss": 2.5153,
      "theoretical_loss": 3.461735350871277,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023035105315947842,
      "loss": 2.4963,
      "theoretical_loss": 3.461724590210575,
      "tokens_seen": 1795424256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023034102306920763,
      "loss": 2.3877,
      "theoretical_loss": 3.4617138300526222,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303309929789368,
      "loss": 2.6344,
      "theoretical_loss": 3.4617030703973777,
      "tokens_seen": 1795555328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230320962888666,
      "loss": 2.4278,
      "theoretical_loss": 3.461692311244798,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 990185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5343732833862305,
      "objective/train/theoretical_loss": 3.4616815525948432,
      "objective/train/tokens_used": 1816146400,
      "theoretical_loss": 3.4616815525948432,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023031093279839518,
      "loss": 2.4829,
      "theoretical_loss": 3.4616815525948432,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023030090270812438,
      "loss": 2.6864,
      "theoretical_loss": 3.4616707944474703,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023029087261785356,
      "loss": 2.5462,
      "theoretical_loss": 3.4616600368026376,
      "tokens_seen": 1795817472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023028084252758275,
      "loss": 2.524,
      "theoretical_loss": 3.4616492796603033,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023027081243731193,
      "loss": 2.7369,
      "theoretical_loss": 3.461638523020426,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002302607823470411,
      "loss": 2.4251,
      "theoretical_loss": 3.4616277668829634,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023025075225677032,
      "loss": 2.6792,
      "theoretical_loss": 3.461617011247874,
      "tokens_seen": 1796079616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002302407221664995,
      "loss": 2.5683,
      "theoretical_loss": 3.461606256115116,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023023069207622868,
      "loss": 2.5207,
      "theoretical_loss": 3.461595501484647,
      "tokens_seen": 1796210688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023022066198595786,
      "loss": 2.5099,
      "theoretical_loss": 3.4615847473564263,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023021063189568707,
      "loss": 2.7973,
      "theoretical_loss": 3.4615739937304113,
      "tokens_seen": 1796341760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023020060180541625,
      "loss": 2.6516,
      "theoretical_loss": 3.4615632406065604,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023019057171514543,
      "loss": 2.4848,
      "theoretical_loss": 3.461552487984832,
      "tokens_seen": 1796472832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002301805416248746,
      "loss": 2.3098,
      "theoretical_loss": 3.4615417358651843,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002301705115346038,
      "loss": 2.699,
      "theoretical_loss": 3.461530984247575,
      "tokens_seen": 1796603904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230160481444333,
      "loss": 2.3491,
      "theoretical_loss": 3.4615202331319637,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023015045135406218,
      "loss": 2.7843,
      "theoretical_loss": 3.4615094825183066,
      "tokens_seen": 1796734976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023014042126379136,
      "loss": 2.4498,
      "theoretical_loss": 3.4614987324065645,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023013039117352054,
      "loss": 2.5009,
      "theoretical_loss": 3.4614879827966933,
      "tokens_seen": 1796866048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023012036108324975,
      "loss": 2.6648,
      "theoretical_loss": 3.461477233688653,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023011033099297893,
      "loss": 2.5386,
      "theoretical_loss": 3.4614664850824,
      "tokens_seen": 1796997120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023010030090270814,
      "loss": 2.5846,
      "theoretical_loss": 3.4614557369778947,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023009027081243732,
      "loss": 2.5519,
      "theoretical_loss": 3.461444989375094,
      "tokens_seen": 1797128192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300802407221665,
      "loss": 2.4595,
      "theoretical_loss": 3.461434242273957,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300702106318957,
      "loss": 2.6782,
      "theoretical_loss": 3.461423495674441,
      "tokens_seen": 1797259264
    },
    {
      "epoch": 0.54,
      "objective/train/docs_used": 990815,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5431771278381348,
      "objective/train/theoretical_loss": 3.461412749576505,
      "objective/train/tokens_used": 1817784800,
      "theoretical_loss": 3.461412749576505,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300601805416249,
      "loss": 2.4925,
      "theoretical_loss": 3.461412749576505,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023005015045135407,
      "loss": 2.493,
      "theoretical_loss": 3.4614020039801074,
      "tokens_seen": 1797390336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023004012036108326,
      "loss": 2.7011,
      "theoretical_loss": 3.4613912588852065,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023003009027081246,
      "loss": 2.6043,
      "theoretical_loss": 3.46138051429176,
      "tokens_seen": 1797521408
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023002006018054164,
      "loss": 2.361,
      "theoretical_loss": 3.4613697701997266,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023001003009027083,
      "loss": 2.4372,
      "theoretical_loss": 3.461359026609065,
      "tokens_seen": 1797652480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023,
      "loss": 2.4974,
      "theoretical_loss": 3.4613482835197322,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022998996990972922,
      "loss": 2.3345,
      "theoretical_loss": 3.461337540931688,
      "tokens_seen": 1797783552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002299799398194584,
      "loss": 2.5875,
      "theoretical_loss": 3.4613267988448913,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022996990972918758,
      "loss": 2.6613,
      "theoretical_loss": 3.4613160572592983,
      "tokens_seen": 1797914624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022995987963891676,
      "loss": 2.6279,
      "theoretical_loss": 3.4613053161748684,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022994984954864594,
      "loss": 2.5921,
      "theoretical_loss": 3.46129457559156,
      "tokens_seen": 1798045696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022993981945837515,
      "loss": 2.6548,
      "theoretical_loss": 3.461283835509332,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022992978936810433,
      "loss": 2.5564,
      "theoretical_loss": 3.4612730959281413,
      "tokens_seen": 1798176768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002299197592778335,
      "loss": 2.7305,
      "theoretical_loss": 3.4612623568479477,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002299097291875627,
      "loss": 2.6617,
      "theoretical_loss": 3.461251618268709,
      "tokens_seen": 1798307840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002298996990972919,
      "loss": 2.3174,
      "theoretical_loss": 3.4612408801903833,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022988966900702108,
      "loss": 2.445,
      "theoretical_loss": 3.4612301426129295,
      "tokens_seen": 1798438912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022987963891675026,
      "loss": 2.5761,
      "theoretical_loss": 3.4612194055363057,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022986960882647944,
      "loss": 2.7026,
      "theoretical_loss": 3.4612086689604706,
      "tokens_seen": 1798569984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022985957873620862,
      "loss": 2.4218,
      "theoretical_loss": 3.461197932885382,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022984954864593783,
      "loss": 2.5596,
      "theoretical_loss": 3.461187197310999,
      "tokens_seen": 1798701056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229839518555667,
      "loss": 2.559,
      "theoretical_loss": 3.4611764622372796,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002298294884653962,
      "loss": 2.6447,
      "theoretical_loss": 3.4611657276641816,
      "tokens_seen": 1798832128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022981945837512538,
      "loss": 2.5122,
      "theoretical_loss": 3.4611549935916646,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 992118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3550384044647217,
      "objective/train/theoretical_loss": 3.4611442600196867,
      "objective/train/tokens_used": 1819423200,
      "theoretical_loss": 3.4611442600196867,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022980942828485458,
      "loss": 2.493,
      "theoretical_loss": 3.4611442600196867,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022979939819458376,
      "loss": 2.6462,
      "theoretical_loss": 3.461133526948206,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022978936810431295,
      "loss": 2.5465,
      "theoretical_loss": 3.4611227943771814,
      "tokens_seen": 1799094272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022977933801404213,
      "loss": 2.676,
      "theoretical_loss": 3.4611120623065705,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297693079237713,
      "loss": 2.4574,
      "theoretical_loss": 3.4611013307363323,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022975927783350052,
      "loss": 2.5421,
      "theoretical_loss": 3.4610905996664254,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297492477432297,
      "loss": 2.5227,
      "theoretical_loss": 3.4610798690968076,
      "tokens_seen": 1799356416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022973921765295888,
      "loss": 2.4198,
      "theoretical_loss": 3.461069139027438,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022972918756268806,
      "loss": 2.4443,
      "theoretical_loss": 3.461058409458275,
      "tokens_seen": 1799487488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022971915747241727,
      "loss": 2.5596,
      "theoretical_loss": 3.461047680389277,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022970912738214645,
      "loss": 2.5931,
      "theoretical_loss": 3.461036951820402,
      "tokens_seen": 1799618560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022969909729187563,
      "loss": 2.6325,
      "theoretical_loss": 3.4610262237516096,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296890672016048,
      "loss": 2.3865,
      "theoretical_loss": 3.4610154961828568,
      "tokens_seen": 1799749632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229679037111334,
      "loss": 2.7893,
      "theoretical_loss": 3.461004769114103,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296690070210632,
      "loss": 2.4781,
      "theoretical_loss": 3.4609940425453067,
      "tokens_seen": 1799880704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022965897693079238,
      "loss": 2.7497,
      "theoretical_loss": 3.460983316476426,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022964894684052156,
      "loss": 2.7175,
      "theoretical_loss": 3.46097259090742,
      "tokens_seen": 1800011776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022963891675025074,
      "loss": 2.6532,
      "theoretical_loss": 3.4609618658382466,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022962888665997995,
      "loss": 2.8623,
      "theoretical_loss": 3.4609511412688647,
      "tokens_seen": 1800142848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022961885656970913,
      "loss": 2.4953,
      "theoretical_loss": 3.4609404171992324,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022960882647943831,
      "loss": 2.4464,
      "theoretical_loss": 3.460929693629309,
      "tokens_seen": 1800273920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002295987963891675,
      "loss": 2.5907,
      "theoretical_loss": 3.4609189705590513,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022958876629889668,
      "loss": 2.6057,
      "theoretical_loss": 3.46090824798842,
      "tokens_seen": 1800404992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022957873620862589,
      "loss": 2.5402,
      "theoretical_loss": 3.460897525917373,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022956870611835507,
      "loss": 2.5154,
      "theoretical_loss": 3.4608868043458676,
      "tokens_seen": 1800536064
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 993246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6612374782562256,
      "objective/train/theoretical_loss": 3.4608760832738636,
      "objective/train/tokens_used": 1821061600,
      "theoretical_loss": 3.4608760832738636,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022955867602808425,
      "loss": 2.5881,
      "theoretical_loss": 3.4608760832738636,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022954864593781343,
      "loss": 2.4863,
      "theoretical_loss": 3.4608653627013193,
      "tokens_seen": 1800667136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022953861584754264,
      "loss": 2.6592,
      "theoretical_loss": 3.4608546426281928,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022952858575727182,
      "loss": 2.6084,
      "theoretical_loss": 3.460843923054443,
      "tokens_seen": 1800798208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229518555667001,
      "loss": 2.4182,
      "theoretical_loss": 3.4608332039800294,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022950852557673018,
      "loss": 2.622,
      "theoretical_loss": 3.460822485404909,
      "tokens_seen": 1800929280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022949849548645936,
      "loss": 2.3605,
      "theoretical_loss": 3.460811767329041,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022948846539618857,
      "loss": 2.5453,
      "theoretical_loss": 3.460801049752384,
      "tokens_seen": 1801060352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022947843530591775,
      "loss": 2.7276,
      "theoretical_loss": 3.4607903326748968,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022946840521564693,
      "loss": 2.521,
      "theoretical_loss": 3.4607796160965374,
      "tokens_seen": 1801191424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294583751253761,
      "loss": 2.464,
      "theoretical_loss": 3.460768900017265,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022944834503510532,
      "loss": 2.6981,
      "theoretical_loss": 3.4607581844370383,
      "tokens_seen": 1801322496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294383149448345,
      "loss": 2.5989,
      "theoretical_loss": 3.460747469355815,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022942828485456368,
      "loss": 2.7925,
      "theoretical_loss": 3.4607367547735546,
      "tokens_seen": 1801453568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022941825476429286,
      "loss": 2.6114,
      "theoretical_loss": 3.460726040690215,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022940822467402207,
      "loss": 2.5517,
      "theoretical_loss": 3.4607153271057554,
      "tokens_seen": 1801584640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022939819458375125,
      "loss": 2.6631,
      "theoretical_loss": 3.460704614020135,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022938816449348043,
      "loss": 2.4198,
      "theoretical_loss": 3.460693901433311,
      "tokens_seen": 1801715712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022937813440320962,
      "loss": 2.603,
      "theoretical_loss": 3.4606831893452425,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293681043129388,
      "loss": 2.6943,
      "theoretical_loss": 3.460672477755889,
      "tokens_seen": 1801846784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229358074222668,
      "loss": 2.621,
      "theoretical_loss": 3.4606617666652078,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293480441323972,
      "loss": 2.7226,
      "theoretical_loss": 3.4606510560731585,
      "tokens_seen": 1801977856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293380140421264,
      "loss": 2.5507,
      "theoretical_loss": 3.4606403459796997,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022932798395185558,
      "loss": 2.3955,
      "theoretical_loss": 3.4606296363847897,
      "tokens_seen": 1802108928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022931795386158478,
      "loss": 2.5324,
      "theoretical_loss": 3.4606189272883876,
      "tokens_seen": 1802174464
    },
    {
      "debugging/Self-BLEU-5": 0.16124925125063172,
      "debugging/distinct-1-grams": 0.9299999999999999,
      "debugging/distinct-2-grams": 0.9897959183673469,
      "debugging/entropy-1-grams": 4.131368988645327,
      "debugging/entropy-2-grams": 4.2431546043494865,
      "debugging/length": 335.5,
      "debugging/num_segments": 2,
      "epoch": 0.55,
      "objective/train/docs_used": 993693,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7326791286468506,
      "objective/train/theoretical_loss": 3.460608218690451,
      "objective/train/tokens_used": 1822700000,
      "theoretical_loss": 3.460608218690451,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022930792377131397,
      "loss": 2.4543,
      "theoretical_loss": 3.460608218690451,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022929789368104315,
      "loss": 2.3685,
      "theoretical_loss": 3.46059751059094,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022928786359077233,
      "loss": 2.3868,
      "theoretical_loss": 3.4605868029898126,
      "tokens_seen": 1802371072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292778335005015,
      "loss": 2.6462,
      "theoretical_loss": 3.4605760958870277,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022926780341023072,
      "loss": 2.3642,
      "theoretical_loss": 3.460565389282543,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292577733199599,
      "loss": 2.582,
      "theoretical_loss": 3.4605546831763188,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022924774322968908,
      "loss": 2.4063,
      "theoretical_loss": 3.460543977568313,
      "tokens_seen": 1802633216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022923771313941826,
      "loss": 2.5718,
      "theoretical_loss": 3.460533272458484,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022922768304914747,
      "loss": 2.4648,
      "theoretical_loss": 3.4605225678467906,
      "tokens_seen": 1802764288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022921765295887665,
      "loss": 2.4982,
      "theoretical_loss": 3.460511863733192,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022920762286860583,
      "loss": 2.5922,
      "theoretical_loss": 3.460501160117647,
      "tokens_seen": 1802895360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229197592778335,
      "loss": 2.7416,
      "theoretical_loss": 3.4604904570001134,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291875626880642,
      "loss": 2.6766,
      "theoretical_loss": 3.4604797543805503,
      "tokens_seen": 1803026432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291775325977934,
      "loss": 2.5336,
      "theoretical_loss": 3.460469052258917,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022916750250752258,
      "loss": 2.4009,
      "theoretical_loss": 3.460458350635172,
      "tokens_seen": 1803157504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022915747241725176,
      "loss": 2.484,
      "theoretical_loss": 3.460447649509274,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022914744232698094,
      "loss": 2.7901,
      "theoretical_loss": 3.4604369488811813,
      "tokens_seen": 1803288576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022913741223671015,
      "loss": 2.6256,
      "theoretical_loss": 3.4604262487508533,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022912738214643933,
      "loss": 2.4941,
      "theoretical_loss": 3.460415549118248,
      "tokens_seen": 1803419648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022911735205616851,
      "loss": 2.5909,
      "theoretical_loss": 3.460404849983325,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291073219658977,
      "loss": 2.6986,
      "theoretical_loss": 3.4603941513460423,
      "tokens_seen": 1803550720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022909729187562688,
      "loss": 2.4432,
      "theoretical_loss": 3.4603834532063598,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022908726178535609,
      "loss": 2.5145,
      "theoretical_loss": 3.460372755564235,
      "tokens_seen": 1803681792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022907723169508527,
      "loss": 2.5648,
      "theoretical_loss": 3.4603620584196273,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022906720160481445,
      "loss": 2.6688,
      "theoretical_loss": 3.4603513617724957,
      "tokens_seen": 1803812864
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 994874,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3781020641326904,
      "objective/train/theoretical_loss": 3.460340665622798,
      "objective/train/tokens_used": 1824338400,
      "theoretical_loss": 3.460340665622798,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022905717151454363,
      "loss": 2.6147,
      "theoretical_loss": 3.460340665622798,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022904714142427284,
      "loss": 2.5169,
      "theoretical_loss": 3.460329969970495,
      "tokens_seen": 1803943936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022903711133400202,
      "loss": 2.3934,
      "theoretical_loss": 3.460319274815543,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002290270812437312,
      "loss": 2.7032,
      "theoretical_loss": 3.460308580157902,
      "tokens_seen": 1804075008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022901705115346038,
      "loss": 2.5181,
      "theoretical_loss": 3.460297885997531,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022900702106318956,
      "loss": 2.2725,
      "theoretical_loss": 3.4602871923343894,
      "tokens_seen": 1804206080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022899699097291877,
      "loss": 2.4235,
      "theoretical_loss": 3.4602764991684345,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022898696088264795,
      "loss": 2.3512,
      "theoretical_loss": 3.460265806499626,
      "tokens_seen": 1804337152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022897693079237713,
      "loss": 2.5665,
      "theoretical_loss": 3.4602551143279228,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289669007021063,
      "loss": 2.4817,
      "theoretical_loss": 3.4602444226532834,
      "tokens_seen": 1804468224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022895687061183552,
      "loss": 2.6488,
      "theoretical_loss": 3.4602337314756664,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289468405215647,
      "loss": 2.672,
      "theoretical_loss": 3.460223040795032,
      "tokens_seen": 1804599296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022893681043129388,
      "loss": 2.4495,
      "theoretical_loss": 3.4602123506113376,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022892678034102306,
      "loss": 2.4111,
      "theoretical_loss": 3.4602016609245423,
      "tokens_seen": 1804730368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022891675025075227,
      "loss": 2.4011,
      "theoretical_loss": 3.460190971734605,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022890672016048145,
      "loss": 2.6372,
      "theoretical_loss": 3.4601802830414856,
      "tokens_seen": 1804861440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022889669007021063,
      "loss": 2.7543,
      "theoretical_loss": 3.4601695948451416,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022888665997993982,
      "loss": 2.5925,
      "theoretical_loss": 3.4601589071455323,
      "tokens_seen": 1804992512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228876629889669,
      "loss": 2.709,
      "theoretical_loss": 3.460148219942617,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002288665997993982,
      "loss": 2.6094,
      "theoretical_loss": 3.4601375332363538,
      "tokens_seen": 1805123584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002288565697091274,
      "loss": 2.4557,
      "theoretical_loss": 3.4601268470267024,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022884653961885657,
      "loss": 2.5044,
      "theoretical_loss": 3.4601161613136213,
      "tokens_seen": 1805254656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022883650952858575,
      "loss": 2.6799,
      "theoretical_loss": 3.4601054760970698,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022882647943831496,
      "loss": 2.6258,
      "theoretical_loss": 3.460094791377006,
      "tokens_seen": 1805385728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022881644934804414,
      "loss": 2.4599,
      "theoretical_loss": 3.4600841071533894,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 995886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.115525245666504,
      "objective/train/theoretical_loss": 3.460073423426179,
      "objective/train/tokens_used": 1825976800,
      "theoretical_loss": 3.460073423426179,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022880641925777332,
      "loss": 2.5756,
      "theoretical_loss": 3.460073423426179,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002287963891675025,
      "loss": 2.5222,
      "theoretical_loss": 3.460062740195333,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022878635907723168,
      "loss": 2.5245,
      "theoretical_loss": 3.4600520574608113,
      "tokens_seen": 1805647872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002287763289869609,
      "loss": 2.6723,
      "theoretical_loss": 3.4600413752225725,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022876629889669007,
      "loss": 2.632,
      "theoretical_loss": 3.460030693480575,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022875626880641925,
      "loss": 2.4997,
      "theoretical_loss": 3.460020012234778,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022874623871614843,
      "loss": 2.6967,
      "theoretical_loss": 3.460009331485141,
      "tokens_seen": 1805910016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022873620862587764,
      "loss": 2.5674,
      "theoretical_loss": 3.459998651231622,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022872617853560682,
      "loss": 2.6531,
      "theoretical_loss": 3.459987971474181,
      "tokens_seen": 1806041088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228716148445336,
      "loss": 2.7024,
      "theoretical_loss": 3.4599772922127765,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022870611835506518,
      "loss": 2.588,
      "theoretical_loss": 3.4599666134473672,
      "tokens_seen": 1806172160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022869608826479437,
      "loss": 2.8305,
      "theoretical_loss": 3.459955935177912,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022868605817452357,
      "loss": 2.4536,
      "theoretical_loss": 3.45994525740437,
      "tokens_seen": 1806303232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022867602808425276,
      "loss": 2.7428,
      "theoretical_loss": 3.459934580126701,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022866599799398194,
      "loss": 2.7283,
      "theoretical_loss": 3.459923903344863,
      "tokens_seen": 1806434304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022865596790371112,
      "loss": 2.4138,
      "theoretical_loss": 3.4599132270588155,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022864593781344033,
      "loss": 2.5059,
      "theoretical_loss": 3.4599025512685166,
      "tokens_seen": 1806565376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002286359077231695,
      "loss": 2.5339,
      "theoretical_loss": 3.459891875973927,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002286258776328987,
      "loss": 2.4039,
      "theoretical_loss": 3.459881201175004,
      "tokens_seen": 1806696448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022861584754262787,
      "loss": 2.6261,
      "theoretical_loss": 3.4598705268717076,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022860581745235705,
      "loss": 2.6975,
      "theoretical_loss": 3.459859853063996,
      "tokens_seen": 1806827520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022859578736208629,
      "loss": 2.3772,
      "theoretical_loss": 3.4598491797518287,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022858575727181547,
      "loss": 2.4887,
      "theoretical_loss": 3.4598385069351654,
      "tokens_seen": 1806958592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022857572718154465,
      "loss": 2.6731,
      "theoretical_loss": 3.459827834613964,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022856569709127383,
      "loss": 2.5727,
      "theoretical_loss": 3.459817162788184,
      "tokens_seen": 1807089664
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 996582,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9981250762939453,
      "objective/train/theoretical_loss": 3.4598064914577846,
      "objective/train/tokens_used": 1827615200,
      "theoretical_loss": 3.4598064914577846,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022855566700100304,
      "loss": 2.6678,
      "theoretical_loss": 3.4598064914577846,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022854563691073222,
      "loss": 2.6893,
      "theoretical_loss": 3.4597958206227246,
      "tokens_seen": 1807220736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002285356068204614,
      "loss": 2.6052,
      "theoretical_loss": 3.459785150282963,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022852557673019058,
      "loss": 2.885,
      "theoretical_loss": 3.459774480438459,
      "tokens_seen": 1807351808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022851554663991976,
      "loss": 2.6922,
      "theoretical_loss": 3.4597638110891715,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022850551654964897,
      "loss": 2.4562,
      "theoretical_loss": 3.4597531422350603,
      "tokens_seen": 1807482880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022849548645937815,
      "loss": 2.4693,
      "theoretical_loss": 3.459742473876083,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022848545636910733,
      "loss": 2.5967,
      "theoretical_loss": 3.4597318060122,
      "tokens_seen": 1807613952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002284754262788365,
      "loss": 2.616,
      "theoretical_loss": 3.4597211386433697,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022846539618856572,
      "loss": 2.7295,
      "theoretical_loss": 3.4597104717695517,
      "tokens_seen": 1807745024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002284553660982949,
      "loss": 2.4501,
      "theoretical_loss": 3.459699805390705,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022844533600802408,
      "loss": 2.7781,
      "theoretical_loss": 3.4596891395067875,
      "tokens_seen": 1807876096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022843530591775326,
      "loss": 2.4805,
      "theoretical_loss": 3.45967847411776,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022842527582748247,
      "loss": 2.4078,
      "theoretical_loss": 3.4596678092235806,
      "tokens_seen": 1808007168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022841524573721165,
      "loss": 2.7349,
      "theoretical_loss": 3.4596571448242086,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022840521564694084,
      "loss": 2.6379,
      "theoretical_loss": 3.459646480919603,
      "tokens_seen": 1808138240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022839518555667002,
      "loss": 2.6609,
      "theoretical_loss": 3.4596358175097235,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283851554663992,
      "loss": 2.4822,
      "theoretical_loss": 3.4596251545945282,
      "tokens_seen": 1808269312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283751253761284,
      "loss": 2.5232,
      "theoretical_loss": 3.4596144921739773,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283650952858576,
      "loss": 2.7273,
      "theoretical_loss": 3.45960383024803,
      "tokens_seen": 1808400384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022835506519558677,
      "loss": 2.5913,
      "theoretical_loss": 3.459593168816644,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022834503510531595,
      "loss": 2.5707,
      "theoretical_loss": 3.4595825078797793,
      "tokens_seen": 1808531456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022833500501504516,
      "loss": 2.4813,
      "theoretical_loss": 3.4595718474373953,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022832497492477434,
      "loss": 2.5099,
      "theoretical_loss": 3.459561187489451,
      "tokens_seen": 1808662528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022831494483450352,
      "loss": 2.5912,
      "theoretical_loss": 3.459550528035905,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 997137,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.588252305984497,
      "objective/train/theoretical_loss": 3.4595398690767176,
      "objective/train/tokens_used": 1829253600,
      "theoretical_loss": 3.4595398690767176,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283049147442327,
      "loss": 2.6841,
      "theoretical_loss": 3.4595398690767176,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022829488465396188,
      "loss": 2.467,
      "theoretical_loss": 3.459529210611847,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282848545636911,
      "loss": 2.5713,
      "theoretical_loss": 3.4595185526412524,
      "tokens_seen": 1808924672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022827482447342027,
      "loss": 2.3152,
      "theoretical_loss": 3.4595078951648937,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022826479438314945,
      "loss": 2.3585,
      "theoretical_loss": 3.4594972381827294,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022825476429287863,
      "loss": 2.4353,
      "theoretical_loss": 3.4594865816947182,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022824473420260784,
      "loss": 2.7453,
      "theoretical_loss": 3.4594759257008207,
      "tokens_seen": 1809186816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022823470411233702,
      "loss": 2.4329,
      "theoretical_loss": 3.4594652702009956,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282246740220662,
      "loss": 2.4256,
      "theoretical_loss": 3.4594546151952015,
      "tokens_seen": 1809317888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022821464393179538,
      "loss": 2.5174,
      "theoretical_loss": 3.4594439606833975,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022820461384152457,
      "loss": 2.5711,
      "theoretical_loss": 3.459433306665544,
      "tokens_seen": 1809448960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022819458375125377,
      "loss": 2.6714,
      "theoretical_loss": 3.4594226531415986,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022818455366098296,
      "loss": 2.6427,
      "theoretical_loss": 3.4594120001115223,
      "tokens_seen": 1809580032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022817452357071214,
      "loss": 2.5115,
      "theoretical_loss": 3.459401347575273,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022816449348044132,
      "loss": 2.6094,
      "theoretical_loss": 3.4593906955328104,
      "tokens_seen": 1809711104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022815446339017053,
      "loss": 2.5625,
      "theoretical_loss": 3.4593800439840936,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002281444332998997,
      "loss": 2.5657,
      "theoretical_loss": 3.459369392929082,
      "tokens_seen": 1809842176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002281344032096289,
      "loss": 2.5791,
      "theoretical_loss": 3.4593587423677348,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022812437311935807,
      "loss": 2.5698,
      "theoretical_loss": 3.459348092300011,
      "tokens_seen": 1809973248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022811434302908725,
      "loss": 2.5696,
      "theoretical_loss": 3.45933744272587,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022810431293881646,
      "loss": 2.7263,
      "theoretical_loss": 3.459326793645271,
      "tokens_seen": 1810104320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022809428284854564,
      "loss": 2.5073,
      "theoretical_loss": 3.4593161450581738,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022808425275827482,
      "loss": 2.6775,
      "theoretical_loss": 3.4593054969645367,
      "tokens_seen": 1810235392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228074222668004,
      "loss": 2.6855,
      "theoretical_loss": 3.45929484936432,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002280641925777332,
      "loss": 2.4663,
      "theoretical_loss": 3.459284202257482,
      "tokens_seen": 1810366464
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 998243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.327003240585327,
      "objective/train/theoretical_loss": 3.459273555643983,
      "objective/train/tokens_used": 1830892000,
      "theoretical_loss": 3.459273555643983,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002280541624874624,
      "loss": 2.6463,
      "theoretical_loss": 3.459273555643983,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022804413239719157,
      "loss": 2.6188,
      "theoretical_loss": 3.4592629095237815,
      "tokens_seen": 1810497536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022803410230692075,
      "loss": 2.7248,
      "theoretical_loss": 3.459252263896837,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022802407221664993,
      "loss": 2.4005,
      "theoretical_loss": 3.4592416187631088,
      "tokens_seen": 1810628608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022801404212637914,
      "loss": 2.6145,
      "theoretical_loss": 3.4592309741225558,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022800401203610832,
      "loss": 2.558,
      "theoretical_loss": 3.459220329975138,
      "tokens_seen": 1810759680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002279939819458375,
      "loss": 2.6414,
      "theoretical_loss": 3.4592096863208144,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022798395185556669,
      "loss": 2.4383,
      "theoretical_loss": 3.4591990431595443,
      "tokens_seen": 1810890752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002279739217652959,
      "loss": 2.5763,
      "theoretical_loss": 3.4591884004912874,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022796389167502508,
      "loss": 2.7086,
      "theoretical_loss": 3.4591777583160024,
      "tokens_seen": 1811021824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022795386158475426,
      "loss": 2.4515,
      "theoretical_loss": 3.4591671166336484,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022794383149448344,
      "loss": 2.6532,
      "theoretical_loss": 3.459156475444186,
      "tokens_seen": 1811152896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022793380140421262,
      "loss": 2.5284,
      "theoretical_loss": 3.4591458347475736,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022792377131394183,
      "loss": 2.4153,
      "theoretical_loss": 3.4591351945437703,
      "tokens_seen": 1811283968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000227913741223671,
      "loss": 2.5878,
      "theoretical_loss": 3.4591245548327363,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002279037111334002,
      "loss": 2.6512,
      "theoretical_loss": 3.45911391561443,
      "tokens_seen": 1811415040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022789368104312937,
      "loss": 2.5825,
      "theoretical_loss": 3.4591032768888117,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022788365095285858,
      "loss": 2.7617,
      "theoretical_loss": 3.4590926386558403,
      "tokens_seen": 1811546112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022787362086258776,
      "loss": 2.7387,
      "theoretical_loss": 3.459082000915475,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022786359077231694,
      "loss": 2.5445,
      "theoretical_loss": 3.4590713636676753,
      "tokens_seen": 1811677184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022785356068204612,
      "loss": 2.727,
      "theoretical_loss": 3.4590607269124005,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022784353059177536,
      "loss": 2.8106,
      "theoretical_loss": 3.459050090649611,
      "tokens_seen": 1811808256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022783350050150454,
      "loss": 2.6116,
      "theoretical_loss": 3.4590394548792642,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022782347041123372,
      "loss": 2.643,
      "theoretical_loss": 3.459028819601321,
      "tokens_seen": 1811939328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002278134403209629,
      "loss": 2.5035,
      "theoretical_loss": 3.459018184815741,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 999229,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2615716457366943,
      "objective/train/theoretical_loss": 3.459007550522482,
      "objective/train/tokens_used": 1832530400,
      "theoretical_loss": 3.459007550522482,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022780341023069208,
      "loss": 2.5426,
      "theoretical_loss": 3.459007550522482,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277933801404213,
      "loss": 2.6329,
      "theoretical_loss": 3.458996916721505,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022778335005015047,
      "loss": 2.6186,
      "theoretical_loss": 3.4589862834127683,
      "tokens_seen": 1812201472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022777331995987965,
      "loss": 2.553,
      "theoretical_loss": 3.458975650596232,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022776328986960883,
      "loss": 2.5496,
      "theoretical_loss": 3.458965018271855,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022775325977933804,
      "loss": 2.5544,
      "theoretical_loss": 3.458954386439598,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022774322968906722,
      "loss": 2.633,
      "theoretical_loss": 3.458943755099419,
      "tokens_seen": 1812463616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277331995987964,
      "loss": 2.6399,
      "theoretical_loss": 3.4589331242512777,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022772316950852558,
      "loss": 2.6812,
      "theoretical_loss": 3.458922493895134,
      "tokens_seen": 1812594688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022771313941825477,
      "loss": 2.6438,
      "theoretical_loss": 3.458911864030947,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022770310932798397,
      "loss": 2.377,
      "theoretical_loss": 3.4589012346586765,
      "tokens_seen": 1812725760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022769307923771316,
      "loss": 2.3926,
      "theoretical_loss": 3.458890605778281,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022768304914744234,
      "loss": 2.623,
      "theoretical_loss": 3.4588799773897216,
      "tokens_seen": 1812856832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022767301905717152,
      "loss": 2.4385,
      "theoretical_loss": 3.4588693494929563,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022766298896690073,
      "loss": 2.4331,
      "theoretical_loss": 3.458858722087945,
      "tokens_seen": 1812987904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002276529588766299,
      "loss": 2.7077,
      "theoretical_loss": 3.4588480951746474,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002276429287863591,
      "loss": 2.5307,
      "theoretical_loss": 3.4588374687530226,
      "tokens_seen": 1813118976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022763289869608827,
      "loss": 2.6907,
      "theoretical_loss": 3.4588268428230307,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022762286860581745,
      "loss": 2.5574,
      "theoretical_loss": 3.4588162173846304,
      "tokens_seen": 1813250048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022761283851554666,
      "loss": 2.6133,
      "theoretical_loss": 3.458805592437782,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022760280842527584,
      "loss": 2.3539,
      "theoretical_loss": 3.458794967982444,
      "tokens_seen": 1813381120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022759277833500502,
      "loss": 2.5913,
      "theoretical_loss": 3.458784344018577,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275827482447342,
      "loss": 2.5597,
      "theoretical_loss": 3.45877372054614,
      "tokens_seen": 1813512192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275727181544634,
      "loss": 2.6439,
      "theoretical_loss": 3.4587630975650923,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275626880641926,
      "loss": 2.4789,
      "theoretical_loss": 3.4587524750753937,
      "tokens_seen": 1813643264
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 999785,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.323129892349243,
      "objective/train/theoretical_loss": 3.4587418530770035,
      "objective/train/tokens_used": 1834168800,
      "theoretical_loss": 3.4587418530770035,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022755265797392177,
      "loss": 2.599,
      "theoretical_loss": 3.4587418530770035,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022754262788365095,
      "loss": 2.6465,
      "theoretical_loss": 3.4587312315698813,
      "tokens_seen": 1813774336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022753259779338013,
      "loss": 2.5155,
      "theoretical_loss": 3.458720610553987,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022752256770310934,
      "loss": 2.6989,
      "theoretical_loss": 3.4587099900292797,
      "tokens_seen": 1813905408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022751253761283852,
      "loss": 2.6475,
      "theoretical_loss": 3.4586993699957187,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275025075225677,
      "loss": 2.5841,
      "theoretical_loss": 3.4586887504532644,
      "tokens_seen": 1814036480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022749247743229689,
      "loss": 2.6628,
      "theoretical_loss": 3.458678131401876,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274824473420261,
      "loss": 2.3507,
      "theoretical_loss": 3.4586675128415125,
      "tokens_seen": 1814167552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022747241725175528,
      "loss": 2.6697,
      "theoretical_loss": 3.4586568947721337,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022746238716148446,
      "loss": 2.539,
      "theoretical_loss": 3.4586462771936994,
      "tokens_seen": 1814298624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022745235707121364,
      "loss": 2.6562,
      "theoretical_loss": 3.4586356601061694,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022744232698094282,
      "loss": 2.4885,
      "theoretical_loss": 3.4586250435095027,
      "tokens_seen": 1814429696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022743229689067203,
      "loss": 2.7783,
      "theoretical_loss": 3.4586144274036594,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274222668004012,
      "loss": 2.5598,
      "theoretical_loss": 3.4586038117885987,
      "tokens_seen": 1814560768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274122367101304,
      "loss": 2.5368,
      "theoretical_loss": 3.4585931966642804,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022740220661985957,
      "loss": 2.7085,
      "theoretical_loss": 3.458582582030664,
      "tokens_seen": 1814691840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022739217652958878,
      "loss": 2.5134,
      "theoretical_loss": 3.4585719678877087,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022738214643931796,
      "loss": 2.4593,
      "theoretical_loss": 3.458561354235375,
      "tokens_seen": 1814822912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022737211634904714,
      "loss": 2.5849,
      "theoretical_loss": 3.458550741073622,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022736208625877632,
      "loss": 2.6529,
      "theoretical_loss": 3.458540128402409,
      "tokens_seen": 1814953984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273520561685055,
      "loss": 2.6592,
      "theoretical_loss": 3.458529516221696,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273420260782347,
      "loss": 2.7447,
      "theoretical_loss": 3.4585189045314424,
      "tokens_seen": 1815085056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273319959879639,
      "loss": 2.6106,
      "theoretical_loss": 3.458508293331609,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022732196589769307,
      "loss": 2.4851,
      "theoretical_loss": 3.458497682622153,
      "tokens_seen": 1815216128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022731193580742225,
      "loss": 2.4403,
      "theoretical_loss": 3.4584870724030363,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1000504,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.408949851989746,
      "objective/train/theoretical_loss": 3.4584764626742173,
      "objective/train/tokens_used": 1835807200,
      "theoretical_loss": 3.4584764626742173,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022730190571715146,
      "loss": 2.5544,
      "theoretical_loss": 3.4584764626742173,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022729187562688064,
      "loss": 2.672,
      "theoretical_loss": 3.4584658534356567,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022728184553660983,
      "loss": 2.516,
      "theoretical_loss": 3.458455244687313,
      "tokens_seen": 1815478272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000227271815446339,
      "loss": 2.5069,
      "theoretical_loss": 3.458444636429146,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022726178535606821,
      "loss": 2.5702,
      "theoretical_loss": 3.458434028661116,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002272517552657974,
      "loss": 2.503,
      "theoretical_loss": 3.4584234213831824,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022724172517552658,
      "loss": 2.5751,
      "theoretical_loss": 3.458412814595305,
      "tokens_seen": 1815740416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022723169508525576,
      "loss": 2.5795,
      "theoretical_loss": 3.458402208297443,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022722166499498494,
      "loss": 2.3569,
      "theoretical_loss": 3.4583916024895567,
      "tokens_seen": 1815871488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022721163490471415,
      "loss": 2.7036,
      "theoretical_loss": 3.458380997171605,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022720160481444333,
      "loss": 2.4939,
      "theoretical_loss": 3.4583703923435483,
      "tokens_seen": 1816002560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271915747241725,
      "loss": 2.5402,
      "theoretical_loss": 3.458359788005346,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271815446339017,
      "loss": 2.6088,
      "theoretical_loss": 3.458349184156958,
      "tokens_seen": 1816133632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271715145436309,
      "loss": 2.565,
      "theoretical_loss": 3.458338580798343,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022716148445336008,
      "loss": 2.7261,
      "theoretical_loss": 3.4583279779294624,
      "tokens_seen": 1816264704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022715145436308926,
      "loss": 2.8093,
      "theoretical_loss": 3.458317375550275,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022714142427281844,
      "loss": 2.564,
      "theoretical_loss": 3.4583067736607402,
      "tokens_seen": 1816395776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022713139418254762,
      "loss": 2.648,
      "theoretical_loss": 3.4582961722608188,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022712136409227683,
      "loss": 2.5645,
      "theoretical_loss": 3.458285571350469,
      "tokens_seen": 1816526848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000227111334002006,
      "loss": 2.6256,
      "theoretical_loss": 3.4582749709296516,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022710130391173522,
      "loss": 2.5501,
      "theoretical_loss": 3.458264370998326,
      "tokens_seen": 1816657920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270912738214644,
      "loss": 2.7177,
      "theoretical_loss": 3.4582537715564525,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270812437311936,
      "loss": 2.6629,
      "theoretical_loss": 3.45824317260399,
      "tokens_seen": 1816788992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270712136409228,
      "loss": 2.6181,
      "theoretical_loss": 3.458232574140899,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022706118355065197,
      "loss": 2.5951,
      "theoretical_loss": 3.4582219761671382,
      "tokens_seen": 1816920064
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1001166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3694236278533936,
      "objective/train/theoretical_loss": 3.4582113786826683,
      "objective/train/tokens_used": 1837445600,
      "theoretical_loss": 3.4582113786826683,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022705115346038115,
      "loss": 2.5311,
      "theoretical_loss": 3.4582113786826683,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022704112337011033,
      "loss": 2.6937,
      "theoretical_loss": 3.4582007816874487,
      "tokens_seen": 1817051136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022703109327983954,
      "loss": 2.5823,
      "theoretical_loss": 3.4581901851814396,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022702106318956872,
      "loss": 2.8356,
      "theoretical_loss": 3.4581795891646,
      "tokens_seen": 1817182208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270110330992979,
      "loss": 2.5063,
      "theoretical_loss": 3.4581689936368907,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022700100300902709,
      "loss": 2.4515,
      "theoretical_loss": 3.45815839859827,
      "tokens_seen": 1817313280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002269909729187563,
      "loss": 2.4596,
      "theoretical_loss": 3.458147804048699,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022698094282848548,
      "loss": 2.7244,
      "theoretical_loss": 3.458137209988138,
      "tokens_seen": 1817444352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022697091273821466,
      "loss": 2.5728,
      "theoretical_loss": 3.4581266164165445,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022696088264794384,
      "loss": 2.3857,
      "theoretical_loss": 3.45811602333388,
      "tokens_seen": 1817575424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022695085255767302,
      "loss": 2.563,
      "theoretical_loss": 3.4581054307401047,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022694082246740223,
      "loss": 2.5681,
      "theoretical_loss": 3.458094838635177,
      "tokens_seen": 1817706496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002269307923771314,
      "loss": 2.7895,
      "theoretical_loss": 3.458084247019057,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002269207622868606,
      "loss": 2.4003,
      "theoretical_loss": 3.458073655891706,
      "tokens_seen": 1817837568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022691073219658977,
      "loss": 2.527,
      "theoretical_loss": 3.458063065253082,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022690070210631898,
      "loss": 2.5022,
      "theoretical_loss": 3.4580524751031456,
      "tokens_seen": 1817968640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022689067201604816,
      "loss": 2.5966,
      "theoretical_loss": 3.4580418854418573,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022688064192577734,
      "loss": 2.5962,
      "theoretical_loss": 3.4580312962691755,
      "tokens_seen": 1818099712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022687061183550652,
      "loss": 2.4891,
      "theoretical_loss": 3.458020707585061,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268605817452357,
      "loss": 2.6084,
      "theoretical_loss": 3.4580101193894732,
      "tokens_seen": 1818230784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268505516549649,
      "loss": 2.5567,
      "theoretical_loss": 3.457999531682373,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268405215646941,
      "loss": 2.8011,
      "theoretical_loss": 3.4579889444637186,
      "tokens_seen": 1818361856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022683049147442327,
      "loss": 2.5099,
      "theoretical_loss": 3.457978357733471,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022682046138415245,
      "loss": 2.7922,
      "theoretical_loss": 3.4579677714915897,
      "tokens_seen": 1818492928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022681043129388166,
      "loss": 2.5979,
      "theoretical_loss": 3.457957185738035,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1002311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.384216070175171,
      "objective/train/theoretical_loss": 3.4579466004727664,
      "objective/train/tokens_used": 1839084000,
      "theoretical_loss": 3.4579466004727664,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022680040120361084,
      "loss": 2.3402,
      "theoretical_loss": 3.4579466004727664,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022679037111334003,
      "loss": 2.7363,
      "theoretical_loss": 3.4579360156957435,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267803410230692,
      "loss": 2.7127,
      "theoretical_loss": 3.4579254314069265,
      "tokens_seen": 1818755072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022677031093279841,
      "loss": 2.3095,
      "theoretical_loss": 3.4579148476062755,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267602808425276,
      "loss": 2.6191,
      "theoretical_loss": 3.4579042642937505,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022675025075225678,
      "loss": 2.6294,
      "theoretical_loss": 3.4578936814693106,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022674022066198596,
      "loss": 2.5632,
      "theoretical_loss": 3.4578830991329164,
      "tokens_seen": 1819017216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022673019057171514,
      "loss": 2.5216,
      "theoretical_loss": 3.4578725172845273,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022672016048144435,
      "loss": 2.5518,
      "theoretical_loss": 3.457861935924104,
      "tokens_seen": 1819148288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022671013039117353,
      "loss": 2.6524,
      "theoretical_loss": 3.4578513550516057,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267001003009027,
      "loss": 2.561,
      "theoretical_loss": 3.457840774666993,
      "tokens_seen": 1819279360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266900702106319,
      "loss": 2.5346,
      "theoretical_loss": 3.457830194770225,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266800401203611,
      "loss": 2.7574,
      "theoretical_loss": 3.4578196153612617,
      "tokens_seen": 1819410432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022667001003009028,
      "loss": 2.5112,
      "theoretical_loss": 3.457809036440064,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022665997993981946,
      "loss": 2.6201,
      "theoretical_loss": 3.4577984580065912,
      "tokens_seen": 1819541504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022664994984954864,
      "loss": 2.5261,
      "theoretical_loss": 3.4577878800608026,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022663991975927782,
      "loss": 2.3837,
      "theoretical_loss": 3.4577773026026595,
      "tokens_seen": 1819672576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022662988966900703,
      "loss": 2.6788,
      "theoretical_loss": 3.457766725632121,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266198595787362,
      "loss": 2.6624,
      "theoretical_loss": 3.457756149149147,
      "tokens_seen": 1819803648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266098294884654,
      "loss": 2.615,
      "theoretical_loss": 3.457745573153698,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022659979939819458,
      "loss": 2.618,
      "theoretical_loss": 3.4577349976457334,
      "tokens_seen": 1819934720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022658976930792378,
      "loss": 2.3105,
      "theoretical_loss": 3.4577244226252137,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022657973921765296,
      "loss": 2.4802,
      "theoretical_loss": 3.457713848092099,
      "tokens_seen": 1820065792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022656970912738215,
      "loss": 2.7722,
      "theoretical_loss": 3.4577032740463483,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022655967903711133,
      "loss": 2.7564,
      "theoretical_loss": 3.4576927004879225,
      "tokens_seen": 1820196864
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1003206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.480213165283203,
      "objective/train/theoretical_loss": 3.4576821274167813,
      "objective/train/tokens_used": 1840722400,
      "theoretical_loss": 3.4576821274167813,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002265496489468405,
      "loss": 2.4943,
      "theoretical_loss": 3.4576821274167813,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022653961885656972,
      "loss": 2.7067,
      "theoretical_loss": 3.4576715548328845,
      "tokens_seen": 1820327936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002265295887662989,
      "loss": 2.4453,
      "theoretical_loss": 3.4576609827361926,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022651955867602808,
      "loss": 2.5459,
      "theoretical_loss": 3.4576504111266653,
      "tokens_seen": 1820459008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022650952858575726,
      "loss": 2.5789,
      "theoretical_loss": 3.457639840004262,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022649949849548647,
      "loss": 2.6551,
      "theoretical_loss": 3.4576292693689443,
      "tokens_seen": 1820590080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022648946840521565,
      "loss": 2.6061,
      "theoretical_loss": 3.4576186992206708,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022647943831494483,
      "loss": 2.7034,
      "theoretical_loss": 3.4576081295594023,
      "tokens_seen": 1820721152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000226469408224674,
      "loss": 2.5043,
      "theoretical_loss": 3.457597560385098,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002264593781344032,
      "loss": 2.6633,
      "theoretical_loss": 3.4575869916977195,
      "tokens_seen": 1820852224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002264493480441324,
      "loss": 2.5928,
      "theoretical_loss": 3.4575764234972253,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022643931795386158,
      "loss": 2.4599,
      "theoretical_loss": 3.4575658557835753,
      "tokens_seen": 1820983296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022642928786359076,
      "loss": 2.6914,
      "theoretical_loss": 3.457555288556731,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022641925777331994,
      "loss": 2.6913,
      "theoretical_loss": 3.4575447218166517,
      "tokens_seen": 1821114368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022640922768304915,
      "loss": 2.531,
      "theoretical_loss": 3.4575341555632972,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022639919759277833,
      "loss": 2.6888,
      "theoretical_loss": 3.4575235897966277,
      "tokens_seen": 1821245440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022638916750250751,
      "loss": 2.6177,
      "theoretical_loss": 3.457513024516604,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002263791374122367,
      "loss": 2.7555,
      "theoretical_loss": 3.457502459723185,
      "tokens_seen": 1821376512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022636910732196588,
      "loss": 2.4866,
      "theoretical_loss": 3.4574918954163314,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022635907723169508,
      "loss": 2.3084,
      "theoretical_loss": 3.4574813315960036,
      "tokens_seen": 1821507584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002263490471414243,
      "loss": 2.5862,
      "theoretical_loss": 3.4574707682621613,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022633901705115347,
      "loss": 2.493,
      "theoretical_loss": 3.4574602054147645,
      "tokens_seen": 1821638656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022632898696088266,
      "loss": 2.7388,
      "theoretical_loss": 3.4574496430537733,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022631895687061186,
      "loss": 2.6106,
      "theoretical_loss": 3.4574390811791478,
      "tokens_seen": 1821769728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022630892678034104,
      "loss": 2.4181,
      "theoretical_loss": 3.4574285197908483,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1003656,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7480885982513428,
      "objective/train/theoretical_loss": 3.4574179588888345,
      "objective/train/tokens_used": 1842360800,
      "theoretical_loss": 3.4574179588888345,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022629889669007023,
      "loss": 2.6481,
      "theoretical_loss": 3.4574179588888345,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262888665997994,
      "loss": 2.727,
      "theoretical_loss": 3.4574073984730678,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022627883650952861,
      "loss": 2.3667,
      "theoretical_loss": 3.4573968385435068,
      "tokens_seen": 1822031872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262688064192578,
      "loss": 2.671,
      "theoretical_loss": 3.4573862791001124,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022625877632898698,
      "loss": 2.6996,
      "theoretical_loss": 3.4573757201428443,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022624874623871616,
      "loss": 2.6263,
      "theoretical_loss": 3.457365161671663,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022623871614844534,
      "loss": 2.6819,
      "theoretical_loss": 3.457354603686529,
      "tokens_seen": 1822294016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022622868605817455,
      "loss": 2.592,
      "theoretical_loss": 3.457344046187401,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022621865596790373,
      "loss": 2.5007,
      "theoretical_loss": 3.457333489174241,
      "tokens_seen": 1822425088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262086258776329,
      "loss": 2.7401,
      "theoretical_loss": 3.4573229326470076,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261985957873621,
      "loss": 2.646,
      "theoretical_loss": 3.457312376605662,
      "tokens_seen": 1822556160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261885656970913,
      "loss": 2.4993,
      "theoretical_loss": 3.457301821050164,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022617853560682048,
      "loss": 2.6385,
      "theoretical_loss": 3.4572912659804738,
      "tokens_seen": 1822687232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022616850551654966,
      "loss": 2.6259,
      "theoretical_loss": 3.4572807113965514,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022615847542627884,
      "loss": 2.7484,
      "theoretical_loss": 3.4572701572983573,
      "tokens_seen": 1822818304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022614844533600802,
      "loss": 2.5092,
      "theoretical_loss": 3.457259603685851,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022613841524573723,
      "loss": 2.5469,
      "theoretical_loss": 3.4572490505589943,
      "tokens_seen": 1822949376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261283851554664,
      "loss": 2.56,
      "theoretical_loss": 3.4572384979177455,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261183550651956,
      "loss": 2.7715,
      "theoretical_loss": 3.4572279457620656,
      "tokens_seen": 1823080448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022610832497492478,
      "loss": 2.5443,
      "theoretical_loss": 3.457217394091915,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022609829488465398,
      "loss": 2.5745,
      "theoretical_loss": 3.4572068429072536,
      "tokens_seen": 1823211520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022608826479438316,
      "loss": 2.6075,
      "theoretical_loss": 3.457196292208042,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022607823470411235,
      "loss": 2.801,
      "theoretical_loss": 3.4571857419942402,
      "tokens_seen": 1823342592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022606820461384153,
      "loss": 2.7428,
      "theoretical_loss": 3.457175192265808,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002260581745235707,
      "loss": 2.4674,
      "theoretical_loss": 3.457164643022706,
      "tokens_seen": 1823473664
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1004868,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.848341226577759,
      "objective/train/theoretical_loss": 3.4571540942648946,
      "objective/train/tokens_used": 1843999200,
      "theoretical_loss": 3.4571540942648946,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022604814443329992,
      "loss": 2.4798,
      "theoretical_loss": 3.4571540942648946,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002260381143430291,
      "loss": 2.5033,
      "theoretical_loss": 3.4571435459923334,
      "tokens_seen": 1823604736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022602808425275828,
      "loss": 2.6432,
      "theoretical_loss": 3.457132998204984,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022601805416248746,
      "loss": 2.835,
      "theoretical_loss": 3.457122450902805,
      "tokens_seen": 1823735808
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022600802407221667,
      "loss": 2.7024,
      "theoretical_loss": 3.4571119040857576,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022599799398194585,
      "loss": 2.7052,
      "theoretical_loss": 3.4571013577538015,
      "tokens_seen": 1823866880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022598796389167503,
      "loss": 2.7647,
      "theoretical_loss": 3.457090811906898,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259779338014042,
      "loss": 2.6682,
      "theoretical_loss": 3.457080266545006,
      "tokens_seen": 1823997952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259679037111334,
      "loss": 2.6396,
      "theoretical_loss": 3.457069721668087,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259578736208626,
      "loss": 2.576,
      "theoretical_loss": 3.4570591772761006,
      "tokens_seen": 1824129024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022594784353059178,
      "loss": 2.4052,
      "theoretical_loss": 3.4570486333690065,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022593781344032096,
      "loss": 2.5424,
      "theoretical_loss": 3.4570380899467663,
      "tokens_seen": 1824260096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022592778335005014,
      "loss": 2.5984,
      "theoretical_loss": 3.4570275470093397,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022591775325977935,
      "loss": 2.6292,
      "theoretical_loss": 3.457017004556687,
      "tokens_seen": 1824391168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022590772316950853,
      "loss": 2.5371,
      "theoretical_loss": 3.457006462588768,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022589769307923771,
      "loss": 2.4147,
      "theoretical_loss": 3.456995921105544,
      "tokens_seen": 1824522240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002258876629889669,
      "loss": 2.562,
      "theoretical_loss": 3.4569853801069743,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022587763289869608,
      "loss": 2.6234,
      "theoretical_loss": 3.4569748395930198,
      "tokens_seen": 1824653312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022586760280842528,
      "loss": 2.3777,
      "theoretical_loss": 3.4569642995636407,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022585757271815447,
      "loss": 2.6203,
      "theoretical_loss": 3.4569537600187976,
      "tokens_seen": 1824784384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022584754262788365,
      "loss": 2.5604,
      "theoretical_loss": 3.4569432209584496,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022583751253761283,
      "loss": 2.4621,
      "theoretical_loss": 3.4569326823825586,
      "tokens_seen": 1824915456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022582748244734204,
      "loss": 2.9212,
      "theoretical_loss": 3.4569221442910845,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022581745235707122,
      "loss": 2.6143,
      "theoretical_loss": 3.4569116066839873,
      "tokens_seen": 1825046528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002258074222668004,
      "loss": 2.6309,
      "theoretical_loss": 3.4569010695612272,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1006237,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4566733837127686,
      "objective/train/theoretical_loss": 3.456890532922765,
      "objective/train/tokens_used": 1845637600,
      "theoretical_loss": 3.456890532922765,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022579739217652958,
      "loss": 2.7237,
      "theoretical_loss": 3.456890532922765,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022578736208625876,
      "loss": 2.7356,
      "theoretical_loss": 3.456879996768561,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022577733199598797,
      "loss": 2.5494,
      "theoretical_loss": 3.4568694610985755,
      "tokens_seen": 1825308672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022576730190571715,
      "loss": 2.5871,
      "theoretical_loss": 3.456858925912768,
      "tokens_seen": 1825374208
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022575727181544633,
      "loss": 2.5298,
      "theoretical_loss": 3.4568483912111008,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257472417251755,
      "loss": 2.6663,
      "theoretical_loss": 3.4568378569935323,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022573721163490472,
      "loss": 2.9006,
      "theoretical_loss": 3.4568273232600237,
      "tokens_seen": 1825570816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257271815446339,
      "loss": 2.6968,
      "theoretical_loss": 3.456816790010536,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022571715145436308,
      "loss": 2.6537,
      "theoretical_loss": 3.4568062572450287,
      "tokens_seen": 1825701888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022570712136409226,
      "loss": 2.65,
      "theoretical_loss": 3.456795724963462,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022569709127382147,
      "loss": 2.4755,
      "theoretical_loss": 3.456785193165797,
      "tokens_seen": 1825832960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022568706118355065,
      "loss": 2.538,
      "theoretical_loss": 3.456774661851995,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022567703109327983,
      "loss": 2.4161,
      "theoretical_loss": 3.4567641310220134,
      "tokens_seen": 1825964032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022566700100300902,
      "loss": 2.4368,
      "theoretical_loss": 3.4567536006758157,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256569709127382,
      "loss": 2.6606,
      "theoretical_loss": 3.456743070813361,
      "tokens_seen": 1826095104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256469408224674,
      "loss": 2.5828,
      "theoretical_loss": 3.4567325414346093,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022563691073219659,
      "loss": 2.5244,
      "theoretical_loss": 3.4567220125395215,
      "tokens_seen": 1826226176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022562688064192577,
      "loss": 2.7648,
      "theoretical_loss": 3.456711484128059,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022561685055165495,
      "loss": 2.7103,
      "theoretical_loss": 3.45670095620018,
      "tokens_seen": 1826357248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022560682046138416,
      "loss": 2.3799,
      "theoretical_loss": 3.456690428755847,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022559679037111336,
      "loss": 2.6946,
      "theoretical_loss": 3.4566799017950194,
      "tokens_seen": 1826488320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022558676028084255,
      "loss": 2.4341,
      "theoretical_loss": 3.4566693753176576,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022557673019057173,
      "loss": 2.544,
      "theoretical_loss": 3.4566588493237234,
      "tokens_seen": 1826619392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002255667001003009,
      "loss": 2.6778,
      "theoretical_loss": 3.4566483238131753,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022555667001003012,
      "loss": 2.5754,
      "theoretical_loss": 3.456637798785975,
      "tokens_seen": 1826750464
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1006887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.665243148803711,
      "objective/train/theoretical_loss": 3.4566272742420825,
      "objective/train/tokens_used": 1847276000,
      "theoretical_loss": 3.4566272742420825,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002255466399197593,
      "loss": 2.5324,
      "theoretical_loss": 3.4566272742420825,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022553660982948848,
      "loss": 2.3525,
      "theoretical_loss": 3.4566167501814586,
      "tokens_seen": 1826881536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022552657973921766,
      "loss": 2.7186,
      "theoretical_loss": 3.4566062266040634,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022551654964894687,
      "loss": 2.5781,
      "theoretical_loss": 3.4565957035098576,
      "tokens_seen": 1827012608
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022550651955867605,
      "loss": 2.5131,
      "theoretical_loss": 3.4565851808988013,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022549648946840523,
      "loss": 2.3874,
      "theoretical_loss": 3.4565746587708563,
      "tokens_seen": 1827143680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254864593781344,
      "loss": 2.5773,
      "theoretical_loss": 3.456564137125981,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254764292878636,
      "loss": 2.4866,
      "theoretical_loss": 3.456553615964138,
      "tokens_seen": 1827274752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254663991975928,
      "loss": 2.4307,
      "theoretical_loss": 3.4565430952852862,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022545636910732198,
      "loss": 2.6753,
      "theoretical_loss": 3.456532575089387,
      "tokens_seen": 1827405824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022544633901705116,
      "loss": 2.6805,
      "theoretical_loss": 3.4565220553764004,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022543630892678034,
      "loss": 2.6008,
      "theoretical_loss": 3.4565115361462873,
      "tokens_seen": 1827536896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022542627883650955,
      "loss": 2.6975,
      "theoretical_loss": 3.456501017399008,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022541624874623873,
      "loss": 2.4678,
      "theoretical_loss": 3.4564904991345236,
      "tokens_seen": 1827667968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022540621865596791,
      "loss": 2.4936,
      "theoretical_loss": 3.4564799813527936,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253961885656971,
      "loss": 2.366,
      "theoretical_loss": 3.4564694640537788,
      "tokens_seen": 1827799040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022538615847542628,
      "loss": 2.5182,
      "theoretical_loss": 3.4564589472374405,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022537612838515548,
      "loss": 2.4221,
      "theoretical_loss": 3.4564484309037384,
      "tokens_seen": 1827930112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022536609829488467,
      "loss": 2.5852,
      "theoretical_loss": 3.456437915052634,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022535606820461385,
      "loss": 2.4977,
      "theoretical_loss": 3.456427399684087,
      "tokens_seen": 1828061184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022534603811434303,
      "loss": 2.3335,
      "theoretical_loss": 3.456416884798058,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022533600802407224,
      "loss": 2.6614,
      "theoretical_loss": 3.456406370394508,
      "tokens_seen": 1828192256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022532597793380142,
      "loss": 2.746,
      "theoretical_loss": 3.456395856473397,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253159478435306,
      "loss": 2.653,
      "theoretical_loss": 3.456385343034686,
      "tokens_seen": 1828323328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022530591775325978,
      "loss": 2.3655,
      "theoretical_loss": 3.456374830078336,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1008261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5732953548431396,
      "objective/train/theoretical_loss": 3.456364317604306,
      "objective/train/tokens_used": 1848914400,
      "theoretical_loss": 3.456364317604306,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022529588766298896,
      "loss": 2.4943,
      "theoretical_loss": 3.456364317604306,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022528585757271817,
      "loss": 2.2379,
      "theoretical_loss": 3.456353805612559,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022527582748244735,
      "loss": 2.5574,
      "theoretical_loss": 3.4563432941030534,
      "tokens_seen": 1828585472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022526579739217653,
      "loss": 2.3713,
      "theoretical_loss": 3.4563327830757506,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252557673019057,
      "loss": 2.6448,
      "theoretical_loss": 3.4563222725306115,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022524573721163492,
      "loss": 2.6507,
      "theoretical_loss": 3.4563117624675965,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252357071213641,
      "loss": 2.6523,
      "theoretical_loss": 3.456301252886666,
      "tokens_seen": 1828847616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022522567703109328,
      "loss": 2.5443,
      "theoretical_loss": 3.456290743787781,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022521564694082246,
      "loss": 2.7512,
      "theoretical_loss": 3.456280235170902,
      "tokens_seen": 1828978688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022520561685055167,
      "loss": 2.6944,
      "theoretical_loss": 3.4562697270359886,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022519558676028085,
      "loss": 2.3703,
      "theoretical_loss": 3.456259219383003,
      "tokens_seen": 1829109760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022518555667001003,
      "loss": 2.5635,
      "theoretical_loss": 3.456248712211905,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022517552657973922,
      "loss": 2.256,
      "theoretical_loss": 3.4562382055226557,
      "tokens_seen": 1829240832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002251654964894684,
      "loss": 2.5049,
      "theoretical_loss": 3.456227699315215,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002251554663991976,
      "loss": 2.4974,
      "theoretical_loss": 3.456217193589544,
      "tokens_seen": 1829371904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022514543630892679,
      "loss": 2.5396,
      "theoretical_loss": 3.4562066883456035,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022513540621865597,
      "loss": 2.5019,
      "theoretical_loss": 3.456196183583354,
      "tokens_seen": 1829502976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022512537612838515,
      "loss": 2.5457,
      "theoretical_loss": 3.4561856793027568,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022511534603811436,
      "loss": 2.5458,
      "theoretical_loss": 3.4561751755037706,
      "tokens_seen": 1829634048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022510531594784354,
      "loss": 2.4303,
      "theoretical_loss": 3.4561646721863584,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022509528585757272,
      "loss": 2.4894,
      "theoretical_loss": 3.4561541693504796,
      "tokens_seen": 1829765120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002250852557673019,
      "loss": 2.861,
      "theoretical_loss": 3.4561436669960948,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022507522567703108,
      "loss": 2.2832,
      "theoretical_loss": 3.4561331651231653,
      "tokens_seen": 1829896192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002250651955867603,
      "loss": 2.6027,
      "theoretical_loss": 3.4561226637316516,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022505516549648947,
      "loss": 2.5105,
      "theoretical_loss": 3.456112162821514,
      "tokens_seen": 1830027264
    },
    {
      "epoch": 0.55,
      "objective/train/docs_used": 1008847,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.592195749282837,
      "objective/train/theoretical_loss": 3.456101662392714,
      "objective/train/tokens_used": 1850552800,
      "theoretical_loss": 3.456101662392714,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022504513540621865,
      "loss": 2.7343,
      "theoretical_loss": 3.456101662392714,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022503510531594783,
      "loss": 2.6602,
      "theoretical_loss": 3.4560911624452113,
      "tokens_seen": 1830158336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022502507522567704,
      "loss": 2.6773,
      "theoretical_loss": 3.456080662978967,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022501504513540622,
      "loss": 2.5267,
      "theoretical_loss": 3.4560701639939424,
      "tokens_seen": 1830289408
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002250050150451354,
      "loss": 2.6395,
      "theoretical_loss": 3.4560596654900975,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022499498495486458,
      "loss": 2.68,
      "theoretical_loss": 3.456049167467393,
      "tokens_seen": 1830420480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022498495486459377,
      "loss": 2.5645,
      "theoretical_loss": 3.45603866992579,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022497492477432297,
      "loss": 2.6441,
      "theoretical_loss": 3.456028172865249,
      "tokens_seen": 1830551552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022496489468405215,
      "loss": 2.362,
      "theoretical_loss": 3.4560176762857315,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022495486459378134,
      "loss": 2.4621,
      "theoretical_loss": 3.456007180187197,
      "tokens_seen": 1830682624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022494483450351052,
      "loss": 2.5536,
      "theoretical_loss": 3.455996684569607,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022493480441323973,
      "loss": 2.5603,
      "theoretical_loss": 3.4559861894329216,
      "tokens_seen": 1830813696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002249247743229689,
      "loss": 2.5924,
      "theoretical_loss": 3.4559756947771025,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002249147442326981,
      "loss": 2.6027,
      "theoretical_loss": 3.45596520060211,
      "tokens_seen": 1830944768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022490471414242727,
      "loss": 2.5458,
      "theoretical_loss": 3.4559547069079044,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022489468405215645,
      "loss": 2.4082,
      "theoretical_loss": 3.455944213694447,
      "tokens_seen": 1831075840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022488465396188566,
      "loss": 2.5023,
      "theoretical_loss": 3.4559337209616987,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022487462387161484,
      "loss": 2.6421,
      "theoretical_loss": 3.45592322870962,
      "tokens_seen": 1831206912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022486459378134402,
      "loss": 2.4118,
      "theoretical_loss": 3.4559127369381715,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002248545636910732,
      "loss": 2.6746,
      "theoretical_loss": 3.4559022456473145,
      "tokens_seen": 1831337984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022484453360080244,
      "loss": 2.5802,
      "theoretical_loss": 3.455891754837009,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022483450351053162,
      "loss": 2.7308,
      "theoretical_loss": 3.4558812645072168,
      "tokens_seen": 1831469056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002248244734202608,
      "loss": 2.707,
      "theoretical_loss": 3.455870774657898,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022481444332998998,
      "loss": 2.7263,
      "theoretical_loss": 3.4558602852890132,
      "tokens_seen": 1831600128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022480441323971916,
      "loss": 2.5049,
      "theoretical_loss": 3.4558497964005244,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1009995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2182343006134033,
      "objective/train/theoretical_loss": 3.455839307992391,
      "objective/train/tokens_used": 1852191200,
      "theoretical_loss": 3.455839307992391,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022479438314944837,
      "loss": 2.3477,
      "theoretical_loss": 3.455839307992391,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022478435305917755,
      "loss": 2.6354,
      "theoretical_loss": 3.455828820064575,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022477432296890673,
      "loss": 2.5823,
      "theoretical_loss": 3.455818332617036,
      "tokens_seen": 1831862272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002247642928786359,
      "loss": 2.3915,
      "theoretical_loss": 3.455807845649735,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022475426278836512,
      "loss": 2.6051,
      "theoretical_loss": 3.4557973591626343,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002247442326980943,
      "loss": 2.6223,
      "theoretical_loss": 3.4557868731556933,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022473420260782348,
      "loss": 2.3165,
      "theoretical_loss": 3.455776387628873,
      "tokens_seen": 1832124416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022472417251755266,
      "loss": 2.7262,
      "theoretical_loss": 3.4557659025821352,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022471414242728187,
      "loss": 2.2977,
      "theoretical_loss": 3.45575541801544,
      "tokens_seen": 1832255488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022470411233701105,
      "loss": 2.4973,
      "theoretical_loss": 3.455744933928748,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022469408224674023,
      "loss": 2.6063,
      "theoretical_loss": 3.45573445032202,
      "tokens_seen": 1832386560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022468405215646942,
      "loss": 2.4983,
      "theoretical_loss": 3.455723967195218,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002246740220661986,
      "loss": 2.5787,
      "theoretical_loss": 3.4557134845483013,
      "tokens_seen": 1832517632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002246639919759278,
      "loss": 2.3768,
      "theoretical_loss": 3.4557030023812323,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022465396188565699,
      "loss": 2.7814,
      "theoretical_loss": 3.4556925206939706,
      "tokens_seen": 1832648704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022464393179538617,
      "loss": 2.5129,
      "theoretical_loss": 3.455682039486478,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022463390170511535,
      "loss": 2.5087,
      "theoretical_loss": 3.455671558758715,
      "tokens_seen": 1832779776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022462387161484456,
      "loss": 2.6861,
      "theoretical_loss": 3.4556610785106425,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022461384152457374,
      "loss": 2.7084,
      "theoretical_loss": 3.455650598742221,
      "tokens_seen": 1832910848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022460381143430292,
      "loss": 2.5102,
      "theoretical_loss": 3.4556401194534123,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245937813440321,
      "loss": 2.386,
      "theoretical_loss": 3.4556296406441764,
      "tokens_seen": 1833041920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022458375125376128,
      "loss": 2.5273,
      "theoretical_loss": 3.4556191623144747,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245737211634905,
      "loss": 2.5334,
      "theoretical_loss": 3.455608684464268,
      "tokens_seen": 1833172992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022456369107321967,
      "loss": 2.4841,
      "theoretical_loss": 3.4555982070935176,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022455366098294885,
      "loss": 2.5046,
      "theoretical_loss": 3.4555877302021836,
      "tokens_seen": 1833304064
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1010488,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.001770257949829,
      "objective/train/theoretical_loss": 3.455577253790228,
      "objective/train/tokens_used": 1853829600,
      "theoretical_loss": 3.455577253790228,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022454363089267803,
      "loss": 2.6839,
      "theoretical_loss": 3.455577253790228,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022453360080240724,
      "loss": 2.4974,
      "theoretical_loss": 3.4555667778576105,
      "tokens_seen": 1833435136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022452357071213642,
      "loss": 2.7012,
      "theoretical_loss": 3.455556302404293,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245135406218656,
      "loss": 2.6997,
      "theoretical_loss": 3.4555458274302353,
      "tokens_seen": 1833566208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022450351053159478,
      "loss": 2.448,
      "theoretical_loss": 3.4555353529354003,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022449348044132397,
      "loss": 2.6957,
      "theoretical_loss": 3.4555248789197472,
      "tokens_seen": 1833697280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022448345035105317,
      "loss": 2.7248,
      "theoretical_loss": 3.455514405383237,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022447342026078236,
      "loss": 2.6065,
      "theoretical_loss": 3.455503932325832,
      "tokens_seen": 1833828352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022446339017051154,
      "loss": 2.5368,
      "theoretical_loss": 3.4554934597474922,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022445336008024072,
      "loss": 2.5952,
      "theoretical_loss": 3.4554829876481787,
      "tokens_seen": 1833959424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022444332998996993,
      "loss": 2.5953,
      "theoretical_loss": 3.4554725160278528,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002244332998996991,
      "loss": 2.4588,
      "theoretical_loss": 3.4554620448864743,
      "tokens_seen": 1834090496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002244232698094283,
      "loss": 2.6593,
      "theoretical_loss": 3.4554515742240057,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022441323971915747,
      "loss": 2.3706,
      "theoretical_loss": 3.4554411040404074,
      "tokens_seen": 1834221568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022440320962888665,
      "loss": 2.4826,
      "theoretical_loss": 3.45543063433564,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022439317953861586,
      "loss": 2.5804,
      "theoretical_loss": 3.4554201651096648,
      "tokens_seen": 1834352640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022438314944834504,
      "loss": 2.7282,
      "theoretical_loss": 3.455409696362443,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022437311935807422,
      "loss": 2.4834,
      "theoretical_loss": 3.4553992280939356,
      "tokens_seen": 1834483712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243630892678034,
      "loss": 2.5916,
      "theoretical_loss": 3.4553887603041034,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243530591775326,
      "loss": 2.5412,
      "theoretical_loss": 3.4553782929929073,
      "tokens_seen": 1834614784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243430290872618,
      "loss": 2.8994,
      "theoretical_loss": 3.4553678261603085,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022433299899699097,
      "loss": 2.7951,
      "theoretical_loss": 3.455357359806268,
      "tokens_seen": 1834745856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022432296890672015,
      "loss": 2.6912,
      "theoretical_loss": 3.455346893930747,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022431293881644933,
      "loss": 2.5882,
      "theoretical_loss": 3.4553364285337063,
      "tokens_seen": 1834876928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022430290872617854,
      "loss": 2.3475,
      "theoretical_loss": 3.4553259636151066,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1011110,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9770421981811523,
      "objective/train/theoretical_loss": 3.45531549917491,
      "objective/train/tokens_used": 1855468000,
      "theoretical_loss": 3.45531549917491,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022429287863590772,
      "loss": 2.7367,
      "theoretical_loss": 3.45531549917491,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242828485456369,
      "loss": 2.6557,
      "theoretical_loss": 3.455305035213076,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022427281845536609,
      "loss": 2.6387,
      "theoretical_loss": 3.455294571729567,
      "tokens_seen": 1835139072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242627883650953,
      "loss": 2.6351,
      "theoretical_loss": 3.455284108724344,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022425275827482448,
      "loss": 2.6886,
      "theoretical_loss": 3.4552736461973668,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022424272818455366,
      "loss": 2.7588,
      "theoretical_loss": 3.455263184148598,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022423269809428284,
      "loss": 2.6698,
      "theoretical_loss": 3.455252722577997,
      "tokens_seen": 1835401216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022422266800401202,
      "loss": 2.5098,
      "theoretical_loss": 3.455242261485527,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022421263791374123,
      "loss": 2.6698,
      "theoretical_loss": 3.4552318008711476,
      "tokens_seen": 1835532288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242026078234704,
      "loss": 2.7006,
      "theoretical_loss": 3.45522134073482,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241925777331996,
      "loss": 2.5054,
      "theoretical_loss": 3.455210881076505,
      "tokens_seen": 1835663360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022418254764292877,
      "loss": 2.5289,
      "theoretical_loss": 3.455200421896165,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022417251755265798,
      "loss": 2.6255,
      "theoretical_loss": 3.45518996319376,
      "tokens_seen": 1835794432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022416248746238716,
      "loss": 2.6999,
      "theoretical_loss": 3.4551795049692515,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022415245737211634,
      "loss": 2.8628,
      "theoretical_loss": 3.4551690472226007,
      "tokens_seen": 1835925504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022414242728184552,
      "loss": 2.6599,
      "theoretical_loss": 3.455158589953768,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241323971915747,
      "loss": 2.5478,
      "theoretical_loss": 3.455148133162715,
      "tokens_seen": 1836056576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241223671013039,
      "loss": 2.5116,
      "theoretical_loss": 3.455137676849403,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241123370110331,
      "loss": 2.6363,
      "theoretical_loss": 3.455127221013793,
      "tokens_seen": 1836187648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022410230692076227,
      "loss": 2.5545,
      "theoretical_loss": 3.4551167656558457,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022409227683049148,
      "loss": 2.2969,
      "theoretical_loss": 3.4551063107755233,
      "tokens_seen": 1836318720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002240822467402207,
      "loss": 2.6308,
      "theoretical_loss": 3.4550958563727856,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022407221664994987,
      "loss": 2.4542,
      "theoretical_loss": 3.4550854024475948,
      "tokens_seen": 1836449792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022406218655967905,
      "loss": 2.6274,
      "theoretical_loss": 3.4550749489999113,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022405215646940823,
      "loss": 2.622,
      "theoretical_loss": 3.4550644960296966,
      "tokens_seen": 1836580864
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1012365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7557899951934814,
      "objective/train/theoretical_loss": 3.455054043536912,
      "objective/train/tokens_used": 1857106400,
      "theoretical_loss": 3.455054043536912,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022404212637913744,
      "loss": 2.7961,
      "theoretical_loss": 3.455054043536912,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022403209628886662,
      "loss": 2.587,
      "theoretical_loss": 3.455043591521518,
      "tokens_seen": 1836711936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002240220661985958,
      "loss": 2.3494,
      "theoretical_loss": 3.4550331399834766,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022401203610832498,
      "loss": 2.7496,
      "theoretical_loss": 3.4550226889227487,
      "tokens_seen": 1836843008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022400200601805417,
      "loss": 2.666,
      "theoretical_loss": 3.455012238339296,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022399197592778337,
      "loss": 2.5364,
      "theoretical_loss": 3.455001788233078,
      "tokens_seen": 1836974080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022398194583751256,
      "loss": 2.6442,
      "theoretical_loss": 3.4549913386040574,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022397191574724174,
      "loss": 2.7147,
      "theoretical_loss": 3.454980889452195,
      "tokens_seen": 1837105152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022396188565697092,
      "loss": 2.77,
      "theoretical_loss": 3.4549704407774517,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022395185556670013,
      "loss": 2.812,
      "theoretical_loss": 3.454959992579789,
      "tokens_seen": 1837236224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002239418254764293,
      "loss": 2.4815,
      "theoretical_loss": 3.4549495448591685,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002239317953861585,
      "loss": 2.4991,
      "theoretical_loss": 3.4549390976155503,
      "tokens_seen": 1837367296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022392176529588767,
      "loss": 2.665,
      "theoretical_loss": 3.4549286508488963,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022391173520561685,
      "loss": 2.5714,
      "theoretical_loss": 3.454918204559168,
      "tokens_seen": 1837498368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022390170511534606,
      "loss": 2.7933,
      "theoretical_loss": 3.4549077587463257,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022389167502507524,
      "loss": 2.4321,
      "theoretical_loss": 3.4548973134103322,
      "tokens_seen": 1837629440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022388164493480442,
      "loss": 2.5651,
      "theoretical_loss": 3.4548868685511467,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002238716148445336,
      "loss": 2.5793,
      "theoretical_loss": 3.4548764241687318,
      "tokens_seen": 1837760512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002238615847542628,
      "loss": 2.5296,
      "theoretical_loss": 3.4548659802630484,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000223851554663992,
      "loss": 2.5009,
      "theoretical_loss": 3.454855536834058,
      "tokens_seen": 1837891584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022384152457372117,
      "loss": 2.7266,
      "theoretical_loss": 3.4548450938817212,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022383149448345035,
      "loss": 2.5885,
      "theoretical_loss": 3.454834651406,
      "tokens_seen": 1838022656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022382146439317953,
      "loss": 2.5711,
      "theoretical_loss": 3.454824209406855,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022381143430290874,
      "loss": 2.5871,
      "theoretical_loss": 3.4548137678842474,
      "tokens_seen": 1838153728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022380140421263792,
      "loss": 2.5451,
      "theoretical_loss": 3.4548033268381397,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1012997,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.479828357696533,
      "objective/train/theoretical_loss": 3.454792886268492,
      "objective/train/tokens_used": 1858744800,
      "theoretical_loss": 3.454792886268492,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237913741223671,
      "loss": 2.7901,
      "theoretical_loss": 3.454792886268492,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022378134403209629,
      "loss": 2.5571,
      "theoretical_loss": 3.4547824461752654,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237713139418255,
      "loss": 2.6999,
      "theoretical_loss": 3.4547720065584215,
      "tokens_seen": 1838415872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022376128385155468,
      "loss": 2.6973,
      "theoretical_loss": 3.454761567417922,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022375125376128386,
      "loss": 2.8487,
      "theoretical_loss": 3.454751128753728,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022374122367101304,
      "loss": 2.7893,
      "theoretical_loss": 3.4547406905658002,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022373119358074222,
      "loss": 2.6675,
      "theoretical_loss": 3.454730252854101,
      "tokens_seen": 1838678016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022372116349047143,
      "loss": 2.5131,
      "theoretical_loss": 3.454719815618591,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237111334002006,
      "loss": 2.5407,
      "theoretical_loss": 3.4547093788592314,
      "tokens_seen": 1838809088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237011033099298,
      "loss": 2.8539,
      "theoretical_loss": 3.4546989425759835,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022369107321965897,
      "loss": 2.6911,
      "theoretical_loss": 3.454688506768809,
      "tokens_seen": 1838940160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022368104312938818,
      "loss": 2.6549,
      "theoretical_loss": 3.4546780714376695,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022367101303911736,
      "loss": 2.6705,
      "theoretical_loss": 3.454667636582525,
      "tokens_seen": 1839071232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022366098294884654,
      "loss": 2.6543,
      "theoretical_loss": 3.4546572022033377,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022365095285857572,
      "loss": 2.7402,
      "theoretical_loss": 3.4546467683000692,
      "tokens_seen": 1839202304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236409227683049,
      "loss": 2.4261,
      "theoretical_loss": 3.4546363348726805,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236308926780341,
      "loss": 2.5562,
      "theoretical_loss": 3.4546259019211334,
      "tokens_seen": 1839333376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236208625877633,
      "loss": 2.5367,
      "theoretical_loss": 3.4546154694453883,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022361083249749247,
      "loss": 2.5013,
      "theoretical_loss": 3.454605037445407,
      "tokens_seen": 1839464448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022360080240722165,
      "loss": 2.736,
      "theoretical_loss": 3.454594605921151,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022359077231695086,
      "loss": 2.4765,
      "theoretical_loss": 3.454584174872582,
      "tokens_seen": 1839595520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022358074222668004,
      "loss": 2.7233,
      "theoretical_loss": 3.454573744299661,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022357071213640923,
      "loss": 2.5706,
      "theoretical_loss": 3.454563314202349,
      "tokens_seen": 1839726592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002235606820461384,
      "loss": 2.5897,
      "theoretical_loss": 3.454552884580607,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022355065195586761,
      "loss": 2.6828,
      "theoretical_loss": 3.4545424554343978,
      "tokens_seen": 1839857664
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1014020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.84538197517395,
      "objective/train/theoretical_loss": 3.454532026763682,
      "objective/train/tokens_used": 1860383200,
      "theoretical_loss": 3.454532026763682,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002235406218655968,
      "loss": 2.7351,
      "theoretical_loss": 3.454532026763682,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022353059177532598,
      "loss": 2.5926,
      "theoretical_loss": 3.4545215985684212,
      "tokens_seen": 1839988736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022352056168505516,
      "loss": 2.4106,
      "theoretical_loss": 3.4545111708485763,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022351053159478434,
      "loss": 2.5012,
      "theoretical_loss": 3.454500743604109,
      "tokens_seen": 1840119808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022350050150451355,
      "loss": 2.5762,
      "theoretical_loss": 3.4544903168349808,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022349047141424273,
      "loss": 2.3674,
      "theoretical_loss": 3.454479890541153,
      "tokens_seen": 1840250880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002234804413239719,
      "loss": 2.7638,
      "theoretical_loss": 3.454469464722587,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002234704112337011,
      "loss": 2.637,
      "theoretical_loss": 3.4544590393792443,
      "tokens_seen": 1840381952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002234603811434303,
      "loss": 2.7113,
      "theoretical_loss": 3.4544486145110858,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022345035105315948,
      "loss": 2.3803,
      "theoretical_loss": 3.454438190118074,
      "tokens_seen": 1840513024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022344032096288866,
      "loss": 2.5349,
      "theoretical_loss": 3.4544277662001694,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022343029087261784,
      "loss": 2.5746,
      "theoretical_loss": 3.454417342757334,
      "tokens_seen": 1840644096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022342026078234702,
      "loss": 2.6191,
      "theoretical_loss": 3.4544069197895286,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022341023069207623,
      "loss": 2.6675,
      "theoretical_loss": 3.454396497296715,
      "tokens_seen": 1840775168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002234002006018054,
      "loss": 2.9166,
      "theoretical_loss": 3.454386075278855,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002233901705115346,
      "loss": 2.6774,
      "theoretical_loss": 3.4543756537359096,
      "tokens_seen": 1840906240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022338014042126377,
      "loss": 2.6759,
      "theoretical_loss": 3.4543652326678402,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022337011033099298,
      "loss": 2.7668,
      "theoretical_loss": 3.454354812074609,
      "tokens_seen": 1841037312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022336008024072216,
      "loss": 2.3496,
      "theoretical_loss": 3.454344391956176,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022335005015045135,
      "loss": 2.7011,
      "theoretical_loss": 3.4543339723125035,
      "tokens_seen": 1841168384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022334002006018055,
      "loss": 2.5368,
      "theoretical_loss": 3.4543235531435537,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022332998996990973,
      "loss": 2.6115,
      "theoretical_loss": 3.4543131344492872,
      "tokens_seen": 1841299456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022331995987963894,
      "loss": 2.5351,
      "theoretical_loss": 3.454302716229665,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022330992978936812,
      "loss": 2.4414,
      "theoretical_loss": 3.45429229848465,
      "tokens_seen": 1841430528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002232998996990973,
      "loss": 2.6435,
      "theoretical_loss": 3.454281881214203,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1014627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6324822902679443,
      "objective/train/theoretical_loss": 3.454271464418285,
      "objective/train/tokens_used": 1862021600,
      "theoretical_loss": 3.454271464418285,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022328986960882649,
      "loss": 2.6505,
      "theoretical_loss": 3.454271464418285,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002232798395185557,
      "loss": 2.7179,
      "theoretical_loss": 3.454261048096858,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022326980942828488,
      "loss": 2.3715,
      "theoretical_loss": 3.4542506322498836,
      "tokens_seen": 1841692672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022325977933801406,
      "loss": 2.6809,
      "theoretical_loss": 3.454240216877323,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022324974924774324,
      "loss": 2.776,
      "theoretical_loss": 3.454229801979138,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022323971915747242,
      "loss": 2.4634,
      "theoretical_loss": 3.45421938755529,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022322968906720163,
      "loss": 2.5873,
      "theoretical_loss": 3.4542089736057404,
      "tokens_seen": 1841954816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002232196589769308,
      "loss": 2.5552,
      "theoretical_loss": 3.4541985601304503,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022320962888666,
      "loss": 2.5137,
      "theoretical_loss": 3.4541881471293827,
      "tokens_seen": 1842085888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022319959879638917,
      "loss": 2.3721,
      "theoretical_loss": 3.4541777346024976,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022318956870611838,
      "loss": 2.6053,
      "theoretical_loss": 3.454167322549757,
      "tokens_seen": 1842216960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022317953861584756,
      "loss": 2.6496,
      "theoretical_loss": 3.4541569109711228,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022316950852557674,
      "loss": 2.6895,
      "theoretical_loss": 3.454146499866556,
      "tokens_seen": 1842348032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022315947843530592,
      "loss": 2.4441,
      "theoretical_loss": 3.4541360892360187,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231494483450351,
      "loss": 2.7234,
      "theoretical_loss": 3.4541256790794725,
      "tokens_seen": 1842479104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231394182547643,
      "loss": 2.626,
      "theoretical_loss": 3.4541152693968784,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231293881644935,
      "loss": 2.636,
      "theoretical_loss": 3.454104860188198,
      "tokens_seen": 1842610176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022311935807422267,
      "loss": 2.398,
      "theoretical_loss": 3.4540944514533933,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022310932798395185,
      "loss": 2.5063,
      "theoretical_loss": 3.454084043192426,
      "tokens_seen": 1842741248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022309929789368106,
      "loss": 2.5572,
      "theoretical_loss": 3.4540736354052575,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022308926780341024,
      "loss": 2.5335,
      "theoretical_loss": 3.454063228091848,
      "tokens_seen": 1842872320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022307923771313943,
      "loss": 2.5483,
      "theoretical_loss": 3.4540528212521617,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002230692076228686,
      "loss": 2.8843,
      "theoretical_loss": 3.4540424148861586,
      "tokens_seen": 1843003392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022305917753259781,
      "loss": 2.6199,
      "theoretical_loss": 3.4540320089938,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000223049147442327,
      "loss": 2.6044,
      "theoretical_loss": 3.454021603575048,
      "tokens_seen": 1843134464
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1015947,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9211390018463135,
      "objective/train/theoretical_loss": 3.454011198629865,
      "objective/train/tokens_used": 1863660000,
      "theoretical_loss": 3.454011198629865,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022303911735205618,
      "loss": 2.5535,
      "theoretical_loss": 3.454011198629865,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022302908726178536,
      "loss": 2.7455,
      "theoretical_loss": 3.454000794158211,
      "tokens_seen": 1843265536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022301905717151454,
      "loss": 2.5731,
      "theoretical_loss": 3.453990390160049,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022300902708124375,
      "loss": 2.4396,
      "theoretical_loss": 3.4539799866353396,
      "tokens_seen": 1843396608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022299899699097293,
      "loss": 2.6189,
      "theoretical_loss": 3.453969583584045,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229889669007021,
      "loss": 2.4353,
      "theoretical_loss": 3.4539591810061268,
      "tokens_seen": 1843527680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229789368104313,
      "loss": 2.7168,
      "theoretical_loss": 3.453948778901547,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229689067201605,
      "loss": 2.6194,
      "theoretical_loss": 3.4539383772702665,
      "tokens_seen": 1843658752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022295887662988968,
      "loss": 2.5682,
      "theoretical_loss": 3.4539279761122472,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022294884653961886,
      "loss": 2.7671,
      "theoretical_loss": 3.4539175754274507,
      "tokens_seen": 1843789824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022293881644934804,
      "loss": 2.476,
      "theoretical_loss": 3.453907175215839,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022292878635907722,
      "loss": 2.5197,
      "theoretical_loss": 3.4538967754773733,
      "tokens_seen": 1843920896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022291875626880643,
      "loss": 2.5077,
      "theoretical_loss": 3.453886376212015,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002229087261785356,
      "loss": 2.3249,
      "theoretical_loss": 3.4538759774197265,
      "tokens_seen": 1844051968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002228986960882648,
      "loss": 2.5589,
      "theoretical_loss": 3.4538655791004693,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022288866599799397,
      "loss": 2.6163,
      "theoretical_loss": 3.4538551812542053,
      "tokens_seen": 1844183040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022287863590772318,
      "loss": 2.5171,
      "theoretical_loss": 3.4538447838808954,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022286860581745236,
      "loss": 2.5118,
      "theoretical_loss": 3.453834386980502,
      "tokens_seen": 1844314112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022285857572718155,
      "loss": 2.5847,
      "theoretical_loss": 3.453823990552986,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022284854563691073,
      "loss": 2.4455,
      "theoretical_loss": 3.4538135945983104,
      "tokens_seen": 1844445184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002228385155466399,
      "loss": 2.4867,
      "theoretical_loss": 3.453803199116435,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022282848545636912,
      "loss": 2.5047,
      "theoretical_loss": 3.453792804107324,
      "tokens_seen": 1844576256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002228184553660983,
      "loss": 2.4894,
      "theoretical_loss": 3.4537824095709366,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022280842527582748,
      "loss": 2.8172,
      "theoretical_loss": 3.4537720155072362,
      "tokens_seen": 1844707328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022279839518555666,
      "loss": 2.6976,
      "theoretical_loss": 3.4537616219161835,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1016542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6252291202545166,
      "objective/train/theoretical_loss": 3.4537512287977403,
      "objective/train/tokens_used": 1865298400,
      "theoretical_loss": 3.4537512287977403,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022278836509528587,
      "loss": 2.5652,
      "theoretical_loss": 3.4537512287977403,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022277833500501505,
      "loss": 2.2833,
      "theoretical_loss": 3.4537408361518693,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022276830491474423,
      "loss": 2.8261,
      "theoretical_loss": 3.4537304439785315,
      "tokens_seen": 1844969472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227582748244734,
      "loss": 2.4424,
      "theoretical_loss": 3.4537200522776885,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227482447342026,
      "loss": 2.4669,
      "theoretical_loss": 3.4537096610493023,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227382146439318,
      "loss": 2.4553,
      "theoretical_loss": 3.4536992702933347,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022272818455366098,
      "loss": 2.3953,
      "theoretical_loss": 3.453688880009747,
      "tokens_seen": 1845231616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022271815446339016,
      "loss": 2.6769,
      "theoretical_loss": 3.4536784901985014,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022270812437311934,
      "loss": 2.8062,
      "theoretical_loss": 3.4536681008595593,
      "tokens_seen": 1845362688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022269809428284855,
      "loss": 2.6986,
      "theoretical_loss": 3.4536577119928835,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022268806419257773,
      "loss": 2.5909,
      "theoretical_loss": 3.4536473235984344,
      "tokens_seen": 1845493760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022267803410230691,
      "loss": 2.7481,
      "theoretical_loss": 3.4536369356761742,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226680040120361,
      "loss": 2.5052,
      "theoretical_loss": 3.4536265482260653,
      "tokens_seen": 1845624832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022265797392176528,
      "loss": 2.7138,
      "theoretical_loss": 3.4536161612480685,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022264794383149448,
      "loss": 2.6972,
      "theoretical_loss": 3.453605774742146,
      "tokens_seen": 1845755904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022263791374122367,
      "loss": 2.6677,
      "theoretical_loss": 3.4535953887082598,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022262788365095285,
      "loss": 2.5929,
      "theoretical_loss": 3.4535850031463715,
      "tokens_seen": 1845886976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022261785356068203,
      "loss": 2.4314,
      "theoretical_loss": 3.453574618056443,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022260782347041124,
      "loss": 2.4304,
      "theoretical_loss": 3.4535642334384358,
      "tokens_seen": 1846018048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022259779338014042,
      "loss": 2.7496,
      "theoretical_loss": 3.4535538492923123,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022258776328986963,
      "loss": 2.413,
      "theoretical_loss": 3.4535434656180337,
      "tokens_seen": 1846149120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225777331995988,
      "loss": 2.4911,
      "theoretical_loss": 3.4535330824155617,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022256770310932801,
      "loss": 2.4833,
      "theoretical_loss": 3.453522699684859,
      "tokens_seen": 1846280192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225576730190572,
      "loss": 2.6135,
      "theoretical_loss": 3.4535123174258864,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022254764292878638,
      "loss": 2.4844,
      "theoretical_loss": 3.4535019356386067,
      "tokens_seen": 1846411264
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1017618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4465034008026123,
      "objective/train/theoretical_loss": 3.4534915543229805,
      "objective/train/tokens_used": 1866936800,
      "theoretical_loss": 3.4534915543229805,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022253761283851556,
      "loss": 2.5404,
      "theoretical_loss": 3.4534915543229805,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022252758274824474,
      "loss": 2.5781,
      "theoretical_loss": 3.4534811734789708,
      "tokens_seen": 1846542336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022251755265797395,
      "loss": 2.7321,
      "theoretical_loss": 3.453470793106539,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022250752256770313,
      "loss": 2.5843,
      "theoretical_loss": 3.453460413205647,
      "tokens_seen": 1846673408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224974924774323,
      "loss": 2.6055,
      "theoretical_loss": 3.453450033776256,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224874623871615,
      "loss": 2.4297,
      "theoretical_loss": 3.453439654818329,
      "tokens_seen": 1846804480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224774322968907,
      "loss": 2.4981,
      "theoretical_loss": 3.4534292763318275,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022246740220661988,
      "loss": 2.6859,
      "theoretical_loss": 3.4534188983167127,
      "tokens_seen": 1846935552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022245737211634906,
      "loss": 2.5984,
      "theoretical_loss": 3.453408520772947,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022244734202607824,
      "loss": 2.6482,
      "theoretical_loss": 3.453398143700492,
      "tokens_seen": 1847066624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022243731193580742,
      "loss": 2.3664,
      "theoretical_loss": 3.4533877670993096,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022242728184553663,
      "loss": 2.5877,
      "theoretical_loss": 3.4533773909693624,
      "tokens_seen": 1847197696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224172517552658,
      "loss": 2.3631,
      "theoretical_loss": 3.4533670153106115,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222407221664995,
      "loss": 2.6233,
      "theoretical_loss": 3.453356640123019,
      "tokens_seen": 1847328768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022239719157472418,
      "loss": 2.565,
      "theoretical_loss": 3.4533462654065463,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022238716148445338,
      "loss": 2.582,
      "theoretical_loss": 3.4533358911611565,
      "tokens_seen": 1847459840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022237713139418256,
      "loss": 2.5824,
      "theoretical_loss": 3.4533255173868103,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022236710130391175,
      "loss": 2.505,
      "theoretical_loss": 3.4533151440834704,
      "tokens_seen": 1847590912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022235707121364093,
      "loss": 2.6098,
      "theoretical_loss": 3.4533047712510982,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223470411233701,
      "loss": 2.5263,
      "theoretical_loss": 3.453294398889656,
      "tokens_seen": 1847721984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022233701103309932,
      "loss": 2.5628,
      "theoretical_loss": 3.453284026999105,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223269809428285,
      "loss": 2.6991,
      "theoretical_loss": 3.4532736555794084,
      "tokens_seen": 1847853056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022231695085255768,
      "loss": 2.603,
      "theoretical_loss": 3.453263284630527,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022230692076228686,
      "loss": 2.6746,
      "theoretical_loss": 3.453252914152423,
      "tokens_seen": 1847984128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022229689067201607,
      "loss": 2.4067,
      "theoretical_loss": 3.4532425441450587,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1018803,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6557962894439697,
      "objective/train/theoretical_loss": 3.4532321746083956,
      "objective/train/tokens_used": 1868575200,
      "theoretical_loss": 3.4532321746083956,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022228686058174525,
      "loss": 2.7492,
      "theoretical_loss": 3.4532321746083956,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022227683049147443,
      "loss": 2.4828,
      "theoretical_loss": 3.4532218055423956,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002222668004012036,
      "loss": 2.241,
      "theoretical_loss": 3.4532114369470213,
      "tokens_seen": 1848246272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002222567703109328,
      "loss": 2.7674,
      "theoretical_loss": 3.453201068822234,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222246740220662,
      "loss": 2.5812,
      "theoretical_loss": 3.453190701167996,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022223671013039118,
      "loss": 2.4916,
      "theoretical_loss": 3.4531803339842693,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022222668004012036,
      "loss": 2.7954,
      "theoretical_loss": 3.453169967271015,
      "tokens_seen": 1848508416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022221664994984954,
      "loss": 2.4893,
      "theoretical_loss": 3.4531596010281964,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022220661985957875,
      "loss": 2.5298,
      "theoretical_loss": 3.4531492352557747,
      "tokens_seen": 1848639488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022219658976930793,
      "loss": 2.3545,
      "theoretical_loss": 3.453138869953712,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022218655967903711,
      "loss": 2.617,
      "theoretical_loss": 3.4531285051219704,
      "tokens_seen": 1848770560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002221765295887663,
      "loss": 2.5443,
      "theoretical_loss": 3.4531181407605116,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022216649949849548,
      "loss": 2.5235,
      "theoretical_loss": 3.453107776869298,
      "tokens_seen": 1848901632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022215646940822468,
      "loss": 2.663,
      "theoretical_loss": 3.4530974134482917,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022214643931795387,
      "loss": 2.5753,
      "theoretical_loss": 3.4530870504974542,
      "tokens_seen": 1849032704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022213640922768305,
      "loss": 2.4768,
      "theoretical_loss": 3.4530766880167474,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022212637913741223,
      "loss": 2.6708,
      "theoretical_loss": 3.453066326006134,
      "tokens_seen": 1849163776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022211634904714144,
      "loss": 2.6227,
      "theoretical_loss": 3.4530559644655754,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022210631895687062,
      "loss": 2.2136,
      "theoretical_loss": 3.453045603395034,
      "tokens_seen": 1849294848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002220962888665998,
      "loss": 2.5871,
      "theoretical_loss": 3.4530352427944715,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022208625877632898,
      "loss": 2.6596,
      "theoretical_loss": 3.45302488266385,
      "tokens_seen": 1849425920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022207622868605816,
      "loss": 2.4942,
      "theoretical_loss": 3.453014523003132,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022206619859578737,
      "loss": 2.8071,
      "theoretical_loss": 3.453004163812279,
      "tokens_seen": 1849556992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022205616850551655,
      "loss": 2.4112,
      "theoretical_loss": 3.452993805091253,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022204613841524573,
      "loss": 2.4684,
      "theoretical_loss": 3.4529834468400162,
      "tokens_seen": 1849688064
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1019318,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.330780267715454,
      "objective/train/theoretical_loss": 3.4529730890585313,
      "objective/train/tokens_used": 1870213600,
      "theoretical_loss": 3.4529730890585313,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002220361083249749,
      "loss": 2.5972,
      "theoretical_loss": 3.4529730890585313,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022202607823470412,
      "loss": 2.3719,
      "theoretical_loss": 3.4529627317467595,
      "tokens_seen": 1849819136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002220160481444333,
      "loss": 2.6447,
      "theoretical_loss": 3.4529523749046627,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022200601805416248,
      "loss": 2.6251,
      "theoretical_loss": 3.4529420185322035,
      "tokens_seen": 1849950208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022199598796389166,
      "loss": 2.6382,
      "theoretical_loss": 3.4529316626293443,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022198595787362084,
      "loss": 2.6216,
      "theoretical_loss": 3.4529213071960463,
      "tokens_seen": 1850081280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022197592778335005,
      "loss": 2.5811,
      "theoretical_loss": 3.4529109522322727,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022196589769307923,
      "loss": 2.7008,
      "theoretical_loss": 3.452900597737984,
      "tokens_seen": 1850212352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022195586760280842,
      "loss": 2.699,
      "theoretical_loss": 3.452890243713144,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002219458375125376,
      "loss": 2.4472,
      "theoretical_loss": 3.452879890157713,
      "tokens_seen": 1850343424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002219358074222668,
      "loss": 2.6153,
      "theoretical_loss": 3.452869537071655,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022192577733199599,
      "loss": 2.5,
      "theoretical_loss": 3.4528591844549306,
      "tokens_seen": 1850474496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022191574724172517,
      "loss": 2.5629,
      "theoretical_loss": 3.4528488323075024,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022190571715145435,
      "loss": 2.6464,
      "theoretical_loss": 3.4528384806293326,
      "tokens_seen": 1850605568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022189568706118356,
      "loss": 2.6449,
      "theoretical_loss": 3.452828129420384,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022188565697091274,
      "loss": 2.6039,
      "theoretical_loss": 3.452817778680617,
      "tokens_seen": 1850736640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022187562688064192,
      "loss": 2.541,
      "theoretical_loss": 3.4528074284099954,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002218655967903711,
      "loss": 2.553,
      "theoretical_loss": 3.45279707860848,
      "tokens_seen": 1850867712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022185556670010028,
      "loss": 2.7543,
      "theoretical_loss": 3.452786729276034,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022184553660982952,
      "loss": 2.71,
      "theoretical_loss": 3.452776380412619,
      "tokens_seen": 1850998784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002218355065195587,
      "loss": 2.7985,
      "theoretical_loss": 3.4527660320181974,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022182547642928788,
      "loss": 2.6417,
      "theoretical_loss": 3.4527556840927307,
      "tokens_seen": 1851129856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022181544633901706,
      "loss": 2.5014,
      "theoretical_loss": 3.452745336636182,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022180541624874627,
      "loss": 2.7388,
      "theoretical_loss": 3.452734989648513,
      "tokens_seen": 1851260928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022179538615847545,
      "loss": 2.4891,
      "theoretical_loss": 3.4527246431296854,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1019593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7899038791656494,
      "objective/train/theoretical_loss": 3.4527142970796625,
      "objective/train/tokens_used": 1871852000,
      "theoretical_loss": 3.4527142970796625,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022178535606820463,
      "loss": 2.5942,
      "theoretical_loss": 3.4527142970796625,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002217753259779338,
      "loss": 2.4552,
      "theoretical_loss": 3.452703951498405,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221765295887663,
      "loss": 2.7336,
      "theoretical_loss": 3.452693606385876,
      "tokens_seen": 1851523072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002217552657973922,
      "loss": 2.7857,
      "theoretical_loss": 3.452683261742038,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022174523570712138,
      "loss": 2.4826,
      "theoretical_loss": 3.4526729175668525,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022173520561685056,
      "loss": 2.5551,
      "theoretical_loss": 3.452662573860281,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022172517552657974,
      "loss": 2.5641,
      "theoretical_loss": 3.452652230622288,
      "tokens_seen": 1851785216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022171514543630895,
      "loss": 2.5581,
      "theoretical_loss": 3.4526418878528333,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022170511534603813,
      "loss": 2.5615,
      "theoretical_loss": 3.4526315455518803,
      "tokens_seen": 1851916288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022169508525576731,
      "loss": 2.6854,
      "theoretical_loss": 3.452621203719391,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002216850551654965,
      "loss": 2.4393,
      "theoretical_loss": 3.452610862355327,
      "tokens_seen": 1852047360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022167502507522568,
      "loss": 2.9684,
      "theoretical_loss": 3.452600521459652,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022166499498495488,
      "loss": 2.9095,
      "theoretical_loss": 3.4525901810323263,
      "tokens_seen": 1852178432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022165496489468407,
      "loss": 2.5196,
      "theoretical_loss": 3.4525798410733133,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022164493480441325,
      "loss": 2.4828,
      "theoretical_loss": 3.4525695015825755,
      "tokens_seen": 1852309504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022163490471414243,
      "loss": 2.7161,
      "theoretical_loss": 3.4525591625600742,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022162487462387164,
      "loss": 2.6665,
      "theoretical_loss": 3.4525488240057722,
      "tokens_seen": 1852440576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022161484453360082,
      "loss": 2.6879,
      "theoretical_loss": 3.4525384859196313,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022160481444333,
      "loss": 2.8713,
      "theoretical_loss": 3.4525281483016146,
      "tokens_seen": 1852571648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022159478435305918,
      "loss": 2.6197,
      "theoretical_loss": 3.4525178111516834,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022158475426278836,
      "loss": 2.5697,
      "theoretical_loss": 3.4525074744698,
      "tokens_seen": 1852702720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022157472417251757,
      "loss": 2.6315,
      "theoretical_loss": 3.4524971382559277,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022156469408224675,
      "loss": 2.4857,
      "theoretical_loss": 3.4524868025100277,
      "tokens_seen": 1852833792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022155466399197593,
      "loss": 2.7993,
      "theoretical_loss": 3.4524764672320627,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215446339017051,
      "loss": 2.4904,
      "theoretical_loss": 3.4524661324219945,
      "tokens_seen": 1852964864
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1020708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6842453479766846,
      "objective/train/theoretical_loss": 3.452455798079786,
      "objective/train/tokens_used": 1873490400,
      "theoretical_loss": 3.452455798079786,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022153460381143432,
      "loss": 2.8054,
      "theoretical_loss": 3.452455798079786,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215245737211635,
      "loss": 2.6826,
      "theoretical_loss": 3.4524454642053994,
      "tokens_seen": 1853095936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022151454363089268,
      "loss": 2.5756,
      "theoretical_loss": 3.4524351307987966,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022150451354062186,
      "loss": 2.5188,
      "theoretical_loss": 3.45242479785994,
      "tokens_seen": 1853227008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022149448345035105,
      "loss": 2.3849,
      "theoretical_loss": 3.452414465388792,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022148445336008025,
      "loss": 2.9179,
      "theoretical_loss": 3.452404133385315,
      "tokens_seen": 1853358080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022147442326980943,
      "loss": 2.5575,
      "theoretical_loss": 3.4523938018494706,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022146439317953862,
      "loss": 2.4219,
      "theoretical_loss": 3.452383470781222,
      "tokens_seen": 1853489152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002214543630892678,
      "loss": 2.7119,
      "theoretical_loss": 3.4523731401805313,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221444332998997,
      "loss": 2.5249,
      "theoretical_loss": 3.4523628100473607,
      "tokens_seen": 1853620224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022143430290872619,
      "loss": 2.5528,
      "theoretical_loss": 3.4523524803816716,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022142427281845537,
      "loss": 2.6649,
      "theoretical_loss": 3.4523421511834282,
      "tokens_seen": 1853751296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022141424272818455,
      "loss": 2.7133,
      "theoretical_loss": 3.4523318224525914,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022140421263791376,
      "loss": 2.7162,
      "theoretical_loss": 3.452321494189124,
      "tokens_seen": 1853882368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022139418254764294,
      "loss": 2.4171,
      "theoretical_loss": 3.452311166392988,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022138415245737212,
      "loss": 2.4683,
      "theoretical_loss": 3.4523008390641463,
      "tokens_seen": 1854013440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002213741223671013,
      "loss": 2.5767,
      "theoretical_loss": 3.452290512202561,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022136409227683048,
      "loss": 2.7395,
      "theoretical_loss": 3.4522801858081946,
      "tokens_seen": 1854144512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002213540621865597,
      "loss": 2.6463,
      "theoretical_loss": 3.4522698598810084,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022134403209628887,
      "loss": 2.7654,
      "theoretical_loss": 3.452259534420967,
      "tokens_seen": 1854275584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022133400200601805,
      "loss": 2.6251,
      "theoretical_loss": 3.4522492094280297,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022132397191574723,
      "loss": 2.4425,
      "theoretical_loss": 3.4522388849021617,
      "tokens_seen": 1854406656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022131394182547644,
      "loss": 2.4743,
      "theoretical_loss": 3.4522285608433236,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022130391173520562,
      "loss": 2.5657,
      "theoretical_loss": 3.4522182372514783,
      "tokens_seen": 1854537728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212938816449348,
      "loss": 2.4172,
      "theoretical_loss": 3.4522079141265882,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1021298,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.786825656890869,
      "objective/train/theoretical_loss": 3.4521975914686163,
      "objective/train/tokens_used": 1875128800,
      "theoretical_loss": 3.4521975914686163,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022128385155466398,
      "loss": 2.5733,
      "theoretical_loss": 3.4521975914686163,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022127382146439317,
      "loss": 2.5356,
      "theoretical_loss": 3.4521872692775237,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022126379137412237,
      "loss": 2.5117,
      "theoretical_loss": 3.4521769475532738,
      "tokens_seen": 1854799872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022125376128385155,
      "loss": 2.5953,
      "theoretical_loss": 3.4521666262958286,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022124373119358074,
      "loss": 2.6571,
      "theoretical_loss": 3.4521563055051505,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022123370110330992,
      "loss": 2.6691,
      "theoretical_loss": 3.452145985181202,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022122367101303913,
      "loss": 2.3793,
      "theoretical_loss": 3.4521356653239454,
      "tokens_seen": 1855062016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212136409227683,
      "loss": 2.78,
      "theoretical_loss": 3.452125345933343,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212036108324975,
      "loss": 2.3841,
      "theoretical_loss": 3.452115027009358,
      "tokens_seen": 1855193088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022119358074222667,
      "loss": 2.4117,
      "theoretical_loss": 3.4521047085519516,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022118355065195585,
      "loss": 2.495,
      "theoretical_loss": 3.4520943905610872,
      "tokens_seen": 1855324160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022117352056168506,
      "loss": 2.513,
      "theoretical_loss": 3.4520840730367266,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022116349047141424,
      "loss": 2.7374,
      "theoretical_loss": 3.452073755978833,
      "tokens_seen": 1855455232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022115346038114342,
      "loss": 2.4252,
      "theoretical_loss": 3.4520634393873673,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002211434302908726,
      "loss": 2.3913,
      "theoretical_loss": 3.452053123262294,
      "tokens_seen": 1855586304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002211334002006018,
      "loss": 2.4919,
      "theoretical_loss": 3.452042807603574,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000221123370110331,
      "loss": 2.4625,
      "theoretical_loss": 3.45203249241117,
      "tokens_seen": 1855717376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022111334002006017,
      "loss": 2.6757,
      "theoretical_loss": 3.452022177685045,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022110330992978935,
      "loss": 2.5698,
      "theoretical_loss": 3.4520118634251613,
      "tokens_seen": 1855848448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022109327983951856,
      "loss": 2.5222,
      "theoretical_loss": 3.452001549631481,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022108324974924777,
      "loss": 2.2605,
      "theoretical_loss": 3.451991236303967,
      "tokens_seen": 1855979520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022107321965897695,
      "loss": 2.5746,
      "theoretical_loss": 3.4519809234425813,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022106318956870613,
      "loss": 2.4827,
      "theoretical_loss": 3.4519706110472868,
      "tokens_seen": 1856110592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210531594784353,
      "loss": 2.6187,
      "theoretical_loss": 3.451960299118046,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022104312938816452,
      "loss": 2.4246,
      "theoretical_loss": 3.451949987654821,
      "tokens_seen": 1856241664
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1022391,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.530734062194824,
      "objective/train/theoretical_loss": 3.4519396766575747,
      "objective/train/tokens_used": 1876767200,
      "theoretical_loss": 3.4519396766575747,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210330992978937,
      "loss": 2.4721,
      "theoretical_loss": 3.4519396766575747,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022102306920762288,
      "loss": 2.6136,
      "theoretical_loss": 3.451929366126269,
      "tokens_seen": 1856372736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022101303911735206,
      "loss": 2.564,
      "theoretical_loss": 3.4519190560608672,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022100300902708125,
      "loss": 2.5945,
      "theoretical_loss": 3.451908746461331,
      "tokens_seen": 1856503808
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022099297893681045,
      "loss": 2.6016,
      "theoretical_loss": 3.4518984373276234,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022098294884653963,
      "loss": 2.3321,
      "theoretical_loss": 3.4518881286597067,
      "tokens_seen": 1856634880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022097291875626882,
      "loss": 2.3731,
      "theoretical_loss": 3.4518778204575433,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220962888665998,
      "loss": 2.5868,
      "theoretical_loss": 3.4518675127210967,
      "tokens_seen": 1856765952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002209528585757272,
      "loss": 2.5601,
      "theoretical_loss": 3.4518572054503283,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022094282848545639,
      "loss": 2.5887,
      "theoretical_loss": 3.4518468986452007,
      "tokens_seen": 1856897024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022093279839518557,
      "loss": 2.3343,
      "theoretical_loss": 3.4518365923056766,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022092276830491475,
      "loss": 2.626,
      "theoretical_loss": 3.4518262864317193,
      "tokens_seen": 1857028096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022091273821464396,
      "loss": 2.566,
      "theoretical_loss": 3.45181598102329,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022090270812437314,
      "loss": 2.5666,
      "theoretical_loss": 3.4518056760803524,
      "tokens_seen": 1857159168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022089267803410232,
      "loss": 2.5676,
      "theoretical_loss": 3.451795371602868,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208826479438315,
      "loss": 2.6498,
      "theoretical_loss": 3.4517850675908006,
      "tokens_seen": 1857290240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022087261785356068,
      "loss": 2.3884,
      "theoretical_loss": 3.451774764044112,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208625877632899,
      "loss": 2.9454,
      "theoretical_loss": 3.4517644609627647,
      "tokens_seen": 1857421312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022085255767301907,
      "loss": 2.2959,
      "theoretical_loss": 3.4517541583467213,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022084252758274825,
      "loss": 2.5481,
      "theoretical_loss": 3.451743856195945,
      "tokens_seen": 1857552384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022083249749247743,
      "loss": 2.6637,
      "theoretical_loss": 3.4517335545103967,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022082246740220664,
      "loss": 2.3302,
      "theoretical_loss": 3.4517232532900413,
      "tokens_seen": 1857683456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022081243731193582,
      "loss": 2.3579,
      "theoretical_loss": 3.45171295253484,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220802407221665,
      "loss": 2.4706,
      "theoretical_loss": 3.451702652244755,
      "tokens_seen": 1857814528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022079237713139418,
      "loss": 2.5423,
      "theoretical_loss": 3.45169235241975,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1022928,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8109171390533447,
      "objective/train/theoretical_loss": 3.4516820530597867,
      "objective/train/tokens_used": 1878405600,
      "theoretical_loss": 3.4516820530597867,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022078234704112337,
      "loss": 2.5516,
      "theoretical_loss": 3.4516820530597867,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022077231695085257,
      "loss": 2.8095,
      "theoretical_loss": 3.4516717541648285,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022076228686058175,
      "loss": 2.7633,
      "theoretical_loss": 3.4516614557348375,
      "tokens_seen": 1858076672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022075225677031094,
      "loss": 2.8056,
      "theoretical_loss": 3.451651157769777,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022074222668004012,
      "loss": 2.6367,
      "theoretical_loss": 3.451640860269608,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022073219658976933,
      "loss": 2.4759,
      "theoretical_loss": 3.4516305632342945,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207221664994985,
      "loss": 2.8115,
      "theoretical_loss": 3.4516202666637987,
      "tokens_seen": 1858338816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207121364092277,
      "loss": 2.5131,
      "theoretical_loss": 3.451609970558083,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022070210631895687,
      "loss": 2.5666,
      "theoretical_loss": 3.4515996749171114,
      "tokens_seen": 1858469888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022069207622868605,
      "loss": 2.5372,
      "theoretical_loss": 3.4515893797408443,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022068204613841526,
      "loss": 2.619,
      "theoretical_loss": 3.4515790850292465,
      "tokens_seen": 1858600960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022067201604814444,
      "loss": 2.4233,
      "theoretical_loss": 3.4515687907822787,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022066198595787362,
      "loss": 2.4398,
      "theoretical_loss": 3.451558496999905,
      "tokens_seen": 1858732032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206519558676028,
      "loss": 2.4826,
      "theoretical_loss": 3.451548203682088,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220641925777332,
      "loss": 2.6631,
      "theoretical_loss": 3.451537910828789,
      "tokens_seen": 1858863104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206318956870612,
      "loss": 2.5146,
      "theoretical_loss": 3.4515276184399717,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022062186559679037,
      "loss": 2.5733,
      "theoretical_loss": 3.451517326515599,
      "tokens_seen": 1858994176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022061183550651955,
      "loss": 2.9118,
      "theoretical_loss": 3.4515070350556334,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022060180541624873,
      "loss": 2.6222,
      "theoretical_loss": 3.4514967440600364,
      "tokens_seen": 1859125248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022059177532597794,
      "loss": 2.5249,
      "theoretical_loss": 3.451486453528773,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022058174523570712,
      "loss": 2.6624,
      "theoretical_loss": 3.4514761634618036,
      "tokens_seen": 1859256320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205717151454363,
      "loss": 2.6334,
      "theoretical_loss": 3.4514658738590915,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022056168505516549,
      "loss": 2.6076,
      "theoretical_loss": 3.4514555847206,
      "tokens_seen": 1859387392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205516549648947,
      "loss": 2.4759,
      "theoretical_loss": 3.451445296046292,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022054162487462387,
      "loss": 2.3641,
      "theoretical_loss": 3.4514350078361296,
      "tokens_seen": 1859518464
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1024242,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7819745540618896,
      "objective/train/theoretical_loss": 3.4514247200900754,
      "objective/train/tokens_used": 1880044000,
      "theoretical_loss": 3.4514247200900754,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022053159478435306,
      "loss": 2.621,
      "theoretical_loss": 3.4514247200900754,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022052156469408224,
      "loss": 2.394,
      "theoretical_loss": 3.4514144328080922,
      "tokens_seen": 1859649536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022051153460381142,
      "loss": 2.6649,
      "theoretical_loss": 3.4514041459901432,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022050150451354063,
      "loss": 2.7791,
      "theoretical_loss": 3.4513938596361906,
      "tokens_seen": 1859780608
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204914744232698,
      "loss": 2.6367,
      "theoretical_loss": 3.451383573746197,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220481444332999,
      "loss": 2.6147,
      "theoretical_loss": 3.4513732883201254,
      "tokens_seen": 1859911680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022047141424272817,
      "loss": 2.3835,
      "theoretical_loss": 3.451363003357939,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022046138415245738,
      "loss": 2.5262,
      "theoretical_loss": 3.4513527188596,
      "tokens_seen": 1860042752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022045135406218656,
      "loss": 2.5624,
      "theoretical_loss": 3.4513424348250705,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022044132397191574,
      "loss": 2.5454,
      "theoretical_loss": 3.4513321512543147,
      "tokens_seen": 1860173824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022043129388164492,
      "loss": 2.5476,
      "theoretical_loss": 3.451321868147294,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204212637913741,
      "loss": 2.4776,
      "theoretical_loss": 3.451311585503972,
      "tokens_seen": 1860304896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204112337011033,
      "loss": 2.5124,
      "theoretical_loss": 3.4513013033243114,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204012036108325,
      "loss": 2.6966,
      "theoretical_loss": 3.4512910216082746,
      "tokens_seen": 1860435968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022039117352056167,
      "loss": 2.5057,
      "theoretical_loss": 3.4512807403558243,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022038114343029085,
      "loss": 2.4553,
      "theoretical_loss": 3.451270459566924,
      "tokens_seen": 1860567040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022037111334002006,
      "loss": 2.7253,
      "theoretical_loss": 3.451260179241536,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022036108324974924,
      "loss": 2.8563,
      "theoretical_loss": 3.451249899379622,
      "tokens_seen": 1860698112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022035105315947842,
      "loss": 2.5078,
      "theoretical_loss": 3.4512396199811466,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022034102306920763,
      "loss": 2.7114,
      "theoretical_loss": 3.451229341046072,
      "tokens_seen": 1860829184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022033099297893684,
      "loss": 2.6588,
      "theoretical_loss": 3.4512190625743604,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022032096288866602,
      "loss": 2.562,
      "theoretical_loss": 3.451208784565975,
      "tokens_seen": 1860960256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002203109327983952,
      "loss": 2.7774,
      "theoretical_loss": 3.4511985070208793,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022030090270812438,
      "loss": 2.4792,
      "theoretical_loss": 3.4511882299390346,
      "tokens_seen": 1861091328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022029087261785357,
      "loss": 2.6692,
      "theoretical_loss": 3.4511779533204043,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1025485,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.364495277404785,
      "objective/train/theoretical_loss": 3.4511676771649524,
      "objective/train/tokens_used": 1881682400,
      "theoretical_loss": 3.4511676771649524,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022028084252758277,
      "loss": 2.4558,
      "theoretical_loss": 3.4511676771649524,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022027081243731195,
      "loss": 2.5241,
      "theoretical_loss": 3.4511574014726403,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022026078234704114,
      "loss": 2.3883,
      "theoretical_loss": 3.4511471262434306,
      "tokens_seen": 1861353472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022025075225677032,
      "loss": 2.5707,
      "theoretical_loss": 3.4511368514772878,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022024072216649953,
      "loss": 2.5177,
      "theoretical_loss": 3.451126577174173,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002202306920762287,
      "loss": 2.2479,
      "theoretical_loss": 3.4511163033340497,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002202206619859579,
      "loss": 2.6098,
      "theoretical_loss": 3.451106029956881,
      "tokens_seen": 1861615616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022021063189568707,
      "loss": 2.6704,
      "theoretical_loss": 3.45109575704263,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022020060180541625,
      "loss": 2.4511,
      "theoretical_loss": 3.451085484591258,
      "tokens_seen": 1861746688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022019057171514546,
      "loss": 2.7473,
      "theoretical_loss": 3.4510752126027295,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022018054162487464,
      "loss": 2.4846,
      "theoretical_loss": 3.451064941077007,
      "tokens_seen": 1861877760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022017051153460382,
      "loss": 2.665,
      "theoretical_loss": 3.451054670014053,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220160481444333,
      "loss": 2.6516,
      "theoretical_loss": 3.4510443994138305,
      "tokens_seen": 1862008832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002201504513540622,
      "loss": 2.619,
      "theoretical_loss": 3.451034129276302,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002201404212637914,
      "loss": 2.5582,
      "theoretical_loss": 3.451023859601431,
      "tokens_seen": 1862139904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022013039117352057,
      "loss": 2.6601,
      "theoretical_loss": 3.45101359038918,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022012036108324975,
      "loss": 2.6482,
      "theoretical_loss": 3.4510033216395124,
      "tokens_seen": 1862270976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022011033099297893,
      "loss": 2.6375,
      "theoretical_loss": 3.4509930533523905,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022010030090270814,
      "loss": 2.5584,
      "theoretical_loss": 3.450982785527777,
      "tokens_seen": 1862402048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022009027081243732,
      "loss": 2.8244,
      "theoretical_loss": 3.4509725181656354,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002200802407221665,
      "loss": 2.5258,
      "theoretical_loss": 3.4509622512659286,
      "tokens_seen": 1862533120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022007021063189569,
      "loss": 2.3938,
      "theoretical_loss": 3.4509519848286185,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002200601805416249,
      "loss": 2.7295,
      "theoretical_loss": 3.4509417188536693,
      "tokens_seen": 1862664192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022005015045135408,
      "loss": 2.7298,
      "theoretical_loss": 3.4509314533410436,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022004012036108326,
      "loss": 2.587,
      "theoretical_loss": 3.450921188290704,
      "tokens_seen": 1862795264
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1026014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4433135986328125,
      "objective/train/theoretical_loss": 3.450910923702613,
      "objective/train/tokens_used": 1883320800,
      "theoretical_loss": 3.450910923702613,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022003009027081244,
      "loss": 2.6583,
      "theoretical_loss": 3.450910923702613,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022002006018054162,
      "loss": 2.5258,
      "theoretical_loss": 3.4509006595767344,
      "tokens_seen": 1862926336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022001003009027083,
      "loss": 2.6364,
      "theoretical_loss": 3.4508903959130306,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022,
      "loss": 2.6577,
      "theoretical_loss": 3.4508801327114647,
      "tokens_seen": 1863057408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002199899699097292,
      "loss": 2.6654,
      "theoretical_loss": 3.450869869972,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021997993981945837,
      "loss": 2.6603,
      "theoretical_loss": 3.450859607694599,
      "tokens_seen": 1863188480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021996990972918758,
      "loss": 2.8481,
      "theoretical_loss": 3.4508493458792246,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021995987963891676,
      "loss": 2.7886,
      "theoretical_loss": 3.4508390845258394,
      "tokens_seen": 1863319552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021994984954864594,
      "loss": 2.5715,
      "theoretical_loss": 3.4508288236344073,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021993981945837512,
      "loss": 2.5184,
      "theoretical_loss": 3.4508185632048907,
      "tokens_seen": 1863450624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002199297893681043,
      "loss": 2.5696,
      "theoretical_loss": 3.4508083032372525,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002199197592778335,
      "loss": 2.6915,
      "theoretical_loss": 3.4507980437314556,
      "tokens_seen": 1863581696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002199097291875627,
      "loss": 2.5399,
      "theoretical_loss": 3.4507877846874635,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021989969909729187,
      "loss": 2.4616,
      "theoretical_loss": 3.450777526105239,
      "tokens_seen": 1863712768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021988966900702105,
      "loss": 2.7334,
      "theoretical_loss": 3.4507672679847445,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021987963891675026,
      "loss": 2.6122,
      "theoretical_loss": 3.450757010325944,
      "tokens_seen": 1863843840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021986960882647944,
      "loss": 2.5584,
      "theoretical_loss": 3.450746753128799,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021985957873620862,
      "loss": 2.6957,
      "theoretical_loss": 3.450736496393274,
      "tokens_seen": 1863974912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002198495486459378,
      "loss": 2.6993,
      "theoretical_loss": 3.4507262401193315,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021983951855566701,
      "loss": 2.7001,
      "theoretical_loss": 3.4507159843069344,
      "tokens_seen": 1864105984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002198294884653962,
      "loss": 2.5239,
      "theoretical_loss": 3.450705728956045,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021981945837512538,
      "loss": 2.7563,
      "theoretical_loss": 3.4506954740666274,
      "tokens_seen": 1864237056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021980942828485456,
      "loss": 2.6029,
      "theoretical_loss": 3.450685219638644,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021979939819458374,
      "loss": 2.6244,
      "theoretical_loss": 3.450674965672059,
      "tokens_seen": 1864368128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021978936810431295,
      "loss": 2.5273,
      "theoretical_loss": 3.4506647121668332,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.56,
      "objective/train/docs_used": 1027424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8552238941192627,
      "objective/train/theoretical_loss": 3.450654459122932,
      "objective/train/tokens_used": 1884959200,
      "theoretical_loss": 3.450654459122932,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021977933801404213,
      "loss": 2.6503,
      "theoretical_loss": 3.450654459122932,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002197693079237713,
      "loss": 2.6675,
      "theoretical_loss": 3.450644206540316,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002197592778335005,
      "loss": 2.5767,
      "theoretical_loss": 3.4506339544189504,
      "tokens_seen": 1864630272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002197492477432297,
      "loss": 2.5396,
      "theoretical_loss": 3.450623702758797,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021973921765295888,
      "loss": 2.8186,
      "theoretical_loss": 3.4506134515598195,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021972918756268806,
      "loss": 2.7416,
      "theoretical_loss": 3.450603200821981,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021971915747241724,
      "loss": 2.4129,
      "theoretical_loss": 3.4505929505452437,
      "tokens_seen": 1864892416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021970912738214642,
      "loss": 2.5009,
      "theoretical_loss": 3.450582700729571,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021969909729187563,
      "loss": 2.3439,
      "theoretical_loss": 3.4505724513749265,
      "tokens_seen": 1865023488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002196890672016048,
      "loss": 2.6238,
      "theoretical_loss": 3.4505622024812728,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219679037111334,
      "loss": 2.678,
      "theoretical_loss": 3.450551954048573,
      "tokens_seen": 1865154560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021966900702106317,
      "loss": 2.7107,
      "theoretical_loss": 3.4505417060767902,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021965897693079238,
      "loss": 2.7215,
      "theoretical_loss": 3.4505314585658877,
      "tokens_seen": 1865285632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021964894684052156,
      "loss": 2.5274,
      "theoretical_loss": 3.450521211515828,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021963891675025075,
      "loss": 2.5396,
      "theoretical_loss": 3.4505109649265755,
      "tokens_seen": 1865416704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021962888665997993,
      "loss": 2.7585,
      "theoretical_loss": 3.4505007187980916,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002196188565697091,
      "loss": 2.582,
      "theoretical_loss": 3.4504904731303405,
      "tokens_seen": 1865547776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021960882647943832,
      "loss": 2.6494,
      "theoretical_loss": 3.4504802279232845,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195987963891675,
      "loss": 2.6581,
      "theoretical_loss": 3.4504699831768875,
      "tokens_seen": 1865678848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195887662988967,
      "loss": 2.6672,
      "theoretical_loss": 3.4504597388911122,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021957873620862589,
      "loss": 2.5101,
      "theoretical_loss": 3.4504494950659215,
      "tokens_seen": 1865809920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195687061183551,
      "loss": 2.5556,
      "theoretical_loss": 3.4504392517012796,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021955867602808428,
      "loss": 2.5388,
      "theoretical_loss": 3.450429008797148,
      "tokens_seen": 1865940992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021954864593781346,
      "loss": 2.7981,
      "theoretical_loss": 3.450418766353491,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021953861584754264,
      "loss": 2.5691,
      "theoretical_loss": 3.450408524370271,
      "tokens_seen": 1866072064
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1028516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4778594970703125,
      "objective/train/theoretical_loss": 3.450398282847452,
      "objective/train/tokens_used": 1886597600,
      "theoretical_loss": 3.450398282847452,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021952858575727182,
      "loss": 2.6057,
      "theoretical_loss": 3.450398282847452,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021951855566700103,
      "loss": 2.7577,
      "theoretical_loss": 3.4503880417849966,
      "tokens_seen": 1866203136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195085255767302,
      "loss": 2.4753,
      "theoretical_loss": 3.450377801182868,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194984954864594,
      "loss": 2.6219,
      "theoretical_loss": 3.4503675610410287,
      "tokens_seen": 1866334208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021948846539618857,
      "loss": 2.6233,
      "theoretical_loss": 3.4503573213594434,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021947843530591778,
      "loss": 2.6314,
      "theoretical_loss": 3.450347082138074,
      "tokens_seen": 1866465280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021946840521564696,
      "loss": 2.5833,
      "theoretical_loss": 3.4503368433768835,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021945837512537614,
      "loss": 2.5654,
      "theoretical_loss": 3.450326605075836,
      "tokens_seen": 1866596352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021944834503510532,
      "loss": 2.7844,
      "theoretical_loss": 3.450316367234894,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194383149448345,
      "loss": 2.6802,
      "theoretical_loss": 3.450306129854021,
      "tokens_seen": 1866727424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194282848545637,
      "loss": 2.5385,
      "theoretical_loss": 3.45029589293318,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194182547642929,
      "loss": 2.4877,
      "theoretical_loss": 3.4502856564723343,
      "tokens_seen": 1866858496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021940822467402207,
      "loss": 2.7638,
      "theoretical_loss": 3.450275420471447,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021939819458375125,
      "loss": 2.5963,
      "theoretical_loss": 3.4502651849304815,
      "tokens_seen": 1866989568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021938816449348046,
      "loss": 2.4143,
      "theoretical_loss": 3.4502549498494,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021937813440320964,
      "loss": 2.5939,
      "theoretical_loss": 3.4502447152281674,
      "tokens_seen": 1867120640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021936810431293882,
      "loss": 2.6595,
      "theoretical_loss": 3.4502344810667456,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219358074222668,
      "loss": 2.9239,
      "theoretical_loss": 3.4502242473650986,
      "tokens_seen": 1867251712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021934804413239721,
      "loss": 2.5543,
      "theoretical_loss": 3.4502140141231887,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193380140421264,
      "loss": 2.8543,
      "theoretical_loss": 3.4502037813409796,
      "tokens_seen": 1867382784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021932798395185558,
      "loss": 2.686,
      "theoretical_loss": 3.450193549018435,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021931795386158476,
      "loss": 2.845,
      "theoretical_loss": 3.4501833171555174,
      "tokens_seen": 1867513856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021930792377131394,
      "loss": 2.6913,
      "theoretical_loss": 3.4501730857521906,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021929789368104315,
      "loss": 2.5314,
      "theoretical_loss": 3.4501628548084167,
      "tokens_seen": 1867644928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021928786359077233,
      "loss": 2.5856,
      "theoretical_loss": 3.4501526243241605,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1029117,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4569618701934814,
      "objective/train/theoretical_loss": 3.4501423942993843,
      "objective/train/tokens_used": 1888236000,
      "theoretical_loss": 3.4501423942993843,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192778335005015,
      "loss": 2.7411,
      "theoretical_loss": 3.4501423942993843,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192678034102307,
      "loss": 2.6731,
      "theoretical_loss": 3.4501321647340513,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192577733199599,
      "loss": 2.7839,
      "theoretical_loss": 3.4501219356281254,
      "tokens_seen": 1867907072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021924774322968908,
      "loss": 2.6182,
      "theoretical_loss": 3.450111706981569,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021923771313941826,
      "loss": 2.4964,
      "theoretical_loss": 3.4501014787943456,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021922768304914744,
      "loss": 2.8061,
      "theoretical_loss": 3.4500912510664192,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021921765295887662,
      "loss": 2.4417,
      "theoretical_loss": 3.450081023797752,
      "tokens_seen": 1868169216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021920762286860583,
      "loss": 2.7699,
      "theoretical_loss": 3.4500707969883084,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219197592778335,
      "loss": 2.4939,
      "theoretical_loss": 3.4500605706380503,
      "tokens_seen": 1868300288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191875626880642,
      "loss": 2.8798,
      "theoretical_loss": 3.450050344746942,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021917753259779337,
      "loss": 2.4515,
      "theoretical_loss": 3.450040119314947,
      "tokens_seen": 1868431360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021916750250752258,
      "loss": 2.6821,
      "theoretical_loss": 3.450029894342028,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021915747241725176,
      "loss": 2.413,
      "theoretical_loss": 3.450019669828148,
      "tokens_seen": 1868562432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021914744232698095,
      "loss": 2.698,
      "theoretical_loss": 3.4500094457732704,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021913741223671013,
      "loss": 2.7474,
      "theoretical_loss": 3.449999222177359,
      "tokens_seen": 1868693504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191273821464393,
      "loss": 2.5893,
      "theoretical_loss": 3.449988999040377,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021911735205616852,
      "loss": 2.5428,
      "theoretical_loss": 3.449978776362288,
      "tokens_seen": 1868824576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191073219658977,
      "loss": 2.7043,
      "theoretical_loss": 3.4499685541430543,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021909729187562688,
      "loss": 2.7631,
      "theoretical_loss": 3.4499583323826397,
      "tokens_seen": 1868955648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021908726178535606,
      "loss": 2.5376,
      "theoretical_loss": 3.4499481110810075,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021907723169508527,
      "loss": 2.7759,
      "theoretical_loss": 3.4499378902381217,
      "tokens_seen": 1869086720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021906720160481445,
      "loss": 2.7428,
      "theoretical_loss": 3.449927669853945,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021905717151454363,
      "loss": 2.3804,
      "theoretical_loss": 3.4499174499284404,
      "tokens_seen": 1869217792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002190471414242728,
      "loss": 2.9239,
      "theoretical_loss": 3.449907230461572,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000219037111334002,
      "loss": 2.7706,
      "theoretical_loss": 3.4498970114533023,
      "tokens_seen": 1869348864
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1029470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3672564029693604,
      "objective/train/theoretical_loss": 3.4498867929035955,
      "objective/train/tokens_used": 1889874400,
      "theoretical_loss": 3.4498867929035955,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002190270812437312,
      "loss": 2.6151,
      "theoretical_loss": 3.4498867929035955,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021901705115346038,
      "loss": 2.5428,
      "theoretical_loss": 3.4498765748124143,
      "tokens_seen": 1869479936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021900702106318956,
      "loss": 2.3047,
      "theoretical_loss": 3.449866357179723,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021899699097291874,
      "loss": 2.6136,
      "theoretical_loss": 3.4498561400054832,
      "tokens_seen": 1869611008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021898696088264795,
      "loss": 2.8139,
      "theoretical_loss": 3.44984592328966,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021897693079237713,
      "loss": 2.7477,
      "theoretical_loss": 3.449835707032216,
      "tokens_seen": 1869742080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021896690070210631,
      "loss": 2.5623,
      "theoretical_loss": 3.4498254912331148,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002189568706118355,
      "loss": 2.6637,
      "theoretical_loss": 3.4498152758923197,
      "tokens_seen": 1869873152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021894684052156468,
      "loss": 2.6128,
      "theoretical_loss": 3.4498050610097932,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021893681043129388,
      "loss": 2.4096,
      "theoretical_loss": 3.4497948465855,
      "tokens_seen": 1870004224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021892678034102307,
      "loss": 2.6036,
      "theoretical_loss": 3.4497846326194033,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021891675025075225,
      "loss": 2.5905,
      "theoretical_loss": 3.449774419111466,
      "tokens_seen": 1870135296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021890672016048143,
      "loss": 2.5652,
      "theoretical_loss": 3.449764206061652,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021889669007021064,
      "loss": 2.3135,
      "theoretical_loss": 3.4497539934699235,
      "tokens_seen": 1870266368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021888665997993982,
      "loss": 2.902,
      "theoretical_loss": 3.4497437813362453,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218876629889669,
      "loss": 2.6873,
      "theoretical_loss": 3.44973356966058,
      "tokens_seen": 1870397440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021886659979939818,
      "loss": 2.6597,
      "theoretical_loss": 3.449723358442892,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021885656970912736,
      "loss": 2.7709,
      "theoretical_loss": 3.449713147683143,
      "tokens_seen": 1870528512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021884653961885657,
      "loss": 2.5725,
      "theoretical_loss": 3.449702937381298,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021883650952858578,
      "loss": 2.7385,
      "theoretical_loss": 3.44969272753732,
      "tokens_seen": 1870659584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021882647943831496,
      "loss": 2.7196,
      "theoretical_loss": 3.449682518151172,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021881644934804414,
      "loss": 2.7064,
      "theoretical_loss": 3.4496723092228176,
      "tokens_seen": 1870790656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021880641925777335,
      "loss": 2.5338,
      "theoretical_loss": 3.449662100752221,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021879638916750253,
      "loss": 2.5252,
      "theoretical_loss": 3.449651892739344,
      "tokens_seen": 1870921728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187863590772317,
      "loss": 2.5859,
      "theoretical_loss": 3.449641685184152,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1030757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7632298469543457,
      "objective/train/theoretical_loss": 3.449631478086607,
      "objective/train/tokens_used": 1891512800,
      "theoretical_loss": 3.449631478086607,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187763289869609,
      "loss": 2.622,
      "theoretical_loss": 3.449631478086607,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187662988966901,
      "loss": 2.632,
      "theoretical_loss": 3.4496212714466727,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021875626880641928,
      "loss": 2.6617,
      "theoretical_loss": 3.449611065264313,
      "tokens_seen": 1871183872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021874623871614846,
      "loss": 2.5802,
      "theoretical_loss": 3.449600859539491,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021873620862587764,
      "loss": 2.7123,
      "theoretical_loss": 3.4495906542721704,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021872617853560682,
      "loss": 2.8172,
      "theoretical_loss": 3.4495804494623146,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021871614844533603,
      "loss": 2.7026,
      "theoretical_loss": 3.4495702451098875,
      "tokens_seen": 1871446016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187061183550652,
      "loss": 2.5698,
      "theoretical_loss": 3.449560041214851,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186960882647944,
      "loss": 2.4917,
      "theoretical_loss": 3.4495498377771705,
      "tokens_seen": 1871577088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021868605817452357,
      "loss": 2.6701,
      "theoretical_loss": 3.4495396347968086,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021867602808425278,
      "loss": 2.7399,
      "theoretical_loss": 3.4495294322737293,
      "tokens_seen": 1871708160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021866599799398196,
      "loss": 2.4671,
      "theoretical_loss": 3.449519230207895,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021865596790371115,
      "loss": 2.4782,
      "theoretical_loss": 3.4495090285992704,
      "tokens_seen": 1871839232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021864593781344033,
      "loss": 2.6291,
      "theoretical_loss": 3.4494988274478184,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186359077231695,
      "loss": 2.4538,
      "theoretical_loss": 3.4494886267535025,
      "tokens_seen": 1871970304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021862587763289872,
      "loss": 2.6177,
      "theoretical_loss": 3.4494784265162863,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186158475426279,
      "loss": 2.7207,
      "theoretical_loss": 3.4494682267361334,
      "tokens_seen": 1872101376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021860581745235708,
      "loss": 2.4203,
      "theoretical_loss": 3.449458027413007,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021859578736208626,
      "loss": 2.3798,
      "theoretical_loss": 3.4494478285468713,
      "tokens_seen": 1872232448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021858575727181547,
      "loss": 2.6976,
      "theoretical_loss": 3.4494376301376892,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021857572718154465,
      "loss": 2.5521,
      "theoretical_loss": 3.4494274321854244,
      "tokens_seen": 1872363520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021856569709127383,
      "loss": 2.6756,
      "theoretical_loss": 3.4494172346900402,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218555667001003,
      "loss": 2.4651,
      "theoretical_loss": 3.449407037651501,
      "tokens_seen": 1872494592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002185456369107322,
      "loss": 2.7643,
      "theoretical_loss": 3.4493968410697695,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002185356068204614,
      "loss": 2.4973,
      "theoretical_loss": 3.449386644944809,
      "tokens_seen": 1872625664
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1031383,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.284093141555786,
      "objective/train/theoretical_loss": 3.449376449276584,
      "objective/train/tokens_used": 1893151200,
      "theoretical_loss": 3.449376449276584,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021852557673019058,
      "loss": 2.6845,
      "theoretical_loss": 3.449376449276584,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021851554663991976,
      "loss": 2.6504,
      "theoretical_loss": 3.449366254065058,
      "tokens_seen": 1872756736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021850551654964894,
      "loss": 2.7234,
      "theoretical_loss": 3.449356059310194,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021849548645937815,
      "loss": 2.6692,
      "theoretical_loss": 3.4493458650119555,
      "tokens_seen": 1872887808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021848545636910733,
      "loss": 2.3069,
      "theoretical_loss": 3.4493356711703065,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021847542627883651,
      "loss": 2.7049,
      "theoretical_loss": 3.4493254777852096,
      "tokens_seen": 1873018880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002184653961885657,
      "loss": 2.6406,
      "theoretical_loss": 3.44931528485663,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021845536609829488,
      "loss": 2.7233,
      "theoretical_loss": 3.4493050923845305,
      "tokens_seen": 1873149952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021844533600802408,
      "loss": 2.6192,
      "theoretical_loss": 3.4492949003688747,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021843530591775327,
      "loss": 2.5628,
      "theoretical_loss": 3.4492847088096257,
      "tokens_seen": 1873281024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021842527582748245,
      "loss": 2.6108,
      "theoretical_loss": 3.4492745177067476,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021841524573721163,
      "loss": 2.5833,
      "theoretical_loss": 3.449264327060204,
      "tokens_seen": 1873412096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021840521564694084,
      "loss": 2.7211,
      "theoretical_loss": 3.4492541368699587,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021839518555667002,
      "loss": 2.6298,
      "theoretical_loss": 3.449243947135974,
      "tokens_seen": 1873543168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002183851554663992,
      "loss": 2.6526,
      "theoretical_loss": 3.4492337578582157,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021837512537612838,
      "loss": 2.7219,
      "theoretical_loss": 3.449223569036646,
      "tokens_seen": 1873674240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021836509528585756,
      "loss": 2.7253,
      "theoretical_loss": 3.449213380671229,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021835506519558677,
      "loss": 2.5892,
      "theoretical_loss": 3.4492031927619275,
      "tokens_seen": 1873805312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021834503510531595,
      "loss": 2.6266,
      "theoretical_loss": 3.4491930053087065,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021833500501504513,
      "loss": 2.4115,
      "theoretical_loss": 3.4491828183115283,
      "tokens_seen": 1873936384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002183249749247743,
      "loss": 2.7947,
      "theoretical_loss": 3.449172631770357,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021831494483450352,
      "loss": 2.6462,
      "theoretical_loss": 3.4491624456851566,
      "tokens_seen": 1874067456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002183049147442327,
      "loss": 2.6578,
      "theoretical_loss": 3.4491522600558904,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021829488465396188,
      "loss": 2.8612,
      "theoretical_loss": 3.4491420748825226,
      "tokens_seen": 1874198528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021828485456369106,
      "loss": 2.4775,
      "theoretical_loss": 3.449131890165016,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1031901,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4496192932128906,
      "objective/train/theoretical_loss": 3.449121705903335,
      "objective/train/tokens_used": 1894789600,
      "theoretical_loss": 3.449121705903335,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021827482447342024,
      "loss": 2.5747,
      "theoretical_loss": 3.449121705903335,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021826479438314945,
      "loss": 2.5933,
      "theoretical_loss": 3.4491115220974424,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021825476429287863,
      "loss": 2.3842,
      "theoretical_loss": 3.449101338747303,
      "tokens_seen": 1874460672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021824473420260782,
      "loss": 2.6334,
      "theoretical_loss": 3.449091155852879,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218234704112337,
      "loss": 2.779,
      "theoretical_loss": 3.4490809734141354,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002182246740220662,
      "loss": 2.6494,
      "theoretical_loss": 3.4490707914310352,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021821464393179539,
      "loss": 2.5599,
      "theoretical_loss": 3.449060609903543,
      "tokens_seen": 1874722816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021820461384152457,
      "loss": 2.539,
      "theoretical_loss": 3.4490504288316215,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021819458375125375,
      "loss": 2.5418,
      "theoretical_loss": 3.4490402482152343,
      "tokens_seen": 1874853888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021818455366098296,
      "loss": 2.5835,
      "theoretical_loss": 3.4490300680543458,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021817452357071214,
      "loss": 2.5644,
      "theoretical_loss": 3.449019888348919,
      "tokens_seen": 1874984960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021816449348044132,
      "loss": 2.5752,
      "theoretical_loss": 3.4490097090989185,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181544633901705,
      "loss": 2.5417,
      "theoretical_loss": 3.4489995303043073,
      "tokens_seen": 1875116032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021814443329989968,
      "loss": 2.5993,
      "theoretical_loss": 3.448989351965049,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181344032096289,
      "loss": 2.577,
      "theoretical_loss": 3.448979174081108,
      "tokens_seen": 1875247104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021812437311935807,
      "loss": 2.6583,
      "theoretical_loss": 3.4489689966524475,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021811434302908725,
      "loss": 2.7963,
      "theoretical_loss": 3.4489588196790315,
      "tokens_seen": 1875378176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021810431293881643,
      "loss": 2.6898,
      "theoretical_loss": 3.448948643160823,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021809428284854564,
      "loss": 2.4347,
      "theoretical_loss": 3.448938467097787,
      "tokens_seen": 1875509248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021808425275827485,
      "loss": 2.5978,
      "theoretical_loss": 3.448928291489886,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021807422266800403,
      "loss": 2.6174,
      "theoretical_loss": 3.448918116337085,
      "tokens_seen": 1875640320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002180641925777332,
      "loss": 2.6622,
      "theoretical_loss": 3.448907941639346,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002180541624874624,
      "loss": 2.5585,
      "theoretical_loss": 3.448897767396635,
      "tokens_seen": 1875771392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002180441323971916,
      "loss": 2.6479,
      "theoretical_loss": 3.448887593608914,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021803410230692078,
      "loss": 2.5061,
      "theoretical_loss": 3.448877420276147,
      "tokens_seen": 1875902464
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1033016,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0129494667053223,
      "objective/train/theoretical_loss": 3.448867247398298,
      "objective/train/tokens_used": 1896428000,
      "theoretical_loss": 3.448867247398298,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021802407221664996,
      "loss": 2.8145,
      "theoretical_loss": 3.448867247398298,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021801404212637914,
      "loss": 2.4844,
      "theoretical_loss": 3.448857074975332,
      "tokens_seen": 1876033536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021800401203610835,
      "loss": 2.6229,
      "theoretical_loss": 3.448846903007211,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021799398194583753,
      "loss": 2.3976,
      "theoretical_loss": 3.448836731493899,
      "tokens_seen": 1876164608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021798395185556671,
      "loss": 2.704,
      "theoretical_loss": 3.44882656043536,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002179739217652959,
      "loss": 2.3667,
      "theoretical_loss": 3.4488163898315585,
      "tokens_seen": 1876295680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021796389167502508,
      "loss": 2.6534,
      "theoretical_loss": 3.4488062196824574,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021795386158475428,
      "loss": 2.6692,
      "theoretical_loss": 3.448796049988021,
      "tokens_seen": 1876426752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021794383149448347,
      "loss": 2.4814,
      "theoretical_loss": 3.448785880748213,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021793380140421265,
      "loss": 2.6567,
      "theoretical_loss": 3.4487757119629965,
      "tokens_seen": 1876557824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021792377131394183,
      "loss": 2.6099,
      "theoretical_loss": 3.4487655436323363,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021791374122367104,
      "loss": 2.5879,
      "theoretical_loss": 3.448755375756196,
      "tokens_seen": 1876688896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021790371113340022,
      "loss": 2.607,
      "theoretical_loss": 3.4487452083345396,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178936810431294,
      "loss": 2.8691,
      "theoretical_loss": 3.44873504136733,
      "tokens_seen": 1876819968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021788365095285858,
      "loss": 2.4729,
      "theoretical_loss": 3.4487248748545314,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021787362086258776,
      "loss": 2.7277,
      "theoretical_loss": 3.448714708796108,
      "tokens_seen": 1876951040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021786359077231697,
      "loss": 2.5751,
      "theoretical_loss": 3.4487045431920236,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021785356068204615,
      "loss": 2.5303,
      "theoretical_loss": 3.4486943780422417,
      "tokens_seen": 1877082112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021784353059177533,
      "loss": 2.4925,
      "theoretical_loss": 3.4486842133467266,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178335005015045,
      "loss": 2.7176,
      "theoretical_loss": 3.4486740491054415,
      "tokens_seen": 1877213184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021782347041123372,
      "loss": 2.4803,
      "theoretical_loss": 3.448663885318351,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178134403209629,
      "loss": 2.7139,
      "theoretical_loss": 3.448653721985418,
      "tokens_seen": 1877344256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021780341023069208,
      "loss": 2.6215,
      "theoretical_loss": 3.448643559106607,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021779338014042126,
      "loss": 2.6654,
      "theoretical_loss": 3.448633396681882,
      "tokens_seen": 1877475328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021778335005015044,
      "loss": 2.5077,
      "theoretical_loss": 3.4486232347112065,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1034169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.636319875717163,
      "objective/train/theoretical_loss": 3.448613073194544,
      "objective/train/tokens_used": 1898066400,
      "theoretical_loss": 3.448613073194544,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021777331995987965,
      "loss": 2.654,
      "theoretical_loss": 3.448613073194544,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021776328986960883,
      "loss": 2.613,
      "theoretical_loss": 3.4486029121318595,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021775325977933802,
      "loss": 2.5286,
      "theoretical_loss": 3.448592751523116,
      "tokens_seen": 1877737472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002177432296890672,
      "loss": 2.336,
      "theoretical_loss": 3.4485825913682775,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002177331995987964,
      "loss": 2.5292,
      "theoretical_loss": 3.448572431667308,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021772316950852559,
      "loss": 2.5878,
      "theoretical_loss": 3.448562272420171,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021771313941825477,
      "loss": 2.6508,
      "theoretical_loss": 3.4485521136268313,
      "tokens_seen": 1877999616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021770310932798395,
      "loss": 2.5287,
      "theoretical_loss": 3.4485419552872525,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021769307923771316,
      "loss": 2.6271,
      "theoretical_loss": 3.4485317974013974,
      "tokens_seen": 1878130688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021768304914744234,
      "loss": 2.6047,
      "theoretical_loss": 3.4485216399692313,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021767301905717152,
      "loss": 2.6241,
      "theoretical_loss": 3.4485114829907175,
      "tokens_seen": 1878261760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002176629889669007,
      "loss": 2.7097,
      "theoretical_loss": 3.44850132646582,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021765295887662988,
      "loss": 2.7325,
      "theoretical_loss": 3.4484911703945027,
      "tokens_seen": 1878392832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002176429287863591,
      "loss": 2.5331,
      "theoretical_loss": 3.4484810147767293,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021763289869608827,
      "loss": 2.5554,
      "theoretical_loss": 3.448470859612464,
      "tokens_seen": 1878523904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021762286860581745,
      "loss": 2.7094,
      "theoretical_loss": 3.4484607049016702,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021761283851554663,
      "loss": 2.7584,
      "theoretical_loss": 3.448450550644313,
      "tokens_seen": 1878654976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021760280842527584,
      "loss": 2.5103,
      "theoretical_loss": 3.4484403968403554,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021759277833500502,
      "loss": 2.361,
      "theoretical_loss": 3.448430243489762,
      "tokens_seen": 1878786048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002175827482447342,
      "loss": 2.5298,
      "theoretical_loss": 3.4484200905924953,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021757271815446338,
      "loss": 2.6147,
      "theoretical_loss": 3.448409938148521,
      "tokens_seen": 1878917120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021756268806419257,
      "loss": 2.5585,
      "theoretical_loss": 3.448399786157802,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021755265797392177,
      "loss": 2.6425,
      "theoretical_loss": 3.4483896346203027,
      "tokens_seen": 1879048192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021754262788365095,
      "loss": 2.6654,
      "theoretical_loss": 3.448379483535987,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021753259779338014,
      "loss": 2.5659,
      "theoretical_loss": 3.4483693329048184,
      "tokens_seen": 1879179264
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1034876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3631134033203125,
      "objective/train/theoretical_loss": 3.4483591827267617,
      "objective/train/tokens_used": 1899704800,
      "theoretical_loss": 3.4483591827267617,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021752256770310932,
      "loss": 2.6625,
      "theoretical_loss": 3.4483591827267617,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021751253761283852,
      "loss": 2.5979,
      "theoretical_loss": 3.4483490330017803,
      "tokens_seen": 1879310336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002175025075225677,
      "loss": 2.6531,
      "theoretical_loss": 3.4483388837298383,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174924774322969,
      "loss": 2.5533,
      "theoretical_loss": 3.4483287349108993,
      "tokens_seen": 1879441408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021748244734202607,
      "loss": 2.7961,
      "theoretical_loss": 3.448318586544928,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021747241725175525,
      "loss": 2.4689,
      "theoretical_loss": 3.448308438631888,
      "tokens_seen": 1879572480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021746238716148446,
      "loss": 2.4354,
      "theoretical_loss": 3.4482982911717435,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021745235707121364,
      "loss": 2.7338,
      "theoretical_loss": 3.4482881441644584,
      "tokens_seen": 1879703552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021744232698094282,
      "loss": 2.687,
      "theoretical_loss": 3.4482779976099964,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000217432296890672,
      "loss": 2.546,
      "theoretical_loss": 3.448267851508322,
      "tokens_seen": 1879834624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174222668004012,
      "loss": 2.5653,
      "theoretical_loss": 3.448257705859399,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174122367101304,
      "loss": 2.8596,
      "theoretical_loss": 3.4482475606631913,
      "tokens_seen": 1879965696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021740220661985957,
      "loss": 2.885,
      "theoretical_loss": 3.448237415919663,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021739217652958875,
      "loss": 2.6868,
      "theoretical_loss": 3.448227271628778,
      "tokens_seen": 1880096768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021738214643931793,
      "loss": 2.6006,
      "theoretical_loss": 3.4482171277905005,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021737211634904714,
      "loss": 2.4538,
      "theoretical_loss": 3.448206984404795,
      "tokens_seen": 1880227840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021736208625877632,
      "loss": 2.7134,
      "theoretical_loss": 3.448196841471624,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173520561685055,
      "loss": 2.4357,
      "theoretical_loss": 3.4481866989909533,
      "tokens_seen": 1880358912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173420260782347,
      "loss": 2.4989,
      "theoretical_loss": 3.4481765569627463,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021733199598796392,
      "loss": 2.5744,
      "theoretical_loss": 3.448166415386966,
      "tokens_seen": 1880489984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173219658976931,
      "loss": 2.5725,
      "theoretical_loss": 3.448156274263578,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021731193580742228,
      "loss": 2.3246,
      "theoretical_loss": 3.4481461335925463,
      "tokens_seen": 1880621056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021730190571715146,
      "loss": 2.5812,
      "theoretical_loss": 3.4481359933738336,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021729187562688065,
      "loss": 2.7822,
      "theoretical_loss": 3.4481258536074053,
      "tokens_seen": 1880752128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021728184553660985,
      "loss": 2.5844,
      "theoretical_loss": 3.448115714293225,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1035977,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.580977439880371,
      "objective/train/theoretical_loss": 3.4481055754312564,
      "objective/train/tokens_used": 1901343200,
      "theoretical_loss": 3.4481055754312564,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021727181544633903,
      "loss": 2.4114,
      "theoretical_loss": 3.4481055754312564,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021726178535606822,
      "loss": 2.5488,
      "theoretical_loss": 3.448095437021464,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172517552657974,
      "loss": 2.5698,
      "theoretical_loss": 3.448085299063812,
      "tokens_seen": 1881014272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172417251755266,
      "loss": 2.6378,
      "theoretical_loss": 3.448075161558264,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021723169508525579,
      "loss": 2.6272,
      "theoretical_loss": 3.448065024504784,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021722166499498497,
      "loss": 2.5578,
      "theoretical_loss": 3.448054887903337,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021721163490471415,
      "loss": 2.6344,
      "theoretical_loss": 3.448044751753886,
      "tokens_seen": 1881276416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021720160481444336,
      "loss": 2.4502,
      "theoretical_loss": 3.448034616056396,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021719157472417254,
      "loss": 2.6808,
      "theoretical_loss": 3.448024480810831,
      "tokens_seen": 1881407488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021718154463390172,
      "loss": 2.7059,
      "theoretical_loss": 3.4480143460171546,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171715145436309,
      "loss": 2.542,
      "theoretical_loss": 3.448004211675331,
      "tokens_seen": 1881538560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021716148445336008,
      "loss": 2.69,
      "theoretical_loss": 3.447994077785325,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171514543630893,
      "loss": 2.6481,
      "theoretical_loss": 3.4479839443471,
      "tokens_seen": 1881669632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021714142427281847,
      "loss": 2.7006,
      "theoretical_loss": 3.4479738113606198,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021713139418254765,
      "loss": 2.6281,
      "theoretical_loss": 3.4479636788258494,
      "tokens_seen": 1881800704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021712136409227683,
      "loss": 2.8042,
      "theoretical_loss": 3.4479535467427525,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021711133400200604,
      "loss": 2.4941,
      "theoretical_loss": 3.4479434151112933,
      "tokens_seen": 1881931776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021710130391173522,
      "loss": 2.6037,
      "theoretical_loss": 3.4479332839314365,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170912738214644,
      "loss": 2.6742,
      "theoretical_loss": 3.447923153203145,
      "tokens_seen": 1882062848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021708124373119358,
      "loss": 2.6185,
      "theoretical_loss": 3.447913022926384,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021707121364092277,
      "loss": 2.6322,
      "theoretical_loss": 3.4479028931011175,
      "tokens_seen": 1882193920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021706118355065197,
      "loss": 2.6673,
      "theoretical_loss": 3.4478927637273094,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021705115346038115,
      "loss": 2.5087,
      "theoretical_loss": 3.447882634804923,
      "tokens_seen": 1882324992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021704112337011034,
      "loss": 2.6426,
      "theoretical_loss": 3.4478725063339244,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021703109327983952,
      "loss": 2.8292,
      "theoretical_loss": 3.4478623783142766,
      "tokens_seen": 1882456064
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1036495,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.446045160293579,
      "objective/train/theoretical_loss": 3.447852250745944,
      "objective/train/tokens_used": 1902981600,
      "theoretical_loss": 3.447852250745944,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021702106318956873,
      "loss": 2.4661,
      "theoretical_loss": 3.447852250745944,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170110330992979,
      "loss": 2.6948,
      "theoretical_loss": 3.4478421236288908,
      "tokens_seen": 1882587136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170010030090271,
      "loss": 2.5946,
      "theoretical_loss": 3.447831996963081,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021699097291875627,
      "loss": 2.7897,
      "theoretical_loss": 3.447821870748479,
      "tokens_seen": 1882718208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021698094282848545,
      "loss": 2.6243,
      "theoretical_loss": 3.4478117449850485,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021697091273821466,
      "loss": 2.6638,
      "theoretical_loss": 3.447801619672754,
      "tokens_seen": 1882849280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021696088264794384,
      "loss": 2.5881,
      "theoretical_loss": 3.44779149481156,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021695085255767302,
      "loss": 2.5933,
      "theoretical_loss": 3.4477813704014313,
      "tokens_seen": 1882980352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002169408224674022,
      "loss": 2.7635,
      "theoretical_loss": 3.44777124644233,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002169307923771314,
      "loss": 2.5949,
      "theoretical_loss": 3.4477611229342227,
      "tokens_seen": 1883111424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002169207622868606,
      "loss": 2.636,
      "theoretical_loss": 3.447750999877072,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021691073219658977,
      "loss": 2.6428,
      "theoretical_loss": 3.447740877270842,
      "tokens_seen": 1883242496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021690070210631895,
      "loss": 2.475,
      "theoretical_loss": 3.4477307551154985,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021689067201604813,
      "loss": 2.7729,
      "theoretical_loss": 3.4477206334110044,
      "tokens_seen": 1883373568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021688064192577734,
      "loss": 2.639,
      "theoretical_loss": 3.4477105121573244,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021687061183550652,
      "loss": 2.4855,
      "theoretical_loss": 3.447700391354423,
      "tokens_seen": 1883504640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168605817452357,
      "loss": 2.8576,
      "theoretical_loss": 3.4476902710022634,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021685055165496489,
      "loss": 2.7365,
      "theoretical_loss": 3.447680151100811,
      "tokens_seen": 1883635712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168405215646941,
      "loss": 2.7015,
      "theoretical_loss": 3.4476700316500293,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021683049147442327,
      "loss": 2.641,
      "theoretical_loss": 3.447659912649883,
      "tokens_seen": 1883766784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021682046138415246,
      "loss": 2.6379,
      "theoretical_loss": 3.447649794100336,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021681043129388164,
      "loss": 2.4592,
      "theoretical_loss": 3.4476396760013532,
      "tokens_seen": 1883897856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021680040120361082,
      "loss": 2.8094,
      "theoretical_loss": 3.447629558352898,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021679037111334003,
      "loss": 2.9019,
      "theoretical_loss": 3.4476194411549352,
      "tokens_seen": 1884028928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167803410230692,
      "loss": 2.8158,
      "theoretical_loss": 3.447609324407429,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1037793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7135426998138428,
      "objective/train/theoretical_loss": 3.4475992081103435,
      "objective/train/tokens_used": 1904620000,
      "theoretical_loss": 3.4475992081103435,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167703109327984,
      "loss": 2.6869,
      "theoretical_loss": 3.4475992081103435,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021676028084252757,
      "loss": 2.5489,
      "theoretical_loss": 3.4475890922636436,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021675025075225678,
      "loss": 2.6551,
      "theoretical_loss": 3.447578976867293,
      "tokens_seen": 1884291072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021674022066198596,
      "loss": 2.8233,
      "theoretical_loss": 3.4475688619212557,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021673019057171514,
      "loss": 2.677,
      "theoretical_loss": 3.4475587474254965,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021672016048144432,
      "loss": 2.4684,
      "theoretical_loss": 3.4475486333799794,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167101303911735,
      "loss": 2.6675,
      "theoretical_loss": 3.4475385197846693,
      "tokens_seen": 1884553216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167001003009027,
      "loss": 2.7593,
      "theoretical_loss": 3.4475284066395293,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166900702106319,
      "loss": 2.7004,
      "theoretical_loss": 3.4475182939445252,
      "tokens_seen": 1884684288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021668004012036107,
      "loss": 2.5641,
      "theoretical_loss": 3.4475081816996207,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021667001003009025,
      "loss": 2.7787,
      "theoretical_loss": 3.4474980699047792,
      "tokens_seen": 1884815360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021665997993981946,
      "loss": 2.8079,
      "theoretical_loss": 3.4474879585599667,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021664994984954864,
      "loss": 2.75,
      "theoretical_loss": 3.447477847665146,
      "tokens_seen": 1884946432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021663991975927782,
      "loss": 2.7527,
      "theoretical_loss": 3.4474677372202827,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216629889669007,
      "loss": 2.5024,
      "theoretical_loss": 3.4474576272253405,
      "tokens_seen": 1885077504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021661985957873621,
      "loss": 2.6632,
      "theoretical_loss": 3.4474475176802835,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166098294884654,
      "loss": 2.5716,
      "theoretical_loss": 3.4474374085850763,
      "tokens_seen": 1885208576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021659979939819458,
      "loss": 2.4872,
      "theoretical_loss": 3.447427299939683,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021658976930792378,
      "loss": 2.7078,
      "theoretical_loss": 3.447417191744069,
      "tokens_seen": 1885339648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021657973921765297,
      "loss": 2.654,
      "theoretical_loss": 3.4474070839981974,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021656970912738217,
      "loss": 2.7711,
      "theoretical_loss": 3.447396976702033,
      "tokens_seen": 1885470720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021655967903711135,
      "loss": 2.5622,
      "theoretical_loss": 3.4473868698555403,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021654964894684054,
      "loss": 2.6452,
      "theoretical_loss": 3.447376763458683,
      "tokens_seen": 1885601792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021653961885656972,
      "loss": 2.5853,
      "theoretical_loss": 3.4473666575114263,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021652958876629893,
      "loss": 2.7104,
      "theoretical_loss": 3.447356552013735,
      "tokens_seen": 1885732864
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1038418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9850575923919678,
      "objective/train/theoretical_loss": 3.4473464469655717,
      "objective/train/tokens_used": 1906258400,
      "theoretical_loss": 3.4473464469655717,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165195586760281,
      "loss": 2.7462,
      "theoretical_loss": 3.4473464469655717,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165095285857573,
      "loss": 2.57,
      "theoretical_loss": 3.447336342366903,
      "tokens_seen": 1885863936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021649949849548647,
      "loss": 2.6572,
      "theoretical_loss": 3.4473262382176912,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021648946840521565,
      "loss": 2.6522,
      "theoretical_loss": 3.447316134517902,
      "tokens_seen": 1885995008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021647943831494486,
      "loss": 2.8715,
      "theoretical_loss": 3.447306031267499,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021646940822467404,
      "loss": 2.7221,
      "theoretical_loss": 3.4472959284664473,
      "tokens_seen": 1886126080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021645937813440322,
      "loss": 2.5299,
      "theoretical_loss": 3.4472858261147112,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002164493480441324,
      "loss": 2.6065,
      "theoretical_loss": 3.447275724212255,
      "tokens_seen": 1886257152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002164393179538616,
      "loss": 2.6699,
      "theoretical_loss": 3.4472656227590424,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002164292878635908,
      "loss": 2.7039,
      "theoretical_loss": 3.4472555217550394,
      "tokens_seen": 1886388224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021641925777331997,
      "loss": 2.6169,
      "theoretical_loss": 3.4472454212002086,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021640922768304915,
      "loss": 2.6877,
      "theoretical_loss": 3.4472353210945155,
      "tokens_seen": 1886519296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021639919759277833,
      "loss": 2.7127,
      "theoretical_loss": 3.4472252214379244,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021638916750250754,
      "loss": 2.6192,
      "theoretical_loss": 3.4472151222304,
      "tokens_seen": 1886650368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021637913741223672,
      "loss": 2.6279,
      "theoretical_loss": 3.447205023471906,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002163691073219659,
      "loss": 2.56,
      "theoretical_loss": 3.4471949251624077,
      "tokens_seen": 1886781440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021635907723169509,
      "loss": 2.5793,
      "theoretical_loss": 3.4471848273018684,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002163490471414243,
      "loss": 2.8797,
      "theoretical_loss": 3.4471747298902535,
      "tokens_seen": 1886912512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021633901705115347,
      "loss": 2.5334,
      "theoretical_loss": 3.447164632927527,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021632898696088266,
      "loss": 2.4944,
      "theoretical_loss": 3.447154536413654,
      "tokens_seen": 1887043584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021631895687061184,
      "loss": 2.6837,
      "theoretical_loss": 3.4471444403485982,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021630892678034102,
      "loss": 2.5929,
      "theoretical_loss": 3.447134344732324,
      "tokens_seen": 1887174656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021629889669007023,
      "loss": 2.6858,
      "theoretical_loss": 3.447124249564797,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162888665997994,
      "loss": 2.4521,
      "theoretical_loss": 3.44711415484598,
      "tokens_seen": 1887305728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162788365095286,
      "loss": 2.5735,
      "theoretical_loss": 3.447104060575839,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1039523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4777672290802,
      "objective/train/theoretical_loss": 3.4470939667543377,
      "objective/train/tokens_used": 1907896800,
      "theoretical_loss": 3.4470939667543377,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021626880641925777,
      "loss": 2.5574,
      "theoretical_loss": 3.4470939667543377,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021625877632898698,
      "loss": 2.6886,
      "theoretical_loss": 3.44708387338144,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021624874623871616,
      "loss": 2.547,
      "theoretical_loss": 3.447073780457112,
      "tokens_seen": 1887567872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021623871614844534,
      "loss": 2.6556,
      "theoretical_loss": 3.4470636879813172,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021622868605817452,
      "loss": 2.6823,
      "theoretical_loss": 3.44705359595402,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162186559679037,
      "loss": 2.5915,
      "theoretical_loss": 3.447043504375185,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162086258776329,
      "loss": 2.5486,
      "theoretical_loss": 3.447033413244777,
      "tokens_seen": 1887830016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002161985957873621,
      "loss": 2.8164,
      "theoretical_loss": 3.4470233225627602,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021618856569709127,
      "loss": 2.4256,
      "theoretical_loss": 3.447013232329099,
      "tokens_seen": 1887961088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021617853560682045,
      "loss": 2.637,
      "theoretical_loss": 3.4470031425437586,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021616850551654966,
      "loss": 2.5112,
      "theoretical_loss": 3.446993053206703,
      "tokens_seen": 1888092160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021615847542627884,
      "loss": 2.4893,
      "theoretical_loss": 3.4469829643178964,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021614844533600802,
      "loss": 2.5477,
      "theoretical_loss": 3.4469728758773037,
      "tokens_seen": 1888223232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002161384152457372,
      "loss": 2.7546,
      "theoretical_loss": 3.44696278788489,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021612838515546641,
      "loss": 2.6119,
      "theoretical_loss": 3.446952700340619,
      "tokens_seen": 1888354304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002161183550651956,
      "loss": 2.9151,
      "theoretical_loss": 3.446942613244455,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021610832497492478,
      "loss": 2.6117,
      "theoretical_loss": 3.4469325265963637,
      "tokens_seen": 1888485376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021609829488465396,
      "loss": 2.5982,
      "theoretical_loss": 3.4469224403963086,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021608826479438314,
      "loss": 2.3824,
      "theoretical_loss": 3.4469123546442546,
      "tokens_seen": 1888616448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021607823470411235,
      "loss": 2.57,
      "theoretical_loss": 3.4469022693401667,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021606820461384153,
      "loss": 2.8983,
      "theoretical_loss": 3.446892184484009,
      "tokens_seen": 1888747520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002160581745235707,
      "loss": 2.6141,
      "theoretical_loss": 3.4468821000757464,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002160481444332999,
      "loss": 2.6643,
      "theoretical_loss": 3.4468720161153428,
      "tokens_seen": 1888878592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002160381143430291,
      "loss": 2.4504,
      "theoretical_loss": 3.446861932602763,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021602808425275828,
      "loss": 2.5322,
      "theoretical_loss": 3.446851849537972,
      "tokens_seen": 1889009664
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1040960,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1707992553710938,
      "objective/train/theoretical_loss": 3.446841766920934,
      "objective/train/tokens_used": 1909535200,
      "theoretical_loss": 3.446841766920934,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021601805416248746,
      "loss": 2.3876,
      "theoretical_loss": 3.446841766920934,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021600802407221664,
      "loss": 2.7003,
      "theoretical_loss": 3.446831684751614,
      "tokens_seen": 1889140736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021599799398194582,
      "loss": 2.4766,
      "theoretical_loss": 3.446821603029976,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021598796389167503,
      "loss": 2.592,
      "theoretical_loss": 3.4468115217559854,
      "tokens_seen": 1889271808
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002159779338014042,
      "loss": 2.5899,
      "theoretical_loss": 3.446801440929606,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002159679037111334,
      "loss": 2.6825,
      "theoretical_loss": 3.4467913605508027,
      "tokens_seen": 1889402880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021595787362086257,
      "loss": 2.5726,
      "theoretical_loss": 3.4467812806195397,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021594784353059178,
      "loss": 2.7512,
      "theoretical_loss": 3.4467712011357827,
      "tokens_seen": 1889533952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021593781344032096,
      "loss": 2.4037,
      "theoretical_loss": 3.4467611220994954,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021592778335005014,
      "loss": 2.6689,
      "theoretical_loss": 3.446751043510642,
      "tokens_seen": 1889665024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021591775325977933,
      "loss": 2.5465,
      "theoretical_loss": 3.4467409653691887,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002159077231695085,
      "loss": 2.5668,
      "theoretical_loss": 3.446730887675099,
      "tokens_seen": 1889796096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021589769307923772,
      "loss": 2.5015,
      "theoretical_loss": 3.4467208104283373,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158876629889669,
      "loss": 2.6471,
      "theoretical_loss": 3.446710733628869,
      "tokens_seen": 1889927168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021587763289869608,
      "loss": 2.5036,
      "theoretical_loss": 3.446700657276658,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021586760280842526,
      "loss": 2.5077,
      "theoretical_loss": 3.4466905813716697,
      "tokens_seen": 1890058240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021585757271815447,
      "loss": 2.5352,
      "theoretical_loss": 3.446680505913868,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021584754262788365,
      "loss": 2.6593,
      "theoretical_loss": 3.446670430903218,
      "tokens_seen": 1890189312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021583751253761286,
      "loss": 2.6334,
      "theoretical_loss": 3.4466603563396845,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021582748244734204,
      "loss": 2.6864,
      "theoretical_loss": 3.4466502822232323,
      "tokens_seen": 1890320384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021581745235707122,
      "loss": 2.6918,
      "theoretical_loss": 3.4466402085538252,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021580742226680043,
      "loss": 2.6715,
      "theoretical_loss": 3.446630135331428,
      "tokens_seen": 1890451456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157973921765296,
      "loss": 2.5339,
      "theoretical_loss": 3.4466200625560064,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157873620862588,
      "loss": 2.4792,
      "theoretical_loss": 3.446609990227524,
      "tokens_seen": 1890582528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021577733199598797,
      "loss": 2.3752,
      "theoretical_loss": 3.446599918345946,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1041567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2605631351470947,
      "objective/train/theoretical_loss": 3.4465898469112366,
      "objective/train/tokens_used": 1911173600,
      "theoretical_loss": 3.4465898469112366,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021576730190571718,
      "loss": 2.6977,
      "theoretical_loss": 3.4465898469112366,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021575727181544636,
      "loss": 2.3716,
      "theoretical_loss": 3.446579775923362,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021574724172517554,
      "loss": 2.8488,
      "theoretical_loss": 3.4465697053822844,
      "tokens_seen": 1890844672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021573721163490472,
      "loss": 2.5274,
      "theoretical_loss": 3.4465596352879704,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157271815446339,
      "loss": 2.4824,
      "theoretical_loss": 3.446549565640384,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157171514543631,
      "loss": 2.4491,
      "theoretical_loss": 3.44653949643949,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157071213640923,
      "loss": 2.6868,
      "theoretical_loss": 3.4465294276852534,
      "tokens_seen": 1891106816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021569709127382147,
      "loss": 2.7121,
      "theoretical_loss": 3.446519359377638,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021568706118355065,
      "loss": 2.7958,
      "theoretical_loss": 3.44650929151661,
      "tokens_seen": 1891237888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021567703109327986,
      "loss": 2.6139,
      "theoretical_loss": 3.4464992241021326,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021566700100300904,
      "loss": 2.7612,
      "theoretical_loss": 3.446489157134171,
      "tokens_seen": 1891368960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021565697091273822,
      "loss": 2.4897,
      "theoretical_loss": 3.4464790906126908,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156469408224674,
      "loss": 2.6425,
      "theoretical_loss": 3.446469024537656,
      "tokens_seen": 1891500032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021563691073219661,
      "loss": 2.6331,
      "theoretical_loss": 3.446458958909031,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156268806419258,
      "loss": 2.5318,
      "theoretical_loss": 3.4464488937267808,
      "tokens_seen": 1891631104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021561685055165498,
      "loss": 2.4587,
      "theoretical_loss": 3.4464388289908703,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021560682046138416,
      "loss": 2.5659,
      "theoretical_loss": 3.4464287647012646,
      "tokens_seen": 1891762176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021559679037111334,
      "loss": 2.6235,
      "theoretical_loss": 3.4464187008579277,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021558676028084255,
      "loss": 2.5082,
      "theoretical_loss": 3.446408637460825,
      "tokens_seen": 1891893248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021557673019057173,
      "loss": 2.5527,
      "theoretical_loss": 3.446398574509921,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155667001003009,
      "loss": 2.6059,
      "theoretical_loss": 3.44638851200518,
      "tokens_seen": 1892024320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155566700100301,
      "loss": 2.6577,
      "theoretical_loss": 3.4463784499465673,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155466399197593,
      "loss": 2.6324,
      "theoretical_loss": 3.4463683883340472,
      "tokens_seen": 1892155392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021553660982948848,
      "loss": 2.6396,
      "theoretical_loss": 3.4463583271675855,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021552657973921766,
      "loss": 2.481,
      "theoretical_loss": 3.446348266447146,
      "tokens_seen": 1892286464
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1042675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5854508876800537,
      "objective/train/theoretical_loss": 3.446338206172694,
      "objective/train/tokens_used": 1912812000,
      "theoretical_loss": 3.446338206172694,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021551654964894684,
      "loss": 2.5146,
      "theoretical_loss": 3.446338206172694,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021550651955867602,
      "loss": 2.5615,
      "theoretical_loss": 3.446328146344194,
      "tokens_seen": 1892417536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021549648946840523,
      "loss": 2.8431,
      "theoretical_loss": 3.44631808696161,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154864593781344,
      "loss": 2.5296,
      "theoretical_loss": 3.4463080280249088,
      "tokens_seen": 1892548608
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154764292878636,
      "loss": 2.66,
      "theoretical_loss": 3.4462979695340534,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021546639919759277,
      "loss": 2.5784,
      "theoretical_loss": 3.4462879114890095,
      "tokens_seen": 1892679680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021545636910732198,
      "loss": 2.6825,
      "theoretical_loss": 3.446277853889742,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021544633901705116,
      "loss": 2.5628,
      "theoretical_loss": 3.4462677967362145,
      "tokens_seen": 1892810752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021543630892678034,
      "loss": 2.3711,
      "theoretical_loss": 3.4462577400283934,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021542627883650953,
      "loss": 2.7311,
      "theoretical_loss": 3.4462476837662424,
      "tokens_seen": 1892941824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154162487462387,
      "loss": 2.5232,
      "theoretical_loss": 3.446237627949727,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021540621865596792,
      "loss": 2.5391,
      "theoretical_loss": 3.4462275725788114,
      "tokens_seen": 1893072896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153961885656971,
      "loss": 2.6001,
      "theoretical_loss": 3.4462175176534613,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021538615847542628,
      "loss": 2.6553,
      "theoretical_loss": 3.44620746317364,
      "tokens_seen": 1893203968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021537612838515546,
      "loss": 2.8628,
      "theoretical_loss": 3.446197409139314,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021536609829488467,
      "loss": 2.6248,
      "theoretical_loss": 3.446187355550448,
      "tokens_seen": 1893335040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021535606820461385,
      "loss": 2.7511,
      "theoretical_loss": 3.4461773024070057,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021534603811434303,
      "loss": 2.5208,
      "theoretical_loss": 3.4461672497089526,
      "tokens_seen": 1893466112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153360080240722,
      "loss": 2.5562,
      "theoretical_loss": 3.4461571974562535,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153259779338014,
      "loss": 2.9379,
      "theoretical_loss": 3.4461471456488733,
      "tokens_seen": 1893597184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153159478435306,
      "loss": 2.4249,
      "theoretical_loss": 3.446137094286777,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021530591775325978,
      "loss": 2.8084,
      "theoretical_loss": 3.4461270433699296,
      "tokens_seen": 1893728256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021529588766298896,
      "loss": 2.4948,
      "theoretical_loss": 3.446116992898295,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021528585757271814,
      "loss": 2.5373,
      "theoretical_loss": 3.4461069428718387,
      "tokens_seen": 1893859328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021527582748244735,
      "loss": 2.6058,
      "theoretical_loss": 3.4460968932905263,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1043306,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.76495099067688,
      "objective/train/theoretical_loss": 3.4460868441543218,
      "objective/train/tokens_used": 1914450400,
      "theoretical_loss": 3.4460868441543218,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021526579739217653,
      "loss": 2.6374,
      "theoretical_loss": 3.4460868441543218,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002152557673019057,
      "loss": 2.7168,
      "theoretical_loss": 3.44607679546319,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002152457372116349,
      "loss": 2.5589,
      "theoretical_loss": 3.4460667472170963,
      "tokens_seen": 1894121472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021523570712136408,
      "loss": 2.5378,
      "theoretical_loss": 3.446056699416005,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021522567703109328,
      "loss": 2.687,
      "theoretical_loss": 3.446046652059882,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021521564694082247,
      "loss": 2.5753,
      "theoretical_loss": 3.4460366051486915,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021520561685055165,
      "loss": 2.6651,
      "theoretical_loss": 3.446026558682398,
      "tokens_seen": 1894383616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021519558676028083,
      "loss": 2.8259,
      "theoretical_loss": 3.446016512660967,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021518555667001004,
      "loss": 2.6779,
      "theoretical_loss": 3.4460064670843638,
      "tokens_seen": 1894514688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021517552657973922,
      "loss": 2.6159,
      "theoretical_loss": 3.445996421952552,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151654964894684,
      "loss": 2.8159,
      "theoretical_loss": 3.445986377265498,
      "tokens_seen": 1894645760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021515546639919758,
      "loss": 2.724,
      "theoretical_loss": 3.445976333023166,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021514543630892676,
      "loss": 2.5926,
      "theoretical_loss": 3.4459662892255203,
      "tokens_seen": 1894776832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021513540621865597,
      "loss": 2.85,
      "theoretical_loss": 3.4459562458725275,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021512537612838515,
      "loss": 2.5194,
      "theoretical_loss": 3.445946202964151,
      "tokens_seen": 1894907904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021511534603811433,
      "loss": 2.3289,
      "theoretical_loss": 3.4459361605003567,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151053159478435,
      "loss": 2.606,
      "theoretical_loss": 3.4459261184811085,
      "tokens_seen": 1895038976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021509528585757272,
      "loss": 2.7977,
      "theoretical_loss": 3.4459160769063724,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021508525576730193,
      "loss": 2.7471,
      "theoretical_loss": 3.445906035776113,
      "tokens_seen": 1895170048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150752256770311,
      "loss": 2.5026,
      "theoretical_loss": 3.4458959950902948,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150651955867603,
      "loss": 2.7625,
      "theoretical_loss": 3.445885954848884,
      "tokens_seen": 1895301120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150551654964895,
      "loss": 2.6651,
      "theoretical_loss": 3.445875915051844,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021504513540621868,
      "loss": 2.7261,
      "theoretical_loss": 3.4458658756991407,
      "tokens_seen": 1895432192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021503510531594786,
      "loss": 2.4207,
      "theoretical_loss": 3.445855836790739,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021502507522567704,
      "loss": 2.6863,
      "theoretical_loss": 3.4458457983266033,
      "tokens_seen": 1895563264
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1044447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.729426383972168,
      "objective/train/theoretical_loss": 3.4458357603066996,
      "objective/train/tokens_used": 1916088800,
      "theoretical_loss": 3.4458357603066996,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021501504513540622,
      "loss": 2.721,
      "theoretical_loss": 3.4458357603066996,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021500501504513543,
      "loss": 2.8457,
      "theoretical_loss": 3.445825722730992,
      "tokens_seen": 1895694336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149949849548646,
      "loss": 2.5188,
      "theoretical_loss": 3.4458156855994457,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149849548645938,
      "loss": 2.7322,
      "theoretical_loss": 3.4458056489120263,
      "tokens_seen": 1895825408
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021497492477432297,
      "loss": 2.5653,
      "theoretical_loss": 3.4457956126686975,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021496489468405218,
      "loss": 2.6462,
      "theoretical_loss": 3.445785576869426,
      "tokens_seen": 1895956480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021495486459378136,
      "loss": 2.6287,
      "theoretical_loss": 3.445775541514175,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021494483450351055,
      "loss": 2.3769,
      "theoretical_loss": 3.445765506602911,
      "tokens_seen": 1896087552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021493480441323973,
      "loss": 2.5482,
      "theoretical_loss": 3.4457554721355983,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149247743229689,
      "loss": 2.548,
      "theoretical_loss": 3.445745438112202,
      "tokens_seen": 1896218624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021491474423269812,
      "loss": 2.6268,
      "theoretical_loss": 3.4457354045326873,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149047141424273,
      "loss": 2.5102,
      "theoretical_loss": 3.4457253713970184,
      "tokens_seen": 1896349696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021489468405215648,
      "loss": 2.3446,
      "theoretical_loss": 3.4457153387051616,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021488465396188566,
      "loss": 2.4853,
      "theoretical_loss": 3.4457053064570813,
      "tokens_seen": 1896480768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021487462387161487,
      "loss": 2.4407,
      "theoretical_loss": 3.4456952746527425,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021486459378134405,
      "loss": 2.4478,
      "theoretical_loss": 3.44568524329211,
      "tokens_seen": 1896611840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021485456369107323,
      "loss": 2.4921,
      "theoretical_loss": 3.44567521237515,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148445336008024,
      "loss": 2.6853,
      "theoretical_loss": 3.445665181901826,
      "tokens_seen": 1896742912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148345035105316,
      "loss": 2.5532,
      "theoretical_loss": 3.4456551518721037,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148244734202608,
      "loss": 2.7782,
      "theoretical_loss": 3.4456451222859483,
      "tokens_seen": 1896873984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021481444332998998,
      "loss": 2.5192,
      "theoretical_loss": 3.4456350931433253,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021480441323971916,
      "loss": 2.4893,
      "theoretical_loss": 3.445625064444199,
      "tokens_seen": 1897005056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021479438314944834,
      "loss": 2.4844,
      "theoretical_loss": 3.445615036188534,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021478435305917755,
      "loss": 2.5211,
      "theoretical_loss": 3.445605008376297,
      "tokens_seen": 1897136128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021477432296890673,
      "loss": 2.4631,
      "theoretical_loss": 3.4455949810074515,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.57,
      "objective/train/docs_used": 1045087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9239680767059326,
      "objective/train/theoretical_loss": 3.445584954081964,
      "objective/train/tokens_used": 1917727200,
      "theoretical_loss": 3.445584954081964,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021476429287863591,
      "loss": 2.5801,
      "theoretical_loss": 3.445584954081964,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002147542627883651,
      "loss": 2.5212,
      "theoretical_loss": 3.4455749275997976,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021474423269809428,
      "loss": 2.5678,
      "theoretical_loss": 3.4455649015609193,
      "tokens_seen": 1897398272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021473420260782348,
      "loss": 2.686,
      "theoretical_loss": 3.4455548759652936,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021472417251755267,
      "loss": 2.6532,
      "theoretical_loss": 3.4455448508128854,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021471414242728185,
      "loss": 2.4101,
      "theoretical_loss": 3.4455348261036596,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021470411233701103,
      "loss": 2.6091,
      "theoretical_loss": 3.445524801837582,
      "tokens_seen": 1897660416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021469408224674024,
      "loss": 2.9209,
      "theoretical_loss": 3.4455147780146174,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021468405215646942,
      "loss": 2.5647,
      "theoretical_loss": 3.4455047546347304,
      "tokens_seen": 1897791488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002146740220661986,
      "loss": 2.6652,
      "theoretical_loss": 3.4454947316978863,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021466399197592778,
      "loss": 2.7885,
      "theoretical_loss": 3.445484709204051,
      "tokens_seen": 1897922560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021465396188565696,
      "loss": 2.6763,
      "theoretical_loss": 3.445474687153189,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021464393179538617,
      "loss": 2.6974,
      "theoretical_loss": 3.4454646655452654,
      "tokens_seen": 1898053632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021463390170511535,
      "loss": 2.8042,
      "theoretical_loss": 3.4454546443802454,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021462387161484453,
      "loss": 2.5861,
      "theoretical_loss": 3.4454446236580942,
      "tokens_seen": 1898184704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002146138415245737,
      "loss": 2.408,
      "theoretical_loss": 3.445434603378777,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021460381143430292,
      "loss": 2.3946,
      "theoretical_loss": 3.445424583542258,
      "tokens_seen": 1898315776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145937813440321,
      "loss": 2.7096,
      "theoretical_loss": 3.4454145641485043,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021458375125376128,
      "loss": 2.6681,
      "theoretical_loss": 3.445404545197479,
      "tokens_seen": 1898446848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021457372116349046,
      "loss": 2.6471,
      "theoretical_loss": 3.4453945266891486,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021456369107321964,
      "loss": 2.6153,
      "theoretical_loss": 3.4453845086234782,
      "tokens_seen": 1898577920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021455366098294885,
      "loss": 2.5885,
      "theoretical_loss": 3.4453744910004325,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021454363089267803,
      "loss": 2.4979,
      "theoretical_loss": 3.4453644738199767,
      "tokens_seen": 1898708992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021453360080240721,
      "loss": 2.5051,
      "theoretical_loss": 3.4453544570820758,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145235707121364,
      "loss": 2.4236,
      "theoretical_loss": 3.4453444407866956,
      "tokens_seen": 1898840064
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1046191,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.293760299682617,
      "objective/train/theoretical_loss": 3.4453344249338005,
      "objective/train/tokens_used": 1919365600,
      "theoretical_loss": 3.4453344249338005,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145135406218656,
      "loss": 2.7453,
      "theoretical_loss": 3.4453344249338005,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021450351053159479,
      "loss": 2.5151,
      "theoretical_loss": 3.445324409523356,
      "tokens_seen": 1898971136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021449348044132397,
      "loss": 2.3986,
      "theoretical_loss": 3.4453143945553277,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021448345035105315,
      "loss": 2.5672,
      "theoretical_loss": 3.4453043800296808,
      "tokens_seen": 1899102208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021447342026078236,
      "loss": 2.5297,
      "theoretical_loss": 3.4452943659463795,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021446339017051154,
      "loss": 2.6135,
      "theoretical_loss": 3.4452843523053898,
      "tokens_seen": 1899233280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021445336008024072,
      "loss": 2.748,
      "theoretical_loss": 3.445274339106677,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144433299899699,
      "loss": 2.6242,
      "theoretical_loss": 3.445264326350206,
      "tokens_seen": 1899364352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021443329989969908,
      "loss": 2.6692,
      "theoretical_loss": 3.4452543140359415,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144232698094283,
      "loss": 2.6706,
      "theoretical_loss": 3.44524430216385,
      "tokens_seen": 1899495424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021441323971915747,
      "loss": 2.5056,
      "theoretical_loss": 3.4452342907338958,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021440320962888665,
      "loss": 2.5756,
      "theoretical_loss": 3.445224279746044,
      "tokens_seen": 1899626496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021439317953861583,
      "loss": 2.66,
      "theoretical_loss": 3.4452142692002603,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021438314944834504,
      "loss": 2.7159,
      "theoretical_loss": 3.44520425909651,
      "tokens_seen": 1899757568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021437311935807422,
      "loss": 2.5147,
      "theoretical_loss": 3.445194249434758,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002143630892678034,
      "loss": 2.6546,
      "theoretical_loss": 3.44518424021497,
      "tokens_seen": 1899888640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021435305917753258,
      "loss": 2.8038,
      "theoretical_loss": 3.4451742314371105,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021434302908726176,
      "loss": 2.7171,
      "theoretical_loss": 3.4451642231011452,
      "tokens_seen": 1900019712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000214332998996991,
      "loss": 2.708,
      "theoretical_loss": 3.4451542152070393,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021432296890672018,
      "loss": 2.5758,
      "theoretical_loss": 3.4451442077547583,
      "tokens_seen": 1900150784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021431293881644936,
      "loss": 2.7232,
      "theoretical_loss": 3.445134200744267,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021430290872617854,
      "loss": 2.7815,
      "theoretical_loss": 3.44512419417553,
      "tokens_seen": 1900281856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021429287863590775,
      "loss": 2.6418,
      "theoretical_loss": 3.4451141880485148,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021428284854563693,
      "loss": 2.5061,
      "theoretical_loss": 3.4451041823631843,
      "tokens_seen": 1900412928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021427281845536611,
      "loss": 2.3891,
      "theoretical_loss": 3.445094177119505,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1047224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.621803045272827,
      "objective/train/theoretical_loss": 3.4450841723174426,
      "objective/train/tokens_used": 1921004000,
      "theoretical_loss": 3.4450841723174426,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002142627883650953,
      "loss": 2.3745,
      "theoretical_loss": 3.4450841723174426,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021425275827482448,
      "loss": 2.5442,
      "theoretical_loss": 3.4450741679569608,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021424272818455368,
      "loss": 2.6669,
      "theoretical_loss": 3.445064164038026,
      "tokens_seen": 1900675072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021423269809428287,
      "loss": 2.6103,
      "theoretical_loss": 3.4450541605606038,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021422266800401205,
      "loss": 2.4476,
      "theoretical_loss": 3.4450441575246584,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021421263791374123,
      "loss": 2.5737,
      "theoretical_loss": 3.4450341549301564,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021420260782347044,
      "loss": 2.4585,
      "theoretical_loss": 3.4450241527770613,
      "tokens_seen": 1900937216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021419257773319962,
      "loss": 2.5462,
      "theoretical_loss": 3.4450141510653403,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141825476429288,
      "loss": 2.7341,
      "theoretical_loss": 3.4450041497949577,
      "tokens_seen": 1901068288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021417251755265798,
      "loss": 2.5915,
      "theoretical_loss": 3.444994148965879,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021416248746238716,
      "loss": 2.699,
      "theoretical_loss": 3.4449841485780697,
      "tokens_seen": 1901199360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021415245737211637,
      "loss": 2.6337,
      "theoretical_loss": 3.444974148631495,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021414242728184555,
      "loss": 2.5859,
      "theoretical_loss": 3.4449641491261196,
      "tokens_seen": 1901330432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021413239719157473,
      "loss": 2.5106,
      "theoretical_loss": 3.4449541500619096,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141223671013039,
      "loss": 2.6398,
      "theoretical_loss": 3.4449441514388304,
      "tokens_seen": 1901461504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021411233701103312,
      "loss": 2.4553,
      "theoretical_loss": 3.4449341532568467,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141023069207623,
      "loss": 2.6247,
      "theoretical_loss": 3.4449241555159245,
      "tokens_seen": 1901592576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021409227683049148,
      "loss": 2.5529,
      "theoretical_loss": 3.4449141582160285,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021408224674022066,
      "loss": 2.5141,
      "theoretical_loss": 3.4449041613571247,
      "tokens_seen": 1901723648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021407221664994984,
      "loss": 2.4631,
      "theoretical_loss": 3.4448941649391784,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021406218655967905,
      "loss": 2.5254,
      "theoretical_loss": 3.444884168962154,
      "tokens_seen": 1901854720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021405215646940823,
      "loss": 2.5354,
      "theoretical_loss": 3.444874173426018,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021404212637913742,
      "loss": 2.8904,
      "theoretical_loss": 3.444864178330735,
      "tokens_seen": 1901985792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002140320962888666,
      "loss": 2.5657,
      "theoretical_loss": 3.444854183676271,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002140220661985958,
      "loss": 2.4487,
      "theoretical_loss": 3.444844189462591,
      "tokens_seen": 1902116864
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1047755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.837980270385742,
      "objective/train/theoretical_loss": 3.44483419568966,
      "objective/train/tokens_used": 1922642400,
      "theoretical_loss": 3.44483419568966,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021401203610832499,
      "loss": 2.4192,
      "theoretical_loss": 3.44483419568966,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021400200601805417,
      "loss": 2.4033,
      "theoretical_loss": 3.4448242023574447,
      "tokens_seen": 1902247936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021399197592778335,
      "loss": 2.6462,
      "theoretical_loss": 3.444814209465909,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021398194583751256,
      "loss": 2.69,
      "theoretical_loss": 3.444804217015019,
      "tokens_seen": 1902379008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021397191574724174,
      "loss": 2.7757,
      "theoretical_loss": 3.44479422500474,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021396188565697092,
      "loss": 2.4102,
      "theoretical_loss": 3.4447842334350374,
      "tokens_seen": 1902510080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002139518555667001,
      "loss": 2.5317,
      "theoretical_loss": 3.4447742423058765,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021394182547642928,
      "loss": 2.5864,
      "theoretical_loss": 3.4447642516172223,
      "tokens_seen": 1902641152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002139317953861585,
      "loss": 2.4731,
      "theoretical_loss": 3.4447542613690416,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021392176529588767,
      "loss": 2.5596,
      "theoretical_loss": 3.444744271561298,
      "tokens_seen": 1902772224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021391173520561685,
      "loss": 2.569,
      "theoretical_loss": 3.4447342821939584,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021390170511534603,
      "loss": 2.4859,
      "theoretical_loss": 3.4447242932669875,
      "tokens_seen": 1902903296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021389167502507524,
      "loss": 2.4988,
      "theoretical_loss": 3.4447143047803506,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021388164493480442,
      "loss": 2.3917,
      "theoretical_loss": 3.444704316734013,
      "tokens_seen": 1903034368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002138716148445336,
      "loss": 2.5758,
      "theoretical_loss": 3.444694329127941,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021386158475426278,
      "loss": 2.7334,
      "theoretical_loss": 3.4446843419621,
      "tokens_seen": 1903165440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021385155466399196,
      "loss": 2.5812,
      "theoretical_loss": 3.4446743552364545,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021384152457372117,
      "loss": 2.5153,
      "theoretical_loss": 3.4446643689509697,
      "tokens_seen": 1903296512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021383149448345035,
      "loss": 2.6954,
      "theoretical_loss": 3.444654383105613,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021382146439317954,
      "loss": 2.7721,
      "theoretical_loss": 3.4446443977003476,
      "tokens_seen": 1903427584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021381143430290872,
      "loss": 2.5136,
      "theoretical_loss": 3.44463441273514,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021380140421263792,
      "loss": 2.6257,
      "theoretical_loss": 3.444624428209956,
      "tokens_seen": 1903558656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137913741223671,
      "loss": 2.5915,
      "theoretical_loss": 3.4446144441247606,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137813440320963,
      "loss": 2.7248,
      "theoretical_loss": 3.4446044604795194,
      "tokens_seen": 1903689728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021377131394182547,
      "loss": 2.6102,
      "theoretical_loss": 3.4445944772741974,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1048883,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8418326377868652,
      "objective/train/theoretical_loss": 3.444584494508761,
      "objective/train/tokens_used": 1924280800,
      "theoretical_loss": 3.444584494508761,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021376128385155465,
      "loss": 2.6386,
      "theoretical_loss": 3.444584494508761,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021375125376128386,
      "loss": 2.6978,
      "theoretical_loss": 3.444574512183175,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021374122367101304,
      "loss": 2.5328,
      "theoretical_loss": 3.444564530297405,
      "tokens_seen": 1903951872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021373119358074222,
      "loss": 2.5151,
      "theoretical_loss": 3.444554548851416,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137211634904714,
      "loss": 2.6034,
      "theoretical_loss": 3.4445445678451745,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137111334002006,
      "loss": 2.5876,
      "theoretical_loss": 3.4445345872786453,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137011033099298,
      "loss": 2.5818,
      "theoretical_loss": 3.444524607151794,
      "tokens_seen": 1904214016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021369107321965897,
      "loss": 2.6437,
      "theoretical_loss": 3.444514627464587,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021368104312938815,
      "loss": 2.6276,
      "theoretical_loss": 3.444504648216988,
      "tokens_seen": 1904345088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021367101303911733,
      "loss": 2.7603,
      "theoretical_loss": 3.4444946694089635,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021366098294884654,
      "loss": 2.7212,
      "theoretical_loss": 3.4444846910404796,
      "tokens_seen": 1904476160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021365095285857572,
      "loss": 2.6578,
      "theoretical_loss": 3.444474713111501,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136409227683049,
      "loss": 2.7305,
      "theoretical_loss": 3.4444647356219935,
      "tokens_seen": 1904607232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021363089267803409,
      "loss": 2.6143,
      "theoretical_loss": 3.444454758571922,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136208625877633,
      "loss": 2.6471,
      "theoretical_loss": 3.4444447819612534,
      "tokens_seen": 1904738304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021361083249749247,
      "loss": 2.6583,
      "theoretical_loss": 3.444434805789952,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021360080240722166,
      "loss": 2.4471,
      "theoretical_loss": 3.444424830057984,
      "tokens_seen": 1904869376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021359077231695084,
      "loss": 2.7115,
      "theoretical_loss": 3.444414854765314,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021358074222668004,
      "loss": 2.6026,
      "theoretical_loss": 3.444404879911909,
      "tokens_seen": 1905000448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021357071213640925,
      "loss": 2.7438,
      "theoretical_loss": 3.4443949054977336,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021356068204613843,
      "loss": 2.488,
      "theoretical_loss": 3.4443849315227535,
      "tokens_seen": 1905131520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021355065195586762,
      "loss": 2.5602,
      "theoretical_loss": 3.444374957986934,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002135406218655968,
      "loss": 2.5924,
      "theoretical_loss": 3.444364984890241,
      "tokens_seen": 1905262592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000213530591775326,
      "loss": 2.5039,
      "theoretical_loss": 3.4443550122326405,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021352056168505519,
      "loss": 2.7195,
      "theoretical_loss": 3.444345040014097,
      "tokens_seen": 1905393664
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1049481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.779414176940918,
      "objective/train/theoretical_loss": 3.444335068234577,
      "objective/train/tokens_used": 1925919200,
      "theoretical_loss": 3.444335068234577,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021351053159478437,
      "loss": 2.6997,
      "theoretical_loss": 3.444335068234577,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021350050150451355,
      "loss": 2.5893,
      "theoretical_loss": 3.4443250968940458,
      "tokens_seen": 1905524736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021349047141424276,
      "loss": 2.4811,
      "theoretical_loss": 3.4443151259924685,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021348044132397194,
      "loss": 2.6594,
      "theoretical_loss": 3.444305155529811,
      "tokens_seen": 1905655808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021347041123370112,
      "loss": 2.5418,
      "theoretical_loss": 3.4442951855060393,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134603811434303,
      "loss": 2.7476,
      "theoretical_loss": 3.4442852159211186,
      "tokens_seen": 1905786880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021345035105315948,
      "loss": 2.5844,
      "theoretical_loss": 3.444275246775015,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134403209628887,
      "loss": 2.384,
      "theoretical_loss": 3.4442652780676926,
      "tokens_seen": 1905917952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021343029087261787,
      "loss": 2.6019,
      "theoretical_loss": 3.444255309799119,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021342026078234705,
      "loss": 2.5389,
      "theoretical_loss": 3.444245341969258,
      "tokens_seen": 1906049024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021341023069207623,
      "loss": 2.5822,
      "theoretical_loss": 3.4442353745780765,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021340020060180544,
      "loss": 2.6002,
      "theoretical_loss": 3.44422540762554,
      "tokens_seen": 1906180096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021339017051153462,
      "loss": 2.4877,
      "theoretical_loss": 3.444215441111613,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002133801404212638,
      "loss": 2.8639,
      "theoretical_loss": 3.444205475036262,
      "tokens_seen": 1906311168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021337011033099298,
      "loss": 2.6008,
      "theoretical_loss": 3.444195509399453,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021336008024072216,
      "loss": 2.6341,
      "theoretical_loss": 3.444185544201151,
      "tokens_seen": 1906442240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021335005015045137,
      "loss": 2.8661,
      "theoretical_loss": 3.444175579441322,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021334002006018055,
      "loss": 2.7532,
      "theoretical_loss": 3.444165615119931,
      "tokens_seen": 1906573312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021332998996990974,
      "loss": 2.668,
      "theoretical_loss": 3.4441556512369442,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021331995987963892,
      "loss": 2.7884,
      "theoretical_loss": 3.4441456877923273,
      "tokens_seen": 1906704384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021330992978936812,
      "loss": 2.6677,
      "theoretical_loss": 3.4441357247860456,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132998996990973,
      "loss": 2.7131,
      "theoretical_loss": 3.4441257622180648,
      "tokens_seen": 1906835456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132898696088265,
      "loss": 2.7602,
      "theoretical_loss": 3.4441158000883507,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021327983951855567,
      "loss": 2.6545,
      "theoretical_loss": 3.444105838396869,
      "tokens_seen": 1906966528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021326980942828485,
      "loss": 2.538,
      "theoretical_loss": 3.4440958771435852,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1050597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4539685249328613,
      "objective/train/theoretical_loss": 3.444085916328465,
      "objective/train/tokens_used": 1927557600,
      "theoretical_loss": 3.444085916328465,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021325977933801406,
      "loss": 2.4239,
      "theoretical_loss": 3.444085916328465,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021324974924774324,
      "loss": 2.576,
      "theoretical_loss": 3.444075955951474,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021323971915747242,
      "loss": 2.6504,
      "theoretical_loss": 3.4440659960125783,
      "tokens_seen": 1907228672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132296890672016,
      "loss": 2.4841,
      "theoretical_loss": 3.4440560365117427,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132196589769308,
      "loss": 2.6357,
      "theoretical_loss": 3.444046077448934,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021320962888666,
      "loss": 2.4913,
      "theoretical_loss": 3.4440361188241173,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021319959879638917,
      "loss": 2.7163,
      "theoretical_loss": 3.444026160637258,
      "tokens_seen": 1907490816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021318956870611835,
      "loss": 2.4764,
      "theoretical_loss": 3.4440162028883226,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021317953861584753,
      "loss": 2.7076,
      "theoretical_loss": 3.444006245577276,
      "tokens_seen": 1907621888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021316950852557674,
      "loss": 2.5148,
      "theoretical_loss": 3.443996288704084,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021315947843530592,
      "loss": 2.7165,
      "theoretical_loss": 3.443986332268713,
      "tokens_seen": 1907752960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002131494483450351,
      "loss": 2.6934,
      "theoretical_loss": 3.443976376271128,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021313941825476429,
      "loss": 2.5539,
      "theoretical_loss": 3.4439664207112948,
      "tokens_seen": 1907884032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002131293881644935,
      "loss": 2.649,
      "theoretical_loss": 3.4439564655891797,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021311935807422267,
      "loss": 2.6691,
      "theoretical_loss": 3.4439465109047473,
      "tokens_seen": 1908015104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021310932798395186,
      "loss": 2.4946,
      "theoretical_loss": 3.4439365566579645,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021309929789368104,
      "loss": 2.4822,
      "theoretical_loss": 3.4439266028487965,
      "tokens_seen": 1908146176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021308926780341022,
      "loss": 2.856,
      "theoretical_loss": 3.4439166494772087,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021307923771313943,
      "loss": 2.4228,
      "theoretical_loss": 3.4439066965431677,
      "tokens_seen": 1908277248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002130692076228686,
      "loss": 2.5725,
      "theoretical_loss": 3.4438967440466386,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002130591775325978,
      "loss": 2.8128,
      "theoretical_loss": 3.443886791987587,
      "tokens_seen": 1908408320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021304914744232697,
      "loss": 2.7004,
      "theoretical_loss": 3.443876840365979,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021303911735205618,
      "loss": 2.6022,
      "theoretical_loss": 3.4438668891817805,
      "tokens_seen": 1908539392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021302908726178536,
      "loss": 2.4233,
      "theoretical_loss": 3.443856938434957,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021301905717151454,
      "loss": 2.613,
      "theoretical_loss": 3.4438469881254745,
      "tokens_seen": 1908670464
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1051023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3824679851531982,
      "objective/train/theoretical_loss": 3.443837038253298,
      "objective/train/tokens_used": 1929196000,
      "theoretical_loss": 3.443837038253298,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021300902708124372,
      "loss": 2.6133,
      "theoretical_loss": 3.443837038253298,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129989969909729,
      "loss": 2.6856,
      "theoretical_loss": 3.443827088818394,
      "tokens_seen": 1908801536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129889669007021,
      "loss": 2.6421,
      "theoretical_loss": 3.4438171398207285,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129789368104313,
      "loss": 2.3945,
      "theoretical_loss": 3.443807191260267,
      "tokens_seen": 1908932608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021296890672016047,
      "loss": 2.6683,
      "theoretical_loss": 3.4437972431369746,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021295887662988965,
      "loss": 2.4986,
      "theoretical_loss": 3.4437872954508175,
      "tokens_seen": 1909063680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021294884653961886,
      "loss": 2.7,
      "theoretical_loss": 3.443777348201762,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021293881644934804,
      "loss": 2.6368,
      "theoretical_loss": 3.443767401389773,
      "tokens_seen": 1909194752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021292878635907722,
      "loss": 2.612,
      "theoretical_loss": 3.4437574550148176,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129187562688064,
      "loss": 2.5467,
      "theoretical_loss": 3.4437475090768603,
      "tokens_seen": 1909325824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002129087261785356,
      "loss": 2.6909,
      "theoretical_loss": 3.4437375635758674,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128986960882648,
      "loss": 2.5512,
      "theoretical_loss": 3.4437276185118053,
      "tokens_seen": 1909456896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021288866599799398,
      "loss": 2.5882,
      "theoretical_loss": 3.4437176738846382,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021287863590772316,
      "loss": 2.6411,
      "theoretical_loss": 3.4437077296943337,
      "tokens_seen": 1909587968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021286860581745234,
      "loss": 2.4097,
      "theoretical_loss": 3.4436977859408566,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021285857572718155,
      "loss": 2.3294,
      "theoretical_loss": 3.443687842624173,
      "tokens_seen": 1909719040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021284854563691073,
      "loss": 2.6796,
      "theoretical_loss": 3.443677899744249,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128385155466399,
      "loss": 2.4161,
      "theoretical_loss": 3.4436679573010496,
      "tokens_seen": 1909850112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021282848545636912,
      "loss": 2.5694,
      "theoretical_loss": 3.4436580152945413,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021281845536609832,
      "loss": 2.3887,
      "theoretical_loss": 3.44364807372469,
      "tokens_seen": 1909981184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128084252758275,
      "loss": 2.6084,
      "theoretical_loss": 3.4436381325914613,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127983951855567,
      "loss": 2.842,
      "theoretical_loss": 3.443628191894821,
      "tokens_seen": 1910112256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021278836509528587,
      "loss": 2.5293,
      "theoretical_loss": 3.443618251634735,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021277833500501505,
      "loss": 2.65,
      "theoretical_loss": 3.443608311811169,
      "tokens_seen": 1910243328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021276830491474426,
      "loss": 2.6,
      "theoretical_loss": 3.443598372424089,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1051849,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.797550916671753,
      "objective/train/theoretical_loss": 3.443588433473461,
      "objective/train/tokens_used": 1930834400,
      "theoretical_loss": 3.443588433473461,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021275827482447344,
      "loss": 2.6537,
      "theoretical_loss": 3.443588433473461,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021274824473420262,
      "loss": 2.5941,
      "theoretical_loss": 3.443578494959251,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127382146439318,
      "loss": 2.417,
      "theoretical_loss": 3.4435685568814245,
      "tokens_seen": 1910505472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212728184553661,
      "loss": 2.5743,
      "theoretical_loss": 3.4435586192399477,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127181544633902,
      "loss": 2.6669,
      "theoretical_loss": 3.443548682034786,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021270812437311937,
      "loss": 2.5211,
      "theoretical_loss": 3.443538745265905,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021269809428284855,
      "loss": 2.642,
      "theoretical_loss": 3.443528808933272,
      "tokens_seen": 1910767616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021268806419257773,
      "loss": 2.4986,
      "theoretical_loss": 3.443518873036852,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021267803410230694,
      "loss": 2.6172,
      "theoretical_loss": 3.44350893757661,
      "tokens_seen": 1910898688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021266800401203612,
      "loss": 2.5106,
      "theoretical_loss": 3.443499002552514,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002126579739217653,
      "loss": 2.6558,
      "theoretical_loss": 3.443489067964528,
      "tokens_seen": 1911029760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021264794383149449,
      "loss": 2.6866,
      "theoretical_loss": 3.4434791338126187,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002126379137412237,
      "loss": 2.6906,
      "theoretical_loss": 3.443469200096752,
      "tokens_seen": 1911160832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021262788365095287,
      "loss": 2.516,
      "theoretical_loss": 3.4434592668168933,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021261785356068206,
      "loss": 2.7298,
      "theoretical_loss": 3.4434493339730095,
      "tokens_seen": 1911291904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021260782347041124,
      "loss": 2.5943,
      "theoretical_loss": 3.4434394015650653,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021259779338014042,
      "loss": 2.6986,
      "theoretical_loss": 3.4434294695930276,
      "tokens_seen": 1911422976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021258776328986963,
      "loss": 2.6512,
      "theoretical_loss": 3.443419538056862,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002125777331995988,
      "loss": 2.8703,
      "theoretical_loss": 3.4434096069565348,
      "tokens_seen": 1911554048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212567703109328,
      "loss": 2.6197,
      "theoretical_loss": 3.4433996762920107,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021255767301905717,
      "loss": 2.595,
      "theoretical_loss": 3.443389746063257,
      "tokens_seen": 1911685120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021254764292878638,
      "loss": 2.8144,
      "theoretical_loss": 3.443379816270239,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021253761283851556,
      "loss": 2.6883,
      "theoretical_loss": 3.4433698869129232,
      "tokens_seen": 1911816192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021252758274824474,
      "loss": 2.5334,
      "theoretical_loss": 3.443359957991275,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021251755265797392,
      "loss": 2.6748,
      "theoretical_loss": 3.4433500295052597,
      "tokens_seen": 1911947264
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1052477,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.106308698654175,
      "objective/train/theoretical_loss": 3.4433401014548446,
      "objective/train/tokens_used": 1932472800,
      "theoretical_loss": 3.4433401014548446,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002125075225677031,
      "loss": 2.8863,
      "theoretical_loss": 3.4433401014548446,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124974924774323,
      "loss": 2.5204,
      "theoretical_loss": 3.4433301738399953,
      "tokens_seen": 1912078336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124874623871615,
      "loss": 2.4401,
      "theoretical_loss": 3.4433202466606767,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021247743229689067,
      "loss": 2.6787,
      "theoretical_loss": 3.4433103199168564,
      "tokens_seen": 1912209408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021246740220661985,
      "loss": 2.6598,
      "theoretical_loss": 3.4433003936084994,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021245737211634906,
      "loss": 2.7497,
      "theoretical_loss": 3.443290467735572,
      "tokens_seen": 1912340480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021244734202607824,
      "loss": 2.6572,
      "theoretical_loss": 3.44328054229804,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021243731193580742,
      "loss": 2.5773,
      "theoretical_loss": 3.4432706172958687,
      "tokens_seen": 1912471552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124272818455366,
      "loss": 2.5005,
      "theoretical_loss": 3.4432606927290257,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124172517552658,
      "loss": 2.4448,
      "theoretical_loss": 3.4432507685974754,
      "tokens_seen": 1912602624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212407221664995,
      "loss": 2.6119,
      "theoretical_loss": 3.443240844901185,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021239719157472418,
      "loss": 2.5317,
      "theoretical_loss": 3.4432309216401196,
      "tokens_seen": 1912733696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021238716148445336,
      "loss": 2.6771,
      "theoretical_loss": 3.4432209988142457,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021237713139418254,
      "loss": 2.4145,
      "theoretical_loss": 3.4432110764235295,
      "tokens_seen": 1912864768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021236710130391175,
      "loss": 2.7353,
      "theoretical_loss": 3.4432011544679364,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021235707121364093,
      "loss": 2.6876,
      "theoretical_loss": 3.443191232947433,
      "tokens_seen": 1912995840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123470411233701,
      "loss": 2.5954,
      "theoretical_loss": 3.4431813118619847,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123370110330993,
      "loss": 2.5936,
      "theoretical_loss": 3.443171391211558,
      "tokens_seen": 1913126912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123269809428285,
      "loss": 2.7314,
      "theoretical_loss": 3.4431614709961185,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021231695085255768,
      "loss": 2.7413,
      "theoretical_loss": 3.4431515512156325,
      "tokens_seen": 1913257984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021230692076228686,
      "loss": 2.4924,
      "theoretical_loss": 3.4431416318700663,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021229689067201604,
      "loss": 2.5894,
      "theoretical_loss": 3.443131712959386,
      "tokens_seen": 1913389056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021228686058174522,
      "loss": 2.5977,
      "theoretical_loss": 3.443121794483557,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021227683049147443,
      "loss": 2.7105,
      "theoretical_loss": 3.4431118764425452,
      "tokens_seen": 1913520128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122668004012036,
      "loss": 2.6165,
      "theoretical_loss": 3.4431019588363174,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1053822,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.537475109100342,
      "objective/train/theoretical_loss": 3.4430920416648396,
      "objective/train/tokens_used": 1934111200,
      "theoretical_loss": 3.4430920416648396,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122567703109328,
      "loss": 2.4024,
      "theoretical_loss": 3.4430920416648396,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021224674022066197,
      "loss": 2.6878,
      "theoretical_loss": 3.443082124928077,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021223671013039118,
      "loss": 2.4149,
      "theoretical_loss": 3.443072208625997,
      "tokens_seen": 1913782272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021222668004012036,
      "loss": 2.6011,
      "theoretical_loss": 3.4430622927585643,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021221664994984954,
      "loss": 2.5171,
      "theoretical_loss": 3.443052377325746,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021220661985957873,
      "loss": 2.6747,
      "theoretical_loss": 3.4430424623275075,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121965897693079,
      "loss": 2.8489,
      "theoretical_loss": 3.4430325477638153,
      "tokens_seen": 1914044416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021218655967903712,
      "loss": 2.5809,
      "theoretical_loss": 3.443022633634635,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121765295887663,
      "loss": 2.5392,
      "theoretical_loss": 3.4430127199399334,
      "tokens_seen": 1914175488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021216649949849548,
      "loss": 2.5682,
      "theoretical_loss": 3.443002806679676,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021215646940822466,
      "loss": 2.3198,
      "theoretical_loss": 3.442992893853829,
      "tokens_seen": 1914306560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021214643931795387,
      "loss": 2.506,
      "theoretical_loss": 3.4429829814623583,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021213640922768305,
      "loss": 2.5054,
      "theoretical_loss": 3.4429730695052303,
      "tokens_seen": 1914437632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021212637913741223,
      "loss": 2.6683,
      "theoretical_loss": 3.442963157982412,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121163490471414,
      "loss": 2.4593,
      "theoretical_loss": 3.4429532468938673,
      "tokens_seen": 1914568704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121063189568706,
      "loss": 2.8143,
      "theoretical_loss": 3.4429433362395643,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120962888665998,
      "loss": 2.758,
      "theoretical_loss": 3.442933426019468,
      "tokens_seen": 1914699776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212086258776329,
      "loss": 2.473,
      "theoretical_loss": 3.442923516233545,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120762286860582,
      "loss": 2.2922,
      "theoretical_loss": 3.4429136068817616,
      "tokens_seen": 1914830848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021206619859578737,
      "loss": 2.5595,
      "theoretical_loss": 3.4429036979640832,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021205616850551658,
      "loss": 2.8599,
      "theoretical_loss": 3.4428937894804763,
      "tokens_seen": 1914961920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021204613841524576,
      "loss": 2.5304,
      "theoretical_loss": 3.442883881430908,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021203610832497494,
      "loss": 2.4018,
      "theoretical_loss": 3.4428739738153427,
      "tokens_seen": 1915092992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021202607823470412,
      "loss": 2.5903,
      "theoretical_loss": 3.442864066633747,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120160481444333,
      "loss": 2.6742,
      "theoretical_loss": 3.442854159886088,
      "tokens_seen": 1915224064
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1054329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7224864959716797,
      "objective/train/theoretical_loss": 3.442844253572331,
      "objective/train/tokens_used": 1935749600,
      "theoretical_loss": 3.442844253572331,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120060180541625,
      "loss": 2.7346,
      "theoretical_loss": 3.442844253572331,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119959879638917,
      "loss": 2.5472,
      "theoretical_loss": 3.4428343476924432,
      "tokens_seen": 1915355136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021198595787362087,
      "loss": 2.7368,
      "theoretical_loss": 3.4428244422463887,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021197592778335005,
      "loss": 2.39,
      "theoretical_loss": 3.4428145372341357,
      "tokens_seen": 1915486208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021196589769307926,
      "loss": 2.5773,
      "theoretical_loss": 3.4428046326556494,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021195586760280844,
      "loss": 2.7433,
      "theoretical_loss": 3.442794728510896,
      "tokens_seen": 1915617280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021194583751253762,
      "loss": 2.2994,
      "theoretical_loss": 3.4427848247998423,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119358074222668,
      "loss": 2.6521,
      "theoretical_loss": 3.4427749215224535,
      "tokens_seen": 1915748352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211925777331996,
      "loss": 2.5983,
      "theoretical_loss": 3.4427650186786964,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002119157472417252,
      "loss": 2.5268,
      "theoretical_loss": 3.442755116268537,
      "tokens_seen": 1915879424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021190571715145438,
      "loss": 2.6755,
      "theoretical_loss": 3.4427452142919415,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021189568706118356,
      "loss": 2.5451,
      "theoretical_loss": 3.442735312748876,
      "tokens_seen": 1916010496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021188565697091274,
      "loss": 2.6053,
      "theoretical_loss": 3.442725411639307,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021187562688064195,
      "loss": 2.6002,
      "theoretical_loss": 3.4427155109632004,
      "tokens_seen": 1916141568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021186559679037113,
      "loss": 2.6922,
      "theoretical_loss": 3.4427056107205223,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118555667001003,
      "loss": 2.6037,
      "theoretical_loss": 3.4426957109112393,
      "tokens_seen": 1916272640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118455366098295,
      "loss": 2.437,
      "theoretical_loss": 3.442685811535317,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118355065195587,
      "loss": 2.6205,
      "theoretical_loss": 3.4426759125927227,
      "tokens_seen": 1916403712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021182547642928788,
      "loss": 2.778,
      "theoretical_loss": 3.4426660140834215,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021181544633901706,
      "loss": 2.5896,
      "theoretical_loss": 3.44265611600738,
      "tokens_seen": 1916534784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021180541624874624,
      "loss": 2.6244,
      "theoretical_loss": 3.442646218364565,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021179538615847542,
      "loss": 2.6638,
      "theoretical_loss": 3.442636321154942,
      "tokens_seen": 1916665856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021178535606820463,
      "loss": 2.3675,
      "theoretical_loss": 3.442626424378477,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117753259779338,
      "loss": 2.6128,
      "theoretical_loss": 3.4426165280351366,
      "tokens_seen": 1916796928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211765295887663,
      "loss": 2.6357,
      "theoretical_loss": 3.4426066321248876,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1055508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6412904262542725,
      "objective/train/theoretical_loss": 3.4425967366476957,
      "objective/train/tokens_used": 1937388000,
      "theoretical_loss": 3.4425967366476957,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021175526579739217,
      "loss": 2.5392,
      "theoretical_loss": 3.4425967366476957,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021174523570712138,
      "loss": 2.5108,
      "theoretical_loss": 3.442586841603527,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021173520561685056,
      "loss": 2.4007,
      "theoretical_loss": 3.4425769469923475,
      "tokens_seen": 1917059072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021172517552657974,
      "loss": 2.7113,
      "theoretical_loss": 3.442567052814124,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021171514543630893,
      "loss": 2.3623,
      "theoretical_loss": 3.4425571590688233,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117051153460381,
      "loss": 2.4786,
      "theoretical_loss": 3.442547265756411,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021169508525576732,
      "loss": 2.4219,
      "theoretical_loss": 3.4425373728768527,
      "tokens_seen": 1917321216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116850551654965,
      "loss": 2.3498,
      "theoretical_loss": 3.4425274804301154,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021167502507522568,
      "loss": 2.6909,
      "theoretical_loss": 3.442517588416165,
      "tokens_seen": 1917452288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021166499498495486,
      "loss": 2.6458,
      "theoretical_loss": 3.442507696834969,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021165496489468407,
      "loss": 2.4792,
      "theoretical_loss": 3.442497805686492,
      "tokens_seen": 1917583360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021164493480441325,
      "loss": 2.5765,
      "theoretical_loss": 3.4424879149707017,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021163490471414243,
      "loss": 2.6024,
      "theoretical_loss": 3.442478024687563,
      "tokens_seen": 1917714432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116248746238716,
      "loss": 2.7212,
      "theoretical_loss": 3.442468134837043,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116148445336008,
      "loss": 2.4751,
      "theoretical_loss": 3.4424582454191084,
      "tokens_seen": 1917845504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021160481444333,
      "loss": 2.7383,
      "theoretical_loss": 3.4424483564337245,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021159478435305918,
      "loss": 2.7107,
      "theoretical_loss": 3.4424384678808586,
      "tokens_seen": 1917976576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021158475426278836,
      "loss": 2.6668,
      "theoretical_loss": 3.4424285797604757,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021157472417251754,
      "loss": 2.5299,
      "theoretical_loss": 3.4424186920725437,
      "tokens_seen": 1918107648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021156469408224675,
      "loss": 2.5064,
      "theoretical_loss": 3.4424088048170276,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021155466399197593,
      "loss": 2.4103,
      "theoretical_loss": 3.4423989179938945,
      "tokens_seen": 1918238720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115446339017051,
      "loss": 2.575,
      "theoretical_loss": 3.4423890316031107,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115346038114343,
      "loss": 2.7556,
      "theoretical_loss": 3.442379145644642,
      "tokens_seen": 1918369792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021152457372116348,
      "loss": 2.6002,
      "theoretical_loss": 3.442369260118455,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021151454363089268,
      "loss": 2.5512,
      "theoretical_loss": 3.442359375024516,
      "tokens_seen": 1918500864
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1056802,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.478919267654419,
      "objective/train/theoretical_loss": 3.4423494903627914,
      "objective/train/tokens_used": 1939026400,
      "theoretical_loss": 3.4423494903627914,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021150451354062186,
      "loss": 2.5599,
      "theoretical_loss": 3.4423494903627914,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021149448345035105,
      "loss": 2.6486,
      "theoretical_loss": 3.442339606133247,
      "tokens_seen": 1918631936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021148445336008023,
      "loss": 2.5275,
      "theoretical_loss": 3.4423297223358507,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021147442326980944,
      "loss": 2.6358,
      "theoretical_loss": 3.442319838970567,
      "tokens_seen": 1918763008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021146439317953862,
      "loss": 2.4775,
      "theoretical_loss": 3.4423099560373633,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002114543630892678,
      "loss": 2.637,
      "theoretical_loss": 3.442300073536206,
      "tokens_seen": 1918894080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021144433299899698,
      "loss": 2.656,
      "theoretical_loss": 3.4422901914670607,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021143430290872616,
      "loss": 2.372,
      "theoretical_loss": 3.4422803098298944,
      "tokens_seen": 1919025152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021142427281845537,
      "loss": 2.7808,
      "theoretical_loss": 3.4422704286246733,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021141424272818455,
      "loss": 2.715,
      "theoretical_loss": 3.442260547851364,
      "tokens_seen": 1919156224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021140421263791373,
      "loss": 2.9344,
      "theoretical_loss": 3.4422506675099322,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113941825476429,
      "loss": 2.6971,
      "theoretical_loss": 3.442240787600345,
      "tokens_seen": 1919287296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021138415245737212,
      "loss": 2.5248,
      "theoretical_loss": 3.4422309081225686,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113741223671013,
      "loss": 2.6926,
      "theoretical_loss": 3.4422210290765687,
      "tokens_seen": 1919418368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021136409227683048,
      "loss": 2.3141,
      "theoretical_loss": 3.442211150462313,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021135406218655966,
      "loss": 2.5164,
      "theoretical_loss": 3.4422012722797666,
      "tokens_seen": 1919549440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021134403209628884,
      "loss": 2.7235,
      "theoretical_loss": 3.442191394528897,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021133400200601808,
      "loss": 2.5599,
      "theoretical_loss": 3.4421815172096695,
      "tokens_seen": 1919680512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021132397191574726,
      "loss": 2.5132,
      "theoretical_loss": 3.4421716403220515,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021131394182547644,
      "loss": 2.4835,
      "theoretical_loss": 3.4421617638660083,
      "tokens_seen": 1919811584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021130391173520562,
      "loss": 2.5466,
      "theoretical_loss": 3.442151887841508,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021129388164493483,
      "loss": 2.6654,
      "theoretical_loss": 3.442142012248515,
      "tokens_seen": 1919942656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211283851554664,
      "loss": 2.6047,
      "theoretical_loss": 3.442132137086997,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112738214643932,
      "loss": 2.6365,
      "theoretical_loss": 3.4421222623569205,
      "tokens_seen": 1920073728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021126379137412237,
      "loss": 2.6211,
      "theoretical_loss": 3.442112388058251,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1057447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.344655990600586,
      "objective/train/theoretical_loss": 3.442102514190956,
      "objective/train/tokens_used": 1940664800,
      "theoretical_loss": 3.442102514190956,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021125376128385158,
      "loss": 2.4843,
      "theoretical_loss": 3.442102514190956,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021124373119358076,
      "loss": 2.5984,
      "theoretical_loss": 3.4420926407550008,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021123370110330994,
      "loss": 2.7614,
      "theoretical_loss": 3.442082767750353,
      "tokens_seen": 1920335872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021122367101303913,
      "loss": 2.4634,
      "theoretical_loss": 3.4420728951769783,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112136409227683,
      "loss": 2.643,
      "theoretical_loss": 3.4420630230348435,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021120361083249752,
      "loss": 2.5298,
      "theoretical_loss": 3.4420531513239148,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111935807422267,
      "loss": 2.5568,
      "theoretical_loss": 3.4420432800441585,
      "tokens_seen": 1920598016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021118355065195588,
      "loss": 2.6299,
      "theoretical_loss": 3.4420334091955413,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021117352056168506,
      "loss": 2.4657,
      "theoretical_loss": 3.44202353877803,
      "tokens_seen": 1920729088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021116349047141427,
      "loss": 2.5842,
      "theoretical_loss": 3.4420136687915903,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021115346038114345,
      "loss": 2.5533,
      "theoretical_loss": 3.442003799236189,
      "tokens_seen": 1920860160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021114343029087263,
      "loss": 2.576,
      "theoretical_loss": 3.441993930111793,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111334002006018,
      "loss": 2.6868,
      "theoretical_loss": 3.4419840614183688,
      "tokens_seen": 1920991232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211123370110331,
      "loss": 2.5897,
      "theoretical_loss": 3.4419741931558816,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111133400200602,
      "loss": 2.6805,
      "theoretical_loss": 3.4419643253242995,
      "tokens_seen": 1921122304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021110330992978938,
      "loss": 2.5411,
      "theoretical_loss": 3.441954457923588,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021109327983951856,
      "loss": 2.5456,
      "theoretical_loss": 3.441944590953714,
      "tokens_seen": 1921253376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021108324974924774,
      "loss": 2.4021,
      "theoretical_loss": 3.4419347244146437,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021107321965897695,
      "loss": 2.396,
      "theoretical_loss": 3.4419248583063435,
      "tokens_seen": 1921384448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021106318956870613,
      "loss": 2.5061,
      "theoretical_loss": 3.4419149926287806,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002110531594784353,
      "loss": 2.4764,
      "theoretical_loss": 3.441905127381921,
      "tokens_seen": 1921515520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002110431293881645,
      "loss": 2.5298,
      "theoretical_loss": 3.4418952625657306,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021103309929789368,
      "loss": 2.6697,
      "theoretical_loss": 3.4418853981801774,
      "tokens_seen": 1921646592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021102306920762288,
      "loss": 2.6126,
      "theoretical_loss": 3.4418755342252267,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021101303911735207,
      "loss": 2.4967,
      "theoretical_loss": 3.441865670700846,
      "tokens_seen": 1921777664
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1058853,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.9420671463012695,
      "objective/train/theoretical_loss": 3.4418558076070003,
      "objective/train/tokens_used": 1942303200,
      "theoretical_loss": 3.4418558076070003,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021100300902708125,
      "loss": 2.8214,
      "theoretical_loss": 3.4418558076070003,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021099297893681043,
      "loss": 2.4985,
      "theoretical_loss": 3.4418459449436574,
      "tokens_seen": 1921908736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021098294884653964,
      "loss": 2.6152,
      "theoretical_loss": 3.4418360827107835,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021097291875626882,
      "loss": 2.6584,
      "theoretical_loss": 3.4418262209083457,
      "tokens_seen": 1922039808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000210962888665998,
      "loss": 2.6807,
      "theoretical_loss": 3.4418163595363094,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021095285857572718,
      "loss": 2.4968,
      "theoretical_loss": 3.4418064985946417,
      "tokens_seen": 1922170880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021094282848545636,
      "loss": 2.7324,
      "theoretical_loss": 3.441796638083309,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021093279839518557,
      "loss": 2.5081,
      "theoretical_loss": 3.441786778002278,
      "tokens_seen": 1922301952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021092276830491475,
      "loss": 2.4723,
      "theoretical_loss": 3.4417769183515157,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021091273821464393,
      "loss": 2.4321,
      "theoretical_loss": 3.4417670591309877,
      "tokens_seen": 1922433024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002109027081243731,
      "loss": 2.3666,
      "theoretical_loss": 3.441757200340662,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021089267803410232,
      "loss": 2.5707,
      "theoretical_loss": 3.4417473419805034,
      "tokens_seen": 1922564096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108826479438315,
      "loss": 2.6881,
      "theoretical_loss": 3.4417374840504795,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021087261785356068,
      "loss": 2.5202,
      "theoretical_loss": 3.4417276265505565,
      "tokens_seen": 1922695168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021086258776328986,
      "loss": 2.2651,
      "theoretical_loss": 3.441717769480701,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021085255767301904,
      "loss": 2.5569,
      "theoretical_loss": 3.4417079128408803,
      "tokens_seen": 1922826240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021084252758274825,
      "loss": 2.7448,
      "theoretical_loss": 3.4416980566310604,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021083249749247743,
      "loss": 2.7542,
      "theoretical_loss": 3.4416882008512073,
      "tokens_seen": 1922957312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021082246740220661,
      "loss": 2.5769,
      "theoretical_loss": 3.4416783455012885,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108124373119358,
      "loss": 2.5305,
      "theoretical_loss": 3.4416684905812707,
      "tokens_seen": 1923088384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000210802407221665,
      "loss": 2.5247,
      "theoretical_loss": 3.4416586360911197,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021079237713139419,
      "loss": 2.7012,
      "theoretical_loss": 3.441648782030802,
      "tokens_seen": 1923219456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021078234704112337,
      "loss": 2.5298,
      "theoretical_loss": 3.4416389284002853,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021077231695085255,
      "loss": 2.6188,
      "theoretical_loss": 3.4416290751995353,
      "tokens_seen": 1923350528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021076228686058176,
      "loss": 2.7531,
      "theoretical_loss": 3.4416192224285194,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1059274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5529448986053467,
      "objective/train/theoretical_loss": 3.4416093700872032,
      "objective/train/tokens_used": 1943941600,
      "theoretical_loss": 3.4416093700872032,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021075225677031094,
      "loss": 2.507,
      "theoretical_loss": 3.4416093700872032,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021074222668004012,
      "loss": 2.6432,
      "theoretical_loss": 3.441599518175554,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107321965897693,
      "loss": 2.1994,
      "theoretical_loss": 3.441589666693538,
      "tokens_seen": 1923612672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021072216649949848,
      "loss": 2.7188,
      "theoretical_loss": 3.4415798156411226,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107121364092277,
      "loss": 2.4937,
      "theoretical_loss": 3.4415699650182736,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021070210631895687,
      "loss": 2.5487,
      "theoretical_loss": 3.4415601148249584,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021069207622868605,
      "loss": 2.6128,
      "theoretical_loss": 3.4415502650611427,
      "tokens_seen": 1923874816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021068204613841523,
      "loss": 2.4784,
      "theoretical_loss": 3.441540415726794,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021067201604814444,
      "loss": 2.7007,
      "theoretical_loss": 3.4415305668218785,
      "tokens_seen": 1924005888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021066198595787362,
      "loss": 2.5684,
      "theoretical_loss": 3.4415207183463625,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002106519558676028,
      "loss": 2.3914,
      "theoretical_loss": 3.4415108703002137,
      "tokens_seen": 1924136960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021064192577733198,
      "loss": 2.56,
      "theoretical_loss": 3.441501022683398,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021063189568706116,
      "loss": 2.4572,
      "theoretical_loss": 3.441491175495882,
      "tokens_seen": 1924268032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021062186559679037,
      "loss": 2.5508,
      "theoretical_loss": 3.4414813287376327,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021061183550651955,
      "loss": 2.6465,
      "theoretical_loss": 3.4414714824086166,
      "tokens_seen": 1924399104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021060180541624873,
      "loss": 2.555,
      "theoretical_loss": 3.4414616365088007,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021059177532597792,
      "loss": 2.6847,
      "theoretical_loss": 3.4414517910381512,
      "tokens_seen": 1924530176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021058174523570715,
      "loss": 2.5332,
      "theoretical_loss": 3.441441945996635,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021057171514543633,
      "loss": 2.5926,
      "theoretical_loss": 3.441432101384218,
      "tokens_seen": 1924661248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021056168505516551,
      "loss": 2.4838,
      "theoretical_loss": 3.441422257200869,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002105516549648947,
      "loss": 2.4362,
      "theoretical_loss": 3.4414124134465522,
      "tokens_seen": 1924792320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021054162487462388,
      "loss": 2.5328,
      "theoretical_loss": 3.4414025701212365,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021053159478435308,
      "loss": 2.5884,
      "theoretical_loss": 3.4413927272248865,
      "tokens_seen": 1924923392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021052156469408227,
      "loss": 2.2933,
      "theoretical_loss": 3.441382884757471,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021051153460381145,
      "loss": 2.6864,
      "theoretical_loss": 3.4413730427189546,
      "tokens_seen": 1925054464
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1060184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.482656717300415,
      "objective/train/theoretical_loss": 3.441363201109305,
      "objective/train/tokens_used": 1945580000,
      "theoretical_loss": 3.441363201109305,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021050150451354063,
      "loss": 2.4938,
      "theoretical_loss": 3.441363201109305,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021049147442326984,
      "loss": 2.5624,
      "theoretical_loss": 3.44135335992849,
      "tokens_seen": 1925185536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021048144433299902,
      "loss": 2.7061,
      "theoretical_loss": 3.4413435191764745,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104714142427282,
      "loss": 2.3504,
      "theoretical_loss": 3.4413336788532263,
      "tokens_seen": 1925316608
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021046138415245738,
      "loss": 2.4946,
      "theoretical_loss": 3.441323838958712,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021045135406218656,
      "loss": 2.6915,
      "theoretical_loss": 3.4413139994928974,
      "tokens_seen": 1925447680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021044132397191577,
      "loss": 2.7609,
      "theoretical_loss": 3.4413041604557506,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021043129388164495,
      "loss": 2.4708,
      "theoretical_loss": 3.441294321847238,
      "tokens_seen": 1925578752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021042126379137413,
      "loss": 2.6703,
      "theoretical_loss": 3.4412844836673253,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104112337011033,
      "loss": 2.5655,
      "theoretical_loss": 3.44127464591598,
      "tokens_seen": 1925709824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021040120361083252,
      "loss": 2.4776,
      "theoretical_loss": 3.4412648085931696,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103911735205617,
      "loss": 2.5822,
      "theoretical_loss": 3.44125497169886,
      "tokens_seen": 1925840896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021038114343029088,
      "loss": 2.4992,
      "theoretical_loss": 3.4412451352330176,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021037111334002006,
      "loss": 2.7437,
      "theoretical_loss": 3.4412352991956094,
      "tokens_seen": 1925971968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021036108324974924,
      "loss": 2.762,
      "theoretical_loss": 3.441225463586603,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021035105315947845,
      "loss": 2.6251,
      "theoretical_loss": 3.4412156284059643,
      "tokens_seen": 1926103040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021034102306920763,
      "loss": 2.4367,
      "theoretical_loss": 3.44120579365366,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021033099297893681,
      "loss": 2.5998,
      "theoretical_loss": 3.441195959329658,
      "tokens_seen": 1926234112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000210320962888666,
      "loss": 2.0921,
      "theoretical_loss": 3.4411861254339238,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103109327983952,
      "loss": 2.6331,
      "theoretical_loss": 3.4411762919664244,
      "tokens_seen": 1926365184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021030090270812439,
      "loss": 2.5926,
      "theoretical_loss": 3.441166458927127,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021029087261785357,
      "loss": 2.5876,
      "theoretical_loss": 3.4411566263159985,
      "tokens_seen": 1926496256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021028084252758275,
      "loss": 2.4106,
      "theoretical_loss": 3.441146794133005,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021027081243731196,
      "loss": 2.5283,
      "theoretical_loss": 3.4411369623781134,
      "tokens_seen": 1926627328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021026078234704114,
      "loss": 2.6416,
      "theoretical_loss": 3.4411271310512914,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1060737,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.610013484954834,
      "objective/train/theoretical_loss": 3.441117300152505,
      "objective/train/tokens_used": 1947218400,
      "theoretical_loss": 3.441117300152505,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021025075225677032,
      "loss": 2.4777,
      "theoretical_loss": 3.441117300152505,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102407221664995,
      "loss": 2.6177,
      "theoretical_loss": 3.441107469681721,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021023069207622868,
      "loss": 2.5486,
      "theoretical_loss": 3.441097639638907,
      "tokens_seen": 1926889472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102206619859579,
      "loss": 2.5257,
      "theoretical_loss": 3.4410878100240287,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021021063189568707,
      "loss": 2.7592,
      "theoretical_loss": 3.4410779808370533,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021020060180541625,
      "loss": 2.5741,
      "theoretical_loss": 3.441068152077948,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021019057171514543,
      "loss": 2.3571,
      "theoretical_loss": 3.4410583237466796,
      "tokens_seen": 1927151616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021018054162487464,
      "loss": 2.5698,
      "theoretical_loss": 3.441048495843215,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021017051153460382,
      "loss": 2.5754,
      "theoretical_loss": 3.4410386683675203,
      "tokens_seen": 1927282688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000210160481444333,
      "loss": 2.4434,
      "theoretical_loss": 3.4410288413195627,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021015045135406218,
      "loss": 2.6349,
      "theoretical_loss": 3.441019014699309,
      "tokens_seen": 1927413760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021014042126379136,
      "loss": 2.7227,
      "theoretical_loss": 3.4410091885067264,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021013039117352057,
      "loss": 2.6315,
      "theoretical_loss": 3.440999362741781,
      "tokens_seen": 1927544832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021012036108324975,
      "loss": 2.6549,
      "theoretical_loss": 3.440989537404441,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021011033099297894,
      "loss": 2.8161,
      "theoretical_loss": 3.4409797124946726,
      "tokens_seen": 1927675904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021010030090270812,
      "loss": 2.6199,
      "theoretical_loss": 3.4409698880124413,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021009027081243732,
      "loss": 2.536,
      "theoretical_loss": 3.440960063957716,
      "tokens_seen": 1927806976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100802407221665,
      "loss": 2.6065,
      "theoretical_loss": 3.4409502403304626,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100702106318957,
      "loss": 2.6439,
      "theoretical_loss": 3.4409404171306477,
      "tokens_seen": 1927938048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021006018054162487,
      "loss": 2.5628,
      "theoretical_loss": 3.440930594358239,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021005015045135405,
      "loss": 2.6979,
      "theoretical_loss": 3.440920772013203,
      "tokens_seen": 1928069120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021004012036108326,
      "loss": 2.5177,
      "theoretical_loss": 3.440910950095506,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021003009027081244,
      "loss": 2.5554,
      "theoretical_loss": 3.440901128605116,
      "tokens_seen": 1928200192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021002006018054162,
      "loss": 2.5751,
      "theoretical_loss": 3.4408913075419987,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100100300902708,
      "loss": 2.6834,
      "theoretical_loss": 3.4408814869061217,
      "tokens_seen": 1928331264
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1062041,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.348745107650757,
      "objective/train/theoretical_loss": 3.440871666697452,
      "objective/train/tokens_used": 1948856800,
      "theoretical_loss": 3.440871666697452,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021,
      "loss": 2.6651,
      "theoretical_loss": 3.440871666697452,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099899699097292,
      "loss": 2.3639,
      "theoretical_loss": 3.440861846915956,
      "tokens_seen": 1928462336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020997993981945837,
      "loss": 2.4818,
      "theoretical_loss": 3.4408520275616006,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020996990972918755,
      "loss": 2.5009,
      "theoretical_loss": 3.4408422086343533,
      "tokens_seen": 1928593408
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020995987963891673,
      "loss": 2.4709,
      "theoretical_loss": 3.4408323901341813,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020994984954864594,
      "loss": 2.5989,
      "theoretical_loss": 3.44082257206105,
      "tokens_seen": 1928724480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020993981945837512,
      "loss": 2.6987,
      "theoretical_loss": 3.440812754414927,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099297893681043,
      "loss": 2.4633,
      "theoretical_loss": 3.4408029371957802,
      "tokens_seen": 1928855552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020991975927783348,
      "loss": 2.4955,
      "theoretical_loss": 3.4407931204035753,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002099097291875627,
      "loss": 2.6414,
      "theoretical_loss": 3.44078330403828,
      "tokens_seen": 1928986624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020989969909729187,
      "loss": 2.3629,
      "theoretical_loss": 3.440773488099861,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020988966900702106,
      "loss": 2.381,
      "theoretical_loss": 3.4407636725882846,
      "tokens_seen": 1929117696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020987963891675024,
      "loss": 2.4939,
      "theoretical_loss": 3.440753857503519,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020986960882647942,
      "loss": 2.6953,
      "theoretical_loss": 3.44074404284553,
      "tokens_seen": 1929248768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020985957873620863,
      "loss": 2.5776,
      "theoretical_loss": 3.440734228614285,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098495486459378,
      "loss": 2.7037,
      "theoretical_loss": 3.4407244148097504,
      "tokens_seen": 1929379840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000209839518555667,
      "loss": 2.5271,
      "theoretical_loss": 3.4407146014318943,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098294884653962,
      "loss": 2.63,
      "theoretical_loss": 3.4407047884806827,
      "tokens_seen": 1929510912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098194583751254,
      "loss": 2.3819,
      "theoretical_loss": 3.4406949759560836,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020980942828485459,
      "loss": 2.5252,
      "theoretical_loss": 3.4406851638580624,
      "tokens_seen": 1929641984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020979939819458377,
      "loss": 2.6187,
      "theoretical_loss": 3.440675352186587,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020978936810431295,
      "loss": 2.3563,
      "theoretical_loss": 3.440665540941625,
      "tokens_seen": 1929773056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020977933801404216,
      "loss": 2.3444,
      "theoretical_loss": 3.4406557301231424,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020976930792377134,
      "loss": 2.717,
      "theoretical_loss": 3.4406459197311063,
      "tokens_seen": 1929904128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020975927783350052,
      "loss": 2.6438,
      "theoretical_loss": 3.440636109765484,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.58,
      "objective/train/docs_used": 1063329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.659806966781616,
      "objective/train/theoretical_loss": 3.440626300226242,
      "objective/train/tokens_used": 1950495200,
      "theoretical_loss": 3.440626300226242,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097492477432297,
      "loss": 2.6807,
      "theoretical_loss": 3.440626300226242,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020973921765295888,
      "loss": 2.6137,
      "theoretical_loss": 3.440616491113348,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097291875626881,
      "loss": 2.6743,
      "theoretical_loss": 3.4406066824267683,
      "tokens_seen": 1930166272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020971915747241727,
      "loss": 2.8564,
      "theoretical_loss": 3.44059687416647,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020970912738214645,
      "loss": 2.6033,
      "theoretical_loss": 3.440587066332421,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020969909729187563,
      "loss": 2.62,
      "theoretical_loss": 3.4405772589245873,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020968906720160484,
      "loss": 2.5546,
      "theoretical_loss": 3.440567451942936,
      "tokens_seen": 1930428416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020967903711133402,
      "loss": 2.63,
      "theoretical_loss": 3.440557645387435,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002096690070210632,
      "loss": 2.5633,
      "theoretical_loss": 3.44054783925805,
      "tokens_seen": 1930559488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020965897693079238,
      "loss": 2.4711,
      "theoretical_loss": 3.4405380335547493,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020964894684052156,
      "loss": 2.54,
      "theoretical_loss": 3.4405282282774987,
      "tokens_seen": 1930690560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020963891675025077,
      "loss": 2.8698,
      "theoretical_loss": 3.4405184234262656,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020962888665997995,
      "loss": 2.6286,
      "theoretical_loss": 3.440508619001018,
      "tokens_seen": 1930821632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020961885656970914,
      "loss": 2.5545,
      "theoretical_loss": 3.440498815001722,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020960882647943832,
      "loss": 2.4653,
      "theoretical_loss": 3.440489011428345,
      "tokens_seen": 1930952704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020959879638916752,
      "loss": 2.5154,
      "theoretical_loss": 3.440479208280853,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095887662988967,
      "loss": 2.4173,
      "theoretical_loss": 3.440469405559215,
      "tokens_seen": 1931083776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095787362086259,
      "loss": 2.3434,
      "theoretical_loss": 3.4404596032633963,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020956870611835507,
      "loss": 2.4023,
      "theoretical_loss": 3.440449801393364,
      "tokens_seen": 1931214848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020955867602808425,
      "loss": 2.4392,
      "theoretical_loss": 3.440439999949087,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020954864593781346,
      "loss": 2.311,
      "theoretical_loss": 3.440430198930531,
      "tokens_seen": 1931345920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020953861584754264,
      "loss": 2.5053,
      "theoretical_loss": 3.440420398337663,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020952858575727182,
      "loss": 2.5986,
      "theoretical_loss": 3.44041059817045,
      "tokens_seen": 1931476992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000209518555667001,
      "loss": 2.6923,
      "theoretical_loss": 3.4404007984288594,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095085255767302,
      "loss": 2.4476,
      "theoretical_loss": 3.440390999112858,
      "tokens_seen": 1931608064
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1064052,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1443674564361572,
      "objective/train/theoretical_loss": 3.4403812002224137,
      "objective/train/tokens_used": 1952133600,
      "theoretical_loss": 3.4403812002224137,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094984954864594,
      "loss": 2.4435,
      "theoretical_loss": 3.4403812002224137,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020948846539618857,
      "loss": 2.6496,
      "theoretical_loss": 3.440371401757493,
      "tokens_seen": 1931739136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020947843530591775,
      "loss": 2.6612,
      "theoretical_loss": 3.4403616037180624,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020946840521564693,
      "loss": 2.511,
      "theoretical_loss": 3.44035180610409,
      "tokens_seen": 1931870208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020945837512537614,
      "loss": 2.4858,
      "theoretical_loss": 3.440342008915542,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020944834503510532,
      "loss": 2.7659,
      "theoretical_loss": 3.440332212152386,
      "tokens_seen": 1932001280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094383149448345,
      "loss": 2.3593,
      "theoretical_loss": 3.440322415814589,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020942828485456368,
      "loss": 2.5773,
      "theoretical_loss": 3.4403126199021186,
      "tokens_seen": 1932132352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094182547642929,
      "loss": 2.3864,
      "theoretical_loss": 3.4403028244149416,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020940822467402207,
      "loss": 2.5653,
      "theoretical_loss": 3.4402930293530245,
      "tokens_seen": 1932263424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020939819458375126,
      "loss": 2.7197,
      "theoretical_loss": 3.440283234716335,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020938816449348044,
      "loss": 2.5872,
      "theoretical_loss": 3.44027344050484,
      "tokens_seen": 1932394496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020937813440320962,
      "loss": 2.5274,
      "theoretical_loss": 3.440263646718507,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020936810431293883,
      "loss": 2.5973,
      "theoretical_loss": 3.4402538533573024,
      "tokens_seen": 1932525568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000209358074222668,
      "loss": 2.6174,
      "theoretical_loss": 3.440244060421194,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002093480441323972,
      "loss": 2.4617,
      "theoretical_loss": 3.440234267910149,
      "tokens_seen": 1932656640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020933801404212637,
      "loss": 2.4787,
      "theoretical_loss": 3.4402244758241336,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020932798395185558,
      "loss": 2.7472,
      "theoretical_loss": 3.4402146841631165,
      "tokens_seen": 1932787712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020931795386158476,
      "loss": 2.6275,
      "theoretical_loss": 3.4402048929270634,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020930792377131394,
      "loss": 2.4598,
      "theoretical_loss": 3.4401951021159416,
      "tokens_seen": 1932918784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020929789368104312,
      "loss": 2.6678,
      "theoretical_loss": 3.4401853117297194,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092878635907723,
      "loss": 2.4818,
      "theoretical_loss": 3.4401755217683627,
      "tokens_seen": 1933049856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092778335005015,
      "loss": 2.6682,
      "theoretical_loss": 3.440165732231839,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092678034102307,
      "loss": 2.5199,
      "theoretical_loss": 3.440155943120116,
      "tokens_seen": 1933180928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020925777331995987,
      "loss": 2.5935,
      "theoretical_loss": 3.440146154433161,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1065201,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.658121347427368,
      "objective/train/theoretical_loss": 3.4401363661709397,
      "objective/train/tokens_used": 1953772000,
      "theoretical_loss": 3.4401363661709397,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020924774322968905,
      "loss": 2.5836,
      "theoretical_loss": 3.4401363661709397,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020923771313941826,
      "loss": 2.722,
      "theoretical_loss": 3.4401265783334205,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020922768304914744,
      "loss": 2.5621,
      "theoretical_loss": 3.4401167909205705,
      "tokens_seen": 1933443072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020921765295887662,
      "loss": 2.7054,
      "theoretical_loss": 3.440107003932357,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002092076228686058,
      "loss": 2.5129,
      "theoretical_loss": 3.440097217368746,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020919759277833499,
      "loss": 2.6229,
      "theoretical_loss": 3.4400874312297063,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091875626880642,
      "loss": 2.4256,
      "theoretical_loss": 3.4400776455152036,
      "tokens_seen": 1933705216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020917753259779338,
      "loss": 2.6267,
      "theoretical_loss": 3.4400678602252066,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020916750250752256,
      "loss": 2.6153,
      "theoretical_loss": 3.4400580753596817,
      "tokens_seen": 1933836288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020915747241725174,
      "loss": 2.439,
      "theoretical_loss": 3.440048290918596,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020914744232698095,
      "loss": 2.826,
      "theoretical_loss": 3.4400385069019164,
      "tokens_seen": 1933967360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020913741223671013,
      "loss": 2.6531,
      "theoretical_loss": 3.4400287233096107,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091273821464393,
      "loss": 2.5306,
      "theoretical_loss": 3.4400189401416466,
      "tokens_seen": 1934098432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091173520561685,
      "loss": 2.5322,
      "theoretical_loss": 3.4400091573979896,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091073219658977,
      "loss": 2.6808,
      "theoretical_loss": 3.439999375078609,
      "tokens_seen": 1934229504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020909729187562688,
      "loss": 2.6656,
      "theoretical_loss": 3.439989593183471,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020908726178535606,
      "loss": 2.5063,
      "theoretical_loss": 3.4399798117125426,
      "tokens_seen": 1934360576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020907723169508527,
      "loss": 2.6026,
      "theoretical_loss": 3.439970030665791,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020906720160481445,
      "loss": 2.5492,
      "theoretical_loss": 3.439960250043184,
      "tokens_seen": 1934491648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020905717151454366,
      "loss": 2.6644,
      "theoretical_loss": 3.4399504698446886,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020904714142427284,
      "loss": 2.4986,
      "theoretical_loss": 3.439940690070272,
      "tokens_seen": 1934622720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020903711133400202,
      "loss": 2.5282,
      "theoretical_loss": 3.4399309107199016,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002090270812437312,
      "loss": 2.5224,
      "theoretical_loss": 3.439921131793544,
      "tokens_seen": 1934753792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002090170511534604,
      "loss": 2.4403,
      "theoretical_loss": 3.439911353291168,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002090070210631896,
      "loss": 2.6333,
      "theoretical_loss": 3.4399015752127386,
      "tokens_seen": 1934884864
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1065618,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.708529233932495,
      "objective/train/theoretical_loss": 3.439891797558225,
      "objective/train/tokens_used": 1955410400,
      "theoretical_loss": 3.439891797558225,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020899699097291877,
      "loss": 2.6588,
      "theoretical_loss": 3.439891797558225,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020898696088264795,
      "loss": 2.5178,
      "theoretical_loss": 3.4398820203275937,
      "tokens_seen": 1935015936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020897693079237713,
      "loss": 2.5303,
      "theoretical_loss": 3.439872243520812,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020896690070210634,
      "loss": 2.5135,
      "theoretical_loss": 3.4398624671378473,
      "tokens_seen": 1935147008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020895687061183552,
      "loss": 2.6333,
      "theoretical_loss": 3.4398526911786664,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089468405215647,
      "loss": 2.5137,
      "theoretical_loss": 3.4398429156432373,
      "tokens_seen": 1935278080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020893681043129389,
      "loss": 2.7928,
      "theoretical_loss": 3.439833140531527,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089267803410231,
      "loss": 2.3823,
      "theoretical_loss": 3.4398233658435027,
      "tokens_seen": 1935409152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020891675025075227,
      "loss": 2.321,
      "theoretical_loss": 3.4398135915791315,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020890672016048146,
      "loss": 2.7536,
      "theoretical_loss": 3.4398038177383805,
      "tokens_seen": 1935540224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020889669007021064,
      "loss": 2.6201,
      "theoretical_loss": 3.439794044321218,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020888665997993982,
      "loss": 2.5268,
      "theoretical_loss": 3.4397842713276106,
      "tokens_seen": 1935671296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020887662988966903,
      "loss": 2.7208,
      "theoretical_loss": 3.439774498757526,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002088665997993982,
      "loss": 2.4016,
      "theoretical_loss": 3.439764726610931,
      "tokens_seen": 1935802368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002088565697091274,
      "loss": 2.5183,
      "theoretical_loss": 3.4397549548877926,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020884653961885657,
      "loss": 2.5715,
      "theoretical_loss": 3.4397451835880792,
      "tokens_seen": 1935933440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020883650952858578,
      "loss": 2.5988,
      "theoretical_loss": 3.439735412711757,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020882647943831496,
      "loss": 2.5444,
      "theoretical_loss": 3.4397256422587947,
      "tokens_seen": 1936064512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020881644934804414,
      "loss": 2.4823,
      "theoretical_loss": 3.4397158722291583,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020880641925777332,
      "loss": 2.6417,
      "theoretical_loss": 3.4397061026228157,
      "tokens_seen": 1936195584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087963891675025,
      "loss": 2.3511,
      "theoretical_loss": 3.439696333439734,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087863590772317,
      "loss": 2.4724,
      "theoretical_loss": 3.439686564679881,
      "tokens_seen": 1936326656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087763289869609,
      "loss": 2.6031,
      "theoretical_loss": 3.4396767963432238,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020876629889669007,
      "loss": 2.5704,
      "theoretical_loss": 3.4396670284297297,
      "tokens_seen": 1936457728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020875626880641925,
      "loss": 2.5832,
      "theoretical_loss": 3.4396572609393656,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1066685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6814911365509033,
      "objective/train/theoretical_loss": 3.4396474938720996,
      "objective/train/tokens_used": 1957048800,
      "theoretical_loss": 3.4396474938720996,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020874623871614846,
      "loss": 2.6284,
      "theoretical_loss": 3.4396474938720996,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020873620862587764,
      "loss": 2.5157,
      "theoretical_loss": 3.439637727227899,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020872617853560682,
      "loss": 2.473,
      "theoretical_loss": 3.439627961006731,
      "tokens_seen": 1936719872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000208716148445336,
      "loss": 2.4475,
      "theoretical_loss": 3.4396181952085625,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020870611835506519,
      "loss": 2.6441,
      "theoretical_loss": 3.4396084298333616,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086960882647944,
      "loss": 2.5619,
      "theoretical_loss": 3.4395986648810943,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020868605817452358,
      "loss": 2.4566,
      "theoretical_loss": 3.43958890035173,
      "tokens_seen": 1936982016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020867602808425276,
      "loss": 2.5837,
      "theoretical_loss": 3.439579136245235,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020866599799398194,
      "loss": 2.6386,
      "theoretical_loss": 3.439569372561577,
      "tokens_seen": 1937113088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020865596790371115,
      "loss": 2.7682,
      "theoretical_loss": 3.439559609300722,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020864593781344033,
      "loss": 2.4205,
      "theoretical_loss": 3.439549846462639,
      "tokens_seen": 1937244160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086359077231695,
      "loss": 2.6013,
      "theoretical_loss": 3.4395400840472954,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086258776328987,
      "loss": 2.5142,
      "theoretical_loss": 3.439530322054658,
      "tokens_seen": 1937375232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086158475426279,
      "loss": 2.6419,
      "theoretical_loss": 3.439520560484694,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020860581745235708,
      "loss": 2.6742,
      "theoretical_loss": 3.4395107993373712,
      "tokens_seen": 1937506304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020859578736208626,
      "loss": 2.4267,
      "theoretical_loss": 3.439501038612657,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020858575727181544,
      "loss": 2.7977,
      "theoretical_loss": 3.4394912783105185,
      "tokens_seen": 1937637376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020857572718154462,
      "loss": 2.3995,
      "theoretical_loss": 3.4394815184309238,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020856569709127383,
      "loss": 2.3923,
      "theoretical_loss": 3.4394717589738395,
      "tokens_seen": 1937768448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000208555667001003,
      "loss": 2.5206,
      "theoretical_loss": 3.4394619999392333,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085456369107322,
      "loss": 2.6874,
      "theoretical_loss": 3.4394522413270727,
      "tokens_seen": 1937899520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020853560682046137,
      "loss": 2.8336,
      "theoretical_loss": 3.4394424831373254,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020852557673019058,
      "loss": 2.5639,
      "theoretical_loss": 3.439432725369958,
      "tokens_seen": 1938030592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020851554663991976,
      "loss": 2.5879,
      "theoretical_loss": 3.439422968024939,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020850551654964894,
      "loss": 2.6598,
      "theoretical_loss": 3.4394132111022353,
      "tokens_seen": 1938161664
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1067245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7488558292388916,
      "objective/train/theoretical_loss": 3.4394034546018144,
      "objective/train/tokens_used": 1958687200,
      "theoretical_loss": 3.4394034546018144,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020849548645937813,
      "loss": 2.5836,
      "theoretical_loss": 3.4394034546018144,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084854563691073,
      "loss": 2.5003,
      "theoretical_loss": 3.4393936985236433,
      "tokens_seen": 1938292736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020847542627883651,
      "loss": 2.4267,
      "theoretical_loss": 3.439383942867691,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084653961885657,
      "loss": 2.4229,
      "theoretical_loss": 3.4393741876339226,
      "tokens_seen": 1938423808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020845536609829488,
      "loss": 2.4619,
      "theoretical_loss": 3.439364432822307,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020844533600802406,
      "loss": 2.3553,
      "theoretical_loss": 3.439354678432812,
      "tokens_seen": 1938554880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020843530591775327,
      "loss": 2.5586,
      "theoretical_loss": 3.439344924465404,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020842527582748245,
      "loss": 2.6031,
      "theoretical_loss": 3.4393351709200513,
      "tokens_seen": 1938685952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020841524573721163,
      "loss": 2.3724,
      "theoretical_loss": 3.439325417796721,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084052156469408,
      "loss": 2.3961,
      "theoretical_loss": 3.4393156650953802,
      "tokens_seen": 1938817024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020839518555667,
      "loss": 2.4523,
      "theoretical_loss": 3.439305912815997,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083851554663992,
      "loss": 2.6276,
      "theoretical_loss": 3.4392961609585386,
      "tokens_seen": 1938948096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020837512537612838,
      "loss": 2.5095,
      "theoretical_loss": 3.439286409522973,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020836509528585756,
      "loss": 2.6432,
      "theoretical_loss": 3.439276658509267,
      "tokens_seen": 1939079168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020835506519558674,
      "loss": 2.4167,
      "theoretical_loss": 3.4392669079173883,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020834503510531595,
      "loss": 2.7195,
      "theoretical_loss": 3.4392571577473046,
      "tokens_seen": 1939210240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020833500501504513,
      "loss": 2.4782,
      "theoretical_loss": 3.439247407998983,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020832497492477434,
      "loss": 2.3517,
      "theoretical_loss": 3.4392376586723916,
      "tokens_seen": 1939341312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020831494483450352,
      "loss": 2.4237,
      "theoretical_loss": 3.439227909767497,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083049147442327,
      "loss": 2.6809,
      "theoretical_loss": 3.439218161284268,
      "tokens_seen": 1939472384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082948846539619,
      "loss": 2.373,
      "theoretical_loss": 3.439208413222671,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082848545636911,
      "loss": 2.6491,
      "theoretical_loss": 3.4391986655826736,
      "tokens_seen": 1939603456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020827482447342027,
      "loss": 2.7416,
      "theoretical_loss": 3.439188918364244,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020826479438314945,
      "loss": 2.677,
      "theoretical_loss": 3.439179171567349,
      "tokens_seen": 1939734528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020825476429287866,
      "loss": 2.5853,
      "theoretical_loss": 3.439169425191957,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1068281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8406989574432373,
      "objective/train/theoretical_loss": 3.439159679238035,
      "objective/train/tokens_used": 1960325600,
      "theoretical_loss": 3.439159679238035,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020824473420260784,
      "loss": 2.7324,
      "theoretical_loss": 3.439159679238035,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020823470411233702,
      "loss": 2.4204,
      "theoretical_loss": 3.4391499337055498,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082246740220662,
      "loss": 2.358,
      "theoretical_loss": 3.4391401885944703,
      "tokens_seen": 1939996672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082146439317954,
      "loss": 2.6976,
      "theoretical_loss": 3.439130443904763,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082046138415246,
      "loss": 2.6848,
      "theoretical_loss": 3.4391206996363968,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020819458375125378,
      "loss": 2.5763,
      "theoretical_loss": 3.439110955789338,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020818455366098296,
      "loss": 2.5247,
      "theoretical_loss": 3.4391012123635534,
      "tokens_seen": 1940258816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020817452357071214,
      "loss": 2.5339,
      "theoretical_loss": 3.4390914693590124,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020816449348044135,
      "loss": 2.5178,
      "theoretical_loss": 3.439081726775682,
      "tokens_seen": 1940389888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020815446339017053,
      "loss": 2.6544,
      "theoretical_loss": 3.4390719846135296,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081444332998997,
      "loss": 2.7183,
      "theoretical_loss": 3.4390622428725224,
      "tokens_seen": 1940520960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081344032096289,
      "loss": 2.5463,
      "theoretical_loss": 3.439052501552628,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081243731193581,
      "loss": 2.3993,
      "theoretical_loss": 3.4390427606538148,
      "tokens_seen": 1940652032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020811434302908728,
      "loss": 2.7173,
      "theoretical_loss": 3.4390330201760495,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020810431293881646,
      "loss": 2.5735,
      "theoretical_loss": 3.439023280119301,
      "tokens_seen": 1940783104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020809428284854564,
      "loss": 2.5581,
      "theoretical_loss": 3.439013540483535,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020808425275827482,
      "loss": 2.5757,
      "theoretical_loss": 3.4390038012687203,
      "tokens_seen": 1940914176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020807422266800403,
      "loss": 2.6524,
      "theoretical_loss": 3.4389940624748236,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002080641925777332,
      "loss": 2.5863,
      "theoretical_loss": 3.438984324101814,
      "tokens_seen": 1941045248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002080541624874624,
      "loss": 2.1903,
      "theoretical_loss": 3.4389745861496577,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020804413239719157,
      "loss": 2.6255,
      "theoretical_loss": 3.438964848618323,
      "tokens_seen": 1941176320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020803410230692078,
      "loss": 2.4995,
      "theoretical_loss": 3.438955111507777,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020802407221664996,
      "loss": 2.6022,
      "theoretical_loss": 3.438945374817988,
      "tokens_seen": 1941307392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020801404212637914,
      "loss": 2.6629,
      "theoretical_loss": 3.4389356385489234,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020800401203610833,
      "loss": 2.521,
      "theoretical_loss": 3.43892590270055,
      "tokens_seen": 1941438464
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1068989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6390552520751953,
      "objective/train/theoretical_loss": 3.4389161672728368,
      "objective/train/tokens_used": 1961964000,
      "theoretical_loss": 3.4389161672728368,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002079939819458375,
      "loss": 2.4885,
      "theoretical_loss": 3.4389161672728368,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020798395185556671,
      "loss": 2.3817,
      "theoretical_loss": 3.4389064322657505,
      "tokens_seen": 1941569536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002079739217652959,
      "loss": 2.4814,
      "theoretical_loss": 3.438896697679259,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020796389167502508,
      "loss": 2.5276,
      "theoretical_loss": 3.4388869635133297,
      "tokens_seen": 1941700608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020795386158475426,
      "loss": 2.7269,
      "theoretical_loss": 3.4388772297679306,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020794383149448347,
      "loss": 2.5119,
      "theoretical_loss": 3.438867496443029,
      "tokens_seen": 1941831680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020793380140421265,
      "loss": 2.5639,
      "theoretical_loss": 3.438857763538593,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020792377131394183,
      "loss": 2.502,
      "theoretical_loss": 3.4388480310545892,
      "tokens_seen": 1941962752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207913741223671,
      "loss": 2.5479,
      "theoretical_loss": 3.4388382989909863,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002079037111334002,
      "loss": 2.6649,
      "theoretical_loss": 3.438828567347752,
      "tokens_seen": 1942093824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078936810431294,
      "loss": 2.6652,
      "theoretical_loss": 3.4388188361248533,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020788365095285858,
      "loss": 2.6039,
      "theoretical_loss": 3.4388091053222585,
      "tokens_seen": 1942224896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020787362086258776,
      "loss": 2.5123,
      "theoretical_loss": 3.4387993749399346,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020786359077231694,
      "loss": 2.7731,
      "theoretical_loss": 3.4387896449778497,
      "tokens_seen": 1942355968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020785356068204615,
      "loss": 2.4798,
      "theoretical_loss": 3.4387799154359717,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020784353059177533,
      "loss": 2.474,
      "theoretical_loss": 3.4387701863142675,
      "tokens_seen": 1942487040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078335005015045,
      "loss": 2.524,
      "theoretical_loss": 3.4387604576127053,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078234704112337,
      "loss": 2.5058,
      "theoretical_loss": 3.438750729331253,
      "tokens_seen": 1942618112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020781344032096288,
      "loss": 2.5036,
      "theoretical_loss": 3.438741001469878,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020780341023069208,
      "loss": 2.5951,
      "theoretical_loss": 3.438731274028547,
      "tokens_seen": 1942749184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020779338014042126,
      "loss": 2.4226,
      "theoretical_loss": 3.43872154700723,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020778335005015045,
      "loss": 2.3182,
      "theoretical_loss": 3.438711820405893,
      "tokens_seen": 1942880256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020777331995987963,
      "loss": 2.6733,
      "theoretical_loss": 3.438702094224504,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020776328986960884,
      "loss": 2.538,
      "theoretical_loss": 3.4386923684630304,
      "tokens_seen": 1943011328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020775325977933802,
      "loss": 2.5628,
      "theoretical_loss": 3.438682643121441,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1070184,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.677218198776245,
      "objective/train/theoretical_loss": 3.4386729181997024,
      "objective/train/tokens_used": 1963602400,
      "theoretical_loss": 3.4386729181997024,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077432296890672,
      "loss": 2.6541,
      "theoretical_loss": 3.4386729181997024,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020773319959879638,
      "loss": 2.6223,
      "theoretical_loss": 3.4386631936977827,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020772316950852556,
      "loss": 2.6795,
      "theoretical_loss": 3.43865346961565,
      "tokens_seen": 1943273472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020771313941825477,
      "loss": 2.5212,
      "theoretical_loss": 3.4386437459532715,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020770310932798395,
      "loss": 2.4562,
      "theoretical_loss": 3.4386340227106147,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020769307923771313,
      "loss": 2.6941,
      "theoretical_loss": 3.438624299887648,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076830491474423,
      "loss": 2.4697,
      "theoretical_loss": 3.438614577484339,
      "tokens_seen": 1943535616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020767301905717152,
      "loss": 2.587,
      "theoretical_loss": 3.438604855500655,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076629889669007,
      "loss": 2.73,
      "theoretical_loss": 3.4385951339365644,
      "tokens_seen": 1943666688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020765295887662988,
      "loss": 2.5232,
      "theoretical_loss": 3.438585412792034,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020764292878635906,
      "loss": 2.6782,
      "theoretical_loss": 3.4385756920670327,
      "tokens_seen": 1943797760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020763289869608824,
      "loss": 2.6119,
      "theoretical_loss": 3.438565971761528,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020762286860581745,
      "loss": 2.351,
      "theoretical_loss": 3.4385562518754864,
      "tokens_seen": 1943928832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020761283851554663,
      "loss": 2.5047,
      "theoretical_loss": 3.4385465324088766,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020760280842527581,
      "loss": 2.3807,
      "theoretical_loss": 3.438536813361667,
      "tokens_seen": 1944059904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207592778335005,
      "loss": 2.5698,
      "theoretical_loss": 3.4385270947338245,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075827482447342,
      "loss": 2.4481,
      "theoretical_loss": 3.4385173765253167,
      "tokens_seen": 1944190976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075727181544634,
      "loss": 2.3967,
      "theoretical_loss": 3.4385076587361123,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075626880641926,
      "loss": 2.5531,
      "theoretical_loss": 3.4384979413661787,
      "tokens_seen": 1944322048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020755265797392177,
      "loss": 2.5264,
      "theoretical_loss": 3.438488224415483,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020754262788365098,
      "loss": 2.6118,
      "theoretical_loss": 3.4384785078839935,
      "tokens_seen": 1944453120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020753259779338016,
      "loss": 2.5489,
      "theoretical_loss": 3.438468791771678,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020752256770310934,
      "loss": 2.6615,
      "theoretical_loss": 3.4384590760785043,
      "tokens_seen": 1944584192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020751253761283853,
      "loss": 2.7968,
      "theoretical_loss": 3.4384493608044404,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075025075225677,
      "loss": 2.4952,
      "theoretical_loss": 3.4384396459494537,
      "tokens_seen": 1944715264
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1070544,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6319706439971924,
      "objective/train/theoretical_loss": 3.4384299315135123,
      "objective/train/tokens_used": 1965240800,
      "theoretical_loss": 3.4384299315135123,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020749247743229692,
      "loss": 2.5516,
      "theoretical_loss": 3.4384299315135123,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074824473420261,
      "loss": 2.5993,
      "theoretical_loss": 3.438420217496584,
      "tokens_seen": 1944846336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020747241725175528,
      "loss": 2.6436,
      "theoretical_loss": 3.438410503898636,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020746238716148446,
      "loss": 2.6677,
      "theoretical_loss": 3.438400790719637,
      "tokens_seen": 1944977408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020745235707121367,
      "loss": 2.4915,
      "theoretical_loss": 3.4383910779595537,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020744232698094285,
      "loss": 2.6308,
      "theoretical_loss": 3.4383813656183557,
      "tokens_seen": 1945108480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020743229689067203,
      "loss": 2.4908,
      "theoretical_loss": 3.438371653696009,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074222668004012,
      "loss": 2.6728,
      "theoretical_loss": 3.4383619421924823,
      "tokens_seen": 1945239552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074122367101304,
      "loss": 2.6868,
      "theoretical_loss": 3.4383522311077432,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074022066198596,
      "loss": 2.3316,
      "theoretical_loss": 3.43834252044176,
      "tokens_seen": 1945370624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020739217652958878,
      "loss": 2.5476,
      "theoretical_loss": 3.4383328101945,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020738214643931796,
      "loss": 2.4615,
      "theoretical_loss": 3.438323100365931,
      "tokens_seen": 1945501696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020737211634904714,
      "loss": 2.473,
      "theoretical_loss": 3.4383133909560213,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020736208625877635,
      "loss": 2.7236,
      "theoretical_loss": 3.4383036819647383,
      "tokens_seen": 1945632768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020735205616850553,
      "loss": 2.4975,
      "theoretical_loss": 3.4382939733920503,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073420260782347,
      "loss": 2.5851,
      "theoretical_loss": 3.4382842652379244,
      "tokens_seen": 1945763840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073319959879639,
      "loss": 2.5837,
      "theoretical_loss": 3.4382745575023295,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020732196589769308,
      "loss": 2.7209,
      "theoretical_loss": 3.4382648501852326,
      "tokens_seen": 1945894912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020731193580742228,
      "loss": 2.5992,
      "theoretical_loss": 3.4382551432866024,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020730190571715146,
      "loss": 2.5857,
      "theoretical_loss": 3.438245436806406,
      "tokens_seen": 1946025984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020729187562688065,
      "loss": 2.4598,
      "theoretical_loss": 3.438235730744611,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020728184553660983,
      "loss": 2.2502,
      "theoretical_loss": 3.4382260251011862,
      "tokens_seen": 1946157056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020727181544633904,
      "loss": 2.4427,
      "theoretical_loss": 3.4382163198760995,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020726178535606822,
      "loss": 2.4369,
      "theoretical_loss": 3.4382066150693182,
      "tokens_seen": 1946288128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002072517552657974,
      "loss": 2.4926,
      "theoretical_loss": 3.43819691068081,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1071829,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.820390224456787,
      "objective/train/theoretical_loss": 3.438187206710543,
      "objective/train/tokens_used": 1966879200,
      "theoretical_loss": 3.438187206710543,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020724172517552658,
      "loss": 2.5973,
      "theoretical_loss": 3.438187206710543,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020723169508525576,
      "loss": 2.3429,
      "theoretical_loss": 3.4381775031584856,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020722166499498497,
      "loss": 2.4582,
      "theoretical_loss": 3.4381678000246056,
      "tokens_seen": 1946550272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020721163490471415,
      "loss": 2.5998,
      "theoretical_loss": 3.43815809730887,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020720160481444333,
      "loss": 2.427,
      "theoretical_loss": 3.4381483950112477,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071915747241725,
      "loss": 2.5285,
      "theoretical_loss": 3.438138693131706,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020718154463390172,
      "loss": 2.4788,
      "theoretical_loss": 3.4381289916702134,
      "tokens_seen": 1946812416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071715145436309,
      "loss": 2.4671,
      "theoretical_loss": 3.4381192906267373,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020716148445336008,
      "loss": 2.7558,
      "theoretical_loss": 3.438109590001246,
      "tokens_seen": 1946943488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020715145436308926,
      "loss": 2.5465,
      "theoretical_loss": 3.4380998897937065,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020714142427281844,
      "loss": 2.5058,
      "theoretical_loss": 3.438090190004088,
      "tokens_seen": 1947074560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020713139418254765,
      "loss": 2.5661,
      "theoretical_loss": 3.438080490632358,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020712136409227683,
      "loss": 2.6772,
      "theoretical_loss": 3.438070791678484,
      "tokens_seen": 1947205632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020711133400200601,
      "loss": 2.5114,
      "theoretical_loss": 3.4380610931424345,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071013039117352,
      "loss": 2.6325,
      "theoretical_loss": 3.4380513950241767,
      "tokens_seen": 1947336704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070912738214644,
      "loss": 2.3946,
      "theoretical_loss": 3.438041697323679,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020708124373119358,
      "loss": 2.3308,
      "theoretical_loss": 3.43803200004091,
      "tokens_seen": 1947467776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020707121364092277,
      "loss": 2.6628,
      "theoretical_loss": 3.4380223031758366,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020706118355065195,
      "loss": 2.6807,
      "theoretical_loss": 3.438012606728427,
      "tokens_seen": 1947598848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020705115346038116,
      "loss": 2.6888,
      "theoretical_loss": 3.43800291069865,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020704112337011034,
      "loss": 2.2714,
      "theoretical_loss": 3.4379932150864723,
      "tokens_seen": 1947729920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020703109327983952,
      "loss": 2.5274,
      "theoretical_loss": 3.4379835198918625,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070210631895687,
      "loss": 2.5729,
      "theoretical_loss": 3.4379738251147884,
      "tokens_seen": 1947860992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020701103309929788,
      "loss": 2.6034,
      "theoretical_loss": 3.4379641307552182,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070010030090271,
      "loss": 2.6871,
      "theoretical_loss": 3.43795443681312,
      "tokens_seen": 1947992064
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1072422,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.84401798248291,
      "objective/train/theoretical_loss": 3.4379447432884613,
      "objective/train/tokens_used": 1968517600,
      "theoretical_loss": 3.4379447432884613,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020699097291875627,
      "loss": 2.6317,
      "theoretical_loss": 3.4379447432884613,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020698094282848545,
      "loss": 2.5283,
      "theoretical_loss": 3.4379350501812107,
      "tokens_seen": 1948123136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020697091273821463,
      "loss": 2.5036,
      "theoretical_loss": 3.437925357491335,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020696088264794384,
      "loss": 2.6699,
      "theoretical_loss": 3.437915665218804,
      "tokens_seen": 1948254208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020695085255767302,
      "loss": 2.6733,
      "theoretical_loss": 3.437905973363584,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002069408224674022,
      "loss": 2.5152,
      "theoretical_loss": 3.437896281925644,
      "tokens_seen": 1948385280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020693079237713138,
      "loss": 2.5011,
      "theoretical_loss": 3.4378865909049514,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020692076228686056,
      "loss": 2.7288,
      "theoretical_loss": 3.4378769003014744,
      "tokens_seen": 1948516352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020691073219658977,
      "loss": 2.5065,
      "theoretical_loss": 3.437867210115181,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020690070210631895,
      "loss": 2.4554,
      "theoretical_loss": 3.43785752034604,
      "tokens_seen": 1948647424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020689067201604813,
      "loss": 2.586,
      "theoretical_loss": 3.437847830994018,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020688064192577732,
      "loss": 2.375,
      "theoretical_loss": 3.4378381420590842,
      "tokens_seen": 1948778496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020687061183550652,
      "loss": 2.497,
      "theoretical_loss": 3.437828453541206,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002068605817452357,
      "loss": 2.4247,
      "theoretical_loss": 3.4378187654403516,
      "tokens_seen": 1948909568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020685055165496489,
      "loss": 2.402,
      "theoretical_loss": 3.437809077756489,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020684052156469407,
      "loss": 2.6273,
      "theoretical_loss": 3.437799390489586,
      "tokens_seen": 1949040640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020683049147442328,
      "loss": 2.438,
      "theoretical_loss": 3.437789703639611,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020682046138415248,
      "loss": 2.6951,
      "theoretical_loss": 3.4377800172065323,
      "tokens_seen": 1949171712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020681043129388166,
      "loss": 2.5612,
      "theoretical_loss": 3.4377703311903174,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020680040120361085,
      "loss": 2.5329,
      "theoretical_loss": 3.437760645590934,
      "tokens_seen": 1949302784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020679037111334003,
      "loss": 2.5634,
      "theoretical_loss": 3.437750960408351,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020678034102306924,
      "loss": 2.7044,
      "theoretical_loss": 3.4377412756425363,
      "tokens_seen": 1949433856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020677031093279842,
      "loss": 2.6433,
      "theoretical_loss": 3.437731591293457,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067602808425276,
      "loss": 2.5622,
      "theoretical_loss": 3.4377219073610825,
      "tokens_seen": 1949564928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020675025075225678,
      "loss": 2.5404,
      "theoretical_loss": 3.4377122238453803,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1073760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.728753089904785,
      "objective/train/theoretical_loss": 3.437702540746318,
      "objective/train/tokens_used": 1970156000,
      "theoretical_loss": 3.437702540746318,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020674022066198596,
      "loss": 2.5697,
      "theoretical_loss": 3.437702540746318,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020673019057171517,
      "loss": 2.7396,
      "theoretical_loss": 3.4376928580638646,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020672016048144435,
      "loss": 2.5568,
      "theoretical_loss": 3.4376831757979875,
      "tokens_seen": 1949827072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020671013039117353,
      "loss": 2.5522,
      "theoretical_loss": 3.437673493948655,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067001003009027,
      "loss": 2.6856,
      "theoretical_loss": 3.437663812515835,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067001003009027,
      "loss": 2.6142,
      "theoretical_loss": 3.437654131499496,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020669007021063192,
      "loss": 2.6384,
      "theoretical_loss": 3.4376444508996054,
      "tokens_seen": 1950089216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066800401203611,
      "loss": 2.4099,
      "theoretical_loss": 3.437634770716132,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020667001003009028,
      "loss": 2.5741,
      "theoretical_loss": 3.4376250909490436,
      "tokens_seen": 1950220288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020665997993981946,
      "loss": 2.3889,
      "theoretical_loss": 3.4376154115983084,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020664994984954864,
      "loss": 2.3937,
      "theoretical_loss": 3.4376057326638945,
      "tokens_seen": 1950351360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020663991975927785,
      "loss": 2.5279,
      "theoretical_loss": 3.4375960541457697,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020662988966900703,
      "loss": 2.6923,
      "theoretical_loss": 3.437586376043902,
      "tokens_seen": 1950482432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020661985957873621,
      "loss": 2.5041,
      "theoretical_loss": 3.43757669835826,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066098294884654,
      "loss": 2.5483,
      "theoretical_loss": 3.437567021088812,
      "tokens_seen": 1950613504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065997993981946,
      "loss": 2.6021,
      "theoretical_loss": 3.4375573442355254,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020658976930792379,
      "loss": 2.4214,
      "theoretical_loss": 3.4375476677983685,
      "tokens_seen": 1950744576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020657973921765297,
      "loss": 2.3999,
      "theoretical_loss": 3.43753799177731,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020657973921765297,
      "loss": 2.6083,
      "theoretical_loss": 3.437528316172317,
      "tokens_seen": 1950875648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020656970912738215,
      "loss": 2.8424,
      "theoretical_loss": 3.437518640983359,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020655967903711136,
      "loss": 2.7623,
      "theoretical_loss": 3.4375089662104035,
      "tokens_seen": 1951006720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020654964894684054,
      "loss": 2.2909,
      "theoretical_loss": 3.4374992918534177,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020653961885656972,
      "loss": 2.5697,
      "theoretical_loss": 3.4374896179123713,
      "tokens_seen": 1951137792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065295887662989,
      "loss": 2.7343,
      "theoretical_loss": 3.4374799443872317,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020651955867602808,
      "loss": 2.5318,
      "theoretical_loss": 3.4374702712779666,
      "tokens_seen": 1951268864
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1074230,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2596077919006348,
      "objective/train/theoretical_loss": 3.4374605985845452,
      "objective/train/tokens_used": 1971794400,
      "theoretical_loss": 3.4374605985845452,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065095285857573,
      "loss": 2.4284,
      "theoretical_loss": 3.4374605985845452,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020649949849548647,
      "loss": 2.5814,
      "theoretical_loss": 3.4374509263069353,
      "tokens_seen": 1951399936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020648946840521565,
      "loss": 2.5534,
      "theoretical_loss": 3.437441254445104,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020647943831494483,
      "loss": 2.541,
      "theoretical_loss": 3.4374315829990207,
      "tokens_seen": 1951531008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020646940822467404,
      "loss": 2.5716,
      "theoretical_loss": 3.437421911968653,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020645937813440322,
      "loss": 2.5626,
      "theoretical_loss": 3.43741224135397,
      "tokens_seen": 1951662080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002064493480441324,
      "loss": 2.5608,
      "theoretical_loss": 3.4374025711549385,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020643931795386158,
      "loss": 2.5735,
      "theoretical_loss": 3.4373929013715276,
      "tokens_seen": 1951793152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020642928786359076,
      "loss": 2.4109,
      "theoretical_loss": 3.437383232003705,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020641925777331997,
      "loss": 2.6246,
      "theoretical_loss": 3.4373735630514393,
      "tokens_seen": 1951924224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020640922768304915,
      "loss": 2.3522,
      "theoretical_loss": 3.4373638945146983,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020639919759277833,
      "loss": 2.6404,
      "theoretical_loss": 3.4373542263934507,
      "tokens_seen": 1952055296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020638916750250752,
      "loss": 2.5344,
      "theoretical_loss": 3.4373445586876636,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020637913741223672,
      "loss": 2.7196,
      "theoretical_loss": 3.4373348913973065,
      "tokens_seen": 1952186368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002063691073219659,
      "loss": 2.5259,
      "theoretical_loss": 3.4373252245223473,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020635907723169509,
      "loss": 2.5407,
      "theoretical_loss": 3.4373155580627532,
      "tokens_seen": 1952317440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020634904714142427,
      "loss": 2.3435,
      "theoretical_loss": 3.437305892018494,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020633901705115345,
      "loss": 2.6979,
      "theoretical_loss": 3.437296226389537,
      "tokens_seen": 1952448512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020632898696088266,
      "loss": 2.6907,
      "theoretical_loss": 3.43728656117585,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020631895687061184,
      "loss": 2.6652,
      "theoretical_loss": 3.4372768963774023,
      "tokens_seen": 1952579584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020630892678034102,
      "loss": 2.7143,
      "theoretical_loss": 3.437267231994161,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062988966900702,
      "loss": 2.6686,
      "theoretical_loss": 3.4372575680260953,
      "tokens_seen": 1952710656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062888665997994,
      "loss": 2.4356,
      "theoretical_loss": 3.437247904473173,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062788365095286,
      "loss": 2.7915,
      "theoretical_loss": 3.437238241335362,
      "tokens_seen": 1952841728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020626880641925777,
      "loss": 2.619,
      "theoretical_loss": 3.4372285786126318,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1075533,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7241132259368896,
      "objective/train/theoretical_loss": 3.437218916304949,
      "objective/train/tokens_used": 1973432800,
      "theoretical_loss": 3.437218916304949,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020625877632898695,
      "loss": 2.6765,
      "theoretical_loss": 3.437218916304949,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020624874623871613,
      "loss": 2.5756,
      "theoretical_loss": 3.4372092544122825,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020623871614844534,
      "loss": 2.3954,
      "theoretical_loss": 3.4371995929346006,
      "tokens_seen": 1953103872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020622868605817452,
      "loss": 2.7799,
      "theoretical_loss": 3.437189931871872,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062186559679037,
      "loss": 2.4535,
      "theoretical_loss": 3.437180271224064,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020620862587763288,
      "loss": 2.5429,
      "theoretical_loss": 3.437170610991146,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061985957873621,
      "loss": 2.5083,
      "theoretical_loss": 3.4371609511730856,
      "tokens_seen": 1953366016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020618856569709127,
      "loss": 2.4477,
      "theoretical_loss": 3.437151291769851,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020617853560682046,
      "loss": 2.6424,
      "theoretical_loss": 3.437141632781411,
      "tokens_seen": 1953497088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020616850551654964,
      "loss": 2.6174,
      "theoretical_loss": 3.4371319742077326,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020615847542627882,
      "loss": 2.3998,
      "theoretical_loss": 3.4371223160487854,
      "tokens_seen": 1953628160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020614844533600803,
      "loss": 2.5626,
      "theoretical_loss": 3.437112658304537,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061384152457372,
      "loss": 2.4472,
      "theoretical_loss": 3.4371030009749566,
      "tokens_seen": 1953759232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061283851554664,
      "loss": 2.5731,
      "theoretical_loss": 3.437093344060011,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020611835506519557,
      "loss": 2.442,
      "theoretical_loss": 3.4370836875596695,
      "tokens_seen": 1953890304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020610832497492478,
      "loss": 2.5425,
      "theoretical_loss": 3.437074031473901,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020609829488465396,
      "loss": 2.7872,
      "theoretical_loss": 3.437064375802672,
      "tokens_seen": 1954021376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020608826479438314,
      "loss": 2.6405,
      "theoretical_loss": 3.4370547205459516,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020607823470411235,
      "loss": 2.4459,
      "theoretical_loss": 3.4370450657037086,
      "tokens_seen": 1954152448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020606820461384156,
      "loss": 2.5474,
      "theoretical_loss": 3.4370354112759114,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020605817452357074,
      "loss": 2.5117,
      "theoretical_loss": 3.4370257572625276,
      "tokens_seen": 1954283520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020604814443329992,
      "loss": 2.4508,
      "theoretical_loss": 3.4370161036635256,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002060381143430291,
      "loss": 2.408,
      "theoretical_loss": 3.437006450478875,
      "tokens_seen": 1954414592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020602808425275828,
      "loss": 2.4234,
      "theoretical_loss": 3.4369967977085416,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002060180541624875,
      "loss": 2.6188,
      "theoretical_loss": 3.436987145352496,
      "tokens_seen": 1954545664
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1076714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.37070894241333,
      "objective/train/theoretical_loss": 3.436977493410706,
      "objective/train/tokens_used": 1975071200,
      "theoretical_loss": 3.436977493410706,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020600802407221667,
      "loss": 2.6842,
      "theoretical_loss": 3.436977493410706,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020599799398194585,
      "loss": 2.5074,
      "theoretical_loss": 3.436967841883139,
      "tokens_seen": 1954676736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020598796389167503,
      "loss": 2.5589,
      "theoretical_loss": 3.436958190769764,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020597793380140424,
      "loss": 2.6223,
      "theoretical_loss": 3.4369485400705497,
      "tokens_seen": 1954807808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020596790371113342,
      "loss": 2.2463,
      "theoretical_loss": 3.4369388897854636,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002059578736208626,
      "loss": 2.6479,
      "theoretical_loss": 3.436929239914475,
      "tokens_seen": 1954938880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020594784353059178,
      "loss": 2.573,
      "theoretical_loss": 3.436919590457552,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020593781344032096,
      "loss": 2.8222,
      "theoretical_loss": 3.436909941414662,
      "tokens_seen": 1955069952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020592778335005017,
      "loss": 2.6818,
      "theoretical_loss": 3.4369002927857744,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020591775325977935,
      "loss": 2.6239,
      "theoretical_loss": 3.436890644570857,
      "tokens_seen": 1955201024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020590772316950853,
      "loss": 2.4656,
      "theoretical_loss": 3.436880996769879,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020589769307923772,
      "loss": 2.5706,
      "theoretical_loss": 3.436871349382808,
      "tokens_seen": 1955332096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020588766298896692,
      "loss": 2.4452,
      "theoretical_loss": 3.436861702409612,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058776328986961,
      "loss": 2.297,
      "theoretical_loss": 3.4368520558502604,
      "tokens_seen": 1955463168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058676028084253,
      "loss": 2.6225,
      "theoretical_loss": 3.4368424097047208,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020585757271815447,
      "loss": 2.4339,
      "theoretical_loss": 3.436832763972962,
      "tokens_seen": 1955594240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020584754262788365,
      "loss": 2.5754,
      "theoretical_loss": 3.4368231186549525,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020583751253761286,
      "loss": 2.5614,
      "theoretical_loss": 3.43681347375066,
      "tokens_seen": 1955725312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020582748244734204,
      "loss": 2.4983,
      "theoretical_loss": 3.436803829260054,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020581745235707122,
      "loss": 2.4508,
      "theoretical_loss": 3.4367941851831016,
      "tokens_seen": 1955856384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058074222668004,
      "loss": 2.6237,
      "theoretical_loss": 3.4367845415197715,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057973921765296,
      "loss": 2.5643,
      "theoretical_loss": 3.4367748982700332,
      "tokens_seen": 1955987456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057873620862588,
      "loss": 2.5149,
      "theoretical_loss": 3.436765255433854,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020577733199598797,
      "loss": 2.4682,
      "theoretical_loss": 3.436755613011203,
      "tokens_seen": 1956118528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020576730190571715,
      "loss": 2.4991,
      "theoretical_loss": 3.436745971002048,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1077107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.2328686714172363,
      "objective/train/theoretical_loss": 3.4367363294063575,
      "objective/train/tokens_used": 1976709600,
      "theoretical_loss": 3.4367363294063575,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020575727181544633,
      "loss": 2.4975,
      "theoretical_loss": 3.4367363294063575,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020574724172517554,
      "loss": 2.656,
      "theoretical_loss": 3.4367266882241,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020573721163490472,
      "loss": 2.7624,
      "theoretical_loss": 3.436717047455245,
      "tokens_seen": 1956380672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057271815446339,
      "loss": 2.6416,
      "theoretical_loss": 3.436707407099759,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020571715145436308,
      "loss": 2.5811,
      "theoretical_loss": 3.4366977671576118,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057071213640923,
      "loss": 2.4292,
      "theoretical_loss": 3.436688127628771,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020569709127382147,
      "loss": 2.3494,
      "theoretical_loss": 3.4366784885132056,
      "tokens_seen": 1956642816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020568706118355066,
      "loss": 2.5874,
      "theoretical_loss": 3.436668849810884,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020567703109327984,
      "loss": 2.6678,
      "theoretical_loss": 3.436659211521774,
      "tokens_seen": 1956773888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020566700100300902,
      "loss": 2.5681,
      "theoretical_loss": 3.4366495736458456,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020565697091273823,
      "loss": 2.4793,
      "theoretical_loss": 3.4366399361830657,
      "tokens_seen": 1956904960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056469408224674,
      "loss": 2.5004,
      "theoretical_loss": 3.436630299133403,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056369107321966,
      "loss": 2.574,
      "theoretical_loss": 3.4366206624968267,
      "tokens_seen": 1957036032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020562688064192577,
      "loss": 2.5185,
      "theoretical_loss": 3.4366110262733045,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020561685055165498,
      "loss": 2.5279,
      "theoretical_loss": 3.4366013904628048,
      "tokens_seen": 1957167104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020560682046138416,
      "loss": 2.6886,
      "theoretical_loss": 3.4365917550652973,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020559679037111334,
      "loss": 2.4815,
      "theoretical_loss": 3.4365821200807494,
      "tokens_seen": 1957298176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020558676028084252,
      "loss": 2.5116,
      "theoretical_loss": 3.4365724855091293,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002055767301905717,
      "loss": 2.4904,
      "theoretical_loss": 3.4365628513504065,
      "tokens_seen": 1957429248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002055667001003009,
      "loss": 2.5409,
      "theoretical_loss": 3.4365532176045486,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002055566700100301,
      "loss": 2.3362,
      "theoretical_loss": 3.436543584271524,
      "tokens_seen": 1957560320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020554663991975927,
      "loss": 2.7036,
      "theoretical_loss": 3.4365339513513025,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020553660982948845,
      "loss": 2.6681,
      "theoretical_loss": 3.4365243188438512,
      "tokens_seen": 1957691392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020552657973921766,
      "loss": 2.4317,
      "theoretical_loss": 3.436514686749139,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020551654964894684,
      "loss": 2.3854,
      "theoretical_loss": 3.4365050550671348,
      "tokens_seen": 1957822464
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1078288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.646674156188965,
      "objective/train/theoretical_loss": 3.4364954237978065,
      "objective/train/tokens_used": 1978348000,
      "theoretical_loss": 3.4364954237978065,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020550651955867602,
      "loss": 2.575,
      "theoretical_loss": 3.4364954237978065,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054964894684052,
      "loss": 2.6456,
      "theoretical_loss": 3.436485792941123,
      "tokens_seen": 1957953536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020548645937813439,
      "loss": 2.557,
      "theoretical_loss": 3.436476162497053,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054764292878636,
      "loss": 2.6655,
      "theoretical_loss": 3.4364665324655643,
      "tokens_seen": 1958084608
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020546639919759278,
      "loss": 2.4246,
      "theoretical_loss": 3.4364569028466265,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020545636910732196,
      "loss": 2.6148,
      "theoretical_loss": 3.4364472736402067,
      "tokens_seen": 1958215680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020544633901705114,
      "loss": 2.5243,
      "theoretical_loss": 3.436437644846275,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020543630892678035,
      "loss": 2.5258,
      "theoretical_loss": 3.4364280164647987,
      "tokens_seen": 1958346752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020542627883650953,
      "loss": 2.5842,
      "theoretical_loss": 3.4364183884957464,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054162487462387,
      "loss": 2.4366,
      "theoretical_loss": 3.4364087609390874,
      "tokens_seen": 1958477824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054062186559679,
      "loss": 2.599,
      "theoretical_loss": 3.43639913379479,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053961885656971,
      "loss": 2.6946,
      "theoretical_loss": 3.436389507062822,
      "tokens_seen": 1958608896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020538615847542628,
      "loss": 2.5156,
      "theoretical_loss": 3.4363798807431527,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020537612838515546,
      "loss": 2.6081,
      "theoretical_loss": 3.436370254835751,
      "tokens_seen": 1958739968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020536609829488464,
      "loss": 2.5615,
      "theoretical_loss": 3.4363606293405837,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020535606820461382,
      "loss": 2.4083,
      "theoretical_loss": 3.4363510042576215,
      "tokens_seen": 1958871040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020534603811434303,
      "loss": 2.3934,
      "theoretical_loss": 3.436341379586832,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053360080240722,
      "loss": 2.6675,
      "theoretical_loss": 3.4363317553281836,
      "tokens_seen": 1959002112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020532597793380142,
      "loss": 2.5839,
      "theoretical_loss": 3.436322131481645,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053159478435306,
      "loss": 2.5786,
      "theoretical_loss": 3.436312508047185,
      "tokens_seen": 1959133184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053059177532598,
      "loss": 2.7457,
      "theoretical_loss": 3.4363028850247717,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000205295887662989,
      "loss": 2.7012,
      "theoretical_loss": 3.436293262414374,
      "tokens_seen": 1959264256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020528585757271817,
      "loss": 2.367,
      "theoretical_loss": 3.4362836402159607,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020527582748244735,
      "loss": 2.5572,
      "theoretical_loss": 3.4362740184294998,
      "tokens_seen": 1959395328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020526579739217653,
      "loss": 2.4684,
      "theoretical_loss": 3.43626439705496,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1078935,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.1671700477600098,
      "objective/train/theoretical_loss": 3.4362547760923103,
      "objective/train/tokens_used": 1979986400,
      "theoretical_loss": 3.4362547760923103,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020525576730190574,
      "loss": 2.408,
      "theoretical_loss": 3.4362547760923103,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020524573721163492,
      "loss": 2.5545,
      "theoretical_loss": 3.4362451555415197,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052357071213641,
      "loss": 2.2538,
      "theoretical_loss": 3.436235535402555,
      "tokens_seen": 1959657472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020522567703109328,
      "loss": 2.4747,
      "theoretical_loss": 3.4362259156753874,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052156469408225,
      "loss": 2.5882,
      "theoretical_loss": 3.4362162963599827,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020520561685055167,
      "loss": 2.4402,
      "theoretical_loss": 3.436206677456312,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020519558676028086,
      "loss": 2.6005,
      "theoretical_loss": 3.436197058964342,
      "tokens_seen": 1959919616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020518555667001004,
      "loss": 2.7186,
      "theoretical_loss": 3.4361874408840425,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020517552657973922,
      "loss": 2.48,
      "theoretical_loss": 3.4361778232153815,
      "tokens_seen": 1960050688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020516549648946843,
      "loss": 2.4153,
      "theoretical_loss": 3.436168205958328,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002051554663991976,
      "loss": 2.8025,
      "theoretical_loss": 3.4361585891128508,
      "tokens_seen": 1960181760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002051454363089268,
      "loss": 2.4854,
      "theoretical_loss": 3.436148972678917,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020513540621865597,
      "loss": 2.5184,
      "theoretical_loss": 3.436139356656498,
      "tokens_seen": 1960312832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020512537612838518,
      "loss": 2.4815,
      "theoretical_loss": 3.43612974104556,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020511534603811436,
      "loss": 2.5678,
      "theoretical_loss": 3.436120125846072,
      "tokens_seen": 1960443904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020510531594784354,
      "loss": 2.3285,
      "theoretical_loss": 3.436110511058004,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020509528585757272,
      "loss": 2.6859,
      "theoretical_loss": 3.4361008966813236,
      "tokens_seen": 1960574976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050852557673019,
      "loss": 2.4694,
      "theoretical_loss": 3.4360912827159997,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050752256770311,
      "loss": 2.4015,
      "theoretical_loss": 3.4360816691620006,
      "tokens_seen": 1960706048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050651955867603,
      "loss": 2.7856,
      "theoretical_loss": 3.436072056019295,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020505516549648947,
      "loss": 2.4995,
      "theoretical_loss": 3.4360624432878524,
      "tokens_seen": 1960837120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020504513540621865,
      "loss": 2.7102,
      "theoretical_loss": 3.4360528309676406,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020503510531594786,
      "loss": 2.3589,
      "theoretical_loss": 3.4360432190586288,
      "tokens_seen": 1960968192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020502507522567704,
      "loss": 2.5998,
      "theoretical_loss": 3.436033607560785,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020501504513540622,
      "loss": 2.5773,
      "theoretical_loss": 3.4360239964740784,
      "tokens_seen": 1961099264
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1080156,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.784773588180542,
      "objective/train/theoretical_loss": 3.4360143857984777,
      "objective/train/tokens_used": 1981624800,
      "theoretical_loss": 3.4360143857984777,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050050150451354,
      "loss": 2.6701,
      "theoretical_loss": 3.4360143857984777,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020499498495486459,
      "loss": 2.6393,
      "theoretical_loss": 3.4360047755339513,
      "tokens_seen": 1961230336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049849548645938,
      "loss": 2.5664,
      "theoretical_loss": 3.4359951656804677,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020497492477432298,
      "loss": 2.5078,
      "theoretical_loss": 3.4359855562379957,
      "tokens_seen": 1961361408
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020496489468405216,
      "loss": 2.3168,
      "theoretical_loss": 3.435975947206505,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020495486459378134,
      "loss": 2.4741,
      "theoretical_loss": 3.435966338585963,
      "tokens_seen": 1961492480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020494483450351055,
      "loss": 2.7658,
      "theoretical_loss": 3.4359567303763385,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020493480441323973,
      "loss": 2.5477,
      "theoretical_loss": 3.435947122577601,
      "tokens_seen": 1961623552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049247743229689,
      "loss": 2.4397,
      "theoretical_loss": 3.4359375151897185,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049147442326981,
      "loss": 2.2555,
      "theoretical_loss": 3.43592790821266,
      "tokens_seen": 1961754624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049047141424273,
      "loss": 2.6149,
      "theoretical_loss": 3.4359183016463946,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020489468405215648,
      "loss": 2.514,
      "theoretical_loss": 3.43590869549089,
      "tokens_seen": 1961885696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020488465396188566,
      "loss": 2.6039,
      "theoretical_loss": 3.435899089746116,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020487462387161484,
      "loss": 2.6722,
      "theoretical_loss": 3.4358894844120407,
      "tokens_seen": 1962016768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020486459378134402,
      "loss": 2.6993,
      "theoretical_loss": 3.435879879488633,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020485456369107323,
      "loss": 2.5826,
      "theoretical_loss": 3.435870274975861,
      "tokens_seen": 1962147840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048445336008024,
      "loss": 2.5479,
      "theoretical_loss": 3.4358606708736943,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048345035105316,
      "loss": 2.6213,
      "theoretical_loss": 3.435851067182101,
      "tokens_seen": 1962278912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020482447342026077,
      "loss": 2.3191,
      "theoretical_loss": 3.435841463901051,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020481444332998998,
      "loss": 2.5946,
      "theoretical_loss": 3.435831861030512,
      "tokens_seen": 1962409984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020480441323971916,
      "loss": 2.49,
      "theoretical_loss": 3.4358222585704525,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020479438314944834,
      "loss": 2.6773,
      "theoretical_loss": 3.4358126565208416,
      "tokens_seen": 1962541056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020478435305917753,
      "loss": 2.5266,
      "theoretical_loss": 3.4358030548816485,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047743229689067,
      "loss": 2.6111,
      "theoretical_loss": 3.435793453652842,
      "tokens_seen": 1962672128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020476429287863591,
      "loss": 2.4898,
      "theoretical_loss": 3.4357838528343896,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.59,
      "objective/train/docs_used": 1081069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3030667304992676,
      "objective/train/theoretical_loss": 3.4357742524262616,
      "objective/train/tokens_used": 1983263200,
      "theoretical_loss": 3.4357742524262616,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047542627883651,
      "loss": 2.5752,
      "theoretical_loss": 3.4357742524262616,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020474423269809428,
      "loss": 2.4142,
      "theoretical_loss": 3.4357646524284258,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020473420260782346,
      "loss": 2.5524,
      "theoretical_loss": 3.435755052840851,
      "tokens_seen": 1962934272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020472417251755267,
      "loss": 2.4496,
      "theoretical_loss": 3.4357454536635066,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020471414242728185,
      "loss": 2.6205,
      "theoretical_loss": 3.435735854896361,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020470411233701103,
      "loss": 2.7057,
      "theoretical_loss": 3.4357262565393825,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046940822467402,
      "loss": 2.7622,
      "theoretical_loss": 3.4357166585925407,
      "tokens_seen": 1963196416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046840521564694,
      "loss": 2.4942,
      "theoretical_loss": 3.4357070610558043,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046740220661986,
      "loss": 2.4942,
      "theoretical_loss": 3.4356974639291415,
      "tokens_seen": 1963327488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020466399197592778,
      "loss": 2.608,
      "theoretical_loss": 3.4356878672125215,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020465396188565696,
      "loss": 2.7798,
      "theoretical_loss": 3.435678270905913,
      "tokens_seen": 1963458560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020464393179538614,
      "loss": 2.7324,
      "theoretical_loss": 3.4356686750092846,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020463390170511535,
      "loss": 2.3989,
      "theoretical_loss": 3.4356590795226056,
      "tokens_seen": 1963589632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020462387161484453,
      "loss": 2.6631,
      "theoretical_loss": 3.4356494844458445,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002046138415245737,
      "loss": 2.6547,
      "theoretical_loss": 3.43563988977897,
      "tokens_seen": 1963720704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002046038114343029,
      "loss": 2.5636,
      "theoretical_loss": 3.435630295521951,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020459378134403207,
      "loss": 2.5722,
      "theoretical_loss": 3.4356207016747566,
      "tokens_seen": 1963851776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020458375125376128,
      "loss": 2.6443,
      "theoretical_loss": 3.4356111082373553,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045737211634905,
      "loss": 2.4593,
      "theoretical_loss": 3.435601515209716,
      "tokens_seen": 1963982848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020456369107321967,
      "loss": 2.3981,
      "theoretical_loss": 3.435591922591807,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020455366098294885,
      "loss": 2.6099,
      "theoretical_loss": 3.435582330383598,
      "tokens_seen": 1964113920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020454363089267806,
      "loss": 2.5885,
      "theoretical_loss": 3.4355727385850576,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020453360080240724,
      "loss": 2.6934,
      "theoretical_loss": 3.435563147196155,
      "tokens_seen": 1964244992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020452357071213642,
      "loss": 2.7076,
      "theoretical_loss": 3.435553556216858,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045135406218656,
      "loss": 2.3334,
      "theoretical_loss": 3.4355439656471356,
      "tokens_seen": 1964376064
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1081685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8482184410095215,
      "objective/train/theoretical_loss": 3.4355343754869576,
      "objective/train/tokens_used": 1984901600,
      "theoretical_loss": 3.4355343754869576,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020450351053159479,
      "loss": 2.7198,
      "theoretical_loss": 3.4355343754869576,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000204493480441324,
      "loss": 2.719,
      "theoretical_loss": 3.435524785736292,
      "tokens_seen": 1964507136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020448345035105318,
      "loss": 2.3114,
      "theoretical_loss": 3.4355151963951083,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020447342026078236,
      "loss": 2.5342,
      "theoretical_loss": 3.4355056074633747,
      "tokens_seen": 1964638208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020446339017051154,
      "loss": 2.6172,
      "theoretical_loss": 3.4354960189410604,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020445336008024075,
      "loss": 2.4908,
      "theoretical_loss": 3.435486430828134,
      "tokens_seen": 1964769280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020444332998996993,
      "loss": 2.4997,
      "theoretical_loss": 3.4354768431245652,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044332998996991,
      "loss": 2.5456,
      "theoretical_loss": 3.435467255830322,
      "tokens_seen": 1964900352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044232698094283,
      "loss": 2.6273,
      "theoretical_loss": 3.4354576689453733,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044132397191575,
      "loss": 2.5549,
      "theoretical_loss": 3.4354480824696885,
      "tokens_seen": 1965031424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020440320962888668,
      "loss": 2.7147,
      "theoretical_loss": 3.4354384964032363,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020439317953861586,
      "loss": 2.484,
      "theoretical_loss": 3.4354289107459848,
      "tokens_seen": 1965162496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020438314944834504,
      "loss": 2.595,
      "theoretical_loss": 3.435419325497904,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020437311935807422,
      "loss": 2.4785,
      "theoretical_loss": 3.4354097406589625,
      "tokens_seen": 1965293568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020436308926780343,
      "loss": 2.3352,
      "theoretical_loss": 3.4354001562291288,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002043530591775326,
      "loss": 2.6406,
      "theoretical_loss": 3.435390572208372,
      "tokens_seen": 1965424640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002043430290872618,
      "loss": 2.634,
      "theoretical_loss": 3.4353809885966613,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020433299899699097,
      "loss": 2.2274,
      "theoretical_loss": 3.4353714053939655,
      "tokens_seen": 1965555712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020432296890672018,
      "loss": 2.4484,
      "theoretical_loss": 3.4353618226002527,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020431293881644936,
      "loss": 2.4279,
      "theoretical_loss": 3.435352240215493,
      "tokens_seen": 1965686784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020430290872617854,
      "loss": 2.5162,
      "theoretical_loss": 3.4353426582396542,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020429287863590773,
      "loss": 2.5951,
      "theoretical_loss": 3.4353330766727064,
      "tokens_seen": 1965817856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002042828485456369,
      "loss": 2.5068,
      "theoretical_loss": 3.4353234955146177,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020427281845536611,
      "loss": 2.5488,
      "theoretical_loss": 3.4353139147653575,
      "tokens_seen": 1965948928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002042627883650953,
      "loss": 2.528,
      "theoretical_loss": 3.4353043344248944,
      "tokens_seen": 1966014464
    },
    {
      "debugging/Self-BLEU-5": 0.3525225322618345,
      "debugging/distinct-1-grams": 0.8026223673467898,
      "debugging/distinct-2-grams": 0.9790296052631579,
      "debugging/entropy-1-grams": 5.169945565512759,
      "debugging/entropy-2-grams": 5.752330724154541,
      "debugging/length": 571.5,
      "debugging/num_segments": 4,
      "epoch": 0.6,
      "objective/train/docs_used": 1082856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.168870210647583,
      "objective/train/theoretical_loss": 3.4352947544931967,
      "objective/train/tokens_used": 1986540000,
      "theoretical_loss": 3.4352947544931967,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020425275827482448,
      "loss": 2.3803,
      "theoretical_loss": 3.4352947544931967,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020424272818455366,
      "loss": 2.3825,
      "theoretical_loss": 3.435285174970235,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020423269809428287,
      "loss": 2.2924,
      "theoretical_loss": 3.435275595855977,
      "tokens_seen": 1966211072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020422266800401205,
      "loss": 2.6049,
      "theoretical_loss": 3.4352660171503913,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020421263791374123,
      "loss": 2.7409,
      "theoretical_loss": 3.4352564388534477,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002042026078234704,
      "loss": 2.4515,
      "theoretical_loss": 3.4352468609651154,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041925777331996,
      "loss": 2.6473,
      "theoretical_loss": 3.435237283485362,
      "tokens_seen": 1966473216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041825476429288,
      "loss": 2.4047,
      "theoretical_loss": 3.4352277064141585,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020417251755265798,
      "loss": 2.5946,
      "theoretical_loss": 3.4352181297514717,
      "tokens_seen": 1966604288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020416248746238716,
      "loss": 2.6541,
      "theoretical_loss": 3.4352085534972723,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020415245737211634,
      "loss": 2.5395,
      "theoretical_loss": 3.4351989776515275,
      "tokens_seen": 1966735360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020414242728184555,
      "loss": 2.5426,
      "theoretical_loss": 3.435189402214208,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020413239719157473,
      "loss": 2.4257,
      "theoretical_loss": 3.435179827185282,
      "tokens_seen": 1966866432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041223671013039,
      "loss": 2.5935,
      "theoretical_loss": 3.4351702525647183,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041123370110331,
      "loss": 2.4648,
      "theoretical_loss": 3.435160678352486,
      "tokens_seen": 1966997504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020410230692076228,
      "loss": 2.4312,
      "theoretical_loss": 3.4351511045485545,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020409227683049148,
      "loss": 2.5017,
      "theoretical_loss": 3.435141531152892,
      "tokens_seen": 1967128576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020408224674022066,
      "loss": 2.4814,
      "theoretical_loss": 3.4351319581654685,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020407221664994985,
      "loss": 2.4076,
      "theoretical_loss": 3.4351223855862525,
      "tokens_seen": 1967259648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020406218655967903,
      "loss": 2.4142,
      "theoretical_loss": 3.4351128134152122,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020405215646940823,
      "loss": 2.532,
      "theoretical_loss": 3.435103241652318,
      "tokens_seen": 1967390720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020404212637913742,
      "loss": 2.5608,
      "theoretical_loss": 3.4350936702975376,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002040320962888666,
      "loss": 2.7132,
      "theoretical_loss": 3.435084099350841,
      "tokens_seen": 1967521792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020402206619859578,
      "loss": 2.3899,
      "theoretical_loss": 3.435074528812197,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020401203610832496,
      "loss": 2.5347,
      "theoretical_loss": 3.4350649586815742,
      "tokens_seen": 1967652864
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1083372,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5426712036132812,
      "objective/train/theoretical_loss": 3.435055388958942,
      "objective/train/tokens_used": 1988178400,
      "theoretical_loss": 3.435055388958942,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020400200601805417,
      "loss": 2.524,
      "theoretical_loss": 3.435055388958942,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020399197592778335,
      "loss": 2.6715,
      "theoretical_loss": 3.4350458196442695,
      "tokens_seen": 1967783936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020398194583751253,
      "loss": 2.3697,
      "theoretical_loss": 3.435036250737525,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002039719157472417,
      "loss": 2.6826,
      "theoretical_loss": 3.4350266822386786,
      "tokens_seen": 1967915008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020396188565697092,
      "loss": 2.6292,
      "theoretical_loss": 3.4350171141476986,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002039518555667001,
      "loss": 2.2238,
      "theoretical_loss": 3.435007546464554,
      "tokens_seen": 1968046080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020394182547642928,
      "loss": 2.5895,
      "theoretical_loss": 3.4349979791892142,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020393179538615846,
      "loss": 2.6439,
      "theoretical_loss": 3.434988412321648,
      "tokens_seen": 1968177152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020392176529588764,
      "loss": 2.5759,
      "theoretical_loss": 3.4349788458618247,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020391173520561685,
      "loss": 2.3609,
      "theoretical_loss": 3.434969279809713,
      "tokens_seen": 1968308224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020390170511534603,
      "loss": 2.7094,
      "theoretical_loss": 3.434959714165282,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020389167502507521,
      "loss": 2.5695,
      "theoretical_loss": 3.4349501489285013,
      "tokens_seen": 1968439296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002038816449348044,
      "loss": 2.5223,
      "theoretical_loss": 3.4349405840993392,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002038716148445336,
      "loss": 2.5507,
      "theoretical_loss": 3.434931019677765,
      "tokens_seen": 1968570368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020386158475426278,
      "loss": 2.417,
      "theoretical_loss": 3.434921455663748,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020385155466399197,
      "loss": 2.5875,
      "theoretical_loss": 3.4349118920572574,
      "tokens_seen": 1968701440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020384152457372115,
      "loss": 2.4222,
      "theoretical_loss": 3.4349023288582616,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020383149448345036,
      "loss": 2.6182,
      "theoretical_loss": 3.4348927660667306,
      "tokens_seen": 1968832512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020382146439317956,
      "loss": 2.665,
      "theoretical_loss": 3.4348832036826322,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020381143430290874,
      "loss": 2.6325,
      "theoretical_loss": 3.434873641705937,
      "tokens_seen": 1968963584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020380140421263793,
      "loss": 2.5463,
      "theoretical_loss": 3.434864080136612,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037913741223671,
      "loss": 2.5241,
      "theoretical_loss": 3.434854518974629,
      "tokens_seen": 1969094656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020378134403209631,
      "loss": 2.5659,
      "theoretical_loss": 3.434844958219955,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037713139418255,
      "loss": 2.4561,
      "theoretical_loss": 3.4348353978725603,
      "tokens_seen": 1969225728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020376128385155468,
      "loss": 2.4299,
      "theoretical_loss": 3.434825837932413,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1084308,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7147793769836426,
      "objective/train/theoretical_loss": 3.4348162783994827,
      "objective/train/tokens_used": 1989816800,
      "theoretical_loss": 3.4348162783994827,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020375125376128386,
      "loss": 2.4171,
      "theoretical_loss": 3.4348162783994827,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020374122367101307,
      "loss": 2.5049,
      "theoretical_loss": 3.4348067192737384,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020373119358074225,
      "loss": 2.6831,
      "theoretical_loss": 3.4347971605551493,
      "tokens_seen": 1969487872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020372116349047143,
      "loss": 2.5267,
      "theoretical_loss": 3.4347876022436843,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037111334002006,
      "loss": 2.7166,
      "theoretical_loss": 3.434778044339313,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037011033099298,
      "loss": 2.4133,
      "theoretical_loss": 3.434768486842005,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203691073219659,
      "loss": 2.6219,
      "theoretical_loss": 3.4347589297517276,
      "tokens_seen": 1969750016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020368104312938818,
      "loss": 2.5346,
      "theoretical_loss": 3.434749373068451,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020367101303911736,
      "loss": 2.4038,
      "theoretical_loss": 3.4347398167921446,
      "tokens_seen": 1969881088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020366098294884654,
      "loss": 2.4857,
      "theoretical_loss": 3.4347302609227772,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020365095285857575,
      "loss": 2.2692,
      "theoretical_loss": 3.434720705460318,
      "tokens_seen": 1970012160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020364092276830493,
      "loss": 2.4699,
      "theoretical_loss": 3.434711150404736,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036308926780341,
      "loss": 2.6742,
      "theoretical_loss": 3.4347015957560005,
      "tokens_seen": 1970143232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036208625877633,
      "loss": 2.6075,
      "theoretical_loss": 3.4346920415140803,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020361083249749248,
      "loss": 2.557,
      "theoretical_loss": 3.4346824876789452,
      "tokens_seen": 1970274304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020360080240722168,
      "loss": 2.4759,
      "theoretical_loss": 3.434672934250564,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020359077231695086,
      "loss": 2.6145,
      "theoretical_loss": 3.434663381228906,
      "tokens_seen": 1970405376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020358074222668005,
      "loss": 2.8568,
      "theoretical_loss": 3.434653828613939,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020357071213640923,
      "loss": 2.4004,
      "theoretical_loss": 3.434644276405635,
      "tokens_seen": 1970536448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020356068204613844,
      "loss": 2.5402,
      "theoretical_loss": 3.434634724603961,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020355065195586762,
      "loss": 2.4123,
      "theoretical_loss": 3.434625173208886,
      "tokens_seen": 1970667520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002035406218655968,
      "loss": 2.5613,
      "theoretical_loss": 3.43461562222038,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020353059177532598,
      "loss": 2.6472,
      "theoretical_loss": 3.4346060716384126,
      "tokens_seen": 1970798592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020352056168505516,
      "loss": 2.3054,
      "theoretical_loss": 3.434596521462952,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020351053159478437,
      "loss": 2.668,
      "theoretical_loss": 3.434586971693968,
      "tokens_seen": 1970929664
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1084897,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.094876527786255,
      "objective/train/theoretical_loss": 3.4345774223314294,
      "objective/train/tokens_used": 1991455200,
      "theoretical_loss": 3.4345774223314294,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020350050150451355,
      "loss": 2.2698,
      "theoretical_loss": 3.4345774223314294,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020349047141424273,
      "loss": 2.4867,
      "theoretical_loss": 3.434567873375306,
      "tokens_seen": 1971060736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002034804413239719,
      "loss": 2.4017,
      "theoretical_loss": 3.4345583248255656,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020347041123370112,
      "loss": 2.5838,
      "theoretical_loss": 3.4345487766821794,
      "tokens_seen": 1971191808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002034603811434303,
      "loss": 2.6591,
      "theoretical_loss": 3.434539228945115,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020345035105315948,
      "loss": 2.5973,
      "theoretical_loss": 3.4345296816143422,
      "tokens_seen": 1971322880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020344032096288866,
      "loss": 2.4645,
      "theoretical_loss": 3.43452013468983,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020343029087261784,
      "loss": 2.4124,
      "theoretical_loss": 3.4345105881715483,
      "tokens_seen": 1971453952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020342026078234705,
      "loss": 2.4656,
      "theoretical_loss": 3.4345010420594653,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020341023069207623,
      "loss": 2.6216,
      "theoretical_loss": 3.4344914963535507,
      "tokens_seen": 1971585024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020340020060180541,
      "loss": 2.6155,
      "theoretical_loss": 3.434481951053774,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033901705115346,
      "loss": 2.7607,
      "theoretical_loss": 3.4344724061601037,
      "tokens_seen": 1971716096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033801404212638,
      "loss": 2.4781,
      "theoretical_loss": 3.4344628616725097,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020337011033099298,
      "loss": 2.4515,
      "theoretical_loss": 3.434453317590961,
      "tokens_seen": 1971847168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020336008024072217,
      "loss": 2.3638,
      "theoretical_loss": 3.4344437739154268,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020335005015045135,
      "loss": 2.541,
      "theoretical_loss": 3.434434230645876,
      "tokens_seen": 1971978240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020334002006018056,
      "loss": 2.4235,
      "theoretical_loss": 3.434424687782279,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020332998996990974,
      "loss": 2.6314,
      "theoretical_loss": 3.4344151453246035,
      "tokens_seen": 1972109312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020331995987963892,
      "loss": 2.4461,
      "theoretical_loss": 3.4344056032728196,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033099297893681,
      "loss": 2.4569,
      "theoretical_loss": 3.4343960616268965,
      "tokens_seen": 1972240384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020329989969909728,
      "loss": 2.6027,
      "theoretical_loss": 3.4343865203868034,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032898696088265,
      "loss": 2.6124,
      "theoretical_loss": 3.4343769795525096,
      "tokens_seen": 1972371456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020327983951855567,
      "loss": 2.4756,
      "theoretical_loss": 3.4343674391239842,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020326980942828485,
      "loss": 2.4054,
      "theoretical_loss": 3.4343578991011965,
      "tokens_seen": 1972502528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020325977933801403,
      "loss": 2.563,
      "theoretical_loss": 3.4343483594841158,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1086100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5159900188446045,
      "objective/train/theoretical_loss": 3.4343388202727114,
      "objective/train/tokens_used": 1993093600,
      "theoretical_loss": 3.4343388202727114,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020324974924774324,
      "loss": 2.5106,
      "theoretical_loss": 3.4343388202727114,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020323971915747242,
      "loss": 2.6027,
      "theoretical_loss": 3.4343292814669524,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032296890672016,
      "loss": 2.4758,
      "theoretical_loss": 3.4343197430668084,
      "tokens_seen": 1972764672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020321965897693078,
      "loss": 2.6489,
      "theoretical_loss": 3.4343102050722485,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020320962888665996,
      "loss": 2.5082,
      "theoretical_loss": 3.434300667483242,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020319959879638917,
      "loss": 2.5294,
      "theoretical_loss": 3.434291130299758,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020318956870611835,
      "loss": 2.6525,
      "theoretical_loss": 3.4342815935217663,
      "tokens_seen": 1973026816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020317953861584753,
      "loss": 2.3976,
      "theoretical_loss": 3.4342720571492356,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020316950852557672,
      "loss": 2.6528,
      "theoretical_loss": 3.434262521182135,
      "tokens_seen": 1973157888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020315947843530592,
      "loss": 2.4516,
      "theoretical_loss": 3.434252985620435,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031494483450351,
      "loss": 2.5379,
      "theoretical_loss": 3.4342434504641037,
      "tokens_seen": 1973288960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020313941825476429,
      "loss": 2.4406,
      "theoretical_loss": 3.4342339157131114,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020312938816449347,
      "loss": 2.5342,
      "theoretical_loss": 3.434224381367426,
      "tokens_seen": 1973420032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020311935807422265,
      "loss": 2.5982,
      "theoretical_loss": 3.434214847427018,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020310932798395186,
      "loss": 2.5049,
      "theoretical_loss": 3.434205313891857,
      "tokens_seen": 1973551104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020309929789368104,
      "loss": 2.507,
      "theoretical_loss": 3.434195780761911,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020308926780341022,
      "loss": 2.6114,
      "theoretical_loss": 3.4341862480371503,
      "tokens_seen": 1973682176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002030792377131394,
      "loss": 2.5662,
      "theoretical_loss": 3.434176715717544,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020306920762286864,
      "loss": 2.6803,
      "theoretical_loss": 3.434167183803061,
      "tokens_seen": 1973813248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020305917753259782,
      "loss": 2.5804,
      "theoretical_loss": 3.4341576522936714,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203049147442327,
      "loss": 2.6837,
      "theoretical_loss": 3.4341481211893443,
      "tokens_seen": 1973944320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020303911735205618,
      "loss": 2.6399,
      "theoretical_loss": 3.4341385904900483,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020302908726178536,
      "loss": 2.6051,
      "theoretical_loss": 3.4341290601957537,
      "tokens_seen": 1974075392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020301905717151457,
      "loss": 2.6293,
      "theoretical_loss": 3.4341195303064294,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020300902708124375,
      "loss": 2.5129,
      "theoretical_loss": 3.4341100008220447,
      "tokens_seen": 1974206464
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1087285,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4230730533599854,
      "objective/train/theoretical_loss": 3.4341004717425694,
      "objective/train/tokens_used": 1994732000,
      "theoretical_loss": 3.4341004717425694,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020299899699097293,
      "loss": 2.4992,
      "theoretical_loss": 3.4341004717425694,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002029889669007021,
      "loss": 2.5992,
      "theoretical_loss": 3.4340909430679725,
      "tokens_seen": 1974337536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020297893681043132,
      "loss": 2.6065,
      "theoretical_loss": 3.4340814147982233,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002029689067201605,
      "loss": 2.6607,
      "theoretical_loss": 3.4340718869332907,
      "tokens_seen": 1974468608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020295887662988968,
      "loss": 2.4895,
      "theoretical_loss": 3.4340623594731454,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020294884653961886,
      "loss": 2.6727,
      "theoretical_loss": 3.4340528324177555,
      "tokens_seen": 1974599680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020293881644934804,
      "loss": 2.6061,
      "theoretical_loss": 3.4340433057670907,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020292878635907725,
      "loss": 2.6989,
      "theoretical_loss": 3.434033779521121,
      "tokens_seen": 1974730752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020291875626880643,
      "loss": 2.4494,
      "theoretical_loss": 3.434024253679815,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020290872617853561,
      "loss": 2.4378,
      "theoretical_loss": 3.4340147282431426,
      "tokens_seen": 1974861824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028986960882648,
      "loss": 2.6928,
      "theoretical_loss": 3.434005203211073,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202888665997994,
      "loss": 2.6238,
      "theoretical_loss": 3.4339956785835755,
      "tokens_seen": 1974992896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020287863590772318,
      "loss": 2.647,
      "theoretical_loss": 3.4339861543606194,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020286860581745237,
      "loss": 2.5911,
      "theoretical_loss": 3.4339766305421744,
      "tokens_seen": 1975123968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020285857572718155,
      "loss": 2.3616,
      "theoretical_loss": 3.4339671071282094,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020284854563691073,
      "loss": 2.8256,
      "theoretical_loss": 3.433957584118695,
      "tokens_seen": 1975255040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020283851554663994,
      "loss": 2.5468,
      "theoretical_loss": 3.4339480615135987,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020282848545636912,
      "loss": 2.7763,
      "theoretical_loss": 3.4339385393128916,
      "tokens_seen": 1975386112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028184553660983,
      "loss": 2.5894,
      "theoretical_loss": 3.4339290175165424,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020280842527582748,
      "loss": 2.5607,
      "theoretical_loss": 3.4339194961245205,
      "tokens_seen": 1975517184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027983951855567,
      "loss": 2.6292,
      "theoretical_loss": 3.4339099751367956,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020278836509528587,
      "loss": 2.7484,
      "theoretical_loss": 3.4339004545533367,
      "tokens_seen": 1975648256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020277833500501505,
      "loss": 2.5818,
      "theoretical_loss": 3.433890934374114,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020276830491474423,
      "loss": 2.4287,
      "theoretical_loss": 3.433881414599096,
      "tokens_seen": 1975779328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020275827482447344,
      "loss": 2.5168,
      "theoretical_loss": 3.4338718952282523,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1087922,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.908144950866699,
      "objective/train/theoretical_loss": 3.4338623762615526,
      "objective/train/tokens_used": 1996370400,
      "theoretical_loss": 3.4338623762615526,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020274824473420262,
      "loss": 2.7696,
      "theoretical_loss": 3.4338623762615526,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027382146439318,
      "loss": 2.6265,
      "theoretical_loss": 3.4338528576989664,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020272818455366098,
      "loss": 2.6013,
      "theoretical_loss": 3.433843339540463,
      "tokens_seen": 1976041472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020271815446339016,
      "loss": 2.8158,
      "theoretical_loss": 3.4338338217860116,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020270812437311937,
      "loss": 2.5984,
      "theoretical_loss": 3.4338243044355825,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020269809428284855,
      "loss": 2.5371,
      "theoretical_loss": 3.433814787489144,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020268806419257773,
      "loss": 2.6438,
      "theoretical_loss": 3.4338052709466664,
      "tokens_seen": 1976303616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020267803410230692,
      "loss": 2.5386,
      "theoretical_loss": 3.433795754808119,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020266800401203612,
      "loss": 2.814,
      "theoretical_loss": 3.4337862390734717,
      "tokens_seen": 1976434688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002026579739217653,
      "loss": 2.5689,
      "theoretical_loss": 3.4337767237426924,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020264794383149449,
      "loss": 2.6309,
      "theoretical_loss": 3.433767208815752,
      "tokens_seen": 1976565760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020263791374122367,
      "loss": 2.5405,
      "theoretical_loss": 3.43375769429262,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020262788365095285,
      "loss": 2.7246,
      "theoretical_loss": 3.4337481801732648,
      "tokens_seen": 1976696832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020261785356068206,
      "loss": 2.6009,
      "theoretical_loss": 3.4337386664576566,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020260782347041124,
      "loss": 2.6754,
      "theoretical_loss": 3.433729153145765,
      "tokens_seen": 1976827904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020259779338014042,
      "loss": 2.3616,
      "theoretical_loss": 3.433719640237559,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025877632898696,
      "loss": 2.7136,
      "theoretical_loss": 3.4337101277330087,
      "tokens_seen": 1976958976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025777331995988,
      "loss": 2.7344,
      "theoretical_loss": 3.433700615632083,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202567703109328,
      "loss": 2.7135,
      "theoretical_loss": 3.433691103934752,
      "tokens_seen": 1977090048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020255767301905717,
      "loss": 2.564,
      "theoretical_loss": 3.4336815926409843,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020254764292878635,
      "loss": 2.757,
      "theoretical_loss": 3.4336720817507507,
      "tokens_seen": 1977221120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020253761283851553,
      "loss": 2.5552,
      "theoretical_loss": 3.433662571264019,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020252758274824474,
      "loss": 2.519,
      "theoretical_loss": 3.4336530611807605,
      "tokens_seen": 1977352192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020251755265797392,
      "loss": 2.3865,
      "theoretical_loss": 3.4336435515009436,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025075225677031,
      "loss": 2.4351,
      "theoretical_loss": 3.433634042224538,
      "tokens_seen": 1977483264
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1089194,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.615063190460205,
      "objective/train/theoretical_loss": 3.433624533351513,
      "objective/train/tokens_used": 1998008800,
      "theoretical_loss": 3.433624533351513,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020249749247743228,
      "loss": 2.4549,
      "theoretical_loss": 3.433624533351513,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024874623871615,
      "loss": 2.5892,
      "theoretical_loss": 3.4336150248818393,
      "tokens_seen": 1977614336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020247743229689067,
      "loss": 2.6093,
      "theoretical_loss": 3.4336055168154846,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020246740220661985,
      "loss": 2.3286,
      "theoretical_loss": 3.43359600915242,
      "tokens_seen": 1977745408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020245737211634904,
      "loss": 2.4211,
      "theoretical_loss": 3.4335865018926146,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020244734202607822,
      "loss": 2.6886,
      "theoretical_loss": 3.433576995036037,
      "tokens_seen": 1977876480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020243731193580743,
      "loss": 2.8142,
      "theoretical_loss": 3.433567488582658,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024272818455366,
      "loss": 2.2784,
      "theoretical_loss": 3.433557982532447,
      "tokens_seen": 1978007552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024172517552658,
      "loss": 2.4793,
      "theoretical_loss": 3.433548476885372,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020240722166499497,
      "loss": 2.7134,
      "theoretical_loss": 3.433538971641405,
      "tokens_seen": 1978138624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020239719157472418,
      "loss": 2.5352,
      "theoretical_loss": 3.433529466800514,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020238716148445336,
      "loss": 2.5487,
      "theoretical_loss": 3.433519962362668,
      "tokens_seen": 1978269696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020237713139418254,
      "loss": 2.4761,
      "theoretical_loss": 3.433510458327838,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020236710130391172,
      "loss": 2.6174,
      "theoretical_loss": 3.433500954695993,
      "tokens_seen": 1978400768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023570712136409,
      "loss": 2.6328,
      "theoretical_loss": 3.4334914514671024,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023470411233701,
      "loss": 2.6043,
      "theoretical_loss": 3.433481948641136,
      "tokens_seen": 1978531840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023370110330993,
      "loss": 2.6289,
      "theoretical_loss": 3.433472446218063,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023269809428285,
      "loss": 2.5472,
      "theoretical_loss": 3.4334629441978537,
      "tokens_seen": 1978662912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020231695085255768,
      "loss": 2.6087,
      "theoretical_loss": 3.4334534425804764,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023069207622869,
      "loss": 2.6761,
      "theoretical_loss": 3.433443941365902,
      "tokens_seen": 1978793984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020229689067201607,
      "loss": 2.6352,
      "theoretical_loss": 3.4334344405541,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020228686058174525,
      "loss": 2.6847,
      "theoretical_loss": 3.4334249401450387,
      "tokens_seen": 1978925056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020227683049147443,
      "loss": 2.4965,
      "theoretical_loss": 3.433415440138689,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020226680040120364,
      "loss": 2.7385,
      "theoretical_loss": 3.43340594053502,
      "tokens_seen": 1979056128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020225677031093282,
      "loss": 2.5433,
      "theoretical_loss": 3.4333964413340015,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1090532,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8386666774749756,
      "objective/train/theoretical_loss": 3.4333869425356025,
      "objective/train/tokens_used": 1999647200,
      "theoretical_loss": 3.4333869425356025,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202246740220662,
      "loss": 2.6262,
      "theoretical_loss": 3.4333869425356025,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020223671013039118,
      "loss": 2.5365,
      "theoretical_loss": 3.4333774441397935,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020222668004012036,
      "loss": 2.8644,
      "theoretical_loss": 3.4333679461465434,
      "tokens_seen": 1979318272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020221664994984957,
      "loss": 2.4381,
      "theoretical_loss": 3.433358448555822,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020220661985957875,
      "loss": 2.5834,
      "theoretical_loss": 3.4333489513675994,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020219658976930793,
      "loss": 2.6442,
      "theoretical_loss": 3.4333394545818443,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020218655967903712,
      "loss": 2.371,
      "theoretical_loss": 3.4333299581985273,
      "tokens_seen": 1979580416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020217652958876632,
      "loss": 2.6461,
      "theoretical_loss": 3.433320462217617,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002021664994984955,
      "loss": 2.448,
      "theoretical_loss": 3.433310966639084,
      "tokens_seen": 1979711488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020215646940822469,
      "loss": 2.6503,
      "theoretical_loss": 3.4333014714628973,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020214643931795387,
      "loss": 2.3793,
      "theoretical_loss": 3.433291976689027,
      "tokens_seen": 1979842560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020213640922768305,
      "loss": 2.5467,
      "theoretical_loss": 3.433282482317442,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020212637913741226,
      "loss": 2.6453,
      "theoretical_loss": 3.4332729883481123,
      "tokens_seen": 1979973632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020211634904714144,
      "loss": 2.5034,
      "theoretical_loss": 3.4332634947810083,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020210631895687062,
      "loss": 2.6764,
      "theoretical_loss": 3.4332540016160986,
      "tokens_seen": 1980104704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002020962888665998,
      "loss": 2.3347,
      "theoretical_loss": 3.4332445088533534,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202086258776329,
      "loss": 2.4928,
      "theoretical_loss": 3.433235016492742,
      "tokens_seen": 1980235776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002020762286860582,
      "loss": 2.5399,
      "theoretical_loss": 3.433225524534235,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020206619859578737,
      "loss": 2.2966,
      "theoretical_loss": 3.4332160329778003,
      "tokens_seen": 1980366848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020205616850551655,
      "loss": 2.5203,
      "theoretical_loss": 3.433206541823409,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020204613841524573,
      "loss": 2.6431,
      "theoretical_loss": 3.4331970510710303,
      "tokens_seen": 1980497920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020203610832497494,
      "loss": 2.5681,
      "theoretical_loss": 3.433187560720634,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020202607823470412,
      "loss": 2.5804,
      "theoretical_loss": 3.4331780707721897,
      "tokens_seen": 1980628992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002020160481444333,
      "loss": 2.6615,
      "theoretical_loss": 3.433168581225667,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020200601805416248,
      "loss": 2.6872,
      "theoretical_loss": 3.433159092081036,
      "tokens_seen": 1980760064
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1091139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.503357410430908,
      "objective/train/theoretical_loss": 3.433149603338266,
      "objective/train/tokens_used": 2001285600,
      "theoretical_loss": 3.433149603338266,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002019959879638917,
      "loss": 2.6682,
      "theoretical_loss": 3.433149603338266,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020198595787362087,
      "loss": 2.637,
      "theoretical_loss": 3.433140114997326,
      "tokens_seen": 1980891136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020197592778335005,
      "loss": 2.3574,
      "theoretical_loss": 3.433130627058187,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020196589769307924,
      "loss": 2.6023,
      "theoretical_loss": 3.433121139520818,
      "tokens_seen": 1981022208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020195586760280842,
      "loss": 2.459,
      "theoretical_loss": 3.433111652385189,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020194583751253763,
      "loss": 2.5056,
      "theoretical_loss": 3.4331021656512695,
      "tokens_seen": 1981153280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002019358074222668,
      "loss": 2.4405,
      "theoretical_loss": 3.433092679319029,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201925777331996,
      "loss": 2.4631,
      "theoretical_loss": 3.4330831933884367,
      "tokens_seen": 1981284352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020191574724172517,
      "loss": 2.7019,
      "theoretical_loss": 3.4330737078594638,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020190571715145438,
      "loss": 2.6094,
      "theoretical_loss": 3.4330642227320793,
      "tokens_seen": 1981415424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020189568706118356,
      "loss": 2.5682,
      "theoretical_loss": 3.433054738006253,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020188565697091274,
      "loss": 2.6574,
      "theoretical_loss": 3.433045253681954,
      "tokens_seen": 1981546496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020187562688064192,
      "loss": 2.6213,
      "theoretical_loss": 3.4330357697591527,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018655967903711,
      "loss": 2.5855,
      "theoretical_loss": 3.4330262862378182,
      "tokens_seen": 1981677568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018555667001003,
      "loss": 2.7213,
      "theoretical_loss": 3.4330168031179213,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018455366098295,
      "loss": 2.7871,
      "theoretical_loss": 3.433007320399431,
      "tokens_seen": 1981808640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020183550651955867,
      "loss": 2.805,
      "theoretical_loss": 3.4329978380823167,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020182547642928785,
      "loss": 2.6677,
      "theoretical_loss": 3.4329883561665486,
      "tokens_seen": 1981939712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020181544633901706,
      "loss": 2.5842,
      "theoretical_loss": 3.432978874652097,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020180541624874624,
      "loss": 2.7074,
      "theoretical_loss": 3.4329693935389303,
      "tokens_seen": 1982070784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020179538615847542,
      "loss": 2.6692,
      "theoretical_loss": 3.4329599128270196,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002017853560682046,
      "loss": 2.516,
      "theoretical_loss": 3.432950432516334,
      "tokens_seen": 1982201856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020177532597793379,
      "loss": 2.6825,
      "theoretical_loss": 3.4329409526068426,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201765295887663,
      "loss": 2.4725,
      "theoretical_loss": 3.4329314730985163,
      "tokens_seen": 1982332928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020175526579739218,
      "loss": 2.6887,
      "theoretical_loss": 3.4329219939913247,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1091959,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4094791412353516,
      "objective/train/theoretical_loss": 3.4329125152852367,
      "objective/train/tokens_used": 2002924000,
      "theoretical_loss": 3.4329125152852367,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020174523570712136,
      "loss": 2.6745,
      "theoretical_loss": 3.4329125152852367,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020173520561685054,
      "loss": 2.7606,
      "theoretical_loss": 3.432903036980223,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020172517552657975,
      "loss": 2.5079,
      "theoretical_loss": 3.4328935590762533,
      "tokens_seen": 1982595072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020171514543630893,
      "loss": 2.485,
      "theoretical_loss": 3.4328840815732966,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002017051153460381,
      "loss": 2.5872,
      "theoretical_loss": 3.4328746044713236,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016950852557673,
      "loss": 2.662,
      "theoretical_loss": 3.4328651277703033,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016850551654965,
      "loss": 2.3874,
      "theoretical_loss": 3.4328556514702058,
      "tokens_seen": 1982857216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020167502507522568,
      "loss": 2.5845,
      "theoretical_loss": 3.432846175571001,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020166499498495486,
      "loss": 2.4019,
      "theoretical_loss": 3.432836700072659,
      "tokens_seen": 1982988288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020165496489468404,
      "loss": 2.6856,
      "theoretical_loss": 3.4328272249751492,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020164493480441322,
      "loss": 2.5427,
      "theoretical_loss": 3.432817750278441,
      "tokens_seen": 1983119360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020163490471414243,
      "loss": 2.7473,
      "theoretical_loss": 3.432808275982505,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016248746238716,
      "loss": 2.6846,
      "theoretical_loss": 3.4327988020873104,
      "tokens_seen": 1983250432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016148445336008,
      "loss": 2.405,
      "theoretical_loss": 3.4327893285928273,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020160481444332997,
      "loss": 2.4401,
      "theoretical_loss": 3.4327798554990254,
      "tokens_seen": 1983381504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020159478435305918,
      "loss": 2.6548,
      "theoretical_loss": 3.432770382805874,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020158475426278836,
      "loss": 2.632,
      "theoretical_loss": 3.4327609105133443,
      "tokens_seen": 1983512576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020157472417251757,
      "loss": 2.5423,
      "theoretical_loss": 3.432751438621405,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020156469408224675,
      "loss": 2.6624,
      "theoretical_loss": 3.4327419671300263,
      "tokens_seen": 1983643648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020155466399197593,
      "loss": 2.604,
      "theoretical_loss": 3.4327324960391783,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020154463390170514,
      "loss": 2.6191,
      "theoretical_loss": 3.43272302534883,
      "tokens_seen": 1983774720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020153460381143432,
      "loss": 2.6875,
      "theoretical_loss": 3.4327135550589514,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015245737211635,
      "loss": 2.5727,
      "theoretical_loss": 3.4327040851695134,
      "tokens_seen": 1983905792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020151454363089268,
      "loss": 2.7693,
      "theoretical_loss": 3.432694615680485,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015045135406219,
      "loss": 2.4202,
      "theoretical_loss": 3.4326851465918358,
      "tokens_seen": 1984036864
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1092471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3928778171539307,
      "objective/train/theoretical_loss": 3.4326756779035357,
      "objective/train/tokens_used": 2004562400,
      "theoretical_loss": 3.4326756779035357,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020149448345035107,
      "loss": 2.7747,
      "theoretical_loss": 3.4326756779035357,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020148445336008026,
      "loss": 2.6961,
      "theoretical_loss": 3.4326662096155554,
      "tokens_seen": 1984167936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020147442326980944,
      "loss": 2.3196,
      "theoretical_loss": 3.4326567417278637,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020146439317953862,
      "loss": 2.6437,
      "theoretical_loss": 3.4326472742404315,
      "tokens_seen": 1984299008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020145436308926783,
      "loss": 2.5044,
      "theoretical_loss": 3.432637807153228,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201444332998997,
      "loss": 2.8101,
      "theoretical_loss": 3.4326283404662226,
      "tokens_seen": 1984430080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002014343029087262,
      "loss": 2.5214,
      "theoretical_loss": 3.4326188741793864,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020142427281845537,
      "loss": 2.5631,
      "theoretical_loss": 3.432609408292689,
      "tokens_seen": 1984561152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020141424272818458,
      "loss": 2.4066,
      "theoretical_loss": 3.4325999428060987,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020140421263791376,
      "loss": 2.7546,
      "theoretical_loss": 3.432590477719587,
      "tokens_seen": 1984692224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020139418254764294,
      "loss": 2.5768,
      "theoretical_loss": 3.432581013033124,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020138415245737212,
      "loss": 2.6458,
      "theoretical_loss": 3.4325715487466777,
      "tokens_seen": 1984823296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013741223671013,
      "loss": 2.5834,
      "theoretical_loss": 3.43256208486022,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013640922768305,
      "loss": 2.5139,
      "theoretical_loss": 3.43255262137372,
      "tokens_seen": 1984954368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013540621865597,
      "loss": 2.75,
      "theoretical_loss": 3.4325431582871473,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020134403209628887,
      "loss": 2.4995,
      "theoretical_loss": 3.432533695600472,
      "tokens_seen": 1985085440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020133400200601805,
      "loss": 2.5506,
      "theoretical_loss": 3.4325242333136643,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020132397191574726,
      "loss": 2.4084,
      "theoretical_loss": 3.4325147714266944,
      "tokens_seen": 1985216512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020131394182547644,
      "loss": 2.7994,
      "theoretical_loss": 3.432505309939531,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020130391173520562,
      "loss": 2.6205,
      "theoretical_loss": 3.432495848852145,
      "tokens_seen": 1985347584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012938816449348,
      "loss": 2.7405,
      "theoretical_loss": 3.432486388164506,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020128385155466399,
      "loss": 2.634,
      "theoretical_loss": 3.4324769278765834,
      "tokens_seen": 1985478656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012738214643932,
      "loss": 2.6564,
      "theoretical_loss": 3.432467467988348,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020126379137412238,
      "loss": 2.6143,
      "theoretical_loss": 3.4324580084997693,
      "tokens_seen": 1985609728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020125376128385156,
      "loss": 2.6157,
      "theoretical_loss": 3.4324485494108177,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1093318,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4800171852111816,
      "objective/train/theoretical_loss": 3.4324390907214624,
      "objective/train/tokens_used": 2006200800,
      "theoretical_loss": 3.4324390907214624,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020124373119358074,
      "loss": 2.631,
      "theoretical_loss": 3.4324390907214624,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020123370110330995,
      "loss": 2.5682,
      "theoretical_loss": 3.4324296324316736,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020122367101303913,
      "loss": 2.604,
      "theoretical_loss": 3.432420174541422,
      "tokens_seen": 1985871872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012136409227683,
      "loss": 2.7056,
      "theoretical_loss": 3.4324107170506757,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012036108324975,
      "loss": 2.7432,
      "theoretical_loss": 3.4324012599594065,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011935807422267,
      "loss": 2.7151,
      "theoretical_loss": 3.432391803267583,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020118355065195588,
      "loss": 2.7701,
      "theoretical_loss": 3.4323823469751766,
      "tokens_seen": 1986134016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020117352056168506,
      "loss": 2.5023,
      "theoretical_loss": 3.432372891082156,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020116349047141424,
      "loss": 2.5672,
      "theoretical_loss": 3.4323634355884916,
      "tokens_seen": 1986265088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020115346038114342,
      "loss": 2.6602,
      "theoretical_loss": 3.4323539804941534,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020114343029087263,
      "loss": 2.5116,
      "theoretical_loss": 3.432344525799111,
      "tokens_seen": 1986396160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011334002006018,
      "loss": 2.5102,
      "theoretical_loss": 3.432335071503335,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201123370110331,
      "loss": 2.7008,
      "theoretical_loss": 3.4323256176067947,
      "tokens_seen": 1986527232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020111334002006017,
      "loss": 2.6081,
      "theoretical_loss": 3.4323161641094604,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020110330992978938,
      "loss": 2.532,
      "theoretical_loss": 3.4323067110113024,
      "tokens_seen": 1986658304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020109327983951856,
      "loss": 2.7515,
      "theoretical_loss": 3.4322972583122904,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020108324974924774,
      "loss": 2.4806,
      "theoretical_loss": 3.4322878060123942,
      "tokens_seen": 1986789376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020107321965897692,
      "loss": 2.7918,
      "theoretical_loss": 3.4322783541115838,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010631895687061,
      "loss": 2.4772,
      "theoretical_loss": 3.432268902609829,
      "tokens_seen": 1986920448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020105315947843531,
      "loss": 2.5503,
      "theoretical_loss": 3.4322594515071003,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010431293881645,
      "loss": 2.8803,
      "theoretical_loss": 3.432250000803368,
      "tokens_seen": 1987051520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020103309929789368,
      "loss": 2.5987,
      "theoretical_loss": 3.4322405504986007,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020102306920762286,
      "loss": 2.5277,
      "theoretical_loss": 3.4322311005927704,
      "tokens_seen": 1987182592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020101303911735207,
      "loss": 2.4674,
      "theoretical_loss": 3.432221651085845,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020100300902708125,
      "loss": 2.6221,
      "theoretical_loss": 3.432212201977796,
      "tokens_seen": 1987313664
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1093806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8140084743499756,
      "objective/train/theoretical_loss": 3.4322027532685926,
      "objective/train/tokens_used": 2007839200,
      "theoretical_loss": 3.4322027532685926,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020099297893681043,
      "loss": 2.9088,
      "theoretical_loss": 3.4322027532685926,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009829488465396,
      "loss": 2.6867,
      "theoretical_loss": 3.432193304958205,
      "tokens_seen": 1987444736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009729187562688,
      "loss": 2.3103,
      "theoretical_loss": 3.4321838570466037,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200962888665998,
      "loss": 2.7303,
      "theoretical_loss": 3.432174409533758,
      "tokens_seen": 1987575808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020095285857572718,
      "loss": 2.6691,
      "theoretical_loss": 3.4321649624196384,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020094282848545636,
      "loss": 2.5858,
      "theoretical_loss": 3.432155515704215,
      "tokens_seen": 1987706880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020093279839518554,
      "loss": 2.5188,
      "theoretical_loss": 3.432146069387458,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020092276830491475,
      "loss": 2.5385,
      "theoretical_loss": 3.4321366234693365,
      "tokens_seen": 1987837952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020091273821464393,
      "loss": 2.4352,
      "theoretical_loss": 3.432127177949821,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009027081243731,
      "loss": 2.6184,
      "theoretical_loss": 3.432117732828882,
      "tokens_seen": 1987969024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002008926780341023,
      "loss": 2.4743,
      "theoretical_loss": 3.4321082881064884,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020088264794383147,
      "loss": 2.5702,
      "theoretical_loss": 3.4320988437826117,
      "tokens_seen": 1988100096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020087261785356068,
      "loss": 2.8154,
      "theoretical_loss": 3.4320893998572215,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020086258776328986,
      "loss": 2.5259,
      "theoretical_loss": 3.432079956330287,
      "tokens_seen": 1988231168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020085255767301905,
      "loss": 2.512,
      "theoretical_loss": 3.4320705132017792,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020084252758274823,
      "loss": 2.5852,
      "theoretical_loss": 3.4320610704716676,
      "tokens_seen": 1988362240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020083249749247743,
      "loss": 2.7334,
      "theoretical_loss": 3.4320516281399227,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020082246740220664,
      "loss": 2.2511,
      "theoretical_loss": 3.4320421862065147,
      "tokens_seen": 1988493312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020081243731193582,
      "loss": 2.6674,
      "theoretical_loss": 3.432032744671413,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200802407221665,
      "loss": 2.558,
      "theoretical_loss": 3.4320233035345877,
      "tokens_seen": 1988624384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020079237713139419,
      "loss": 2.6091,
      "theoretical_loss": 3.4320138627960097,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007823470411234,
      "loss": 2.8171,
      "theoretical_loss": 3.4320044224556483,
      "tokens_seen": 1988755456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020077231695085258,
      "loss": 2.7118,
      "theoretical_loss": 3.4319949825134737,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020076228686058176,
      "loss": 2.608,
      "theoretical_loss": 3.4319855429694566,
      "tokens_seen": 1988886528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020075225677031094,
      "loss": 2.6126,
      "theoretical_loss": 3.4319761038235663,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1094923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.71999192237854,
      "objective/train/theoretical_loss": 3.431966665075773,
      "objective/train/tokens_used": 2009477600,
      "theoretical_loss": 3.431966665075773,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020074222668004015,
      "loss": 2.6978,
      "theoretical_loss": 3.431966665075773,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020073219658976933,
      "loss": 2.4806,
      "theoretical_loss": 3.4319572267260474,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007221664994985,
      "loss": 2.4445,
      "theoretical_loss": 3.431947788774359,
      "tokens_seen": 1989148672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007121364092277,
      "loss": 2.6405,
      "theoretical_loss": 3.431938351220678,
      "tokens_seen": 1989214208
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007021063189569,
      "loss": 2.5439,
      "theoretical_loss": 3.431928914064975,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020069207622868608,
      "loss": 2.9438,
      "theoretical_loss": 3.4319194773072192,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020068204613841526,
      "loss": 2.579,
      "theoretical_loss": 3.4319100409473813,
      "tokens_seen": 1989410816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020067201604814444,
      "loss": 2.6039,
      "theoretical_loss": 3.4319006049854313,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020066198595787362,
      "loss": 2.3897,
      "theoretical_loss": 3.4318911694213394,
      "tokens_seen": 1989541888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020065195586760283,
      "loss": 2.4248,
      "theoretical_loss": 3.431881734255076,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200641925777332,
      "loss": 2.5772,
      "theoretical_loss": 3.431872299486611,
      "tokens_seen": 1989672960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002006318956870612,
      "loss": 2.4511,
      "theoretical_loss": 3.4318628651159138,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020062186559679037,
      "loss": 2.6177,
      "theoretical_loss": 3.4318534311429554,
      "tokens_seen": 1989804032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020061183550651958,
      "loss": 2.4933,
      "theoretical_loss": 3.4318439975677055,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020060180541624876,
      "loss": 2.4908,
      "theoretical_loss": 3.431834564390135,
      "tokens_seen": 1989935104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020059177532597794,
      "loss": 2.6168,
      "theoretical_loss": 3.4318251316102124,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020058174523570713,
      "loss": 2.501,
      "theoretical_loss": 3.4318156992279096,
      "tokens_seen": 1990066176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002005717151454363,
      "loss": 2.6036,
      "theoretical_loss": 3.4318062672431964,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020056168505516551,
      "loss": 2.2804,
      "theoretical_loss": 3.431796835656042,
      "tokens_seen": 1990197248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002005516549648947,
      "loss": 2.9173,
      "theoretical_loss": 3.431787404466417,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020054162487462388,
      "loss": 2.5834,
      "theoretical_loss": 3.431777973674292,
      "tokens_seen": 1990328320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020053159478435306,
      "loss": 2.3708,
      "theoretical_loss": 3.431768543279637,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020052156469408227,
      "loss": 2.6348,
      "theoretical_loss": 3.431759113282422,
      "tokens_seen": 1990459392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020051153460381145,
      "loss": 2.6176,
      "theoretical_loss": 3.431749683682617,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020050150451354063,
      "loss": 2.7718,
      "theoretical_loss": 3.4317402544801925,
      "tokens_seen": 1990590464
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1095442,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.825226068496704,
      "objective/train/theoretical_loss": 3.4317308256751184,
      "objective/train/tokens_used": 2011116000,
      "theoretical_loss": 3.4317308256751184,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004914744232698,
      "loss": 2.9433,
      "theoretical_loss": 3.4317308256751184,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200481444332999,
      "loss": 2.6397,
      "theoretical_loss": 3.4317213972673652,
      "tokens_seen": 1990721536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004714142427282,
      "loss": 2.5756,
      "theoretical_loss": 3.4317119692569023,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020046138415245738,
      "loss": 2.5589,
      "theoretical_loss": 3.431702541643701,
      "tokens_seen": 1990852608
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020045135406218656,
      "loss": 2.4167,
      "theoretical_loss": 3.431693114427731,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020044132397191574,
      "loss": 2.6904,
      "theoretical_loss": 3.4316836876089623,
      "tokens_seen": 1990983680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020043129388164495,
      "loss": 2.5691,
      "theoretical_loss": 3.431674261187365,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020042126379137413,
      "loss": 2.5874,
      "theoretical_loss": 3.4316648351629095,
      "tokens_seen": 1991114752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004112337011033,
      "loss": 2.7404,
      "theoretical_loss": 3.4316554095355665,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004012036108325,
      "loss": 2.6739,
      "theoretical_loss": 3.4316459843053053,
      "tokens_seen": 1991245824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020039117352056167,
      "loss": 2.7318,
      "theoretical_loss": 3.4316365594720963,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020038114343029088,
      "loss": 2.6483,
      "theoretical_loss": 3.43162713503591,
      "tokens_seen": 1991376896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020037111334002006,
      "loss": 2.6675,
      "theoretical_loss": 3.431617710996717,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020036108324974925,
      "loss": 2.6493,
      "theoretical_loss": 3.431608287354487,
      "tokens_seen": 1991507968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020035105315947843,
      "loss": 2.4218,
      "theoretical_loss": 3.4315988641091906,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020034102306920763,
      "loss": 2.9455,
      "theoretical_loss": 3.4315894412607966,
      "tokens_seen": 1991639040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020033099297893682,
      "loss": 2.4546,
      "theoretical_loss": 3.431580018809277,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200320962888666,
      "loss": 2.7482,
      "theoretical_loss": 3.431570596754601,
      "tokens_seen": 1991770112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020031093279839518,
      "loss": 2.6001,
      "theoretical_loss": 3.4315611750967396,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020030090270812436,
      "loss": 2.7418,
      "theoretical_loss": 3.431551753835662,
      "tokens_seen": 1991901184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020029087261785357,
      "loss": 2.5811,
      "theoretical_loss": 3.4315423329713397,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020028084252758275,
      "loss": 2.5115,
      "theoretical_loss": 3.4315329125037417,
      "tokens_seen": 1992032256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020027081243731193,
      "loss": 2.5093,
      "theoretical_loss": 3.4315234924328393,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002607823470411,
      "loss": 2.5344,
      "theoretical_loss": 3.431514072758602,
      "tokens_seen": 1992163328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020025075225677032,
      "loss": 2.6597,
      "theoretical_loss": 3.4315046534810003,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1096582,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.356750249862671,
      "objective/train/theoretical_loss": 3.4314952346000047,
      "objective/train/tokens_used": 2012754400,
      "theoretical_loss": 3.4314952346000047,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002407221664995,
      "loss": 2.5186,
      "theoretical_loss": 3.4314952346000047,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020023069207622868,
      "loss": 2.5944,
      "theoretical_loss": 3.4314858161155852,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020022066198595786,
      "loss": 2.7713,
      "theoretical_loss": 3.431476398027712,
      "tokens_seen": 1992425472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020021063189568704,
      "loss": 2.6936,
      "theoretical_loss": 3.431466980336355,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020020060180541625,
      "loss": 2.6147,
      "theoretical_loss": 3.4314575630414854,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020019057171514543,
      "loss": 2.6669,
      "theoretical_loss": 3.4314481461430724,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020018054162487461,
      "loss": 2.5813,
      "theoretical_loss": 3.4314387296410875,
      "tokens_seen": 1992687616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002001705115346038,
      "loss": 2.6617,
      "theoretical_loss": 3.4314293135355003,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200160481444333,
      "loss": 2.695,
      "theoretical_loss": 3.431419897826281,
      "tokens_seen": 1992818688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020015045135406218,
      "loss": 2.668,
      "theoretical_loss": 3.4314104825133995,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020014042126379137,
      "loss": 2.5273,
      "theoretical_loss": 3.431401067596827,
      "tokens_seen": 1992949760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020013039117352055,
      "loss": 2.703,
      "theoretical_loss": 3.431391653076533,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020012036108324973,
      "loss": 2.6377,
      "theoretical_loss": 3.431382238952488,
      "tokens_seen": 1993080832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020011033099297894,
      "loss": 2.9293,
      "theoretical_loss": 3.4313728252246634,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020010030090270812,
      "loss": 2.6579,
      "theoretical_loss": 3.4313634118930274,
      "tokens_seen": 1993211904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002000902708124373,
      "loss": 2.5968,
      "theoretical_loss": 3.431353998957552,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020008024072216648,
      "loss": 2.5687,
      "theoretical_loss": 3.4313445864182066,
      "tokens_seen": 1993342976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020007021063189571,
      "loss": 2.7676,
      "theoretical_loss": 3.4313351742749623,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002000601805416249,
      "loss": 2.5629,
      "theoretical_loss": 3.4313257625277886,
      "tokens_seen": 1993474048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020005015045135408,
      "loss": 2.7135,
      "theoretical_loss": 3.4313163511766565,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020004012036108326,
      "loss": 2.7204,
      "theoretical_loss": 3.4313069402215355,
      "tokens_seen": 1993605120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020003009027081247,
      "loss": 2.6783,
      "theoretical_loss": 3.431297529662397,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020002006018054165,
      "loss": 2.7427,
      "theoretical_loss": 3.43128811949921,
      "tokens_seen": 1993736192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020001003009027083,
      "loss": 2.6275,
      "theoretical_loss": 3.431278709731946,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 2.6253,
      "theoretical_loss": 3.431269300360575,
      "tokens_seen": 1993867264
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1097915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4239437580108643,
      "objective/train/theoretical_loss": 3.431259891385067,
      "objective/train/tokens_used": 2014392800,
      "theoretical_loss": 3.431259891385067,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999899699097292,
      "loss": 2.5314,
      "theoretical_loss": 3.431259891385067,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999799398194584,
      "loss": 2.5537,
      "theoretical_loss": 3.431250482805392,
      "tokens_seen": 1993998336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019996990972918758,
      "loss": 2.427,
      "theoretical_loss": 3.4312410746215214,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019995987963891676,
      "loss": 2.4997,
      "theoretical_loss": 3.4312316668334253,
      "tokens_seen": 1994129408
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019994984954864594,
      "loss": 2.5844,
      "theoretical_loss": 3.431222259441073,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019993981945837515,
      "loss": 2.4487,
      "theoretical_loss": 3.4312128524444363,
      "tokens_seen": 1994260480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019992978936810433,
      "loss": 2.6315,
      "theoretical_loss": 3.431203445843485,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999197592778335,
      "loss": 2.6465,
      "theoretical_loss": 3.4311940396381884,
      "tokens_seen": 1994391552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001999097291875627,
      "loss": 2.6503,
      "theoretical_loss": 3.4311846338285186,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019989969909729187,
      "loss": 2.6847,
      "theoretical_loss": 3.4311752284144443,
      "tokens_seen": 1994522624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019988966900702108,
      "loss": 2.7742,
      "theoretical_loss": 3.4311658233959372,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019987963891675026,
      "loss": 2.3197,
      "theoretical_loss": 3.4311564187729675,
      "tokens_seen": 1994653696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019986960882647945,
      "loss": 2.6643,
      "theoretical_loss": 3.431147014545505,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019985957873620863,
      "loss": 2.529,
      "theoretical_loss": 3.4311376107135203,
      "tokens_seen": 1994784768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019984954864593783,
      "loss": 2.6297,
      "theoretical_loss": 3.4311282072769833,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019983951855566702,
      "loss": 2.5736,
      "theoretical_loss": 3.4311188042358656,
      "tokens_seen": 1994915840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001998294884653962,
      "loss": 2.7233,
      "theoretical_loss": 3.4311094015901364,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019981945837512538,
      "loss": 2.571,
      "theoretical_loss": 3.431099999339767,
      "tokens_seen": 1995046912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019980942828485456,
      "loss": 2.402,
      "theoretical_loss": 3.431090597484727,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019979939819458377,
      "loss": 2.523,
      "theoretical_loss": 3.4310811960249867,
      "tokens_seen": 1995177984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019978936810431295,
      "loss": 2.481,
      "theoretical_loss": 3.431071794960517,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019977933801404213,
      "loss": 2.5778,
      "theoretical_loss": 3.4310623942912892,
      "tokens_seen": 1995309056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001997693079237713,
      "loss": 2.6169,
      "theoretical_loss": 3.431052994017272,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019975927783350052,
      "loss": 2.7874,
      "theoretical_loss": 3.431043594138436,
      "tokens_seen": 1995440128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001997492477432297,
      "loss": 2.4037,
      "theoretical_loss": 3.431034194654753,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.6,
      "objective/train/docs_used": 1098508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5603280067443848,
      "objective/train/theoretical_loss": 3.4310247955661923,
      "objective/train/tokens_used": 2016031200,
      "theoretical_loss": 3.4310247955661923,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019973921765295888,
      "loss": 2.7142,
      "theoretical_loss": 3.4310247955661923,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019972918756268806,
      "loss": 2.5046,
      "theoretical_loss": 3.4310153968727244,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019971915747241724,
      "loss": 2.5945,
      "theoretical_loss": 3.43100599857432,
      "tokens_seen": 1995702272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019970912738214645,
      "loss": 2.6585,
      "theoretical_loss": 3.4309966006709494,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019969909729187563,
      "loss": 2.7669,
      "theoretical_loss": 3.4309872031625828,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019968906720160481,
      "loss": 2.846,
      "theoretical_loss": 3.4309778060491913,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000199679037111334,
      "loss": 2.6425,
      "theoretical_loss": 3.4309684093307444,
      "tokens_seen": 1995964416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001996690070210632,
      "loss": 2.6876,
      "theoretical_loss": 3.430959013007213,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019965897693079238,
      "loss": 2.3786,
      "theoretical_loss": 3.430949617078568,
      "tokens_seen": 1996095488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019964894684052157,
      "loss": 2.5628,
      "theoretical_loss": 3.4309402215447795,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019963891675025075,
      "loss": 2.7516,
      "theoretical_loss": 3.430930826405817,
      "tokens_seen": 1996226560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019962888665997993,
      "loss": 2.8643,
      "theoretical_loss": 3.4309214316616528,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019961885656970914,
      "loss": 2.4167,
      "theoretical_loss": 3.4309120373122557,
      "tokens_seen": 1996357632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019960882647943832,
      "loss": 2.6299,
      "theoretical_loss": 3.4309026433575966,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001995987963891675,
      "loss": 2.6006,
      "theoretical_loss": 3.4308932497976468,
      "tokens_seen": 1996488704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019958876629889668,
      "loss": 2.709,
      "theoretical_loss": 3.4308838566323754,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001995787362086259,
      "loss": 2.5178,
      "theoretical_loss": 3.4308744638617537,
      "tokens_seen": 1996619776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019956870611835507,
      "loss": 2.6574,
      "theoretical_loss": 3.4308650714857523,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019955867602808425,
      "loss": 2.771,
      "theoretical_loss": 3.4308556795043414,
      "tokens_seen": 1996750848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019954864593781343,
      "loss": 2.5936,
      "theoretical_loss": 3.4308462879174915,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019953861584754264,
      "loss": 2.6743,
      "theoretical_loss": 3.4308368967251726,
      "tokens_seen": 1996881920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019952858575727182,
      "loss": 2.5113,
      "theoretical_loss": 3.4308275059273563,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199518555667001,
      "loss": 2.7091,
      "theoretical_loss": 3.430818115524012,
      "tokens_seen": 1997012992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019950852557673018,
      "loss": 2.4939,
      "theoretical_loss": 3.430808725515111,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019949849548645936,
      "loss": 2.7181,
      "theoretical_loss": 3.430799335900623,
      "tokens_seen": 1997144064
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1099395,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.039354085922241,
      "objective/train/theoretical_loss": 3.4307899466805187,
      "objective/train/tokens_used": 2017669600,
      "theoretical_loss": 3.4307899466805187,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019948846539618857,
      "loss": 2.8081,
      "theoretical_loss": 3.4307899466805187,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019947843530591775,
      "loss": 2.5789,
      "theoretical_loss": 3.430780557854769,
      "tokens_seen": 1997275136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019946840521564693,
      "loss": 2.601,
      "theoretical_loss": 3.4307711694233443,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019945837512537612,
      "loss": 2.5397,
      "theoretical_loss": 3.430761781386215,
      "tokens_seen": 1997406208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019944834503510532,
      "loss": 2.7686,
      "theoretical_loss": 3.430752393743351,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994383149448345,
      "loss": 2.5694,
      "theoretical_loss": 3.430743006494724,
      "tokens_seen": 1997537280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019942828485456369,
      "loss": 2.8073,
      "theoretical_loss": 3.4307336196403035,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019941825476429287,
      "loss": 2.6282,
      "theoretical_loss": 3.4307242331800607,
      "tokens_seen": 1997668352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019940822467402205,
      "loss": 2.4357,
      "theoretical_loss": 3.4307148471139657,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019939819458375126,
      "loss": 2.7161,
      "theoretical_loss": 3.430705461441989,
      "tokens_seen": 1997799424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019938816449348044,
      "loss": 2.4551,
      "theoretical_loss": 3.4306960761641014,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019937813440320962,
      "loss": 2.7165,
      "theoretical_loss": 3.4306866912802736,
      "tokens_seen": 1997930496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993681043129388,
      "loss": 2.7222,
      "theoretical_loss": 3.430677306790475,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199358074222668,
      "loss": 2.6222,
      "theoretical_loss": 3.4306679226946777,
      "tokens_seen": 1998061568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993480441323972,
      "loss": 2.541,
      "theoretical_loss": 3.430658538992851,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019933801404212637,
      "loss": 2.8188,
      "theoretical_loss": 3.430649155684966,
      "tokens_seen": 1998192640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019932798395185555,
      "loss": 2.6648,
      "theoretical_loss": 3.4306397727709936,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019931795386158476,
      "loss": 2.7879,
      "theoretical_loss": 3.4306303902509034,
      "tokens_seen": 1998323712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019930792377131397,
      "loss": 2.6134,
      "theoretical_loss": 3.4306210081246666,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019929789368104315,
      "loss": 2.7031,
      "theoretical_loss": 3.430611626392254,
      "tokens_seen": 1998454784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019928786359077233,
      "loss": 2.4931,
      "theoretical_loss": 3.4306022450536355,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992778335005015,
      "loss": 2.4753,
      "theoretical_loss": 3.4305928641087817,
      "tokens_seen": 1998585856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019926780341023072,
      "loss": 2.7274,
      "theoretical_loss": 3.4305834835576636,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992577733199599,
      "loss": 2.8174,
      "theoretical_loss": 3.430574103400252,
      "tokens_seen": 1998716928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019924774322968908,
      "loss": 2.8066,
      "theoretical_loss": 3.4305647236365164,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1100643,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.4168500900268555,
      "objective/train/theoretical_loss": 3.430555344266428,
      "objective/train/tokens_used": 2019308000,
      "theoretical_loss": 3.430555344266428,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019923771313941826,
      "loss": 2.4536,
      "theoretical_loss": 3.430555344266428,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019922768304914744,
      "loss": 2.8393,
      "theoretical_loss": 3.4305459652899577,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019921765295887665,
      "loss": 2.7206,
      "theoretical_loss": 3.4305365867070754,
      "tokens_seen": 1998979072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019920762286860583,
      "loss": 2.6495,
      "theoretical_loss": 3.430527208517752,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019919759277833501,
      "loss": 2.6067,
      "theoretical_loss": 3.4305178307219584,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991875626880642,
      "loss": 2.6372,
      "theoretical_loss": 3.430508453319665,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991775325977934,
      "loss": 2.5988,
      "theoretical_loss": 3.430499076310842,
      "tokens_seen": 1999241216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019916750250752258,
      "loss": 2.4068,
      "theoretical_loss": 3.4304896996954604,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019915747241725177,
      "loss": 2.5308,
      "theoretical_loss": 3.4304803234734904,
      "tokens_seen": 1999372288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019914744232698095,
      "loss": 2.5663,
      "theoretical_loss": 3.430470947644903,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019913741223671013,
      "loss": 2.6179,
      "theoretical_loss": 3.4304615722096687,
      "tokens_seen": 1999503360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019912738214643934,
      "loss": 2.4336,
      "theoretical_loss": 3.4304521971677584,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019911735205616852,
      "loss": 2.5062,
      "theoretical_loss": 3.430442822519142,
      "tokens_seen": 1999634432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991073219658977,
      "loss": 2.5648,
      "theoretical_loss": 3.4304334482637904,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019909729187562688,
      "loss": 2.4919,
      "theoretical_loss": 3.4304240744016745,
      "tokens_seen": 1999765504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001990872617853561,
      "loss": 2.7186,
      "theoretical_loss": 3.430414700932765,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019907723169508527,
      "loss": 2.6309,
      "theoretical_loss": 3.4304053278570317,
      "tokens_seen": 1999896576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019906720160481445,
      "loss": 2.5464,
      "theoretical_loss": 3.430395955174446,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019905717151454363,
      "loss": 2.8268,
      "theoretical_loss": 3.4303865828849784,
      "tokens_seen": 2000027648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019904714142427284,
      "loss": 2.5693,
      "theoretical_loss": 3.4303772109885995,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019903711133400202,
      "loss": 2.6107,
      "theoretical_loss": 3.43036783948528,
      "tokens_seen": 2000158720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001990270812437312,
      "loss": 2.6075,
      "theoretical_loss": 3.43035846837499,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019901705115346038,
      "loss": 2.8012,
      "theoretical_loss": 3.4303490976577007,
      "tokens_seen": 2000289792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019900702106318956,
      "loss": 2.5413,
      "theoretical_loss": 3.430339727333383,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019899699097291877,
      "loss": 2.7185,
      "theoretical_loss": 3.4303303574020063,
      "tokens_seen": 2000420864
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1101272,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.723714590072632,
      "objective/train/theoretical_loss": 3.4303209878635426,
      "objective/train/tokens_used": 2020946400,
      "theoretical_loss": 3.4303209878635426,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019898696088264795,
      "loss": 2.6807,
      "theoretical_loss": 3.4303209878635426,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019897693079237713,
      "loss": 2.5977,
      "theoretical_loss": 3.430311618717962,
      "tokens_seen": 2000551936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019896690070210632,
      "loss": 2.7514,
      "theoretical_loss": 3.4303022499652354,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019895687061183552,
      "loss": 2.6319,
      "theoretical_loss": 3.4302928816053324,
      "tokens_seen": 2000683008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001989468405215647,
      "loss": 2.6566,
      "theoretical_loss": 3.4302835136382255,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019893681043129389,
      "loss": 2.5698,
      "theoretical_loss": 3.4302741460638844,
      "tokens_seen": 2000814080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019892678034102307,
      "loss": 2.7461,
      "theoretical_loss": 3.430264778882279,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019891675025075225,
      "loss": 2.5688,
      "theoretical_loss": 3.430255412093381,
      "tokens_seen": 2000945152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019890672016048146,
      "loss": 2.7893,
      "theoretical_loss": 3.430246045697161,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019889669007021064,
      "loss": 2.5474,
      "theoretical_loss": 3.4302366796935893,
      "tokens_seen": 2001076224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019888665997993982,
      "loss": 2.4261,
      "theoretical_loss": 3.4302273140826367,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198876629889669,
      "loss": 2.6797,
      "theoretical_loss": 3.430217948864274,
      "tokens_seen": 2001207296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001988665997993982,
      "loss": 2.6395,
      "theoretical_loss": 3.430208584038472,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001988565697091274,
      "loss": 2.5858,
      "theoretical_loss": 3.430199219605201,
      "tokens_seen": 2001338368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019884653961885657,
      "loss": 2.5791,
      "theoretical_loss": 3.430189855564432,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019883650952858575,
      "loss": 2.612,
      "theoretical_loss": 3.4301804919161354,
      "tokens_seen": 2001469440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019882647943831493,
      "loss": 2.7955,
      "theoretical_loss": 3.430171128660283,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019881644934804414,
      "loss": 2.6103,
      "theoretical_loss": 3.4301617657968437,
      "tokens_seen": 2001600512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019880641925777332,
      "loss": 2.6902,
      "theoretical_loss": 3.4301524033257893,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987963891675025,
      "loss": 2.8248,
      "theoretical_loss": 3.4301430412470904,
      "tokens_seen": 2001731584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019878635907723168,
      "loss": 2.7335,
      "theoretical_loss": 3.4301336795607176,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987763289869609,
      "loss": 2.6604,
      "theoretical_loss": 3.4301243182666417,
      "tokens_seen": 2001862656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019876629889669007,
      "loss": 2.5757,
      "theoretical_loss": 3.4301149573648333,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019875626880641925,
      "loss": 2.8235,
      "theoretical_loss": 3.430105596855263,
      "tokens_seen": 2001993728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019874623871614844,
      "loss": 2.6769,
      "theoretical_loss": 3.430096236737902,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1102668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.969188690185547,
      "objective/train/theoretical_loss": 3.4300868770127213,
      "objective/train/tokens_used": 2022584800,
      "theoretical_loss": 3.4300868770127213,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019873620862587762,
      "loss": 2.7601,
      "theoretical_loss": 3.4300868770127213,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019872617853560683,
      "loss": 2.5029,
      "theoretical_loss": 3.43007751767969,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198716148445336,
      "loss": 2.6986,
      "theoretical_loss": 3.4300681587387807,
      "tokens_seen": 2002255872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987061183550652,
      "loss": 2.4671,
      "theoretical_loss": 3.430058800189963,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019869608826479437,
      "loss": 2.7092,
      "theoretical_loss": 3.430049442033208,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019868605817452358,
      "loss": 2.6391,
      "theoretical_loss": 3.4300400842684864,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019867602808425276,
      "loss": 2.5652,
      "theoretical_loss": 3.4300307268957693,
      "tokens_seen": 2002518016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019866599799398194,
      "loss": 2.606,
      "theoretical_loss": 3.430021369915027,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019865596790371112,
      "loss": 2.3649,
      "theoretical_loss": 3.4300120133262304,
      "tokens_seen": 2002649088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986459378134403,
      "loss": 2.6451,
      "theoretical_loss": 3.43000265712935,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986359077231695,
      "loss": 2.5005,
      "theoretical_loss": 3.429993301324357,
      "tokens_seen": 2002780160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986258776328987,
      "loss": 2.3839,
      "theoretical_loss": 3.4299839459112222,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019861584754262787,
      "loss": 2.3469,
      "theoretical_loss": 3.4299745908899157,
      "tokens_seen": 2002911232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019860581745235705,
      "loss": 2.5795,
      "theoretical_loss": 3.4299652362604087,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019859578736208626,
      "loss": 2.3145,
      "theoretical_loss": 3.4299558820226723,
      "tokens_seen": 2003042304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019858575727181544,
      "loss": 2.604,
      "theoretical_loss": 3.429946528176677,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019857572718154462,
      "loss": 2.7513,
      "theoretical_loss": 3.4299371747223937,
      "tokens_seen": 2003173376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019856569709127383,
      "loss": 2.7356,
      "theoretical_loss": 3.429927821659793,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019855566700100304,
      "loss": 2.4723,
      "theoretical_loss": 3.429918468988845,
      "tokens_seen": 2003304448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019854563691073222,
      "loss": 2.5805,
      "theoretical_loss": 3.4299091167095215,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985356068204614,
      "loss": 2.7427,
      "theoretical_loss": 3.4298997648217933,
      "tokens_seen": 2003435520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019852557673019058,
      "loss": 2.5396,
      "theoretical_loss": 3.42989041332563,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019851554663991976,
      "loss": 2.7923,
      "theoretical_loss": 3.4298810622210043,
      "tokens_seen": 2003566592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019850551654964897,
      "loss": 2.592,
      "theoretical_loss": 3.4298717115078854,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019849548645937815,
      "loss": 2.6418,
      "theoretical_loss": 3.429862361186245,
      "tokens_seen": 2003697664
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1103172,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 3.0667786598205566,
      "objective/train/theoretical_loss": 3.4298530112560535,
      "objective/train/tokens_used": 2024223200,
      "theoretical_loss": 3.4298530112560535,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019848545636910733,
      "loss": 2.8253,
      "theoretical_loss": 3.4298530112560535,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019847542627883652,
      "loss": 2.5881,
      "theoretical_loss": 3.429843661717282,
      "tokens_seen": 2003828736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019846539618856572,
      "loss": 2.5367,
      "theoretical_loss": 3.4298343125699002,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001984553660982949,
      "loss": 2.6517,
      "theoretical_loss": 3.4298249638138807,
      "tokens_seen": 2003959808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019844533600802409,
      "loss": 2.4982,
      "theoretical_loss": 3.4298156154491934,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019843530591775327,
      "loss": 2.4354,
      "theoretical_loss": 3.4298062674758087,
      "tokens_seen": 2004090880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019842527582748245,
      "loss": 2.6345,
      "theoretical_loss": 3.429796919893698,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019841524573721166,
      "loss": 2.7756,
      "theoretical_loss": 3.4297875727028324,
      "tokens_seen": 2004221952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019840521564694084,
      "loss": 2.7421,
      "theoretical_loss": 3.429778225903182,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019839518555667002,
      "loss": 2.5208,
      "theoretical_loss": 3.429768879494718,
      "tokens_seen": 2004353024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983851554663992,
      "loss": 2.4708,
      "theoretical_loss": 3.4297595334774114,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983751253761284,
      "loss": 2.6362,
      "theoretical_loss": 3.4297501878512326,
      "tokens_seen": 2004484096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983650952858576,
      "loss": 2.5659,
      "theoretical_loss": 3.429740842616153,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019835506519558677,
      "loss": 2.6981,
      "theoretical_loss": 3.429731497772143,
      "tokens_seen": 2004615168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019834503510531595,
      "loss": 2.6496,
      "theoretical_loss": 3.4297221533191737,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019833500501504513,
      "loss": 2.4691,
      "theoretical_loss": 3.429712809257216,
      "tokens_seen": 2004746240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019832497492477434,
      "loss": 2.4916,
      "theoretical_loss": 3.4297034655862406,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019831494483450352,
      "loss": 2.4388,
      "theoretical_loss": 3.429694122306218,
      "tokens_seen": 2004877312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983049147442327,
      "loss": 2.4666,
      "theoretical_loss": 3.4296847794171197,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019829488465396188,
      "loss": 2.369,
      "theoretical_loss": 3.4296754369189166,
      "tokens_seen": 2005008384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982848545636911,
      "loss": 2.688,
      "theoretical_loss": 3.4296660948115782,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019827482447342027,
      "loss": 2.6304,
      "theoretical_loss": 3.4296567530950774,
      "tokens_seen": 2005139456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019826479438314945,
      "loss": 2.6191,
      "theoretical_loss": 3.429647411769384,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019825476429287864,
      "loss": 2.5642,
      "theoretical_loss": 3.4296380708344687,
      "tokens_seen": 2005270528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019824473420260782,
      "loss": 2.7494,
      "theoretical_loss": 3.429628730290303,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1104281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.303032159805298,
      "objective/train/theoretical_loss": 3.4296193901368577,
      "objective/train/tokens_used": 2025861600,
      "theoretical_loss": 3.4296193901368577,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019823470411233703,
      "loss": 2.4994,
      "theoretical_loss": 3.4296193901368577,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982246740220662,
      "loss": 2.5643,
      "theoretical_loss": 3.429610050374103,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982146439317954,
      "loss": 2.5647,
      "theoretical_loss": 3.4296007110020104,
      "tokens_seen": 2005532672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019820461384152457,
      "loss": 2.6505,
      "theoretical_loss": 3.4295913720205506,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019819458375125378,
      "loss": 2.6079,
      "theoretical_loss": 3.4295820334296945,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019818455366098296,
      "loss": 2.6911,
      "theoretical_loss": 3.429572695229413,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019817452357071214,
      "loss": 2.712,
      "theoretical_loss": 3.429563357419677,
      "tokens_seen": 2005794816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019816449348044132,
      "loss": 2.5878,
      "theoretical_loss": 3.4295540200004577,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981544633901705,
      "loss": 2.7795,
      "theoretical_loss": 3.4295446829717258,
      "tokens_seen": 2005925888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981444332998997,
      "loss": 2.4653,
      "theoretical_loss": 3.429535346333452,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981344032096289,
      "loss": 2.6854,
      "theoretical_loss": 3.429526010085607,
      "tokens_seen": 2006056960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019812437311935807,
      "loss": 2.5015,
      "theoretical_loss": 3.4295166742281626,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019811434302908725,
      "loss": 2.5845,
      "theoretical_loss": 3.4295073387610895,
      "tokens_seen": 2006188032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019810431293881646,
      "loss": 2.6788,
      "theoretical_loss": 3.4294980036843583,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019809428284854564,
      "loss": 2.4856,
      "theoretical_loss": 3.4294886689979394,
      "tokens_seen": 2006319104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019808425275827482,
      "loss": 2.6205,
      "theoretical_loss": 3.4294793347018047,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198074222668004,
      "loss": 2.6135,
      "theoretical_loss": 3.4294700007959245,
      "tokens_seen": 2006450176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019806419257773319,
      "loss": 2.7924,
      "theoretical_loss": 3.4294606672802703,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001980541624874624,
      "loss": 2.6292,
      "theoretical_loss": 3.4294513341548125,
      "tokens_seen": 2006581248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019804413239719157,
      "loss": 2.6206,
      "theoretical_loss": 3.4294420014195226,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019803410230692076,
      "loss": 2.7882,
      "theoretical_loss": 3.4294326690743713,
      "tokens_seen": 2006712320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019802407221664994,
      "loss": 2.6205,
      "theoretical_loss": 3.429423337119329,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019801404212637915,
      "loss": 2.5879,
      "theoretical_loss": 3.429414005554367,
      "tokens_seen": 2006843392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019800401203610833,
      "loss": 2.4881,
      "theoretical_loss": 3.429404674379457,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979939819458375,
      "loss": 2.6203,
      "theoretical_loss": 3.429395343594569,
      "tokens_seen": 2006974464
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1104977,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5677521228790283,
      "objective/train/theoretical_loss": 3.4293860131996743,
      "objective/train/tokens_used": 2027500000,
      "theoretical_loss": 3.4293860131996743,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979839518555667,
      "loss": 2.5348,
      "theoretical_loss": 3.4293860131996743,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979739217652959,
      "loss": 2.3496,
      "theoretical_loss": 3.429376683194744,
      "tokens_seen": 2007105536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019796389167502508,
      "loss": 2.5631,
      "theoretical_loss": 3.4293673535797486,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019795386158475426,
      "loss": 2.8612,
      "theoretical_loss": 3.4293580243546598,
      "tokens_seen": 2007236608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019794383149448344,
      "loss": 2.769,
      "theoretical_loss": 3.4293486955194483,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019793380140421262,
      "loss": 2.6009,
      "theoretical_loss": 3.4293393670740846,
      "tokens_seen": 2007367680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019792377131394183,
      "loss": 2.5624,
      "theoretical_loss": 3.42933003901854,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197913741223671,
      "loss": 2.721,
      "theoretical_loss": 3.429320711352786,
      "tokens_seen": 2007498752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001979037111334002,
      "loss": 2.6829,
      "theoretical_loss": 3.429311384076793,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019789368104312937,
      "loss": 2.5737,
      "theoretical_loss": 3.4293020571905317,
      "tokens_seen": 2007629824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019788365095285858,
      "loss": 2.5885,
      "theoretical_loss": 3.429292730693974,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019787362086258776,
      "loss": 2.5488,
      "theoretical_loss": 3.42928340458709,
      "tokens_seen": 2007760896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019786359077231694,
      "loss": 2.6592,
      "theoretical_loss": 3.4292740788698515,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019785356068204612,
      "loss": 2.6761,
      "theoretical_loss": 3.4292647535422285,
      "tokens_seen": 2007891968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978435305917753,
      "loss": 2.8611,
      "theoretical_loss": 3.4292554286041934,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019783350050150451,
      "loss": 2.539,
      "theoretical_loss": 3.429246104055716,
      "tokens_seen": 2008023040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978234704112337,
      "loss": 2.4985,
      "theoretical_loss": 3.4292367798967676,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978134403209629,
      "loss": 2.6664,
      "theoretical_loss": 3.4292274561273195,
      "tokens_seen": 2008154112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019780341023069208,
      "loss": 2.6354,
      "theoretical_loss": 3.4292181327473426,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977933801404213,
      "loss": 2.643,
      "theoretical_loss": 3.429208809756808,
      "tokens_seen": 2008285184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019778335005015047,
      "loss": 2.7474,
      "theoretical_loss": 3.429199487155686,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019777331995987965,
      "loss": 2.8321,
      "theoretical_loss": 3.4291901649439493,
      "tokens_seen": 2008416256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019776328986960884,
      "loss": 2.5811,
      "theoretical_loss": 3.429180843121567,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019775325977933802,
      "loss": 2.587,
      "theoretical_loss": 3.4291715216885112,
      "tokens_seen": 2008547328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019774322968906723,
      "loss": 2.7327,
      "theoretical_loss": 3.429162200644753,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1105995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8363006114959717,
      "objective/train/theoretical_loss": 3.429152879990263,
      "objective/train/tokens_used": 2029138400,
      "theoretical_loss": 3.429152879990263,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977331995987964,
      "loss": 2.7213,
      "theoretical_loss": 3.429152879990263,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977231695085256,
      "loss": 2.6055,
      "theoretical_loss": 3.429143559725013,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019771313941825477,
      "loss": 2.4768,
      "theoretical_loss": 3.429134239848973,
      "tokens_seen": 2008809472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019770310932798398,
      "loss": 2.4651,
      "theoretical_loss": 3.4291249203621144,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019769307923771316,
      "loss": 2.6996,
      "theoretical_loss": 3.4291156012644084,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019768304914744234,
      "loss": 2.6138,
      "theoretical_loss": 3.429106282555826,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019767301905717152,
      "loss": 2.5054,
      "theoretical_loss": 3.4290969642363387,
      "tokens_seen": 2009071616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976629889669007,
      "loss": 2.7562,
      "theoretical_loss": 3.429087646305917,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976529588766299,
      "loss": 2.5785,
      "theoretical_loss": 3.4290783287645317,
      "tokens_seen": 2009202688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976429287863591,
      "loss": 2.5327,
      "theoretical_loss": 3.429069011612155,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019763289869608827,
      "loss": 2.3833,
      "theoretical_loss": 3.429059694848757,
      "tokens_seen": 2009333760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019762286860581745,
      "loss": 2.5658,
      "theoretical_loss": 3.4290503784743085,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019761283851554666,
      "loss": 2.6364,
      "theoretical_loss": 3.429041062488782,
      "tokens_seen": 2009464832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019760280842527584,
      "loss": 2.4989,
      "theoretical_loss": 3.429031746892147,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019759277833500502,
      "loss": 2.5929,
      "theoretical_loss": 3.429022431684375,
      "tokens_seen": 2009595904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001975827482447342,
      "loss": 2.7043,
      "theoretical_loss": 3.4290131168654385,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019757271815446339,
      "loss": 2.5098,
      "theoretical_loss": 3.4290038024353064,
      "tokens_seen": 2009726976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001975626880641926,
      "loss": 2.6543,
      "theoretical_loss": 3.4289944883939514,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019755265797392178,
      "loss": 2.6128,
      "theoretical_loss": 3.428985174741344,
      "tokens_seen": 2009858048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019754262788365096,
      "loss": 2.6826,
      "theoretical_loss": 3.428975861477455,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019753259779338014,
      "loss": 2.4216,
      "theoretical_loss": 3.4289665486022565,
      "tokens_seen": 2009989120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019752256770310935,
      "loss": 2.5222,
      "theoretical_loss": 3.4289572361157186,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019751253761283853,
      "loss": 2.5833,
      "theoretical_loss": 3.428947924017813,
      "tokens_seen": 2010120192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001975025075225677,
      "loss": 2.5798,
      "theoretical_loss": 3.42893861230851,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974924774322969,
      "loss": 2.6958,
      "theoretical_loss": 3.428929300987782,
      "tokens_seen": 2010251264
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1107334,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.3123385906219482,
      "objective/train/theoretical_loss": 3.428919990055599,
      "objective/train/tokens_used": 2030776800,
      "theoretical_loss": 3.428919990055599,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974824473420261,
      "loss": 2.4117,
      "theoretical_loss": 3.428919990055599,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019747241725175528,
      "loss": 2.4052,
      "theoretical_loss": 3.428910679511932,
      "tokens_seen": 2010382336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019746238716148446,
      "loss": 2.713,
      "theoretical_loss": 3.428901369356754,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019745235707121364,
      "loss": 2.6387,
      "theoretical_loss": 3.4288920595900336,
      "tokens_seen": 2010513408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019744232698094282,
      "loss": 2.429,
      "theoretical_loss": 3.4288827502117436,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019743229689067203,
      "loss": 2.6196,
      "theoretical_loss": 3.4288734412218544,
      "tokens_seen": 2010644480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974222668004012,
      "loss": 2.617,
      "theoretical_loss": 3.4288641326203377,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974122367101304,
      "loss": 2.5864,
      "theoretical_loss": 3.428854824407164,
      "tokens_seen": 2010775552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019740220661985957,
      "loss": 2.4082,
      "theoretical_loss": 3.428845516582305,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019739217652958878,
      "loss": 2.7139,
      "theoretical_loss": 3.4288362091457314,
      "tokens_seen": 2010906624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019738214643931796,
      "loss": 2.4859,
      "theoretical_loss": 3.428826902097415,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019737211634904714,
      "loss": 2.6979,
      "theoretical_loss": 3.428817595437326,
      "tokens_seen": 2011037696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019736208625877632,
      "loss": 2.6452,
      "theoretical_loss": 3.4288082891654366,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973520561685055,
      "loss": 2.6798,
      "theoretical_loss": 3.428798983281717,
      "tokens_seen": 2011168768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019734202607823471,
      "loss": 2.4223,
      "theoretical_loss": 3.4287896777861384,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973319959879639,
      "loss": 2.5275,
      "theoretical_loss": 3.4287803726786734,
      "tokens_seen": 2011299840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019732196589769308,
      "loss": 2.69,
      "theoretical_loss": 3.4287710679592913,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019731193580742226,
      "loss": 2.3991,
      "theoretical_loss": 3.428761763627964,
      "tokens_seen": 2011430912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019730190571715147,
      "loss": 2.4716,
      "theoretical_loss": 3.4287524596846635,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019729187562688065,
      "loss": 2.7218,
      "theoretical_loss": 3.42874315612936,
      "tokens_seen": 2011561984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019728184553660983,
      "loss": 2.4327,
      "theoretical_loss": 3.4287338529620244,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197271815446339,
      "loss": 2.5212,
      "theoretical_loss": 3.428724550182629,
      "tokens_seen": 2011693056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972617853560682,
      "loss": 2.4857,
      "theoretical_loss": 3.4287152477911436,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972517552657974,
      "loss": 2.4221,
      "theoretical_loss": 3.4287059457875406,
      "tokens_seen": 2011824128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019724172517552658,
      "loss": 2.6512,
      "theoretical_loss": 3.428696644171791,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1107921,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.8925347328186035,
      "objective/train/theoretical_loss": 3.428687342943866,
      "objective/train/tokens_used": 2032415200,
      "theoretical_loss": 3.428687342943866,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019723169508525576,
      "loss": 2.5676,
      "theoretical_loss": 3.428687342943866,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019722166499498494,
      "loss": 2.5206,
      "theoretical_loss": 3.428678042103736,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019721163490471415,
      "loss": 2.4546,
      "theoretical_loss": 3.428668741651373,
      "tokens_seen": 2012086272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019720160481444333,
      "loss": 2.5279,
      "theoretical_loss": 3.4286594415867477,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971915747241725,
      "loss": 2.3926,
      "theoretical_loss": 3.4286501419098316,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971815446339017,
      "loss": 2.5142,
      "theoretical_loss": 3.428640842620596,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019717151454363087,
      "loss": 2.4849,
      "theoretical_loss": 3.4286315437190122,
      "tokens_seen": 2012348416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019716148445336008,
      "loss": 2.5248,
      "theoretical_loss": 3.4286222452050508,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019715145436308926,
      "loss": 2.4075,
      "theoretical_loss": 3.428612947078684,
      "tokens_seen": 2012479488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019714142427281844,
      "loss": 2.6365,
      "theoretical_loss": 3.428603649339882,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019713139418254763,
      "loss": 2.5042,
      "theoretical_loss": 3.4285943519886164,
      "tokens_seen": 2012610560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019712136409227683,
      "loss": 2.7437,
      "theoretical_loss": 3.4285850550248584,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019711133400200602,
      "loss": 2.6937,
      "theoretical_loss": 3.42857575844858,
      "tokens_seen": 2012741632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971013039117352,
      "loss": 2.4053,
      "theoretical_loss": 3.4285664622597514,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019709127382146438,
      "loss": 2.3314,
      "theoretical_loss": 3.428557166458344,
      "tokens_seen": 2012872704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019708124373119356,
      "loss": 2.4055,
      "theoretical_loss": 3.4285478710443296,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970712136409228,
      "loss": 2.4736,
      "theoretical_loss": 3.428538576017679,
      "tokens_seen": 2013003776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019706118355065198,
      "loss": 2.6516,
      "theoretical_loss": 3.4285292813783634,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019705115346038116,
      "loss": 2.5161,
      "theoretical_loss": 3.4285199871263545,
      "tokens_seen": 2013134848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019704112337011034,
      "loss": 2.4738,
      "theoretical_loss": 3.4285106932616225,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019703109327983955,
      "loss": 2.6949,
      "theoretical_loss": 3.4285013997841403,
      "tokens_seen": 2013265920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019702106318956873,
      "loss": 2.78,
      "theoretical_loss": 3.428492106693878,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970110330992979,
      "loss": 2.6619,
      "theoretical_loss": 3.428482813990807,
      "tokens_seen": 2013396992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970010030090271,
      "loss": 2.6549,
      "theoretical_loss": 3.4284735216748983,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001969909729187563,
      "loss": 2.6112,
      "theoretical_loss": 3.4284642297461243,
      "tokens_seen": 2013528064
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1109221,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7111470699310303,
      "objective/train/theoretical_loss": 3.428454938204455,
      "objective/train/tokens_used": 2034053600,
      "theoretical_loss": 3.428454938204455,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019698094282848548,
      "loss": 2.5129,
      "theoretical_loss": 3.428454938204455,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019697091273821466,
      "loss": 2.4437,
      "theoretical_loss": 3.4284456470498625,
      "tokens_seen": 2013659136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019696088264794384,
      "loss": 2.5913,
      "theoretical_loss": 3.4284363562823175,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019695085255767302,
      "loss": 2.4567,
      "theoretical_loss": 3.4284270659017917,
      "tokens_seen": 2013790208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019694082246740223,
      "loss": 2.5343,
      "theoretical_loss": 3.4284177759082564,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001969307923771314,
      "loss": 2.6057,
      "theoretical_loss": 3.4284084863016826,
      "tokens_seen": 2013921280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001969207622868606,
      "loss": 2.4175,
      "theoretical_loss": 3.428399197082042,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019691073219658977,
      "loss": 2.4655,
      "theoretical_loss": 3.4283899082493052,
      "tokens_seen": 2014052352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019690070210631898,
      "loss": 2.3493,
      "theoretical_loss": 3.428380619803444,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019689067201604816,
      "loss": 2.6867,
      "theoretical_loss": 3.4283713317444295,
      "tokens_seen": 2014183424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019688064192577734,
      "loss": 2.5775,
      "theoretical_loss": 3.4283620440722333,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019687061183550652,
      "loss": 2.5527,
      "theoretical_loss": 3.4283527567868264,
      "tokens_seen": 2014314496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001968605817452357,
      "loss": 2.5232,
      "theoretical_loss": 3.4283434698881807,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019685055165496491,
      "loss": 2.5703,
      "theoretical_loss": 3.4283341833762666,
      "tokens_seen": 2014445568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001968405215646941,
      "loss": 2.5425,
      "theoretical_loss": 3.4283248972510556,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019683049147442328,
      "loss": 2.6937,
      "theoretical_loss": 3.42831561151252,
      "tokens_seen": 2014576640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019682046138415246,
      "loss": 2.6852,
      "theoretical_loss": 3.42830632616063,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019681043129388167,
      "loss": 2.6108,
      "theoretical_loss": 3.428297041195357,
      "tokens_seen": 2014707712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019680040120361085,
      "loss": 2.5361,
      "theoretical_loss": 3.4282877566166734,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019679037111334003,
      "loss": 2.7598,
      "theoretical_loss": 3.428278472424549,
      "tokens_seen": 2014838784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001967803410230692,
      "loss": 2.5369,
      "theoretical_loss": 3.4282691886189562,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001967703109327984,
      "loss": 2.8461,
      "theoretical_loss": 3.4282599051998663,
      "tokens_seen": 2014969856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001967602808425276,
      "loss": 2.4007,
      "theoretical_loss": 3.4282506221672504,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019675025075225678,
      "loss": 2.2288,
      "theoretical_loss": 3.4282413395210796,
      "tokens_seen": 2015100928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019674022066198596,
      "loss": 2.5265,
      "theoretical_loss": 3.4282320572613254,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1109727,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.594752073287964,
      "objective/train/theoretical_loss": 3.428222775387959,
      "objective/train/tokens_used": 2035692000,
      "theoretical_loss": 3.428222775387959,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019673019057171514,
      "loss": 2.6172,
      "theoretical_loss": 3.428222775387959,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019672016048144435,
      "loss": 2.8836,
      "theoretical_loss": 3.4282134939009525,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019671013039117353,
      "loss": 2.5722,
      "theoretical_loss": 3.4282042128002765,
      "tokens_seen": 2015363072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001967001003009027,
      "loss": 2.4788,
      "theoretical_loss": 3.4281949320859026,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966900702106319,
      "loss": 2.5346,
      "theoretical_loss": 3.4281856517578024,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019668004012036107,
      "loss": 2.3295,
      "theoretical_loss": 3.4281763718159466,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019667001003009028,
      "loss": 2.6508,
      "theoretical_loss": 3.428167092260307,
      "tokens_seen": 2015625216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019665997993981946,
      "loss": 2.4898,
      "theoretical_loss": 3.4281578130908548,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019664994984954865,
      "loss": 2.452,
      "theoretical_loss": 3.428148534307562,
      "tokens_seen": 2015756288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019663991975927783,
      "loss": 2.6279,
      "theoretical_loss": 3.4281392559103994,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019662988966900703,
      "loss": 2.606,
      "theoretical_loss": 3.428129977899338,
      "tokens_seen": 2015887360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019661985957873622,
      "loss": 2.5161,
      "theoretical_loss": 3.4281207002743503,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966098294884654,
      "loss": 2.5059,
      "theoretical_loss": 3.4281114230354066,
      "tokens_seen": 2016018432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019659979939819458,
      "loss": 2.6513,
      "theoretical_loss": 3.428102146182479,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019658976930792376,
      "loss": 2.3186,
      "theoretical_loss": 3.428092869715538,
      "tokens_seen": 2016149504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019657973921765297,
      "loss": 2.488,
      "theoretical_loss": 3.4280835936345566,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019656970912738215,
      "loss": 2.5518,
      "theoretical_loss": 3.4280743179395046,
      "tokens_seen": 2016280576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019655967903711133,
      "loss": 2.4907,
      "theoretical_loss": 3.4280650426303536,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965496489468405,
      "loss": 2.7798,
      "theoretical_loss": 3.428055767707076,
      "tokens_seen": 2016411648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019653961885656972,
      "loss": 2.4509,
      "theoretical_loss": 3.428046493169643,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965295887662989,
      "loss": 2.6387,
      "theoretical_loss": 3.428037219018025,
      "tokens_seen": 2016542720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019651955867602808,
      "loss": 2.523,
      "theoretical_loss": 3.428027945252194,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019650952858575726,
      "loss": 2.3161,
      "theoretical_loss": 3.4280186718721217,
      "tokens_seen": 2016673792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019649949849548644,
      "loss": 2.6161,
      "theoretical_loss": 3.428009398877779,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019648946840521565,
      "loss": 2.7656,
      "theoretical_loss": 3.428000126269138,
      "tokens_seen": 2016804864
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1110721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6471118927001953,
      "objective/train/theoretical_loss": 3.427990854046169,
      "objective/train/tokens_used": 2037330400,
      "theoretical_loss": 3.427990854046169,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019647943831494483,
      "loss": 2.5266,
      "theoretical_loss": 3.427990854046169,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019646940822467401,
      "loss": 2.3048,
      "theoretical_loss": 3.4279815822088455,
      "tokens_seen": 2016935936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964593781344032,
      "loss": 2.4839,
      "theoretical_loss": 3.427972310757136,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964493480441324,
      "loss": 2.6557,
      "theoretical_loss": 3.4279630396910146,
      "tokens_seen": 2017067008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019643931795386158,
      "loss": 2.5356,
      "theoretical_loss": 3.4279537690104513,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019642928786359077,
      "loss": 2.3971,
      "theoretical_loss": 3.427944498715418,
      "tokens_seen": 2017198080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019641925777331995,
      "loss": 2.4222,
      "theoretical_loss": 3.427935228805886,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019640922768304913,
      "loss": 2.5776,
      "theoretical_loss": 3.4279259592818265,
      "tokens_seen": 2017329152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019639919759277834,
      "loss": 2.3483,
      "theoretical_loss": 3.4279166901432117,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019638916750250752,
      "loss": 2.6028,
      "theoretical_loss": 3.427907421390012,
      "tokens_seen": 2017460224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001963791374122367,
      "loss": 2.5011,
      "theoretical_loss": 3.4278981530221997,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019636910732196588,
      "loss": 2.554,
      "theoretical_loss": 3.4278888850397458,
      "tokens_seen": 2017591296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001963590772316951,
      "loss": 2.6265,
      "theoretical_loss": 3.4278796174426223,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019634904714142427,
      "loss": 2.5131,
      "theoretical_loss": 3.4278703502308003,
      "tokens_seen": 2017722368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019633901705115345,
      "loss": 2.5812,
      "theoretical_loss": 3.427861083404251,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019632898696088263,
      "loss": 2.5695,
      "theoretical_loss": 3.4278518169629466,
      "tokens_seen": 2017853440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019631895687061187,
      "loss": 2.4982,
      "theoretical_loss": 3.4278425509068575,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019630892678034105,
      "loss": 2.5819,
      "theoretical_loss": 3.4278332852359563,
      "tokens_seen": 2017984512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019629889669007023,
      "loss": 2.739,
      "theoretical_loss": 3.4278240199502137,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962888665997994,
      "loss": 2.4579,
      "theoretical_loss": 3.427814755049601,
      "tokens_seen": 2018115584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962788365095286,
      "loss": 2.5369,
      "theoretical_loss": 3.4278054905340913,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962688064192578,
      "loss": 2.5394,
      "theoretical_loss": 3.4277962264036543,
      "tokens_seen": 2018246656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019625877632898698,
      "loss": 2.6787,
      "theoretical_loss": 3.427786962658262,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019624874623871616,
      "loss": 2.6912,
      "theoretical_loss": 3.427777699297886,
      "tokens_seen": 2018377728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019623871614844534,
      "loss": 2.4824,
      "theoretical_loss": 3.4277684363224985,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1111403,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.609846591949463,
      "objective/train/theoretical_loss": 3.4277591737320696,
      "objective/train/tokens_used": 2038968800,
      "theoretical_loss": 3.4277591737320696,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019622868605817455,
      "loss": 2.5465,
      "theoretical_loss": 3.4277591737320696,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019621865596790373,
      "loss": 2.6632,
      "theoretical_loss": 3.427749911526572,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962086258776329,
      "loss": 2.666,
      "theoretical_loss": 3.427740649705976,
      "tokens_seen": 2018639872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961985957873621,
      "loss": 2.6497,
      "theoretical_loss": 3.4277313882702547,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019618856569709127,
      "loss": 2.6862,
      "theoretical_loss": 3.4277221272193783,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019617853560682048,
      "loss": 2.5275,
      "theoretical_loss": 3.4277128665533194,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019616850551654966,
      "loss": 2.5799,
      "theoretical_loss": 3.427703606272048,
      "tokens_seen": 2018902016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019615847542627885,
      "loss": 2.5253,
      "theoretical_loss": 3.427694346375537,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019614844533600803,
      "loss": 2.4757,
      "theoretical_loss": 3.4276850868637574,
      "tokens_seen": 2019033088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019613841524573723,
      "loss": 2.5717,
      "theoretical_loss": 3.4276758277366803,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019612838515546642,
      "loss": 2.7281,
      "theoretical_loss": 3.4276665689942782,
      "tokens_seen": 2019164160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961183550651956,
      "loss": 2.5779,
      "theoretical_loss": 3.427657310636522,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019610832497492478,
      "loss": 2.6734,
      "theoretical_loss": 3.4276480526633835,
      "tokens_seen": 2019295232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019609829488465396,
      "loss": 2.692,
      "theoretical_loss": 3.4276387950748335,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019608826479438317,
      "loss": 2.6337,
      "theoretical_loss": 3.4276295378708452,
      "tokens_seen": 2019426304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019607823470411235,
      "loss": 2.6574,
      "theoretical_loss": 3.427620281051388,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019606820461384153,
      "loss": 2.3515,
      "theoretical_loss": 3.4276110246164353,
      "tokens_seen": 2019557376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001960581745235707,
      "loss": 2.7298,
      "theoretical_loss": 3.4276017685659577,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019604814443329992,
      "loss": 2.3992,
      "theoretical_loss": 3.4275925128999267,
      "tokens_seen": 2019688448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001960381143430291,
      "loss": 2.4295,
      "theoretical_loss": 3.4275832576183145,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019602808425275828,
      "loss": 2.542,
      "theoretical_loss": 3.4275740027210917,
      "tokens_seen": 2019819520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019601805416248746,
      "loss": 2.3853,
      "theoretical_loss": 3.4275647482082308,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019600802407221664,
      "loss": 2.4787,
      "theoretical_loss": 3.4275554940797033,
      "tokens_seen": 2019950592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019599799398194585,
      "loss": 2.4927,
      "theoretical_loss": 3.42754624033548,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019598796389167503,
      "loss": 2.5054,
      "theoretical_loss": 3.427536986975533,
      "tokens_seen": 2020081664
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1112399,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5939695835113525,
      "objective/train/theoretical_loss": 3.427527733999834,
      "objective/train/tokens_used": 2040607200,
      "theoretical_loss": 3.427527733999834,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019597793380140421,
      "loss": 2.5832,
      "theoretical_loss": 3.427527733999834,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001959679037111334,
      "loss": 2.7047,
      "theoretical_loss": 3.427518481408354,
      "tokens_seen": 2020212736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001959578736208626,
      "loss": 2.6675,
      "theoretical_loss": 3.427509229201066,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019594784353059178,
      "loss": 2.5713,
      "theoretical_loss": 3.4274999773779395,
      "tokens_seen": 2020343808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019593781344032097,
      "loss": 2.6257,
      "theoretical_loss": 3.4274907259389478,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019592778335005015,
      "loss": 2.3682,
      "theoretical_loss": 3.427481474884061,
      "tokens_seen": 2020474880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019591775325977933,
      "loss": 2.4105,
      "theoretical_loss": 3.4274722242132523,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019590772316950854,
      "loss": 2.4216,
      "theoretical_loss": 3.4274629739264926,
      "tokens_seen": 2020605952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019589769307923772,
      "loss": 2.4957,
      "theoretical_loss": 3.427453724023753,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958876629889669,
      "loss": 2.4841,
      "theoretical_loss": 3.427444474505006,
      "tokens_seen": 2020737024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019587763289869608,
      "loss": 2.3829,
      "theoretical_loss": 3.427435225370222,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958676028084253,
      "loss": 2.6363,
      "theoretical_loss": 3.427425976619374,
      "tokens_seen": 2020868096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019585757271815447,
      "loss": 2.5876,
      "theoretical_loss": 3.427416728252433,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019584754262788365,
      "loss": 2.7509,
      "theoretical_loss": 3.4274074802693706,
      "tokens_seen": 2020999168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019583751253761283,
      "loss": 2.4665,
      "theoretical_loss": 3.427398232670158,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019582748244734204,
      "loss": 2.4997,
      "theoretical_loss": 3.4273889854547672,
      "tokens_seen": 2021130240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019581745235707122,
      "loss": 2.4513,
      "theoretical_loss": 3.4273797386231704,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958074222668004,
      "loss": 2.5081,
      "theoretical_loss": 3.4273704921753385,
      "tokens_seen": 2021261312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019579739217652958,
      "loss": 2.4459,
      "theoretical_loss": 3.427361246111243,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019578736208625876,
      "loss": 2.3134,
      "theoretical_loss": 3.427352000430856,
      "tokens_seen": 2021392384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019577733199598797,
      "loss": 2.2538,
      "theoretical_loss": 3.4273427551341493,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019576730190571715,
      "loss": 2.4287,
      "theoretical_loss": 3.427333510221094,
      "tokens_seen": 2021523456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019575727181544633,
      "loss": 2.1502,
      "theoretical_loss": 3.4273242656916616,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019574724172517552,
      "loss": 2.5855,
      "theoretical_loss": 3.4273150215458243,
      "tokens_seen": 2021654528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019573721163490472,
      "loss": 2.573,
      "theoretical_loss": 3.4273057777835536,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1113663,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.6541056632995605,
      "objective/train/theoretical_loss": 3.4272965344048214,
      "objective/train/tokens_used": 2042245600,
      "theoretical_loss": 3.4272965344048214,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957271815446339,
      "loss": 2.4076,
      "theoretical_loss": 3.4272965344048214,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019571715145436309,
      "loss": 2.6509,
      "theoretical_loss": 3.427287291409599,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019570712136409227,
      "loss": 2.4748,
      "theoretical_loss": 3.4272780487978576,
      "tokens_seen": 2021916672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019569709127382145,
      "loss": 2.3505,
      "theoretical_loss": 3.42726880656957,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019568706118355066,
      "loss": 2.5894,
      "theoretical_loss": 3.4272595647247073,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019567703109327984,
      "loss": 2.5309,
      "theoretical_loss": 3.42725032326324,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019566700100300902,
      "loss": 2.2852,
      "theoretical_loss": 3.4272410821851422,
      "tokens_seen": 2022178816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956569709127382,
      "loss": 2.5787,
      "theoretical_loss": 3.427231841490384,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956469408224674,
      "loss": 2.6625,
      "theoretical_loss": 3.427222601178937,
      "tokens_seen": 2022309888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956369107321966,
      "loss": 2.4305,
      "theoretical_loss": 3.4272133612507734,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019562688064192577,
      "loss": 2.5424,
      "theoretical_loss": 3.427204121705864,
      "tokens_seen": 2022440960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019561685055165495,
      "loss": 2.4979,
      "theoretical_loss": 3.427194882544182,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019560682046138413,
      "loss": 2.3548,
      "theoretical_loss": 3.4271856437656982,
      "tokens_seen": 2022572032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019559679037111334,
      "loss": 2.7106,
      "theoretical_loss": 3.4271764053703837,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019558676028084252,
      "loss": 2.4364,
      "theoretical_loss": 3.4271671673582116,
      "tokens_seen": 2022703104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955767301905717,
      "loss": 2.6361,
      "theoretical_loss": 3.4271579297291526,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955667001003009,
      "loss": 2.5168,
      "theoretical_loss": 3.4271486924831787,
      "tokens_seen": 2022834176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019555667001003012,
      "loss": 2.6896,
      "theoretical_loss": 3.4271394556202615,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955466399197593,
      "loss": 2.4339,
      "theoretical_loss": 3.4271302191403725,
      "tokens_seen": 2022965248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019553660982948848,
      "loss": 2.5328,
      "theoretical_loss": 3.4271209830434843,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019552657973921766,
      "loss": 2.5134,
      "theoretical_loss": 3.427111747329567,
      "tokens_seen": 2023096320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019551654964894684,
      "loss": 2.508,
      "theoretical_loss": 3.4271025119985943,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019550651955867605,
      "loss": 2.451,
      "theoretical_loss": 3.4270932770505365,
      "tokens_seen": 2023227392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019549648946840523,
      "loss": 2.4146,
      "theoretical_loss": 3.4270840424853657,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019548645937813441,
      "loss": 2.7325,
      "theoretical_loss": 3.427074808303054,
      "tokens_seen": 2023358464
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1114303,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.5418269634246826,
      "objective/train/theoretical_loss": 3.427065574503572,
      "objective/train/tokens_used": 2043884000,
      "theoretical_loss": 3.427065574503572,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954764292878636,
      "loss": 2.6134,
      "theoretical_loss": 3.427065574503572,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954663991975928,
      "loss": 2.3325,
      "theoretical_loss": 3.427056341086893,
      "tokens_seen": 2023489536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019545636910732198,
      "loss": 2.2752,
      "theoretical_loss": 3.427047108052988,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019544633901705117,
      "loss": 2.3761,
      "theoretical_loss": 3.427037875401828,
      "tokens_seen": 2023620608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019543630892678035,
      "loss": 2.3682,
      "theoretical_loss": 3.4270286431333856,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019542627883650953,
      "loss": 2.5437,
      "theoretical_loss": 3.4270194112476324,
      "tokens_seen": 2023751680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019541624874623874,
      "loss": 2.714,
      "theoretical_loss": 3.4270101797445403,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019540621865596792,
      "loss": 2.5144,
      "theoretical_loss": 3.4270009486240807,
      "tokens_seen": 2023882752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953961885656971,
      "loss": 2.6693,
      "theoretical_loss": 3.4269917178862253,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019538615847542628,
      "loss": 2.5547,
      "theoretical_loss": 3.426982487530947,
      "tokens_seen": 2024013824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953761283851555,
      "loss": 2.2857,
      "theoretical_loss": 3.426973257558216,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019536609829488467,
      "loss": 2.6552,
      "theoretical_loss": 3.4269640279680043,
      "tokens_seen": 2024144896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019535606820461385,
      "loss": 2.5667,
      "theoretical_loss": 3.4269547987602844,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019534603811434303,
      "loss": 2.6005,
      "theoretical_loss": 3.4269455699350275,
      "tokens_seen": 2024275968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019533600802407224,
      "loss": 2.4742,
      "theoretical_loss": 3.4269363414922056,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019532597793380142,
      "loss": 2.2798,
      "theoretical_loss": 3.426927113431791,
      "tokens_seen": 2024407040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953159478435306,
      "loss": 2.7805,
      "theoretical_loss": 3.4269178857537543,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019530591775325978,
      "loss": 2.5698,
      "theoretical_loss": 3.426908658458068,
      "tokens_seen": 2024538112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019529588766298896,
      "loss": 2.6588,
      "theoretical_loss": 3.426899431544704,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019528585757271817,
      "loss": 2.5665,
      "theoretical_loss": 3.426890205013634,
      "tokens_seen": 2024669184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019527582748244735,
      "loss": 2.4983,
      "theoretical_loss": 3.4268809788648293,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019526579739217653,
      "loss": 2.3965,
      "theoretical_loss": 3.426871753098262,
      "tokens_seen": 2024800256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019525576730190572,
      "loss": 2.6834,
      "theoretical_loss": 3.4268625277139044,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019524573721163492,
      "loss": 2.4772,
      "theoretical_loss": 3.426853302711727,
      "tokens_seen": 2024931328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001952357071213641,
      "loss": 2.3622,
      "theoretical_loss": 3.4268440780917038,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1115580,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.87967848777771,
      "objective/train/theoretical_loss": 3.426834853853804,
      "objective/train/tokens_used": 2045522400,
      "theoretical_loss": 3.426834853853804,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019522567703109329,
      "loss": 2.6964,
      "theoretical_loss": 3.426834853853804,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019521564694082247,
      "loss": 2.3944,
      "theoretical_loss": 3.4268256299980013,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019520561685055165,
      "loss": 2.5885,
      "theoretical_loss": 3.4268164065242663,
      "tokens_seen": 2025193472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019519558676028086,
      "loss": 2.463,
      "theoretical_loss": 3.4268071834325715,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019518555667001004,
      "loss": 2.4696,
      "theoretical_loss": 3.4267979607228893,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019517552657973922,
      "loss": 2.4109,
      "theoretical_loss": 3.42678873839519,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951654964894684,
      "loss": 2.5744,
      "theoretical_loss": 3.4267795164494466,
      "tokens_seen": 2025455616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951554663991976,
      "loss": 2.4823,
      "theoretical_loss": 3.42677029488563,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951454363089268,
      "loss": 2.6751,
      "theoretical_loss": 3.4267610737037133,
      "tokens_seen": 2025586688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019513540621865597,
      "loss": 2.5659,
      "theoretical_loss": 3.426751852903667,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019512537612838515,
      "loss": 2.2632,
      "theoretical_loss": 3.426742632485464,
      "tokens_seen": 2025717760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019511534603811433,
      "loss": 2.7408,
      "theoretical_loss": 3.4267334124490754,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019510531594784354,
      "loss": 2.5399,
      "theoretical_loss": 3.426724192794473,
      "tokens_seen": 2025848832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019509528585757272,
      "loss": 2.5215,
      "theoretical_loss": 3.426714973521629,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950852557673019,
      "loss": 2.5889,
      "theoretical_loss": 3.4267057546305155,
      "tokens_seen": 2025979904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019507522567703108,
      "loss": 2.3907,
      "theoretical_loss": 3.426696536121104,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950651955867603,
      "loss": 2.3409,
      "theoretical_loss": 3.426687317993366,
      "tokens_seen": 2026110976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019505516549648947,
      "loss": 2.5523,
      "theoretical_loss": 3.4266781002472744,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019504513540621865,
      "loss": 2.5526,
      "theoretical_loss": 3.4266688828828,
      "tokens_seen": 2026242048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019503510531594784,
      "loss": 2.634,
      "theoretical_loss": 3.4266596658999147,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019502507522567702,
      "loss": 2.7185,
      "theoretical_loss": 3.426650449298591,
      "tokens_seen": 2026373120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019501504513540622,
      "loss": 2.3544,
      "theoretical_loss": 3.4266412330788008,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950050150451354,
      "loss": 2.451,
      "theoretical_loss": 3.4266320172405154,
      "tokens_seen": 2026504192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001949949849548646,
      "loss": 2.2233,
      "theoretical_loss": 3.426622801783707,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019498495486459377,
      "loss": 2.4833,
      "theoretical_loss": 3.426613586708347,
      "tokens_seen": 2026635264
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1116297,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.148904800415039,
      "objective/train/theoretical_loss": 3.426604372014408,
      "objective/train/tokens_used": 2047160800,
      "theoretical_loss": 3.426604372014408,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019497492477432298,
      "loss": 2.4015,
      "theoretical_loss": 3.426604372014408,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019496489468405216,
      "loss": 2.4955,
      "theoretical_loss": 3.426595157701861,
      "tokens_seen": 2026766336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019495486459378134,
      "loss": 2.4932,
      "theoretical_loss": 3.426585943770679,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019494483450351052,
      "loss": 2.544,
      "theoretical_loss": 3.426576730220833,
      "tokens_seen": 2026897408
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001949348044132397,
      "loss": 2.3913,
      "theoretical_loss": 3.4265675170522956,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001949247743229689,
      "loss": 2.54,
      "theoretical_loss": 3.426558304265038,
      "tokens_seen": 2027028480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001949147442326981,
      "loss": 2.6172,
      "theoretical_loss": 3.426549091859032,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019490471414242727,
      "loss": 2.4007,
      "theoretical_loss": 3.426539879834251,
      "tokens_seen": 2027159552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019489468405215645,
      "loss": 2.4588,
      "theoretical_loss": 3.4265306681906647,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019488465396188566,
      "loss": 2.5263,
      "theoretical_loss": 3.4265214569282465,
      "tokens_seen": 2027290624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019487462387161484,
      "loss": 2.5669,
      "theoretical_loss": 3.4265122460469675,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019486459378134402,
      "loss": 2.5223,
      "theoretical_loss": 3.4265030355468005,
      "tokens_seen": 2027421696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001948545636910732,
      "loss": 2.5324,
      "theoretical_loss": 3.4264938254277166,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019484453360080239,
      "loss": 2.7043,
      "theoretical_loss": 3.426484615689688,
      "tokens_seen": 2027552768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001948345035105316,
      "loss": 2.5643,
      "theoretical_loss": 3.4264754063326865,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019482447342026077,
      "loss": 2.519,
      "theoretical_loss": 3.426466197356685,
      "tokens_seen": 2027683840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019481444332998998,
      "loss": 2.4932,
      "theoretical_loss": 3.426456988761654,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019480441323971916,
      "loss": 2.6981,
      "theoretical_loss": 3.4264477805475657,
      "tokens_seen": 2027814912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019479438314944837,
      "loss": 2.6065,
      "theoretical_loss": 3.426438572714393,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019478435305917755,
      "loss": 2.5499,
      "theoretical_loss": 3.4264293652621065,
      "tokens_seen": 2027945984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019477432296890673,
      "loss": 2.7336,
      "theoretical_loss": 3.4264201581906795,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019476429287863592,
      "loss": 2.5706,
      "theoretical_loss": 3.426410951500083,
      "tokens_seen": 2028077056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019475426278836512,
      "loss": 2.6224,
      "theoretical_loss": 3.4264017451902893,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947442326980943,
      "loss": 2.5307,
      "theoretical_loss": 3.4263925392612697,
      "tokens_seen": 2028208128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019473420260782349,
      "loss": 2.5467,
      "theoretical_loss": 3.426383333712997,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.61,
      "objective/train/docs_used": 1117380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.373683452606201,
      "objective/train/theoretical_loss": 3.4263741285454428,
      "objective/train/tokens_used": 2048799200,
      "theoretical_loss": 3.4263741285454428,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019472417251755267,
      "loss": 2.4702,
      "theoretical_loss": 3.4263741285454428,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019471414242728185,
      "loss": 2.4963,
      "theoretical_loss": 3.4263649237585794,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019470411233701106,
      "loss": 2.5271,
      "theoretical_loss": 3.426355719352378,
      "tokens_seen": 2028470272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019469408224674024,
      "loss": 2.647,
      "theoretical_loss": 3.4263465153268116,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019468405215646942,
      "loss": 2.4455,
      "theoretical_loss": 3.426337311681851,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001946740220661986,
      "loss": 2.3802,
      "theoretical_loss": 3.4263281084174695,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001946639919759278,
      "loss": 2.3561,
      "theoretical_loss": 3.4263189055336376,
      "tokens_seen": 2028732416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000194653961885657,
      "loss": 2.4821,
      "theoretical_loss": 3.426309703030328,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019464393179538617,
      "loss": 2.4998,
      "theoretical_loss": 3.426300500907513,
      "tokens_seen": 2028863488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019463390170511535,
      "loss": 2.4807,
      "theoretical_loss": 3.4262912991651637,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019462387161484453,
      "loss": 2.6129,
      "theoretical_loss": 3.426282097803253,
      "tokens_seen": 2028994560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019461384152457374,
      "loss": 2.407,
      "theoretical_loss": 3.426272896821753,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019460381143430292,
      "loss": 2.5702,
      "theoretical_loss": 3.4262636962206345,
      "tokens_seen": 2029125632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001945937813440321,
      "loss": 2.4284,
      "theoretical_loss": 3.42625449599987,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019458375125376128,
      "loss": 2.6041,
      "theoretical_loss": 3.4262452961594323,
      "tokens_seen": 2029256704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001945737211634905,
      "loss": 2.5972,
      "theoretical_loss": 3.4262360966992924,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019456369107321967,
      "loss": 2.5437,
      "theoretical_loss": 3.426226897619423,
      "tokens_seen": 2029387776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019455366098294885,
      "loss": 2.5895,
      "theoretical_loss": 3.426217698919795,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019454363089267804,
      "loss": 2.6835,
      "theoretical_loss": 3.426208500600382,
      "tokens_seen": 2029518848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019453360080240722,
      "loss": 2.1924,
      "theoretical_loss": 3.426199302661155,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019452357071213642,
      "loss": 2.6225,
      "theoretical_loss": 3.4261901051020858,
      "tokens_seen": 2029649920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001945135406218656,
      "loss": 2.5567,
      "theoretical_loss": 3.426180907923147,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001945035105315948,
      "loss": 2.566,
      "theoretical_loss": 3.4261717111243106,
      "tokens_seen": 2029780992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019449348044132397,
      "loss": 2.3171,
      "theoretical_loss": 3.426162514705548,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019448345035105318,
      "loss": 2.5239,
      "theoretical_loss": 3.426153318666832,
      "tokens_seen": 2029912064
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1118053,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.0985701084136963,
      "objective/train/theoretical_loss": 3.426144123008134,
      "objective/train/tokens_used": 2050437600,
      "theoretical_loss": 3.426144123008134,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019447342026078236,
      "loss": 2.2497,
      "theoretical_loss": 3.426144123008134,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019446339017051154,
      "loss": 2.4297,
      "theoretical_loss": 3.426134927729427,
      "tokens_seen": 2030043136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019445336008024072,
      "loss": 2.5705,
      "theoretical_loss": 3.426125732830682,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944433299899699,
      "loss": 2.3437,
      "theoretical_loss": 3.426116538311871,
      "tokens_seen": 2030174208
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944332998996991,
      "loss": 2.5784,
      "theoretical_loss": 3.4261073441729666,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944232698094283,
      "loss": 2.4762,
      "theoretical_loss": 3.426098150413941,
      "tokens_seen": 2030305280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019441323971915747,
      "loss": 2.5347,
      "theoretical_loss": 3.4260889570347652,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019440320962888665,
      "loss": 2.5842,
      "theoretical_loss": 3.426079764035413,
      "tokens_seen": 2030436352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019439317953861586,
      "loss": 2.6653,
      "theoretical_loss": 3.4260705714158544,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019438314944834504,
      "loss": 2.6429,
      "theoretical_loss": 3.426061379176063,
      "tokens_seen": 2030567424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019437311935807422,
      "loss": 2.5834,
      "theoretical_loss": 3.42605218731601,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943630892678034,
      "loss": 2.6261,
      "theoretical_loss": 3.426042995835668,
      "tokens_seen": 2030698496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019435305917753259,
      "loss": 2.6343,
      "theoretical_loss": 3.4260338047350087,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943430290872618,
      "loss": 2.3454,
      "theoretical_loss": 3.426024614014004,
      "tokens_seen": 2030829568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019433299899699097,
      "loss": 2.5377,
      "theoretical_loss": 3.426015423672627,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019432296890672016,
      "loss": 2.4818,
      "theoretical_loss": 3.426006233710848,
      "tokens_seen": 2030960640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019431293881644934,
      "loss": 2.5236,
      "theoretical_loss": 3.425997044128641,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019430290872617855,
      "loss": 2.6375,
      "theoretical_loss": 3.425987854925977,
      "tokens_seen": 2031091712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019429287863590773,
      "loss": 2.7242,
      "theoretical_loss": 3.4259786661028278,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942828485456369,
      "loss": 2.5854,
      "theoretical_loss": 3.4259694776591663,
      "tokens_seen": 2031222784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942728184553661,
      "loss": 2.6129,
      "theoretical_loss": 3.425960289594964,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942627883650953,
      "loss": 2.6483,
      "theoretical_loss": 3.4259511019101936,
      "tokens_seen": 2031353856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019425275827482448,
      "loss": 2.6114,
      "theoretical_loss": 3.4259419146048264,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019424272818455366,
      "loss": 2.6194,
      "theoretical_loss": 3.425932727678835,
      "tokens_seen": 2031484928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019423269809428284,
      "loss": 2.5611,
      "theoretical_loss": 3.425923541132191,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.62,
      "objective/train/docs_used": 1119152,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/original_loss": 2.7294652462005615,
      "objective/train/theoretical_loss": 3.4259143549648674,
      "objective/train/tokens_used": 2052076000,
      "theoretical_loss": 3.4259143549648674,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019422266800401202,
      "loss": 2.5141,
      "theoretical_loss": 3.4259143549648674,
      "tokens_seen": 2031616000
    }
  ],
  "max_steps": 50354,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 1.036806782976e+18,
  "trial_name": null,
  "trial_params": null
}